S
Saccadic Reaction Time
A saccade is a fast movement of an eye (or head or other
part of an animal body or device) toward the onset of
a visual (or auditory, or tactile) stimulus. Saccadic
reaction time is defined as the time between the onset
of the target and the onset of the saccade as defined by
setting some velocity criteria.

Satisficing
The notion that a decision maker stops their search and
selects a solution that satisfies some criteria, rather than
continuing to search for an optimal solution.

Scaffold
▶ Schema-Based Problem Solving

Sagacity
▶ Experiencing Wisdom Across the Lifespan

Scaffolding
Sample Matching
Sample matching is a newly developed methodology
for selecting “representative” samples from nonrandomly selected pools of respondents (Rubin 2006).

References
Rubin, D. (2006). Matched sampling for causal effects. Cambridge:
Cambridge University Press.

Sapience
▶ Experiencing Wisdom Across the Lifespan

Satisfaction
▶ Aristotle on Pleasure and Learning

JANET MANNHEIMER ZYDNEY
Curriculum & Instruction: Instructional Design
and Technology, University of Cincinnati,
Cincinnati, OH, USA

Synonyms
Adaptive support; Instructional support

Definition
Scaffolding has been defined as a “process that enables
a child or novice to solve a problem, carry out a task or
achieve a goal which would be beyond his unassisted
efforts” (Wood et al. 1976, p. 90). Scaffolding provides
a temporary structure or support to assist a learner in
a task and can be gradually reduced and eventually
removed altogether once the learner can carry out the
performance on his or her own (Pea 2004). In order to
determine the adjustable level of support that meets the
learner’s needs at a particular time, the scaffolding
process involves an ongoing diagnosis of a learner’s
proficiency in the task (Pea 2004).

N. Seel (ed.), Encyclopedia of the Sciences of Learning, DOI 10.1007/978-1-4419-1428-6,
# Springer Science+Business Media, LLC 2012

2914

S

Scaffolding

Theoretical Background
The concept of scaffolding grew out of Vygotsky’s
learning construct, ▶ the zone of proximal development (ZPD). This construct established the notion
that coaches or peers who are more capable could
help learners move beyond their actual developmental
level to one that meets their learning potential
(Vygotsky 1935/1978). The actual term scaffolding
was later defined by Wood et al. (1976). In this article,
Wood et al. (1976) explained that “scaffolding consists
essentially of the adult ‘controlling’ those elements of
the task that are initially beyond the learner’s capacity,
thus permitting him to concentrate upon and complete
only those elements that are within his range of competence” (p. 90). This assistance by the adult would
enable the learner to develop competence at a much
faster rate than if unassisted.
Although the original concept of scaffolding
involved mentoring by a more experienced person
(e.g., Palincsar 1986; Wood et al. 1976), more recently,
educators have become interested in scaffolding provided through computer-based tools. Computer-based
scaffolding focuses on learners working with software
tools that distribute the cognitive task between the
learner and the computer system (Salomon 1993). As
the learner uses the tool, they begin to “internalize” the
guidance it provides; this guidance eventually leaves
a “cognitive residue” on the learner, allowing them to
complete an increasing amount of the work themselves
without the need of the supports offered by the computer (Salomon 1993, pp. 130–131). Once the guidance
has been internalized, the learner is better able to selfregulate his or her own cognitions. This leads to
improved ability to get more out of the tool and eventually improved competence even when the tool is no
longer used.
Although scaffolding is described in many ways
throughout the literature, there are several common
features associated with scaffolded instruction:
1. Recruiting and maintaining learner’s attention
toward a goal: Central to scaffolded instruction is
recruiting and enlisting interest (Wood et al. 1976).
This can be accomplished by developing a common
goal or shared understanding of what is to be
accomplished (Puntambekar and Kolodner 2005).
Once the common goal is established, it is important to keep the learner focused by sustaining

2.

3.

4.

5.

interest and maintaining direction toward the
objective (Palincsar 1986; Pea 2004; Wood et al.
1976). Another way to focus a learner’s attention
is by highlighting critical features of the activity
(Pea 2004).
Simplifying the task: Another important aspect of
scaffolding is to simplify the task by reducing the
degrees of freedom or the size of the activity to
make it more manageable for an inexperienced
person (Palincsar 1986; Pea 2004; Wood et al.
1976). Sometimes, simplifying the task can involve
off-loading some of the task onto someone or
something else (Pea 2004; Salomon 1993).
Modeling and demonstrating: Much of the literature
on scaffolding also highlights the need to model the
task or demonstrate the solution to the learner
(Palincsar 1986; Pea 2004; Wood et al. 1976). This
involves demonstrating the target performance
along with providing explanation of the steps
(Wood et al. 1976).
Ongoing diagnosis and assessment: In order to determine the appropriate level of support that the
learner needs, it is critical to continually diagnosis
and assess the learner’s current level of proficiency
(Pea 2004; Puntambekar and Kolodner 2005). One
method to accomplish this is to note the discrepancy between the learner’s performance and the
expected outcomes related to the critical features
of the task (Wood et al. 1976).
Fading support and eventual transfer: The final feature of scaffolding is to reduce the support over
time and eventually remove it so that the learner
is performing the task independently (Pea 2004;
Puntambekar and Kolodner 2005). This can be
achieved by the learner internalizing the process of
using the scaffold in such a way that eventually he
or she no longer requires its use (Salomon 1993).

Important Scientific Research and
Open Questions
As technology-based scaffolding becomes more prevalent, it raises many important issues and unanswered
questions. One of the major areas of debate in the
literature is around the concept of ▶ fading or adapting
the support over time. Many researchers describe all
types of computer-based tools as scaffolds regardless of
whether or not they include fading. However, Pea
(2004) has argued that a scaffold without fading is

Scaffolding

really more of a ▶ distributed intelligence or ▶ distributed cognition, in which the task is shared between the
learner and the tool, and the tool enables the work to be
accomplished more efficiently and effectively. This type
of tool is not designed to be removed, but instead
learners are taught how to perform better through the
use of it. For example, when students are taught to use
a word processor for writing, they are not expected to
stop using this tool, but are expected to learn how to
write better through the use of it.
On a similar note, Salomon (1993) discussed that
there are two cases of distributed cognition: one where
the task is off-loaded onto a tool or person, and the
other which provides qualitative scaffolding in which
the task is guided by a tool or person. He explained,
“The tools, teamwork, or teacher behaviors that are
more akin to the off-loading type are less likely to
cultivate desired cognitions than those tools and
other partnerships that are of the qualitative scaffolding kind. The former, it might be pointed out, are more
likely to lead to deskilling than to the cultivation of skill
mastery” (p. 133).
Although scholars have argued a clear need for
scaffolding with fading in order to promote transfer
of learning, there is surprisingly little empirical
research on the use of technology-based scaffolds
which fade. This is often due to technical limitations
of adding a mechanism to adaptively fade the tool.
Moreover, another issue is that in order to know
when to fade, the tool needs to be able to provide
diagnostic assessment of the learner’s performance,
which for complex tasks may not be technologically
feasible (Pea 2004). Given these limitations, researchers
have explored a few different ways to fade support.
The most optimal alternative is to design the tool
in such a way to provide different levels of support.
For example, some scaffolding provides different levels
of prompts for students based on assessments at specific milestones (Puntambekar and Kolodner 2005).
Another option is to provide multiple scaffolding
tools to provide different kinds of help for the same
task and then remove some of the tools to fade the
support (Puntambekar and Kolodner 2005). Finally,
a third option is to allow learners to choose whether
or not they need a particular scaffold. However, Pea
(2004) raised an important point regarding this option:
“how is it that they [the learners] are expected to
know how to decide among nonadaptive choices for

S

scaffolding? They need scaffolds for the scaffolds”
(p. 433). One solution to this problem may be to have
the teacher provide this type of scaffolding to enable
learners to make these types of choices.
This leads to another area of discussion in the
literature on technology-based scaffolds around the
role of the teacher. In addition to providing sophisticated diagnosis and making decisions regarding what
scaffolding tools should be available, other types of
teacher roles include monitoring participation, offering explanations, clarifying misconceptions, modeling,
helping students to reflect, and coordinating the whole
environment (Puntambekar and Kolodner 2005).
Although teacher scaffolding is clearly needed in connection with technology-based scaffolds, more research
is required to “comparatively examine whether both
software and teacher scaffolding functions could be
achieved by software alone, by teacher alone, and
what this says about whether it is important that the
teacher’s support is human and socially interactive or
simply that it needs to be interactive and adaptive in
nature” (Pea 2004, p. 437). This need can be further
expanded as ▶ Web 2.0 technologies afford the ability
to have social interactions remotely. Thus, future
research might also wish to examine whether in-person
scaffolding versus online scaffolding provided by
a person are equally effective.
The final area of debate in the literature on scaffolding to be discussed in this chapter is how to deal
with the use of scaffolding in classroom environments
to meet the needs of a wide range of students all
working within their own ZPDs. Puntambekar and
Kolodner (2005) have found that one tool alone may
not provide enough assistance to support the wide
range of learner needs. These researchers suggested
the use of ▶ distributed scaffolding or scaffolding that
is distributed over many trials and numerous forms of
support, such as tools, artifacts, resources, and people
within the classroom (Puntambekar and Kolodner
2005). Multiple scaffolding tools may allow each
learner in the classroom to take advantage of scaffolding at a time and, in such a manner, that best supports
his or her individual learning needs. These researchers
recommended for future research that each tool be
analyzed separately to determine how each support is
used by students at different developmental levels. Pea
(2004) concurred that more research is needed on the
use of scaffolding for students at differing levels.

2915

S

2916

S

Scaffolding Discovery Learning Spaces

Cross-References

Definition

▶ Adaptive Game-Based Learning
▶ Cognitive Apprenticeship Learning
▶ Collaboration Scripts
▶ Computer-Based Learning Tools
▶ Distributed Learning
▶ Reciprocal Learning
▶ Scaffolding for Learning
▶ Scaffolding for Learning by the Use of Visual
Representations
▶ Vygotsky’s Philosophy of Learning
▶ Zone of Proximal Development

The scaffolding of discovery learning spaces originates
from the benefits of strategic minimal guidance in order
to enhance self-directed learners’ opt-in metacognition,
decision-making, self-regulation, self-assessments, selfdirected learning, and skills acquisition. With the explosion of publicly available information through the Web
and Internet, information technology (IT) affordances
may be deployed to help discovery learners maximize
their learning opportunities. Scaffolding may be executed in face-to-face, purely online spaces, mixed-reality,
and augmented reality spaces in order to promote discovery of known information as well as novel findings.

References
Palincsar, A. S. (1986). The role of dialogue in providing scaffolded
instruction. Educational Psychologist, 21, 73–98.
Pea, R. D. (2004). The social and technological dimensions of
scaffolding and related theoretical concepts for learning, education, and human activity. The Journal of the Learning Sciences, 13,
423–451.
Puntambekar, S., & Kolodner, J. L. (2005). Toward implementing
distributed scaffolding: Helping students learn science from
design. Journal of Research in Science Teaching, 42, 185–217.
Salomon, G. (1993). No distribution without individuals’ cognition:
A dynamic interactional view. In G. Salomon (Ed.), Distributed
cognitions: Psychological and educational considerations
(pp. 111–138). New York: Cambridge University Press.
Vygotsky, L. S. (1935/1978). Interaction between learning and development. In M. Cole, V. John-Steiner, S. Scribner, & E.
Souberman (Eds.), Mind in society (pp. 79–91). Cambridge,
MA: Harvard University Press.
Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in
problem solving. Journal of Child Psychology, Psychiatry, &
Applied Disciplines, 17, 89–100.

Scaffolding Discovery Learning
Spaces
SHALIN HAI-JEW
Information Technology Assistance Center (iTAC),
Kansas State University, Manhattan, KS, USA

Synonyms
Automatic learning; Independent learning; Learnerdriven learning; Self-directed learning; Self-regulated
learning

Theoretical Background
Learning for individuals and organizations is seen as
salutary, health-producing, and necessary for growth,
innovations, and strategic changes. Discovery learning
uses cases occur in both formal and informal contexts.
In work and higher education, discovery learning leads
to value-added knowledge and skill-building outside
the confines of formal training programs and courses;
discovery learning enhances both the non-formal
and informal methods of knowledge acquisition.
In personal lives, learners may delve into hobbies and
personal areas of interest as amateurs (people who just
want to dabble), novices (people who want to sometime
attain expertise), and lifelong learners. Traditionally,
three models have been linked to self-directed learning:
“linear, interactive, and instructional” (Merriam and
Caffarella 1999, p. 293).
Dedicated discovery learning affects a minority of
active learners who have low learning dependencies or
need for supervision. Such discovery learners tend to be
purposive and focused, self-disciplined, open-minded,
and both strategic and tactical in the pursuit of knowledge and skills. The ability to engage in discovery
learning may come with practice and maturity in
adult learners. Discovery learners take initiative to set
learning goals; scan the subject-domain environment;
identify resources; create a strategy to acquire relevant
information, experiences, and skills; assess their progress and revise strategies; and engage in short-term
and long-term learning activities in a self-disciplined
way. Long-term endeavors require grit and persistence,
particularly for more complex achievements; they also
require progressively greater challenges to maintain
learner interest and creative flow. The new learning

Scaffolding Discovery Learning Spaces

may augment formal learning (for workplaces and
institutions of higher education); it may be part
of lifelong learning, self-enrichment learning, hobbies,
entertainment, and other idiosyncratic interests.
Discovery learning follows a general trajectory,
which is nonlinear and recursive (Fig. 1). The first
cause or inspiration for discovery learning may stem
from any number of contexts. Discovery learners often
frame their inquiry as a real or artificial problemsolving, informational needs, or curiosity-fulfilling
endeavor. The work may begin with casual browsing
and early probes into the topic. They set learning goals;
they determine the standards for information relevance; It may begin with a simple question they set
the schedule and pace of the learning. They may choose
to partner and collaborate with others in this endeavor.
They pursue learning experiences and information in
a variety of environments (online, mixed-reality, and
real spaces) and are rewarded with signs of progress and
development. They may follow a structured process or
an unstructured approach. They make decisions in this
process of what resources and experiments and social
interactions to pursue. Discovery learners vet the foraged data and judge them for veracity; they evaluate

2917

concepts for generalization and transference; they identify the far limits of analogies; they appraise processes
for applicability to different situations. They apply the
new information to their particular situations; they
work on fresh insights. They coordinate their own
discovery learning with a set of tools and resources
that they are comfortable with, and they may develop
and build new learning abilities. Some go farther and
generate new knowledge and innovations. A rare few
may evolve the learning into something of value for
the larger society, knowledge structures, or markets.
They engage their own metacognition to understand
how they are learning, in order to self-regulate. Discovery learning may be applied in any of the various stages
of skill development as conceptualized by H. Dreyfus
(1982): novice, advanced beginner, competence, proficiency, and expertise.
The aim of scaffolding is to build up discovery
learners and their learning efficacy. How learners see
themselves affects their level of initiative and risktaking in learning, so their sense of confidence and
ability should be supported. They should have the
proper information necessary to make the appropriate
decisions for their own learning. They should have

First Cause,
Inspiration
Value-Added
Contributions
to Society
Generating New
Knowledge and
Innovations

S

Framing
the Inquiry,
Problem, or Project
Early Probes,
Casual Browsing

A General Trajectory of Discovery Learning

Application of
New Learning to
Context

Levels of Expertise

Evaluation of
the Experiences,
Information, and
Learning
Pursuing
Learning
Experiences

(Subject Matter) Expertise
Proficiency
Competency
Advanced Beginner
Novice (Dreyfus, 1982)

Partnering,
Collaboration

S
Setting Learning
Goals

Determining
Standards for
Information
Relevance
Scheduling and
Pacing the
Learning

Scaffolding Discovery Learning Spaces. Fig. 1 A general trajectory of discovery learning

2918

S

Scaffolding Discovery Learning Spaces

a clear sense of the progress they have made and the
learning gaps that still exist – as well as the available
learning resources. Higher education has long crossed
the Rubicon of information technology (IT) integration and acceptance, and these added functionalities
must be designed into scaffolding in face-to-face learning, online learning, mixed-reality, and real spaces.
Pure discovery learning exists free-form and
unstructured. Guided discovery learning involves minimal or weak support. Scaffolding refers to the support
tools used to enhance learning for the outliers in
a learning community – the novices and the experts.
IT affordances enable guided discovery learning based
on the realization that discoveries are not usually fully
chance ones but benefit from purposive help, whether
the scaffolding is structural, tool-based, artificial-intelligence-agent based, or human facilitated. For example,
automated explanatory feedback is stronger for scaffolding learning than merely corrective feedback
(Moreno 2004).
While prior discovery learning has focused on
bounded learning of developed fields of study, newer
discovery learning now also involves generative learning, the surfacing of novel knowledge. This new focus
means the promotion of open learning through serendipity, recombinant insights, chance encounters, stochastic (undeterministic) simulation modeling, free
association, and accidental sagacity – based on a history
of field-changing, paradigm-shifting changes based
on unexpected learning. “Chaos” and ambiguity have
been identified as necessary conditions in formulating
insights, innovations, and entrepreneurial activities.
Innovators must fully understand the current intellectual order but also challenge that order with new conceptualizations (Csikszentmihalyi 1996).
A range of factors may affect the quality of the
discovery learning, including human perceptual limitations, biases, selectivity, irrationality in response to
cues (Kruschke 2003), and cognitive dissonance. To
head off the risks of negative learning, discovery
learners may benefit from support at various stages.
The inferences learners make may amplify concepts and
understandings, but they may also lead to dangerous
misconceptions. If a full understanding of a particular
field’s logic of discovery is not understood well, foundational gaps may be missing from learner mental models.
The self-guiding of amateurs and novices provides
plenty of risk for false discovery and wasted energies,

time, and resources. Researchers have observed that even
with contextual help and manuals, learners may end up
in situations of unrecoverable errors.

Scaffolding Discovery Learning
Spaces
The different types of scaffolds for discovery learning
may be conceptualized as three main types: (1) structures/environments, (2) dedicated tools, and (3) virtual
learning communities. Figure 2, General Types of
Scaffolds for Discovery Learning, shows some of the
elements that fit into these categories.

Structures/Environments and Ecologies
A scaffolded structure, environment, or ecology creates
a space for discovery learning. These may involve discovery learners at various stages of their pursuit but
often offer tools to help users frame their questions.
These are built to create a sense of learning context
often aligning with the particular domain. Supportive
scaffolds involve advanced organizers which preface
contents with schemas or outlines. Various learning
elements may be unified into a cohesive space which
structures time and the pacing of learning; sequential
learning experiences, and sensory inputs. These spaces
may bring together those with similar interests for live
interactivity (such as through human-embodied avatars) and intercommunications (audio, video, textual,
and other sensory-rich combinations).
Immersive spaces offer a range of synthetic worlds,
gaming spaces, and artificial life ecosystems for learner
engagement and self-learning. Underlying these spaces
are semi-hidden models of physics, natural worlds, and
artificial lives – which in combination create complexity. These spaces are enhanced by the presence of other
people in human-embodied avatars. Some role plays
and simulations may be derived from hypothetical
events with “meager” information – near-histories,
scenarios (March et al. 1991, p. 4), which expand
insights and the human imagination.
Knowledge-based environments may structure the
information with concept maps, ontologies, taxonomies,
or node structures. These are ways that information are
described, categorized, set in interrelationships, and
represented visually. Information may be represented
in the context of whole-part relationships and according
to different levels of abstraction or granularity. Accurate
semantics and metadata are critical for knowledge

Scaffolding Discovery Learning Spaces

Personalized feedback

Advance organizers/outlines/concept mapping
Knowledge structures/ontologies/taxonomies/node structures
Search tools (with customization and serendipity)

Executive discovery learning functions
Self reflection and documentation tools
Conceptualization tools

Task sequencing/processes/procedures

Pedagogical agentry and intelligent robots

Institutional memory and reputation
Rewards and reinforcements

Smart and customized search tools

Repeatability and review

Priming and debriefing
Simulations
Modeling
Designed cues
Immersive and experiential learning

2919

Contextualized help
Feedback loops
Curriculum and study guides

Framing the issues
Defining the context/situation
Introducing the domain

Decision support tools

S

Decision support tools

Structures/
environments

Dedicated
tools

Intercommunications tools

Virtual learning
communities
Icebreakers
Trust building tools
Dyadic partnering
Group collaboration
Reputation management (both individual and team)
Pay and exchanges
Multisensory intercommunications (asynchronous)
Multisensory intercommunications (synchronous)
Conceptualization tools
Visualization tools
Knowledge archival and selective dissemination/sharing
Intellectual property protections
Private/semi-public/public spaces
Connections to larger societal resources for dissemination/
rewards

Scaffolding Discovery Learning Spaces. Fig. 2 General types of scaffolds for discovery learning

management. A knowledge domain involves learning
about related terminology, theories and principles,
history, biographies of contributors to the field, values
and ethics, professional roles, professional practices,
standards for new learning, applied decision-making
and problem-solving, technologies, and other factors.
Information may be understood as stand-alone or as
part of a larger context or domain. Information may be
raw and unstructured or processed and integrated with
a larger construct.
In knowledge environments, search is a central part
of discovery through various text effective and accurate
audio, visual, and computer-enhanced customization
tools. People engage in different types of browsing:
“opportunistic browsing” is intentional but without
a particular goal in mind; “involuntary browsing”
involves unintentional eye gazing, without even any
latent goals; “search browsing” is intentional and goal
directed (De Brujin and Spence 2008, pp. 5:3–5:4).

Stimulated browsing may raise the breadth of what is
searched and promote serendipitous discoveries.
Scaffolded ecologies support novices by offering
built-in extrinsic rewards (such as greater standing
and resources in educational games) as encouragement
for efforts; these rewards should be nuanced so as not
to minimize intrinsic motivations. They may offer optin practices and interactive learning, along with rich
explanatory feedback. The programmed learning may
build off of swarm intelligence by offering unique
learning sequences based on the learner’s profile
and similar profiles of prior successful learners’ sequential experiences; powerful datamining and machine
intelligence tools enable the capture of human experience to scaffold discovery learning for others.
Performance-based branched learning may also be
delivered in customized ways. Such systems strive to
maintain learner interest. If the learning is too easy,
learners may become bored; if the learning is too

S

2920

S

Scaffolding Discovery Learning Spaces

difficult, learners may feel frustrated and anxious.
Cognitive overload occurs when a learner has reached
absorptive capacity.
In real spaces, latitudes and longitudes may be geotagged and involve the delivery of digitized information
and explorative augmented reality experiences. Digital
installations may involve tangible smart devices and
wearable computing to interact with a computerized
system (White et al. 2007, p. 1); some even use natural
human gestures as communications with a computer
system, which uses cameras and image recognition for
machine comprehension and feedback.
Discovery learning environments may include task
sequencing to support the learning process. These
learning sequence strategies may be developmental,
time-based, problem-solving/project-based, socially
coordinated, or mixed methods (Hai-Jew 2008).
These may maintain institutional memories of people’s
contributions and unique individual and group reputations for past behaviors, achievements, and contributions, as in wikis and virtual learning environments.
There may be tools to surface latent learning through
reflection and thinking prompts.
Learners may be able to repeat and review their
online experiences. They may be able to access decision-support tools – whether these are delivered in the
virtual environments or on hand-held devices in ubiquitous learning situations. Those who take part in
immersive simulations, digital laboratories, and experiential models may also experience pre-immersion
priming for specific prerequisite learning and then a
debriefing afterward to reinforce the important learning.
Lastly, such spaces need to be longitudinal and persistent for value, which means that they must exist in a
perpetual state for learning value. These learning spaces
may become part of learners’ lifelong learning endeavors,
if the spaces may evolve with the learners’ needs.

Dedicated Tools
Dedicated tools may be loosely coupled and standalone, such as Web services delivered through cloud
computing. Tools may also be tightly coupled with
discovery learning systems.
Socio-technical spaces and learning/course management systems often use contextualized help to
enable users to access processes and mental models of
various aspects of their tools. Games and automated
learning sequences involve rich feedback loops.

Automated online learning often uses curriculum and
study guides to enhance the accessibility of the learning. Automated curriculums also offer feedback based
on learner actions, and more sophisticated systems
offer personalized feedback based on learner performance and profiling and stated preferences. Human
help may be accessed through help desks that are usually available 24/7/365.
Such tools may enhance the executive function of
discovery learning – which involves scheduling and
follow-through through calendaring and reminders,
decision-making through workflows, and inspiration
through new contents, live events, hedonic enjoyment,
and digital game “hard fun.” Tools that respond to
learners’ respective strengths and needs enhance their
self-enthusiasm, which supports autonomous learning.
Optimally, scaffolding should address learners “self-systems” in terms of how they approach the learning: the
stable internal structures for that domain, mental models,
sociocultural beliefs about the field, affective schemata,
habitual behavioral patterns, and defensive routines.
Self-reflection, conceptualization, and documentation tools enable learners to doodle, draw, sketch, write,
think-aloud, journal, note-take, and create digital
objects to express their ideas and thinking, in order to
enhance deep learning. Some “modding” enables coding by learners to change their game and simulation
experiences. They may express their own mental
models of a particular knowledge domain and compare
that against the conceptual models of experts.
Discovery learners may locate and store relevant
research as well as “encountered information” which
may have future relevance (Marshall and Jones 2006).
They may meet others with similar interests and collaborate around digital libraries and repositories and
other knowledge structures. They may interact with
pedagogical agents who work as guides or tutors
through a knowledge space. Their work may be
enhanced by machine-identified resources, some of
which are tailored to the searcher’s preferences and
behaviors (such as their degree of openness to following serendipitous links).
They may use various assessments to get a sense of
their level of learning, and they may access other types
of decision-support tools to help them make critical
choices at certain junctures. They may use various
software programs to evaluate qualitative, quantitative,
and mixed methods research – of the primary and raw

Scaffolding Discovery Learning Spaces

kind. Various tools may also provide back-end hidden
information through datamining for a meta-perspective. The activity logs, the build-up of shared knowledge,
and people’s behaviors in online and mixed-reality
spaces may be captured in the “digital enclosures” (as
suggested by M. Andrejevic in iSpy: Surveillance and
Power in the Interactive Era) through various types of
digital tracking and powerful computers with facial
and voice recognition tools, and the use of protected
personal information (PPI).
Rich intercommunications tools and social networking sites enable global interconnectivity with
others who may be helpful informants or collaborators.

Virtual Learning Communities
Virtual learning communities collaborate around
shared topics and endeavors of interest. Members
often share “symmetry of ignorance,” with co-learners
all having a piece of knowledge but not the complete
picture. In others, amateurs, novices, and subject matter experts all commingle, with fresh insights and
shared guidance among them. Learning for novices
often needs to be convergent (to particular “correct”
answers); learning for subject matter experts often
needs to be divergent (to lead to creative innovations).
These human connections enhance boundaryspanning beyond the organizational limits of businesses, institutions of higher education, government
agencies, and nonprofit organizations. These also mitigate cultural blindnesses (limits of observation and
analysis based on social conventions) by offering
a global range of wide scale participants with different
viewpoints. Individual experiences are necessarily limited, so the exchange of relevant information broadens
the participants’ insights. Such communities may
encourage crowd-sourcing or referring to the larger
public for knowledge.
Scaffolding in virtual learning communities
includes a range of tools. Expertise finder systems
locate individuals with particular skill sets or knowledge. Back-end technologies may identify ad hoc virtual groups with shared interests and cyber affinities.
Knowledge may be built more effectively in social
groups, according to constructivist theory. With the
popularity of social networking spaces, people may
use these to express achievement, affiliation, and power.
Rich intercommunication tools enhance human
sociability through digital icebreakers and trust-building

S

2921

tools (such as profiles and institutional memories
of users’ past behaviors and contributions and popularity rankings). Telepresence (individual) and social
presence (group) communicate live embodiment and
actions in the digital space. The creation of both public
and private spaces enables dyadic and group collaborations. Individuals and teams may influence their
reputations through management tools. Payment and
exchange devices may be deployed for freelance business
arrangements. Multisensory intercommunications –
both asynchronous and synchronous – enable individual
and shared design and visualization. Mobile devices may
deploy applications that show the users’ respective
micro-blogs and real-time locations – to connect them
to others with shared interests or from the same social
networks.
The archival of knowledge and group-created
knowledge includes intellectual property protections
and data integrity – for learner protections. The sanctity of private spaces must be protected for easeful
work, and semiprivate spaces enhance select sharing
among group members. Public spaces are needed to
showcase the virtual community’s work and to help
members connect with larger societal levers for work
dissemination, credit, and rewards.

The Usability Design of Scaffolds
Present-day learners maintain high expectations that
even complex socio-technical spaces and tools have
intuitive user interfaces. These interfaces must bridge
the so-called “gulf of execution” which exists between
user understandings of what needs to be done to
achieve a particular aim vs. what actually needs to be
done. Some tools may be encapsulated and invisible to
users (such as nuanced constraints to aid the learning
or opaque underlying modeling for simulations and
immersive experiential learning), and others need to be
visible for opt-in usage. There is a fine balance between
explicit and highly how intrusive recommender services
should be to users and whether services should be
solicited or unsolicited. Usable scaffolds will have to fit
learners at their particular stage of needs. They have
to be unobtrusive but accessible and functional when
required. Scaffolds should “fall away” as they are not
needed. They have to be customizable and adaptable to
meet the unique needs of particular learners. They must
add value by aligning with human perception, cognition, memory, analysis, and creativity.

S

2922

S

Scaffolding for Learning

Important Scientific Research and
Open Questions
Scaffolding discovery learning spaces is a fairly new
phenomenon, originating with the advent of information technology (IT) popularization in higher education in the 1990s and then the building of the semantic
Web and social networking technologies that have
enabled varied types of self-learning. As such, a range
of research questions may be asked about scaffolding
discovery learning.
● What are practical and applied strategies for scaf-

●
●
●

●

●
●
●
●

●

folding discovery learning for different types of
learners?
What are the salient points of learner profiling and
tracking for the most effective learner feedback?
What are some optimal ways to scaffold for novel
discovery in automated discovery learning contexts?
What mixes of technological affordances may
enhance discovery learning in face-to-face spaces?
In mixed reality? In online spaces? In augmented
reality spaces?
What lifelong tracking of learners and their choices
and learning trajectories will surface insights beneficial to other learners?
What unique cultural differences exist between different learner groups from different backgrounds?
How may their respective needs be better met?
What strategies enhance collaborative discovery
learning especially in virtual contexts?
What sorts of disruptive innovations may be forthcoming in terms of information technology (IT),
and new methods of virtual collaborations?
What are the needs of pre-K discovery learners?
K-12 discovery learners? Young adult discovery
learners? Older adult discovery learners? How may
their respective needs be met at each developmental
phase?

Cross-References
▶ Active Learning
▶ Cognitive Load Theory
▶ Cognitive Self-Regulation
▶ Communities of Practice
▶ Design of Learning Environments
▶ Discovery Learning
▶ Guided Discovery Learning
▶ Knowledge Organization

▶ Mental Models
▶ Mixed Reality Learning
▶ Scaffolding
▶ Self Efficacy and Learning
▶ Task Sequencing and Learning

References
Csikszentmihalyi, M. (1996). Creativity: Flow and the psychology of
discovery and invention. New York: HarperCollins.
De Brujin, O., & Spence, R. (2008). A new framework for theory-based
interaction design applied to serendipitous information retrieval.
ACM Transactions on Computer-Human Interaction, 5(1), 1–38.
Hai-Jew, S. (2008). Scaffolding discovery learning spaces. Journal of
Online Learning and Teaching, 4(4), 1–13.
Kruschke, J. K. (2003). Category learning (Chapter for volume edited
by Lamberts and Goldstone), pp. 1–57. http://cognitrn.psych.
indiana.edu/rgoldsto/courses/concepts/kruschkeconcepts.pdf.
Accessed 8 Apr 2011.
March, J. G., Sproull, L. S., & Tamuz, M. (1991). Learning from
samples of one or fewer. Organization Science, 2(1), 1–13.
Marshall, C. C., & Jones, W. (2006). Keeping encountered information. Communications of the ACM, 49(1), 66–67.
Merriam, S. B., & Caffarella, R. S. (1999). Learning in adulthood:
A comprehensive guide (2nd ed.). San Francisco: Jossey-Bass.
Moreno, R. (2004). Decreasing cognitive load for novice students:
Effects of explanatory versus corrective feedback in discoverybased multimedia. Instructional Science, 32, 99–113.
White, S., Lister, L., & Feiner, S. (2007). Visual hints for tangible
gestures in augmented reality. Proceedings of the 2007 6th IEEE
and acm international symposium on mixed and augmented reality. Washington, DC.
Further Readings
Gallagher, W. (2009). Rapt: Attention and the focused life. New York:
The Penguin.

Scaffolding for Learning
K. ANN RENNINGER1, ALEXANDRA LIST2
1
Department of Educational Studies, Swarthmore
College, Swarthmore, PA, USA
2
Department of Human Development, University of
Maryland, College Park, MD, USA

Synonyms
Feedback; Fixed/adaptive scaffolds; Guided participation; Hard/soft scaffolding; Implicit/explicit scaffolds;
Scripts; Supporting student-centered learning tasks/
environments; Zone of proximal development

Scaffolding for Learning

Definition
Scaffolding is a reciprocal feedback process in which
a more expert other (e.g., teacher, or peer with greater
expertise) interacts with a less knowledgeable learner,
with the goal of providing the kind of conceptual
support that enables the learner, over time, to be able
to work with the task, content, or idea independently.
Scaffolding is a sustained interactive process that
involves the fading of assistance/gradual task modifications by an expert. It is informed by nature of the
context, task requirements, and the learners’ interest,
strengths, and needs. Scaffolding should be distinguished from one-time forms of feedback and from
programmed scripts used to prompt learners.

Theoretical Background
Wood et al. (1976) first used the term “scaffolding” to
describe a tutor’s support of a tutee working with
a puzzle: an exchange between a more knowledgeable
other and a less knowledgeable other to support learning. Wood et al. (1976) described scaffolding instructional as interactive and developmental; a support
intended to promote learner autonomy, involving the
gradual reduction of expert guidance over time. Meyer
(1993) extended the use of the scaffolding metaphor to
encompass classroom structure and practice. She identified “three theoretical tenants” of scaffolding that
suggest that the process of scaffolding requires the
teacher (or person doing the scaffolding) to appreciate
learner knowledge and its development as a constructivist process that is influenced by context and evolves
through interaction. Kao and Lehman (1996) further
extended the scaffolding construct to include facilitative tools and skills that enable learners to perform
beyond their present capacities.
Central to each of the above discussions of scaffolding is the reciprocal and evolutionary/iterative nature
of expert-learner interaction(s). Further, these scaffolding interactions change in relation to the learning context and task as well as to the learner’s interests,
strengths, and needs. The term scaffolding, however,
is frequently, and inappropriately, applied to any form
of verbal interaction that is undertaken with learning as
a goal. Scaffolding should be distinguished from feedback that is a one-time, discrete directive or a hardwired script intended to enhance performance. Scaffolding constitutes an ongoing process of providing
feedback with the goal of supporting learning.

S

Discussions about the provision of scaffolding typically reference Vygotsky’s (1978) Zone of Proximal
Development (ZPD). A learner’s ZPD is the theoretical
zone within which he or she can be supported to engage
particular content; where one end point of the ZPD is
the point at which the learner is currently able to
complete the task independently and the other end
point of the ZPD is his or her capacities when provided
with support. The process of providing scaffolding
then can be described as the expert’s adjustment of
his or her interactions within the learner’s ZPD such
that the learner becomes increasingly able to work
independently. Stone (1998) describes the provision
of scaffolding as having four key features: the task and
nature of the supports that enable the learner to
become involved in the task; the learner’s present
capacity to engage the task; the range of supports
offered; and the assumption that the provision of scaffolding will be temporary.
When effective, scaffolding enables independent
activity (Granott 2005); it stretches the learner to
think, develop principled knowledge, and ask questions
(Tabak 2004). Several distinctions for considering types
of scaffolding have been suggested. Granott (2005)
described the process of effective scaffolding as including both vertical variability (where the expert continuously adjusts his or her support in order to promote
independent activity) and horizontal variability (where
the expert shifts strategies if his or her initial efforts to
scaffold are ignored or misunderstood).
Brush and Saye (2002) distinguished between
soft and hard formats of scaffolding. Soft scaffolds
describe supports evolving from purposeful interactions wherein the provider of scaffolding diagnoses students’ needs and provides individualized and timely
support. Hard scaffolds describe scaffolding that is
built into curricular design (e.g., the requirement to
respond in writing to feedback). Soft scaffolds are
dynamic, situation-specific aides provided by a teacher
or peer that can be implemented in both real-time and
online (e.g., experts providing scaffolding feedback to
learners remotely). Hard scaffolds, in contrast, are static
supports that can be anticipated and planned in advance
based upon the expected needs of learners working with
a task. They can be embedded within classroom instructional practices or in multimedia and hypermedia software to provide support to learners at differing points
of entry.

2923

S

2924

S

Scaffolding for Learning

Hannafin et al. (1999) categorized types of scaffolds as including: (1) contextual, providing hints or
prompts; (2) metacognitive, providing specific support
for the particular task; (3) procedural, providing support for resources that could aid task completion; and
(4) strategic, providing different techniques or models.
Hadwin and Winne (2001) further distinguished
between tacit, or implicit, scaffolds and explicit scaffolds in order to suggest how scaffolding can support
the development of self-regulation. Tacit scaffolds refer
to embedded tools that serve to draw students’ attention to their learning behaviors without explicitly
instructing them on task completion through four
phases: task understanding, goal setting, metacognitive
monitoring, and metacognitive evaluation and adaption. In contrast, explicit scaffolds provide direct
instruction to students about how to improve their
learning and strategies for working with the tools provided and afford students opportunities to request
additional support.
Just as interactions in the classroom may involve
one or more types of scaffolding that are sequenced
and made tacit or explicit according to learner needs
and contextual factors, multimedia and hypermedia
environments are designed to include multiple scaffolds that can be made more or less explicit. The
scope and effectiveness of scaffolding differs if
the learner is engaged in a one-on-one tutoring context, in a classroom that rarely allows for individualized interactions, or working with software that
gradually guides him or her through a task (Davis
and Miyake 2004). While it might appear that the
multimedia and hypermedia environments are best
equipped to provide the individualized support and
iterative adjustment that scaffolding requires, many
learners are not ready to work independently with
these tools and require at least initial support to do
so (Azevedo et al. 2005).
Further, multimedia and hypermedia are typically
limited in the interactions and iterations of support
that they can provide learners. In these contexts, the
process of scaffolding involves the gradual modification of a task, or restructuring of a problem, in order to
increase accessibility. The emphasis is on task modification, or task parameters, rather than the learner’s
solution procedure – meaning that the process of scaffolding is not likely to require the learner to modify his

or her strategies in the process of acquiring the skill or
concept, nor does it necessitate the kind of reflection
and consolidation of information that would facilitate
in transfer (Quintana et al. 2004).

Important Scientific Research and
Open Questions
What is the most effective way to provide scaffolding
that will result in reflection and transfer?
For the practitioner and the researcher, the answer
to this question also involves answering questions
such as:
● What are the goals for scaffolding? For the learner?

For the experienced other?
● What are the parameters of the task being

scaffolded?
● How do task parameters support engagement – for
whom and when?
● How do each of the identified distinctions in types
of scaffolding impact learning?
In Bruner’s (1990) description of scaffolding, the
“teacher” or “parent” assumes a role in supporting the
learner not only through task completion, but also
through the process of valuing, engaging, and setting
goals for working with the task. Although much of the
research on scaffolding has focused on supporting selfregulation, provision of scaffolding involves not only
supporting the skills necessary to engage the content of
the task but also the support of student interest and
engagement (Azevedo et al. 2005; Hadwin and Winne
2001).
Discussions of scaffolding have stressed that the
learner needs to be the focus of the interaction. It
appears that learner characteristics that may determine
students’ scaffolding needs and preferences merit further examination. There is evidence that learners’ abilities to work with different types of feedback is
dependent on their phase of interest. Feedback offered
to learners in an earlier phase of interest would be
understood differently by learners in later phases of
interest because of differences in knowledge and feelings about and value for the task (Lipstein and
Renninger 2007; Sansone 1986). Differences in phases
of interest impact students’ conceptualization of the
task, questions, learners’ wants and needs for feedback,
and self-regulatory skills (Renninger 2010).

Scaffolding for Learning

Phase of interest may also determine the scaffolding
learners may need. In earlier phases of interest development, when the goals, knowledge, and value for
the task are less developed, the learner needs scaffolding
to provide what Azevedo et al. (2005) term regulatory
support to engage. Once the learner is engaged, and in
a later phase of interest development, he or she can be
expected to self-regulate: asking questions, seeking
and reflecting on answers. In later phases of interest,
effective scaffolding will change to support the learner
to reflect on and think about the content of the task
and engage in meta-cognitive assessment and adaption (as opposed to simply encouraging engagement)
(Renninger 2010).
The following questions need to be considered as
next steps for research and practice:
● How is the learner in earlier and later phases of

interest positioned to respond to scaffolding?
What are the contributions of the learner’s sociocultural context, age- and experience-related stages
of development, their epistemic beliefs about
a given domain?
● What does the person providing scaffolding
and/or the software designer need to know or
find out about the learner in order to provide
effective scaffolding? How can this information
be acquired?
● How can scaffolding provide support to both selfregulate and engage and also to work with task
content? Does this type of scaffolding vary
depending on whether it is occurring in person or
online?

Cross-References
▶ Adjustment and Learning
▶ Cognitive Apprenticeship-Learning
▶ Feedback and Learning
▶ Feedback in Instructional Contexts
▶ Feedback Strategies
▶ Interests and Learning
▶ Reciprocal Learning
▶ Scaffolding
▶ Scaffolding Discovery Learning Spaces
▶ Scaffolding Learning by Use of
Representations
▶ Self-Regulated and Motivation Strategies

Visual

S

2925

▶ Self-Regulated Learning
▶ Social Interactions and Learning
▶ Student-Centered Learning

References
Azevedo, R., Cromley, J. G., Winters, F. I., Moos, D. C., & Greene, J. A.
(2005). Adaptive human scaffolding facilitates adolescents’ selfregulated learning with hypermedia. Instructional Science, 33,
381–412. Special issue on scaffolding self-regulated learning
and metacognition: Implications for the design of computerbased scaffolds.
Bruner, J. (1990). Acts of meaning. Cambridge: Harvard University
Press.
Brush, T. A., & Saye, J. W. (2002). A summary of research exploring
hard and soft scaffolding for teachers and students using
a multimedia supported learning environment. The Journal of
Interactive Online Learning, 1(2), 1–12.
Davis, E. A., & Miyake, N. (2004). Explorations of scaffolding in
complex classroom systems. The Journal of the Learning Sciences,
13(3), 265–272.
Granott, N. (2005). Scaffolding dynamically toward change: Previous
and new perspectives. New Ideas in Psychology, 23, 140–151.
Hadwin, A., & Winne, P. (2001). CoNoteS2: A software tool for
promoting self-regulation. Educational Research and Evaluation,
7(2/3), 313–334.
Hannafin, M., Land, S., & Oliver, K. (1999). Open learning environments: Foundations, methods, and models. In C. M. Reigeluth
(Ed.), Instructional design theories and models (Vol. 2, pp. 115–140).
Mahwah, NJ: Erlbaum.
Kao, M., Lehman, J., & Cennamo, K. (1996). Scaffolding in hypermedia assisted instruction: An example of integration. Paper
presented at the convention of the Association for Educational
Communications and Technology. Bloomington, IN. (ERIC
Document Reproduction Service No. ED397 803).
Lipstein, R., & Renninger, K. A. (2007). “Putting things into
words”: The development of 12–15-year-old students’ interest
for writing. In P. Boscolo & S. Hidi (Eds.), Motivation and
writing: Research and school practice (pp. 113–140). New York:
Elsevier.
Meyer, D. K. (1993). What is scaffolded instruction? Definitions,
distinguishing features, and misnomers. In D. J. Leu &
C. K. Kinzer (Eds.), Examining central issues in literacy research,
theory, and practice: Forty-second yearbook of the National
Reading Conference (pp. 41–53). Chicago: National Reading
Conference.
Quintana, C., Reiser, B. J., Davis, E. A., Krajcik, J., Fretz, E., Duncan,
R. G., et al. (2004). A scaffolding design framework for software
to support science inquiry. Journal of the Learning Sciences, 13(3),
305–336.
Renninger, K. A. (2010). Working with and cultivating interest, selfefficacy, and self-regulation. In D. Preiss & R. Sternberg (Eds.),
Innovations in educational psychology: Perspectives on learning,
teaching and human development (pp. 158–195). New York:
Springer.

S

2926

S

Scaffolding Hypermedia

Sansone, C. (1986). A questions of competence: The effects of competence and task feedback on intrinsic interest. Journal of Personality and Social Psychology, 51(5), 918–931.
Stone, A. (1998). The metaphor of scaffolding: Its utility for the
field of learning disabilities. Journal of Learning Disabilities,
3(4), 344–364.
Tabak, I. (2004). Synergy: A complement to emerging patterns of
distributed scaffolding. The Journal of the Learning Sciences,
13(3), 337–386.
Vygotsky, L. S. (1978). Mind in society: The development of higher
psychological processes. Cambridge, MA: Harvard University
Press.
Wood, D., Bruner, J., & Ross, G. (1976). The role of tutoring in
problem solving. Journal of Child Psychology and Psychiatry and
Allied Disciplines, 17, 89–100.

Scaffolding Hypermedia
▶ Metacognition and Hypermedia Learning: How Do
They Relate?

Scaffolding Learning by the Use
of Visual Representations
ILONCA HARDY1, SUSANNE KOERBER2
1
Department of Education, Goethe-University
Frankfurt, Frankfurt, Germany
2
Department of Psychology, University of Education
Freiburg, Freiburg, Germany

Synonyms

Theoretical Background
In this entry, we will focus on instructional uses of
visual representations within the fields of mathematics
and science learning that are motivated by socioconstructivist views of learning. There is a broad
literature on the use of visual representations within
multimedia learning, which will not be considered in
detail.
Distinguishing characteristics of visual representations. Within the broad category of external representations we can distinguish between descriptions (written
language/symbols) and depictions, with depictions further subdivided into visual-spatial depictions (manipulatives such as a balance beam) and visual-graphical
depictions (i.e., realistic pictures such as photographs,
analogical pictures such as use of realistic elements in
metaphorical intention, and logical pictures such as
graphs and diagrams; Schnotz 1992), see Fig. 1. These
forms of external representation differ with respect to
their mode of visualization, affecting their effectiveness
(representational power). Both realistic pictures and
analogical pictures display a given concept in a rather
iconic way whereas logical pictures are not characterized by this visual similarity (iconicity) with the
depicted situation; rather, they are based on arbitrary,
context- independent representations and conventions.
This more abstract mode of visualization allows the
representation of elements and structures across a
variety of domains, thus increasing representational
power. Within the last years, visual-graphical representations increasingly gained attention in educational
research because they combine the advantages of depictions and descriptions: visualizing concepts in a
context-independent way – they offer a wide range of
possibilities to display qualitative and quantitative
information; exploiting spatial relations for displaying

Graphical representation; Model; Scientific inscription;
Visual display
external representations

Definition
Visual representations are depictions that use space to
represent nonspatial concepts, offering a wide range of
possibilities to display qualitative and quantitative
information in scientific contexts. Due to their facilitation of information processing, their support of
knowledge construction, and their modeling of expert
reasoning, they are employed as means of scaffolding in
instructional contexts.

depictional

visual-spatial
manipulative

descriptional

visual-graphical
realistic

analogical

logical

Scaffolding Learning by the Use of Visual Representations. Fig. 1 Types of external representations

Scaffolding Learning by the Use of Visual Representations

structural relations – they also support efficient
processing of information.
Instructional uses of visual representations. An
important goal of science and mathematics education
is students’ familiarization with discipline norms of
reasoning and argumentation. Models and representations are one of these widely employed tools by experts.
In the perspective of modeling science, the construction of conceptual knowledge is regarded as closely tied
to modeling practices within meaningful contexts
(Lehrer and Schauble 2003). Modeling here refers to
students’ identification of variables and their interrelations in a given system, their synthesis as well as their
(empirical) validation. Thus, the process of modeling is
inherently tied to the use of external representations
and artifacts, allowing the construction and negotiation of multiple points of view. Using external representations in complex and authentic settings, students
will be familiarized with a scientific cycle where multiple explanatory models are contrasted in a search of
alternative explanations. A further instructional goal of
representational activities concerns students’ ability to
make sense of different forms of conventional and selfdesigned representation. Here, students need to be
attuned to different ways of expressing structural relationships – a competence that has also been called
metarepresentational (diSessa 2004). To this end, visual
representations in instruction should ideally (1) scaffold the process of reasoning, (2) support studentinitiated conceptual understanding, and (3) allow the
transfer between multiple forms of representation.
Mechanisms of scaffolding. For meaningful learning
to take place, students need to actively access relevant
prior knowledge and to relate this knowledge to the
problem at hand to create a mental model. Elements
of scaffolding are meant to support this process of
meaning making. The functioning of scaffolding
has been described by the mechanisms of focusing
and modeling (Pea 2004). While focusing reduces the
complexity of a task, for example, by reducing the
degrees of freedom during problem solution, modeling
situates a problem within a broader disciplinary context. This may be achieved by modeling expert solution
processes and reasoning, thereby producing a discrepancy between current student conceptions and discipline norms.
Visual representations fulfill these two functions of
scaffolds. In the function of focusing, they make salient

S

2927

relevant quantities and relationships between constructs while other (irrelevant) variables are not
represented. For example, in a Cartesian graph we
may represent the density of an object as the proportional relation between mass and volume, while the
characteristics of form or surface structure (which are
commonly associated with an object’s floating and
sinking in naı̈ve conceptions) are not considered. The
visualization of quantities therefore provides an attentional focus to students that reduces the salience of
objects less central to the construct under consideration. In their function of modeling, visual representations may be regarded as ways of familiarizing students
with disciplinary norms of reasoning, thus modeling
more advanced conceptualizations. For example,
a Cartesian graph may support students’ realization of
the multiplicative strategy in proportional reasoning
when students are confronted with a discrepancy
between their (inadequate) additive conceptualization
of ratios and their construction of a slope conveying
the same concepts in a multiplicative relation. This
way, students may come to be dissatisfied with their
existing knowledge, one important precondition for
conceptual change. Beyond experiencing a discrepancy,
students are also provided with a model of more
advanced reasoning with proportional concepts when
using a Cartesian graph.

Important Scientific Research and
Open Questions
There are three main strands of research regarding
students’ uses of visual representations, focusing on
the developmental prerequisites and stages of graph
comprehension, the cognitive effects of visual representations in text comprehension, and the instructional
uses of visual representations especially in the fields of
mathematics and science.
Developmental research. The developmental research
on visual-graphical competencies is concerned with
(1) early competencies with regard to the comprehension of graphical conventions and (2) natural constraints of graph uses due to internal mapping
mechanisms. It has been found that already 4-yearolds are able to acquire rules for the interpretation of
conventional graphs and are able to use information
of bar diagrams to solve simple arithmetic problems.
This ability increases with age from more qualitative
to quantitative uses of bar graphs. Similar evidence

S

2928

S

Scaffolding Learning by the Use of Visual Representations

of early competencies has been found for the interpretation of covariation data. Generally, children from
preschool age to elementary school age develop the
capability to differentiate between relevant and irrelevant attributes of an external representation. For example, a child comes to understand that the color used to
represent a street in a map does not correspond to the
color of the street in reality. Despite positive findings on
early competencies, there is also evidence on persisting
misconceptions on graphs even in secondary school.
In order to be classified as competent representational
users, students should have at least acquired a form
of “correspondence mastery” where they are capable
to (explicitly) refer to the mapping conventions and
conventions of interpretation when using an external
representation. A controversy concerns the degree to
which children access a form of privileged knowledge
when interpreting graphs. In the hypothesis of
a “structural similarity constraint,” it is postulated
that children are endowed with a structurally sensitive mapping mechanism from nonspatial concepts
to spatial concepts, where conceptual elements are
mapped onto spatial elements, conceptual relations
are mapped onto spatial relations, and secondorder conceptual relations onto second-order spatial
relations. This constraint would help explain both
early competencies of graph comprehension and the
occurrence of instructionally resistant misconceptions
of graph interpretation in cases where intuitive mapping mechanisms are violated.
Cognitive research. In the cognitive research on the
use of visual representations, the effects of different
modes of presentation (text/text plus picture, animated/static, etc.) are compared with regard to different learning outcomes. In the so-called multimedia
principle, it has been established that learners improve
in achievement especially when verbal and pictorial
information is presented in a coordinated way. This
can be explained by the dual coding and processing of
information, where the visual system allows a more
efficient processing of spatial information and images
than the verbal system. Also, the effects of cognitive
load on learners, having to deal with information and
procedures more or less central to task solution, have
been established in this strand of research. It has been
found that individual differences play a critical role in
learning from visual representations. For example,
learners with lower spatial ability (especially female

students) may compensate their deficiencies by
working with animated displays. Also, the degree
of prior knowledge impacts the way in which a learner
can profit from a given visual display. Finally, the
effects of visual representations and animations differ
according to their function in the learning process
(representational versus decorational) and the content
to be learned. These differential effects have led to a
variety of methodical interventions with regard to the
prompting of cognitive and metacognitive processing
of information in computer-supported learning environments (Höffler and Leutner 2007).
Educational research. The educational research on
the use of visual representations is concerned with
instructional methods facilitating student understanding of conceptual content with conventional and
student-constructed external representations, usually
based on a socio-constructivist view of learning. For
example, it is investigated whether it is better to provide students with conventional representations or to
have them construct their own forms of representation.
Here, experience-near forms of representation may be
distinguished from experience-distant forms, differing
in the degree to which rules of construction and interpretation need to be established explicitly in instruction. Since students’ naı̈ve conceptions usually only
permit the incomplete representation of a situation,
most student-designed representations at first will be
missing some of the scientifically relevant quantities
and structures. Also, the interpretability of studentsdesigned representations for other students will be
constrained by idiosyncratic ways of representation.
Therefore, the teacher needs to find forms of adequate
support during the process of model construction and
the collective negotiation of meaning such as specific
cognitive and metacognitive prompting. In the long
run, modeling activities can be considered helpful
only if a representation’s idiosyncratic character is
given up in favor of expressing scientifically correct
quantities and relations. This may be achieved in
classroom discussions where norms of representing
are negotiated – a process which moves a representation from a “model of ” (a given situation) to
a “model for” (scientific reasoning), see Gravemeijer
et al. (2002).
Overall, teachers need to know about their students’
cognitive prerequisites with regard to the processing of
visual information, they need to know about the

Schedules of Reinforcement

specific effects of different forms of visual displays, and
they need to settle their specific uses of representation
within an instructional theory of modeling. As is often
the case, the three strands of research have developed
rather independently so that theoretical views are rarely
transferred between them. Especially with regard to the
development of students’ visual-graphical competencies within complex instructional settings, it will
become necessary to tackle issues such as the mechanisms of development and processing with different
theoretical lenses in order to allow instructionally adequate interventions at an individual and group level.

S

2929

Scene Interpretation
Scene interpretation is an automatic visual information
extraction process for general purpose from an image.
It can find figural objects and background in the level of
recognition or categorization.

Scene Perception
▶ Attention and the Processing of Visual Scenes

Cross-References
▶ Animation and Learning
▶ Concept Learning
▶ Constructivist Learning
▶ Learning with External Representations
▶ Mathematical Learning
▶ Mental Models
▶ Model-Based Teaching
▶ Scaffolding
▶ Visual Communication and Learning

References
diSessa, A. (2004). Metarepresentation: Native competence and target
for instruction. Cognition and Instruction, 22(3), 291–331.
Gravemeijer, K., Lehrer, R., van Oers, B., & Verschaffel, L. (2002).
Symbolization, modeling and tool use in mathematics education.
Dorchedt: Kluwer.
Höffler, T., & Leutner, D. (2007). Instructional animation versus static
pictures: A meta-analysis. Learning and Instruction, 17, 722–738.
Lehrer, R., & Schauble, L. (2003). Origins and evolution of modelbased reasoning in mathematics and science. In R. Lesh & H. M.
Doerr (Eds.), Beyond constructivism: A models and modeling
perspective on mathematics problem-solving, learning, and teaching (pp. 59–70). Mahwah: Lawrence Erlbaum.
Pea, R. (2004). The social and technological dimensions of scaffolding and related theoretical concepts for learning, education, and
human activity. Journal of the Learning Sciences, 13(3), 423–451.
Schnotz, W. (2002). Towards an integrated view of learning from
text and visual displays. Educational Psychology Review, 14(2),
101–120.

Scenario-Based Learning
▶ Action-Based Learning

Scene Understanding
▶ Model-Based Scene Interpretation by Multilayered
Context Information

Schedules of Reinforcement
KENNON A. LATTAL
Department of Psychology, West Virginia University,
Morgantown, WV, USA

Synonyms
Contingencies of reinforcement; Operant behavior;
Operant conditioning; Positive reinforcement; Negative reinforcement

Definition
Schedules of reinforcement are procedures for arranging reinforcers in time and in relation to responses.

Theoretical Background
As tools for the study and change of behavior, they are
widely used in both basic and applied research in
psychology and in application. Procedurally, they are
the wellspring of individual-subject research designs
in psychology. Conceptually, they are fundamental in
understanding the role that the environment plays
in controlling behavior. Practically, they are ubiquitous
in the everyday life of all living organisms.

S

2930

S

Schedules of Reinforcement

History: Schedules were discovered by B. F. Skinner
in his earliest experimental analyses of behavior. In one
experiment he “decided to reinforce a response only
once a minute and to allow all other responses to go
unreinforced” (1956, p. 226). The result was a periodic
(or, later, a fixed-interval) schedule. By varying the
response requirement in a study of the relation between
it and deprivation, he subsequently discovered the
fixed-ratio schedule. These two schedules generated
different patterns of responding, and in turn led to
discoveries of numerous other schedules and combinations of schedules, many of which are summarized by
Ferster and Skinner (1957).
The taxonomy of reinforcement schedules developed by Ferster and Skinner (1957) is the standard
description in most behavioral research. A different
taxonomic system, based strictly on temporal variables,
was proposed by Schoenfeld and Cole (1972), and
received some use in earlier times, but it is rarely used
today. Mechner (1959) proposed a flexible system for
diagramming and describing reinforcement schedules
that was the basis for developing a widely used computer programming system for arranging reinforcement schedules.
Reinforcement: Reinforcers are stimuli or other
events that, when made dependent on a response,
develop or maintain that response as a result of either
their presentation (positive reinforcement) or removal
(negative reinforcement). The critical procedural feature of reinforcement is the dependency between the
response and reinforcer. Reinforcers can be scheduled
in the absence of a response-reinforcer dependency, at
either fixed or variable times since the previous reinforcer. Under these conditions, responding can
develop, but often it is idiosyncratic and unstable.
Indeed, delivering response-independent reinforcers
has been shown to be an effective means of reducing
problem behavior.
Arranging schedules of reinforcement: Schedules
embed the basic response-reinforcer dependency that
is critical in the reinforcement process in a broader
temporal and response-requirement context. Reinforcers can be scheduled after a number of responses
have occurred or after a single response has occurred
and a predetermined period of time (an interreinforcer
interval or IRI) has passed. The former describe ratio
schedules and the latter, interval schedules. In addition,
response requirements and time periods may be fixed

or variable, which, combined with the time or response
requirements yields a 2 
 2 matrix defining four basic
schedules: fixed- and variable-interval (FI and VI) and
fixed- and variable-ratio (FR and VR).
Reinforcement also can be scheduled dependent on
sequences of responses. Reinforcing two responses that
occur within a certain time of one another is sometimes
called an interresponse time (IRT) reinforcement
schedule. If two responses are reinforced only if they
are less than t seconds from one another, it is an IRT < t
schedule, and if they are reinforced only if they are at
least t seconds apart, it is an IRT > t schedule.
These schedules also sometimes are labeled, respectively, differential-reinforcement-of- high (or low)rate (DRH or DRL).
The schedules described above have been combined
in other ways, suitable for studying specific research
and applied problems. Two or more single schedules
can be alternated, either in the presence of the same (a
mixed schedule) or different (a multiple schedule) discriminative stimuli; made available at the same time
(a concurrent schedule); arranged sequentially (one
after the other) in the presence of the same (a tandem
schedule) or different (a chained schedule) discriminative stimuli; arranged in an interdependent fashion such
that the requirements of two or more schedules must
be met for the response to be reinforced (a conjunctive
schedule); or interconnected such that with both a time
and a response requirement, responding changes the
time requirement (an interlocking schedule).
Other scheduling arrangements are possible, but
the ones described in this section are the ones used
most often in behavioral research to this time. From
their inception until the advent of digital computers,
schedules were restricted by technological limitations
in their programming, which, in the beginning, was
accomplished by electromechanical devices. With the
advent of the digital computer, more complex scheduling arrangements became possible. For example,
schedules have been arranged to reinforce specific patterns of responses as defined by a computer algorithm,
such as linear, or positive or negatively accelerated
responding across the interreinforcer interval (Hawkes
and Shimp 1975). The analysis of such dynamic interactions between responding and its consequences, as
enhanced by digital computer technology, is a logical
extension of the more static arrangements that characterize earlier research on reinforcement schedules.

Schedules of Reinforcement

Important Scientific Research and
Open Questions
Effects of reinforcement schedules on behavior: Behavior
typically conforms to the requirements of the reinforcement schedule, other things being equal. The
cumulative response graphs in Fig. 1 illustrate the different rates and patterns of responding that are
obtained from FI, FR, VI, VR, and Sidman avoidance
schedules. The pattern of responding across individual
FIs is described as positively accelerated; however, this
pattern can change to a more bi-valued “break-andrun” pattern in the steady state. That is, after reinforcement there is a pause followed by a relatively high,
steady response rate. The latter pattern also is common
of FR schedules, in contrast to VI and VR schedules,
which generate more evenly distributed responding in
time (albeit at very different rates, with the VR controlling considerably higher rates). Schedules also may
induce other forms of behavior, such as excessive
drinking or aggression.
Responding on mixed and tandem schedules is an
amalgam of the responding maintained by the separate
schedules, but multiple and chained schedules, because
of the presence of the discriminative stimuli, maintain
the characteristic performance appropriate to each of
the component schedules. In concurrent schedules, the
allocation of responding to the different choice alternatives depends on the schedules in effect (see entries

2931

Free-operant avoidance

Responses

The schedules described thus far in principle could
be used with both positive and negative reinforcers;
however, the predominance of research involving such
schedules been with the former. The most frequently
used schedules of negative reinforcement involve the
postponement or deletion of scheduled aversive stimuli, such as brief electric shocks. The first such schedule
was named Sidman avoidance in recognition of its
creator (also called free-operant avoidance). In the
presence of an environment without exteroceptive
stimuli ever changing, each response postpones for
a fixed, specified time an otherwise inevitable electric
shock. Continued responding postpones shock indefinitely. Variations on Sidman avoidance schedules program shocks to be delivered at fixed or variable times
from one another, and a single response within
each intershock interval cancels the shock. As in
Sidman avoidance, the exteroceptive stimuli are constant throughout each experimental session.

S

FI

FR

VI

VR

Time

Schedules of Reinforcement. Fig. 1 Cumulative
response graphs showing accumulated responses on the
y-axis as a function of time on the x-axis under different
schedules of reinforcement. The graphs are not on the
same scale, so the number of responses and time depicted
in each is indicated in parentheses in the following
descriptions. Top: Free-operant or Sidman avoidance
(1,000 responses/60 min). Bottom: FI (500 responses/18
min); FR (780 responses/10 min); VI (1,950 responses/20
min); and VR (1,050 responses/8.35 min). The vertical
deflections on the top graph indicate shock deliveries and
on the bottom graphs food presentations

on ▶ Choice, ▶ Matching). When the choice is
between VI schedules, reinforcement rate is a primary
determinant of the time and responses allocated to
either alternative.
Research on schedules involving escape from aversive stimuli is insufficient to allow conclusions to be
drawn about the functional similarities of positive and
negative reinforcement schedules. Schedules arranging
for aversive stimuli to be avoided or postponed control
behavior efficiently, to the point that only a small
proportion of the total number of possible negative
reinforcers (aversive events) that could occur, actually

S

2932

S

Schedules of Reinforcement

do occur. An example of the rather evenly spaced
steady-state responding on a Sidman avoidance procedure also is shown in Fig. 1 (top graph).
Responding on reinforcement schedules is affected
by two types of variables: those imposed by the algorithm by which the schedule is programmed (labeled
direct variables) and those that emerge from the interaction of the programmed schedule and behavior
(labeled indirect variables). Direct variables include
the nature of the response required (e.g., the force of
the response); parameters of reinforcement such as its
rate, magnitude, and immediacy after the response;
and the degree of restriction of access to the reinforcer
(e.g., whether the same reinforcer is available from
other sources and for less cost). Indirect variables are
variables such as the time between reinforcers under FR
or VR schedules, the obtained delay (as compared to
the programmed delay) between reinforcer delivery
and the response that produces it when delays are
unsignaled, and the actual interresponse time that is
followed by reinforcement (unless it is an IRT schedule
as noted above).
Significance of schedules: Reinforcement schedules
are foundational tools in behavioral research. In the
laboratory, they serve as baselines for the study of
a variety of behavioral processes. For example, two
concurrently available chained schedules have been
useful in the analysis of both conditioned reinforcement and foraging and other schedules underlie the
analysis of economic concepts, choice, and decision
making.
Methodologically, research on schedules of reinforcement was largely responsible for the presence of
single-subject research designs in psychology. Skinner’s
early research was conducted largely in the experimental physiology laboratory of W. J. Crozier at Harvard
University. Experimental physiology research typically
involved repeated measurements using only a few subjects. Skinner’s schedules similarly involved the study of
repetitive cycles of responding followed by reinforcement, replicated across sessions with a few subjects.
This replication of schedule effects across sessions
evolved into the use of schedules as baselines for studying other behavioral phenomena. As Skinner’s research
on schedules of reinforcement began to permeate
experimental psychology, many also adopted the single-subject research methods that were so integral to
his research.

Schedules of reinforcement are the basis of several
single-case research designs commonly used in applied
behavior analysis to assess the effects of interventions
on enhancing constructive behavior (and thus reducing problem behavior). The alternating treatment and
simultaneous treatment designs (Kazdin 1982), for
example, derives from the multiple schedule; the
changing-criterion design derives from schedules
involving adjusting contingencies as a function of the
organism’s behavior, and the “simultaneous availability
of all conditions” treatment design (Kazdin) derives
from the concurrent schedule.
Schedules have significance beyond their utilitarian,
methodological value. Morse and Kelleher (1977, p.
197) observed that “schedule control is the single
most important property of operant behavior.” They
are the fundamental determinants of behavior in two
respects. First, other behavioral processes, such as punishment, conditioned reinforcement and punishment,
and stimulus control, all involve the maintenance of
behavior, which defines schedules of reinforcement.
Second, the form and function of behavior derives
from the organism’s history, which is to say from the
ways in which reinforcement has been arranged –
scheduled – in the past as well as in the present. Whenever behavior occurs, a schedule of reinforcement was
and is present. Whether the subject matter is a single
well-defined response or multielement behavioral repertoires of adult humans, schedules of reinforcement
are responsible for shaping and maintaining the behavior. It is for these reasons that schedules of reinforcement are said to be ubiquitous.

Cross-References
▶ Feedback and Learning
▶ Feedback Strategies
▶ Law of Effect
▶ Operant Behavior
▶ Operant Learning
▶ Reinforcement Learning
▶ Skinner, B.F.

References
Ferster, C. B., & Skinner, B. F. (1957). Schedules of reinforcement.
New York: Appleton.
Hawkes, L., & Shimp, C. P. (1975). Reinforcement of behavioral
patterns: Shaping a scallop. Journal of the Experimental Analysis
of Behavior, 23, 3–16.

Schema(s)
Kazdin, A. (1982). Single-case research designs. Oxford: Oxford University Press.
Mechner, F. (1959). A notation system for the description of behavioral procedures. Journal of the Experimental Analysis of Behavior,
2, 133–150.
Morse, W. H., & Kelleher, R. T. (1977). Determinants of reinforcement and punishment. In W. K. Honig & J. E. R. Staddon (Eds.),
Handbook of operant behavior (pp. 98–124). New York: Prentice
Hall.
Schoenfeld, W. N., & Cole, B. K. (1972). Stimulus schedules: The t-tau
system. New York: Harper & Row.
Skinner, B. F. (1956). A case history in scientific method. The American Psychologist, 11, 221–233.

Schema(s)
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany

Synonyms
Frame; Plan; Scheme; Script

Definition
The word schema comes from the Greek word “swήma”
(skhēma), which means shape, or more generally, plan.
The plural is “swήmata” (skhēmata). The term
“schema” (plural: schemata [UK], or sometimes
schemas [USA]) is used in the sciences of learning and
cognition to designate a psychological construct that
accounts for the molar forms of human knowledge.
A schema represents the generic and abstract knowledge
a person has acquired in the course of numerous individual experiences with objects, people, situations, and
events. Schemas organize knowledge about specific
stimulus domains and guide both the processing of
new information and the retrieval of stored information. They can be viewed as structured expectations
about people, situations, and events.

Theoretical Background
Some authors have argued (e.g., Neisser 1976) that all
human knowledge – everything from knowledge about
the form of the letter A to abstract knowledge about
astrophysics or political ideologies – is organized by
schemas. They facilitate the ▶ assimilation of new

S

information into knowledge structures. As soon as a
schema can be activated, it runs automatically and regulates information processing in a “top-down” manner.
This allows information to be processed very quickly,
a function which is vital for humans as it enables them
to adapt to their environment more quickly.
However schema activation and schema construction are two different problems. While it is possible to
activate existing schemata with a given topic, it does
not necessarily follow that a learner can use this activated knowledge to develop new knowledge and skills.
This can be achieved through the construction of mental models (i.e., a central means of ▶ accommodation).
The concept of schemas has been a central theoretical concept of cognitive and developmental psychology for decades. From a historical perspective, the
concept of schemas can be traced back to Plato, who
elaborated the Greek doctrine of ideal types – such as
the perfect circle that exists in the mind but which no
one has ever seen. The German philosopher Immanuel
Kant (1724–1804) further developed this notion and
introduced the word schema to characterize innate
structures which organize our world. Kant comprehends
a “schema” (of a concept) in general as a “procedure of
imagination which provides a concept with an image”
(Kant, Critique of Pure Reason, 1781/1929 B, p. 179 f.).
For example, he describes the “dog” schema as a mental
pattern which “can delineate the figure of a four-footed
animal in a general manner without limitation to any
single determinate figure as experience or any possible
image that I can represent in concreto.” Such perceptual
schemas defined by means of particular codes of practice are necessary whenever concepts are applied to
infinite processes and therefore cannot be determined
through the designation of a finite sequence of exemplifying illustrations. This is called “schematism” and is
described by Kant as “an art concealed in the depths of
the human soul, whose real modes of activity nature is
hardly likely ever to allow us to discover” (Kant 1781/
1929, pp. A141–B181).
Another German philosopher who used the term
“schema” is Johann Gottlieb Fichte (1762–1814). Occasionally, he defines schema as the shape or form of an
object. This corresponds to Kant’s schematism of
empirical concepts. Sometimes Fichte also uses the
term “schema” in the sense of a “compendium of
a knowledge domain, proofing, order” when he speaks
about a “schema of investigation.”

2933

S

2934

S

Schema(s)

Since Kant and Fichte, “schema” has become
a technical term of philosophy and the philosophy
of science that appears mostly in composite words,
such as “▶ axiom schema,” “▶ definition schema,” or
“action schema.” These different uses of “schema” are
united by the idea that the option of an action or
a procedure is contrasted with the corresponding realizations. Particular actions occur as realizations of an
action schema. Wright defines action schemas in terms
of generic acts and their realizations as individual acts.
Another distinction is made between type (= action
schema) and token (= realization). One and the same
action event can refresh/activate different action
schemas.
Under the influence of Kant, German psychologists,
such as Otto Selz (1913) and Karl Bühler (1918),
applied the schema concept to the fields of productive
thinking and cognitive development. In the UK, Bartlett (1932) used the term in his experiments on social
psychology. “The War of the Ghosts” used in Bartlett’s
experiments can be found in many textbooks on cognitive psychology.
For the second half of the twentieth century, Jean
Piaget (1897–1980) can probably be considered the
most influential epistemologist and developmental
psychologist. In his seminal work, the concept
“schema” plays an important role. One of the central
concerns of Piaget’s work was the development of
schemas and their functions for ▶ equilibration. He
suggested that children learn by using existing
schemas that are accommodated or assimilated.
Piaget also used the term “affective” or “emotional”
schemas – not in contrast to cognitive schemas but
rather as totalities of primarily affective responses and
feelings that can be transferred to analogous situations
(Eckblad 1981).
The key theoretical development of schema theory was made in several fields, including linguistics,
anthropology, psychology, and artificial intelligence.
The heyday of schema theory was probably in the
1970s and 1980s as an outcome of the paradigm
shift in psychology (the so-called cognitive revolution) in the 1960s that resulted in cognitive constructivism. Its central assertion is that the content which is
available in memory at a given point in time is the
product of a constructive process in which consciousness itself is first built up as awareness. Cognitive
constructivism also posits that thinking is a process

of operating with symbols which makes people capable of representing their subjective experiences, ideas,
thoughts, and feelings. Research interest in this field
thus concentrates on the question of how knowledge
is represented. Since cognitive psychology deals with
the relationship between learning and retaining, it is
guided by the assumption that memory includes all
of the knowledge a person has and represents past
experiences in cognitive structures. These structures
are referred to as “schemas” and are understood as
the building blocks of a constructive concept of
consciousness.
Consequently, schema theorists (e.g., Mandler
1984) argue that schemas represent the generic and
abstract knowledge a person has acquired through
numerous individual experiences with objects, people,
situations, and actions. The major assumptions of
schema theory can be summarized as follows
(Rumelhart et al. 1986):
1. Human knowledge about the world is represented
in memory as the sum of structured units called
schemas.
2. Most schemas are acquired through learning; however, some primitive schemas are assumed to be
innate.
3. There are schemas of objects (e.g., a CHAIR
schema), of persons (e.g., a TEACHER schema),
of a state of affairs (e.g., a PEACE schema), of
abstract concepts (e.g., a COMMUNISM schema),
of relationships between objects (e.g., a RELATION: CHAIR-TABLE schema), etc. Schemas of
actions and events are often labeled as scripts,
a metaphorical use of the corresponding cinematic
term. Furthermore, a distinction has been made
between situational schemas (= “frames”), action
schemas (= “plans”), and “grammars,” which represent the regular structure of narratives and stories
(Mandler 1984).
4. Schemas are not direct, image-like copies of phenomena in the external world but the result of
cognitive processing of information. The mind neither copies the world, passively accepting it as
a readymade given, nor does it ignore the world.
5. Schemas have different levels of complexity and
abstraction. One can have an abstract schema
CHAIR as well as a concrete schema of one specific
chair, for example, in a particular dining room.

Schema(s)

6. Schemas have variables which can take on various
values. The elements that make up a schema can be
called slots. Any important element or schema
within a schema may be thought of as a slot that
can accept any of the range of values that are compatible with its associated schemata. For example,
a very general schema for BUY may have four fairly
abstract slots: a seller, a buyer, an object, and
money. On different occasions these slots in the
BUY schema will be filled differently. In one case
the buyer will be a woman, in another a man, the
object bought can be either a Hamburger or a book,
etc., that is, the concrete slot fillers vary in different
environments and situations but fit into the same
schema BUY.
7. A schema can integrate new data into its organizational structure. It can put new information into
the existing human database.
8. Schemas can embed with each other (e.g., CHAIR
as a subschema of FURNITURE). This characteristic has to do with the abstract intensification of
attributes and the hierarchical organization of
schemas of various generalization and abstraction
resulting from this process.
Schemas serve several cognitive functions, which
can be characterized as integrating information, regulating attention, and making inferences. However, since
Piaget the major cognitive function of schemata is seen
in assimilation. Schemata provide learners with the
cognitive framework for “matching” information
from stimuli with content from knowledge memory,
thus allowing them to select the information that is
consistent with the schema. Anderson (1984, p. 5)
captures the essence of these functions of schemas
when he remarks: “Without a schema to which an
event can be assimilated, learning is slow and
uncertain.”
Types of cognitive schemas are as follows:
● Social schemas are about general social knowledge.
● Person schemas are about the attributes (skills, com-

petencies, values) of a particular individual. Person
schemas often correspond with the personality
which is attributed to a person. Idealized person
schemas are called prototypes. This word, however,
is also used for any generalized schema.
● Self-schemas are about the self. They are generalizations about the self abstracted from the present

S

2935

situation and past experiences. Self-efficacy, for
instance, is a type of self-schema that applies to
a particular task. People also hold idealized or
projected selves, or possible selves.
● Role schemas are about proper behaviors in given
situations. Role schemas contain sets of role expectations on how an individual will occupy a certain
role to behavior. They are used for various purposes, such as evaluation of others, role playing,
identification of roles, and prediction of behaviors.
● Event schemas (or scripts) are about what happens in
specific situations. Event schemas correspond with
processes, practices, or ways in which people typically approach tasks and problems. They are the
programs of action.

Important Scientific Research and
Open Questions
The concept of a schema is sometimes so general that it
raises and leaves many important questions. As a consequence, there are psychologists (e.g., Brown 1979)
who reject the concept of schemas as unsuitable for
theoretical purposes. Indeed, a major criticism of schema
theories is that they are basically assimilation models
(Brown 1979) which fail to answer questions on how
existing conceptions are modified in the face of inconsistent input and how such theories deal with novelty.
In cognitive psychology, schemas are defined as
large-scale slot-filler structures that play critical roles
in the interpretation of input data, the guiding of
action, and the storage of knowledge in memory. Connectionist and PDP models, however, do not work with
concepts such as schemas but rather conceptualize
memory – consistent with recent brain research – by
means of “▶ constraint networks” resulting from the
sum of excitatory and inhibitory influences on a unit.
Input comes into the cognitive system and activates
a set of units which tend to act together as a response
to certain patterns of the input. At a first glance, it does
not seem possible to reconcile the concept of schemas
with connectionist models of auto-associative memory,
which posit that knowledge memory consists of
a closely connected network of neuronal nodes, each
of which includes a certain amount of activity. Knowledge is spread over the nodes and the connections between
them and stored knowledge does not correspond directly
to a schema. Obviously, connectionistic models refer to
a completely different level of information processing

S

2936

S

Schema Development

than the psychological conception of schemas:
Connectionistic models study the microstructure of
cognition whereas psychological constructs such as
schemas refer to the comprehensive data structures
which play a critical role in interpreting new information, regulating actions, and constructing knowledge
and organizing it in memory. According to Rumelhart
et al. (1986), however, this does not necessarily rule out
the possibility that the concept of schemas corresponds
to connectionist model approaches. These authors
argue that schemata may be understood as characteristics of complex networks which are composed of
many small neuronal cycles. These tightly
interconnected units may correspond to what has
been called a schema in cognitive psychology. “What
is stored is a set of connection strengths which, when
activated, have implicitly in them the ability to generate
states that correspond to instantiated schemata”
(Rumelhart et al. 1986, p. 21).

Cross-References
▶ Emotional Schema(s)
▶ Mental Models
▶ Mental Representations
▶ Schema Development
▶ Schema Therapy
▶ Schema-based Learning
▶ Schema-dependent Neocortical Connectivity During
Information Processing
▶ Schematic Influences on Category Learning and
Recognition Memory

References
Anderson, R. C. (1984). Some reflections on the acquisition of
knowledge. Educational Researcher, 13(9), 5–10.
Bartlett, F. C. (1932). Remembering: A study in experimental and social
psychology, Cambridge: University Press.
Brown, A. L. (1979). Theories of memory and the problems of
development: Activity, growth, and knowledge. In L. S. Cermak
& F. I. M. Craik (Eds.), Levels of processing in human memory
(pp. 225–258). Hillsdale: Erlbaum.
Bühler, K. (1918). Die geistige Entwicklung des Kindes. Jena: Verlag
Gustav Fischer.
Eckblad, G. (1981). Scheme theory. A conceptual framework for cognitive-motivational processes. London: Academic.
Kant, I. (1781/1929) Critique of pure reason. Trans. N. K. Smith.
New York: St. Martin’s Press.
Mandler, J. M. (1984). Stories, scripts, and scenes: Aspects of schema
theory. Hillsdale: Erlbaum.

Neisser, U. (1976). Cognition and reality. San Francisco: Freeman.
Rumelhart, D. E., Smolensky, P., McClelland, J. L., & Hinton, G. E.
(1986). Schemata and sequential thought processes in PDP
models. In J. L. McClelland, D. E. Rumelhart, & The PDP
research group (Eds.), Parallel distributed processing. Explorations
in the microstructure of cognition (Psychological and biological
models, Vol. 2, pp. 7–57). Cambridge, MA: MIT Press.
Selz, O. (1913). Über die Gesetze des geordneten Denkverlaufes. Eine
experimentelle Untersuchung. Stuttgart: Speemann.

Schema Development
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany

Synonyms
Acquisition of schemas; Construction of schemas; Formation of schemas

Definition
Schemas are acquired and constructed through experiences with specific instances. Physiologically speaking,
they start as simple networks and develop into more
complex structures. From the perspective of psychology, the development of schemas starts with the construction of simple behavioral action schemas, which
are learned through organizational socialization and
concrete experiences, and proceeds to cognitive schemas
by means of the functional incorporation of the regular
structure of actions into the memory. Cognitive
schemas, such as scripts or frames, can be acquired
either directly through a long-term process of learning
and confirmation through repetition or indirectly
through adaptation to stories, myths, films, movies,
conversations, and role models.

Theoretical Background
The biological development of the brain initially causes
the functions for perception and motor activities
and their coordination to be developed in the cortex.
Thus, the ideas which small children develop with
regard to the world are dependent on their sensory
and motor skills. In order to “grasp” the world in a
very literal sense, small children develop sensory-motor
or enactive or action schemas by repeating actions.

Schema Development

Actually, isolating a common deep structure from the
manifold and various actions in the world is the essence
of the “condensation” of more abstract schemas. Therefore, enactive or action schemas are the building blocks
of sensory-motor intelligence and cognitive development. In Piaget’s epistemology, cognitive schemas are
acquired and formed through a process of internalization conceived of as a functional incorporation of
the regular structure of actions into the memory
(Piaget 1954).
Schemas are higher-level cognitive units that are
acquired through slow learning. However, some primitive schemata are assumed to be innate. The gradual
and lasting development of schemas is associated with
processes of ▶ abstraction and ▶ generalization. Consequently, schemas have different levels of complexity
and abstraction. Mature schemas are more extensive,
more organized, more abstract, and contain more characteristics than less mature schemas. An important
aspect of the organization of schemas is that simpler
schemas can be “embedded” within more complex
schemas.
The development of schemas is a long process
which requires many different kinds of learning experiences and is dependent on favorable conditions for
psychological development. However, once a schema
has been built up, it immediately and often unexpectedly begins to exert decisive influence on the acquisition of knowledge as well as on the spontaneous
retrieval of attributes in defined memory traces. Information processing difficulties may occur when schemas

S

2937

cannot be activated or matched with input information. In this case, schemas must be newly learned,
revised, and/or replaced. Rumelhart and Norman
(1978) suggest a classification of learning in terms of
accretion, tuning, and restructuring of schemas.
● Accretion: It refers to the expansion of databases

through the accumulation of new knowledge and
the enlargement of cognitive structure.
● Tuning: A learner may use new information to
“tune” an existing schema so it is more accurate,
complete, or useful.
● Restructuring: A learner may use new information
to reorganize or restructure a schema. If this is not
possible or helpful, the learner has to construct
a new schema or a mental model.
These three modes of cognitive learning can be
integrated into a cognitive architecture grounded on
Piaget’s epistemology of equilibration and the interplay
between assimilation and accommodation (see Fig. 1).
Piaget (1954) argued that infants acquire knowledge of
the world by repeatedly executing action schemas. This
activity was assumed to be innately rewarding. Piaget
introduced assimilation of new experience into extant
schemas and accommodation of schemas to experiences
that do not quite “fit” as the principal learning methods
for infants.
Schemas are defined as coherent slot-filler structures of the human mind that provide an individual
with cognitive structures which allow a prompt interpretation of new information. If a schema does not fit

S
Accommodation

Assimilation

Accretion
Schema
modification
Task
problem

Schema
activation

Tuning
Reorganization

Failure

Success
Success

Mental model
Failure

Solution

Revision

Success

Solution

Schema Development. Fig. 1 Schema modification as the result of accommodation (Seel et al. 2009, p. 18)

2938

S

Schema Development

immediately with the requirements of a new task it can
be adjusted to meet the new requirements by means of
accretion, tuning, or reorganization. However, if accretion or tuning is not successful or if no schema is
available, accommodation must take place in order to
reorganize and structure an individual’s knowledge
concerning the construction of a mental model.
In general, accretion, tuning, and restructuring represent three modes of cognitive learning that have
different implications for the status of a schema. However, new information can also have other effects on
current knowledge, including two that are important
for schema development: A learner may use new information to create and maintain ▶ cognitive dissonance.
When new information does not fit with what is
known, that information can be considered as an
anomaly – something apparently real but presently
incomprehensible. In this case, people activate their
schemas to define what is “known” and then actively
seek known anomalies in order to test and update the
known. A learner may also use new information to
clarify or increase his or her degree of certainty in
a schema and thus confirm it.

Important Scientific Research and
Open Questions
In cognitive and educational psychology, one can find
the idea that schemas are stored in memory and can be
activated whenever input information fits with their
defaults. In the case of PDP models, however, schemas
are conceived of as properties of entire neural networks
rather than as single units or small circuits. Accordingly, schemas are not “things” and there is no representational object in the mind which is a schema
(cf. Rumelhart et al. 1986). What is stored is a set of
connection strengths which generate states that correspond to instantiated schemas. Schemas emerge at the
moment they are needed from the interaction between
large numbers of much simpler elements all working in
concert with one another.
In general, schemas are assumed to be “working
models” composed of thematically related belief systems which are assembled during the entire course of
life experiences. Beck defined a schema as “a structure
for screening, coding, and evaluating the stimuli that
impinge upon an organism . . . On the basis of the
matrix of schemas, the individual is able to orient
himself in relation to time and space and to categorize

and interpret experiences in a meaningful way” (Beck
1967, p. 419). Developmental psychologists regularly
emphasize the central role of early experiences in the
formation of schemas, but cognitive-oriented psychotherapy argues that schemas may be modified at any
time during life.
However, the development of schemas is in general
a lengthy process which requires many different kinds
of learning experiences and is dependent on favorable
conditions for cognitive development. However, at the
moment we do not know how many repetitions of
experiences with specific instances are necessary and
sufficient for the development of a new schema.
Another interesting question in current research on
human cognition aims at determining the conditions
under which a schema is revised, supplemented,
replaced, or abandoned. Clearly, such changes are at
the core of learning. When important schemas closely
related to one’s sense of self (“core schemas”) are superseded or shattered, extreme confusion, suffering, and
meaninglessness can result. This means that they evoke
affects and emotions and constitute the fundamental
basis of emotional schemas.
Finally, since psychological work on schemas has
come about quite separately from neurological or
neuropsychological research, it becomes necessary
to seek out parallels between the former and the latter
in order to postulate a neuropsychological substrate
for the schemas. On the one hand, cognitive psychology has focused on generating plausible models and
then checking for experimental and, more rarely,
neuropsychological viability. On the other hand, neuropsychological analysis of psychological constructs
such as schemas has been limited due to insufficient
methods and techniques of assessment as well as the
tendency of neuropsychological research to focus
upon “lower-order,” observable, or measurable brain
processes.
However, there have been some neuropsychological
findings which have to some extent gone beyond the
limits encountered by Piaget thanks to the development of new diagnostic and exploratory technology
and methodology. These newer sorts of studies show
“up-close” neuropsychological mechanisms which may
well be implicated in the formation and maintenance of
cognitive schemas. However, current understanding of
the neurological substrate for intentionality, motivation, and consciousness suggests that such high-level

Schema Therapy

cognitive processes are not seated in any single area of
the brain but rather seem to arise from complex feedback loops based on neural pathways linking numerous
neuropsychological functions. Eigen and Schuster
(1979) proposed a cyclic reaction scheme, termed
▶ hypercycle, in which each replicator would aid in
the replication of the next one in a regulatory cycle
closing in on itself. Hypercycles can be conceived of as
dynamic patterns of energy which may play a role in the
formation of schemas based on patterned tendencies
built into the brain.

Cross-References
▶ Action Schema(s)
▶ Emotional Schema(s)
▶ Generalization versus Discrimination in Learning
▶ Schema(s)
▶ Schema-Based Learning

S

2939

Schema Focused Therapy (SFT)
▶ Schema Therapy

Schema Therapy
JOAN M. FARRELL1, HEATHER FRETWELL2, NEELE REISS3
1
Indiana University School of Medicine, Schema
Therapy Institute – Midwest Indianapolis Center,
Indianapolis, IN, USA
2
Midtown Community Mental Health Center, Indiana
University School of Medicine, Indiana, USA
3
Department of Psychiatry and Psychotherapy,
University Medical Center Mainz, Mainz, Germany

Synonyms
References
Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical
aspects. New York: Harper & Row.
Eigen, M., & Schuster, P. (1979). The hypercycle: A principle of natural
self-organization. Berlin/Heidelberg/New York: Springer.
Piaget, J. (1954). The construction of reality in the child. New York:
Basic.
Rumelhart, D. E., & Norman, D. A. (1978). Accretion, tuning, and
restructuring: Three models of learning. In J. U. Cotton & R. L.
Klatzky (Eds.), Semantic facts in cognition (pp. 37–54). Hillsdale:
Erlbaum.
Rumelhart, D. E., Smolensky, P., McClelland, J. L., & Hinton, G. E.
(1986). Schemata and sequential thought processes in PDP
models. In J. L. McClelland, D. E. Rumelhart, & The PDP
research group (Eds.), Parallel distributed processing. Explorations
in the microstructure of cognition (Volume 2: Psychological and
biological models, pp. 7–57). Cambridge, MA: MIT Press.
Seel, N. M., Ifenthaler, D., & Pirnay-Dummer, P. N. (2009). Mental
models and problem solving: Technological solutions for measurement and assessment of the development of expertise. In
P. Blumschein, W. Hung, D. Jonassen, & J. Strobel (Eds.), Modelbased approaches to learning: Using systems models and simulations to improve understanding and problem solving in complex
domains (pp. 17–40). Rotterdam: Sense Publ.

Schema Focused Cognitive
Therapy
▶ Schema Therapy

Schema; Schema focused cognitive therapy; Schema
Focused Therapy (SFT); Schematherapy

Definition
Schema Therapy (ST) is an integrative psychotherapy
approach that grew out of efforts to more effectively
treat patients with personality disorders who had limited response or high relapse rates with other treatments. ST is based on an integration of insights from
cognitive and experiential therapies and the theory and
research in the areas of object relations, attachment,
and developmental psychology into a system of therapy
with a coherent conceptual model that translates readily into clinical practice. The major developer of the
theory is Jeffrey Young, Ph.D. A key collaborator,
Arnoud Arntz, Ph.D., is the leader of research to empirically validate this approach to treatment of personality
disorder patients (Arntz and van Genderen 2009). In
comparison with standard cognitive therapy, schema
therapy probes more deeply into the childhood origins
of distorted thinking, relies more on imagery and
emotion-focused techniques, and is of somewhat longer term.
The two major constructs differentiating it from
other types of psychotherapy are schemas and modes
of operation. The major stages of the therapy are
(1) identification of major schemas and modes, (2) cognitive change work, (3) experiential change work, and

S

2940

S

Schema Therapy

(4) breaking maladaptive behavioral patterns. The
overall goal of the therapy is to develop the Healthy
Adult mode of a person to meet their underlying needs
in adaptive ways. The therapist attitude is one of high
involvement and empathy. In ST terms, the therapist
stance is described as “limited reparenting.” Applications of ST include the treatment of personality disorders including borderline personality disorder, Cluster
C disorders, forensic populations, treatment-resistant
depression, and couples therapy. ST can be administered individually or in a group setting (see separate
entry for Group ST).

Theoretical Background
Schema therapy (ST) was developed conceptually in
the late 1980s and early 1990s, a time when numerous
investigators had training in both psychoanalytically
derived models and cognitive behavioral therapies,
and recognized the limitations of narrow approaches
and were moving toward psychotherapy integration
(Edwards and Arntz 2011). ST derives principally
from Cognitive Therapy (established by Aaron Beck),
attachment theory (particularly the work of John
Bowlby), Gestalt therapy, and the idea that corrective
emotional experiences are needed within therapy
(Ferenczi and Alexander) (summarized in Edwards
and Arntz, 2011).
Maladaptive ▶ schemas are psychological constructs that include beliefs that we have about ourselves, the world, and other people which result from
interactions of unmet core childhood needs, innate
temperament, and early environment. They are comprised of memories, bodily sensations, emotions, and
cognitions that originate in childhood and are elaborated through a person’s lifetime. These schemas often
have an adaptive role in childhood (e.g., in terms of
survival in an abusive situation it engenders more
hope for a child if they believe they are defective
as opposed to the adult). By adulthood, maladaptive
schemas are inaccurate, dysfunctional, and limiting,
but strongly held and frequently not in the person’s
conscious awareness. Young (1990) identified 18 maladaptive schemas in patients with personality disorders
that represent the following psychological constructs:
I: Disconnection and rejection, II: Impaired autonomy and performance, III: Impaired limits, IV:
Other-directedness, V: Over-vigilance and inhibition.
A self-report questionnaire was developed to evaluate

the strength of schemas endorsed – the Young Schema
Questionnaire.
An important development in ST was the addition
of the concept of “modes of operation” or schema
modes (Young et al. 2003). A schema mode is defined
as the current emotional, cognitive, and behavioral
state that a person is in. Dysfunctional modes occur
when multiple maladaptive schemas are triggered.
Generally speaking, four categories of modes are
defined. Primary Child modes (e.g., vulnerable child,
angry child) are said to develop when certain basic
emotional needs (such as safety, nurturance, or autonomy) are not adequately met in a patient’s childhood.
These “child modes” are defined by intense feelings,
such as fear, helplessness, or rage and represent the
innate reactions a child would have. Dysfunctional
Parent modes such as Punitive Parent or Demanding
Parent mode define the second category of modes.
Dysfunctional Parent modes reflect the internalization
of negative aspects of attachment figures (e.g., parents,
teachers, and peers) during childhood and adolescence.
When a patient is in a dysfunctional parent mode they
experience self-devaluation, self-hatred, and/or they
put extremely high pressure upon themselves.
Dysfunctional coping modes, a third category of
modes, are defined by an overuse of coping strategies
such as overcompensation, avoidance, or surrender
with a goal of protecting vulnerable modes from further pain. In a fourth category healthy modes, such as
healthy adult mode or happy, joyful child mode, are
found. The Healthy Adult mode includes functional
thoughts and balanced behaviors, and the Happy Joyful
Child mode is a resource for playful and enjoyable
activities, especially in social networks. Modes are
often triggered by events that patients experience as
highly emotional. Modes can switch rapidly in patients
suffering from severe personality disorders such as
BPD, which results in sudden changes in behavior or
seemingly disproportionate reactions. In general terms,
the goal of schema therapy is to correct the maladaptive
schemas via cognitive and experiential techniques and
strengthen the healthy modes. A solid Healthy Adult is
able to nurture and set limits with child modes, banish
dysfunctional parent modes, replace the need for dysfunctional coping modes, has access to the Happy
Child, and is able to function adaptively in the world
and have a good quality of life. A self-report questionnaire, the Schema Mode Inventory, was developed by

Schema Therapy

Young and colleagues at the Schema Therapy Institute of
New York and later tested and refined psychometrically
by Arntz and Lobbestael. This questionnaire is a helpful
therapy tool and is used in schema therapy outcome
studies to assess decreases in maladaptive modes and
increases in or strengthening of healthy modes.
The therapist in ST must offer not only a collaborative adult relationship, as in standard cognitive therapy, but also a parenting relationship to the client’s
child side. This is seen as necessary to correct dysfunctional schemas and to allow healthy new schemas to
form in the context of a healthy reparenting relationship in therapy. According to Young (1990), this would
involve finding out what basic emotional needs were
not met in childhood and meeting them to a reasonable
degree in therapy. The ST term “reparenting” describes
the therapists’ caring attitude toward their clients and
their acting as a “good parent” would early in therapy,
in particular when dealing with the Child modes. In
experiential work with the Vulnerable Child the therapist provides validation and nurturing within the
bounds of a professional relationship. This position
has been the source of some controversy and confusion.
The basic position of ST is that patients who experienced depriving and/or abusive early caregivers must
experience positive parenting before they can learn to
do this for themselves. The goal of ST is autonomy, so
this early focus on the therapist doing the reparenting is
ultimately replaced by a developed and strengthened
Healthy Adult mode of the patient. Reparenting targets
the need of the mode the patient is in, so the Angry or
Impulsive Child modes are provided with the empathic
confrontation or limit-setting needed in those modes.
The therapist in ST is also active in helping the patient
replace maladaptive Coping modes like detachment
with more adaptive coping and in assisting the patient
in banishing the dysfunctional parent modes. This
empathic, active involvement, and modeling by the
therapist is termed “limited reparenting.”
The cognitive techniques of ST are often familiar to
those trained in cognitive therapy, and include monitoring thoughts, identifying distortions, examining
evidence for and against a negative core belief, and
reframing the belief. However, unlike many cognitive
therapies, the therapist also assists the patient in linking
distorted thoughts to the underlying maladaptive
schema from the beginning of therapy. This enables
a “bottom-up” approach in contrast to a “top-down”

S

2941

approach focusing more on surface thoughts. A major
innovation of ST in relation to CBT is the central place
given to experiential techniques. These addressed
a problem Young had identified – that clients in spite
of working with rational analysis, changing negative
thoughts and even experimenting with new behaviors,
still failed to achieve change at the emotional level.
Experiential techniques used in ST include imagery
work, chair work, and mode dialogues in which the
patient improves conscious access to various modes.

Important Scientific Research and
Open Questions
Applications
Schema therapy was originally developed to treat
patients with personality disorders (PD) (particularly
borderline and narcissistic PD) and treatment-resistant
depression in an individual therapy setting. The effectiveness of ST for treating borderline personality disorder (BPD) was tested in a randomized controlled trial
in the Netherlands led by Arnoud Arntz (Giesen-Bloo
et al. 2006). In this trial, 2 years of individual ST was
compared to the same amount of individual psychodynamic psychotherapy for patients with BPD. Patients
in the ST arm showed improvement in all major
domains of BPD symptoms, had higher retention
rates, and were more likely to no longer meet the
criteria for BPD by the end of the study. ST was also
demonstrated to be significantly more cost effective
than the comparator arm and therapists and patients
reported higher satisfaction rates for ST compared to
the psychodynamic comparator. An implementation
trial conducted in general health care, also in the Netherlands, demonstrated similar effectiveness of individual ST for BPD. ST has also been tested with positive
outcome in mixed personality disorders, cluster
C personality disorders, and in forensic settings with
patients with antisocial personality disorder.
ST has been adapted for groups by Farrell and Shaw
and tested in an RCT with BPD patients (Farrell et al.
2009). There is evidence to suggest that curative factors
of groups may accelerate treatment effects and group
delivery of treatment has the potential to make ST even
more accessible and cost effective. A large international
multisite RCT to further test the group model is underway led by Arntz and Farrell (see separate entry on
Group ST). In various studies, many patients with

S

2942

S

Schema Therapy Group

BPD feel that a schema therapy experience was “the
first time I felt understood” and readily identify with
the mode model that describes their own inner experience of rapid shifts in moods and behaviors. In additional clinical applications, ST has become popular
with couples therapy, as patients find the concept of
marital friction being worsened by cross-activation of
various schemas and modes to be very helpful.

Conclusion
Schema therapy has become increasingly popular,
reflected in its implementation in many clinical settings
and the number of studies focusing on this particular
type of therapy. Patients who have not responded to
other therapy approaches often engage well to the
warmth and empathy of the limited reparenting therapeutic stance, and relate to the user-friendly concepts
and strategies utilized by ST. Future directions include
additional empirical validation for the group STmodel;
development of treatment and preventative work for
children and adolescents at risk for personality disorders; trials testing the effectiveness of ST for eating
disorders, substance disorders, and other conditions
considered “difficult to treat.”

Cross-References
▶ Cognitive-Behavioral Family Therapy
▶ Group Schema Therapy
▶ Maladaptive Schema(s)

References
Arntz, A., & van Genderen, H. (2009). Schema therapy for borderline
personality disorder. New York: Wiley.
Edwards, D., & Arntz, A. (2011). A brief history of schema therapy.
In M. vanVreeswijk, M. Nadort, & J. Brierson (Eds.), Handbook
of schema therapy. New York: Wiley.
Farrell, J. M., Shaw, I. A., & Webber, M. A. (2009). A schema-focused
approach to group psychotherapy for outpatients with borderline personality disorder: A randomized controlled trial. Journal
of Behavior Therapy and Experimental Psychiatry, 40, 317–328.
Giesen-Bloo, J., van Dyck, R., Spinhoven, Ph, van Tilburg, W.,
Dirksen, C., van Asselt, Th, Kremers, I., Nadort, M., &
Arntz, A. (2006). Outpatient psychotherapy for borderline personality disorder: A randomized trial of schema-focused therapy
vs. transference-focused psychotherapy. Archives of General Psychiatry, 63, 649–658.
Young, J. E. (1990). Cognitive therapy for personality disorders:
A schema-focused approach. Sarasota: Professional Resource
Press.
Young, J. E., Klosko, J., & Weishaar, M. E. (2003). Schema therapy:
A practitioner’s guide. New York: Guilford.

Schema Therapy Group
▶ Group Schema Therapy

Schema-Based Architectures of
Machine Learning
GIOVANNI PEZZULO1,2, MARTIN V. BUTZ3
1
Istituto di Linguistica Computazionale “Antonio
Zampolli”, National Research Council, Pisa, Italy
2
Istituto di Scienze e Tecnologie della Cognizione,
National Research Council, Roma, Italy
3
Department of Cognitive Psychology III, University of
Würzburg, Würzburg, Germany

Synonyms
Cognitive architectures; Society of mind

Definition
Schema-based architectures (SBAs) consist of collections of modularly and hierarchically organized
▶ schemas, which constitute building blocks for perception, cognition, and action. An SBA organizes these
schemas in such a way so that action selection, motor
coordination, and cognition in the general sense interact effectively. SBAs were mainly inspired by theories of
▶ sensorimotor adaptation and ▶ cognitive development and learning. Particularly ▶ Jean Piaget’s research
on and theories of cognitive development in infants
and children inspired the design of SBAs. ▶ Machine
learning develops algorithms to learn, structure, and
continuously adapt SBAs. Various forms of representations are used to develop SBAs, including symbolic
representations, rule-based representations, as well as
neural network representations.

Theoretical Background
Schema-based architectures have two main features.
First, the computation is distributed among independent components (the schemas), in contrast to one
centralized process. Second, SBAs encode a complete
perception-action loop, in the sense that they include
both perceptual and motor elements (plus, possibly,

Schema-Based Architectures of Machine Learning

internal representations). Thus, SBAs realize the control of a whole system.
Due to these coordinated interactions of schemas,
the success of an SBA not only depends on the quality
of individual schemas but also on the quality of
their coordination and their collective performance.
Individual schemas in an SBA are specialized components that are responsible for a particular type of
operation – such as the recognition of a specific feature (e.g., red) or of a specific class of objects (e.g.,
apples), the control of a particular action (e.g., grasping), or even the realization of sequences of perceptions or actions (e.g., a concatenation of reaching,
grasping, and lifting an object). The coordination
of schema interactions ensures that the schemas successfully cooperate in order to achieve complex behaviors beyond the abilities of individual schemas and, at
the same time, that they effectively compete for the
chance to be active. An SBA regulates the schema activities and thus determines which schemas take control
of sensory processing and effector control. This feature
of SBAs is usually referred to as cooperative competition
(Arbib 2003).
SBAs may be further classified into reactive SBAs,
which consist of a collection of stimulus-action
schemas, and anticipatory SBAs, which consist of a
collection of action-effect and condition-action-effect
schemas. The former class of schemas has been widely
used in the realization of behavior-based architectures,
in which reactivity is an essential feature that distinguishes them from traditional, symbolic, deliberative,
and logic-based AI architectures. In these architectures,
reactivity ensures quick responses to rapidly changing
environments but it also reveals interesting emergent
behaviors, which partially exceed the capabilities of
traditional AI and robotic architectures. However, reactive architectures were found to be difficult to apply to
more complex problems, which require additional
capabilities such as, most notably, the ability to represent and reason about the future. Moreover, an increasing number of psychological experiments exhibit
▶ latent learning and generally ▶ anticipatory learning
in animals – neither of which are possible with reactive
architectures or reactive ▶ reinforcement learning
principles. Thus, anticipatory SBAs take expected
future states – be they action effects or other changes
in the environment – into account for the determination of current schema activations.

S

2943

Finally, SBAs may be classified into flat and hierarchical SBAs. In flat SBAs, sensory information is transferred into motor commands in a flat architecture
without further abstractions. While the transfer process
may still be reactive or anticipatory, there are no
abstractions in the involved schemas in that higherlevel schemas may initiate or coordinate the activation
of lower-level schemas. In hierarchical SBAs, such hierarchical schema interactions additionally enable the
coordination of complex sequences of movements as
well as the representation of more abstract concepts of
perceptions and actions, such as something edible in
contrast to a particular apple. Hierarchical architectures theoretically enable SBAs to realize more complex
and flexible behavior patterns because current priorities and goals can be assigned to more abstract schemas,
which in turn activate more concrete schemas dependent on the current availability in the environment. For
example, a “hungry” SBA may activate the search for
something edible, which may result in an apple being
found. Next, that SBA may need to coordinate
a consumption schema, which will first activate an
appropriate grasping action, coordinated movements
to the mouth and movements of the mouth, etc.
Regardless of which type of SBA is used, the biggest
challenge is to design or learn appropriate schema
interactions. Each schema usually has an own-activity
level, which is influenced by the interacting schemas as
well as by the current sensory and motor activities.
How these interactions are realized strongly depends
on the particular SBA being used, the machine learning
mechanisms involved, and the available representations, sensory perceptions, and motor activities.

Important Scientific Research and
Open Questions
Various SBAs have been developed in machine learning. One of the pioneering systems in this field is
Drescher’s (1991) schema architecture: a schemabased agent that operates in a simplified environment,
with schemas having the form of action-effect rules.
One interesting aspect of this architecture is that the
agent is able to interactively enlarge its ontology by
learning new synthetic items. If the system recognizes
that there was a common cause of a set of related
interactions, it postulates that an object caused this
effect. For instance, if the agent recognizes that pushing
left resulted in a stop, looking left resulted in seeing red,

S

2944

S

Schema-Based Architectures of Machine Learning

etc., it postulates that a (red) object was in the left.
Importantly, learning new synthetic items can be taken
as a basis for the incremental construction of increasingly complex ontologies and behaviors, since the agent
can then postulate new objects from a combination of
old objects, and learn the effects of its actions upon
them.
Anticipatory learning classifier systems (ALCS)
(Butz 2002) also develop sets of condition-action-effect
rules. They are learned by means of combinations of
statistical, evolutionary, and rule-based mechanisms
and are based on the principle of anticipatory learning.
Moreover, ALCSs include generalization capabilities in
the learning mechanism so that overspecialized rules
can be generalized to be applicable in a broader context.
Generally, in ALCSs as well as in Drescher’s schema
architecture a rule is activated given its condition is
sufficiently satisfied. Once activated, it competes with
other currently active rules. The SBA chooses the rule
with the most promising effect based on the anticipated
effects. Such effects may be chained to base action
selection on priorities that lie even further in the future.
The most promising action is executed and the process
starts anew.
Pezzulo and Calvi (2007) have proposed a schemabased architecture that uses the self-organizing principles of local activation (that is, schemas that transfer
activation to each other) and global inhibition (that is,
limited activation resources) for realizing cooperation
and selection of schemas. Self-organizing neural networks are a good candidate to also shape hierarchical
SBAs, but research in this direction is still in its infancy.
It is particularly difficult to assess when a hierarchy
should be established, since this decision may not
only depend on the structure of the available perceptions, but also on the possible interactions with the
environment as well as on the priorities of the system.
Hierarchical reinforcement learning approaches may
currently provide the best framework and most studies
into the formation of such hierarchical SBAs – albeit
mainly only in reactive architectures.
Also neural network-base SBAs have been developed for the coordination of schemas. Particularly the
forward-inverse framework of Wolpert and Kawato
(1998) builds an interesting approach to schema activity coordination: each schema contains a neural forward model of action-dependent system changes.

If the forward model of a schema fits well with the
currently encountered changes, then it is also a good
candidate to be activated for control. If additional
current system priorities also favor the activation of
this schema, its associated inverse controller may be
activated.
The extension of SBAs’ capability beyond the sensorimotor domain is certainly another crucial aspect of
schema learning. One important example is to learn
about other entities in the environment, since this
knowledge constitutes one essential element of
advanced conceptual abilities in humans and it can be
associated with priorities and behaviorally relevant
aspects, such as food, water, or mates. The detection
and distinction of entities in the environment must
thus be guided by machine learning methods that
take into account internal rewards or priorities,
external perceptual structures, as well as sensorimotor
contingencies. Learning such concepts leads to a
schema-based representation of affordances since
schemas will associate entities directly with those interactions that are meaningful and possible.
Another aspect of schema learning, beyond the
sensorimotor domain, is the acquisition of linguistic
and communicative abilities through schemas. Roy
(2005) has proposed how sensorimotor and linguistic
processes may be coordinated based on an SBA. First,
representations of objects may be built based on
encountered sensorimotor interactions with them and
schemas may be associated to manipulate them. Later
on, these schemas may be associated with linguistic
representations of the same objects and interactions.
Such an SBA thus integrates schemas that can recognize, manipulate, and linguistically refer to objects.
How such an SBA may be learned with machine learning techniques remains one of the big challenges in the
further development of SBAs in machine learning.
In general, how hierarchical, anticipatory SBAs may
be represented – most likely with probabilistic, Bayesian-based representations that associate schemas in
various ways conditioned on current internal system
priorities, current context, and current perceptions and
motor activities – and how such representations may be
learned rigorously (most likely with ▶ Bayesian learning, ▶ connectionist theories of learning, or ▶ association learning techniques) are highly challenging
research questions, which still lead to much debate

Schema-Based Instruction

and controversy. In addition to the theoretical and
methodological difficulties, one of the main challenges
may lie in the definition of a sufficiently general problem setting that closely resembles natural problems.
Once such SBAs are designed and improved by
machine learning techniques, SBAs may yield even
more power for the design and development of cognitive systems and robots. The resulting systems will not
only be able to exploit and recombine their schemas to
achieve their goals in (partially) novel environments and
situations, but also to autonomously enlarge their basic
schema repertoire by learning novel schemas at different
hierarchical levels where necessary. These two processes
are particularly relevant for schema learning, and correspond to the two main aspects of adaptation in ▶ Jean
Piaget’s theory: Assimilation and Accommodation.

Cross-References
▶ Anticipatory Learning Mechanisms
▶ Anticipatory Schemas
▶ Associationism
▶ Bayesian Learning
▶ Connectionist Theories of Learning
▶ Development and Learning
▶ Machine Learning
▶ Piaget, Jean
▶ Reinforcement Learning
▶ Schema(s)
▶ Sensorimotor Adaptation

References
Arbib, M. A. (2003). Schema theory. In The handbook of brain theory
and neural networks (Vol. 2). Boston: MIT Press.
Butz, M. V. (2002). Anticipatory learning classifier systems. Boston:
Kluwer.
Drescher, G. L. (1991). Made-up minds: A constructivist approach to
artificial intelligence. Boston: MIT Press.
Pezzulo, G., & Calvi, G. (2007). Schema-based design and the AKIRA
schema language: An Overview. In M. V. Butz, O. Sigaud,
G. Pezzulo, & G. Baldassarre (Eds.), Anticipatory behavior in
adaptive learning systems: From brains to individual and social
behavior (Lecture notes in artificial intelligence, Vol. 4250,
pp. 128–152). Berlin: Springer.
Roy, D. (2005). Semiotic schemas: a framework for grounding
language in action and perception. Artificial Intelligence, 167,
170–205.
Wolpert, D. M., & Kawato, M. (1998). Multiple paired forward
and inverse models for motor control. Neural Networks, 11,
1317–1329.

S

2945

Schema-Based Decision Making
▶ Schemas and Decision Making

Schema-Based Expectations
▶ Anticipatory Schemas

Schema-Based Instruction
SANDRA P. MARSHALL
Department of Psychology, San Diego State University,
San Diego, CA, USA

Synonyms
Problem solving; Visual representation

Definition
▶ Schema-based instruction is a method of teaching
problem solving that emphasizes both the semantic
structure of the problem and its mathematical structure. It utilizes recognition of key words (as does
a simple key-word strategy) but goes further than simple recognition to stress understanding of the situation
represented in the problem.

Theoretical Background
Most schema-based instruction has targeted arithmetic
word problems or algebra word problems. Research has
shown that students have difficulty solving these problems for many reasons. A large number of errors can be
traced to improper understanding of the problem itself,
rather than inability to do the correct arithmetic. In
the 1980s and 1990s, many psychological and mathematical educational researchers became interested in
the basic cognitive issues required in problem solving
(see ▶ schema-based problem solving) and began to
investigate the importance of visual representation of
a schema.
Little research has been done on schema-based
instruction and most of it to date focuses on problem

S

2946

S

Schema-Based Learning

solving. At the heart of such instruction are graphical
representations of the situations that may occur in
problems (Chandler and Sweller 1991; Marshall
1995). These visual representations attempt to capture
the basic relationships that exist in the problem and
help students to construct mental models of the problem. For example, simple diagrams have been used
successfully in a series of studies comparing traditional
instruction with schema-based instruction for word
problems at the third grade and ratios and proportions
at the seventh grade (Jitendra et al. 2009; Griffin and
Jitendra 2008). Students given schema-based instruction generally performed better than students with
traditional instruction.

Important Scientific Research and
Open Questions
Students in many fields (e.g., science, mathematics, and
engineering) are encouraged to make visual representations of problems they encounter, but they often
have no existing cognitive structures that help them
do so. Schema-based instruction is designed to remedy
this lack. Thus far, most formal schema-based instruction has been directed at elementary school mathematics. Further research is required to determine how
well this method of instruction fares in more advanced
arenas.

Cross-References
▶ Assimilation Theory of Learning
▶ Generative Teaching: Improvement of Generative
Learning
▶ Schema(s)
▶ Schema-Based Learning
▶ Schema-Based Problem Solving

Schema-Based Learning
JUNGMI LEE, NORBERT M. SEEL
Department of Educational Science, University of
Freiburg, Freiburg, Germany

Synonyms
Schema-Centered Learning; Schema-Oriented Learning

Definition
Schema-based learning is a central theoretical approach
in cognitive and educational psychology as well as in
artificial intelligence. Schemas allow learners to reason
about unfamiliar learning situations and interpret these
situations in terms of their generalized knowledge. In
cognitive and educational psychology, schema-based
learning is grounded in capturing and using expertgenerated schemas as frameworks for teaching and
learning. Schemas can be learned to promote the acquisition of new scientific knowledge and skills. At the core
of this approach is a generalization and abstraction
method designed to extract and condense as much
information as possible from a single example of successful task completion or problem solving.
In the field of artificial intelligence, schema-based
learning is conceived as a generalized framework for the
design of integrated adaptive autonomous agents
aiming at the incorporation of general principles of
adaptive organization and coherence maximization.
Schema-based learning allows the development of
increasingly complex patterns of interaction between
the agent and its environment by confining statistical
estimation to a narrow criterion.

Theoretical Background
References
Chandler, P., & Sweller, J. (1991). Cognitive load theory and the
format of instruction. Cognition & Instruction, 8, 293–332.
Griffin, C., & Jitendra, A. (2008). Word-problem solving instruction
in inclusive third-grade mathematics classrooms. Journal of Educational Research, 102, 187–201.
Jitendra, A., Star, J., Starosta, K., Leh, J., Sood, S., Caskie, G., Hughes,
C., & Mack, T. (2009). Improving seventh grade students’ learning of ratio and proportion: The role of schema-based instruction. Contemporary Educational Psychology, 34, 250–264.
Marshall, S. (1995). Schemas in problem solving. New York: Cambridge University Press.

Schema-based learning builds on the schema theory.
The theoretical assumption of schema-based learning
is that newly gained knowledge is assimilated into
preexisting knowledge and organized to form schemas.
However, it is not enough to simply have a collection of
passive schemas: one must also know how to use them.
Anderson cites two types of situations involving
schemas, namely: activation of preexisting schemas
and construction of new schemas. However, Bransford
(1984) points out that while it is possible to activate
existing schemas with a given topic, this does not

Schema-Based Learning

necessarily mean that a learner can use this activated
knowledge to develop new knowledge and skills.
Accordingly, Bransford stresses the importance of helping learners to “activate various preexisting ‘packets’ of
knowledge” and to “reassemble and construct this
knowledge into an integrated new schema” (Bransford
1984, p. 264). Bransford and colleagues (1981) also
argue that the new learning situation should help
learners to fully understand the “significance of new
facts” as it helps them to reorganize the previously
unrelated facts in a more meaningful way.
Schema-based learning emphasizes three issues:
● It is advantageous (efficient) to cope with new

experiences on the basis of positive results from
previous similar experiences.
● Schema-based learning allows incremental development of complex cognitive structures through
aggregation from a restricted stock of schemas
(i.e., stable units of composition) to more complex
interactive structures.
● Schema-based learning provides the learner with
instantiation, accretion, tuning, and reconstruction
of knowledge. As a consequence, new schemas can
be formed.
At the heart of this approach is a generalization and
abstraction method which is designed to extract and
condense as much information as possible from single
examples of successful task completion or problem
solving.
Essential constraints from integral parts of explanations of success must be incorporated into the results
of generalization. There are two types of essential
constraints:
● Internal-schema constraints connect component

schemas together to form a complex schema. Technically, these are the assertions which support goal
achievement by identifying parameters of instances
of schematic forms with instances of ▶ prototypes.
● Intra-schema constraints ensure that each component of a complex schema is complete. Technically,
these are the immediate supporters of the “self ”
constraints which support the achievement of
the goal.
Optional constraints are “forced” or implied by
essential constraints. They need not be included but
should be. They do not alter generality but improve

S

efficiency. Technically, any assertion which has some
justification depending purely on essential constraints
is a member of this class. Extraneous constraints include
most instantiation bindings and all implications based at
least in part on extraneous constraints. Schema-based
learning allows the development of increasingly complex
patterns of interaction between the person (or an agent)
and its environment by starting with a limited stock of
simple schemas that allow efficient learning by confining
statistical estimations of future events within the realm
of a relatively narrow space.
The literature on schema-based learning theories
describes three types of learning: accretion, tuning,
and cognitive restructuring (Rumelhart and Norman
1978). Research conducted in the area of cognitive
restructuring is generally categorized under the heading of ▶ conceptual change. A traditional method used
to facilitate conceptual change is to provide the learner
with examples that contradict their “naı̈ve theories,”
which is referred to as the anomalous data approach.
Schema construction is a central element in schemabased learning. Learning is triggered when predictions
made by anticipatory schemas do not match the
observed results and the system does not reach the goal
expectations. The dynamics of the system first try to
reduce the error by tuning or accreting the current
stock of schemas, and when this fails, a new schema
must be constructed to remedy the error. In the long
term, the reliability of the predictions will increase, making the system increasingly better at predicting the
results of actions for a given context. In many cases,
the agent will construct both a new schema and its
corresponding predictive schema. Actually, most predictive/anticipatory schemas have to be learned by the agent
through continuous interactions with the environment.
In the field of informatics and artificial intelligence,
schema-based learning is a data-driven, constructivist
approach used to discover probabilistic action models
within environments that serve as a generalized framework for designing integrated adaptive autonomous
agents and predicting their actions by incorporating
general principles of adaptive organization and coherence maximization (Corbacho 1998). A schema is
defined as an experience-based recurrent pattern of
interaction with the environment, and coherence is
a measure of the congruence between the result of an
interaction with the environment and the expectations
the agent has for that interaction.

2947

S

2948

S

Schema-Based Learning

Schema learning comprises two basic phases: discovery, in which context-free action/result schemas are
found, and refinement, in which context is added to
increase reliability.

Important Scientific Research and
Open Questions
In cognitive and educational psychology, schema-based
learning is grounded on the concept of equilibration in
Piaget’s epistemology. According to Piaget, equilibration
involves both assimilation and accommodation. During
each stage of development, people conduct themselves
with certain logical internal mental structures that allow
them to make adequate sense of the world. When the
newly learned information packet does not match with
the learner’s existing internal mental structures (existing
schema), equilibration helps the learner to make sense of
the world around him by assimilating new information
into pre-existing mental schemes and accommodating it
when necessary through logical thinking. This internal
mental process enables the learner to construct more
sophisticated schemas, which in turn leads to cognitive
development.
Pascual-Leone retained Piaget’s view of the
human cognitive process as a highly dynamic and
self-reflective system which passes through stages of
stability and disequilibrium in the course of cognitive
development. Pascual-Leone and Goodman (1979)
distinguish several operators which are involved in
the construction of new knowledge based on the
assumption that the construction of new knowledge
can be understood as a process of mental representation of environmental patterns. This process is strongly
influenced by existing assimilative schemas. Accordingly, field forces compete with operators to determine
which of the many schemas will be activated to regulate
the processing of new information. During the course
of processing, the activated schemas are enriched or
restructured by internal operators. Pascual-Leone and
Goodman distinguish between L-operators, which represent former learning experiences, A-operators, which
represent emotional and affective side-effects, and Boperators, which represent stable personality traits.
Finally, they add an M-operator, a moderator for
using information processing capacity gained during
development.
Ausubel’s theory of meaningful learning (1968)
is also meant to help students activate their

pre-existing knowledge so that it can be assimilated,
tuned, and restructured into new schemas. Hence,
Ausubel introduced the theory of “advance organizers” to support the learner’s pre-existing knowledge. Advance organizers can help learners to
activate schemas more effectively and allow them to
use their pre-existing knowledge in a more effective manner. However, Bransford (1984) argues that
advance organizers should be written differently
depending on whether they are to be used for schema
activation or schema construction. He states that an
advance organizer can be effective if the learner has
already acquired the necessary schemas for the given
problem. However, it will not be of much help for
schema construction.
Bruner’s (1966) cognitive structure (i.e., schema,
mental models) is consistent with schema-based learning in that it describes learning as an active process in
which learners construct new ideas or concepts based
on their existing knowledge. The learner does not passively respond to stimuli but actively selects and transforms information, constructs hypotheses, and makes
decisions based on his or her cognitive structure.
Bruner asserts that the cognitive structure supports
the learner, actively generates meaning for real-world
experiences, and allows the individual to process information in a meaningful way. He emphasizes that the
learning situation should help learners to actively reorganize the new information, allowing them to build on
existing knowledge in a meaningful way and use the
newly gained meaningful knowledge effectively in the
future.
Wittrock’s concept of generative learning asserts
that the learner actively constructs the whole process
of learning. Wittrock (1991) states with regard to the
generative model that “learners must construct
between stored knowledge, memories of experience,
and new information for comprehension to occur.”
Accordingly, an important aspect of generative teaching is knowing learners’ preconceptions, how to modify
these preconceptions, and how to induce them to generate a new model of the phenomenon by revising or
transforming their models (Wittrock 1991).
Cognitive load theory assumes that learning consists primarily of the acquisition of schemas. Sweller
(1988) understands schemas as the cognitive structures
that compose an individual’s knowledge. Learning
requires a change in the schematic structures of long-

Schema-Based Problem Solving

term memory and is demonstrated by performance
that progresses from slow and difficult (novice-like)
to smooth, fast, and effortless automation (expertlike). The change in performance occurs as the learner’s
schemas are increasingly associated (activated) with the
learning material. Consequently, it requires instructional techniques that take the optimal level of cognitive load into account and do not interfere with schema
acquisition (Sweller 1988).
In the 1980s, many knowledge-based AI systems
used schemas (so-called knowledge packets) as the
fundamental basis for computational models of
understanding, planning, and problem solving. In
this field of application, schemas usually serve as
structured beliefs which involve causal-predictive cycles
of action and perception. Such models have been used
to interpret basic speech acts and linguistic expressions in an agent’s physical environment in terms of
grounded schemas (Roy 2005). Another application
of schema-based learning is generalization by means
of knowledge chunking (DeJong and Mooney 1986).
The idea of generalizing with schemas was closely
related to the development of explanation-based systems, which are supposed to be capable of learning new
schemas which can be used to generalize examples
(O’Rorke 1984). This form of generalization is possible
because schemas themselves are generalized knowledge
chunks. However, up to now, only very few of these
artificial systems are capable of generating their own
schemas.

Cross-References
▶ Ausubel, David P.
▶ Cognitive Load Theory
▶ Early Maladaptive Schemas: The Moderating Effects
of Optimism
▶ Generative Learning
▶ Schema Development
▶ Schema(s)
▶ Schema-Based Architectures of Machine Learning

References
Ausubel, D. P. (1968). Educational psychology: A cognitive view.
New York: Holt, Rinehart, & Winston.
Bransford, J. D., Stein, B. S., Shelton, T. S., & Owings, R. A. (1981).
Cognition and adaptation: The importance of learning to learn.
In J. H. Harvey (Ed.), Cognition, social behavior and the environment. (pp. 93–110). Hillsdale, NJ: Erlbaum.

S

2949

Bransford, J. D. (1984). Schema activation and schema acquisition: Comments on Richard C. Anderson’s remarks. In
R. C. Anderson, J. Osborn, & R. J. Tierney (Eds.), Learning to
read in American schools: Basal readers and content texts.
Hillsdale: Lawrence Erlbaum.
Bruner, J. (1966). Toward a theory of instruction. Cambridge, MA:
Harvard University Press.
Corbacho, F. J. (1998). Schema-based learning. Artificial Intelligence,
101(1–2), 337–339.
DeJong, G., & Mooney, R. (1986). Explanation-based Learning: An
alternative view. Machine Learning, 1, 145–186.
O’Rorke, P. (1984). Generalization for explanation-based schema
acquisition. In Proceedings of the national conference on artificial
intelligence (pp. 260–263). Austin: Morgan-Kaufmann.
Pascual-Leone, J., & Goodman, N. (1979). Intelligence and experience: A neo-Piagetian approach. Instructional Science, 8,
314–367.
Roy, D. (2005). Semiotic schemas: A framework for grounding
language in action and perception. Artificial Intelligence, 167,
170–205.
Rumelhart, D. E., & Norman, D. A. (1978). Accretion, tuning, and
restructuring: Three models of learning. In J. U. Cotton & R. L.
Klatzky (Eds.), Semantic facts in cognition (pp. 37–54). Hillsdale:
Erlbaum.
Sweller, J. (1988). Cognitive load during problem solving: Effects on
learning. Cognitive Science, 12, 257–285.
Wittrock, M. C. (1991). Generative teaching of comprehension. The
Elementary School Journal, 92(2), 169–184.

Schema-Based Problem Solving
SANDRA P. MARSHALL
Department of Psychology, San Diego State University,
San Diego, CA, USA

S
Synonyms
Frame; Plan; Scaffold; Script

Definition
Schema-based problem solving requires use of specific
schemas or templates to recognize, understand, and
solve problems. Broadly, such schemas have four characteristics: they are organized so that an individual can
quickly identify new instances that are similar to those
on which the schema was founded; they are general
templates but also have links to specific individual
experiences that match the current template; they
guide an individual’s efforts to draw inferences, make

2950

S

Schema-Based Problem Solving

estimates, and create plans to solve problems; and they
connect with essential skills or procedures the individual already has.

Theoretical Background
Problem-solving schemas have most often appeared in
arithmetic or algebraic contexts. Hinsley et al. (1977)
were among the first researchers to look seriously at the
way schemas are used by problem solvers. They focused
on the question of whether individuals actually use
schema knowledge to solve problems. If so, they reasoned that the following claims must be true: individuals can recognize problem types based on their
schema knowledge, individuals can identify problem
types without first developing equations to represent
the problems, and individuals can use their problem
identifications to formulate plans for solving the problems (p. 92). These claims have guided much of the
research about how schemas are used in problem
solving.

Important Scientific Research
and Open Questions
Research has focused on the underlying relationships
within problem-solving schemas and on the types of
knowledge that comprise a schema. A problem-solving
schema should capture essential information about the
situation represented by the problem and provide
a scaffold for learning how to recognize it. Thus,
a first research question has to do with the nature of
problem situations. Riley et al. (1983) focused on the
structure of arithmetic story problems such as those
found in elementary school classrooms. They discovered that three relationships – Change, Combine, and
Compare – capture most of the problems that require
addition and subtraction. The importance of their
research is that they shifted the focus away from the
equations or algorithms needed to solve a problem to
the underlying situation represented by the problem.
Marshall (1995) later expanded the original three situations to a set of five to cover story problems that
appear in pre-algebra classrooms.
The research on story problems made it clear
that situational information is a key constituent of a
problem-solving schema. However, problem solving is
more than problem recognition, and a schema is more
than a situation description. Active solving is required,

and a schema should capture both the situational features and the solution strategies of problem solving.
Building on Hinsley et al.’s earlier approach, Marshall (1995) hypothesized that four types of knowledge
were essential for schema-based problem solving: identification, elaboration, planning, and execution. To
solve problems using a schema, an individual must
first recognize the features of the problem that are
relevant to the schema (i.e., identification knowledge).
Next, the solver needs to figure out how these features
relate to past experiences of problem solving (i.e., elaboration knowledge). These two types of information are
critical but by themselves insufficient to characterize the
schema. The solver must also draw upon two additional
types of schema knowledge. Specifically, the solver must
formulate plans about how to reach a solution (i.e.,
planning knowledge) and then must call upon the skills
required to carry out the plans (i.e., execution knowledge). A well-formed schema will have these four types
of knowledge linked in memory so that retrieval of one
component automatically activates the others.
Understanding the types of knowledge involved in
schema-based problem solving has diagnostic value.
Often, the nature of a problem-solving error indicates
that the root of the difficulty lies in a faulty schema. An
individual can fail to solve a problem for many reasons.
The failure may be traced to the attempt to use an
incorrect schema. Or, there may be evidence that the
appropriate schema was activated but some knowledge
component was faulty. In either case, schema-based
instruction may remedy the failure.

Cross-References
▶ Problem Solving
▶ Problem-Based Learning
▶ Schema(s)
▶ Schema-Based Instruction

References
Hinsley, D., Hayes, J., & Simon, H. (1977). From words to equations: Meaning and representation in algebra word problems.
In M. Just & P. Carpenter (Eds.), Cognitive processes in comprehension (pp. 89–106). Hillsdale, NJ: LEA.
Marshall, S. (1995). Schemas in problem solving. New York: Cambridge University Press.
Riley, M., Greeno, J., & Heller, J. (1983). Development of children’s
problem-solving ability in arithmetic. In H. Ginsberg (Ed.), The
development of mathematical thinking (pp. 153–196). New York:
Academic Press.

Schema-Based Reasoning

Schema-Based Reasoning
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany

Synonyms
Case-based reasoning; Mental logic

Definition
Schema-based reasoning is a generalized form of
case-based reasoning widely used in cognitive science
and artificial intelligence. In both fields of application, the basic assumption is that generalized knowledge is stored in schemas that contain the records of
single cases of past successful problem solving. These
explicitly represented schemas can be combined with
a flexible reasoning mechanism, resulting in an adaptive and context-sensitive approach to real-world problem solving. Schema-based reasoning expands on the
idea of case-based reasoning by referring to generalized
cases (schemas) rather than specific cases. It thus relies on
the effective use of generic contextual knowledge in order
to be transferred to a current problem (Turner 1994).

Theoretical Background
In the 1980s, psychologists and cognitive scientists used
previous studies on syllogistic reasoning to develop
various theoretical approaches to deductive reasoning.
These approaches may be divided roughly into two
main classes, the first being syntactic approaches,
which assume that people apply syntactic rules which
constitute a mental logic when practicing deductive
reasoning: “In evaluating a deductive argument, people
construct a mental proof or derivation, doing informally what logic students are taught to do formally”
(Rips 1984, p. 123). In contrast, semantic or pragmatic
approaches are based on the use of so-called judgment
schemas. Semantic approaches include different theories, such as the theory of pragmatic reasoning schemas
(Cheng and Holyoak 1985), the two-factor theory by
Evans (1982), the theory of mental logic by Braine
(1990), and the theory of mental models by JohnsonLaird (1983).
Cheng and Holyoak (1985) use the concept of pragmatic reasoning schemas, which may be applied to

S

specific types of situations (e.g., promise, obligation,
permission) to explain the influence the content of the
premises has on the conclusion. Cheng and Holyoak
argue that people possess classes of linguistically based
permission schemas that have a fixed internal structure
which is determined by pragmatic considerations.
The effectiveness of permission schemas (“If you
want to do P, then you must do Q”) could be measured
in selection tasks, where the success rate has been much
superior to what is usually observed in logical tasks.
Cheng and Holyoak define permission schemas by
a set of production rules that give the same answers
to problems of conditional inference as those of
formal logic.
In sum, pragmatic reasoning schemas provide
a convincing explanation of why we generally find it
easy to deduce from the pair of propositions – “The
father promised the boy five dollars if he would
mow the lawn” and “The boy did not receive the five
dollars” – that the boy did not mow the lawn. However,
this approach is not capable of explaining why it is
more difficult to solve problems which have the following logical form: “A rule specifies that a 7 is on the card
if an A is on it, and there is no 7 on the card. Is there an
A on the card?”
The two-factor theory by Evans (1982) is designed to
make it understandable why individuals often have a
tendency to choose inconsistent answers. When examined more closely, this theory proves to be a conflict
model “in which logical and non-logical influences
compete for control of the subjects’ responses” (Evans
1982, p. 125). The logical component establishes the
extent to which a person’s answers are directed toward
the logical structure of the problem. The logical processes a person adopts are conceived of as interpretational processes which may run away and lead to
logically incorrect answers. Evans classifies all influences on answers which come about as a result of
factors which are logically irrelevant as nonlogical components. Among other things, he emphasizes people’s
tendency to accept or choose conclusions which are
consistent with their previous knowledge and preconceptions rather than conclusions which may be logically correct but are not consistent with their previous
knowledge (belief bias).
The theory of mental logic is closely related to Rips’
(1984) system of natural reasoning and includes three
main elements: (1) a set of judgment schemas,

2951

S

2952

S

Schema-Based Reasoning

(2) a program of reasoning which selects the schemas to
be used in a line of argumentation, and (3) a set of
independently motivated pragmatic principles which
influence the interpretation of external attributes and
invite one to make nonlogical inferences. A judgment or
inferential schema defines a special type of inference by
making explicit reference to the conclusion which can
be inferred from premises with a certain form. The
second component of the theory, the program of reasoning, creates a model of how people select a schema to
apply at a certain point in the chain of argumentation.
It includes a direct, standard procedure of reasoning
and a set of strategies which may be applied if the direct
procedure fails. The direct procedure is a simple program. It uses a schema which can be applied to a set of
premises (with conditions which prevent continuous
loops) and adds the inferences to the set (i.e., it either
reuses or evaluates the conclusion). Proponents of this
theory claim that this program is available to all
humans and that it is used routinely in the processing
of texts and language as well as in solving logical problems. The strategic component of the program of reasoning only comes into use if the direct procedure fails,
which, however, is often the case in deductive reasoning. According to the theory of mental logic, propositional logic problems can be solved on the basis of
the direct reasoning routine. However, deductive inferences may require skills which are beyond the scope of
this form of mental logic and are thus relatively difficult. The theoretical approach of mental models has
been considered as a promising procedure of reasoning
for such cases (cf. Johnson-Laird 1983), but this
approach does not belong to the area of schemabased reasoning.

Important Scientific Research and
Open Questions
The various approaches of schema-based reasoning
have been contrasted with the theory of mental
models (Johnson-Laird 1983) and its application to
deductive reasoning for several decades, resulting in
an interesting debate. Proponents of schema-based
approaches, such as Braine and colleagues, conclude
from their studies that the theory of mental logic is
superior to the theory of mental models, whereas proponents of the theory of mental models come to the
conclusion that their theory is superior to the
approach of mental logic. This is not the place to

justify the two approaches and the results of studies
on them (for an overview, see Braine and O’Brien
1998). However, it is clear that these theories have
contrasting strengths and weaknesses. Remaining
within the realm of the theory of mental logic, the
idea of inferential schemas provides a direct method
for making inferences related to real-world problems.
A person can perform deductive reasoning either by
applying the rules of a system of mental logic or by
going through the various means of interpreting the
premises of a proposition. Of prime significance in
this process is how the premises are represented in the
mind. It is on the basis of this mental representation
that one can determine how difficult various deductive problems are. Researchers are not yet in agreement as to how much of problem solving is performed
through logical-rational processes and to what extent
it is influenced by pragmatic aspects which seem irrelevant from the perspective of formal logic and could
even lead to systematic, nonlogical distortions. All in
all, it must be stated that deductive reasoning has not
yet been studied sufficiently from a psychological
point of view, seen by the fact that it is still unclear
which cognitive processes are involved in deductive
thinking.
Another area of research on the implementation
of schema-based reasoning is the field of artificial
intelligence (AI). Actually, the integration of various
reasoning modalities into intelligent systems and
autonomous agent architectures is one of the most
interesting and challenging aspects of current AI. The
exploitation of the synergy between different reasoning
methods is a major feature of multi-modal reasoning
systems, which integrate and combine different reasoning modalities and different knowledge representation
formalisms. Case-based reasoning, and in consequence
schema-based reasoning, is often considered a fundamental modality in several multi-modal reasoning systems. The basic idea of schema-based reasoning in the
area of AI is simple and corresponds to a large extent
to the approaches of cognitive science as described
above: The relevant knowledge of an agent must be
represented explicitly as declarative knowledge structures called schemas, and these schemas can then be
activated and used to guide the agent’s problem solving. These contextual schemas represent not only the
task-relevant knowledge but also commitments to perform future actions. They therefore allow the agent

Schema-Dependent Neocortical Connectivity During Information Processing

to go beyond purely reactive decision making and
proceed to adaptive and context-sensitive solutions to
problems. Contextual schemas represent patterns
appearing in the world as well as ones that have been
experienced in former problem solving. This means
that schema-based reasoning is sensitive to changes in
the agent’s environment but enables the agent to automatically adapt to situations. At the moment, the relevant contextual schemas are still given to the agent by
other people with experience and expertise. However,
in the future it should be possible for the agents themselves to create and modify these schemas based on
their own experience and evolving expertise. This is
a promising field of current AI research (see, e.g.,
Portinale et al. 2004).

Cross-References
▶ Deductive Reasoning
▶ Default Reasoning
▶ Inductive Learning and Reasoning
▶ Inferential Learning
▶ Logical Reasoning and Learning
▶ Model-Based Reasoning
▶ Pragmatic Reasoning Schema(s)

References
Braine, M. D. (1990). The natural logic approach to reasoning. In
W. Overton (Ed.), Reasoning, necessity, and logic: Developmental
perspectives (pp. 133–157). Hillsdale, NJ: Lawrence Erlbaum
Assoc.
Braine, M. D. S., & O’Brien, D. P. (Eds.). (1998). Mental logic.
Mahwah, NJ: Lawrence Erlbaum Associates.
Cheng, P. W., & Holyoak, K. J. (1985). Pragmatic reasoning schemas.
Cognitive Psychology, 17, 391–426.
Evans, J. S. B. T. (1982). The psychology of deductive reasoning.
London: Routledge.
Johnson-Laird, P. N. (1983). Mental models. Towards a cognitive
science of language, inference, and consciousness. Cambridge:
Cambridge University Press.
Oaksford, M., & Chater, N. (1996). Rational explanation of the
selection task. Psychological Review, 103(2), 381–391.
Portinale, L., Magro, D., & Torasso, P. (2004). Multi-modal
diagnosis combining case-based and model-based reasoning:
A formal and experimental analysis. Artificial Intelligence,
158(2), 109–153.
Rips, L. J. (1984). Reasoning as a central ability. In R. J. Sternberg
(Ed.), Advances in the psychology of human intelligence (Vol. 2,
pp. 105–147). Hillsdale, NJ: Lawrence Erlbaum Associates.
Turner, R. M. (1994). Adaptive reasoning for real-world problems:
A schema-based approach. Hillsdale, NJ: Lawrence Erlbaum
Associates.

S

2953

Schema-Centered Learning
▶ Schema-Based Learning

Schema-Dependent Memory
Consolidation
▶ Schema-Dependent Neocortical Connectivity During Information Processing

Schema-Dependent Neocortical
Connectivity During
Information Processing
MARLIEKE T. R. VAN KESTEREN1,2, MARK RIJPKEMA1,
DIRK J. RUITER2, GUILLÉN FERNÁNDEZ1,3
1
Donders Institute for Brain, Cognition, and
Behaviour, Radboud University Nijmegen, Nijmegen,
The Netherlands
2
Department of Anatomy, Radboud University
Nijmegen Medical Centre, Nijmegen, The Netherlands
3
Department of Cognitive Neuroscience, Radboud
University Nijmegen Medical Centre, Nijmegen,
The Netherlands

Synonyms
Preexisting Knowledge; Schema-Dependent Memory
Consolidation; Schemas in the Brain

Definition
A schema, in the form of preexisting knowledge
acquired throughout life, generating expectations and
prospects toward external input, is known to have
a positive influence on remembering newly learned
information related to this schema. Schemas have
been widely studied in cognitive psychology (Johnson-Laird 1983), but theories as to how this memory
enhancement arises are still hard to relate directly to
underlying neural processes. Recently, however, neuroscientific studies on schemas, both in rodents and
humans, started to shed some light on schema-related
mnemonic mechanisms in the brain.

S

2954

S

Schema-Dependent Neocortical Connectivity During Information Processing

Theoretical Background
The brain is an interactive and plastic organ. It does not
just passively await incoming – bottom-up – information passing through the different senses to be
processed to initiate an adequate response. Contrarily,
the brain is continuously anticipating and predicting –
top-down – how its environment and internal states are
possibly going to evolve, so adequate responses are
readily available when necessary (Friston 2005). Consequently, the brain is able to function in an optimal
manner, spending as little energy as possible on
processing the unimportant parts of the continuous
stream of incoming information, but readily picking
out important features. To continuously influence this
information processing, the brain uses internal topdown signals to actively evaluate novel information
and mediate how it is best processed and assimilated
into its existing schemas.
A schema consists of all sorts of representations
about the world, interconnected in an intricate network
constituted to form a realistic idea of what the world is
like. Using these schemas, the brain can predict future
events in the world, leading to differential processing of
schema-related or unrelated information. When information is consistent with a schema, it will be passed
through a fast, efficient processing mode. On the other
hand, information that is inconsistent will yield
a prediction error (Friston 2005) and will consequently
be more elaborately processed in order to fit best
preexisting schemas or to build new schemas. Moreover, inconsistent information that is encountered
repeatedly might in the end lead to schema adjustment,
influencing information processing in the long run. In
this view, the strength of the prediction error as
a reaction to novel information thus influences how
the brain will process this information.
By manipulating a prior schema, and testing information processing during the different stages that
a new mnemonic trace undergoes, we can gain insights
into not only how a schema mediates information
processing, but also which different mnemonic mechanisms are affected. Remembering information over
time is a complex process consisting of several interrelated subprocesses. Learned information will first
reside in short-term working memory, which helps to
encode the information into long-term memory. To best
merge the new information with the prior information
stored in the brain, the information is assimilated and

consolidated into the schema over a longer period of
time (up to years). Stored information can then be
retrieved and might even be reconsolidated after its
retrieval. All these mnemonic processes have individual
characteristics, but they also show many similarities.
For example, encoding of information is found to
recruit many of the same brain regions as later retrieval
of the same information does. All these mnemonic
processes and associated brain activity may be equally
or differentially affected by schemas. Research into
schemas in the brain can thus help improving our ability
to tease these different mnemonic processes apart.
During systems consolidation of a memory trace, the
hippocampus is found to initially bind different parts of
a memory trace, stored throughout the neocortex in
a distributed manner. With time, a memory becomes
less hippocampally dependent because it is better consolidated into neocortical modules. The hippocampus
thus becomes less necessary to retrieve this specific
memory over time. Concomitantly, the medial prefrontal cortex (mPFC) increasingly shows more of the
characteristics that were initially dedicated to the hippocampus, activating and binding different parts of
a memory trace. Consequently, it has been proposed
that hippocampal and mPFC influence related to
a certain memory trace is balanced, and that this balance shifts from the hippocampus toward the mPFC
with time (Takashima et al. 2006). This switch in balance, as a feature of systems consolidation, can presumably be used to probe the effect of a schema on the
consolidation process.

Important Scientific Research and
Open Questions
Up to now, only few studies have examined influences of
a schema on mnemonic processing in the brain. In
rodents, it has been discovered that when a prior associative schema was present in the brain, novel memory
traces related to this schema became hippocampally
independent faster (Tse et al. 2007). This was tested by
overtraining the rodents on a spatial event arena, in
which they could find differently flavored food at different locations. Subsequently, the rodents learned novel
information that was either related or unrelated to the
spatial properties of the event arena. After 48 h, their
hippocampus was removed. When testing on memory
afterwards, the information that was related to the event
area turned out to be remembered, while unrelated

Schema-Dependent Neocortical Connectivity During Information Processing

information was forgotten. The novel information that
was related to the spatial schema was thus faster consolidated and lost its hippocampal dependence already
within 48 h, while unrelated information at that point
in time was still hippocampally dependent.
In humans, the schema-effect on memory has been
investigated using functional Magnetic Resonance
Imaging (fMRI) both during encoding, post-encoding
rest and retrieval of memories that were either related
or unrelated to an existing schema. During encoding,
and persisting during post-encoding rest, participants
with a coherent schema about the storyline of a movie
showed a decreased interaction between the hippocampus and the mPFC while watching the last part of this
movie and during a rest period thereafter, compared to
participants that did not have a coherent schema (van
Kesteren et al. 2010a). The participants with the schema
thus needed less interaction between these regions compared to the participants without a schema to integrate
adequately the novel information, presumably because
it was better related to their schema. During retrieval,
mPFC activity and connectivity with somatosensory
cortex, related to the tactile part of a memory trace,
was found to be enhanced for visuo–tactile information
congruent with common knowledge versus information
incongruent with common knowledge (van Kesteren
et al. 2010b). In this experiment, congruency with
common knowledge resulted in more mPFC involvement during successful retrieval of the memory trace.
These findings credibly fit into the model discussed
above, where the balance between the hippocampus
and the mPFC shifts with systems consolidation, facilitated by a prior schema. Additionally, these data can
account for the view that the influence of a schema on
processes related to the assimilation of new information can quite reliably be investigated by looking at this
hippocampal-mPFC shift.
The question which specific mnemonic processes
are affected by prior schema, however, is still partly
unanswered. The interaction between the hippocampus and the mPFC has previously merely been attributed to consolidation processes. However, recent
insights have revealed that the mPFC is already active
during encoding of new information, contradicting
this hypothesis. Therefore, we propose that this process, in which the hippocampal function is gradually
transferred to the mPFC, is already active during
encoding of novel information and is thus already

S

facilitated by a prior schema in very early stages of
information processing. Whether other processes specific to encoding are also facilitated by prior schema,
and whether these are crucial for facilitated processing
later on, remains to be investigated. The same holds for
retrieval and reconsolidation processes. As explained
above, these issues can be investigated by looking at
connectivity between the hippocampus, the mPFC
and other neocortical regions representing parts of
the (consolidated) memory trace.
Additionally, more research on how a schema influences long-term memory can be beneficial for educational strategies such as developing curricula and
detecting why students have trouble comprehending
certain issues. Understanding more about how the
brain forms schemas and how it processes information
either related or unrelated to these schemas will help
understanding which learning methods work and why
and how they work. By investigating more specific problems that are related to real-life classroom situations,
cognitive neuroscience might thus help educational sciences to gain more insight into how students can acquire
new knowledge most optimally in classroom settings,
from books, and through multisensory and multimedia
learning. Therefore, educational sciences and cognitive
neuroscience must work closer together in order to best
detect and investigate these current issues.
In summary, research into the effect of a prior
schema on mnemonic processes in the brain is thus far
showing promising results. By investigating the hippocampal-mPFC balance, but also by looking at connectivity traces from these brain areas to areas representing
specific features of memory traces, it appears that a prior
schema facilitates information processing throughout
different mnemonic stages, such as encoding, consolidation, and subsequent retrieval. Nevertheless, there are
questions that remain open. Next to the possible
differentiating effects of mnemonic processes such as
encoding, consolidation, and retrieval, schema formation periods can be a unique window when investigating schema formation and its effect on future
learning. Furthermore, the question why a prior schema
facilitates information processing needs further elaboration by formulating theories and building computational models that mimic this process and are able
to predict consequences to be tested in further empirical research, both in animals and in humans. Finally,
these fundamental insights into learning and memory

2955

S

2956

S

Schema-Like Representations of Categories

processes in the brain can be used for better build-up of
educational programs and an enhancement of educational strategies (van Kesteren et al. 2010b).

Schema-Oriented Learning
▶ Schema-Based Learning

Cross-References
▶ Anticipatory Schema(s)
▶ Categorical Learning
▶ Categorical Representation
▶ Memory Consolidation and Reconsolidation
▶ Representation, Presentation, and Conceptual
Schemas
▶ Retrieval Cues and Learning
▶ Schema(s)
▶ Schema-Based Learning
▶ Schema-Based Reasoning
▶ Schematic Influences on Category Learning and
Recognition Memory

References
Friston, K. (2005). A theory of cortical responses. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences,
360(1456), 815–836.
Johnson-Laird, P. N. (1983). Mental models: Towards a cognitive science of language, inference, and consciousness. Cambridge, MA:
Harvard University Press.
Takashima, A., Petersson, K. M., Rutters, F., Tendolkar, I., Jensen, O.,
Zwarts, M. J., et al. (2006). Declarative memory consolidation in
humans: A prospective functional magnetic resonance imaging
study. Proceedings of the National Academy of Sciences of the
United States of America, 103(3), 756–761.
Tse, D., Langston, R. F., Kakeyama, M., Bethus, I., Spooner, P. A.,
Wood, E. R., et al. (2007). Schemas and memory consolidation.
Science, 316(5821), 76–82.
van Kesteren, M. T., Fernandez, G., Norris, D. G., & Hermans, E. J.
(2010a). Persistent schema-dependent hippocampal-neocortical
connectivity during memory encoding and postencoding rest in
humans. Proceedings of the National Academy of Sciences of the
United States of America, 107(16), 7550–7555.
van Kesteren, M. T., Rijpkema, M., Ruiter, D. J., & Fernandez, G.
(2010b). Retrieval of associative information congruent with
prior knowledge is related to increased medial prefrontal
activity and connectivity. The Journal of Neuroscience, 30(47),
15888–15894.

Schemas and Decision Making
SANDRA P. MARSHALL1, NORBERT M. SEEL2
1
Department of Psychology, San Diego State
University, San Diego, CA, USA
2
Department of Educational Science, University of
Freiburg, Freiburg, Germany

Synonyms
Naturalistic decision making; Schema-based decision
making

Definition
Decision-making schemas coincide widely with problem
solving. Both have the same general characteristics and
both serve as mental templates against which a decision
maker or problem solver maps a specific situation.
Both also draw heavily on prior experiences, using
them to shape the current interpretation of the decision
or problem. However, schema-based decision making
is distinct from schema-based problem solving. The
key issue in problem solving is recognizing the underlying nature of the problem so that a known solution
strategy may be applied if one exists, whereas, in decision making, the problem is usually known, but the
solution strategy is not. The key issue here is how to
select one strategy from many available alternative
strategies. From a schema perspective, both decision
making and problem solving utilize the same types of
long-term knowledge and cognitive processes, but the
relative importance of the basic schema knowledge
components varies.

Theoretical Background

Schema-Like Representations
of Categories
▶ Schematic Influences on Category Learning and
Recognition Memory

Research about schemas and decision making has been
predominantly focused on military settings. It has been
closely related to the study of situation awareness and
its importance in tactical decision making (Endsley
2000). Situation awareness refers to how well an individual perceives and comprehends the elements in

Schemas and Decision Making

a given environment which is constantly changing so
that a decision maker must constantly adapt and adjust
the view of the situation to accommodate these changes
(Endsley 1995). Actually, military operators and decision makers require situation awareness to understand
the goals and decision tasks for a specific set of circumstances (Craig 2001; Klein et al. 1986). Individuals with
good situation awareness are believed to make better
and more efficient decisions. Thus, situation awareness
is analogous to problem recognition in schema terms
and is an essential part of a decision-making schema. It
spans two schema knowledge components: identification knowledge and elaboration knowledge (see the
entry on schema-based problem solving).
Schema research grew out of the need to better
understand situation awareness, mental models, and
critical thinking in relation to individuals working in
stressful situations. In particular, research about
schemas in decision making has focused on the identification and elaboration components of schema
knowledge in tactical settings, attempting to further
understand how these two types of knowledge intersect. Based on this research, the Recognition Primed
Decision (RPD) model has been developed in the field
of natural decision making (Klein 1989). The starting
point of this model is an action of the decision maker
that is based on both the perception of a known or
prototypical situation and a serial strategy of the evaluation of behavioral options. That is, a person conceives a situation in terms of familiarity with past
experiences. The rating of familiarity in relation to
a set of known cases provided the person with the
identification of achievable goals, relevant details,
expectations, and plausible actions of decision making. Based on an activated schema the decision maker
can refer to experiences in order to generate a possible
option of action. The rating of this option occurs
by means of a mental simulation (thought experiment) with the aim to check if there are some risks
for realizing the option. The RPD model of schemabased decision making can be summarized as depicted
in Fig. 1.
The RPD model contains aspects of problem solving and judgment formation in accordance with processes of natural decision making. Moreover, in its
emphasis on mental simulations of behavioral options,
it also corresponds with the theory of mental models
(Kuipers 1994). Additionally, the RPD model includes

S

2957

the concept of situation awareness as it plays an important role in the fields of military or pilot training since
the 1980s.

Important Scientific Research and
Open Questions
Research on schemas and decision making is relatively
sparse because this is a relatively new field of interest. It
is an area of some importance in applied areas such as
military decision making because these decisions are
made in real-world settings and have significant consequences. Two ongoing areas of research are the differences between different types of settings (such as
operational versus tactical) and the relative importance
of the four knowledge components (i.e., identification,
elaboration, planning, and execution).
Schema-based decision making is closely related
with the field of naturalistic decision making which
investigates how people make decisions under risk in
everyday situations. Decision making under risk is in
general characterized by dynamically changing conditions, the challenge to respond immediately to these
changes, ill-defined tasks, time pressure, and farreaching personal consequences in case of mistakes.
Since the 1980s, researchers from different fields have
developed theoretical models to explain how experienced and inexperienced decision makers make decisions under risk. Some of these models are related to
the expected utility theory (EUT) which aims at
weighting evidence and choosing an optimal action.
Kahneman and Tversky (1979) have developed the
prospect theory as a variant of EUT by arguing that
a decision maker weights achievable results with higher
scores than results which are only probable. Furthermore, decision makers show a tendency of preferring
strongly weighted low probabilities as well as they are
more sensitive with regard to uncertainty than to risk.
Although the relevance of the prospect theory for decision making in general is widely accepted in the literature, its applicability for decision making under risk
has been questioned because it does not take into
account important factors involved in natural decision
in complex situations. Klein and Calderwood (1991)
argue that analytical methods of decision making
under risk eventually fail because they need too much
time and lack the flexibility to respond to rapidly
changing conditions of situations. In accordance with
the idea of schema-based decision making, it can be

S

2958

S

Schemas and Decision Making

Experience the Situation in a Changing Context

Reassess
Situation

No
Is the situation
familiar?

Seek More
Information

Yes

(Recognition
of match to
prototype)

Activation of information from memory
Yes
Are expectancies
violated?

Plausible
Goals

Relevant
Cues

Expectancies

Actions
1...n

No

Mental Simulation of Action (n)

Yes, but
Modify

Will it work?

No

Yes
Implement

Schemas and Decision Making. Fig. 1 The complex RPD strategy (Adopted from Klein and Calderwood 1991)

argued that the activation of a specific schema brings
about enormous time advantages in mastering challenging situations if they are similar and belong to the
same category (Falzer 2004; Marshall 1995). However,
in the case of novel phenomena and problems, the
available schemas are eventually inappropriate and
must be replaced by mental models. Indeed, the theoretical approach of mental models emphasizes cognitive processes of generating plausibility and of
probabilistic reasoning (Gigerenzer et al. 1991) that
are involved in decision making under risk. Therefore,
natural decision making on the basis of mental models
can be considered as a theoretically sound alternative to
schema-based decision making.

Cross-References
▶ Default Reasoning
▶ Model-Based Reasoning

▶ Schema(s)
▶ Schema-Based Problem Solving
▶ Schema-Based Reasoning

References
Craig, P. A. (2001). Situational awareness: Controlling pilot error.
New York: McGraw-Hill.
Endsley, M. R. (1995). Toward a theory of situation awareness in
dynamic systems. Human Factors, 37(1), 32–64.
Endsley, M. (2000). Theoretical underpinnings of situation awareness: A critical review. In M. Endsley & D. Garland (Eds.),
Situation awareness analysis and measurement (pp. 3–28).
Mahwah: Lawrence Erlbaum.
Falzer, P. R. (2004). Cognitive schema and naturalistic decision making in evidence-based practices. Journal of Biomedical Informatics, 37(2), 86–98.
Gigerenzer, G., Hoffrage, U., & Kleinbölting, H. (1991). Probabilistic
mental models: A Brunswikian theory of confidence. Psychological Review, 98(4), 506–528.

Schematic Influences on Category Learning and Recognition Memory
Kahneman, D., & Tversky, A. (1979). Prospect theory: An analysis of
decisions under risk. Econometrica, 47, 263–291.
Klein, G. (1989). Recognition-primed decisions. Advances in ManMachine Research, 5, 47–92.
Klein, G. (1998). Sources of power: how people make decisions.
Cambridge: MIT Press.
Klein, G., & Calderwood, R. (1991). Decision models: Some lessons
from the field. IEEE Transactions on Systems, Man, and Cybernetics, 21(5), 1018–1026.
Kuipers, B. (1994). Qualitative reasoning: Modeling and simulation
with incomplete knowledge. Cambridge, MA: MIT Press.
Marshall, S. (1995). Schemas in problem solving. New York:
Cambridge University Press.

Schemas in the Brain
▶ Schema-Dependent Neocortical Connectivity During
Information Processing

Schematherapy
▶ Schema Therapy

S

2959

Definition
The title, schematic influences on category learning
and recognition memory, means that our mental representations of categories, such as dogs and friends, are
schema-like. A ▶ schema is a knowledge structure that
provides a set of expectations based on past experience.
For example, a person may have a schema for libraries.
When this schema is activated, it makes properties
like books and desks available. ▶ Category learning is
the process of forming representations of categories.
▶ Classification learning is one type of category learning, in which the learner acquires information about
categories through classifying examples. For example,
a child may classify an animal as a bird using its properties, such as small and flying. The child may learn that
the animal is indeed a bird, strengthening the hypothesis about which properties predict the bird category.
Alternatively, the child may learn that the animal is
a mammal, leading to a modification of the hypothesis.
One way to study how learners represent categories is
by measuring their recognition memory for information about the categories and their members. Interestingly, recognition results in category learning studies
parallel those in schema studies, suggesting that people
develop schema-like representations of categories.

Theoretical Background

Schematic Bases of Category
Learning
▶ Schematic Influences on Category Learning and
Recognition Memory

Schematic Influences on
Category Learning and
Recognition Memory
YASUAKI SAKAMOTO
Howe School of Technology Management, Stevens
Institute of Technology, Hoboken, NJ, USA

Synonyms
Schema-like representations of categories; Schematic
bases of category learning

People use categories to generalize from past experience to the current situation that they may not have
encountered previously. For example, once people categorize a never-before-seen animal as a dog, they can
infer that the animal may bark and bite. This ability to
use categories is critical because it allows people to
process information efficiently and prepare their
actions in time. Categories guide the encoding of new
information and retrieval of stored information that is
relevant to the present task.
Schemas serve similar functions. For example, a
library schema, when activated, provides people with
a certain set of expectations such as being quiet and
borrowing books. Like categories, schemas guide the
encoding and retrieval of information, and inform
people how to act.
The two literatures address a related set of issues. In
particular, both literatures are concerned with how
people process information in relation to existing
knowledge structures, such as categories and schemas.
Category learning researchers are also interested in how

S

2960

S

Schematic Influences on Category Learning and Recognition Memory

people acquire and represent categories. Consequently,
measures of recognition memory for studied information figure prominently in both category learning
research and schema research.
Sakamoto and Love (2004) drew parallels between
these two literatures and imported ideas from work in
schemas to the category learning literature. Memory
for schema-consistent and schema-inconsistent information is well studied in schema research. For example,
whereas seeing desks in a library is schema-consistent,
seeing beds is schema-inconsistent. A meta-analysis
by Rojahn and Pettigrew (1992) revealed that when
measures of recognition memory were corrected for
false alarm rate, there was a recognition advantage
for schema-inconsistent information over schemaconsistent information. For a library schema, a common false alarm might be reporting to have seen a desk
when in fact a desk did not appear in any studied scene.
In category learning research, Palmeri and
Nosofsky (1995) found a recognition advantage for
inconsistent information over consistent information.
In their studies, subjects learned to classify 16 geometric stimuli, one by one, into one of two mutually
exclusive categories through numerous trials with corrective feedback. Most stimuli could be classified by
a simple rule: For example, the rule might be “the
large items are in category A, and the small items are
in category B.” These rule-following stimuli are analogous to schema-consistent information. However,
two stimuli were inconsistent with the rule: There
was a large item belonging to category B, and a small
item belonging to category A. These rule-violating
items are analogous to schema-inconsistent information. Following learning, subjects completed a recognition memory test consisting of studied items and
foil items. The main finding was that the subjects
recognized the two rule-violating items better than
the rule-following items; they selectively attended to
the rule during learning, and the two rule-violating
items stood out.
Like in the schema research, there is a recognition advantage for items that deviate from a salient
regularity in the category learning research. The
schema literature further suggests that the strength of
the regularity, measured by the proportion of items
that conform to the regularity, modulates the advantage for deviant items. For example, Rojahn and
Pettigrew (1992) found in their meta-analysis that the

memory advantage for the schema-inconsistent items
became greater as the proportion of the schemainconsistent items became smaller.
Similarly, Sakamoto and Love (2004) found in
a category learning experiment that the memory
advantage for the rule-violating item was greater
when the violated rule was stronger. In Sakamoto and
Love, as in Palmeri and Nosofsky (1995), most members of categories A and B followed an imperfect rule.
The strength of the rule was manipulated by varying
the frequency of rule-following items: Whereas category A contained eight rule-following items, category
B contained four. Each category contained a ruleviolating item. To master the classification task, subjects needed to differentiate category B’s rule-violating
item from the eight rule-following items in category A,
and category A’s rule-violating item from the four rulefollowing items in category B. After learning, these
rule-violating items were remembered better than the
rule-following items, replicating Palmeri and Nosofsky.
Furthermore, as in the schema research, memory for
the category B’s rule-violating item, which violated the
more frequent rule of category A, was enhanced relative
to that for the category A’s rule-violating item.
The parallel results from the schema and category
learning research indicate that the learners develop
schema-like representations of categories, and challenge many existing models of category learning. For
example, exemplar models (e.g., Kruschke 1992),
which have enjoyed a long history of explaining key
psychological phenomena in the category learning
research, cannot account for the enhanced recognition
of rule-violating items without modifications. Exemplar models store every studied item in memory as
a separate trace. An item is classified into category
A or B depending on the item’s relative similarity to
all stored exemplars belonging to categories A and B.
The likelihood of recognizing an item as a studied item
is proportional to the sum of the item’s similarity to all
stored exemplars. The exemplar models predict no
recognition advantage for rule-violating items because
these items share the same similarity relations with
stored exemplars as rule-following items do. The exemplar models’ recognition memory is not sensitive to the
deviant nature of rule-violating items.
Rule-based (hypothesis-testing) models of category learning cannot account for the entire pattern
of the results either. A rule-plus-exception model

Schematic Influences on Category Learning and Recognition Memory

(Nosofsky et al. 1994) constructs rules and stores
exceptions to the rules. Rules represent the rulefollowing items. Information about inconsistent items
is explicitly stored as exceptions. Classification decisions are based on first matching the item to stored
exceptions. If there is no match, then the model
applies the rule. The likelihood of recognizing a test
item is determined by summing the output from the
rule and the output from the exceptions. The separate
storage of inconsistent information allows the ruleplus-exception model to predict a memory advantage
for rule-violating items. However, this model underpredicts the recognition memory for studied rulefollowing items because rules encode very little
information about these items. Furthermore, rulebased models cannot account for the greater recognition advantage for rule-inconsistent items violating a
stronger rule without also incorrectly predicting the
enhanced recognition of rule-following items that conform to the stronger rule.
The failures of the rule-based models and exemplar
models suggest that humans store more than just rules
to represent rule-following items, and they do not store
every individual item as a separate memory trace.
Instead, they build schema-like representations, in
which rule-following items are encoded as a set of
expectations, and rule-violating items are stored separately. One candidate is cluster-based representations.
The clustering model (Love et al. 2004) can predict
the memory advantage of rule-violating items. It represents categories by one or more clusters. The first
stimulus item encountered becomes the initial cluster.
The clustering model tries to group the next item
into this cluster. If this attempt leads to a classification error, the model recruits a new cluster for the
error-producing item. The model tries to assign subsequent items into the clusters that are most similar to
the items. In this way, the clustering model joins similar
items together and encodes rule-violating items separately when they elicit classification errors. An item is
classified according to the category label associated
with the cluster it is assigned to. Recognition response
is determined by summing the output of all clusters. In
the clustering model, whereas the rule-following items
will tend to cluster with one another, each ruleviolating item will be isolated in its own cluster.
Because of this differential storage, rule-violating
items will be more distinctive in memory and result

S

2961

in a recognition advantage. Furthermore, unlike the
rule-plus-exception model, the clustering model can
correctly predict some recognition memory for studied
rule-following items because rule-consistent clusters
encode more information than just the rule. The clustering model can also predict enhanced memory for
the item that violates the stronger (i.e., more frequent)
regularity by recruiting more clusters to represent
the stronger regularity. The rule-inconsistent cluster
representing the item that violates the stronger regularity is likely similar to a number of rule-consistent
clusters. To differentiate itself, more specific information needs to be encoded for the rule-inconsistent
cluster, which gives a recognition advantage for the
rule-inconsistent item.

Important Scientific Research and
Open Questions
Schema research examines pre-established knowledge
structures, and thus it is difficult to monitor the development of schemas or study how they are acquired. In
contrast, in category learning research, subjects acquire
novel categories during an experimental session, and
researchers can develop models of how representations
of categories are formed. These models make precise
predictions and may prove to be valuable tools for
research in schemas. At the same time, applying category learning models to schema research will require
extending these models to incorporate factors such as
prior knowledge and one trial learning that are important in everyday learning but are often ignored in
category learning research.
To build such models, a stronger link between
schema research and category learning research is
needed. More work should examine memory for consistent and inconsistent information in category learning research. There are many ways people learn about
categories other than classification learning. For example, in inference learning, the learner acquires category
knowledge by inferring properties of category members. A child inferring characteristics of birds, and
a medical student inferring symptoms of patients
with a certain disease are two examples of inference.
Based on the corrective feedback, they learn the category-property associations. Inference learning is essentially the opposite of classification learning. Whereas
the inference learner predicts the properties associated
with the given category, the classification learner

S

2962

S

Schematization

predicts the category label for an item with the given
properties. It is unclear how people would process
consistent and inconsistent information in inference
learning, and whether existing models of category
learning could capture aspects of inference learning
behavior. This line of work will give researchers the
opportunities to extend and refine current theories of
category learning.

Scheme
▶ Schema(s)

Scholasticism
▶ Epistemology and Learning in Medieval Philosophy

Cross-References
▶ Categorical Learning
▶ Categorical Representation
▶ Computational Models of Human Learning
▶ Concept Learning
▶ Exemplar-Based Learning
▶ Mental Representations
▶ Recognition of Prior Learning
▶ Rule Formation
▶ Rule Learning
▶ Schema Developement
▶ Schema-Based Learning
▶ Schema(s)
▶ Selective Learning

References
Kruschke, J. K. (1992). ALCOVE: An exemplar-based connectionist
model of category learning. Psychological Review, 99, 22–44.
Love, B. C., Medin, D. L., & Gureckis, T. M. (2004). SUSTAIN:
A network model of human category learning. Psychological
Review, 111, 309–332.
Nosofsky, R. M., Palmeri, T. J., & McKinley, S. C. (1994). Rule-plusexception model of classification learning. Psychological Review,
101, 53–79.
Palmeri, T. J., & Nosofsky, R. M. (1995). Recognition memory for
exceptions to the category rule. Journal of Experimental Psychology. Learning, Memory, and Cognition, 21, 548–568.
Rojahn, K., & Pettigrew, T. F. (1992). Memory for schema-relevant
information: A meta-analytic resolution. The British Journal of
Social Psychology, 31, 81–109.
Sakamoto, Y., & Love, B. C. (2004). Schematic influences on category
learning and recognition memory. Journal of Experimental Psychology: General, 133, 534–553.

Schematization
▶ Exemplar Learning and Schematization in Language
Development

School Atmosphere
▶ School Climate and Learning

School Choice
▶ Homeschooling and Teaching

School Climate and Learning
ERIN MORAN1, JOHN S. CARLSON1, BETTY TABLEMAN2
1
College of Education Erickson Hall, Michigan State
University, East Lansing, MI, USA
2
School of Social Work Baker Hall, Michigan State
University, East Lansing, MI, USA

Synonyms
Academic outcomes; School atmosphere; School culture; School environment

Definition
School climate refers to characteristics of the school
building or classroom setting. School climate is often
closely associated with school culture (i.e., the values,
attitudes, and expectations that are characteristic of
a particular school district). Factors associated with
a school’s climate can have an impact on student learning. A positive school climate fosters learning and student success through a caring, safe, and supportive
environment. A negative school climate can inhibit

School Climate and Learning

student achievement. The nature of a particular
school’s climate may be evident through a close examination of physical, social, affective, and academic
conditions.

Theoretical Background
Bronfenbrenner’s Bioecological Perspective evaluates
the interaction between person and environmental factors by providing a model depicting the various systems which influence an individual’s development and
learning. Such system levels include the individual’s
immediate environments including family, social
groups, and school, the interaction of these environments, external environments such as the parent’s
workplace, and larger cultural contexts, all combining
to affect development. The school is an important
factor within this bioecological system, because
a child spends a large portion of his/her day within
this context and is greatly influenced by the culture of
this institution and the social interactions created
within it.
Abraham Maslow proposed a Hierarchy of Needs to
understanding the essential needs of individuals in
order to achieve higher self-actualization and learning.
This theory is represented as a pyramid, with the most
basic needs at the bottom, representing the importance
of fulfilling these basic needs in order to achieve higher
actualization. While physiological needs (e.g., breathing, food, water, and sleep) are the most basic, more
complex needs that must be met include safety and
social needs. If an individual does not feel safe, have
a sense of belonging within a school setting, and have
a healthy sense of themself, then learning and growth
cannot occur because these essential needs are not
being met. When left unsatisfied, the need for safety
and belongingness will preempt focus and motivation
over school achievement. Thus, in order for students to
be able to learn at their greatest potential, schools must
be able to create a school climate which fosters a feeling
of safety and belonging through cultivating positive
relationships and support.
B.F. Skinner, in his behaviorist model of learning,
examined the critical role that environmental rewards
and punishments play in influencing behavior and
learning through his operant conditioning model, in
which behavior is reinforced with rewards or punishments in order to affect the future likelihood of the
behavior. With this work, Skinner showed that the

S

likelihood that a behavior would be repeated increased
when a reward was provided for this behavior and that
the likelihood would decrease if a punishment followed
the behavior. This model creates an important building
block in understanding how the individual’s environment affects individual behavior and learning.
Similar to learning theory, social learning theory
emphasizes the importance of the interaction between
social environmental influences and the individual’s
internal processes in learning and behavior. Social
learning theory proposes that individuals learn
through reinforcement, punishment, and observational learning. Reinforcement and punishment can
easily be enforced through student–student interactions or through teacher–student interactions within
the school setting. Observational learning suggests that
much of what an individual learns is through the
observation and imitation of others with whom they
are in close contact or whom they view as role models.
Within school settings, most learning of expectations
and appropriate behavior occurs through watching
peers or role models, such as teachers or older students.
A social and emotional learning (SEL) education
framework promotes teaching a core set of social and
emotional skills to help children build resilience and
handle challenging situations. The SEL program
teaches students the knowledge, skills, and attitudes
they need to create positive relationships, make appropriate decisions, and care for others (Collaborative for
Academic, Social, and Emotional Learning 2009). SEL
focuses on five categories of skills: self-awareness, selfmanagement, social awareness, relationship skills, and
responsible decision-making skills. This type of framework works to include students, create an atmosphere
of warmth and respect, as well as build positive relationships and skills for student–student and teacher–
student interactions. The inclusion of such a program
and its goals within schools have allowed students to
develop better skills in social interaction and therefore
improve the social atmosphere of the school climate.
Each of these theories and research supports the
importance of identifying and creating positive school
climates in affecting learning. In order to understand
the learning process, one must look not only at the
individual but also at the environmental influences on
learning. A positive, caring, safe, and suppportive
school climate provides a nurturing place for students
to learn and can increase their chances for success.

2963

S

2964

S

School Climate and Learning

Important Scientific Research and
Open Questions
There is currently an increased focus on school performance and schools are being held responsible for
improving academic performance. However, it is
important for schools to not only make changes to
curriculum and classroom instruction, but to take
note of the school climate and school culture, which
can have important effects on learning (Tableman
2004). School climate refers to the characteristics and
organization of the school building and classroom,
whereas the term school culture refers to the districtwide values and attitudes that provide a set of standard
expectations. While a positive school climate can be
created within a building or classroom, the school
culture of the district can play an important role in
positively or negatively influencing the school climate.
Often positive changes to the school climate cannot be
achieved without similar and supportive changes to the
school culture of the district. Tableman (2004) presents
the symbols, values, and assumptions/beliefs that constitute a school culture that supports learning and one
that impedes learning. A school culture supporting
learning will arrange the building in a way that best
serves the needs of children, values the opinions and
participation of administrators, parents, teachers, and
students, and is a place where all students are able to
learn and where parents play an active role in education
and want their children to succeed. On the other hand,
a school culture that impedes learning will not emphasize child accomplishments, the decision process will
not include parents, teachers, and students, and the
school will make the assumptions that parents do not
care about their child’s education and that some students are incapable of learning.
While school climate has been discussed in efforts
to reform school systems or improve academic performance, it can also serve as a resolution to such issues as
bullying, student conflicts, and character building.
Tableman (2004) identified four aspects of a school
climate that can have positive or negative effects
on the school and learning: a physical environment
that is welcoming and conducive to learning, a social
environment that promotes communication and interaction, an affective environment that promotes a sense
of belonging and self-esteem, and an academic environment that promotes learning and self-fulfillment.
For example, a physical environment that supports

learning might include a limited number of students,
open and welcoming classrooms, and a building in
which students feel safe and comfortable, whereas a
physical environment that hinders learning might
include a large number of students, classrooms that are
not visible and welcoming, and a building in which
students may worry about being bullied in the classrooms or hallways. Much research has been conducted
that highlights the value of promoting a positive school
climate, including higher grades and achievement, better
socioeconomic health, higher attendance, fewer suspensions, lower dropout rates, higher self-esteem, less anxiety and substance abuse, and higher standardized test
scores (Scales and Leffert 1999).
Research has also been conducted specifically
looking at the size of the school as an important indicator of school climate. The number of students in
a school can affect the physical and social environment
of the school, and it has been found in research that
smaller schools allow for better student–teacher interactions and more involvement from students (Cotton
1996). Smaller schools, generally defined as falling
within the range of 300–400 students in elementary
settings and 400–800 students in secondary settings,
allow for more involvement in extracurricular activities, better attitudes about school from students and
teachers, fewer behavior problems, and a higher sense
of belongingness among students.
The concept of school climate has gained importance within research on school effectiveness and learning over the past two decades. A safe school
environment that fosters relationships where students
feel respected, engaged with their work, and competent
is essential to the learning process (National School
Climate Center 2009). However, a gap exists between
the research on the importance of creating positive
school climates and the implementation of this knowledge into school practice. In order to effectively
improve or change a school climate, the school climate
must first be identified. There are several measures
available at the elementary school, middle school, and
high school level to determine school climate outcomes. Some elementary school measures with high
reliability, validity, and scoring methods include the
School as Caring Community Profile – II, the Sense
of School as Community Scale, and the Liking for
School Scale. At the middle school level, some measures receiving high ratings in these areas include the

School Climate and Learning

Kettering Scale of School Climate and the School Climate Questionnaire. At the high school level, a highly
rated measure includes the School Culture Scale
(Adapted from the Character Education Partnership,
as cited in Tableman 2004). School Climate can also
be assessed by measuring the physical, social, affective,
and academic environmental factors as identified by
Tableman (2004), which can be modified into a survey
checklist.
Once the school climate and areas of weakness have
been identified, then change is possible. Change to the
school climate will take time and will involve many
players within the school system. While positive
changes to a school climate may temporarily change
the attitudes and behaviors of students and teachers,
these changes will not be effective or lasting without the
support and change to the district-wide school culture
(Tableman 2004). School culture and climate change
begins with the district superintendent who will make
decisions about the district’s mission, goals, and staff,
whereas on a school level, the principal is responsible
for promoting change, setting expectations, and getting
teachers involved in the decision-making process. In
promoting change, a school system can promote a safe
and orderly environment (i.e., keep buildings in good
physical condition, implement consequences for inappropriate behaviors, create anti-bullying and conflict
resolution workshops, or create time-out areas for students to use throughout the day), facilitate conditions
that promote relationships and interaction (i.e., build
smaller schools, use smaller teacher–student ratios, and
promote small group activities), and promote
a positive affective environment (i.e., have students
attend summer school rather than retention, create
a cooperative environment rather than a competitive
one, and ensure that students have a positive relationship with at least one adult figure in the school)
(Tableman 2004). An interaction of some or all of
these suggestions to change can work to create a more
positive school climate and culture.
It is important to note that although teachers can
establish a positive school climate within their classroom (or principals within their building), unless they
are supported by the district through a supportive
school culture, they often burn out and leave. In 2007,
the Office of Superintendent of Public Instruction in
the state of Washington constructed an education manual highlighting nine research-based characteristics of

S

2965

high-performing schools (Shannon and Bylsma 2007).
These nine characteristics incorporate actions at the
federal, state, district, building, and classroom levels.
Among the nine characteristics are the following elements of a positive school climate and school culture:
a clear and shared focus, high standards and expectations for all students, effective school leadership, high
levels of collaboration and communication, a supportive learning environment that is safe, healthy, engaging,
and intellectually stimulating, and high levels of family
and community involvement.
Although significant research has been done in the
field of school climate and culture, and their effects on
learning, there still are many areas that need to be
further addressed. While school culture and school
climate have been discussed, there still remains the
issue of what the impact of school culture at a district
level is on school climate within individual schools,
which is the responsibility of principals and teachers.
This area is important to investigate because it will help
to shed light on how changes in the school culture can
lead to changes in the school climate in order to
improve student learning. Further investigation into
effective methods of incorporating learning and training in developing a positive school climate and culture
into academic and in-service preparation for teachers,
principals, and superintendents would benefit schools
in putting theory into practice.

Cross-References
▶ Affective Dimensions of Learning
▶ Choreographies of School Learning
▶ Climate of Learning
▶ Conditions of Learning
▶ Mediators of Learning

References
Center for Social and Emotional Education (2009). National School
Climate Center. In undefined. Retrieved January 4, 2010, from
http://nscc.csee.net/aboutnscc/
Collaborative for Academic, Social, and Emotional Learning. (2009).
Social and emotional learning and bullying prevention. Washington, D.C.: National Center for Mental Health Promotion and
Youth Violence. Prevention, Education Development Center.
Retrieved Jan 6, 2010, from http://www.casel.org/pub/reports.
php
Cotton, K. (1996). School climate, school size, and student performance. School improvement research series: Series, 1995–1996,
20, 49–75.

S

2966

S

School Culture

Scales, P. C., & Leffert, N. (1999). Developmental assets. Minneapolis,
MN: Search Institute. Retrieved January 6, 2010.
Shannon, G. S., & Bylsma, P. (2007). The nine characteristics of highperforming schools: A research-based resource for schools and districts to assist with improving student learning (2nd ed.). Olympia,
WA: OSPI.
Tableman, B. (2004). School climate and learning. Best Practice Briefs,
31, 1–10. Retrieved Jan 4, 2010, from www.outreach.msu.edu/
bpbriefs

School Culture
Culture consists of shared understandings between
actors that provide a sense of school identity, and
shapes how the school will seek to fulfill its mission.
Culture is comprised of deeply rooted values, beliefs,
symbols, behavioral norms, organizational history, and
myths that serve to shape individuals’ behavior within
schools.

Cross-References
▶ School Climate and Learning

School Engagement Profile
▶ Styles of Engagement in Learning

School Environment
▶ School Climate and Learning

School Motivation
DENNIS M. MCINERNEY
Educational Psychology Psychological Studies
Department, Block D1, 2Fl, Room 15, The Hong Kong
Institute of Education, Tai Po, Hong Kong

Synonyms
Achievement motivation; Engaged learning

Definition
Motivation (derived from the Latin motives, a moving
cause) may be defined as an internal state that instigates, directs, and maintains behavior. School motivation may be defined by four qualities: CHOICE,
motivated students choose to do some activities rather
than others. ENERGY, motivated students invest high
energy, characterized by involvement, enthusiasm, and
interest in activities. STANDARDS, motivated students
seek high personal standards in activities. CONTINUING MOTIVATION, motivated students return to
activities voluntarily, time and again, because they
enjoy and feel rewarded through them.

Theoretical Background
Historically, the dominant paradigm of school motivation until the early 1980s was reinforcement theory. B. F.
Skinner (1954) has had a powerful impact on the world
of education. Skinner’s work is now commonly considered an extension of Thorndike’s ▶ law of effect. In
general, Skinner believed that all animals, including
humans, learn things by having certain aspects of
their behavior reinforced while other aspects are not.
▶ Positive reinforcement occurs when something, such
as a reward, is added to the situation that makes the
performance of a particular behavior more likely in the
future. ▶ Negative reinforcement occurs when something unpleasant is removed from the situation contingent upon the performance of the behavior. Under this
paradigm, reinforcement acts as the motivating force
behind continued involvement in an activity. For children, reinforcement may be: material, such as toys or
some enjoyable activity; token, such as stamps and gold
stars; or social, such as the goodwill and recognition
of the teacher or competition. Often this style of
reinforcement is called extrinsic reinforcement as it is
externally applied, usually by a teacher. Positive reinforcement is considered more effective than negative
reinforcement.
While reinforcement is pervasive in classrooms as
a motivational tool, the current predominant theoretical paradigm of school motivation is cognitive. Cognitive theories emphasize mental processes and
perception as important elements of motivation and
the personal construction of the meaning of experiences which impacts on an individual’s level of motivation. Implied in such views of motivation are
a concern with personal beliefs and values, perceptions

School Motivation

of self, including perceptions about self-worth, abilities
and competencies, and goals and expectations for success or failure. This shift from elementary and basic
theories of motivation based on behavioral models to
cognitive models has had a strong impact on the way in
which we look at classrooms and schools and
how educators may most effectively engage students
in learning. Among cognitive theories are ▶ attribution
theory, ▶ expectancy-value theory, goal theory, ▶ selfdetermination theory, personal investment theory,
self-worth theory, and self-related constructs such as
▶ self-concept, self-regulation, and ▶ self-efficacy theories. Each of these theories implicitly or explicitly
addresses choice, energy, standards, and continuing
motivation.
Expectancy-value theory. This theory proposes that
in order to be motivated, students need to have
a strong expectation that they will be successful in
a learning activity. If they believe success is beyond
their capacity, they will be less motivated. However,
even if students possess a strong expectation of success
but do not value an activity/outcome, they will not be
motivated (Wigfield and Eccles 2000). Hence, the
essence of the motivational state that drives engagement is believed to be a judicious blend of expectation
of success and valuing success. Expectancy and value
are considered to be influenced by task-specific beliefs
such as ability beliefs (i.e., the individuals’ perceptions
of their current competence for a given activity), perceived difficulty of the task, and the individual’s goals
for schooling, sense of self, and affective memories
for similar tasks. These social cognitive variables are
influenced by an individual’s perceptions of their previous experiences and a variety of wider socialization
influences. The theory holds that students’ expectancy
and task values directly predict their achievement
outcomes, including performance, persistence, and
choices of which activities to do. Empirical support
for these proposed linkages has been found in longitudinal studies of children ranging in age from 6 to
18 years. Even when level of previous performance is
controlled, students’ competence beliefs strongly predict their performance in different domains, including
math, reading, and sports. It appears, however, that
while valuing a task may be important in the initial
choice of activity, expectancy of success is more
important to motivation and performance than valuing after that.

S

2967

Attribution Theory
When students have successes or failures, they automatically search for reasons. The motivational importance of such perceived causal control over one’s
successes and failures has been a focus of Bernard
Weiner’s attribution theory (Weiner 2004). The theory rests on three basic assumptions. First, it assumes
that people attempt to determine the causes of their
own behavior and that of others. Second, it assumes
that the reasons people give to explain their behavior
govern their future behavior in predictable ways
from one situation to the next. And third, it assumes
that the specific causes attributed to a particular
behavior will influence subsequent emotional and
cognitive behavior in predictable ways. Hence, the
essence of the motivational state that drives continuing engagement or disengagement in an activity is
believed to be an individual’s explanation and interpretation of the causes of their successes and failures.
In other words, individuals ask questions such as,
“Why did I fail the exam?” or “Why did I achieve so
well?” It is, however, more likely that these types of
questions are asked after failure, rather than after
success experiences.
Weiner originally postulated four causes that are
perceived as most responsible for success and failure in achievement-related contexts: Ability, Effort,
Task Difficulty, and Luck. Ability refers to a person’s
perceived performance capacity on a particular task.
Effort refers to the energy expended on the task. Task
difficulty refers to the perceived difficulty of the task,
i.e., tasks that most people can perform are considered
easy, while tasks that few can master are considered
difficult. Last, luck refers to the variables that lie
outside personal control that may affect the performance on the task (other than the first three mentioned), e.g., such as being unwell. Dimensions of
internality and externality, and controllability are
believed to influence the motivational impact of
attributions. Ability and effort are considered internal dimensions, task difficulty and luck are considered external dimensions. Ability, task difficulty, and
luck are considered uncontrollable, effort is considered controllable. According to the theory attributions to the internal controllable dimension, effort is
more efficacious than attributions to either internal
or external uncontrollable dimensions (ability, luck,
task difficulty).

S

2968

S

School Motivation

Achievement Goal Theory
▶ Achievement goal theory posits that there is an integrated pattern of beliefs (goal orientations) that lead
students to approach, engage, and respond to achievement tasks and situations in specific ways (Schunk et al.
2008). Goals represent the purposes that students have
in different achievement situations and are presumed
to guide students’ behavior, cognition, and affect as
they become involved in academic work (Ames 1992).
Hence, the essence of the motivational state that drives
engagement in an activity is believed to be an individual’s self-thoughts about the purpose(s) of achieving
(or being successful) in the activity. Two academic
goals have been the focus of much theorizing: mastery
goals (sometimes called learning goals or task goals)
and performance goals (sometimes called ego goals or
relative ability goals).
Central to a ▶ mastery goal is the belief that effort
leads to success: the focus of attention is the intrinsic
value of learning. With a mastery goal, individuals are
oriented toward developing new skills, trying to understand their work, improving their level of competence,
or achieving a sense of mastery. In other words, students feel successful if they believe they have personally
improved or have come to understand something.
Their performance relative to others is irrelevant; of
greater importance to them is completing the task. The
latest development of achievement goal theory has
bifurcated mastery goal orientation into two forms,
mastery-approach and mastery-avoidance (Elliot and
McGregor 2001). While a mastery-approach goal orientation is essentially identical to the mastery goal
orientation explained above, a mastery-avoidance
goal focuses on avoiding showing misunderstanding,
or avoiding not learning or not mastering the task.
Central to a performance goal is a focus on one’s
ability and sense of self-worth. Ability is shown by doing
better than others, by surpassing norms, or by achieving success with little effort. Public recognition for
doing better than others is an important element of
a performance-goal orientation. Performance goals
and achievement are “referenced” against the performance of others or against external standards such as
marks and grades. Consequently, “self-worth” is determined by one’s perception of ability to perform relative
to others. Hence, when students try hard without being
completely successful (in terms of the established
norms), their sense of self-worth may be threatened.

Performance goals have also been bifurcated into
performance-approach and performance-avoidance
goals. Students who hold a ▶ performance-approach
goal orientation want to do better than their classmates
so that they will be recognized as competent by their
peers, teachers, and parents. Students who hold
a ▶ performance-avoidance goal orientation do their
academic work primarily because they fear appearing
incompetent (Elliot and McGregor 2001).
Adaptive and maladaptive achievement goals.
Research suggests that mastery-approach is an adaptive
motivator. Students adopting a mastery-approach goal
orientation tend to use high levels of deep cognitive
strategies, such as elaboration, as well as metacognitive
and self-regulatory strategies. A similar pattern of findings has been found across cultures. The few studies
that have examined a mastery-avoidance goal orientation have demonstrated that this orientation is mostly
unrelated to cognitive strategies but negatively related
to intrinsic motivation, perceived competence and
classroom grades, and positively related to negative
emotions such as test anxiety and worry, help-seeking
threat, and to less adaptive approaches to learning.
Research has also demonstrated the adaptive
effects of a performance-approach goal orientation on
valued educational outcomes such as deep cognitive
strategies, positive affects, positive peer relationships,
and classroom grades. It should also be noted, however,
that a performance-approach goal orientation has
also been associated with negative outcomes such as
anxiety, disruptive behavior, and low retention of
knowledge.
In contrast, students who adopt a performanceavoidance goal orientation are more likely to use surface cognitive strategies such as rote memorization and
rehearsal. Furthermore, a performance-avoidance goal
orientation also appears to diminish intrinsic motivation for learning and is related to low levels of task
engagement and persistence, avoidance of help seeking,
anxiety, procrastination, and low grades.
A third type of goal orientation, work avoidance,
has received less attention by researchers but appears to
be quite important in influencing students’ attitudes
toward their school work (Dowson and McInerney
2001). ▶ Work-avoidance (or work-avoidant) goals
represent a type of goal orientation, where students
deliberately avoid engaging in academic tasks and/or
attempt to minimize the effort required to complete

School Motivation

academic tasks. This orientation, although distinct
from both performance and mastery orientation,
may nevertheless combine with these orientations to
affect students’ cognitive engagement and academic
achievement.

Effective Goal Setting
The above section describes specific types of goals
referred to as achievement goals, i.e., the psychological
purposes a student has for engaging in a task. However,
the term “goal” is also used in an alternative manner to
refer to the nature (quantity or quality) of an outcome
(performance) that individuals are trying to accomplish (Locke and Latham 2002). From this perspective,
the essence of the motivational state that drives continuing engagement in an activity is believed to be the
specificity, appropriateness, and achievability of
desired goals. Effective goal setting is considered essential to effective motivation and involves establishing
quantitative and qualitative standards or objectives to
serve as the aim of one’s actions. Setting appropriately
challenging levels of goals, divided according to different phases of attainment, is considered crucial to motivating students to engage in self-regulated learning
(Schunk 1990). Goals help give structure to student
learning, and a set of benchmarks by which students
and teachers can evaluate progress. Knowledge that
progress is being made toward desired goals is considered motivational, enhances students’ self-efficacy, and
leads students to select new, challenging goals.
There are three important aspects of goal setting.
First, goal setting is believed to be more effective when
goals are proximal (short-term) rather than distal
(long-term). This is because students are less likely to
become discouraged or unmotivated during long-term
tasks when they set short-term goals (Locke and
Latham 2002). Although the distal goal is important
for students to keep in mind, progress toward a longterm goal is sometimes difficult for students to gauge.
Short-term goals can raise self-efficacy simply by making a task appear more manageable, and they can also
enhance perceptions of competence by giving continual feedback that conveys a sense of mastery. Second, it
is believed that goals that incorporate specific performance standards are more likely to enhance motivation
and ▶ self-regulation than broad goals such as “you
must do your best” or “you should try hard.” Specific
goals raise performance because they specify the

S

2969

amount of effort required for success and boost selfefficacy by providing a clear standard against which to
determine progress. Third, it is believed that goals
should be moderately difficult or challenging. Unlike
the first two points above, goal difficulty does not
follow a linear relationships to performance. Overly
easy goals do not motivate, neither do goals perceived
to be impossible to achieve.

Self-Efficacy
Students’ self-efficacy refers to their perceptions or
beliefs of capability to learn and perform particular
academic tasks at particular levels at a particular
point in time. Efficacy beliefs influence how people
feel, think, motivate themselves, behave, and overall
achievement (Bandura 1997). For example, if students
enter an achievement situation with high self-efficacy
perceptions, they believe that they can accomplish what
the achievement situation requires them to do; as
a consequence, they are likely to approach tasks with
confidence and engage in them willingly, actively, and
persistently. On the contrary, students who are low on
self-efficacy beliefs are unsure what they can achieve or
even convinced that they cannot do the task; as a result,
they doubt their own capabilities for success and,
hence, are more likely to try to avoid the situation, or
if this is not possible, to give up easily when they
encounter frustration and failure. From this perspective, the essence of the motivational state that drives
engagement in an activity is believed to be the level of
self-efficacy the student has for the task at hand.
Research has operationally defined and assessed
self-efficacy with reference to the capabilities needed
to succeed in particular achievement situations. In this
regard, the term has a more specific meaning than such
terms as academic confidence or academic self-concept. In general, this research has documented the
positive role of self-efficacy beliefs on students’ effort,
persistence, and choice of activities, academic performance, interest, management of academic stressors,
and growth of cognitive competencies.

Self-Regulation
Drawn from self-efficacy theorizing, self-regulation is
considered to be a process that involves self-generated
thoughts, feelings, and actions for attaining academic
goals (Zimmerman 2006). The prototypical selfregulated learner views learning as a systematic and

S

2970

S

School Motivation

controllable process and accepts responsibility for
achievement outcomes. Prototypical self-regulated
learners approach tasks with confidence, diligence and
resourcefulness, and proactively seek out information
when needed, attempting to take the necessary steps to
master it. They are metacognitively, motivationally,
and behaviorally active participants in their own learning. From this perspective, the essence of the motivational state that drives engagement in an activity is
believed to be the effective utilization of a range of
metacognitive and self-regulatory learning strategies
the student has for the task at hand.
Research has demonstrated that self-regulated
learners value the importance of effort and intrinsic
interest in the task, appear to be high on selfattribution (i.e., they accept responsibility for successes
and failures), and also report high self-efficacy (i.e.,
a belief in themselves as learners). Students’ selfefficacy beliefs can enhance their motivation, which
in turn leads the students to continue learning in
a self-directed manner (Zimmerman 2006). Research
has also indicated that students who are academic high
achievers practice a greater range of self-regulatory
strategies, and more often, than low achievers. They
may also make greater use of the full range of strategies
such as self-evaluating, goal setting, and planning,
keeping records, and monitoring and reviewing.

Self-Determination
Self-determination theory (SDT) posits three universal, fundamental, and broad ranging psychological
needs which are considered to motivate goal-oriented
pursuits. These needs are need for autonomy (i.e., selfdetermination in deciding what to do and how to
do it), need for competence (i.e., developing and
exercising skills for manipulating and controlling the
environment), and need for relatedness (i.e., affiliation
with others through prosocial relationships) (Deci and
Ryan 2000). From the SDT perspective, the essence
of the motivational state that drives engagement in
an activity is believed to be the satisfaction of the
three basic needs that give people the freedom to
engage in self-determined activities. Thus, the theory
postulates that students are likely to be intrinsically
motivated to be engaged in learning activities when
they are placed in a learning environment that supports
satisfaction of their autonomy, competence, and relatedness needs.

There are four types of behavioral regulation that
can be ordered along a continuum between
amotivation at one end and intrinsic motivation at
the other end: external regulation, introjected regulation, identified regulation, and integrated regulation.
External regulation occurs when students’ actions are
regulated by external rewards, pressures, or constraints
(e.g., a student who does an assignment with an aim to
obtain teacher’s praise or to avoid parental confrontation). Introjected regulation occurs when students act
because they think they should or would feel guilty if
they did not (e.g., a student adopts this regulation when
he studies primarily because he knows that he will get
a bad grade and disappoint his parents if he does not).
Identified regulation occurs when the regulation or
value is adopted by the self as personally important
and valuable (e.g., a student who willingly does extra
work in a particular academic task because he believes
it is important for his self-selected future goal of
attending college or entering a particular occupation).
Lastly, integrated regulation, which is the most selfdetermined form of behavioral regulation and loosely
related to intrinsic motivation, occurs as a result from
the integration of identified values and regulations into
one’s coherent self (e.g., a student with two conflicting
identifications of being a good student and a successful
athlete who harmoniously integrates the two with each
other; when this state is achieved, the student’s behavior is an expression of what is valued by and considered
important by the student). The theory posits that students can gradually internalize extrinsic reasons for
completing necessary, but uninteresting, tasks and
learning activities and, thus, infuse agency into daily
learning activities.

Self-Concept
The notions of self-concept and ▶ self-esteem are often
used interchangeably. Researchers, however, argue that
self-concept refers to descriptive information about
oneself such as height, hair, skin color, ability in academics, sports, and so on, whereas self-esteem is the
evaluative component of self-concept, which refers to
how one feels about these objective qualities of selfdescription. Thus, in this regard, self-esteem reflects
the components of self-concept judged to be important
by a particular individual. The term self-concept is used
in this discussion to cover both the descriptive and
evaluative dimensions.

School Motivation

Research has consistently indicated that the more
positive students’ academic self-concept, the higher
their achievement. Individuals who have a positive
self-concept are expected to be more motivated to
perform particular activities than those who have
a poor self-concept, and success in performing certain
activities is believed to enhance self-concept. Hence,
the relationship between self-concept and behaviors,
particularly in terms of success and failure, is presumed
to be reciprocal (Marsh and Craven 1997). From a selfconcept perspective, the essence of the motivational
state that drives engagement in an activity is believed
to be the student’s positive feelings about their capacities and aptitudes.
Self-concept is formed through social interaction
and social comparison. Feedback from significant
others such as parents, siblings, teachers, and peers is
influential in the growth of one’s self-concept. Social
frames of reference indicate to us what our capacities
and qualities are under particular circumstances.
For example, on an objective criterion, some students
may be quite good at mathematics, yet when they use
others who are superior at mathematics as a frame of
reference, they may develop a relatively negative
mathematics self-concept (Marsh and Craven 1997).
This is sometimes referred to as an external frame of
reference. Furthermore, according to Marsh, we tend
to compare our self-perceived skills in one area (such
as mathematics) with our self-perceived skills in
another (such as English) and use this internal, relativistic impression as a second basis for arriving at our
self-concept in particular areas. This is often referred
to as an internal frame of reference. Hence, students
who are good at both mathematics and English may,
nevertheless, have a more negative self-concept in
mathematics if they perceive that they are better at
English and vice versa. This explains why even slow
learners differentiate their self-concepts across subjects and hold high self-concepts in some areas, even
though their objective performance may be poor. This
effect, called internal-external frame of reference, seems
to be universal among students from a wide range
of cultures.
There is also a relationship between a student’s selfconcept, average school performance, and the student’s
school achievement. This relationship is often referred
to as the big-fish-little-pond effect. That is, it is often
better for a bright student to be a “big fish” in a “little

S

2971

pond” (i.e., doing well among a mixed-ability group)
than to be a “little fish” in a “big pond” (i.e.,
performing at an average level in a high-ability
group). In the former case, it is easier for students to
establish and maintain positive feelings about their
academic accomplishments, which serve to reinforce
further academic pursuits. In selective educational
environments, where the average ability of students is
high, it is more difficult to establish and maintain these
positive feelings, and high-ability students may choose
less demanding coursework and have lower academic
self-concepts, lower achievement scores, lower educational aspirations, and lower occupational aspirations
than similar students in nonselective educational
environments.

Important Scientific Research and
Open Questions
There are a number of contemporary cognitive theories
of motivation that attempt to provide insights into and
explanations of the dynamics of school motivation. No
one of these theoretical perspectives is sufficient, in and
of itself, to explain the complex nature of school motivation. Nevertheless, drawing the various perspectives
together may give an enhanced capacity to educators to
facilitate the active engagement of students in learning.
Further research should be directed at “harmonizing”
the various cognitive theories with the purpose of
examining whether there are some underlying principles that might characterize the essence of motivation
from a cognitive perspective.
Although used to guide instructional practices in
many non-Western settings, most theories of school
motivation are based upon Western theorizing. There
is a need to consider whether these theories have validity and applied value in non-Western settings. It is also
important to develop “indigenous” theories of school
motivation that might be more heuristic in nonWestern educational contexts.

Cross-References
▶ Academic Motivation
▶ Achievement Motivation and Learning
▶ Attribution Theory of Motivation
▶ Cognitive and Affective Learning Strategies
▶ Cognitive Self-Regulation
▶ Conditioning

S

2972

S

Schooling

▶ Engagement in Learning
▶ Law of Effect
▶ Motivation and Learning: Modern Theories
▶ Motivation Enhancement
▶ Motivation to Learn
▶ Motivation, Volition, and Performance
▶ Motivational Variables in Learning
▶ Self-Efficacy & Learning
▶ Self-Concept and Learning
▶ Self-Confidence and Learning
▶ Self-Determination of Learning
▶ Self-Efficacy for Self-Regulated Learning
▶ Self-Regulation and Motivation Strategies

References
Ames, C. (1992). Classrooms: Goals, structures, and student motivation. Journal of Educational Psychology, 84, 261–71.
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
Freeman.
Deci, E. L., & Ryan, R. M. (2000). The “what” and “why” of goal
pursuits: Human needs and the self-determination of behaviour.
Psychological Inquiry, 11, 227–268.
Dowson, M., & McInerney, D. M. (2001). Psychological parameters of
students’ social and work avoidance goals: A qualitative investigation. Journal of Educational Psychology, 93, 35–42.
Elliot, A. J., & McGregor, H. A. (2001). A 2 
 2 achievement
goal framework. Journal of Personality and Social Psychology, 30,
957–971.
Locke, E. A., & Latham, G. P. (2002). Building a practically useful
theory of goal setting and task motivation: A 35-year Odyssey.
American Psychologist, 57, 705–717.
Marsh, H. W., & Craven, R. (1997). Academic self-concept: Beyond
the dustbowl. In G. Phye (Ed.), Handbook of classroom assessment: Learning, achievement, and adjustment. San Diego:
Academic.
Schunk, D. H. (1990). Goal setting and self-efficacy during selfregulated learning. Educational Psychologist, 25, 71–86.
Schunk, D. H., Pintrich, P. R., & Meece, J. L. (2008). Motivation in
education: Theory, research, and applications (3rd ed.). Upper
Saddle River: Pearson Prentice Hall.
Skinner, B. F. (1954). The science of learning and the art of teaching.
Harvard Educational Review, 24, 86–97.
Weiner, B. (2004). Attribution theory revisited: Transforming cultural plurality into theoretical unity. In D. M. McInerney &
S. Van Etten (Eds.), Big theories revisited (pp. 13–29). Greenwich:
Information Age Press.
Wigfield, A., & Eccles, J. S. (2000). Expectancy–value theory of
achievement motivation. Contemporary Educational Psychology,
25, 68–81.
Zimmerman, B. J. (2006). Self-regulation and effective learning.
In D. M. McInerney & V. McInerney (Eds.), Educational psychology: Constructing learning (pp. 190–191). Sydney: Pearson
Education.

Schooling
▶ Formal Learning

Schools-Within-Schools
▶ Community of Learners

Science for the Internet
▶ Web Science

Science for the Web
▶ Web Science

Science Inquiry
▶ Technology-Enhanced Learning Environments for
Science Inquiry

Science of Behavior
▶ Behaviorism and Behaviorist Learning Theories

Science, Art, and Learning
Experiences
MARK GIROD
Western Oregon University, Monmouth, OR, USA

Synonyms
Aesthetic experience; Transformative experience

Science, Art, and Learning Experiences

Definition
Any definition of a set of terms as widely diverse as
science, art, and learning experiences may be viewed as
simplistic and exclusionary. Rather, one should view
these terms and their connections as suggesting there
are some salient qualities about learning experiences that
emerge at the intersection of art and science. Even this,
however, runs the risk of an overly prescriptive definition. Scientists, artists, philosophers, historians, and
psychologists have written about these terms but there
is an emerging unified discourse that draws mainly from
the aesthetic theory of (John Dewey 1934). From this
one might define these learning experiences as those
which transform the perceptions of the learner, yield
emotional qualities similar to those experienced when
viewing great art, and educate conceptual understanding of important subject matter. These kinds of learning experiences may also be described as transformative
learning experiences or aesthetic experiences.

Theoretical Background
Consideration of connections between art and science
is not new but recognition of the unique nature of these
connections and their role in facilitating powerful
learning experiences is an emerging field. Philosophers
have long recognized that science and art share some
underlying values such as parsimony, form, symmetry,
pattern, and unity. Similarly, scientists have acknowledged the role these values play in judgments about
science and in pursuit of scientific inquiry (Fischer
1999). The science education community has begun
to see efforts to build theoretical models, pedagogical
strategies, and generate empirical evidence to systematically support these linkages in ways that promote
powerful learning experiences.
The best-articulated lines of research in this field
follow (Dewey 1934) and the translation of Deweyan
aesthetic theory into psychological and educational
principles (Jackson 1998). Emphasis is on aesthetic
experience and transaction between experiencer (person) and experienced (world). This transaction is characterized by both doing and undergoing, and acting
and changing. It is infused with appropriate emotion,
bears anticipatory qualities, and consummates fully as
opposed to ending abruptly. Scholars in this area have
used these key aesthetic constructs as a roadmap for the
articulation of educational theories that yield similar
powerful learning experiences.

S

2973

Key constructs include expansion of perception,
motivated use, and experiential value. Expansion of perception refers to seeing anew through important and
powerful scientific and artistic lenses. Powerful learning
experiences at the intersection of science and art yield
new ways of seeing the world and the learner, and his or
her position in the world. Motivated use refers to the
desire to utilize or exercise newfound knowledge, skills, or
dispositions. Experiential value refers to finding enjoyment in what new learning provides or how it enriches
ordinary experience. Similar conceptual elements may be
described as transformative, aesthetic experiences, or aesthetic understanding.
Pedagogical models for facilitation of learning experiences that link science and art through aesthetic experiences include modeling expansion of perception,
increased value, interest, and positive attitude toward
science, art, and new learning. Curricular structures that
focus on important and powerful subject matter learned
through metaphor, promotion of awe and wonderment,
use of sublime, story-telling, and powerful imagery is also
suggested. A key Deweyan construct here is that of the
idea (Dewey 1933). An idea can be viewed as a lens or
a plan of action that illuminates and fosters powerful,
artful learning.

Important Scientific Research and
Open Questions
Research related to science, art, and learning experiences may be spread across a wide range of literatures
including those in educational psychology exploring
motivation, engagement, interest, attitude, flow, transfer, efficacy, and identity and may derive from cognitive
conceptual frameworks focused on conceptual change
or socio-cultural frameworks focused on situative
understanding and applications of new knowledge in
authentic settings. In teacher education literature only
a few key references exist describing pedagogical strategies useful for facilitating powerful learning experiences at the intersection of art and science (Pugh and
Girod 2007; Wickman 2006).
To date, it is fairly well established that teaching in
ways that link art and science can have a positive impact
on student conceptual understanding, reduce forgetting, increase motivation and engagement with science,
facilitate positive identity affiliations toward science,
and promote transfer of learning to out-of-school environments (Girod et al. 2010; Pugh et al. 2010).

S

2974

S

Scientific Curiosity

A wide range of open questions remain including
descriptive studies that explore the prevalence of naturally occurring learning experiences at the intersection
of art and science both in schools and beyond as well as
deep analyses of the psychological nature and educational value of these kinds of learning experiences.
Correlational studies remain necessary to explore relationships between a wide range of contextual and
learner-level variables and outcomes associated with
science, art, and learning. Causal studies are needed
to explore high valence pedagogical strategies, curricular models, and to establish empirically validated theoretical linkages between constructs of interest.
Measurement studies are needed to further demonstrate the quality, quantity, and predictive validity of
learning experiences at the intersection of science
and art.
Much work remains to be done in this field both
from the theoretical perspectives described in this entry
and from beyond.

Cross-References
▶ Aesthetic Learning
▶ Affective Dimensions of Learning
▶ Dewey, John
▶ Engagement in Learning
▶ Experimental Learning Environments
▶ Laboratory Learning
▶ Meaningful Verbal Learning

References
Dewey, J. (1933). How we think: A restatement of the relation of
reflective thinking to the educative process. Boston: Heath.
Dewey, J. (1934/1980). Art as experience. New York: Berkley.
Fischer, E.P. (1999). Beauty and the beast: The aesthetic moment in
science (trans: Oehlkers, E.). New York: Plenum Publishing
Corporation.
Girod, M., Twyman, T., & Wojcikiewicz, S. (2010). Teaching and
learning science for transformative, aesthetic experience. Journal
of Science Teacher Education, 21, 801–824.
Jackson, P. (1998). John Dewey and the lessons of art. New Haven: Yale
University Press.
Pugh, K., & Girod, M. (2007). Science, art and experience:
Constructing a science pedagogy from Dewey’s aesthetics.
Journal of Science Teacher Education, 18(1), 9–27.
Pugh, K. J., Linnenbrink-Garcia, L., Koskey, K. L. K., Stewart, V. C., &
Manzey, C. (2010). Motivation, learning, and transformative
experience: A study of deep engagement in science. Science Education, 94, 1–28.
Wickman, P. (2006). Aesthetic experience in science education.
Mahway: Lawrence Erlbaum.

Scientific Curiosity
▶ Epistemic Curiosity

Scientific Discovery
▶ Theory Construction

Scientific Inscription
▶ Scaffolding Learning by the Use of Visual
Representations

Scientific Method
This term refers to a body of techniques for investigating phenomena, acquiring new knowledge, or
correcting and integrating previous knowledge. To be
termed scientific, a method of inquiry must be based
on gathering observable, empirical and measurable
evidence subject to specific principles of reasoning.
A scientific method consists of the collection of data
through observation and experimentation, and the formulation and testing of hypotheses.

Cross-References
▶ Methodologies of Research on Learning (Overview
Article)

Scientific Realism
▶ Naturalistic Epistemology

Scope of Attention
▶ Behavioral Capacity Limits

Seamless Learning

Script And/Or Schema
Acquisition Through External
Representations

S

2975

Scripted Cooperation
▶ Collaboration Scripts

▶ Supplantation Effect on Learning

Seamless Learning
Script(s)
NORBERT SEEL
Faculty of Economics and Behavioral Sciences,
Department of Education, Freiburg, Germany

Scripts are a subtype of schema that is used for
representing procedural knowledge. More specifically,
a script is a predetermined, stereotyped sequence of
actions that define a well-known situation. As do
schemas, scripts also contain slots and default values
that can be filled. The resulting structure is an
interconnected whole, and what is in one slot affects
what can be in another.
The classic example of a script involves the typical
sequence of events that occur when a person dines in
a restaurant: finding a seat, reading the menu, ordering
drinks from the staff . . ..
Scripts were developed in the early AI work by Roger
Schank and Robert P. Abelson and their research group.
These authors basically assume that procedural knowledge from everyday situations is stored in the human
mind in the form of scripts. They are very similar to
frames, except that the values that fill the slots of a script
must be strictly ordered. Understanding a familiar situation involves activating a stored script of this situation.

Cross-References
▶ Knowledge Representation
▶ Scaffolding for Learning
▶ Schema(s)
▶ Schema-Based Problem Solving

References
Abelson, R. P. (1981). Psychological status of the script concept. The
American Psychologist, 36, 715–726.
Schank, R. C., & Abelson, R. P. (1977). Scripts, plans, goals, and
understanding: An inquiry into human knowledge structures.
Hillsdale: Erlbaum.

CHEE-KIT LOOI, HYO-JEONG SO, WENLI CHEN, BAOHUI
ZHANG, LUNG-HSIANG WONG, PETER SEOW
National Institute of Education, Nanyang
Technological University, Singapore, Singapore

Synonyms
Blended learning; Flexible learning; Ubiquitous
learning

Definition
▶ Seamless learning bridges private and public learning
spaces where learning happens through both individual
and collective efforts, and across time and different contexts (such as in-school vs after-school, formal vs ▶ informal learning, physical world vs. virtual reality or
cyberspace). Traditionally, ▶ formal learning is defined
as learning that happens at a fixed time following
predefined curricula or plan. Informal learning, on
the other hand, means a mode of learning driven by
self-interest outside of school environment, and is
emergent in nature.

Theoretical Background
Seamless learning is marked by the continuity of learning experiences across different scenarios or contexts.
The challenge is to enable learners to learn whatever
they are curious about and to seamlessly switch
between different contexts, such as between formal
and informal contexts, between individual and social
learning, and by extending the social spaces in which
learners interact with each other. Technology plays an
important role in mediating the switching between
these different spaces.
Our students live in a digital world, and the use
of technologies such as instant messaging, video sharing, photo sharing, social network tools, podcasting,
and blogging are integrated into their lifestyles.
Smartphones are used not only for making calls, but

S

2976

S

Seamless Learning

for taking pictures and uploading them to shared
spaces, creating mobile blogs, or accessing the web on
the move. The use of these technologies facilitates
communication, collaboration, sharing, and learning
in informal settings with their peers, friends and family,
unbounded by time and location. Students spend more
time in such “informal” settings than “formal” settings
in the school. One of the fundamental challenges for
the twenty-first century learners is that there is a need
to understand not only what they learn, but how and
when they learn. A deep understanding of how learners
learn informally can be used to inform the facilitation
of formal and informal learning practices.
A seamless learning environment bridges private
and public learning spaces where learning happens
through both individual and collective efforts and
across different contexts (such as in-school vs afterschool, formal vs informal, physical world vs. virtual
reality or cyberspace). When thinking about learning
scenarios in schools and other places of learning, people often conjure up mental images of a classroom with
all seats facing the teacher. The presumption is that
learning happens at fixed times and fixed places. However, with the diffusion of technology, the notions of
place, time, and space for learning have changed. The
learning space is no longer defined by the “class” but by
“learning” unconstrained by scheduled class hours or
specific locations. With mobile technologies at hand,
students can learn seamlessly – both in classroom and
out of classroom, both in school time and after school
time. While learning can be facilitated or scaffolded by
teachers or peers, at other times it could be studentinitiated, impromptu, and emergent.
Personal, portable, wirelessly networked technologies
will become ubiquitous in the lives of learners. With
quick and ready access to these technologies, we enter
into new phase in the evolution of technology-enhanced
learning (TEL), characterized by “seamless learning
spaces” and marked by the continuity of learning experiences across different scenarios or contexts, and emerging
from the availability of one device or more per student
(“one-to-one”) (Chan et al. 2006). These developments,
supported by theories of social learning, situated learning, and knowledge construction, will influence the
nature, process, and outcomes of learning. One-toone TEL will push the frontier of technology use in
formal and informal learning. The ingenious,

emergent, or pervasive use of one-to-one devices in
some usage contexts may be close to the tipping point
in terms of effecting fundamental shifts in the ways
students learn in schools and outside of schools. Previous ▶ mobile learning research, however, has typically
focused on either formal or informal settings and failed
to examine the integrated and synergetic effects of
linking these two contexts or environments of learning
(Sharples 2006).
Mobile technology has the potential to mediate
seamless learning, challenge the traditional dichotomous distinction between formal learning and informal learning by creating seamlessly connected learning
experiences (Looi et al. 2010). While research on cognition and learning during the past decades has emphasized the importance of linking learning in the
classroom and learning in the field, the dominant characteristic of school learning still has a strong focus on
individual cognition, pure mental activity without tool
use, and overly context-general learning. Moreover,
there are tensions between formal learning, which is
based on fixed curricula enacted in classroom environments, and informal learning where learners are participating in intentional or unintentional experiences
outside school settings. We believe that the two forms
of learning should not be seen as dichotomous and
conflicting situations (Sharples 2006). Instead, by utilizing the affordances of mobile technology, we can
bridge the gap between formal and informal learning,
and encourage students to learn in naturalistic settings
for developing context-specific competencies.
A suitable lens for interpreting seamless learning
activities is the distributed cognition (Hollan et al.
2001). In a seamless learning environment, learning
takes place through individual learning in private
spaces, collaborative learning in public spaces, and
interactions with the environment and articfacts across
time, context and physical or virtual spaces mediated
by technology (Sharples 2006; Chan et al. 2006). Distributed cognition can provide a framework for understanding how learning occurs through the interactions
of students, artifacts, and the environment mediated by
technology over space and time. Though early studies
aim to understand how teams comprising individuals
with different skills or assigned roles function together
to achieve a specific task, the notion of distributed
cognition is applicable to learning as students’ expertise

Seamless Learning

and interactions are also distributed in the classroom
(Pea 1993). Hollan et al. (2001) proposed three principles in which cognitive processes occur: They are distributed: (1) across the members of the social group;
(2) over time; and (3) coordination between material
or environmental structures in the system. Learners
interact with the environment in the community, artifacts, activity, and space through the cognitive tools to
form a joint learning system (Kim and Reeves 2007).
Our seamless learning framework is based on our identification of components in a seamless learning environment and the theory of distributed cognition,
namely, space, time, context, community, and tools.
Space: Seamless learning suggests that the learners
can move seamlessly between different spaces –
physically and virtually. The physical space where students use to conduct inquiry with the mobile device
can be used as a resource for learning (Squire and
Klopfer 2007).
Time: Over time, when learners operate on artifacts,
collaborate with peers, teachers and experts or make
discovery, they acquire and construct knowledge. Time
can play an important role in shaping and evolving
inquiry, and developing deeper understanding as they
interact in a seamless learning environment.
Context: The context of the designed or emergent
activities in which the learner is engaged and the environment in which these activities occur impacts their
learning, application, and plans.
Community: The community in a seamless learning
environment comprises learners, teachers, and domain
experts. Individual learners in a seamless learning environment can move from individual learning to community learning and from private cognition to public
cognition and vice versa.
Tools: As students use the mobile device to record
data, capture images, upload data to the online portal
and reference them, mobile devices and online portal
become cognitive tools where they are able to offload
tasks, recall information over time, and modify their
initial thoughts.

Important Scientific Research and
Open Questions
A large research gap exists in the area of bridging
formal and informal settings in order to construct
a seamless learning environment. There is also a lack

S

of longitudinal studies to explore the affordances of
such learning environments in promoting twenty-first
century knowledge, skills, and positive attitudes toward
learning. Regarding the use of methodological issues
for seamless learning research, the design experiment
methodology is typically used to design and implement
seamless learning research. The choice of design experiment is ideal as this method stresses upon systemic
thinking on the interdependence of design elements,
and the importance of examining emerging issues
through progressive refining processes (Collins et al.
2004). More importantly, in order to design sustainable twenty-first century learning environments, as
researchers, we need to make a commitment to conduct sustainable research, and this necessitates the use
of theoretical and methodological lens that are congruent with the goals of this research.
An important consideration in seamless learning
research with mobile devices is to understand the
enactment of learning activities, which unfold in various situations. Previous research that examined the use
of mobile devices in informal settings has shown both
promises and challenges (e.g., Sharples 2006; Squire
and Klopfer 2007). Mobile technologies with portability, connectivity, and versatility enable learning to be
ubiquitous in and out of classrooms, provide potential
opportunities for collaborative learning, and enrich
learning experiences with the support of technologies.
For instance, Price and Rogers (2004) suggest that
mobile devices can be used to help students explore
digitally augmented physical environments where contextually relevant information and resources are provided. In such digitally enhanced settings, students
using mobile devices can explore, capture, and manipulate both physical and virtual (or digital) objects for
active understanding. From design and research perspectives, however, studying mobile learning in informal settings is challenging because students are “on the
move” across different modes of space (both physical
and virtual) and time. Thus, an ethnographic approach
(Anderson-Levitt 2006) can be integrated into design
research for observing how students are engaged in
informal and formal learning settings in their interaction with their handheld devices, peers, teachers, and
other people in their learning community.
Studies that focus on examining short-lived learning experiences such as user satisfaction surveys and

2977

S

2978

S

Seamless Learning

strict comparisons of test measures, fail to provide
comprehensive perspectives on learners’ meaningful
experiences across settings over time. Indeed, mobile
learning researchers face methodological challenges in
terms of the scales of space and time (Lemke 2000):
how to record learning across different physical spaces
and different technological media, and how to examine
learning in the longer timescale including informal
learning outside school contexts. However, because of
the novelty of the proposed study, there is no “off-theshelf ” methodology for us to adopt. For data collection, the learning trajectory of students using mobile
technologies for learning across subjects and over time
needs to be recorded.
Possible data sources include but are not limited to
observations, field notes, audio and video recordings,
interviews, student artifacts, self-documentation by
participants, and log files on computers. There are
also methodological issues involved in observations
such as distorted behaviors and artificial tasks (Gardner 2000), and also ethics and privacy issues for observing students outside school settings. Researchers
should aim to minimize potential problems by
employing unobtrusive methods such as log files,
which provide an authentic, time-efficient means of
recording student learning behaviours. In situ sampling of the students’ daily experiences with mobile
devices can be captured using the experience sampling
method (ESM) (Csikszentmihalyi and Larson 1987).
This method may provide us with a better understanding and natural assessment of how students are engaged
in informal learning everyday with mobile devices as
they are using it.
By employing ethnographic methods, in situ selfreport procedures, constant comparisons, and
sustained observations as well as analyzing quantifiable
measures, we can critically examine how learners use
mobile technology across subject areas and how different user experiences and motivation levels affect learning over time.
Important assessment issues loom in the space of
seamless learning. With students’ use of devices for informal learning, what are the indicators of learning? Or what
accounts for learning at the first place. One well-cited
definition of learning is “changing through experience . . .
acquiring relatively permanent change in understanding,
attitude, knowledge, information, ability, and skill
through experience” (Wittrock 1977, p. ix). To us, the

more important change might be in student value and
character change, which can gauge students as lifelong
learners and persons-to-be. Therefore, challenges exist in
assessing the skills, knowledge, identity, values, and epistemology (Shaffer 2007) as students become adept in
using the mobile device as routine practice in the classroom and out of the classroom.
One approach for assessment in seamless learning
environment is to adopt a preparation for future learning
(PFL) framework (Bransford and Schwartz 1999) to
emphasize assessment for learning. The purpose of
PFL is to promote deep understanding and knowledge
transfer in multiple contexts, and mobile devices can
act as a mediating tool enabling such learning transfer.
Traditional approaches of assessment focus on measuring student abilities to directly apply their previous
knowledge to new problems without help or resources.
This type of direct application, however, fails to measure the zone of proximal development (Vygotsky
1978), that is, students’ potential abilities to learn in
knowledge- and resource-rich environments. In seamless learning research, researchers can explore different
sequencing of learning conditions such as (a) formal vs.
informal learning, (b) intentional vs. unintentional
learning, and (c) abstract context-general vs. and concrete context-specific settings in order to identify
enabling conditions that better prepare students for
future learning.
The motivation for promoting informal learning
probably first started from the training work place skills
because it involves obvious costs. Another term we have
not mentioned is “nonformal” learning, which refers to
learning that happens in formal learning settings but is
not tested or assessed in traditional ways. So, formal,
informal, and ▶ nonformal learning are all learning. As
we have argued, learning can happen at any situation and
context. However, how to capture learning that is not
planned, not fixed and probably without validated instruments to measure, and usually individualized poses great
challenges to learning science researchers. We need also to
collect multiple data sets over time for triangulation
purposes. When considering the linkage between formal
and informal learning, we might be able to infer the
effectiveness of informal learning through assessing the
conceptual equivalents specified in formal curricula. On
the other hand, performances in informal settings can be
a result of formal learning in terms of preparing the
students for future learning.

Second Language Acquisition

Studying school-based learning and following
through with after-school learning will enable the
exploration of a theory of mobile learning for seamless
learning tied strongly to empirical evidence. For
instance, the PFL perspective can be adopted to frame
the use of mobile devices in informal settings as
enabling students to familiarize with a problem and
its context before in-school learning of formal concepts. Our findings will be used for further understanding of the application of the PFL framework, as well as
providing evidence of the efficacy of different sequencing of formal and informal learning activities.
Research into seamless learning needs a strong focus
on pedagogy, professional development of teachers, codesign of lessons with teachers, a design research perspective, and low-cost affordable mobile learning devices.
International collaboration and innovation can contribute toward the broader research agenda. By organizing
and sharing information across design experiments in
diverse settings, a collaboration of researchers can more
rapidly and systematically explore the design space (Hawkins 1997). For instance, the same-grade classrooms
across different countries can implement mobile learning
devices for all subject areas, allowing a broad examination
of solutions and challenges. By collaborating across the
globe, researchers could take advantage of different student device preferences, exchange curriculum ideas,
understand cultural differences, and better address issues
of scale.

Cross-References
▶ Blended Learning
▶ Formal Learning
▶ Informal Learning
▶ Lifelong Learning
▶ Mobile Learning

References
Anderson-Levitt, K. M. (2006). Ethnography. In P. B. Elmore, G.
Camilli, & J. Green (Eds.), Handbook of complementary methods
in education research (pp. 279–298). Washington D.C/Mahwah:
AERA & Lawrence Erlbaum Associates.
Bransford, J. D., & Schwartz, D. L. (1999). Rethinking transfer:
A simple proposal with multiple implications. Review of Research
in Education, 24, 61–100.
Chan, T.-W., et al. (2006). One-to-one technology-enhanced learning: An opportunity for global research collaboration. Research
and Practice of Technology Enhanced Learning, 1(1), 3–29.

S

Collins, A., Joseph, D., & Bielaczyc, K. (2004). Design research:
Theoretical and methodological issues. Journal of the Learning
Sciences, 13(1), 15–42.
Csikszentmihalyi, M., & Larson, R. (1987). Validity and reliability of
the experience sampling method. The Journal of Nervous and
Mental Disease, 175(9), 526–536.
Gardner, F. (2000). Methodological issues in the direct observation of
parent–child interaction: Do observational findings reflect the
natural behavior of participants? Journal of Clinical Child and
Family Psychology Review, 3(3), 185–198.
Hawkins, J. (1997). The national design experiments consortium: Final
report. New York: Center for Children and Technology, Educational Development Center.
Hollan, J., Hutchins, E., & Kirsch, D. (2001). Distributed
cognition: Toward a new foundation for human-computer
interaction research. In J. M. Carroll (Ed.), Human computer
interaction in the new millennium (pp. 75–94). New York: ACM
Press.
Kim, B., & Reeves, T. C. (2007). Reframing research on learning with
technology: in search of the meaning of cognitive tools. Instructional Science, 35(3), 207–256.
Lemke, J. L. (2000). Across the scales of time: Artifacts, activities, and
meanings in ecosocial systems. Mind, Culture, and Activity, 7(4),
273–290.
Looi, C. K., Seow, P., Zhang, B., So, H. J., Chen, W.-L., & Wong, L. H.
(2010). Leveraging mobile technology for sustainable seamless
learning. British Journal of Educational Technology, 41(2),
154–169.
Pea, R. (1993). No distribution without individuals’ cognition:
A dynamic interactional view. In G. Salomon (Ed.), Distributed
cognition. Psychological and educational perspectives (pp. 111–138).
Cambridge: Cambridge University Press.
Price, S., & Rogers, Y. (2004). Let’s get physical: the learning benefits
of interacting in digitally augmented physical spaces. Computers
& Education, 43, 137–151.
Shaffer, D. W. (2007). How computer games help children learn.
New York: Palgrave.
Sharples, M. (2006). How can we address the conflicts between
personal informal learning and traditional classroom education.
In M. Sharples (Ed.), Big issues in mobile learning (pp. 21–24).
Nottingham: LSRI, University of Nottingham.
Squire, K., & Klopfer, E. (2007). Augmented reality simulations on
handheld computers. Journal of the Learning Sciences, 16(3),
371–413.
Vygotsky, L. S. (1978). Mind in society: The development of
higher psychological processes. Cambridge: Harvard University
Press.
Wittrock, M. (Ed.). (1977). Learning and instruction. Berkeley:
McCutchan.

Second Language Acquisition
▶ Second Language Learning

2979

S

2980

S

Second Language Learning

Second Language Learning
ANGELIKA RIEDER-BÜNEMANN
English Department, University of Vienna,
Vienna, Austria

Synonyms
Foreign language learning; L2 acquisition; Second
language acquisition

Definition
Second language learning (SLL) is concerned with the
process and study of how people acquire a second
language, which is often referred to as L2 or target
language, as opposed to L1 (the native language).
Generally, the term second language in this context can
refer to any language (also a third or fourth language)
learned in addition to the native language. However,
second language learning would be contrasted with
a bilingual learning situation, in which a child acquires
two languages simultaneously (e.g., when the parents
speak two different languages). We only speak of second language acquisition if another language is
acquired after the first language.
The terms learning and acquisition are frequently
treated as synonyms in the literature. Some researchers,
however, distinguish between acquisition and learning,
stating that acquisition refers to the gradual subconscious development of language abilities by using the
language naturally (similar to the process of children
acquiring their first language), whereas learning refers
to the conscious process of accumulating knowledge of
a language, typically as the product of formal instruction (Krashen 1982). In the same vein, the term second
language learning is sometimes associated specifically
with the learning of languages through formal instruction, while the term second language acquisition is commonly used as a generic cover term to refer to learning
a target language in all situations, including the classroom. Another relevant term in this context is foreign
language learning, which some researchers use to differentiate between learning environments. In their
view, foreign language learning refers specifically to the
learning of a nonnative language in the environment of
one’s native language, while the term second language
acquisition to them refers to language learning in the

country where this language is spoken (which may or
may not take place within a classroom setting). Not all
researchers, however, agree with this distinction.

Theoretical Background
The study of second language learning/acquisition is a
fairly recent phenomenon, belonging to the second half
of the twentieth century. In the various periods of SLL
research, there were conflicting views about the nature
and development of the learner’s language knowledge,
depending on the prevailing linguistic theory of the
time (for a detailed discussion, see Ellis 2008).
Around the 1950s, the established theory was
▶ behaviorist learning theory, which saw language
learning, like any other learning, as a habit formation.
Most scholars claimed that knowing a language means
knowing a number of items and their potential
arrangements, and a second language learner was seen
as exhibiting imperfect knowledge of these items and
rules, with gaps in L2 knowledge naturally filled by the
grammar of the learner’s L1. Consequently, it was
predicted that the main cause of errors in the L2 were
interferences of the L1 (called negative transfer). With
a method called contrastive analysis, the linguists of the
time contrasted the L1 and L2 language systems and
looked for potential areas of difficulty due to differences between the language systems.
As research showed, however, not all learner errors
can be attributed to the phenomenon of transfer.
Rather, learners appear to be actively involved in
constructing their own rules, and behaviorists were
criticized for ignoring the creative side of language.
Subsequently, researchers’ attitude toward learner
errors changed, and contrastive analysis was replaced
by error analysis, where errors were not seen as examples of non-learning from the language norm any more
but were regarded as examples of learners’ active testing
of hypotheses.
In the 1960s and 1970s with Noam Chomsky, a new
research paradigm emerged, stressing the innate properties of the human mind that shape language learning,
which provided a ▶ mentalist theory of language learning. With his innateness hypothesis, Chomsky (1965)
claimed that humans possess a faculty for learning
languages referred to as Language Acquisition Device
(LAD), which is equipped with Universal Grammar
(UG), i.e., principles and parameters underlying all
grammars of all languages. Drawing directly from these

Second Language Learning

mentalist views of L1 acquisition, researchers then
claimed that the learner does not move from L1 to L2
in an incomplete language system, but that there exists
a learner language system called interlanguage (Selinker
1972), which is a dynamic system independent of L1
and L2 with rules of its own.
A question which has frequently been considered by
second language researchers is whether we still have
access to the LAD postulated by Chomsky in L2 acquisition. Basically, four possibilities have been proposed:
the “no-access position” argues that L2 learners acquire
the L2 grammar without any reference to UG but
through other faculties of the human mind, whereas
the “complete access position” claims that L2 learners
learn in exactly the same ways as L1 learners. In the
“partial access position,” then, L2 learners have access
to UG through what they know of the L1, starting with
L1 parameter settings rather than with the initial neutral state, while in the “dual access position,” L2 learners
have continued access to UG, but also have access to
a general problem-solving module, with both systems
competing in L2 acquisition (see Ellis 2008).
Since the 1980s, our understanding of the many
facets of second language learning has been greatly
enriched by perspectives from related disciplines. Two
theories which have become particularly popular
recently here are connectionism (viewing SLL from
a neuropsychological perspective) and Vygotskyan
learning theory (operating within a sociocultural
framework; for details and a survey of SLL theories
see Mitchell and Myles 2004).
▶ Connectionism, which is cognitively oriented, is
not a particularly new approach, but progress in computer technology has increased its popularity since
most studies involve hypothesis testing via computer
simulation. Like other cognitive theories, it sees second
language learning/knowledge as similar to other types
of learning/knowledge (in contrast to a linguistic theory, which claims that linguistic knowledge is unique).
Here, the brain is conceived as consisting of neural
networks, with (language) learning occurring on the
basis of associations, by which links between information nodes become strengthened or weakened. Thus,
connectionism differs from other theories insofar as it
does not believe that language acquisition involves rule
learning, but postulates associative processes instead.
Vygotskyan ▶ sociocultural theory (a general
theory of learning associated with Lev S. Vygotsky

S

2981

(1896–1934) applied to SLL by James Lantolf), like
connectionism, claims that language learning and general learning are subject to identical mechanisms. In
contrast to cognitive approaches, however, it argues
that social interaction plays a fundamental role in the
development of cognition. While sociocultural theory
rests on a variety of theoretical principles, its core
construct is that of mediation, i.e., that mental activity
in general and learning in particular is a process which
is mediated through mental tools (e.g., language) and
through social interaction; the theory thus rejects the
conventional distinction between social and psychological components of cognition.
Shifting the focus from SLL theories to the language
learner and learning results, a characteristic phenomenon frequently observed is that an L2 is rarely acquired
to a degree that is comparable to native-speaker proficiency in the L1, and that the results achieved vary
widely; in most cases the process of learning will stop
at some point, and learner language will fossilize at
a stage when it still contains certain features which do
not correspond to the target language. One goal of SLL
research is to identify the external and internal factors
determining why language learners acquire language in
the way they do and why different proficiency levels are
attained by different L2 learners.
Several external and internal factors appear to come
into play here:
● The social context forms a decisive external factor.

Even in first language acquisition, social factors
concerned with the relation of child and caretaker
will play a role. In second language learning, the social
context will have even more importance because it is
increasingly complex. On the one hand, it will play
a role in influencing the attitude of the language
learner toward the second language, while on the
other hand, it will also determine the social provision
of learning situations and opportunities.
● ▶ Motivation represents an internal factor which
may be influenced by the social context, since the
initial motivation to learn the L2 will be created
by the attitude of the learner toward the language
and/or the community. Two main types of motivation commonly distinguished here are integrative
motivation (i.e., language learning due to a desire to
integrate with the L2 community) and instrumental
motivation (i.e., language learning in order to

S

2982

S

Second Language Learning

receive a good grade, no particular interest in communication). In addition to this general motivation
to learn, motivation can also arise from success in
the learning process, and can thus be both a cause
and the result of success.
● Personal characteristics of the language learner will
also have an influence on the learning process and
outcome. Age will play a role, insofar as adults
appear to be less likely to achieve native-like pronunciation in the target language. Another potential influencing factor is the learner’s general
▶ personality, which manifests itself in factors like
the degree of anxiety he/she exhibits. The particular
learning style and the ▶ learning strategies the
learner favors and develops will also be an influence
in this context. Finally, apart from age and personality, learners will also vary with regard to their
▶ language aptitude, i.e., their natural disposition
for learning an L2.
● The amount and characteristics of the learning
opportunities form another external influencing factor which is crucial for the acquisition process.
Whatever the language learner brings to the task
in terms of aptitude, motivation, etc., the outcome
of language learning depends to a large extent on
the amount and kind of exposure to the target
language. Here we can firstly differentiate learning
opportunities in terms of time of exposure to the
language, and secondly distinguish between different kinds of learning, i.e., formal instruction vs.
natural acquisition. Both of these variables will
influence the route and rate of second language
learning.

Important Scientific Research and
Open Questions
Since the initial stages, SLL research has developed
into an independent and interdisciplinary field, with
investigations becoming increasingly diverse and sophisticated. Competing theories of second language learning
exhibit substantial differences since researchers
approach the question how second languages are learned
from a wide range of perspectives and research backgrounds, such as linguistics, psychology, or education.
Apart from differences related to the scientific field of
origin, theories differ for instance in the type of learning
they cover (e.g., naturalistic vs. instructed acquisition),
in research approach (e.g., inductive vs. deductive

reasoning), or in the theory’s core elements (e.g., the
role attributed to influencing factors). Although all theories have in common that they aim at explaining how
second language learning works, the field of SLL is still
struggling with the multitude of competing explanations
(Mitchell and Myles 2004).
Another open question in this context concerns the
relation between research on second language learning
on the one hand and language teaching on the other. To
this end, researchers are still debating in what way and
to what degree second language learning research actually impacts on second language teaching. While early
SLL research was frequently pedagogically motivated,
much of the current research is no longer directly
concerned with language teaching issues. Nevertheless,
we can assume that research results are at least indirectly relevant to language pedagogy, since second language learning is the process that language teaching
aims at facilitating. Various parallel methods exist for
teaching a foreign language, reaching from the early
grammar translation and audio-lingual method to the
communicative approach, which are usually conceptualized on the basis of a specific view of what language is,
how it is learned, and what conditions facilitate learning. Nevertheless, the problem of a gap between SLL
research and language pedagogy due to the former’s
interest in technical knowledge, in contrast to the latter’s concern with practical knowledge, is frequently
addressed in the literature, and attempts are made to
bridge this gap (Ellis 1997).

Cross-References
▶ Behaviorism and Behaviorist Learning Theories
▶ Beliefs About Language Learning
▶ Bilingualism and Learning
▶ Connectionist Theories of Learning
▶ Language Acquisition and Development
▶ Learner Characteristics
▶ Motivation and Learning: Modern Theories
▶ Personality and Learning
▶ Psycholinguistics and Learning
▶ Sociocultural Research on Learning

References
Chomsky, N. (1965). Aspects of the theory of syntax. Cambridge, MA:
MIT Press.
Ellis, R. (1997). SLA research and language teaching. Oxford: Oxford
University Press.

Selective Associations
Ellis, R. (2008). The study of second language acquisition (2nd ed.).
Oxford: Oxford University Press.
Krashen, S. (1982). Principles and practice in second language acquisition. London: Pergamon.
Mitchell, R., & Myles, F. (2004). Second language learning theories
(2nd ed.). London: Arnold.
Selinker, L. (1972). Interlanguage. International Review of Applied
Linguistics, 10, 209–241.

S

2983

Segmentation
▶ Statistical Learning in Perception

Selection of Learning Methods
▶ Meta-learning

Second Language Vocabulary
Acquisition
▶ Vocabulary Learning in a Second Language

Secondary Task Techniques
A measure of cognitive effort in which learners are
asked to perform a primary task, such as reading
a text passage or viewing a video tape, while simultaneously performing a secondary task, such as pressing
a button immediately after they hear a tone or see
a light flash. Also known as dual-task techniques, secondary task measures of cognitive effort assume that
there is a limit to the learner’s cognitive capacity, and
that when a great deal of cognitive capacity is consumed by the primary task, less capacity is available
to devote to the secondary task. The difference in
reaction time to the secondary task between a baseline
condition, where the learner only responds to the secondary task, and the experimental condition, where the
learner performs both the primary and secondary task,
is assumed to be reflective of the amount of cognitive
effort expended on the primary task.

Second-Order Factor
Dimensions of personality can be analyzed at different
levels of generalization. In ordinary factor analysis
observed (measured) variables are grouped into one
or more first-order factors corresponding to a first
level of generalization. Second-order factors are clusters of correlated first-order factors and correspond to
a higher level of generalization. There are also third-,
fourth-, and higher-level factors.

Selection Task
The Wason selection task is one of the most famous
tasks in the psychology of reasoning. It is a logic puzzle
in which subjects must assess the truth or falsity of
a general rule from specific instances. In particular,
the subjects face a problem of optimal data selection
because they must decide which of four cards (p, not-p,
q, or not-q) is likely to provide the most useful data to
test a conditional rule, if p then q. The “logical” solution
is to select the p and the not-q cards. Oaksford and
Chater (1996) argue that this solution presupposes
a falsification of a kind described by Popper in which
only data which can disconfirm hypotheses are relevant
for the selection task. However, due to the fact that they
implement rational analysis using a Bayesian approach,
Oaksford and Chater consider selection tasks as inductive rather than deductive reasoning tasks.

References
Oaksford, M., & Chater, N. (1996). Rational explanation of the
selection task. Psychological Review, 103, 381–391.

Selective Associations
STANLEY J. WEISS
Department of Psychology, American University,
Washington, DC, USA

Synonyms
Cue-to-consequence learning; Selective attention;
Stimulus–reinforcer interaction

S

2984

S

Selective Associations

Definition
Selective associations have been one of the more intensively studied examples of animals’ predispositions to
learn “more” about certain contingencies than others.
The differential effectiveness of a stimulus in controlling responding when it is paired with different reinforcers is the evidence for one kind of selective
association, a “stimulus–reinforcer interaction.”

Theoretical Background
The study of conditioning and learning is concerned
with the associative processes underlying learned modifications of behavior. In classical conditioning, these
modifications are produced by contingencies arranged
between a stimulus and an outcome. In operant learning, they are produced by contingencies arranged
between a response and an outcome. Traditional
behavior theory concentrated on the discovery of general laws of learning where stimuli, responses, and
reinforcers might be viewed generically, with different
instances within any class of events interchangeable.
This equipotentiality assumption of general-process
learning theory has been brought into question by
findings demonstrating what appear to be biological
constraints on learning and the associability of various
events and has continued to be a source of debate and
investigation (Domjan 1983).
One of the early examples of a biological constraint
on instrumental learning showed that racoons could
only be taught to drop tokens into a slot for positive
reinforcement (food) with extreme difficulty, if at all.
Innate patterns of behavior seem to supercede the
contingencies of reinforcement imposed by the trainer.
Does this mean that the empirical law of effect should
be abandoned? Not necessarily.
If more than the experimenter-defined aspects of
the situation are considered (token release for food),
what was called “misbehavior” can be explained in
terms of stimulus–reinforcer contingencies relevant to
obtaining and consuming food. The racoon’s behavior
is consistent with generally applicable, if more complex, general learning principles, but the “influence
. . .of. . .the conditioned motivational state in which
the instrumental conditioning was conducted and the
motivational state that was conditioned by presentations of the reinforcer [must be considered]” (Domjan
1983, p. 264). From this perspective, “misbehavior”
and other apparent biological constraints on learning

have strengthened general-process theory by encouraging it to deal functionally with the complete learning
situation. Generalizations thus developed are
concerned with more detailed features of a learning
situation, rather than the simplistic interchangeability
of cues, responses, and reinforcers (Domjan 1983). The
present entry will relate the phenomenon of “selective
associations” to general learning principles.

Important Scientific Research and
Open Questions
In Garcia and Koelling’s (1966) seminal selective association experiment, a compound conditioned stimulus
(CS) consisting of a taste and an audiovisual stimulus
was paired with a lithium chloride unconditioned
stimulus (US) in one group of rats and with an electric
shock US in another. In testing, the taste and the
audiovisual stimulus were presented separately. The
taste stimulus controlled a stronger conditioned
response (CR), suppression of drinking, when lithium
chloride was the US, while the audiovisual stimulus
controlled a stronger conditioned response when electric shock was the US. This first demonstration of
a selective association revealed that the choice of the
CS used in conditioning experiments was not as arbitrary as believed – and that experimental outcomes
might critically depend on the particular combinations
of CSs and USs used. More specifically, it showed that
when a set of stimuli have been given equal opportunity
to control a response, the US may determine which
stimulus is most effective. Such instances of selective
associations have also been called “stimulus–reinforcer
interactions” and “cue-to-consequence learning.”
Garcia and Koelling’s (1966) experiment employed
the conditioned taste aversion paradigm wherein nausea, caused by lithium chloride, following a novel flavor
might be considered intrinsically related. However,
selective associations have been demonstrated in more
traditional and often encountered learning paradigms
in pigeons by Foree and LoLordo (1973) and in rats by
Schindler and Weiss (1982). In a discrete-trials operant
procedure, Foree and LoLordo trained different groups
of food-deprived pigeons to depress a foot treadle in
the presence of a 5 s compound stimulus consisting of
a 440 Hz tone and a red houselight. For one group these
treadle presses avoided electric shock, and for the other
group they produced grain. When the pigeons were
effectively avoiding shock or earning food in the

Selective Associations

compound, the tone and light were presented separately as a test of control by the stimulus elements.
After appetitive training, the red light exerted strong
control over treadle pressing, while few responses were
emitted during the tone. By contrast, after training with
shock avoidance in the compound, the tone controlled
more treadle pressing than the light. This “stimulus–
reinforcer interaction”, that was systematically replicated in pigeons by Weiss et al. (1993b), is plotted in
the right frame of Fig. 1.
Schindler and Weiss (1982) expanded the generality
of this selective association phenomenon over species
to rats and to free-operant avoidance (FOA) contingencies. In a tone-plus-light (TL) compound, one
group of rats bar pressed to produce food on
a variable-interval (VI) schedule and another pressed
to postpone shock on an FOA schedule. In the absence
of tone and light (T L), responding had no scheduled
consequences [extinction (EXT)]. When the subjects
were responding at least ten times as rapidly in TL as in
T L, a test was administered to assess the degree
of control exerted by the tone and light elements of
the compound training stimulus. Consistent with the
findings of Foree and LoLordo, on a stimulus–element
test (a) the light gained almost exclusive control of
responding when bar pressing was maintained by

Schindler & Weiss (1982)

80
60

Tone

40

Light

20

100

Foree & LoLordo (1973)

80
Tone
60
40

Light

20
0

0
Food

2985

food, and (b) the tone gained considerable control
under the shock-avoidance schedule (see left frame of
Fig. 1). A similar selective association was revealed over
these two species even though pigeons are clearly more
visual animals than rats.
Selective associations appear to be a basic attentional process that generalizes over species and experimental conditions. Unfortunately, in these studies
reporting stimulus–reinforer interactions in traditional
operant (Foree and LoLordo 1973; Schindler and Weiss
1982) and classical [see Shapiro, Jacobs and LoLordo
(1980) cited in Weiss et al. (1993a)] situations, the
physical nature of the stimulus responsible for reinforcement (food and shock) was covarying with what
might be considered the conditioned hedonic value
that would have been conditioned to the TL compound. Therefore, whether these selective associations
were determined by the physical nature of the reinforcer or the hedonic state they produced remained to
be determined.
In both, the Foree and LoLordo and the Weiss and
Schindler experiments, when the subjects were working
for food during TL they were in a hedonically positive
condition that would be preferred over the nonreinforcement associated with the absence of these
stimuli. In comparison, TL would have become

% Element Responses

% Element Responses

100

S

Shock

Food

Shock

Group
Selective Associations. Fig. 1 Left frame: The percent of total element responses emitted in tone (filled circles) and in
light (open circles) by rats that had been trained by Schindler and Weiss (1982) in tone-plus-light to earn food on an
intermittent schedule or to avoid shock. Right frame: Same measure for the pigeons of Foree and LoLordo (1973, adapted
from their Fig. 3). Responding came primarily under the control of the light when food was the reinforcer and tone when
shock avoidance was the reinforcer. Because different classes of reinforcer, appetitive or aversive, maintained responding
in both groups of each study, these profiles have been termed “stimulus–reinforcer interactions”

S

2986

S

Selective Associations

a hedonically negative (nonpreferred) condition, compared to shock-free T L, when TL was associated with
shock avoidance. This means that in these studies
reporting “stimulus–reinforcer interactions” food and
shock were confounded with relative hedonic value
conditioned to the TL compound stimulus that
would be revealed by schedule-component preference.
Following this train of reasoning, one could contend
that when TL signals a preferred condition, the light
will be selectively attended to and when it signals
a nonpreferred condition the tone will be attended
to – what might be termed a “stimulus–preference
interaction” In this analysis, attention is related to
a contingency-generated psychological state – rather
than food- or shock-related situations in particular
that are just one means of creating these states.
Weiss et al. (1993a) eliminated this covariation by
creating TL compounds of positive and negative
hedonic value (i.e., positive stimuli the organism
is drawn toward or negative stimuli it is repelled
from, respectively) using only appetitive contingencies.
Konorki’s appetitive–aversive interaction theory of
motivation, that proposed the affective equivalence of
excitors and inhibitors of contrasted affective value
(appetitive vs. aversive) was their point of departure.
According to Konorski, (1) a stimulus signaling the
absence of food has many negative properties in common with aversive stimuli such as electric shock, and
(2) a stimulus differentially associated with the absence
of shock has positive (attractive) properties that parallel those produced by appetitive reinforcement. What
one is, therefore, concerned with here is not the physical character (e.g., food or shock) of the event ultimately responsible for the conditioned association, but
rather, functionally, with whether the organism is
drawn toward (positive) or repelled from (negative)
the situation associated with the TL compound. Operationally, the organism will work to produce the former
and terminate the latter condition.
In the Weiss et al. (1993a) experiment, bar pressing
in TL was ultimately maintained by food in two groups
of rats but in a manner that broke the covariation
described above between the physical nature of the
reinforcer and the conditioned affective value of the
TL compound. To accomplish this, contingencies were
arranged such that, relative to T L, TL became hedonically positive (a conditioned appetitive excitor) for
one group and hedonically negative (a conditioned

appetitive inhibitor) for the other. This latter group
was meant to be hedonically comparable to the groups
reported by Foree and LoLordo (1973) and Schindler
and Weiss (1982) where shock-avoidance-associated
TL was a conditioned aversive excitor.
To accomplish this, both Weiss et al. (1993a) groups
responded in TL and ceased responding in T L, just as
the food and shock groups presented in Fig. 1 did.
However, one group (TL+) earned all their food reinforcers on a VI schedule in TL for bar pressing while T L
was associated with extinction (EXT), like the food
group in the Schindler and Weiss (1982) study. In
contrast, their TL group received reinforcers only in
T L, where responding ceased because it delayed food
delivery. But, they were on a chain schedule wherein
bar pressing in TL produced, on a VI schedule, the T L
component wherein all food was received. This was an
implicit measure that TL was the nonpreferred component because the rats were working therein to produce
T L. After the subjects of the TL+and TL groups were
brought under stimulus control, with responding in TL
indistinguishable over groups (see Fig. 1 in Weiss et al.
(1993a)), they received a stimulus–element test. On
this test, tone and light were presented separately for
the first time to determine which modality controlled
primarily responding in TL (auditory or visual).
All rats in the TL group emitted more responses
in the tone than the light test element, while no rat in
the TL+group did this. On average, rats in the TL+
group emitted 65% of their element test responses in
the presence of the light while the TL group emitted
a significantly lower 40% of their element responses
in the light. The distribution of element responses
for the TL group, for whom no food was received
in TL, resembles that of animals that avoided shock
in TL in the Schindler and Weiss (1982) study. This
can be appreciated when Fig. 2 is compared to the left
frame of Fig. 1.
The stimulus–element test in Fig. 2, wherein TL+
and TL were generated solely with food-related contingencies, produced a selective-association interaction
profile similar to those reported by Schindler and Weiss
(1982) and Foree and LoLordo, where TL+was food
related and TL was shock related. This demonstrated
that the dynamics underlying selective associations
might be better understood in terms of the equivalence
of excitors and inhibitors of contrasted affective value –
an explanation confirmed by Weiss et al. (1993b) who

Selective Associations

Weiss, Panlilio, & Schindler (1993a)

% Element Responses

70
Tone
60
50

40

Light

30
FoodRelated
TL+

FoodRelated
TL−

Selective Associations. Fig. 2 The percent of element
responses to the tone (filled circles) and the light (open
circles) by the TL+group and by the TL group on
a stimulus element test. In training, see text, a food
contingency maintained responding in TL in both groups.
However, all food was received in TL by the TL+group
while no food was received in TL by the TL group. When
the underlying food reinforcement contingency created
a TL+(preferred), responding was primarily controlled by
the light component of the TL compound. In contrast,
when a TL (nonpreferred) was created the tone
component gained control [Adapted from Fig. 3 of Weiss
et al. (1993a)]

reported selective associations when responding in
both TL+and TL were maintained by shock-related
contingencies.
The relative hedonic value of TL (as measured by
component preference or observing behavior) appear
to be a determinant of stimulus control in these
situations. This should encourage us to look for
biological constraints (influences) on learning at
a higher level of processing than heretofore since
selective associations appear to be a product of
a conditioned psychological hedonic state. If this
turns out to be the case, it could mean that the
variables involved in fundamental psychological processes such as choice behavior in general, conditioned
preference, and appetitive–aversive interactions might
also contribute to selective associations.
This strongly suggests that selective associations be
viewed from the broader perspective of conditioned
hedonic states. That selective associations can be

S

2987

produced by manipulating contingencies involving
solely food or solely shock should increase our appreciation of the power of contingencies themselves in
fostering comparisons between schedule components
that influence stimulus selectivity. Food and shock
appear to have been the most straightforward and
direct ways of producing these positive and negative,
respectively, hedonic states in early selective association
research. That seemed to divert attention from the
fundamental hedonic processes responsible for the
phenomenon in traditional conditioning and learning
situations. As the many cross-references listed below
reveal, this integrative connection relates selective associations to fundamental processes in conditioning and
learning.

Cross-References
▶ Associative Learning
▶ Attention and Implicit Learning
▶ Avoidance Learning
▶ Biological and Evolutionary Constraints of Learning
▶ Comparative Psychology and Ethology
▶ Contingency in Learning
▶ Discriminative Learning
▶ Emotion-Based Learning
▶ Motivational Variables in Learning
▶ Selective Learning

References
Domjan, M. (1983). Biological constraints on instrumental and classical conditioning: Implications for general process theory. In
G. Bower (Ed.), The psychology of learning and motivation (Vol.
17, pp. 215–277). New York: Academic Press.
Foree, D. D., & LoLordo, V. M. (1973). Attention in the pigeon:
Differential effects of food-getting versus shock-avoidance procedures. Journal of Comparative and Physiological Psychology, 85,
551–558.
Garcia, J., & Koelling, R. A. (1966). Relation of cue to consequence in
avoidance learning. Psychonomic Science, 4, 123–124.
Schindler, C. W., & Weiss, S. J. (1982). The influence of positive and
negative reinforcement on selective attention in the rat. Learning
and Motivation, 13, 304–323.
Weiss, S. J., Panlilio, L. V., & Schindler, C. W. (1993a). Selective
associations produced solely with appetitive contingencies: The
stimulus-reinforcer interaction revisited. Journal of the Experimental Analysis of Behavior, 59, 309–322.
Weiss, S. J., Panlilio, L. V., & Schindler, C. W. (1993b). Singleincentive selective associations produced solely as a function of
compound-stimulus conditioning context. Journal of Experimental Psychology: Animal Behavior Processes, 19, 284–294.

S

2988

S

Selective Attention

Selective Attention
Selective attention is the ability to learn to actively
process and/or ignore a subset of the available information during the course of information processing.

Cross-References
▶ Selective Associations

Selective Attention in Social
Learning of Vervet Monkeys
ERICA VAN DE WAAL
Centre for Social Learning and Cognitive Evolution,
and Scottish Primate Research Group, School of
Psychology, University of St-Andrews, St-Andrews,
Scotland, UK

Synonyms
BIOL (Bonding and Identification-Based Observational Learning); Observation-based Learning Rather
than Experienced Individual Learning; Role Model

Definition
Selective attention in social learning: social learning is
defined as learning by observing and copying other
group members. Selective attention is defined as observing specific models rather than any group members.

Theoretical Background
Efficient ▶ social learning plays a major role in human
life as it provides the basis for traditions and culture
(Plotkin 2007). Consequently, studying the roots of
culture in other animals has been a key research topic
for decades (Whiten 2009). Laboratory experiments
demonstrated that a variety of vertebrate species
might be able to learn socially, and many field studies
documented either the spread of innovations are
important differences between populations that seem
to be based on traditions. However, experiments are
needed to test rules for social learning. Theoretical
studies on social learning rules suggest that individuals
should be selective when deciding both when to learn
socially and who to choose as a model (Boyd and

Richerson 1985). The hypothesis that individuals
should copy group members that are currently successful at a given task has repeatedly received experimental
support (Kendal et al. 2009). In contrast to a scenario
where individuals constantly monitor each other and
compare relative gains, the “social model hypothesis,”
also known as BIOL (bonding and identification-based
observational learning), predicts that primates living in
structured social groups are most likely to learn from
social models that have general qualities such as high
relatedness (the mother), knowledge, high age, and/or
high rank (de Waal 2001). In contrast, actions of young
or less familiar members may be largely ignored, even
in situations where such individuals have found the
best solution.

Important Scientific Research and
Open Questions
The BIOL hypothesis has been tested with a social
learning experiment on six wild vervet monkey groups,
where models were either a dominant female or
a dominant male. In this species, females remain in
their group of birth, while males migrate once sexually
mature. Thus, social links to females should be stronger
and more important than social links to males, genetic
relatedness to females should be higher than to males,
and females should on average be more knowledgeable
about food sources in the territory. Together, these
factors lead to the prediction that individuals should
be more likely to copy females than males. In the
experiment, “artificial fruit” boxes that had doors on
opposite, differently colored ends for access to food
were presented to the monkeys. One option was
blocked during the demonstration phase, creating consistent demonstrations of one possible solution by
a monopolizing dominant individual. Following 25
correct demonstrations, all other group members
were exposed to the box with both doors functional.
We observed a significantly higher participation rate
(“stimulus enhancement,” def: Hoppitt and Laland
2008) in groups with female models compared to
groups with male models, and only in the former we
found a significant preference of subjects to manipulate
the same-door manipulation as the model (“local
enhancement,” def: Hoppitt and Laland 2008). These
findings implied a different copying rate of group
members, depending on the identity of the model.
Analyses of the context of the demonstrations revealed

Selective Learning

that the bias is not due to lack of tolerance or aggressiveness of male models. However, when observing
gazes toward the models, significant results were
found providing evidences for selective attention
toward female models. These results demonstrated
the favored role of dominant females as a source for
“directed” social learning in a species with female
philopatry. Females should often be better than immigrants as sources for social learning, at least in the
context of foraging. Based on our findings, we hypothesize that in species in which members of one sex form
the core of stable groups, the migration of individuals
of the other sex leads to proper exchange of genetic
adaptations but much less to the exchange of socially
acquired adaptive information. Our findings imply
that migration does not necessarily lead to an exchange
of socially acquired information within populations,
potentially causing highly localized traditions.
These results suggest the importance of the
philopatric sex for the transmission of knowledge.
A comparison between species with different migration
patterns would be very instructive. The study described
above did not address other potential key variables
like rank, age, or relatedness. Furthermore, one should
test social learning patterns in contexts other than
feeding, like social behaviors or communication. Such
tests would allow answering the question whether
philopatry matters across contexts or whether role
models are chosen according to their abilities in each
specific domain. Combined analyses would yield a
clearer picture concerning the scale of traditions. Typically, researchers have assumed that differences mainly
occur between populations, but the results on vervet
monkeys imply the prevalence of distinctive traditions
within populations.

Cross-References
▶ Animal Learning
▶ Field Research on Learning
▶ Imitative Learning in Humans and Animals
▶ Learning from Animals
▶ Social Cognition in Animals

References
Boyd, R., & Richerson, P. J. (1985). Culture and the evolutionary
process. Chicago: University of Chicago Press.
de Waal, F. B. M. (2001). The ape and the sushi master: Cultural
reflections of a primatologist. New York: Basic Books.

S

2989

Hoppitt, W., & Laland, K. N. (2008). Social processes influencing
learning in animals: A review of the evidence. Advances in the
Study of Behavior, 38, 105–165.
Kendal, J. R., Rendell, L., Pike, T. W., & Laland, K. N. (2009). Ninespined sticklebacks deploy a hill-climbing social learning strategy. Behavioral Ecology, 20, 238–244.
Plotkin, H. (2007). The power of culture. In R. I. M. Dunbar &
L. Barrett (Eds.), The Oxford handbook of evolutionary psychology
(pp. 11–19). Oxford: Oxford University Press.
Whiten, A. (2009). The identification and differentiation of culture in
chimpanzees and other animals: From natural history to diffusion experiments. In K. N. Laland & B. G. Galef Jr. (Eds.), The
question of animal culture (pp. 99–124). Cambridge: Harvard
University Press.

Selective Learning
GERRI HANTEN
Cognitive Neuroscience Laboratory, Dept. of Physical
Medicine and Rehabilitation, Baylor College of
Medicine, Houston, TX, USA

Synonyms
Selective remembering; Strategic learning; Strategic
memory; Value-directed memory

Definition
Consider this situation: you are on the way to the
grocery store, but you have forgotten your list. The
complete list of 18 items included butter, carrots, celery, cheese, chocolate, eggs, flour, garlic, juice, lemons,
lettuce, milk, peppers, potatoes, salt, soup, walnuts,
and yogurt. Some of these are everyday items that
need to be restocked soon (low immediate value), and
others are items you need to bake a cake for a friend’s
birthday this evening (high immediate value). Which
ones will you remember, how, and why? The answers to
these questions are explored in the study of selective
learning. Thus, selective learning is the value-based
strategic memory for a subset of items most important
to achieving a current goal.

Theoretical Background
Selective learning, the ability to select items from among
other items to learn based on value, requires the appropriate allocation of cognitive resources, allowing for

S

2990

S

Selective Learning

preferential focusing on those items of greater value
while taking into consideration the limitations of memory span. Although the degree to which memorial processes can be wilfully controlled has been studied in
adults since the early 1970s, these studies generally
have investigated the ability of individuals to ignore
irrelevant (or no-value) items in favor of relevant
items, as in the directed forgetting paradigms developed
by Robert Bjork and his colleagues, and other studies of
the effects of incentive on memory.
Selective learning paradigms address the bias to
remember important items in preference to items of
lesser importance when all items have value. Typically
two measures are taken: total words recalled, for
an estimation of memory span, and a selective efficiency score. The selective efficiency score is calculated
by the formula (actual score – chance score)/(ideal
score – chance score) where actual score = sum of
the point values of items recalled; chance score = the
score predicted in the absence of selection given the
number of items recalled (i.e., if half the words
recalled were high value and the other half were low
value); ideal score = the maximum score that could
be achieved given the number of items recalled. Thus,
the index of selective efficiency is based on the maximum score possible for an individual given his/her
total words recalled. The range of the index extends
from 1 to 1, with a score of 0 indicating chance
performance (that is, equal numbers of high and
low items).
Selective learning is a skill that is highly applicable
to everyday memory function in both adults and children. Given the limits of human memory, the ability
to respond appropriately to value with regard to what
one recalls is highly adaptive. Evidence of an innate
ability to select information for learning based upon
the inherent importance or relevance to personal objectives is reflected in behavioral observations of conditioned responses in animals and humans, which
suggests the ability to select personally relevant information for learning may be instinctively driven. Thus,
it may be that humans are naturally inclined toward
strategic learning of important information.
Some key elements involved in selective learning are:
1. Memory span, typically immediate memory for
word lists although studies on selective learning
of items within connected text have been done.

The relation of selective efficiency to memory
span is not straightforward, and several studies
have reported dissociations between memory span
and selective efficiency in normal and clinical
populations. (Castel et al. 2007; Hanten et al. 2007).
2. Inhibitory processing, or the ability to suppress
retrieval of less relevant items. Studies examining
the ability to select information for learning based
upon extrinsically assigned value (i.e., holds little to
no personal significance) have emphasized the role
of controlled cognitive processes (Hanten et al.
2002). A number of studies show that cognitive
control develops throughout childhood and into
early adulthood and is closely tied to frontal lobe
development. The frontal lobes show a protracted
development period that extends from childhood
into late adolescence or early adulthood. Hence,
if cognitive control is a primary factor, selective
learning proficiency may rely on frontal lobe
development.
3. Incentive, or response to reward. Although motivation is a very likely contributor to selective efficiency, under some circumstances it fails. For
example, when the value (including monetary
reward) of the items is designated after presentation
of the items, selective efficiency is greatly diminished, if not eliminated altogether (Hanten et al.
2002), thus suggesting some limits of cognitive
control on incentive.
4. Metacognition. Because instructions for selective
learning tasks typically emphasize that the goal is
to earn a high number of points, but do not provide
explicit directions for achieving that goal, selfawareness of one’s memorial limitations and the
ability to implement an effective strategy is an
advantage. If a person can remember all items,
low and high value, he or she need not select to
remember in order to make a high score. However,
a person who is aware that there are more items
than can be remembered may devise a strategy to
remember only high value items, thus allowing
a high degree of selectivity.

Important Scientific Research and
Open Questions
The field of selective learning is understudied. A few
studies relating to this topic (under various rubrics of

Selective Learning

selective learning, strategic learning, selective remembering, and value-directed learning or memory) have
been largely limited to clinical or developmental studies, although there are a few studies exploring the
mechanisms of selective learning.
In non-clinical populations, studies have been
performed that address selective learning in children
and in adults. Castel and colleagues (e.g., Castel et al.
2007) investigated the differences in selective learning
in older and younger adults, found both groups (older
and younger adults) were equally good at recalling the
range of points assigned to high-value words, but
younger adults outperformed older adults when
recalling specific values. Both groups exhibited poorer
recall of negative value information. This was
interpreted to suggest that older adults rely more on
gist-based operations, whereas younger adults remember specific numeric value information. In another
study, Castel and his colleagues found that older adults
recalled fewer words overall than younger adults, but
showed greater efficiency in recalling high value words.
At the other end of the developmental spectrum,
Hanten et al. (2007) explored selective learning using
auditory word lists in typically developing children,
and found memory span to have a different developmental trajectory than selective efficiency, with the two
measures not significantly related to one another. In
a different study of selective learning in children for
visually presented stimuli, Hanten and her colleagues
reported that the mode of presentation was an important factor: there is an advantage for the simultaneous
presentation of all words in one list rather than
presenting each word singly and sequentially, and an
advantage for selecting on the basis of physical characteristics of the stimuli (type of print used for the words)
rather than the semantic characteristics of the stimulus
set (i.e., natural category membership).
Efficiency of some types of selective learning
appears to depend on the relative incentive associated
with stimulus items, and to increase with age. Studies
of the trade-off between the ability to select for specific
items and the total number of words recalled indicate
that the cost (in terms of the number of words recalled)
associated with more efficient selection appears to
decrease with age (Miller et al. 1991).
Selective learning paradigms have been found useful
in examining areas of cognitive impairment and relative
preservation in clinical populations. In a number of

S

studies, Hanten and colleagues have demonstrated that
children with traumatic brain injury are impaired in
selective learning efficiency for word lists relative
to typically developing children in verbal selective learning in both auditory and visual modalities, and for
connected text, with total word recall relatively preserved
(Hanten et al. 2002, 2004). Castel and colleagues have
revealed that children with ADHD, especially Combined Type ADHD, have shown much the same pattern, with no between-group differences between
children with ADHD and controls on total word recall,
but deficits in selective efficiency in children ADHD
(Castel et al. 2011). Castel and colleagues have also
studied selective learning in patients with Alzheimer’s
Disease, who were found to be impaired in selective
efficiency. Other clinical populations with reported deficits in selective learning are children with spina bifida,
who demonstrate impaired word recall relative to typically developing children, but relatively preserved selective efficiency. However, children with spina bifida did
report the use of an inefficient strategy to remember all
the words as compared to the control group, who indicated they tried to remember the high value words.
Selective learning is a skill important to everyday
function, and especially salient in children. The field of
study is in its infancy and would benefit from more
studies of the perceptual and conceptual parameters
that affect selective efficiency, the nature of strategy
employed, development of aspects of selective learning,
as well as studies of cognition in clinical populations.
Further, to date, there is a paucity (or complete
absence) of studies illuminating brain-behavior relationships in selective learning, or in the role of selective
learning in social-cognitive domains.

Cross-References
▶ Attentional Modulation of Spread of Activation
▶ Capacity Limitations of Memory and Learning
▶ Changes in Memory and Learning Across the
Lifespan
▶ Conditional Learning
▶ Directed Forgetting
▶ Inhibition and Learning
▶ Memory Dynamics
▶ Metacognition and Learning
▶ Motivation, Volition and Performance
▶ Short-term Memory
▶ Strategic Learning

2991

S

2992

S

Selective Remembering

References
Castel, A. D., Farb, N. A. S., & Craik, F. I. M. (2007). Memory for
general and specific value information in younger and older
adults: Measuring the limits of strategic control. Memory and
Cognition, 35, 689–700.
Castel, A. D., Lee, S. S., Humphreys, K. L., & Moore, A. N. (2011).
Memory capacity, selective control, and value-directed remembering in children with and without attention–deficit/hyperactivity disorder (ADHD). Neuropsychology, 25, 15–24.
Hanten, G., Zhang, L., & Levin, H. S. (2002). Selective learning in
children after traumatic brain injury: A preliminary study. Child
Neuropsychology, 8, 107–120.
Hanten, G., Chapman, S. B., Gamino, J. F., Zhang, L., Benton, S. B.,
Stallings-Roberson, G., Hunter, J. V., & Levin, H. S. (2004).
Verbal selective learning after traumatic brain injury in children.
Annals of Neurology, 56, 847–853.
Hanten, G., Li, X., Chapman, S. B., Swank, P., Benton, S. B., Roberson,
G., & Levin, H. S. (2007). Development of verbal selective learning. Developmental Neuropsychology, 32, 585–596.
Miller, P. H., Seier, W. L., Probert, J. S., & Aloise, P. A. (1991). Age
differences in the capacity demands of a strategy among spontaneously strategic children. Journal of Experimental Child Psychology, 52, 149–165.

Selective Remembering
▶ Selective Learning

Selective Trust
▶ Children’s Critical Assessment of the Reliability
of Others

Self Concept
Broadly based individual beliefs about self in physical,
social, and academic domains.

Self Efficacy
One’s perceptions of one’s own ability to succeed on
valued tasks.

Self Handicapping
Purposely placing obstacles that prevent optimal performance, or from being responsible for failure typically for the purpose of protecting one’s self-esteem.

Self Processes and Learning
▶ Identity and Learning

Self-Adaptation of Cultural
Learning Parameters
DARA CURRAN
Computer Science Department, Cork Constraint
Computation Centre (4C), University College Cork,
Cork, Ireland

Synonyms
Evolutionary computation

Definition
Self-adaptation of parameters is a technique employed
by evolutionary computation researchers to evolve
populations to solve a particular problem while, at
the same time, adjusting its own parameter set to
improve performance and fitness.
Self-adaptation is related to coevolution in the sense
that the values for chosen parameters are evolved at the
same time as the population itself. The adaptation of
parameters for cultural learning is focused on evolving
good parameter values for the purposes of producing
effective cultural learning within a population.

Theoretical Background
Some research has been performed on the selfadaptation of parameters employed in evolutionary
computation. Angeline (1995) divided such adaptation
into three groups: population level, individual level,
and component level.
Population-level adaptation dynamically adjusts a
parameter that is subsequently used globally across
the population, such as a global crossover value

Self-Determination and Learning

S

(Spears 1995). Individual level-adaptation reduces the
impact of a parameter to an individual level, such as
adapting crossover points (Rosca 1994). Componentlevel adaptation dynamically changes the way in which
individual components of an individual will be altered
independent of one another (Back and Schwefel 1993).
Researchers have applied self-adaptive parameters
at varying levels to evolutionary programming (Fogel
et al. 1991) and to genetic algorithms (Back 1992;
Hinterding 1995; Smith and Fogarty 1996).

Smith, J., & Fogarty, T. (1996). Self-adaptation of mutation rates in
a steady state genetic algorithm. In IEEE international conference
on evolutionary computation (pp. 318–323). New York.
Spears, W. M. (1995). Adapting crossover in evolutionary algorithms.
In J. R. McDonnell, R. G. Reynolds, & D. B. Fogel (Eds.), Proceedings of the fourth annual conference on evolutionary programming (pp. 367–384). Cambridge, MA: MIT Press.

Important Scientific Research and
Open Questions

▶ Meta-learning

The model proposed by Curran et al. shows that the
addition of cultural learning to evolutionary learning
has beneficial effects in terms of fitness, as seen in previous work. Furthermore, results obtained from the examination of the self-adaptive parameters illustrate the
relative quality of teachers (and, therefore, the impact
of cultural learning) at different stages in the experiment.
The fact that the values for these parameters changes
significantly throughout the experiment shows the usefulness of self-adaptive cultural parameters for both
performance and analysis of cultural learning.

2993

Self-Adaptive Systems

Self-Awareness
▶ Self-Reflecting Methods of Learning Research

Self-Centered Learning
▶ Narcissistic Learning

Cross-References
▶ Computational Models of Human Learning
▶ Cultural Learning
▶ Metalearning
▶ Population Learning

Self-Control
▶ Self-regulation and Motivation Strategies

References
Angeline, P. J. (1995). Adaptive and self-adaptive evolutionary computations. In M. Palaniswami & Y. Attikiouzel (Eds.), Computational intelligence: A dynamic systems perspective (pp. 152–163).
Piscataway: IEEE Press.
Back, T. (1992). Self-adaption in genetic algorithms. In Proceedings of
the first european conference on artificial life (pp. 263–271). Cambridge, 1992.
Back, T., & Schwefel, H. P. (1993). An overview of evolutionary
algorithms for parameter optimization. Evolutionary Computation, 1(1), 1–23.
Fogel, D., Fogel, L., & Atmar, J. (1991). Meta-evolutionary programming. In Proceedings of the conference on signals, systems, and
computers (pp. 540–545). San Jose.
Hinterding, R. (1995). Gaussian mutation and self-adaption in
numeric genetic algorithms. In IEEE international conference on
evolutionary computation (pp. 384–389). New York, 1995.
Rosca, J. P. (1994). Hierarchical self-organization in genetic programming. In Proceedings of the eleventh international conference on
machine learning (pp. 251–258). Los Altos.

Self-Determination and
Learning
ARIANE S. WILLEMS, DORIS LEWALTER
TUM School of Education, Fachgebiet für
Gymnasialpädagogik, Technical University of Munich,
Munich, Germany

Synonyms
Autonomy; Self-direction; Self-regulation

Definition
The concept of self-determination is used in various
theories of learning and instruction. Over the past

S

2994

S

Self-Determination and Learning

decades, Self-Determination Theory (SDT) (Ryan and
Deci 2008) has evolved as an influential framework,
especially in education and learning. According to
SDT, students who feel that their learning is self-determined/autonomous experience their studying as a selfchosen, volitional act that reflects their individual goals
and values. Experiencing autonomy/self-determination
in learning involves a sense of unpressured willingness. It
is defined as full and nonconflictive endorsement to
engage and proceed in learning. In that sense, autonomy
concerns the extent to which students authentically and
genuinely concur with the (external) forces underlying
specific learning situations. Self-determination/autonomy encompasses ideas such as self-direction, self-governance, self-regulation, coherence with the self, and the
experience of freedom and volition. It excludes concepts
such as individualism, independence, power, dominance,
or selfishness (Assor and Kaplan 2001).

Theoretical Background
A key concept in theories about processes of teaching
and learning is motivation. Students often struggle to
mobilize their efforts, to find energy to start and persist
at learning tasks, and to direct their endeavors toward
the achievement of learning goals. Complementarily,
teachers are concerned with fostering their students’
learning processes and seek ways to facilitate their
motivation. This challenge is particularly evident as
academic learning is often driven by external factors
such as grades, evaluations, and imposed learning
goals. Schools, as one predominantly outcome-oriented
context of formal learning, have always imposed structures and controls. But with the strong emphasis on
educational accountability and high-stakes testing, the
intensity and consequences of such controls have
recently increased (Reeve 2002; Ryan and Deci 2009).
SDT represents a framework for analyzing human
motivation and personality. At its core lies the investigation of the interplay between extrinsic forces
influencing human behavior and intrinsic motives
and needs inherent in human nature.
SDT formulates a meta-theory for framing studies
of motivational development in different contexts,
including the fields of teaching and learning. The theory follows an integrative organismic and dialectical
approach (Ryan and Deci 2008): SDT emanates from
the assumption that humans are active organisms, with
evolved tendencies toward learning, mastering

challenges, and integrating new experiences into a
coherent sense of self. Additionally, SDT postulates
that these prevalent natural developmental tendencies
do not operate automatically. Instead, they require
external supports. Here, SDT focuses on the analysis of
contextual factors that either support or thwart the
natural tendencies toward active engagement and learning. Regarding institutionalized academic learning processes, the dialectic between students as active organisms
and their social context is the basis for SDT’s predictions
about the students’ behavior, experience, motivation,
and development in learning (Reeve 2002).
Based on this framework, SDT comprises four
mini-theories, each of which was developed to explain
different motivationally based phenomena. For academic learning, the Cognitive Evaluation Theory
(CET) and the Organismic Integration Theory (OIT)
are vital to understanding the quality and outcomes
of learning processes (Ryan and Deci 2008).
At the center of CET lies the concept of intrinsic
motivation which is understood as the prototype of
self-determined regulation. Intrinsically motivated
behaviors stem from the inherent satisfaction of the
behaviors per se rather than from contingencies that
are operationally separable from the activity itself
(Ryan and Deci 2008). The aim of CET is to identify
situational factors of learning environments that
account for the variability in students’ intrinsic motivation. Here, the basic psychological needs for autonomy,
competence, and relatedness are important. CET states
that the more a learning environment facilitates the
students’ basic needs, the more likely intrinsic motivation develops. This postulate is seen as universal across
age, gender, culture, context, and background.
The need for autonomy refers to being the origin
and source of one’s own behavior. It is defined as feeling
volitional and congruent with respect to what one does.
Experiencing autonomy also means feeling selfgovernance concerning the initiation and direction of
behaviors (Assor and Kaplan 2001). Historically, this
need goes back to Richard de Charms’ concept of
personal causation and is also related to Julian B. Rotter’s concept of an internal locus of control. The need for
competence refers to the need to feel effective in the
interaction with the social environment and to experience opportunities to exercise, expand, and express
one’s capacities. It is derived from Robert White’s concept of effectance motivation. Finally, relatedness refers

Self-Determination and Learning

to the need to feel connected with others, to care for
and to be cared for by others, and to have a sense of
belongingness and respect within one’s own community (Ryan and Deci 2008). It is rooted in Abraham
Maslow’s need of relatedness and David McClelland’s
need for affiliation.
Whereas modern theories of motivation (expectancy-value theories, social cognitive theories) treat
motivation as unitary concept that typically varies in
amount but not in kind, SDT focuses on different types
of motivation, all of which differ in quality rather than
quantity. Based on a classic dichotomy of intrinsic vs.
extrinsic motivation, SDT further distinguishes
between different kinds of extrinsic motivation. Instead
of arguing that motivation per se is a key factor that
influences learning, OIT asserts that the quality of
motivation is decisive for the quality and outcomes of
learning processes. OIT proposes a taxonomy of four
different types of extrinsic motivation that vary in their
degree of relative autonomy. The description of properties, determinants, consequences, and correlates of
these types of extrinsic motivation are the major elements of OIT (Ryan and Deci 2009).
Extrinsic motivation by definition comprises
instrumentality: Externally motivated behavior aims
toward outcomes which are separable from the behavior itself (Ryan and Deci 2008). Yet there are distinct
forms of instrumentality. In OIT, these types of instrumentality are treated as underlying attitudes and goals
that give rise to an action and regulate behavior. These
types of regulation are eventually covered by the different qualities of extrinsic motivation. Following
OIT’s argumentation, such subtypes of extrinsic motivation are arranged along a continuum of relative selfdetermination reflecting a process of internalization.
This process is defined as a sequence of assimilating
non-intrinsically motivated behaviors and values from
the social environment into one’s own value and regulatory system: The more internalized the extrinsic motivation, the more autonomous the person will be and the
more a certain regulation becomes part of the individual
self. In line with findings from CET, OIT investigates the
factors that conduce toward people either, resisting,
partially adopting, or deeply internalizing values and
goals. Much like intrinsic motivation, internalization is
expected to occur spontaneously within certain learning
environments: When students feel socially accepted and
related to significant others, they are inclined to

S

2995

internalize the values of those around them. When the
support for the personal need for relatedness is further
combined with supports for autonomy and competence, the process of internalization goes beyond mere
value-assimilation and finally leads to a process of integration. Here, students more fully transform and incorporate the external regulation into their own regulatory
system (Ryan and Deci 2009).
The different types of extrinsic motivation are
defined as follows (Ryan and Deci 2008):
1. External regulation: This form of extrinsic motivation is the least autonomous one. It entails being
motivated to obtain rewards or avoid punishments.
Behaviors based on external regulations are continued only as long as external demands are
maintained.
2. Introjected regulation: Behaviors based on this type
of motivation involve an external regulation that
has been internalized but not fully accepted as one’s
own. Introjected-based behaviors are performed in
order to avoid guilt and shame, or conversely, to
gain approval. Students, for example, aim at behaving like teachers, peers, parents, or society, expect
them to act.
3. Identified regulation: This motivational quality
is even more self-determined. It involves a conscious valuation of a behavioral goal or regulation.
Behaviors based on identified regulations reflect the
acceptance of the behavior as personally important.
4. Integrated regulation: This form of motivation provides the basis for the most autonomous extrinsic
motivation. It represents the highest state in the
process of transforming external regulation into
true self-regulation. It results when the proceeding
identification has been brought into congruence
with the personal value system.
On the motivational continuum, these types of
extrinsic motivation are framed by the state of
amotivation (as status of lacking an intention to act)
on the one side and intrinsic motivation on the other.
Regarding academic learning, this distinction accounts
for the fact that most of the tasks worked on in formal learning contexts are not intrinsically motivated,
but are – to varying degrees – externally driven. Furthermore, the continuum explains that not all forms of
extrinsic motivation represent impoverished motivational states. Some forms even cover active and agentic

S

2996

S

Self-Determination and Learning

states. As educators cannot always rely on intrinsic
motivation to foster learning, understanding these different types of extrinsic motivation, their correlates,
and the factors that foster each of them, is essential
(Ryan and Deci 2009).

Important Scientific Research and
Open Questions
Early work within SDTexamined whether offering people
extrinsic rewards for doing intrinsically motivated activities affected their level of intrinsic motivation. Multiple
findings primarily gained in laboratory experiments illustrated that tangible rewards diminish intrinsic motivation
whereas positive, performance-relevant feedback maintains or enhances intrinsic motivation. To explain these
mechanisms, SDT draws upon the functional character of
the basic needs: When using rewards to prompt activities,
people lose their sense of autonomy and therefore intrinsic motivation decreases. In contrast, when people experience an absence of external pressure, they experience
self-determination which increases their intrinsic motivation. Similarly, effectance promoting feedback and optimal challenges help maintaining feelings of competence
and thereby enhance intrinsic motivation. Negative feedback, especially when given in a controlling manner, or
excessive demands undermine the experience of competence and decrease intrinsic motivation. These interpretations were confirmed by further studies: External events
such as punishments, deadlines, imposed goals, pressure
toward performance, competition, and evaluations
undermine intrinsic motivation in learners because they
thwart their needs for autonomy and competence (Ryan
and Deci 2008). These findings were later transferred to
the analyses of different forms of extrinsic motivation:
Field studies conducted in educational settings
supported two major conclusions: Autonomously
motivated students thrive in educational settings and
students benefit when teachers support their autonomy
(Reeve 2002). Such educational benefits refer to both
cognitive and affective components of learning: Selfdetermined motivation, for example, involves higher
academic achievements, greater conceptual understanding and flexibility in thinking, higher amounts
of active information processing, higher beliefs of
competence, efficacy, and control, higher rates of retention, and higher senses of well-being and more positive
emotionality. Autonomous forms of motivation are
also empirically associated with greater engagement,

less dropping out, more enjoyment, and better coping
styles (Ryan and Deci 2009). These findings led to the
analysis of autonomy-supportive teacher behaviors.
Such behaviors are, for example, fostering understanding and interest, providing meaningful rationales,
allowing criticism and encouraging independent thinking, and providing meaningful and true choices (Assor
and Kaplan 2001). Based on these results, more recent
studies heighten the ambivalent impact of choices on
autonomy and self-determined motivation (Patall et al.
2008). Research shows that autonomy support can be
manifested in diverse ways. One proposition is to
distinguish between organizational, procedural, and
cognitive autonomy (Stefanou et al. 2004). How far
such different qualities of autonomy support also
result in different forms of motivation is a mainly
unresolved problem. Although research in SDT has
a long and influential tradition in the field of learning
and instruction, several questions are still unsolved.
These open questions, for example, concern developmental and differential aspects of autonomous motivation: Whereas research has mainly focused on
the effects of immediate contextual conditions that
either support or thwart the basic needs, studies
approaching the idea that basic need support also
stems from individuals’ inner resources which potentially support ongoing feelings of autonomy and competence are at their infancies. How far contextual
aspects and individual characteristics are intertwined
in specific learning situations and how both factors
function in relation to each other and in dependence
to further constructs is a relevant question to answer.
For future research, it is further important to understand how motivation emanating from the experiences in specific learning situations develops into
a stabilized motivation that can be maintained across
different learning situations.

Cross-References
▶ Interests and Learning
▶ Learning-Related Motives and the Perception of the
Motivational Quality of the Learning Environment
▶ Motivation and Learning: Modern Theories
▶ Motivational Variables in Learning
▶ School Motivation
▶ Self-Regulation and Motivation Strategies
▶ Understanding Intrinsic and Extrinsic Motivation:
Age Differences and Meaningful Correlates

Self-directed Learning and Learner Autonomy

S

References

Definition

Assor, A., & Kaplan, A. (2001). Mapping the domain of autonomy
support: Five important ways to enhance or undermine students’
experience of autonomy in learning. In A. Efklides, J. Kuhl, &
R. M. Sorrentino (Eds.), Trends and prospects in motivation
research (pp. 101–120). Dordrecht: Kluwer.
Patall, E. A., Cooper, H., & Robinson, J. C. (2008). The effects of choice
on intrinsic motivation and related outcomes: A meta-analysis of
research findings. Psychological Bulletin, 134(2), 270–300.
Reeve, J. (2002). Self-determination theory applied to educational
settings. In E. L. Deci & R. M. Ryan (Eds.), Handbook of selfdetermination research (pp. 183–203). Rochester: University of
Rochester Press.
Ryan, R. M., & Deci, E. L. (2008). Self-determination theory and
the role of basic psychological needs in personality and
the organization of behavior. In O. P. John, R. W. Robins, &
L. A. Pervin (Eds.), Handbook of personality: Theory and research
(pp. 654–678). New York: Guilford Press.
Ryan, R. M., & Deci, E. L. (2009). Promoting self-determined school
engagement. In K. R. Wentzel & A. Wigfield (Eds.), Handbook of
motivation at school (pp. 171–195). New York: Routledge.
Stefanou, C. R., Perencevich, K. C., DiCintio, M., & Turner, J. C.
(2004). Supporting autonomy in the classroom: Ways teachers
encourage student decision making and ownership. Educational
Psychologist, 39(2), 97–110.

Self-directed learning is a concept that came out of
the humanistic tradition of adult education in the
second half of the twentieth century. One of the first
researchers on the topic, the Canadian scholar Allen
Tough, viewed self-directed learning largely as an
expression of human personal agency.

Self-Determined Motivation
▶ Understanding Intrinsic and Extrinsic Motivation:
Age Differences and Meaningful Correlates

"

2997

Man [sic], according to this view, can be active, energetic, free, and aware. He often chooses his goals,
direction, and behavior; he is not always pushed and
pulled by his environment and by unconscious inner
forces. (Tough 1979, p. 45)

Theoretical Background
SDL as Process
Interestingly, Tough’s concept of SDL adhered more or
less stringently to a linear process of learning that is
similar to other types of programmed learning. For
example, he described the self-directed learning project’s “stages” as being,
1. Decide on a learning goal
2. Determine a learning sequence and a learning
schedule
3. Secure the physical and financial resources to
pursue the learning project
4. Select a suitable place to learn

Self-Directed Learning
▶ Independent Learning
▶ Scaffolding Discovery Learning Spaces
▶ Self-organized Learning

5. Select resources and materials
6. Find appropriate resource persons
7. Resolve motivation issues
8. Overcome learning difficulties
9. Minimize self-doubt

Self-directed Learning and
Learner Autonomy
PAUL BOUCHARD
Department of Education Educational Studies and
Adult Education, Concordia University, Montreal,
QC, Canada

Synonyms
Self-guided learning

10. Set subsequent learning goals at the end of
a learning sequence

The framework proposed by Tough has been
applied to a large number of studies during the 1970s
and 1980s that assumed more or less implicitly that
self-directed learning is a process that can be accurately
described and analyzed over time as a sequential series
of events. Today, there are a large number of selfdirected “how-to” methods to be found in the educational world, particularly in the area of workplace
learning, that adhere to this philosophy.

S

2998

S

Self-directed Learning and Learner Autonomy

SDL as Personality
During that same period, Lucy Gulglielmino was working on a model of individuals’ propensity toward selfdirection in learning. She derived a multidimensional
model of assessment that is synthesized in a self-report
Likert-scale test, called Self-Directed Learning Readiness Scale (SDLRS).
Over the years, the SDLRS has been used to
establish correlations between readiness for selfdirection and numerous other variables. For instance,
Sabbaghian (1979) established a positive correlation
between SDL and self-perception. Torrance and
Mourad (1978) found that self-directed learners have
a marked propensity for “right-hemisphere” tasks,
such as creativity, analogy, and problem solving. Overall, according to Guglielmino’s view, not all persons
exhibit the same predisposition for self-direction in
learning, just as they differ in other psychological abilities such as creativity, problem-resolution, mathematical reasoning, etc.
There was some debate over the validity of the
SDLRS, mostly on the grounds that it tends to measure
a propensity toward learning in general, rather than for
self-directed learning in particular. Field (1989) and
Bonham (1991) objected to the vagueness of the concept, self-direction, in Guglielmino’s work. Nevertheless, researchers in adult education are still using the
SDLRS today, and the notion of self-direction as
a personality trait is still very much alive.

Situating their work in the light of Kurt Lewin’s
“field theory,” Mocker and Spear introduced the notion
of the “organizing circumstance” to shed some doubt
on the linear character that theorists would impose on
all types of learning, and especially self-directed learning. In particular, they contend that the task of planning a learning sequence is a rather specialized task for
which a learner in the natural setting is not likely to be
prepared. In the end, they conclude that it is not reasonable to assume that self-directed learning projects
can be planned in a similar way as formal learning
projects. This realization is being made today by proponents of self-managed learning networks or
connected “personal learning environments.”

Dimensions of Learner Autonomy
Certainly the most complete discussion of selfdirection in learning is to be found in Candy’s classic
work published in 1991. In that book, Candy (p. 101)
introduces the dual nature of learner autonomy,
both as an expression of self-determination (the personal disposition to pursue learning) and of selfmanagement (the ability to exert control over one’s
learning process). In order for self-directed learning
to occur, both conditions of learner autonomy must
be present.
Candy defines the pursuit of a SDL project as,
"

Being able to pursue a learning goal with equal vigour
and determination without being adversely affected by
external factors including the increase or decrease of
rewards for pursuing or attaining the goal (. . .)

"

Conceiving of goals, policies and plans independently
of pressures to do so, or not do so. (Candy 1991, p. 181)

SDL as Environment
George Spear and Donald Mocker (1984) introduced
the notion of SDL as an environmentally determined
phenomenon. While Tough and Guglielmino
contended that self-directed learners have the will and
capacity for carrying out personal learning projects,
Spear and Mocker pointed out that this was often not
the case at all. Their research revealed that learning was
determined by the learners’ surrounding circumstances
much more than by their “determination” or their
“inner predisposition.” Indeed, respondents to Spear
and Mocker’s research stated that they had not planned
any specific tasks or sequence in their learning.
"

Self-directed learners, rather than pre-planning their
learning projects, tend to select a course from limited
alternatives which occur fortuitously in their environment. (Spear and Mocker 1984, p. 4)

And further:
"

Being aware of alternative choices, both as to learning
strategies and to interpretations or value positions
being expressed, and making reasoned choices about
the route to follow in accordance with personally significant ideas and purposes. (Candy 1991, p. 261)

The first quote refers the learners achieving some
kind of freedom from the interfering forces of teachers
and institutions, while the second alludes to learners’
actual ability to structure and conduct their own learning. These are the first two “dimensions” of learner
autonomy.

Self-directed Learning and Learner Autonomy

The Conative Dimension
The first expression of learner autonomy resides in the
motivational-intentional forces that drive the learner
to apply some determination (or “vigor”) to the act of
learning. This area embodies the conative functions of
learning, which are at the foundation of learner initiative, motivation, and personal involvement. Most
often, adult learners harbor life-goals that are related,
but distinct from the actual learning goals (e.g., career
advancement, good parenting, or better health), as part
of the conative baggage they carry. Other possible
drives include the pleasure one derives from the act of
learning in itself, and the satisfaction obtained from
becoming part of a particular culture of knowledge
(Houle 1961).
Learner autonomy entails the possibility for individuals to make choices (or exercise control), beginning of course with the choice of whether to learn
anything at all in the first place (Candy 1991; Long
1993). This can be construed as the opposite of some
forms of “other-directed” learning such as mandatory
schooling, or of some instances of workplace learning,
where people have no choice about whether to learn, or
what to learn. So, the first area where learners may
exercise control over their learning lies in their option
to initiate and pursue learning, or to choose to do
something else instead. Interestingly, this means that
two learners sitting side by side in the same classroom
can be said to exercise contrastingly different levels of
learner autonomy: the first student having chosen to be
there, while the other has not.

The Algorithmic Dimension
The second set of elements that define learner autonomy involve control over the aspects of learning usually
taken over by a teacher or by a managed learning
environment (MLE). They include defining learning
goals, deciding on a learning sequence, choosing
a workable sequencing and pacing of learning activities,
and selecting learning resources (Hrimech and Bouchard 1998). These elements can be grouped under the
heading of the algorithmic dimension of learner autonomy. In traditional learning environments, most of the
algorithms are the responsibility of a teacher or
a teaching institution. Learning goals, student workload, and methods of evaluation are usually stipulated
at the outset and little participation in their formulation is expected from the learner. But some of the

S

2999

“teaching tasks” normally reserved for the instructor
can also be devolved to the learner, on top of the
expected “learning tasks.” In this sense, we can say
that autonomy is directly related to the number and
magnitude of the “teaching tasks” that are appropriated by the learner (Tough 1965).

Important Scientific Research and
Open Questions
Learner Autonomy in Context
The first two dimensions of learning that allow for the
possible exercise of learner agency, or autonomy, are
the conative (psychological) and the algorithmic (procedural) dimensions. However, these do not account
for the considerable weight of contextual forces that
inevitably influence the learning process. Let us consider for example the fact that learning can occur
within or without a formal learning institution, knowledge can be pursued alone, in small or large groups,
using prescribed texts or self-researched online
resources, etc. In order to take into consideration the
various contexts of learning, we have proposed two
other dimensions of learner autonomy (Bouchard
2009).

The Semiotic Dimension
Just a few years ago the resources for learning were
limited to interacting with printed text, or face-toface with other people. Today, the means of communicating information are very diversified and include
print, nonprint, multimedia, hypertext, internet web
pages, RSS feeds, and social networking. While these
may or may not affect the actual goals or outcomes of
learning (the algorithms), they will alter significantly
the learning process. Furthermore, individual learners
vary in their preferences for different environments
(British Library and JISC 2008). For instance, hypertext
can be selected as a means of aggregating information
or shunned as a mish-mash of poorly organized data.
Similarly, poring over texts in a library offers a very
different experience from lively participation in a blog
post. Each medium and its ultimate organizational
form carries its own set of pragmatics that will largely
determine the quality of the learning experience. These
examples among others illustrate the importance for
learners of making choices within the semiotic dimension of learner autonomy.

S

3000

S

Self-directed Learning and Learner Autonomy

The Economic Dimension
The traditional access to knowledge has been through
academically sanctioned programs in schools and colleges. The alternative was self-directed learning – to
pursue learning on one’s own or with the help of
informal mentors. It can be argued that both processes
lead to similar results in terms of learning achievement.
The main difference lies in the institution’s delivery of
an officially sanctioned, recognized accreditation.
While there may be little consequence, pedagogically speaking, of choosing one alternative over the
other, the outcome is of some importance economically when a learner enters the job market. In today’s
exploding “knowledge economy,” the types of choices
that have economic impact are no longer binary (formal vs nonformal) but are multiplied exponentially.
The value of learning is no longer the simple calculation of the cost of schooling against the benefits of
the degree conferred. The value of learning must be
assessed in light of numerous factors. For example,
the “perceived” value conferred by academic credentials can be matched or surpassed by the “actual” value
of knowledge in the marketplace. The cost of acquiring
credentials also varies from one institution to the
next, especially in their online offerings. In short,
learners in today’s complex environment typically
must exercise choices in the economic dimension of
learner autonomy.

Conclusions
Self-directed learning has been defined as a process,
a personality construct, and an environmentally determined phenomenon. In the end, any tangible occurrence of self-directed learning undoubtedly involves
the interaction of all three aspects, in that it will entail
(1) the application of some actions or procedures,
(2) by a person who is not psychologically averse to
the experience, (3) in an environment, which at the
very least does not preclude the emergence of selfdirected learning.
Learner autonomy is a multidimensional concept
that involves the exercise of learner-control in four
areas. The conative dimension represents the inner
motives and self-determination of the learner; the algorithmic dimension refers to the actual procedures
involved in the learning process; the semiotic dimension includes the symbolic and pragmatic features
of the learning environment; finally, the economic

dimension allows for the estimation of the value of
learning in its particular context.
In the course of pursuing their self-directed project,
learners will typically exercise some degree of control in
each of the dimensions of learner autonomy.

Cross-References
▶ Adult Learner Characteristics
▶ Adult Learning/Andragogy
▶ Adult Learning Styles
▶ Adult Learning Theory
▶ Cognitive and Affective Learning Strategies
▶ Ecology of Learning
▶ Epistemic Curiosity
▶ Experiential Learning Theory
▶ Learner Control
▶ Lifelong Learning
▶ Metacognition and Learning
▶ Self-determination and Learning
▶ Self-efficacy for Self-regulated Learning
▶ Self-organized Learning

References
Bonham, A. (1991). Guglielmino’s self-directed learning readiness
scale: What does it Measure? Adult Education Quarterly, 41(2),
92–99.
Bouchard, P. (2009). Some factors to consider when designing semiautonomous learning environments. European Journal of
e-learning, 7(2), 93–100.
British Library and JISC (2008). Information behaviour of the
researcher of the future. 11 January 2008. Retrieved www.ucl.
ac.uk/slais/research/ciber/downloads/
Candy, P. (1991). Self-direction in learning: A comprehensive guide to
theory and practice. San Francisco: Jossey Bass.
Field, L. (1989). An investigation into the structure, validity and
reliability of Guglielmino’s SDLRS. Adult Education Quarterly,
39(3), 125–139.
Houle, C. O. (1961). The inquiring mind. Madison: The University of
Wisconsin Press.
Hrimech, M., & Bouchard, P. (1998). Spontaneous learning strategies
in the natural setting: learning to use computers. In Huey B. Long
and Associates (Eds.), Developing paradigms for self-directed
learning. Oklahoma Research Center, University of Oklahoma.
Long, H. B. (1993). Emerging perspectives of self-directed learning.
Research Center for Professional and Continuing Education,
University of Oklahoma.
Spear, G. E., & Mocker, D. W. (1984). The organizing circumstance:
Environmental determinants in self-directed learning. Adult
Education Quarterly, 35(1), 1–10.
Tough, A. M. (1979). Choosing to learn. Toronto: The Ontario Institute for Studies in Education (OISE).

Self-Efficacy for Self-Regulated Learning

Self-Direction
▶ Self-Determination and Learning

Self-Efficacy
▶ Hope Theory and Hope Therapy

Self-Efficacy for Self-Regulated
Learning
ELLEN L. USHER
Educational, School, and Counseling Psychology,
University of Kentucky, Lexington, KY, USA

Synonyms
Academic self-efficacy; Motivation to learn;
Self-regulatory self-efficacy; Study skills self-efficacy

Definition
Self-efficacy for self-regulated learning refers to the
beliefs individuals hold in their capabilities to think
and behave in ways that are systematically oriented
toward or associated with their learning goals. Students
with a robust sense of efficacy in their self-regulatory
capabilities believe they can manage their time effectively, organize their work, minimize distractions, set
goals for themselves, monitor their comprehension, ask
for help when necessary, and maintain an effective
work environment.

Theoretical Background
In 1986, psychologist Albert Bandura advanced
a theory of human functioning that emphasized the
role of cognitive, vicarious, self-regulatory, and selfreflective processes in human adaptation and change.
This view highlights the primary role of humans as
agentic forces in their own development and stands in
contrast to previous psychological theories, which
viewed people as reactive organisms primarily shaped
by their environments or driven by hidden inner
impulses. Social cognitive theorists explain human

S

functioning as the result of the interaction between
personal (i.e., cognitive, affective, and biological),
behavioral, and social/environmental factors. These
three factors influence one another in a process of
triadic reciprocality. For this reason, humans, through
their own cognition and behaviors, are thought to
play a primary role in determining the course their
lives will take.
Social cognitive researchers have noted the importance of human agency in the learning process. To be
successful, learners must monitor their own behavior,
manage changing environmental conditions, and regulate their own cognition so as to select, initiate, and
engage appropriate strategies throughout the learning
process. Self-regulated learning involves metacognitive
awareness that leads to goal-directed activities such as
attending to instruction, selecting relevant information, connecting new learning to prior knowledge,
motivating one’s self, and establishing productive
work routines and environments (Schunk and
Zimmerman 2003).
Zimmerman (2000) described a three-phase, cyclical model of self-regulation by which individuals
engage in (a) a forethought phase that sets the stage
for action, (b) a performance phase in which learning
tasks take place, and (c) a self-reflection phase during
which learners reflect on their strategy use and prior
performance. The decision to engage in self-regulatory
learning strategies depends in part on one’s beliefs in
one’s capability to employ the strategies required to
achieve desired learning outcomes, or one’s self-efficacy
for self-regulated learning. Learners must not only know
what strategies are needed to achieve desired learning
outcomes, they must also possess a belief that they are
capable of employing such strategies. For example,
a writer might know that outlining, drafting, incubation, and revision are all important steps to producing
a good manuscript, but unless she feels confident that
she is capable of accomplishing these steps, she is
unlikely to initiate them.
During the performance phase, students who
believe they are capable of accomplishing given learning tasks make better use of cognitive strategies and
persist longer than do those who harbor self-doubts.
After they have performed, students interpret the
results of their actions, the social cues they receive
from others, and their own physiological and emotional states to revise their efficacy judgments, which

3001

S

3002

S

Self-Efficacy for Self-Regulated Learning

are then used to initiate a new self-regulatory cycle.
Throughout this cyclical process, learners set goals,
plan their strategies, observe and learn from models,
assess their self-efficacy, evaluate their goal progress,
and alter their task strategies as needed. Developing
and maintaining a sense of self-efficacy helps to motivate students and to promote their self-regulated learning (Zimmerman and Cleary 2009). By assessing
learners’ self-efficacy for self-regulated learning, teachers
will know with greater certainty whether learners will in
fact regulate their learning appropriately.

Important Scientific Research and
Open Questions
A key determinant of whether learners employ selfregulatory strategies rests in the beliefs they hold
about their capabilities to do so (see Zimmerman and
Cleary 2006). Self-regulatory efficacy beliefs influence a
number of outcomes. First, students who possess a
robust belief in their self-regulatory capabilities engage
successful self-regulatory skills and strategies across
academic domains. Second, students’ self-efficacy for
self-regulated learning has been shown to be related to
key motivation variables, such as academic self-efficacy,
self-concept, perceived value of school, and achievement goal orientation (i.e., positive relationship with
mastery goals and negative relationship with performance-avoidance goals). Third, self-regulatory selfefficacy is related to students’ academic performance
in domains such as writing, science, mathematics problem-solving, and general academics. Self-regulatory
efficacy beliefs are also predictive of students’ school
achievement such as course grades. Self-efficacy for
self-regulated learning is inversely related to some
affective variables. For example, students who report
higher academic and subject-specific anxiety report less
confidence in their self-regulatory capabilities.
Students’ self-efficacy for self-regulated learning has
been found to vary among different groups of students.
Girls typically report higher levels of self-efficacy for
self-regulated learning than do boys (Pajares 2002).
Some scholars have reported that students’ confidence
in their self-regulatory capabilities declines as they
progress through school, even though older students
use more self-regulatory strategies (Usher and Pajares
2008). In lower school grades, many academic tasks
and activities are structured, guided, and closely monitored so as to instruct and instill the self-regulatory

habits that will serve youngsters for years ahead. As
students approach secondary school, they face more
demanding academic work that is no longer
chaperoned by external aids. Older students are often
expected to regulate their academic work and study
habits on their own, and many lose confidence when
their self-regulatory skills are insufficient for new task
demands. Successful learners are able to self-regulate
across the situational demands they face. As Bandura
(2006) contended, “The issue is not whether one can do
the activities occasionally, but whether one has the
efficacy to get oneself to do them regularly in the face
of different types of dissuading conditions” (p. 311).
Because self-efficacy refers to a task-specific judgment of what one can do, it is likely that self-efficacy for
self-regulated learning will prove maximally predictive
of academic outcomes when it is measured in a manner
that is specific to the academic task at hand. For example, beliefs in one’s capability to minimize distractions
while reading complex information may be related to
behavioral or motivational outcomes in the domain of
reading but not in the domain of mathematics.
Researchers must be attentive to the local contexts in
which self-efficacy is being assessed. One context in
which self-efficacy for self-regulation might play an
increasingly important role is in digital learning environments. When learning activities require the use of
digital learning devices, one’s beliefs that one can stay
on track by reducing distractions and staying focused
may become even more consequential. This area offers
a promising avenue for future research.

Cross-References
▶ Procrastination and Learning
▶ Self-Determination and Learning
▶ Self-Regulated Learning
▶ Self-Regulation and Motivation Strategies

References
Bandura, A. (2006). Guide for constructing self-efficacy scales.
In F. Pajares & T. Urdan (Eds.), Adolescence and education (Selfefficacy and adolescence, Vol. 5, pp. 307–337). Greenwich, CT:
Information Age Publishing.
Pajares, F. (2002). Gender and perceived self-efficacy in self-regulated
learning. Theory into Practice, 41, 116–225.
Schunk, D. H., & Zimmerman, B. (2003). Self-regulation and learning. In W. M. Reynolds & G. E. Miller (Eds.), Handbook of
psychology (Vol. 7): Educational psychology (pp. 59–78). Hoboken, NJ: Wiley.

Self-Esteem and Learning
Usher, E. L., & Pajares, F. (2008). Self-efficacy for self-regulated
learning: A validation study. Educational and Psychological Measurement, 68, 443–463.
Zimmerman, B. J. (2000). Attaining self-regulation: A social cognitive
perspective. In M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.),
Handbook of self-regulation (pp. 13–39). San Diego: Academic
Press.
Zimmerman, B. J., & Cleary, T. J. (2006). Adolescents’ development of
personal agency. In F. Pajares & T. Urdan (Eds.), Adolescence and
education (Self-efficacy beliefs of adolescents, Vol. 5, pp. 45–69).
Greenwich, CT: Information Age Publishing.
Zimmerman, B. J., & Cleary, T. J. (2009). Motives to self-regulate
learning. In K. R. Wentzel & A. Wigfield (Eds.), Handbook of
motivation at school (pp. 247–264). New York, NY: Routledge.
Further Reading
Bandura, A. (1997). Self-efficacy: The exercise of control. New York:
Freeman.
Pajares, F. (2007). Motivational role of self-efficacy beliefs in selfregulated learning. In B. J. Zimmerman & D. H. Schunk (Eds.),
Motivation and self-regulated learning: Theory, research, and applications (pp. 111–140). New York: Erlbaum.
Zimmerman, B. J., & Kitsantas, A. (2005). The hidden dimension of
personal competence: Self-regulated learning and practice. In
A. Elliot & C. S. Dweck (Eds.), Handbook of competence and
motivation (pp. 509–526). New York: Guilford Press.

Self-Esteem and Learning
DORIS LEWALTER, ARIANE S. WILLEMS
TUM School of Education, Fachgebiet
Gymnasialpädagogik, Technical University Munich,
Munich, Germany

Synonyms
Self-regard; Self-respect; Self-worth

Definition
Self-esteem describes a subjective way of experiencing
one’s own self. It encompasses cognitive, emotional, and
evaluative dimensions and generally describes how much
value people assign to themselves. Self-esteem in this
sense reflects a person’s overall appraisal of his or her
own worthiness, competence, and personality. Theoretically, personal self-esteem applies to different levels
of generalization: As individual trait variable, self-esteem
is treated as a general or global psychological construct
functioning across various life domains. On a more

S

3003

specific and content-related level, it is specified as
a multidimensional concept including four major areas:
the intellectual, social, emotional, and physical aspect of
self-esteem. These factors describe different dynamic
parts of a person’s self-esteem as a state variable that
operates within particular situations. Within each of
these factors a distinction between several facets can be
made. For example, the intellectual self-esteem of a person, which is activated in learning situations, is usually
differentiated with regard to a specific subject matter or
content domain. Although self-esteem is related to the
concept of self-efficacy it differs from it. Self-efficacy in
contrast to self-esteem implies a future time component
and is conceptually restricted to subjective competence
perceptions. Hence, self-efficacy in general relates to
a person’s perception or belief of his or her future ability
to reach a goal or produce a certain level of performance
in a particular task within a specific situation.

Theoretical Background
The term self-esteem has a long history and is still used
in various ways (Kernis 2006). In earlier theories selfesteem was conceptualized as a basic human need or
motivation. For example, Abraham Maslow (1987)
included two different forms of self-esteem in his hierarchy of needs, a lower and a higher form of selfesteem. He differentiates between both, the need to be
respected by others and the need for self-respect or
inner self-esteem. The former describes the need for
recognition, attention, acceptance, and appreciation.
The latter one involves ideas such as the individual
need for strength, competence, self-confidence, and
freedom. Maslow assumed that inner self-esteem is
the higher form because it is both more personal and
stable, as it particularly refers to the valuation of inner
competencies. The need to be respected by others on
the opposite side represents the lower form of selfesteem as it seems to be more fragile and dependent
on the valuation of others. According to Maslow’s
proposed hierarchy of needs, the fulfillment of the
needs for love and belonging is a necessary precondition for perceiving self-esteem and confidence. The
fulfillment of the need for self-esteem in turn is
required for an individual’s personal growth and for
obtaining self-actualization. Further conceptualizations define self-esteem in terms of a one-dimensional
approach of worthiness (e.g., Morris Rosenberg).
Other approaches define self-esteem as a direct

S

3004

S

Self-Esteem and Learning

function of a person’s experienced competence and
effective behavior in important life domains as well as
a kind of personal response to the approval gained by
significant others (e.g., Susan Harter). According to
this definition self-esteem is based on success and
achievement.
In most of the current theories and research
approaches, self-esteem includes a combination of both
aspects; competence and worthiness (Mruk 2006). Selfesteem is defined according to these two interrelated
aspects and entails a sense of personal efficacy and
personal worth. It is conceptualized as the integrated
sum of self-confidence and self-respect and describes
the belief that one is competent to live and worthy of
living (Mruk 2006).
Self-esteem can be conceptualized either as an
underlying dispositional tendency influencing a person’s behaviors and experiences across certain life
domains – the so-called global or general self-esteem
(trait self-esteem) – or as a more transitional psychological condition mainly operating within particular
domains (state self-esteem). Both conceptualizations
of self-esteem are substantially correlated with one
another. Yet, the mutual relation between a global
self-esteem and a more domain-specific self-esteem is
discussed controversially (Kernis 2006): It is an open
question whether domain-specific factors of selfesteem emerge from a differentiation of experiences
related to the global self-esteem or whether the global
self-esteem emerges by processes of generalization
based on experiences affecting domain-specific components of self-esteem.
Self-esteem in general refers to a person’s subjective
belief about herself or himself. Baumeister et al. (2003)
point out that such subjective beliefs do not necessarily
correspond to a person’s real characteristics. Thus, high
self-esteem may stem from an accurate and warrantable
appreciation of a person’s own worthiness or his or her
success and competencies in an important domain, but
it can also be built on an inflated, unjustified sense of
imagined superiority over others. Accordingly, low selfesteem can be either a realistic, well-founded understanding of one’s insufficiency as a person or a biased,
even pathological sense of insecurity and inferiority. As
a person’s subjective sense of self-esteem does not imply
any definitional requirements of accuracy, it is rather
perception than reality (Baumeister et al. 2003). Therefore, level and quality of self-esteem are two distinct

characteristics although they are correlated. Regarding
the quality of self-esteem, various differentiations are
made like, for example, fragile vs. secure self-esteem,
conscious vs. unconscious self-esteem, or implicit vs.
explicit self-esteem (Kernis 2006).
Furthermore, self-esteem is not only based on an
absolute appraisal but rather on a social comparison.
A person’s self-esteem stems from the assumed difference between him or her and the group of reference, for
example, a school class. The effects of social comparisons on self-esteem are particularly strong for the intellectual part of a person’s self-esteem, which is activated
in specific learning situations. As self-esteem stems
from social comparisons which in turn are particularly
important in learning situations in order to develop
a realistic intellectual self-esteem, empirical research on
the development and the impact of self-esteem is
important.

Important Scientific Research and
Open Questions
In modern Western cultures, global self-esteem has been
regarded as a psychological attribute of major importance. It was assumed that high self-esteem will cause
many positive outcomes and benefits, like, for example,
academic achievement, interpersonal and vocational success, well-being, health, and happiness (Baumeister et al.
2003). The underlying belief was: the higher person’s selfesteem the better. As a consequence, self-esteem enhancing projects were conducted. Regarding learning and
school performance there are plausible reasons for that
assumption (Baumeister et al. 2003). Learners with high
self-esteem presumably set higher aspirations, may be
more willing to persist when confronted with difficult
tasks, problems or initial failure, and may be more
confident to tackle demanding tasks. Students with
high self-esteem should have less feeling of self-doubt
and incompetence when facing a challenging task than
learners with low self-esteem. Consequently, students
with high self-esteem should reach high academic
achievements und successes.
Yet, various research findings gained over the last
two decades illustrate that this simplified conclusion is
not adequate (Baumeister et al. 2003). The empirical
evidence that self-esteem improves performance in
academic settings is weaker than expected. Correlations
are both unstable and only modest on average. These
modest correlations between self-esteem and school

Self-explanation, Feedback and the Development of Analogical Reasoning Skills

performance do not indicate that high self-esteem leads
to good academic performance. Students with high
self-esteem outperform students with low self-esteem
only to a small extent. Recent research even partly
indicates that high self-esteem as well as inflating students’ self-esteem has no significant positive effect on
grades (Baumeister et al. 2003).
Additionally, correlations by nature do not carry
any information about causality. Recent research
approaches aiming at disentangling causal directions
via complex path analyses indicated that there is no
direct causal path from self-esteem to learning, academic achievement, or performance. Such findings
lead to the assumption that self-esteem and academic
achievement mutually influence each other and are
probably both results of a set of third variables like for
example, ability or family background. Scientific
results indeed support the hypothesis that high selfesteem is partly the result of school performance.
Therefore, self-esteem seems to be an effect rather
than a reason of high performance. In addition, interventions designed to raise self-esteem may either fail to
influence academic performance or actually undermine it, for example, by encouraging a sense of complacency (Baumeister et al. 2003).
Major open questions concern the conceptualization
of self-esteem, its measurement, and the explanations of
the role of self-esteem for other variables like, for example, learning, academic achievement, and success. In current research theories, there is a noticeable lack of a widely
accepted conceptualization of what self-esteem actually is
(Kernis 2006). There is no agreement even on basic
aspects of self-esteem across the multiple theories available. This lack of correspondence has important implications on research and application. One consequence is
a variety of different instruments of measurements
reflecting somewhat different constructs of self-esteem.
This is the case for trait self-esteem as well as state selfesteem. This leads to an inconsistent body of research
results which entails divergent appraisals of the role of
self-esteem for variables such as learning processes, persistence, academic achievement, or success.

Cross-References
▶ Development of Self-Consciousness
▶ Learner Characteristis
▶ Motivation and Learning: Modern Theories
▶ Narcissistic Learning

S

3005

References
Baumeister, Rf, Campbell, J. D., Krueger, J. I., & Vohs, K. D. (2003).
Does high self-esteem cause better performance, interpersonal
success, happiness, or healthier life styles? Psychological Science in
the Public Interest, 4, 1–44.
Kernis, M. H. (2006). Self-esteem issues and answers. New York:
Psychology Press.
Maslow, A. H. (1987). Motivation and personality. New York: Harper
& Row.
Mruk, C. J. (2006). Self-esteem. Research, theory and practice. New
York: Springer.

Self-explanation, Feedback and
the Development of Analogical
Reasoning Skills
ANDREA S. TOWSE, LINDEN J. BALL, CHARLIE N. LEWIS
Department of Psychology, Lancaster University,
Lancaster, UK

Synonyms
Justification; Think-aloud protocols; Verbalization

Definition
Self-explanation is a general term used for verbalizations that participants provide during experiments.
These explanations can be concurrent (e.g., given while
learning about a particular phenomenon or during
problem solving) or retrospective (e.g., made after
learning has taken place or a problem has been solved).
They can be elaborations of given information during
learning, verbalizations of accessible thought processes,
or justifications of why an answer has been given in
a problem-solving situation. The main feature that selfexplanations have in common is that they are verbalizations that a participant provides about their
own learning or problem solving.
Analogical reasoning is reasoning based on understanding one thing in terms of another. To investigate
analogical reasoning ability, proportional analogy
problems are often employed. These problems are
where two related items are presented together and
a third item is presented on its own. The participant
has to generate or select an additional item that relates
to the third item in the same way in which the first two

S

3006

S

Self-explanation, Feedback and the Development of Analogical Reasoning Skills

items are related. This is often depicted as A is to B as
C is to? A simple example is: cat is to kitten as dog is to?;
the relation between these items is the name of the
young offspring for the type of animal. The correct
answer, therefore, is puppy.

Theoretical Background
The Self-explanation Effect
Asking participants to provide self-explanations in
learning and problem-solving situations enhances performance in both adults and children. There are two
main theories that try to account for the benefits of selfexplanation: a processing account and a mental models
account. The processing account suggests that providing self-explanations during learning or problem solving helps participants to avoid focusing too much on
surface level features of a problem, the overall goal and
rules of a problem, but to concentrate on strategies,
subgoals, and evaluation of progress (e.g., BerardiColetta et al. 1995). On the other hand, the mental
models account suggests that self-explanations help
participants to identify gaps in their knowledge or
understanding (e.g., Chi 2000). Once participants
have identified these gaps, they can then make the effort
to gain the information required to complete their
understanding.

Development of Analogical
Reasoning Skills
There is evidence to suggest that analogy is used to
reason spontaneously from infancy with a significant
increase in children’s ability to reason analogically over
the first few years at school (Siegler and Svetina 2002).
The three main factors that have been posited to
explain this development include: (1) the way children
are able to represent relations, (2) an increase in
domain knowledge, and (3) maturational changes in
high-level cognitive ability.
The first factor concerns the idea is that children
undergo a “relational shift” (e.g., Ratterman and
Gentner 1998), whereby their analogical reasoning
changes over time from being based initially on the
surface features of stimuli (e.g., objects and attributes)
to the inclusion of information about the relations
between objects, eventually incorporating whole systems of relations including hierarchically nested relational structures.

The occurrence of this relational shift may be
related to the second factor, that is, the idea that as
children gain more general knowledge their ability to
identify relations increases (e.g., knowing that a kitten
is the offspring of a cat) such that they are able to use
analogy to reason more successfully. Children who lack
relevant domain knowledge will incorrectly rely on
surface features of a problem rather than the necessary
relations.
It has also been claimed that working memory plays
a pivotal role in the acquisition of analogical reasoning
skills. This includes maturational changes in its capacity as well as attentional and inhibitory mechanisms
associated with executive functioning. These changes
facilitate children’s ability to represent multiple dimensions, which, in turn, gives rise to improved performance on tasks involving more complex analogical
mappings.

Important Scientific Research and
Open Questions
The Confound of Feedback
Feedback is emerging as an important factor in participants’ problem-solving performance. Participants
who receive feedback outperform those who do not
(e.g., Ball et al. 2010; Cheshire et al. 2005). It is common in studies that elicit self-explanation for some
kind of feedback to be provided to participants,
whether it is implicit or explicit. For example, when
participants are thinking aloud while learning from
official texts, they know the information is correct,
and when participants are asked to provide
a justification for an answer they have provided, they
are often corrected or asked to explain why a different
answer is correct. Therefore, an important question to
ask is: How do we know that self-explanation per se is
the factor that is enhancing performance and not the
associated feedback?

How do Self-explanation and
Feedback Affect the Development of
Analogical Reasoning?
Several ▶ microgenetic studies have examined the
development of children’s analogical reasoning skills
using complex proportional analogy tasks. Results provide compelling evidence that self-explanation and

Self-explanation, Feedback and the Development of Analogical Reasoning Skills

feedback enhance the development of children’s analogical reasoning skills (e.g., Ball et al. 2010; Cheshire
et al. 2005; Siegler and Svetina 2002). Children shift
from focusing on surface level features of items in the
problems to using strategies based on relations between
items. These findings are in line with the idea of
a relational shift in the development of children’s reasoning skills.
Self-explanation has been shown to have limited
long-term benefits. Enhancements to children’s performance is observed at the time of providing explanations, but the level of performance drops off when
participants are no longer required to provide explanations (e.g., Cheshire et al. 2005). This suggests that selfexplanation may enhance the processing aspects of
problem solving; verbalizations may help guide participants to focus on the less superficial aspects of the
problem (e.g., item similarity) and employ the most
successful strategies based on relations between the
items. These findings provide support for the
processing account of the benefits of self-explanation
(e.g., Berardi-Coletta et al. 1995).
Feedback seems to have more long-term benefits.
When no longer given feedback, participants continue
to perform at the same levels as when they were given
feedback (e.g., Cheshire et al. 2005). This suggests that
it is feedback (rather than self-explanation) that promotes the completion of a participant’s mental model
of proportional analogy. Participants might be keen to
improve their own performance and analyze why they
were wrong. A more complete mental model means
that participants’ understanding is long lasting enough
for them to perform well even when they are not being
given trial-by-trial feedback. However, feedback is most
useful when it is combined with explanation.
The combined benefits of self-explanation and
feedback lead participants to the greatest levels of
understanding (Cheshire et al. 2005; Siegler and
Svetina 2002). This is unsurprising if self-explanation
is helping participants processing abilities and feedback
is enhancing their mental representations. These factors are likely to interact to produce high levels of
performance and are unlikely simply to be additive.
Furthermore, asking participants to explain why an
answer is incorrect has been show to be even more
useful that explaining why an answer is correct (e.g.,
Ball et al. 2010). It is possible that explaining erroneous
answers requires higher levels of analysis and, in turn,

S

3007

leads to deeper understanding of how to make effective
relational mappings.

Remaining Challenges
There are still many unanswered questions in the area
of self-explanation, feedback, and the development of
analogical reasoning skills. A content analysis of selfexplanations may reveal further insights into how verbalization helps participants to focus on the essential
elements of a problem. Furthermore, similar research
needs to be carried out on other analogical reasoning
tasks to see if the impacts of explanation and feedback
are similar in other areas of analogy and not just proportional problems.

Cross-References
▶ Analogical Reasoning
▶ Analogical Reasoning by Young Children
▶ Feedback and Learning
▶ Feedback Strategies
▶ Metacognition and Learning
▶ Modeling Microgenetic Data
▶ Qualitative Learning Research
▶ Self-Reflecting Methods of Learning Research

References
Ball, L. J., Hoyle, A. M., & Towse, A. S. (2010). The facilitatory
effect of negative feedback on the emergence of analogical reasoning abilities. British Journal of Developmental Psychology, 28,
583–602.
Berardi-Coletta, B., Buyer, L. S., Dominowski, R. L., & Rellinger, E. R.
(1995). Metacognition and problem solving: a process-oriented
approach. Journal of Experimental Psychology. Learning, Memory,
and Cognition, 21, 205–223.
Cheshire, A., Ball, L. J., & Lewis, C. N. (2005). Self-explanation,
feedback and the development of analogical reasoning skills:
Microgenetic evidence for a metacognitive processing account.
In: Peer reviewed paper in B. G. Bara, L. Barsalou, & M.
Bucciarelli (Eds.), Proceedings of the twenty-seventh annual conference of the Cognitive Science Society (pp. 435–441). Mahwah,
NJ: Lawrence Erlbaum Associates.
Chi, M. T. H. (2000). Self-explaining expository texts: The dual
processes of generating inferences and repairing mental
models. In R. Glaser (Ed.), Advances in instructional psychology
(pp. 161–238). Hillsdale, NJ: Lawrence Erlbaum Associates.
Ratterman, M. J., & Gentner, D. (1998). More evidence for a relational
shift in the development of analogy: children’s performance on
a causal mapping task. Cognitive Development, 13, 453–478.
Siegler & Svetina, 2002.
Siegler, R. S., & Svetina, M. (2002). A microgenetic/cross-sectional
study of matrix completion: Comparing short-term and longterm change. Child Development, 73, 793–809.

S

3008

S

Self-explanations

Self-explanations
Self-explanations involve the learner generating statements (i.e., elaborations) during learning by explaining
the content of the learning materials to themselves. As
an example, we could tell learners that “positive reinforcement increases the likelihood of the behavior that
it follows.” A self-explanation could be: “Positive reinforcement would need to be something attractive that
I would want to receive again.”

Self-Guided Learning
▶ Self-directed Learning and Learner Autonomy

Self-Organization in Learning
▶ Self-organized Learning

Self-organized Learning
RENAE LOW, PUTAI JIN
School of Education, The University of New South
Wales, Sydney, NSW, Australia

Synonyms
Self-directed learning; Self-organization in learning

Definition

Self-Handicapping
▶ Procrastination and Learning

Self-Inquiry
▶ Creative Inquiry

Self-Knowledge
▶ Development of Self-consciousness

Self-Maintenance
▶ Physiological Homeostasis and Learning

Self-Management
▶ Self-regulation and Motivation Strategies

The conceptual framework of self-organized learning
was originally proposed and developed by E. Sheila
Harri-Augstein, Laurie F. Thomas, and their associates
of the Centre for the Study of Human Learning at
Brunel University (Harri-Augstein and Thomas 1979,
1983). It was initially a reaction to “cookbooks” for
study (taking “good notes,” following the “algorithmic”
instructions on the laboratory card, etc.) that flooded
the education market, which, according to them, might
result in students’ dependency upon the content and
structure of their enrolled courses. As pointed out by
Harri-Augstein and Thomas (1983), many students are
unaware of the processes that attribute personal meaning to the displays of public knowledge offered in
classrooms, laboratories, libraries, and workplaces.
They assert that student should learn how to learn
and unfreeze their fixed habits of reading, listening,
talking, writing, thinking, feeling, and doing in order
to conduct conscious reflection and constructive
review.
According to Harri-Augstein and Thomas (1991),
self-organized learning refers to the conscious and
ongoing activity of learners who are able to apply
constructive conversations with themselves and others
about the learning process and are skillful in observing,
searching, analyzing, formulating, reflecting, and
reviewing on the basis of such engagement. Because
a number of students, if unaided, are unable to generate
effective learning conversations with themselves,
supporting tools can be used to facilitate a systematic

Self-organized Learning

review and remolding of the existing learning strategies
and skills. Through learning conversations, students
can become self-organized learners, who not only possess the dispositions to improve their own learning
processes and outcomes, but also render support to
the learning of others.
It should be pointed out that the concept of selfdirected learning in a sense is related to, but broader
than self-organized learning. Most often self-directed
learning is considered as a series of complex educational processes in which learners (a) take initiatives
to identify their own learning needs, (b) set up their
own learning goals, (c) search resources (including
instructors/peers and material) for learning, (d) choose
and implement their own approaches to and strategies
of learning, and (e) evaluate learning outcomes
(Knowles 1975). Basically, there are two types of selfdirected learning: autodidaxy and learner control.
Whereas autodidaxy is a self-taught, independent practice in which individuals select their own topic,
method, and standards for learning, learner control
refers to the process that responsibilities of learning
are transferred step by step from experts to the individual student, a continual process closely in line with
the description of self-organized learning highlighted
by Harri-Augstein and Thomas (1979, 1983, 1991).
Readers need to be aware that in the literature,
a proportion of researchers have used the term “selforganized learning” loosely or even interchangeably
with self-directed learning and self-regulated learning.
For instance, in a special issue on “Technology Support
for Self-Organized Learners” in the journal of Educational Technology & Society (Kalz et al. 2009),
a collection of articles using terms like “self-regulation,” “self-regulated online learning,” “self-directed
learning,” “self-direct intentional learning,” “selfdevelopment of competences,” “self-regulating strategies,” and “self-organized learning/learners” were
included. This phenomenon of cross-boundary usage
of related terms in some research reports should be
noted for further conceptual clarification.

Theoretical Background
The practice of self-organized learning has its deep
roots in humanistic, cognitive, and behavioral perspectives (see Harri-Augstein and Thomas 1979). The
humanistic underpinnings are largely based on the
person-centered approach developed by Carl Rogers,

S

whose theory of therapy, personal growth, and interpersonal relationships emphasizes three critical conditions for meaningful and resourceful learning to occur:
congruence, acceptance, and empathic understanding.
Congruence or realness can be achieved by letting students construct their environment for learning and
create a system for open reactions (mutual feedback)
so that both the student and the tutor (who takes the
role as a facilitator) can view the whole picture of this
proactive learning state. Acceptance of or respect to
students can encourage individuals to express their
learning targets. Empathic understanding can help
ascertain learners’ problems, constraints, and potentials.
According to Harri-Augstein and Thomas (1979,
1983), the self-organized learning approach is further
supported by George Kelly’s cognitive theory that
focuses on personal efforts to obtain meanings of various encounters by the ways in which a person anticipates certain events that are likely to occur. In other
words, ordinary people, just like scientists, continuously construct, test, revise, and develop personal theories of self as well as the world (associated activities)
surrounding self. Whereas the personal construct theory was originally used to deal with thoughts and
feelings, Harri-Augstein and Thomas (1979, 1983)
have applied this theory to education and drawn specific attention to the mechanisms of learning by which
personal meaning is searched, constructed, and
remolded.
The framework of self-organized learning also benefits from the conversation theory developed by Gordon Pask. Based on systems research, the theory is
focused on conversations between human beings as
participants or between relevant perspectives in one’s
human brain. Learning can be attained in a program
(i.e., a set of instructions and exchanges) via a language,
which is either a natural language or a system of symbolic behaviors. In this interactive process, the conversation is a vehicle for proposing, commanding,
questioning, expressing, and exchanging. To facilitate
learning, the subject matter should be conversed in
the form of explicit entailment structures that exhibit
what has been learned and indicate what is to be
learned. While acknowledging Gordon Pask’s contributions to bridge the gap between thought and action,
Harri-Augstein and Thomas (1979) consider that
the formality of his approach places learners in an
over-constrained position. Thus in their self-organized

3009

S

3010

S

Self-organized Learning

learning approach, they encourage learners to gradually
take control of their own learning and assist them in
building up a repertoire of their personal language and
tactics to achieve this. In other words, in the process of
learning conversations, the teacher and/or peers assist
the learner to clarify essential components of effective
learning and support the learner in establishing
a personal learning contract (i.e., a self-controlled,
structured, achievable outline within an intentional
study period). Ultimately, learners are able to act on
their new internalized reality, try out and demonstrate
new or changed behaviors.

Important Scientific Research and
Open Questions
Self-organized learning per se is a type of Lewinian
action research in educational settings, in which individuals interact with educational resources available to
them, identify criteria with rigor and validity, and build
up personal autonomy. In this learning conversation
process, learners reflect on their learning experiences,
and their tutor mirrors the reflection and heightens
learners’ awareness, thus enabling learners to develop
their skills for exploration and to move toward greater
competency and creativity.
The action research projects of learning conversation have been carried out at three levels of discourse:
(a) the learning-to-learn conversation, in which learning habits can be identified, evaluated, altered,
reviewed, and rebuilt; (b) the tutorial conversation, in
which the longer-term strategic issues of learning, such
as the planning of goals, deployment of resources,
negotiation of benchmarks, and execution of purposes
are discussed; and (c) the life conversation, in which the
learner attains greater competence in self-organized
learning that can be used to improve quality of life
and contribute to individual enterprises and society.
For instance, the repertory grid techniques and talkback strategies have been adopted to assist learners to
explore, examine, and develop their self-organized
capacity to learn from print (Harri-Augstein and
Thomas 1979, 1983, 1991).
The repertory grid was originally designed by
George Kelly in psychotherapy and social psychology
to provide clients/subjects with an opportunity to generate and assess their own constructs based on a set of
elements listed by those participants. In learning conversational action research, this grid-based method is

used to help a learner tap hidden feelings, prejudgments, and habituated assumptions that influence
learning. Initially, the learner is aided to explore the
area of concern (e.g., having difficulties in taking useful
notes); a set of representative elements that are central
and meaningful to the problem area are listed;
a construct is indicated by emerging clusters (each
contains several elements); consensus is made through
exchanges between the learner and the consultant; the
grids, mapping personal constructs in the reorganized
form, are fed back to the participant, who sustains this
exploratory process to reflect his or her whole system of
feelings and thoughts in relation to particular difficult
learning encounters; and a reappraisal of the learner’s
approach to the area of concern is thus on the agenda.
In order to facilitate self-organized learning, HarriAugstein and Thomas (1979, 1983, 1991) have also
used talk-back strategies, such as a behavioral record
of reading and a flow diagram of a text, in learning
conversations. While reading, many individuals are
unaware of the actual cognitive/affective processes
that underlie their reading behavior. By means of the
computerized device reading recorder, the learner is
given the opportunity to check the time spent on
each passage, changes in pace, hesitations, skipping,
skimming, backtracking, pause, note-making, and so
forth. Furthermore, the advisor who actively participates in this learning conversation should offer appropriate interpretations and help the learner acquire
relevant perceptual skills to recognize personally significant events in the processes of reading. In addition, the
flow-diagram technique can be employed to help the
learner to display a multilevel, multidimensional structure of meaning that is relevant to personal experience
and understanding. By mapping this complex structure
of meanings and relationships onto the records of
reading episodes, the learner becomes explicitly aware
of the processes in which various interrelated meanings
are attributed to corresponding parts of the text. In
particular, the underlying factors that cause the learner
to slow down, pause, review back, and skip forward can
be investigated. The learner and the teaching participant work together to interpret the records of reading
behavior and the flow diagram by referring to the
original text and the learner’s purpose for reading.
Such experiential learning raises the level of selfawareness and brings forth an opportunity for the
learner to become a purposeful, self-organized reader.

Self-Reflecting Methods of Learning Research

S

In the era when the mission of education is
redefined as lifelong learning and the growth of technological innovations is at an exponential speed, selforganized learning networks and techniques offer tremendous potentials for formal and informal education
to be transformed to a learner-centered and selfcontrolled enterprise, which is far beyond the courses
and curricula managed by expert providers or
constrained by technological tools. For instance, in
the online and distributed sectors, Sharma and Fiedler
(2007) report that the personal Web publishing system,
with its inherent features of personal ownership, individual customization, and public conversation, can
foster reflective practices with self and others and thus
support self-organized learning. A five-phase procedure can be used to encourage students to combine
private and public learning conversations through
Weblogs: establishment, introspections, reflective
monologue, reflective dialogue, and constructing
knowledge artifacts. Such a type of social software
allows learners to have multiple roles (being authors,
audience, commentators, feedback receivers, revisers,
etc.) and thus to take multifaceted perspectives in
a wider self-organized learning community. The evidence-based knowledge of human cognitive architecture accumulated in recent decades can also be used to
enhance self-organized learning. For instance, the
learner and the teaching participant can be involved
in a joint venture of customized instructional design, in
which a learner’s mental load, preferences, prior knowledge, and other characteristics are taken into account
(see, for example, ▶ Modality Effect, ▶ Redundancy
Effect, and ▶ Stress and Learning), so that the learner
becomes a “scientist” on learning and gains increasing
insights into the human information processing. In
sum, integration of age-proven pedagogical principles
with digital technologies has opened the gate for new
forms of self-organized learning.

References

Cross-References

MICHAELA GLÄSER-ZIKUDA
Department of Education, Institute for Educational
Science, Friedrich-Schiller University of Jena,
Jena, Germany

▶ Achievement Motivation and Learning
▶ Action Research
▶ Calibration
▶ Lifelong Learning
▶ Metacognition and Learning
▶ Self-Regulated Learning
▶ Stress and Learning
▶ Technological Learning in Organizations

3011

Harri-Augstein, E. S., & Thomas, L. F. (1979). Learning conversations: A person-centred approach to self-organised learning.
British Journal of Guidance & Counselling, 7(1), 80–91.
Harri-Augstein, E. S., & Thomas, L. F. (1983). Developing selforganized learners: A reflective technology. New Directions for
Adult and Continuing Education, 1983(19), 39–48.
Harri-Augstein, E. S., & Thomas, L. F. (1991). Learning conversations:
The self-organized learning way to personal and organizational
growth. London: Routledge.
Kalz, M., Koper, R., & Hornung-Prähauser, V. (2009). Technology
support for self-organized learners (guest editorial). Educational
Technology & Society, 12(3), 1–3.
Knowles, M. (1975). Self-directed learning: A guide for learners and
teachers. New York: Cambridge Book.
Sharma, P., & Fiedler, S. (2007). Supporting self-organized learning
with personal Webpublishing technologies and practices. Journal
of Computing in Higher Education, 18(2), 3–24.

Self-Organizing Maps
▶ Connectionist Theories of Learning

Self-Recognition
▶ Development of Self-consciousness

Self-Reference
▶ Development of Self-consciousness

S

Self-Reflecting Methods of
Learning Research

Synonyms
Introspection; Learning diary; Learning journal; Portfolio; Protocol; Self-awareness; Self-reflection

3012

S

Self-Reflecting Methods of Learning Research

Definition
Self-reflection or introspection means self-observation
and report of one’s thoughts, desires, and feelings. It is
a conscious mental process relying on thinking, reasoning, and examining one’s own thoughts, feelings,
and, ideas. It is contrasted with extrospection, the
observation of things external to one’s self.
In the past years, there has been a growing interest
in introspective or self-reflecting methods, such as the
“thinking-aloud” interview or stimulated recall, in
which a subject engaged in a task, speaks his/her
thoughts aloud. This allows studying thoughts without
influencing the subject to think too long of what he/she
is asked, for example, in questionnaires. Further,
written forms of self-reflection are learning diary,
learning protocol, and portfolio. From the perspective
of learning psychology and educational science, these
approaches contribute to the paradigm shift from
teaching to learning. Empirical studies in the context
of education focus mainly on learning diary, respectively learning journal, protocol, and on portfolio.
Learning diaries, for example, have a broad and
long tradition in different disciplines. Diaries are
applied in clinical psychology, developmental psychology, and medicine. They are research objects in biographical research, developmental, and educational
research. Pedagogical diaries, for example, are well
known for description and reflection of everyday
school life and instructional planning. The use of diaries by researchers shows their versatility as a research
tool. Diaries are applied in the evaluation and interpretation of the practice of teaching, training, and learning, especially in action research. The advantages to
analyze, for example, students’ learning strategies or
teachers’ beliefs are discussed in numerous studies.
Moreover, diaries are used in studies on personal identity and life transition, and health. In the past few years,
diaries are implemented as learning tools on school and
university level, as well. A learning diary allows and
supports continuous documentation and reflection of
the learning process. Because of its characteristic of
continuity the diary may be applied as a time-sampling
method to capture courses and changes, for example, of
learning strategies, attention, motivation, or emotional
experiences over time (Schmitz and Wiese 2006).
Complementary to a diary or learning protocol,
a portfolio combines more perspectives and functions.

Characteristics of a portfolio are, besides self-reflection,
documentation of processes and learning outcomes, as
well as their evaluation and assessment (Paulson et al.
1991). Generally, five portfolio types may be differentiated (Spandel 1997). (1) The working-portfolio is
used to document strengths and weaknesses of
a learning process. Therefore, it is mainly used for
diagnostic purposes, and may serve for a consultation
with respect to learning and achievement. (2) Learning
progress and improvement are focused in a developmental portfolio. Learners may easier observe and evaluate their own learning processes, and by that, plan and
organize further learning steps more adequately. (3)
The presentation portfolio is a collection of the best
learning documents or products of a learner he/she
selected to demonstrate personalabilities in one or
multiple subjects/domains. (4) The fourth type is the
evaluation or assessment portfolio. This type of a portfolio helps to document the performance of a learner
and has a more formal character. It may serve for
information of parents, teachers, and principals of
other schools. (5) Finally, the application portfolio
focuses on the documentation of qualifications and
performance as a certificate. The portfolio is also
discussed as an alternative instrument for assessment.
The portfolio approach focuses on multiple aims.
A portfolio may intend to enhance self-regulated learning by enabling and supporting planning, transposing,
observing, and evaluating students learning by themselves. Self-regulated learning is an important condition of lifelong learning, and presumes metacognitive
knowledge, and metacognitive strategies. By emphasizing student orientation in school and university, the
portfolio approach contributes to the shift from teaching to learning.

Theoretical Background
Both, learning diaries and portfolios, focus on the
participation of learner in planning, realizing, and
evaluating their own learning process. Furthermore,
communication and refection are important conditions for self-dependent and self-regulated learning.
It is assumed that they contribute to lifelong learning.
Generally, two main directions of learning diary
and portfolio in education are discussed. First, the
enhancements of (self-regulated) learning, and second, the aspects of assessment are linked with both

Self-Reflecting Methods of Learning Research

self-reflecting methods. The application of learning
diaries and portfolios may be seen as an example for
the shift from teaching to learning. From this point of
view educational institutions are seen no longer as
a place for knowledge transfer but rather as a wellprepared learning environment to support individual
learning processes. Main characteristics are problemcentered and competence-oriented learning demands
and tasks, high quality of interactions between learners
and teachers, possibilities for cooperative learning, and
a well-balanced relation between structured-guided
phases and open-individualized phases.
The described demands of self-regulated learning
emphasize the importance of the ability to reflect on
one’s own learning process. Two dimensions concerning reflection may be differentiated. The first
dimension describes reflection as a process of a person’s
thinking of something. Self-reflection as well as reflection in interaction with other persons is important. The
second dimension refers to the object of reflection. It
may refer to aspects of the person itself or to aspects
that are relevant for other persons, as “external reflection,” or for a whole group, as for example thinking of
collective goals or socio-emotional climate. Also the
learning environment may be an object of reflection.
A third dimension is the aim of reflection which may be
retrospective or prospective.
As a cognitive process reflection does not need any
support by media. But the representing of thoughts by
symbols (language, written language, and pictures)
gives the possibility to pass on one’s own thoughts
and to imagine one’s own thoughts at a later date. In
the tradition of Socrates, introspection was applied by
the German physiologist Wundt in his experimental
psychology laboratory in 1879. Wundt assumed by
using introspection in his experiments he would gather
information about how minds of human beings are
working. In education different approaches to motivate
learners to document and to reflect on their learning
process have a long tradition. By the end of nineteenth
and the beginning of twentieth centuries, the
“Reformpädagogik” had already highlighted the potential of learning diary, portfolio, or further reflexive
instruments.
A person reflects upon himself/herself and the
learning process and its outcome to become aware of
the sense and purpose of learning. This may foster the

S

ability of self-regulation, of personal responsibility to
make learning more effective. What exactly is reflected
has to be specified. It may be personal learning goals,
motives, topics and subjects, learning methods and
strategies, progress, problems or obstructions, or learning results that are reflected upon. Following the three
phases of self-regulated learning as a model of complete
action includes planning, transposition, and evaluation. As cognitive, metacognitive, motivational, volitional, and emotional factors interact in all phases of
the learning process, continuous reflection is needed to
understand and regulate the whole process effectively.
Further, it is assumed that the personality of the learner
is influenced favorably.
A learner may reflect in thought or communicate
dialectically. A third option is written reflection.
Procedures of continuous, written, and self-reflective
documentation of one’s own learning process are
assumed to have positive effects on information
processing and cognition, and acquisition of knowledge. From a constructivist perspective, writing is an
active process of knowledge acquisition, and may be
understood as thinking and learning tool. In addition,
it may be understood as problem-solving process.
Empirical studies revealed, for example, that students who did reflect on what they were doing were
more effective in solving tasks in a short time, in adaptive learning, elaborating, and in testing hypotheses
than students who were not trained in reflection or
avoided reflecting for personal reasons. Recent investigations show a positive effect of reflection on learning
strategies and achievement, mainly in the learning process, i.e., the actional phase.
Negative effects of reflection were described, as
well. Not adequately developed abilities for reflection and a lack of support were related to excessive
demand of the learner. Too extended or intensive
reflection caused lack of spontaneity or even incapability of action.
Finally, the quality of reflection documented in
learning diaries or portfolio depends on the learner’s
personality, epistemology, individual goals and
motives, as also on conditions of the learning environment. It will be more demanding to establish reflexive
elements in teacher-centered instruction than in an
open and student-oriented learning environment. Consequently, learning environment as well as learning and

3013

S

3014

S

Self-Reflecting Methods of Learning Research

teaching culture has to be considered for an integration
of reflexive elements. Furthermore, support, interaction,
and communication with classmates, peers, and teachers
have been described as relevant for successful implementation of self-reflexive instruments.
Support and guidance to fill-in a learning diary or
a portfolio are relevant suppositions for the realization
and quality of reflection. Central aspects, such as aims
of reflection, and reflection objects are helpful for the
learner. One possibility is to offer a written instruction
with the main objectives. A second possibility is
to formulate open guiding questions or “prompts”
focusing on the purpose and aim of reflection. In
most of the studies, prompts point to phases, steps, or
specific aspects of the individual learning process.
Interindividual differences have an influence on the
effectiveness of prompts. For example, prompts are
not helpful for persons skilled at reflection, but they
are experienced as helpful by learners who are not
accustomed to reflect (Berthold et al. 2007).
The use of learning diaries and portfolios in
research studies draws the following picture. Studies
applying learning diaries focus first of all on the
enhancement of self-regulated learning, and analyze
and support the development and use of learning strategies. The improvement of achievement by application
of learning diaries is discussed in these studies, as well.
Meanwhile, learning diaries are used in school, university education, teacher education, and in-service training. In the past years, digital, electronic, and Web-based
diaries and portfolios are applied increasingly often
because of multiple advantages of filling-in and analyzing the instrument. How intensively self-reflecting
methods are discussed, shows the implementation of
electronic portfolios in different countries at university
level. The aim is to support students’ competence and
lifelong-oriented learning.
Especially the portfolio is implemented in research
on learning to assess competences and performance in
numerous areas, but first of all in writing education
(Bangert-Drowns et al. 2004). Furthermore, portfolio
as an alternative form of assessment in comparison
to traditional approaches of assessment is investigated
(Tillema 2001). As a tool for self-assessment, the
portfolio is meanwhile investigated, as well. Moreover,
some studies analyze the portfolio in examination
situations. The majority of research on portfolio as
an instrument for improvement of learning and

achievement takes the students’ and teachers’ perspective likewise into account.

Important Scientific Research and
Open Questions
The utilization of self-reflecting methods in learning
research is a comparably new research area. Therefore,
systematical analyses are needed to clarify individual,
social, and environmental conditions, influences and
effects of learning diaries and portfolios on affective,
cognitive, and social factors of learning. Different formats have to be tested, such as open, less or highly
structured instruments, paper–pencil, and digital or
Web-based versions. The existing perception of potentials of learning diary and portfolio needs to be systematically expanded with respect to different groups of
learners, domains, and institutions. Furthermore, it is
of interest how implementation of these instruments
may be transposed in different educational settings
and contexts, as well as in in-service trainings. Up to
now, less is known about aggravating or facilitating
circumstances. Finally, a comparison of different selfreflecting methods, mainly of learning diary and portfolio, is indispensable to understand and capture the
potential and limitation of these approaches.

Cross-References
▶ Action Research
▶ Assessment in Learning
▶ Diagnosis of Learning
▶ Evaluation of Student Progress in Learning
▶ Learning Criteria, Learning Outcomes and Assessment Criteria
▶ Methodologies of Learning Research
▶ Peer Learning and Assessment
▶ Qualitative Learning Research

References
Bangert-Drowns, R. L., Hurley, M. M., & Wilkinson, B. (2004). The
effects of school-based writing-to-learn interventions on academic achievement: A meta-analysis. Review of Educational
Research, 74(1), 29–58.
Berthold, K., Nückles, M., & Renkl, A. (2007). Do learning protocols
support learning strategies and outcomes? The role of cognitive and metacognitive prompts. Learning & Instruction, 17(5),
564–577.
Paulson, F. L., Paulson, P. R., & Meyer, C. A. (1991). What makes
a portfolio a portfolio? Educational Leadership, 48(5), 60–63.

Self-Regulated Learning
Schmitz, B., & Wiese, B. S. (2006). New perspectives for the evaluation of training sessions in self-regulated learning: Time-series
analyses of diary data. Contemporary Educational Psychology,
31(1), 64–96.
Spandel, V. (1997). Reflections on portfolios. In G. D. Phye (Ed.),
Handbook of academic learning. Construction of knowledge (The
educational psychology series, pp. 573–591). San Diego:
Academic.
Tillema, H. (2001). Portfolios as developmental assessment
tools. International Journal of Training and Development, 5(2),
126–135.

Self-Reflection
▶ Self-Reflecting Methods of Learning Research

Self-Regard
▶ Self-Esteem and Learning

Self-Regulated Comprehension
▶ Comprehension Monitoring

Self-Regulated Learning

S

goal-setting, planning, organizing, strategy use, selfmonitoring, and feedback seeking, has been triggered
(cf. Schunk 2004). Learning, after all, is regarded as
a kind of complex human activity to be done by students rather than to be done for students. In the mid1980s, educational psychologists and pedagogical
experts used the term self-regulated learning more frequently and worked on projects associated with its
conceptual framework. During the recent two decades,
research on self-regulated learning has been flourishing
in various areas, ranging from classroom activities to
homework accomplishment, from conventional
instruction to e-learning, and from training in sports
to health education (Hadwin et al. 2010; Jin and Low
2009; Pintrich et al. 1993; Zimmerman 2008).
Despite the diversity of studies on self-regulated
learning, researchers in general accept the following
generic definition developed by Barry J. Zimmerman,
Dale H. Schunk, Paul R. Pintrich, Philip H. Winne, and
others: self-regulated learning refers to the phenomenon that students systematically activate and sustain
metacognitive, motivational/affective, and behavioral
processes in order to attain their learning goals in
a particular context. As such, self-regulated learning
usually involves four distinctive phases, namely, planning, monitoring, control, and evaluation; and the
associated activities are ongoing and spiral toward the
ultimate goals. Although at the earlier stage of research
self-regulated learning was mainly examined in the
individual learning context, it should be pointed out
that this type of learning can be implemented and
tested in both individual and collaborative settings.

Theoretical Background
RENAE LOW, PUTAI JIN
School of Education, The University of New South
Wales, Sydney, NSW, Australia

Synonyms
Self-regulating processes;
regulatory learning

Self-regulation;

Self-

Definition
Concomitant with the trend of education that started
in the 1970s shifting from teacher-centered approaches
to student-centered approaches, research attention to
individual self-regulatory processes in learning, such as

3015

According to Schunk (2004), the theoretical framework
of self-regulated learning is mainly derived from behavioral, information processing, and social cognitive perspectives. Behavioral research shows that operant
behavior (a) is likely to occur if reinforced with positive
consequences, and (b) can be fostered in the presence
of discriminative stimuli. In a learning context, desirable behavior (e.g., completing a quality assignment
before the due date) can be nurtured by self-instruction
(e.g., establishing a rule of using the period from
Monday to Friday between 7 and 9 pm for study
only rather than playing the online game World of
Warcraft), self-monitoring (e.g., keeping a diary and
checking the utilization of planned study periods), and

S

3016

S

Self-Regulated Learning

self-reinforcement (e.g., rewarding oneself a weekend
ski trip after the completion of a satisfactory assignment). These three subprocesses form the key behavioral components of self-regulated learning.
Information processing theories view human cognitive architecture as an operational model in terms of
the organization of information handling and storage.
The model comprises sensory memory–working memory–long-term memory components. Working memory (named short-term memory previously), which
can hold about five to seven chunks of information
for a very short period, is the gateway between sensory
input and long-term information storage. From the
perspective of evolutionary psychology, the limited
capacity and duration of working memory may be
necessary to allow only essential but not random, dysfunctional information to be sent to long-term store,
thus maintaining the cognitive system’s relative stability and incremental progress. The utility of one’s
existing long-term memory (which has no known
limit) is a major, effective way to relax the limits in
working memory for information processing; and
learning occurs meaningfully when there is progressive
alterations in the learner’s long-term memory. In other
words, learning can be facilitated if relevant prior
knowledge stored in long-term memory is activated
and then effectively integrated with new information
in working memory. If learners possess a high level of
metacognitive awareness (what is to be learned, when
and how it is to be learned, and so forth) and a set of
useful strategies (analyzing, planning, organizing,
implementing, adjusting, etc.), the information
processing can be enhanced by their purposeful selfregulatory activities. For instance, a student with
a certain degree of constructive metacognition may
acquire the note-taking skill, use it in lecture, and
carefully check the task demands before action. Such
initiatives contribute to effective self-regulation.
From a social cognition perspective, learning is
a process where learners transform their mental abilities and epistemological beliefs into the acquisition
of specific knowledge or skills. Self-regulated learning
is characterized by its proactive features – the learner
not only displays personal initiatives, adaptability,
and perseverance, but also actively seeks feedback
and help from peers, instructors, and social network.
Zimmerman (2008) theorizes that there are several
steps that can lead to the desired level of self-regulated

learning: (a) observation, in which a learner adopts
vicarious learning from a skilled model; (b) imitation,
in which a learner replicates a modeled skill while
receiving social feedback; (c) self-control, in which
a learner independently uses a demonstrated skill
on a structured task; and (d) self-regulation, in which
a learner is capable to adaptively employ skills to deal
with changing tasks in various contexts. Self-regulators
tend to have higher self-efficacy, demonstrate more
intensive goal-setting activities, and report less negative
academic procrastination problems than externally
regulated students (Jin and Low 2009; Schunk 2004;
Zimmerman 2008).

Important Scientific Research and
Open Questions
Research in the area of self-regulated learning is one of
the fruitful areas in modern educational psychology
and pedagogy. The following sections attempt to summarize the major progresses and future directions in
this domain.
Constructs and instruments – Because of the
multidimensional features of self-regulated learning,
its constructs have been expressed as the degree to
which learners proactively pursue their goals through
metacognitive, motivational, and behavioral engagement. Basically, there are two categories of measures,
aptitude-related and event-related. The aptitude measures are employed to gauge a learner’s relatively enduring attributes related to self-enhanced learning
behavior, motivation, and metacognition. The scales
in these inventories are often designed to calibrate
a person’s responses over time, such as “most of time”
or that “is typical of me” (Schunk 2004; Zimmerman
2008). The event-related approach represents an alternative assessment of self-regulated learning, which is
regarded as a set of multifaceted, sequential responses
over a particular period and in an authentic context. To
some extent, an event-related approach that adopts
multipoint measures can reflect the sequential dependency of learners’ responses and thus are helpful in
evaluating the effectiveness of interventions (e.g.,
metacognitive skills training) and drawing causal
inferences.
The most frequently used aptitude-related measures include self-reporting inventories, structured
interview, and teachers’ judgments (Pintrich et al.
1993; Zimmerman 2008). Commonly used inventories

Self-Regulated Learning

include the Learning and Study Strategies Inventory
(LASSI) assessing skill, will and self-regulation strategies, the Motivated Strategies for Leaning Questionnaire
(MSLQ) assessing learning strategies and motivation,
and other shorter questionnaires. Those tailored inventories are useful in collecting cross-sectional information about self-regulated learning on a relatively large
scale. Structured interview methodology, with its
unique feature to elicit students’ prospective answers,
can be used to obtain an insightful view on students’
internal states as well as overt responses. For instance,
a team led by Zimmerman (2008) has developed the
Self-regulated Learning Scales (SRLIS) as a structured
interview format to assess students’ motivation (selfevaluation reactions and self-consequences), metacognition (goal-setting and planning, organizing and
transforming, seeking information, and rehearsing
and memorizing), and behavior (environmental structuring, keeping records and monitoring, reviewing,
and seeking assistance from others), In addition,
teachers’ ratings and comments are another type of
rich sources to examine the consistency of students’
self-regulated learning and to verify information gathered from other approaches.
There is a wide range of event-related measures
available for self-regulated learning research, such as
think-aloud method, error detection in task performance, trace method, and observation in naturalistic
settings. In view of the advancement of information
technology, Zimmerman (2008) proposes a number of
innovative ways to scrutinize self-regulated learning,
including trace logs (frequency, duration, and topics)
in computer-assisted learning environments, using
a think-aloud protocol in hypermedia environments
to record students thoughts and associated cognitive
activities while performing a task, structured diary to
identify learning processes, observation with multimedia means, and micro-analytic measures and cyclical
analyses.
Effectiveness of self-regulated learning and student
training – On the whole, research endorses the merits
of self-regulated learning in terms of personal development as well as learners’ acquisition of academic domain
knowledge and skills (Schunk 2004; Zimmerman 2008).
For instance, in a recent study undertaken in a secondary school, Grigorenko et al. (2009) found that, in
addition to the grade prior to entry and standard
admission test scores, aspects of self-regulated learning,

S

such as academic self-efficacy, academic motivation,
academic locus of control, and reasoning skills were
also powerful predictors of academic success. The
researchers further speculate that high levels of selfregulated learning and analytical abilities may even
lead to post-schooling life success. Research indicates
that more effort is needed to promote self-regulated
learning for ill-structured tasks. There are also preliminary findings in relation to gender, disciplinary, and
age differences in self-regulated learning. More and
more researchers, instructors, and educational administrators are investigating motivational issues and
working on educational programs to enhance motivation and self-regulation skills. In general, efforts are
exerted in the aspects of motivational analysis, training,
and maintenance. In should be noted that it is more
economical and effective to have those interventions
embedded in daily teaching and learning practices than
separate actions.
Teachers’ (and parents’) attitude and teacher training –
Students’ engagement in self-regulated learning by no
means implies any reduction in teaching load. On the
contrary, teachers have an increased responsibility to
nurture students’ self-regulated learning competence,
and they themselves need to understand critical issues
in relation to self-regulated learning, in both theory
and practice. Effort has been made to develop a selfreport teacher scale to assess teachers’ beliefs about
introducing self-regulated learning in school education. Investigation in preservice teachers’ professional
growth shows that the trainees under the condition combining e-learning with self-regulated learning exhibited high competence in self-regulated
learning (indicated by behavioral, metacognitive, and
motivational factors), sound pedagogical knowledge
(e.g., designing a learning unit), and strong studentcentered orientations (e.g., self-construction of
knowledge). Research also indicates that it is feasible
to promote students’ self-regulatory activities at
various levels (from kindergarten children to senior
students) by the implementation of self-regulated
learning programs for teachers. Likewise, children’s
self-regulation is also influenced by parental attitude
and support. Concrete work on parental education and
school–family–community linkage is much needed to
create a pro-self-regulation climate.
The role of self-regulated learning in digital
environments – The popularity of computer-based

3017

S

3018

S

Self-Regulated Learning Strategies

education provides both opportunities and threats to
the implementation of effective self-regulated learning.
On the one hand, students can access a large amount of
course information almost without boundaries and at
such a high speed that former generations never
dreamed of. On the other hand, students may be distracted by nonessential information and lost in the
cyber space. We need to construct effective e-learning
systems based on the evidence accumulated in cognitive, motivational, and behavioral research, and students should take multiple roles (as users, testers,
informants, and design partners) in this process. In
this regard, Winne and associates have made a successful attempt to produce a versatile software system
(gStudy) and create an interactive digital environment
for self-regulated learning (e.g., Hadwin et al. 2010).
Detailed trace data, such as concept maps, learners’
notes and uploaded material, can be unobtrusively
collected. Data collected from interactive environments
are conducive to ascertaining the “blind spots” in selfregulated learning studies, such as active strategies
being adopted at different stages, social aspects of
learning processes revealed from online dialogue, and
indicators of changes in students’ interaction throughout e-learning. Research in this direction is promising
and challenging.

Hadwin, A. F., Oshige, M., Gress, C. L. Z., & Winne, P. H. (2010).
Innovative ways for using gStudy to orchestrate and research
social aspects of self-regulated learning. Computers in Human
Behavior, 26(5), 794–805.
Jin, P., & Low, R. (2009). Enhancing motivation and self-regulated
learning in multimedia environments. In R. D. Koo, B. C. Choi,
M. R. D. Lucas, & T. C. Chan (Eds.), Education policy, reform, and
school innovations in the Asia-Pacific Region (pp. 525–547). Hong
Kong: Association for Childhood Education International –
Hong Kong & Macao (ACEI-HKM).
Pintrich, P. R., Smith, D., Garcia, T., & McKeachie, W. (1993). Predictive validity and reliability of the Motivated Strategies for
Learning Questionnaire (MSLQ). Educational and Psychological
Measurement, 53, 801–813.
Schunk, D. H. (2004). Learning theories: An educational perspective
(4th ed.). Upper Saddle River: Pearson Education.
Zimmerman, B. J. (2008). Investigating self-regulation and motivation: Historical background, methodological development, and
future prospects. American Educational Research Journal, 45(1),
166–183.

Self-Regulated Learning
Strategies
▶ Cognitive and Affective Learning Strategies

Cross-References
▶ Achievement Motivation and Learning
▶ Autonomous Learning and Effective Engagement
▶ Independent Learning
▶ Introspective Learning and Reasoning
▶ Learner Control
▶ Learning to Learn
▶ Metacognition and Hypermedia Learning: How Do
They Relate?
▶ Metacognition and Learning
▶ Modality Effect
▶ Redundancy Effect
▶ Scaffolding Discovery Learning Spaces
▶ Self-Organized Learning

References
Grigorenko, E. L., Jarvin, L., Diffley, R., III, Goodyear, J., Shanahan,
E. J., & Sternberg, R. J. (2009). Are SSATS and GPA enough?
A theory-based approach to predicting academic success in
secondary school. Journal of Educational Psychology, 101(4),
964–981.

Self-Regulating Processes
▶ Self-Regulated Learning

Self-regulation
Responsibility for learning outcomes assumed by the
learner, including self-generated thoughts, feelings, and
actions for attaining academic goals.

Cross-References
▶ Extraversion, Social Interaction, and Affect Repair
▶ Physiological Homeostasis and Learning
▶ Self-Determination and Learning
▶ Self-Regulated Learning

Self-regulation and Motivation Strategies

Self-regulation and Motivation
Strategies
MARY NIEMCZYK
Department of Technological Entrepreneurship &
Innovation Management, Arizona State University,
Mesa, AZ, USA

Synonyms
Self-control; Self-management

Definition
In general, self-regulated learning focuses on the manner in which individuals orient their behaviors, cognition, and affect in order to attain a desired achievement
goal. Self-regulated learners initiate and direct their
efforts to learn and are, therefore, active participants
in their own learning processes.

Theoretical Background
Self-regulation is a fairly new focus area first appearing
in psychological research journals during the 1980’s.
One aspect of this construct, self-regulated learning,
has been described from various theoretical perspectives, the most prominent being: operant, information
processing, developmental, social constructivist, and
social cognitive.

Operant Theory
The operant theory of self-regulation is based on the
work of B.F. Skinner. Operant behavior is behavior
whose occurrence depends on the resulting environmental consequences. A behavior becomes more likely to
occur if it results in either positive or negative reinforcement and less likely to reoccur if it results in punishment.
Behaviors can also be influenced by events occurring prior to the behavior. Prior events, or antecedent
stimuli, that are predictive of reinforcement can
become discriminative stimuli. If an individual correlates the discriminative stimuli to subsequent reinforcement, the discriminative stimuli can cause the
individual to behave in a particular way due to the
increased likelihood of reinforcement.
Self-regulation as viewed by operant theorists is
therefore based on two types of controlling stimuli:
those that occur before the behavior, antecedent

S

3019

stimuli, and those that result because of the behavior,
reinforcement, or punishment. In this perspective,
the individual determines the behaviors to regulate
corresponding to the discriminative stimuli. The
reoccurrence of these behaviors is based on whether
the performance matches the discriminative stimuli,
and subsequent reinforcement received. Those behaviors that are reinforced are more likely to occur again.

Information Processing Theory
Information processing theorists depict learning as
the transfer of information from short-term memory
into more permanent storage of long-term memory.
To do this effectively, individuals need to utilize organizational learning strategies to connect new information to prior knowledge, making learning more
meaningful. From this standpoint, self-regulatory
processes consists of the individual comprehensively
assessing the learning task to determine the desired
outcomes, evaluate their personal capabilities and
current knowledge, then determine the appropriate
cognitive strategies that will effectively transfer information into long-term memory. In this perspective,
the individual maintains continuous awareness, or
metacognition, of their progress so that he or she can
adjust his or her performance in order to attain the
desired learning goal.

Developmental Theory
Self-regulation in this perspective focuses on the evolution of control that learners execute on their
thoughts, behaviors, and affect. An individual’s progression toward becoming self-regulated is presumed
to be dependent on the growth of their cognitive skills.
Developmental theorists suggest that selfregulatory behaviors are introduced to the individual
through external, social sources, such as parents. As the
individual’s cognitive development progresses, these
skills tend to become internalized and controlled by
the individual. By modeling behaviors of parents and
teachers, novice learners can begin to acquire and utilize self-regulatory behaviors. After continued guidance, feedback and reinforcement, these skills tend to
become more automatic. Instead of being an exact
replication, individuals may exhibit behaviors that generally imitate the models behaviors, thereby indicating
a transition to self-control. As self-regulatory skills
advance, the individual may seek out teachers and

S

3020

S

Self-regulation and Motivation Strategies

mentors to assist them in refining these skills; however,
over time, this may occur less frequently.

Social Constructivist Theory
The social constructivist perspective of self-regulation
is founded on the theories of cognitive development
regarding change and continuity. A central tenet of this
perspective is that learning is considered to be an
inherent need. Individuals continuously seek new
information and attempt to develop understanding by
imposing meaning based on prior knowledge and
experiences. This understanding is continuously
transforming through input of new information, cognitive reorganization, reflection, experience, and social
guidance. Learning may be constrained at times, however, due to the individual’s current level of knowledge
and skill. It is at this point, characterized by Vygotsky as
the “zone of proximal development,” that the individual requires assistance from a teacher or advanced peer
in order to proceed. As the individual continues to
reflect and reorganize information, they are intrinsically motivated to acquire new information. This
reexamination of self, behavior, and knowledge are
consistent with the major components of self-regulated
learning.

Social Cognitive Theory
The social cognitive perspective of self-regulation
focuses on the interaction of personal, behavioral, and
environmental processes. Because these factors are continuously changing and affecting the learning situation,
the individual needs to constantly monitor feedback,
be metacognitively aware, and formulate strategic
adjustments. This triadic depiction of self-regulation
provides for a cyclical, open feedback loop that allows
for both reactive and proactive increases in learning
and performance.
Self-regulatory processes in this perspective consist
of forethought, performance, or volitional control, and
self-reflection. The forethought phase consists of task
analysis goal setting, strategic planning, as well as
assessment of self-motivational beliefs such as self-efficacy, and outcome expectations. The performancevolitional control phase involves aspects of self-control
and self-observation. The third phase, self-reflection,
includes the processes of self-evaluation and selfjudgment.

Social cognitive theorists believe that self-regulation
is domain specific consequently processes may vary
across learning situations. While the overall processes
of self-regulation are maintained, individuals need to
have flexibility in adapting these processes to specific
learning situations. Inherent in this framework is the
individual’s confidence, or self-efficacy, in their ability to
adapt their behaviors to unique situations.

Achievement Goals
Achievement goals are the underlying reasons or purposes for an individual’s learning behaviors. Essentially,
goals represent the importance that an individual
assigns to a learning activity. Achievement goals provide individuals with a means not only to define their
successes and failures but also how they may possibly
react to the outcomes of their efforts.
Highly self-regulated individuals typically establish
both process goals as well as outcome goals. Accomplishment of process goals provides the individual
feedback on their performance and serves as indicators
of progress toward the overall achievement goal.
The selection of a particular achievement goal by an
individual creates a structure of consequent achievement behaviors. Depending on the goal adopted, individuals utilize varying methodologies to gather
situational data to determine the appropriate strategies
for success. An individual’s goal orientation establishes
a framework of thoughts and behaviors, such as asking
relevant questions and seeking information that
will enable the individual to attain the goal. Goal orientations also elicit different motivational processes.
In general, goals are considered to be the cognitive
link between motivation and behavior.

Motivation
Psychologists posit that motivation is a core construct
and a component of most human activity. The psychological processes inherent in motivation consist of individuals being moved into action because of a need or
desire that is at least partially unfulfilled or below
expectations. This discrepancy initiates action in
a particular direction or toward a goal. The intensity
of effort is dependent on the goal. While goals serve as
the key driver for action, goal difficulty and importance
are associated with the psychological environment that
exists at the time, which consists of perceptions, knowledge, skills, abilities, and attributions.

Self-Regulatory Self-Efficacy

There are many motivational theories related to
learning, each with its own description of the constructs. Despite the variations in terminology, many
similarities exist. In general, motivation influences an
individual’s behavior in regard to direction, effort, persistence, and accomplishment of goals. Motivation is
not a stable trait but is dynamic in that it is affected by
both internal and external forces, varies among and
within individuals, and is context dependent. As such,
an individual’s motivation may vary as a function of
subject matter domains and even by instructors and
classroom situations.
Regulation of motivational strategies is an important facet of self-regulated learning. Similar to the regulation and selection of cognitive strategies appropriate
for accomplishing a learning task, individuals also regulate motivational strategies that influence choice,
effort, and persistence. In addition, the particular motivational strategies adopted may also influence the use
of specific cognitive strategies. Self-regulated learning is
often considered to be the synthesis and utilization of
both cognitive and motivational strategies.
As described, self-regulation finds its basis in various theoretical frameworks; therefore, many of the
processes may be referred to by different terms. The
general context remains the same, however. Selfregulation is a systematic process of aligning behavior,
cognitive strategies, goal establishment, self-assessment
and feedback to attain a desired learning or achievement goal. The processes of self-regulation are interrelated and function in a cyclical manner, whereby
feedback from previous experiences has been found to
influence current efforts.

Important Scientific Research and
Open Questions
While there has been a great deal of research focusing
on self-regulation, there is still confusion as to the key
components of the construct and related subprocesses.
Work continues to focus on determining these processes as well as the distinct steps or phases involved
in self-regulated learning. There has also been much
discussion on the development of self-regulatory skills,
and the demographic, cultural, and socioeconomic
implications that may influence acquisition of these
skills. Additionally, self-regulatory processes are linked
with content domains; therefore, investigations continue to focus on the nuances of self-regulation in

S

3021

various content areas, and the transferability of these
skills to unique situations. As studies in this construct
progress, it may be beneficial to refine research designs
and statistical techniques in order to capture the
dynamic interactions of the processes of self-regulated
learning.

Cross-References
▶ Academic Motivation
▶ Metacognition and Learning
▶ Motivation to Learn
▶ Self-Regulated Learning

Further Reading
Boekarts, M., Pintrich, P., & Zeidner, M. (Eds.). (1999). Handbook of
self-regulation. New York: Academic.
Pintrich, P. (Ed.). (1995). Understanding self-regulated learning. San
Francisco: Jossey-Bass.
Schunk, D. H., & Zimmerman, B. J. (Eds.). (1994). Self-regulation of
learning and performance. Hillsdale: Lawrence Erlbaum.
Zimmerman, B. J., & Schunk, D. H. (Eds.). (1989). Self-regulated
learning and academic achievement: Theory, research and practice.
New York: Springer.

Self-Regulation of Learning
▶ Learning Strategies

Self-Regulatory Learning
▶ Self-Regulated Learning

Self-Regulatory Processes
▶ Cognitive and Affective Learning Strategies

Self-Regulatory Self-Efficacy
▶ Self-Efficacy for Self-Regulated Learning

S

3022

S

Self-Respect

Self-Respect
▶ Self-Esteem and Learning

Self-schema
The term self-schema refers to the beliefs and thoughts
people have about themselves in order to organize
information about the self. Self-schemas are generalizations about the self that are abstracted from past
experiences and acting in a present situation. Selfschemas correspond to a large extent with one’s selfconcept which results from perceptions of oneself in
terms of aptitudes, competencies, and values. Thus,
self-schemas vary from person to person but within
a person they are relatively stable and persistent.

Further Reading
Cervone, D., & Pervin, L. (2008). Personality theory and research.
Hoboken: Wiley.
Markus, H. (1977). Self-schemata and processing information about
the self. Journal of Personality and Social Psychology, 35, 63–78.

Self-Worth
▶ Self-Esteem and Learning

Semantic Classification
▶ Concept Formation: Characteristics and Functions

Semantic Knowledge
Conceptual knowledge, or knowledge about the meaning or understanding of information. Semantic knowledge is not linked to any specific episodic event, but
rather is similar to procedural knowledge in that it is
typically accessed without conscious awareness. An
example of semantic knowledge would be the ability
to indicate if an apple is a fruit or a vegetable.

Semantic Memory in Profound
Amnesia
BRENDAN D. MURRAY, ELIZABETH A. KENSINGER
Department of Psychology, Boston College,
Chestnut Hill, MA, USA

Synonyms
Category learning; Memory loss

Definition
Semantic memory refers to memory for general factual
knowledge: for example, knowing the definitions of
words, the names of world leaders, and so forth.
Semantic memory can be contrasted with episodic
memory, which refers to memory for specific personal
events: a meeting we went to last Tuesday, what we had
for breakfast this morning, or a movie we saw last
month, for example. Semantic information differs
from episodic information in that semantic memory
is devoid of context; although we can recall specific
who/what/when/where details about our vacation to
France (episodic), no such details need come to mind
when we retrieve the meaning of a French word or
indicate that the Eiffel tower is in France (semantic).
The memory for this semantic information does not
require us to remember the context in which we learned
the information and, in fact, we often cannot remember the details associated with our acquisition of this
type of general factual knowledge.
Semantic memories can be further divided into personal and public semantic memories. Public semantic
memory is knowledge that certain public events
occurred: knowing that Barack Obama was elected president of the United States, for example. Personal semantic memory is knowledge that personal events occurred:
knowing the city in which you were born, or knowing
that you graduated from college, for example. These
memories are semantic, rather than episodic, because
they are stripped of context: knowing that Barack
Obama was elected does not require that you remember
how many votes he received, where he gave his acceptance speech, and so forth. Similarly, knowing that you
are a college graduate probably does not require that you
recollect your graduation day, and knowing the city in
which you were born does not indicate that you can

Semantic Memory in Profound Amnesia

remember the episodic details of your first days of life.
However, there is some disagreement over the extent to
which personal semantic memories are devoid of episodic details. Some researchers argue that personal
semantic memories (e.g., knowing that you graduated
from college) necessarily entail some episodic detail or, at
least, are the product of some episodic knowledge (e.g.,
you know you are a college graduate because you
attended the ceremony and received the diploma).
Other researchers, however, state that it is possible to
recall such knowledge about oneself in the absence of
any memory for the contextual detail.
One way to examine the relation between episodic
and semantic information is to look at how the two are
affected in conditions such as amnesia. Amnesia refers
to a state in which memory for information or experience is damaged or lost. Amnesia can arise from either
organic means (such as traumatic head injury or
▶ neurodegenerative disease) or from psychogenic
means (such as memory loss from depression).
Organic types of amnesia, which this article will focus
on, are often the result of damage to parts of the brain
called the ▶ medial temporal lobes.
A distinction is made between anterograde amnesia and retrograde amnesia. In anterograde amnesia,
the ability to acquire new information is impaired,
but access to information that was already learned
remains intact. In retrograde amnesia, access to information that had been previously learned is impaired
or lost, although new learning can still occur. It is
believed that the most common form of amnesia is
anterograde with some degree of retrograde amnesia;
these patients are profoundly impaired at learning
new information, and often show some (lesser) degree
of impairment for recollecting old information. This
article will focus on this latter type of amnesia as most
of the extant scientific literature describes this type of
amnesia.
Medial temporal-lobe amnesia can differentially
affect the acquisition and recollection of episodic and
semantic information. Typically, these amnesic patients
show profound impairment for both learning and
retrieval of episodic detail (although the magnitude of
the retrieval deficit is debated), and also some impairment for learning new semantic information (such as
the names of people who became famous after the
onset of amnesia). However, the retrieval of semantic
information is usually relatively preserved.

S

3023

Theoretical Background
Semantic memory can be probed in a number of ways,
one of which is to test naming ability. The Boston
Naming Test, for example, asks people to name objects
depicted by line drawings. These items may range from
easy to name (e.g., “bed” or “tree”) to difficult (e.g.,
“protractor” or “abacus”). Vocabulary tests, which ask
people to either define words or to select synonyms for
words, are also frequently used. General world knowledge may also be tested. The Comprehension subtest of
the Wechsler Adult Intelligence Scale, for example, asks
participants what should be done if one finds a sealed,
stamped, and addressed envelope lying on the sidewalk.
While both acquisition and retrieval of episodic
information appear to be profoundly impaired in
amnesic patients, acquisition of semantic knowledge
is typically more impaired than retrieval. Indeed,
amnesic patients often perform normally (or, no
worse than they did before their amnesia) on tests of
verbal or lexical intelligence: providing definitions,
completing sentences, generating synonyms, and generating exemplars in a category (for example, given the
category “zoo animals,” being able to name animals
typically found in zoos), to name a few.
However, patients with amnesia have shown varying
degrees of difficulty in learning new semantic information. In 1988, John Gabrieli and colleagues showed that
a group of amnesic patients could not learn the meaning
of English words that they did not know prior to their
amnesia. Bradley Postle and Suzanne Corkin showed
that if new vocabulary words were only encountered
infrequently, amnesic patients showed no retention of
those words. Conversely, William Hirst and colleagues
describe a patient, C.S., who was able to learn a new
language despite having anterograde amnesia. Other
researchers have shown that patients with anterograde
can learn computer terminology even with no prior
computing experience, and other patients show the ability to recognize famous persons who were not famous
before the onset of amnesia.
Although researchers may disagree about the degree
to which semantic learning is impaired in amnesia,
there is agreement that the acquisition of new semantic
information is a slower and more laborious process for
those with amnesia than for those without it. This
behavioral dissociation between impaired semantic
learning and relatively spared semantic memory
retrieval suggests that the medial temporal lobes are

S

3024

S

Semantic Memory in Profound Amnesia

necessary for normal semantic learning but not necessary for the retrieval of semantic information (as will be
discussed in more detail in the following section).

Important Scientific Research and
Open Questions
Much of the important scientific research on amnesia
has come from patients with damage to medial temporal lobes. Possibly the most famous such case is that of
patient H.M., who in 1953 had his medial temporal
lobes removed to try and treat his severe epilepsy, which
had not responded to any other type of treatment. As
later reported by H.M.’s surgeon, William Beecher
Scoville, and neuropsychologist Brenda Milner, H.M.
exhibited profound anterograde amnesia, as well as
limited retrograde amnesia for events in the few years
preceding his surgery. H.M. could no longer commit
personal events to memory, and he showed impairment
in learning new semantic information. However, H.M.’s
retrieval of semantic information was relatively intact.
In 2001, Elizabeth Kensinger, Michael Ullman, and
Suzanne Corkin showed that on tests of semantic memory – assessing vocabulary, general world knowledge,
the ability to identify famous historical figures, etc. – H.
M. performed as well as age- and education-matched
control subjects. Given the same tests over the course of
several decades, H.M.’s performance did not decline,
further evidencing that his ability to retrieve semantic
knowledge was not disrupted by his surgery. Martial
Van der Linden and colleagues also demonstrated preserved semantic memory retrieval in patient A.C, who
was able to name persons who were famous before the
onset of his amnesia as well as control participants
could.
Importantly, H.M. and A.C., as well as other amnesic patients, also demonstrate some ability to learn new
semantic information. H.M., for example, could learn
that Woody Allen was an actor and comedian, even
though Woody Allen did not become famous until
after H.M.’s surgery. In the mid-1990s, A.C. could
identify the current president of the United States and
Prime Minister of Belgium (Bill Clinton and Jean-Luc
Dehaene), even though those men were not known
political figures at the onset of A.C.’s amnesia in 1977.
The magnitude of information learned by medial temporal-lobe amnesic patients is usually far from normal,
however, and so current research is examining what
features of the information may affect whether or not

amnesic patients can retain the new knowledge and
what processes may support the learning in these
patients.
In terms of the information that can be learned, one
possibility is that amnesic patients have an easier time
learning new semantic information when that information is personally relatable and can be anchored to
preexisting knowledge. As noted above, some
researchers had little success in teaching patients new
vocabulary words. However, Brian Skotko and others
showed that H.M. was able to learn that a vaccine
developed by Jonas Salk (which was not produced
until after H.M.’s surgery) could prevent polio because
polio was still common when H.M. was young and
therefore personally relevant. Similarly, A.C.’s family
indicated that he was always very interested in politics,
which may have made the names of new political figures easier to retain in memory. To this end, it has been
suggested that the presence of ▶ schemas – mental
constructs that help to organize information – for
certain domains of information may facilitate semantic
learning in that domain. Amnesic patients may retain
schemas for information that was well known to them
prior to the onset of their amnesia and that is personally relevant, and they may be able to anchor new
information to those pre-existing schemas.
The cognitive and neural processes that support the
learning and retrieval of semantic information are still
under investigation. At a cognitive level, it may be that
amnesic patients cannot learn semantic information
through the one-trial learning processes that assist
healthy individuals. Amnesic patients may need extensive repetition of information to learn the new knowledge, and it is even possible that they are benefiting
more from ▶ implicit learning of the information than
from ▶ explicit learning. At a neural level, there is some
evidence that the ▶ lateral temporal cortex and
▶ parahippocampal gyrus may play a role. These
regions were intact in patients H.M. and A.C. but
were damaged in patient E.P., tested by Peter Bayley
and Larry Squire. E.P. could not demonstrate any new
semantic learning; however, E.P.’s damage was so diffuse as to make it difficult to conclusively ascribe
semantic learning to these regions.

Cross-References
▶ Amnesia and Learning
▶ Categorical Learning

Semantic Networks

▶ Categorical Representation
▶ Neural Mechanisms of Extinction Learning and
Retention
▶ Retention and Learning

Further Reading
Bayley, P. J., & Squire, L. R. (2002). Medial temporal lobe amnesia:
gradual acquisition of factual information by nondeclarative
memory. The Journal of Neuroscience, 22, 5741–5748.
Hirst, W., Johnson, M. K., Phelps, E. A., & Volpe, B. T. (1988). More
on recognition and recall in amnesics. Journal of Experimental
Psychology. Learning, Memory, and Cognition, 14, 758–762.
Postle, B. R., & Corkin, S. (1998). Impaired word-stem completion
priming with novel words: evidence from the amnesic patient H.
M. Neuropsychologia, 36, 421–440.
Scoville, W. B., & Milner, B. (1957). Loss of recent memory after
bilateral hippocampal lesions. Journal of Neurology, Neurosurgery
and Psychiatry, 20, 11–21.
Skotko, B. G., Kensinger, E. A., Locascio, J. J., Einstein, G., Rubin,
D. C., Tupler, L. A., Krendl, A., & Corkin, S. (2004). Puzzling
thoughts for H.M.: can new semantic information be anchored
to old semantic memories? Neuropsychology, 18, 756–769.
Van der Linden, M., Bredart, S., Depoorter, N., & Coyette, F. (1996).
Semantic memory and amnesia: a case study. Cognitive Neuropsychology, 13, 391–413.

Semantic Networks
PABLO PIRNAY-DUMMER, DIRK IFENTHALER,
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany

Synonyms
Active structural network; Associative networks; Conceptual graphs

Definition
Semantic networks are a logic-based formalism for
knowledge representation. Semantic networks are graphs
which are constructed from both a set of vertices (or
nodes) and a set of directed and labeled edges. The
vertices or nodes represent concepts, and the edges represent semantic relations between the concepts. Knowledge about accepted meanings should be processed in
adjacent regions of the semantic network. Therefore,

S

3025

semantic networks are often termed “associative networks.” They directly address issues of information
retrieval, since the associations between concepts
define access paths for traversing a structured knowledge base. Another important feature of this kind of
representation schema is the availability of organizational principles, such as generalization, instantiation,
and aggregation. A third characteristic feature is the
graphical notation that enhances their comprehensibility. Indeed, the chief advantage of graph notations
is the ability to show direct connections between
vertices.
Semantic networks are widely applied in the field
of Artificial Intelligence (especially language processing) but they also found acceptance as psychological
models of the human semantic memory. However, they
must be distinguished from concepts maps which are
defined as external network structures aiming at the
organization of knowledge for teaching and learning
(Novak 1998).

Theoretical Background
The idea of depicting semantic structures, which
include concepts and relations between concepts, has
its source in two fields: semantics (especially propositional logic) and linguistics. From a psychological
point of view, semantic networks are representation
schemas which operate on the hypothesis, that intelligent systems memorize knowledge by means of specific
propositions about the conceptual information that is
inherent in the issue to be remembered. More specifically, semantic networks are schemas for representing
declarative knowledge which is considered as a set of
facts that can be represented as a data structure. The
facts describe elements of the world (objects, events),
relations between the elements as well as states of the
elements.
At the beginning, there was Quillian’s (1968) work
in Semantic Memory. Quillian used a network structure
to model human verbal memory. His idea was to get
a “humanlike use” by processing a network of associations with a spreading activation search. The nodes of
Quillian’s networks represent concepts which were
described by their relations with other concepts.
Quillian distinguishes several types of relations, such
as set operators (subclass, modification), semantic
cases (subject, object), and logical operators (and, or,
not). Some years later, semantic networks had been

S

3026

S

Semantic Networks

developed also by Anderson and Bower (1973) as
a representation schema for modeling human associative memory, and similarly, Norman and Rumelhart
(1978) operated with semantic networks in modeling
semantic memory. A major characteristic of these early
approaches as well of follow-up approaches (e.g., Aebli
1981) was the “linguisticizing” of semantic networks to
adjusting their notation to linguistics (Mylopoulos and
Levesque 1984). This corresponds with the idea of
Richens in 1956 to use semantic networks as an
“interlingua” for the computation of natural languages.
The emergence of Artificial Intelligence as a prospering
field of informatics then saw a rapid increase in the
development of and research on semantic networks as
“tools” of conceptual modeling in cognitive science,
machine learning, and programming languages (see
for an overview Brodie et al. 1984). Today, semantic
networks come in such a wide variety of forms and are
used in so many ways that it is difficult to describe what
is common to all of them. Actually, many different
representation schemas are called semantic networks
but they minimally share a common notation based on
graph theory.

Features of Semantic Networks
In its most basic form, a semantic network represents
declarative knowledge in terms of a collection of objects
(nodes) and binary associations (directed labeled
edges), the former standing for individuals (or concepts of some sort), and the latter standing for binary
relations over these. According to this view, a knowledge structure or knowledge base is a collection of
objects and relations defined over them, and modifications to the knowledge base occur through the insertion or deletion of objects and the manipulation of
relations. Representation of knowledge by means
of a semantic network corresponds with a graphical
representation where each node denotes an object or
concept, and each labeled edge (n1, R, n2) indicates
that (n1, n2) 2 R with R being one of the
relations used in the knowledge structure. Despite the
differences between semantic networks, three types of
edges are usually contained in all network representation schemas: (a) Generalization (is-a, ako) connects a
concept with a more general one (e.g., bird – is a –
animal). The generalization relation between concepts
is a partial order and organizes concepts into a hierarchy. (b) Individualization (instance of, member of)

connects an individual (token) with its generic type
(e.g., Laura – instance of – bird). (c) Aggregation
(part-of, has-as-part) connects an object with its attributes (parts, functions) (e.g., wings – part of – bird).
Another method of organizing semantic networks
is partitioning which involves grouping objects and
elements or relations into partitions that are organized
hierarchically, so that if partition A is below partition B,
everything visible or present in B is also visible in
A unless otherwise specified. Partitions are useful in
representing time, hypothetical worlds, and belief
spaces.

Development of Semantic Networks
Despite their differences, semantic networks conform
to their foundation on predicators (i.e., predicateargument structures) which can be depicted in conceptual graphs. A predicate is a general function which
specifies the relations which exist between a certain
number of domains (or arguments): P (Domain1,
Domain2, Domain3, . . . Domainn). The term predicate
refers to the function itself, whereas the term domain
(or argument) refers to the modality of concepts which
can be applied within predicates. Predicates refer to
coherences and concepts to individuals of a domain.
The concepts can be considered as constants that take
the place of the variables of the predicates. Thus, concepts create depictions of single events or states of the
cognized world.
Actually, the nodes of semantic networks are
abstract parameters which appear at that place where
the set of relations coincide to constitute a concept.
Usually, the nodes and relations are denoted by expressions of natural language that constitute a form of
information and must also be represented in some
way. This linguistic information is often enclosed in
a vocabulary. If a concept represented by a node has
a name in terms of natural language then a relation
traverses to exactly the content of the vocabulary that
contains the applicable words. This can be illustrated as
in Fig. 1.
The interpretation of a relation which connects
two nodes depends on both the label and the direction of passing through the relation. As a consequence, it is possible to apply various measures
from graph theory to track the development of knowledge representations (see for example Ifenthaler
et al. 2010).

Semantic Networks

S

3027

Vocabulary
Ball
Hit
Name

House
John

Name

Type

Agens

< >

Object

Type

< >

Mary
Name
Type
Agens
Type

Object

< >

Type

Instrument

Type

Type
Name

Type

(

)

Agens

Object

(

)

Type

Semantic Networks. Fig. 1 Example of a semantic network with vocabulary (Rumelhart and Norman 1978, p. 55)

Important Scientific Research and
Open Questions
Graph-based knowledge representation is undoubtedly
the most popular modality of representing semantic
knowledge in current cognitive science and informatics
(Chein and Mugnier 2009). On the one hand, semantic
networks and conceptual graphs aim at modeling
semantic memory due to the basic assumption that
information can be represented with the help of active
structural networks (Norman and Rumelhart 1978).
On the other hand and due to their origins in linguistics, semantic networks find a widespread application
in the field of natural language processing in Artificial
Intelligence.
In early semantic networks the nodes stood for
linguistic terms. This resulted in a propagation of
types of edges and, in consequence, complicated their
processing. A solution was found by (a) a limitation of
edge types to few but semantically exactly defined relations and depiction of content-related edges through
nodes, and (b) by the introduction of so-called semantic primitives. They play an important role within the
model of Conceptual Dependency Structure of Schank
(1980) which, however, is not a semantic network.
Rather, it operates with conceptualizations which attribute actions to cases.
Since their emergence in the 1970s, semantic networks experienced manifold variations and elaborations.

Although there is no standard procedure for developing
semantic networks, most of them agree on the general
principles as described, for instance, in detail by
Sowa (1984) who furthermore introduced a taxonomy
of network representational schemas:
(a) Definitional networks are based on is-a relations
between the nodes and organize a generalization
or subsumption hierarchy with its major characteristics of inheritance of features from types to tokens.
(b) Assertional networks are designed to assert propositions. They often serve as models of the conceptual structures underlying the semantics of natural
languages.
(c) Implicational networks operate with the relation of
implication for connecting nodes. They may be used
to represent beliefs, causalities, or simple inferences.
These three network representation schemas
can be found in nearly each field of application in
artificial intelligence and psychology. However,
network representation schemas in the field of AI
usually go beyond these types of semantic networks
because intelligent systems, natural or artificial,
must be able to draw inferences and to learn.
Therefore, the taxonomy of network representational schemas include also the following:
(d) Executable networks which contain mechanisms
for passing messages, searching for patterns and

S

3028

S

Semantic Networks

associations, and most notably for performing
inferences. Often these inferences are based on
default reasoning.
(e) Learning networks develop or extend their representations by acquiring knowledge from examples.
The resulting new declarative knowledge may
change the former network by adding and deleting
nodes and edges or by modifying numerical values,
so-called weights that are associated with the nodes
and edges.
Finally, there are hybrid networks which combine
two or more of the previous techniques, either in a
single network or in separate, but closely interacting
networks.

Learning Networks
From the perspective of Artificial Intelligence, learning
systems are necessary for adjusting representational
schemas to changing environments. More specifically,
learning networks adaptively respond to new information by modifying its representations so that the intelligent system, natural or artificial, is capable to adjust
more effectively to its environment. Systems that use
network representations can modify the networks in
three ways: First, by means of rote learning that simply
adds new information to an available semantic network
without substantial changes to it. Second, in some
networks the vertices and edges have weights that can
be changed. In implicational networks, weights might
represent probabilities which could increase due to
repetitions of the same type of network. Third, learning
can aim at restructuring a network with the aim to
basically change the overall structure of the network.
This form of learning is certainly the most complex
form of learning networks and also the most difficult
because the extent and quality of structural changes are
infinite.
Since the seminal work of Collins and Quillian
(1969) who investigated the interaction between a
quite simple structure for semantic networks and complementary processes the research on the development
of semantic exploded. Models of associative semantic
networks regularly contained some kind of spreadingactivation process and were widely applied to predict
performance in experimental memory retrieval tasks as
well as to explain priming and interferences in information processing (Anderson 2000). As a result of this

research the processes involved in the formation and
search of semantic memory are well understood now
whereas research on the large-scale structure of semantic memory and its interactions with knowledge acquisition and retrieval is at its beginning. Steyvers and
Tenenbaum (2005), for example, explored the largescale structure of three types of semantic networks
(word associations, WordNet, and Roget’s Thesaurus)
and could show that these semantic networks have
a small-world structure which is characterized by
sparse connectivity, short average path lengths, and
strong local clustering. Furthermore, Steyvers and
Tenenbaum described a simple model for semantic
growth, in which each new concept is connected to
an existing network by differentiating the connectivity
pattern of an existing node. This model also provides
a mechanism for the effects of learning history and
performance in semantic tasks.
Another line of current research on semantic networks is concerned with the comparison of ontology
with semantic networks for knowledge representation
(e.g., Salem and Alfonse 2008).

Cross-References
▶ Activityand
Taxonomy-Based
Representation
▶ Concept Maps
▶ Knowledge Representation
▶ Mental Representation
▶ Representation, Presentation, and
Schemas

Knowledge

Conceptual

References
Aebli, H. (1981). Denken: das Ordnen des Tuns. Band II: Denkprozesse.
Stuttgart: Klett-Cotta.
Anderson, J. R. (2000). Learning and memory: An integrated approach
(2nd ed.). New York: Wiley.
Anderson, J. R., & Bower, G. H. (1973). Human associative memory.
Washington, DC: Winston.
Brodie, M. L., Mylopoulos, J., & Schmidt, J. W. (Eds.). (1984).
On conceptual modelling. Perspectives from artificial intelligence,
databases, and programming languages. New York: Springer.
Chein, M., & Mugnier, M. L. (2009). Graph-based knowledge
representation: Computational foundations of conceptual graphs.
New York: Springer.
Collins, A. M., & Quillian, M. R. (1969). Retrieval time from
semantic memory. Journal of Verbal Learning and Verbal Behavior, 8, 240–248.
Hartley, R. T., & Barnden, J. A. (1997). Semantic networks: Visualizations of knowledge. Trends in Cognitive Science, 1(5), 169–175.

Semantic Technologies and Learning
Ifenthaler, D., Pirnay-Dummer, P., & Seel, N. M. (Eds.). (2010).
Computer-based diagniostics and systematic analysis of knowledge. New York: Springer.
Mylopoulos, J., & Levesque, H. J. (1984). An overview of knowledge
representation. In M. L. Brodie, J. Mylopoulos, & J. W. Schmidt
(Eds.), On conceptual modelling. Perspectives from artificial
intelligence, databases, and programming languages (pp. 3–17).
New York: Springer.
Norman, D. A., & Rumelhart, D. E. (1978). Gedächtnis und Wissen. In
D. A. Norman & D. E. Rumelhart (Hrsg.), Strukturen des Wissens.
Wege der Kognitionsforschung (S. 21–47). Stuttgart: Klett-Cotta.
Novak, J. D. (1998). Learning, creating, and using knowledge: Concept
maps as facilitative tools in schools and corporations. Mahwah:
Lawrence Erlbaum.
Quillian, R. (1968). Semantic memory. In M. Minsky (Ed.), Semantic
information processing. Cambridge, MA: MIT Press.
Richens, R. H. (1956). Preprogramming for mechanical translation.
Mechanical Translation, 3(1), 20–25.
Salem, A-B. M., & Alfonse, M. (2008). Ontology versus semantic
networks for medical knowledge representation. In Proceedings
of the 12th WSEAS international conference on computers
(pp. 769–774). Heraklion, 23–25 July 2008.
Schank, R. C. (1980). Language and memory. Cognitive Science,
4, 243–284.
Sowa, J. F. (1984). Conceptual structures: Information processing in
mind and machine. Reading: Addison-Wesley.
Steyvers, M., & Tenenbaum, J. B. (2005). The large-scale structure of
semantic networks: Statistical analyses and a model of semantic
growth. Cognitive Science, 29, 41–78.

Semantic Technologies and
Learning
THANASSIS TIROPANIS1, HUGH C. DAVIS1, STEFANO A. CERRI2
1
Web and Internet Science group, Electronics and
Computer Science, University of Southampton,
Southampton, UK
2
LIRMM: Laboratory of Informatics, Robotics and
Microelectronics, University Montpellier2 & CNRS,
Montpellier Cedex 5, France

Synonyms
Knowledge representation and reasoning and learning;
Linked data and learning; Semantic web and learning

Definition
Semantic technologies provide for the expression of
meaning of resources such as content, software, systems, people, and communities in machine processable

S

3029

formats with the help of ontologies. Ontologies provide
the vocabulary for describing resources and their relationships (Gruber 1993; Musen 1992). Based on the
meaning of resources, semantic technologies can draw
conclusions and enable better resource discovery and
matching.
In a learning context, the potential of semantic
technologies translates to a number of affordances.
The types of resources involved in formal or informal
learning environments could include learners, teachers,
software, services and content. Given the description
(or annotation) of these resources with the help of
ontologies, it is possible to support more efficient discovery of learning resources, sequencing of relevant
learning material, adaptation of learning experiences
to match learner profiles, formation of groups for
teaching learning activities, collaborative construction
of knowledge, recommendation of learning resources
and activities, and assessment. A number of ontologies
are developed in the learning domain in order to support these applications with the help of reasoning
software.

Theoretical Background
Semantic technologies for learning draw from work in
Knowledge Representation and Reasoning and have
been significantly influenced by developments in the
Semantic Web community (Berners-Lee et al. 2001).
The Semantic Web vision has been considered “inevitable” and its impact on education significant (Ohler
2008).
With the help of ontologies, semantic technologies
can support knowledge representation and reasoning.
There are a number of ontology languages to address
the tradeoff between how expressive an ontology language is and the performance of reasoning based
on resources annotated with this ontology language.
Resource descriptions in more expressive ontology languages could potentially let us perform advanced reasoning and draw more conclusions about resources,
at the cost, however, of increased computation and
potentially reduced performance.
For example, some ontology languages let us use
transitive properties when describing how resources
relate to each other, which can provide for more powerful reasoning. Consider an ontology for the domain
of language education. We could identify concepts such
as student and class and relationships that express that

S

3030

S

Semantic Technologies and Learning

a student attends one class only or that a student is
a classmate of another student; one could also specify
that the relationship “is a classmate of ” its transitive
(i.e., if A is a classmate of B and B is a classmate of C then
A is a classmate of C). Given this ontology, one can
annotate (mark up) specific resources as instances of
ontology concepts. For example, one could mark up
each of “Alice,” “Bob,” and “Eve” as instances of a
student and “French” as an instance of a class; one
could also state that “Alice” is a classmate of “Bob,”
that “Bob” is a classmate of “Eve,” and that “Eve” attends
“French.” Given this annotation and the ontology, it is
possible to draw conclusions with the help of software;
in this example, software could infer that “Alice” attends
“French” and that “Alice” is a classmate of “Eve” even
if this information is not explicitly stated in the
annotation.
The power of knowledge representation and reasoning bears significant promise for learning. In addition, the well-formed description of resources with the
use of lightweight or more expressive ontologies bears
potential for interoperability and data integration. This
has led to research on the use of semantic technologies
in a number of areas related to learning.

Semantic Technology Affordances
The promise of semantic technologies in learning has
been identified as an important aspect of Technology
Enhanced Learning (TEL) in both formal and informal
learning settings. The ability of semantic technologies
to match people, content, and communities that are
involved in learning processes is central in all these areas.
Matching learners to learning resources has led to
work on the integration of semantic technologies
Learning Management Systems (LMS), Learning Content Management Systems (LCMS), or Virtual Learning Environments (VLE). The role of semantic
technologies in these domains is on discovering,
sequencing, or adapting learning material based on
the learner profile and learning outcomes. Semantic
technologies have been considered in the context of
collaborative building of knowledge and learning material by means of collaborative ontology building, topic
maps and semantic wikis. Collaborative ontology
building tools have also been developed to that end.
The use of semantic technologies to match learners
to other learners or teachers has been considered in
group-formation systems in formal learning and

informal learning settings. In vocational training,
semantic technologies have been proposed to form
groups of learners with similar or supplementary competencies. In informal learning, successfully matching
learners based on their profiles and their learning
objectives using semantic technologies has been
pursued.
Another significant advantage of semantic technologies is in matching content and people involved in
learning processes on a large scale: a larger pool of
people and content increases the chances of accurate
matching. In this respect, the semantic description of
resources available in repositories of reference material
has been used for advanced content discovery. Semantic applications to support critical thinking and argumentation by enabling the discovery of material in
support or against a certain argument are also being
explored. At the same time, aggregating and processing
information about people and their activity using
semantic technologies has led to work on expert
matching tools and query answering. Ontologies for
describing relationships among people such as the
Friend-of-a-Friend (FOAF) ontology enable additional
tools on discovering people and supporting informal
learning processes.
Semantic technologies can support a number of
processes in formal education institutions such as
assessment. In the environment of assessment semantic
technologies can help in preparing appropriate assessment material that matches certain assessment criteria
or in the analysis of text that is submitted as part
of assessment; there are examples of use of semantic
technologies for both summative and formative
assessment.
Semantic technologies can also support wellformed description of resources in a formal education
environment and address interoperability and data
integration requirements inside and across education
institutions. In the context of formal education, the
data interoperability and integration affordances of
semantic technologies can support efficient systems
for curriculum design, development of learning material, and collaboration among the people involved in
teaching and learning activities.

Building Semantic Technologies
Different approaches can be followed when it comes to
developing semantic technologies in a specific domain.

Semantic Technologies and Learning

One can consider a top-down approach where the use
of an ontology or a set of ontologies are agreed first
before the annotation of resources and the deployment
of applications. An alternative approach gives priority
to exposing data sources using lightweight vocabularies
first, in order to enable data interoperability and integration before considering more complex ontologies in
the context of specific applications. The latter view is
central in the linked data movement (http://linkeddata.
org/), which proposes practices for exposing and
connecting structured data on the Web (Bizer et al.
2009). The Linked Data movement supports the exposure of data sources in lightweight vocabularies such as
RDF (http://www.w3.org/RDF/) in order to allow for
the emergence of intelligent applications that will make
use of the exposed data.
The value of semantic technologies and linked data
in terms of well-formed description of resources, data
interoperability, and data integration in the UK higher
education sector has been highlighted in a number of
reports (Tiropanis et al. 2009). The types of semantic
technologies that are related to learning and which can
benefit from the linked data approach include:
● Collaborative authoring and annotation tools
● Searching and matching tools based on linked data

resources (e.g., expert finders)
● Repositories and VLEs that import or export their

data in linked data formats to support learning
resource discovery
In parallel to the Semantic Web vision, there
are proposals for a Pragmatic Web (http://www.
pragmaticweb.info), in which the description of the
meaning of resources is based not just on the semantic
constructs and their relations as defined in ontologies,
but also on the context in which resources are found
and used. This contextual information could provide
for more accurate mechanisms for learning resource
discovery and use.

Important Scientific Research and
Open Questions
Research on semantic technologies in learning has been
developing in several directions. A non-exhaustive list of
such directions and open research questions includes:
● Collaborative knowledge representation and main-

tenance in learning environments

S

3031

● Supporting personal and group knowledge con●

●
●
●
●
●

●

struction with semantic technologies
Establishing the underpinning pedagogy of semantic applications or introducing pedagogy in semantic applications
Knowledge extraction on resources related to
learning
Transition from taxonomies and hierarchies to
more expressive ontologies for learning
Semantic applications for critical thinking and
argumentation
Exploring the potential of semantic technologies for
exploratory learning and problem-based learning
Investigating the potential of sharing and combining linked data inside or across educational
institutions
Semantic technologies to capture and enhance
informal learning processes

There are systems that can support personal or
collaborative knowledge construction, which are often
Wiki-based or Topic-map based. In many cases, the
constructed knowledge is not available in machine
processable formats or formats that are appropriate
for integration with other relevant applications. This
opens siginificant questions on the feasibility, scalability, and efficiency of semantic technology based
approaches to learning resource authoring and to
data interoperability and integration for learning systems. In addition, the potential of semantic technologies to support pedagogy in technology-enhanced
learning requires further research. Semantic applications seem promising in supporting critical thinking
and argumentation given the availability of relevant
resources and applications. Applications that combine
resources described using lightweight vocabularies on
a large scale, across educational institutions and organizations, could potentially support novel processes
for formal, informal, exploratory, and problem-based
learning.

Cross-References
▶ Collaborative Knowledge Building
▶ Knowledge Acquisition: Constructing Meaning from
Multiple Information Sources
▶ Knowledge Organization
▶ Knowledge Representation
▶ Ontology and Semantic Web

S

3032

S

Semantic Web and Learning

▶ Ontology Development and Learning
▶ Ontology of Learning Objects Repository for
Knowledge Sharing

References
Berners-Lee, T., Hendler, J., & Lassila, O. (2001). The Semantic web.
Scientific American, 284(5), 29–37.
Bizer, C., Heath, T., & Berners-Lee, T. (2009). Linked data-the story so
far. International Journal on Semantic Web and Information Systems, 5(3), 1–22.
Gruber, T. R. (1993). A translation approach to portable ontology
specification. Knowledge Acquisition, 5, 199–220.
Musen, M. A. (1992). Dimensions of knowledge sharing and reuse.
Computers and Biomedical Research, 25, 435–467.
Ohler, J. (2008) The Semantic web in education. Educause Q, 31(4).
http://net.educause.edu/ir/library/pdf/EQM0840.pdf
Tiropanis, T., Davis, H., Millard, D., Weal, M., White, S. and Wills, G.
(2009). Semantic technologies in learning and teaching
(SemTech) – JISC Report. London, England: Joint Information
Systems committee (JISC). http://www.jisc.ac.uk/media/documents/projects/semtech-report.pdf

The term “semiotics” can be traced back to the
ancient Greek semeiοn (semeion), which means “sign”
or “index.” As the theory of signs, semiotics investigates
the intentional use of signs to represent something – for
instance using a red light to represent (and communicate) danger. Semiotics as a discipline is concerned with
the use of all forms of signs: from human gestures to
different forms of animal communication and information processes within living cells or molecules. The
relevance of semiotics for the study of learning results
from the fact that virtually all learning takes place
through text and speech and presupposes the use of
linguistic signs and symbols. In addition, in many
learning situations, picture-like signs are also used for
representation and communication purposes. Accordingly, semioses (i.e., sign processes) are considered to
be an important ingredient for cognition, learning, and
teaching.

Theoretical Background
"

Semantic Web and Learning
▶ Semantic Technologies and Learning

Semiotics and Learning
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany

Synonyms
Sign systems; Symbolism

Definition
Cognitive psychology and conceptions of cognitive
learning argue that human learning and thinking
necessitate computational processes on symbolic notations of mental representations and the use of signs and
sign systems. Semiotics is the study of signs as elements
of representation and communication, and therefore
cognitive learning and semiotics are closely related with
each other.

“Let me tell you first what I and my friends thought the
[cognitive] revolution was about there in the late 1950s
. . . It focused upon the symbolic activities that human
beings employed in constructing and in making sense
not only of the world, but of themselves” (Bruner
1990, p. 2).

Accordingly, cognitive and educational psychologists, such as Aebli, Ausubel, Salomon, and many
others, have recognized the central importance of sign
processes (semioses) for human learning and thinking.
It has been argued that learning and thinking depends
not only on the availability of domain-specific knowledge and cognitive skills but also on individuals’ semiotic competence in using specific sign systems for
representational purposes (Seel and Winn 1997). This
argumentation corresponds to the basic assumption in
cognitive psychology that thinking is a process of symbol manipulation which enables learners to form and
express subjective experiences, ideas, thoughts, and
feelings. Accordingly, the question of how knowledge
about facts, events, actions, and plans is represented by
means of signs and sign systems has been recognized as
central to the construction of theories of human
cognition.
However, the idea that cognition presupposes the
capacity to represent knowledge by means of signs and
rules for manipulating these signs has a longer history

Semiotics and Learning

than cognitive psychology. It can be traced back to the
ancient Stoics and the Indian Nyaya Sutras (approximately 200 BC) and can also be found in the work of
John of Saint Thomas, Locke, Leibniz, Peirce, Morris,
and others who were concerned with semiotics and all
forms of signs that can serve communicational and
representational purposes. The basic understanding of
the use of signs has been described by Goodman
(1968), who explains that the introduction and use of
a sign is always a definition by an individual who
intentionally takes something as a specific sign in
order to represent something else. More specifically,
Goodman argues that (1) anything can be a representation of something else; (2) it is the cognitive system
that takes something as a representation of something
else; and (3) it follows from (1) and (2) that a representation is always what a cognitive system defines or
takes intentionally as a representation. Nothing can be
a representation without the intentional activity of
representing. Several decades earlier, Külpe (1923)
and Cassirer (1929) had developed similar conceptions
by pointing out that a world of self-constructed signs is
faced with so-called objective reality. As ways of worldmaking (Goodman 1968), semioses presuppose representational means as signs that refer to the objects to be
represented.
Traditionally, semiotics differentiates between the
intentional use of a sign as an index, icon, and/or
symbol, and it usually locates the prototype of a sign
in verbal language. This basic distinction has been
adapted to some extent in the area of cognitive psychology. For instance, Bruner distinguishes between
enactive, iconic, and symbolic forms of representation.
A good example for enactive representation is dancing,
which may serve as a nonverbal language for imagining
and learning. Anderson (1983) distinguishes between
temporal strings, spatial images, and propositions; and
Johnson-Laird (1983) makes a distinction between
images, mental models, and propositions. Besides
some differences in detail, there is agreement that
there are picture-like (“images”) and language-like
(propositions) forms of representation.
Some decades earlier, Piaget (1959) introduced
a particular epistemological view on semiotic functions
he considered necessary for cognition since they provide the means of mentally representing objects and
structures in the real and imagined world. According to
Piaget, thinking presupposes that cognitive operations

S

follow through on their commitment to perceptions
and concrete actions. They are executed mentally on
the basis of gesticulatory, pictorial, linguistic, or other
(for instance symbolic) signs that function as media of
thinking. This requires the functional incorporation of
communication codes into the space of mental representations. This internalization, or adaptation of the
external for internal use (Salomon 1979), involves
abstracting and transposing imitative actions into
so-called semiotic functions comprising the totality
of denotation functions that are separated from the
content of perception and recognition of objects. Semiotic functions include all imaginative and depictive
actions of the human mind. They are not components
of the thinking process itself but rather aids for
representing knowledge about the world. Naturally,
the emergence of semiotic functions follows a developmental sequence. At the beginning of this sequence are
imitative social behaviors and at the end are linguistic
signs. Accordingly, the development of semiotic functions starts with imitation and symbolic play, followed
by delayed imitation (i.e., imitation in absence of the
model),“drawings, painting, modeling, and the use of
internal images, and culminating in verbal language
and the use of linguistic signs” (Piaget 1959). The role
of verbal language as a thinking tool has been discussed
for a long time in philosophy and psychology. Generally, it can be argued that thoughts can be expressed
through language and that internal speech supports
thinking. However, an individual who uses language
and linguistic signs for knowledge representation must
be able to store and retrieve these signs just as knowledge. Therefore, it is possible to differentiate between
two stages of language use: In the first, language is used
to communicate automatically, whereas the second
requires that the individual knows explicitly how to
operate with language in order to represent and communicate knowledge.
Piaget’s ontogenetic view on semiotic functions
presupposes the internalization as well as the mental
manipulation of signs. Basically, internalization has
two aspects, namely: (a) a figurative aspect based upon
the continuous transposition of imitation activities
into picture-like or language-like signs, and (b) an
interpretative aspect which assigns meanings to signs.
On the whole, the process of internalization constitutes
the fundamental basis for mental representations
by bringing about the transition from concrete

3033

S

3034

S

Semiotics and Learning

experiences to imagination. Representational codes
and semiotic functions evolve in accordance with the
individuals’ capability to differentiate between a sign
and its denotation, between meaning and phenomenon, between linguistic expressions and their content.
Evidently, the use of signs and symbols follows the
gradual developmental of intellectual capacities. This
can be illustrated by ▶ The Harvard Project Zero: Following a period of mundane symbol mastery in
infancy, young children rapidly attain initial competence in dealing with a broad range of sign systems in
their culture (e.g., language, gestures, pictures, numbers, and music) mediated by everyday life experiences.
By school age, the major task is the mastery of notations (for instance, in mathematics) that entail learning
features of reduction, systematicity, and legibility. In
consequence, there are authors who aim to create multiple semiotic worlds in the classroom (Hicks 1995) by
focusing on the use of different signs and symbols for
representation and communication purposes.

Important Scientific Research and
Open Questions
In the course of the past four decades, semiotics has
established itself as an independent scientific discipline
that can be divided into a philosophical theory of communication concerned with the creation of sign processes for the acquisition and mediation of knowledge
and an information theory which considers semioses as
natural processes and thus as natural objects of empirical investigation. However, from the point of view of
philosophical semiotics the principal decision to follow
an empirical approach is subordinate to the traditional
reflective approach.
The role of sign processes in mental representation
was a central issue of the so-called imagery debate of the
1980s in cognitive psychology. The controversy had to
do with the question as to whether mental representations and computation are only based on the use of
abstract symbols or also on picture-like images. The
major result of the imagery debate was that there is
now general agreement among cognitive psychologists
on the point that humans primarily (but not exclusively)
use picture-like and language-like signs for representation and communication purposes. This theoretical
conception has been elucidated recently in Mayer’s
framework of cognitive learning as well as in the cognitive load theory.

Parallel to the imagery debate, educational psychologists were concerned with the idea of internalization
and transposition of communication codes for representation purposes. These researchers focused especially on the interaction between media, cognition,
and learning (Salomon 1979). The central idea of this
research (for an overview see: Seel and Winn 1997) was
that individuals prefer to represent information mentally in the same modality which was originally used to
communicate it to them. In consequence, it should be
possible to influence the acquisition and application of
different sign systems for representing knowledge in
dependence on specific experiences by using signs in
communication situations with specific characteristics.
This argumentation culminated in Salomon’s (1979)
supplantation hypothesis, which assumes that the sign
systems and coding elements of communication may
affect information processing and related semioses by
way of two basic mechanisms. Either the sign systems
used in communication activate preexisting mental
skills and cultivate skill mastery through exercise, or
they overtly supplant mental skills into which the signs
and codes of communication (such as pictorial and
linguistic signs) are then functionally incorporated
through modeling. Accordingly, the signs used by
delivery systems and media can, once internalized, be
used as tools for thought.
Another important line of research on semioses
followed the assumption that individuals acquire semiotic competence in the course of school learning.
Indeed, from the perspective of semiotics, instruction
does not aim only at the acquisition of knowledge and
cognitive skills but rather also involves the functional
incorporation of signs and symbols for representation
and communication purposes. This raises the question
as to whether and how we can facilitate semioses
through instruction. Indeed, an important goal or side
effect of school learning consists in the mediation of
particular symbol systems. “Children progress through
a regular sequence of stages in their transition from
dependence on concrete material to the ability to apprehend the meaning of abstract propositions presented
symbolically . . . At this stage of development, therefore,
properly arranged reception of symbolic material is
highly meaningful” (Ausubel and Robinson 1969,
p. 109). This can be illustrated by learning in the classroom in specific disciplines, such as mathematics,
physics, chemistry, biology, and so on, where students

Semiotics and Learning

have to learn certain symbolic systems and models
that are often radically different from their everyday
experiences. Greeno (1989) provided some examples
to demonstrate that, unsurprisingly, individuals in
fact do not use the symbol systems taught in schools
when they have to solve everyday problems. This corresponds to Bruner’s (1968) assumption that individuals can use alternative sign systems for knowledge
representation. The use of a particular sign system
for representing and communicating knowledge is primarily a matter of adaptation to a particular case – but
not in the classroom where the particular demands
of the various disciplines constrain the use of sign
systems.
In the area of subject matter learning, cognitive
processes are traditionally separated from the manipulation of symbolizations or inscriptions, as the example
of mathematical learning illustrates (Meira 1995).
Recent research in the field of mathematics learning
considers mathematizing as a form of modeling that
usually involves using specialized languages, symbols,
graphs, pictures, concrete materials, and other notational systems to develop mathematical descriptions
and explanations that obviously make great demands
on learners’ representational capabilities (Lesh and
Doerr 2000). Similarly, research in the area of biology
and biochemistry learning shows that experts can be
defined especially well by the specialized symbol systems they use to represent and communicate domainspecific knowledge (Kindfield 1999).
On the whole, research on semioses within the
realm of subject matter learning is still in its infancy.
This holds also true with regard to the interplay
between semiotics, cognitive theory, and learning in
virtual environments as investigated initially by Winn
et al. (1999).

Acknowledgments
In memory of William D. Winn (1945–2006)

Cross-References
▶ Cognitive Learning
▶ Cognitive Load Theory
▶ Dancing: A Nonverbal Language for Imagining and
Learning
▶ Imagery and Learning
▶ Information Literacy

S

3035

▶ Knowledge Representation
▶ Learning Numerical Symbols
▶ Linguistic Factors of Learning
▶ Mental Imagery
▶ Multimodal Learning Through Media
▶ Psychosemiotic Perspective of Learning
▶ Role Play and the Development of Mental Models
▶ Supplantation Effect

References
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.
Ausubel, D. P., & Robinson, F. G. (1969). School learning. An introduction to educational psychology. New York: Holt.
Bruner, J. S. (1968). Toward a theory of instruction. Cambridge:
Belknap.
Bruner, J. (1990). Acts of meaning. Cambridge, MA: Harvard University Press.
Cassirer, E. (1929). Études sur la pathologie de la conscience
symbolique. Journal de Psychologie normale et pathologique, 26,
289–336 & 523–566.
Goodman, N. (1968). Languages of art. An approach to a theory of
symbols. Indianapolis: Bobbs-Merrill Comp.
Greeno, J. G. (1989). Situations, mental models, and generative knowledge. In D. Klahr & K. Kotovsky (Eds.), Complex
information processing (pp. 285–318). Hillsdale: Lawrence
Erlbaum.
Hicks, D. (1995). Discourse, learning, and teaching. Review of
Research in Education, 21, 49–95.
Johnson-Laird, P. N. (1983). Mental models. Towards a cognitive
science of language, inference, and consciousness. Cambridge:
Cambridge University Press.
Kindfield, A. C. H. (1999). Generating and using diagrams to learn
and reason about biological processes. Journal of Structural
Learning and Intelligent Systems, 14(2), 81–124.
Külpe, O. (1923). Vorlesungen über Logik. Leipzig: Hirzel.
Lesh, R., & Doerr, H. M. (2000). Symbolizing, communicating, and
mathematizing: Key components of models and modeling.
In P. Cobb, E. Yackel, & K. McClain (Eds.), Symbolizing and
communicating in mathematics classrooms. Perspectives on discourse, tools, and instructional design (pp. 361–383). Mahwah:
Lawrence Erlbaum.
Meira, L. (1995). The microevolution of mathematical representations in children’s activity. Cognition and Instruction, 13,
269–313.
Piaget, J. (1959). La formation du symbole chez l’enfant. Imitation, jeu
et rêve – image et représentation. Neuchâtel: Delachaux et Nestlé
S.A.
Salomon, G. (1979). Interaction of media, cognition and learning. San
Francisco: Jossey Bass.
Seel, N. M., & Winn, W. D. (1997). Research on media and
learning. Distributed cognition and semiotics. In R. D.
Tennyson, F. Schott, S. Dijkstra, & N. M. Seel (Eds.), Instructional design: International perspectives (Theories and models of

S

3036

S

Semi-supervised Learning

instructional design, Vol. 1, pp. 293–326). Mahwah: Lawrence
Erlbaum.
Winn, W., Hoffman, H., & Osberg, K. (1999). Semiotics, cognitive
theory and the design of objects, actions and interactions in
virtual environments. Journal of Structural Learning and Intelligent Systems, 14(1), 29–49.
Further Reading
Cobb, P., Yackel, E., & McClain, K. (Eds.). (2000). Symbolizing and
communicating in mathematics classrooms. Perspectives on discourse, tools, and instructional design. Mahwah: Lawrence
Erlbaum.

Semi-supervised Learning

Sensitization: Incremental
Plasticity
▶ Habituation and Sensitization

Sensori-Emotional Learning
▶ Aesthetic Learning

Sensorimotor Adaptation

Obtaining labelled data for supervised learning can be
costly, but often large amounts of unlabeled data can be
obtained cheaply. Semi-supervised learning exploits
both at once and is useful when only limited labelled
data is available.

OTMAR BOCK
German Sport University, Institute of Physiology and
Anatomy, Köln, Germany

Cross-References

Synonyms

▶ Supervised Learning

Motor learning

Definition

Senior Citizens and Health
Learning
▶ Empowering Health Learning for the Elderly (EHLE)

Sense-Making
▶ Learning as Meaning Making

Sensitive Period
▶ Neuroeducational Approaches on Learning

Sensorimotor adaptation is the ability to gradually
modify our motor commands in order to compensate
for changes in our body and in the environment. Examples from everyday life include adaptation to the visual
distortions caused by prescription glasses, to the unfamiliar driving characteristics of a rental car, and to the
changes of limb length and muscle force during maturation. Adaptation can be understood as a form of
motor learning, accomplished by the modification of
existing behavior rather than by the acquisition of new
behavior. It is achieved by evaluating sensory information about performance errors, and using it to induce
changes at two hierarchical levels: a lower level
recalibration of sensorimotor pathways, and higher
level adjustments of motor strategies. The Latin word
“adaptare” means to fit together, and is derived from the
Greek word “aptos” = appropriate.

Theoretical Background

Sensitization
▶ Perceptual Learning

In 1867, Helmholtz was the first to report that subjects
wearing wedge prisms before their eyes will initially
misreach for visual targets, but will overcome their
errors with practice. This phenomenon spurred

Sensorimotor Adaptation

vigorous research by experimental psychologists in the
1950s and 1960s, and became popular again since the
1990s as an easily observable manifestation of neuronal
plasticity. It was found that adaptation is not limited to
prismatic displacements and other visual distortions;
rather, subjects adapted as well to acoustic, tactile,
▶ proprioceptive, and mechanical (i.e., force field) distortions. Furthermore, adaptation was found for different types of motor response, such as pointing,
grasping, and tracking movements of the arm, saccadic
eye movements, walking, and balancing. In modern
research, adaptation is typically evaluated by asking
subjects to point at visual targets while visual feedback
of their hand is distorted. Readers can easily replicate
this paradigm by trying to point at objects on
a computer screen with a computer mouse that is
rotated 90 clockwise: at first, they will dramatically
misdirect the mouse pointer, but their performance
will gradually improve within about 20 min.
Adaptive improvement under a distortion does not
follow a simple exponential function, but rather is
characterized by at least three time constants, in the
order of seconds, minutes, and weeks. These time constants reflect distinct mechanisms, with different functional properties. Once the improvement is completed,
it is possible to observe various aftereffects: generalization of adaptive behavior to new locations in the
workspace and to new movement types (e.g., from
pointing to tracking), transfer of adaptation to another
limb or to a new distortion, persistence of adaptive
behavior even if it is no longer adequate since the
distortion was removed, retention of the adaptive
change such that it emerges again when the same distortion is reintroduced up to 1 year later, and
metalearning of the ability to adapt (e.g., easier adaptation to left-right reversal following an adaptation to
up-down reversal).
Adaptive improvement is achieved both by
a recalibration of sensorimotor pathways, and by strategic adjustments such as postural changes, associative
learning of stimulus–response pairs, and cognitive
reinterpretations of sensory feedback. In contrast,
aftereffects are due to recalibration alone, since strategies are effortful, situation-specific, and short-lived,
and therefore quickly dissipate when no longer needed.
This differential role of strategies is reflected by the
differential effect of several factors on adaptive
improvement versus aftereffects. Thus, old age and

S

3037

poor cognition were found to degrade adaptive
improvement but not aftereffects, while a happy
mood facilitated improvement but not aftereffects;
this indicates that emotions, cognitions, and aging are
more closely related to strategies than to recalibration.
In contrast, dopamine depletion in Parkinson’s disease degraded improvement and aftereffects, which
suggests that the dopaminergic system is involved in
recalibration.
The most popular distortion in adaptation research
is a rotation of visual feedback about the starting point.
Adaptation to this rotation is achieved by a gradual
rotation of an internal reference frame from its original
orientation through intermediate angles up to the
required value. In consequence, adaptation to a given
angle of rotation facilitates the subsequent adaptation
to a larger rotation, and hinders the subsequent adaptation to an opposite rotation. The internal reference
frame cannot be rotated by more than 90 ; adaptation
to larger distortions is achieved by an abrupt axis inversion (which corresponds to a 180 rotation), followed
by a gradual “backward” rotation of the internal reference frame.

Important Scientific Research and
Open Questions
What type of error is used to drive adaptation? Error
information about an ongoing movement could be
derived from visual and from proprioceptive signals.
It has been shown that either modality can drive adaptation, but if both are available, proprioception is
suppressed unless it provides relevant information
not available in the visual system. Other work
documented that adaptive changes are based on the
discrepancy between intended and observed movements (i.e., sensory prediction errors), rather than
between observed movements and targets (i.e., target
errors). Since sensory prediction errors are processed
mainly in the cerebellum, it is not surprising that adaptation depends critically on the integrity of that brain
structure. However, it remains unclear whether the
cerebellum passes the error information on to adaptive
mechanisms located elsewhere, or rather is itself the site
of adaptive change. Since the adaptation deficits of
cerebellar patients are fully explainable by their motor
control deficits, the cerebellum seems not seem to
possess separate mechanisms devoted only to adaptation and not to motor control.

S

3038

S

Sensorimotor Adaptation

Where in the brain is adaptation located? Most neuroimaging studies attempted to capture adaptationrelated activation as the difference between activations
in an adaptation task and in a distortion-free control
task, and found that differential activity to be widely
scattered throughout the brain. However, the adaptation task differed from the control task not only by the
presence of adaptation but also by the presence of
movement errors, and the reported activity therefore
included error-related processes such as surprise, attention focusing, and corrective eye and arm movements.
Studies using more adequate control tasks yielded
much less activation, mainly in the inferior parietal
cortex, the dorsolateral premotor cortex, and a small
area in the superior cerebellum, which is known to be
connected with the other two.
Is adaptation based on one common, or on multiple,
task-specific mechanisms? Adaptation achieved in a
given sensorimotor task was found to transfer to a different sensory modalitiy, movement type, end-effector
(e.g., arm to arm, and arm to eyes), and distortion (e.g.,
position-dependent to velocity-dependent, and visual
to mechanical). This indicates that different sensorimotor tasks can access a common adaptive mechanism.
However, the magnitude of transfer was often moderate
or small, which suggests that access to the common
mechanism is limited. From this, it has been concluded
that adaptation is based neither on one common
nor on multiple task-specific, but rather on multiple
overlapping mechanisms. This work is relevant for rehabilitation, which builds upon a transfer from therapeutic
interventions to real life.
Can we adapt to several distortions at the same time?
We can adapt to two distortions if each is associated
with a different arm, arm posture, target location, gaze
direction, or background color. This has been taken as
evidence that adaptation is localized in multiple modules, which the sensorimotor systems can access
depending on contextual cues. It remains open how
many of those modules exist, and thus, to how many
distortions we can concurrently adapt.
What cognitive functions contribute at which time
during adaptation? This question can be answered by
a dual-task approach, with subjects adapting while
concurrently engaging in a cognition-demanding task:
interference between both tasks could be taken as evidence that they compete for a common pool of
resources. An alternative approach is to correlate each

subject’s adaptive progress with their cognitive performance scores established in a separate session. Both
approaches have been rarely used in literature, but
preliminary evidence suggests that the cognitive
demand changes during adaptation, quantitatively
and qualitatively.
Do astronauts adapt to weightlessness? The absence
of gravity changes the functioning of several sensory
modalities, and modifies the behavioral outcome of
motor commands; it thus represents a combined sensory and motor distortion, to which astronauts must
adapt. Indeed, movements executed in weightlessness
are slower, less accurate, and/or cognitively more
demanding than on Earth, and these deficits gradually
diminish with time. Gross motor tasks such as postural
responses normalize within days, while fine motor
tasks such as handwriting are not fully compensated
even after several months of spaceflight. Early adaptation of fine motor performance draws on cognitive
resources related to sensory functions, while late adaptation requires predominantly resources related to
movement planning. The pronounced resource-dependence, as well as the scarcity of aftereffects, suggests that
much of the adaptation to weightlessness is achieved by
strategic adjustments rather than by recalibration.

Cross-References
▶ Adaptation to Weightlessnes
▶ Aging Effects on Motor Learning
▶ Sensorimotor Schema

References
Bastian, A. (2008). Understanding sensorimotor adaptation and
learning for rehabilitation. Current Opinion in Neurology, 21,
628–633.
Bock, O., Abeele, S., & Eversheim, U. (2003). Human adaptation to
rotated vision: interplay of a continuous and a discrete process.
Experimental Brain Research, 152, 528–532.
Bock, O., & Girgenrath, M. (2005). Relationship between sensorimotor adaptation and cognitive functions in younger and older
subjects. Experimental Brain Research, 169, 400–406.
Krakauer, J., Ghilardi, M.-F., Mentis, M., Barnes, A., Veytsman, M.,
Eidelberg, D., & Ghez, C. (2003). Differential cortical and subcortical activations in learning rotations and gains for reaching:
A PET study. Journal of Neurophysiology, 91, 294–333.
Redding, G., & Wallace, B. (1996). Adaptive spatial alignment and
strategic perceptual-motor control. Journal of Experimental Psychology: Human Perception and Performance, 22(2), 379–394.
Shadmehr, R. (2004). Generalization as a behavioral window to the
neural mechanisms of learning internal models. Human Movement Science, 23, 543–568.

Sensorimotor Schema

Sensorimotor Learning
▶ Anticipatory Learning
▶ Learning to Sing Like a Bird: Computational Developmental Mimicry
▶ Song Learning and Sleep

Sensorimotor Schema
FRANK GUERIN
Department of Computing Science, University of
Aberdeen, Aberdeen, Scotland, UK

Synonyms
Enactive or action schemas; Infant habitual action

Definition
Within constructivist theories, the sensorimotor
schema is held to be the principal unit of knowledge
in use during infancy. A sensorimotor schema is
a psychological construct which gathers together the
perceptions and associated actions involved in the performance of one of the habitual behaviors in the
infant’s repertoire. The schema represents knowledge
generalized from all the experiences of that behavior. It
includes knowledge about the context in which the
behavior was performed as well as expectations about
the effects. Sensorimotor schemas are central to Jean
Piaget’s explanation of infant development.

Theoretical Background
The development of the psychological construct known
as the sensorimotor schema has its origins primarily in
the work of Jean Piaget and, in particular, in his books
on sensorimotor development (Piaget 1936, 1937).
Sensorimotor schemas are a special case of the more
general notion of schemas. Like schemas they are
knowledge structures which represent the generic
knowledge abstracted from a number of experiences
in a similar situation. Sensorimotor schemas are simply
those schemas which are in use during the sensorimotor period (from birth to roughly the end of the second
year). A typical example of a sensorimotor schema is
the schema of shaking a rattle. This schema gathers

S

together the visual and tactile perceptions of the rattle,
as well as the motor actions involved in the shaking,
and the expectation of visual and auditory perceptions
during the execution of the action. Because it is generic,
the same schema could be said to operate in the infant’s
interactions with a class of objects which are broadly
similar to the rattle, and where the infant’s action is
broadly similar. If the infant is presented with a new
object, and the infant seems to treat it like the rattle,
shaking it in a similar way, then Piaget would say this
new object had been assimilated to the schema of the
rattle. If the rattle-schema is adjusted to account for
slight differences in this new object, then Piaget would
say the rattle-schema had accommodated to the new
object. If the accommodation is significant, then Piaget
would say that the schema had been differentiated to
produce a new schema appropriate to the new situation. In other situations, it may be that two independent schemas are found to be closely correlated and
become combined as a new (higher-level) schema. In
this way, Piaget describes the whole of infant development from the perspective of the operation of sensorimotor schemas. In comparison with the schemas in use
in adults, sensorimotor schemas tend to be less abstract
and more directly tied to the actions, objects, and
contexts in the infant’s experiences. Nevertheless,
there are major differences in the types of schemas in
use at the start of the sensorimotor period and at its
end, with a clear progression toward more abstract and
higher-level schemas.
Piaget defined six sequential stages during sensorimotor development (Piaget 1936), with a qualitative
difference between the sensorimotor schemas in use in
each stage. The first of these stages is the reflex stage
(roughly the first month), where the reflex actions such
as sucking and grasping, and papillary and palpebral
reactions to light are exercised. Each of these actions is
associated with its own global schema which generalizes from experiences, where the action happens, and
gradually learns to recognize the situations where the
action is triggered, and the expectation of what sensory
impressions arise while the action is in progress. Furthermore, different modes of operation become recognized, for example, the difference between sucking the
nipple for feeding or sucking the fingers.
The second stage (roughly from 1 to 4 months)
begins with schemas which arise from the fortuitous
discovery of the results of actions on the infant’s own

3039

S

3040

S

Sensorimotor Schema

body, for example, the schema of sucking the thumb.
Piaget identifies these schemas with what J. M. Baldwin
has called the circular reactions. In this stage, schemas
are also formed which integrate different modalities,
for example, auditory, tactile, and visual. By the end of
this stage, a schema for reaching for, and grasping, seen
objects has been formed.
During the third stage (roughly from 4 to
8 months), it is again a case of repeating results fortuitously discovered, but these schemas are now of
a higher order because they act on objects in the environment. Piaget calls these the secondary circular reactions. A typical example of a schema discovered at this
stage is that of pulling a string to shake a rattle. Note
that the infant at this stage is incapable of understanding the connection between the string and rattle, and
will not be capable of discriminating, in advance, situations where the string is connected or not connected.
This schema is a global ensemble, incorporating parts
whose connection is not fully understood. During this
stage, there is a rapid growth in the number of schemas
in the infant’s repertoire as new schemas are differentiated from previous ones in order to repeat interesting
discoveries. Example actions include squeezing, shaking, striking, scraping, rubbing, and pulling.
During the fourth stage (roughly from 8 to 12
months), intentional means-end sequences of actions
are performed. For example, the infant will intentionally displace an obstacle in order to retrieve a desirable
object which is visible behind it. In this type of situation, two distinct sensorimotor schemas can be distinguished; one for the means action, which displaces the
obstacle, and one for the end, which is to grab the
desired object. The schema of the end is able to subordinate the means schema, and to direct its operation.
This implies that the sensorimotor schemas must now
incorporate relatively advanced knowledge of the world
because the effect of an action on an object is understood, and also on the relationships between objects
(e.g., the relationship “in front”). In contrast to the
previous stage, at this stage, the schemas can be combined in new situations and are not confined to the
situations where the combination was discovered fortuitously. For this reason, Piaget refers to these as
mobile schemas.
During the fifth stage (roughly from 12 to 18
months), schemas become ▶ experimental; they are
not only repeated, but are intentionally varied so that

the relationships between initial conditions and effects
can be studied. Piaget calls these schemas tertiary circular reactions. This new experimental ability can now
be employed to generate new means schemas when the
infant’s existing repertoire does not happen to have the
appropriate means schema. An example of this is the
effort to retrieve an out-of-reach object with a stick;
even if the infant does not have schemas for performing
the appropriate actions to displace the object, the
required schemas can be discovered through experimental groping.
The sixth stage (roughly from 18 to 24 months)
could be said to not properly belong to the sensorimotor period because there is significant evidence of internal representation of objects, actions, and effects. This
gives rise to covert planning, where rather than groping
in the world, the infant carries out the necessary experimentation mentally and then exhibits a complete correct sequence of actions. This has been achieved by
internalization of the sensorimotor schemas acquired
previously.
Through all this progression, there is a gradual
increase in the abstractness and objectiveness of the
sensorimotor schemas; while the earliest schemas capture a very subjective knowledge locked in particular
contexts, the later schemas have abstracted away from
these contexts and capture knowledge about relationships between objects and actions in the world. In this
development, schemas are a tool for learning because
their effort to repeat and assimilate events leads them to
generalize and gradually embody more abstract information about the world.
The term “sensorimotor schema” is not widely used
in contemporary psychological research. Piaget’s theory paints a picture of a very close coupling between the
development of perceptual competences and action
competences, and even representation and reasoning,
with each being built up in tandem by the same process.
For this reason, the idea of a sensorimotor schema
which combined all elements made sense as the unit
of knowledge. A classic example is in the case of the
means-end action of retrieving a hidden obstacle. Piaget held that it was through experiences with acting
on objects in relationships such as “in front” that
the perceptual competence and representational competence to understand about hidden objects was
constructed. Later research has suggested that perceptual and representational competence may be more

Sensory Curiosity

independent from action competence (Bremner 1994),
and that in most cases perceptual competence precedes
action competence but in some cases the reverse may be
true (Kellman and Arterberry 1998). Although
a consensus has not yet been reached on many details,
the general picture emerging is considerably more
complex than Piaget’s account, and the sensorimotor
schema may be at too coarse a level for its proper
description.

Important Scientific Research and
Open Questions
Since Piaget’s early work on infancy new experimental
methods have been developed to probe infants’ knowledge of the world, most notably the infant habituation
method, and the violation of expectation paradigm.
This measures infants’ looking times to determine
if the event displayed to them violates their expectations. This method has been used to determine such
things as knowledge that objects continue to exist while
occluded. This work has uncovered a curious gap
between the advanced knowledge which infants
seem to demonstrate by looking and the relatively
impoverished knowledge that seems to be available
to them when they are required to act. There is considerable controversy about many of these results,
and especially their interpretation, with some scientists contradicting them outright (Cohen and Cashon
2003).
Neuroscientific evidence suggests that there are in
fact two parallel processing streams for objects; the
ventral or “what” stream, and the dorsal or “action”
stream. The fact that there might be multiple partial
object representations (Mareschal et al. 2007), independent of a particular behavior, suggests the need for
a much more fine grained unit of organization than
Piaget’s sensorimotor schema; although there must be
cells integrating this sensory and motor information,
these cells must be integrating relatively sophisticated
fragments, rather than building up the knowledge from
scratch as in Piaget’s theory.
Within Artificial Intelligence, there have been
some efforts to computationally model sensorimotor
schemas; this requires a representation to capture
schema knowledge. There is a consensus in a number
of works that a schema should be a three-part structure
with an initial state description, and action, and
a predicted resulting state description. There are also

S

3041

ongoing efforts to define algorithms which can exhibit
cognitive development by building higher-level
schemas out of lower-level ones (e.g., Chaput 2004).
Many questions about sensorimotor schemas
remain to be answered. For example, is the sensorimotor schema really a useful concept to serve as a generic
knowledge structure, or is it maybe an approximation
for great diversity of different things which only seem
to share similar features from a “zoomed out” view but
may turn out to be quite different on closer inspection.
Similarly, it is not known if sensorimotor schemas
really do share a common structure with later more
abstract adult schemas (as suggested by Piaget) and if
they are reusing the same machinery in the brain, or if
on closer inspection they may be quite different.

Cross-References
▶ Development and Learning
▶ Developmental Cognitive Neuroscience
Learning
▶ Infant Learning and Development
▶ Schema Development
▶ Schema(s)

and

References
Bremner, J. G. (1994). Infancy. Cambridge, MA: Blackwell.
Cohen, L. B., & Cashon, C. H. (2003). Infant perception and cognition, in Comprehensive handbook of psychology. In R. Lerner,
A. Easterbrooks, & J. Mistry (Eds.), Developmental psychology. II.
Infancy (Vol. 6, pp. 65–89). New York: Wiley.
Kellman, P., & Arterberry, M. (1998). The cradle of knowledge.
Cambridge, MA: MIT Press.
Mareschal, D., Johnson, M. H., Sirois, S., Spratling, M., Thomas, M.,
& Westermann, G. (2007). Neuroconstructivism (How the brain
constructs cognition, Vol. I). Oxford: Oxford University Press.
Piaget, J. (1936). La Naissance de l’Intelligence chez l’énfant.
Neuchâtel: Delachaux et Niestlé.
Piaget, J. (1937). La Construction du Réel chez l’Enfant. Neuchâtel:
Delachaux et Niestlé.
Further Reading
Chaput, H. H. (2004). The constructivist learning architecture: a model
of cognitive development for robust autonomous robots. PhD thesis,
AI Laboratory, The University of Texas at Austin.

Sensory Curiosity
▶ Curiosity and Exploration

S

3042

S

Sensory Learning

Sensory Learning
▶ Visual Perceptual Learning

Sensory Learning Styles
A categorization of learning styles according to the
sensory modalities that are applied for learning information. Generally, four modalities are distinguished:
Visual, Aural (sound), Read/write, and Kinesthetic
(touch). Learners usually have a preference for a certain
sensory style of learning.

Sensory Memory
ZHONG-LIN LU
Departments of Psychology, University of Southern
California, Los Angeles, CA, USA

Synonyms
Sensory registers

Definition
Stimulation of human sense organs is initially
represented in sensory memory for a brief period by
a literal, labile, and modality-specific neural copy. The
term iconic memory stands for the initial representation of visual stimuli, and echoic memory is its counterpart for auditory stimulation (Neisser 1967).
Sensory memory is often contrasted with short-term
memory and working memory which are assumed to
be less modality specific, and all of these are distinct
from long-term memory. In functional terms, sensory
memory is comparable to a register in a computer. In
a human, if you want to know what the last sensory
input was, you examine the sensory register. The
prevailing view of human memory systems is that
there are one or more modality-specific sensory registers in each sensory modality, plus a register or registers
for short-term and/or working memory, plus functionally distinct long-term memory systems for episodic,
semantic, procedural, and perhaps others form of longterm memory.

Theoretical Background
Visual sensory memory, i.e., iconic memory, was first
demonstrated by George Sperling using the partialreport paradigm (Sperling 1960): After a brief presentation of a 3 
 3 or 3 
 4 array of letters, observers often
can report all the letters in any cued row if the cue
occurs immediately after the visual presentation (“partial-report”), even though they can only report four to
five letters when asked to recall all the items in the
display (“whole-report”). Because all rows are cued
with equal probability, reporting all the items in
a randomly cued row implies that observer has access
to all the items in the array at the termination of
the display. The partial-report superiority effect –
performance advantage of partial-report over wholereport – suggests that there exists a fast-decaying iconic
memory that can initially hold at least 9–12 items.
Numerous studies have established that iconic memory
has a large capacity, decays rapidly, and is destroyed by
post-stimulus masking (Coltheart 1980; Gegenfurtner
and Sperling 1993; Neisser 1967; Sperling 1960). The
duration of iconic memory has been estimated to be
about 300–500 ms for young adult observers.
Auditory sensory memory, also called “echoic memory” (Neisser 1967), is essential for integrating acoustic
information presented sequentially over an appreciable
period of time (Crowder 1976; Glanzer and Cunitz
1966; Treisman 1964; Wickelgren 1969). Psychophysical experiments suggest that echoic memory resides at
a central rather than a peripheral site (Massaro 1975).
Electrophysiological investigations of auditory memory functions in animals provide evidence that similar
short-term memory functions are served by sensory
areas of cerebral cortex (Weinberger et al. 1990).
In biological sensory processing, there is a considerable amount of memory – perhaps more aptly called
retention – that is inherent in the processing itself. So,
in a broad sense, sensory memory exists and may differ
for different stimuli in every stage of sensory
processing. For example, visual sensory memory exists
in the receptors and bipolar cells of the retina as an
afterimage. These cells do not constitute a register in
the usual meaning of the word, but they can retain
a representation of the input image long after it has
been removed. The afterimage is a negative of the
original stimulus that is perceived upon subsequent
stimulation. The original stimulus can itself persist in
the retina as “persistence of sensation.”

Sensory Memory

A complete understanding of a sensory memory
system requires specification of (1) its brain location,
(2) the time constant for which it retains information,
and (3) the format (or encoding) of the information it
retains. The magnetoencephalography (MEG) habituation paradigm developed by Lu, Williamson, and
Kaufman (Lu et al. 1992a) offers a means to specify
the first two components of a sensory memory (location
and time constant) and perhaps the third (encoding) as
well. In this paradigm, the same stimulus is presented
repetitively with different repetition rates. The recovery
of the MEG response with increasing inter-stimulus
interval (ISI) serves as a measure of the decay of the
initial activation. In its first application, Lu et al. (1992a)
used auditory stimulation with various repetition rates
to characterize the temporal dynamics of neural activation in the primary and secondary auditory cortices
of human observers in terms of simple exponential
decay functions. They later found that the lifetime of
a cortical activation trace in primary auditory cortex
predicts the lifetimes of behaviorally determined auditory sensory memory of the loudness of a tone for each
individual subject (Lu et al. 1992b). Using the same
paradigm with flickering checkerboard stimuli,
Uusitalo, Williamson, and Seppa (Uusitalo et al.
1996) were able to detect 10 brain locations with different recovery time constants in the visual system. Each
one of these brain locations represents a process with an
inherent retention time, a process that could become
a component of the overall retention time observed in
a psychophysical sensory memory experiment.

Important Scientific Research and
Open Questions
Vision and audition each have several levels of sensory
memory, and constructing experiments to selectively
measure one or another of these memories has been an
elusive goal. There have been many attempts in the
literature to distinguish different forms of visual sensory memory (for review, see Coltheart 1980, Long
1980). One well-recognized distinction is between
eye-specific visual persistence and non-eye-specific
iconic memory. Many have debated interpretations of
results from information-based assessments (e.g., the
partial-report procedure) and some more direct sensory measures of visual sensory memory. Coltheart
(1980) concluded that the partial-report procedure
mostly reflects retinal nonspecific iconic memory

S

3043

(but see Massaro and Loftus 1996). Beyond visual sensory (iconic) and visual short-term memory, deriving
the taxonomy of visual memory systems has not been
possible by means of psychophysical tasks alone.
The taxonomy of auditory sensory memory systems
is probably even more complicated because auditory
signals are subjected to much more complicated processing prior to reaching auditory temporal cortex area
A1 than are visual signals before reaching occipital cortex
area V1. Cowan’s reviews conclude that there are at least
two forms of auditory sensory memory: a very shortlived memory such as the memory for tonal loudness
described above and a second auditory sensory memory
with a duration on the order of seconds (Cowan 1995).
Indeed, it is not obvious that the next-higher auditory
memory can be characterized by a time constant in
seconds. It might be a limited capacity memory based
on retroactive interference – new inputs push older material out of memory. In the absence of new auditory input
(i.e., silence), periods of silence might themselves be
recorded as memory events. This is a much more complex process than simple, exponential passive decay with
time, although it may be difficult to distinguish from
passive decay when new stimuli arrive at a constant rate.

Cross-References
▶ Attention and the Processing of Visual Scenes
▶ Habituation
▶ Human Information Processing
▶ Memory Codes
▶ Perceptual Learning
▶ Short-Term Memory and Learning
▶ Visual Perceptual Learning
▶ Working Memory

References
Coltheart, M. (1980). Iconic memory and visual persistence. Perception and Psychophysics, 27(3), 183–228.
Cowan, N. (1995). Attention and memory: An integrated framework.
New York, NY: Oxford University Press.
Crowder, R. G. (1976). Principles of learning and memory. Hillsdale,
NJ: Erlbaum.
Gegenfurtner, K. R., & Sperling, G. (1993). Information transfer in
iconic memory experiments. Journal of Experimental Psychology:
Human Perception & Performance, 19, 845–866.
Glanzer, M., & Cunitz, A. R. (1966). Two storage mechanisms in free
recal. Journal of Verbal Learning and Verbal Behavior, 5, 351–360.
Long, G. M. (1980). Iconic memory: A review and critique of
the study of short-term visual storage. Psychological Bulletin,
88, 785–820.

S

3044

S

Sensory Plasticity

Lu, Z. L., Williamson, S. J., & Kaufman, L. (1992a). Human auditory
primary and association cortex have differing lifetimes for activation traces. Brain Res, 572(1–2), 236–241.
Lu, Z. L., Williamson, S. J., & Kaufman, L. (1992b). Behavioral
lifetime of human auditory sensory memory predicted by physiological measures. Science, 258(5088), 1668–1670.
Massaro, D. W. (1975). Experimental psychology and information
processing. Chicago, IL: Rand McNally.
Massaro, D. W., & Loftus, G. R. (1996). Sensory and perceptual
storage: data and theory. In E L Ba R A Bjork (Ed.), Memory.
San Diego, CA: Academic Press.
Neisser, U. (1967). Cognitive psychology. East Norwalk: AppletonCentury-Crofts.
Sperling, G. (1960). The information available in brief visual presentation. Psychological Monographs, 74, 1–29.
Treisman, A. (1964). Monitoring and storage of irrelevant messages
in selective attention. Journal of Verbal Learning and Verbal
Behavior, 3, 449–459.
Uusitalo, M. A., Williamson, S. J., & Seppa, M. T. (1996). Dynamical
organisation of the human visual system revealed by lifetimes of
activation traces. Neuroscience Letters, 213, 149–152.
Weinberger, N. M., Ashe, J. H., Metherate, R., McKenna, T. M.,
Diamond, D. M., & Bakin, J. (1990). Retuning auditory cortex
by learning: a preliminary model of receptive field plasticity.
Concepts in Neuroscience, 1, 91–132.
Wickelgren, W. A. (1969). Associative strength theory of recognition
memory for pitch. Journal of Mathematical Psychology, 6, 13–61.

Sensory Plasticity
▶ Visual Perceptual Learning

Sensory Preconditioning
Learning of an association between two stimuli that are
not biologically significant at the time of training. To
show that learning has taken place, it is usually necessary to later train one of the neutral stimuli to predict
an innately significant stimulus and subsequently test
the response to the other neutral stimulus.

Sensory Registers
▶ Sensory Memory

Sentiment
▶ Attitudes – Formation and Change

Sequence Learning
▶ Learning and Consolidation in Autism
▶ Procedural Learning
▶ Task Sequencing and Learning

Sequence Learning Without
Awareness
▶ Implicit Sequence Learning

Sequence Skill Consolidation in
Normal Aging
DANIEL COHEN1, NICHOLA RICE COHEN2,
ALVARO PASCUAL-LEONE1, EDWIN ROBERTSON1
1
Department of Neurology, Berenson-Allen Center for
Non-invasive Brain Stimulation, Beth Israel Deaconess
Medical Center, Harvard Medical School, Boston,
MA, USA
2
Department of Psychology, Volen Center for Complex
Systems, Brandeis University, Waltham, MA, USA

Synonyms

Sensory Register
Component of the information-processing model of
cognitive learning concerned with initial awareness of
information through the senses (vision, hearing, touch,
smell, and taste).

Offline learning; Procedural memory consolidation;
Skill enhancement

Definition
Following practice, a new skill memory continues to be
processed offline during a period of consolidation.
Consolidation stabilizes a memory, making it more

Sequence Skill Consolidation in Normal Aging

resistant to decay from the passage of time, competing
information, or physical changes to the nervous system. For some types of skills, such as the ability to
produce a sequential pattern of movements, consolidation can be expressed as an improvement in skill that
develops between practice sessions, a phenomenon that
has been called “offline learning.” Following an offline
interval, usually at least several hours in duration, an
individual is able to increase the speed, accuracy, or
efficiency in which the learned sequence is executed
compared to a novel or random pattern of movements.
Deficits in consolidation therefore undermine the benefits of practice and training. Recent evidence has
suggested that offline learning is impaired in normal
aging. With an aging population and technological
advances that constantly introduce new skills, understanding the changes in sequence skill consolidation as
a result of normal aging has emerged as an important
topic of research. Possible mechanisms accounting for
impaired consolidation in aging include disruption of
sleep, changes in local synaptic plasticity, as well as
alterations of interhemispheric interactions and other
shifts in neural network dynamics with aging.

Theoretical Background
Sequence learning refers to knowledge about the
underlying predictive structure between temporally
ordered or contingent events (for a review, see
Keele et al. 2003). Everyday activities such as writing,
cooking, typing, and driving can unfold relatively
automatically without attention to the individual
movements that make up these actions. The neural
substrate for sequential knowledge depends on multiple factors, such as the information format (i.e., sensory or motor codes) and reference frames (i.e.,
egocentric or allocentric coordinates) that define the
individual elements of a sequence, as well as the underlying predictive structure in which downstream elements are predicted by preceding elements. For
example, patients with medial temporal lobe pathology
demonstrate normal sequential learning when there is
a first-order predictive structure to the sequence (each
element is predicted with a certain probability by the
immediately preceding element) but show some deficits with higher-order associations (each element is
predicted by combinations of preceding elements).
Recent evidence suggests that there are multiple
mechanisms involved in offline learning depending

S

on the underlying nature of the sequential information
and the corresponding neural networks representing
that information (Robertson and Cohen 2006). For
example, electrophysiological events unique to sleep
may promote strengthening of neural connections
within a memory network for some types of skills.
Sleep may be particularly important for offline learning
in sequential tasks that depend on the hippocampus
(Spencer et al. 2007). Even within the same task, multiple components of a sequential skill may be acquired
in parallel and yet undergo consolidation preferentially
in distinct physiological states. For example, after
learning a motor sequence in response to visual cues
in a finger tapping task, known as the serial reaction
time task (SRTT), offline learning for the spatial layout
of targets (allocentric coordinates) preferentially occurs
over a night with sleep, and offline learning for the
specific series of muscle contractions (egocentric coordinates) preferentially occurs over the waking day
(reviewed in Robertson and Cohen 2006).
Sleep undergoes characteristic changes with the
aging process. Alterations of sleep may represent one
mechanism causing sequence skill consolidation to
decline with normal aging. In healthy elderly individuals, there is a decline in nightly total sleep time, an
increase in the rate of arousal, reduced amount of slow
wave activity (EEG oscillations in the 0.5–4.5 Hz frequency bands and a marker for sleep depth),and
reduced sleep spindles. Slow wave sleep in particular
has been implicated in the neuronal reactivation or
replay of neuronal firing patterns activated during
recent wakefulness, a mechanism in which sleep may
promote use-dependent synaptic plasticity and memory enhancement. In addition, skill learning increases
subsequent slow wave sleep locally in cortical regions
involved in the skill memory network. The induced
slow wave changes are postulated to contribute to synaptic homeostasis, tuning the connectivity within
a skill memory network and improving the signal-tonoise ratio. Sleep spindles have also been implicated in
processes of memory consolidation. Therefore, the
changes in sleep with aging would generally predict
impaired memory consolidation. However, there are
multiple offline learning mechanisms depending on
the anatomical networks involved, and some offline
mechanisms do not require sleep, so it remains possible
that offline learning could be preserved, at least for
some tasks, during the aging process.

3045

S

3046

S

Sequence Skill Consolidation in Normal Aging

Long-term potentiation (LTP) and long-term
depression (LTD) represent core models of local synaptic plasticity and are widely believed to be key molecular correlates of memory processes. LTP/LTD reflect
a cascade of molecular events that unfold over time
following an initial stimulation protocol, ultimately
leading to more or less effective synaptic communication within elements of a neural circuit. Late phases of
LTP/LTD are protein synthesis dependent and act on
the scale of hours, similar to the time course of offline
learning. While it remains unclear whether LTP/LTD or
specific phases of LTP/LTD coincide with behavioral
observations of memory consolidation, these models
make it clear that some mechanisms of synaptic plasticity are impaired with aging. The decline in synaptic
plasticity with age is potentially related to a decline in
positive modulators or an increase in negative modulators (Lynch et al. 2006). One negative modulator of
LTP is adenosine, a product of neuronal energy metabolism, whose concentrations increase in the cerebral
extracellular space with increased wakefulness and
decline with sleep. Adenosine concentration increases
in aging might be possibly explained, at least in part, by
the changes in sleep physiology mentioned above.
Interestingly, in early aging, there appears to be a
regionally specific decline in the capacity to induce
LTP/LTD. Such regional specificity for changes in plasticity would also suggest that offline learning might be
impaired for some types of skills and relatively preserved for others during the aging process.
In addition to changes in synaptic and local plasticity, aging is associated with changes in activity along
distributed bihemispheric brain networks. Functional
neuroimaging studies have shown that older adults,
even when task difficulty is controlled for, tend to
over-recruit brain regions during perceptual, motor,
and cognitive tasks. Older individuals also show a loss
in functional specialization and hemispheric lateralization. These findings suggests a shift in functional brain
network dynamics that some have argued might represent an adaptive compensatory strategy, but more
recent evidence suggests that it might be a manifestation of neural inefficiency and noisier brain activity
(Grady 2008). Sequence learning and consolidation represent dynamic changes across bihemispheric brain networks, and reduced neural efficiency, increased noise, or
other changes in network dynamics with aging might
lead to significant behavioral consequences.

Important Scientific Research and
Open Questions
Offline learning has generally been studied in healthy
young adults below the age of 35 with only recent
attention focused on sequence skill consolidation in
normal aging (Spencer et al. 2007; Brown et al. 2009).
Both of these studies compared healthy older participants with a mean age of approximately 59 years to
a group of healthy young participants with a mean age
of approximately 20 years using modified versions of
the SRTT. The SRTT can dissociate sequence-specific
skill learning from general improvements in
visuomotor reaction time by measuring the reaction
time cost of switching from blocks of patterned or
sequential trials to random trials. Participants may be
told there is an underlying pattern (explicit condition),
but sequence-specific learning can be demonstrated
even when participants are not informed that
a pattern exists (implicit condition). Sleep is required
for offline learning to occur when there is explicit
awareness of the underlying sequence (Spencer et al.
2007; Brown et al. 2009). These studies showed that
older adults retain the ability to learn a new sequential
skill. However, the results suggest that both sleepdependent and sleep-independent offline sequence
learning is impaired in healthy aging.
The study of sequence skill consolidation in normal
aging is at its infancy. Future behavioral work is necessary to determine the properties of skill consolidation
in the elderly across a range of tasks. Structural and
functional imaging as well as noninvasive electrophysiological studies can shed more light on the underlying
changes in connectivity and dynamics in skill memory
networks that occur as we age. More animal data is
required to assess the specific molecular and cellular
mechanisms that account for reductions in local synaptic plasticity with aging. Such work can help to
determine the following: (1) Are there certain types
of sequence skills for which offline learning is completely spared or perhaps even enhanced in aging?
(2) Can pharmacological agents, for example, adenosine antagonists, rescue offline learning in normal
aging? (3) Do specific neuroanatomical changes such
as white matter integrity or functional connectivity
contribute to alterations in consolidation with aging?
(4) Do measures to improve sleep quality, duration, or
depth improve sleep-dependent offline learning in
the elderly?.

Sequential Learning

Cross-References
▶ Aging Effects on Motor Learning
▶ Changes in Memory and Learning Across the
Lifespan
▶ Developmental Cognitive Neuroscience and
Learning
▶ Implicit Sequence Learning
▶ Memory Codes
▶ Memory Consolidation and Reconsolidation
▶ Procedural Learning
▶ Production Systems and Operator Schemas for
Representing Procedural Learning
▶ Reactivation and Consolidation of Memory During
Sleep
▶ Sequential Learning
▶ Verbal Learning and Aging

S

3047

language, music, animal communication, and motor
skills). The cognitive and neural processes involved in
learning about the proper ordering of events and stimuli
are called sequential learning. Sequential learning can be
applied toward encoding a specific sequence in its
entirety (e.g., A-C-E-F) as well as learning the distributional properties among multiple sequences, in which
the learner induces the fragments or underlying regularities common to all exemplars (e.g., noticing that the
fragment A-B occurs in the following three sequences:
A-B-D-E, C-A-B-A, D-A-B-C). Hierarchical sequential
patterns can also be acquired, a type of learning that
may be crucial for much of human cognition and
behavior, especially language processing. Sequential
learning can occur implicitly or explicitly, under supervised or unsupervised conditions, or through any combination thereof.

References
Brown, R. M., Robertson, E. M., et al. (2009). Sequence skill acquisition and off-line learning in normal aging. PLoS ONE, 4(8),
e6683.
Grady, C. L. (2008). Cognitive neuroscience of aging. Annals of the
New York Academy of Sciences, 1124, 127–144.
Keele, S. W., Ivry, R., et al. (2003). The cognitive and neural architecture of sequence representation. Psychological Review, 110(2),
316–339.
Lynch, G., Rex, C. S., et al. (2006). Synaptic plasticity in early aging.
Ageing Research Reviews, 5(3), 255–280.
Robertson, E. M., & Cohen, D. A. (2006). Understanding consolidation through the architecture of memories. The Neuroscientist,
12(3), 261–271.
Spencer, R. M., Gouw, A. M., et al. (2007). Age-related decline of sleepdependent consolidation. Learning & Memory, 14(7), 480–484.

Sequential Learning
CHRISTOPHER M. CONWAY
Department of Psychology, 222 Shannon Hall Saint
Louis University, St. Louis, MO, USA

Synonyms
Sequential processing; Serial learning; Serial order
behavior

Definition
For most higher organisms, the order in which events
occur is of paramount importance (e.g., spoken

Theoretical Background
Although research exploring serial order behavior has
a long tradition, beginning with the early work of
Ebbinghaus, it was Lashley (1951) seminal paper that
elevated the importance of sequential processing in the
psychological sciences. Lashley was one of the first to
fully recognize and appreciate the ubiquity of sequencing in human behavior, thought, and language. For
example, the sentence “please pass me the bottle that
is near the glass,” has a vastly different meaning from
“please pass me the glass that is near the bottle,” despite
the fact that the only difference between the two
sentences is the order in which the words bottle and
glass occur. To take one of Lashley’s own examples, in
English, an adjective generally precedes the noun. “I see
the red ball” is grammatical but “I see the ball red” is
not. In French, the opposite is true.
In addition to language and communication, other
domains in which sequential order is important
include motor and skill learning, music perception
and production, problem solving, and planning.
Clearly, however, not all serial order behaviors are
learned. Many sequential behaviors in humans and
other species appear to be largely pre-wired, such as
grooming movements, simple locomotion, and respiration. Thus, sequential learning encompasses the cognitive and neural mechanisms involved in the process
of acquisition itself, which may be independent from
other issues relevant to serial order behavior more
generally, such as how serial patterns are perceived

S

3048

S

Sequential Learning

(but not necessarily learned), represented in the mind
and brain (whether they were originally pre-wired or
learned sequences), or produced.
There are at least three different modes of sequential learning, each corresponding to a different type of
sequence pattern (Conway and Christiansen 2001):
fixed, statistical or probabilistic, and hierarchical.
Fixed sequential learning is perhaps the simplest
type, involving learning of any arbitrary serial pattern
or list (e.g., A-E-G-K), such as a phone number. In
language, the use of fixed sequences can be found at
different levels, including idioms and stock phrases
(e.g., “once upon a time”) or even words themselves,
which can be construed as fixed sequences of phonemes. This type of sequential learning is informed by
a vast amount of previous research in areas such as list
learning, Hebb repetition effects, short-term memory,
and working memory (e.g., Marshuetz 2005).
Many situations facing humans and other higher
organisms involve more complex patterns that consist
of statistical patterns of co-occurring elements within
multiple exemplars. For example, if one were to consider the entire corpus of English speech that a listener
may perceive over the course of a year, the sound
sequences “fun-ny” and “ro-bot” are each likely to
occur more frequently than “ny-rob.” This is true
because funny and robot are English words, whereas
nyrob would generally only be heard across a word
boundary (e.g., “I see the funny robot”). Being sufficiently sensitive to the frequently co-occurring sounds
in continuous speech, a process known as word segmentation, allows a language learner to induce which
speech sounds constitute a word versus which sounds
just happen to co-occur because they are the final and
initial sounds of two separate words that were spoken
consecutively.
Word segmentation is an example of the more
general phenomenon of statistical or probabilistic
sequential learning, which involves inducing the common underlying distributional patterns from among
multiple exemplars. In language, learners can use the
statistical or probabilistic properties of linguistic input
to discover the structure not just of words, but also of
other properties such as phonology (sound patterns)
and syntax (Saffran 2003). Although language provides
a wealth of examples of statistical sequential learning,
many nonlanguage domains also have probabilistic
structure, in which a given element of a sequence does

not perfectly predict the next element, but rather may
predict it in a probabilistic fashion. Thus, statistical
learning is likely used in many aspects of perceptual
and motor processes, such as learning to imitate the
complex movements of a dancer, or learning the
melodic, rhythmic, and harmonic structures of different styles of music.
The previous two forms of sequential learning –
fixed and statistical – are both similar in that they
involve learning the relationships between adjacent
elements, a process sometimes referred to as
“chaining.” However, as Lashley (1951) made clear,
even though a sequence may appear as a chain of linear
elements, there may exist an underlying hierarchical
structure. For example, in the repeating sequence “13-2-3-1-2-1-3-2-3-1-2. . .,” each item can be followed
by one of two possible items (e.g., “1” is followed by
either “3” or “2” with equal probability). Only by
taking into account the previous context in which an
item occurs can one accurately predict the subsequent
item (e.g., knowing that the “1” is preceded by “2”
allows one to accurately predict “3”). With these
more complex patterns, it is necessary to encode not
just the preceding element in a sequence, but the previous two or more elements. Such a strategy of
encoding long-distance or nonadjacent relationships
among items in a sequence might provide the basis
for hierarchical sequential learning, in which primitive
units are combined to create more complex units
(forming a “chunk”), which in turn can be combined
to create even more complex units in a recursive fashion. Learning hierarchical structure has several advantages over merely learning the “flat” linear structure,
including it being easier to self-repair in the event of
failure and providing easier and more efficient access to
subroutines of sequences (Bapi et al. 2005).
In human language, hierarchical structure is especially evident in the realm of syntax. In the
sentence,“The cat chased the mouse,” there are two
phrases, a noun phrase (“the cat”) and a verb phrase
(“chased the mouse”), with the latter containing
a secondary noun phrase within it (“the mouse”).
Once one has learned that a particular phrase is
a noun phrase (e.g., “the mouse”) and assuming one
has also learned the basic structure of English syntax,
the phrase forms a single unit that can then be inserted
into other situations that require a noun phrase (“The
mouse ate the cheese” or “I see the mouse”). In addition

Sequential Learning

to language, other domains such as complex motor
skills and problem solving can have hierarchical
structure.

Important Scientific Research and
Open Questions
A key question is to what extent other animal species,
such as nonhuman primates, demonstrate equivalent
sequential learning abilities to humans. By exploring
the abilities and limitations that other primates have
for processing sequential information, we can begin to
understand the evolutionary origins of such capabilities in humans as well as the unique aspects of human
sequential learning.
Unfortunately, few studies have provided direct
comparisons between nonhuman primates and
humans. The evidence that is available suggests
a number of commonalities as well as crucial differences between the species (Conway and Christiansen
2001). In terms of fixed sequential learning, primates
appear to be capable of encoding, storing, and recalling
arbitrary fixed sequences consisting of motor actions as
well as visual stimuli, with proficiency comparable to
that of human preschoolers. In terms of statistical
sequential learning, one particular species of primate,
the cotton-top tamarin, appears to have some of the
same basic capabilities, at least on a par with human
infants. However, in the case of hierarchical learning,
there is growing consensus that nonhuman primates
lack human-like proficiency. Most studies have demonstrated that apes and monkeys rarely use hierarchical
routines in their spontaneous and learned actions.
Whether this suggests a genuine cognitive limitation
for learning hierarchical sequential patterns or is due to
other methodological or contextual discrepancies is
not entirely clear. Even so, the apparent species differences especially in hierarchical learning may be one
crucial reason that nonhuman primates are generally
incapable of acquiring human-like language.
In general, the pattern of performance differences
across species suggests that some aspects of human
sequential learning (fixed and statistical) derives from
evolutionarily old cognitive substrates, whereas hierarchical sequential learning may be a more recent development. A related key issue, then, is to understand the
neural bases of sequential learning. In humans, findings from neuroimaging studies indicate that the frontal cortex (e.g., prefrontal cortex, premotor cortex,

S

supplementary motor areas, etc.), subcortical areas
(e.g., basal ganglia), and the cerebellum play essential
roles in sequence learning and representation (Bapi
et al. 2005). The ways in which these different brain
areas interact appear to be complex and may partly
depend on the nature of the task demands. For example, in some cases, the prefrontal cortex appears necessary for learning new sequences while the basal ganglia
only become active once the sequences become wellpracticed. In other cases, learning appears to occur first
in the basal ganglia, which then guides learning in
prefrontal cortex. A third possibility is that the basal
ganglia contribute to reinforcement learning while the
cortex is specialized to handle unsupervised learning
situations.
One issue that neuroscience studies can help resolve
is to what extent the three types of sequential learning
(fixed, statistical, and hierarchical) are in fact distinct
mechanisms. Consistent with the behavioral findings
with nonhuman primates, some evidence does suggest
that hierarchical processing in humans relies on an
evolutionarily newer brain region, Broca’s area
(Friederici et al. 2006). A second issue that neuroimaging studies can help illuminate is whether different
domains such as language, music, and skill learning
are mediated by a common pool of neural regions or
whether domain-specificity exists. The evidence here is
mixed; on the one hand, there is evidence suggesting
that Broca’s area is used across multiple domains
including both music and language and therefore may
act as a kind of “supramodal” sequence processor; on
the other hand, a handful of studies have shown that
sensory-specific brain regions (e.g., occipital cortex)
are involved in the acquisition of sequential patterns
in a modality-specific manner.
Because the majority of sequential learning research
has been devoted to a handful of domains (spoken
language acquisition, motor skill learning, and
music), there is a need to explore connections to
other areas of cognition and behavior. Very few studies
have also investigated how these three types of sequential learning skills develop in humans. One possibility is
that there is a progression of developmental stages,
with fixed sequential learning developing first, then
statistical learning, and finally hierarchical learning.
Another possibility is that the different facets of
sequential learning develop more or less in parallel, or
in an interactive fashion. In the realm of education,

3049

S

3050

S

Sequential Processing

there is a need to further explore how sequential learning is used in areas such as mathematics, reading, and
writing, and whether individual differences in sequential learning may relate to various aspects of cognitive,
linguistic, and educational outcomes.

Serial Position Curve
NEELIMA RANJITH
Union Christian College, Aluva, Ernakulam,
Kerala, India

Cross-References
▶ Analytic Learning
▶ Animal Intelligence
▶ Implicit Sequence Learning
▶ Statistical Learning in Perception

References
Bapi, R. S., Pammi, C., Miyapuram, K. P., & Ahmed (2005). Investigation of sequence processing: A cognitive and computational
neuroscience perspective. Current Science, 89, 1690–1698.
Conway, C. M., & Christiansen, M. H. (2001). Sequential learning in
non-human primates. Trends in Cognitive Sciences, 5, 529–546.
Friederici, A. D., Bahlmann, J., Heim, S., Schubotz, R. I., & Anwander,
A. (2006). The brain differentiates human and non-human
grammars: Functional localization and structural connectivity.
Proceedings of the National Academy of Sciences, 103, 2458–2463.
Lashley, K. S. (1951). The problem of serial order in behavior. In
L. A. Jeffress (Ed.), Cerebral mechanisms in behavior. New York:
Wiley.
Marshuetz, C. (2005). Order information in working memory: An
integrative review of evidence from brain and behavior. Psychological Bulletin, 131, 323–339.
Saffran, J. R. (2003). Statistical learning: Mechanisms and constraints.
Current Directions in Psychological Science, 12, 110–114.

Sequential Processing
▶ Sequential Learning

Serial Learning
▶ Sequential Learning

Serial Order Behavior
▶ Sequential Learning

Synonyms
U-shaped learning curve

Definition
Serial position curve is a “U”-shaped learning curve
that is normally obtained while recalling a list of words
due to the greater accuracy of recall of words from the
beginning and end of the list than words from the
middle of the list. First described by Nipher (1878),
the serial position curve can be defined as a “U-shaped
relationship between a word’s position in a list and its
probability of recall.” This occurs due to a phenomenon known as Serial Position Effect. The serial position
effect consists of two phenomena viz. primacy effect
and recency effect. Primacy effect refers to the better
recall of items from the beginning of list (first three or
four items), whereas recency effect refers to the better
recall of items from the end of the list (last three or four
items) than middle items of the list.

Theoretical Background
The theoretical construct of serial position curve is that
the primacy effect represents recall from a more remote
memory or ▶ long-term memory which is better consolidated than the recency effect which represents recall
from a more recent memory or working memory
(Glanzer and Cunitz 1966). One of the ways to demonstrate serial position curve is by means of using a free
recall task using a verbal learning test such as Rey
Auditory Verbal Learning Test (RAVLT). In this
method, the subject is given a list of various items
(words of the same length) that should be remembered.
The items to be remembered are presented on
a manner of reading the list to the subject. Each item
is introduced at a regular interval. The subject is then
asked recall the items that he/she remembered in any
order. The frequency of recall is plotted against the
position an item takes in a list. The thus obtained
graph, Fig. 1 Nipher (1878), has become known as the
serial position curve of single-trial free recall. Typical is

Serial Position Curve

Probability of recall

1.00

.50

.00

Primacy
effect

1

5

Recency
effect

10

15

Serial position of item

Serial Position Curve. Fig. 1 Serial position curve

that the last and first few items – the recency and
primacy effect – SPE’S – are more readily recalled
than items in the middle of the list, which gives the
graph its typical U shape as represented below. The
primacy effect corresponds to the tail of the U on
the left. It is called the primacy effect because these
items were the ones presented first to the subject in the
memory experiment. The recency effect corresponds to
the tail of the U on the right. It is called the recency effect
because these items were the ones presented most
recently to the subject in the memory experiment.

Important Scientific Research and
Open Questions
A common explanation of the primacy and recency
effects were introduced by Atkinson and Shiffrin
(1968). According to this viewpoint, the primacy effect
is a result of the greater amount of attention and
rehearsal allocated to the first few items on a list. This
advantage in processing given to those items allows
them to be transferred into the long-term memory
store and thus have a higher probability of being
retrieved out of long-term memory. They (Atkinson
and Shiffrin 1968) attributed the recency effect to signify output from what they referred to as primary
memory in the form of a short-term memory buffer.
Thus, the most recent items viewed in a list are still in
short-term memory and are recalled there. Evidence
that the primacy effect is due to a greater amount of
rehearsal to the first few items is clear in a study done by
Rundus (1971). In this study, subjects were asked to
rehearse out loud, and it was recorded. After reviewing
the recordings, Rundus (1971) found that participants
devoted more overt rehearsal to the first few items on
the list. Research done by Glanzer and Cunitz (1966)

S

3051

also showed that primacy is reduced when the items are
presented at a faster rate, thus eliminating opportunity
for extensive rehearsal by the participants.
Research has also been done to demonstrate the use
of short-term memory in explaining the recency effect.
Because the recency effect is explained by a retrieval of
items from short-term memory, it should be eliminated if a person is asked to do another task before
they are asked to recall the items on the list. This was
demonstrated in experiments by both Postman and
Phillips (1965) and Glanzer and Cunitz (1966). Both
studies provided evidence in support of the short-term
memory account for the recency effect by having their
participants perform a “distractor activity” after the
last item on the list but before they were signaled to
begin recalling the list items.
The reason for recency effect is also explained in
terms of the retroactive interference during encoding,
i.e., the earlier items suffering interference from the
later ones in the list (Oberauer 2003). Thus, the items
toward the end of the list interfere with the recall of
midlist items producing a recency effect. Retroactive
interference during a series of outputs (output interference), on the other hand, provides an advantage for items
retrieved first because the early items interfere with the
later items in the output sequence. This should generate
a primacy effect over output position: Items retrieved
earlier interfere with items yet to be retrieved, regardless
of input or spatial position. Interestingly, the serial position curve can be obtained in a free recall task of word list
in all individuals even in the cognitively impaired.

Cross-References
▶ Selective Associations
▶ Short-term Memory
▶ Working Memory and Information Processing

References
Atkinson, R. C., & Shiffrin, R. M. (1968). Human memory:
A proposed system and its control processes. In K. W. Spence &
J. T. Spence (Eds.), The psychology of learning and motivation
(Vol. 2, pp. 89–195). New York: Academic.
Glanzer, M., & Cunitz, A. R. (1966). Two storage mechanisms in free
recall. Journal of Verbal Learning and Verbal Behavior, 5, 351–360.
Nipher, F. E. (1878). On the distribution of errors in numbers written
from memory. Transactions of the Academy of Science of St. Louis,
3, CCX–CCXI.
Oberauer, K. (2003). Selective attention to elements in working
memory. Experimental Psychology, 50(4), 257–269.

S

3052

S

Serial Probe Recognition Memory

Postman, L., & Phillips, W. (1965). Short term temporal changes
in free recall. The Quarterly Journal of Experimental Psychology,
17, 132–138.
Rundus, D. (1971). Analysis of rehearsal processes in free recall.
Journal of Experimental Psychology, 89, 63–77.

Serial Probe Recognition
Memory

Set Learning
▶ Deutero-learning

Setting
▶ Cooperative Learning Groups and Streaming

▶ List Memory and Change-Detection Memory in
Animals

Sexual Conditioning
Serial Reaction Time (SRT) Task
The SRT task is a standard experimental paradigm
to investigate implicit learning. On each trial, a target
appears at one of a number of locations on a monitor and the participant is asked to press the key
corresponding to the location of the target. Unknown
to the subjects, the sequence of the target locations is
repeated after a certain number of (e.g., 20) trials.
Although the observers do normally not realize there
is a repeated sequence, they become increasingly faster
in pressing the keys corresponding to the repeated
locations as opposed to random locations.

▶ Erotic/Sexual Learning

Shaping
▶ Robot Learning from Feedback
▶ Shaping of New Responses

Shaping by the Method of
Successive Approximations
Serial-Order Learning

▶ Shaping of New Responses

▶ Procedural Learning

Shaping of New Responses
Serious Emotional Disturbance
▶ Achievement Deficits of Students with Emotional
and Behavioral Disabilities

Serious Game
▶ Learning with Games

BERTRAM O. PLOOG
Department of Psychology, Center for Developmental
Neuroscience, College of Staten Island, and Graduate
School of CUNY, Staten Island, NY, USA

Synonyms
Shaping by the method of successive approximations;
Shaping

Shaping of New Responses

Definition
“Shaping by the method of successive approximations” –
“shaping” for short – generally refers to a technique
in operant conditioning in which seemingly novel
behavior (i.e., behavior that does not obviously exist
yet in an animal’s or person’s repertoire) is created
by a careful balance of reinforcing (strengthening)
responses (an operant) that approximate the terminal
target behavior and extinguishing (weakening)
responses that are either too dissimilar from the target
behavior or are not a sufficiently close approximation
of the target behavior relative to some prior performance. Shaping is a practical procedure, which
is conducted by a person in order to create a new
behavior (e.g., an adult teaching a child with a developmental disability to get dressed independently).
Shaping may also occur as a desirable or undesirable
by-product of natural, that is, not explicitly planned,
consequences (e.g., improving one’s pronunciation of
the French word “pain” until one is understood by the
baker in Paris and receives the requested bread rather
than blank stares). There is also “auto-shaping” (Brown
and Jenkins 1968), which results in the creation of
seemingly new behavior (such as a pigeon’s pecking at
a purple hexagon), but it primarily involves classical
conditioning, and it is not clear whether auto-shaping
really involves shaping at all or if it is simply the
acquisition of a conditioned response (pecking) that
was in the animal’s repertoire all along (albeit occurring
under novel circumstances, i.e., pecking at the hexagon
instead of grain). New behavior can also be generated
through imitation, prompting, verbal instruction
(especially in humans), and experiential exposure, but
these do not involve shaping per se. This section limits
its discussion to the Skinnerian/operant notion of
shaping.

Theoretical Background
B. F. Skinner described shaping in his book Science and
Human Behavior (Skinner 1953, p. 91):
"

Operant conditioning shapes behavior as a sculptor
shapes a lump of clay. Although at some point the
sculptor seems to have produced an entirely novel
object, we can always follow the process back to the
original undifferentiated lump, and we can make the
successive stages by which we return to this condition
as small as we wish. At no point does anything emerge

S

3053

which is very different from what preceded it. The final
product seems to have a special unity or integrity of
design, but we cannot find a point at which this suddenly appears. In the same sense, an operant is not
something which [sic] appears full grown in the behavior of the organism. It is the result of a continuous
shaping process.

Shaping (by the method of successive approximations) consists of systematically applying reinforcement and extinction to some behavior (e.g., touching
a tooth brush) in order to modify it such that it gradually approaches a target behavior (e.g., brushing
teeth). This balance between reinforcement and extinction is known as differential reinforcement. Reinforcement might consist of food or praise that is presented
when a desired response, specifically an acceptable
approximation of the target behavior, occurred. For
example, touching the toothbrush and picking it up
may be reinforced by praise. The result is that the
reinforced behavior will be more likely to occur in the
future under circumstances that are similar to those
when a reinforcer was earned previously. Extinction is
the omission of reinforcement of behavior that was
previously reinforced. For example, at some point during shaping, simply touching the toothbrush is not
reinforced anymore – only touching it, picking it up
and, perhaps, moving it close to the face is. The result of
extinction is that the extinguished behavior (simply
touching the toothbrush) is less likely to occur in the
future. A by-product of extinction is that behavior
becomes more variable. To stay with the toothbrush
example: If simply touching the toothbrush is
extinguished, the child might scream (“Look, I am
touching it! Give me my token!”), touch the toothbrush with increased force, and start waving it around –
the latter being a behavior that can be reinforced in
order to approximate actual brushing. This increase
in variability is a critical point as will become apparent
in a moment when further discussing shaping.
Generally, with shaping, behaviors that resemble
more and more the target behavior are reinforced. In
contrast, behaviors that are too dissimilar to the target
behavior, or behaviors that do not meet an increasingly
stringent criterion of approximating the target behavior, are extinguished. This process is illustrated in
Fig. 1, with the example of shaping the target behavior
“standing on the right,” when originally this was a very

S

S

Shaping of New Responses
Step 1: Position at M is most likely
(mode). When shaping, movements
to positions to the left of the criterion
(solid line) are extinguished (EXT)
and movements to the right of the
criterion are reinforded (RF).

EXT RF

M
EXT RF

Frequency of Target Behavior

3054

Step 2: M shifted to the right
as a result of shaping. Thus,
the criterion was shifted to the
right. EXT and RF apply to the
new criterion.

M
EXT RF
Step 3: M shifted further
to the right, as did the
criterion. EXT and RF
continue to apply to the
new criterion.

M

Step N: Steps 1 through N –1 were
performed. M and the criterion line
continued to shift to the right, and
EXT and RF continued to apply to
the current criterion. At the last step,
criterion may remain at M.

Left

...

Middle

EXT RF

M
...

Right

Position
Shaping of New Responses. Fig. 1 A schematic
illustration of the mechanisms of shaping. Behavior is
variable with regard to its topography (represented as
a normal distribution here). The topography of a given
behavior tends to occur around a mode. By extinguishing
instances of this behavior with a topography that falls short
of a criterion and by reinforcing those that exceed the
criterion, the mode of the behavior gradually shifts in the
direction of the reinforced instances

low-probability behavior. A speaker might be standing
habitually on the left side of a stage (Step 1, Position M).
Quite often, the speaker stands a little bit to the left and
to the right of M, and on very rare occasions, she might
even be standing quite far from her typical left position
(represented by the normal distributions over M).
Should the listeners in the audience show attention to
the speaker (e.g., smile, nod, make eye contact) whenever she stands just a little bit further to the right than
usual and start withdrawing attention (e.g., yawn,
crinkle paper, close eyes, have private conversations)
whenever she stands again more to the left than at an

earlier step, gradually her most likely position (M) will
shift toward the right until she stands habitually at
the right (Steps 2 through N). Figure 1 also illustrates
the workings of a so-called percentile schedule of reinforcement (Galbicka 1994), which works by providing
reinforcement whenever a response occurs that falls on
or above a percentile that is set as a criterion.

Important Scientific Research and
Open Questions
Several considerations regarding shaping are of interest. In the example above, the change in behavior
occurred – at least when only considering a given
phase of the shaping procedure – in what is referred
to as one response class. For example, there are many
ways to brush your teeth (with little or a lot of force or
movement; for 1 or 2 min; with your left or right hand,
holding the brush with your fist or between thumb,
index, and middle fingers; etc.) A response class is
defined as “a set of responses that either are similar
on at least one response dimension [force, duration, or
topography of brushing], or share the effects of reinforcement and punishment [reinforcement increases
touching, lifting, and moving the toothbrush], or
serve the same function [all responses in the
brushing-teeth class produce praise]” (Malott 2008
p. 129, brackets added). It cannot be precisely defined
at what point exactly a response is considered to change
its membership to a different class or loses continuity
with other behaviors (Skinner 1953, p. 91–95). For
example, touching a toothbrush may be considered
discontinuous from sticking the brush into one’s
mouth and sliding it over one’s teeth. However, it is
also logically defendable – as described in Skinner’s
quote above – to see the change through shaping,
from touching the brush to actually brushing one’s
teeth, as a series of continuous behaviors, each linked
through a small change in response dimension. Adjacent steps certainly belong in the same response class;
early and late steps perhaps not.
The differentiation between response classes may be
of mostly theoretical interest but of little practical
interest. When shaping is used as a deliberate training
procedure to create new, complex behavior, it matters
little if touching the brush belongs to a different
response class than brushing teeth. In most real-life
situations, shaping is combined with a technique
known as behavioral chaining. The chain of brushing

Shaping of New Responses

teeth can be viewed as a series of links (applying tooth
paste to toothbrush, touching brush, lifting brush,
inserting brush into mouth, moving brush up and
downward, etc.) Link 1 has to be performed before
Link 2, and Link 2 has to be performed before Link 3,
etc. Each link may be trained separately by shaping.
Once all links are established, these links can be combined to form a chain – a process that can be done in
a forward or backward manner, resulting in forward or
backward chaining, respectively, the latter being the
most common and probably most efficient way of
training. (The rationale for this statement is beyond
the scope of this section and is not critical for understanding how chaining combined with shaping can be
used to create new responses.) Here is an example of
backward chaining: Once moving the brush up and
down in the mouth is established (Link N), one
might require that the child first pick up the brush
and insert it into his mouth (Link N – 1) before
performing Link N, which then leads to praise. When
this two-link chain is performed, Link N – 2 may be
required (e.g., put tooth paste on the brush), and so on.
Shaping the behavior of one link occurs most likely
within one response class (e.g., changing the response
dimension of force within the response class “lifting
tooth brush”). Shaping the behaviors of more than one
link occurs most likely across response classes (e.g.,
shaping the behavior of putting tooth paste on the
brush and then shaping the behavior of lifting the
brush to the mouth).
There are two types of shaping: variable- and fixedoutcome shaping (Malott 2008, p. 150). In fixedoutcome shaping, the amount of the reinforcer stays
the same for each successful performance of the critical
response (e.g., one token, one sip of soda, or one word
of praise with the same level of enthusiasm). In variable-outcome shaping, the amount of the reinforcer is
proportional to the level of performance of the critical
response (e.g., a “jack-pot” of five tokens, a half a can of
soda, or many words of particularly enthusiastic praise)
for a particularly well-done job. Variable-outcome
shaping mimics most closely an intuitive shaping
procedure that a parent might use naturally in child
rearing. Very little formal research has been conducted
to find out the most efficient ways of shaping. Thus,
some authors have referred to shaping as an “art” (Pryor
1999; but see Galbicka 1994, who argues against this).
To improve the efficiency of shaping systematically,

S

Pryor (1999; p. 38–39) has proposed “Ten Laws of
Shaping,” including “Raise criterion in sufficiently
small increments,” “Work on one behavior at a time,”
and “Go ‘back to kindergarten’ (i.e., step back to an
easier level) when the progression stalls.” It is clear that
some parents, teachers, or animal trainers have
a particular knack for teaching new behavior to their
students. Those who are most successful probably
acquired their own shaping skills without much awareness of naturally occurring contingencies that in turn
shaped their very own shaping skills! For example,
should a teacher let pass too much time between the
occurrence of the student’s critical response and its
reinforcer, the process of shaping might stall because
the greater the delay to reinforcement, the less effective
it becomes. Thus, the teacher’s behavior (shaping skill)
of presenting a reinforcer as quickly as possible is naturally shaped itself by reward, that is, greater and faster
success in shaping the student’s behavior. (This, in
effect, is an example of a naturally occurring variableoutcome shaping procedure.) Delivering a reinforcer
with too much delay, on the other hand, will be
extinguished because the shaping procedure will
not result in success. In other words, naturally occurring differential reinforcement shapes the teacher’s
shaping skills.
Unplanned differential reinforcement may also
shape undesirable behavior. Consider a child in a toy
store, demanding a toy with ever increasing force.
When the child is merely whining, the parent might
try to ignore (extinguish) the whining. But, as extinction results in an increase in variability, the child might
now exhibit whining with increased force, which is
hard to ignore. So, finally, the parent gives in and
actually just reinforced whining with increased force –
an example of shaping a response dimension (force)
within a response class (whining). Such a situation
might result in accidentally shaping serious temper
tantrums in every toy store the child visits with his
parents from now on.
As a final note, so-called “clicker-training” (Pryor
1999) will be mentioned, which is a powerful technique
to make shaping more efficient. For example, consider
a situation in which a trainer wants to shape “jumpingthrough-a-hoop” in a dolphin as a target behavior.
The hoop (and dolphin) may be 30 ft away from the
trainer. If the dolphin performs a first approximation
of the target behavior (e.g., swimming close to the

3055

S

3056

S

Shared Cognition

hoop), and the trainer offers a fish as a reward, the
dolphin has to swim toward the trainer in order to
receive the fish. In effect, the trainer reinforces swimming toward the trainer and not, as would be necessary,
approaching the hoop. However, in clicker-training,
a clicker (or whistle) has been paired previously with
a reinforcer, which established the clicker sound as a socalled conditioned reinforcer (by force of classical conditioning). A conditioned reinforcer is a stimulus that
has a similar effect on strengthening a response as any
reinforcer, but it does not have to be consumed at the
moment of presentation. In the dolphin example,
instead of offering a fish to the dolphin, the approximation of swimming close to the hoop is followed
immediately by the presentation of the clicker.
The clicker alone will strengthen the approximation,
and other behaviors that occur afterward (such as
swimming toward the trainer) are not reinforced accidentally. It is important that the conditioned reinforcement properties of the clicker be occasionally
maintained by following the sound with an established
reinforcer (fish in the example above).

Cross-References
▶ Differential Reinforcement
▶ Extinction Learning
▶ Imitation
▶ Internal Reinforcement Hypothesis
▶ Operant Conditioning
▶ Pavlovian Conditioning
▶ Reinforcement Learning
▶ Reinforcement Learning in Animals

References
Brown, P. L., & Jenkins, H. M. (1968). Auto-shaping of the pigeon’s
key-peck. Journal of the Experimental Analysis of Behavior,
11, 1–8.
Galbicka, G. (1994). Shaping in the 21st century: Moving percentile
schedules into applied settings. Journal of Applied Behavior Analysis, 27, 739–760.
Malott, R. W. (2008). Principles of behavior (6th ed.). Upper Saddle
River: Pearson Prentice Hall.
Pryor, K. (1999). Don’t shoot the dog (revised ed.). New York:
Bantam Books.
Skinner, B. F. (1953). Science and human behavior. New York:
Macmillan.
Further Reading
Cooper, J. O., Heron, T. E., & Heward, W. L. (2007). Applied behavior
analysis (2nd ed.). Upper Saddle River: Pearson Prentice Hall.

Shared Cognition
RIM RAZZOUK, TRISTAN JOHNSON
Department of Educational Psychology and Learning
Systems, College of Education, Learning Systems
Institute, Florida State University – Learning Systems
Institute, Tallahassee, FL, USA

Synonyms
Distributed cognition; Group cognition; Shared
knowledge; Shared mental model; Shared understanding; Team cognition

Definition
Shared cognition is the collective cognitive activity
from individual group members where the collective
activity has an impact on the overall group goals and
activities. In a complex setting, higher levels of shared
cognition are associated with more similar problem
conceptualizations and solution approaches. Shared
cognition is closely linked to knowledge such that
knowledge possessed by effective teams has been
referred to as shared knowledge, shared mental model,
team knowledge, and shared understanding (Klimoski
and Mohammed 1994). Shared cognition includes the
knowledge that team members hold, which enables
them to form accurate explanations and expectations
for the task and in turn to coordinate their actions and
adapt their behavior to demands of the task and other
team members (Cannon-Bowers et al. 1993). Shared
cognition is potentially different from the sum of individual cognition; rather, it is a result of collective cognitive, behavioral, and attitudinal activities during the
group dynamic interactions which influence coordination and lead to better team performance (CannonBowers et al. 1993; Klimoski and Mohammed 1994).
Shared cognition consists of two types of knowledge:
team knowledge and task knowledge. Team knowledge is
the knowledge associated with the general team processes and characteristics (e.g., team members’ preferences, strengths, tendencies, attitudes, preferences)
(Cannon-Bowers et al. 1993; Johnson et al. 2007).
Task knowledge is specific knowledge that is needed to
successfully perform tasks like procedures, sequences,
actions, and strategies (Cannon-Bowers et al. 1993;
Cooke et al. 2000). Several methods can be used to

Shared Cognition

measure shared cognition. These methods include
different elicitation techniques, such as cognitive
interviewing, observations, causal mapping, card
sorting, and representation techniques that use aggregate methods such as multidimensional scaling and
distance ratio formulas (Cooke et al. 2000; Klimoski
and Mohammed 1994).

Theoretical Background
The first theoretical perspective that underlies shared
cognition is social cognition which involves social process that relate to the acquisition, storage, transmission,
manipulation, and use of information for the purpose
of creating a group-level interactive and collective
product. One area of social cognition that represents
shared cognition is team cognition. The team cognition
is the theoretical framework that describes the factors
that play into a team’s ability to think. The underlying
assumption is that each team member possesses cognitions such as, thoughts, understandings, interpretations, beliefs, schemas, and mental models regarding
some aspect of the team’s work. It is the overlap or
similarity of these understandings, beliefs, mental
models, etc., that researchers can assess in order to
understand what is “shared” by team members.
Team cognition is more than the sum of the cognition
of each individual within the team as it emerges from
the interaction of the individual cognition of each
team member and team process. Within the team
cognition model, the majority of research has focused
on the sharing of mental models. Shared mental
models are knowledge structure(s) held by each member of a team that enables them to form accurate
explanations and expectations for the team and
task, and in turn, to coordinate their actions and
adapt their behavior to demands of the task and other
team members. This shared knowledge generates
from individual cognitive processes represented by the
individual mental models that refer to a general class
of cognitive construct that has been invoked to explain
how knowledge and information are represented in
the mind.

Important Scientific Research and
Open Questions
Mathieu, Heffner, Goodwin, Salas, and CannonBowers (2000) examined the effects of shared mental

S

models on team process and team performance. They
studied 56 pairs of undergraduate students performing
a flight simulation. They assessed teammates’ mental
models using ratings and analyzed mental model convergence by a network-analysis program. They found
the process to improve from Time 1 to Time 2, but not
any convergence of mental models over time. Mathieu
and colleagues (2000) confirmed the distinction
between taskwork mental models and teamwork mental models and found that the task mental models had
an indirect effect on team performance through team
processes. They found that the impact of team mental
model convergence on team performance was fully
mediated by team process.
In 1996, Volpe, Cannon-Bowers, Salas, and Spector
examined cross-training as a way of providing team
members with interpositional knowledge. They
conducted a 2 
 2 factorial between-subjects design
where 80 male college students participated in a flight
simulation. They found that cross-training resulted in
more effective teamwork, more efficient communication, and better performance than those who had not
been trained in their teammates’ responsibilities.
Smith-Jentsch, Campbell, Milanovich, and Reynolds (2001) examined teamwork mental models.
This refers to the understanding of the components of
teamwork that are critical for effective performance as
well as the relationships between those components.
They proposed that people with greater experience
working with teams have more knowledge and
a better understanding of what makes effective teamwork. They assessed 176 navy personnel with a cardsorting task to assess their mental models of teamwork.
They found a significant correlation between similarity to an expert model of teamwork (accuracy) and
navy rank. Additionally, they found that there was
greater similarity of mental models within highranking groups and within groups where people had
been in the service for a long time. Next, they designed
a training program to try to teach people to have
mental models of teamwork more similar to the
expert model. They assessed 42 civilian government
employees that had participated in a computer-based
training program. Results revealed that accuracy
improved after training, as did similarity to each
other’s models.
Even though, research has shown that shared cognition among team members is an important factor

3057

S

3058

S

Shared Knowledge

linked to team performance; there is only a small
amount of relevant empirical research relative to team
cognition and performance, most of which is indirect.
In the team performance area, the concept of shared
cognition or shared mental models is often merely
invoked post hoc, to help describe and interpret team
phenomena. While researchers have admitted the need
for direct experimental evidence, few have provided
any. Most empirical studies of shared cognition have
been conducted in experimental or artificial environments and with teams that were put together specifically for the purpose of the study. Finally, there is a need
to study the lack of knowledge with regard to the
cognitive, affective, and social mechanisms underlying
effective team performance.

Smith-Jentsch, K. A., Campbell, G. E., Milanovich, D. M., & Reynolds, A. M. (2001). Measuring teamwork mental models to
support training needs assessment, development, and evaluation: two empirical studies. Journal of Organizational Behavior,
22, 179–194.
Volpe, C. E., Cannon-Bowers, J. A., Salas, E., & Spector, P. E. (1996).
The impact of cross-training on team functioning: An empirical
investigation. Human Factors, 38, 87–100.

Shared Knowledge
▶ Shared Cognition

Shared Mental Model
Cross-References
▶ Collective Learning
▶ Distributed Learning
▶ Group Cognition and Collaborative Learning
▶ Group Dynamics and Learning
▶ Group Learning
▶ Knowledge Integration
▶ Learning in the Social Context
▶ Mental Model
▶ Peer influences on Learning
▶ Peer Learning
▶ Situated Cognition
▶ Team Learning
▶ Workplace Learning

References
Cannon-Bowers, J. A., Salas, E., & Converse, S. (1993). SMMs in
expert team decision making. In N. J. Castellan (Ed.), Individual
and group decision making (pp. 221–246). Hillsdale: Lawrence
Erlbaum.
Cooke, N. J., Salas, E., Cannon-Bowers, J. A., & Stout, R. J.
(2000). Measuring team knowledge. Human Factors, 42(1),
151–173.
Johnson, T. E., Lee, Y., Lee, M., O’Connor, D. L., Khalil, M. K., &
Huang, X. (2007). Measuring sharedness of team-related
knowledge: Design and validation of a shared mental model
instrument. Human Resource Development International, 10(4),
437–454.
Klimoski, R., & Mohammed, S. (1994). Team mental model: Construct or metaphor? Journal of Management, 20, 403–437.
Mathieu, J. E., Heffner, T. S., Goodwin, G. F., Salas, E., & CannonBowers, J. A. (2000). The influence of shared mental models on
team process and performance. The Journal of Applied Psychology,
85(2), 273–283.

▶ Role-Play and the Development of Mental Models
▶ Shared Cognition

Shared Understanding
▶ Shared Cognition

Shock Studies
▶ Learning of Obedience to Authority

Short Term Storage
▶ Short-Term Memory

Short-Term Memory
RONNY GEVA
Department of Psychology, The Gonda Brain Research
Center, Bar Ilan University, Ramat Gan, Israel

Synonyms
Related term: Working memory; Transient recollection; Short term storage

Short-Term Memory

Definition
Short-term memory (STM) is a process or processes
that enable holding on-line a representation of
a specific set of stimuli for a brief interval in order to
process it, manipulate it, or store it for a short period of
time. A more inclusive term often used is “working
memory,” which refers to an activated system that
keeps products of perceptual analysis and maintains
their traces in active state for on-line processing of
various cognitive operations and socioemotional functions. It may be triggered either by external stimulation,
by internally driven operations, or by both.

Theoretical Background
STM has been traditionally thought of as being limited
to a short interval and a limited number/complexity of
items. Nevertheless STM’s capacity is not thought to be
fixed; rather it is thought to be restricted by the
momentary available resources of attention that
may be allotted to tend each of the items in the context of its presentation. For example, this system is
susceptible to lifelong developmental changes. The
recall span for digits that are presented serially aurally
at a one-item-per-second rate typically increases linearly with age during childhood years; it typically
asymptotes to a range of four to nine items during
young adulthood, and then gradually diminishes in
old age.
Depending on strategies used, the complexity of the
stimuli and the individual’s resources at the time of
task, STM’s capacity may be extended dramatically
(Chase and Ericsson 1992) or be markedly reduced
(Gathercole et al. 1994).
Short-term/working memory allows for on-line
processing of temporal information, such as for
comprehending a long sentence, as this one, or for
solving a mathematical problem; and allows for on-line
processing of a visuospatial problem that is presented
simultaneously, by enabling a prolongation of the
processing phase, to permit a step-by-step analysis of
a complex set of stimuli and inhibit wrong yet predominant or salient responses.
There seems to be support for multiple short-term/
working memory systems, rather than a unitary system. This evidence is often based upon findings
of people who exhibit a greater capacity for verbal
STM than for visual STM, or vice versa, and on
individual differences in peoples’ predominant

S

dispositions to process stimuli in verbal/visual STM,
when both are relevant.
Short-term/working memory occurs due to the
neural system’s activity that maintains a set of signals
active, for example, by evoking a heightened or dampened activity in a distributed neural network. These
may be expressed by changes in firing rates or by
changes in graded potentials (Hebb 1949).
Short-term memory decays rapidly with time and is
sensitive to interferences and blurring. Forgetting of
materials held in STM may be accounted by the return
of the activity of this network back to its baseline or
default levels.
Rehearsal strategies may be employed to refresh
traces of a memorized set of stimuli to postpone rate
of decay and increase its accuracy.
Retrieval from STM of recently experienced stimulation or of internally represented experiences is based
on drawing short-term activity traces that are generated by the working memory system(s). It has been
suggested that during the first years of life, children
use no controlled strategy in the recall process (Palmer
2000); this is followed by a period in which a visual
strategy prevails, followed by a period of dual visualverbal coding before the adult-like strategy of verbal
coding finally emerges.
Baddeley and Hitch’s seminal model of working
memory offered a structural view of working memory.
According to this framework, the operating working
memory system is comprised of three integral divisions: a central executive controller of attention
resources, and two content-based processors: an articulatory loop and a visuospatial sketch pad (Baddeley and
Hitch 1974). More recent scientific developments proposed that the articulatory loop that holds on mostly
language-like stimuli is comprised of an “in-house”
short-term articulatory control process, which acts as
a momentary translator into modalities, and
a phonological store which encodes and registers the
phonetic data (Nairne 1996).
According to this model, the phonological loop
seems to be the most likely candidate to be activated
in the processing of both auditory and visual verbal
stimuli. Nevertheless, visual stimuli, even of verbal
nature, may activate a bottom-up visual perceptual
system that is related to the inner scribe component
of the visuospatial sketch pad. This component is
responsible for active rehearsal of information held

3059

S

3060

S

Short-Term Memory

within a passive visuospatial cache, and is involved with
extraction of information for execution of voluntary
motor acts.

Important Scientific Research and
Open Questions
STM is highly susceptible to any number of transient
conditions that compromise resources, such as fatigue,
physical illness, or nutritional deficiencies such as iron
deficiency. These effects are often refractory, and STM
capacity improves when this condition is amended. At
the same time, certain processes may divert normal
development of STM capacities in a more chronic
manner. An understanding of the mechanisms
involved in transient STM deficits may generate novel
research directions in populations who exhibit chronic
deficits.
STM deficits are often diagnosed in people who are
intellectually challenged, people with attention deficit
disorders, with or without hyperactivity, and often
accompany various verbal and nonverbal learning disabilities. STM is susceptible to an array of specific
genetic, structural, and nutritional aberrations such as
a result of intrauterine growth restriction (Geva et al.
2006), specific genetic syndromes, as Trisomy 21 and
Williams syndrome (Purser and Jarrold 2005; Vicari
and Carlesimo 2006), neuronal degenerative diseases,
and more. Certain focal neurological insults may also
result in lowered STM capacities. The mechanisms
involved in these clinical conditions are not yet fully
understood. An array of conditions that are related to
alterations in neural activity, neural integrity, and neural maturation or in the spread of activation may result
in STM deficits.
In order to better characterize the system(s) that
enable STM/working memory processing to improve
learning, it may be important to explore the following
directions:
Firstly, to date the literature on STM/working memory
mostly concentrates on studies using verbal stimuli
that are presented sequentially and or use of complex visuospatial information that is often
presented simultaneously. It may important to
find ways to better equate similar encoding opportunities of auditory and visual stimuli. Using
a presentation of sequential visual items and use
of nonverbal auditory spans may be important to

better characterize the interrelationships between
the executive controller network, the articulatory
loop, and the visuospatial sketch pad.
Secondly, the interrelationships between sensory reception systems (i.e., their sensitivity, range of reactivity, and accuracy) and the initial registration
mechanisms in STM as well as studies of factors
that affect de-blurring and reactivation of previous
recollections’ traces may advance our understanding of the mechanisms involved in rehabilitation of
affected STM/working memory systems.
Finally, in view of recent developmental studies of highrisk populations, it seems that assumption of an
immediate and automatic processing at the phonological loop needs to be carefully studied and validated in children or adults who are diagnosed with
an atypical neural-developmental process (Geva
et al. 2006). Thus, a careful study of the factors
that affect STM/working memory functions in various clinical populations who experience learning
deficits may be fruitful both for theoretical purposes and for intervention purposes to improve
learning.

Cross-References
▶ Capacity Limitations of Memory and Learning
▶ Controlled Information Processing
▶ Development and Learning (Overview)
▶ Examination Stress and Components of Working
Memory
▶ Memory Codes
▶ Mental Activities of Learning
▶ Neuropsychology of Learning
▶ Perceptual Learning
▶ Phonological Representation
▶ Working Memory
▶ Working Memory and Information Processing

References
Baddeley, A. D., & Hitch, G. (1974). Working memory. In G. H.
Bower (Ed.), The psychology of learning and motivation (Vol. 8,
pp. 47–89). San Diego: Academic.
Chase, W. G., & Ericsson, K. A. (1992). Skill and working memory. In
G. H. Bower (Ed.), The psychology of learning and motivation
(Vol. 16, pp. 1–58). San Diego: Academic.
Gathercole, S. E., Adams, A. M., & Hitch, G. J. (1994). Do children
rehearse? An individual differences analysis. Memory & Cognition, 22, 201–207.

Signal Detection Models
Geva, R., Eshel, R., Leitner, Y., Fattal-Valevski, A., & Harel, S. (2006).
Memory functions of children born with asymmetric intrauterine growth restriction. Brain Research, 1117, 186–194.
Hebb, D. O. (1949). Organization of behavior. New York: Wiley.
Nairne, J. S. (1996). Short-term/working memory. In E. L. Bjork &
R. A. Bjork (Eds.), Memory, handbook of perception and cognition
(2nd ed.). New York: Academic.
Palmer, S. (2000). Working memory: A developmental study of
phonological recoding. Memory, 8, 179–193.
Purser, H. M., & Jarrold, C. (2005). Impaired verbal short-term
memory in Down syndrome reflects a capacity limitation rather
than atypical rapid forgetting. Journal of Experimental Child
Psychology, 91, 1–23.
Vicari, S., & Carlesimo, G. A. (2006). Short-term memory deficits are
not uniform in Down and Williams syndromes. Neuropsychology
Review, 16, 87–94.

S

3061

Sign Systems
▶ Semiotics and Learning

Signal Detection Models
MICHAEL J. WENGER1, TAMARYN MENNEER2,
JENNIFER L. BITTNER3
1
Department of Psychology, The University of
Oklahoma, Norman, OK, USA
2
School of Psychology, University of Southampton,
Southampton, UK
3
Department of Psychology, Indiana University,
Bloomington, IN, USA

Short-Term Memory Capacity
▶ Behavioral Capacity Limits

Synonyms
Theory of signal detectability; Theory of signal
detection

Short-Term Memory Span
The largest amount of information that can be held in
short-term memory at a given time.

Sidman Equivalence
▶ Learning of Equivalence Classes

Definition
Signal detection models are probabilistic representations of the structure of the information and the rules
for decision-making that can be applied to a specific
decision problem. A completely specified signal detection model requires (a) the definition of the dimension
(s) of the evidence or information, (b) selection of the
probability distribution(s) for the random variable
used to represent the evidence or information, and
(c) the rule for selecting the criterion values used to
translate the values of that random variable into discrete decisions.

Theoretical Background

Sight Word Spellings
▶ Mental Graphemic Representations

Sign Language
▶ Verbal Behavior and Learning

The formal foundations for signal detection models
were developed as a part of the effort to develop radar
and sonar technologies, with much of this work occurring during and after World War II. The approach
is general enough to have allowed for a wide range of
applications, including statistical signal processing,
telecommunications, statistical decision-making, quality control, and human perception, memory, and
cognition, with special applications to training and
evaluating diagnostic abilities in medicine.

S

S

Signal Detection Models

The application of signal detection models to
human performance dates to the mid-1950s, with critical initial work in human psychophysics (W. P. Tanner
and Swets 1954). The introduction of signal detection
models to psychophysics was in part motivated by
a concern that classical approaches to psychophysics
did not make a distinction between the total amount of
perceptual evidence available in two or more different
states of the environment and the manner in which that
evidence was used. These two aspects of human performance are referred to as sensitivity and bias within
the literatures on signal detection theory (Green and
Swets 1966).
The basic elements of a signal detection model can
be illustrated by considering the general problem faced
by any system (e.g., a human observer, a radar system,
etc.) that needs to detect the presence of a specific type
of signal (a target) that is embedded in a background of
irrelevant noise. Thus, the environment in which the
system operates can be described as existing in one of
two possible states at any point in time: one in which
the target signal is present along with the background
noise and one in which only the background noise is
present. The system generates one of two possible
responses to the environment: one indicating that it
judges the target to be present and one indicates that it
judges the target to be absent. This is represented schematically in Fig. 1a, along with the standard labels for
the outcomes associated with each combination of true
state of the environment and system response.
Assume that the system has some method for
encoding the level of evidence available on any encounter with the environment with respect to the presence
of the signal, and that the level of evidence encoded is

a combined function of signal strength and the level of
noise. Let X be the random variable representing the
level of evidence and assume that X is distributed as
a Gaussian random variable, with f ðxjs þ nÞ denoting
the pdf for this random variable when the signal is
present in the noise, and f ðxjnÞ denoting the pdf
when the signal is absent. Assume that these pdf ’s are
ordered with respect to their means, such that
msþn > mn , and that the variances for the two pdf ’s are
equal, with ssþn ¼ sn .
At this point, the model is specified in terms of the
dimension of interest (evidence for the presence of the
signal) and the applicable probability distribution. In
order to use this model to make predictions for behavior or performance, what remains is to specify the rule
for generating a detection response. Assume that there
is a critical level of evidence, xc , the specific value of
which is set by the system, such that when X  xc the
system judges the signal to be present, otherwise judging the signal to be absent. This decision rule is applied
to both probability distributions, allowing for both
types of responses to be generated for both states of
the environment. The model for this (example) situation is now complete.
The model (shown schematically in Fig. 1b) can be
used to make predictions for performance in the following ways. First, predictions regarding probabilities of
each of the four outcomes in Fig. 1a can be obtained
by integrating the pertinent pdf relative to the value of
xc . For example, the probability of a “hit”
R1
can be obtained as f ðxjs þ nÞdx. Predicted sensitivxc

ity of the system to the presence of the signal can be
quantified in terms of the distance separating the pdfs,

System Response

True State of
the Environment

3062

"Signal
present"

"Signal
absent"

Signal present

hit

miss

Signal absent

false
alarm

correct
rejection

xc (criterion)
f(x | n)

f(x | s + n)

X (evidence variable)

Signal Detection Models. Fig. 1 (a) Outcomes associated with the pairing of each possible true state of the environment
with each possible response. (b) Schematic representation of a fully specified signal detection model for detecting the
presence of an arbitrary signal embedded in noise

Signal Detection Models

S

3063

The application of signal detection models has been
quite widespread, including both natural and
engineered systems, due in no small part to the generality of the approach. The approach has included incorporation of models for learning and applications to
multidimensional (multivariate) issues (Ashby and
Townsend 1986), though unidimensional models are
far more common, particularly with respect to human
performance.

Incorporating Learning into Signal
Detection Models
Signal detection models for learning have often been
represented within a dynamic systems framework. In
early models (Atkinson 1963) sensory processes were
modeled through the changes in activation of a sensory
state that corresponded to a particular aspect of the
experimental paradigm. Decisional processes were then
incorporated as a set of responses assigned to the various sensory states. These models were used to examine
the effects of various manipulations on response bias,
including types of feedback and presentation schedules
(Tanner et al. 1970). The majority of models addressed
learning in unidimensional applications, though
a more recent extension to multidimensional contexts
(Ashby 2000) incorporates random walk theory in
order to address within-trial dynamics.

Connecting Theory and Data
An ongoing issue with respect to the application of
signal detection models concern methods for
connecting theory and data. The problem is one of
identifying a unique model for a given set of experimental data. This issue is of particular concern for
multidimensional signal detection models, given the
increase in complexity and number of potentially-free
parameters relative to the degrees of freedom in the
data.
To illustrate these issues, consider the extension of
the unidimensional model in Fig. 1 to two dimensions.
A common and simple way of visualizing this situation

y

Important Scientific Research and
Open Questions

y

and the predicted bias of the system is captured by the
location of the critical value of the evidence variable xc .

x

x

Signal Detection Models. Fig. 2 Illustration of mean shift
integrality (left) and an alternative configuration (right). Left
panel: both distributions and the decision bound have
shifted up by the same amount. Right panel: negative
bivariate correlations in all distributions give rise to the
same pattern of response data as the shifted configuration
on the left

is in terms of contours of equal likelihood, with a linear
decision criterion for each dimension. Models of this
form allow for characterization of issues such as interactions among dimensions (Ashby and Townsend
1986). These interactions can then be identified at the
level of the perceptual evidence (parameters of the
multivariate distributions) and at the level of the rules
for generating responses (parameters of the decision
bounds).
A simple example in which model identifiability is
a problem occurs when both the multivariate distributions and the decision bounds shift. In Fig. 2, the
distributions on the right are higher than those on the
left, but their positions relative to the decision criterion
are the same as those on the left. In terms of response
data, such a situation is indistinguishable from the
configuration in the right panel of Fig. 2. Such ambiguity prevents conclusions being drawn as to whether
dimensional interdependencies exist at the perceptual
or decisional level.
The statistical issues associated with model estimation and identifiability have been noted frequently
(Ashby and Townsend 1986). A variety of methods
have been proposed (Thomas 2001), although to date
no comprehensive comparisons of relative robustness
have been published. However, there are indications
that by using multiple statistical methods, inferential
errors can be minimized. Such multi-method
approaches are consistent in spirit with seminal

S

3064

S

Signal Detection Theory

research using multi-measure approaches to estimate
multidimensional signal detection models (e.g., Ashby
and Townsend 1986).

Cross-References
▶ Choice Reaction Time and Learning
▶ Mathematical Models/Theories of Learning
▶ Model-based Learning with Systems Dynamics

References
Ashby, F. G. (2000). A stochastic version of general recognition
theory. Journal of Mathematical Psychology, 44, 310–329.
Ashby, F. G., & Townsend, J. T. (1986). Varieties of perceptual independence. Psychological Review, 93, 154–179.
Atkinson, R. C. (1963). A variable sensitivity theory of signal detection. Psychological Review, 70(1), 91–106.
Green, D. M., & Swets, J. A. (1966). Signal detection theory and
psychophysics. New York: Wiley.
Tanner, W. P., & Swets, J. A. (1954). A decision-making theory of
visual detection. Psychological Review, 61(6), 401–409.
Tanner, T. A., Rauk, J. A., & Atkinson, R. C. (1970). Signal recognition
as influence by information feedback. Journal of Mathematical
Psychology, 7, 259–274.
Thomas, R. D. (2001). Characterizing perceptual interactions in face
identification using multidimensional signal detection theory. In
M. J. Wenger & J. T. Townsend (Eds.), Computational, geometric,
and process perspectives on facial cognition: Contexts and challenges (pp. 193–228). New Jersey: Lawrence Erlbaum.

Signal Detection Theory
Signal detection theory is a principled explanation for
decision making (detection, discrimination, recognition, classification, and identification) under noisy
conditions. In psychophysical experiments, if certain
assumptions are met, it permits to separate
a participant’s sensitivity from his or her response bias.

Signal Processing
▶ Phonetics and Speech Processing

Significant Learning
▶ Person-Centered Learning

Similarities in and Differences
Between Nonhuman Ape and
Human Cognition: The Cultural
Intelligence Hypothesis
ESTHER HERRMANN
Department of Developmental and Comparative
Psychology, Max Planck Institute for evolutionary
Anthropology, Leipzig, Germany

Definition
Despite sharing many cognitive abilities with other
animals, humans also have a specific set of cognitive
skills, such as language, mathematics, and the ability to
create complex technologies and social institutions,
skills which are not even possessed by their closest
primate relatives, the great apes. The cultural intelligence hypothesis argues that this is mainly due to
a specialized set of social cognitive abilities for participating and exchanging information in cultural groups.
These social cognitive skills emerge early in ontogeny
and enable young children growing up within a cultural group to benefit from the preexisting knowledge
and skills of others around them. Children learn and
are taught to use the already existing symbols and
artifacts such as language, tools, and Arabic numerals.
This species’ unique sociocultural world then supports further development in many different cognitive
domains.

Theoretical Background
Nonhuman great apes (orangutans, gorillas, chimpanzees, and bonobos; hereafter great apes) are humans’
closest relatives. They all shared a common ancestor
around 15 Ma ago, with bonobos, chimpanzees, and
humans also sharing a common ancestor around 6 Ma
ago and 98% of their genotype. In addition, based on
their close evolutionary history, they also share many
morphological and physiological characteristics and
the majority of human cognitive abilities, such as an
understanding of space, the ability to discriminate and
quantify objects, to recognize individual conspecifics
and third party relationships, and to understand conspecifics as goal directed and intentional beings
(Tomasello and Call 1997; Tomasello et al. 2005). However, humans have certain cognitive skills which are not

Similarities in and Differences Between Nonhuman Ape and Human Cognition: The Cultural Intelligence Hypothesis

found in their closest relatives, such as language, symbolic art, mathematics, and the ability to generate
complex technologies such as computers or airplanes
and create social institutions such as religions. The
questions of how and why specific cognitive skills
evolved in primates, and especially the very distinctive
cognitive skills of humans, have long been of interest,
and various hypotheses have been proposed to answer
these questions.
One proposal to explain uniquely human cognition
is the general intelligence hypothesis: humans’ extralarge brains enable them to perform all kinds of cognitive operations better and faster than other species
based on more memory, faster learning, faster perceptual processing, more robust inferences, longer-range
planning, and so on. However, other theorists suggest
that cognitive abilities evolve in response to relatively
specific environmental challenges. Therefore, it has
been proposed that primate physical cognition evolved
mainly in the context of navigation and foraging for
patchily distributed and seasonal food, such as ripe
fruits, new leaves, flowers, and extracting food that is
not directly perceptible (including the use of tools).
Besides this ecological intelligence hypothesis, other
researchers have proposed that living in a social group
and dealing with conspecifics creates additional cognitive challenges beyond dealing with and the manipulation of inanimate objects. Conspecifics behave
spontaneously on their own and are therefore
unpredictable. Moreover, the behavior of a conspecific
can be influenced, controlled, or exploited in order to
gain useful information. Therefore, the “social intelligence hypothesis” proposes that primates’ special cognitive abilities evolved mainly in response to the
especially challenging demands of a complex social
life characterized by constant competition and cooperation with conspecifics (Byrne and Whiten 1988). In
a modified version of the social intelligence hypothesis
Whiten and van Schaik (2007) recently proposed
interdependence between having multiple traditions
and intelligence, focusing mainly on social learning
and innovation. The common idea shared by the different social intelligence hypotheses is that the evolution of social cognitive skills in primates has been
mainly fuelled by constant competitive and exploitative
processes between conspecifics. However, these social
intelligence hypotheses cannot fully explain the differences in cognitive abilities between great apes and

S

humans. In the case of humans, one variant of the
social intelligence hypothesis emphasizes the role of
cooperation. According to this theory, humans are
extraordinarily social, and their distinct cognitive skills
evolved mainly in response to their complex sociocultural life which depends highly not only on competing
with others but also on complex forms of collaboration
(Vygotsky 1978). To successfully live and exchange
information within the cultural world human children
are born into, it is necessary to possess sophisticated
social-cognitive skills such as social learning, communication, understanding the intentions of others, and
the motivation to share knowledge with other individuals. These skills enable young children to benefit from
the accumulated skills and knowledge of their fellow
human beings. They learn and are even explicitly
taught to use already existing artifacts and symbols
such as language, tools, and Arabic numerals (Vygotsky
1978; Tomasello et al. 2005). However, other great ape
species such as chimpanzees and orangutans also have
special traditions and culture, in which a variety of
behaviors are transmitted across generations. These
great apes are able to learn new behaviors by observing
others; however, there is little evidence that learning is
assisted by teaching. In general, apes also do not rely on
participating in such cultural interactions to the same
extent as humans. The “cultural intelligence hypothesis” therefore suggests that the distinct cognitive skills
of human adults which are not found to the same
degree in other primates are a product of specialized
social cognitive skills which emerge early in ontogeny
and make it possible to participate in the cultural
world, correspondingly allowing the distinctively complex development of human physical and social cognition (Herrmann et al. 2007).

Important Scientific Research and
Open Questions
To test and compare two alternative hypotheses about
the evolution of human cognitive abilities – the cultural
versus the general intelligence hypothesis – Herrmann
et al. (2007) presented a comprehensive battery of
cognitive tests to large numbers of two of human’s
closest primate relatives, chimpanzees, and orangutans,
as well as to 2.5-year-old human children. The test
battery consisted of 16 different nonverbal tasks
assessing all kinds of cognitive skills, involving both
physical and social problems relevant to primates in

3065

S

S

Similarities in and Differences Between Nonhuman Ape and Human Cognition: The Cultural Intelligence Hypothesis

their natural environment. While the tests of the physical world consisted of problems concerning space,
quantities, tools, and causality, the tests of the social
world required subjects to imitate another’s solution to
a problem, communicate nonverbally with others, and
read the intentions of others from their behavior. If it is
the case that humans have greater general intelligence
than great apes, children should differ from chimpanzees and orangutans uniformly across all the physical
and social cognitive tasks. The findings did not support
this hypothesis. Two-and-half-year-old human children and chimpanzees had very similar cognitive skills
for dealing with the physical world; however, the same
children, who were using language but were still years
away from reading, counting, or going to school, had
already more sophisticated cognitive skills than either
one of the ape species for dealing with the social world
(Fig. 1).
In addition, Herrmann et al. (2010) examined the
structure of individual differences in the cognitive abilities of human children and chimpanzees in this comprehensive range of cognitive tasks. Even if human
children were not more intelligent overall than other
great apes but only more skillful in solving social cognitive problems, it is still possible that a general intelligence factor could explain the individual differences
within a species. One hypothesis is that only humans
exhibit a general intelligence factor, which would mean

Physical domain

a
1.00

that children who are great at solving physical tasks are
also great at dealing with social tasks. But this was not
the case. Neither the childrens’ nor the chimpanzees’
performances could be explained by a general intelligence factor since individual performance on one task
did not predict performance on other tasks across the
board. Rather, results showed a similar factor for spatial
cognition for both species – but beyond these commonalities, the two species showed different factor
structures. While one factor comprising both physical
and social cognitive tasks was found for chimpanzees,
children had two distinct factors for physical cognition
and social cognition.
Thus, 2.5-year-old children not only showed more
sophisticated cognitive skills in the social domain alone
rather than across the board, but there was also no
general intelligence factor found for children. Instead,
a factor for social cognition was established. Together,
these findings suggest that humans share many cognitive skills with their closest living relatives – especially
for dealing with the physical world – but in addition,
they have evolved some specialized and more integrated social cognitive skills. With these special skills
of social cognition, emerging early in ontogeny, young
children are well equipped to acquire all kinds of
knowledge and skills through communicating, collaborating, and socially learning from others in their cultural group. Being able to participate in the species

0.80
0.60
0.40
0.20
0.00

Human

Chimpanzee Orangutan

Social domain

b
Proportion of correct responses

Proportion of correct responses

3066

1.00
0.80
0.60
0.40
0.20
0.00

Human

Chimpanzee Orangutan

Similarities in and Differences Between Nonhuman Ape and Human Cognition: The Cultural Intelligence Hypothesis.
Fig. 1 Physical (a) and social (b) cognitive performance of human children, chimpanzees, and orangutans (Taken from
Herrmann et al. 2007)

Similarity Learning

unique sociocultural world then facilitates all their
physical and social cognitive skills. Hence, this set of
species specific skills suggests that humans compared
to other animals have a specific biological adaptation
for the complex social life and cultural world that they
themselves have created.
Such large-scale comparisons provide a first step
toward systematically investigating similarities in and
differences between humans and other great apes and
open the way for mapping out in a systematic way the
cognitive skills of other primate species across the evolutionary spectrum. This will then provide the kind of
detailed description of primate cognition needed to
reconstruct both the biological and cultural evolution
of human cognition. Furthermore, much remains
unknown about the physical and social cognitive development of great apes and how it compares to human
children’s development. This kind of research is necessary to gain a better understanding of developmental differences in cognition across humans and great
apes and especially of the way in which ontogenetic
differences influence the cognitive differences found
between human adults and our closest living relatives.
Finally, another important avenue for future research
is to investigate the similarities and differences in
temperament and emotions that might contribute
to the extraordinary social nature of humans and
other great apes, including their special social cognitive skills.

Cross-References
▶ Animal Culture
▶ Group Cognition and Collaborative Learning
▶ Human Cognition and Learning
▶ Linguistic and Cognitive Capacities of Apes
▶ Social Cognition in Animals

References
Byrne, R. W., & Whiten, A. (1988). Machiavellian intelligence. Social
expertise and the evolution of intellect in monkeys, apes, and
humans. New York: Oxford University Press.
Herrmann, E., Call, J., Hernández-Lloreda, M. V., Hare, B., &
Tomasello, M. (2007). Humans have evolved specialized skills of
social cognition: The cultural intelligence hypothesis. Science, 317,
1360–1366.
Herrmann, E., Hernández-Lloreda, M. V., Call, J., Hare, B., &
Tomasello, M. (2010). The structure of individual differences in
the cognitive abilities of children and chimpanzees. Psychological
Science, 21, 102–110.

S

3067

Tomasello, M., & Call, J. (1997). Primate cognition. New York: Oxford
University Press.
Tomasello, M., Carpenter, M., Call, J., Behne, T., & Moll, H. (2005).
Understanding and sharing intentions: The origins of cultural
cognition. The Behavioral and Brain Sciences, 28, 675–691.
Vygotsky, L. S. (1978). Mind and society: The development of higher
mental processes. Cambridge, MA: Harvard University Press.
Whiten, A., & van Schaik, C. P. (2007). The evolution of animal
‘cultures’ and social intelligence. Philosophical Transactions of
the Royal Society B, 362, 603–620.

Similarity Learning
MARCEL WORRING
Intelligent Systems Lab Amsterdam, University of
Amsterdam, Amsterdam, The Netherlands

Synonyms
Interactive image retrieval; Metric learning; Relevance
feedback learning

Definition
When describing images, humans often resort to similarity defining the characteristics of the image in relative terms rather than absolute terms. Subtle
differences between images can be indicated by
a human easily while completely describing a single
image is a challenging task. Cognitive evidence also
suggests that we interpret objects by relating them to
prototypical examples stored in our brain. So similarity
is a fundamental property and of great importance in
retrieval and categorization tasks alike.
Similarity learning is the process of determining
a function d(a,b), which finds the optimal relation
between two different data items a and b in
a quantitative way. In general, defining similarity for
structured symbolic or numeric data is rather straightforward, but defining similarity from a human perspective is difficult (Tversky 1977), this is especially
true for visual data for which there is a semantic gap
between what a system can extract from the sensory
data and the interpretation the same data has for
a human observer (Smeulders et al. 2000). In this
respect, similarity learning is viewed as determining
a function d(a,b) best matching the user interpretation
of the data items and their relations.

S

3068

S

Similarity Learning

Theoretical Background
The starting point in defining similarity is a feature
function fa, which takes the sensory data of a and
summarizes it by a feature vector describing perceptual
properties like color, texture, and local shape. Function
d(a,b) can then be expressed in terms of the feature
vectors fa and fb, yielding a perceptual distance function. This function in many cases is such that it follows
the metric properties dða; aÞ ¼ 0; dða; bÞ ¼ dðb; aÞ and
the triangular inequality dða; bÞ þ dðb; cÞ  dða; cÞ.
However, there are also many applications where
d is nonsymmetric or where the triangular inequality
only holds locally. And indeed, given a set of features,
many functions are in use to compute the similarity like
Histogram intersection, Minkowski distance, Euclidean distance, or Earth Movers Distance (Smeulders
et al. 2000). As the user interpretation in general is
subjective, learning the similarity requires relevance
feedback RF from the user. So similarity learning is an
iterative process defined as follows:
RFi

d i ða; bÞ!d iþ1 ða; bÞ
where d 0 ða; bÞis a perceptual distance function. Based
on some initial query on the dataset, the system presents a result set to the user conforming to d 0. The user
is then asked to give feedback on the result, typically
identifying correct and incorrect elements. From there,
in an iterative manner, the system updates its internal
definition of similarity and presents a next set of
results, until the user is satisfied with the results or
until the user reaches the conclusion that such a result
cannot be achieved. This process is illustrated in Fig. 1.

Due to the semantic gap, learning the similarity
function is required for two reasons:
● The user interpretation of the images can be sub-

jective and hence observer dependent.
● The mapping from perceptual features to interpre-

tation in terms of concepts is complex.
Because of this distinction, we have at least two
forms of similarity learning that play a role. To learn
the subjective interpretation of the user requires shortterm learning where typically the user is interacting in
a single session with the system. Long-term learning on
the other hand is based on several interactive sessions
from a larger number of users. As in the latter case
subjective interpretations are averaged out, we can
start to learn the complex mapping from perceptual
distance functions to conceptual similarity.
One of the first methods for similarity learning
in image retrieval and conceptually the simplest is by
Rui et al. (1998) and is based on learning a similarity
function as linear combination of distances for individual elements of a feature vector. After that many
other methodologies have been defined (Smeulders
et al. 2000, Zhou and Huang 2003) based on different
similarity functions, different types of feedback, and
different ways of updating the similarity function.
Apart from updating the similarity function itself it is
also possible to update the interpretation of the similarity function in terms of revising the probability
associated to certain concept occurrence.
Recent methods have taken the importance of prototypical data elements (Pekalska et al. 2006) or images

Distance indicates
(dis)similarity

Before feedback

User selection of positive
elements (dark), others assumed
negative

After feedback positives
closer together others
further away

Similarity Learning. Fig. 1 Simplified overview of the concept of interactive similarity learning

Simon, Herbert A. (1916–2001)

for defining similarity a step further. In these methods,
a small set of prototypical images from the collection are
selected, for example, by taking the center elements of
data-driven clusters. Now a new feature vector is defined
based on the distances of an element to each of those
prototypes. This new feature space has good performance in terms of classification of the elements. When
projected to two dimensions for visualization, this space
allows the user to interact directly with the similarity
between different images by changing the positions of
individual images on the screen (Nguyen et al. 2007).

Important Scientific Research and
Open Questions

S

3069

Pe˛kalska, E., Duin, R. P. W., & Paclik, P. (2006). Prototype selection
for dissimilarity-based classifiers. Pattern Recognition, 39(2),
189–208.
Rui, Y., Huang, T. S., Ortega, M., & Mehrotra, S. (1998). Relevance
feedback: A power tool for interactive content-based image
retrieval. IEEE Transactions on Circuits and Systems for Video
Technology, 8(22), 644–655.
Smeulders, A. W. M., Worring, M., Santini, S., Gupta, A., & Jain, R.
(2000). Content based image retrieval at the end of the early
years. IEEE Transactions on Pattern Analysis and Machine Intelligence, 22(12), 1349–1380.
Tversky, A. (1977). Features of similarity. Psychological Review, 84,
327–352.
Zhou, X. S., & Huang, T. S. (2003). Relevance feedback in image
retrieval: A comprehensive review. ACM Multimedia Systems
Journal, 8, 536–544.

Although methods for learning similarity have been
around for quite some time, there are still many
challenged ahead. Some we highlight here:
● Scalability: image collections are growing rapidly in

size, but interactive learning of similarity requires
interactive response times.
● Implicit Feedback: indicating positive and negative
examples are the de facto standard in learning similarity, yet few users are willing to do this at large
scale. Implicit feedback mechanisms seem more
appropriate.
● Bootstrapping: learning similarity hinges on a good
starting point, making wrong choices here might
lead to biased interaction processes not leading to
the right result.
● Learning Semantics: truly understanding the relation between low-level perceptual distance functions and high-level conceptual relations and
obtaining this relation through interaction remains
a challenge.

Cross-References
▶ Active Learning
▶ Discriminative Learning
▶ Feedback Strategies
▶ Interactive Learning
▶ Measures of Similarity
▶ Simulation-Based Learning

References
Nguyen, G. P., Worring, M., & Smeulders, A. W. M. (2007). Interactive Search by direct manipulation of dissimilarity space. IEEE
Transactions on Multimedia, 9(7), 1404–1415.

Similarity-Based Problem
Solving by Animals
▶ Analogical Reasoning in Animals

Simile
▶ Analogy/Analogies: Structure and Process

Simon, Herbert A. (1916–2001)
NORBERT M. SEEL
Faculty of Economics and Behavioral Sciences,
Department of Education, University of Freiburg,
Freiburg, Germany

Life Dates
Herbert Alexander Simon was born in Milwaukee,
Wisconsin, on June 15, 1916. During his school education he developed a strong interest in science. Later in
his life, Simon was influenced by his mother’s younger
brother, Harold Merkel, who had published books
on economics and psychology. Thus, Simon became

S

3070

S

Simon, Herbert A. (1916–2001)

interested in the social sciences. He received his Ph.D.
in political science from the University of Chicago in
1943. Then he served as professor of political science
at the Illinois Institute of Technology until 1949,
where he began intensive study of institutional economics. Simon then accepted a position as professor
of administration and chair of the Department
of Industrial Management at Carnegie Institute of
Technology, where he remained until his death,
teaching subjects as diverse as psychology and computer science.
As is already evident from his teaching career,
Simon never stopped studying new subjects. For example, he studied mathematical economics from 1950 to
1955, developing and proving the Hawkins-Simon theorem with David Hawkins on the “conditions for the
existence of positive solution vectors for input-output
matrices” as well as theorems on near-decomposability
and aggregation. Interestingly, Simon also applied
these theorems to organizations. Not surprisingly, he
also became highly interested in problem solving and
the emerging field of computer science, and in the mid
1950s, he concluded that these two fields of interest
could be linked through simulations of problem solving with computer programs. Altogether, it can be said
that Herbert Simon was an ingenious researcher who
was accordingly awarded with numerous accolades,
among which the 1978 Nobel Memorial Prize in Economics for his pioneering research into decision making within economic organizations deserves special
attention.
Simon died on February 9, 2001.

Theoretical Background
During his entire life, Herbert Simon was a true interdisciplinary scientist who was affiliated with several
Carnegie Mellon departments, such as the School of
Computer Science, the Tepper School of Business, and
the departments of Philosophy, Social and Decision
Sciences, and Psychology. Accordingly, he contributed
to various fields of interest, such as decision making,
problem solving, development of expertise, and artificial intelligence.
Already as a young scientist, Simon became popular
due to his research on industrial organization and
decision making. As a contrast to neo-classical theories
of rational decision making, Simon coined the term
bounded rationality, according to which individuals (as

parts of organizations) tend to maximize utility functions under current constraints in pursuit of their
self-interests. According to Simon’s theory of subjective
expected utility, bounded rationality refers to a rational
choice among alternatives that takes into consideration
not only the organizational constraints of decision
making but also the cognitive limitations of the decision maker. Bounded rationality became a central issue
within the realm of behavioral economics, and Simon’s
research on decision making became an important
paradigm in industrial organization and organizational
psychology.
Simon was a pioneer in the field of artificial intelligence. Together with Allen Newell, he developed the
General Problem Solver (GPS) in 1957, a technique for
separating problem solving strategies from particular
problems that also includes a theory for simulating
human problem solving based on production systems.
Related to this seminal work on problem solving was
Simon and Feigenbaum’s development of the EPAM
(Elementary Perceiver and Memorizer) theory, one of
the first theories of learning to be implemented as
a computer program.
Closely related to the study of human problem
solving was Simon’s interest in studying the role of
knowledge in the development of expertise, especially
in the area of chess (Simon and Chase 1973). He found
that characteristics of becoming an expert include
a minimum of 10 years of consistent practice and
an early phase of learning characterized by excitement
and enjoyment without outcome-oriented objectives.
Ericsson continued this research on expertise development and focused on the role of deliberate practice,
which encourages novice learners to pursue new levels
of performance in a field of practice (Ericsson et al.
1993).
Simon believed that the study of human problem
solving required new kinds of assessment and measurement. Therefore, together with Ericsson he developed
the experimental technique of verbal protocol analysis
(Ericsson and Simon 1980), which is still the standard
today.

Contributions to the Field of Learning
Herbert Simon was, of course, also interested in how
humans learn, and he thus developed the EPAM (Elementary Perceiver and Memorizer) theory with
Edward Feigenbaum (Feigenbaum and Simon 1984).

Simulation and Learning

EPAM operates with discrimination nets to regulate the
flow of information processing. It was one of the first
computer programs which could simulate verbal learning and was also influential in formalizing the concept
of a chunk. Later, the theory was extended to concept
learning and the acquisition of expertise. Basically, it
assumes that learning consists in the growth of discrimination nets. Additionally, EPAM contains a limited
short-term memory and several attention mechanisms.
Information processing occurs in EPAM through various steps: First, a stimulus is perceived and its features
are identified. Second, the features are arranged within
the discrimination net in such a way that an association
with particular content from long-term memory can be
generated. Third, simultaneously with the arrangement
of the features in the discrimination net a process
occurs in the net in order to decide whether learning
is necessary (e.g., adding additional features). Finally,
the computational system acts or the next stimulus is
processed.
Related computational cognitive models of thinking and learning are CHREST and SOAR, which are
described in separate entries of this encyclopedia.
In addition to his pioneering work on EPAM and
the development of expertise, Simon also contributed
to educational issues. For instance, when the discussion
on situated learning and education culminated in the
1990s in fundamental criticism of cognitive theories of
learning, especially in mathematics education, Simon
took issue with the overstatements of the situated
learning approach and emphasized the relevance of
traditional cognitive research on learning and transfer
(Anderson et al. 1996).
To sum up Simon’s contributions to several disciplines, it can be stated that he was a real polymath and
one of the most influential economic and social scientists of the twentieth century.

3071

References
Anderson, J. R., Reder, L. M., & Simon, H. A. (1996). Situated
learning and education. Educational Researcher, 25(4), 5–11.
Ericsson, K. A., & Simon, H. A. (1980). Verbal reports as data.
Psychological Review, 87, 215–251.
Ericsson, K. A., Krampe, R Th, & Tesch-Römer, C. (1993). The role of
deliberate practice in the acquisition of expert performance.
Psychological Review, 100, 363–406.
Feigenbaum, E. A., & Simon, H. A. (1984). EPAM-like models of
recognition and learning. Cognitive Science, 8, 305–336.
Simon, H. A., & Chase, W. G. (1973). Skill in chess. American
Scientist, 61, 394–403.

Simple Discrimination
An experimental procedure in which two or more
discriminative stimuli are presented in a series of trials
and one stimulus is always designated as correct (and
responses to it are reinforced). Thus, the stimuli have
an invariant relation to reinforcement and extinction.

Simplified Input
Simplified input is when the speech of NS (native
speaker) is characterized by less complexity but
increased length similar to foreigner talk, whereas in
the latter NS provides an opportunity for NS–NNS
(non-native speaker) interaction (Long 1998).

References
Long, M. (1996). The role of the linguistic environment in second
language acquisition. In W. Ritchie & T. Bhatia (Eds.), Handbook
of second language acquisition (pp. 413–68). San Diego:
Academic.

Cross-References
▶ Computational Models of Human Learning
▶ Deliberate Practice and Its Role in the Development
of Expertise
▶ Development of Expertise
▶ Development of Expertise and High Performance
▶ Learning in the CHREST Cognitive Architecture
▶ Long-Term Learning in SOAR
▶ Problem Solving
▶ Situated Learning

S

Simulation
▶ Models and Modeling in Science Learning

Simulation and Learning
▶ Simulation-Based Learning

S

3072

S

Simulation and Learning: The Role of Mental Models

Simulation and Learning: The
Role of Mental Models

happens when a student compares his own mental
model of a system with that of a simulation.

Theoretical Background
FRANCO LANDRISCINA
Allianz, E-Learning, Trieste, Italy

Synonyms
Conceptual change; Mental models; Mental simulation; Model-based learning; Modeling and simulation;
Organizational learning theory; Perceptual simulations; Thought experiments

Definition
The etymology of the word simulation (Latin: simulō,
imitate) shows that it could also mean to pretend and
therefore have a negative connotation. Simulation is
often seen as a false representation of the real world. In
many Sci-Fi films, the protagonist finds him lost in
a simulated world, created to hide the real world. This
negative connotation can hark back to Plato’s conception of the “mίmZsiς” (mimesis, Greek term for simulation) as the imperfect copy or fictitious replica of
reality. On the contrary, according to Aristotle, the
mimesis is a means to know nature through representations which can be valid and acceptable. This shifting
of the emphasis from imitation to representation corresponds to the shifting of the focus from the superficial aspects of a simulation, for example, interactivity,
to the underlying model, as happens currently in
model-based instruction. In this wider concept of simulation, mental models assume a central role. In his
dialogue “Politics” (Statesman), Plato describes the
cognitive role of the models, or “paradeigmata”
(“paradeίgmati”). In order to illustrate the nature of
the statesman, one of the characters of the dialogue
refers to the model of the weaver, where the technique
of the politician is compared to that of the weaver who
weaves fibers of different nature to create a single
fabric. In general terms, Plato describes the usefulness
of the model as a process of identification of similitudes
and differences. Examining the similitudes and the
conceptual differences between the model and the phenomenon being studied, the subject transforms his
initial ideas, confused and approximate, into a more
precise and rigorous comprehension. This is also what

The role of mental models in simulation was first
highlighted by organizational learning theorists.
According to Senge (1990, p. 8), mental models are
“deeply ingrained assumptions, generalizations, or
even pictures and images that influence how we
understand the world and how we take action.”
Senge fostered the idea of computerized simulations,
called microworlds, which could allow the managers
of an organization to interpret their real roles and
understand how these interact among themselves.
The realization of microworlds of this type is one
of the practical applications of system dynamics, a
computer-aided approach to policy analysis and design
in organizations. System dynamics provides a series of
methods for the elicitation, articulation, and description of the tacit knowledge contained in the mental
models of the experts and for the building of computer-based simulations on the basis of such models
(Ford and Sterman 1998). This corresponds to the
shifting from the concept of model as representation of reality to a concept of the model as a cognitive
artifact.
The relation between mental model and education has been examined in depth by Seel (2003)
who has formulated a learning and teaching theory
based on models (Model-Centered Learning and
Instruction). Model-centered learning can be described
as a progression of mental models, from an initial
state, characterized by the student’s preconceptions,
to a desired final state, of causal explanation. One
of the general functions assigned by Seel (1991) to
mental models is mental simulation. In this interpretative framework, computer-based simulation is
seen as a teaching methodology, able to facilitate
model-centered learning and is particularly effective when the learning objective requires a restructuring of the individual mental models of the
students (seen here as cognitive structures in longterm memory). This is the case, for example, of
the preconceptions with which the student starts
learning scientific concepts, or the mental models
which oppose the change of strategies and behaviors
in the organizations.

Simulation and Learning: The Role of Mental Models

Seel distinguishes between:
● Mental model, the internal and subjective model of

a system
● Conceptual model, which is objective and shared by

a scientific community
● Design and instructional model, used for building

the user interface and the learning tasks
In order to be used as a simulation model, a conceptual model must be formalized in mathematical
terms (such as rules and equations) which can be
developed as computer programs. In a simulationbased learning environment, the student interacts
with such a model only in a mediated way. Faced with
a learning task, he must then build his own mental
model of the phenomenon being studied and use it as
a basis for prediction, inference, and explanation. The
field of Human-Computer Interaction (HCI) offers
a series of design principles that can be applied to
help the students build a correct mental model of the
system. This requires a preliminary analysis to define
the characteristics of the intended users and how they
will be using the software. In the words of Senge (1991,
p. 72): “Simulations with thousands of variables and
complex arrays of details can actually distract us from
seeing patterns and major interrelationships.” In the
case of “black-box simulation models,” the student
can explore the behavior of the system, but the underlying conceptual model remains hidden and can only
be inferred by what is happening on the screen. This
could lead the students to believe their partial conclusions are undisputable assumptions. For example,
a simulation game such as SimCity® presents thousands of possible scenarios and applications, but it
doesn’t show the rules put in the game by its creators.
Facing the events which occur as the consequence of its
own actions, the players tend to automatically attribute
some rules to the system, which can coincide with those
present in the program or be simply the result of their
mental models. Instead, “glass-box models” simulations openly show the relationships between the variables and therefore the hypothesis at the basis of the
model. This too might not be sufficient. It is tempting
to identify the equations of a model with the model
itself, but it has to be borne in mind that the same
model can be implemented using different equations,
and models have different properties than equations.

S

3073

The system model can also be shown in a visual way.
For example, in a simulation of an ideal gas, the underlying physical “billiard ball model” can be shown
through the animation of molecules represented as
rigid spheres, clashing with each other and with the
walls of the container in a perfectly elastic way. For the
purposes of the conceptual change, the construction of
a simulation model is an activity potentially more
effective than the simple exploration of the behavior
of a preexisting model as the students (and the teacher)
can represent, share, and test their mental models.
In the building of a simulation to solve a problem
the student must:
1. Decide how to describe the system to be studied in
terms of objects, relationships, and quantities
2. Recognize which principles to apply to solve the
problem
3. Infer qualitative rules between the values of the
variables
These activities are similar to those described
by VanLehn and van de Sande in their study of the
acquiring of conceptual expertise from modeling in
elementary physics. The expertise in this domain is
characterized by a qualitative understanding of real
world situations. The capacity of solving conceptual
problems arises from extensive practice of a certain
kind and develops following a sequence from a superficial understanding, to a semantic understanding, and
finally to a qualitative understanding. VanLehn and van
de Sande (2009, p. 374) suggest that the students
should not circumvent the model-construction and
model-interpretation processes and focus on increasing their semantic understanding of the models.
In many cases, the restructuring of the mental
models remains nevertheless an objective which is difficult to achieve and, therefore, the extra-technological
factors which characterize the context become decisive,
among which are, for example, the attitudes and the
expectations of the participants, the organizational and
management models, the social relations and the relational systems underlying the use of technologies. Seel
stresses the importance of the relationships between
internal models and external models, where the former
are implicit and individual, while the latter are explicit
and can be shared. From this point of view, the task of
the teacher is that of aiding the students to externalize

S

3074

S

Simulation and Learning: The Role of Mental Models

and discuss their own mental models and internalize
the external ones. In order to do this, he can use the
tools of cognitive mediation between the student and
the simulation model.
Examples of such tools are:
● Verbal language, to give explanations, compare

ideas, and make decisions.
● Images and animations, to visually represent the

change of a system over time.
● Causal maps, to describe the cause–effect relation-

ship among the variables.
● Graphs and diagrams, to study the time-related

behavior of the variables.
In order to favor reflection and self-regulation in
the learning process, educational techniques such as
the following can be used:
● Self-explanation (the student has to explain to him

out loud what he has understood)
● Forecasting (the student must foresee what will hap-

pen in the next step of the simulation)
● Alternation of observation-practice activities carried

out in pairs (the students have to work in pairs and,
in turn, one of them performs the simulation while
the other one observes)
In order to evaluate the change in the mental
models of the student within simulation-based learning environments, many different approaches have
been suggested based on the think-aloud method, cognitive task analysis, concept maps, Bayesian networks,
or on the automated analysis of textual data (Shute
et al. 2009).

Important Scientific Research and
Open Questions
Some questions about the design and use of simulation-based learning environments are directly related
to fundamental issues in the cognitive sciences. Significant cognitive science research supports the hypothesis
that simulation is a fundamental form of computation
in the human brain, and this capability of simulation
could be the basis of many capabilities: from perception
to memory, from language to problem-solving
(Barsalou 2008). Mental simulation is a type of reasoning based on the manipulation of a mental model “in
the mind’s eye.” It can be part of everyday reasoning,
such as when we have to decide how to move a sofa

from one room to the next, or be used by scientists to
carry out “thought experiments” (Trickett and Trafton
2007). According to Nersessian (1999), simulative
modeling is a form of model-based reasoning that
includes mental models that are dynamical in nature,
whereas analogical modeling and visual modeling may
employ static representations. Mental simulations
ostensibly imply the reactivation of patterns of neuronal activity in the perceptive and motor parts of the
brain initially activated by the direct experience of the
event and are therefore also called “perceptual simulations.” The representations involved are in this case of
a “modal” type as they preserve the information relative to the modalities through which their external
referents were experienced. For this reason, a promising research field is that of the role of perceptual factors
in the use of simulations and on the relationship
between imagistic simulations, as used by experts and
scientists in their thought experiments and the type of
visualizations used in simulation-based learning environments to represent the simulated processes (Clement 2009). Abstract concepts can be represented by
spatial and kinesthetic structures, known in cognitive
semantics as “image-schemas.” For example, imageschemas of concepts such as “expansion” and “compression” can play a role in the construction of mental
models of physical or biological phenomena supported
by the simulation (Craig et al. 2002). Perceptual factors
of this nature can be manipulated through stories,
diagrams, and animations, or utilizing new interfaces
based on force-feedback devices and visuohaptic
technologies.
Mental simulation is also one of the possible mechanisms at the base of the “Theory of Mind” (ToM), the
ordinary ability people have to understand their own
minds and those of others. Two contrasting arguments
have been suggested to explain this ability. In the “Theory-Theory” perspective, the ToM is seen as a naive
theory (Folk psychology) with posits, axioms, and rules
of inferences. The “Simulation Theory” (Gordon
2009), on the other hand, posits that man uses his
own mental resources to simulate the psychological
causes of others’ behavior by means of role taking,
that is, by “putting oneself in the other’s place.” The
term “simulation” in this case is used to denote automatic mirroring responses such as the subliminal mimicry of facial expressions and bodily movements (“lowlevel” simulation) or the use of one’s own behavior

Simulation of Dynamic Systems

control system as a manipulable model to predict and
anticipate the behavior of others (“high-level” simulation), in analogy with scientific simulation.
However, mental simulation shows clear limits, the
most important being that it relies on qualitative relations rather than on precise numerical representation.
In simulative reasoning, the inferences derive from the
use of the knowledge embedded in the constraints of
a mental model to produce new states of the model.
Where the situations are more distant from sensorial
experience, there are fewer guarantees that the simulation process will yield success. This is particularly evident in the case of nonlinear and self-organizing
systems, where also very simple equations or rules can
determine complex and unforeseeable behaviors. For
example, by observing the causal map of a system with
feedback circuits, it is not possible to infer the timerelated behavior of the system. Only computer-based
simulation manages to show these behaviors, sometimes counterintuitive or unexpected also for those
who built the model. The other limit of mental simulation is that it cannot be shared with others. The
building of computer-based simulation models can be
useful in this case too, through the shifting from
implicit metal models to explicit conceptual ones. Simulation models can thus extend our biological capacity
to carry out mental simulations and simulative reasoning. Between student and simulation, a form of “cognitive partnering” can then be set up, where the mental
and the conceptual models modify each other in real
time, a circular interaction thanks to which the computer can become a real tool for thinking.

Cross-References
▶ Mental Model
▶ Mental Models and Lifelong Learning
▶ Mental Models in Improving Learning
▶ Model-based Learning
▶ Model-Based Learning with System Dynamics
▶ Model-based Teaching
▶ Models and Modeling in Science Learning
▶ Simulation-Based Learning

S

Craig, D. L., Nersessian, N. J., & Catrambone, R. (2002). Perceptual
simulation in analogical problem solving. In L. Magnani & N. J.
Nersessian (Eds.), Model-based reasoning: Science, technology, &
values (pp. 167–191). New York: Kluwer/Plenum.
Ford, D., & Sterman, J. (1998). Expert knowledge elicitation for
improving mental and formal models. System Dynamics Review,
14, 309–340.
Gordon, R. M. (2009). Folk psychology as mental simulation. In E. N.
Zalta (Ed.), The stanford encyclopedia of philosophy (Fall 2009
Edition).
http://plato.stanford.edu/archives/fall2009/entries/
folkpsych-simulation/. Accessed 3 Mar 2011.
Nersessian, N. J. (1999). Model-based reasoning in conceptual
change. In N. J. Nersessian & P. Thagard (Eds.), Model-based
reasoning in scientific discovery (pp. 5–22). New York: Kluwer/
Plenum.
Seel, N. M. (1991). Weltwissen und mentale Modelle [World knowledge
and mental models]. Göttingen: Hogrefe.
Seel, N. M. (2003). Model-centered learning and instruction. Technology, Instruction, Cognition and Learning, 1, 59–85.
Senge, P. M. (1991). The fifth discipline. The art and practice of the
learning organization. New York: Doubleday/Currency.
Shute, V. J., Jeong, A. C., Spector, J. M., Seel, N. M., & Johnson, T. E.
(2009). Model-based methods for assessment, learning, and
instruction: Innovative educational technology at Florida State
University. In M. Orey, V. J. McClendon, & R. M. Branch (Eds.),
Educational media and technology yearbook, 1 (Vol. 34,
pp. 61–79). New York: Springer.
Trickett, S. B., & Trafton, J. G. (2007). “What if...”: the use of conceptual simulations in scientific reasoning. Cognitive Science, 31,
843–875.
VanLehn, K., & van de Sande, B. (2009). Expertise in elementary
physics, and how to acquire it. In K. A. Ericsson (Ed.), The
development of professional performance: Toward measurement
of expert performance and design of optimal learning environments
(pp. 356–378). Cambridge: Cambridge University Press.

Simulation Model
▶ Computer Simulation Model

Simulation Model of AnalogyMaking
▶ Dynamic Modeling and Analogies

References
Barsalou, L. W. (2008). Grounded cognition. Annual Review of Psychology, 59, 617–645.
Clement, J. J. (2009). The role of imagistic simulation in scientific
thought experiments. Topics in Cognitive Science, 1(4), 686–710.

3075

Simulation of Dynamic Systems
▶ Dynamic Modeling and Analogies

S

3076

S

Simulation-Based 3D Learning Environments

Simulation-Based 3D Learning
Environments
▶ Virtual Reality Learning Environments

Simulation-Based Instruction
▶ Model-Based Learning with System Dynamics

Simulation-Based Learning
CLAUDE FRASSON, EMMANUEL G. BLANCHARD
Department of Computer Science and Operational
Research, University of Montréal, Montréal, QC,
Canada

Synonyms
Learning with games; Similarity learning; Simulation
and learning; Spatial learning; Visual perception
learning

Definition
The term “simulation-based learning” refers to the use
of simulation(s) for learning purposes. Definitions for
the two parts of this term, i.e., “learning” and “simulation,” are provided in the following paragraphs.
Learning is a complex topic as illustrated by the
variety of entries in the current encyclopedia. A broad
definition is adopted in the current entry with learning
defined as the process of acquiring or enhancing
knowledge and skills, whether they are cognitive, physical, social, etc.
Simulations have been used in many if not all
scientific disciplines, which has led to a variety of definitions sometimes reflecting influences and considerations of a specific discipline. For instance, some
researchers suggest that simulations are necessarily
related to computers. However simulations have
emerged ways before the eve of computers’ age; to
some extent, the old Chinese game of GO can be seen

as a war simulation. Other researchers such as Aldrich
(2009) conceive that simulations’ primary goal is oriented toward a learning purpose. Even though the
current entry aims at discussing simulations for learning purposes, it will be later shown that simulations
whose goal is not primary educational can also provide
indirect support to learning, as illustrative elements for
instance. Indeed, it would not be adequate to refer to
a definition that would not consider the learning
potential of simulations that aim at establishing theoretical proofs, or at representing natural phenomena to
cite but a few.
Hartmann’s (1996) definition of simulation is an
interesting step toward a generic definition of this
concept. He states that “a simulation imitates one process by another process.” A key aspect of this definition
lies in the temporal nature of simulations since processes that he refers to are perceived as elements that
evolve over time.
Yet some controversies remain on the very idea
of considering all simulations as genuinely dynamic.
Hughes (1999, see also Grüne-Yanoff and Weirich
2010) insists that simulations sometimes mainly focus
on the imitation of the structural aspects of a model
rather than on its dynamic characteristics, thus making
such simulations relatively static elements.
Another noticeable feature for stressing the genuine
nature of the simulation concept emerges from the
previous point: a simulation is a manifestation of a
model, the latter referring to a set of intertwined elements (such as rules, properties, or lower-order models
for instances) providing a description of a target.
Grüne-Yanoff and Weirich (2010), however, state that
there is not always a clear-cut distinction between
a model and its associated simulation.
Accordingly, in this entry, simulating is defined as
the action of imitating a model’s dynamics and/or
structure with its resulting element, i.e., a simulation
possibly including static and/or dynamic features.
Several kinds of simulations exist and can contribute to learning activities at different levels, in a direct
manner (as the primary educational element) or in an
indirect one (as an illustration or as a positive sidefeature of the learning objective). It is also worth noticing that game-based learning is frequently linked with
simulations: business games are a very popular way to
teach various aspects of economy, business, and management for instance. However it is important to note

Simulation-Based Learning

that there exist games that are not based on simulations
and simulations that are not games.

Theoretical Background
Simulation-based learning can occur in two ways: (a)
through observation, the simulation thus being an
illustration of the content to be learned or understood,
or (b) through experiential learning induced by direct
interactions with the simulation. Specific characteristics of a given simulation can make the former, the
latter, or both learning approaches possible.
Furthermore simulations have been applied in various context that shape their inner properties and
objectives. For instance, simulations have been used:
● In human and social sciences to understand civili-

●

●

●
●
●
●
●

●

zations and represent their past, present, and future
evolutions, or to study social phenomena
In natural sciences to illustrate behaviors of complex systems such as societies of insects or swarms
of birds, or to study the impact of a specific phenomenon on an ecosystem
In cognitive sciences to analyze intertwined mental
processes at stake within the mind of individual
human beings
In medicine to represent virtual patients that would
express symptoms of a specific illness
In chemistry and pharmacology to analyze pros and
cons of interactions between molecules
In physics to test theories and apparatus addressing
atomic to cosmological issues
In engineering to test structures and mechanics,
and ensure their efficiency and safety
In business, management, economy, and finance to
deal with processes occurring at the micro or macro
level through actors of economical systems
In videogames as virtual worlds where a game
occurs that could possibly be populated by simulated inhabitants

Several features have been discussed in the literature
to classify simulations, and influence simulation-based
learning opportunities.

Realization
A same model (i.e., a set of rules and constraints) can
lead to differently realized simulations. Three main
realization classes can be distinguished: (a) simulations

S

3077

involving humans, (b) mechanical simulations, and (c)
computer simulations.
(a) The business simulation game is a well-known
example of simulation involving humans. It is
a learning practice where (teams of) students act
as if they were in charge of a real business. They
may have to deal with stocks, to negotiate with
suppliers in order to get the best deals to purchase
goods, to compete with other teams of students
that run companies with similar objectives, and so
on. Similarly, virtual patient simulations have been
used in medical education for years. Medical students confronted to trained actors playing the role
of a patient with a specific disease can learn
the proper way to detect and cure the targeted
disease. One of the learning interests of humanbased simulations lies in the genuine adaptive
skills of trained human beings who are able to
develop coherent and realistic decision making in
unplanned situations.
(b) Mechanical simulations refer to the use of artificial
apparatus to reproduce natural (waves for
instance) or artificial (machines for instance) phenomena or systems. Some mechanical simulations
such as professional flight simulators can simulate
physical sensations that students could experience
during the real activity. Among other things,
mechanical simulations have been extensively
used as reproductions of complex, costly, and dangerous systems such as spacecrafts, or airplanes,
thus providing situated, secured, and safe learning
environments for people who will have to maneuver them.
(c) Computers are frequently used to process and display models related to various aspects of life, with
equation-based and agent-based simulations being
the two main branches of computer simulations.
As pointed out by Neumann (2010), in equationbased simulations, a centralized approach is
adopted that implies to have a clear a priori theory
about the phenomenon to be simulated so that all
its characteristics are transcribed into a system of
equations. In agent-based simulations, effects,
knowledge, and decision-making abilities are
described and distributed among several entities
known as agents. Hence characteristics of the
global phenomenon are not known a priori but

S

3078

S

Simulation-Based Learning

will emerge from interactions between the different
entities. Several computer simulations also have
both equation-based and agent-based characteristics. Computer simulations nowadays benefit of
the steady computing power progression, which
allows an always greater amount of information
to be considered at the same time and lead to
simulated result closer to reality.
Each kind of realization has specific interests for
learning purpose, such as the ability to adapt to
unpredicted situations, to reproduce physical experiences, or to deal with models that require lots of information. More and more hybrid solutions are developed
to combine various positive properties. For instance,
professional flight simulators use computers to display
cockpit views and to transcribe pilot actions into
commands to a “six degrees of freedom” motion platform that reproduce physical changes (cockpit inclinations). A human simulator officer is also frequently in
charge of the simulation and can trigger events such as
material failures that could strengthen the learning
experience.

Fidelity
Simulations can be characterized on a continuum
going from low to high fidelity. Fidelity refers to the
extent to which the model correctly represents the
element it is supposed to illustrate. From a learning
point of view, high and low fidelity simulations each
have pros and cons.
High fidelity simulations more coherently represent the content to be learned, which results in a
limited deformation between this content and the
simulated counterpart that the learner is dealing
with. Still, high fidelity simulations do not always
provide the best pedagogical opportunities since
many potentially nonrelevant details may have been
modeled that may distract learners from learning
objectives or just reveal too complex for their current
understanding abilities.
Hence, a lower fidelity simulation that would focus
only on pedagogically relevant aspects of the domain to
be learned is sometimes more adequate, especially for
beginners that have to deal with lots of novel information. However lower fidelity is not necessary due only
to an incomplete model, and fantasy elements may
have been added (Hollan et al. 1987). While they

might raise the fun level of a simulation serious game,
and improve the users’ motivation, fantasy elements
can also corrupt the learning message and confuse
learners who could hardly distinguish true learning
information from a fantasized one.

Interactivity
Interactivity refers to the possibility for users not only
to be passive observers of the simulation, but also to
act on it and to influence its evolution. Simulations
can be interactive in different ways. The user could
for instance have the possibility to modify simulation parameters, and learning would result from
observations of changes induced by these modifications. Another possibility is to have the user
directly involved in the simulation, as an inner actor
that could even have its representation, frequently
referred to as “avatar.” Hence the user would directly
interact with characteristics of the simulated phenomenon, and could then extrapolate some experiential
learning.
Constraints sometimes reduce interactivity. For
instance, navigating from one location to another in
a virtual world is sometimes restricted to a limited set
of paths. But in other virtual worlds, freedom of movement is total. Such world simulations are sometimes
labeled as “open world.” Depending on the case, constraints can be perceived as guidance or as restrictions.
Interactivity can be beneficial to simulation-based
learning because it is an efficient way of promoting
motivation for an activity. As such, it can influence
learners’ perseverance in the activity and their willingness to reuse the simulation-based learning system in
the future. Simulations can also propose interactive
ways to answer to learners’ needs of guidance and
explanations. However interactive techniques can also
distract learners from their objectives when they are not
used with caution.

Immersiveness
Virtual environments sometimes provide rendering
fidelity and accuracy in simulated behaviors, particularly if they are delivered with features that trigger
immersive feelings. This frequently results in a better
perception and more adequate reactions. In addition,
pedagogical capabilities can be incorporated into
autonomous agents (Johnson et al. 1998). CAVEs

Simulation-Based Learning

(cave automatic virtual environment) are specific
immersive virtual reality environments where several
projectors are oriented to a curved screen (concave
wall) or to a cube (three to six walls). In such
a setting the user is thus located in the center of
a fully immersive environment with high-resolution
projectors displaying images on each screen. Sometimes images are projected onto mirrors to reflect the
images onto the projection screens, reducing the space
required for the CAVE. Using stereoscopic glasses,
images are separated for each user’s eye that can see
3D stereoscopic objects floating in the air. Users can
evolve around these objects with a dynamic tracking
of their positions which increases the interaction
and simulation between user and environment. They
learn rapidly how to set up, improve, and test the
environment.
The major advantage is that various prototypes
(car, plane, building) or interfaces can be developed
and tested, without spending any money on industrial
development of the complete environment. Reactions
of both the environment and the user into the environment can be tested in advance.

Intelligence
Simulation-based learning can also be enhanced by
intelligent components such as intelligent agents
(Lester et al. 1999) that are able to react to user’s
interaction and to propose various constructive alternatives such as counterexamples, corrections to the
manipulation, or guidance toward correct gestures
and actions. In addition, the system can react intelligently to user’s interactions in order to progressively
present new decision situations. This is particularly
important for instance in medicine where specific
emergency situations can be created, in military strategic cases in which correct decisions must be taken, in
aeronautics, fire-fighters training, and other situations
which require a decision-making process. Simulation
environments can be enhanced by life-like animated
agents such as Cosmo (Lester et al. 1999) that provides
customized advices to support students’ problem solving. A key problem posed by these life-like agents that
inhabit artificial worlds is deictic believability and for
that they have to move through their environment,
pointing judiciously to objects as they provide problem-solving advice. These types of simulations have
a strong contribution to learning.

S

3079

Important Scientific Research and
Open Questions
Simulation-based learning should increase in the
future resulting from improvements of technology
and software. Important questions remain such as
what is the impact of virtual reality environments on
human behavior and decisions? Visualization process is
more and more important and a particular effort is
driven on facilities to produce a simulation.
To increase the realism of simulation (believability),
affective considerations begin to be introduced and
tested with affective agents or avatars. They mimic
emotional reactions, an important part of human
behavior. Several works are under way on this topic.
Other research is focusing on the introduction
of learning through a large variety of games (funny
games, 3D real time game of strategy and adventure).
As it becomes more easy and funny to learn through
these subtleties a lot of work try to investigate ways to
introduce learning into games or games into learning
programs.
Other works are focusing on the introduction of
intelligence into games. This can take into consideration the learning style or the motivation of the learner,
but also several cognitive components such as attention, focus, speed, spatial reasoning, problem solving,
stress, or reaction time. This aspect of intelligence
means that the game should adapt to each user not
only at the beginning of the game but also dynamically
along the game.

Cross-References
▶ Guided Discovery Learning
▶ Imagery and Learning
▶ Simulation and Learning: The Role of Mental
Models
▶ Visualizations and Animations in Learning Systems

References
Aldrich, C. (2009). Virtual worlds, simulations, and games for education: A unified view. Innovate, 5(5). http://www.
innovateonline.info/index.php?view=article&id=727.
Grüne-Yanoff, T., & Weirich, P. (2010). The philosophy and epistemology of simulation: A review. Simulation and Gaming, 41(1), 20–50.
Hartmann, S. (1996). The world is a process: Simulations in the
natural and social sciences. In R. Hegselmann, U. Mueller, & K.
Troitzsch (Eds.), Modelling and simulation in the social sciences
from the philosophy of science point of view (pp. 77–100).
Dordrecht: Kluwer.

S

3080

S

Simulator Model

Hollan, J. D., Hutchins, E. L., & Weitzman, L. M. (1987). STEAMER:
An interactive, inspectable, simulation-based training system. In
Artificial intelligence and instruction: Application and methods.
Boston: Addison-Wesley Longman.
Hughes, R. I. G. (1999). The Ising model, computer simulation,
and universal physics. In M. S. Morgan & M. Morrison (Eds.),
Models as mediators: Perspectives on natural and social science
(pp. 97–145). Cambridge: Cambridge University Press.
Johnson, L. W., Rickel, J., Stiles, R., & Munro, A. (1998). Integrating
pedagogical agents into virtual environments. Presence:
Teleoperators and Virtual Environments, 7(6), 523–546. Cambridge, MA: MIT press.
Lester, J. C., Stone, B. A., & Stelling, G. D. (1999). Lifelike pedagogical
agents for mixed-initiative problem solving in constructivist
learning environments. User Modeling and User-Adapted Interaction, 9(1–2), 1–44.
Neumann, M. (2010). An epistemological gap in simulation technologies and the science of society. In E. G. Blanchard & D. Allard
(Eds.), Handbook of research on culturally-aware information
technology: Perspectives and models (pp. 114–135). Hershey: IGI
Global.

Simulator Model
▶ Computer Simulation Model

Simultaneous Discrimination
Learning in Animals
THOMAS R. ZENTALL
Department of Psychology, University of Kentucky,
Lexington, KY, USA

Synonyms
Contrast; Difference; Distinction

Definition
In a simultaneous discrimination, two (or more) stimuli
are presented simultaneously, and only one is correct
((the S+) and choice of it is reinforced). Choice of the
other(s) (the S) is not. In a successive discrimination,
two (or more) stimuli are presented successively, and
responses to one are reinforced whereas responses to
the other are not. Of particular interest is the effect that
one stimulus has on the other.

Theoretical Background
Psychologists have been studying instrumental discrimination learning in animals systematically for
almost 100 years. In most cases, two stimuli are
presented, and choice of one (often referred to as the
S plus or S+) is correct (reinforced), whereas choice of
the other (often referred to as the S minus or S) is not.
In the case of successive discriminations, each stimulus
is presented individually, and the measure of discrimination learning is typically the ratio of responses to the
S+ relative to total responses to both the S+ and the S
(often referred to as the discrimination ratio), whereas
with simultaneous discriminations, both stimuli are
presented together, a choice is made between them,
and the measure of discrimination learning is generally
the proportion or percentage of choices of the S+.
There are two forms of stimulus interactions that
should be distinguished. The first is based on the properties of the stimuli themselves (the similarity between
them). The more similar the stimuli to be discriminated, the more difficult it is to learn the discrimination. The second is based on the properties of the
procedure. When stimuli are presented simultaneously,
what is learned about one may affect the other because
it is present at the same time. Most of the research on
discrimination learning that has been published in the
past 50 years has used a successive discrimination procedure, perhaps because it is simpler (only one stimulus is presented at a time) and, consequently, it is easier
to interpret the meaning of the rate of responding to
the stimuli. However, recent research using simultaneous discrimination procedures suggests that the
stimulus interactions that occur with this procedure
may be quite different from those that occur when
discriminations are acquired by successive procedures.
In successive discriminations, the effect that one
stimulus has on the other is generally described as
contrast. That is, if an S+ stimulus is alternated with
an S stimulus, the tendency to respond to the S+
stimulus is generally greater than if both of the stimuli
are S+ stimuli. This is known as positive contrast.
Similarly, if two S+ stimuli are alternated and if one
of them improves in value, responding to the other
tends to decrease. This is known as negative contrast.
In other words, one of the stimuli appears to be valued
relative to the other and the better one makes the
poorer one appear worse, while the poorer one makes
the better one appear even better.

Simultaneous Discrimination Learning in Animals

According to traditional theories of discrimination
learning (e.g., Spence 1937), successive and simultaneous discriminations should involve similar learning.
Research has shown, however, that in simultaneous
discriminations, the effect that one stimulus has on
the other is likely to be more complex. That is, it may
increase or decrease the value of the other depending
on the experiences the animal has had with each.

Important Scientific Research and
Open Questions
Major strides in understanding stimulus interactions in
simultaneous discriminations have been made in the
past 20 years. The process began with Fersen, Wynne,
Delius, and Staddon’s (1991) suggestion that in
a simultaneous discrimination, some of the value associated with the S+ transfers (or generalizes) to the S
that is presented with it (an induction-like process)
such that the S+ increases the value of S with which
it was paired. If the notion of value transfer in
a simultaneous discrimination is correct, it would
suggest that stimulus interactions in simultaneous discriminations are opposite to the contrast effects typically found in successive discriminations. A direct test
of value transfer in simultaneous discriminations
was conducted by Zentall and Sherburne (1994). In
one discrimination, A100B0, responses to A were
reinforced on all trials, whereas responses to B were
never reinforced. In the other discrimination, C50D0,
responses to C were reinforced on 50% of the trials,
whereas responses to D were never reinforced. On test
trials, when the pigeons were tested with B and D, they
almost always preferred B over D. Thus, A appeared to
give more value to B than C could give to D, evidence
for value transfer.
An important difference between the successive and
simultaneous procedures is the experience that the
animals have with the S stimuli. With the successive
procedure, the S stimulus is presented by itself so the
animal has ample opportunity to learn about the consequences responding to it. In the simultaneous procedure, however, the animal quickly learns that the S+ is
the better of the two stimuli, and it does not learn as
much about value of the S stimulus. Clement and
Zentall (2000) asked if that could explain the difference
in results between the two procedures. To test this
hypothesis, Clement and Zentall trained pigeons with
A100B0 and C50D0, but they included presentations of

S

3081

each of the two S stimuli, B0 and D0, alone. In this
case, on test trials, when they gave the pigeons a choice
between B and D, the pigeons showed a strong preference for D, evidence for contrast.
Thus, it appears that the effect that the stimuli in
a simultaneous discrimination have on each other
depends on the kinds of experiences that have been
had with the S stimulus. Under typical simultaneous
discrimination conditions in which there is little experience with the S stimulus because learning is quite
rapid, there appears to be little opportunity to build up
inhibition to or avoidance of the S stimulus and value
will transfer from the S+ to the S. However, if inhibition is allowed to build up because of extended experience with the consequences of responding to the S
stimulus, it should result in contrast, resulting in
both an increase in the relative value of the S+ with
which it appears and a decrease in the relative value of
the S itself. That is, extended experience with the
S stimulus in a simultaneous discrimination effectively converts a simultaneous discrimination into
a successive discrimination.
An interesting parallel can sometimes be seen in
human social contexts. If one is introduced to the
friend of a friend, one tends to ascribe some of the
attributes of one’s friend to the new person (value
transfer). Once one gets to know the new person,
however, one is more likely to compare the two individuals and focus on their differences (contrast).

Cross-References
▶ Animal Intelligence
▶ Associative Learning
▶ Behaviorism and Behaviorist Learning Theories
▶ Discrimination Learning Model
▶ Extinction Learning
▶ Generalization Versus Discrimination in Learning
▶ Inferential Theory of Learning
▶ Pavlovian Conditioning
▶ Relational Learning
▶ Signal-Detection Models

References
Clement, T. S., & Zentall, T. R. (2000). Determinants of value transfer
and contrast in simultaneous discriminations. Animal Learning
& Behavior, 28, 195–200.
Spence, K. W. (1937). The differential response in animals to
stimuli varying within a single dimension. Psychological Review,
44, 430–444.

S

3082

S

Situated Cognition

von Fersen, L., Wynne, C. D. L., Delius, J. D., & Staddon, J. E. R.
(1991). Transitive inference formation in pigeons. Journal of
Experimental Psychology. Animal Behavior Processes, 17, 334–341.
Zentall, T. R., & Sherburne, L. M. (1994). Transfer of value from S+ to
S in a simultaneous discrimination. Journal of Experimental
Psychology. Animal Behavior Processes, 20, 176–183.
Further Reading
Riley, D. A. (1968). Discrimination learning. Boston, MA: Allyn and
Bacon.

Situated Cognition
MURAT ATAIZI
Department of Communication, School of
Communication, Anadolu University, Eskisehir,
Turkey

Synonyms
Anchored instruction; Cognitive approach; Communities of practice; Situated learning

Definition
Humans are socially curious beings and learn mostly
through social interaction with others. This social
interaction involves context, culture, activity, discourse, people, and so on. Situated cognition is the
study of human learning that takes place when someone is doing something in both the real and virtual
world, and therefore learning occurs in a situated activity that has social, cultural, and physical contexts.

Theoretical Background
Situated cognition is a theoretical approach to human
learning that supports the idea that learning takes place
when an individual is doing something. Situated cognition has been positioned as an alternative to information
processing theory. Situated cognition theory promises to
complete the symbolic-computation approach to cognition, as information processing theory neglects conscious
reasoning and thought (Wilson and Myers 2000).
Researchers are divided into two camps of situated
cognition theory:
1. Anthropologists like Jean Lave (1988, 1991) and
Lucy Suchman (1993) are interested in the cultural
construction of meaning. These anthropologists

combine anthropology and critical theory with the
socioculturalism of Vygotsky.
2. Allan Collins, John Seeley Brown, Don Norman,
and Bill Clancey are interested in cognition at
individual and social levels. According to these
researchers, situated cognition has strong links to
artificial intelligence, neuroscience, linguistics, and
psychology due to their focus on understanding the
individual mind (pp. 65–66).
Another viewpoint on situated cognition comes
from Kirshner and Whitson (1997). They analyze situated cognition as a research approach that relates
social, behavioral, psychological, and neural perspectives of knowledge and action in their edited collection
of chapters on situated cognition. The contributors
examine the problems of situated cognition using conceptual resources from a broader range of social theory.
They emphasize theories such as situated cognition
theory, situativity theory, and other positions that
seek to better reflect the fundamentally social nature
of learning and cognition and provide new opportunities to reorient and perhaps redesign education.
In the field of educational psychology, situated cognition gained researchers’ attention in late twentieth
century. In the early 1980s, the Soviet scientist Lev
Vygotsky and his colleagues and students became interested in studying human cognitive capabilities, and in
this context they focused on scientific thinking.
“Learning by doing” became the key concept in the
field of learning and education at this time. Brown,
Collins, and Duguid were adapters of this concept in
the field of education (1989). Jean Lave and Etienne
Wenger (1991) came to situated learning and cognition
from the anthropological perspective of communities
of practice. Clancey (1997) studied the changing design
of intelligent machines and examined the implications
of situated action from the perspective of artificial
intelligence scientists interested in building robots
and seeking to relate descriptive models to neural and
social views of language. Clancey (1997) posits that
situated cognition developed not as a discipline within
artificial intelligence or psychology or educational
technology but as a way of thinking by the some of
the well-known scientists of the twentieth century in
psychology, biology, sociology, psychiatry, and philosophy. He states that situated cognition is the study of
how human knowledge develops as a means of

Situated Cognition

coordinating activity within activity itself. The meaning of this explanation is that feedback is very important because it is occurring internally and within the
environment over time. Therefore, knowledge has
a dynamic aspect in both formation and content. This
is a shift from knowledge as stored artifact to knowledge as constructed capability in action (p. 4).
According to this assumption, learning occurs through
feedback inside and outside the brain, but this is not
a simple process because people do not simply make
plans and then do things. They continuously adjust and
invent. Managing this process means managing learning, not managing application (Clancey 1995, p. 39).
Gee (1997) focuses on construction of meaning and
on thinking from a sociocultural approach. Here, construction of meaning is related to the specific context
and purpose. Learning certain types of patterns and
futures in a particular context requires a special culture;
Gee calls this type of cultural interaction “discourses.”
“Discourses are sociohistorical coordinations of people, objects (props), ways of talking, acting, interacting,
thinking, valuing and (sometimes) writing and reading
that allow for the display and recognition of socially
significant identities like being a (certain sort of) African American, Turkish worker in Germany, lawyer,
street-gang member, schoolchild, teacher, feminist
and so on” (pp. 255–256). Discourses help people to
exchange thoughts and explanations. “If you destroy
a Discourse (and they do die), you also destroy its
cultural models, situated meanings, and its concomitant identities” (p. 256). Therefore, discourse plays
a vital role in human learning and cultural development when looking from a situated cognition perspective. New research on situated cognition theory focus
on communities of practice, online communities, and
artificial intelligence with instructional design practices
since discourse focuses attention on human learning.

Important Scientific Research and
Open Questions
Both cognitivism and behaviorism explain learning
and knowledge as resulting from human experience in
the objective world. According to these theories, the
world is a stable place where human knowledge and
values are objective. Learning and knowledge are not
created in an individual’s thoughts, but are determined
by the nature of reality in the objective world.

S

3083

According to Kirshner and Whitson (1997), these
objective limitations and stability create dissatisfaction
among educators, anthropologists, psychologists, and
social theorists. They focus on the opportunity to
explore learning and knowledge as processes that
occur in a local, subjective, and socially constructed
world that is dreadfully limited by both behaviorist
and cognitivist approaches (p. vii, viii).
Situated cognition theory loses some of its potential
to inspire in the arena of education if it is not connected
with real-world situations by instructional design
activities. The work of the Cognition and Technology
Group at Vanderbilt University is a good example of
anchored instruction that has close relation to the
situated learning approach and situated cognition theory. The goal of anchored instruction is to create technology-rich learning environments with video to help
learners to solve authentic, real-world problems. The
students first identify a problem situation through
a collaborative problem formulation and then solve
the problem with their team. Research on situated
cognition theory takes place mostly in real-world settings and, therefore, mixed methods and qualitative
approaches are used by researchers. The application
of this theory into practice has great promise for current and future learning. The next step for situated
cognition theorists is to develop new, innovative
models to explore knowledge and learning as fundamentally more social and cultural, rather than objective
and stable, and to connect these models with practice
to create worthwhile learning environments, both virtual and real.

Cross-References
▶ Cognitive Apprenticeship Learning
▶ Communication and Learning in the Context of
Instructional Design
▶ Constructivist Learning
▶ Culture of Learning
▶ Situated Learning
▶ Transformational Learning

References
Bransford, J., & CTGV. (1990). Anchored instruction and its relationship to situated cognition. Educational Researcher, 19(6), 2–10.
Brown, J. S., Collins, A., & Duguid, S. (1989). Situated cognition and
the culture of learning. Educational Researcher, 18(1), 32–42.

S

3084

S

Situated Learning

Clancey, W. J. (1995). Practice can not be reduced to theory:
Kowledge, representations, and change in the workplace. In
S. Bagnara, C. Zuccermaglio, & S. Stuckey (Eds.), Organizational
learning and technological change (pp. 14–46). New York:
Springer.
Clancey, W. J. (1997). Situated cognition: On human knowledge and
computer representations. Cambridge: Cambridge University
Press.
Gee, J. P. (1997). Thinking, leraning, and reading: The situated
sociocultural mind. In D. Kirshner & J. A. Whitson (Eds.),
Situated cognition. Mahwah: Lawrence Erlbaum Associates.
Kirshner, D., & Whitson, J. A. (1997). Situated cognition. Mahwah:
Lawrence Erlbaum Associates.
Lave, J. (1988). Cognition in practice. Cambridge, England: Cambridge University Press.
Lave, J. (1991). Situated learning in communities of practice. In L. B.
Resnick, L. M. Levine, & Teasley (Eds.), Perspectives on socially
shared cognition (pp. 63–82). Washington, DC: American Psychological Association.
Lave, J., & Wenger, E. (1991). Situated learning: Legitemate peripheral
participation. Cambridge: Cambridge University Press.
Suchman, L. (1993). Response to Vera and Simon’s Situated Action: A
Symbolic Interpretation. Cognitive Science, 17(1), 71–75.
Wilson, B. G., & Myers, K. M. (2000). Situated cognition in theoretical and practical context. In D. H. Jonassen, & S. M. Land (Eds.),
Theoretical foundations of learning environments (pp. 57–88).
Mahwah: Lawrence Erlbaum Associates.

Situated Learning
MURAT ATAIZI
School of Communication, Department of
Communication, Anadolu University, Eskisehir,
Turkey

Synonyms
Cognitive apprenticeship; Communities of practice;
Situated cognition

Definition
Humans are socially curious beings and learn mostly
through social interaction with others. Learning does
not take place in an individuals’ mind, it is situated in
a context in which participation of individuals to the
communities of practice plays a vital role on situated
learning process. Situated learning occurs generally
when an individual is not intended or planned to
learn. Participation and doing take main place in

situated learning. Situated learning take place when
learning is specific to the situation in which it is
learned.

Theoretical Background
Jean Lave and Etienne Wenger are the first presenters of
situated learning, and they argue that learning is situated in certain forms of social coparticipation rather
than asking what kinds of cognitive processes and conceptual structures are involved during the learning
processes. They paraphrase learning as a social engagements in a specific context in which learning takes place
(p.14. Lave and Wenger 1991). The important point
here is that learning is not an individual mind process,
as explained in classical intellectualist theory that the
individual acquires mastery over the processes of reasoning and description, by internalizing and manipulating structures (p.15), whereas Lave and Wenger
argue that the learning takes place when an individual
participates in communities of practice, and learning is
distributed among participants of those communities.
In their view, learning is an integral part of generative
social practice in the lived-in world. Learning can be
explored as legitimate peripheral participation. When
a person become into communities of practice gradually transformed into practitioner, a newcomer becoming old-timer, a member of a community of practice in
which all the task, skills, and knowledge can be learned.
Brown et al. (1989) focus on school learning and
explaining that the conventional school learning
influenced by the culture of school. They are remarking
coherent, meaningful, and purposeful activities as
authentic activities and expressing that these activities
are not seen and found much in conventional schooling
systems. According to them, authentic activities simply
defined as the ordinary practices of the culture (p.34).
Authentic activities are very important for learners
because it is the only way to gain access to the standpoint of practitioners. As a part of situated learning
model, Brown et al. (1989) added “JPFs” – Just Plain
Folks – beside the concepts of students and practitioners. The JPFs shows some similarities during reasoning with the expert rather than the students. The
students reason with Laws with the influence of school
culture but the JPFs reason with casual stories and the
practitioners reason with casual models. In a normal
daily life when JPFs desire to learn a particular set of
practice, they have two options. The first option is

Situated Learning

enculturation through apprenticeship. For instance,
craft apprentices acquire and develop knowledge by
apprenticeship in real-world environments all the
time. The apprentices’ behavior and the JPFs’ behavior
are very similar in this sense. The second option is to
enter into a school as a student. The school setting is
pretty much different than the outside world. School
activities are much consists of precise, well-defined
problems, formal definitions, and symbol manipulations (p.35). The students themselves can acquire and
develop knowledge by the help of teachers and coaches
in the school settings. They are also remarking the idea
of cognitive apprenticeship. Cognitive apprenticeship
claims that social interaction and social construction of
knowledge improve learning both inside and outside of
the school. Cognitive apprenticeship methods try to
enculturate students into authentic practices through
activity and social interaction that are similar with the
craft apprentices’ acquiring and developing knowledge
by apprenticeship in real world environments.
McLennan (1996) arranged and published a situated learning model based upon the notion that
knowledge is contextually situated and is fundamentally influenced by the activity, context, and culture in
which it is used (Brown et al. 1989). The key components of the situated learning model are:
1.
2.
3.
4.
5.
6.
7.
8.

Stories
Reflection
Cognitive apprenticeship
Collaboration
Coaching
Multiple practice
Articulation of learning skills
Technology

1. Stories are playing a very important role for situated learning and for the social construction of
knowledge because they help people keep track of
their discoveries. Also, stories provide a meaningful
structure of remembering for what has been learned
before.
2. Reflection is another significantly important component of situated learning. Norman (1993) expresses
that there is reflective and experiential cognition of
human being. The experiential thinking is faster and
immediate, but the reflective thinking is deeper.
Situated learning combines both reflective and experiential cognition in the communities of practice.

S

3. Cognitive apprenticeship is the other element of
situated learning model. Brown et al. (1989) propose the concept of cognitive apprenticeship. “Cognitive apprenticeship methods try to enculturate
students into authentic practices through activity
and social interaction in a way similar to that evident-in craft apprenticeship (p.37).”
4. Collaboration is another vital element of situated
learning model. Brown et al. (1989) proposed the
four major features of collaborative learning:
(1) collective problem solving, (2) displaying multiple roles, (3) confronting ineffective strategies and
misconceptions, and (4) providing collaborative
work skills.
5. Coaching plays central role between cognitive
apprenticeship and situated learning. Coaching
consists of observing and supporting students
while they are practicing and learning in collaboration. Coaching also provides scaffolding for learning whenever and wherever necessary.
6. Articulation encloses two aspects. The first aspect
is the concept of separating or articulating different component skills in order to learn them more
effectively. Second, articulation refers the goal of
getting students to articulate their knowledge, reasoning, or problem-solving processes in a domain
(McLennan 1996, p. 12).
7. Technology today plays vital and core role to support and create situated-learning environments and
also contributes for enhancing reflective thinking
of students. The most well-known multimedia
program that produced by The Cognition and Technology Group at Vanderbilt is an example of technology supported situated-learning application.
The situated-learning model is a good guide for
instructional designers to create effective situatedlearning environments.

Important Scientific Research and
Open Questions
Situated-learning environments like Algebra Project,
the work of the Cognition and Technology Group at
Vanderbilt University, Shoenfeld’s approach to teaching mathematics, the Foxfire Project, and other new
situated-learning projects both online and offline suggest that situated learning has great promise for learning needs for today and tomorrow. In this regard,

3085

S

3086

S

Situated Prompts in Authentic Learning Environments

situated learning and instructional design areas need
to develop better mixed learning environments to
consider the situated-learning model. A great deal
of work is needed to a fine-tune theory and assess
how to transform this theory into practice McLennan
(1996).
Winn (1996) debates that instructional design
assumes that what people learn is relatively stable
across the situations in which it is used and that people
apply what they have learned in logical and planned
ways. Since situated-learning remarks that human
action is independent on the context in which it occurs,
it is impossible to foresee every situation when designing instruction. Therefore, it is impossible to design
instruction that can prepare students to act appropriately in all situations (p. 59). Arguments show that
situated learning and instructional design have some
discrepancies in their point of view to learning, but
innovative learning environments can be designed by
the help of multiple perspectives of both theories and
researchers.

Cross-References
▶ Authenticity in Learning Activities and Settings
▶ Cognitive Apprenticeship Learning
▶ Constructivist Learning
▶ Culture of Learning
▶ Learning in Practice (Heidegger and Schön)
▶ Model-Facilitated Learning
▶ Open Instruction and Learning
▶ Situated Cognition
▶ Situated Prompts in Authentic Learning
Environments
▶ Social Construction of Learning
▶ Transformational Learning

References
Brown, J. S., Collins, A., & Duguid, S. (1989). Situated cognition and
the culture of learning. Educational Researcher, 18(1), 32–42.
Lave, J., & Wenger, E. (1991). Situated learning: Legitemate peripheral
participation. Cambridge, UK: Cambridge University Press.
McLellan, H. (1996). Situated learning perspectives. Englewood Cliffs,
NJ: Educational Technology.
Norman, D. (1993). Things that make us smart. Reading, MA:
Addison-Wesley.
Winn, W. (1996). Instructional design and situated learning:
Paradox or partnership? In H. McLellan (Ed.), Situated learning
perspectives (pp. 57–66). Englewood Cliffs, NJ: Educational
Technology.

Situated Prompts in Authentic
Learning Environments
HOLGER HORZ
Department of Psychology, Educational Psychology
and Head of Interdisciplinary College of University
Didactics, Johann Wolfgang Goethe-University,
Frankfurt, Germany

Synonyms
Computer-supported learning; Instructional help;
Instructional support; Prompting; Situated learning

Definition
Prompts
Prompts are messages on a computer screen that
require users to react in a specific way. Prompts are
commonly used in computer-based learning environments to enable learning processes and to dint of provide instructional support.

Authentic Computer-Based Learning
Environment
A computer-based learning environment is called
authentic if learners use learning environments to simulate realistic contexts that reflect the way the knowledge will be used in real life.

Situated Prompts
Situated prompts are embedded in the realistic context
of learning environments so that learners do not perceive them as “additional” instructions but rather as an
integral part of the authentic learning environments’
scenario.

Cognitive Load
It is assumed that learning is associated with cognitive
load in working memory, which has a limited capacity.
It is important for effective learning that cognitive load
triggered by learning materials is not higher than the
available working memory capacity.

Theoretical Background
A main point of criticism concerning school and university learning is the separation between knowing and

Situated Prompts in Authentic Learning Environments

doing something. This separation may lead to “inert”
knowledge where learners can express the learned
information but cannot use it. Brown et al. (1989)
argue that authentic activities of learners are necessary
to acquire the active use of a domain’s competencies.
This idea led to the theory of situated learning, which
proposes that learning should happen in contexts
reflecting a useful or meaningful way to apply acquired
knowledge and competencies in real life.
However, a well-known problem is a lack of regular
classrooms’ appropriateness to create authentic learning environments. A possible solution may be computer-based authentic learning environments, which
simulate real-life tasks as an incentive for learning.
The theoretical assumption is that computer-based
authentic learning environments offer opportunities
to apply acquired theoretical knowledge to real life
tasks and, by doing so, prevent inert knowledge.
Although computer-based authentic learning environments are not full-fledged substitutes of experiences,
which learners could have made in praxis, it is widely
accepted that these environments are a useful alternative to real-life setting. Computer-based learning environments (1) allow integration of different authentic
media, (2) make authentic adaptive interactive trainings more likely than other media, and (3) facilitate the
simulation of realistic complex relations between different objects in a learning environment.
Nearly all kinds of learning environments use
instructional support to enhance the learning processes. The rationale behind integrating instructional
support is to offer assistance which can be used when
needed. Following the approach of computer-based
authentic learning environments, not only the learning
content should be situated and embedded in the learning context, but additional instructions should be
linked to the demands of the specific task in the learning situation as well. In addition to the positive effects
of situated learning in general, it is assumed that situating instructional support (e.g., by prompts) will
reduce appearance probability of insufficient use of
instructional support in computer-based learning environments. Because learners often perceive instructional
support as “additional” or “less important” (meaning
not required), a well-designed and adequate situated
prompt (or other kinds of presenting situated instructional support) will reduce such counterproductive
impressions of instructional support.

S

3087

However, integrating situated prompts may further raise cognitive demands on learners. It leads to
an authenticity increase of a learning environment,
which predominantly causes a higher learning environments’ complexity because of integrating additional
non-essential information. Therefore, as a product of
cognitive activities in the working memory (Sweller,
van Merriënboer, and Paas 1998), a higher cognitive
load results from learning with authentic learning environments compared to learning environment without
any added information to enhance the learning environments’ authenticity.
Situated prompts may hamper their positive effects
with respect to the learning outcome. This is because of
restricted processing capacity of human mind, especially
if learning time is limited (van Merriënboer et al. 2006).
In that case, learners will not manage all affordances of
an authentic learning environments and its situated
prompts because the cognitive load is simply too high
(cognitive overload). As a result, learning efficiency
diminishes because of increasing learning time and
learners’ limited endurance.

Important Scientific Research and
Open Questions
The assumption about positive effects of computerbased authentic learning environments and situated
prompts in terms of learners’ acquisition of more elaborated and active knowledge is empirically supported
(Cognition & Technology Group at Vanderbilt 1993).
However, it is even empirically demonstrated that
enhancing authenticity of computer-based learning
environments or integrating situated prompts to prevent inert knowledge (Horz et al. 2009) imposes high
regulatory demands on learners and may hamper their
learning outcomes.
Different factors are identified, influencing the
effectivity of computer-based authentic learning
environments and/or instructional support like situated prompts. A review by Clarebout and Elen (2006)
showed that the type of support device plays an
important role. First, learners mainly use instructional support that delivers additional information, while support devices aimed at supporting
metacognitive skills by inviting students to reflect
on their activities were seldom used. Second, learners’
prior knowledge influences the use of instructional
support.

S

3088

S

Situated Prompts in Authentic Learning Environments

Learners with low domain and/or low computer
knowledge may be overwhelmed by task complexity
in computer-based authentic learning environments
and become uncertain how to deal with them. This
often leads to disorientation problems (“lost in hyperspace”; see Dillon and Gabbard 1998) so that the learning achievement may be disrupted, resulting in worse
learning outcomes. One possibility to solve these problems is to integrate situated instructions in learning
environments by the use of prompts.
Because of learners’ limited cognitive resources in
the working memory, any evaluation of an authentic
learning environment should also take into account the
potential interactions between the support type and
individual learners’ prerequisites (aptitude–treatment
interactions). The proper cognitive resource allocation
is critical to learning, if learners are requested to devote
mental resources to activities not directly linked to
information processing and integration of knowledge.
In this way, the learning process may be disturbed
(Sweller, van Merriënboer, and Paas 1998).
In complex learning environments learners with
low prior knowledge show mostly a reduced learning
success compared to learners with high prior knowledge. When learners with low prior knowledge encounter problems while working with a computer-based
authentic learning environment, they are probably
unable to increase their already high cognitive load
any further and process additional support information as presented by situated prompts. They are simply
“maxed out” by the additional information and the
consequent increasing cognitive load. As a consequence, learning regulation may fail, and the learning
outcome may decline as a whole (cf. Horz et al. 2009).
Based on the reasoning previously outlined,
a dilemma for the implementation of situated instructional prompts arises, which will be a starting point for
future research in that field. Situated prompts have to
be a part of a fully authentic learning environment, and
authenticity leads to a higher cognitive load as outlined
above. Situated prompts must be tailored to increase
learning outcome of the learner. However, if a learning
environment is set up to be fully authentic it must
necessarily have some features that are not relevant
for the learning task, so that learners get additional
information about their role in the learning environments far in excess of the minimum facts, which should
be learned. This leads to an ambiguous result. Added

features, which are integrated to increase the authenticity, need to be processed by the learners and do
increase their cognitive load during learning. In contrast, the added features do support learning activities.
As a consequence, these learning environments
should be designed in a way that learners become
aware of their comprehension failures and use additional information when necessary. Therefore, it is
beneficial to increase and facilitate the use of additional
information by situated prompts. However, it must be
considered that situated prompts may increase cognitive load, especially with learners who have low prior
knowledge. These learners tend not benefit from situated prompts. In future, it will be a challenging technological and psychological research issue to find the
right and perfectly balanced way in which to add information for a higher authenticity in computer-based
learning environments and to regulate added cognitive
load of any kind of instructional support.

Cross-References
▶ Authenticity in Learning Activities and Settings
▶ Cognitive Load Theory
▶ Computer-Based Learning Environments
▶ Interactive Learning Environments
▶ Knowledge Acquisition: Constructing Meaning from
Multiple Information Sources
▶ Learning Technology
▶ Situated Learning
▶ Virtual Reality Learning Environments

References
Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition and
the culture of learning. Educational Research, 18, 32–42.
Clarebout, G., & Elen, J. (2006). Tool use in computer-based learning
environments: Towards a research framework. Computers in
Human Behavior, 22, 389–411.
Cognition & Technology Group at Vanderbilt. (1993). Anchored
instruction and situated cognition revisited. Educational Technology, 33, 52–70.
Dillon, A., & Gabbard, R. (1998). Hypermedia as an educational
technology: a review of the quantative research literature in
learner comprehension, control, and style. Review of Educational
Research, 68, 322–349.
Horz, H., Winter, C., & Fries, S. (2009). Differential benefits of
inserted instructional prompts. Computers in Human Behavior,
25, 818–828.
van Merriënboer, J. J. G., Kester, L., & Paas, F. (2006). Teaching
complex rather than simple tasks: balancing intrinsic and germane load to enhance transfer of learning. Applied Cognitive
Psychology, 20, 343–352.

Skill Learning

Situational Interest
Situational interest refers to the interestingness of
a situation or context; that is, it is generated by the
characteristics, conditions, or features of the social and
nonsocial environment. Situational interests emerge
when the characteristics of an activity, toy, material,
person, or event attract a child’s attention, arouse curiosity or surprise, and invite engagement. Situations
that have elements of novelty, exploration, or surprise
often evoke situational interest. Situational interests are
considered to be context-specific and relatively temporary or transient but may develop into more longlasting personal interests.

S

3089

Skill Acquisition
▶ Task Sequencing and Learning

Skill Acquisition in the Elderly
▶ Aging Effects on Motor Learning

Skill Development
▶ Ability Determinants of Complex Skill Acquisition

Situational Interests
▶ Interest-Based Child Participation in Everyday
Learning Activities

Skill Development During
Adolescence
▶ Adolescent Learners’ Characteristics

Situational Models in Discourse
Processing
▶ Mental Models in Discourse Processing

Skeletal Knowledge
▶ Folk Knowledge and Academic Learning

Skepticism
▶ Children’s Critical Assessment of the Reliability of
Others

Skill
Overlearned behavioral routine resulting from
practice.

Skill Enhancement
▶ Sequence Skill Consolidation in Normal Aging

Skill Growth
▶ Ability Determinants of Complex Skill Acquisition

Skill Improvement
▶ Ability Determinants of Complex Skill Acquisition

Skill Learning
▶ Learning and Consolidation in Autism
▶ Procedural Learning
▶ Routinized Learning of Behavior

S

3090

S

Skill Transfer

Skill Transfer
▶ Model-Based Imitation Learning

Skinner, Burrhus F. (1904–1990)
FREDERICK TOATES
Department of Life Sciences, The Open University,
Milton Keynes, UK

Life Dates
Burrhus Frederic Skinner was born on March 20, 1904,
in the small Pennsylvania railroad town of Susquehanna, where he grew up. Known to most as “Fred,”
his childhood was characterized by a strong curiosity
about the natural world. He spent much time in exploration of nature and in devising gadgets of various
sorts, which was to serve him well later. After leaving
school, Skinner studied at Hamilton College and, on
leaving, decided to become a writer (though only later
was this aim realized and in the form of psychology
texts).
Skinner’s reading of Bertrand Russell and John
Watson exerted a profound influence on him in terms
of philosophical foundations. At 24 years of age, Skinner became a graduate student in the Psychology
Department of Harvard University. He studied the
behavior of rats and pigeons and devised a number of
laboratory gadgets to facilitate this research. Among
these was the apparatus that was later to be called the
“Skinner box” and the cumulative recorder, a piece of
equipment in which a roll of paper was drawn across
a pen. When the rat pressed the lever in the Skinner box
or a pigeon pecked the key, a click on the pen was
triggered such that a record (“cumulative record”) of
the animal’s behavior was formed.
In 1936, Skinner married Yvonne Blue and he
moved to Minnesota to take up a university teaching
position at the University of Minnesota.
His first book, entitled The Behavior of Organisms,
was published in 1938.
During the Second World War, Skinner investigated
the possibility of training pigeons to guide missiles. At
this time, Skinner also devised a heated crib for his

daughter. To hostile commentators, there was confusion between the crib and the Skinner box and
unfounded rumors circulated that his daughter had
committed suicide.
At the end of the war, Skinner embarked on writing
a Utopian novel, entitled Walden Two. This described
a community of people who shun materialism and
punitive means of control, leading a life based upon
egalitarianism and mutual positive reinforcement. It
was something of a Green proclamation, years before
the political notion of “Green” was in vogue.
In 1945, Skinner moved to take up the Chair of
Psychology at the University of Indiana in Bloomington. A society dedicated to studying behavior within
the Skinnerian tradition, termed the Society for the
Experimental Analysis of Behavior, was formed and
held its first meeting in Indiana in 1946. A journal
dedicated to the publication of articles within this
tradition, the Journal of the Experimental Analysis of
Behavior, first appeared in 1958.
Skinner moved back to Harvard University in 1948
to take up a teaching position. In 1953, he published
the book Science and Human Behavior, followed in
1957 by Schedules of Reinforcement with Charles Ferster.
The year of 1953 saw an important (“epiphany”)
moment in Skinner’s life when he attended a class in
which his daughter was being taught mathematics. He
was struck by the pace of the class, which was too fast
for some pupils and too slow for others. Generic feedback, which was given to the whole class, was inevitably
delayed relative to each individual pupil’s behavior and
often inappropriate to the individual. From this observation, he was moved to devise a teaching machine, in
which each pupil paced himself or herself. Feedback
was thereby appropriate to each individual pupil and
was presented immediately after the response by the
pupil. Problems were broken down into small parts so
that answers were reinforced before going onto the next
question.
In 1957, he published his book on language entitled
Verbal Behavior, which formed the target of a famous
and highly critical review written by Noam Chomsky.
Skinner was concerned that behavioral psychology
be used to solve human problems on a national and
global scale. His conviction was that social problems
arise in large part from the failure of public authorities
to utilize positive reinforcement and instead to use
punitive controls. In 1971, he produced a best seller

Skinner, Burrhus F. (1904–1990)

entitled Beyond Freedom and Dignity. This argued for
abandoning the notion of autonomy and accepting
a scientific understanding of human behavior and the
principles of reinforcement. The book sparked outrage
from many offended individuals including the US Vice
President Spiro Agnew.
Starting in 1976, Skinner published three volumes
of autobiography.
In 1989, doctors discovered that Skinner had leukemia. Nonetheless, he carried on working right up to
the day of his death, giving a presentation to the American Psychological Association 10 days prior to this. He
died on August 18, 1990.

Contribution(s) to the Field of
Learning
In human and nonhuman species, Skinner identified
many of the important parameters that determine
learning, such as partial reinforcement (a schedule in
which only certain responses are reinforced). He was an
early advocate of the need to treat each learner as an
individual. What might be positively reinforcing for
one learner might be punishing to another. That is to
say, reinforcement was defined, not in terms of any
intrinsic physical properties of reinforcers, but by its
effect on behavior. Skinner drew attention to the everpresent process of reinforcement that is often
overlooked or denied. For example, viewed within
this perspective, teachers need to observe carefully the
effects of any interventions in the teaching situation. In
teaching, Skinnerian approaches identify hidden and
subtle sources of reinforcement. For example, what
might intuitively be felt to be punishing (i.e., lowering
the probability of the instance of behavior being
shown), might, in reality, be positively reinforcing
(i.e., increasing the probability of the behavior being
shown). Criticism given to bad behavior, though
thought to be punishing, might, in reality, be positively
reinforcing, via attention seeking.
Reinforcers are usually seen to be such things as
food, water, and praise. However, behavior itself might
exert a reinforcing effect. For example, in smoking cigarettes, the action of smoking might form part of the
reinforcement, in addition to any reinforcing component arising from the absorption of nicotine. Intrinsic
biological changes might also constitute part of reinforcement, as in the adrenalin surge on engaging in risky
activities. It is often suggested that social reinforcers are

S

3091

crucially important as in forming part of a social group,
as in criminal gangs.
Skinner was the inspiration behind using tokens as
reinforcers, as in mental hospitals. Patients who were
hitherto unresponsive could earn tokens for desired
behavior. The tokens would later be exchanged for
rewards.
Skinnerian psychology acknowledges that some
behavior is not the consequence of reinforcement, as
in following rules. However, an analysis of behavior can
often show where a process of reinforcement produces
behavior that acts in opposition to rules and sanctions.
Addiction to drugs illustrates the power of such
reinforcement.
Skinner was a committed humanitarian who
campaigned tirelessly against the use of aversive controls of behavior. This was on the grounds of their
unethical nature and their relative ineffectiveness as
controls of behavior.
Skinner’s invention of the teaching machine was an
early forerunner of what is now embedded in the much
more sophisticated versions that utilize personal computers but the principles of self-pacing and guidance
through feedback is the same.

Theoretical Background
In terms of the specifics of the principles of learning,
the intellectual framework for Skinner was provided
by the work of Pavlov, Watson, and Thorndike. The
theoretical model current at the time of the initiation
of his investigation was one described by the term
“stimulus-response,” in which an instance of behavior,
the response, was said to be triggered by an immediately preceding event, defined as the “stimulus.” Skinner felt that the kind of behavior that he studied in the
Skinner box could not be accounted for by such
a model. Rather, behavior was more closely associated
with the consequence of earlier behavior, i.e., the
history of reinforcement. Such behavior, which “operated” upon the environment, was termed “operant
behavior.” It is a special case of “instrumental behavior,” the broad class being defined in terms of behavior
being instrumental in the outcome, e.g., behavior is
instrumental in getting to food in a maze. Where the
animal freely paces its own activity, the process of
changing behavior by means of the consequences
that follow from behavior is known as “operant
conditioning.”

S

3092

S

Sleep or Offline Learning

Cross-References
▶ Behaviorism and Behaviorist Learning Theories
▶ Conditioning
▶ Insight Learning or Shaping
▶ Operant Behavior
▶ Thorndike, Edward L. (1874–1949)
▶ Watson, J.B. (1878–1958)

References
Bjork, D. W. (1997). B.F. Skinner: A life. Washington: American
Psychological Association.
Skinner, B. F. (1976). Particulars of my life. London: Jonathan Cape.
Skinner, B. F. (1979). The shaping of a behaviorist. New York: Alfred
A Knopf.
Toates, F. (2009). Burrhus F Skinner. Basingstoke: Palgrave Macmillan.

Sleep or Offline Learning
▶ Song Learning and Sleep

Sleep-Dependent System
Consolidation of Memory
▶ Reactivation and Consolidation of Memory During
Sleep

Small Group Learning
▶ Collaborative Learning
▶ Cooperative Learning

Small Group Learning
Strategies
▶ Collaborative Learning Strategies

Small Groups
▶ Cooperative Learning Groups and Streaming

Small Learning Communities
▶ Community of Learners

Small Schools
▶ Community of Learners

Smaller Learning Communities
▶ Community of Learners

Smart Phone
This is a mobile phone that offers more advanced
computing ability and connectivity than a contemporary basic mobile phone. It is an intelligent mobile
phone that has basic mobile phone functions and additional functions such as Personal Digital Assistants
function, Internet ability, remote function, and touch
screen interface. Smartphone also allows the user to
install and run applications based on a specific platform.

Social Affective Learning
▶ Learning the Affective Value of Others

Social and Cognitive
Underpinnings of Adolescents’
Learning
▶ Adolescent Learners’ Characteristics

Social and Cultural Resources at
Home
▶ Family Background and Effects on Learning

Social Cognitive Learning

Social and Emotional
Experiential Learning
▶ Learning by Feeling

Social Bookmarking
With social bookmarking Internet users organize bookmarks of web resources. As opposed to bookmarking
on one’s own computer, you can share digital resources
by publishing the references of the resources online
with social bookmarking. Usually social bookmarks
also contain metadata and key words (called tags) so
that other users may immediately understand the content of the resource.

Social Brain
▶ Machiavellian Intelligence Hypothesis

Social Cognition
▶ Learning in the Social Context

Social Cognitive Learning
EUGENE SUBBOTSKY
Department of Psychology, University of Lancaster
University – Lancaster University, Lancaster, UK

Synonyms
Imitation; Observational learning; Social learning

Definition
Social cognitive learning occurs when an individual
learns from other members of the group by observing

S

3093

and imitating their behavior. In order to imitate
another person an individual has to establish a correspondence between this person’s behavior and that of
his or her own, and this requires a certain level of
cognitive development. Some researchers argue that
certain species of animals, like chimpanzees, are capable of imitation, and human infants are able to do this
from the earliest hours following birth, for instance, by
imitating their mother’s facial expressions.

Theoretical Background
Social cognitive learning theory is a theory that aspired
to make for the drawbacks of the traditional learning
theory based on the concept of classical conditioning
(Eysenk 1960; Skinner 1971). According to social learning theorists, learning, particularly in the area of social
behavior, cannot be satisfactorily explained by the concept of social reinforcement of new forms of behavior,
because it does not explain how these new forms of
behavior appear in the child’s behavioral repertoire for
the first time. These theorists postulate that certain
forms of behavior, such as imitation, are selfreinforced, and therefore children can learn directly
from observing other people’s behavior. Bandura and
Walters (1964) indeed demonstrated that direct observation of models elicit apparently unconditional imitation in children. The theoretical explanation of this
“unconditional” imitation can be found in Mowrer’s
(1950) theory of “primary identification.” According to
this theorist, primary identification emerges in the first
3 years of life, when the child’s caregivers’ behavior
becomes associated in the child’s mind with all sorts
of positive emotional experience that the child gets
from his or her caregivers in the process of routine
interactions, such as feeding, caressing, etc. This association brings in the hidden motive of the apparently
“unconditional” imitation: through imitating the caregivers’ behavior, the child reexperiences the pleasant
emotions that had been previously associated with the
caregivers. The mechanism of “primary identification”
can explain the motivational underpinning of learning
in both cognitive and moral domains, yet it is limited in
regard to the explanation of causes of antisocial behavior in cognitively sophisticated individuals. According
to Bandura (1991), the most sophisticated moral
thinking and reasoning cannot provide a direct link
to moral conduct: Along with creating an invariant
control system within a person that prompts the person

S

3094

S

Social Cognitive Learning

to cherish moral values, cognitive development also
creates mechanisms by which self-sanctions can be
disengaged from antisocial conduct. Bandura (1999)
provided a detailed description of those mechanisms,
such as moral justification of reprehensive conduct,
diffusion of responsibility, minimizing detrimental
effects and consequences of immoral acts, and dehumanization of victim.

Important Scientific Research and
Open Questions
Bandura et al. (1996), studied 10–15-year-old adolescents’ proneness to moral disengagement (such as to
resort to moral justification, euphemistic labeling,
advantageous comparison, displacement and diffusion
of responsibility, distortion of consequences, and
dehumanization of victim) and found high correlations between moral disengagement and delinquent
behavior. In his more refined version of social cognitive
learning theory, Bandura (1991) postulates the concept
of “moral agency” – a mechanism through which
“moral reasoning is translated into actions through
self-regulatory mechanisms rooted in moral standards
and self-sanctions” (Bandura 1999, p. 2). In the course
of moral development, the emergence of moral agency
can provide gradual substitution of external sanctions
and demands by symbolic and internal moral motivation. In the above study (Bandura et al. 1996), children
who were prone to moral disengagement exhibited
a higher level of interpersonal aggression and delinquent behavior than individuals who maintained
a high level of moral agency. Other theorists, however,
argue that social cognitive learning theory alone cannot
exhaustively account for learning processes such as
moral development. According to these theorists,
moral development cannot be reduced to cognitive
mechanisms, such as observation and imitation, but
has to include emotional processes as well. For
instance, Baumrind (1967) found that children who
displayed most competent and mature social and
moral behavior had loving, demanding, and understanding parents. In contrast, restrictive, punitive, and
unaffectionate parents mostly had disphoric and
socially immature children. Although the connection
between child-rearing practices and children’s psychological development is a complex one and depends on
children’s individual differences (Kochanska 1991;
Kochanska et al. 1989; Lamb 1982), the view prevails

that parental warmth, cooperation, and helping in
treating their children; understanding children’s
needs; and respecting the child’s self-esteem facilitates
children’s moral development (Damon, 1988; Dunn et
al. 1995; Higgins, 1989; Maccoby, 1980; Maccoby and
Martin, 1983; Walker and Talor, 1991). While it is true
that children acquire the positive moral image of themselves through the mechanisms of social cognitive
learning, this image can only become attractive if it
acquires an emotionally positive value for the children.
This positive emotional value can only be achieved if
children are treated in a trusting and loving way, if they
have a fair share of their parents’ time and personal
attention to their needs.

Cross-References
▶ Moral Learning
▶ Prosocial Learning in Adolescence: The Mediating
Role of Prosocial Values
▶ Social Construction of Learning
▶ Social Interactions and Effects on Learning
▶ Social Learning
▶ Social Learning Theories
▶ Socialization-related Learning

References
Bandura, A. (1991). Social cognitive theory of moral thought and
action. In W.M. Kurtines, & J.L. Gewirtz (Eds.) Handbook of
moral behavior and Development (Vol 1: Theory, pp. 45–101).
Hillsdale, NJ: Lawrence Erlbaum.
Bandura, A. (1999). Moral disengagement in the perpetration
of inhumanities. Personality and Social Psychology Review, 3,
193–209.
Bandura, A., & Walters, R. H. (1964). Social learning and personality
development. New York/Toronto/London: Holt.
Bandura, A., Barbarinelli, C., Caprara, G. V., & Pastorelli, C.
(1996). Mechanism of moral disengagement in the exercise of
moral agency. Journal of Personality and Social Psychology, 71,
364–374.
Baumrind, D. (1967). Child care practices anteceding three patterns
of preschool behavior. Genetic Psychology Monographs, 75, 43–88.
Damon, W. (1988). The moral child. New York: Free Press.
Dunn, J., Brown, J. R., & Maquire, M. (1995). The development of
children’s moral sensibility: Individual differences and emotion
understanding. Developmental Psychology, 31, 649–659.
Eysenk, A. J. (1960). The development of moral values in children:
The contribution of learning theory. The British Journal of Educational Psychology, 30, 11–21.
Higgins, E. T. (1989). Continuities and discontinuities in self-regulatory and self-evaluative process: A developmental theory relating
self and affect. Journal of Personality, 57, 407–444.

Social Construction of Learning
Kochanska, G. (1991). Socialization and temperament in the
development of guilt and conscience. Child Development, 62,
1379–1392.
Kochanska, G., Kuczynski, L., & Radke-Yarrow, M. (1989). Correspondence between mothers’ self reported and observed childrearing practices. Child Development, 60, 56–63.
Lamb, M. E. (1982). What can “research experts” tell parents about
effective socialization? In E. Zigler, M. E. Lamb, & I. L. Child
(Eds.), Socialization and personality development. Oxford:
Oxford University Press.
Maccoby, E. E. (1980). Social development. Psychological growth and
the parent-child relationship. New York: Harcourt Brace.
Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of
the family: Parent-child interaction. In P. H. Mussen (Series Ed.)
& E. M. Hetherington (Vol.Ed.), Handbook of child psychology:
Vol.4. Socialization, personality, and social development (4th ed.,
pp.1–101). New York: Wiley.
Mowrer, O. H. (1950). Learning theory and personality dynamics.
New York: Ronald Press.
Skinner, B. F. (1971). Beyond freedom and dignity. New York/Toronto/
London: Bantam and Vintage Books.
Walker, L. J., & Taylor, J. H. (1991). Family interactions and the
development of moral reasoning. Child Development, 62(2),
64–283.

Social Cognitive Theory
▶ Social Learning
▶ Social Learning Theory

Social Competence
▶ Socialization-Related Learning

Social Construction of Learning
CURT DUDLEY-MARLING
Lynch School of Education, Boston College, Chestnut
Hill, MA, USA

S

3095

Definition
Researchers who emphasize a social construction of
learning present learning as a social and cultural process
that occurs in the context of human relationships and
activity and not just “in the heads” of individual
learners. In this “social” formulation of learning, the
▶ sociocultural context is not merely the location of
learning. The sociocultural context affects how people
learn (through participation in ▶ cultural activities)
and what is learned (▶ social practices), and is itself
part of what is learned. Crucially, psychological (learning) processes are not independent of the sociocultural
context; indeed, they are constituted by the context of
which they are a part (Cole 1996; Gee 2008).

Theoretical Background
In the fields of psychology and education, cognition
has tended to be situated in the heads and bodies of
autonomous individuals for whom learning is largely
a solitary, mental activity. Dominant metaphors for
thinking and learning which conceive of the mind as
a storage device or an information-processing machine
reflect this “in-the-head” conception of human learning. This highly individualistic conception of learning
is, however, being challenged by developments in the
fields of cultural and clinical psychology (e.g., activity
theory, situated learning, narrative therapy), the learning sciences (e.g., sociocultural studies of mind, constructivism), sociology (e.g., social constructivism),
anthropology (e.g., ethnography, ethnomethodology),
sociolinguistics (e.g., discourse analysis), and related
disciplines (e.g., literary criticism, cultural studies, disability studies).
Social constructions of learning differ from primarily psychological, in-the-head conceptions of learning
in terms of what is learned; where and how it is learned;
and, where knowledge (what has been learned) resides
or is stored. It is important to acknowledge, however,
that there are a range of perspectives on the social
construction of learning that vary according to the
degree to which they emphasize the social and psychological dimensions of learning. The perspective
presented here reflects the social end of the psychological–social continuum.

Synonyms

What Is Learned

Distributed cognition; Situated learning; Social constructivism; Socialization of intelligence

Conventionally, learning is understood to be a matter
of acquiring knowledge and skills. Children learn to

S

3096

S

Social Construction of Learning

talk, read, and write, for example, by acquiring a set of
skills associated with effective reading, writing, and
speaking. However, a social construction of learning
holds that people do not learn to read, write, or speak
“once and for all.” Instead, they learn to read, write, and
speak in particular ways, for particular purposes and
audiences, in particular social settings (Gee 2008). Specifically, reading, writing, and speaking are sets of social
practices that enable participation in particular cultural
activities in various cultural communities. Crucially,
these social practices involve more than mastering
a set of linguistic skills (vocabulary, syntax, etc.). Language practices, for example, always involve ways of
behaving, thinking, interacting, valuing, believing,
and speaking, in short, becoming particular kinds of
people (Gee 2008). The use of language as part of
a religious ritual, for instance, includes saying and
doing just the right things in just the right way and
the right time in the appropriate setting. Moreover, the
right sort of clothing, demeanor, and attitude and
beliefs will also be necessary for this social practice to
“count” as a religious expression. In other words, participating in this social practice involves being
a particular kind of (religious) person. Other social
practices, entailing different purposes, settings, participants and so on, involve learning to become different
sorts of people. In this sense, learning implies becoming a certain kind of person, indeed, becoming different
sorts of people (i.e., different identities) in various
social and cultural settings. In other words, what is
learned are different relationships among and between
people in various contexts; therefore, learning always
involves the construction of identities (Lave and
Wenger 1991). Ultimately, the community will adjudicate learning, that is, whether people demonstrate they
have learned by showing that they have learned to be
the “right sort of person” in particular social/cultural
settings.
A social construction of learning indicates that
what people learn is largely about participating in various communities. This is also where meaning resides
since people always make sense from a particular (cultural) perspective. But this is not just about learning in
(a social) context. The sociocultural context is not just
where learning takes place or where knowledge is
constructed; it is part of what is learned (Cole 1996).
So what is learned is always social and, therefore, all
knowledge carries social and cultural meanings. Even

what appears to be solitary learning has social character
since people learn as members of social and cultural
groups and what they learn is related to social practices
that are enacted in communities. Bakhtin famously
argued that language – perhaps the most powerful of
cultural tools – is far from a neutral medium since is it
is “overpopulated with the intentions of others”
(Bakhtin 1981, p. 294). Language, the learning tool
extraordinaire, is saturated with social, cultural and
historical meanings; therefore, learning mediated
by language – as is usually the case for higher-order
learning – will always have a social character. Given that
the primary tool for learning, language, is imbued with
a social and cultural quality, even what appears to be
solitary learning has a social character.

Where and How Learning Occurs
Conventionally, it is understood that learning occurs in
the minds and bodies of autonomous individuals. In
contrast, a social construction of learning emphasizes
the profoundly social/cultural character of learning. In
this formulation, learning is a process that takes place
in participation frameworks and not in the minds of
individuals. It is in the context of engagement with
participation frameworks that people learn to engage
in cultural activities (social practices) that make them
into particular kinds of people (Lave and Wenger
1991). However, people cannot participate or learn to
participate in cultural activities or practices on their
own. Effective participation in cultural practices
requires coordinated activity among members of the
cultural community and it is this coordinated activity
that is learned. Moreover, coordinated action is how
learning occurs and where it occurs.
The metaphor of a dance is a useful, if somewhat
oversimplified, way to explicate this coordinated action
in which various members of the community work
(dance) together to perform social practices that enact
particular (learning) identities for each of the participants. Consider, for example, the context of formal
school learning. The teaching–learning interaction can
be likened to an intricate dance to which both students
and teachers contribute. This dance is mediated by the
curriculum, school policies, the culture of the school and
classroom (and schooling more generally), the language
of students and teachers, and so on. Take the classic
initiation (teacher asks a question) – response (student
answers the question) – evaluation (teacher evaluates

Social Construction of Learning

the correctness of the student’s response) (IRE) paradigm, for example. In the context of the IRE paradigm
learning has a very particular meaning but for students
to succeed – or fail – requires the coordination of
a particular set of moves by the teacher and the student.
It also takes someone with the authority to adjudicate
the student’s response, in this case the teacher. Importantly, different moves by the teacher (or the student)
will alter the shape of the dance, potentially
transforming students’ and teachers’ identities. If,
instead of responding to an unexpected response by
a student with an evaluation (the “E” in I-R-E), the
teachers asks, “what do you mean?” she transforms her
relationship with her student and, ultimately the meaning of learning. In either case what students and
teachers learn is a socializing routine, a particular set
of moves or practices that “count” as learning in the
classroom. They also learn an identity. In the case of
I-R-E, students are positioned as “getters of the answer
the teacher has in mind.” In the alternative example,
students are positioned as thinkers or explainers.
The meaning of coordinated actions – in this case
school learning – also relies on just the right people
doing just the right moves in a particular time and
place. The I-R-E paradigm, for example, would signal
a very different meaning performed outside the context
of a classroom (in say, a bar) or if the evaluative move
was performed by someone who had not been granted
the authority to adjudicate the performance of the
other party.
The teaching–learning “dance” in a classroom has
a particular character that sets it apart from other kinds
of learning. But the concept of learning as coordinated
action among particular people performing just the
right action at the right time and place applies to all
forms of learning from learning to drive a car, learning
to cook, learning to read, write, or speak, or learning to
make love, to learning to solve a math problem. These
are all social/cultural practices that are learned through
the process of coordinated actions in contexts where
these practices are being performed. From the perspective of a social construction of learning, learning is very
much context specific, guided by others, and mediated
by particular cultural tools and artifacts. Returning to
the dance metaphor, learning the dance depends on
people working together to accomplish a particular
set of moves that will be recognized as dancing. These
moves cannot be accomplished by a person acting on

S

3097

their own or without the aid of music and, perhaps,
props (certain dances require specific clothing, spaces,
etc.). Again, learning is never a solitary activity accomplished by autonomous individuals. Things (i.e., neurological activity) do go on in peoples’ heads, but those
processes are inextricably linked to social, cultural, and
historical factors. Learning is in relations among people
in coordinated activity in, with, and emerging from the
socially and culturally constructed world (Lave and
Wenger 1991). This is what is learned, how it is learned,
and where learning occurs.

Where Learning Resides
Social/cultural practices are what people learn, how
they learn, and the location of learning. Therefore,
learning resides in the practice; it is distributed across
cultural practices in the actions of people as they participate in (cultural) activities, in the cultural tools (like
language) that have been created as part of the practices, the social setting, and so on. Returning to the
metaphor of the dance, the “memory” of the dance
simultaneously inhabits the dancers, the music, the
venue created for the performance of the dance, even
the dancers’ shoes and clothing. In other words, the
memory of the dance is in the dance. Learning to
participate in the school practice known as InitiationResponse-Evaluation (see above) resides in teachers, students, curriculum, and testing materials, the physical
organization of the classroom (teacher standing in front
of students seated in rows), and most of all in the language teachers and students draw on to enact this practice. Ultimately, learning does not belong to individual
persons, but to the various cultural activities or practices
of which they are a part (McDermott 1993). In this
way, learning is distributed across people as they participate in culturally relevant activities (Robbins 2005).

Important Scientific Research and
Open Questions
Experimental and quasi-experimental approaches to
research based on assumptions about learning as the
possession of autonomous individuals are inadequate
for investigating questions animated by a social construction of learning. Ethnographic research methods
are better suited to capturing the immediate context of
learning but still fall short of accounting for the nearly
infinite complexity of social, cultural, and historical
factors that comprise the sociocultural context of

S

3098

S

Social Constructivism

learning. So there is a need for the development of
innovative research methodologies that capture the
complexity of human learning in and through social
and cultural practices within particular learning contexts. This project might also include the search for new
metaphors for understanding learning. A social construction of learning rejects container or computerprocessing metaphors for human learning but language
in cultures that valorize the autonomous individual
is generally inadequate for describing sociocultural
models of learning.

Cross-References
▶ Activity Theories of Learning
▶ Cultural–Historical Theory of Development
▶ Culture of Learning
▶ Learning Identity
▶ Situated Cognition
▶ Social-Cultural Research on Learning
▶ Socio-constructivist Models of Learning

and, consequently, shapes the actions and reactions of
group members. Social constructionist explanations
for phenomena (for example, behavioral differences
between men and women which result from negatively
stereotyping the later) are at odds with essentialism.

Cross-References
▶ Learning in the Social Context
▶ Social Construction of Learning

Social Curiosity
▶ Interpersonal Curiosity

Social Emotional Learning
▶ Social Learning of Fear

References
Bakhtin, M. M. (1981). The dialogic imagination: Four essays. H.
Michael (Ed.), Trans. caryl emerson and Michael Holquist
(pp. 259–422). Austin/London: University of Texas Press.
Cole, M. (1996). Cultural psychology: A once and future discipline.
Cambridge: Belknap.
Gee, J. P. (2008). Social linguistics and literacies: Ideology in discourses
(3rd ed.). New York: Routledge.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral
participation. New York: Cambridge University Press.
McDermott, R. P. (1993). The acquisition of a child by a learning
disability. In C. Chaiklin & J. Lave (Eds.), Understanding practice:
Perspectives on activity and context (pp. 269–305). New York:
Cambridge University Press.
Robbins, J. (2005). Contexts, collaboration, and cultural tools:
A sociocultural perspective on researching children’s thinking.
Contemporary Issues in Early Childhood, 6(2), 140–149.

Social Constructivism
This primarily sociological theory of knowledge and
reality posits that no characteristics are inherent in
objects or individuals, but that meaning is negotiated
through interaction with others who, together, create
a social reality. This reality, constantly formed and
changed through choices, reinforcements and encounters with others, shapes both what is perceived as real

Social Graphs
▶ Social Networks Analysis and the Learning Sciences

Social Influence and the
Emergence of Cultural Norms
MICHAEL MÄS1, JAMES A. KITTS2
1
Department of Sociology/ICS, University of
Groningen, Groningen, The Netherlands
2
Graduate School of Business, Columbia University,
New York, NY, USA

Synonyms
Clustering; Opinion dynamics; Polarization

Definition
It is a truism that everyday social interaction leads to
interpersonal influence. We often see that influence
leads interaction partners to grow more similar in
opinions, values, and behaviors. This pattern may be

Social Influence and the Emergence of Cultural Norms

S

because interaction partners exchange arguments and
persuade each other directly, or may be due to imitation or social learning among peers, or may also be
because interaction partners are exposed to common
external influences, such as groups that pressure members to conform to shared norms. Cognitive consistency theories provide one account for this increasing
similarity among friends, by assuming that people have
a motivational drive to be similar to people that they
like or respect. In order to resolve the dissonance
caused by disagreeing with friends, actors may try to
convince friends to adopt similar opinions or may
change their own opinions to conform to their friends.
Social influence can lead to complex dynamics
in groups where several individuals influence each
other repeatedly (Mason et al. 2007; Latane et al.
1994). Researchers have developed ▶ formal models
to investigate these dynamics in ▶ social networks,
and applied them to theoretical problems like the
emergence of shared ▶ cultural norms, the persistence
of cultural ▶ diversity, the development of organizational culture, conflicts in work teams, and clustering
and ▶ polarization of political opinions.

has a very simple influence network – a line. That is,
actor 1 is only influenced by actor 2. Actor 2 is only
influenced by actors 1 and 3. Actor 3 is only influenced
by actors 2 and 4 and so on. At the beginning of the
process, actors in the interior of the line do not change
their opinions because the influences by their two
neighbors balance each other out. However, actors 1
and 10 are each influenced in only one direction by
a single neighbor – who has a less extreme opinion – so
they develop more moderate opinions. This compromise by actors 1 and 10 causes the opinions of actors 2
and 9 to become less extreme, triggering opinion
adjustments of actors 3 and 8, and so on.
As Fig. 1 illustrates, the opinions of the actors
converge toward a common value (a shared norm) in
the long run. In fact, as long as the influence network is
strongly connected (i.e., there exists a directed path
connecting each pair in the influence network), the
classical influence model always generates convergence
to global uniformity. In other words, the group will
converge toward a uniform opinion unless the network
is broken into subsets that have no influence on one
another (Mason et al. 2007).

Theoretical Background

Important Scientific Research and
Open Questions

When two individuals influence each other, they typically become more similar. In social groups, this influence can lead to self-reinforcing dynamics that drive
the group toward homogeneity. To illustrate this, Fig. 1
shows a typical linear influence process in a group of
ten individuals, labeled 1–10. At the outset of the influence process, each individual holds a numerical opinion corresponding to its label. Furthermore, the group

10
9
8
Opinion

7
6
5
4
3
2
1
Time

Social Influence and the Emergence of Cultural Norms.
Fig. 1 Classical social influence model

The pervasive drive toward uniformity is puzzling
because it seems to contradict the high diversity of
opinions and behavior that we observe empirically,
even in small groups where the network is strongly
connected. In these cases, social separation cannot
explain the diversity of opinions, if influence operates
as described in the classical model.
A common extension of the basic influence model
is the assumption of homophily, represented by the
adage “birds of a feather flock together.” This principle
of social network formation – that individuals develop
and maintain social ties to others who are sufficiently
similar to themselves – is analogous to ▶ Hebbian
Learning in neural networks. In tandem with the classical influence model, homophily implies a positive
feedback loop: If relatively similar actors interact and
influence each other, they become even more similar,
which increases interaction and thus greater similarity.
However, a second set of actors that is sufficiently
dissimilar from the first set (but similar to each other)
may converge toward their own norm, while ignoring
or forgetting the first set. In this way, distinct clusters

3099

S

S

Social Influence and the Emergence of Cultural Norms

Opinion

Opinion

3100

Time

Social Influence and the Emergence of Cultural Norms.
Fig. 2 Social influence model with homophily

of similar actors can emerge and persist over time
(Hegselmann and Krause 2002; Axelrod 1997).
To illustrate how the interplay of selection and
influence triggers cluster formation, Fig. 2 shows a scenario with 100 individuals having random opinions at
the outset. Individuals have social ties to (and are
influenced by) others that are relatively similar but
ignore others that are very different. In contrast to the
classical social influence models (see Fig. 1), distinct
clusters form in these conditions. Dynamics settle
when the clusters are internally homogeneous.
Thus, the interplay of homophily and social influence offers an explanation for the emergence and persistence of subgroups with distinct opinions or
behaviors. However, this theory fails to explain polarization, the development of increasingly different subgroups. Trying to explain polarization, researchers have
gone beyond the homophily assumption and assumed
that actors actually differentiate themselves from others
who are very different (Macy et al. 2003). Figure 3
illustrates the implications of such negative influence
tendencies. In this model, individuals are positively
influenced by similar others (“friends”) as before,
but form negative ties to very dissimilar others (“enemies”), which convey negative influence. At the outset,
opinions are uniformly distributed. Individuals with
moderate opinions do not change their opinions much
at the beginning of the influence process because they
receive balancing forces of influence in both directions.
However, individuals who hold relatively extreme opinions are driven to differentiate themselves from enemies
at the opposite extreme. As more individuals adopt very

Time

Social Influence and the Emergence of Cultural Norms.
Fig. 3 Social influence model with differentiation

extreme opinions, the influence of the extremists on the
moderate individuals increases in strength. As
a consequence, moderate individuals gradually turn
more extreme. Opinion dynamics settle when all
individuals hold an opinion on one of the two poles of
the scale. Of course, uniformity remains an equilibrium
under this model, but this tendency to differentiate from
enemies creates a powerful force toward polarization
that in most conditions will prevent convergence toward
uniformity in heterogeneous groups.
Formal models of social influence in networks
are highly abstract, as they are based on strong simplifying assumptions about individual behavior. Some
researchers are validating these models by bringing
them into dialog with empirical data. For example,
social psychologists have studied social influence
through networks implemented in experimental laboratories, and have also explored the conditions under
which pressures to uniformity will be realized in small
groups. Field researchers study influence networks
through e-mail communication or contacts on social
networking Web sites.

Cross-References
▶ Agent-Based Modeling
▶ Cross-cultural Factors in Learning and Motivation
▶ Cultural Influences on Personalized e-Learning
Systems
▶ Learning and Evolution of Social Norms
▶ Multi-Agent Model-Based Reinforcement Learning
▶ Prosocial Learning in Adolescence: The Mediating
Role of Prosocial Values

Social Interaction Learning Styles

S

References

Definition

Axelrod, R. (1997). The dissemination of culture - A model with local
convergence and global polarization. Journal of Conflict Resolution, 41, 203–226.
Hegselmann, R., & Krause U. (2002). Opinion dynamics and
bounded confidence models, analysis, and simulation. Journal
of Artificial Societies and Social Simulation, 5.
Latane, B., Nowak, A., & Liu, J. (1994). Measurement of emergent
social phenomena: dynamism, polarization, and clustering as
order parameters of social systems. Behavioral Sciences, 39, 1–24.
Macy, M. W., Kitts, J. A., Flache, A., & Benard, S. (2003). Polarization
in dynamic networks: a hopfield model of emergent structure.
Dynamic social network modeling and analysis (pp. 162–173).
National Academies Press: Washington, DC.
Mason, W. A., Conrey, F. R., & Smith, E. R. (2007). Situating social
influence processes: dynamic, multidirectional flows of influence
within social networks. Personality and Social Psychology Review,
11, 279–300.

Social interaction learning styles are models that
learners strategize in their acquisition of knowledge
and information. Learning is the process by which
behavior is modified according to the experience
and exposure presented to learners in different settings. This process involves interaction between the
teacher and the learner in different settings, which
allows for information dissemination and, in turn,
how the information is received by the learner. Such
an activity is termed as social interaction which
comes with scales of varying relationship promoted
among individuals that are in the interaction. Simultaneously, learning styles include the personal quality that influences a learner’s ability to acquire
knowledge.
As early as the nineteenth century, the concept of
social interaction stems from social behavior which has
two components, stipulated by which are the actions
and the orientation the actions are being attached to by
the actor(s). Later toward the last century, the social
interaction scientist, Kurt Lewin proposed the concept
of group dynamics. This interaction sees the relationship between individuals as well as between individuals
and the group.
Learning styles are the learning preference a learner
adopts in ensuring the acquisition of knowledge and
information in a learning setting. They can also be
personal quality that influences a learner’s ability to
acquire information. This is conducted in the way a
learner interacts with his peers and his teachers and the
way he participates in a learning experience. Experts
have presented the learning styles in the way a learner
approach a learning environment. Teachers need to
understand the importance of learning styles as these
provide guidance for improved teaching and learning
process.
Further understanding of the learning styles
enables teacher to strategize teaching approach to
suit the differing manner of learning among the
learners, hence, reducing frustration in teaching.
When such an understanding is established, it will
then propose the utilization of social interaction
learning styles which provide educational institutions
with a better approach to handle differing learner’s
learning preferences and therefore, assist in the development of curriculum.

Social Intelligence
▶ Machiavellian Intelligence Hypothesis
▶ Social-Emotional Learning Scale

Social Interaction
▶ Argumentation and Learning
▶ Socio-emotional Aspects of Learning

Social Interaction Learning
Styles
1

2

KAMARUZAMAN JUSOFF , SITI AKMAR ABU SAMAH
1
Faculty of Forestry, Universiti Putra Malaysia,
Serdang, Selangor, Malaysia
2
University Publication Centre, Universiti Teknologi
MARA, Shah Alam, Malaysia

Synonyms
Communication; Group dynamics; Inflections; Interaction of individuals; Learning preferences; Personal
quality; Relationship among individuals; Ways to seek
knowledge

3101

S

3102

S

Social Interaction Learning Styles

Theoretical Background
Grasha and Riechmann Learning Style
Theory (1974)
This theory looks at the social and affective perspective
that deals with patterns of preferred styles that are
adopted by learners when interacting with teachers
and among the learners themselves. It revolves around
three dimensions, namely, learner’s attitudes toward
learning, learner’s views of teachers and their peers,
and learner’s reactions toward classroom procedures.
While this exploration is partaken by the learners,
the patterns of preferred styles for interaction with
teachers and peers are further classified as social interaction scales identified as Avoidant-Participant,
Competitive-Collaborative, and Dependent-Independent. However, for the discussion, the following presents the learner’s preference in learning, each of this
includes (Riechmann and Grasha 1974):
● Avoidant: Usually classified as low achiever, this

●

●

●

●
●

learner may have a different focus in his learning
as he takes less responsibility in his learning.
Participative: This is the type that enjoys pleasing the
teacher. He is interested in class activity and is keen to
complete his lesson within the time frame of learning.
Competitive: The learner tends to be wary of his
peers and that he will be in a mode to do better
than the others.
Collaborative: This learner enjoys group discussion
in which he is able to gather information through
sharing and cooperating with his teachers and team
members.
Independent: The learner prefers independence and
self-paced activity as he enjoys working alone.
Dependent: As the term suggests, this type of learner
prefers having the presence of expert as a source for
guidance in deciding what to do.

Curry’s Onion Model (1983)
This social interaction model is one of the learning
styles models which suggest for consideration to vary
the strategies adopted by the learner to suit his specific
environment and social context (Curry 1983).

Kolb’s inventory includes four dimensions namely
diverging, assimilating, converging and accommodating. Learner who has diverging learning style prefers to
feel and look at things. This type of learner prefers
gathering information and uses imagination to solve
problems. In doing so, they work well in groups, and
the divergent learner enjoys brainstorming learning
approach. The next in the inventory list is the assimilating learning preference sees the learner who enjoys
ideas and concepts rather than people. The learner is
more percept to ideas than practical application.
Hence, he learns well in lecture, exploring analytical
model and in reading in which there is time for him to
think. The next type of learner prefers using his learning to look for solution. In his way of learning, a learner
practices what he learns to work out ways to solve
problems and to make decision, and this is known as
the converging learning style. Finally, in Kolb’s inventory, the next type of learner prefer doing and feeling in
his learning or known also as accommodating. This
learner prefers experiential and practical ways to solve
problems. He is a team player and sees learning well
using other learner’s analysis and ideas (Kolb 1984).

Gardner’s Multiple Intelligences (1985)
Perhaps the most identifiable of learning preference is
that proposed by Howard Gardner. His concept of
multiple intelligences identifies learner as the following
(Gardner 1985):
● Visual: The learner has to have the presence of his

teacher as he takes to learning from the teacher’s
facial expression and body language. He learns best
visually, thus taking down notes and prefers to be in
front of the class.
● Auditor : As the term suggests, this learner listens
well and is attentive in lecture and discussion.
Although taking down notes is not a preference,
learning is best when there is reading aloud or
recording his voice.
● Kinesthetic: The learner prefers learning actively and
physically. Having short attention span, he has to be
given the liberty to “move around” and to be provided with hands-on activities.

Kolb’s Learning Styles Inventory (1984)

Gregorc’s Mind Styles Model (1985)

Kolb Learning Styles Inventory is probably the most
distinguishable among the rest of the proponents.

As in Kolb’s, Gregorc’s Mind Styles Model identifies
four types of learning preference. The first is known as

Social Interaction Learning Styles

Concrete Sequential which identifies learner who prefers instruction, following logical sequence and getting
facts. A practical person, this learner works well in
a structure environment and does not enjoy abstract
ideas nor working with ambiguity. In Gregorc’s inventory, a learner who enjoys personalized environment,
given instruction and does not like competition, is
known as Abstract Random. This learner prefers working in a friendly situation in which strict teacher and
negative criticism are not welcomed. The next is
Abstract Sequential which describes a learner as one
who prefers stimulating environment that allows for
appreciative listening and presence of reference to work
out solutions. This type of learner does not learn well in
a pressured environment of repetitive activity and
undiplomatic team members. The last type of learning
preference in this Model is known as Concrete Random
sees a learner who enjoys to experiment, to find
answers, and to solve problems independently. He has
competitive spirit in his learning and does not learn
well with restrictions and repetitive activity.

S

3103

● Analytic: This learner draws from his learning the

important things that can add to his world of
knowledge. In his reflective mode, he wants to
develop intellectually.
● Common sense: When a learner goes for answers in
an active and straightforward approach, he also
looks for practical and useful ways that can make
things happen.
● Dynamic: This is the kind of learner that tends to
use his “gut feelings” to look for possibilities in his
quest for learning. He is daring, loves adventure and
spontaneous.

Important Scientific Research and
Open Questions
Wratcher et al. (1997)
That learner takes learning differently is widely
accepted and teacher has to be aware of the varieties
he has in his lesson in order to strategize for effective
teaching. This awareness helps a teacher to better
understand the different needs of his learners.

Myers–Briggs Type Indicator (1985)
In Myers–Briggs Indicator, there are eight classifications that identify a learner’s style. The list includes
Introversion in which a learner’s interest is in the
inner world of concepts and ideas. The next is Extroversion in which preference in learning goes for actions,
objects, and team work. A learner who prefers immediate and practical information is in the classification
of Sensing. Next, in the preference for Intuition,
a learner tends to perceive meaning of experiences
and possibilities to establish his purpose. When one
prefers to make decisions rationally, the classification is
known as Thinking, while the one known as Feeling,
this type of learner will make decision based on his
subjective emotion. Judging sees the learner to tend to
take action in a planned and definite approach and
finally Perceiving identifies the learner as one who prefers to take action in a flexible but flighty manner.

McCarthy’s Learning Styles (1990)
There are four learning styles as identifies in
McCarthy’s and the following are (McCarthy 1990):
● Innovative: This type of learner tends to enjoy the

values he learns from his lesson via social interaction and enjoys cooperation.

Diaz et al. (1999)
The study focuses on the discovery of relationships
between learning styles and specific learner’s achievement outcomes. These outcomes encompass dropout
rate, completion rate, attitudes about learning and predictors of high risks.

Cross-References
▶ Adaptation to Learning Styles
▶ Adult Learning Styles
▶ Cross-Cultural Learning Styles
▶ Extraversion, Social Interaction, and Affect Repair
▶ Jungian Learning Styles
▶ Learning Style(s)
▶ Mimicry in Social Interaction: Its Effect on Learning
▶ Social Interactions and Learning
▶ Social Learning

References
Armstrong, T. (1993). Multiple intelligences in the classroom. Alexandria: Association for Supervision & Curriculum Development
(ASCD).
Curry, L. (1983). An organization of learning styles theory and constructs. ERIC documents.
Gardner, H. (1985). Frames of mind: The theory of multiple intelligences. New York: Basic Books.

S

3104

S

Social Interactions and Learning

Gardner, H. (1993). Multiple intelligences: The theory in practice.
New York: Harper Collin.
Grasha, A. F. (1984). Learning styles: The journey from Greenwich
observatory to the college classroom. Improving College and
University Teaching, 22, 46–52.
Grasha, A. F. (1996). Teaching with style: A guide to enhancing learning
by understanding teaching and learning styles. Pittsburg: Alliance.
Hamidah, J. S., Sarina, M. N., & Kamaruzaman, J. (2009). The social
interaction learning styles of science and social science students.
Asian Social Science, 5(7), 58–64.
Kolb, D. A. (1984). Experiential learning: Experience as the source of
learning and development. Englewood Cliffs: Prentice-Hall.
McCarthy, B. (1990). Using the 4NAT system to bring learning styles
to schools. Educational Leadership, 48(2), 31–36.
Miell, D. (1996). Social interaction and personal relationships. London:
The Open University.
Muzafer, S., & Brannigan, A. (2009). Social interaction. New Brunswick: Transaction.
Riechmann, S. W., & Grasha, A. F. (1974). A rational approach to
developing and assessing the construct validity of a student learning style scales instrument. Journal of Psychology, 87, 213–223.
Turner, J. H. (1988). A theory of social interaction. Stanford: Stanford
University Press.

Social Interactions and
Learning
SANDRA Y. OKITA
Dept. of Mathematics, Science and Technology,
Teachers College, Columbia University, New York,
NY, USA

Synonyms
Collaborative learning; Peer learning

Definition
Social interaction plays an important role in learning.
Interacting with other people has proven to be quite
effective in assisting the learner to organize their
thoughts, reflect on their understanding, and find
gaps in their reasoning. Underneath the broad
umbrella of social interactions and learning, variants
can range from peer learning, reciprocal teaching,
learning by teaching, learning by observation, learning
by doing, and self–other monitoring. These areas overlap in scholarship and are often an optimal way to
help students learn. Different forms of collaborative

learning can create ideal circumstances when examining the impact of social interactions on learning.

Theoretical Background
Vygotsky believed that culture, history, and social interactions play a critical role in the cognitive development
of children. Through observation, Vygotsky found that
children develop higher mental functions such as identifying speech patterns, learning a language, and deriving meaning from symbols, when interacting with
parents and other adults within the community.
Vygotsky referred to language, numbers, signs, and
symbols as cultural tools that help integrate the child
into the culture. Vygotsky believed that the internalization of these cultural tools led to higher thinking
skills. Children first learn how to use these cultural
tools through the social interactions with parents,
teachers, or more experienced peers, and later internalize the skills so they can perform independently. This is
different from Jean Piaget’s understanding of child
development where development precedes learning.
Vygotsky’s Zone of Proximal Development (ZPD)
is a theory about the dynamic relationship between
learning and development. ZPD is the area between
the learner’s independent performance level and the
level that can be achieved with assistance of a more
knowledgeable peer. ZPD not only reveals the learner’s
potential but also shows that with assistance, a higher
performance level can be achieved.
Social interaction is also a critical component for
other theories. Vygotsky’s theories were further elaborated upon by other researchers and implemented into
practical applications. Some examples are Situated
Learning, when learning occurs in the same context in
which it is applied. Learning is a social process that is
co-constructed through the involvement in “community of practice” where members of the community
share information and learn from one another (Lave
and Wenger 1990). The novice learner embodies beliefs
and behaviors through social interactions with more
experienced members of the community. With time,
the learner moves from the periphery of the community to the center, becoming more engaged and active
within the culture, and eventually takes the role of the
expert or senior member. Another example, Cognitive
Apprenticeship (Collins et al. 1989), further develops
the theory of knowledge construction through social
interactions like coaching, scaffolding, modeling, and

Social Interactions and Learning

reflection. Reciprocal teaching (Palincsar and Brown
1984) is when the teacher or peer provides the learner
with guided practice using four strategies of summarizing, question generating, clarifying, and predicting,
when reading a piece of common text. The learner and
teacher (or peer) take turns playing the lead role as
a teacher, and use the four strategies to support their
discussion on segments of the text. Over time, children
begin to internalize the processes until the strategies
become a natural part of their internal reading and
listening skills. The strategies help the learner and
teacher (or peer) develop deeper understanding of the
text and better reading comprehension skills.
These theories have also been applied in the context
of technology-based learning activities. Peer learning
and collaborative learning was once only possible in
shared physical space, but now learners can participate
remotely via the Internet and technology-mediated
tools.

Important Scientific Research and
Open Questions
People learn from various sources. Traditional sources
involve learning from humans or objects (e.g., books),
while recent sources may involve computerized people
(e.g., pedagogical agents and avatars) and/or computerized instructions (e.g., intelligent tutoring systems).
Social interactions also occur in various settings. Traditional settings involve face-to-face interactions in
both formal and informal environments (e.g., classroom and private tutor), while recent settings can
involve online learning environments (e.g., video conferencing systems like Adobe Connect and virtual reality environments like Second Life). Under this broad
umbrella, the following may be considered: (1) learning
in social interactions with others, (2) learning in social
interactions with others through ▶ computermediated communication (CMC), and (3) learning in
social interactions with technology.
Learning in social interactions with others: People
often turn to others for learning. Social interaction
plays an important role in learning, and has proven to
be quite effective in peer learning, reciprocal teaching,
and behavior modeling. Such forms of collaborative
learning are often an optimal way to help people learn
(Chi et al. 2001). For example, Learning by teaching
and explaining to others can be an effective way to learn
(Palincsar and Brown 1984). Another situation may be

S

learning by observing other people. In tutoring, one
observes whether their pupil applies what they were
taught during problem solving. Their pupils’ performance can reveal gaps in what the tutor taught and
perhaps understands. The performance of the pupil can
provide alternatives the tutor did not think of. Even if
these alternatives are not correct, they may slow down
the tutor’s natural inertia to keep thinking in the same
way. Studies have shown that learning among peers can
be very useful in several ways. Learning can occur by
comparing ourselves to peers, or observing others to
develop a better understanding of the self. For example,
even if a student cannot solve a math problem, observing someone else may help you learn how to solve the
problem. This is because the person they are observing
can provide a model of competent performance. In
other situations, interacting with somebody who
knows about the same as (or knows less than you) can
be beneficial. For example, in reciprocal teaching, students may spontaneously compare their understanding
to what they observe in another person, and any discrepancies can alert them to think more deeply about
who is right. This implies that observing a peer, under
the right circumstances, can trigger learning and reflection. In other cases, just anticipating a social interaction can lead to more learning. For example, preparing
to teach others influences students to learn more compared to students who study for themselves (e.g., study
for exam). In this case, learning occurs just with the
“thought” of a social interaction.
Learning in social interactions with others through
computer-mediated communication tools: There is no
need to be physically present to learn in person.
Through the use of the Internet technology and computer-mediated communication tools, real-time social
interactions are possible. Many synchronous online
learning (or distance learning) environments use
video conferencing tools that allow face-to-face interaction via technology mediation (e.g., Adobe Connect). More recent forms of online learning may
involve virtual reality (e.g., Second Life) where your
peers are represented by a computer graphic character
that they remotely control in a virtual reality environment (e.g., an avatar). Such technological tools allow
real-time exchange of audio, video, text, and graphical
information between learners (Dede et al. 2002). Successful virtual reality environments such as Second Life
and Active Worlds provide space to support online

3105

S

3106

S

Social Interactions and Learning

group activities. There are some concerns that social
interactions are limited in online learning, compared to
the traditional face-to-face learning experience. Others
attest that technology-mediated tools can elicit social
responses and create unique social interactions with
interesting implications for learning. For example, children can build their own simulated world (e.g., Ecosystem) rather than passively partake in a given situation. This may allow children to directly experience the
causal chains from their actions and help visualize and
reason about the situation. Another distinct feature in
virtual reality is that the learner’s environment can be
manipulated based on their needs. For example, the
teacher can be represented differently to communicate
with the learner in the most optimal way (e.g., with or
without eye contact), allow the learner to experience
different points of view (e.g., first person, third person,
and birds-eye view), and the seating in virtual classrooms can even be positioned based on the learner’s
attention level.
Learning in social interaction with technological
tools: Applying educational content and pedagogy to
technology is not new. The first testing and teaching
machine by Pressey appeared in 1926, and since then
people have had high hopes for technology in restoring
personalized instruction. Technology has the potential
to provide a wide range of tools tailored to each student’s learning needs. However, much of the learning in
the initial stage focused on machine learning, intelligent
expert systems, and computer modeling of human
behavior. Expert systems were successful in their
intended domain, but often evaluated unfairly, because
of the high expectation of the Turing Test. Some have
argued that by making machines smarter, good teaching and tutoring strategies can be implemented for the
learner. However, interactions with intelligent
machines do not always guarantee learning. Learning
can be difficult without a meaningful interaction
between the human and machine. Recently, development has shifted the focus from intelligent to directable
technologies in assisting human learning. Computerized people and instructions still consist of intelligent
behaviors, but more emphasis is placed on human-like
features for eliciting social responses. Technological
tools such as pedagogical agents, tutoring agents, and
humanoid robots, consist of strong social components
that enable students to share knowledge and build
peer-like relations. However, not all technologies put

emphasis on direct social exchange with humans (e.g.,
industrial robots). Most fall somewhere in between,
and partake of both machine-like and human-like features (e.g., pedagogical agents and humanoid robots).
Some systems may tacitly draw on social schemas,
but not include a real social presence or metaphor.
For example an intelligent tutor that is a computational model may represent student thinking and cognition, but its appearance may be a disembodied text
with no visual character. Other systems build on
explicit social metaphors of interaction and appearances to invite social interaction. An example may
be a socially explicit pedagogical agent taking on the
role of a peer learner. Students learn by teaching this
pedagogical agent. Based on what the agent is taught,
the agent can answer questions. Students can observe
their agent’s answers and revise the agent’s understanding (and their own). The learner can structure their
thoughts through the social interactions with the
agent, and even develop metacognitive skills (Biswas
et al. 2001).
Aside from content, advancement in sensors and
audio-visual tools has helped detect human behavior
(e.g., physiological sensors). Automation and expressive tools have helped technological tools respond to
humans. Sensors and behavior models implemented
into the system have improved some aspects in the
quality of social interactions between human and
machines. However, technological tools still fall short
when coming across unfamiliar content, and do not
easily afford the wide range of possible social interactions. Unlike a human peer or teacher, technology presents limitations, where the learner may often times be
constrained by what the tool (e.g., pedagogical agent)
or environment (e.g., Second Life, avatars) can do in
response. Until technological tools have both the intelligence and flexibility to respond to the learner’s interactive bids, examining the social exchange and
interactive styles that guide learning is crucial.

Cross-References
▶ Cognitive Apprenticeship Learning
▶ Learning by Teaching
▶ Reciprocal Learning
▶ Situated Learning
▶ Observational Learning
▶ Online Learning
▶ Pedagogical Agents

Social Learning

▶ Peer Learning and Assessment
▶ Vygotsky’s Philosophy of Learning
▶ Zone of Proximal Development

References
Biswas, G., Schwartz, D. L., & Bransford, J. D. (2001). Technology
support for complex problem solving: From SAD environments
to AI. In K. Forbus & P. Feltovich (Eds.), Smart machines in
education (pp. 71–98). Menlo Park, CA: AAAI/MIT Press.
Chi, M. T. H., Silver, S. A., Jeong, H., Yamauchi, T., & Hausmann,
R. G. (2001). Learning from human tutoring. Cognitive Science,
25, 471–533.
Collins, A., Brown, J. S., & Newman, S. E. (1989). Cognitive apprenticeship: Teaching the craft of reading, writing and mathematics.
In L. B. Resnick (Ed.), Knowing, learning, and instruction: Essays
in honor of Robert Glaser (pp. 453–494). Hillsdale, NJ: Lawrence
Erlbaum.
Dede, C., Whitehouse, P., & Brown-L’Bahy, T. (2002). Designing and
studying learning experiences that use multiple interactive media
to bridge distance and time. In C. Vrasid & G. Glass (Eds.), Current
perspectives on applied information technologies (Distance Education, Vol. 1, pp. 1–30). Greenwich, CT: Information Age Press.
Lave, J., & Wenger, E. (1990). Situated learning: Legitimate peripheral
participation. Cambridge, UK: Cambridge University Press.
Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of
comprehension-fostering and comprehension monitoring activities. Cognition and Instruction, 1, 117–175.

Social Interest Behaviors
▶ Altruism and Health

Social Learning
EYLEM SIMSEK
Department of Communication, Anadolu University,
Eskisehir, Turkey

Synonyms
Imitation; Modeling; Observational learning; Social
cognitive theory; Vicarious learning

Definition
Social learning is the synthesis of cognitive psychology
and behaviorism, emphasizing the role of personal

S

3107

cognitive capability and the environment. On the
one hand, neither classical nor operant conditioning
of behaviorism explains the unreinforced behaviors.
On the other hand, cognitive theories mostly ignore
the role of environment. As a theory of interwoven,
social learning assumes that one can learn by observing
the behaviors of the other under certain environmental conditions; this process is often called vicarious
learning.
According to the behavioral psychology, the imitated behaviors are those reinforced. This approach
implies operant conditioning of behaviorism. Skinner
focused on the extrinsic reinforcement for learning.
In contrast, Bandura (2001a) underlines in his social
cognitive theory psychological mechanisms of the self
system and socio-structural factors which together
constitute behaviors. Learning is not as simple as stimuli, extrinsic reinforcement, conditioned stimulus, and
imitation; one can learn by observing and modeling so
that social learning depends largely on reasoning the
consequences of the actions.

Theoretical Background
The term of social learning is often attributed to Albert
Bandura, who is one of the important pioneers in the
field of social learning. However John W. Watson, Neal.
E. Miller, John Dollard, and B. F. Skinner are former
contributors in this area.
The first social learning theorist is John B. Watson
(1878–1958), who is also the founder of behaviorism.
In opposition to the previous “tabula rasa” view, he
argues that the behaviors are learned and this development process is continuous without clear-cut stages.
He discusses that behaviors usually precede the learning
of rules and laws, except some human reflexes. Under
this conception, three main social learning approaches
have been developed: neo-Hullian Approach, Skinner’s
operant-learning model, and Bandura’s cognitive social
learning model (Shaffer 1988).
Neo-Hullian Theory explains social learning with
habits. This theory departed from the Freudian view
claiming that human behavior depends not only on
instincts but also on the primary and secondary drives.
Primary drives are inherent, whereas secondary drives
are learned by experience. Behaviors could be changed
by stimulus and response interaction. However, reinforcers are those responses who reduce drives. Primary
reinforcers, often acquired inherently, are basic needs

S

3108

S

Social Learning

for survival such as water and food. On the other hand,
secondary reinforcers are associated with higher needs
such as love, status, etc. After the association of the
neutral stimulus and consequence of behavior, secondary reinforcers are formed. From the social learning
perspective, Miller and Dollard focused on reinforcement for acquiringstable social behavioral patterns,
called habits which are accepted as a part of personality.
They claimed that secondary reinforcers are more
influent on the development of habits during the
daily life (Shaffer 1988).
Operant-learning theory contributes to social
learning by underlining the importance of environment on behaviors. Skinner, one of the most prominent behavioral psychologists, was interested in
learning within the scheme of operant conditioning
and volunteered behavior instead of reflexes. Of
course, reinforcement plays a crucial role in this process. Reinforcer is defined as an external or internal
stimulus that strengthens the behavior. Secondary
reinforcers are used much more in this context than
primary ones, similar to neo-Hullian perspective.
Operant conditioning gives more attention to external
stimuli in order to increase the frequency or probability of a response. Positive and negative reinforcers
along with punishment are instruments to shape
and/or change certain behaviors. Antecedent and
consequent events act as reinforcer and shape the
behavior. Shaping is the key factor explaining social
learning patterns in terms of behaviorism (Glover and
Bruning 1988).
Finally, cognitive social learning theory combines
the neo-Hullian and the behaviorist approach, adding
cognitive perspective to the equation. Outcome expectations as acting reinforcers mostly cause the behavior
in social learning. Operant conditioning depends on
behaviors that are performed. This is the main difference between cognitive social learning approach and
the previous approaches. In many cases of social learning, one could learn by observing without any stimulus,
reinforcement, response, or drives.
Cognitive social learning theory was adopted from
social cognitive theory. Bandura (2001b) indicates that
human behavior is explained generally by unidirectional view such as by environmental influences or by
internal dispositions. Social cognitive theory, however,
follows triadic reciprocal causation for human functioning. In this transactional point of view, personal,

behavioral, and environmental determinants operate
bidirectionally (Fig. 1).
Reciprocal determinism is one of the key terms of
social learning. In contrast to environmental determinism, behavior is the result of interaction results between
prior learning and present environmental conditions.
According to social cognitive approach, society and
personal interactions have influences on behavior as
well as social learning. Bandura assigns people as an
active role rather than passive recipients throughout
self-reinforcement mechanisms of learning. People are
self-organizing, proactive, self-reflecting, and selfregulating organisms rather than being shaped only
by intrinsic and extrinsic forces. Symbolizing capability, self-regulatory capability, self-reflective capability,
and vicarious capability of people together enable
social learning and modeling via complex cognitive
mechanisms.
Bandura’s modeling process is operated by four subprocesses. attention, retention, reproduction, and motivation (reinforcement). As well known, one could not
be able to learn without paying attention. Attention
process is related to the interest, values, attitudes, or
personality. People pay more attention to the models
who are similar to themselves. People having high status,
competence, and expertise are more likely to be imitated. Moreover, incentives and distinctive cues increase
attention. Retention is required for future reproduction
of the behavior; symbolic coding- (imaginative or verbal) is needed for memorizing the models behavior.
Motoric representation refers to self-corrective adjustments of the modeled behavior. Finally reinforcement
process enlightens why all observed and learned behaviors are not modeled (Gage and Berliner 1987).
Opposite to behaviorism, performance and learning are different phenomena in social learning. People

Personal
determinants

Behavioral
determinants

Environmental
determinants

Social Learning. Fig. 1 Schematization of triadic
reciprocal causation in the causal model of social cognitive
theory (Bandura 2001b)

Social Learning in Animals

do not perform everything they have learned. The
distinction between observational learning and performance-modeling could be clarified by anticipatory
mechanisms, which generate positive expectations or
incentives. Learning does not guarantee successful
behavior or performance. Expected consequences as
well as actual ones determine the responses to the
observation or modeling (Glover and Bruning 1988).

Important Scientific Research and
Open Questions
Social learning theory is applied to understand aggression with well known “Bobo doll experiment.” Bandura
has conducted the series of experiments between 1961
and 1965. The main research question was whether
behaviors like aggression are learned by observing or
imitating in light of the social learning theory. The
children were exposed to aggressive and non-aggressive
behaviors to a bobo doll by adult models. Physical and
verbal aggression, the frequency of aggression besides
hitting the Bobo doll, and non-imitative forms of
aggression have been evaluated. As expected, children
exposed to aggressive models showed more aggressive
behavior than children exposed to non-aggressive
models. This experiment proves observational or social
learning theory even without any reinforcement on
both models and children (Bandura et al. 1961).
Bandura’s other experiment in 1963 was designed
to show aggressive behavior via video playback. Compared to the real model, video playback is less effective
for modeling. The shortcoming of the experiment was
that children could hit the Bobo doll only for fun. To
sum up, the Bobo doll experiments proved that without any reinforcement one can learn by observing and
modeling. The results have implementation for the
treatment of phobias, explain the routes of criminal
behaviors and measurement of negative television
effects (Bandura 2001a, b).
There are some criticisms to the social learning
approach. First, social learning is criticized for
oversimplifying genetic characteristics of people
which directly determine the learning capacity. Second,
social learning approach is thought as a nondevelopmental theory. Development theories explain
specific age-related changes, whereas social learning is
a process approach. In other words, the social learning
theory could not identify the age and stage development (Shaffer 1988).

S

3109

Cross-References
▶ Argumentation and Learning
▶ Emotional Learning
▶ Field Research on Learning
▶ Imitation: Definitions, Evidence, and Mechanisms
▶ Imitative Learning in Humans and Animals
▶ Learning Through Social Media
▶ Learning with and from Blogs
▶ Observational Learning of Complex Action (Dance)
▶ Observational Learning: The Sound of Silence
▶ Social Learning in Animals
▶ Social Cognitive Learning
▶ Social Learning of Fear
▶ Theory of Conformist Social Learning
▶ Visual Communication and Learning

References
Bandura, A. (2001a). Social cognitive theory: An agentic perspective.
Annual Review of Psychology, 52, 1–26.
Bandura, A. (2001b). Social cognitive theory of mass communication. Media Psychology, 3, 265–299.
Bandura, A., Ross, D., & Ross, S. A. (1961). Transmission of aggressions through imitation of aggressive models. Journal of Abnormal and Social Psychology, 63, 575–582.
Gage, N. L., & Berliner, D. C. (1987). Educational psychology (3rd ed.).
Boston: Houghton Mifflin.
Shaffer, D. R. (1988). Social and personality development (2nd ed.).
Pasific Grave: Brooks/Cole.
Further Reading
Glover, J. A., & Bruning, R. H. (1987). Educational psychology: Principles and applications (2nd ed.). Boston: Little, Brown.

Social Learning and
Communication
▶ Song Learning and Sleep

Social Learning in Animals
LUDWIG HUBER
Department of Cognitive Biology, University of
Vienna, Vienna, Austria

Synonyms
Observational learning; Socially biased learning

S

3110

S

Social Learning in Animals

Definition
Many species live in social groups, which offers the
possibility to learn from the behavior of others. Examples of social influences on the adaptive modification of
behavior are widespread and diverse, ranging from
food selection and predator avoidance to learning of
songs, routes, and motor skills. It can also lead to
synchrony in the performance of established behavior
and the transmission of new behavior patterns throughout a group. These affects are mostly (though not necessarily and not always) beneficial to the observer, either
immediately or long term, and are therefore assumed to
outweigh potential costs in terms of vigilance, attention,
memory, or special learning mechanisms.
The term social learning usually refers to learning
that is influenced by observation of, or interaction
with, another animal or its products. Many people
use “individual learning” as a complementary term, to
refer to learning that does not involve social interaction
or any information that is provided by others. It would,
however, be less confusing to use the term “asocial
learning,” because in all cases it is ultimately individuals
who learn.
Social learning may be direct or indirect. Direct
social learning refers to learning by observing others,
whereas indirect social learning describes processes in
which social partners aid the cognitive development of
younger animals through the alteration of the physical
environment. When formal research on the subject
began a century ago, the focus was to identify evidence
of social learning in the animal kingdom. It focused on
the demonstration of this capacity in a given species
and the role of social learning in the ontogeny of
adaptive behavior. Much later, psychologists, fueled
by the investigation of imitation in human children,
became interested in the underlying mechanisms,
investigating whether different mechanisms are
responsible for social and asocial learning, and whether
nonhuman animals are capable of the most cognitively
advanced forms of social learning shown by humans. In
many cases there is no need to invoke a unique learning
mechanism to explain the acquisition process. Indeed,
there is great heterogeneity in mechanisms, causation,
and possible functions of social learning.

Theoretical Background
What is the impact of social learning on the development of adaptive behavioral repertoires of naturally

living animals? It has been suggested that social learning may fill an important niche between species-typical, genetically predisposed behavior and individual
learning. Learning by observing others may provide
more flexibility than is possible with species-typical
behavior, but it may also shortcut the many iterations
of trial and error necessary for most individual learning, and move directly to solutions previously found by
others.
Broadly speaking, learning can be viewed as an
adaptive modification of those neuronal circuits that
give rise to behavior. It occurs during the whole lifetime
of an individual, but is particularly important during
early development. It allows youngsters to fine-tune
their behavior to the rapidly changing properties of
their local circumstances. Natural selection, acting on
genetic variation in a population, is too slow to furnish
individual organisms with those adaptations.
Social learning can have effects on the small and the
large scale. On the one hand, it can aid the development
and fine-tuning of important behaviors in the individual. Here the information transmitted between demonstrator and observer is usually of only transient
value. On the other hand, social learning can contribute to group- or population-specific behavior. This
kind of information transmission would lead to behavioral homogeneity in the group that extends beyond
the period of interaction. Generated in this way and
maintained over several generations these processes can
ultimately result in the emergence of traditions and
cultures.
A further important question concerns the direction of transmitted information. On the one hand,
information can spread through animal populations
horizontally, sometimes even between unrelated individuals, as in bird flocks or fish schools. On the other
hand, it can be transmitted vertically, from generation
to generation, mostly from parents to offspring. Most
social learning seems to fall in the highly horizontal
category. The transmission of food preferences in rats is
one of the best-studied examples of this; the forming of
song dialects in song birds or cetaceans is evidence of
the vertical category.
Mathematical models have been developed to study
the costs and benefits of social learning, the effects of
social dynamics, the factors that influence the spread
of information in populations, the adaptive advantages
of social information, and the evolutionary

Social Learning in Animals

consequences of a capacity for social learning. Furthermore, the debate of animal culture has contributed to
the development of “dual-inheritance” models to
investigate the complex interaction between culturally
and genetically transmitted information.

Important Scientific Research and
Open Questions
Learning about novel food sources is a critical task for
most animals. To avoid potential poisoning most species are neophobic toward unknown food items, either
avoiding them completely or eating only small quantities initially. During infancy, when virtually all food
items are novel, it would be safer to exploit the experience of other conspecifics and to learn about the palatability of unknown items socially instead of through
trial and error. Social mediation of food preferences has
been found in birds (e.g., pigeons and blackbirds) and
mammals (e.g., mice, dogs, and bats). The best-studied
example is the Norway rat. It not only learns from elder
conspecifics which foods to eat, but is reluctant to
ingest foods which older members of its colony have
not introduced it to. This social influence can last for
a month and can be so strong that it can even reverse
conditioned food aversions.
Social learning about food is often actively promoted. Food sharing is not only a means of providing
infants with additional nutrition or with items that
infants cannot obtain by themselves, but also facilitates
the transmission of food preferences from adults to
young animals. In common marmosets, for instance,
food transfers preferably include food that is novel to
the infants, thereby fostering learning about it. The
most advanced form is tutoring, where the adults track
what is novel to the infants. However, evidence for this
active demonstrator role is poor in nonhuman animals.
Extractive foraging may be the most challenging
situation to foster social learning. However, the extent
to which social learning helps acquisition of the motor
patterns that an animal needs to process encapsulated,
or otherwise secured, food remains a contentious issue.
Roof rats of the pine forests of Israel only learn to
efficiently strip scales from a pinecone (to reveal the
pine seeds) if it is demonstrated by skilled conspecifics.
Social learning may also be involved in the development of novel feeding techniques in birds, as, for example, in the spread of puncturing the top of milk bottles
by great tits.

S

Concerning the most advanced foraging technique,
tool use, the evidence for a major role of social learning
is even more ambiguous. Among tool-using birds, two
species have been intensively investigated, and the evidence is far from conclusive. In Galápagos woodpecker
finches, the use of modified twigs or cactus spines to
pry arthropods out of tree holes seems to depend on
very specific learning that involves trial-and-error during a sensitive phase in ontogeny. Social learning is not
essential for development of this behavior. In Caledonian crows, like in chimpanzees, the use and manufacture of probing tools may depend on prey distribution
as well as on genetic differences between populations,
though social transmission cannot be ruled out. From
a comparative point of view, the question of social
learning of tool use in animals is further complicated
by the fact that several species of ants use, and even
manufacture, foraging tools.
Although the social learning of foraging skills is
perhaps the most obvious or at least most studied
function of social learning, researchers have found
social learning effects in many other domains. These
include learning how to get to a resource (e.g., in
wrasses, grunts, and guppies), how to hunt (e.g., in
archer fish, hyenas, killer whales, and chimpanzees),
what to fear (e.g., in rhesus monkeys, blackbirds,
and minnows), about mate-choice (in guppies and
quails), how to court (e.g., in cowbirds), learning of
nonfunctional behavior (such a stone clapping in Japanese macaques or fire making in orangutans), greeting
gestures (gray parrot), and finally song learning in
songbirds and cetaceans. These examples show social
learning can occur via many different modalities,
including the use of auditory and olfactory cues.
Learning from others is not always adaptive. Animals may ignore social information under specific circumstances or switch conditionally between reliance
on social and asocial information according to its reliability. Observers are selective in many respects. They
do not blindly copy what they see, but take into account
the what (goals, actions, results), when (when uncertain etc.) and from whom (the majority, or the age, sex,
dominance status, affiliation, and knowledge of the
model) to learn. This selective nature of social learning
is confirmed by theoretical and ecological models of the
adaptive advantages of social learning.
Social learning that results in long-term effects on
the behavioral repertoire or the knowledge of the

3111

S

3112

S

Social Learning in Animals

individual should be distinguished from (transient)
social influences on underlying motivational and perceptual processes of the observer. The former influences
are generally called social facilitation, the increased
probability of performing a class of behaviors (e.g.,
feeding) in the (mere) presence of a conspecific (usually mediated by an increase in arousal or fear reduction). A similar phenomenon is contagion, referring to
the unconditioned release of an instinctive behavior or
a specific response (response facilitation) in one animal
by the performance of the same behavior in another
animal. Possible cases of contagion include synchronized predator evasion in flocks and herds of animals,
synchronized movements, chorusing by birds, synchronous courtship behavior, contagious laughing, and
contagious yawning. However, behavioral synchrony
is itself not indicative of any influence through observation of a conspecific.
The classical phenomena related to social influences
on perceptual processes are local and stimulus enhancement. They refer to the animal’s shift of attention to a
particular location or a particular object as a consequence of observing a conspecific engaged in rewarding
activity at this location/with that object. This observation may result in an increase in the observer’s interaction with that stimulus and subsequent learning about
the distinctive value of that location or object.
Here we have some overlap between social influences on learning and social learning itself. While the
(short-term) change of (directed) attention would not
itself imply learning, stimulus enhancement could be
regarded as a subset of single-stimulus learning, where
exposure to a stimulus results in a change in the subject’s responsiveness to that stimulus.
A clear-cut social learning phenomenon is observational conditioning. Since the observer does not actually
experience the unconditioned stimulus, but only
observes it, a form of higher order conditioning is
involved. Examples of observational conditioning are
the learned snake fear in Rhesus macaques and social
transmission of mobbing in blackbirds.
In the cases of social learning described so far, the
response or action of the observers is relatively similar
to that shown by the demonstrator. But animals may
learn by about affordances, functional properties or
operating mechanisms of objects by observing others
and subsequently use their own, preferred manipulative movements to achieve similar effects as seen.

This is called emulation and is regarded as intelligent
problem-solving behavior because it signifies the
observer’s ability to understand a change of state in
the world produced by manipulations of another.
Finally, imitation is social learning about (some
aspects of) the behavior of others, rather than its consequences. It is considered an important and intriguing
neuro-cognitive process in humans, which promotes
empathy and cooperation. Further, it provides a channel of evolutionary, cultural inheritance that makes us
distinctively human. However, there is now ample evidence that animals can learn imitatively, though possibly not with the same complexity as humans.

Future Research
As social learning is considered a major component
(prerequisite) of cultural processes, it has become
a multidisciplinary research focus. From a biological
point of view, a complete understanding requires the
integration of answers to four different questions
(Tinbergen’s four whys); in addition to the adaptive
function (ultimately survival value) and the physiological (neuronal) mechanisms (proximate causation)
described here, one needs to consider how an ability
arises during an individual’s development (ontogeny)
and, finally, how a particular behavior evolved, how it
came into existence and changed during the history of
the species (phyolgeny). An answer to the latter has
been complicated by the fact that solitary species, octopuses and tortoises, can learn by the observation of
others. This challenges all current theories about the
evolution of social learning that are based on the
assumption that social learning is an adaptive specialization for social living.

Cross-References
▶ Adaptation and Learning
▶ Animal Culture
▶ Animal Intelligence
▶ Associative Learning
▶ Collaborative Learning
▶ Collective Learning
▶ Cultural Learning
▶ Imitation Learning from Demonstration
▶ Imitation: Definitions, Evidence, and Mechanisms
▶ Imitational Learning (of Robots)
▶ Imitative Learning in Humans and Animals

Social Learning of Fear

▶ Learning (and Evolution) of Social Norms
▶ Learning in the Social Context
▶ Prosocial Learning
▶ Small Group Learning
▶ Social Cognition in Animals
▶ Social Construction of Learning
▶ Social Interaction Dynamics in Supporting Learning
▶ Social Learning
▶ Social Learning of Fear
▶ Social-Cognitive Influences on Learning
▶ Social-Emotional Learning Scale
▶ Socialization-Related Learning
▶ Societal Influences on Learning
▶ Theory of Conformist Social Learning

Further Reading
Galef, B. G. J., & Heyes, C. M. (2004). Social learning and imitation.
Learning and Behavior, 32(1) (special issue).
Galef, B. G. J., & Laland, K. N. (2005). Social learning in animals:
Empirical studies and theoretical models. BioScience, 55(6),
489–499.
Heyes, C. M. (1994). Social learning in animals: Categories and
mechanisms. Biological Reviews, 69, 207–231.
Heyes, C. M., & Galef, B. G. J. (Eds.). (1996). Social learning in
animals: The roots of culture. San Diego: Academic.
Hoppitt, W., & Laland, K. N. (2008). Social processes influencing
learning in animals: A review of the evidence. Advances in the
Study of Behavior, 38, 105–165.
Zentall, T. R., & Galef, B. G., Jr. (Eds.). (1988). Social learning:
Psychological and biological perspectives. Hillsdale: Lawrence
Erlbaum.

Social Learning of Fear
ANDREAS OLSSON
Division of Psychology, Department of Clinical
Neuroscience, Karolinska Institutet, Stockholm,
Sweden

S

observing, and listening to others. Because of its reliance on socially transmitted information, SLoF is an
indirect (or vicarious) learning procedure in contrast to
direct learning procedures, such as classical (Pavlovian)
fear conditioning, in which the animal learns to fear
a conditioned stimulus (CS) or context through direct,
individual experiences of its pairing with an aversive
unconditioned stimulus (US). In comparison to direct
learning procedures, SLoF provides a less risky and
faster way to acquire adaptive avoidance behavior of
potentially dangerous events in the surrounding. SLoF
is commonly measured through the learner’s expressed
fear of a CS in the absence of the individual(s) who
transmitted the fear. This feature differentiates social
fear learning from related social behaviors, among
them imitation and socially facilitated and contagious
fear responses, such as those seen in flocks behaving in
unison, schools, and herds of animals. Socially acquired
fear – like conditioned fear – can be expressed as
changes in neural, endocrine, and behavioral systems.
Social learning of fear can be further differentiated by
reference to the specific means of fear transmission.
Two kinds of SLoF have been particularly investigated:
(a) observational (sometimes referred to as “vicarious
fear conditioning”) and (b) instructed (verbal) fear
learning:
(a) In an observational fear learning procedure, the
learner (observer) attains fear-relevant information by observing a learning model’s (demonstrator’s) expressions of fear, pain, or other distress.
This form of SLoF is the most widely studied and
has been documented across several social species,
such as birds, rodents, and primates.
(b) In an instructional fear learning procedure, which
due to its reliance of symbolic language is uniquely
human, the learner receives fear-relevant information through verbal instructions. In humans,
observational and instructed fear learning often
co-occur.

Theoretical Background
Synonyms
Social emotional learning; Vicarious fear conditioning

Definition
Social learning of fear (SLoF) refers to the acquisition
of fear through social means, such as interacting,

3113

The powerful effect of socially mediated fear has been
discussed by philosophers and capitalized on by the
arts since the origin of the Greek tragedies. In modern
time, British Enlightenment philosophers, notably
David Hume (1711–1776) and Adam Smith (1723–
1790), argued that vicarious emotions (“passions”)

S

3114

S

Social Learning of Fear

were critical to the individual’s social functioning and
to moral development and thus to a well-functioning
society. Hume’s prophetic speculations about the
underlying mechanisms of vicarious fear and anxiety
included both accounts of emotional “mirroring” and
of “associations” between the present situation and
earlier encounters of similar events.
The background of empirical research in SFoL can
be traced to three different, but partially overlapping,
strains of research: (1) experimental research in
nonhuman animals, (2) clinical observations in
humans, and (3) experimental research on normal
healthy human children and adults.
1. Since the first half of the 1900s, fear learning
through social observation has been identified
in both humans and several other social species,
such as birds, cats, rodents, and rhesus monkeys.
It has been shown that the learning model’s expressions of distress (facial fear expression in primates)
can serve as a powerful US. In a seminal series
of experiments, Mineka and others (e.g., Mineka
and Cook 1993) showed that cage-reared rhesus
monkeys quickly acquired a long-lasting fear
response toward snakes after as little as one episode
of observing a conspecific’s fear response toward
the snake. The relationships between the strength
of a learning model’s expressed distress, the
observer’s immediate response to the model’s distress, and the resulting fear learning in the observer
at a later point were comparable to the relationship
reported between US, UR, and CR in classical
fear conditioning. These and subsequent studies,
predominantly in rodents and humans, support
the view that observational fear rely on learning
processes similar to those underlying classical
conditioning.
2. Based on clinical observation, Rachman (1968) and
Bandura (1969) suggested that observational and
instructed fear learning needed to be added to
classical conditioning as important routes to the
development of phobias. Supported by an extensive literature on retrospective reports of the
antecedents of phobias, this research has been producing viable experimental models of phobic
fear acquisition through both observational and
instructed fear. These have recently been used to

learn more about how specific anxious populations
(e.g., phobic and anxiety patients) respond to
socially transmitted threats.
3. Since the 1960s, experimental research on healthy
humans by social and biological, and more recently
developmental, psychologists have provided independent support for the idea that classical conditioning processes might be partly responsible for
SLoF. Importantly, this line of work has also
emphasized that social fear transmission in humans
is dependent on social characteristics of the model
or instructor (e.g., emotional expressions and spoken language), which has shown to affect the ensuing learning. Early studies of observational fear
learning demonstrated that the learning is dependent on the attribution of emotions and behavioral
dispositions to the model during the learning procedure (e.g., Berger 1962).

Important Scientific Research and
Open Questions
Current research on SLoF study both humans and
nonhumans to explore unanswered questions
remaining from past experimental research. This
research has continued to use measures of behavior
(e.g., avoidance), cognitions (e.g., verbal expressions),
and psychophysiology (e.g., skin conductance and
potentiation of the eye-blink startle reflex) in both
healthy subjects and patients (e.g., Askew and Field
2008). SLoF has also been investigated by more recently
developed methods of neuroimaging, such as functional magnetic resonance imaging (fMRI) and eventrelated potentials (ERP), as well as with the help of
pharmacological manipulations and genotyping in
both humans and nonhumans (e.g., Olsson and Phelps
2007). In light of these new developments and past
research, it is now widely believed that SLoF shares
many features of classical conditioning. For example,
in observational fear learning – like in classical fear
conditioning – the observer’s learned response (conditioned response, CR) to the CS is not of the same kind
as the unconditioned response (UR) to the model’s
emotional expression (the social US). Rather, research
suggests that there is an overlap between the observer’s
and the model’s UR and CR, respectively. Moreover,
fear acquired through both observation and verbal
instruction is expressed through the same brain

Social Learning of Fear

systems (critically including the amygdala) and peripheral neural systems.
Much research has been spent on investigating the
significance of the model in observational fear learning.
In the past, the view that the model’s emotional expression constitutes an US has been questioned. At least
two alternative explanations exist: (a) the learner’s
acquired fear might be due to attention to the model’s
US rather than the model’s responses per se and/or (b)
a result of social inference, in which the model’s fear
expression is a CS (and not an US) that was previously
associated with a directly experienced aversive event
(US). Research in both humans and nonhumans showing that the learner’s UR is present in the absence of the
model’s US, and the fact that the learner’s UR in observational fear learning shows characteristics similar to
the analogous response in classical conditioning argues
against (a) and (b), respectively.
Adding to the view that SLoF relies on the same
learning processes as direct learning procedures is the
view that social cognitions also play an important role
in the learning. Research has demonstrated that human
learners do not need to observe the model’s expression
of fear to acquire a fear response. Vicarious fear learning is dependent on the observer’s interpretation of the
model, as well as the learning context. Research on
vicariously instigated emotions and empathy has
described how an observer’s emotional response to
a distressed other is dependent on the observer’s bond
to the other individual. For example, a potential future
competitor or someone who has defected in a previous
social game elicits less empathy (and in some cases even
reward-related responses) in the observer. However, it
is unknown how these variations in instigated emotions are affecting the observer’s learning.
Recent research using neural measures in human
and nonhuman animals supports the view that observational and instructed fear learning partly rely on the
same neural processes as classical conditioning, but
that these processes are modifiable by social cognitive
processes (Olsson and Phelps 2007). Current research
has begun to investigate the role of mental attributions
to the learning model (or instructor) in SLoF.
A growing understanding for the neural mechanisms
of mental state attributions, such as the attributions of
intentions and emotions, as well as the increasing mapping of the processes underlying empathy are likely to

S

3115

enable a better appreciation of the role of social cognitions in SLoF. Continued research on SLoF is likely to
enhance our understanding for two related fundamental questions related to its nature; what are the specific
contributions to SLoF by (1) data-driven (perceptual)
information and top-down (abstract) information, and
the interaction of the two; and (2) mechanisms of
classical conditioning and social cognition and the
interaction of the two. The enhanced understanding
of the specific mechanisms of different kinds of SLoF
(such as observational and instructed fear learning)
will underscore the importance to study how these
forms interacts among themselves and with other pathways to fear, such as classical fear conditioning, and
anxious dispositions believed to underlie the development of anxiety disorders. Another important question
for future research will be to explore fear learning
through the observation of others’ instrumental
actions. This would constitute an indirect learning
analogue to instrumental conditioning in which punishments of others actions can produce avoidance
and fear.

Cross-References
▶ Associative Learning
▶ Facial Expression Learning
▶ Fear Conditioning in Animals and Humans
▶ Learning in the Social Context
▶ Mimicry in Social Interaction: Its Effect on Learning
▶ Neural Substrates of Avoidance Learning
▶ Observational Learning: The Sound of Silence
▶ Social Learning
▶ Stress, and Learning

S
References
Askew, C., & Field, A. P. (2008). The vicarious learning pathway to
fear 40 years on. Clinical Psychology Review, 28, 1249–1265.
Bandura, A. (1969). Principles of behavior modification. New York:
Holt, Rineheart and Winston.
Berger, S. M. (1962). Conditioning through vicarious instigation.
Psychological Review, 69, 450–466.
Mineka, S., & Cook, M. (1993). Mechanisms involved in the observational conditioning of fear. Journal of Experimental Psychology:
General, 122, 23–38.
Olsson, A., & Phelps, E. A. (2007). Social learning of fear. Nature
Neuroscience, 10, 1095–1102.
Rachman, S. (1968). Phobias: Their nature and control. Springfield:
Thomas.

3116

S

Social Learning Theory

Social Learning Theory
SHERRY D. LYONS1, ZANE L. BERGE2
1
Continuing & Professional Studies - English Language
Institute, University of Maryland Baltimore County
(UMBC), Baltimore, MD, USA
2
Department of Education, University of Maryland
Baltimore County (UMBC), Baltimore, MD, USA

Synonyms
Observational Learning; Social Cognitive Theory

Definition
Social learning theory, or SLT, is predicated on the
notion that learning occurs through social observation and subsequent imitation of modeled behavior.
According to SLT, humans learn from observing the
actions and resulting consequences of others. By doing
so, individuals can learn to imitate the observed behavior, and thus reap the rewards, or they can learn not to
imitate a particular action and thereby avoid the disagreeable consequences. Often seen as a bridge between
both behaviorist and cognitive learning theories, social
learning theory involves reciprocal interaction between
cognitive, behavioral, and environmental influences
(Pajares 2002).

Theoretical Background
Associated most notably with Albert Bandura, SLT has
its roots in Rotter’s Social Learning and Clinical
Psychology (1954, for related content see http://psych.
fullerton.edu/jmearns/rotter.htm). A quarter century
later, Bandura expanded and refined Rotter’s ideas in
Social Learning Theory (1977). SLT is also related to
Vygotsky’s social development theory (1978, for related
content see http://tip.psychology. org./vygotsky.html).
SLT posits that learning best takes place in social contexts through observation, imitation, and modeling. At
its inception, social learning theory challenged the traditional principles of behaviorism and its perceived
limitations as a learning theory. It called into question
whether true learning could only result from the experiences of reinforcement or punishment, which was the
prevailing behaviorist view of the time. Social learning
theory also deviated from the basic tenets of behaviorism by espousing the importance of cognition, which

had been excluded from the central ideas of behavioral
theories, and by supporting the major role that cognitive approaches play in learning processes.
Social learning theory is based on Bandura’s premise
that learning does not always occur as a result of firsthand experiences alone, but through harnessing the
power of observation and imitation (Martinez 2010).
Bandura states that through observing others, humans
have the capacity to develop ideas about how new
behaviors are performed. This information is then
coded, stored into memory, and serves as a guide for
action either immediately after the observation or for
use on later occasions. The four major components,
according to Bandura (1977), that comprise observational learning or modeling include:
1. Attention – In order for learning to take place during observation, individuals must pay attention to
the modeled behavior. Characteristics of both the
observer and the person being observed (the
model) can influence how much attention is given
to the modeled activities. For example, an observer
who is sleepy, sick, or otherwise distracted will not
likely have the same level of attentiveness as an
observer who is completely focused on the model.
2. Retention – If individuals are to learn from observed
behavior, they must, in turn, remember the modeled
activities. Retention and recall can be aided through
the use of imagery and descriptive language, thus
increasing the likelihood that the modeled behavior
can be reproduced by the observer.
3. Reproduction – At this stage, the observer translates
the modeled behavior into their appropriate individual actions. Reproducing the observed behavior
involves converting the retained imagery and language, provided by the model, into a response that
is in line with the modeled pattern. Behavioral
reproduction improves as the observer practices
the new behavior.
4. Motivation – Reproducing observed behavior
requires some motivation to do so. Without some
reason for imitating the modeled behavior, it is
unlikely that individuals would make the effort.
Although the cornerstone of observational learning
is attention, memory, and motivation, Bandura began
to add to his work on modeling by also exploring the
cognitive concept of self-efficacy beliefs as they relate to
advancing the understanding of human behavior

Social Learning Theory

(Martinez 2010). Self-efficacy is the belief in one’s own
capabilities. These beliefs can significantly influence
one’s environment and outcomes, and often determine
whether or not someone is able to successfully perform
specific actions (Foster 2006; Martinez 2010). Bandura’s
focus on self-efficacy pushed his social learning theory
deeper into the cognitive realm; thus, “He coined the
term social cognitive theory, holding that a person’s
behavior, environment and inner qualities interact,
rather than one of them being predominant in explaining how people function” (Foster 2006, para. 10).

Important Scientific Research and
Open Questions
Before the advent of social learning theory, behaviorist
theories influenced the ways in which learning (i.e.,
behavior) was viewed. As the basic tenets of behaviorism were called into question, however, particularly
those relating to reinforcement, punishment, and the
exclusion of cognitive approaches, other views about
learning began to emerge and directly challenged the
foundation of behaviorist perspectives (Martinez
2010). Albert Bandura, whose name has become synonymous with social learning theory, was one of the
first to expand on the earlier ideas that cognition, and
not merely behavior, is an important element of learning. In fact, according to Martinez (2010), Bandura’s
social learning theory requires neither obvious behavior
on the part of the learner, nor any reinforcement or
punishment, all of which are associated with behaviorist theories. The learning actually takes place through
social observation.
A prime example of social observation at work is
Bandura’s Bobo doll experiment. During the experiment, children were exposed to both aggressive and
nonaggressive models that played with the dolls either
violently or placidly. The results revealed that the “subjects in the aggression condition reproduced a good deal
of physical and verbal aggressive behavior resembling
that of the models. . .” (Bandura et al. 1961, Results
section, para. 1). The same was true for the subjects
exposed to nonaggressive models. In other words, the
children who were exposed to the nonaggressive model
imitated similar nonaggressive behavior when playing
with their Bobo doll. This experiment revealed that
merely observing the behavior of a model was enough
to influence behavior, especially among children
(Bandura et al. 1961, as cited in Martinez 2010). This

S

idea of observed/modeled behavior, particularly as it
relates to influencing children’s behavior, has sparked
questions over whether exposure to violent/aggressive
imagines in today’s media (news, television programs,
movies, video games, and social networks, to name
a few) actually causes aggressive behavior in children.
According to Kevin O’Rorke (2006), mass media plays
a highly influential role in today’s society, and thus,
exposure to models represented in the media can have
a powerful impact on shaping human behavior, especially aggressive behavior in children. In fact, research
has shown that exposure to televised violence can
increase children’s aggressiveness within their lifetime,
thus illustrating the significant role that social learning
theory plays in influencing behavior and learning
through observation (O’Rorke 2006).
The work of the early pioneers of SLT has sparked
more recent research in how social learning theory
relates to such diverse areas as criminology and deviant
behavior, advertising and consumer behavior, health
and exercise, as well as technology and social media,
to name a few. Currently, as SLT relates to criminology
and deviant behavior, researchers continue to expound
on the earlier work of Edwin Sutherland, Robert
Burgess, and Ronald Akers. For example, Brauer
(2009) states that through SLT, we can better understand how individuals learn criminal and deviant
behavior, which is learned in much the same way as
noncriminal behavior. In other words, through associations or social interactions, over time, with those who
serve as a model for criminal and deviant behavior,
individuals can learn to imitate the observed behavior,
particularly in the absence of opposing positive influences or models to the contrary. Whether the deviant
behavior is imitated or rejected by the observer is
a result of either directly or vicariously experiencing
the rewards or punishments of the modeled behavior.
Furthermore, researchers are presently examining
SLT as it relates to advertising. Specifically, Kinard and
Webster (2010) have explored the impact of advertising
and Bandura’s self-efficacy construct on unhealthy consumption behavior of adolescents, namely, their use of
tobacco and alcohol consumption. Advertising has
long been thought to have a significant influence on
individual behavior by using favorable stereotypes that
are attractive, successful, and healthy. These positive
images are widely used to depict essentially risky and
unhealthy consumer behaviors, such as smoking and

3117

S

3118

S

Social Network

alcohol consumption (Pechmann and Knight 2002;
Pechmann and Shih 1999, as cited in Kinard and Webster 2010). Adolescents tend to be particularly susceptible to these images, and thus are especially open to
initially engaging in smoking cigarettes and drinking
alcohol as well as maintaining these behaviors throughout adolescence and into adulthood. Moreover, advertising has been shown to act as an additional model of
behavior that reinforces the social influences of adolescents’ parents and peers who smoke and drink. The
added reinforcement provided by advertising increases
the likelihood that the observed behaviors (i.e., smoking
and drinking alcohol) will be imitated by adolescents.
Researchers continue to examine the established
tenets of social learning theory and its application
across a broad spectrum of areas, and the basic principles of SLT continue to hold true. In other words, to
one degree or another, a great deal of individual learning takes place within a social context through observing and eventually imitating the behavior (positive or
negative) of others. Although other complex factors,
such as self-efficacy and the perceived rewards and/or
punishments of the behavior, can influence the degree
to which, if at all, the modeled behavior will be imitated
by the observer(s), research (past and present) has
shown that regardless of the domain, observation
plays a powerful role in the learning process.

Cross-References
▶ Imitation and Social Learning
▶ Learning in the Social Context
▶ Self-Efficacy for Self-Regulated Learning
▶ Social Learning
▶ Socio-Constructivist Models of Learning
▶ Socio-Cultural Research on Learning

References
Bandura, A. (1977). Social learning theory. New York: General Learning Press.
Bandura, A., Ross, D., & Ross, S. A. (1961). Transmission of aggression through imitation of aggressive models. Journal of Abnormal
and Social Psychology, 63, 575–582. http://psychclassics.asu.edu/
Bandura/bobo.htm#f2. Accessed 19 Nov 2010. (First published in
Journal of Abnormal and Social Psychology.).
Brauer, J. R. (2009). Testing social learning theory using reinforcement’s residue: A multilevel analysis of self-reported theft and
marijuana use in the national youth survey. Criminology, 43(3),
929–970. http://onlinelibrary.wiley.com.proxy-bc.researchport.
umd.edu/doi/10.1111/ j.1745-9125.2009.00164.x/pdf. Accessed
17 Nov 2010.

Foster, C. (2006 September/October). Confidence man. Stanford
Magazine.
http://www.stanfordalumni.org/news/magazine/
2006/sepoct/features/bandura.html. Accessed 19 Nov 2010.
Kinard, B. R., & Webster, C. (2010). The effects of advertising, social
influences, and self-efficacy on adolescent tobacco use and alcohol consumption. The Journal of Consumer Affairs, 44(1), 24–43.
http://onlinelibrary.wiley.com/doi/10.1111/j.17456606.2010.01156.x/pdf. Accessed 26 Nov 2010.
Martinez, M. E. (2010). Learning and cognition: The design of the
mind. Upper Saddle River: Person Education.
O’Rorke, K. (2006). Social learning theory & mass communication.
ABEA Journal, 25, 72–74. http://lpru.asu.edu/v25/v25v22n2.pdf.
Accessed 3 Dec 2010.
Pajares, F. (2002). Overview of social cognitive theory and of selfefficacy. http://www.des.emory.edu/mfp/eff.html. Accessed 19
Nov 2010.
Pechmann, C., & Knight, S. J. (2002). An experimental investigation
of the joint effects of advertising and peers on adolescents’
beliefs and intentions about cigarette consumption. Journal
of Consumer Research, 29, 5–19. http://web.gsm.uci.edu/
antismokingads/articles/JCR_File.pdf. Accessed 26 April 2011.
Pechmann, C., & Shih, C. F. (1999). Smoking scenes in movies and
antismoking advertisements before movies: Affects of youth.
Journal of Marketing, 63, 1–13. http://web.gsm.uci.edu/
antismokingads/articles/pechshih.pdf. Accessed 26 April 2011.
Rotter, J. B. (1954). Social learning and clinical psychology. New York:
Prentice-Hall.
Vygotsky, L. S. (1978). Mind and society: The development of higher
mental processes. Cambridge, MA: Harvard University Press.
Further Reading
Rotter. http://psych.fullerton.edu/jmearns/rotter.htm. Accessed 23
Nov 2010.
Vygotsky. http://tip.psychology.org/vygotsky.html. Accessed 23 Nov
2010.

Social Network
▶ Network Communities
▶ Trust into e-Learning

Social Network Analysis (SNA)
▶ Social Networks Analysis and the Learning Sciences

Social Network Learning
▶ Learning Through Social Media

Social Networks Analysis and the Learning Sciences

Social Networks Analysis and
the Learning Sciences
JOSEPH PSOTKA
Basic Research Unit, US Army Research Institute for
the Behavioral and Social Sciences, Arlington, VA, USA

Synonyms
Dynamic network analysis (DNA); Friendship networks; Social graphs; Social network analysis (SNA)

Definition
Social networks provide a linkage analysis of relationships among people, where the links can be almost any
kind of relationship, such as friendship, trust, seeking
or providing advice, references, dominance, or interest
and the nodes can be things, ideas, papers, people,
groups, organizations, or larger entities. From simple
graphs and trees to complex networks, these relationships can be visualized, factor analyzed, or implicit, but
they offer educators new insights into the sometimes
hidden attitudes, motivation, knowledge, and the ability of their students, parents, peers, and beyond in
informal and commercial settings. Educational networking uses social network technologies for educational purposes.

Theoretical Background
As this is written, social network sites throughout the
world are increasingly popular: some (such as
Facebook) have become the most visited sites on the
Internet. These are sites where friends and acquaintances (and some strangers) share daily and even
minute by minute updates on their activities, interests,
plans, and attitudes. They comprise Wikis and Blogs
and photo galleries and games and short messages.
Students are known to stay up all night sharing their
tweets and messages. Since this phenomenon has arisen
dramatically in the last few years, it is not surprising that
so far it has had little impact on teaching and learning.
As a result, this short article is short on research findings
and conclusions and long on speculation.

Social Networks and Students
Many authors have suggested ways to adapt students’
existing habits on social network sites to the

S

3119

educational realm; for instance, such habitual activities
as messaging each other, checking each other’s profiles,
and browsing upcoming parties might all be adapted to
education. The most frequent and self-evident suggestion for educators thinking about how to use social
networks in education is to reach out to students and
invite their participation in asynchronous and synchronous discussion groups on these popular sites.
Participation and engagement could be encouraged
with incoming classes, such as college applicants, or
in classes of already enrolled pupils. The central idea is
that this engagement could motivate and support students and increase their likelihood of staying through
to graduation, and promote more interchange of ideas
in the classrooms. Boosting engagement has been
shown to have a clear effect on how well-integrated
students feel within their social and academic environments. However, it is still very unclear how well students’ social networking habits translate directly to
teaching and learning environments. In particular
mentoring students with feedback and advice may be
just as time consuming as face-to-face mentoring, and
less effective. Peer to peer sharing and support may be
more easily exploited for education, and that may
already be happening without formal educational
support.
On the negative side, social networks have the
potential to distract students, and fill their academic
time with an overabundance of irrelevant detail, and
possibly even encourage cheating among classmates.

Social Networks and Teachers
Within many popular social network environments
many sites have grown up to link faculty and students,
or faculty to faculty. If this continues and remains
successful beyond the initial flurry of creative excitement, these sites could serve many purposes such as
continuing professional development, mentoring, and
enrichment. One such site reports that in a first year
reading program, they have had over 100 staff, faculty,
and students join. However, it soon became a not very
active page, despite what was thought to be attractive
and inviting discussion threads. Students did not really
talk to each other on the site, and it became apparent
that educators may overstate the usefulness of social
networking technology for teaching.
Other sites are building social networking software
designed especially for education, designed from the

S

3120

S

Social Networks Analysis and the Learning Sciences

ground up to support learning. The main difference is
that these sites are not a centralized service. Whereas
most commercial networks are one space branded in
one particular way, special purpose education social
network sites will allow organizations to host their
own systems customized to their specific requirements,
but still retain connectivity from site to site. This is an
attempt to build a truly distributed learning landscape
which allows for local networks of personal acquaintances, with the wider benefit of being able to make
relevant connections across sites.

Social Network Analysis (SNA)
At the same time as social network, Internet sites have
become increasingly popular, the technology and statistics of social network analysis (SNA) has matured.
A brief description of these tools is provided, followed
by a short analysis of their use in education. For
research purposes several free public domain SNA programs are widely available but specific tools and environments for educators to display their students’ social
and knowledge networks are not yet widely available.

Definitions
Several concepts haven proven useful in the SNA of
educational groups.
Centrality: This measure indexes the representativeness and proximity of a node to all other nodes.
The most popular and well-connected person would
have high centrality, high degree, and high closeness.
Degree: The number of links to other nodes in the
network. Indegree and Outdegree measure the number
of links from and to others, respectively.

Important Scientific Research and
Open Questions
In research on social networks in the learning sciences
often only the easiest metrics with simple tools are
analyzed. For instance, the density of a network is at
a maximum when all the nodes are connected to each
other. However, there is a confounding with size. The
larger a group is, the harder it is for the network to be
fully connected, with everyone linked to everyone else.
However, if density is an important measure for how
well a class is interacting with each other, it must
somehow be normalized on class size, making the density measure more complex and creating a difficulty for
how to normalize it.

Increasingly, social networks are being computed
not directly from surveys of students’ preferences, but
instead from communication data in virtual environments, commercial social network sites, help and
advice networks, and preference networks in online
purchasing and rental stores.
For instance, an affiliation network of students can
be computed from conversation threads. Two social
networks can easily be derived from the same discussion group: a participant network that shows the connections among students who have participated in the
same conversation threads, and a network of threads
that have common students. Each provides information about students’ interests and abilities. Within
these networks the flow of knowledge and its increase
with teaching and learning can be assessed with more
complex analyses.
Expertise and trust have been measured in online
discussion environments and diverse question-answer
forums. In these, people network with each other to
share their expertise, technical knowledge, advice, and
opinions. Research has revealed that experts tend not to
reserve their expertise for the really difficult questions,
but instead answer more questions than others, including the simple questions. As attributes of experts are
uncovered in this research, guidance can be formulated
for teachers to better understand how to encourage
students to provide and seek advice and become
experts in their domains of personal interest.

Social Network Principles and
Findings
As a young science drawing from many disciplines,
SNA has few general findings that illuminate education
and learning. Here are two potentially useful insights
that SNA has produced.
Dunbar Number. Research has suggested that there
is a limit to the number of people who can be understood well enough to be called friends and be part of
a community. Because of the complexity and great
number of variables that need to be understood about
a person to call them a friend, the cognitive load of
developing a complete mental model of each limits
their number to about 150. One can see this limit in
operation in the social network data of many organizations: a company in the army, a community neighborhood in a city, and the size of villages in agricultural
communities.

Social Psychology of Music Instruction and Learning

Small World: Early experiments using mail systems
discovered that any person on the US East coast could
be found from the West coast using five or six intermediary people who each knew each other personally. This
has been encapsulated by the phrase: “six degrees of
separation.” More recent explorations of the phenomenon using email arrive at very similar results. In comparison with the hub-based structure of Internet
servers, where certain sites, such as Google or Facebook
have a preponderance of links, the links among people
are much more evenly distributed, but both distributions may be best described by power laws.

Cross-References
▶ Asynchronous Learning
▶ Asynchronous Learning Networks
▶ Community of Practice
▶ Interests and Learning
▶ Motivational Variables in Learning
▶ Semantic Networks
▶ Shared Cognition

S

3121

Social Psychology of Music
▶ Social Psychology of Music Instruction and Learning

Social Psychology of Music
Education
▶ Social Psychology of Music Instruction and Learning

Social Psychology of Music
Instruction and Learning
C. VICTOR FUNG, CLINT RANDLES
Center for Music Education Research, School of Music,
MUS 101, College of The Arts, University of South
Florida, Tampa, FL, USA

Further Reading
Adamic, L. A., Zhang, J, Bakshy, E, Mark S., & Ackerman, M.S.
(2008). Knowledge sharing and yahoo answers: Everyone
knows something. Proceeding of the 17th international conference
on World Wide Web, Beijing, China, 21–25 April 2008.
Barabasi, A.-L. (2002). Linked. New York: Penguin.
Toikkanen, T., & Lipponen, L. (2009). The applicability of social
network analysis to the study of networked learning. Interactive
Learning Environments, 17, 1–15. iFirst article.
Wasserman, S., & Faust, K. (1998). Social network analysis: Methods
and applications. Cambridge: Cambridge University Press.

Social Norms
▶ Culture in Second Language Learning

Social Practices
Social practices are the means by which people participate in cultural activities. Literacy, for example, can be
defined as a set of practices that enable participation in
various discourses associated with cultures. Social
practices are also the means by which people take on
and enact various (learning) identities.

Synonyms
Social psychology of music; Social psychology of music
education

Definition
Social psychology of music instruction and learning is
the study of relationships among individuals, groups of
individuals, and the society in music instructional and
learning contexts. It contains a tradition in psychology,
which tends to focus on the effect of social factors on
individuals, and a tradition in sociology, which tends to
view individuals within a larger social network. Since
the vast majority of music instructional and learning
activities involve two or more individuals in a social
context, this is an enormously broad area of study
that covers many aspects of such interactions and
relationships.

Theoretical Background
Social psychology of music instruction and learning
developed as a parallel with social psychology since
the early part of the twentieth century. This was mainly
due to the work of Paul R. Farnsworth (1899–1978),
who was trained as a psychologist but with an intense
interest in music and a strong desire to work with

S

3122

S

Social Psychology of Music Instruction and Learning

musicologists, and later with music educators (LeBlanc
2001). He published many music articles in psychology
journals since the 1920s. Since the late 1930s, his work
began to appear in music journals. His last book, The
Social Psychology of Music (first published in 1958,
second edition in 1969), was seminal in establishing
the field of social psychology of music. His work covered many different topics in this field. The most extensive and well-known ones were musical eminence and
musical taste.
Other social psychologists, such as Daniel Berlyne
(1924–1976) and David Hargreaves (1948–), built
upon the work of Farnsworth and consolidated the
field of social psychology of music. Berlyne was
known for his experimental aesthetics which suggested
an inverted-U relationship between liking and certain
qualities of the stimuli, such as complexity. Hargreaves
and Adrian North coedited a book, The Social Psychology of Music, published in 1997 and coauthored
another book, The Social and Applied Psychology of
Music, published in 2008. Also, McDonald, Hargreaves,
and Miell coedited a book in 2002 titled Musical Identities. The tables of contents of these books reflected to
some degree the scope of the field. Hargreaves and
North (1997) covered the areas of individual differences, social groups and situations, social and cultural
influences, developmental issues, musicianship, and
real world applications including education, therapy,
and commerce. North and Hargreaves (2008) included
the areas of composition and musicianship; musical
preference and taste; “problem music” and subcultures;
music, business, and health; and musical development
and education. McDonald and others (2002) addressed
how musical identities were developed in the school
and family environments and in various age-groups
and performance contexts. It also tackled how identities were developed through music in different gender,
youth, and national groups.
Music educators saw the relevance of Farnsworth’s
and other social psychologists’ work and expanded
on the concept of social psychology to include
specific situations applicable in music instruction and
learning. These music education researchers included
Harold F. Abeles in sex-stereotyping of musical instruments, Ed Asmus in music attribution, Manny Brand
in parental involvement in music learning, Lucy Green
in informal learning in music, Albert LeBlanc in music
preference, Anthony Kemp in music and personality,

Charles P. Schmidt in applied music instructions, and
so forth. Below are highlights of current interests in
some of these areas.
Identity has become an important topic in music
education research. Adding to the complexity of the
topic is the idea that there are differences among individuals, which entail many variables relevant to music
instruction and learning. Such different variables
include sex, age, ethnicity, culture, personality, and so
forth. The literature suggests that these variables not
only have an impact on musical development and
musical responses, but also help shape how music
learners perceive themselves as music makers.
Another layer of musical identity issues is the disconnect between music learner identity and teacher
identity. Roberts (2010) cites the path of a music
teacher, from being solely a performer to being primarily
a teacher of performers, as being a source of concern for
the music education profession. Other authors have
explored the possibility that being a musician and
a music teacher might include performing different
diverse roles including composer, improviser, producer,
or conductor. Along with this emerging strain of
research is the idea that for much of the history of
music teaching and learning, performance has been the
primary mode of music instruction, to the neglect of
other modes of music making, such as composition,
improvisation, various forms of small-group music
making, and the inclusion of various musical styles
that do not go along well with the traditional composing-performing-listening sequence. Collaborative music
making and simultaneous composing-performinglistening modes have led to a new set of identities for
both the learner and the instructor. Social psychologists
in music have also considered the role of the instructor
and learner within the large ensemble – band, choir, and
orchestra – setting. These researchers in music education
often cite McCall and Simmons (1966) coauthored book
Identities and Interactions: An Examination of Human
Associations in Everyday Life, to account for the interactions of various identities within these settings.
Attribution of success and failure in music has been
another important topic guiding the work of music
education researchers. Some have applied Albert
Bandura’s social learning theory, Bernard Weiner’s
attribution theory, or Edward L. Deci and Richard M.
Ryan’s Self-Determination Theory to music studies.
Others have used James Connell’s self-system processes

Social Psychology of Music Instruction and Learning

to account for motivation and persistence in music.
Within Connell’s theory, self-system processes are
divided into three groups based on the following psychological needs:
1. Competence: Including strategies – knowing what it
takes to do well, and capacities – possessing the
skills to do well.
2. Autonomy: Including self-regulation – wanting to
do well because doing so has personal meaning.
3. Relatedness: Including self – accepting and caring
about oneself, and others – feeling accepted and
cared about by others.
These processes are carried out in social settings
and, therefore, have a number of interesting layers.
Some researchers extend these layers to include school
community, regional identification, nationality, and
global perspectives. One could easily find social psychology of music instruction and learning crossing
over with other subfields of psychology and other
neighboring fields, such as sociology and anthropology.

● Which factors contribute to differences in musical

behavior among members involved in a music
teaching and learning setting?
● What is the role of gender in music instruction and
learning, given the rapidly changing social climate
and availability of various technologies?

3123

● What is the effect of a single-sex music instruction

●

●

●

●

●

Important Scientific Research and
Open Questions
Scientific research in the field may be analogous to
a kaleidoscope. Researchers, be they psychologists or
music educators, tend to study topics that they see
most relevant and important. As the field continues
to develop, psychologists, musicians, and educators
should continue to be more aware of each other’s
work. They might learn from each other’s theories,
perspectives, methodologies, and findings. Questions
that remain open include those that are based on value
judgments. For example, what role should music preference play in the music instruction and learning process? How should individuals determine for themselves
what musical diet is most appropriate for themselves?
To what extent should some people, such as music
instructors, have authority over others’, such as music
learners’, musical choice? Open questions that could be
subjected to empirical observations include:

S

●

and learning environment compared to that of
a coeducational environment?
What theories are best in explaining the attribution
of success and failure in instruction and learning
specific to music?
Based on a current assessment of how and by which
means adolescents identify with music, how might
the traditional roles of music teachers change to
better reach the twenty-first century adolescent
music student? Which genres of music and which
modes of music making present the greatest potential to reach the adolescent musician and beyond?
What theories are best to account for the manifestation of identity within the teaching and learning
of music of both teacher and students?
What sort of power struggles take place in the
instruction and learning setting of music among
peers, teachers, parents, and the larger school
community?
How do students perceive the role of the ensemble
director in the learning exchange? How does the
director perceive the role of students in the learning
exchange? How do these perceptions interact? What
implications do these interactions have for music
instruction and learning in each particular setting?
How have musical experiences outside of educational institutions contributed to the learning of
music within these institutions, and vice versa?

Furthermore, much research in the social psychology of music instruction and learning has been focused
on studying the first 20-some years of live of the study
participants. Researchers are still exploring the roles
music plays beyond the early adult years and into the
final years of life. Given the lengthening life expectancy,
many may continue to learn and experience music and
find meaning in it for half a century or more beyond
the first 20-some years.

Cross-References
▶ Learning in the Social Context
▶ Learning Through Social Media
▶ Social Cognitive Learning
▶ Social Construction of Learning
▶ Social Influence and the Emergence of Cultural
Norms
▶ Social Interaction Learning Styles

S

3124

S

Social Rule

▶ Social Interactions and Effects on Learning
▶ Social Interactions and Learning
▶ Social Learning
▶ Social Learning Theory

References
Hargreaves, D. J., & North, A. C. (1997). The social psychology of
music. New York: Oxford University Press.
LeBlanc, A. (2001). Paul Farnsworth: Pioneer scholar of music listening preference. Bulletin of the Council for Research in Music
Education, 149, 3–12.
MacDonald, R. A., Hargreaves, D. J., & Miell, D. (2002). Musical
identities. New York: Oxford University Press.
McCall, G. J., & Simmons, J. L. (1966). Identities and interactions: An
examination of human associations in everyday life. New York:
The Free Press.
North, A. C., & Hargreaves, D. J. (2008). The social and applied
psychology of music. New York: Oxford University Press.
Roberts, B. (2004). Who’s in the mirror? Issues surrounding the
identity construction of music educators. Action, Criticism, and
Theory for Music Education, 3(2). Retrieved 26 Oct 2010, from
http://act.maydaygroup.org/articles/Roberts3_2.pdf.

Social Theory / Model
Social Theory is a communication model by which the
audience or reader is recognized as an active maker of
meaning, negotiating message content relative to prior
knowledge or experiences, and one’s individual identity
within the social context of the message transmission.

Social, Cultural, and Study
Resources at Home
▶ Family Background and Effects on Learning

Social-Constructivist Learning
Theory
▶ Activity Theories of Learning

Social Rule
▶ Learning and Evolution of Social Norms
▶ Normative Reasoning and Learning

Social-Emotional Competence
▶ Social-Emotional Learning Scale

Social Stressor
A social stressor is an evaluated performance situation
in a domain of personal importance in which one is
motivated to succeed.

Social Tagging / Folksonomy
A tag is a keyword assigned to a piece of information.
So tagging is the activity to assign tags. Social tagging
means to do this online and collectively, that is, within
a principally open group of Internet users in order to
share digital content. Social tagging is also called
folksonomy. The term folksonomy should express the
special characteristic that not experts but all or many
Internet users (the “folk”) make their own taxonomies.

Social-Emotional Learning
Scale
STEPHANIE D. H. EVERGREEN, CHRIS L. S. CORYN
The Evaluation Center, Western Michigan University,
Kalamazoo, MI, USA

Synonyms
Emotional intelligence; Emotional quotient; Socialemotional competence; Social intelligence

Definition
The Social-Emotional Learning Scale (SELS) is a 20-item
questionnaire that measures the latent constructs Task

Social-Emotional Learning Scale

S

Articulation, Peer Relationships, and Self-Regulation
through self-report. It is designed to assess socialemotional learning in late elementary students (ages
9–11 in the USA) to inform needs assessment, program
planning, and evaluation. Respondents endorse items
on a five-point scale, which ranges from strongly disagree
to strongly agree. The SELS instrument is a mixed
model, in that it taps into both personality and intelligence constructs.

leaving researchers and program planners without the
ability to meaningfully measure their findings against
other programs with common objectives. Therefore,
while measures of social-emotional learning are prolific, they are generally insufficient for the purposes that
the SELS sought to address.

Theoretical Background

Operationalization of social-emotional learning is still
a subject of debate. Some scholars believe that what is
labeled emotional intelligence is a product of mental
processes, in other words something that can be
learned. Such theories are bolstered by Gardner’s
Multiple Intelligences, in which interpersonal and
intrapersonal intelligences describe social-emotional
orientation (Gardner 1983). However, tests of discriminant validity have failed to show the construct as
operationally distinct from others, such as verbal intelligence (Caruso et al. 2002; Ciarrochi et al. 2000;
Roberts et al. 2001). Other theorists disagree with the
notion that intelligence is the basis for one’s level of
social-emotional understanding, suggesting instead
that it is a trait of one’s personality, such as aggression
or empathy (Bar-On 1997). While considerable
research has been conducted to investigate the distinction, definitive evidence is not yet available. Whether it
is a component of intelligence (i.e., something than can
be learned) or personality (i.e., something innate), or
some combination of the two, remains an open question for further research.

Social-emotional learning (SEL) refers to a child’s ability to “recognize and manage emotions, develop caring
and concern for others, establish positive relationships,
make responsible decisions, and handle challenging
situations constructively” (CASEL 2008, p. 1).
The SELS instrument was created to assist program
planners and evaluators in selecting an aspect of SEL as
a programmatic focus for populations of late elementary
students using reliable and valid measure. The need for
a common tool for curricula or program planning is
clear, because the instruments typically used to address
social-emotional development through these avenues
are dispersed. Program evaluations of well-regarded
SEL interventions have typically designed instruments
particular to the given program, whether they address
comprehensive or specific aspects of SEL, and regardless
of whether they are in a needs assessment or program
evaluation stage.
The development of the SELS is based in the work
of the Collaborative for Academic, Social, and Emotional Learning (CASEL), which acts as a bridge
between the theoretical and practitioner worlds.
CASEL was a reference for the four states that developed SEL guidelines for education and from which the
items for the SELS instrument were developed. Yet,
even with major theorists in collaboration with
CASEL, their list of suggested assessment tools did
not address program planning needs for a general education audience. The SELS was designed as an improvement of existing instruments, which tend to overfocus
on specific individual deficiencies or come buried in
a larger school climate survey, by providing a more
general tool having evidence of validity and reliability.
Creating a program-specific instrument, even if
culled from a combination of existing instruments,
is both costly and time consuming. Moreover, in
such situations, instruments are not norm-referenced,

3125

Important Scientific Research and
Open Questions

Cross-References
▶ Emotional Intelligence and Learning
▶ Emotional Learning
▶ Learning by Feeling
▶ Learning the Affective Value of Others
▶ Socio-Emotional Aspects of Learning

References
Bar-On, R. (1997). Bar-on emotional quotient inventory: Technical
manual. Toronto, ON: Multi-Health Systems.
Caruso, D. R., Mayer, J. D., & Salovey, P. (2002). Relation of an ability
measure of emotional intelligence to personality. Journal of
Personality Assessment, 79(2), 306–320.
Ciarrochi, J. V., Chan, A. Y. C., & Caputi, P. (2000). A critical evaluation of the emotional intelligence construct. Personality and
Individual Differences, 28, 539–561.

S

3126

S

Socialization

Collaborative for Academic, Social, and Emotional Learning (2008).
Basics, definition. Retrieved April 16, 2008, from http://www.
casel.org/basics/definition.php
Coryn, C. L. S., Spybrook, J. K., Evergreen, S. D. H., & Blinkiewicz,
M. V. (2009). Development and evaluation of the socialemotional learning scale. Journal of Psychoeducational Assessment, 27(4), 283–295.
Gardner, H. (1983). Frames of mind: The theory of multiple intelligences. New York: Basic Books.
Roberts, R. D., Zeidner, M., & Matthews, G. (2001). Does emotional
intelligence meet traditional standards for an intelligence? Some
new data and conclusions. Emotion, 1(3), 196–231.

Socialization
▶ Enculturation and Acculturation
▶ Extraversion, Social Interaction, and Affect Repair

Socialization of Intelligence
▶ Social Construction of Learning

Socialization-Related Learning
THOMAS G. REIO, JR.
Department of Leadership and Professional Studies,
Florida International University, Miami, FL, USA

Synonyms
Academic socialization; Adaptation; Informal learning;
Social competence; Workplace learning

Definition
Socialization in its broadest sense is the process where
inexperienced individuals are taught the skills, behaviors, norms, attitudes, values, and motivations required
to be able to function competently as a member of
a culture. Socialization entails transmitting culture
from one generation to the next, counting training
and development activities linked to specific tasks
or roles in groups, organizations, and occupations

(Maccoby 2007). Socialization-related learning, then,
is the learning associated with becoming a functioning
member of a culture.

Theoretical Background
Socialization encompasses a wide variety of activities
that influence the thinking and behavior of novices that
result in important societal outcomes. An important
consideration is the proactivity of the individual to
learn what is necessary to adapt and fit in to a culture,
and continuously so over time. Without proactively
learning during the socialization process, the novice is
far less likely to acquire the information necessary to
learn how to think and act in accordance with cultural
norms. The learning associated with being socialized,
i.e., socialization-related learning, is a curiosity-driven
process where individuals seek the new information
needed to situationally perform specific tasks and
roles in light of cultural norms and expectations (Reio
and Callahan 2004). Curiosity, the desire for new information and sensory experiences, stimulates observation, consultation, and directed thinking types of
exploratory behaviors (Berlyne 1960). Because individuals should not expect to be merely taught what they
need to know on a consistent basis, he or she must be
proactively curious and exploratory if they ever hope to
learn how to be and remain a fully functioning member
of a culture. Although formal learning in classroom
settings, for example, can be useful for facilitating
socialization-related learning (e.g., a children’s lesson
on how to behave properly in a museum and other
public places) socialization-related learning occurs
through engaging in mostly informal learning types
of activities, including coaching and mentoring, trialand-error experimentation, reflection, self-directed
study, etc.
With children, parents are considered to be primary
socialization agents where they teach the appropriate
social skills, attitudes, and behaviors to be able to
continuouosly adapt to a culture, serving ultimately
as a determinant of being an independent and successful adult (Taylor et al. 2004). In school settings, both
parents and teachers are the primary academic socialization agents that shape children’s thinking, behavior,
and academic success through supporting positive
school experiences. To be sure, however, there are a
number of other socialization agents or sources of
information and learning that influence children’s

Socialization-Related Learning

socialization-related learning and successful socialization-related outcomes. Older siblings, peers, mass
media (e.g., Internet, Twitter, podcasts, wikis, television, and magazines to name a few), and neighborhood
contexts also serve to shape children’s thinking and
behaviors related to performing well in school. Again,
children must be proactively curious and exploratory
about their environment in general if they hope to learn
and adapt to the daily contingencies associated with
being a successful member of their culture.
For adolescents, parents and teachers continue to be
powerful socialization agents, yet peers can also be
quite influential during this time (Ryan 2001). As
with children, mass media, older siblings, and neighborhood contexts can impact adolescent socialization,
but work contexts contribute to their socialization as
well. At school, peers communicate expectations for
valued forms of behavior. For example, most middleclass adolescents report that peers convey expectations
of cooperativeness and helpfulness in classroom settings and engagement at least to some degree in
prosocial behaviors like volunteering outside of school.
Peers also influence educational goal setting for academic accomplishment and other positive forms of
behaviors. The possibility always remains though for
peers to negatively influence socialization-related outcomes. Those who have not successfully navigated the
socialization process are less likely to have proactively
sought the information needed to fit in and function well at school. As a result, such adolescents are
predisposed to being socially rejected by their peers and
apt to being less adjusted and more withdrawn than
many of their classmates (Wentzel and Looney 2007).
In adulthood, individuals continue the socialization-related learning useful for adapting to societal
demands consistent with this stage of the lifespan.
Friends, peer groups, romantic and social partners,
media, and workplace contexts serve as socialization
agents during this time. Through socialization-related
activities, adults learn how to make productive relationships with family and friends, take on coaching and
mentoring roles, assume positions of leadership, and
acquire a sense of themselves as independent, competent societal members. These endeavors require being
proactively curious and exploratory to learn and adjust
to life’s challenges and stressors. Socialization-related
learning suffers when a culture does not embrace such
information seeking.

S

3127

In workplace settings, socialization helps organizational “outsiders” become organizational “insiders.”
The socialization process applies not only to organizational beginners like apprentices and trainees, but also
experienced workers at a company who are newcomers
to a workgroup or new facility. Further, socialization
occurs when workers get a new supervisor or attend
school. In other words, the socialization process in
organizations is a continuous one, regardless of one’s
organizational status, experience, or time on the job.
Socialization-related learning allows workers to adapt
to evolving contingencies in their work environment by
facilitating their adjustment to the organization’s
values and norms, clarifying role identities, developing
job- and performance-related skills and capacities, and
helping them learn to whom to turn for the information needed to interpret organizational uncertainties
(Reio and Callahan 2004). This process by which organizational members learn and adapt to their workplace
surroundings has been found to be consistent in union
and nonunion, profit and nonprofit, and government
and nongovernment settings.

Important Scientific Research and
Open Questions
Mounting research evidence supports the relevance of
socialization-related learning to adapting to challenges
and opportunities presented by striving to be
a competently functioning member of one’s culture
across the lifespan. A number of socialization agents
have been identified (e.g., teachers, peers, parents,
mentors), but an understanding of the agents’ degree
of salience during specific stages or phases of the
lifespan remains unexplored. Moreover, there is not
simply a one-way transfer of information from these
agents to those who are less experienced. For example,
mentors impart important information to mentees
concerning career and social development issues, yet
mentees also contribute their insights and prior knowledge into enriching the mentor’s understandings of
these same organizational issues. It is likely, therefore,
that there exists a bidirectional learning relationship
between those who are more and less experienced.
More complete understandings of the socialization
process will require acknowledging this bidirectionality
in research designs. Finally, it is clear that proactive
information seeking in the form of curiosity and exploration is a key component of the socialization process

S

3128

S

Socially Biased Learning

as it motivates learning and adaptation, yet researchers
have little sense of how and to what degree to promote
this proactivity across a broad range of environmental
settings.

Socio-Cognitive Conflict
▶ Cognitive Conflict and Learning

Cross-References
▶ Adaptation and Learning
▶ Adaptation and Unsupervised Learning
▶ Coaching and Mentoring
▶ Organizational Learning

References
Berlyne, D. E. (1960). Conflict, arousal and curiosity. New York:
McGraw-Hill.
Maccoby, E. E. (2007). Historical overview of socialization research
and theory. In J. E. Grusec & P. D. Hastings (Eds.), Handbook of
socialization: Theory and research (pp. 13–41). New York:
Guilford.
Reio, T. G., Jr., & Callahan, J. (2004). Affect, curiosity, and socialization-related learning: A path analysis of antecedents to job performance. Journal of Business and Psychology, 19, 3–22.
Ryan, A. M. (2001). The peer group as a context for the development
of young adolescent motivation and achievement. Child Development, 72, 1135–1150.
Taylor, L. C., Clayton, J. D., & Rowley, S. J. (2004). Academic socialization: Understanding parental influences on children’s schoolrelated development in the early years. Review of General Psychology, 8, 163–178.
Wentzel, K. R., & Looney, L. (2007). Socialization in schools settings.
In J. E. Grusec & P. D. Hastings (Eds.), Handbook of socialization:
Theory and research (pp. 382–403). New York: Guilford.

Socio-Constructivist Models of
Learning
VASILIKI K. SIMINA
University of Manchester Institute of Science and
Technology (UMIST), Manchester, UK
Thessaloniki, Greece

Synonyms
Sociocultural model; Sociological model

Definition
The term is used to describe instructional design
models drawing from the learning theory of social
constructivism. Since learning is considered to be an
active process with the learners constructing their
knowledge on their own based on experience and
reflecting on this experience, social constructivism
focuses on the social and cultural context which shapes
the construction of knowledge.

Theoretical Background

Socially Biased Learning
▶ Social Learning in Animals

Socially Transmitted “Memes”
▶ Animal Culture

Society of Mind
▶ Schema-Based Architectures of Machine Learning

Teaching and learning theory has seen a paradigm
change in which the learner becomes the center of
learning and does not passively receive knowledge.
Consistent with that direction is social constructivism,
which is a variable of the well-known theory of constructivism. The basic tenet of constructivism is that
individuals construct their knowledge on their own by
associating new with prior information through their
interaction with their social and physical environment
and by reflecting on their experiences. Elements of the
constructivist theory can be detected not only in the
antiquity (e.g., in Socrates’ maieutic method) but also
in discovery or experiential learning. However, the
theories of Jean Piaget and Lev Vygotsky are the most
influential for the structure of cognitive and social
constructivist theory respectively.

Socio-Constructivist Models of Learning

Piaget (1953) focused on cognitive development
and developed a theory of the different cognitive stages
children come to know the world. He asserted that
mind has mental structures, called schemata, which
are modified and enlarged and which then become
more complicated with mental development through
the processes of assimilation and accommodation from
infancy to adulthood. Learning, here, is based on discovery, experience, and the predisposition of the individual to adapt to his environment, something which
he struggles to succeed by establishing equilibrium
between schemes and the environment.
With Vygotsky (1978), the emphasis is not only on
the cognitive aspect of learning, but also on the social
one because the individual cannot be separated from
the society and, therefore, thinking develops under
particular sociocultural conditions. One of his fundamental ideas is that cognition develops out of social
interaction and negotiation of meaning. In that sense,
the construction of knowledge is a shared, rather than
an individual, experience. Through dialogue, information is transmitted helping the individual to internalize
knowledge, thinking to be developed and learning to
occur. Language, which is socially and culturally oriented, plays a significant role in this process since it is
the necessary tool for conveying and negotiating meaning. In addition, the notion of the “zone of proximal
development” is basic in the social construction of
understanding because through the process of scaffolding the learner is guided in such a way so as to perform
at a level beyond the one at which he could perform
on his own. Therefore, social interaction promotes
collaboration for learners to achieve self-reliance
(Wood et al. 1976).
Another fundamental assumption of social constructivism is that cognitive processes, thinking and
learning included, are situated in social and physical
contexts. Social interaction is a critical component of
situated learning. According to Brown et al. (1989),
learners are involved in a “community of practice”
which embodies certain beliefs and behaviors to be
acquired. They also argue that learning, both inside
and outside school, advances through collaborative
social interaction and social construction of knowledge. Knowledge is contextually situated and is fundamentally influenced by the activity, context, and
culture in which it is used. Therefore, knowledge cannot be taught in abstract but within context. For that

S

reason, in situated learning knowledge is presented in
an authentic context, that is, in settings that normally
involve that knowledge. Approaches such as taskbased, project-based, and content-based try to integrate learners in authentic environments.
Context is very important for the constructivists
who believe that types of learning cannot be identified
independent of the content and context of learning. In
a socio-constructivist model of learning, it is not only
the sociocultural conditions in which learning occurs
that should be taken into consideration, but also the
social and cultural context which each learner carries
within the learning process. All these contexts create
a variety of perspectives which learners can have access
to through social interaction and negotiation of meaning and help learners construct knowledge on their
own. Therefore, learning must be situated in rich,
reflective real world contexts for the constructive learning process to occur.
Taking all the above into consideration, a socioconstructivist model of learning should consist of the
following principles. First of all, learning should be
considered as an active process in which the learner is
placed in the center while the teacher plays the role of
the facilitator or the guide. The learner should be able
to make associations with his prior knowledge and
experience. Moreover, the focus is on meaning and
not on structure. Hence, the learning environment
should encourage collaboration, negotiation, and
social interaction and foster the sharing of multiple
perspectives, providing the stimuli for the construction
of knowledge. Therefore, learners should be engaged in
context-rich experience based and complex authentic
activities, which are meaningful to them, and where
authenticity and complexity appear within a proximal
range of the learner’s knowledge and prior experiences
(Bednar et al. 1992).
One type of online application, such as WebQuests
(Dodge 1995), is a potential example of good practice
of the socio-constructivist model of learning. In
WebQuests, learning is situated in authentic environments because they are based on scenarios which create
a situation in which the completion of a task is meaningful to the learners. Moreover, learners usually work
in groups promoting not only collaborative and cooperative learning, but also the sharing of meaning.
WebQuests also support scaffolding since the information is given to the learner little by little through

3129

S

3130

S

Socio-Constructivist Models of Learning

the different steps of the task, helping him/her gradually become self-reliant. The multimodal aspect of
WebQuests provides a wide variety of resources which
learners can use in order to construct knowledge on
their own. Finally, learners have the opportunity to
reflect on what they have learned and discuss possible
extensions and applications of the gained knowledge.
Therefore, all these characteristics make WebQuests an
ideal socio-constructivist learning environment.

Important Scientific Research and
Open Questions
The term “social constructivism” is used in almost
every discipline. However, it is remarkable how frequently it appears in the discourse of educational
research, theory, and policy, even if it is simply used
to characterize any innovative approach to learning
where the learner plays a more active role and there is
some kind of interaction. Throughout the literature,
despite the fact that most researchers are thoroughly
engaged with the broad theory of social constructivism
and few papers succeed in developing applied models
of teaching, the interest is focused on (a) what can be
done within the classroom environment, (b) computer-assisted learning, and (c) curriculum theory.
To begin with, a general socio-constructivist framework can be effective for teaching students of all ages
and levels of ability (Watson 2001). Social constructivism can also provide crucial direction for teacher education since the paradigm shift in education demands
the transformation of the traditional teacher. Thus, the
socio-constructivist model of learning can be applied
in all levels of education, but what is still there to be
examined is how effective such a model can be on
a long-term level rather than simply within the framework of a project or task.
Moreover, computer-based learning seems to be
consistent with the principles of social constructivism.
Most attention is drawn to the online learning because
it fosters collaborative discourse and the individual
development of meaning through construction and
sharing of texts and other social artifacts. According
to Jonassen et al. (1999) multimedia and hypertext can
provide a learning environment where learners can
construct knowledge on their own. In addition, the
use of computer-supported collaboration and communication tools can extend the ways in which students
and instructors interact within and beyond classroom.

The claims for synchronous and asynchronous online
discussions are found to be frequently based on socioconstructivist principles. Chat rooms, blogs, wikis, and
podcasts can act as social mediation tools in the learning process because they enable the sharing of ideas and
experiences (Chandra and Chalmers 2010). Finally, the
scaffolding mechanisms can be effective in online environments such as learning management systems.
Therefore, computer technology offers so many potentials for applying the socio-constructivist model of
learning leading learners to self-reliance and instructors to the role of the facilitator.
Regarding the design of the curriculum, the way
students learn needs to be carefully considered and
the socio-constructivist model of learning can be an
effective way of conceptualizing learning. Over the past
years, curriculum theory has worked out many educational implications of social constructivism. Many
scholars have suggested the socio-constructivist framework in their claims for curriculum reform in many
countries. However, since there is a more holistic
approach toward the design of curricula, there are not
many curricula that can be merely characterized as
social constructivist. For example, in Turkey, there is
an attempt but only in subjects related to science (Fer
2009). However, we need to wait for the results of such
an endeavor in order to be able to discuss how interactive learning can help learners construct knowledge on
their own on long-term practical base.
Although social constructivism is a term which is
frequently used and teachers try to apply its implications in their everyday teaching, still the role of social
interaction in the learning process raises some questions. As Kittleson and Southerland (2004) said, we will
always wonder about what the co-construction of
knowledge looks like. They claimed that social factors
such as status, gender, and leadership style can have an
important role in shaping opportunities for knowledge
construction, but how can we make sure that the learning opportunities available to students in groups are
equally distributed across groups? What happens in the
cases where there is competitiveness between learners
or unwillingness to participate? Therefore, there must
be a clarification of the interaction between learners. It
is often presumed that it facilitates learning, but we
need to find the most suitable tools in order to measure
the learning outcomes associated with participation in
social interaction.

Sociocultural Research on Learning

Cross-References
▶ Collaborative Learning
▶ Computer-Based Learning
▶ Constructivist Learning
▶ Cooperative Learning
▶ Experiential Learning Theory
▶ Online Collaborative Learning
▶ Scaffolding Learning
▶ Situated Cognition
▶ Situated Learning
▶ Social Construction of Learning
▶ Social Learning
▶ Zone of Proximal Development

S

3131

immediately present, including historical factors, all
of which give meaning to learning. Crucially, learners
and learning are not merely situated in a sociocultural
context. They help shape the context and are simultaneously constituted by the context of which they
are part.

Sociocultural Model
▶ Socio-Constructivist Models of Learning

References
Bednar, A. K., Cunningham, D., Duffy, T. M., & Perry, J. D. (1992).
Theory and practice: How do we link? In T. M. Duffy & D. H.
Jonassen (Eds.), Constructivism and the technology of instruction:
A conversation (pp. 17–34). Hillsdale: Lawrence Erlbaum.
Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition and
the culture of learning. Educational Researcher, 18(1), 32–42.
Chandra, V., & Chalmers, C. (2010). Blogs, wikis and podcasts –
Collaborative knowledge building tools in a design and technology course. Journal of Learning Design, 3(2), 35–49.
Dodge, B. (1995). Some thoughts about WebQuests [WWW page].
Retrieved 30 Sep 2010, from http://webquest.sdsu.edu/about_
webquests.html.
Fer, S. (2009). Social constructivism and social constructivist curricula in turkey to meet the needs of young people learning science:
overview in light of the PROMISE project. In T. Tajmel & K. Starl
(Eds.), Science education unlimited: Approaches to equal opportunities in learning science (pp. 179–200). Münster: Waxmann.
Jonassen, D. H., Peck, K., & Wilson, B. (1999). Learning with technology: A constructivist perspective. Upper Saddle River: Prentice Hall.
Kittleson, J. M., & Southerland, S. A. (2004). The role of discourse in
group knowledge construction: A case study of engineering students. Journal of Research in Science Teaching, 41(3), 267–293.
Piaget, J. (1953). The origin of intelligence in the child. London:
Routledge & Kegan Paul.
Vygotsky, L. S. (1978). Mind in society: The development of higher psychological processes. Cambridge, MA: Harvard University Press.
Watson, J. (2001). Social constructivism in the classroom. Support of
Learning, 16(3), 140–147.
Wood, D., Bruner, J. S., & Ross, G. (1976). The role of tutoring in
problem solving. Journal of Child Psychology and Psychiatry,
17(2), 89–100.

Socio-cultural Context
The overall context of learning including the range of
social and cultural factors, many of which are not

Sociocultural Psychology
▶ Cultural-Historical Theory of Development

Sociocultural Research on
Learning
CAROLYN P. PANOFSKY
Department of Educational Studies, Rhode Island
College, Providence, RI, USA

Synonyms
Cultural historical activity theory (“CHAT”) research;
Cultural historical research; Socio-cultural-historical
research; Socio-historical research

Definition
Sociocultural research refers to work in the tradition
initiated in the 1920s by Russian thinker L. S. Vygotsky
(1896–1934) and his coworkers, especially A. R. Luria
and A. L. Leontiev. Any introduction to sociocultural
research must be framed within the key concepts of this
unique perspective. Although the key themes of this
approach were first set down by Vygotsky, his work was
cut short by his early death (from tuberculosis) at the
age of 37 and was also suppressed for many years by the
Soviet regime. Vygotsky’s work was first translated into
English in the 1960s and his original ideas continue
to be elaborated by an international community of

S

3132

S

Sociocultural Research on Learning

scholars, with some of his works still not translated into
English. The key texts containing Vygotsky’s work are
Thought and language (1962; 1986; also translated as
Thinking and speech [1987]) and Mind in society (1978).
The term “sociocultural” (or cultural–historical, or
socio–historical) highlights the prominence given in
this framework to social, cultural, and historical sources
of human development, mind, and learning, in contrast
to other theories that either see biological sources of
development as primary, or that view experience and
environment as having importance but do not foreground social, cultural, and historical contexts. In fact,
this approach goes further than taking social, cultural,
and historical contexts as influences on learning and
development. Rather, for researchers in this tradition,
the cultural activity setting itself – rather than the individual learner – comprises the unit of analysis in
research. As Rogoff explains,
"

In contrast to theories of development that focus on
the individual and the social or cultural context as
separate entities (adding or multiplying one and the
other), the cultural-historical approach assumes that
individual development must be understood in, and
cannot be separated from, its social and culturalhistorical context. According to Vygotsky’s theory, the
efforts of individuals are not separate from the kinds of
activities in which they engage and the kinds of institutions of which they are a part. (2003, p. 50)

Theoretical Background
Given this commitment to a holistic unit of analysis,
sociocultural researchers of learning examine human
activities in cultural contexts, mediated by language
and other symbol systems, and developing historically
through time. Many different kinds of contexts are
explored and considerable attention is given, in this
research, to the analysis of contextual specificity to
understand human activity-in-context.
The sociocultural approach reflects several key
themes (Wertsch 1985) articulated in Vygotsky’s
work: (a) the social sources of learning and development; (b) the central role of semiotic mediation in
human functioning; (c) the importance of genetic analysis. These three themes will be discussed first, followed
by several related topics.
Sociocultural–historical research views all learning
and development as occurring through the processes of

social interaction. Any individual development, in its
cognitive, social, and emotional aspects, results from
the learner’s interaction with others. Thus, the unit of
analysis must be social, rather than focused on the
individual. Vygotsky captured this idea in his “genetic
law of cultural development”:
"

Every function in the child’s cultural development
appears twice: first, on the social level, and later, on
the individual level; first, between people
(interpsychological), and then inside the child
(intrapsychological). This applies equally to voluntary
attention, to logical memory, and to the formation of
concepts. All the higher functions originate as actual
relations between human individuals. (Vygotsky 1978,
p. 57)

Sociocultural research has explored the workings of
social interaction for learning and development in
numerous ways, across a range of cultural contexts,
for learners of all ages, and in multifarious activities:
studies have analyzed the social sources of the earliest
language acquisition of infants in interaction with parents in Western professional families (Bruner 1983);
the formal and informal apprenticeships of
nonschooled children and adults in a variety of traditional and contemporary cultures (see Rogoff 1990);
the social sources of highly creative and productive
individuals in the arts, sciences, and humanities (see
John-Steiner 1985); the peer interaction and collaboration of children in and out of school settings (see
Cole 1996; Daniels 2007). As these examples suggest,
much of the research has been conducted in contexts of
naturally occurring activity (rather than experimentally contrived), in informal settings such as homes or
after-school programs, as well as formal learning in
schools, and in an impressive range of settings, rural
and urban, Western and non-Western. In any of this
work, a full examination of context is key. As Rogoff
writes, summarizing this view for child development,
"

[T]he individual child, social partners, and the cultural
milieu [are] inseparable contributors to the ongoing
activities in which child development takes place. The
concept of guided participation attempts to keep the
roles of the individual and the socio-cultural context in
focus. Instead of working as separate or interacting
forces, individual efforts, social interaction, and the
cultural context are inherently bound together in the

Sociocultural Research on Learning

overall development of children into skilled participants in society. (1990, pp. 18–19)

For Vygotsky, all thinking and all action are mediated, either by symbols or by tools, or both. Semiotic
mediation is the idea that thinking is made possible by
the use of symbolic tools, such as oral and written
language, in both individual and social functioning;
consciousness does not preexist but comes into existence through the use of semiotic tools of various kinds
(e.g., written language, musical notation, or the Cartesian coordinate system). In addition, it is the use of
these semiotic means that not only contributes to the
creation of internal reality, but also connects the external and the internal reality:
The use of notched sticks and knots, the beginnings of
writing and simple memory aids all demonstrate that
even at early stages of historical development humans
went beyond the limits of the psychological functions
given to them by nature and proceeded to a new culturally-elaborated organization of their behavior. . . .
Even such comparatively simple operations as tying
a knot or marking a stick as a reminder change the
psychological structure of the memory process
[extending] the operation of memory beyond the biological dimensions of the human nervous system . . . to
incorporate artificial, or self-generated, stimuli, which
we call signs. (1978, p. 39)

Just as physical tools are directed toward the external world, psychological tools are directed internally,
toward the solution of mental challenges. And just as
physical tools have been developed through culture and
historical time, in the context of specific cultural activities and practices (such as the handheld hoe used in
farming), so, too, psychological tools are the products
of sociocultural activities and appropriated by individuals in the context of those practices. Studies have
examined many processes of mediation, from the
impact of oral and written language systems or structures of questioning, to the effect of physical objects or
spatial arrangements on memory or calculation (see
Daniels 2007; Kozulin 1996). Such appropriation or
internalization of “sign-using activity in children is
neither simply invented nor passed down by adults”
but arises from a series of “qualitative transformations”
through a dialectical process (Vygotsky 1978, p. 46).
Much research has shown that the forms and functions

3133

of language and other sign systems used in instruction
affect learning in a range of qualitative and quantitative
ways and for all kinds of learners (see Daniels 2007;
Kozulin 1996; Wells 1999).
The process of transformation whereby the learner’s
participation in an activity shifts from dependence on
an adult or more capable peer to initiating his or her
own use of some mediator is examined through
“genetic analysis.” Genetic analysis refers to the genesis
of psychological functions, their origins and history
(and is not related to biological issues of heredity).
The interest in sociocultural–historical approaches is
primarily in the process, rather than the product, of
learning and development. Vygotsky wrote:
"

"

S

To study something historically means to study it in the
process of change; that is the dialectical method’s basic
demand. To encompass in research the process of
a given thing’s development in all its phases and
changes . . . fundamentally means to discover its
nature, its essence. . . . Behavior can be understood
only as the history of behavior. (1978, p. 65)

In the sociocultural approach, learning and development are most often examined in the social and
cultural contexts in which they are routinely produced.
This approach rejects any universal scheme for understanding learning or development. Rather, psychological systems that unite or separate various functions
vary across contexts, in terms of culture, history, and
other shared dimensions of human social life. Studies
of parents and children engaged in a task such as
puzzle-solving for the first time have demonstrated
the genesis of the child’s participation in the task,
from initial dependence on the adult for guidance
to the child as eventual initiator of independent problem-solving (Wertsch 1985). Studies of oral or written
private speech (or “speech for the self ”), for example,
have documented the use of semiotic mediators and
their transformative affects among school children
learning math, among adults learning a new language,
among creative thinkers in the process of producing
and capturing new thoughts (see John-Steiner 2007;
Wertsch 2007).

Important Scientific Research and
Open Questions
Vygotsky contrasted his view of learning and development with those of other theories. He saw learning and

S

3134

S

Sociocultural Research on Learning

development as both distinct and interdependent, and
he criticized Piaget’s view in which “maturation is
viewed as a precondition of learning but never as
a result of it” (1978, p. 80).
"

Learning awakens a variety of internal developmental
processes that are able to operate only when the child
is interacting with people in his environment and in
cooperation with his peers. . . . Learning is not development; however, properly organized learning results
in mental development and sets in motion a variety of
developmental processes that would be impossible
apart from learning. Thus learning is a necessary and
universal aspect of the process of developing culturally
organized, specifically human, psychological functions.
(1978, p. 90)

In exploring the difference between learning and
development, Vygotsky identified the “zone of proximal development,” defined as “the distance between the
actual developmental level as determined through
independent problem solving and the level of potential
development as determined through problem-solving
under adult guidance or in collaboration with more
capable peers” (1978, p. 86). The zone of proximal
development (“zpd”) captures the dynamism of the
process by which the learner moves from dependence
or interdependence to independence, through which
regulation by another becomes regulation by the self,
and the activity that could only be accomplished with
another can now be accomplished alone and the function may be considered to be learned, appropriated, or
“internalized.”
However, the concept of “internalization” is controversial among sociocultural scholars: one critique
finds the concept of internalization behavioristic in
the sense of a mere copying. Yet the process of internalization may also be seen as beyond simple transmission or copying from one person to another. Rather, the
process is a complex and dynamic one, and “while
children grow into the life of those around them”
(1978, p. 88), the child’s life context and life experience
are always different from the adult’s, so that what is
taken in or taken up by the child undergoes transformation and is not merely copied. The zpd includes not
only other people, but also tools and other artifacts of
existence, and the dynamism and infinite variation in
life experiences, in this view, preclude any simple transmission or copying.

A second critique of “internalization” is that it
is mentalistic. Critics who take this view prefer to
emphasize the outward aspect of change – what the
learner can do. In contrast, to accept the idea of internalization may be to see it as a unification of the
material and ideal, the outer reality with a no-less-real
subjective reality.
School learning is an important source of development. Vygotsky argued that “the only good learning is
in advance of development” (1978, p. 89). Because he
viewed development, in the sense of transformation of
functions, to be the goal of schooling, rather than mere
learning of material, Vygotsky was particularly interested in the way that academic (or “scientific”) concepts differ from everyday (or “spontaneous”) concepts
(Vygotsky 1978, 1986). Two key differences distinguish
the two kinds of concepts. First, academic or scientific
concepts are organized in a system of relations, comprising levels of abstraction and principles of order (of
vertical and horizontal relations) and generality,
whereas everyday concepts are not ordered in
a system. Second, academic concepts are deliberately
and explicitly taught and connected through instruction with other concepts, whereas everyday concepts
tend to be learned simply through exposure in everyday
life. To memorize biological taxonomy might show
learning, but development is shown by the transformation in thinking when one comes to use the structure
and relations of the taxonomic system as a “tool” for
taxonomic categorization.
Often in school learning, there is interference of
everyday concepts in the learning of academic ones,
or academic knowledge is learned as inert facts. But in
the ideal case, the two forms of knowledge inform each
other, with academic concepts reordering the structure
of the more concrete everyday knowledge and everyday
concepts giving specificity and meaning to the abstract
concepts. As suggested above, the academic knowledge
provides tools for thinking, as when the periodic table
or the taxonomic categories of biology are used as
problem-solving devices in the solution of scientific
problems.
Another controversial issue in sociocultural research
is the place of constructivism or social constructivism in
the framework. The emphasis in the framework on
social interaction, the interpsychological process, and
collaborative activity in the zpd, may lead some to see
the sociocultural framework as “social constructivist”

Socio-Cultural-Historical Research

and to favor “guided discovery” kinds of pedagogy.
However, many would consider this to be
a misreading of epistemological assumptions: pedagogy that relies on learners to discover scientific or
academic concepts has the consequence of focusing
on the empirical learning associated with spontaneous
or everyday concepts, rather than on the theoretical
teaching that aims to develop scientific and academic
concepts. Academic concepts are organized on abstract
principles that cannot be readily “discovered” by
learners, even with assistance, but need to be part of
a curriculum carefully designed to teach those abstract
principles in a way that also forges connections with
everyday knowledge of the world. This view recognizes
that such concepts cannot be taught directly, but that
extensive work is needed to develop curricula and
pedagogy effective in theoretical learning and teaching.
A growing body of research is devoted to exploring
theoretical and empirical teaching and learning in
a variety of academic subjects (see Daniels 2007;
Kozulin 1996).
Sociocultural–historical research has roots in
the revolutionary thinking of an earlier time but
remains a fertile source of ideas for the current age.
It challenges other theories, some time-honored,
others new, but shares a commitment to theory and
education in the service of human liberation and
societal advance.

S

3135

John-Steiner, V. (1985). Notebooks of the mind. Albuquerque: University of New Mexico Press.
John-Steiner, V. (2007). Vygotsky on thinking and speaking. In H.
Daniels, M. Cole, & J. Wertsch (Eds.), The Cambridge companion
to Vygotsky (pp. 136–152). Cambridge: Cambridge University
Press.
Kozulin, A. (1996). Psychological tools. Cambridge, MA: Harvard
University Press.
Rogoff, B. (1990). Apprenticeship in thinking: Cognitive development
in social context. New York: Oxford University Press.
Rogoff, B. (2003). The cultural nature of human development. Oxford:
Oxford University Press.
Wells, G. (1999). Dialogic inquiry: Toward a sociocultural practice and
theory of education. Cambridge: Cambridge University Press.
Wertsch, J. (1985). Social formation of mind. Cambridge: Cambridge
University Press.
Wertsch, J. (2007). Mediation. In H. Daniels, M. Cole, & J. Wertsch
(Eds.), The Cambridge companion to Vygotsky (pp. 178–192).
Cambridge: Cambridge University Press.
Primary Works
Vygotsky, L. S. (1978). In M. Cole, V. John-Steiner, S. Scribner, &
E. Souberman (Eds.), Mind in society (pp. 1–133). Cambridge,
MA: Harvard University Press.
Vygotsky, L. S. (1962). In E. Hanfmann & G. Vakar (Eds. & Trans.),
Thought and language (pp. 1–153). Cambridge, MA: MIT Press.
Vygotsky, L. S. (1986). In A. Kozulin (Ed. & Trans.), Thought and
language (pp. 1–256). Cambridge, MA: MIT Press.
Vygotsky, L. S. (1987). Thinking and speech. In R. W. Rieber & A. S.
Carton, (Eds.), N. Minick (Trans.), Collected works, (Vol. 1,
pp. 39–285). New York: Plenum.

Cross-References
▶ Activity Theories of Learning
▶ Collaborative Learning
▶ Communities of Practice
▶ Cultural–Historical Theory of Development
▶ Learning in the Social Context
▶ Mediated Learning and Cognitive Modifiability
▶ Scaffolding Learning
▶ Situated Cognition
▶ Socio-Constructivist Models of Learning
▶ Zone of Proximal Development

Socio-cultural Theory
Theoretical perspective that focuses on the roles of
social processes and culture in promoting learning
and cognitive development. A research approach drawing largely upon the work of Vygotsky, in which higherorder thought is thought to emerge from social interaction. This approach investigates how knowledge is
transmitted within a cultural context, and it emphasizes the role of material and symbolic artifacts.

References
Bruner, J. (1983). Child’s talk. Cambridge, MA: Harvard University
Press.
Cole, M. (1996). Cultural psychology: A once and future discipline.
Cambridge, MA: Harvard University Press.
Daniels, H. (2007). Pedagogy. In H. Daniels, M. Cole, & J. Wertsch
(Eds.), The Cambridge companion to Vygotsky (pp. 307–331).
Cambridge: Cambridge University Press.

Socio-Cultural-Historical
Research
▶ Sociocultural Research on Learning

S

3136

S

Socio-economic Status

Socio-economic Status
This is a composite measure of the social and economic
status of an individual’s family of origin. It combines
three concepts to measure overall socioeconomic background: (a) educational attainment of the parent(s),
(b) family income, and (c) social prestige of the job
held by the parent(s).

Socio-Economic Status of
Family
▶ Family Background and Effects on Learning

Socioemotional and Academic
Adjustment Among Children
with Learning Disorders
MICHAL AL-YAGON1, MALKA MARGALIT1,2
1
School of Education, Tel-Aviv University,
Tel-Aviv, Israel
2
School of Social Studies, Peres Academic Center,
Rechovot, Israel

Theoretical Background
Learning disorders (LD), as defined above, are one of
the most common childhood disorders, occurring in
approximately 2–10% of children and adolescents,
depending on the nature of the definitions applied
(American Psychiatric Association 2000). Most studies
investigating children with LD have demonstrated
their higher levels of academic challenges and their
socioemotional and behavioral difficulties (Al-Yagon
2007; Bryan 1997; Culbertson 1998; Margalit 2010).
Following diagnostic evaluation (often including psycho-educational, neurological, and psychiatric components), these children may be eligible for appropriate
testing accommodations, specific interventions, and
support from the school psycho-educational staff and
from external resources.
General consensus from large numbers of studies
suggests that children with LD exhibit heterogeneous,
sometimes overlapping difficulties in three major academic domains – dyslexia, dysgraphia, and dyscalculia
– and in the socioemotional domain. These studies also
indicated the different cognitive, perceptual, linguistic,
and neuropsychological processes that underlie these
different domains of LD (American Psychiatric Association 2000; Fletcher et al. 2007).
The next sections provide an overview of each
of these major domains of difficulty, in terms of
their features, underling processing deficits, and major
aspects regarding instruction and remediation processes.

Academic Adjustment of Children
with LD
Synonyms
Dyscalculia; Dysgraphia; Dyslexia; Learning disabilities

Definition
As suggested by the DSM-IV-TR (American Psychiatric
Association 2000), children with learning disorders
manifest an average IQ level and evidence substantially
lower achievements on standardized tests (in reading,
writing, and/or mathematics) than expected for age,
schooling, and level of intelligence. Their learning
problems significantly interfere with academic achievement or activities of daily living that require reading,
writing, and mathematics skills. If a sensory deficit is
present, the learning difficulties must be in excess of
those usually associated with the deficit. Learning disorders may persist into adulthood.

Research studies examining the academic adjustment
of children with LD pinpointed these children’s lower
school achievement levels, greater academic performance difficulties, and poorer learning skills compared
to children with typical development. The school dropout rate for children and adolescents with LD is
reported at nearly 40%, approximately 1.5 times the
average (American Psychiatric Association 2000).
Findings from follow-up investigations suggest that
these academic difficulties persist into adulthood;
lower rates of adults with LD attend postsecondary
school and lower rates graduate, compared to non-LD
adults.
Much of the research on academic adjustment in
children with LD has focused on dyslexia. These studies
highlighted the underlying reading difficulties and the

Socioemotional and Academic Adjustment Among Children with Learning Disorders

intervention approaches most effective in addressing
the central areas of reading deficit (see Toste 2007 for
a review). Children with dyslexia manifest a variety of
dysfunctional reading performance skills in two major
areas: decoding skills and/or comprehension skills
(Fletcher et al. 2007; Toste 2007). In examining the
decoding problems of children with dyslexia (e.g.,
slow reading speed, poor reading fluency, and mispronunciation of words in oral reading), studies examined
the possible role of deficits in phonological processing
and other cognitive processes like working memory
deficits, rapid automatized naming deficits, and visualperceptual deficits. Children with poor primary comprehension skills have difficulties in comprehending the
content of the material and may also have linguistic
deficits involving the semantic processing of written
language.
Data on these processes underlying reading difficulties have led to diverse intervention studies examining the best practices in reading instruction (see Toste
2007 for a review). For example, recent studies indicated that phonics instruction enhanced reading acquisition among younger as well as older readers. Such
research also emphasized that reading instruction and
remediation is most effective when highly explicit and
intensive, offered in a small and interactive group format, set up in a way that controls task difficulty, and
offers instruction in basic elements of reading as well as
in metacognitive strategies.
Children with dyscalculia manifest a developmental
arithmetic disorder in one or more of the basic skills
involved in mathematics such as deficits in counting,
computational skills, ability to solve word problems,
and understanding numbers (e.g., American Psychiatric Association 2000; Fletcher et al. 2007). Research
studies suggested several explanations for these difficulties, such as the role of sensorimotor processes
in the development of mathematic disorders and
the contribution of other cognitive processes such as
verbal and nonverbal neurocognitive processes. Such
neurocognitive processes may include, for example,
verbal-auditory discrimination ability, long-term memory for general information, and visuospatial processes.
In exploring the efficacy of mathematical instruction and remediation, researchers have highlighted the
importance of diverse techniques. For example, several
studies underscored the contribution of teaching the
relevant links between different kinds of problems,

S

mathematical procedures, and real world applications.
In addition, studies also emphasized that children with
dyscalculia may benefit from the explicit teaching of selfregulation, self-monitoring, and procedural knowledge
(i.e., understanding the sequential set of steps required
for solving a problem). Such studies also suggested the
role of cognitive strategies that incorporate cueing,
modeling, verbal rehearsal, and explicit instruction for
more effective mathematical instruction.
Children with dysgraphia demonstrate difficulties
in several major aspects of written expression skills
(American Psychiatric Association 2000; Fletcher et al.
2007) such as knowledge of rules for spelling and
grammatical usages as well as difficulties in handwriting, completing written assignments, and performing
writing compositions (e.g., organization of text and
sentence construction). In investigating the possible
sources of these written expression disorders, studies
have proposed variables such as disordered receptive or
expressive language skills and/or neuropsychological
processing (e.g., visuospatial, executive-coordination,
and dyspraxia – fine motor and graphomotor abilities).
Research studies focusing on effective instruction,
intervention, and remediation programs for children
with dysgraphia lag far behind those focusing on dyslexia and dyscalculia. However, the relatively few available studies did reveal the role of intervention methods
such as strategic training, rule-based lexical information, using lexical visual information of whole-word
forms, occupational therapy, and using diverse technologies to facilitate intensive repetitive practice.
Moreover, beyond this differentiation into three
major domains of academic difficulties, some researchers
have pinpointed a subgroup of nonverbal learning disorders, which comprise neurocognitive and adaptive
deficits manifested in tactile perception, visual perception, psychomotor skills, limited exploratory behaviors,
understanding/using the functional dimension of language, and so on. Children with nonverbal LD often
demonstrate strong reading and spelling difficulties,
weak reading comprehension, and poor arithmetic skills.

Socioemotional Adjustment of
Children with LD
Beyond documenting the effects of LD on academic
functioning, research has also indicated these children’s
diverse socioemotional difficulties (Al-Yagon 2007;
Bryan 1997; Margalit 2010). Data from numerous

3137

S

3138

S

Socioemotional and Academic Adjustment Among Children with Learning Disorders

studies have underscored that children with LD evidence a wide range of socioemotional difficulties such
as poor social competence, poor pragmatic communication abilities, impairment in adaptation to novel
situations, peer rejection and loneliness, difficulties
establishing and maintaining satisfying social relationships, higher depression and anxiety, more withdrawn
behaviors, and a lower prevalence of secure attachment
with parents and teachers, all in comparison to typically developing children (e.g., Al-Yagon 2007; Bryan
1997; Margalit 2010).
However, it should be noted that not all of the
children with LD experience a high level of
socioemotional difficulties. Thus, in line with resilience
approaches (Margalit 2010), research has also identified subgroups of children with isolated academic
problems alongside well-adjusted social and emotional
functioning.
Examination of the possible explanations for the
socioemotional difficulties of children with LD have
focused on two major sources: the central processing
deficit and the secondary emotional reaction. The first
approach has pinpointed the role played by these children’s individual-level characteristics such as internal
neurological factors, including information-processing
difficulties, impulsivity, and performance and production deficits (Bryan 1997). Such factors have been
examined for their effect not only on these children’s
academic skills, but also on their perceptions and interpretations of feelings and social situations, which, in
turn, may impair their social, emotional, and behavioral skills (Margalit 2010). In contrast, the second
approach assumes that these socioemotional difficulties manifested by children with LD comprise secondary emotional reactions related to these children’s
primary learning and academic disorders. In this case,
their difficulties may be seen as products of these children’s repeated experiences with frustration, academic
failures, and high levels of stress.
Intervention studies designed to improve the
socioemotional and behavioral functioning of children
with LD generally include individual/family psychotherapy or social skills training, spanning a wide range
of intervention durations, settings, and techniques
(Al-Yagon 2007; Margalit 2010). For example, social
skills training programs may consist of cognitive
behavior modification or metacognition training such
as coaching, modeling, role-playing, feedback, and

mnemonic strategies to train children in efficient
interpersonal problem-solving skills. Recent studies
also suggested the possible role of intervention programs focusing on these children’s close relationships
with significant others such as parents and teachers
(Al-Yagon 2007).

Important Scientific Research and
Open Questions
The study of LD has grown substantially over the last
decades, bringing major developments in understanding this disorder’s genetic, neuropsychological, and
other etiologies, as well as its features, diagnosis, and
intervention approaches. Nevertheless, some important questions regarding academic and socioemotional
adjustment of children with LD require additional
empirical exploration. First, data from many LD studies demonstrated the high prevalence of comorbid disorders such as attention-deficit/hyperactivity disorder
(ADHD), conduct disorder, and major depression disorder (e.g., American Psychiatric Association 2000).
For example, the high comorbidity found between LD
and ADHD suggests a possible shared etiology. It is
beyond the scope of the current discussion to provide
a thorough review regarding these groups of children
with comorbid disorders; however, these issues must
be acknowledged and pursued in future research
endeavors.
Second, despite growing awareness among educators and clinicians alike regarding the importance of
exploring the clinical features and the underlying
processing deficits for the socioemotional functioning
of children with LD, research studies addressing these
difficulties lag far behind those addressing academic
difficulties, thus calling for additional exploration. For
example, relatively few LD researchers have investigated
the extent to which socioemotional and behavioral
difficulties extend into adolescence and adulthood,
despite growing awareness about the importance of
social contexts and close relations with significant
others to adjustment throughout life. Furthermore,
considering the possible shared etiology for children’s
academic and socioemotional disorders, further
research should explore the prevalence and types of
socioemotional difficulties associated with each of the
different major academic LD domains described above.
Third, diagnostic evaluation procedures for LD
have been widely questioned and criticized; therefore,

Socio-emotional Aspects of Learning

future research would do well to continue searching for
comprehensive criteria and methods such as the use of
the response to instruction/intervention (RTI) model,
for diagnosing heterogeneous domains of LD.

Cross-References
▶ Abilities to Learn: Cognitive Abilities
▶ Anxiety Disorders in People with Learning
Disabilities
▶ Dyscalculia in Young Children – Cognitive and
Neurological Bases
▶ Human Information Processing
▶ Individual Differences in Learning
▶ Language-Based Learning Disabilities/Impairments
▶ Mathematical Learning Disability

References
Al-Yagon, M. (2007). Socioemotional and behavioral adjustment
among school-age children with learning disabilities: The moderating role of maternal personal resources. Journal of Special
Education, 40, 205–217.
American Psychiatric Association. (2000). Diagnostic and statistical
manual of mental disorders (text revision). Washington, DC:
Author.
Bryan, T. (1997). Assessing the personal and social status of students
with learning disabilities. Learning Disabilities Research & Practice, 12, 63–76.
Culbertson, J. L. (1998). Learning disabilities. In T. H. Ollendick, &
M. Hersen (Eds.). Handbook of child psychopathology
(pp. 117–156). New York: Plenum Press.
Fletcher, J. M., Lyon, G. R., Fuchs, L. S., & Barnes, M. A. (2007).
Learning disabilities: From identification to intervention. New
York: Guilford.
Margalit, M. (2010). Lonely children and adolescents: Self-perceptions,
social exclusion and hope. New York: Springer.
Toste, J. R. (2007). Developments in reading research: Processes and
instruction approach. Thalamus, 25, 18–22.

Socio-emotional Aspects of
Learning
SANNA JÄRVELÄ
Department of Educational Sciences and Teacher
Education, University of Oulu, Oulu, Finland

Synonyms
Emotion; Motivation; Social interaction

S

3139

Definition
Human thought and action are considered as products
of a dynamic interplay among personal, behavioral,
and environmental influences. Therefore, the basic
assumption involved in the focus on socio-emotional
aspects of learning is that learning situations are not
purely cognitive situations but are also emotionally and
motivationally loaded and situated within a social
context.

Theoretical Background
Learning research has privileged cognitive processes,
particularly
the information-processing view.
Although information processing has generally
excluded the consideration of emotion and social relationships in cognition, recent research has confirmed
the interaction of cognition and emotion. Also, in the
past several decades, research has demonstrated that
emotional and social processes are intertwined with
their cognitive counterparts and must be considered if
one is to understand learning (Leary 2007).
In recent years, increased attention has been devoted
to the social basis of cognition, taking into consideration
how social processes affect learning and performance.
A key focus on a socially shared learning approach has
been to avoid the dichotomy of the individual versus
the group and, instead, to define the ways in which
group life depends on individual participation, while
individual life depends on the impact of groups (Levine
et al. 1993). It has become clear that while trying to
understand social learning, we consider an extremely
complex set of variables: cognitive, social, emotional,
motivational, and contextual variables, interacting
with one another in a systemic and dynamic manner.
The interrelationships of emotional, social, and cognitive processes form the core of individual competence
building. The ▶ affective reactions individuals have in
response to positive and negative outcomes in different
settings reflect an investment in both attaining and
avoiding competence (Elliot and Dweck 2005).
Many of the motivational concepts derive from
experimental psychology and originally have focused
on the self, while the contextual aspects have been
in the background. Today, researchers are interested
how ▶ motivation is influenced, developed, and constructed within learning contexts. They also focus on
reciprocal relationships among individual motivation and peers, learning environment, and culture

S

3140

S

Sociohistorical Psychology

(Järvelä et al. 2010). Conceptualizing motivation in
learning contexts builds upon the situated learning paradigm, viewing the process of learning as distributed
across the learner and the environment in which knowing occurs, as well as the activity in which the learner
is participating. This conceptual framework provides
a useful foundation for understanding students’ goals,
intentions, and emotions across situations – in real
contexts and real time, and the context-person mutual
influences have been highlighted.
Students’ feelings about learning, the experience of
emotions, the level of emotional arousal, classroom
structure, and emotional atmosphere within the classroom, all contribute to their motivation to learn
(Meyer and Turner 2002). Many researchers have
tended to discuss emotions as intrapsychological phenomena. They are defined as subjective experiences
having a positive or negative quality. Yet, an argument
can be made that emotions are social phenomena
involving social experiences. For example, love and
sadness can be considered social rather than personal
emotions, with their antecedent or indicators found at
the social level, outside of a particular person. Other
emotions also arise in social contexts and, as regulators
of behavior, have social consequences (Weiner 2007).
In sum, individual emotions and motivation are
formed at the interface of personal, contextual, and
social aspects. Finally, in successful learning, emotions
and motivation are essential aspects of constructive
self-regulated learning.

Important Scientific Research and
Open Questions
The development of socio-emotional strengths will
become increasingly important in a rapidly changing
society, which demands coping with multiple challenges,
stressful situations, and competing goals. Therefore,
understanding increasingly dynamic social learning
environments will challenge the research field in terms
of conceptual and methodological development.
Methodological development has advanced the
research field. Researchers in process-oriented qualitative studies have targeted more intense and fine-grained
relations of emotions and other aspects of learning
process (Meyer and Turner 2002) in a variety of learning contexts, ranging from classroom interaction studies to technology-enhanced learning contexts (Järvelä
et al. 2010).

To conclude, research on socio-emotional aspects
of learning has taken new avenues in conceptual and
methodological development in order to grasp the
dynamics of motivation and emotions in multiple contexts and, thus, move closer to actual practices. There
are, however, still challenges regarding conceptual clarity and the generation of rigorous empirical designs in
field research to study motivation and emotions across
contexts and over time.

Cross-References
▶ Affective Dimensions of Learning
▶ Emotion Regulation
▶ Emotions and Learning
▶ Motivation and Learning: Modern Theories
▶ Motivational Variables in Learning
▶ Social Interaction Dynamics Supporting Learning

References
Elliot, A. J., & Dweck, C. S. (2005). Competence and motivation:
Competence as the core of achievement motivation. In A. J.
Elliot & C. S. Dweck (Eds.), Handbook of competence and motivation (pp. 3–12). New York: Guilford.
Järvelä, S., Volet, S., & Järvenoja, H. (2010). Research on motivation
in collaborative learning: Moving beyond the cognitive-situative
divide and combining individual and social processes. Educational Psychologist, 45(1), 15–27.
Leary, M. R. (2007). Motivational and emotional aspects of the self.
Annual Review of Psychology, 58, 317–344.
Levine, J. M., Resnick, L. B., & Higgins, E. T. (1993). Social foundations of cognition. Annual Review of Psychology, 44, 585–612.
Meyer, D. K., & Turner, J. C. (2002). Discovering emotion in classroom motivation research. Educational Psychologist, 37, 107–114.
Weiner, B. (2007). Examining emotional diversity in the classroom:
An attribution theorist considers the moral emotions. In P. A.
Schutz & R. Pekrun (Eds.), Emotion in education (pp. 75–88).
New York: Elsevier.

Sociohistorical Psychology
▶ Cultural-Historical Theory of Development

Socio-Historical Research
▶ Sociocultural Research on Learning

Socio-technical Learning

Sociological Model
▶ Socio-Constructivist Models of Learning

Sociological Model of Learning
▶ Collective Development and the Learning Paradox

Socio-technical Learning
ISA JAHNKE
Department of Applied Educational Science,
Umeå University, Umeå, Sweden

Synonyms
Socio-technical learning communities; Technologyenhanced learning

Definition
Socio-technical learning is the process of research-based
online learning that combines individual and cooperative learning with opportunities to interact with other
community members online or face-to-face. The
approach focuses on socio-technical learning communities within higher education. The word socio-technical interrelates to technical systems as well as social
structures – human communication and learning
is integrated into a technical platform. A special case
of socio-technical learning is experimental online
learning.

Theoretical Background
Learning Paradigm
Socio-technical learning follows the constructivism
approach. It means learning processes are cognitive
constructed and socially framed. Learning is defined
as a proactive process of constructing rather than
acquiring knowledge. Individuals create sense of their
own world. Everything they come in contact with is

S

3141

constructed by their own models of their experience.
Hence, learning is not defined as simply the transmission of data from one individual to another, but a social
process whereby knowledge is co-constructed in
a situation within a community of practice (cf. Lave
and Wenger 1991). Teaching or instructions have the
task to support and scaffold (giving structures) this
construction rather than communicating knowledge.
Current discussions in higher education focus on
shifting the focus from the teacher’s teaching to the
student’s learning. Promoting concepts for the shift
from teacher-centered teaching to student-centered
learning concepts are not new; however, discussions
about pedagogical learning approaches got a new
drive as new community platforms based on Web 2.0
technologies emerged, for instance, platforms for usergenerated content like wikis, blogs, and social networking platforms like Facebook or Myspace (Jahnke 2009).
The socio-technical approach has the claim to support
teaching and learning differently. It says that a new
balance between teaching and learning is essential for
supporting creativity and best learning effects. Learning-centered approaches promote a re-orchestration of
teaching and learning – information-generating, pushing-and-pulling arrangements for learners – where
learning is regarded from the viewpoint of the learners.

Exploratory and Research-Based
Learning: Foundation for
Socio-technical Learning
Exploratory learning is an active process in which
a learner constructs his own meaning based on his
own experience. This means learners explore something (e.g., artifacts, hypotheses, ideas, and results)
without having or giving a solution by the teachers.
Learners “interact with the world by exploring and
manipulating objects, wrestling with questions and
controversies, or performing experiments” (Bruner
1961). However, exploratory learning does not mean
unguided learning (Kirschner et al. 2006). Exploratory
learning concepts (also known as discovery learning)
encourage the learner to do experiments and to
uncover relationships. Learners get the opportunity to
discover unknown and unexpected object properties,
characteristics, and theoretical models by following
various learning paths. Exploratory learning often
follows Kolb’s “experiential learning theory” (Kolb
and Boyatzis 2000) covering four steps: concrete

S

3142

S

Socio-technical Learning

experiences (being involved in a situation, doing something), active experimenting (testing a theory by making a plan and following it), reflective observing
(looking at an experience and thinking about it), and
abstract concept-making (forming theories about why
an experience happened the way it did). A pedagogical
approach which includes appropriate structures for the
teaching and learning process is called research-based
learning (Jenkins et al. 2003) where students undertake
research and inquiry. Teaching and learning is structured by the process of research phases (building
hypothesis, delivering theoretical framework, making
research design, doing inquiry, describing results, making conclusion).
A special case of socio-technical learning is
experimental learning. It is defined as combined
forms of research-based and experiential learning
that take place within remote laboratories using an
online learning platform with an Internet-based
access.

Socio-technical Learning in the Age
of Web 2.0
In a former typical one-room schoolhouse 100 years
ago, “learning was social, not didactic,” writes John
Seely Brown. To foster learning as social process, one
approach focuses on learning communities of practices. In words with Digital Natives, TechnologyEnhanced Learning support social learning by using
new media like Social Networking, Forums, or Blogs.
Such Web 2.0 platforms offer new possibilities to easily
enable social learning in groups (e.g., Jahnke and Koch
2009). The availability of web access from anywhere at
any time has made it easier to engage students in learning communities and can also link weakly coupled
learners. In the Web 2.0 age, some academic staff developers stress that socio-technical learning scenarios in
higher education need more attractive concepts (Collins
and Halverson 2009), for example, concepts that supports problem-solving without having any standard
solutions by using Web 2.0 platforms or socio-technical
learning communities.

Socio-technical Learning
Communities
Socio-technical learning communities are forms of
communities of practice – introduced by Lave and

Wenger (1991) as well as Wenger et al. (2002). They
are generated through social relationships among individuals “who share a concern, a set of problems, or a
passion about a topic, and who deepen their knowledge
and expertise in this area by interacting on an ongoing
basis” (Wenger et al. 2002, p. 4). With Preece (2000),
socio-technical communities differ in the following
four areas:
● Group size (e.g., in research on communities,

groups with 25 members are considered small,
whereas groups with 1,700 are considered very
large)
● Primary content (e.g., discussion boards about
Harry Potter books and movies, discussions about
sports like marathon training, communication
about stock exchanges, and information sharing
about lectures at a university)
● Life span (e.g., several years or just for one topic)
● Presence (e.g., either pure online communication,
face-to-face, or mixed communication)
For the design of socio-technical learning, the analysis of the appropriate interplay between social and
technical parts is needed. On the one hand, sociotechnical learning communities consist of actors who
use technical systems to communicate and share
knowledge. On the other hand, the technical system
influences the interaction between community members (human–computer interaction).

Social Structures for Learning
In contrast to work groups in companies where the
group members are formally bound, socio-technical
communities consist of more informal than formal
connections between members. Formal structures are
characterized by conventional forms of behavior, and
established conventions, for example, behavior which
is formally bound by a work contract, or a formal role
represented by a job/task description (e.g., formal
moderator). Informal structures are rather casual,
unofficial, loose, and not triggered by any rules (e.g.,
activities of informal moderation).
Social structures are patterns or interrelationships
of social elements (e.g., human behavior and relationships within socio-technical communities) that can
be called “roles.” To observe the shape of roles in
an online community, observable categories are

Socio-Technical Learning Communities

S

needed. According to Jahnke’s role model (2009), four
categories are useful for analyzing and designing sociotechnical learning processes:

Humanities, Social Sciences) look like? How can we
measure the success, effect, and impact of sociotechnical learning models?

(a) Learner’s position within the community; relations
to other members. Questions for designing sociotechnical learning processes are how to bring the
learners from outside to the middle of the core
members and what methods are helpful for
teachers to support this process.
(b) Learner’s tasks/activities within the learning
process. Questions for designing socio-technical
learning processes are how to support different
activities.
(c) Tacit, implicit, and explicit expectations of
learners. Questions for designing socio-technical
learning are how to support conflicting expectations or problems of learners within the researchbased learning process.
(d) Interactions/role-playing (e.g., problem that students do not regarded themselves as researchers).
Questions for designing socio-technical learning
are how to give a structure for learners by having
enough room for move, how to support rolechanging, and what methods are useful.

Cross-References

Important Scientific Research and
Open Questions
Based on mentioned theoretical background, a sociotechnical learning model has the following dimensions:
● Social design for socio-technical learning (e.g., com-

munication, different social modes, cooperation)
● Technical design (e.g., Web 2.0, technical platforms,

usability)
● Pedagogical design (e.g., model which guided

exploratory, research-mode learning)
and an appropriate interplay of all three dimensions.
The guided questions for designing are: what
socio-technical design for research-based learning is
needed? Derived questions are: what is an appropriate
balance between teaching objects and learning activities in socio-technical environments, how to make
learner-centered learning, or in other words, what is
an attractive learning model from the student’s perspective? What does an attractive exploratory,
research-based learning model in higher education
in special cases (e.g., Faculties of Engineering,

3143

▶ Communities of Practice
▶ Computer-Based Learning
▶ Computer-Supported
Collaborative
Learning
(CSCL)
▶ e-Learning and Digital Learning
▶ Online Learning
▶ Social Networks Analysis and the Learning Sciences

References
Bruner, J. S. (1961). The act of discovery. Harvard Educational Review,
31(1), 21–32.
Collins, A., & Halverson, R. (2009). Rethinking education in the age of
technology: The digital revolution and schooling in America. New
York: Teachers College Press.
Jahnke, I. (2009). Socio-technical communities: From informal to
formal? In B. Whitworth & A. de Moor (Eds.), Handbook of
research on socio-technical design and social networking systems
(pp. 763–778). Hershey: Information Science Reference, IGI
Global. Chapter L.
Jahnke, I., & Koch, M. (2009). Web 2.0 goes academia: Does Web 2.0
make a difference? International Journal of Web Based Communities, 5(4), 484–500.
Jenkins, A., Breen, R., Lindsay, R., & Brew, A. (2003). Re-shaping
higher education: Linking teaching and research. London:
Routledge Falmer/SEDA.
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal
guidance during instruction does not work an analysis of the
failure of constructivist, discovery, problem-based, experiential,
and inquiry-based teaching. Educational Pschologist, 41(2),
75–86.
Kolb, D., & Boyatzis, R. (2000). Experiential learning theory: Previous
research and new directions. In R. J. Sternberg & L. F. Zhang
(Eds.), Perspectives on cognitive, learning, and thinking styles.
Mahwah: Lawrence Erlbaum.
Lave, J., & Wenger, E. (1991). Situated learning: Legitimate peripheral
participation. New York: Cambridge University Press.
Preece, J. (2000). Online communities: Designing usability, supporting
sociability. Chichester: Wiley.
Wenger, E., McDermott, R., & Snyder, W. M. (2002). Cultivating
communities of practice: A guide to managing knowledge. Boston:
Harvard Business School Press.

Socio-Technical Learning
Communities
▶ Socio-technical Learning

S

3144

S

Socio-technological Change of Learning Conditions

Socio-technological Change of
Learning Conditions
INES LANGEMEYER
InterMedia, Faculty of Education, University of Oslo,
OSLO, Norway

Synonyms
Changed conditions for learning

Definition
The idea of investigating socio-technological changes
of learning conditions was enunciated by historicalmaterialist approaches to learning such as the Cultural-historical School (Vygotsky, A. N. Leont’ev,
Luria and others, originated in the 1920s) and the
Berlin school of Critical Psychology (Holzkamp,
Osterkamp, Schurig, F. Haug and others, established
in the 1970s). This idea was nurtured by a particular
Hegelian or Marxian notion of learning defined as
productive human activity by which not only cognitive
schemas, behavior, and personality develop but also the
human potential (Gattungswesen) at a certain culturalhistorical level. From this angle, learning is seen as
a dimension of the entire sociocultural development
in which technological innovations play an important
role. Rather than in terms of their impact on learning,
socio-technological changes are explored as conditions
and demands to which humans actively and creatively
respond. Furthermore, these changes are recognized as
largely connected to the drive of capitalist relations to
increase productivity, to rationalize the mode and the
relations of production, and to augment the capitalist
control of the production forces. Therefore, these
changes are not interpreted as a linear progress to
higher levels of societal life, but studied as contradictory
developments under which learning is realized or
under which people are expected to learn.

Theoretical Background
Conceptualizing the socio-technological change of
learning conditions requires the reciprocal interpretation
of the following three dimensions: (1) the sociotechnological change, (2) the significance of societal conditions for learning, as well as (3) their change as driven
by a socio-technological one. Historical-materialist

approaches engage with these three dimensions by analyzing the shifts of the mode of production, the changing
division of labor, and the alternation of the theory–praxis
relation which is dominant at a certain period of time.
This three-dimensional perspective is uncommon
to many theories of learning and analyses of sociotechnological change. While learning is often conceptualized in mentalist terms regardless of its material
societal conditions, socio-technological change is
largely identified with either the gradual substitution
of manpower by technology and thus suspected of
diminishing the scope of human development; or it is
interpreted as a new kind of man–machine interaction
to which the labor force simply needs to adjust. Only
the use of computer and Internet technologies as
“tools” for information processing and knowledge production inspired researchers of other theoretical traditions to think of these advancements more in terms of
a new societal potential for learning activities.
The advantage of historical-materialist approaches
to learning and socio-technological change lies in their
capacity to connect those three dimensions, and
thereby seek to overcome individualist and mentalist
notions of learning as well as any reductionist technological determinism. The dialectical conception of
human praxis or “activity” is a key to this. In general,
technologies – including tools – are conceived of as social
forms of action rather than isolated material objects.
Furthermore, historical-materialist approaches conceptualize learning as a specific kind of “object-related activity” by which humans individually or collectively gain
control of their societal living conditions and life’s material production (Holzkamp 1993; Stetsenko 2005). In so
doing, these approaches do not outplay the objective
against the subjective side of learning. Not only is
“individual skill and knowledgeability” addressed as
a “currency of activity” (Sawchuk 2006, p. 608), also
the self is seen – within “a unique constellation of
activities in the real world” – as a “person’s ‘leading
activity’” (Stetsenko and Arievitch 2004, p. 477).
The fact, for example, that humans are making
tools for satisfying human needs is seen as the germ
cell of theoretical and experimental knowledge which is
to encounter even in natural religions and which forms
the basis of modern natural sciences. This is illuminated by the following explanation: Hence humans
would start making tools no longer spontaneously but
on purpose and collaboratively, this tool production

Socio-technological Change of Learning Conditions

would presuppose generalized knowledge of human
needs, theoretical insights of using certain tools in
a specific environment and of their impact on certain
material objects. By actively transforming the environment and by deliberately “building” more and more
invariant characteristics into it (which are already cognized), humans would create the material basis for
experimenting with unknown aspects of nature
(Holzkamp 1973, p. 126). Consequently, not only the
scope of epistemic progress of one generation also
processes of individual learning and personality development would depend on the cultural-historical level
of life’s material production.
Against
economic
reductionism,
sociotechnological changes are not identified with technological developments only. Shifts in the division of
labor and, correspondingly, alternations of the theory–praxis relation (cf. PAQ 1987) are considered as
well. Among numerous socio-technological changes,
historical-materialist approaches regard the one related
to automation-, information-, and communicationtechnologies as paramount, because it superseded the
radical division of intellectual and physical labor which
largely determined – until the 1970s – the prevailing
mode of rationalization.

Important Scientific Research and
Open Questions
The derogating effects of industrialization on learning
and human development can be studied in relation to
Frederick W. Taylor’s Principles of Scientific Management (1911) in an exemplary way. These principles
corresponded to the paradigm of mechanizing technologies such as the assembly line which – combined with
the great machinery – successfully facilitated mass production. Simultaneously, these socio-technological
conditions created a mass of unskilled workers on the
one hand and a small, techno-scientific elite of engineers and managers on the other. In so doing, they
represented a point of culmination for the radical division of manual and intellectual labor. Facing this “progressive division of labor,” Vygotsky (1930/1994,
pp. 176, 181) echoed Marx’s critique that “every new
level in the development of the production forces of the
society has not just failed to raise humanity as a whole
and each individual human personality to a higher
level, but has led to an ever deeper degradation of
the human personality and its growth potential.”

S

Therefore, he emphasized the necessity “to overcome the division between physical and intellectual
work and to reunite thinking and work which have
been torn asunder during the process of capitalist
development.” Vygotsky hoped for mitigation by
poly-technical education, but could not foresee any
socio-technological change which portended an end
of the de-intellectualization of manual labor and the
saturation of science and intellectual work with partial
political and private economic interests.
In the 1970s and 1980s, critical psychologists and
sociologists launched a research project (“Projekt
Automation und Qualifikation”) and scrutinized the
shifts of the mode of production at numerous sites of
automated production plants (PAQ 1987). Due to the
fact that manual and routinized labor was increasingly
replaced by automation technologies and that new
work activities emerged in the area of research and
development (cf. Bell 1973), programming, controlling, improving, optimizing, and inventing collaborative uses of these technologies, the research group
concluded that labor would undergo an intellectualization and be subjected to a new type of scientification
(Verwissenschaftlichung). This development would
break with the Taylorist domination of science/theory
over praxis since steps of planning, inventing, executing, and controlling would now merge into integral
work tasks. Neither the active involvement of humans
in this new labor process could now be completely
standardized, nor could workers be entirely controlled
by a hierarchical division of labor. Instead, the intellectual and collaborative capacity of problem solving and
inventing would come to the fore. The research group
PAQ therefore did no longer identify “scientification”
with the scientific constructing of production technologies and the standardization of work knowledge and
work routines only. It rather emphasized the challenge
to cope with information which is presented and
processed in a scientific manner. It stressed the requirements of analyzing errors and solving problems that
occur in automated production and work processes.
Thus, the theory–praxis relation would be subjected to
a radical change: Large parts of the relevant knowledge
for labor and production would emerge along with
activities of experimenting, analyzing, and learning
during the work process. Since science/theory could
no longer be separated from, and would no longer
“reign” over praxis, the former division of labor

3145

S

3146

S

Socio-technological Change of Learning Conditions

would have become obsolete. A glance of hope for
humanizing work and for democratizing the relations
of production appeared. However, a new arena for
conflict and contradiction has come forward with
respect to the issue of learning (cf. Langemeyer 2005;
Sawchuk 2006).
According to the needs of scientificated labor,
Langemeyer and Ohm (2009) see collaborative and
experimental forms of learning emerging which are
facilitated by (mobile) communication-, controlling-,
visualizing-, and simulation-technologies as well as
“intelligent” software agents and other computational
artifacts. However, since learning has become an integral part of intellectualized work activities, such technological devices are also invented to reassure the
capitalist control over labor, production, and productivity. This is achieved for example by new technologies
of (self-) surveillance.
Another issue is the digital form of information
goods, or more precisely, their unlimited duplicability.
Since software, e-books, e-journals, etc., are often made
for commercial purpose intellectual property rights,
copy protection and other barriers are created to ensure
that these products are purchasable. Such innovations
have however quite contradictory effects with regard to
the new scientification of labor. Different from open
software, commercial software entails an inaccessible
programming code and therefore is made unusable (or
very partially usuable), for instance, for programmers’
learning and experimenting activities.
Further difficulties arise with the augmenting complexity of computational systems, the new relations of
production, and the emerging global division of labor.
Since most of the computing systems used in the beginning of the second millennium can no longer be (re-)
produced and comprehended by an individual person
only, the scope of learning activities necessarily becomes
fragmentary and essentially dependent on better and
more sustainable forms of cooperation and organization. Globalized labor markets reinforce competitive
relations, and the global division of labor fosters not
just internationally but also regionally a new segregation
between a skilled and an unskilled workforce.
These socio-technological changes are not simply
creating or enhancing conditions of learning. They
simultaneously demand and limit, restrain or impair
learning at work. Thus, the question arises how professionals cope with scientificated labor. This issue also

leads to query other relevant societal conditions such as
education (system), work culture and social class (cf.
Sawchuk 2003), work contracts, gender relations, social
welfare, or globalization (cf. Livingstone et al. 2008)
which influence the dynamics of a new scientification
of labor and its challenges of learning.

Cross-References
▶ Activity Theories of Learning
▶ Advanced Learning Technologies
▶ Artificial Intelligence
▶ Automatic Information Processing
▶ Collaborative Knowledge Building
▶ Collaborative Learning
▶ Collaborative Learning and Critical Thinking
▶ Collaborative Learning Supported by Digital Media
▶ Collective Learning
▶ Complex Problem Solving
▶ Computer-Based Learning
▶ Computer-Based Learning Environments
▶ Conditions of Learning
▶ Contradictions in Expansive Learning
▶ Creativity, Problem Solving and Feeling
▶ Cultural-Historical Theory of Development
▶ Human–Computer Interaction and Learning
▶ Learning Activity
▶ Learning: A Process of Enculturation
▶ Problem-Based Learning
▶ Simulation-Based Learning
▶ Sociocultural Research on Learning

References
Bell, D. (1973). The coming of post-industrial society. A venture in social
forecasting. New York: Basic Books.
Holzkamp, K. (1973). Sinnliche Erkenntnis. Historischer Ursprung und
gesellschaftliche Funktion der Wahrnehmung. Frankfurt/M:
Athenäum.
Holzkamp, K. (1993). Lernen. Subjektwissenschaftliche Grundlegung.
Frankfurt/M: Campus.
Langemeyer, I. (2005). Contradictions in expansive learning: towards
a critical analysis of self-dependent forms of learning in relation
to contemporary socio-technological change [43 paragraphs].
Forum: Qualitative Social Research, 7(1), Art. 12.
Langemeyer, I., & Ohm, C. H. R. (2009). Verwissenschaftlichung
von Arbeit. In D. Dumbadze et al. (Eds.), Erkenntnis und
Kritik. Zeitgenössische Positionen (pp. 269–292). Bielefeld:
Transcript.
Livingstone, D., Mirchandani, K., & Sawchuk, P. H. (2008). The future
of lifelong learning and work: Critical perspectives. Rotterdam:
Sense Publishing.

Socratic Questioning
Projektgruppe Automation und Qualifikation. (1987). Widersprüche
der Automationsarbeit. Argument (PAQ): Berlin/W.
Sawchuk, P. H. (2003). Adult learning and technology in working-class
life. New York: Cambridge University Press.
Sawchuk, P. H. (2006). ‘Use-value’ and the re-thinking of skills,
learning and the labour process. The Journal of Industrial Relations, 48(5), 593–617.
Stetsenko, A. (2005). Activity as object-related: Resolving the dichotomy of individual and collective types of activity. Mind, Culture,
and Activity, 12(1), 70–88.
Stetsenko, A., & Arievitch, I. (2004). The self in cultural-historical
activity theory: Reclaiming the unity of social and individual
dimensions of human development. Theory & Psychology,
14(4), 475–503.
Vygotsky, L. S. (1930/1994). The socialist alteration of man. In R. van
der Veer & J. Valsiner (Eds.), The Vygotsky Reader (pp. 175–184).
London: Blackwell.

S

3147

Socratic Interrogation
▶ Socratic Questioning

Socratic Irony
▶ Socratic Questioning

Socratic Method
▶ Socratic Questioning

Socratic Cross-Examination
▶ Socratic Questioning

Socratic Practices
▶ Socratic Questioning

Socratic Disputation

Socratic Questioning

▶ Socratic Questioning
AYTAC GOGUS
Center for Individual and Academic Development
(CIAD), Sabanci University, Istanbul, Turkey

Socratic Education
▶ Socratic Questioning

Socratic Elenchus
▶ Socratic Questioning

Socratic Framework
▶ Socratic Questioning

Synonyms
Socratic cross-examination; Socratic disputation;
Socratic education; Socratic elenchus; Socratic framework; Socratic interrogation; Socratic irony; Socratic
method; Socratic practices; Socratic rhetoric; Socratic
teaching; Socratic techniques

Definition
Socratic Questioning is a dialectical method of inquiry
and debate by means of a carefully constructed series
of leading questions to arrive at logical responses and
to stimulate rational thinking. Socratic Questioning
involves the use of systematic questioning, inductive
reasoning, universal definitions, and a disavowal
of knowledge (Carey and Mullan 2004). Socratic

S

3148

S

Socratic Questioning

Questioning refers to a procedure in which people
attempt to change others’ minds, as well as a process
that allows people to change their own minds (Carey
and Mullan 2004). Socratic Questioning is defined as
“the dialectical method supposedly employed by the
historical Socrates, and displayed in Plato’s earlier
dialogues. . . the teacher should by patient questioning
bring the pupil to recognize some true conclusion,
without the teacher’s telling the pupil that that conclusion is true” (Carey and Mullan 2004, p. 221). Socratic
Questioning is used as a teaching method to facilitate
the process of learning through probing student thinking and reasoning in complex problems, and structuring a problem-solving process (Rhee 2007). Socratic
Questioning is a form of active learning pedagogy
that allows the learner to develop higher order thinking
skills such as analysis, synthesis, and evaluation.

Theoretical Background
Socratic Questioning comes from the teaching style of
Socrates (469–399 BC) who was a Greek philosopher
and one of the founders of Western Philosophy.
Socratic Questioning introduces a problem and directs
the conversation back to key points to allow students
discover the answers of the problem and the content
while avoiding lecturing. Ellerman et al. (2001) state
that Socratic Questioning or Socratic Method engages
students in interacting with the teacher and other
students:
"

The classical archetype for the teacher in the active
learning model was Socrates. Socrates could not “transmit” knowledge as from teacher to student because he
had what is now known as “Socratic ignorance”.
Instead of “disseminating knowledge”, Socrates
would engage his interlocutors in a dialogue about
the topic at hand. He would ask pointed questions to
show the shortcomings in the conventional wisdom
and thus he would try to catalyze his dialogical partners
into thinking for themselves to take an active role in
thinking through a question or problem rather than
just reciting the orthodox views poured into people by
the ambient society and its formative institutions. Socrates also helped demonstrate the fallacy of assuming
that people want to learn. Thus the Athenian citizens
whom Socrates pursued with his questions eventually
had him put to death for his troubles. Plato had more
success in communicating Socrates’ thinking than

Socrates himself, by using a combination of narrative
and abstract modes of communication in his dialogues.
(Ellerman et al. 2001, pp. 171–172)

The Socratic method uses a carefully constructed
series of leading questions to draw information out of
learners, rather than providing learners with the information or answers of problems. This one-on-one
engagement leads learners to arrive at logical responses
and to the point where knowledge, application, synthesis, and evaluation are integrated as a form of active
learning pedagogy (Rhee 2007).
Neenan (2009) states that Socratic Questioning is
a cornerstone of cognitive behavioral coaching for
guided discovery and cognitive behavior therapy that
highlights how changing a person’s thinking leads to
emotional and behavioral change. Socratic
Questioning raises awareness, promotes reflection,
and improves problem-solving thinking (Neenan
2009). According to Neenan (2009), the characteristics
of good Socratic questions are being concise, clear,
open, purposeful, constructive, focused, tentative, and
natural. By considering coach–coachee dialogues, the
characteristics of good Socratic questions are defined
by Neenan (2009) as follows:
● Concise – keeping the focus on the coachee
● Clear – reducing potential coachee confusion or

misunderstanding by avoiding prolixity or jargon
● Open – inviting participation and exploring ideas
● Purposeful – explain the reasons for the questions
●
●
●
●

[that coach is] asked
Constructive – promoting insight and action
Focused – on the coachee’s current concerns
Tentative – not assuming the coachee can answer
[the coach’s] question
Neutral – not signaling the answer which would
indicate your viewpoint (Neenan 2009, p. 252)

Neenan (2009) emphasizes that “asking good
Socratic questions is an essential skill for all coaches
to encourage coachees to reflect on their thinking and
actions in order to develop new problem-solving perspectives, improve performance, achieve goals and take
their lives in often unanticipated directions” (p. 263).
Beside the importance of the asking good questions,
teaching a skill, giving feedback, clarifying points,
pointing out perceived inconsistencies in the person’s
thinking, offering advice, and making suggestions are

Socratic Questioning

S

offered by Neenan (2009) as important techniques for
a successful practice of the Socratic Questioning.

● Keeping participants actively engaged in the

Important Scientific Research and
Open Questions

Rhee (2007) emphasizes that Socratic Questioning
is traditionally perceived as a teaching tool but it is also
a concrete analytic tool for the student. Rhee (2007)
states that the root of Socratic Questioning/Method is
grounded in philosophical inquiry followed by many
law teachers and also adapted to a scientific framework
for solving problems:

Paul and Elder (2007) focus on the forms of Socratic
Questioning that can be used in analyzing reasoning:
“the purpose of the reasoning, the questions being
asked, the information being used, the beliefs being
taken for granted or assumed, the points of view
embedded in the reasoning, the concepts guiding the
reasoning, the inferences being made, and the implications of the reasoning” (p. 32). In addition, Paul and
Elder (2008) state that there are universal intellectual
standards used for thinking by educated and reasonable
persons. The universal intellectual standards “include,
but are not limited to, clarity, precision, accuracy, relevance, depth, breadth, logicalness, and fairness” (Paul
and Elder 2008, p. 32). According to Paul and Elder
(2008), skilled thinkers explicitly use intellectual standards on a daily basis without being aware of these
standards.
Paul and Elder (2008) distinguish three general
categories of Socratic Questioning: spontaneous,
exploratory, and focused. Paul and Elder (2008) state
that all three questioning categories can cultivate student thinking:
"

All three types of Socratic discussion we have discussed
require instructors to become more skilled over time in
of the art of questioning. They require the instructor to
develop familiarity with a wide variety of intellectual
moves. The cultivation of critical thinking can be
enhanced by adapting spontaneous or unplanned,
exploratory, and focused modes of orientation and
applying the formal mechanics of Socratic Questioning
within each mode. (Paul and Elder 2008, p. 35)

Paul and Elder (2007) suggest that students can use
Socratic Questioning in group discussions. The benefits of using Socratic Questioning are summarized in
four categories (Paul and Elder 2007):
● Keeping participants focused on the elements of

thought
● Keeping participants focused on systems for

thought
● Keeping participants focused on standards for

thought

3149

discussion

"

Many years ago, the mathematician George Polya
showed generations of math teachers and students
that problem-solving can be learned through the heuristic of simple questions that stimulate curiosity and
creativity. There is a connection between the Socratic
Method and techniques of mathematical problemsolving. The Socratic Method facilitates this process of
learning. Classes can have the feel of a dialogue: the
right questions, level of student participation, challenges, and disagreements among students and professor. At times, however, the method becomes stale or
mechanical; an air of routinization can set in. But sometimes the dialogue seems to lack a grander design. It is
easy to see how a repeated diet of the Socratic Method
can lead to a feeling of a monochromatic routine. (Rhee
2007, p. 884)

According to Rhee (2007), students, who often fail
to connect the abstract concepts with the real-life uses,
may face to follow the mechanical, repetitive, and routine tasks of the Socratic Questioning. Therefore, the
Socratic Questioning should be complemented with
the techniques of problem-solving.

Cross-References
▶ Active Learning
▶ Discovery Learning
▶ Learning from Questions
▶ Plato (429–347 BCE)

References
Carey, T. A., & Mullan, R. J. (2004). What is Socratic questioning?
Psychotherapy: Theory, Research, Practice, Training, 41(3),
217–226.
Ellerman, D., Denning, S., & Hanna, N. (2001). Active learning and
development assistance. Journal of Knowledge Management, 5(2),
173–179.
Neenan, M. (2009). Using Socratic questioning in coaching. Journal
of Rational-Emotive & Cognitive-Behavior Therapy, 27(4),
249–264.

S

3150

S

Socratic Rhetoric

Paul, R., & Elder, L. (2007). Critical thinking: The art of Socratic
questioning, Part II. Journal of Developmental Education, 31(1),
36–37.
Paul, R., & Elder, L. (2008). Critical thinking: The art of Socratic
questioning, Part III. Journal of Developmental Education, 31(3),
36–37.
Rhee, R. J. (2007). The Socratic method and the mathematical heuristic of George Polya. St. John’s Law Review, 81, 881–898.

Socratic Rhetoric
▶ Socratic Questioning

Solo Taxonomy
▶ Deep Approaches to Learning in Higher Education

Song Learning and Sleep
MICHAEL LUSIGNAN, DANIEL MARGOLIASH
Department of Organismal Biology and Anatomy,
University of Chicago, Chicago, IL, USA

Synonyms
Sensorimotor learning; Sleep or offline learning; Song
learning behavior; Social learning and communication

Socratic Teaching
▶ Socratic Questioning

Socratic Techniques
▶ Socratic Questioning

Software for Human Learning
Research
▶ Web-Based Experiment Control for Research on
Human Learning

Solfège
The system of syllables traditionally used to represent
the steps of the musical scale: do (originally ut), re, mi,
fa, sol (or so), la, ti (or si). Solfège is often used in sightsinging, and has two approaches: (a) the Fixed-Do
system in which Do is always the note C, and (b) the
Moveable-Do system, in which Do is always the first
note of the scale, whatever the scale happens to be.

Definition
REM: rapid eye moment sleep
NREM: non-REM sleep
Wiener entropy: Wiener entropy describes the distribution of power in a sound. Much like Shannon entropy,
which indicates the degree of randomness in
a probability distribution, a low value of entropy indicates a highly ordered distribution. In the case of Wiener entropy, a low value of entropy represents a signal
in which a majority of the power resides in a subset of
the frequencies available. Quantitatively, Wiener
entropy
R
exp log Sðf Þ
R
Sðf Þ
is the ratio of the geometric mean to the arithmetic
mean of the power spectrum, and ranges from 0 (low
entropy, high order) to 1 (high entropy, low order).
Sometimes, the log of the ratio is the quantity used, in
which case the value ranges from 0 to negative infinity,
with values of larger magnitude (i.e., less negative)
indicating more order.
Variance of Wiener entropy (entropy variance):
This

2
refers to the statistical variance (i.e., E ð
x  xÞ ) of
the Wiener entropy calculated over a discrete song segment. It is always a positive value. Since Wiener entropy
reflects the distribution in the power spectrum, higher
variance of entropy implies a more rapidly changing
power distribution, hence a more complex sound. In
the songs of developing juvenile birds, low Wiener variance songs are those that have less structure.

Song Learning and Sleep

Theoretical Background
Sleep is a behavior broadly distributed in the animal
kingdom. Definitions of sleep include three criteria:
specific postures, increased sensory thresholds, and
specific electrophysiological properties. There is no
universally accepted definition of sleep hence its phylogenetic distribution is unresolved, but it is broadly
observed in vertebrates and there are well-defined
sleep-like states observed in worms and flies. There is
also no single function that sleep serves, nor are the
functions of sleep and the adaptive advantages of sleep
limited to the brain. In birds and mammals, however,
there is compelling evidence in some species that sleep
contributes to learning.
In humans, investigations of the role of sleep in
learning have examined memory consolidation, the
process whereby a newly formed memory trace is
transformed from a labile state, where it is susceptible
to interference or decay, into a more stabile memory. In
a typical form of the experiment, groups of subjects
learn a task, and a given group’s performance is then
tested at one or more periods in the following days.
A common pattern of results is that performance
improves immediately after training, drops throughout
the first day, has returned to levels achieved immediately after training on the following morning, and is
stable thereafter (hence the memory is “consolidated”).
This pattern has been observed in simple perceptual
and motor learning tasks, generalized speech perceptual learning, and sensorimotor learning. A similar
pattern of sleep-related learning has been observed for
birds on a simple perceptual learning task. Sleep has
also been implicated in facilitating insightful learning,
and other higher level cognitive tasks. Critical questions are why initial memories are labile, how daytime
experience degrades performance, and how sleep stabilizes memories and creates new information.
There are two common theoretical treatments of
how sleep interacts with neuronal plasticity. One
emphasizes homeostatic mechanisms to regulate overall neural excitability (Tononi and Cirelli 2006) and the
other emphasizes coordinated reactivation (of brain
regions) or replay (of single and populations of neurons) to transfer information across brain regions and
stabilize memories (Diekelmann and Born 2010).
These concepts are not mutually exclusive.
There are only a few animal model systems that
have been developed to study sleep and learning

S

phenomena, one of which is birdsong learning. The
oscine birds (or “true” songbirds) are a suborder of
passerine birds comprising some 4,000 species. Notable
within oscine birds is the tremendous variability in the
structure of songs at all levels of analysis, both across
species and across individuals within species. Song is
also used in a very broad range of behaviors. One
central feature of song, perhaps shared universally
across oscine birds, is that song is learned. It has been
thought that song learning was restricted to oscine
birds and not to all passerine birds, but recent results
point to the possibility that some aspects of songs (such
a temporal patterning in duetting species) might also
be learned in neotropical “suboscine” species. This is
a group of approximately 1,500 species for which little
research has been conducted.
Another distinction between oscine birds and other
avian species may be in sleep structure. Most bird
species have little REM and do not show a clear pattern
of nocturnal variation in NREM and REM. In the
intensively studied zebra finches (Taeniopygia guttata),
however, a surprisingly mammalian-like pattern of
sleep structure has been observed, with clear episodes
of slow wave sleep within NREM, decreasing NREM
and increasing REM throughout the night, and long
and frequent REM near the onset of the diurnal period.
This revives questions about the evolution of sleep,
given these complex sets of traits are shared by mammals and apparently oscine birds in general, but are
separated by other bird taxa and reptiles that seemingly
lack these traits.
In the typical song learning pattern, early in development juvenile songbirds acquire a sensory model of
song from exposure during a critical period to one or
more adult male “tutors” (in most species song is
sexually dimorphic), and this exposure triggers
a second, sensorimotor stage when birds slowly learn
to imitate the memorized song. Access to the tutor song
(assessed by isolate rearing) and access to auditory
feedback during practice singing (assessed by deafening
prior to subsong or rearing in high noise levels) are
both essential for normal song development. A central
hypothesis has been that the song memory is stored as
an “acquired sensory template” that is then used to
evaluate auditory feedback during singing (Konishi
2004). In the absence of tutoring, a less precise “innate
template” is used in the evaluative process. The “template” concept could be implemented by direct real-

3151

S

3152

S

Song Learning and Sleep

time comparison of auditory feedback with sensory
memory, and as well – as recent data suggest – by offline
activity especially during sleep. The innate and
acquired templates remain hypothesized constructs
whose anatomical localization and functional organization are not known, but recent observations (below)
implicate particular pathways in giving rise to auditory
input to the song control circuitry.
Sensorimotor song learning typically progresses
through three developmental stages. Subsong is characterized by low amplitude vocalizations of highly variable duration and low repeatability. Following
subsong, juvenile songbirds enter a “plastic” song
stage consisting of distinct, repeatable segments of
song, delivered at a higher amplitude than subsong.
As individual syllables emerge during plastic song,
their order and morphology may change. Lastly, at
sexual maturity, the plastic song crystallizes into an
adult song, with the degree of plasticity in the pattern
and morphology of syllables throughout the rest of life
dependent on species, age, season, social interactions,
and other factors. In zebra finches, common laboratory
birds, the subsong, plastic, and adult song stages begin
at roughly 25, 40, and 90 days respectively after
hatching.

Important Scientific Research and
Open Questions
Recent work has investigated how controlling the introduction of the tutor song during development influences song learning (Derégnaucourt et al. 2005). When
a juvenile zebra finch raised in isolation of adult song
finally gains access to a tutor song, a change in spectral
features of song can be observed. In particular, the
variance of the Wiener entropy, a measure of spectral
complexity, gradually increases from the day following
the introduction of the tutor song and continues to the
onset of adulthood, at which point the song crystallizes.
Over the course of a day, a shorter time scale pattern in
the entropy variance also manifests. Entropy variance
starts at a lower value each morning than the previous
(or following) day’s peak. Although this nightly effect
of sleep appears to be in opposition to the more long
term ontogenetic song learning process, it is adaptive:
larger fluctuations across sleep correlate with greater
efficacy in song copying.
The daily changes observed in entropy variance
appeared to be a direct result of sleep rather than

spontaneous degradation due to lack of practice or
a function of the time of day. Birds prevented from
singing in the afternoon even for long intervals did not
exhibit degradation, and birds not permitted to sing in
the morning sang degraded songs when starting to sing
in the afternoon. When sleep was drug induced for 2–3
h during the day, upon awakening in the afternoon the
first songs were degraded. These results demonstrate
a causal relationship between sleep and variation in
singing behavior (degradation followed by improvement), as measured by entropy variance.
Not only has song learning behavior been well
studied in zebra finches, but substantial physiological
data has been collected on the underlying neural control of these behaviors, both in adults and juveniles.
Song production is driven by activity in two telencephalic pathways that converge at the robustus nucleus
of the arcopallium (RA), the primary forebrain output
nucleus of the song system. Nucleus RA, analogous to
motor cortex, directly innervates the brainstem nuclei
controlling the syrinx, the avian vocal organ, as well as
brainstem respiratory nuclei, to coordinate respiration
and singing. RA also projects to other cortical structures and to thalamus.
The neural activity patterns in this and other
song system nuclei show state-dependent changes.
For example, in sleeping adult zebra finches, neurons in RA spontaneously discharge individual or
sequences of bursts that closely mimic the patterns
those neurons show when the birds sing (Dave and
Margoliash 2000). This “replay” phenomenon has
been hypothesized to be tied to plastic mechanisms
for maintaining adult song, which remains an open
question. Many burst patterns of individual RA neurons change over sleep, and it is thought that these
changes have to be coordinated else the adult song
would rapidly degrade.
In juvenile birds at the plastic song stage, there is
strong evidence for physiological changes accompanying the daily fluctuations in singing (Shank and
Margoliash 2009). Spontaneous neural activity in
sleeping juvenile birds in nucleus RA prior to the introduction of a tutor song exhibits predominantly tonic
firing. On the night immediately following the introduction of a tutor song, bursting activity substantially
increases. Bursting continues to increase on subsequent
nights. This physiological change leads the change
observed in song production behavior, which is

Sparse Feature Learning

observed the following day. Still, it is not known which
specific nuclei in the song system or auditory system
pathways are important for driving these processes, and
a causal relation between sleep dependent learning and
nighttime bursting remains to be fully established.
This spontaneous bursting activity also exhibits
a clear sensory character. Distributions of intervals
between spontaneous spiking events in RA from
birds with the same tutor song exhibit greater similarity than across groups. This strongly implies that
acoustic features of the tutor song influence the spontaneous activity of nucleus RA shortly following the
introduction of tutor song. How auditory memories
are transferred to the motor pathway remains to be
elucidated.
The functional role of the degradation in syllable
morphology across sleep is still unexplained. One
hypothesis is that the degradation and subsequent
readjustment pushes the song production system
through a greater range of motor behavior and resultant sensory feedback. This would be consistent with
general studies in learning theory, which imply more
efficacious learning with a greater sampling of the
decision space.

Cross-References
▶ Learning and Recall under Hypnosis
▶ Learning to Sing Like a Bird
▶ Reactivation and Consolidation of Memory During
Sleep

References
Dave, A. S., & Margoliash, D. (2000). Song replay during sleep and
computational rules for sensorimotor vocal learning. Science,
290, 812–816.
Derégnaucourt, S., Mitra, P. P., Feher, O., Pytte, C., & Tchernichovski,
O. (2005). How sleep affects the developmental learning of bird
song. Nature, 433, 710–716.
Diekelmann, S., & Born, J. (2010). The memory function of sleep.
Nature reviews, 11, 114–126.
Konishi, M. (2004). The role of auditory feedback in birdsong. Annals
of the New York Academy of Sciences, 1016, 463–475.
Margoliash, D., & Schmidt, M. F. (2010). Sleep, off-line processing,
and vocal learning. Brain and Language, 115, 45–58.
Shank, S. S., & Margoliash, D. (2009). Sleep and sensorimotor integration during early vocal learning in a songbird. Nature, 458,
73–77.
Tononi, G., & Cirelli, C. (2006). Sleep function and synaptic homeostasis. Sleep Medicine Reviews, 10, 49–62.

S

3153

Song Learning Behavior
▶ Song Learning and Sleep

Sorting
▶ Learning by Chunking

Source Analysis
▶ Historical Thinking

Source-Based Learning
▶ Resource-Based Learning

Sourcing
▶ Historical Thinking

Space Adaptation
▶ Adaptation to Weightlessnes

Spaced Practice
▶ Trial-Spacing Effect in Associative Learning

Sparse Feature Learning
▶ Learning Hierarchies of Sparse Features

S

3154

S

Sparse Processing

Sparse Processing
Activation of a limited number of units in a neural
network in response to complex inputs. The hippocampus, among other associative structures, has
a predilection for this form of processing.

Spatial Ability
Spatial ability refers to the perception and/or mental
manipulation of visual stimuli. It may also include
mentally rehearsing a visual experience or update
one’s orientation and location in space.

Cross-References
▶ Mental Rotation and Functional Learning

Spatial Cognition in Action
(SCA)
MARILYN PANAYI1,2, DAVID M. ROY3
1
City University, London, UK
2
The School Room Paediatric Neurosciences,
King’s College Hospital, London, UK
3
Ensomatica, London, UK

as explicit action or internally as simulated or imagined
action. Neurodynamics is an interdisciplinary area that
included examination of neural oscillations involved in
neural fields. Characteristics of neural field activity are
linked to cognitive functions, for example, perception,
action, and learning. Neuro-semiotics is a term first
coined in the late 1970s, which initially referred to the
study of sign in the context of neurological process and
dysfunction. It has evolved as a wider sub-domain of
biosemiotics (from the Greek “bios” meaning “life” and
“semeion” meaning “sign”), including cognitive neuroscience. Within SCA, the concept of sign develops
beyond traditional concepts of code. The term
“noema” comes from the Greek “nόZsiς” (noesis) to
designate correlated elements of structure of any intentional act. Advances from such diverse knowledge
domains as Field Theory (from physics and refers to
the mathematical properties of physical phenomena)
and biosemiotics are beginning to converge and significantly influence how we now think about the science of
intentionality. Our ability to understand ourselves and
others in space and time is fundamental to the evolution
of our learning and cognition.

Theoretical Background
The SCA model explores how we can develop a deeper
understanding of human intentionality in terms of child
spatial cognition embodied in their actions. It takes
a biological systems approach that is informed by
neurodynamics and neuro-semiotics. Two principal
questions underpin the model:
● What can gesture reveal about neuro-atypical and

Synonyms
Deed; Involving; Relating to

Definition
Spatial Cognition in Action (SCA) is concerned with the
corporeal, neurodynamics, and neuro-semiotics of
human capacity to perceive, simulate, synthesize, and
execute knowledge (noesis) of action of “self ” and with
“other” (i.e., other person or artifact) in veridical, imaginary, and hybrid space. Corporeal relates to embodiment and comes from the Latin “corporeus,” from
“corpus” meaning “body.” The nature of such action
can be explored through gesture; where gesture is considered an intentional action expressed either corporeal

neuro-typical child spatial cognition?
● Is there the potential to influence gestural capacity in

children?
Roy (1996) considered human action as a “complex
non-linear dynamic system” having both fixed (steady
states) where some are “attractive,” that is, neighboring
states that tend to converge toward the fixed points
termed “attractors.” Unpredictable behavior patterns
may be termed “chaotic”; thus, such systems allow for
behavior to be emergent. Affordance (environmentally
or internally driven) provides one means of reducing
the system complexity. Dynamic System Theory (DST)
and neural network architecture (ANN) were successfully applied, establishing the “proof of concept” for the

Spatial Cognition in Action (SCA)

computer recognition of dynamic cerebral palsy gesture (1994).
Intentionality of action has a long history philosophically (e.g., Aristotle; Aquinas 1272; Husserl 1800s;
Merleau-Ponty 1963/1964), in performance (e.g.,
Adams 1891; Bacon 1881; Decroux 1961), in the
domains of child motor development (e.g., Kelso
1987; Thelen et al. 1994), and in the neuroscience of
action and motor control (including Bernstein 1967;
Keslo 1984; Fuster 1997/2007; Rizzolatti et al. 2001;
Rizzolatti and Arbib 1998; Jeannerod 1994 cited in
Panayi 2010). In the last two decades researchers
have successfully applied a range of mathematical
modeling techniques to the dynamic and learning
aspects of intentionality. Advances in both vision
research and computing power have made it possible
to both recognize aspects of human action and to create
believable synthetics characters (avatars) and autonomous agents (robots). Such technologies are now being
used to enhance a diverse range of human activity,
for example, learning, medicine, rehabilitation, and
leisure.
The SCA Model considers not only the dynamics,
driven by environmental stimuli and affordance but
also the significance of relevance and salience based on
the “experiences” of the organism (child) (Panayi et al.
2005). Field Theories are being applied to Perception,
Action, and Cognition (PAC), for example, Dynamic
Field Theory (DFT) to simulate child cognition (Spencer and Schöner 2006), Child Spatial Cognition in
Action (Panayi et al. in preparation), and Dynamic
Systems Theory is being revisited in the context of
neo-Piagetian learning (Rose and Fischer 2009). Similarly, Quantum Field Theory (QFT) is being applied to
the modeling of the neurodynamics (e.g., Freeman and
Vitiello 2007). Likewise, researchers in the field of
neuro-semiosis (e.g., Bouissac 2006; Hoffmeyer 2007;
Barbieri 2010; and others) offer a means of bringing
together the neurology and salience of action. SCA deals
with cross-domain phenomena that are multi-functional and causal. Dynamic System Theory has the
power to deal with both gestural development and
motor complexity, thus making it a good candidate
for validating and extending such models. The work
of the authors supports that of others and suggests
that the underlying mechanism that supports our
actions, whether simulated (imagined) or executed,

S

3155

is supra-modal in nature (Panayi and Roy 2005).
Such diverse knowledge domains are beginning to
converge and significantly influence how we now
think about the science of intentionality (for review
see Panayi 2010). The SCA Model aims to extend our
thinking on gestural embodiment, by illustrating how
the intersection of biological and physical science in
the twenty-first century is impacting on the science of
learning.

Important Scientific Research and
Open Questions
Gesture Embodiment, Neuro-atypical
Learning, and Field Theories
Research into child gesture has established not only its
phylogenetic and ontogenetic role, but also the role of
gesture in learning. The majority of work addresses cospeech gesture, developmentally, socially, and in pedagogic contexts (e.g., Goldin- Meadow 1993; Goodwin
and Goodwin 1986; Hostetter and Alibali 2008;
Goldin- Meadow and Cook 2009 cited in Panayi
ibid). The SCA Model puts forward two concepts for
examination: “the ecology of interaction” and the relationship of actors within those ecologies, that is, “self
and other,” within the construct of the entity – “Gesture-As-Action” (GAA). The SCA Model has the capacity to capture properties of a natural system, proposing
y (theta) to describe “an open, but bound gesture system”
where:
y1, y2 and y3 describe children’s capacity for the
following:
y1 Conception, access, and control of the body action
schema, self and other
y2 Use of cross-modality to code for action in multisensory space
3
y Ability to execute dynamic transitions between the
physical and conceptual world
Empirically, gestural repertoires are elicited during
children’s engagement in a range of ludic interactions.
Children with severe speech and motor impairment
due to cerebral palsy (SSMI-CP) have experienced significant damage to the basal ganglia (BG) and the
cerebellum (CB) at or before birth. Both areas are
implicated in action control and integration. Such

S

3156

S

Spatial Cognition in Action (SCA)

damage affects the nature of children’s movement profile and levels of control, often resulting in:
● Body and cognitive “de-conditioning”
● Quality of life with a high dependence on others for

their everyday needs
However, given “scaffolded” opportunities to engage
corporeally, both neuro-typical (NT) and neuroatypical (NAT) children produce significant repertoires
of intentional gestural actions, using a range of spatial
action strategies, including:
● Abstraction
● Modulation (Panayi 1998)

SCA Model allows for the gesture system y (theta),
to be driven by physical constraints and underlying
motivation context, rather than by linguistically motivated structures (see ▶ neuro-dynamics section).
These spatial cognitive abilities can be exploited by
children, to develop their interaction, that is, with
inter-actor(s) or artifact(s) in real, imaginary, and
hybrid world (see Roy 1996; Panayi ibid). The value
of examining neuro-atypical Spatial Cognition in
Action lies not only in the potential to inform pedagogic and rehabilitation practice, but also to add to our
theoretical knowledge of how the corporeal “brain–
body” works (Panayi 2005).
The most recent developments of the model
re-visit advances in Field Theories that deal with
neurodynamics and biosemiotics to examining
intentionality of action of the “action ready body,”
(Panayi 2010) that could be code independent (Panayi
2011).
The ephemeral nature of gesture provides an exciting perspective from which to explore how we perceive, simulate, execute, and evolve our intentionality
across ontogenetic, microgenetic, and phylogenetic
time scales. Figure 1 illustrates the SCA Model in
terms of the proposed components of an open, selforganizing system for intentional action and emergent
behavior, together with its learning and evolutionary
capacity.
Theoretical frameworks, in particular, have the
advantage of being able to condense and provide coherence to bodies of related knowledge. Such models are
able to describe, interrogate, and simulate aspects of
complex ideas and processes. “We need the theory and
models or we do not understand what we find” Fuster

2007. Equally, the influence of our practice contributes
to the development of our theoretical models. “Rather
than conceiving practice as an application of theory, as its
consequence, it is, on the contrary, the forerunner that
inspires theory, the creator of a theory yet to come. . .”
Deleuze
What are needed are theories and analytical tools
informed by practice, which both:
● Support the re-examination of body–brain (corpo-

real) complexity
● Guide our deeper understanding of how our embodied

experience influences our actions

SCA Neurodynamics of Field
Theory I and II
Pattern Complexity of Perception,
Action, and Cognition
Perception, simulation, and execution of action in primates are known to be activated by distributed cortical
networks. Spatial Cognition in Action is examined
empirically and theoretically at both microscopic
(actual sensory inputs) and mesoscopic and macroscopic (through the examination of executed gesture)
levels. Collectively these data provided evidence to
inform the “body–brain system” from corporeal
perspectives.
Fuster (1997, 2007 cited in Panayi ibid) proposed
that structure–function relationships and interconnectivity across ontogenetic and phylogenetic time scales
of the Perception-Action-Cycle (PAC) results from
“active engagement moment to moment.” The SCA
model proposes that this “moment-to-moment (gestural) action” is supra-modal. Brain system complexity
can be described as a network with high distributed
connectivity. Such a biological system is “open” and
operates “at a distance” from where energy and matter
may be exchanged, that is, “dissipative structures” which
result from a “breaking of symmetry.” This breakdown
allows for emergent, complex, and/or chaotic forms
(Prigogine and others). Freeman proposes the use of
“neural power dynamics,” where with increasing (neural) power results in increasing order (opposite to
thermodynamics), that is, “as neurons interact they
constrain each other.” He showed that the “neural patterns of meaning” are by nature “mesoscopic wave
packets.” Stimuli deactivate the cortex, activate arrays

Spatial Cognition in Action (SCA)

S

3157

S

Spatial Cognition in Action (SCA). Fig. 1 SCA: Neuro-semiotics of Dynamic, Self-organizing Open System for
Intentionality Action and Emergent behavior with learning and “evolutionary” capacity (Panayi 2010. Developed after Roy
1996 and informed by Uexküll 1909; Deleuze and Guattari 1980; Sebeok and Umiker-Sebock 1992; Bouissac 2006;
Hoffmeyer 1996/2007; Barbieri 2010)

3158

S

Spatial Cognition in Action (SCA)

of synaptic networks of neurons, and “transmit” neural
energy as “waves.” These wave packets follow trajectories that are demarcated by phase transitions. The system tends to evolve toward maximum energy and
order. Amplitude modulation (AM) of 100 ms is
thought by Freeman to signify elicitation. Collectively
the brain thus has “neural capacity for energy-dependent, self-organizing cortical phase transitions.”
Significantly for SCA-learning (simulation and execution of action), these neural arrays could embody
aspects of knowledge that are not directly linked to the
microscopic sensory inputs. This activity takes place in
both sensory and motor cortex and large domains of
the human cortex – the brain becoming a “dynamical
processor with global distribution of energy.” Importantly, the “dynamics of the cortex is independent of
size”; this phenomenon phylogenetically supports the
implications that mammalian brains “work in the same
way” (Fuster 2007; Freeman 2007 cited in Panayi ibid).

Human Interaction Ecologies of Complex
External Environmental Patterns

Figure 2 illustrates components of pattern complexity
(SCA). Figure 3 illustrates simple neurodynamic features involved in PAC (SCA).

SCA Neurodynamics of Field Theory III
Connectivity, Plasticity, and
Development
The next stage of SCA was informed key two Field
Theories: Dynamic Field Theory (DFT) and Quantum
Field Theory (QFT). DFT offers a conceptual and
descriptive (mathematical) framework with two main
advantages: firstly, as a bridge between “motor control
and development into cognitive function,” and secondly,
“thinking about representation-in-the-moment.” Neural
populations (cell assemblies) are conceptualized
as “neural fields” that represent perceptual features,
actions, and cognitive decisions (Spencer and Schöner).
DFT provides a platform for the interrogation and

Simulation / Execution of Intentional Action
Neuro-atypical & Neuro-typical
Sequences and sub-movements including
manipulations and micro-gestures

e.g. Ludic
Ordered / serial patterns or
disordered ‘chaotic’ patterns
Perception

Action
Neural Cyto-architecture
and Mechanism
With capacity for
transforming ‘Thought
action’ to ‘Action in
space and time’ and vis
versa as spatial patterns
either veridical,
imaginary or hybrid
space

Distributed Neural Connectivity [DNC]
Microscopic
Mesoscopic
Macroscopic including:
Cerebral Cortex (Motor [Mi; Premotor
[PM]; Basal ganglia [BG] and cerebellum
[CB] coupling allowing for shared decision
coarse-fine movements.
Excitatory/Inhibitory pathway
connectivity to Thalamus, hippocampus and
amyggala

(Recognition / Activation)
Event exposure and interaction
creates coherence and stability
over time

Phyletic Memory
Gradient
(Recall / Binding / Release)
Candidate for LTM / WM
Synaptic stores and
spiky neurons

Analysis
(‘Coding / Decoding’)
At neurodynamic level neuron energy as
‘carrier waves’ within networks of connectivity

Spatial Cognition in Action (SCA). Fig. 2 SCA Neurodynamics I: Postulated components for dealing with pattern
complexity connectivity. (Schematic informed by the work of Merleau-Ponty 1963/1964; Fuster 1974, 2007; Thelen et al.
1995, Freeman and Vitiello 2007, 2009; Spencer and Schöner 2009; Dreyfus 2007; Sporns 2007; Panayi et al. 2005, 2010 in
preparation)

Spatial Cognition in Action (SCA)

S

3159

Modes such
‘waves’ or
‘oscillations’

Power
increase,
order
increases
(Neurons
interact and
constrain
each other)

Analysis of
‘global’ neural
energy
‘Free’
connectivity
and dynamical
processors’

Dissipative
structure
Result of
‘breaking of
symmetry’

‘Dissipative
structures’
Emergent,
Complex or
chaotic form of
action 3–10 time
per sec (Theta
and alpha
range)

Spatial Cognition in Action (SCA). Fig. 3 SCA Neurodynamics II: Field Theory Action Perception Cycle (Developed after
Freeman and others)

S
simulation of intentional gestural action, particularly of
neuro-atypical children (Panayi et al. in preparation). In
contrast, Quantum Field Theory (QFT) enables the
SCA Model to go beyond “representation as coding” in
the traditional sense (see Fig. 4). Quanta can be
described as “discrete” energy packets that can exist at
different states of energy. At the neuronal level, neurons
are considered as vast assemblies of interacting “particles”; these particles “can occupy the same place in
space,” and thus they are capable of “carrying force;”
specifically electromagnetic force (related to action
potential firing, that is, “neural forces”) (Freeman and
Vitiello 2009).

Likewise, wave-based processes have also been
postulated at both whole organism and at cellular
metabolic level, for example, locomotion, Purkinje
cells, and enzyme catalysis. Correspondingly, graphical theoretical techniques are being applied to brain
networks with structure (simple) to function (repertoire) together with their interactivity and granularity
of these interactions (see review Sporns 2007). Features
of interactivity that are of relevance to the SCA Model
included:
● Centrality of “Hubs” and their functional connec-

tivity could relate to both neuro-typicality and

3160

S

Spatial Cognition in Action (SCA)

Experience Mediated by (external & Internal) Environment
Neuro-atypical &
Neuro-typical Movement profiles
Spontaneous/background activity
Of Spatial Temporal Patterns
Macroscopic e.g. EEG, MEG
Mesoscopic e.g local field potentials ECoG
Microscopic e.g. single cell action potential
Attractor Iandscapes:‘cumulative correlation maps’ making
possible ‘Intentional Arc’
Abstractions and generalizations

Resembles vortices of tornadoes
Every point in the dynamic field
has amplitude and direction of
change in space and time

Stimuli destabilizes cortex
Stimulus and stimuli
independant
Force creates
neuronal activity

Intentional
Action
Simulated
or
Executed
Neuroatypical &
Neurotypical
Gesture
Repertoires

Gestural Action as
Executed Action
GA-EA

Gestural Action as
Simulated Action
GA-SA

SCA

Thus, meaning
Expression and
of sensory
transmission of
information
sensory information
is expressed
and transmitted as
Spatial Patterns of
Cortical Activity
(SpCA).
These resemble
frames in a
cinematographic
movie

System
Dynamics

LTM/STM
Rapid and
repeated
bursts
retrieve
memories
and bind
them with
sensory
information
into percept

Retieval/Binding

Self-organizing
Patterns

Structure for Phase Transitions
‘Gamma wave packet’ is a ‘perceptual carrier’
Phase Transitions create formation of ‘chaotic’ Electroencephalogram (ECoG) spatial
temporal oscillations at frequencies of 12-80 Hz (beta and gamma waves)
Percept is accessed by calculating Amplitude Modulation (AM) feature vector
Spatial patterns of Amplitude Modulation (AM) self-organized,
Co-existing as‘quanta of energy’ able to interact and making higher order patterns e.g.
motor action and at cellular level
‘Activation of Knowledge Base’ e.g. gesture
Proposes: null spike as ultimate marker for the onset of a new percept/memory
Results from ‘Breaking of Symmetry’ (BoS) and re-ordering
Spreads across cortex, thus supports supra-modal integration
SCA without the need for coding
Potential speculative candidate structure ‘konicortex’ Freeman 2009

Spatial Cognition in Action (SCA). Fig. 4 Schematic for spatial cognition in action: SCA Neurodynamics III: From stimuli
to neural waves and intentional action (Schematic informed by the work of Merleau-Ponty 1963/1964; Fuster 1974, 2007;
Thelen et al. 1994, 1995, Freeman and Vitiello 2007, 2009; Spencer and Schöner 2009; Panayi et al. in preparation)

Spatial Cognition in Action (SCA)

neuro-atypicality communication, movement profiles and child developmental learning patterns.
● Segregation relates to degrees of “clustering,” important for enabling “gestural specificity,” for example,
“pretend to sad,” “happy” (illustrate affect gesture
where typically salience is expressed in the head/face
region).
● Integration relates to cross-connetions co-ordination, important in creating states of “gestural coherence” through compression.
These nodes could become instrumental to the
stability/instability of the dynamic system. At a neuronal level this “small-world connectivity” or “corticalmicro-states” can support both spontaneous activities
(across temporal scales) and complex dynamics in
brain-body systems providing a mechanism for brain
selection; “dependent on both need and input” (Sporns
ibid and in press).
Such “small-world connectivity” has been explored
through the examination of children’s gestural repertoires. Importantly for SCA, such “wave phenomena,”
concepts and network mathematics could provide the
mechanism and tools that could mediate learning (see
also the section on Human praxis and ontogeny of
learning). These structural mechanisms could encompass the neuro-structural constraints of neuro-atypical
learners. Thus, target gestures are produced with levels
of abstract conceptual coherence and performed corporeally. Both neuro-atypical and neuro-typical children are able to modify features of their gestural
repertoires and use a range of dynamic strategies, for
example, space–time parameters such as velocity,
deceleration, boundaries, rhythmic and synchronous
features in terms of “self ” and “other.” Interestingly,
the typical rate of gesture production ranges is 3–6 s,
thus gesture serves as a robust example of “moment-tomoment” dynamic activity. As with DST, both DFT and
QFT can accommodate the existence of emergent phenomena in children’s gestural repertoires, for example,
the creation of gestures to explain novel experience such
as describing “neologism” related to “inventions” during
the re-telling and enaction of an animated cartoon.
Extracts from the Child Gesture Corpus (Roy and
Panayi) are illustrated in the figures that follow where:
Exemplar 1 Target Gesture: “pretend to play the
violin,” Gesture Execution: Typically involves interaction with the body (self) taking role of the “other,” that

S

is, “violin player” and interacting at varying degrees of
complexity with a primary imaginary object : “violin,”
secondary imaginary object: “bow” and tertiary object:
“other” – “audience.”
Exemplar 2 Target Gesture: “lasso that steer,” Gesture Execution: Involves the body (self) taking the role
of “other,” that is, “Cowboy” and interaction at varying
degrees of complexity with a primary imaginary object:
“lasso,” secondary imaginary object: “steer,” and tertiary
object: “other” – audience. Co-ordination movement
strategies: coarse sequence of movement: “lasso the
steer” (head, arm, and upper body) and finer submovements (grasp of lasso rope, rotation, and throw
of lasso, emotional salience (facial), see Fig. 5).
Exemplar 1 Extract Stimulus: First 7 min of animation (Claymation) cartoon “The Wrong Trousers” Dir.
N. Park, Aardman Productions.
Novel Gesture Repertoire: Child describes a series
of novel inventions, for example, “Buzzing for Breakfast,” “Flipping bed,” and “dressing robot,” see Fig. 6.
Row One: “Buzzing for Breakfast” Gesture Sequence
(a), (b), (c) illustrates an example of compression of
gestures to convey both “mise-en-scene” and the “complex idea/novel situation.” The gestures embody dynamics and salient narrative features to describe interaction
in imaginary story space. In (a) Gesture (closed fist on
chest) indicates that (a) “You/I” are now upstairs (b)
bimanual “upstairs” gesture with one arm hand
representing “stairs,” other hand/arm represents
“you” going up stairs, where Wallace (character change
POV) and change of role to Wallace, who is pressing the
buzzer c).
Row Two: “Flipping Bed” Child becomes “novel
object” and enacts most salient feature, that is, “flipping bed” framing and dynamics. Facial expression
illustrates the emotional nature of gesture.
Row Three: “Dressing Robot”
Child takes both the role of Wallace and the “dressing robot” to enact Wallace having shirt (arms), (a) and
(b) pullover put on over his head. Body maintains role
of Wallace, hands take role of robot arms.

From Models to Practice
Human Praxis and the Ontogeny of
Learning
The “organic code hypothesis” of Barbieri (2008 cited in
Panayi ibid) offers further potential for development,

3161

S

3162

S

Spatial Cognition in Action (SCA)

Spatial Cognition in Action (SCA). Fig. 5 Emergent child gesture: Neuro-atypical SCA – Dealing with complexity:
Conceptual integration and compression from left to right. Target gesture: pretend to lasso the steer

that is, opportunities for experiences and learning that
perhaps distinguishes as human. Deleuze and
Guattari’s philosophical concept “image of thought”
describes learning through “rhizome” connectivity.
This connectivity could support both understanding of complexities, for example, space and time that
support our actions and give opportunity for “alternative thinking and learning.” At a phylogenetic level,
Hoffmeyer proposes the existence of evolutionary
advantage through what he terms “semiogenic scaffolding,” that is, semiotic components instrumental in
learning and interaction (Hoffmeyer 2007).
At a more tangible level, the value of understanding
the science behind learning lies in its power to inform
a vast array of human activity, for example, from pedagogy, rehabilitation therapies to the development of
future interactive technologies. The relevance of the
SCA model lies in the theoretical and pragmatic contribution to informing change in practice. The authors
argue for an inclusive re-evaluation of how we

encourage children’s thinking and action in space.
Freire advocated for pedagogy to embrace “dialogue
as indispensable to the act of cognition” and essential to
“inquiry and creative transformations.” p71. In the SCA
model, the term “dialogue” has been extended to
include “corporeal dialogue” as action that examines
both the nature of gesture and children’s capacity to
modulate their action. From an interaction paradigm
and pedagogic perspective the SCA Model is informed by
two fundamental requirements:
● Need to understand and include both variability

and mechanism for developmental transitions
● Imagery/simulation as a critical component of the
system dynamics
Specifically, Vygotskian “décalage –variability” is
described as a mechanism of transition (zones of proximal development). In neo-Piagetian “developmental
webs,” identify variability and stability as vital to the
“dynamic ways people’s actions differ and change”

Spatial Cognition in Action (SCA)

S

3163

Spatial Cognition in Action (SCA). Fig. 6 Emergent child gesture: Neuro-typical SCA – Dealing with complexity;
Conceptual integration and compression from left to right, top to bottom. Re-telling and extract of a cartoon narrative
with novel gesture to express neologisms and inventions e.g. ‘Buzz for breakfast’ (row 1), ‘Flipping bed’ (row 2), ‘Dressing
robot’ (row3/4)

(Rose and Fischer 2009, p. 400). Sinha supports the
importance of imagery in spatial cognition with semiotic
and neo-constructivist theories and arguments, where
“radical imaginary is the social imaginary, the foundational source not only of the Vygotskian semiotic mediation of individual cognition, but also of the construction
of all social institutions and social objects.” He goes on to
suggest that “Such an alternative account of human development is both necessary and possible, and its construction
could contribute to both a theoretical and a practical emancipation of imaginative reason . . .” (Sinha 2007, p. 32)

Neuro-atypical Learning Capacity
Neuro-atypical children have the capacity to access
pathways “alternatively,” that is, exploit their cortical

area specialization for learning through “practice” or
“simulation,” rather than by physical practice in the
real (veridical world). (For a derived neurology that
underlies gesture see Panayi 2001 and ibid). Such
mechanisms could make additional neural resources
available to neuro-atypical and neuro-typical children.
Coupled with Field Theories and empirical findings,
the SCA Model illustrates how the body–brain has
capacity for distributed responsiveness and plasticity.
The postulated neurodynamic and neurosemiotic system components and their interactivity are brought
together schematically in SCA. Figures 1–6 illustrate
how, through control of their spatial cognition, they
can deal with patterns of complexity and interaction
that result in intentional action as gesture.

S

3164

S

Spatial Cognition in Action (SCA)

Knowledge exists independent of the knower.
The knower having capacity for ‘alternative
pathways’ for internalisation and transition
across their ‘developmental web’

Existing Knowledge Practices
Systematic State
Traditional learning

Explorative State
Explorative Learning

Paradigm shifts and Innovation in Learning beyond
the 21st Century
SCA
Spatial Cognition in Action
Examines:
Emergent gesture through a bio-semiotic approach to
embodiment of ‘Self & Other’ and the
Dynamic and narrative (semantic, procedural and
episodic) nature of such corporeal (somatic) action
Within the context of the
Neuro-dynamics of the Physicality & Tangibility of
Interaction [PTI] of our Veridical, Imaginary & Hybrid
worlds (creative imagery)
Neuroscience to
Pedagogy

Interactivity
Technology ‘Beyond
the Desktop’

Explorative State
Explorative/Ludic learning
Informed by Informed by Vygotskian,
Neo-Piagetian and Neo-constructivism

Dynamic System & Field Theory
Informs System Change
Dynamic Process of Knowing
Interactive Ecologies of Learning
Individual and social
Lead to rich conceptual development
Deeper understanding through
Co-Construction of Knowledge and
Fluid Intelligence

Spatial Cognition in Action (SCA). Fig. 7 SCA ontogeny of learning

Existing pedagogy and therapy for neuro-atypical
communicators and movers have largely followed
models influenced by the traditional medical model
and pedagogic thinking, often such practices being
inaccessible to dynamic assessment. We need to consider how we can develop both thinking and practice
that tend toward and drive the direction of pedagogy,
therapy, and future interactivity, in ways that can capitalize on both:

● The supra-modal nature of our spatial cognition and

corporeal action
● The diversity capacity of children
The implications of such research which will undoubtedly change not only the way we understand how
we interact with our veridical, imaginary, and hybrid
worlds, but also how we will evolve these interactions (see Fig. 7, SCA – Ontogeny of Learning).

Spatial Cognition in Action (SCA)

From a neuro-dynamic perspective further research is
needed to understand both the nature and the significance of pre-motor activity and its implication for our
spatial cognitive development and learning. Three key
areas are proposed for future scientific research and
open questions with regard to the neurobiology of
underlying systems of gesture rooted in action:
– Performed action as imagined (simulated action)
from an allocentric (action of others) or egocentric
(action of self) or explicitly executed
– Language understanding, in particular, for semantics relating to action (verb or noun) in relation to
self or other (including artifacts)
– Exploration of the role of physical and imaginary or
simulated experience
These domains together with the impact of the
advances in the nature of our technology-mediated
interaction hold a potential to change our pedagogy
and rehabilitation and future technology design practices, now and beyond the twenty-first century.

Cross-References
▶ Action Learning
▶ Biological and Evolutionary Constraints of Learning
▶ Cognitive Models of Learning
▶ Embodied Cognition
▶ Imagery and Learning
▶ Motor Learning
▶ Motor Schema(s)
▶ Pragmatic Reasoning Schemas
▶ Schema(s)
▶ Self-organized Learning
▶ Spatial Learning
▶ Technology-Enhanced Learning Environments

References
Adams Fowle, F. A. (1891). Gesture and pantomimic action. New York:
ES Werner.
Aristotle poetics I (S. H. Butcher, Trans.). http://classics.mit.edu/
Aristotle/poetics.1.1.html
Aquinas (1272) See Stump, E. (2003). Freedom, action, intellect and
will, Chapter 2. From Aquinas. London: Routledge.
Bacon, A. M. (1881). A manual of gesture: Embracing a complete
system of notation, together with the principles of interpretation
and selection for practice. Chicago: J C Buckbee.
Barbieri, M. (2008). The code model of semiosis the first steps
towards a biosemiotics. The American Journal of Semiotics, 24
(1/3), 23–37.

S

Barbieri, M. (2010). Introduction to biosemiotics: The new biological
synthesis (pp. 1–67). Berlin: Springer.
Bentham (1700) cited Husserl, E. (1962) Ideas: General introduction
to pure phenomenology. Collier Books.
Bernstein, N. A. (1967). The co-ordination and regulation of movements. Oxford: Pergamon.
Bouissac, P. (2006). Gestures in evolutionary perspective.
Retrieved June 2010, from http://www.semioticon.com/virtuals/
evolutionofgestures.2.pdf
Cooks-Wagner, S., & Goldin-Meadow, S. (2006). The role of gesture
in learning: Do children use their hands to change their minds?
Journal of Cognition and Development, 7(2), 211–232.
Decroux, E. (1985). Words on mime. Associated University Press.
Deleuze, G., Guattari, F. (1980). A thousand plateaus (B. Massumi,
Trans.). New York: Continuum.
Freeman, W. J., & Vitiello, G. (2007). The dissipative quantum model
of brain and laboratory observations. Electronic Journal of Theoretical Physics, 4, 1–18.
Freeman, W. J. (2007b). In L. Perlovsky & R. Kozma (Eds.),
Neurodynamics of cognition and consciousness, Proposed cortical
“shutter” mechanism in cinematographic perception (pp. 11–38).
Heidelberg: Springer.
Freeman, W., & Vitiello, G. (2009). Dissipative neurodynamics in
perception forms cortical patterns that are stabilized by vortices.
Journal of Physics: Conference Series. doi:10.1088/1742-6596/174/
1/012011, http://iopscience.iop.org
Fuster, J. M. (1997). The prefrontal cortex: Anatomy, physiology, and
neuropsychology of the frontal lobe (2nd ed.). Philadelphia:
Lippincott, Williams and Wilkins.
Fuster, J. M. (2007). Cortical memory. Scholarpedia, p. 11609.
Goldin-Meadow, S., & Wagner, S. M. (2005). How our hands help us
learn. Trends in Cognitive Sciences, 9, 234–241.
Goldin-Meadow, S. (2009). How gestures promote learning throughout childhood. Child Development Perspectives, 3(2), 106–111.
Goldin-Meadow, S., Alibali, M. W., & Church, R. B. (1993). Transitions in concept acquisition: Using the hand to read the mind.
Psychological Review, 100, 279–297.
Goodwin, C. (1986). Gesture as a resource for the organization of
mutual orientation. Semiotica, 62(1–2), 29–49.
Hoffmeyer, J. (2007). Semiotic scaffolding in living systems. In
M. Barbieri (Ed.), Introduction to biosemiotics. The new biological
synthesis (pp. 149–166). Dordrecht: Springer.
Hostetter, A. B., & Alibali, M. W. (2008). Visible embodiment: Gestures as
simulated action. Psychonomic Bulletin and Review, 15, 495–514.
Husserl, E. (1800). Experience and judgement, 1973 [1939], (trans:
Churchill, J. S., and Ameriks, K.). London: Routledge.
Jeannerod, M. (1994). The representing brain: neural correlates of motor
intention and imagery. Behavioural Brain Science, 17, 187–245.
Kelso, J. A. S. (1984). Phase transitions and critical behavior in human
bimanual coordination. American Journal of Physiology: Regulatory, Integrative and Comparative, 15, R1000–R1004.
Kelso, J. A. S., Schöner, G., Scholz, J. P., & Haken, H. (1987). Phaselocked modes, phase transitions and component oscillators in
biological motion. Physica Scripta, 35, 79–87.
Merleau-Ponty, M. (1942/1963). The structure of behavior (A. L.
Fisher, Trans.) Boston: Beacon Press.

3165

S

3166

S

Spatial Contiguity

Merleau-Ponty, M. (1964). The primacy of perception with J.M. Edie
(ed.), (R. McCleary, Trans.). Evanston: Northwestern University
Press.
Panayi, M. (2010). Spatial cognition in action: SCA towards a model of
dynamics neuro-atypical and neuro-typical child embodied gesture.
Unpublished Ph.D. thesis, City University, Northampton Square,
London.
Panayi, M., & Roy, D. M. (2005). Spatial cognition in action – SCA
model: Children’s gestural imagery in action in modelling language. In A. Cangelosi, & G. Bugmann, R. Borisyuk (Eds.),
Cognition and Action: Vol. 16. Proceedings of the Ninth Neural
Computation and Psychology Workshop, 8–10 September 2004.
Plymouth/Singapore: University of Plymouth/World Scientific.
Panayi, M. et al. (in preparation). Using Dynamic Field Theory to
simulate neuro-atypical action, extending the SCA Model.
Panayi, M. (2011). What, why, where & how do children think?
Towards a dynamic model of spatial cognition as action SCA
international gesture workshop, Athens, Greece, 2011. Proceedings Series: LNCS/LNAI.
Rizzolatti, G., & Arbib, M. A. (1998). Language within our grasp.
Trends Neuroscience, 21, 188–194.
Rizzolatti, R., Fogassi, L., & Gallese, V. (2001). Neurophysiological
mechanism underlying the understanding and imitation of
action. Nature Reviews Neuroscience, 2, 661–670.
Rose, L. T., & Fischer, K. W. (2009). Dynamic development: A nonpiagetian approach. In U. Müller, J. I. M. Carpendale, & L. Smith
(Eds.), The Cambridge companion to piaget. Cambridge: Cambridge University Press.
Roy, D. M. (1996). Gestural human-machine interaction using neural
networks for people with severe speech and motor impairment due
to cerebral palsy. Unpublished Ph.D. thesis, City University,
Northampton Square, London.
Roy, D. M. et al. (1994). Gestural human-machine interaction for
people with severe speech and motor impairment due to cerebral
palsy. Conference on human factors in computing systems, Boston,
313–314.
Sebeok, T. A., & Umiker-Sebock, J. (Eds.). (1992). Advances in visual
semiotics. The semiotic web 1992. Berlin: Mouton de Gruyter.
Sinha, C. (2007). Self, symbol and subject. A commentary on Lyra On
abbreviation: dialogue in early life. International Journal of Dialogical Science, 2, 45–50.
Spencer, J. P., & Schöner, G. (2006). An embodied approach to
cognitive systems: A dynamic neural field theory of spatial working memory. Proceedings of the 28th Annual Conference of the
Cognitive Science Society (CogSci 2006)Society (CogSci 2006),
Vancouver, Canada, 2180–2185.
Sporns, O., Honey, C. J., & Kötter, R. (2007). Identification and
classification of hubs in brain networks. PLoS one, 2, e1049.
Thelen, E., & Smith, L. B. (1994). A dynamic systems approach to
the development of cognition and action. Cambridge, MA: MIT
Press.
Uexküll, T.v.(1909). Thure von (1987). The sign theory of Jakob von
Uexküll. In M. Krampen, K. Oehler, R. Posner, T. A. Sebeok, &
T. V. Uexküll (Eds.), Classics of semiotics (pp. 147–179). New York:
Plenum Press. See Kull, K. (Eds.). (2001). Jakob von Uexküll: a
paradigm for biology and semiotics. Semiotica, 134(1/4).

Spatial Contiguity
▶ Split-Attention Effect

Spatial Learning
DAVID R. BRODBECK
Department of Psychology, Algoma University,
Sault Ste. Marie, ON, Canada

Synonyms
Cognitive mapping; Spatial memory

Definition
Spatial learning refers to the process by which an
organism acquires a mental representation of its environment. Spatial learning has been found in both vertebrate and invertebrate species.

Theoretical Background
The first systematic studies of learning by psychologists
such as Thorndike, Skinner, and Pavlov concentrated
on simple stimulus–response (SR) or stimulus–stimulus (SS) connections. In such a framework, an animal
would learn to associate a neutral stimulus with
a biologically relevant one (classical or Pavlovian conditioning) or would associate a behavior with its outcome (operant conditioning).
However, when an animal learns its position in
space, such simple associative processes are not sufficient
to explain how organisms learn locations. The organism
then usually forms some sort of representation of the
environment consisting of the relationship of specific
stimuli in the environment to the animal and to each
other.
This representation was originally termed a cognitive map by Tolman (1948). Tolman found that when
given experience on a maze without receiving a reward,
rats showed similar performance to another group of
rats that were given a reward upon completion of the
maze. Experience in the environment in question then is
the necessary factor in spatial learning, not simple SR or
SS associations.
These cognitive maps are thought to be not unlike
real maps, showing geometric relationships between

Spatial Learning

objects in the environment, with features added. The
animal can then in a sense look up where it is on
its map and determine where to go next, where home
is, where food is, etc. So, the animal learns both distance and direction to and from various landmarks in
the environment. The geometric part of the representation may be served by a specialized module.
While the simple SS or SR approaches to learning
are not that useful when modeling spatial learning, it
may be possible to think of different spatial locations
and features in the environment as stimuli not unlike
those used in mathematical models of associative learning such as the Rescorla–Wagner model. The animal
could weight various sources of information such as
compass direction, proprioceptive cues, etc., differently, allowing it to learn where some goal is in space
(Cheng et al. 2007).

Important Scientific Research and
Open Questions
While Tolman’s (1948) work was groundbreaking
and quite against the more popular behaviorist notions
of the time, it was not until the 1970s that research on
spatial learning really took off. At that time, Olton
and Samuelson (1976) tested rats on a radial maze.
This maze, which has baited arms that radiate out
from a central platform like the spokes of a wheel
was then used in literally hundreds of articles after
1976. Typically rats, and most every other species
tested, performed very well on such a maze, usually
learning to visit arms in a seemingly haphazard order
(rather than say clockwise or counterclockwise until
all the arms are visited) yet made few mistakes. These
results led many to postulate that spatial memory was
somehow different than other forms of learning.
Indeed, when the predictions of and SR type model
and a more general cognitive mapping model are put in
opposition, the cognitive mapping model has won out.
There are many possible stimuli that can be learned
about in spatial learning. Beacons are stimuli that are
directly navigated to. For example, if you wanted to
travel to a given building in a city, the building itself is
a beacon. Landmarks are stimuli that are relatively close
to the goal but are not themselves the goal; instead,
landmarks indicate direction and distance to the goal.
The question of what the content of learning is in
a spatial situation has been addressed especially by
Ken Cheng and his colleagues. When rats are placed

S

in a rectangular arena that has a buried piece of food in
one corner, they are more likely to make any search
errors to the rotational opposite side (e.g., if the baited
corner was the top left of the box, they make errors to
the bottom right) rather than a reflection (in this case
the bottom left). Such errors were present even when
the features of the box, patterns on the walls for example, were different for each wall. This lead Cheng (1986)
to conclude that geometry was encoded first, and
responded to preferentially over other types of cues.
Similar results have been found with young children
(around 3 years old) searching for a toy in a room with
three white walls and one blue wall. However, adults,
when confronted with this task in the presence of a blue
wall, use that cue over geometry (Hermner and Spelke
1994). Such results have lead many researchers to conclude that spatial learning is served by a specialized
geometric module in many species (Cheng et al. 2007).
It is becoming clear that spatial learning is not mediated solely by geometry, landmarks, or beacons, but by
a whole host of cues (e.g., optic flow, proprioceptive
cues) that, for any given location, are more or less useful.
A weighted average of the metric values of the various
cues can be used to point to a particular location. Cheng
et al. (2007) suggest that the weightings follow Bayesian
principles.
Spatial learning is mediated quite clearly by the hippocampus. The discovery of place cells which fire only
when a rat is in a given spatial location is one of the many
pieces of evidence suggesting this (O’Keefe and Nadel
1978). Species that rely extensively on spatial learning
in their natural environment have not only been found
to prefer to solve learning problems spatially, but they
also have a larger hippocampus than animals that rely
less on spatial learning in their natural environment.
Hippocampal lesions in many species disrupt
spatial learning, usually making it impossible. These
lesions do not always affect nonspatial learning, though
this is not universal. Indeed, in humans, the hippocampus not only serves a spatial function but is also crucially important in episodic memory consolidation.

Cross-References
▶ Animal Intelligence
▶ Associative Learning
▶ Cognitive Models of Learning
▶ Comparative Psychology and Ethology
▶ Place Learning and Spatial Navigation

3167

S

3168

S

Spatial Learning and Memory Enhancers

▶ Simulation-Based Learning
▶ Tolman, Edward

References
Cheng, K. (1986). A purely geometric module in the rat’s spatial
representation. Cognition, 23, 149–178.
Cheng, K., Shettleworth, S. J., Huttenlocher, J., & Rieser, J. (2007).
Bayesian integration of spatial information. Psychological Bulletin, 133, 625–637.
O’Keefe, J., & Nadel, L. (1978). The hippocampus as a cognitive map.
Oxford: Oxford University Press.
Olton, D. S., & Samuelson, R. J. (1976). Remembrance of places
passed: Spatial memory in rats. Journal of Experimental Psychology: Animal Behavior Processes, 2, 1–16.
Tolman, E. C. (1948). Cognitive maps in rats and men. Psychological
Review, 55, 189–208.
Further Reading
Hermner, L., & Spelke, E. S. (1994). A geometric process for spatial
representation in young children. Nature, 370, 57–59.

Spatial Learning and Memory
Enhancers
▶ Pharmacological Enhancement of Synaptic Efficacy,
Spatial Learning, and Memory

Specific Exploration
▶ Curiosity and Exploration
▶ Play, Exploration, and Learning

Specific Reading Disorder/
Disability
▶ Language-Based Learning Disabilities

Speech
▶ Communication Theory

Speech and Language
Disorders
▶ Language-Based Learning Disabilities

Speech Category Learning
Spatial Learning in Animals

▶ Acoustic and Phonological Learning

▶ Place Learning and Spatial Navigation

Spatial Memory
▶ Covert Reorganization / Spatial Learning
▶ Learning Spatial Orientation
▶ Memory for “What,” “Where,” and “When” Information in Animals
▶ Place Learning and Spatial Navigation
▶ Spatial Learning

Spatial Navigation
▶ Covert Reorganization / Spatial Learning

Speech Coding
▶ Phonetics and Speech Processing

Speech Perceptual Learning
▶ Acoustic and Phonological Learning

Speech Science
▶ Phonetics and Speech Processing

Spinal Learning

Speedlearning
▶ Superlearning

Spiking Neural Networks
▶ Reinforcement Learning in Spiking Neural Networks
▶ Supervised Learning in Spiking Neural Networks

Spinal Learning
JAMES GRAU
Department of Psychology, Texas A&M University,
College Station, TX, USA

Synonyms
Spinal plasticity

Definition
The spinal cord has been traditionally viewed as a hardwired system that functions to transmit neural information to and from the brain. From this view, neurons in
the spinal cord have little capacity to adapt and instead
respond in a rigid manner over time. Experimental
evidence has discounted this view and demonstrated
that neural mechanisms within this region of the central
nervous system can adapt to new environmental relations and, in this way, meet the criteria for learning.
These criteria require evidence that a behavioral/
physiological modification (1) depends on the organism’s experiential history, (2) is neurally mediated, and
(3) outlasts (extends beyond) the environmental contingencies used to induce it (i.e., the experience has a lasting
effect on performance). Research has shown that spinal
neurons can support behavioral phenomena that meet
these criteria, and in this way, demonstrate spinal learning (for further information on this, and other topics
discussed below, see Patterson and Grau 2001).

Theoretical Background
Researchers have typically used three techniques to
study whether a system is capable of learning: stimulus

S

preexposure, Pavlovian (classical) conditioning, and
instrumental conditioning. Each of these techniques
refers to a kind of methodology, designed to ascertain
whether the system is sensitive to a particular kind of
environmental event, or relation, and whether the
experience has a lasting impact on how the system
operates.
To study spinal learning, researchers typically isolate the lower (lumbar-sacral) spinal cord from the
brain by surgically cutting (transecting) the cord
below the shoulders (a mid-thoracic transection).
This procedure blocks neural communication along
the sensory/motor pathways that carry signals to and
from the brain, producing a paraplegia that blocks
(brain-dependent) conscious sensations from stimuli
applied below the waist and the ability to voluntarily
perform a motor response.
Studies examining the consequences of stimulus
preexposure typically monitor how the behavioral
response elicited by a particular stimulus changes
as a function of experience (defined as the duration
of exposure and/or number of presentations). As
a result of this experience, the behavioral/physiological
response to the stimulus may become weaker (habituation) or stronger (sensitization). In its simplest form,
the only factor that matters is experience with the
stimulus alone – its relation to other stimuli, or the
behavioral response, is irrelevant. Because there are no
associations with other events or responses, this type of
learning is often characterized as nonassociative in
nature.
Over a century ago, Sherrington (1905) showed that
spinal reflexes exhibit habituation. Decades later,
Groves and Thompson (1970) built on this foundation,
demonstrating that neurons within the spinal exhibit
both habituation and sensitization, and that these phenomena abide by many of the same principles that
govern their action in other behavioral paradigms.
The lumbosacral spinal cord contains an inner
region of neurons (the gray matter) that can organize
and execute some simple behaviors in the absence of
input from the brain. For example, a weak shock to
a rear paw can elicit a withdrawal (flexion) response in
a spinally transected animal. Similarly, a noxious thermal stimulus applied to the tail will elicit a withdrawal
response. In both cases, physiological experiments have
shown that the behavioral response depends on
a neural (reflex) arc that passes through the spinal

3169

S

3170

S

Spinal Learning

cord. Because the behavioral response depends on spinal neurons, and can be elicited after communication
with the brain has been severed, it is called a spinal
reflex. Using this preparation, researchers have shown
that repeated presentation of a weak shock causes
a lasting, neurally mediated, decline in response magnitude (habituation). Conversely, exposure to an
intense shock can cause an increase in response magnitude (sensitization). More recently, researchers discovered that a variety of noxious events (e.g., peripheral
inflammation or injury) can sensitize pain-like (withdrawal) behavior. This enhanced behavioral reactivity
has been linked to an increase in neural excitability
within pain (nociceptive) fibers and spinal neurons
(central sensitization) and is thought to contribute to
the development and maintenance of persistent pain.
Whether spinal neurons can support Pavlovian
conditioning has proven much more controversial.
Pavlovian conditioning provides a method to assay
whether a system is sensitive to stimulus–stimulus
(S-S) relations. In the prototypic case, this is studied
by pairing a cue (the conditioned stimulus [CS]) that
has relatively little behavioral effect with a biologically
significant event (the unconditioned stimulus [US]).
For example, Pavlov showed that pairing an auditory
cue (the CS) with food (the US) endowed the CS with
the capacity to generate a conditioned salivation
response (the conditioned response [CR]). Pavlov
assumed that this learning depended on higher-level,
cortical, processes. E. Culler challenged this view
a decade later, presenting evidence that spinal neurons
may be sensitive to S-S relations (reviewed in Patterson
2001). But questions arose and few took the work
seriously until Thompson and his students reexamined
the issue 3 decades later. In spinally transected animals,
they paired a weak shock to the thigh (the CS) with an
intense shock to the toe pads (the US). Over trials, the
paired CS elicited a progressively stronger flexion
response (the CR). When the CS was presented alone
(extinction), the CR waned. Subsequent studies confirmed that learning depended on the CS-US pairing
and demonstrated a range of Pavlovian phenomena,
including differential conditioning, latent inhibition,
blocking, and overshadowing.
While most would agree that the spinal cord is
sensitive to S-S relations, questions remained. Those
wedded to the philosophical school of associationism
sought evidence that a truly neutral cue, that has no

capacity to elicit a CR-like behavior prior to training,
could support conditioning and acquire the ability to
produce a CR as a result of its association with the US.
From this view, any example of conditioning that
involved the modification of a preexisting reflex did
not demonstrate true associative learning and was relegated to the domain of conditioning artifacts with the
label of alpha conditioning. This view held sway for
many years and raised doubts as to whether spinal
neurons could support “true” Pavlovian conditioning,
because in the spinal preparation, the CSs routinely
elicited a CR-like response prior to training. This view
began to shift in the mid-1980s, with evidence of Pavlovian conditioning in the invertebrate Aplysia. Here
was an example of learning that depended on CS-US
pairing and for which Eric Kandel and his associates
were uncovering the neurobiological machinery. Yet,
purists were troubled because learning in this preparation depended upon the sensitization of a preexisting
(CR-like) response. At the same time, electrophysiological studies showed that paradigms thought to meet
the behavioral ideal of true conditioning utilized
a CS that generated CR-like neural activity prior to
training. As the power of these data became clear, the
distinction between true associative learning and alpha
conditioning loss sway. A more modern view
(neurofunctionalism) sees Pavlovian conditioning as
a powerful methodology, which can be used to demonstrate a system is sensitive to S-S relations, but recognizes that there are many ways in which a S-S relation
can be encoded by the nervous system and that there is
little reason to view one as superior to others (Grau and
Joynes 2005). As this view has gained strength, so too
has the acceptance that the spinal cord can learn.
The idea that spinal neurons are sensitive to
response–outcome (R-O; instrumental) relations has
proven equally controversial. In instrumental learning,
there is a relationship between a particular R (e.g., pressing a bar) and an O (e.g., obtaining a food pellet). As
a result of this R-O relation, a rat will typically exhibit an
increase in bar pressing. Similarly, if extending one hind
leg (the R) causes a pain-inducing shock (the O), the
subject will quickly learn to maintain the leg in a flexed
position that minimizes net shock exposure. What is
surprising to many is that the second example of learning does not require a brain.
Demonstrations of spinally mediated instrumental
learning typically use three groups. Master rats receive

Spinal Learning

shock whenever one hind limb is extended (controllable shock). Another group is experimentally coupled
(yoked) to the master rats and receives shock at the
same time, but independent of limb position (uncontrollable shock). A third group remains unshocked
during training. Master rats exhibit a progressive
increase in flexion duration over the course of 30 min
of training. Yoked rats respond, but do not exhibit an
increase in flexion duration (the index of learning).
Subjects are then tested with controllable stimulation
applied to the opposite leg. Master rats typically exhibit
some benefit from the earlier training (a form of positive transfer) relative to the previously unshocked controls. Yoked rats generally fail to learn, a form of
negative transfer reminiscent of learned helplessness.
The term “instrumental conditioning” has its roots
in the reflexive tradition of Edward L. Thorndike and
Clark L. Hull. An alternative term, operant learning,
was coined by B. F. Skinner, who saw this category of
behavior as being emitted, not elicited (the latter he
called respondent behavior). In an ideal operant situation, a range of behaviors could be arbitrarily shaped
using a variety of outcomes (reinforcers, both appetitive and aversive). In reality, most examples of operant
behavior are more biologically constrained than Skinner realized. R-O learning in the spinal cord depends
on an elicited behavior and lacks flexibility – it is highly
biologically constrained. It does, nonetheless, depend
on the R-O relation and meets the other criteria for
learning. Extracting ourselves from this conceptual
morass requires that we recognize: (a) that instrumental conditioning refers to a methodology that can be
used to demonstrate that the R-O relation matters; and
(b) that the terms instrumental and operant should not
be treated as synonyms. All examples of operant behavior will likely meet the behavioral criteria for instrumental behavior, but the converse is not true. It appears
that neural systems can encode R-O relations in many
ways and that simpler systems are generally more biologically constrained in their behavioral repertoire.
Over the last 20 years, researchers have uncovered
many of the neurobiological mechanisms that underlie
spinal cord plasticity. Much of this work has focused on
the identification of the neurochemical systems involved
in central sensitization. These studies have shown that
electrophysiological phenomena (e.g., long-term potentiation [LTP]) and neurochemical mechanisms (e.g.,
NMDA receptor-mediated plasticity) discovered within

S

3171

brain structures involved in learning and memory (e.g.,
the hippocampus) are also observed within the spinal
cord (Ji et al. 2003). Paradoxically, these neurobiological discoveries have led to a shift in thinking that
50 years of behavioral work could not accomplish,
leading many to assume that spinal neurons must be
capable of learning – because the required neurobiological mechanisms are present.

Important Scientific Research and
Open Questions
Current work in the field of spinal learning is focusing
on three issues: further detailing the behavioral limits
of spinal learning, examining the underlying neurobiological systems, and evaluating the clinical implications. New behavioral work has shown that spinal
neurons can discriminate regular (predictable) versus
irregular (unpredictable) stimulation. Further, learning
about temporal regularity appears to engage a protective effect that can enable subsequent learning and
counter the development of neuropathic pain. Studies
examining the neurochemical systems involved in central sensitization are leading to the development of new
drug treatments to counter pain. Other work has shown
that stimulation after a spinal cord injury can influence
recovery; just as uncontrollable shock impairs subsequent learning in transected rats, it inhibits recovery
after a contusion injury (a bruising of the spinal cord
caused by a moderate impact that emulates the type of
injury often observed in humans). Researchers have also
discovered that neurons within the lumbosacral spinal
cord can support stepping behavior and that this system
can be retrained after injury. This discovery has led to
new rehabilitative techniques that emphasize behavioral
training (e.g., stepping on a treadmill) after injury and
has been shown to foster recovery.

Cross-References
▶ Habituation
▶ Learned Helplessness
▶ Operant Behavior
▶ Pavlovian Conditioning

References
Grau, J. W., & Joynes, R. L. (2005). A neural-functionalist approach
to learning. International Journal of Comparative Psychology, 18,
1–22.
Groves, P. M., & Thompson, R. F. (1970). Habituation: A dualprocess theory. Psychological Review, 77, 419–450.

S

3172

S

Spinal Plasticity

Ji, R. R., Kohno, T., Moore, K. A., & Woolf, C. J. (2003). Central
sensitization and LTP: Do pain and memory share similar mechanisms? Trends in Neurosciences, 26, 696–705.
Patterson, M. M. (2001). Classical conditioning of spinal reflexes: The
first seventy years. In J. E. Steinmetz, M. A. Gluck, & P. R.
Solomon (Eds.), Model systems and the neurobiology of associative
learning: A Festschrift in honor of Richard F. Thompson (pp. 1–22).
Mahwah: Lawrence Erlbaum.
Patterson, M. M., & Grau, J. W. (2001). Spinal cord plasticity: Alterations in reflex function. London: Kluwer.
Sherrington, C. S. (1905). The integrative action of the nervous system.
New Haven: Yale University Press.

Spinal Plasticity
▶ Spinal Learning

Spirituality
Use of this term is currently very popular. It has, however, no universally accepted definition. Some authors
understand spirituality in contrast to religion and religiosity emphasizing the experience of transcendence.
Other authors sustain that religiosity and spirituality
are not to be seen as in contrast to each other but rather
as overlapping.

Split-Attention Effect
PAUL AYRES1, GABRIELE CIERNIAK2
1
School of Education, University of New South Wales,
Sydney, Australia
2
Knowledge Media Research Centre, Tübingen,
Germany

Synonyms
Spatial contiguity; Split-attention principle; Temporal
contiguity

Definition
Split-attention occurs when learners are required to
split their attention between two or more mutually

dependent sources of information (e.g., text and diagram), which have been separated either spatially or
temporally. If information of each source is essential for
understanding the topic, all information given must be
mentally integrated by the learner for learning to occur.
However, this forced integration process increases
demands on the learner’s working memory (WM)
and can impact negatively on learning. To create effective learning environments instructional designers
must avoid split-attention by externally integrating
the different sources of information together into a
single integrated source of information. For example,
an integrated format can be achieved by embedding
written instructions within a diagram (avoiding spatial
separation) or aligning spoken text with the targeted
picture during a multimedia presentation (avoiding
temporal separation). The split-attention effect occurs
when a single integrated source of information
enhances knowledge acquisition better than separated
sources of information. The split-attention principle
says that several separated sources of information
should be replaced with a single integrated source of
information.

Theoretical Background
Cognitive load theory (CLT, see Chapter ▶ Cognitive
Load Theory) has defined three types of cognitive load:
intrinsic, extraneous, and germane (Sweller et al. 1998).
Intrinsic cognitive load is the load imposed on WM by
the complexity of the materials to be learned. Complexity depends on the number of interacting elements.
Element interactivity is measured by the number of
elements of information that must be simultaneously
processed in WM for learning. Materials with high
element interactivity are difficult to learn. Extraneous
cognitive load is the load placed on WM by the way the
instructional designers construct the learning materials. Inefficient and badly designed materials can be
harmful to learning as scarce WM resources are taken
up by unnecessary processing. Germane cognitive load
consists of the WM resources directly invested in dealing with intrinsic cognitive load (schema creation). By
keeping extraneous load to a minimum and maximizing germane load, effective instructional designs can be
created.
Research into CLT has identified a number of
strategies that decrease extraneous cognitive load.
One such strategy is the integrated format to avoid

Split-Attention Effect

split-attention. Split-attention was first identified by
Tarmizi and Sweller (1988) who found that conventionally structured worked examples (see Chapter on
the ▶ Worked Example Effect) were an ineffective
method to learn geometry. This finding was unexpected because up to that point, worked examples
had been found to be highly effective in other mathematical domains such as algebra. To explain this failure,
Tarmizi and Sweller argued that when the textual solutions were written below the diagram, cognitive load was
increased. Because of the spatial separation between
diagram and text, learners studying the worked examples were forced to mentally integrate information on
the diagram with the written solution steps. It is important to note that both sources of information were
unintelligible by themselves. Both were needed to fully
understand the solution. The mental integration of the
two sources increased WM load (creating an extraneous
cognitive load) because of this unnecessary processing.
In particular, learners were forced to search for locations
in both the text and diagram and match them. Search
processes require WM resources. Traditional forms of
geometry solutions usually involve both a diagram and
separated lines of mathematical text (solution steps).
However, Tarmizi and Sweller overcame this problem
by integrating the text into the diagram. Using this
integration strategy, worked examples were then shown
to be effective.
Other CLT-based research extended the findings
into others domains of mathematics and science (see
Sweller et al. 2011). One particular study of importance
was conducted by Chandler and Sweller (1991) into
learning about the human heart and lungs. They found
that an integrated format was not superior to a splitattention format if redundant information was
included. In their experiment, both the diagram and
text contained information that was identical. In this
case the text was not needed to understand the diagram. Only when the redundant text parts were
removed did the integrated format become superior
to a split-attention format. The redundancy effect is
another CLT effect that has been widely researched
(see Sweller et al. 2011). The redundancy finding was
important because it demonstrated that the splitattention effect is severely compromised by redundant
information.
The research reported so far describes experiments
that were conducted solely with paper-based materials.

S

Later research into the split-attention effect also used
computers. In a number of experiments completed by
Sweller and Chandler (1994, 1996) computer manuals
were shown to be a common source of split-attention.
Reading from a manual while attending to information
on the computer screen while performing various computer actions inevitably leads to split-attention. Sweller
and Chandler demonstrated that successful integration
could be achieved by disregarding the computer altogether by using a modified manual that included integrated texts and diagrams of the computer screen and
keyboard. Further research demonstrated that the integrated materials could also be solely displayed on the
computer screen without a manual. A second finding in
the Sweller and Chandler studies was that the splitattention effect only occurred when the materials
were high in element interactivity (high intrinsic cognitive load). For low element interactivity materials,
the extraneous cognitive load caused by unnecessary
processing of separated sources was not sufficient to
overload WM and to interfere with learning.
Much of the CLT-based research into the splitattention effect has been conducted by John Sweller
and his colleagues; however, Richard E. Mayer has
also made a significant contribution to understanding
this effect. Earlier findings by Mayer showed that in
learning about how brakes work or how lightning
forms, words and pictures should be integrated
together rather than kept separately (for summaries
see Mayer 2005). In showing this effect, split-attention
scenarios were created by positioning text below the
related pictures diagrams or on a separate page. In
contrast, the integrated formats consisted of placing
text, which describes the action, close to the diagram
depicting the action. On transfer tasks, the integrated
formats were found to be superior to the split-attention
formats with large effect sizes. In Mayer’s studies, evidence was collected that integrating words and pictures
leads to superior learning compared with more spatially separated formats. Mayer called this the Spatial
Contiguity Principle: “People learn more deeply from
a multimedia message when corresponding text and
pictures are presented near rather than far from each
other on the page or screen.” (p. 195).
So far, the research reported above has only
included static presentations, whether they are text
and pictures, or a computer manual and a computer.
Moreover, parallel research has been conducted into

3173

S

3174

S

Split-Attention Effect

dynamic representations or animations. With more
dynamic pictures that show events over time, pictures
come and go as the scenarios unfold. Similarly, any
accompanying text, be it written or spoken, will also
be transitory. Whereas, transitory information provided through animations presents special challenges
to the learner; the synchronization of text and pictures
is crucial if split-attention is to be avoided. In cases
where text and pictures are separated on the screen, i.e.,
where both are not visible at the same time, WM
resources are particularly stretched. If important information, in one mode or the other, has disappeared
from the computer screen, the learner must try to
hold that information in WM for a period of time,
and then match it with the accompanying source
of information when it is eventually presented. Hence,
information needs to be held in WM as well as relevant
searches completed, to match and integrate connected
components. It may be that split-attention in an
animated environment may have a greater impact
than in a static one, where the sources of information can be more easily revisited. Therefore, multiple
sources of information that need to be integrated but
are separated in time also create an extraneous cognitive load.
The temporal separation of information has been
extensively researched by Mayer and his colleagues, and
is referred to as temporal contiguity. In particular,
research into the split-attention effect was extended to
include the inclusion of the spoken word (sound). How
narratives and pictures are presented together is
a crucial consideration in animated multimedia learning environments. In a series of studies (for summaries
see Mayer 2005) across different learning domains,
Mayer and his collaborators demonstrated that words
(narrations) and pictures were more effective if shown
simultaneously rather than temporally distant. It did
not matter if narration followed the relevant pictures,
or vice versa, narrating at the same time as the pictures
were shown (integrated format) led to superior learning. These findings, also with large effect sizes led
Mayer to articulate the Temporal Contiguity Principle:
“People learn more deeply from a multimedia message
when corresponding animation and narration are
presented simultaneously rather than successively”
(Mayer 2005, p. 195).
In summary, it can be concluded that separated
formats of different sources of information, whether

spatially or temporally separated, lead to inferior learning compared with integrated formats. The splitattention effect has very strong implications for
instructional design in multimedia environments.
Whether the medium is animated or static, contains
written or spoken text, an integrated approach to the
different information sources leads to the best learning
outcomes.

Important Scientific Research and
Open Questions
The meta-analysis by Ginns (2006) confirmed the
robustness of the split-attention effect. Split-attention
was equally likely to occur for spatial and temporal
separated materials, both having large effect sizes.
No differences were found between static and dynamic
materials. Interestingly, learning materials with high
element interactivity produced larger effect sizes than
those with low element interactivity. As Ginns points
out, the studies with low element interactivity tended
to lack statistical power, but some evidence did emerge
that the effect could occur with low element interactivity. This finding contrasts with the initial studies of
Sweller and Chandler who predicted that the splitattention effect, like most effects hypothesized within
the CLT framework, would disappear for low element
interactivity material. Consequently, this is one area of
research that merits further investigation: To what
extent does element interactivity moderate the splitattention effect?
A basic assumption made by CLT researchers is
that the split-attention effect is caused by learners
having to complete unnecessary processing by trying
to align the two or more separated materials together.
Thus, harmful extraneous cognitive load is created
from the number of visual search processes that
need to be completed during this cognitively demanding alignment. Although, the negative influence of
searching is a very plausible explanation, little direct
evidence confirming this hypothesis has been collected.
However, with the more recent development of sophisticated eye-tracking apparatus researchers are starting
to investigate the visual search patterns of learners
when confronted with split-attention materials. Hence,
another open question focuses on visual search: To
what extent is the split-attention effect caused by visual
search within and between the different sources of
information?

Spread of Activation Theory

Cross-References
▶ Cognitive Load Theory
▶ Example-Based Learning
▶ Multimedia Learning
▶ Worked Example Effect
▶ Working Memory

References
Chandler, P., & Sweller, J. (1991). Cognitive load theory and the
format of instruction. Cognition and Instruction, 8, 293–332.
Chandler, P., & Sweller, J. (1996). Cognitive load while learning to
use a computer program. Applied Cognitive Psychology, 10(2),
151–170.
Ginns, P. (2006). Integrating information: A meta-analysis of the
spatial contiguity and temporal contiguity effects. Learning and
Instruction, 16, 511–525.
Mayer, R. E. (2005). Principles for reducing extraneous processing
in multimedia learning: Coherence, signalling, redundancy, spatial contiguity and temporal contiguity principles. In R. E. Mayer
(Ed.), The Cambridge handbook of multimedia learning (pp. 169–
182). New York: Cambridge University Press.
Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory.
New York: Springer.
Sweller, J., & Chandler, P. (1994). Why some material is difficult to
learn. Cognition and Instruction, 12(3), 185–233.
Sweller, J., van Merriënboer, J. J. G., & Paas, F. (1998). Cognitive
architecture and instructional design. Educational Psychology
Review, 10, 251–296.
Tarmizi, R., & Sweller, J. (1988). Guidance during mathematical
problem solving. Journal of Education & Psychology, 80, 424–436.

Split-Attention Principle
▶ Split-Attention Effect

Spoken Language
▶ Verbal Behavior and Learning

Spontaneous/Commonsense
Ways of Reasoning
▶ Preconceptions and Learning

S

3175

Sport
▶ Effects of Exercising During Learning

Spread of Activation Theory
FABIO CRESTANI
University of Lugano, Lugano, Switzerland

Synonyms
Spreading activation theory

Definition
Spread of activation theory refers to mathematical theory of the spread of activation on associative networks,
neural networks, or semantic networks. Spread of activation is a method for searching associative networks,
neural networks, or semantic networks that is based on
supposed mechanisms of human memory operations
(Collins and Loftus 1975).

Theoretical Background
The Spread of Activation (SA) model in its pure form is
quite simple. It is made up of a conceptually simple
processing technique on a network data structure.
The network data structure consists of nodes
connected by links, as depicted in Fig. 1. Nodes model
objects or features of objects of the real world to be
represented. Nodes are usually labeled with the name of
the objects they intend to represent. Links model relationships between nodes and they can be labeled and/or
weighted. The connectivity pattern reflects the relationships between objects and/or features of objects of
the real world to be represented. A link usually has
a direction, a label, and/or a weight assigned according
to a specific direction. This representation structure is
very similar to a Semantic Network, but it is more
general than a Semantic Network and it could represent
a more generic Associative Network.
The processing technique is defined by a sequence of
iterations like the one schematically described in Fig. 2.
Each iteration is followed by another iteration until the
process is halted by the user or by the triggering of some
termination condition. An iteration consists of one or

S

3176

S

Spread of Activation Theory

i

Wij
j

Spread of Activation Theory. Fig. 1 The network
structure of a SA model

In the pre-adjustment and post-adjustment phases,
which are optional, some form of activation decay can
be applied to the active nodes. These phases are used to
avoid retention of activation from previous pulses,
enabling to control both activation of single nodes
and the overall activation of the network. In fact, they
implement a form of “loss of interest” in nodes that are
not continually activated.
The spreading phase consists of a number of passages of activation weaves from one node to all other
nodes connected to it. There are many ways of spreading the activation over a network (for an overview see
Preece 1981). In its more simple form, on a single unit
level, Spread of Activation consists first in the computation of the unit input, calculated as:
Ij ¼

Start

n
X

Oi  wij

i¼1

Pre-adjustment

Spreading
Pulse
Post-adjustment

Termination
condition

Not satisfied

Satisfied
Stop

Spread of Activation Theory. Fig. 2 The pure SA model

more pulses and a termination check. What distinguishes the pure Spread of Activation model from
other more complex models of searching on networks
is the sequence of actions which composes the pulse. In
fact, a pulse is made up of three phases:
1. Pre-adjustment
2. Spreading
3. Post-adjustment

where Ij is the total input of node j, Oi is the output of
unit i connected to node j, and wij is a weight associated
to the link connecting node i to node j.
The input and the weight are usually real numbers;
however, their numerical type is determined by the
specific requirements of the application to be modeled.
In particular, they can be binary values (0 or 1), excitatory/inhibitory values (+1 or 1), or they can be real
values indicating the strength of the relation between
nodes. Usually the first two of these options are used in
connection with networks with labeled links, like for
examples Semantic Networks, where the semantic
value of the relation represented by the link determines,
in the context of the application, the value to be associated to the link. The last option is mainly used for
Associative Networks, where there is only one generic
type of association that needs to be properly weighted.
After a node has computed its input value, its output
value must be determined. The numerical type of the
output of a node is also determined by the requirements
of the application. The two most used cases being the
binary active/non-active type (0 or 1) and the real
value type. In Spread of Activation models, there is
usually no distinction between “activation” or “output”
of a unit. The activation level of a unit is its output value.
This is usually computed as a function of the input value:
Oj ¼ f ðIj Þ
There are many different functions that can be used
in the evaluation of the output; some examples are

Spread of Activation Theory

Linear function

Step function

S

3177

Sigmoid function

Spread of Activation Theory. Fig. 3 Some commonly used activation functions

depicted in Fig. 3. The most commonly used function
in pure Spread of Activation models is the threshold
function. It is used to determine if the node j has to be
considered active or not, and thus produce an output.
The application of the threshold function to the above
formula in the case of binary value units gives:

0 Ij < kj
Oj ¼
1 Ij > kj
where kj is the threshold value for unit j. The threshold
value of the activation function is application dependent and can vary from node to node, therefore the
notation kj for the unit threshold has been used.
After the node has computed its output value, it
spreads it to all the nodes connected to it, usually
sending the same value to each of them.
Pulse after pulse, the activation spreads over the
network reaching nodes that are far from the initially
activated ones. After a determined number of pulses
have been fired, a termination condition is checked. If
the condition is verified the Spread of Activation process stops, otherwise it goes on for another series of
pulses. The Spread of Activation process is therefore
iterative, consisting of a sequence of pulses and termination checks.
The result of the Spread of Activation process is
the activation level of nodes reached at termination
time. The interpretation of the level of activation of
each node depends on the application and, in particular, on the characteristics of the object being modeled
by that node.
The pure Spread of Activation model presents some
important drawbacks:
● Unless controlled carefully by means of the pre-

adjustment and the post-adjustment phases the
activation ends up spreading all over the network.

● The information provided by the labels associated

to the links is not used, that is, there is no use of the
semantics of the associations. Thus, it is difficult to
implement some form of inference based on the
semantics of associations.
These problems can find solutions by taking into
account in the processing technique the diverse semantics of the relations among units. This can be achieved
using the information provided by the labels on the
links and by processing links in different ways
according to their semantics. It is possible in this way
to implement some form of heuristics, or to spread
activation on the network according to some inference
rules. A common way of implementing a processing
technique which spreads the activation according to
rules is by means of constraints on the Spread of Activation. Here are some constraints commonly used:
Distance constraints: The spreading of activation
should cease when it reaches nodes that are far away
in terms of links covered to reach them from the initially activated ones. This corresponds to the simple
heuristic rule that the strength of the relation between
two nodes decreases with their semantic distance. Relations can be classified according to their distance in
term of links. Relations between two nodes directly
connected are called first-order relations. Relations
between two nodes connected by means of an intermediate node are called second-order relations, and so on.
It is common to consider only first-, second- and, at
most, third-order relations, although this is application
dependent.
Fan-out constraints: The spreading of activation
should cease at nodes with very high connectivity, or
fan-out, that is at nodes connected to a very large
number of other nodes. The purpose of this constraint
is to avoid a too wide spreading which could derive

S

3178

S

Spread of Activation Theory

from nodes with a very broad semantic meaning and
therefore connected to many other nodes.
Path constraints: Activation should spread using
preferential paths, reflecting application-dependent
inference rules. This can be modeled using the weights
on links or, if links are labeled, diverting the activation
flow to particular path while stopping it from following
other less meaningful paths.
Activation constraints: Using the threshold function
at a single node level, it is possible to control the
spreading of the activation on the network. This can
be achieved by changing the threshold value in relation
to the total level of activation over the entire network at
any single pulse. Moreover, it is possible to assign
different threshold levels to each unit or set of units
in relation to their meaning in the context of the
application. Although this may cause an increase in
the number of computations, it makes possible to
implement various complex inference rules.
Referring to Fig. 2 these constraints can be seen as
acting during the pre-adjustment phase (this is the case
for distance, fan-out, and path constraints) or during
the post-adjustment phase (mainly for activation level
constraints). Therefore, they can be considered as an
enhancement of the pure Spread of Activation model.
Another more practical advantage deriving from
the fact that the activation does not spread over the
entire network is that it permits a reduction of
the computational effort of the spreading, because
only a small portion of the units become active and
send activation to other units.
A further enhancement of the pure SA model can be
obtained by means of feedback from an external source.
In this case, an external evaluation of the activation
level of some units or of the entire network provides
some constraints that would be difficult or impossible
to implement in the form of automatic rules. This
external feedback can be due to another process or
can be provided by the user of the system. The user
evaluates the activation level reached by some nodes
and modifies it according to his or her requirements.
This may result in a following spreading of activation
over particular user-selected paths that differ from
those specified by the predetermined path constraints.
From this point of view, the Spread of Activation model
adapts itself to the specific user’s needs.
This model is particularly useful in application
where there would be too many inference rules to be

represented in the form of constraints and where it is
necessary to provide an external control by means of
a user’s evaluation of the results. The use of feedback
from user can be implemented either in the preadjustment phase, so that the user directs the spreading
of activation of the pulse, or in the post-adjustment
phase, enabling the user to evaluate the result of the
spreading and direct the following pulse accordingly
(Preece 1981).

Important Scientific Research and
Open Questions
The Spread of Activation theory has been applied with
different level of success in different areas of research,
namely Cognitive Science, Databases, Artificial Intelligence, Psychology, Biology, and Information Retrieval
(IR). The application of the Spread of Activation theory to Information Retrieval is one of the most challenging and promising (see Crestani 1997 for a perhaps
out-of-date review of this area of research).
The use of Spread of Activation in IR is based on the
existence of maps specifying relations between terms or
between documents, as the case may be. Nodes correspond to terms, documents, articles, journals, subject
classifications, authors, and so forth. There is no
homogeneity in the network. A node can represent
anything. Links indicate the association of a node
with another node, as, for instance, an author with
a document he/she wrote or a document with
a document it cites. Specific link types include term
occurrence, document publication, term assignment by
indexing, document authorship, document assignment
to classification, document citation, and so forth. The
set of node types and link types is determined by the
data available and by the purpose of the application.
The representation structure is therefore application
specific and the same structure cannot be applied to
different applications. It is also important to note that
relationships could actually be expressed by a pairs of
links. Authorship, for instance, can be represented by
both “authored by” and “is author of ” links. Both links
in such pairs connect the same two nodes, but their
source and destination roles are reversed. Specific
processing rules may inhibit activation in either direction, use them in different ways, or associate different
weights with the different directions.
Given such a complex representation structure, it
has been demonstrated that the best retrieval results

Stability and Change in Interest Development

can be obtained by a Spread of Activation process
which uses some of the constraints described above
and, in particular, path constraints (Preece 1981).
Much of the effectiveness of the process is, however,
crucially dependent on the availability of a representative network. The problem of building a network which
effectively represents the useful relations in terms of the
Information Retrieval aims, has always been the critical
point of many of the attempts to use Spread of Activation in Information Retrieval. These networks are very
difficult to build, to maintain, and keep up to date.
Their construction requires in-depth application
domain knowledge that only experts in the application
domain can provide. Furthermore, their construction
is a very expensive and time-consuming process that
results almost impossible for large collections and/or
collections spanning over a large application domain.

S

3179

St Thomas
▶ Aquinas, Thomas (1225–1274)

Stability and Change in Interest
Development
DOMINIC R. PRIMÉ, TERENCE J. G. TRACEY
Counseling Psychology, Arizona State University,
Tempe, AZ, USA

Synonyms
Career interests; Structure of interests; Vocational
interests

Cross-References
▶ Associationism
▶ Measures of Association
▶ Semantic Networks

References
Collins, A., & Loftus, E. (1975). A spreading-activation theory of
semantic processing. Psychological Review, 82(6), 407–428.
Crestani, F. (1997). Applications of spreading activation techniques
in Information Retrieval. Artificial Intelligence Review, 11(6),
453–582.
Preece, S. (1981). A spreading activation model for Information
Retrieval. PhD thesis, University of Illinois, Urbana-Champaign

Spreading Activation Theory
▶ Spread of Activation Theory

SR Learning
▶ Human Contingency Learning

Definition
Interest development is an area of scientific inquiry,
which seeks to explain the normative developmental
processes underlying how people become interested
and subsequently choose to pursue different vocations
and careers. Vocational interests date back to Parsons’
(1909) model of assisting individuals in matching their
interests and abilities to specific occupations and are
generally viewed as underlying motivational factors
that manifest in preferences for different activities.
The common assumption with regard to vocational
interests is that the better the match of one’s interests
and abilities with an occupation, the better the person’s satisfaction and the greater the job performance.
Interest development is most often evaluated through
the use of psychological interest measures, which
assess preferences (i.e., liking) or aversion (i.e.,
disliking) related to topics such as, school subjects,
occupational preferences, occupational expectations,
anticipated occupations, activities, and general interests. Researchers who examine stability and change in
interest development are primarily concerned with
changes in interest over time and throughout the
lifespan.

Theoretical Background

S-R Learning
▶ Routinized Learning of Behavior

Interests are represented using two very different
approaches: one empirical and the other theoretical.
The Strong Interest Inventory (Harmon et al. 1994) is

S

3180

S

Stability and Change in Interest Development

perhaps the best example of the empirical approach, as
it uses a wide variety of empirically derived items to
match individuals with occupations. Conversely, the
theoretical approach measures theory-based constructs
derived from a much smaller number of aspects that are
then used to understand interests and apply these to the
world of work. John Holland’s theoretical model
(1985) serves as the most prominent example of this
approach and is widely adopted throughout the field of
interest development. Holland posits that six different
personality/interest types exist – the six types are Realistic, Investigative, Artistic, Social, Enterprising, and
Conventional, commonly referred to using the initials
RIASEC – and that individuals express differing
amounts of each of these types.
There are two general approaches to differentiating
interests: person-environment models (Holland 1985)
view interests as a fairly stable trait-like expression to be
assessed and matched to an environment, while developmental models (e.g., Savickas 1999) view interests as
a function of maturation and interaction with the
environment. Thus, one set of models stresses stability
(person-environment) while the other set of models
(developmental) stresses change. Although proponents
of each theoretical school interpret the relevance of
stability and change in interests in a somewhat different
manner, most vocational theorists agree that the developmental focus of interest development necessitates an
examination of both stability and change because it is
not necessarily true that stability is defined by the
absence of change.
Any examination of change or stability over time is
not complete unless the following five types of change
are examined. Change is most often defined by
a change in mean scores and the examination of how
an individual’s scores increase or decrease over time –
this type of examination is also known as absolute
stability. This is different from the examination of relative stability, which is usually assessed using test-retest
correlations to study changes (or the lack thereof) in
one’s relative position in a distribution over time.
Where absolute stability is concerned with changes in
an individual’s scores, rank-order stability is concerned
with one’s position within a distribution and whether
that position has remained relatively similar to others
within the distribution. This concern with the rank
ordering of an individual is also often referred to as
rank-order stability. A third-type of change is the

change of relative ordering of different scale scores
relative to one another (i.e., profile stability). Profile
stability focuses on the relative profile shape of an
individual and whether changes in relative subscale
order remain similar upon consecutive administrations
of a measure. For example, if an individual’s profile
scores remain relatively unchanged upon consecutive
administrations, then this result is interpreted as evidence of profile stability. Another type of change is
structural change, which examines the interaction
between an individual’s scores and the structure of
a specific psychological measure. Structural change is
concerned with the relations among the scales and
items found on a particular measure and its proponents presuppose that evidence of change resides
within the instrument used to assess a particular construct. The findings of Tracey and Ward (1998) are an
example of a change in the structure of a scale where the
factor analyses of young children were different from
those of adolescents and, thus, there were observed
changes in the relation of the items in a scale. The
final type of change is external to any scale and involves
change in the relation of the scale with outside variables;
in career assessment this is often achieved by examining
relations of interest scales with outside variables such
as, self-efficacy, career decision-making, and career
choice.

Important Scientific Research and
Open Questions
Scientific research examining stability and change in
interest development indicates that there is evidence of
stability in adult populations. Conversely, in the period
of childhood and adolescence there is evidence of both
stability and change. In general, most researchers
studying interest development agree that interests
become more stable the older people become, eventually reaching a plateau around the ages of 18–20 years
of age. The commonly held view is that interests are
highly stable and change little after the ages of 25–30
years. However, questions remain about normative
developmental trajectories in childhood and adolescence with regards to interest development. Longitudinal research in this area has shown at least moderately
stable differences in interests across time when examined from both interindividual and intraindividual
(i.e., profile) perspectives. When considered alongside
the relatively stable scores witnessed among children

Stability and Change in Interest Development

and adolescents – especially at the younger ages – these
results may be viewed as support for the trait-like view
of interests.
Important mean differences in interests over time
have also been witnessed. There appear to be a pattern
of increasing interest scores on all scales during the
elementary school years, usually followed by a drop in
middle school, and an increase through high school.
Interestingly, times of transition often induce drops in
interest scores – entry into middle school and the
senior year of high school have both been shown to
be times in which scores typically drop. These drops are
greater for girls than boys, especially in the areas of
scientific interests, which may indicate that girls alter
their interests in response to environmental change
more than do boys.
Interesting structural differences between children
and adults also exist. Young children tend to view
interests using the dimensions of sex-typing (i.e., is
the activity a boy activity or a girl activity?) and activity
locus (i.e., does the activity occur in school or out of
school?). However, as children age, a pattern has been
witnessed using aggregate examinations as well as individual ideographic examinations toward increasing fit
with RIASEC structure and types – a valid and widely
endorsed proxy for normative interest development in
adolescents and adults. Furthermore, structural stability tends to be stable by eighth grade and is similar
across gender and ethnicity, enabling the valid use of
RIASEC scores with all groups, at least in the United
States (see Rounds and Tracey 1996, for a review of
the structure of RIASEC scores in other countries and
cultures).
In younger populations, the relation between interest and competence perceptions (i.e., self-efficacy) has
emerged as one with several implications for assessment and intervention. In adult populations, interest
and competence perceptions appear to be interrelated
and seem to affect each other increasingly over time.
Conversely, in younger populations, children do not
always display this same way of thinking about interest
and competence perceptions (i.e., they do not always
indicate that they are interested in the things they
believe they are good at and vice versa). Since the
relationship between interests and competence perceptions has been shown in all age groups to be bidirectional and equal in magnitude – the effect of competence
perceptions on interest is equal to the effect of interest

S

on competence perceptions – interventions geared
toward increasing both interest and competence perceptions appear to be warranted. Similarly, whereas adults
are able to identify and differentiate those activities (i.e.,
their likes and competence) that appeal to them from
those that do not, children oftentimes tend to like or
dislike everything in an undifferentiated manner (i.e.,
they may indicate that they like most everything or
dislike most everything equally). This trend also
decreases with age and research has shown that the
older individuals become, the more discriminating
they tend to be about their preferences.
Another area of inquiry in stability and change in
interest development is adherence to normative structure (i.e., the extent to which an individual thinks
about interests in the same manner as most people)
and its effect on career exploration. Findings in this
area indicate that children engage in less career exploration than their adolescent and adult counterparts,
and individuals in college whose thinking deviates
from the normative structure of interests exhibit more
career indecision than those who do adhere to the
normative structure. These findings lend themselves
to further areas of inquiry into the effect of structural
adherence on interests and career exploration at key
points over time (e.g., how does structural adherence
affect career exploration at times when interest is low –
for instance the first year of middle school), which may
be key in guiding interventions designed to have an
effect on variables such as the aforementioned relationship between interests and self-efficacy.
Although a great deal is known about stability
and change in interest development, there are still
many questions to be answered, specifically about the
finer discriminations that occur as a result of normative
and nonnormative developmental trajectories. Recent
research has provided important insights into the
process of interest development – indeed, insights
that provide a more thorough and multidimensional
representation of the process of interest development –
however, it is essential that more is learned about what
should be expected to change at what times and to
which populations interventions should be geared.

Cross-References
▶ Interests and Learning
▶ Self-Efficacy and Learning
▶ Vocational Learning

3181

S

3182

S

Stabilize

References
Harmon, L. W., Hansen, J. C., Borgen, F. H., & Hammer, A. L. (1994).
Strong interest inventory. Stanford: Stanford University Press.
Holland, J. L. (1985). Making vocational choices: A theory of vocational
personalities and work environments (2nd ed.). Englewood Cliffs:
Prentice-Hall.
Parsons, F. (1909). Choosing a vocation. Boston: Houghton Mifflin.
Rounds, J., & Tracey, T. J. (1996). Cross-cultural structural equivalence of RIASEC models and measures. Journal of Counseling
Psychology, 43, 310–329.
Savickas, M. L. (1999). The psychology of interests. In M. L. Savickas
& A. R. Spokane (Eds.), Vocational interest: Meaning, measurement, and counseling use (pp. 19–56). Palo Alto: Davies-Black.
Tracey, T. J. G., & Ward, C. C. (1998). The structure of children’s
interests and competence perceptions. Journal of Counseling Psychology, 45, 290–303.

State-Dependency
▶ Mood-Dependent Learning

Statement
▶ Communication Theory

Static Mapping
▶ Initial State Learning

Stabilize
▶ Memory Consolidation and Reconsolidation

Stages of Internalization
Phases (or steps) of transformation of initially external
action mediated by material or materialized tools into
a mental (ideal) plan. Galperin introduced six stages of
internalization: (1) formation of a motivation base of
action; (2) formation of an orientation base of action;
(3) formation of the material (materialized) form of
action; (4) formation of the external socialized verbal
form of action (overt speech); (5) formation of the
internal verbal form of action (covert speech); (6) formation of the mental action; final changes, automatization, and synchronization of the action.

Star-Learning
▶ AQ Learning

State Anxiety
▶ Effects of Anxiety on Affective Learning

Statistical Learning in
Perception
NICHOLAS B. TURK-BROWNE
Department of Psychology, Princeton University,
Princeton, NJ, USA

Synonyms
Associative learning; Contextual cueing; Segmentation;
Unsupervised learning

Definition
In cognitive psychology and cognitive neuroscience,
statistical learning (SL) refers to the extraction of regularities in how features and objects co-occur in the
environment over space and time. Such learning may
be important for detecting and representing higherorder units of perception, such as words, scenes, and
events. SL is defined by three criteria: First, it can
operate over undifferentiated input, where only spatial
and temporal probabilities can be used to determine
which parts of the environment go together; other
segmentation cues, such as grouping, are not required.
Second, SL occurs incidentally as a by-product of perception, without intentional effort or conscious awareness. Third, SL is concerned with extracting how

Statistical Learning in Perception

particular features and objects co-occur, resulting in
knowledge about relationships between specific stimuli. These properties make SL well suited to the continuous and noisy sensory input we receive from the
world.

Theoretical Background
Perception is concerned with interpreting information conveyed by our senses about the external environment, whether listening to music, enjoying a good
meal, or recognizing a friend’s face. Because we live in
natural environments that have been stable for thousands of years, and we tend to inhabit the same artificial environments for years at a time, our perceptual
systems are repeatedly confronted with very similar
sensory information. Over time, repeated aspects
of the environment – regularities – tune perception
to the types of information most frequently encountered. Such learning may be critical for handling the
deluge of sensory information that we experience
from moment to moment, allowing for faster and
more veridical perception of features and objects
that appear briefly (e.g., due to eye movements or
motion), or under variable or degraded conditions
(e.g., due to changes in lighting or occlusion by
other objects).
There are many types of regularities. Nature contains regularities in the layout of information (e.g.,
landscapes have horizontal horizons) and in the laws
governing physical interactions (e.g., detached objects
must be supported from below). Such natural regularities have altered perceptual systems over evolutionary time. For example, neurons in primary visual
cortex are tuned to the properties of images regularly
encountered by our species. Regularities also exist in
terms of which types of information appear together.
For example, academic robes are worn at commencements but not at the beach. Such semantic regularities
are learned during our lifetime, and facilitate the recognition of objects that appear in appropriate contexts.
Finally, regularities exist in terms of which specific
tokens of information appear together. For example, a
consistent sequence of landmarks is passed when navigating to a location: “the restaurant is after the cherry
blossom tree, the independent bookstore, the busker
playing saxophone, on the left”. Importantly, such
groupings are arbitrary (no semantic relationship
between the objects), and the resulting knowledge is

S

specific to the learned exemplars (a different tree, store,
or person would not be helpful). The focus of this entry
is on SL of regularities of this latter kind because:
they are prevalent, they can be learned quickly in
a laboratory setting, and learning is uncontaminated
by preexisting knowledge of natural and semantic
regularities.
Research on SL has roots in language acquisition.
Early work focused on whether the boundaries between
morphemes (smallest unit of meaning in a language)
could be recovered from how phonemes (basic sounds
in a language) are distributed/co-occur over time. The
use of such regularities in language acquisition was first
empirically tested in developmental psychology
(Saffran et al. 1996). In particular, the speech input
we get as listeners does not have reliable markers for
where words start and stop, and so a key challenge for
infants is not only to learn words and their meanings,
but to locate the words in a speech stream in the first
place. To test whether infants rely on statistical regularities to find word boundaries, a fake language can
be constructed from arbitrary three-syllable words,
such as tupiro, golabu, bidaku, padoti, which are then
combined into a speech stream without pauses between
words (e.g., bidakupadotigolabubidakutupiro. . .). Without such segmentation cues, learning of words
requires extracting the transitional probabilities
between syllables, that is, detecting that transitions
within a word (da ! ku) have higher probabilities
than those spanning two words (ku ! pa). After two
minutes of exposure, infants can tell actual words
apart from non-words – syllable sequences that did
not occur during listening (e.g., tulaku) or occurred
with lower probability (e.g., dakupa). Importantly,
each individual syllable in the words and non-words
is equally familiar, yet the infants expressed additional
familiarity with the words. This study was used to argue
that infants have powerful learning abilities, and
that language acquisition may be partly experiencedependent.
Just as words can be defined by regularities of syllables, there is an analogous “language” of vision:
Meaningful units of visual perception, such as scenes
and events, can be defined by regularities of objects.
Learning of visual regularities has been investigated for
both temporal sequences and spatial configurations
(e.g., Fiser and Aslin 2001). Temporal sequences arise
in vision because events and actions (e.g., catching

3183

S

3184

S

Statistical Learning in Perception

a flight) are defined by a specific progression of information (e.g., parking garage, ticket agent, metal detector, gate, seat), and because our eyes get information
from one part of space at a time – resulting in
a temporal sequence as we move our eyes around. SL
of temporal sequences in vision is tested by showing
a continuous stream of nonsense shapes (e.g., each
letter is a shape: ABCGHIDEFABCJKL. . .) that, without observers’ knowledge, has been constructed from
shape triplets (e.g., ABC, DEF, GHI, JKL). After a few
minutes of exposure to the stream, familiarity is higher
for triplets (e.g., ABC) than for recombinations of the
same shapes (e.g., AEI). Again, since each individual
shape is equally familiar, additional familiarity for triplets indicates statistical learning of the shape transitions. Spatial configurations (e.g., the layout of
objects in a room) contain regularities in the relative
positions of objects (e.g., a lamp appears next to the
sofa), and learning of such regularities may facilitate
visual search and spatial navigation. SL of spatial configurations in vision is tested by showing grids of multiple shapes that, without observers’ knowledge, have
been constructed from shape pairs (if one member of a
pair appears on the grid, its associate also appears in a
fixed relation; e.g., to its left). Since many shapes appear
in each grid, pairs can only be learned by extracting
probabilities. After brief exposure to the grids, familiarity is higher for pairs than for recombinations of the
same shapes into new arrangements, providing evidence of SL.
SL is ubiquitous in perception, occurring: (1) in
auditory, visual, and tactile modalities; (2) for a range
of statistics, including frequency and conditional probability; (3) throughout human development; (4) in
other species, such as rats and monkeys; and (5) for
many stimuli, from lower-level features to higher-level
objects, scenes, and actions. The broad scope of SL
suggests that it is a core cognitive function.

Important Scientific Research and
Open Questions
Beyond demonstrating the scope of SL, research has
begun uncovering the underlying nature of this process. For example, SL can occur without conscious
intent or awareness, but is constrained by selective
attention; SL involves memory systems linked to
other types of implicit learning, such as artificial

grammar learning and motor sequence learning; and
SL happens quickly, after only a handful of repetitions
of a regularity. This is a relatively nascent field of
research, and thus three areas of active study are
highlighted here.
First, while initial studies of SL used a single set
of auditory or visual regularities, the environment contains multiple concurrent sets of regularities. For example, words in multiple languages, or the layouts of
different cities. At a more basic level, sensory information arrives at the same time from multiple modalities
(e.g., faces and voices), and about different feature
types within the same modality (e.g., shape and
color). It is important to assess how SL copes with
this complexity to determine whether it is suited to
more ecological settings outside of the laboratory.
Studies have examined SL of regularities appearing
simultaneously in different modalities or features, finding that if modalities/features are correlated (e.g.,
shapes appear in reliable colors) SL proceeds over
multi-modal/multi-feature objects. In addition to
appearing simultaneously, regularities can also appear
sequentially, such as when learning a second language
after already knowing a first. When presented with
a speech stream constructed from one set of threesyllable words, followed immediately by a second
stream constructed from a recombination of the
same syllables into new words, learning occurs only
for the first language (Gebhart et al. 2009) – in essence,
SL of one set of regularities blocks learning of a second
set. When two languages are separated in time by
a pause, learning occurs for both languages. This
raises an interesting possibility that only one set of
regularities can be learned in a given “context,” but
that pauses and other contextual shifts may allow SL
to restart. What counts as a context, and how much
evidence about new regularities is required to overcome previously acquired knowledge remain open
questions.
Second, most studies of SL test learning with the
exact same regularities that were trained. For example,
if presented with a triplet of shapes ABC during the
initial learning phase, the test phase would examine
familiarity for ABC. Without making changes between
learning and test, it is unclear what has actually been
learned. On one hand, SL could be very specific,
resulting in knowledge that A is followed 1 s later by

Statistical Learning Theory

B and then 1 s later by C. Alternatively, SL could be
more abstract, resulting in knowledge that A, B, and
C “belong together.” These possibilities can be tested
using transfer logic: if regularities can be recognized
despite a change between learning and test, then the
changed aspect is not an important part of what has
been learned. Some changes do not matter: SL readily
transfers to new temporal orders (training on ABC,
testing on CBA), and even after the removal of all
temporal information (training on ABC presented
sequentially, testing on ABC presented simultaneously), suggesting that SL produces abstract knowledge that generalizes flexibly (Turk-Browne and Scholl
2009). Other changes do matter: after training on large
spatial groups (shape triplets or quadruplets), SL does
not transfer to the embedded subcomponents (constituent shape pairs), suggesting that SL of spatial configurations operates hierarchically at the largest spatial
scale (Orbán et al. 2008). Characterizing what is
learned during auditory SL remains an open and
important question. For example, the transfer of visual
SL to new temporal orders would be disastrous for
word learning – reversing the syllables in any word
changes its meaning, and most often produces
a nonsense word.
Third, what purpose does SL serve beyond increasing familiarity? For auditory perception, segmentation of otherwise fluent speech during language
acquisition could be a valuable consequence of SL.
Segmentation is also important in visual perception,
both for extracting edges, surfaces, and objects from
the background, and for finding boundaries between
events and actions – and again SL may be helpful.
Another fundamental consequence of SL for perception is that the brain can use knowledge about regularities to anticipate what will appear next in the
environment, allowing for more efficient processing
when the predicted sensory information arrives.
Objects that reliably predict what will appear next in
a temporal sequence activate the anterior hippocampus
and prime regions of visual cortex that are selective for
the category of the predicted object. For example, faces
that are predictive of scenes elicit anticipatory activity
in scene-selective parahippocampal cortex (TurkBrowne et al. 2010). While the hippocampus is
involved in explicit types of future-oriented behavior,
such as planning and prospection, and category-

S

3185

selective visual regions are activated by effortful imagery, SL occurs without conscious awareness and thus
triggers implicit perceptual anticipation. What consequences SL has for other cognitive processes, such as
attention and working memory, remains an open and
actively studied question. In sum, there are reciprocal
interactions between SL and perception: SL occurs
incidentally as a result of perception; in turn, learned
regularities facilitate perception by allowing the brain
to anticipate the future.

Cross-References
▶ Attention and Implicit Learning
▶ Exposure-Based Perceptual Learning
▶ Implicit Sequence Learning
▶ Incidental Learning
▶ Language Acquisition and Development
▶ Relational Learning
▶ Spatial Learning
▶ Task-Irrelevant Perceptual Learning
▶ Unconscious Learning
▶ Visual Perceptual Learning

References
Fiser, J., & Aslin, R. N. (2001). Unsupervised statistical learning of
higher-order spatial structures from visual scenes. Psychological
Science, 12, 499–504.
Gebhart, A. L., Aslin, R. N., & Newport, E. L. (2009). Changing
structures in midstream: Learning along the statistical garden
path. Cognitive Science, 33, 1087–1116.
Orbán, G., Fiser, J., Aslin, R. N., & Lengyel, M. (2008). Bayesian
learning of visual chunks by human observers. Proceedings of
the National Academy of Sciences, 105, 2745–2750.
Saffran, J. R., Aslin, R. N., & Newport, E. L. (1996). Statistical learning
by 8-month-old infants. Science, 274, 1926–1928.
Turk-Browne, N. B., & Scholl, B. J. (2009). Flexible visual statistical learning: Transfer across space and time. Journal of Experimental Psychology. Human Perception and Performance, 35,
195–202.
Turk-Browne, N. B., Scholl, B. J., Johnson, M. K., & Chun, M. M.
(2010). Implicit perceptual anticipation triggered by statistical
learning. Journal of Neuroscience, 30, 11177–11187.

Statistical Learning Theory
▶ PAC Learning

S

3186

S

Statistical Learning Theory and Induction

Statistical Learning Theory and
Induction
GILBERT HARMAN1, SANJEEV KULKARNI2
1
Department of Philosophy, Princeton University,
Princeton, NJ, USA
2
Department of Electrical Engineering, Princeton
University, Princeton, NJ, USA

Synonyms
Nondeductive reasoning; Pattern classification; Pattern
recognition; Supervised learning

Definition
Induction is here taken to be a kind of reasoning from
premises that may not guarantee the truth of the conclusion drawn from those premises. It is to be distinguished from “mathematical induction” which is
a kind of deductive reasoning. The philosophical problem of induction is whether and how inductive reasoning can be justified.
Statistical learning theory (SLT) is a mathematical
theory of a certain type of inductive reasoning involving learning from examples. SLT makes relatively minimal assumptions about an assumed background
probability distribution responsible for connections
between features of examples and their correct classification, the probability that particular examples will
occur, etc. The theory seeks to describe various learning
methods and say how well they can be expected to do at
producing rules with minimum expected error on new
cases.
Among the topics discussed in statistical learning
theory are various nearest neighbor rules, related kernel
rules, feed-forward neural networks, empirical risk
minimization as a kind of inductive generalization,
ways of balancing data-coverage against “simplicity,”
PAC learning, support vector machines, and boosting.

Theoretical Background
A valid deductive argument has a kind of conditional
reliability: The conclusion is guaranteed to be true if
the premises are true. Inductive reasoning typically
lacks such a guarantee. The philosophical “problem of
induction” asks whether any sort of justification is
possible for inductive reasoning.

Clearly there cannot be a deductive demonstration
that inductive reasoning has the same sort of conditional reliability as deduction, because then there
would be no difference between induction and deduction. But there clearly is a difference.
It might be argued that induction is justified
because it works. How do we know it works? Because
it has worked in the past – not always, but much of the
time? It will be objected that this sort of justification
appeals to induction and is therefore circular. (Similarly, a deductive justification of deduction would be
circular.) But maybe it is possible to give some sort of
deductive justification of induction.
Of course, nothing can be deduced about induction
in the absence of substantive assumptions. So the only
serious philosophical problem of induction asks what
sorts of assumptions yield interesting conclusions.
SLT offers a kind of deductive answer to the problem of induction by characterizing various kinds of
induction and then deriving theorems about these versions. While basic results in SLT theory go back to the
1960s, this continues to be an active area and further
research is ongoing (e.g., see, Devroye et al. 1996;
Vapnik 2000; Hastie et al. 2009; and references therein).
In addition to offering a response to the philosophical
problem, SLT has many practical applications.
SLT applies to inductive pattern recognition, function estimation, as well as other inductive issues. We
concentrate here on pattern recognition, which
involves using data in order to learn to classify examples on the basis of their features. We also suppose for
simplicity that the relevant classification is YES/NO: Is
this character an “e”? Does this picture show a face?
The data consist in a set of labeled examples. Each
example has a set of its features and a label giving the
correct classification of the example. In one basic case,
it is also assumed that there is a single background
statistical probability distribution that characterizes the
probabilities that particular instances will arise and the
probabilistic relations between features and correct classifications. Furthermore, it is assumed that the same
probability distribution is responsible both for the data
and for new cases that arise. Minimal assumptions are
made about this probability distribution; for example, in
the basic case it might be assumed that the instances are
independent and identically distributed.
Given the unknown probability distribution as the
background for a pattern recognition problem, there is

Statistical Learning Theory and Induction

a minimum expected error R for rules associating
labels with features. (A rule with this minimum
expected error rate R is called a Bayes Rule.) It follows
immediately that induction cannot arrive at a rule with
a smaller expected error than R and the question is
then how close can one or another inductive rule come
to that minimal error rate.
Items are assumed to have some number d of features, each of which can take several, perhaps infinitely
many, values. We can identify the possible values of
features with real numbers. Then to specify the values
of all the relevant features of a particular item is to
specify a point in a d-dimensional feature space.
A simple classification rule is the l-nearest neighbor
rule. Given n labeled examples, the rule assigns to any
new instance the same label as the nearest labeled
example. There are also k-nearest neighbor rules
which assign to any new instance the label that
a majority of the k-nearest neighbors have (assuming
k is an odd number). And there are kn-nearest neighbor
rules, where the number of nearest neighbors considered is a function of n.
It can be shown that no matter what the background probability distribution, the limit of the
expected error of the 1-nearest rule as n ! 1 is no
more than 2R . And, it can be shown that the expected
error of a kn-nearest neighbor rule approaches R as
n ! 1 if kn ! 1 and kn/n ! 0 as n ! 1, no matter
what the background probability distribution. So that
rule is said to be universally consistent. However, there is
not uniform convergence to R , since there will be probability distributions for which convergence is arbitrarily slow. Here we have two interesting results
about this sort of inductive method.
A different method of reasoning, which philosophers sometimes call enumerative induction, starts
with the selection of a class C of rules and chooses
from C a rule with minimum error on the data. (This
learning method is also called empirical risk minimization.) Vapnik and Chervonenkis (1968) prove that
(subject to mild measurability conditions) if and only
if the rules in C satisfy a certain condition, namely that
they have finite VC dimension (which we explain in
a moment), then with probability approaching 1 the
expected error of the rules endorsed by enumerative
induction will converge uniformly to the minimum
expected error of rules in C. In other words if, and
only if, the VC dimension of C is finite, this sort of

S

enumerative induction satisfies the PAC (probably
approximately correct) criterion.
The VC dimension of the collection of rules C is the
largest number N such that some set of N points in
feature space is shattered by rules in C, so that for any
possible labeling of those N points, some rule in C fits
perfectly the points so labeled. If for any finite number
N there exists some set of N points that is shattered by
the rules C, then the VC dimension of C is infinite.
Subject to some mild measurability conditions on the
rules in C, enumerative induction is an instance of PAC
learning if and only if the VC dimension of rules in C is
finite.
Learning using standard feed-forward artificial
neural networks satisfies the PAC criterion, since the
rules representable by such a network with fixed architecture and variable weights have finite VC dimension.
A different inductive learning method, sometimes
called structural risk minimization balances empirical
adequacy of the data against some other feature of rules
in C, sometimes mistakenly identified with simplicity.
Let C ¼ C1 [ C2 [    Cn [    , where Ci 2 Ciþ1 ,
where the VC dimension of each Ci is finite and the
VC dimension of C is infinite. A particular version of
structural risk minimization selects a rule r in C that
minimizes a function f(e, i) of the error e on the data
and the least i such that r is in Ci. Something like this
fits much scientific practice but cannot satisfy the PAC
criterion (since the VC dimension of the rules C is
infinite). Under certain conditions it does allow universal consistency.
So, SLT sheds light on a variety of inductive
methods by specifying various assumption different
kinds and proving that particular inductive methods
have different desirable properties under the appropriate assumptions. This is discussed further in Harman
and Kulkarni (2007).

Important Scientific Research and
Open Questions
The field of statistical learning theory is quite active.
Some fairly recent developments include support vector machines (Vapnik 2006) and Boosting (Schapire
1999). There are many extensions and generalizations
of the basic results in which the different assumptions
of the theory are relaxed, stronger results are obtained,
more detailed analysis is provided, or different settings
are considered.

3187

S

3188

S

Statistical Models for Longitudinal Data

Cross-References
▶ Bayesian Learning
▶ Connectionist Theories of Learning
▶ Learning in Artificial Neural Networks
▶ PAC Learning
▶ Probability Theory in Machine Learning

References
Devroye, L., Györfi, L., & Lugosi, G. A. (1996). Probabilistic theory of
pattern recognition. New York: Springer.
Harman, G., & Kulkarni, S. (2007). Reliable reasoning: Induction and
statistical learning theory. Cambridge, MA: MIT Press.
Hastie, T., Tibshirani, R., & Friedman, J. (2009). The elements of
statistical learning: Data mining, inference, and prediction
(2nd ed.). New York: Springer.
Schapire, R. E. (1999). A brief introduction to boosting. Proceedings
of the Sixteenth International Joint Conference on Artificial Intelligence, Stockholm, Sweden, August 1999.
Vapnik, V. (2000). The nature of statistical learning theory (2nd ed.).
New York: Springer.
Vapnik, V. (2006). Estimation of dependencies based on empirical data:
Empirical inference science (2nd ed.). New York: Springer.
Vapnik, V., & Chervonenkis, A. Ja. (1968). On the uniform convergence of relative frequencies of events to their probabilities (in
Russian). Doklady Akademii Nauk USSR 181; translated into
English in Theory of Probability and its Applications 16, 264–280
(1971).

Statistical Models for
Longitudinal Data
▶ Measurement of Change in Learning

Steiner, Rudolf (1861–1925)
JENNIFER M. GIDLEY
Global Cities Research Institute, RMIT University,
Melbourne, VIC, Australia

Life Dates
Rudolf Steiner was born in Kraljevec, Austria-Hungary
(currently Croatia) on February 27, 1861, of Austrian
parents. A quiet, introverted child, he spent most of his
childhood and youth in a beautiful mountainous
region of Austria, surrounded by natural beauty but

immersed in the new technologies of telegraph and
railway through his father. He attended high school in
Wiener-Neustadt, near Vienna. In addition to his
school subjects Steiner privately studied high-level
mathematics and physics, Ancient Greek and Latin,
and the scientific and philosophical works of the
major German idealists and Romantics. Between 1879
and 1883, he completed his undergraduate studies at
the Vienna Institute of Technology, during which time
he worked as a private tutor and editor. In 1891,
he completed his Doctorate at Rostock University,
Germany, on Truth and Knowledge: The Fundamentals
of a Theory of Cognition with Special Reference to Fichte’s
Scientific Teaching. While still an undergraduate student, Steiner commenced his in-depth studies of the
scientific works of Goethe. Influenced by his exposure
to what he referred to as the living thinking of Goethe,
Steiner developed a new epistemology in his first
authored book Theory of Knowledge Implicit in Goethe’s
World Conception, published in 1886. In 1893, he
published what is perhaps his most significant philosophical treatise The Philosophy of Freedom. He worked
extensively from the age of 21, editing numerous volumes of Goethe’s works between 1882 and 1897, spending the last 7 years of this period working at the
Schiller-Goethe Archives in Weimar (Steiner 1925/
1928). Steiner’s productivity during this time in
Weimar was prodigious. In addition to editing
Goethe’s work, he put in order the archives of Friedrich
Nietzsche at the request of Nietzsche’s sister, edited
and published the complete works of Schopenhauer
in 12 volumes, and the works of Jean Paul in eight
volumes, researched and published some of his most
fundamental philosophical works.
In 1897, Steiner moved to Berlin, which was to
become his base until 1913, just before the outbreak
of World War I. In Berlin, Steiner combined his editing
work with his own writing and lecturing, and also
worked as an instructor at the Berlin “Workers’ School
of Education.” As editor of the weekly Das Magazin für
Litteratur Steiner came into close social contact with
the intellectual and artistic élite of Berlin at the time, his
acquaintances including poets, playwrights, novelists,
and political activists. In 1900, he began what has
become known as his anthroposophical work, which
he continued until his death. In 1902 he met Marie von
Sievers who would become his constant collaborator
and eventually his second wife. During this time he

Steiner, Rudolf (1861–1925)

worked in collaboration with the Theosophic Society,
including becoming the General Secretary of the
German Theosophic Society. However, as a result of
fundamental differences of direction, he publicly distanced himself from the Theosophic Society in 1913
and moved to Dornach in Switzerland to commence
the building of the first Goetheanum. The building was
tragically burned to the ground by an arsonist in 1922
shortly after its completion. Steiner almost immediately announced plans for a second Goetheanum,
built of concrete, which still stands today in Dornach.
Both buildings are considered architectural landmarks
of the twentieth century.
Steiner’s writing is generally regarded to have
occurred in two major phases. In the first phase until
approximately 1904, he primarily wrote from a purely
philosophical or natural scientific perspective. In the
second phase, he claimed to be speaking and writing
from his intuitive spiritual-scientific research. He spent
much of his adult life traveling and lecturing extensively throughout Europe. From 1888 to 1924 he gave
over 5,000 lectures on a wide variety of themes in 96
European cities within 17 countries. His lecturing took
him the length and breadth of Europe mostly by train,
from Oslo, Stockholm and Helsinki in the north to
Milan, Bologna, and Trieste in the south; from Oxford,
Torquay, and Paris in the west to Vienna, Prague, and
Budapest in the east. In the last few years before his death
he gave an average of 400 lectures per year. Steiner died at
age 64, after a 6-month illness, on March 30, 1925.

Contribution(s) to the Field of
Learning
In the mode of a “Renaissance universal man” Steiner
was a respected and well-published scientific, literary,
and philosophical scholar and artist, as well as an
accomplished linguist, classical scholar, mathematician, and historian (McDermott 2009). Also a futurist,
he had a macrocosmic perspective on time in relation
to the evolution of human culture and consciousness.
His ideas were post-conventional, innovative, and
futures oriented. With foresight he initiated many
significant cultural and social projects, which are still
active throughout the world 100 years later – including
innovative and holistic approaches to education,
agriculture (biodynamics), architecture, medicine,
social reform, community development, and the arts.
He authored over two dozen books, including essays,

S

3189

plays, verse, and autobiography and his collected
lectures make up over 300 volumes. His complete corpus covers almost every imaginable theme, including art,
history, religion, education, evolution, natural and hermetic sciences, meditative practices, psychology, physiology, social and community development, agriculture,
and medicine. On a more personal basis, Steiner remains
somewhat enigmatic, having been described as charismatic, yet humble, erudite yet matter-of-fact, as an idealist and also pragmatic, as both seriously earnest and
extremely funny. Boris Bugayev, a distinguished Russian
symbolist poet who wrote under the pen name of Andrei
Belyi, wrote a book about Rudolf Steiner in Russian in
1928, noting a more personal feature:
"

[Steiner] had, as it were, a therapeutic smile; the countenance blossomed. . .one felt that one had nothing of
the kind to give in return. He had the gift of the smile,
the faculty of direct expression from the heart. . .Many
know his sunny smile; we spoke of it. One must speak
about it, for not a single photograph of his reflects it.
(Cited in (Easton 1980, pp. 215–216))

In spite of his large following, he discouraged any
tendency of others to view him as a guru, believing that
such a role was no longer appropriate in the modern
age when the spirit of individual freedom is vital.
Steiner is probably best known in the world today for
the independent schooling system that he initiated in
the early twentieth century. Steiner founded his first
school in 1919, in Stuttgart, Germany for the children
of the workers of the Waldorf cigarette factory. Notably,
he not only put his educational ideas into practice, he
took personal interest in the teachers and the children
including apparently knowing every child in the school.
He appointed, guided, and advised the first teachers
and delivered numerous series of pedagogical lectures.
The name Waldorf is still used today for many of the
schools, particularly in Germany and the USA. In the
UK, Australia, and elsewhere they are more commonly
called Steiner schools. The hyphenated term Steiner–
Waldorf is becoming increasingly acceptable as an overriding term. As of 2010 there were approximately 1,000
Steiner–Waldorf schools across 60 countries, even
more kindergartens and dozens of institutions for special education worldwide. The most extensive concentrations of the schools, and also many teacher
education courses and colleges, are in Western Europe,
North America, and Australia/New Zealand.

S

3190

S

Steiner, Rudolf (1861–1925)

It is beyond the scope of this brief biography
to discuss Steiner–Waldorf education in any detail.
However, it should be noted that many of the key features of Steiner pedagogy are to be found in the emerging innovative pedagogies of the late twentieth century
and early twenty-first century. These include the importance of integration, imagination, aesthetics, the whole
child, ecology, spirituality to name a few. The unique
architecture of Steiner–Waldorf schools reflects the striving to create a fully integrated learning environment.
While some commentators claim that Steiner’s pedagogical work was ahead of its time, a contrasting view is that
Steiner pedagogy was actually very much a product of
his time and was closely aligned to what emerged in 1919
as the New Education in Germany after World War I.
The substance of much of the critique of Steiner education involves the perceived rigidity of application of
Steiner pedagogy in contemporary Steiner–Waldorf
schools. While this is actually a critique of the contemporary application of Steiner’s pedagogy rather than
a critique of Steiner’s actual ideas, the two are often
conflated. On the other hand, there is emerging support for Steiner pedagogy today, particularly from
holistic and integral educators (Gidley 2007).

Important Scientific Research and
Open Questions
The primary terms Steiner used to refer to his work
were spiritual science and anthroposophy. The term
anthroposophy – coined in the seventeenth century by
Rosicrucian, Thomas Vaughan – is derived from the
Greek anthropos, meaning human being, and sophia,
meaning wisdom. In Steiner’s own words in Anthroposophical Leading Thoughts: “Anthroposophy is a path of
knowledge, to guide the Spiritual in the human being
to the Spiritual in the universe.” The world movement
that has arisen from Steiner’s anthroposophy is based
on the notion that there is a spiritual world comprehensible to pure thought but accessible only through
advanced cognitive faculties.
In addition to the educational institutions, numerous medical clinics, therapeutic facilities, banking
institutions, and hundreds of farms in many parts of
the world bear witness to the wide dissemination
and establishment of Steiner’s ideas. Through his biodynamic farming methods, he anticipated the growth of
organic farming, believing that a farm should be a selfcontained ecological entity. He was also instrumental

in the founding of Weleda, which is still producing
homeopathic and herbal medicines today. Steiner’s aesthetic approach inspired artists such as Piet Mondrian,
Wassily Kandinsky, and Joseph Beuys. He is also
regarded as one of the founders of “organic architecture.”
One of the most striking issues surrounding
Steiner’s prolific – albeit heterodox – contribution to
so many fields of knowledge in the early twentieth
century is not that there is a discourse of academic
critique but rather that there has been such a lack of
response to his work, other than from proponents of
Steiner’s approach. Steiner is frequently referred to as
one of the most underappreciated figures of the twentieth century (Lachman 2007). Although proponents
of Steiner’s work have written a great deal of material
about it, the majority has been published by “in-house”
Steiner Presses, with much of this being unrealistically
uncritical. Steiner’s integralism and his opposition to
materialism and the dry intellectualism of modernity
led to attacks from upholders of the status quo. One
reason given for Steiner’s rejection by the mainstream
academy in his time is that as a forerunner of the
contemporary holistic/integral movement he was
decades ahead of his time. Another stated reason why
Steiner’s work was marginalized by the academy is his
unashamed esotericism.
Critics of Steiner often focus on the fact that he was
a spiritual researcher and because of his early association with the Theosophical Society, link his work with
the occultism of theosophists. Owen Barfield proposed
that it may have been Steiner’s use of the term occult –
in the title of one of his principal source books – that
has played a great part in his rejection by the academy.
Barfield, cited in Reilly (1971/2006) argued, however,
that an objective scholarly approach to Steiner’s work
would demonstrate that the word occult, as used by
Steiner, “signifies no more than what a more conventionally phrased cosmogony would determine as ‘nonphenomenal,’ ‘noumenal,’ ‘transcendental.’” Barfield
contrasted this with the more mystical, trance-like
Theosophic form of occultism, which Steiner eschewed.
According to Steiner himself, in his autobiography,
even as an 8-year-old child he was aware of the existence of a supersensible world which to him was just as
real as the physical, but his interest as an adult was to
further develop such perceptions, consciously through
intense, metacognitive activity. Steiner’s signature contribution to this field of study was to apply the rigorous

Stimulus–Reinforcer Interaction

thinking and methods of the natural sciences to thinking itself, referring to this process as “sense-free thinking.” This formed the basis of much of his methodical
research of psychological and spiritual phenomena.
There have been several challenges to adequate
reception of Steiner’s work in the wider academic
world. Firstly, his writings were originally written or
delivered in German and many of them have still not
been translated into English. Secondly, although
Steiner did write a substantial number of actual
authored books, the vast majority of his available
work is actually based on shorthand transcripts of his
lectures, many of which he never had time to revise.
Thirdly, his works are so extensive and interrelated that
an extraordinary amount of time and effort is required
even to access them, let alone comprehend them.
Although Steiner’s stated aim was to bring spiritual–scientific research out from the forum of secret
societies and place it firmly on the footing of rational
and scientific thinking, his achievements fell short of
what he apparently hoped for. Ironically, much of his
lecturing was enacted outside the academy in private
circles of students or in the public sphere. Arguably, he
did not manage to make a successful bridge between his
spiritual–scientific research and the academy during
his lifetime. A challenge for the future for those academics and educators who see untapped potential in
his work is to demythologize his body of work and
disseminate it in a more scholarly fashion.
In summary, in spite of the relative ignorance of
Steiner’s writing in most fields of academic discourse to
which his work could contribute, he made a seminal
contribution to pedagogical theory and practice as well
as many other fields. Unlike most spiritual researchers
and spiritually oriented philosophers, Steiner never lost
sight of the relationship between his spiritual research
and its application to the life-world.

S

Reilly, R. J. (1971/2006). Romantic religion: A study of Barfield, Lewis,
Williams and Tolkien. Herndon: Lindisfarne Books.
Steiner, R. (1925/1928). The story of my life (GA 28) (trans: Collison,
H.) London: Anthroposophical Publishing Co. (Original work
published 1925) [Electronic version].

Step-by-Step Learning
▶ Resistance to Learning and the Evolution of
Cooperation

Stimuli
▶ Cue Summation and Learning

Stimulus Configuration
▶ Configural Cues in Associative Learning

Stimulus Equivalence
▶ Learning of Equivalence Classes

Stimulus Generalization
In stimulus generalization, the behavioral response
learned for an arbitrary sensory stimulus (e.g., through
operant conditioning) is also evoked by other sensory
stimuli.

References
Easton, S. C. (1980). Rudolf Steiner: Herald of a new epoch. Spring
Valley: Anthroposophic.
Gidley, J. (2007). Educational imperatives of the evolution of consciousness: The integral visions of Rudolf Steiner and Ken
Wilber. International Journal of Children’s Spirituality, 12(2),
117–135.
Lachman, G. (2007). Rudolf Steiner: An introduction to his life and
work. London: Jeremy P. Tarcher.
McDermott, R. (2009). The new essential Steiner: An introduction to
Rudolf Steiner for the 21st Century. Herndon: Lindisfarne Books.

3191

Stimulus Preexposure Effect
▶ Latent Inhibition

Stimulus–Reinforcer Interaction
▶ Selective Associations

S

3192

S

Stimulus–Response (S–R) Association

Definition

Stimulus–Response (S–R)
Association
▶ Habit Learning in Animals

Stimulus–Response Binding
▶ Priming, Response
Suppression

Learning,

and

Repetition

Stimulus-Response Learning
▶ Naı̈ve Reinforcement Learning

Stimulus-Substitution Theory
The idea that, through pairings of a neutral conditioned stimulus (CS) with an unconditioned stimulus
(US), the CS comes to elicit the same responses as – i.e.,
be a substitute for – the US.

A stochastic process is a process which develops dependent on time and in accordance with probabilistic
principles. This means that future behavior cannot be
predicted with certainty but rather only probabilities as
to various possible states for the future. Stochastic
models of learning are based on appropriate mathematical models and are constructed to assess and predict states of learning processes accordingly.

Theoretical Background
Bartholomew’s (1967) seminal work on stochastic processes provides the theoretical foundation for stochastic models of learning. In a metaphorical sense, a
learning process is considered to be comparable to the
growth of different generations within families and
societies. First, a simple mathematical model for the
development of a single family line or a learning process is considered. The fundamental requirement in
such a mathematical model is that the way in which
changes occur must be specified. It is assumed that
these changes can be characterized by transition probabilities which are dependent on time. pij denotes the
probability that the “son” (= second generation of
learning) of a “father” (first generation of learning) in
class i has been in class j since the system is closed,
where k is the number of classes (see Ifenthaler and
Seel 2005).
k
X

Pj ¼ 1

j¼1

Stochastic Learning Theory
▶ Naı̈ve Reinforcement Learning

Stochastic Models of Learning
DIRK IFENTHALER, NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany

Synonyms
Probabilistic models; Stochastic process; Transition of
learning

P denotes the matrix of transition probabilities. If only
family lines in which each “father” has exactly one son
(a rather unrealistic assumption) are considered, the
class history of the family will be a Markov chain. By
interpreting the “population of possible learning
states” as composed of such family lines we can make
inferences about the changing structure of this “population.” However, in practice the requirement that each
“father” should have exactly one “son” is not met. As
a result some lines become extinct and others branch
off. However, in a population whose size remains constant over a period of time, each “father” will have one
“son” on average.
Supposing that the probability of the initial progenitor of a family line (in the case of learning processes
discussed in terms of preconceptions) being in class j at

Stochastic Models of Learning

time zero is pj(0), the probability of the line being in
class j at time T (T = 1, 2, 3, . . .) is pj(T). The probabilities {pj(T)} can be computed recursively from the
fact that
pi ðT þ 1Þ ¼

k
X

pi ðTÞpij

ð1Þ

i¼1

In matrix notation these equations may be written as
pðT þ 1Þ ¼ pðTÞP

ð2Þ

where p(T) = (p1(T), p2(T), . . ., pk(T)). Repeated
application results in
pðTÞ ¼ pð0ÞPT

ð3Þ

The elements of p(T) may also be interpreted as the
proportion of the population expected to be in each of
the various classes at time T. If the original classes of the
family lines are known, the vector p(0) represents the
initial class structure. The matrix PT plays
a fundamental role in the theory of Markov chains. It
can be used to obtain the “state” probabilities from
Eq. 3, but its elements also have a direct probabilistic
interpretation. If we let pij(T) denote (i,j), the element in
PT, then Eq. 3 may be written
pj ðTÞ ¼ pi ð0Þpij ðTÞ ðj ¼ 1; 2; . . . ; kÞ

ð4Þ

pij(T) is the probability of a family line going from class
i to class j in T generations. The case i = j is of special
interest as the probabilities pij(T) can be made the basis
of measures of change (Petermann 1978).
In the case of social systems, mutual (and even
causal) relationships exist. If a characteristic feature is
changed, corresponding changes occur in other parts of
the system as a consequence. If the relevant changes can
be quantitatively assessed, the mutual relationships can
be described – at least in principle – with the help of
mathematical equations. A system of equations that
serves the purpose of describing the system’s behavior
is a mathematical model. Its appropriateness is evaluated in accordance with its potential to provide an
effective prognosis of the results of changes in the social
system described by the model. Furthermore, the
appropriateness of the model depends on whether it
can be adapted to include relevant past changes of the
system or not.

S

3193

Important Scientific Research and
Open Questions
A process which develops dependent on time and in
accordance with probabilistic principles is a stochastic
process. This means it is not possible to predict
with certainty its future behavior but rather only
probabilities as to various possible states for the
future. Bartholomew (1967) introduced the application of stochastic models for describing social processes, specifically the growth of different generations
within families and societies. In this context, Ifenthaler
and Seel (2005) considered the progression of mental
models to be comparable to the growth of such social
processes.
Thus, it is assumed that changes in cognitive structures can be characterized by transition probabilities
which develop over time (see Ifenthaler and Seel in
press). In order to model and analyze the likelihood
that one given state of a cognitive structure (e.g., mental model or schemata) will be followed by another,
transition probabilities from one state to another are
computed. The results can be presented in
a transitional probability matrix (see Eq. 5):

P¼

:77
:38

:23
:62


ð5Þ

In matrix P, the entries in each row add up to 1. For
example, there is a.38 probability that a less elaborated
cognitive structure will increase in size or a.23 probability that an elaborated cognitive structure will
decrease in size. These transition probabilities can be
illustrated by means of a state transition diagram,
which is a diagram showing all states and transition
probabilities (see Fig. 1). Possible missing arrows

.77

.62

.23
+S

.38

-S

Stochastic Models of Learning. Fig. 1 State transition
diagram

S

3194

S

Stochastic Pi-Calculus

indicate zero probability; the density of the arrows
indicates the potency of probability.
In order to identify which transition probability
deviates significantly from its expected values, a z-score
is computed to test significance. A z-score larger than
1.96 absolute is then regarded as statistically significant
at the.5 level (see Bakeman and Gottman 1997). The
above-described stochastic models provide the mathematical basis for precisely computing learningdependent changes in mental models and schemata
(Ifenthaler and Seel 2005, in press).
There are several examples of how stochastic
models of learning can be applied. Such applications
demand a formulation of models by means of stochastic concepts (Ifenthaler and Seel 2005, in press):
1. The first function of the model is to provide insight
into the phenomenon to be investigated and to
contribute to its understanding. This is the activity
which is characteristic for a scientist. The investigation starts with the assessment of data for the
process and the formulation of a model which
typifies the attributes observed in the system.
This stage is the phase of model building. The next
step is to predict prognoses for the system by means
of the model. Following this, the predictions are
tested and conclusions are derived from the
model. This phase is called model solution. The last
phase consists of a comparison of the conclusions
with reality in order to determine whether the
model needs to be modified. This is the phase of
model testing.
2. The second objective of model building is the application of the models with the goal of a prognosis.
Scientists involved in social planning, for example,
wish to know what is likely to happen as a result of
the realization of specific policies. The manager
wishes to know in advance about the consequences
of different hiring or promoting practices in the
organization. A model which adequately describes
the behavior of the system facilitates the answering
of these questions.
3. Closely related to the problem of prognosis is the
planning of social systems and their operation
modus. The management of a company, for example, should be structured in a manner which
ensures that an adequate number of skilled workers
can be employed to carry out the necessary

production processes and that these skilled workers
are sufficiently experienced to carry out the functions of the organization. Whether it is possible to
realize this goal or not depends on an improvement
in the company’s performance. It is therefore necessary to evaluate to which degree the competing
policies contribute to the achievement of the fixed
objectives. However, in the social sciences this is
desirable but only rarely possible. To a certain
degree the model replaces the natural phenomenon
the scientist is studying. It is possible to carry out
(thought) experiments in a “world” created in and
through a model. If reality can be substantiated by
the model, the results of the experiments can be
applied to real phenomena.
4. The fourth way in which stochastic models can
contribute to social investigations has to do with
measurability. At first glance, this seems to contradict the notion that the lack of appropriate measurements prohibits the development of stochastic
models. Both assertions are correct, but they must
be viewed from different viewpoints.

Cross-References
▶ Bayesian Learning
▶ Mathematic Models/Theories of Learning
▶ Measurement of Change in Learning

References
Bakeman, R., & Gottman, J. M. (1997). Observing interaction. An
introduction to sequential analysis. Cambridge: Cambridge University Press.
Bartholomew, D. J. (1967). Stochastic models for social processes.
New York: Wiley.
Ifenthaler, D., & Seel, N. M. (2005). The measurement of change:
Learning-dependent progression of mental models. Technology,
Instruction, Cognition and Learning, 2(4), 317–336.
Ifenthaler, D., & Seel, N. M. (in press). A longitudinal perspective on
inductive reasoning tasks. Illuminating the probability of change.
Learning and Instruction.
Petermann, F. (1978). Veränderungsmessung. Stuttgart: Kohlhammer.

Stochastic Pi-Calculus
▶ Evolutionary Learning and Stochastic Process
Algebra

Stories in Psychotherapy

Stochastic Process
▶ Stochastic Models of Learning

Storage
▶ Psychology of Learning (Overview Article)

Stories in Psychotherapy
GEORGE W. BURNS
Milton H Erickson Institute of Western Australia,
Darlington, WA, Australia

Synonyms
Healing stories; Metaphor therapy; Metaphors; Teaching tales; Therapeutic metaphors; Therapeutic stories

Definition
Stories are used in psychotherapy as metaphors, or
indirect suggestions, to help effectively communicate
a therapuetic message that will commonly provide the
means or pathways by which a client may resolve their
problems. By the time many clients commence therapy, well-meaning people including relatives, friends,
and professionals, have often offered them some
sound and helpful advice. That this advice has not
been accepted or acted on means that any similar
approach in therapy is also likely to be met with
resistance, and result in a similar unsuccessful outcome. Therapeutic stories can help bypass such resistance, particularly when the therapeutic metaphors
are generated by the client, come from the client’s
own story, or are built collaboratively with the client
(Burns 2001, 2007).
In the original Greek, the term metaphor meant “to
carry something across” or “to transfer.” In communication it refers to carrying one image or concept across
to another. Most dictionaries or textbooks define metaphor as a comparison between two things, based on
resemblance or similarity. For Aristotle it meant the act

S

3195

of giving a thing a name that belongs to something else,
such as by saying, “His vulture eyes followed their every
move.” Vulture is imaginative and thus not literal if we
are talking about another human being, but it does
imply characteristics, images, and meaning not present
if the author had simply stated, “His eyes followed their
every move.” For this reason, Diomedes described metaphor as the transferring of things and words from
their proper signification to an improper similitude –
something that was and is done in language and literature for the sake of beauty, necessity, polish, or
emphasis.
Metaphor is thus a form of language, a means of
communication that is expressive, creative, perhaps
challenging, and powerful. As therapy is a languagebased process of healing, heavily reliant on the effectiveness of communication between client and therapist, the use of metaphor in therapeutic stories can link
with familiar language structures of the client and assist
the client’s process of change.

Theoretical Background
As a species, we have evolved and developed in
response to our experiences. Experience has taught
us the skills that we need as people and even shaped
our physiology in a way that most functionally suits
those skills. It has, over the generations, educated us in
the most appropriate forms of mental and emotional
adaptation. Experience is what facilitates new learning, broadens our understanding, and deepens our
knowledge. Basically, the more we experience, the
richer, more functional and more enjoyable is life’s
journey.
Though experience may be first hand, it may also
be second hand and this is where teaching tales
and healing stories come into the picture. Through
them one person can share their learning experiences
with another. In this way stories short-circuit the laborious and long-term process of each individual having
to learn something anew. Those who have had the
experience are able to communicate to others –
through their stories – how to face a similar challenge,
what tools are needed to deal with the situation, and
how they can enjoy the rewards of achievement.
Stories and metaphors are thus efficient and meaningful methods for communicating about experience
and sharing what we have learnt with others in a manner
that will hopefully help make their journey easier and

S

3196

S

Story-Based Learning Environments

more enjoyable. A therapeutic metaphor is, therefore, a
story with an express purpose of assisting clients reach
their goals in the most effective and efficient way. It as
about filling the experiential gap between what is and
what can be.
Given this, it is little wonder that every culture
throughout history has used stories to communicate
values, morals and standards, or that the great teachers
of history chose stories, parables, and metaphors as
their preferred medium of education. Buddha, Jesus,
Mohammed, and Lao Tzu did not lecture, but they told
stories. They did not quote facts, statistics, or evidencebased data, but related tales of life. Instead of teaching
the rhythmic rote learning of times tables, Lord’s
prayers or chants, they offered a parable that opened
the range of experiences and interpretations available
to their followers. Their stories have lived on for as long
as two and a half thousand years, and are still retold by
their followers. Stories are a means of communicating
par excellence: Not only do we learn from them, but we
love them.
What is surprising is that teaching tales took so long
to be recognized for their usefulness in modern day
psychotherapy and counseling, and what is not surprising, is the popularity with which they have been incorporated. It is in the work of psychiatrist and
hypnotherapist, Milton H. Erickson that we see the
origins of systematic, structured, and intentional metaphor stories that were actively employed to create
therapeutic gain. Their benefits in the therapeutic
field are that they are interactive, teach by attraction,
bypass resistance, engage and nurture imagination,
elicit a search process, develop problem-solving skills,
create outcome possibilities, and invite independent
decision-making.

Important Scientific Research and
Open Questions
There is very little scientific research, let alone good
scientific research, into the efficacy of using metaphors as a therapeutic tool. This is in part due to the
fact that metaphors tend to be used therapeutically in
very individualized ways and are not readily accessible
by the standardized, controlled variables desired by
empirical research. Much of the “evidence” about the
applications and effectiveness of metaphors in therapy
therefore comes from case studies. Around four
decades ago, Haley recorded and published a broad

and comprehensive collection of Erickson’s case
studies that included the use of therapeutic stories
(Haley 1973). Since then, the literature (professional
journals and published books) has reported the
application of metaphors in a wide range of therapeutic approaches, including cognitive therapy, counseling, couples therapy, ecopsychology, education,
Ericksonian psychotherapy, family therapy, hypnotherapy, Jungian psychotherapy, psychoanalytic
psychotherapy, solution-focused therapy, strategic
therapy, systems thinking, transactional analysis,
and more (Burns 2001, 2005). There has also been
a wide variety of clinical problems for which metaphors have been used: abusive relationships, anxiety,
asthma education, bulimerexia, cancer, childhood
and adolescent problems, depression, divorce, family
relationships, pain management, pediatric hospice
care, posttrauma stress disorders, somatic complaints,
and traumatic memories to name just a few (Burns
2007). Nonetheless, the field is still open to good,
innovative research that can help inform practitioners
on the effectiveness of using stories in therapy and
the variables that are likely to help or hinder that
effectiveness.

Cross-References
▶ Analogy Therapy
▶ Learning Metaphors
▶ Metaphor Therapy
▶ Nature-Guided Therapy

References
Burns, G. W. (2001). 101 healing stories: Using metaphors in therapy.
New York: Wiley.
Burns, G. W. (2005). 101 healing stories for kids and teens: Using
metaphors in therapy. Hoboken: Wiley.
Burns, G. W. (Ed.). (2007). Healing with stories: Your casebook collection for using therapeutic metaphors. Hoboken: Wiley.
Haley, J. (1973). Uncommon therapy: The psychiatric techniques of
Milton H Erickson. New York: Norton.

Story-Based Learning
Environments
▶ Narrative-Centered Learning Environments

Strategic Learning

Strategic Learning
ALEXANDER BODEN, BERNHARD NETT, THOMAS VON
REKOWSKI, VOLKER WULF
Information Systems and New Media, University of
Siegen, Siegen, Germany

Definition
Strategic Learning has been discussed in a variety of
fields, foremost in the context of:
1. Individual/Education
Strategic Learning can be seen as encompassing
strategies for learning on the individual level,
aiming at positively affecting the learner’s autonomy. Strategic Learning means learning about
learning in order to develop the learner’s full learning potential.
2. Groups/Organizations
On the organizational level, Strategic Learning
describes learning of organizations about developing the strategic orientation of the organization
(usually a company). The strategic orientation
of an organization is understood as resembling
the knowledge of its members regarding its longterm goals, as well as the strategies and means of
how to reach them (i.e., the ability of self-organized
work).

Theoretical Background
Strategic Learning is related to different theories from
education and organizational management, dependent
on whether the term is understood as learning in a
strategic way (educational view), or as learning about
strategies (organizational view).

Individual/Education
The term “Strategic Learning” signifies an educational
demand for “learning in a strategic way.” Students are
understood as active, self-determined individuals
processing information and constructing knowledge
(Weinstein 1994). Hence, it is suggested that students
need to learn effective strategies for learning, i.e., for
reading and understanding texts, or for memorizing
codified knowledge in a structured and efficient way
(e.g., Simpson and Nist 2000).

S

3197

Strategic Learning resembles certain aspects that are
covered by various (learning) concepts/theories, such
as metacognition or self-regulated learning. Congruent
elements that apply to this variety of theories are
“self-reflection” or “awareness-creation,” allowing the
student to better position her/his learning efforts
within her/his general learning objectives, and
affording better opportunities to align and to adapt
those efforts accordingly. Within this scope, Strategic
Learning is seen as a system to help students develop
conditional knowledge and reflect it with regard to its
practical applicability.
In case the (strategic) learning process is triggered
by external input, such as by a (strategic) teacher, it
can be seen subordinate to ▶ cognitive instruction
theories in general. Related questions like “what,”
“how,” and “when” to learn, being aware of different types of content knowledge and mediating strategies of learning/teaching, can be seen as elements of
▶ Strategic Teaching. In this sense, Strategic Learning
can be regarded as successful outcome of Strategic
Teaching.

Groups/Organizations
Strategic Learning on a group/organizational level can
be understood as learning about the underlying strategic orientations of the group or organization, e.g., in
a military or business context. As goals and strategies
themselves are subject of Strategic Learning, the concept is related to a meta-level, this is: strategic selfdevelopment. In this regard, Strategic Learning is
sometimes focused on the level of superior executives
who are in charge of leveraging Strategic Learning in
their organizations, and of sharing and institutionalizing their knowledge on the corporate level.
To some extent, Strategic Learning resembles the
conception of double-loop learning in the sense of
Argyris et al. (1985). In their view, learning can be
achieved when members of an organization compare
the (unexpected) outcomes of their actions with the
expectations and theories that guided their planning.
In this regard, learning entails two different layers
which need to be covered: single-loop learning refers
to the operational level (“Are we doing things right?”),
while double-loop learning comprises learning on the
strategic level (“Are we doing the right things?”). In this
sense, Strategic Learning is related to learning beyond
mere efficiency (“knowing how” to reach an aim), but

S

3198

S

Strategic Memory

is a creative way to question basic assumptions in order
to optimize the very strategies of a company, not only
their application.
As Strategic Learning entails learning that is related
to the aims of the company, it attempts to improve
the abilities of self-organization of the staff. Hence,
it cannot be understood merely in the way of
▶ instructionism, but aims at improving the abilities
of the staff to take strategic decisions that develop the
strategic orientation of the company.

Important Scientific Research and
Open Questions
Corresponding to the great variety of fields that Strategic Learning is related to, we find a great diversity of
“learners” trying to adapt (to) emerging circumstances
in their specific field. How such a diversity of learners is
able to cope with, and can be efficiently supported in
their Strategic Learning efforts, still bears important
research issues. Recent research on Strategic Learning
formulates a “[. . .] dilemma that currently exists
within the paradigm of strategic learning: the dilemma
of whether a strategy should be seen as an action
dependent on the specific knowledge of an educational
actor – strategic knowledge – or whether it is dependent on a planned instructional context – strategic
context” (Monero 2007).
On the group/organizational level, there is
a discourse on whether and how companies can codify,
exchange, and institutionalize strategic knowledge
(Kuwada 1998), especially with regard to dealing with
tacit knowledge of workers. On a general level, there are
discussions on how companies can leverage Strategic
Learning in practice. As far as practical questions are
concerned, related research questions are focused on
how Strategic Learning can be organized, e.g., if it can
or should be formalized, and how the results can be
institutionalized in practice. In this regard, there
is a discussion whether Strategic Learning can better
be supported by external advisors, or by internal members of the organization. The question whether it is
more beneficial to involve “external gatherers to collect
very deep and complex content” (Thomas et al. 2001),
or rather to rely on “internal organizational members
to maximize the understanding of local content and
ambiguity” (Kuwada 1998), is a major issue for further
research. In so far external advisors are involved, the
discourse on Strategic Learning also relates to ▶ Action

Research approaches (such as Cooperative Method
Development or Business Ethnography).
Strategic Learning also relates to the area of design,
e.g., in the context of software development. For
designers, Strategic Learning appears as an explorative
potential of users to self-develop their media competences. To make design supportive for Strategic Learning is to increase learnability, and, therefore, the
usability of software products. Hence, for e-Learning
systems dedicated to explorative learning, designing for
Strategic Learning is crucial, but there is still an ongoing debate on how this can be achieved.

Cross-References
▶ Cognitive Learning
▶ Double-Loop Learning
▶ Metacognition and Learning
▶ Organizational Learning
▶ Selective Learning
▶ Self-Regulated Learning
▶ Tacit Knowledge

References
Argyris, C., Putnam, R., & Smith, D. M. (1985). Action science.
San Francisco: Jossey-Bass.
Kuwada, K. (1998). Strategic learning: The continuous side of
discontinuous strategic change. Organization Science, 9(6),
719–736.
Monereo, C. (2007). Towards a new paradigm of strategic learning:
The role of social mediation, the self and emotions. Electronic
Journal of Research in Educational Psychology. N. 13, 5(3),
497–534.
Simpson, M. L., & Nist, S. L. (2000). An update on strategic learning:
It’s more than textbook reading strategies. Journal of Adolescent
and Adult Literacy, 43(6), 528–541.
Thomas, J. B., Sussman, S. W., & Henderson, J. C. (2001). Understanding “strategic learning”: Linking organizational learning,
knowledge management, and sensemaking. Organization Science, 12, 331–345.
Weinstein, C. E. (1994). Strategic learning/strategic teaching: Flip
sides of a coin. In P. R. Pintrich, D. R. Brown, & C. E. Weinstein
(Eds.), Student motivation, cognition, and learning (pp. 257–273).
Hillsdale: Erlbaum.

Strategic Memory
▶ Selective Learning

Streaming Media

Strategic Teaching
This conception describes the teacher’s role as strategist. Strategic Teaching is a model of cognitive instruction, focusing on a teacher engaging as a strategist,
adapting the teaching process to the student’s Strategic
Learning abilities.

Strategies That Reflect
Volitional Control
▶ Volitional Learning Strategies

S

3199

information located on the Internet is using a technology called World Wide Web (WWW). WWW contains a set of protocols to transmit different types of
information. Hypertext transfer protocol (HTTP) is
the set of rules for transporting information (text,
graphic images, sound, video, and other multimedia
files) on the World Wide Web. Streaming media content can be transported using HTTP or other proprietary protocols. Such transported content can be
viewed using a streaming media player. However the
speed and the quality of streaming media content
depend upon the Internet bandwidth. In the context
of the Internet, bandwidth is defined as the maximum
amount of information that can be transmitted over
the Internet per second. It is measured in bits per
second.

Use of Streaming Media for Education
● Videos which are converted into streamable format

Streaming
▶ Streaming Media

Streaming Media
●

NIPAN J. MANIAR
School of Creative Technologies, University of
Portsmouth, Portsmouth, UK
●

Synonyms
Broadcast; Multimedia; Streaming; Video

Definition
The word “streaming media” refers to the technology
that allows the continuous flow of multimedia over the
Internet. Video is a type of multimedia which combines
both visual and audible components. Streaming media
in higher education refers to the real-time delivery of
audio/video over the networks.

●

●

Theoretical Background
In computer science, network is a group of two or more
computers linked together. A connected group of networks is called Internet. One way of accessing shared

and delivered via Internet ease the problem of distributing video-based learning content compared
with other media storage format such as tape, CD,
and DVD, which have to be physically borrowed
from content distributors. Streaming media technology offers instant global delivery of learning
content, i.e., students can access learning content
anytime, anywhere.
Use of streaming media enables lecturers to search,
play, and embed streaming media content within
any virtual learning environment (Maniar and Bennett 2007)
Streaming media technology allows institutions to
broadcast live events such as lectures, seminars,
workshops, and other education-related content.
Use of this technology gives opportunity to synchronize virtual learning environment into live
event (Maniar and Bennett 2007)
Streamable videos can easily be segmented by
adding bookmarks in real time using appropriate
tools, which enables teachers to breakdown full
length video content into small chunks, with
a personalized title and description (Maniar and
Bennett 2007)
Along with bookmarking, streaming media technology enables a simultaneous synchronization of
events such as Powerpoint slides, lecture notes, captions, URLs, comments, discussions, etc. (Maniar
and Bennett 2007).

S

3200

S

Streaming Media

● Delivery of streaming media content can be

protected using technologies such as digital rights
management (Maniar 2007), which helps to track
and protect the delivery of learning content over the
Internet

Streaming Media Application
Protocols

media using the Adobe Flash Player software. The
RTMP default port number is 1935. TCP and UDP
delivery method is used in conjunction with streaming
media application protocols to deliver streaming
media. RTSP, MMS, and RTMP can be used in conjunction with either UDP or TCP, while HTTP streaming only supports TCP at this time.

Streaming Media Method
HTTP Streaming
HTTP streaming is a mechanism for sending data from
a Web server to a Web browser in response to an event.
HTTP streaming includes the delivery of market data
distribution (stock tickers), live chat/messaging systems, online betting and gaming, sport results, monitoring consoles, and sensor network monitoring. It
uses ▶ TCP (transmission control protocol) rather
than ▶ UDP (user datagram protocol), since reliability
is critical for Web pages with text. HTTP protocol
usually uses ▶ port number80 or 8080. However,
HTTP streaming is not a preferred delivery protocol
for streaming video-based content.

Following are the three methods which can be used
with the above-mentioned transmission types to
deliver streaming media over Internet.
On Demand

On-demand streaming provides “anytime” access to
a prerecorded content. In this scenario, each user can
independently request access to the stored content via
the Internet. Each user has full control over the content
playback, i.e., rewind, pause, fast-forward, and stop.
This is because ▶ unicast on-demand publishing
points provide a unique data path for every client that
requests content.

Real-Time Streaming Protocol (RTSP)

Live

RTSP was developed by the Multiparty Multimedia
Session Control Working Group to establish and control media sessions between end points. Clients of
media servers issue VCR-like commands, such as play
and pause, to facilitate real-time control of playback of
media files from the server (Quicktime.com 2010).
Most streaming servers use the real-time transport
protocol (RTP) for media stream delivery; however
some vendors implement proprietary transport protocols. The RTSP default port number is 554.

Live events such as lectures, seminars, conferences, and
tutorials can be streamed over the Internet with the
help of streaming media server live broadcasting
software.

Microsoft Media Services (MMS)
MMS is a proprietary transport protocol developed by
Microsoft that is primarily used with Microsoft Windows Media Server to transport media stream over the
Internet to the Windows Media Player (Microsoft.com
2010). The MMS default port number is 1755.

Real-Time Messaging Protocol (RTMP)
RTMP is a proprietary transfer protocol developed by
Adobe Systems (formerly developed by Macromedia)
that is primarily used with Adobe Flash Media Server to
transport media stream over the Internet on students’
computers (Adobe.com 2010). Students’ can play flash

Simulated Live

Streaming media server can simulate live broadcasts by
creating playlists of prerecorded media files. As with
live broadcasts, all users connecting to the stream see
the same point in the simulated live broadcast at the
same time, which creates a simulation of watching
a live event. Since the event actually is not live, broadcasting software is not required.

Streaming Media Server Software
A few examples of industry standard streaming media
servers are given in Table 1.

Preparing Video-Based Content for
Streaming
Videos are by default captured using video camera in
an uncompressed file format, which are not streamable.
Uncompressed video files require massive bandwidth,
making it impossible to send them over the Internet for

Streaming Media

Streaming Media. Table 1 Streaming media server
software

Supported
Supported streaming file
video codec format

Supported streaming
file format

Video encoders

Windows media server
(microsoft.com 2010)

Advanced system
format (ASF)

Windows media encoder Windows
(microsoft.com 2010)
media 9

Audio-video interleave
(AVI)

Windows media
video (WMV)

Motion picture expert
group (MPG)

Motion picture
expert group
(MPG)

Flash media server (adobe.com Flash video (FLV)(F4V)
2010)
Flash media encoder
(adobe.com 2010)

VP6

QuickTime (M4V)

Sorenson
spark

MPG-4 part14 (MP4)

H.264

H.264

streaming. Hence it is necessary to compress or encode
videos before streaming. The process of compressing
uncompressed video files into a streamable format is
called video encoding. This process is carried out by
a software application referred to as video encoder.

Video Encoder

Advanced
system format
(ASF)
Audio-video
interleave (AVI)

Windows media video
(WMV)

QuickTime (MOV)

Sorenson squeeze
(sorensonmedia.com
2010)

Flash video
(FLV)(F4V)

Windows
media 9

Advanced
system format
(ASF)

VP6

Audio-video
interleave (AVI)

Sorenson
spark

Windows media
video (WMV)

H.264

Motion picture
expert group
(MPG)

Video encoders are assembled with video compression
codec which reduces the quality and size of the original
video in order to make it streamable. Table 2 gives a few
examples of industry standard video encoders, video
codec, and streaming media file formats.

Flash video
(FLV)(F4V)

Important Scientific Research and
Open Questions

MPG-4 part14
(MP4)

Streaming media has reached a new level in terms of
quality and performance. It is called high definition
(HD) streaming. HD streaming requires a stable and
uninterrupted Internet connection speed of around
4 mbps or more. A server and a client’s computer
which support HD streaming are essential in order to
experience HD streaming. How long it will take for HD
streaming to be a standard streaming medium is
a debatable question.
Secure delivery of audio-video streams has made
this technology very attractive for content distributors.

3201

Streaming Media. Table 2 Video encoders

Streaming media server
software

QuickTime streaming server
(apple.com 2010)

S

QuickTime
(MOV)
QuickTime
(M4V)

Digital rights management is a proven platform that
makes it possible to protect and securely stream content for playback on a computer, portable devices,
mobile phones, and network devices. Using digital
rights management, audio-video files can be protected
to ensure its delivery to authenticate users in an
authorized environment. However standardization of
streaming media platform and digital rights management is much needed.

S

3202

S

Strengthening

Moving images delivered using streaming media
technology provide endless possibilities for teaching
and learning activities. However there is a need to
create a generic theoretical framework via research for
enriching and improving the learning process carried
out using streaming media technology. Such research
may lead to pedagogical practices which could be followed for creating a successful streaming media learning
resource in terms of its learning outcomes. Following are
few open questions, which may help relate streaming
media with teaching and learning activities:

students’ learning experience, ELearning Steering Group, University of Portsmouth, UK. http://www.creativehampshire.co.
uk/ct/research/streaming.doc. Accessed 22 June 2010.
Microsoft.com, Windows Media Services Deployment Guide.
http://technet.microsoft.com/en-us/library/cc730848(WS.10).
aspx. Accessed 27 Mar 2010.
Sorensonmedia.com. http://www.sorensonmedia.com/video-encoding/.
Accessed 27 Mar 2010.

● Can a constructive learning approach be the key for

▶ Memory Dynamics

●
●
●
●

●

the use of asynchronous audio and video in WWWbased learning environments?
How to deliver collaborative learning activities
using streaming media technology?
Is streaming media technology suitable to deliver
blended learning?
Can high definition streaming deliver a better learning experience compared to traditional streaming?
Security remains one of the main challenges for
streaming media technology. Is DRM the best way
to keep the content secure?
What are the minimum hardware, software, and skill
requirements for viewing or listening to the streaming media files for teaching and learning purpose?

Cross-References
▶ Asynchronous Learning
▶ Audiovisual Learning
▶ Blended Learning
▶ Cognitive Learning Strategies for Digital Media
▶ Collaborative Learning Supported by Digital Media
▶ Constructivist Learning
▶ Multimedia Learning
▶ Online Learning
▶ Video-Based Learning
▶ Web Science

References
Adobe.com, Adobe Flash Media Server Family. http://www.adobe.
com/products/flashmediaserver/. Accessed 27 Mar 2010.
Apple.com, Quick Time Streaming. http://www.apple.com/
quicktime/technologies/streaming/. Accessed 27 Mar 2010.
Maniar N (2007) Digital Rights Management for ERA plus License,
British Universities Film and Video Council, Viewfinder,
Issue No. 68.
Maniar, N., & Bennett, E. (2007). Investigation into using video
podcasting, video streaming and live broadcasting to enhance

Strengthening

Strengths Identification
▶ Rapid Learning and Strength Spotting

Strengths-Based Learning and
Development
▶ Rapid Learning and Strength Spotting

Stress
▶ Coping with Stress

Stress and Burnout
▶ Burnout in Teaching and Learning

Stress and Learning
PUTAI JIN
School of Education, The University of New South
Wales, Sydney, NSW, Australia

Synonyms
Learning under stress; Test anxiety

Stress and Learning

Definition
The word stress comes from the Latin words “strictus”
(which means “tight” or “narrow”) and “stringere”
(which means “to tighten”). These word roots carry
the meaning of restriction and limitedness and reflect
individual psychosomatic states under physical pressure or mental demands. In the 1930s, researchers in
physiology and psychology, such as Walter Cannon and
Hans Selye, developed the concept of stress in line with
the proliferation of behavioral approaches that focus
on the framework of stimulus–response (S-R) processes. For instance, Walter Cannon considered stress
as the fight-or-flight reaction that could cause certain
disturbances of homeostasis (i.e., the maintenance of
coordinated, balanced physiological states of an organism). Hans Selye further proposed and attempted to
differentiate two related terms: (1) stressors,
representing any pleasant (e.g., attending an academic
achievement award ceremony) or unpleasant (failure in
a high-stake examination) demands that are imposed
upon the individual, and (2) stress, designating
nonspecific consequences of any demands upon the
body (e.g., fast heart beating and insomnia due to the
excitement or frustration). According to this notion,
regardless of the nature of stressors, an organism may
display a three-stage general adaptation syndrome,
namely, the alarm reaction stage, the resistance stage,
and the exhaustion stage. This conceptual framework,
which triggered a number of behaviorist studies on life
events (beginning or ending school, change in schools,
etc.), was later challenged and modified by researchers
who adopted emerging cognitive perspectives.
According to this new Stimulus-Organism-Response
(S-O-R) framework, stress is defined as “a particular
relationship between the person and the environment
that is appraised by the person as taxing or exceeding
his or her resources and endangering his or her wellbeing” (Lazarus and Folkman 1984, p. 19). This
definition, which is widely accepted by the research
community, emphasizes that an encounter is stressful
when and only when perceived and cognitively
appraised as being threatening to one’s well-being.
Learning, on the other hand, is also defined and
researched via two major approaches: behaviorist
and cognitive. Although systematic empirical research
on learning (of new associations) was conducted as
early as 1885 by a German psychologist Hermann
Ebbinghaus, there has been no single definition used

S

3203

by all researchers and educators. In general, learning
is regarded as changes in one’s task performance from
behaviorist perspectives and viewed as changes in one’s
mental representations from cognitive approaches.
However, as pointed out by Shuell (1986), both sides
appear to have similarities in defining learning in terms
of three criteria: (1) a change in a person’s behavior or
ability, (2) this change arises as a result of practice and/
or experience, and (3) the change is relatively enduring.
Broadly speaking, learning is linked to changes in
learners’ behavior, emotion, thoughts, values, and
belief system as a result of individual experience. In
education and training, learning is often practically
considered as the processes of acquisition of new
knowledge and skills. In that sense, researchers in information processing assert that learning occurs with the
alteration of long-term memory in human cognitive
architecture (Low et al. 2009).
The relationship between stress and learning is complex and can be identified in the following aspects:
(1) learning under stress, (2) impact of learning on
stress, and (3) factors that affect both stress and learning.

Theoretical Background
Learning under stress – Animal and human models have
been proposed to depict the effects of stress on learning
(see Jöels et al. 2006). Stress tends to activate two
psychoneuroendocrine systems, namely, the autonomic nervous system (ANS) and the hypothalamopituitary-adrenal axis (HPA). From an evolutionary
perspective, these two essential systems exert
a streamlined impact on the organism’s information
processing and thus learning. On the one hand, in
response to stressful situation, ANS quickly releases
catecholamines (adrenaline, noradrenaline, and dopamine) and peptides to enhance short-lasting excitability (i.e., increased synaptic responses and electrical
firing of neurons). The consequences are: (1) the performance of working memory is to some extent
enhanced (e.g., shortened reaction time), and (2) working memory performance becomes less accurate due to
the elevated arousal level. On the other hand, HPA
increases the secretion of corticosteroid hormones in
response to stress (cortisol in humans, also known
as stress hormone). While catecholamines from ANS
predominantly alter neural activities immediately
after stress, the HPA action mainly occurs with
a considerable delay and deploys a longer-lasting effect.

S

3204

S

Stress and Learning

In a threatening situation, the elevated HPA functioning tends to suppress irrelevant information sometime
later by raising the threshold for synaptic enhancement
of input from other sources. As pointed out by Jöels
et al. (2006), the combined ANS and HPA activities
“preserve an appropriate priority in the reaction to
challenges” (p. 156) and thus form an optimized functioning for learning under stress, and excessive stress
can cause hyperactivities of both ANS and HPA and
may lead to psychosomatic symptoms (e.g., heart palpitations, hypertension, forgetfulness, and inability to
concentrate) and thus impair memory and learning.
Impact of learning on stress – From the perspective
of evolutionary educational psychology, there are two
types of learning: (1) the learning of biologically primary knowledge for which human beings have evolved
over millennia to acquire easily, quickly, and automatically, and (2) the learning of biologically secondary
knowledge that is only developed and required at the
recent, advanced stage of civilization (Low et al. 2009).
Biologically primary knowledge, such as listening to
and speaking native language or the recognition of
facial expressions, is essential for the human species
to survive and can be learned without conscious effort
simply by immersion into a community. Biologically
secondary knowledge, however, is specifically
demanded by modern societies and needs to be
acquired with great conscious effort and explicit
instruction. Most academic knowledge (e.g., in the
categories of written language, mathematics, physics,
and biology) taught in educational institutions is biologically secondary, and the learning of such knowledge
is evolutionarily unprepared, cognitively difficult, and
intrinsically stressful. The stressful learning processes
for the acquisition of biologically secondary knowledge
can be exacerbated by inappropriate instruction and
ineffective learning strategies.
Factors that affect both stress and learning – In order
to include other factors that significantly influence individuals’ stress level and learning outcomes, a broader
theoretical framework is needed. Very often, a triangular model that consists of social, physiological, and
psychological variables is required to identify complex
interrelations among various factors, and appropriate
model-building methodology can be employed to evaluate its fit and proximity to real-world processes. For
instance, poverty may be considered as an important
contributing factor to pupils’ stress and learning. In this

case, variables like nutrition, physical health, mental
health, study conditions, self-esteem, achievement
motivation, and school performance can be included
in a representative model and thus their integrations
can be examined.

Important Scientific Research and
Open Questions
In line with animal research (mainly using rodents
and chickens) in learning sciences and neurobiology,
there is a trend that more and more studies on stress
and learning have recruited human volunteers. Such
studies are often of an interdisciplinary nature and are
conducted with methodological rigor, employing
instruments with psychometric verifications, physiological equipment, biochemical measures, computerized technology, and true experiments with random
assignment (Jöels et al. 2006). It is found that whether
stress facilitates or impairs (or has no significant effect
on) learning depends on the source of stress, stressor
timing, stressor intensity, stressor duration, and learning task. Research in this direction is needed. For
instance, Schwabe and Wolf (2010) assessed the impact
of learning at the onset of socially evaluated cold pressor
test (i.e., subjects were watched by an experimenter of
the opposite sex and videotaped during hand immersion
into ice water) in healthy young men and women and
reported a memory-impairing effect under stress.
Various learning conditions can induce stress. As
summarized by Low et al. (2009), instructional designs
that violate cognitive principles (e.g., split-attention
effect, redundancy effect, and modality effect) derived
from research on human cognitive architecture tend
to increase cognitive demands and can elicit stress in
learners. For instance, when learners have gained a certain level of expertise in English as a second language,
they found that too many explanatory notes in reading
passages were hard to ignore and thus annoying. Such
a type of redundancy effect has been replicated in experiments conducted in other domains (Low et al. 2009).
Since assessment (including streamlining learners
according to their pre-study knowledge levels, measuring learning outcomes, providing feedback to learners,
etc.) is an essential part of learning processes and many
educational opportunities (as well as life prospects) are
associated with the results of high-stake assessment, it
is not surprising that tests and examinations are perceived by many learners as one of the most stressful

Stress Management

experiences (Stöber and Pekrun 2004). Research on
test anxiety (sometimes under broader or more specific constructs such as examination stress, performance
anxiety, competitive anxiety, maths anxiety, and
test emotions) has a long and fruitful history, and
reliable scales have been developed and used in different educational settings. Current research in this
area is advancing in the following new directions:
multidimensionality of test anxiety, learning phases
associated with test anxiety, how test anxiety relates to
attention and memory, impact of test anxiety on performance, and processes of emotional regulation (see
Stöber and Pekrun 2004).
In research that attempts to incorporate both stress
and learning in a comprehensive framework, there are
two noticeable trends. One approach is to adopt an
ecological model, such as the impact of social problems, digital or mobile study environments, or parental
orientations on stress and learning. The other type of
approach is to focus on individual differences and
effective personalized interventions, for example, how
learning disability influences an individual’s stress level
and learning processes and what coping strategies can
be used.

Cross-References
▶ Achievement Motivation and Learning
▶ Emotion Regulation
▶ Emotions and Learning
▶ Modality Effect
▶ Redundancy Effect
▶ Stress Management

References
Jöels, M., Pu, Z., Wiegert, O., Oitzl, M. S., & Krugers, H. J. (2006).
Learning under stress: How does it work? Trends in Cognitive
Sciences, 10(4), 152–158.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping.
New York: Springer.
Low, R., Jin, P., & Sweller, J. (2009). Cognitive architecture and
instructional design in a multimedia context. In R. Zheng
(Ed.), Cognitive effects of multimedia learning (pp. 1–16).
Hershey: IGI Global.
Schwabe, L., & Wolf, O. T. (2010). Learning under stress impairs
memory formation. Neurobiology of Learning and Memory,
93(2), 183–188.
Shuell, T. J. (1986). Cognitive conceptions of learning. Review of
Educational Research, 56(4), 411–436.
Stöber, J., & Pekrun, R. (2004). Advances in test anxiety research.
Anxiety, Stress, and Coping, 17(3), 205–211.

S

3205

Stress Management
PUTAI JIN
School of Education, University of New South Wales,
Sydney, NSW, Australia

Synonyms
Managing stress; Stress reduction; Stress relief; Stress
treatment

Definition
The term “stress management” refers to a fine array of
evidence-based techniques together with proactive
procedures to handle stress. Earlier research on stress
emphasized the negative consequences of stress caused
by excessive demands or dramatic encounters, such as
the fight-or-flight psychosomatic reactions (Cannon
1929) and general adaptation syndromes (Selye 1936);
the aim of many stress management programs has been
mainly the reduction of stress effects. However, the
current trend is that more and more researchers and
stress management program providers are adopting
a broader spectrum of both theoretical and practical
perspectives to design, test, modify, and implement
stress management techniques and procedures, ranging
over preventative, reaction-focused, and wellbeing oriented measures. The objectives of effective stress management are in fact not only for the amelioration of
psychophysiological reactions to stressors but for the
formation of various positive states of prosperity. In
other words, stress management has transferred from
simply an antidote to unpleasant feelings to part of
a holistic, comprehensive regimen or well-being promotion scheme. In this sense, stress management is
widely used in individual cases (e.g., a tailored procedure to treat early symptoms of depression), group
settings (e.g., family therapy to deal with conflicts
between the parents and their teenage child), and organizational contexts (e.g., employee assistance programs
to train staff to protect themselves from sexual harassment in the workplace).

Theoretical Background
There are three major theoretical frameworks that
form the foundations of various stress management
programs: behavior-focused, cognitively oriented, and

S

3206

S

Stress Management

psychoneuroimmunologically based. The first is behavioral theoretical framework, which emphasizes the
stimulus-response patterns. Many stress management
programs of this type focus on how to take appropriate
actions to control environmental or internal stimuli so
that the behaviors that are related to stress can be
modified. For instance, in order to quit smoking, one
step of the procedure is to remove the ashtray from the
living room so that the habitual behavior of smoking
can be somehow inhibited. Here, the ashtray acts as an
environmental stimulus (a reminder or a convenient
knick-knack) for smoking and thus this negative factor
ought to be removed in order to foster desired behavioral changes. According to the behavioral model of
stress management, the stimuli that lead to certain
responses can be from internal channels as well. For
instance, one of the anger management practices is to
roll one’s tongue and count from one to ten when one
feels the rage and is about to lose one’s temper. Here,
the awareness of a surge of anger (e.g., trembling
hands) serves as an internal stimulus. Theoretically,
the temporary delay of this surge may be able to slow
down or even stop the acceleration of sympathetic
reactions (e.g., increased heart rate and elevated adrenaline secretion for a forthcoming “fight”) to allow the
parasympathetic system to kick in so that the homeostasis can be restored (e.g., enabling cardiovascular
functioning and catecholamine levels to move back to
the balanced, normal state). Therefore, this proactive
step may help a person with problems of controlling
emotion to reduce the habitual explosion of anger.
The second theoretical framework for stress management is derived from cognitive approaches.
According to Lazarus and Folkman’s (1984) seminal
work, stress is determined by an integral cognitive
process of two aspects: (a) how the stressor is continuously appraised by a person, and (b) what the resources
and coping skills a person possesses. The appraisal can
be primary, secondary, tertiary, and so forth; the
resources can be adequate social support actively
sought by an individual and the learned coping skills
can be appropriate reinterpretation of stressful encounters. During the recent decades, this cognitive appraisal
theory has gradually replaced the traditional, seemingly
oversimplified stimulus-response theoretical framework. Various problem-focused and emotion-focused
stress management programs have been proposed and
widely used.

In line with the trend of proliferations of interdisciplinary research, a number of researchers have used
the psychoneuroimmunological framework in stress
management (e.g., Jin 1994). Psychoneuroimmunology emphasizes the complex interactive functioning
between psychological, physiological, endocrine, and
neuroimmune factors. According to research on
the hypothalamo-pituitary-adrenal axis (HPA), both
cognitive and emotional disturbances can trigger the
release of cortisol and other stress-related hormones
in humans, which in turn affect the functioning of
cardiovascular, digestive, skeletal-muscular, and
immune systems. The elevated level of cortisol can
suppress immune functioning and then a weakened
immune system may lead to poor health. An effective
stress management program, be it a type of cognitive/
emotional training (e.g., meditation), physical exercise
(e.g., brisk walking), or a combination of both (e.g., Tai
Chi Chuan – a form of Chinese moving meditation),
may help clients/trainees reduce the excessive level of
stress-related hormones such as cortisol, and consequently the decrease in stress hormones may be beneficial to restore or enhance immune system. This
theoretical framework predicts both the short-term
and long-term effects of established stress management
techniques and associated procedures on mental and
physical health.

Important Scientific Research and
Open Questions
Basically, there are two lines of research in stress management: one line of research assesses the efficacy of
stress management that uses behavioral change strategies, and the other examines the effectiveness of stress
management that employs cognitive restructuring
strategies (Smith 1993). Some studies further attempt
to evaluate the additive gains from both behavioral
training and cognitive treatment (Jin 1994).
As summarized by Smith (1993), a number of
empirical studies have been conducted to ascertain
the utility of stress management with behavioral orientations. On the whole, research has endorsed the usefulness of a large repertoire of relaxation techniques
and behavior modification procedures. For instance,
Smith (1993) recommended a hierarchy of relaxation
techniques (from basic to profound) that are to some
extent evidence-based: progressive relaxation ! Yoga
stretching ! deep breathing exercises ! autogenic

Stress Management

“warmth/heaviness” exercises ! autogenic hints
targeted to internal organs (lower abdomen, heart,
etc.) ! imagery exercises with a simple stimulus !
imagery exercises with unconscious answers !
nonanalytic focusing meditation (counting one’s own
breaths, silently chanting a mantra, etc.) ! Zen mindfulness meditation (openness to the flow of all arising
stimuli). Research in stress reduction also supports the
application of biofeedback, moderate physical exertion, and time management schemes.
Cognitive coping strategies and associated models
have been tested and refined in empirical studies on
stress management (see Smith 1993). Cognitive
appraisal training has been identified as the key element of various cognitive restructuring programs, such
as cognitive therapy and rational emotive therapy. Typically, those programs help clients identify irrational
thinking and test their beliefs against reality through
their experience of personal experimentation. The procedures are varied, depending on the nature of the
problem, features of the stressful situation, and the
client’s personality and previous experiences. Nevertheless, cognitive restructuring processes are more or
less in accordance with an A-B-C-D-E paradigm. Specifically, A refers to the activation of experience; B refers
to the beliefs that a person currently holds; C refers to
the consequences resulted from irrational beliefs; D
refers to the client’s effort (with the necessary assistance
from the therapist) at disputing irrational beliefs; and E
refers to the client’s new experience regarding the effect
of successfully challenging and correcting irrational
beliefs. The effect can be incremental and thus the
whole procedure can be ongoing and spiral. Such cognitive restructuring interventions, in combination with
behavioral modification strategies, have been successfully applied in many stress management programs,
such as desensitization, assertiveness training, and
self-efficacy enhancement.
In the learning context, stress management has
been implicitly or explicitly learned and practiced –
how to deal with pre-exam stress and uncertainty,
meet the deadline of an important assignment,
get along with peers in a team or class, confront
possible campus bullying, adjust oneself in a new
learning environment, and so forth. There is growing
evidence to demonstrate that students need to be
provided with stress management service/training,
not merely by a stand-by counseling center but also

S

by a general education curriculum. For instance, in
a recent study aiming to enhance students’ coping
with test anxiety, a preventive stress management project comprising lectures and peer coaching was
conducted at the University of Würzburg in Germany
(Neuderth et al. 2009). These types of early interventions were well received by the participants. Since stress
reaction in general includes three stages, namely, alarm,
resistance, and exhaustion (Selye 1936; Smith 1993),
it will be helpful to design and implement a clientoriented, carefully tailored stress management education program pinpointing relevant cognitive appraisal
and coping strategies at different phases of stressful
encounters.
Because stress management is an emerging research
area, it is not surprising that this area contains
a variety of research designs and interpretations.
Whereas there are excellent studies in the accumulated
literature of psychosomatic research and stress management, it is not uncommon for some studies to
adopt convenient samples or own-control designs.
Since stress management programs, whether prevention-oriented or intervention-oriented, require a considerable amount of resources, methodological rigor
should be ensured in order to obtain outcomes
that are valid and reliable. Studies on the efficacy of
stress management tend to be conclusive if they
have employed true experimentation, well-controlled
quasi-experimental approach, or survey with representative samples and validated instruments (see Jin
1994). In particular, the placebo effect should be identified or ruled out. Stress management research should
also pay more attention to the groups/subpopulations
having special needs, such as the elderly, students with
learning disabilities, and survivors of natural disasters
or traumatic events. In addition, researchers can take
advantage of rapidly upgraded information technology by designing interactive multimedia material for
the learning of coping skills and using the Web site for
stress diagnosis.

Cross-References
▶ Affective and Emotional
Learning
▶ Coping with Stress
▶ Emotion Regulation
▶ Emotions and Learning
▶ Stress and Learning

Dispositions

of/for

3207

S

3208

S

Stress Reduction

References
Cannon, W. B. (1929). Bodily changes in pain, hunger, fear, and rage.
New York: Appleton.
Jin, P. (1994). Theoretical perspectives on a form of physical and
cognitive exercise, Tai Chi. In G. Davison (Ed.), Applying psychology: Lessons from Asia-Oceania (pp. 135–153). Brisbane:
Australian Academic Press.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping.
New York: Springer.
Neuderth, S., Jabs, B., & Schmidtke, A. (2009). Strategies for reducing
test anxiety and optimizing exam preparation in German university students: A prevention-oriented pilot project of the
University of Würzburg. Journal of Neural Transmission, 116(6),
785–790.
Selye, H. (1936). A syndrome produced by diverse nocuous agents.
Nature, 138, 32.
Smith, J. C. (1993). Understanding stress and coping. New York:
Macmillan.

Stress Reduction

memory stores approximately seven items or chunks
of information for 20–30 s. Long-term memory, which
has unlimited capacity, holds information from
minutes to years.

Structural Determinism
Our structure determines the way in which we respond
to perturbations in our environment. The medium or
environment in which we operate is also a structurally
determined system. Recurrent interactions of both living system and medium will result in changes in both
system and medium. Those changes are determined by
the structure and cannot be other than what they are.
Thus, living system and the medium mutually specify
each other so that each contributes to creating the
world of the next instant, and so on.

▶ Stress Management

Stress Relief
▶ Stress Management

Stress Treatment
▶ Stress Management

Stress-Related Growth
▶ Posttraumatic Growth

Structural Components
Structural components in the Atkinson–Shiffrin model
are places for storing information such as the sensory
registers, short-term memory, and long-term memory.
The sensory registers store sensations for brief periods
so they can be recognized as patterns. Short-term

Structural Equivalence
▶ Analogy/Analogies: Structure and Process

Structural Learning
▶ Learning via Linear Operators

Structural Learning in
Sensorimotor Control
DANIEL A. BRAUN, DANIEL M. WOLPERT
Computational and Biological Learning Lab,
Department of Engineering, University of Cambridge,
Cambridge, UK

Synonyms
Hierarchical learning; Learning-to-learn; Structure
learning

S

Structural Learning in Sensorimotor Control

▶ Structural learning in motor control is the learning
of a general form of the rules that govern a set of related
sensorimotor tasks (such as ice skating and
rollerblading), as opposed to learning a particular sensorimotor mapping that is specific for a single task
environment. Thus, structural learning can be conceived as an adaptive process that extracts invariants
of sensorimotor mappings that are valid for a number
of task environments. These invariants can be exploited
for adaptation to novel but related tasks and lead to
facilitated learning. Such ▶ facilitation is sometimes
called ▶ learning-to-learn.

Theoretical Background
The term “structural learning” has its origins in adaptive control theory (Åström and Wittenmark 1995) and
the theory of Bayesian networks (Pearl 1988). Adaptive
control is a branch of control theory aimed at problems
in which some of the properties of the system that we
wish to control are unknown. Conceptually, two levels
of “not knowing” can be discerned: we might know the
structure of the control task (e.g., the form of the
equations of motion governing an object that we wish
to manipulate), but not know the setting of the particular parameters of the structure that are currently in
play (e.g., the mass of the object). This situation constitutes a parametric adaptive control problem, which
consists of simultaneously estimating the unknown
parameters and using these parameter estimates in the
control process. However, if the structure of the control
task itself is unknown, one deals with the much harder
problem of structural adaptive control. Such structural
learning, in its most general form, needs to develop
a suitable representation of the control process by
determining the dimensionality and form of the control problem, the functional relationships between the
control variables, their range of potential values, and
the variability associated with them. Due to the difficulty of the structural learning problem, control theory
has primarily focused on the parametric control
problem.
The problem of structural adaptive control can be
illustrated in a simple toy example (Braun et al. 2010).
Consider operating a machine with two adjustable
dials, each of which controls a single parameter
(Fig. 1a). The machine is performing well at a particular
task (red setting) and you want to adjust it for a new

a

b
m

R1

Parameter 2

Definition

3209

R2

R3

RN

---

m

P1

P2

Parameter 1

Structural Learning in Sensorimotor Control. Fig. 1
Schematic diagram of structural learning. (a) The task
space is defined by two parameters, but for the given task,
only certain parameter combinations occur (black line).
This relationship is indicated by the curved structure,
which can be parameterized by a one-dimensional metaparameter m. However, a parametric learner that is ignorant
of the structure has to explore the full two-dimensional
space when readjusting the parameter settings. (b)
Structural learning in a Bayesian network. The nodes of the
network represent random variables such as the N sensory
input variables R1 to RN and the two control parameters P1
and P2. Learning the full joint distribution of the network
without the hidden meta-parameter m would imply
a network structure where all the R-nodes are connected to
all the P-nodes. Knowing the structure with the metaparameter m thus simplifies the learning problem
substantially, since only this meta-parameter has to be
adapted if a novel task with the same structure is
encountered (Reprinted from Braun et al. (2009). With
permission)

task (blue setting). There are two ways of approaching
this problem. One solution is to adjust independently
both dials, thereby, exploring the full two-dimensional
space. However, you might find through experience
that for a set of related tasks the correct dial settings
are not randomly scattered in the full two-dimensional
space, but they have a certain functional relationship
that is possibly probabilistic or nonlinear (black line).
Such a functional relationship corresponds to a lower
dimensional structure in the parameter space and
allows building a meta-dial (with setting m) that controls and restricts all other dials to this subspace
(Fig. 1a). Such structural learning would speed up the
adaptation process considerably, as searching a lowdimensional subspace is always faster than exploring
the full-dimensional space. In biological organisms

S

3210

S

Structural Learning in Sensorimotor Control

that are faced with an adaptive control problem, controlling the dials corresponds to controlling internal
parameters that determine how sensory inputs are
transformed into motor outputs. Discovery of structural relationships could therefore provide a basis for
learning-to-learn in biological organisms.
In sensorimotor control, facilitation of learning in
generalization tasks is a well-known phenomenon (e.g.,
generalizing from ice skating to rollerblading). While
such enhanced learning might not always be the consequence of structural learning (there could also be
processes that increase adaptability nonspecifically), it
has been shown that random experiences of different
sensorimotor tasks that share the same structure lead to
structure-specific learning-to-learn effects (Braun et al.
2010). For example, it has been shown that learning
a novel rotational mapping between hand movements
and cursor movements on a screen (e.g., a rotation
angle of 90º means that hand and cursor movements
are orthogonal to each other) can be facilitated if many
different rotation angles have been experienced before
(Braun et al. 2009). Importantly, this facilitation does
not occur if many different unstructured random mappings have been experienced before. This suggests that
the observed facilitation effects are structure-specific.
The term “structural learning” also plays an important role in the theory of Bayesian networks (Pearl
1988). A Bayesian network is a graphical method to
represent the joint distribution of a set of random
variables (Fig. 1b). The random variables are
represented as nodes in the network and could, for
example, stand for sensory inputs or motor outputs.
The dependencies between these variables are
expressed by arrows in the network that indicate the
relation between any variable and its direct causal antecedents. These dependencies – that is, the presence or
absence of arrows in the network – determine the
network structure, while the probabilities that specify
the actual strength of the dependencies are the parameters of the structure. Therefore, structural learning
refers to learning the topology of a Bayesian network,
whereas parametric learning refers to learning the
quantitative relationship between the variables given
by a structure. The toy example above can also be
cast as a Bayesian structure learning problem, where
the relationship between the control variables has to
be learned. In cognitive science, investigators have
employed Bayesian network models extensively to

study how humans learn causal structures (Kemp and
Tenenbaum 2009). Similarly, hierarchical Bayesian
models have also been used to study structural learning
of different sensorimotor mappings.

Important Scientific Research and
Open Questions
Structural learning has been shown to play an important role both in sensorimotor control (Braun et al.
2010) and in human cognition (Kemp and Tenenbaum
2009). This raises the question whether and in what
ways the two kinds of learning interact or built upon
each other. For example, one of the earliest accounts of
facilitated transfer between tasks with a similar structure stem from experiments on animal cognition in
which monkeys had to learn to categorize novel objects
based on a random exposure to similar objects (Harlow
1949). In psychology it has been hypothesized that the
animals had formed learning sets. Yet, this facilitation of
learning after random exposure can be conceptually
recast as a structural learning process. Thus, structural
learning might provide a connection between sensorimotor learning and more cognitive forms of learning.
Scalable motor structures could also be considered as
building blocks of motor concepts or motor schemas.
Some of the important open questions with respect to
structural learning in sensorimotor control are as follows: What is the minimum experience necessary to
induce structural learning? How can structural learning
be enhanced? What is the prior expectation over different structures? In neurobiological terms, one of the
most important questions is how neural networks can
achieve structural learning.

Cross-References
▶ Action Learning
▶ Adaptive Learning
▶ Adaptability and Learning
▶ Schema(s)
▶ Sensorimotor Adaptation
▶ Sensorimotor Schema(s)

References
Åström, K. J., & Wittenmark, B. (1995). Adaptive control (2nd ed.).
Reading: Addison-Wesley.
Braun, D. A., Aertsen, A., Wolpert, D. M., & Mehring, C. (2009).
Motor task variation induces structural learning. Current Biology,
19, 1–6.

Student-Centered Learning
Braun, D. A., Mehring, C., & Wolpert, D. M. (2010). Structure
learning in action. Behavioural Brain Research, 206, 157–165.
Harlow, H. F. (1949). The formation of learning sets. Psychological
Review, 56, 51–65.
Kemp, C., & Tenenbaum, J. B. (2009). Structured statistical models of
inductive reasoning. Psychological Review, 116, 20–58.
Pearl, J. (1988). Probabilistic reasoning in intelligent systems: Networks
of plausible inference. San Mateo: Morgan Kaufmann.

S

3211

Student Traits
▶ Individual Differences

Student-Centered Learning
Structure Learning
▶ Structural Learning in Sensorimotor Control

MICHAEL J. HANNAFIN
Department of Educational Psychology & Instructional
Technology, University of Georgia, Athens, GA, USA

Synonyms

Structure of Interests
▶ Stability and Change in Interest Development

Structure-Fading Effect
▶ Guidance-Fading Effect

Student Effort
▶ Engagement in Learning

Student Engagement
▶ Academic Learning Time
▶ Engagement in Learning

Learner-centered instruction

Definition
Student-centered learning has (SCL) been defined circumstances where the individual determines the learning goal, learning means, or both the learning goals and
means. Accordingly, the individual may establish specific individual and pursue learning goals with few or
no external boundaries as typical during spontaneous,
self-initiated informal learning. Alternatively, the individual may have access only to specific, defined
resources to pursue individual learning goals, such as
during free-time learning in formal settings. In cases
where learning goals are externally established as in
most formal school settings, the individual determines
how they will be pursued. In essence, the cognitive
demands shift from externally mediated selecting,
processing, and encoding during directed learning to
individually anticipating, seeking, and assessing relevance based on unique needs and goals.

Theoretical Background

Student Learning Outcomes
▶ Outcomes of Learning

Student Modeling
▶ Automated Learning Assessment and Feedback

The perspectives of learning and cognition researchers
and theorists vary with respect to the underlying
assumptions of student-centered learning. To scholars
who emphasize externally defined learning outcomes,
SCL principles and practices are criticized as lacking
empirical foundation and being misguided (see, for
example, Clark and Feldon 2005). These criticisms are
bolstered by research indicating the need for and effectiveness of explicit guidance over general advice
(Kirschner et al. 2006) and the consequences of

S

3212

S

Student-Centered Learning

stimulus overload in loosely or ill-structured learning
environments (Mayer et al. 2001).
Although these criticisms have compelling face
validity, the conclusions have been based largely on
externally mediated learning. This, the research
assumptions and context bear little similarity to the
type or nature of learning advanced by SCL advocates.
Perhaps most fundamentally, all learning is not mediated by engineered instruction; instead, individuals
learn and interact continually, seeking knowledge,
negotiating meaning and understanding and learning
within their everyday environments. This is most evident in how and why we access the Internet to identify
a wide range of learning resources, including to locate
resources for formal school lessons and projects, plan
travel, identify activities of interest for children, plan
for retirement, shop comparatively online, and
a virtually unlimited number of planned and spontaneous learning tasks. Instruction comprises one significant option to promote and support learning, and in
many cases it may be the best option, but clearly not the
sole or exclusive approach.
Hannafin et al. (2009) contrasted basic differences
between time-tested cognitive principles supporting
externally mediated learning and student-centered
learning noting “fundamental shifts in cognitive
requirements as well as the foundations and assumptions underlying their design and use” (p. 196). The
locus and nature of knowledge, the role of context in
learning, and the role of prior experience are central to
both externally mediated and student-centered
approaches, but the associated assumptions and implications vary considerably. Among externally mediated
researchers and theorists, knowledge has been traditionally viewed as existing independent of individuals
to be acquired and understood according to canonical
conventions, contexts comprise stimulus elements and
their proximal relationships, and prior knowledge and
experience establishes and reifies strength of association
and relationship within complex schemata. In contrast
for student-centered learning researchers and theorists,
knowledge and meaning does not exist independent
from each other but are constructed dynamically by
individuals, context and knowledge are inextricably
tied and are mutually interdependent, and prior knowledge and experience influence initial beliefs and understanding and must be acknowledged and addressed
for learning to become meaningful to the individual.

Unlike the time-tested principles underlying externally mediated instruction, the research and theory
based underlying SCL is beginning to emerge. Some
have suggested that learning demands become increasingly complex since individual “meaning” is influenced
more by the diversity between than the singularity
across learners. Among advocates, SCL goals have
been described as important but elusive. According to
Land (2000, pp. 75–76) without effective support,
"

misperceptions, misinterpretations, or ineffective strategy use. . .can lead to significant misunderstandings
that are difficult to detect or repair. . .metacognitive
and prior knowledge are needed to ask good questions
and to make sense.

Cognitive psychologists characterize prior knowledge as schema which represents the organization of,
and relationships among, an individual’s network of
knowledge and skill. While there is unanimous agreement on the significance of prior knowledge, the locus,
nature, and meaning of knowledge are often disputed.
Among SCL theorists, understanding and sensemaking are uniquely honed by the individual’s prior
knowledge and experience and influence both what is
known and how it is known. Initial understandings,
including canonically accepted conventions as well as
misconceptions, influence the ability to detect, interpret, and synthesize knowledge. In SCL, canonical perspective does not supplant initial conceptions, but
rather the learner is guided to challenge initial assumptions as well as to test and refine initial conceptions.
Thus, prior knowledge and experience influence the
individual’s ability to mediate their own learning –
a central assumption of student-centered learning.
Explicit guidance can minimize cognitive load
demands during instruction by, for example, reducing
or eliminating extraneous load while amplifying aspects
considered central (germane) to defined goals (Kirschner
et al. 2006). Although well documented for directed
learning during instruction, student-centered learning
emphasizes the individual’s capacity to identify relevant
resources and mediate cognitive load. Consequently,
explicit guidance is rarely appropriate when the individual determines the learning goal and/or means during
SCL. Since neither goals nor means (or both) may be
explicitly specified a priori, scaffolding often assumes the
form of self-checking, navigation guidance, prompts to
reassess and evaluate progress, prompts to reexamine

Student-Centered Learning

goals and progress, reflection on state of understanding,
and support for resetting and refining goals or strategies.
Whereas externally directed approaches tend to
stimulate processing associated with to-be-learned
canonical beliefs, SCL typically focuses on mediating
individual’s perceptions and beliefs. Student-centered
scaffolding supports the individual’s efforts to identify
relevant goals, pursue and monitor efforts toward those
goals, and reconcile differences between existing understanding and to-be-learned concepts and constructs.
Typical SCL approaches identify initial understanding
in order to build from and refine, rather than to impose
canonically correct or generally accepted views on,
existing beliefs and dispositions. Not surprisingly,
among different cognitive styles, field-independent
learners who tend to self-organize and work independently perform well in less-structured learning environment, while field-dependent learners often struggle,
solicit correct answers rather than ideas and guidance,
and become frustrated by the absence of explicit structure and external expectations. To facilitate meaning
making and minimize inert knowledge, SCL environments often emphasize authentic experiences or realistic
simulations to stimulate interaction. These contexts help
students to identify learning goals, formulate and test
predictions, and situate understanding within the individual student’s experiences and enable them to understand ordinary practices from a real-world perspective.

Important Scientific Research and
Open Questions
Some have suggested that while SCL has the potential
to deepen learning when strategies are followed, the
available strategies are unutilized or underutilized.
For example, few researchers have documented conclusive evidence for effective metacognition scaffolding
during student-centered learning. To be effective during SCL, students need key domain knowledge and the
ability to regulate cognition as they formulate and
modify plans, reevaluate goals, and monitor individual
cognitive efforts. This knowledge and skill is necessary
but often insufficient, however, as students often fail to
invoke and regulate existing cognitive skills when
confronting learning tasks that are too easy or too
difficult, where they lack motivation to engage the
tasks, or when they perceive a lack of relevance.
The situated learning paradox suggests that prior
knowledge considered critical during SCL often reflects

S

incomplete and inaccurate misconceptions (Land
2000). Without adequate background knowledge and
support, learners often fail to detect inaccurate information or reject erroneous hypotheses when they
encounter contradictory evidence. Rather than building from and refining initial beliefs based in personal
experience, the misconceptions are inadvertently
reified. Lacking appropriate guidance and support,
misinformation and disinformation may go
undetected as beliefs associated with misunderstandings are strengthened rather than reconciled.
Scaffolding research and theory holds promise for
refining the guidance and support provided during SCL.
Soft scaffolding, provided dynamically and adaptively
by teachers, peers, and other human resources to
accommodate real-time changes in needs and cognitive
demands, has often proven inconsistent in implementation frequency, quality, and impact on student learning.
Similarly, technology-enhanced support (hard scaffolding) has proven effective in learning basic information,
but often ineffective in promoting the generalizable
reasoning and thinking valued in student-centered
learning. Research is needed to study how SCL components are utilized and negotiated individually, meaning
is differentiated based on unique needs and goals, and
individual needs are (and are not) addressed.
Metacognition, the ongoing monitoring of one’s
cognitive processes, is among the most important
yet also potentially most problematic cognitive construct associated with SCL. Since student-centered
learning emphasizes learning in less-structured or illstructured environments, where students regulate their
individual learning, the ability to monitor one’s cognitive processes is fundamental to evaluating progress
toward meeting individual learning goals and means.
Students who have, or develop, metacognitive strategies tend to perform more successfully than those who
do not. Thus, research is needed to clarify the extent to
which learners must possess initially, require advance
training prior to, or can develop the requisite skills
needed to monitor their progress during studentcentered learning.
Given the cognitive demands associated with student-centered learning, it is important to better understand how, when, and if individuals manage cognitive
load. Intrinsic cognitive load reflects the difficulty
inherent in the information to be learned, germane
cognitive load reflects the effort needed to create

3213

S

3214

S

Student-Controlled Instruction

relevant schemas and models for future learning, and
extraneous cognitive load reflects non-relevant cognitive requirements associated with the instructional
materials, method, and environment. Recently, de
Jong (2010) challenged both the research and assumptions underlying and cognitive load theory, noting that
in practice the different types are indistinguishable
from one another; variations in instructional format
influence both the nature and distribution of cognitive
load, individual learner differences are rarely accounted
for, and that efforts to measure cognitive load often fail
to provide valid differentiated estimates. He further
proposed that cognitive load efforts be designed to
measure perceived “difficulty of the subject matter. . .of
interacting with the environment itself. . .helpfulness of
the instructional measures used” (p. XXX). These
issues are especially relevant during student-centered
learning where distinctions between and among different types of cognitive load are typically individually
differentiated based on prior related knowledge and
experience and the individual learning goals pursued.
It is problematic to anticipate which resources and
activities are extraneous, intrinsic, or germane independent of the individual’s learning goals, their particular background knowledge and experience. Since
students must assess veracity and relevance while
attempting to address their individual learning goals
and monitoring their understanding, research is
needed to examine how cognitive load theory and
constructs vary as learners become increasingly facile
with, or frustrated by, their individual learning tasks.

Cross-References
▶ Cognitive Models of Learning
▶ Design of Learning Environments
▶ Guided Learning
▶ Learner-Centered Teaching
▶ Metacognition and Learning
▶ Open Learning
▶ Open Learning Environments
▶ Self-Regulated Learning
▶ Situated Learning

References
Clark, R. E., & Feldon, D. (2005). Five common but questionable
principles of multimedia learning. In R. Mayer (Ed.), The Cambridge handbook of multimedia learning (pp. 97–115). New York:
Cambridge University Press.

de Jong, T. (2010). Cognitive load theory, educational research, and
instructional design: some food for thought. Instructional Science, 38(2), 105–134.
Hannafin, M. J., West, R., & Shepherd, C. (2009). The cognitive
demands of student-centered, web-based multimedia: Current
and emerging perspectives. In R. Zheng (Ed.), Cognitive effects of
multimedia learning (pp. 194–216). New York: Information Science References.
Kirschner, P. A., Sweller, J., & Clark, R. E. (2006). Why minimal
guidance during instruction does not work: An analysis of the
failure of constructivist, discovery, problem-based, experiential,
and inquiry-based teaching. Educational Psychologist, 41(2),
75–86.
Land, S. (2000). Cognitive requirements for learning with openended learning environments. Educational Technology Research
& Development, 48(3), 61–78.
Mayer, R., Heiser, J., & Lonn, S. (2001). Cognitive constraints on
multimedia learning: When presenting more material results in
less understanding. Journal of Educational Psychology, 93(1),
187–198.

Student-Controlled Instruction
▶ Learner Control

Students’ Attitudes Toward
Math Learning
JAIME R. S. FONSECA
Higher Institute for Social and Political Sciences,
Centre for Public Administration and Policies,
Technical University of Lisbon, Lisbon, Portugal

Synonyms
Attitude toward mathematics (ATM) is the student’s
organized predisposition to think, feel, perceive, and
behave toward mathematics (Jovanovic and King
1998); ATM is an aggregated measure of “a liking or
disliking of mathematics, a tendency to engage in or
avoid mathematical activities, a belief that one is good
or bad at mathematics, and a belief that mathematics is
useful or useless” (Neale 1969, p. 632); ATM Scale
measures an individual’s feelings, interests, and predispositions toward mathematics (Askar 1986).

Students’ Attitudes Toward Math Learning

Definition
Following Kulm (1980, p. 358), it is probably not possible to offer a definition of ATM that would be suitable for
all situations, and for Daskalogianni and Simpson
(2000), the definition of attitude assumes the role of
a working definition, because they view the attitude
construct as functional to the researcher’s self-posed
problems.
However, there are basically definitions by dimensions, as follows:
1. Multidimensional definition, which deals with
three components: emotional response, beliefs
regarding the subject, and behavior related to the
subject. From this point of view:
(a) ATM is defined by the emotions that she/he
associates with mathematics, having a positive
or negative value, by the individual’s beliefs
toward mathematics, and by how she/he
behaves (cf. Hart 1989).
2. Bidimensional definition, in which behaviors do
not appear explicitly:
(a) ATM is seen as the pattern of beliefs and emotions toward mathematics (Daskalogianni and
Simpson 2000).
3. Unidimensional definition, which describes it as
positive or negative degree of affect associated
with a certain subject:
(a) ATM is just a positive or negative emotional
disposition toward mathematics (McLeod
1992; Haladyna et al. 1983).

Theoretical Background
One of the most important properties of people is that
they have a learning ability. We learn the behaviors
which are necessary in our lives, through the effects of
our environment and our inherited intelligence; the
learning skills of an individual affect her/his life style
continuously (Orhun 2007), and for this reason, modern societies try to improve the learning methods in
their educational systems continuously.
Human beings spent a long time growing up and
learning how to act in the world compared with other
creatures, and nowadays, the necessity of lifelong learning extends this long learning time even more (Tennyson and Schott 1997). Instructional design is a field of
study concerned with improving student learning.
Because of the low uptake of mathematics and science

S

3215

and the negative attitudes toward these subjects, the
decline in the number of science-based students as
a proportion of all students eligible for higher education has raised cause for concern about the nation’s
economic future (Dearing 1996).
The underlying metaphor of learning as an act of
increasing individual possession – as an acquisition of
entities such as concepts, knowledge, skills, mental
schemas – comes to this scholarly discourse directly
from everyday expressions, such as acquiring knowledge, forming concepts, or constructing meaning
(Sfard 2006). Learning style is a central personality
dimension between learning preference and cognitive
form. There are several learning style definitions, and in
Table 1 we can see four of them.
Slavin (1991) asserted that positive effects of cooperative learning have been consistently found on such
diverse outcomes as students’ self-esteem, intergroup
relations, acceptance of academically handicapped students, attitudes toward school, and ability to work
cooperatively, and contemporary ideas on the nature
of human learning also contribute to the instructional
practice of cooperative learning (Antil et al. 1998).

Students’ Attitudes Toward Math Learning. Table 1
Some learning style definitions
Authors

Learning style

Orhun
(2007)

Personally preferred approximation for
collecting and processing information,
forming an idea, decision-making, and
attitudes and interests

Le Fever
(1995)

The way in which a person sees or perceives
things best and then processes or uses
what has been seen. Each person’s
individual learning style is as unique as
a signature

Keefe
(1987)

Characteristic of cognitive, effective, and
psychological behaviors that serve as
relatively stable indicators of how learners
perceive things, interact with each other,
and respond to the learning environment

Kolb (1984) The way we process the possibilities of each
new emerging event which determines the
range of choices and decisions we see, and
the choices and decisions we make, for
some events to determine the events we
live through which influence our future
choices.

S

3216

S

Students’ Attitudes Toward Math Learning

The promotion of favorable attitudes toward science, scientists, and learning science, which has always
been a component of science education, is increasingly
a concern for science education (Osborne et al. 2003).
Table 2 displays what several authors meant by attitudes toward science (ATS).
Developing positive students’ ATS is a critical part
of science learning (Head 1989); students’ attitudes are
linked to their achievement in science and their

Students’ Attitudes Toward Math Learning. Table 2
Attitudes toward science
Authors

What is meant by ATS?

Klopfer (1971)

● The manifestation of
favorable attitudes toward
science and scientists
● The acceptance of scientific
enquiry as a way of thought
● The adoption of scientific
attitudes
● The enjoyment of science
learning experiences
● The development of
interests in science and
science-related activities
● The development of an
interest in pursuing a career in
science or science-related
work

Gardner (1975),
Haladyna et al. (1982),
Oliver and Simpson
(1988), Talton and
Simpson (1986),
Breakwell and Beardsell
(1992), Talton and
Simpson (1987), Crawley
and Black (1992), Piburn
(1993), Koballa (1995)

● The perception of the

science teacher
● Anxiety toward science
● The value of science
● Self-esteem at science
● Motivation toward science
● Enjoyment of science
● Attitudes of peers and
friends toward science
● Attitudes of parents toward

science
● The nature of the classroom
environment
● Achievement in science
● Fear of failure on course

motivation to persist in science courses in high school
and beyond (Kahle and Meece 1994).
The current wave of school mathematics reform
presents conceptions of content, teaching, and learning
that are quite different from the telling model; mathematical content, activities, teaching, and learning are
no longer based on the view of students as recipients of
knowledge transmitted directly from the teacher (Butty
2002).
Advocates of science education reform believe that
if students are given opportunities to do science, positive ATS will be fostered (Kahle et al. 1993). ATM is the
student’s organized predisposition to think, feel, perceive, and behave toward mathematics (e.g., see
Table 3).

Important Scientific Research and
Open Questions
Higgins (1997) found that middle school students
trained in problem solving techniques following
NCTM recommendations had more positive ATM, and
were more persistent in seeking solutions than were
students in more traditional classrooms (Butty 2002).
Evertson and her colleagues conducted numerous
studies of mathematics instruction at the junior high
school level (e.g., Evertson et al. 1980); these studies

Students’ Attitudes Toward Math Learning. Table. 3
Students’ feelings regarding mathematics
Feelings regarding
mathematics
Positive

I have found math to
be easy for me
throughout school

Result
I think statistics
would be fun

For some reason math This gave me the
was easy to me as
confidence to tackle
I was growing up
areas of mathematics
that were more
challenging
Negative I have always had
a mental block
dealing with
mathematical
formulas

I was terrified when
I learned that I would
have to take Statistics

I had a couple of
I would feel stupid
mathematics teachers and helpless
that were sarcastic

Students’ Attitudes Toward Math Learning

3217

50
Percentage of students

revealed that much more overlap exists between the
processes associated with mathematics achievement
and those associated with positive academic attitudes
(Butty 2002). Classroom organization and instructional variables correlated strongly with achievement,
and measures of teachers’ personal qualities correlated
highly with student attitudes; investigating the relationship between instructional practices and students’
ATM, they reported that classroom organization and
instruction variables correlate more strongly with
mathematics achievement, while measures of teachers’
personal qualities correlate higher with students’ ATM.
Butty’s (2002) findings showed that 10th grade
students with better ATM had a significantly higher
achievement score than those with poorer ATM; in
addition, students with good ATM in the 10th grade
achieved better mathematics scores in 12th grade.
In many studies it was found that there is a positive
relationship between attitude and achievement in
mathematics because students’ mathematical achievement can be influenced by students’ beliefs and attitudes. Recent studies have indicated that male students
have positive ATM, more self-confidence in their abilities to learn mathematics, and less mathematics anxiety than female students (Gasiorowski 1998).
Concerning sex, Ernest (1976) did analyze an interesting question, concerned whom they got help from in
the various subjects; for both girls and boys it is true
that the mother helps more than the father, except in
the higher grades, where the father helps more in
mathematics (and, to a lesser extent, in science).
Figure1 indicating this pattern through the grades,
for mathematics, is quite striking. Beginning in the sixth
grade the father becomes the “authority” on mathematics and continues this role through high school.
The results of Orhun (2007) study suggest that
there were differences among learning modes preferred
by female and male students, their mathematical
achievements, and their ATM; in general, students’
ATM are governed by their perceptions regarding the
usefulness of mathematics, and their confidence in
their ability to learn it; therefore, learning preference
and ATM are related to each other.
At the high school level, Telese (1997) also found
a positive correlation between achievement and attitude among Hispanic students; examining Hispanic
students’ ATM and their classroom experiences, he
found a higher correlation coefficient associated with

S

40
30
20
10
0
2

4

6
8
Grade in school

10

12

To whom do students go for help in mathematics
Father
Mother

Students’ Attitudes Toward Math Learning. Fig. 1 To
whom do students go for help in mathematics

negative attitudes and traditional activities than positive attitudes, indicating that negative ATM are positively correlated with traditional teaching activities.
Fonseca (2007) concluded that students with better
ATM until the ninth year (compulsory) of schooling
had a significantly higher achievement score on quantitative methods subject than those with poor ATM at
the secondary level, but the same did not happen with
the data analysis’ performance at University. Math and
quantitative methods teaching and learning methodology was quite the same, much more theoretical than
practice, and so students who did not like math tended
not like quantitative methods; the used methodology in
teaching and learning data analysis’ subject, based on
computer-assisted teaching and learning, associated to
some basic learning skills such as time planning, reading, working in cooperative groups or oral presentation, also contributed for their nonnegative attitude
toward knowledge acquisition.
However, this methodology goes on with the following question: What does the identification of the
powerful ICT-learning potential mean for that large
majority of children and students around the world who
have no access to computers? Are we facing a technologyfacilitated form of social exclusion (Skovsmose 2006)?

Cross-References
▶ Attitudes – Formation and Change
▶ Computer-Based Learning

S

3218

S

Students’ Attitudes Toward Math Learning

▶ Cooperative Learning
▶ Cross-Cultural Factors in Learning and Motivation
▶ Gendered Perceptions of Learning
▶ Learning by Doing
▶ Mathematical Learning

References
Antil, L. R., Jenkins, J. R., Wayne, S. K., & Vadasy, P. F. (1998).
Cooperative learning: Prevalence, conceptualization, and relation between research and practice. American Educational
Research Journal, 35, 419–454.
Askar, P. (1986). Matematik dersine yönelik tutumölc¸en Likert tipi
birölc¸eğin gelis_tirilmesi. Eğitim ve Bilim, 11, 3–36.
Breakwell, G. M., & Beardsell, S. (1992). Gender, parental and peer
influences upon science attitudes and activities. Public Understanding of Science, 1, 183–197.
Butty, J. L. M. (2002). Teacher instruction, student attitudes, and
mathematics performance among 10th and 12th grade black and
Hispanic students. The Journal of Negro Education, 70, 19–37.
Crawley, F. E., & Black, C. B. (1992). Causal modelling of secondary
science students intentions to enroll in physics. Journal of
Research in Science Teaching, 29, 585–599.
Daskalogianni, K., & Simpson, A. (2000). Towards a definition of
attitude: The relationship between the affective and the cognitive
in pre-university students. Proceedings of PME 24, 2, 217–224,
Hiroshima, Japan.
Dearing, R. (1996). Review of qualifications for 16–19 year olds. London: School Curriculum and Assessment Authority.
Evertson, C., Emmer, E., & Brophy, J. (1980). Predictors of effective
teaching in junior high mathematics classrooms. Journal for
Research in Mathematics Education, 11, 167–178.
Gardner, P. L. (1975). Attitudes to science. Studies in Science Education, 2, 1–41.
Gasiorowski, J. H. (1998). The relationship between student characteristics and mathematics achievement. Doctoral dissertation,
Morgantown.
Haladyna, T., Olsen, R., & Shaughnessy, J. (1982). Relations of student, teacher, and learning environment variables to attitudes to
science. Science Education, 66, 671–687.
Haladyna, T., Shaughnessy, J., & Shaughnessy, M. (1983). A causal
analysis of attitude toward mathematics. Journal for Research in
Mathematics Education, 14, 19–29.
Hart, L. (1989). Describing the affective domain: Saying what we
mean. In D. Mc Leod & V. Adams (Eds.), Affect and mathematical
problem solving (pp. 37–45). New York: Springer.
Head, J. (1989). The affective constraints on learning science. In J.
Bliss, P. Adey, J. Head, & M. Shayer (Eds.), Adolescent development and school science. New York: Falmer Press.
Higgins, K. M. (1997). The effect of year-long instruction in mathematical problem solving on middle-school students’ attitudes,
beliefs, and abilities. The Journal of Experimental Education, 66,
5–28.
Jovanovic, J., & King, S. S. (1998). Boys and girls in the performancebased science classroom: Who’s doing the performing? American
Educational Research Journal, 35, 477–496.

Kahle, J. B., & Meece, J. (1994). Research on gender issues in the
classroom. In D. L. Grabel (Eds.), Handbook of research in science
teaching and learning. New York: Macmillan.
Kahle, J. B., Parker, L. H., Rennie, L. J., & Riley, D. (1993). Gender
differences in science education: Building a model. Educational
Psychologist, 28, 379–404.
Keefe, J. (1987). Learning styles theory and practice. Reston: National
Association of Secondary School Principals.
Klopfer, L. E. (1971). Evaluation of learning in science. In J. T.
Hastings, B. S. Bloom, & G. F. Madaus (Eds.), Handbook of
formative and summative evaluation of student learning. London:
McGraw-Hill.
Koballa, T. R., Jr. (1995). Children’s attitudes toward learning science.
In S. Glyyn & R. Duit (Eds.), Learning science in the schools.
Mahwah: Lawrence Erlbaum.
Kolb, D. A. (1984). Experiential learning: Experience as the
source of learning and development. Englewood Cliffs: Prentice
Hall.
Kulm, G. (1980). Research on mathematics attitude. In R. J. Shumway
(Ed.), Research in mathematics education (pp. 356–387). Reston:
NCTM.
Le Fever, M. D. (1995). Learning styles: Reaching everyone god gave you
to teach. Colorado Springs: David C. Cook Publishing.
McLeod, D. (1992). Research on affect in mathematics education: A
reconceptualization. In D. Grows (Ed.), Handbook of research on
mathematics teaching and learning (pp. 575–596). New York:
MacMillan.
Neale, D. C. (1969). The role of attitudes in learning mathematics.
The Arithmetic Teacher, 16, 631–640.
Oliver, J. S., & Simpson, R. D. (1988). Influences of attitude toward
science, achievement motivation, and science self concept on
achievement in science: A longitudinal study. Science Education,
72, 143–155.
Orhun, N. (2007). An investigation into the mathematics achievement and attitude towards mathematics with respect to learning
style according to gender. International Journal of Mathematical
Education in Science and Technology, 38, 321–333.
Osborne, J., Simon, S., & Collins, S. (2003). Attitudes towards Science: A review of the literature and its implications. International
Journal of Science Education, 25, 1049–1080.
Piburn, M. D. (1993). If I were the teacher. Qualitative study of
attitude toward science. Science Education, 77, 393–406.
Sfard, A. (2006). Participationist discourse on mathematics learning.
In J. Maasz & W. Schloeglmann (Ed.), New mathematics education research and practice. Rotterdam: Sense Publishers.
Skovsmose, O. (2006). Introduction to the section: Mathematics,
culture and society. In J. Maasz & W. Schloeglmann (Eds.),
New mathematics education research and practice. Rotterdam:
Sense Publishers.
Slavin, R. E. (1991). Synthesis of research on cooperative learning.
Educational Leadership, 1–8.
Talton, E. L., & Simpson, R. D. (1986). Relationship of attitudes
toward self, family, and school with attitude toward science
among adolescents. Science Education, 70, 365–374.
Talton, E. L., & Simpson, R. D. (1987). Relationships of attitude
toward classroom environment with attitude toward and

Study Strategies
achievement in science among tenth grade biology students.
Journal of Research in Science Teaching, 24, 507–525.
Telese, J. A. (1997). Hispanic students’ attitudes toward mathematics
and their classroom experience. Paper presented at the annual
meeting of the Southwest Educational Research Association,
Austin, TX. (ERIC Document Reproduction Service No. ED
407–256).
Tennyson, R. D., & Schott, F. (1997). Instructional design: International perspectives. In F. Schott, R. D. Tennyson, N. M. Seel, &
S. Dijkstra (Eds.), Mahwah: Lawrence Erlbaum.

S

3219

Study Strategies
ELIZABETH A. WEBSTER, ALLYSON F. HADWIN
University of Victoria, Victoria, BC, Canada

Synonyms
Learning strategies; Study skills; Study tactics

Definition

Study Habits
▶ Learning Strategies

Study of Psychological Symbols
in Socialization
▶ Psychosemiotic Perspective of Learning

Study of Symbol-Mediated
Thinking and Learning
▶ Psychosemiotic Perspective of Learning

Study of Symbols as Thinking
Instruments
▶ Psychosemiotic Perspective of Learning

Study Skills
▶ Cognitive and Affective Learning Strategies
▶ Learning to Learn
▶ Study Strategies

Study Skills Self-Efficacy
▶ Self-Efficacy for Self-Regulated Learning

Strategies are repertoires of methods and techniques
applied purposefully for specific tasks and task conditions (McKeachie 1988). In contrast to tactics, methods,
or discrete study skills, strategies are targeted arrays of
behaviors, cognitions, or beliefs directed toward
a learning goal or outcome. Strategies are (a) purposeful
selections of tactics, (b) deliberately chosen to satisfy
specific tasks and self-conditions (IF-THEN), and (c)
systematically adapted or fine-tuned in response to
their effectiveness in that situation (ELSE; Winne and
Hadwin 1998).

Theoretical Background
Although study skills texts often classify strategies
according to type of task (e.g., reading, note taking,
and test taking), prominent researchers in the field
encourage the use of a processing approach to study
strategies. From this approach, strategies are considered tactics purposefully selected to address the interaction of three conditions: (a) task context/condition
(e.g., type of task, purpose of task, and complexity of
material), (b) processing goals (cognitive, metacognitive,
affective, and self-management), and (c) selfconditions (e.g., prior background knowledge about
the content, past experience with the type of task, and
self-efficacy). Hence, the same study skill or tactic can
be used for different types of tasks depending on the
individual’s assessment of both the processing required
for the task and his/her abilities in relation to the task
demands. For example, learners may create a concept
map as a tactic to structure information. This technique can be used for a variety of task contexts (e.g.,
taking notes while reading, taking notes during lecture,
and preparing to write an essay) and processing goals
(e.g., activating prior knowledge, organizing, and
rehearsing). Concept mapping is considered a strategy
when learners select it from among alternative

S

3220

S

Study Strategies

techniques to meet their goals for the particular type of
task under specific task and self-conditions.
As a more concrete illustration, consider a learner
who is faced with the task of reading a chapter in his
history textbook (task context). Following the IFTHEN-ELSE model for strategies, IF the learner is familiar with the topic and has strong background knowledge
of the content (self-condition) and IF he wants to activate his prior knowledge (processing goal), THEN he
chooses to create a concept map (tactic) of the information he already knows. ELSE, if he has trouble bringing
ideas to mind using the concept map, the learner then
incorporates a different tactic of surveying the chapter
before reading to get a better sense of the content. This
learner is strategically selecting and changing tactics to
address the task and self-conditions as well as his goal for
processing the information.
While strategies may be individualized to match the
task, self, and processing conditions, they engage
a limited array of empirically supported cognitive,
metacognitive, and affective or self-management processes. Brief descriptions of these processes follow.
Cognitive processing strategies are directed toward
encoding, organizing, and retrieving information.
Research consistently identifies five key processes
involved in information processing, including activating prior knowledge, selecting and isolating key information, organizing and structuring, elaborating and
generative processing, and rehearsal/repetition.
Activating prior knowledge involves bringing to
mind information previously learned in order to facilitate the processing of new information. For example,
before attending a lecture, a learner may first brainstorm ideas about what is known about the topic. In
this instance, brainstorming creates a foundation for
making connections between the information presented
in the lecture and existing knowledge, facilitating understanding, and encoding new information. Other techniques for activating prior knowledge include concept
mapping, K-W-L charts (what I already Know, what
I Want to find out, what I have Learned), partial
outlining, and previewing the textbook.
Selecting and isolating key information is the process
of distinguishing important ideas or concepts from
extraneous and seductive details. Although it does not
directly ensure deeper understanding of the information, searching and selecting is a necessary first step to
focusing attention on important pieces of information

for further processing. A tactic commonly used for
selecting and isolating key information is underlining
or highlighting. Empirical support for the effectiveness
of this tactic is scarce; however, its effectiveness has not
always been examined with respect to a selecting or
isolating process. Furthermore, this tactic may serve
useful functions in combination with elaborative
processing tactics.
Organizing and structuring information imposes
a structure on the information to be learned in order to
recognize and understand patterns and connections
among ideas and concepts. All information sources
have an inherent structure, including description, cause
and effect, classification, compare/contrast, timeline,
sequence, enumeration, and generalization. Recognizing
the inherent structure of information and translating that
information into summaries that make those inherent
structures salient facilitates encoding, storage, and
retrieval. Structuring applies to information presented
during lectures, through written text, or across multiple
information sources. Learners may strategically organize
and structure information by applying one of the following techniques/tactics: compare/contrast charts, concept
or knowledge mapping, outlining hierarchies, flowcharts,
cause and effect diagrams, and timelines.
Elaborating and generative processing helps learners
to build a deeper understanding and memory for ideas
and concepts and also facilitates transfer. This process
involves extending beyond provided information to
draw inferences and create meaningful connections
between new information and existing knowledge. For
instance, learners may generate their own examples of
a particular concept to help understand and remember
the meaning of that concept. Learners can also use techniques such as summarizing, self-questioning, and elaborative verbal rehearsals.
Rehearsing and repeating exposure to information
enhances retention of information. Research consistently shows that (a) repetition facilitates recall and
(b) repetition combined with deeper processing facilitates better recall than repetition alone. The more
exposure one has to material, the greater chance one
has of remembering and recalling it. Some methods of
rehearsal involve rereading or reviewing material, reciting information, reworking material, and retrieving
information from memory. To enhance understanding,
rehearsal and repetition should be combined with
deeper processes such as elaborating on the material.

Study Strategies

Metacognition refers to knowing about cognition as
well as monitoring and regulating cognition.
Metacognitive knowledge includes cognitions about
oneself, the task, and strategies for completing the
task. Metacognitive processes involve monitoring and
evaluating one’s knowledge and learning as well as regulating the factors that influence achievement of one’s
goals for learning. Tactics for monitoring and evaluating
learning may include activating prior knowledge, selfquestioning, self-testing, and tracking time spent studying. For example, a learner might write down what is
known about the topic in a textbook chapter and create
questions about what is not known. During or after
reading, questions can be answered and understanding
is reassessed. The regulatory aspect of metacognition
involves planning, evaluating, and adapting or changing
strategies based on information provided through monitoring. For example, after monitoring comprehension
and evaluating it as low (because he or she cannot answer
those questions), the learner may augment a strategy by
creating a concept map to relate main ideas in the textbook and lecture.
Affective and self-management processes include
focusing attention on the task, maintaining or increasing
motivation, goal-setting and planning, and managing
one’s time, resources, and environment. Examples of
techniques to achieve these purposes include self-talk
focused on thinking positively, enhancing efficacy, or
confidence; reminding oneself of the reasons for doing
the task; regulating emotions; and managing time, task
goals, or the environment. For instance, a learner may
regulate anxiety by practicing deep breathing exercises in
order to better focus on the task. Alternatively, a learner
may purposefully generate anxiety by thinking about the
importance of the task in order to encourage better
preparation and performance. Another learner might
create subgoals that break down the task into smaller
chunks in order to increase motivation for starting
the task.

Important Scientific Research and
Open Questions
Study strategy instruction. Researchers in the field recommend training in study strategies should (a) be
based on a model of learning in order to enhance
transfer, (b) provide empirical evidence for the effectiveness of strategies, (c) include the opportunity for
students to practice applying the strategies to their

S

3221

authentic academic tasks, and (d) promote
metacognitive awareness and self-regulation of strategies, including evaluating the effectiveness and appropriateness of strategies and adjusting or adapting
strategies to meet learners’ needs.
However, most study skills books and Web sites (a)
categorize strategies according to topic area (e.g., time
management, reading, note taking, anxiety and stress,
and motivation) without identifying the underlying
processes (i.e., cognitive, metacognitive, affective) and
(b) emphasize the enactment of study tactics without
providing support in the strategic self-regulation of
tactic use. Study skills instruction could be enhanced
by emphasizing processes essential for successful strategy use, including understanding the task and its purpose, setting task-specific goals, enacting study tactics,
and evaluating the usefulness of tactics and adapting
them to match task conditions and learner needs.
Effectiveness of specific study strategies. Empirical
evidence for the effectiveness of study skills is weak or
mixed at best. The learning processes described above
have consistent empirical support but the nature of
strategies requires that they be carefully tuned to task,
self, and processing demands, which vary greatly
among individuals. Of the plethora of study skills,
only a handful have been tested, with findings that
some are more effective than others. Following is
a snapshot of findings regarding some popular strategies, some of which are less effective than many people
assume:
● Underlining/highlighting. Although an oft-used

strategy, underlining/highlighting has received
mixed empirical evidence. Caverly et al. (2000)
summarized the research on underlining/highlighting, indicating that this tactic is ineffective if (a)
learners are unable to identify main ideas in the
first place, (b) the material is long and complex, (c)
assessments do not measure concepts that were
underlined/highlighted, and (d) learners do not regularly review the underlined content. If learners do
choose underlining/highlighting, they should use it
in conjunction with other more sophisticated strategies that promote deeper processing of the information, such as annotating in their own words.
● Concept mapping. Research on concept mapping
indicates that this is an effective strategy when (a)
learners need to engage in deeper processing of the

S

3222

S

Study Strategies

material and retain main ideas rather than memorize details, (b) learners receive instruction in concept mapping and persist in their use of the strategy,
and (c) learners can assess task demands and their
background knowledge. Nesbit and Adesope (2006)
found in their meta-analysis that studying concept
maps was somewhat more advantageous for
retaining information than studying text passages,
lists, and outlines. This finding suggests that the
format of a concept map (i.e., arranging nodes in
relation to one another and labeling the links) is
more effective than other summary formats. More
evidence is needed, however, for the effects of using
concept maps on higher-level learning, such as
transferring knowledge from one context to another.
● Reading comprehension: SQ3R. SQ3R (Survey,
Question, Read, Recite, Review) is a commonly
taught strategy for monitoring comprehension.
Despite its popularity, several reviews have found
little evidence for the effectiveness of the strategy as
a whole, although the individual parts have received
some support. Caverly et al. (2000) concluded that
SQ3R requires extensive instruction and practice,
especially for learners with low or medium reading
ability, along with an awareness of the effort
involved in using it.
● Self-questioning. Self-questioning involves generating and answering questions about what is to be
learned in the task. Research with this tactic consistently provides support for its positive effect on
retention and comprehension. As with any tactic,
instruction and training are vital; however, selfquestioning may be easier to learn than other
more complex tactics, such as concept mapping.

Future Directions
Further empirical research is needed to investigate
the effectiveness of a variety of strategies, systematically
examining the interaction among task conditions,
self-conditions, processing goals, and the learning context or environment. For example, what strategies are
effective for collaborative learning and for working in
computer-based learning environments? Are the strategies used in traditional contexts appropriate for these
alternative and increasingly prevalent contexts? How
can we best teach learners to experiment with strategies
in these contexts? In addition, there is little research

examining the effectiveness of individual strategies for
regulating motivation and affect.
Finally, Weinstein et al. (2000) offer these questions
for future directions related to the development, training, and transfer of strategy use:
● What are the precursors of effective strategy use?
● How can we facilitate the development of these
●

●
●
●

skills at differing ages?
What can we do to help teachers incorporate learning-to-learn activities into their classroom
teaching?
How do we facilitate high-level transfer across tasks
and content areas?
How do we help students learn to cue themselves to
transfer strategies?
How can we help students learn to take more control of their own learning processes and outcomes?

Cross-References
▶ Approaches to Learning and Studying
▶ Chunking Mechanisms and Learning
▶ Cognitive and Affective Learning Strategies
▶ Cognitive Learning Strategies for Digital Media
▶ Concept Maps
▶ Elaboration Effects on Learning
▶ Learning by Chunking
▶ Learning from Questions
▶ Learning from Text
▶ Learning to Learn
▶ Metacognition and Learning
▶ Role of Prior Knowledge in Learning Processes
▶ Self-Regulated Learning
▶ Self-Regulation and Motivation Strategies
▶ SOAR – Learning
▶ Strategic Learning

References
Caverly, D. C., Orlando, V. P., & Mullen, J. L. (2000). Textbook study
reading. In R. F. Flippo & D. C. Caverly (Eds.), Handbook of
college reading and study strategy research (pp. 105–147).
Mahwah: Lawrence Erlbaum.
McKeachie, W. J. (1988). The need for study strategy training. In C. E.
Weinstein, E. T. Goetz, & P. A. Alexander (Eds.), Learning and
study strategies: Issues in assessment, instruction, and evaluation
(pp. 3–9). San Diego: Academic.
Nesbit, J. C., & Adesope, O. O. (2006). Learning with concept and
knowledge maps: A meta-analysis. Review of Educational
Research, 76(3), 413–448.

Styles of Engagement in Learning
Weinstein, C. E., Husman, J., & Dierking, D. R. (2000). Selfregulation interventions with a focus on learning strategies. In
M. Boekaerts, P. R. Pintrich, & M. Zeidner (Eds.), Handbook of
self-regulation (pp. 727–747). San Diego: Academic.
Winne, P. H., & Hadwin, A. F. (1998). Studying as self-regulated
learning. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.),
Metacognition in educational theory and practice (pp. 277–304).
Mahwah: Lawrence Erlbaum.
Further Reading
Hadwin, A. F., & Winne, P. H. (1996). Study strategies have meager
support: A review with recommendations for implementation.
The Journal of Higher Education, 67(6), 692–715.
Hadwin, A. F., & Winne, P. H. (in press). Promoting learning skills in
undergraduate students. In M. J. Lawson & J. R. Kirby (Eds.), The
quality of learning: Dispositions, instruction, and mental structures. New York: Cambridge University Press.

Study Tactics
▶ Study Strategies

Studying Expert Solutions
▶ Worked Example Effect

Styles of Engagement in
Learning
MARY AINLEY
Psychological Sciences, University of Melbourne,
Melbourne, VIC, Australia

Synonyms
Engagement profile; School engagement profile; Taxonomy of engagement

Definition
The phrase styles of engagement focuses attention on
multidimensional variability in groups of students in
regard to the way they connect with learning. When

S

3223

a set of motivation or engagement variables are measured for the same group of students identification of
distinct response patterns within the set of variables
have been referred to as styles of engagement, as school
engagement profiles, or as a taxonomy of student
engagement.

Theoretical Background
The central assumption behind the styles of engagement
concept is that across any set of attitudinal and motivational variables, combinations of variables provide
information that is complementary to descriptions and
analyses based on isolating individual variables. Traditional experimental and quasi-experimental designs
focus on identifying how separate variables relate to
each other and to outcome variables. This has been
referred to as a variable-centered approach. In contrast
researchers focusing on individuals as complex systems
recommend designs that identify how specific variables
combine in different members of the population
and how the combinations relate to outcomes. This
approach has been referred to as a person-centered
approach. Analytic techniques such as cluster analysis,
pattern analysis, or profile analysis are used to identify
subgroups within a sample or population. The assumption is that while traditional variable-centered
approaches, such as regression, allow assessment of
the contribution of specific predictor variables to an
outcome variable, they are not as effective in accounting for the combined effects over a set of variables.
Key questions concerning students’ motivation and
engagement lend themselves to application of these
person-centered approaches. For example, achievement goal theory in a variety of forms is widely used
to describe students’ beliefs about the purpose of their
learning. Sometimes the learning referent is a specific
domain; sometimes learning in general. The psychology of education literature has a wealth of papers
describing how aspects of students’ achievement are
related to mastery and performance goals, both
approach and avoidance. Work avoidance goals also
feature in some of this literature. Most commonly
achievement goals are conceptualized and measured
as dimensions. However, investigation of the overall
levels of each dimension separately and its relation to
learning outcomes or learning strategies, only provides
limited insight into the ways students’ connect with

S

3224

S

Styles of Engagement in Learning

their learning. Each student relates to learning in specific domains, or to learning in general, in terms of a
composite of these dimensions. In response, a number
of researchers have adopted analytic strategies to identify specific styles, patterns or profiles of achievement
goals that might shed further insight into the relations
between beliefs about the purpose of learning and
students’ achievements. This research is important
because it employs analytic strategies “that preserve
the multidimensional character of student engagement
with learning” (Ainley 1993, p. 395). This development
in approach to researching engagement in learning is
consistent with the outcome of the extensive review of
the literature by Fredericks et al. (2004), who argued
that engagement encompasses behavioral, cognitive,
and emotional components and recommended that it
be studied as a multifaceted construct.

Important Scientific Research and
Open Questions
Two early examples of research studies that identified
profiles of student motivation and investigated the
relation between these profiles and achievement outcomes are Ainley (1993) and, Meece and Holt (1993).
Using an approach to learning measure that distinguishes deep, achieving and surface motivation dimensions, and a measure of general ability, Ainley identified
six profiles (styles of engagement) in 11th Grade female
students. Students’ styles of engagement were shown to
be related to learning strategies when preparing for
examinations and to academic achievement in the
form of final school grades assessed more than 12
months later. Of particular theoretical interest was the
demonstration of the multidimensional character of
engagement. The more engaged students expressed
relatively high levels of both deep and achieving
approaches, and demonstrated a different pattern of
strategy use when preparing for examinations than did
students with less engaged styles. At a different developmental level (5th and 6th grade students), Meece and
Holt (1993) used measures of mastery, ego, and workavoidant goal orientations to identify goal combinations that were significantly related to students’
achievement in science. They argued that using cluster
analysis to identify goal orientation profiles generated
“a more distinctive and internally consistent set of
findings than did pattern analyses that were based on
median split procedures” (p. 582). More recently,

similar profiles of achievement goals have been
reported in a study of 6th grade students (Tapola and
Niemivirta 2008) and in a study of the achievement
goal profiles of undergraduate students (Daniels et al.
2008). Tapola and Niemivirta (2008) used latent-class
analysis to identify multidimensional achievement goal
profiles which were reported to be significantly related
to students’ preferences for different types of instructional environments. Daniels et al. (2008) used cluster
analysis with scores on mastery and performance
achievement goal measures to define profiles, capturing
the different ways students connected with their learning. These profiles were then assessed in relation to
students’ course achievement and to broader features
of how they responded to their academic experience,
for example, expected achievement, perceived success,
and a number of achievement emotions. The analytic
strategies adopted by these studies assume that student
motivation and engagement is multidimensional and
an approach that preserves the multidimensional character of students’ personal orientations to learning will
make a significant contribution to understanding the
dynamics of student engagement with learning. While
the styles of engagement or profiles of goal orientation
that have emerged from these studies have demonstrated substantial consistency, the detail of the profiles
cannot necessarily be generalized beyond the participating sample and all of the studies reported have been
careful not to interpret their profiles as types. Further
research is required to establish where the similarities
between the profiles identified are in fact general patterns that can be used as a typology or taxonomy.
All of the studies referred to so far have based their
identification of styles or profiles of student connection
with learning on measures of motivation that distinguish goals as beliefs about the purpose of learning
(Daniels et al. 2008; Meece and Holt 1993; Tapola and
Niemivirta 2008), or approaches to learning (Ainley
1993). These studies have focused primarily on cognitive and to some extent emotional engagement in the
sets of dimensions investigated. In contrast, Linnakylä
and Malin (2008) using PISA 2003 data focused on
engagement in school life and included behavioral
engagement indicators in their styles of engagement.
Their measures included peer acceptance, student–
teacher relations and students’ attitudes to schooling
“the extent to which students identify with and
value schooling and participate in academic and

Styles of Engagement in Learning

nonacademic learning activities” (p. 585). Of particular
interest in the findings was the lower achievement level
of a group characterized by strong peer acceptance and
negative attitudes to school.
In other research using the styles of engagement
concept the purpose has been development of
a taxonomy of engagement related to a specific domain
of student learning, For example, Bangert-Drowns and
Pike (2002) proposed a taxonomy describing 4th Grade
students’ styles of engagement with interactive educational software. Unlike the previously described studies
based on students self-reports of their motivational
orientations, Bangert-Drowns and Pyke identified students’ styles of engagement based on teacher observations and ratings of how students went about working
with interactive educational software. The purpose of
the research was to develop a taxonomy of levels of
engagement to support teachers in distinguishing “different qualities of student engagement, so they can
better anticipate and respond to different qualities of
student learning” (Bangert-Drowns and Pike 2002,
p. 23). The taxonomy defines seven levels of engagement that form a hierarchical structure. At the base of
the hierarchy are three types of engagement that are
described as problematic: disengagement, unsystematic
engagement, and frustrated engagement. A second level
within the taxonomy represents competent engagement where students are responding to the structure
of the problem. Another three types of engagement
represent the third level of the engagement hierarchy
and are referred to as self-regulated interest, critical
engagement, and literate thinking. Fundamentally,
these styles represent differences in students’ skill base
and in the ways they use their skills to interpret information and to apply efficient thinking strategies to
evaluate the structure of the interactive tasks. To use
the taxonomy, teachers rate each student on the frequency that they display each of the seven levels of
engagement in their interaction with educational software. For each student, this generates a profile of their
style of engagement. In the validation study, students’
style of engagement was significantly correlated with
reading test scores.
To some extent, reference to styles of engagement
begs the question concerning the nature of engagement. The research referred to used different types of
indicators of motivation and engagement. The studies
by Ainley (1993), Meece and Holt (1993), Tapola and

S

3225

Niemivirta (2008), and Daniels et al. (2008), have used
measures of students’ motivational orientations to
define ways that they connect with their learning. On
the other hand, Bangert-Drowns and Pike (2002) have
used ratings of students’ participation in specific classroom activities as indicators of engagement, while
Linnakylä and Malin (2008) using data from PISA
2003 have used students’ self-ratings on issues of
“peer acceptance, student-teacher relations and perceptions of the value of school and education for the
future” (Linnakylä and Malin 2008, p. 583) as indicators of engagement. The variety of indicators used by
these researchers is characteristic of the literature on
student engagement and was highlighted by Fredericks
et al. (2004) and is likely to be debated further. In this
context, the styles of engagement concept emphasizes
the need to preserve the multidimensional character of
students’ ways of connecting with learning.

Cross-References
▶ Academic Motivation
▶ Achievement Motivation and Learning
▶ Approaches to Learning and Studying
▶ Measurement of Student Engagement in Learning
▶ Motivation and Learning: Modern Theories
▶ Motivation, Volition and Performance
▶ School Motivation
▶ Volitional Learning
▶ Volitional Learning Strategies

References
Ainley, M. (1993). Styles of engagement with learning:
Multidimensional assessment of their relationship with strategy
use and school achievement. Journal of Educational Psychology,
85(3), 395–405.
Bangert-Drowns, R. L., & Pike, C. S. (2002). Teacher ratings of
student engagement with educational software: An exploratory
study. Educational Technology, Research and Development, 50(2),
23–37.
Daniels, L. M., Haynes, T. L., Stupinsky, R. H., Perry, R. P., Newall,
N. E., & Pekrun, R. (2008). Individual differences in achievement
goals: A longitudinal study of cognitive, emotional, and achievement outcomes. Contemporary Educational Psychology, 33(4),
584–608.
Fredericks, J. A., Blumenfeld, P. C., & Paris, A. H. (2004). Student
engagement: Potential of the concept, state of the evidence.
Review of Education Research, 74(1), 59–109.
Linnakylä, P., & Malin, A. (2008). Finnish students’ school engagement profiles in the light of PISA 2003. Scandinavian Journal of
Educational Research, 52(6), 583–602.

S

3226

S

Styles of Learning and Thinking: Hemisphericity Functions

Meece, J. L., & Holt, K. (1993). A pattern analysis of students’
achievement goals. Journal of Educational Psychology, 85(4),
582–590.
Tapola, A., & Niemivirta, M. (2008). the role of achievement goal
orientations in students’ perceptions of and preferences for classroom environment. British Journal of Educational Psychology,
78(2), 291–312.

Styles of Learning and
Thinking: Hemisphericity
Functions
KALPANA VENGOPAL
Regional Institute of Education, Mysore,
Karnataka, India

the students, such students are frequently seen to be
uninterested in the class, feel bored, and reject the
learning activity. Since any subject can be taught in
any way that is compatible with any style, students
will seek learning activities that are compatible with
their own preferred styles. Both teachers and students
tend to exploit their preferred styles, which may or may
not match. Therefore, it is important for the teachers to
know the students-preferred styles, so that the teachers
can capitalize the opportunities for students’ learning.
Styles like abilities are not etched in stone at birth.
They are in large part developed by environmental
condition and by way of nurturing children by their
parents and teachers. Some individuals may have one
preferred style at one stage and another preferred style
at some other stage. Styles are not fixed, but changeable. Research tools are readily available to identify the
individual’s preferred style of learning.

Synonyms

Hemisphericity

Ways of learning and thinking

Hemisphericity is the cerebral dominance of an individual in retaining and processing modes of information in his own style of learning and thinking. The
human left cerebral hemisphere is to be specialized
for primarily verbal, analytical, abstract, temporal,
and digital operation and that the right cerebral hemisphere is to be specialized for primarily non-verbal
holistic, concrete, creative, analogic, and aesthetic
functions. The specialized functions of each hemisphere appear well lateralized and established early in
life and barring special intervention or insult, continue
essentially unaltered throughout the normal life span.
Tools have been developed to study the “Style” of
Learning and Thinking and hence it would be possible
to infer the dominance of an individual (Venkataraman
1994).

Definition
The differences in preference of the two hemispheres
for retaining and processing of information in one’s
own style of learning and thinking is known as styles of
learning and thinking. Hemisphericity is the cerebral
dominance of an individual in retaining and processing
modes of information in his own style of learning and
thinking.

Theoretical Background
Styles depend upon cerebral dominance of an individual
in retaining and processing different modes of information in his own styles of learning and thinking. Styles
indicate the hemisphericity functions of brain, and students learning strategy and information processing are
based on the preferences of the brain area. Styles are
propensities rather than abilities. They are the ways of
directing the intellect which an individual finds comfortable. The style of learning and thinking are as important as levels of ability, and we ignore to identify and
develop students’ thinking styles at their earlier and
appropriate stage.
It is foremost important for the teachers to focus
their attention on students’ favored thinking styles
before imparting the subject matter. If mismatch exists
between the preferred styles of the teacher and that of

SOLAT
The differences in preference of the two hemispheres
for retaining and processing of information in one’s
own style of learning and thinking is known as styles
of learning and thinking (SOLAT) by Torrance
(Reynolds and Torrance 1978). The SOLAT tool is
based on the hemisphericity functions of the brain. It
identifies hemisphericity dominance by way of studying the hemisphere functions. It indicates a student’s
learning strategy and brain hemisphere preference in
problem-solving.

Styles of Learning and Thinking: Hemisphericity Functions

Psychological Activities Related to
Hemispheric Function
Right Hemisphere
The right hemisphere has remained underestimated
and even today neurophysiologists cling to the view
that the right hemisphere is a mere unconscious automation, while we live in left hemisphere. It has a great
neuronal capacity to deal with informational complexity. This hemisphere matures earlier than the left hemisphere and the balance between two in children is not
similar to one found in adult. The right hemisphere has
a greater ability to process many modes of information
within a single cognitive task, while the left hemisphere
is superior in tasks which require fixation upon a single
mode of representation or execution.
● Language: Language function is somewhat more

●

●

●

●

equally shared between the hemispheres before the
age of 5. The primary expressive mode of the right
hemisphere speculates to be metaphorical in
nature; however, in general, verbally communicative ability of the right hemisphere is relatively
limited and dependent upon the left hemisphere,
and complimentary non-verbal functions are carried out by the right hemisphere which is dominant
in left-handed people. It coordinates the voluntary
motor activities of the left side of the body.
Visual Patterns: The interpretation of complex
visual patterns is predominantly the right hemisphere function. The retention of visual patterns,
such as geometric designs and graphs, is believed to
be in the domain of the right hemisphere.
Facial Identification: The right cerebral hemisphere
is involved in face processing. Right hemisphere
superiority is greater for upright than inverted
faces and that it involves both perceptual and memory components.
Motor: Awareness of body position, spatial orientation, and the perception of fine and gross motor
activities all come within the realm of the right
hemisphere. Tactile perception is also a right hemisphere function.
Creativity: The clinical studies do not enable one to
conclude about the specific location of creativity in
the brain, i.e., in the right or the left hemisphere.
However, it is clear that creativity will not occur
without the full participation of a well-developed

S

3227

right hemisphere. The left hemisphere also provides
variety of elements, for instance, information which
is indispensable for creativity. There is a general
agreement in creativity researches also that information is a basal factor in creative thinking. The
right hemisphere may be more intuitive, imaginative, insightful; has a rudimentary verbal conceptual
scheme, aesthetic experiences; produces visual
imagery; sees things in a broader perspective; uses
the information from the left hemisphere to elaborate, to form new combinations, and to attribute
new meanings to it.
● Music: Research has stressed the importance of music
and musical instruction in stimulating the right
hemisphere. The brain is organized at birth to obey
musical stimulation in the right hemisphere. Speech
is a function primarily of the left hemisphere, while
song is a function of the right (Kane and Kaane 1979).
● Problems: The right hemisphere is able to overcome
the most difficult logical and systematic problems
which we would conjecture, relaxing the right standards of thought of the left hemisphere. The right
cerebral hemisphere makes an important contribution to human performance. It is the neural basis of
our ability to take in fragmentary sensory information and from it construct a coherent outside world,
a sort of cognitive spatial map within which we plan
our actions.

Left Hemisphere
Each hemisphere is capable of functioning in a manner
different from the other. For many years attention
was focused on the left hemisphere in which speech
was localized the so-called “dominant,” “leading,”
or “major” hemisphere. It was hypothesized that
this hemisphere was primarily responsible for the
processing of language and planning, the two functions
which clearly distinguished men from the rest of the
animal kingdom. It has been found to be anatomically
larger than the right hemisphere, as evidenced by neonatal studies. It is considered to be more active than the
right hemisphere in most adults, as indicated by EEG
analysis. The left hemisphere apparently specializes in
sequential logical, variable, symbolic, convergent production, and logic functioning.
● Speech: Short-term memory is primarily the func-

tion of left hemisphere. Expression of language

S

3228

S

Styles of Learning and Thinking: Hemisphericity Functions

through speech is exclusively processed in the left
hemisphere. Disturbed speech follows left but not
right hemisphere lesions.
● Learning: The left hemisphere is involved in learning
the 3 R’s, and reading is considered to be principally
a left hemisphere function. The mathematical functions, particularly calculations and algebra are left
hemisphere operations. The left hemisphere is well
suited for the “education of relation” which is essentially the ability to analyze the common aspects of
a task and formulate systematic relationship among
these tasks.
● Analytical Thinking: The left brain functioning is
involved in sequential operations, analytical/logical
thinking, explaining, describing, recalling the verbal content and the operations related with words.
This hemisphere is considered to be a rational linear
mind specializing in sequential processing, logical,
and analytical thinking.

Class Preference

Right hemisphere: Get clarity while learning experimentally; learning everything by synthesizing, like concrete learning, slow acquisition of habits, not interested
in games and sports.
Left hemisphere: Get clarity through logical reasoning, understanding better while learning critically and
analytically, likes to learn in an abstract way, interested
in games and sports.
Learning Preference

Right hemisphere: Divergent, concentrates on several
things simultaneously, competitive, unsocial, mysterious, greater tolerance, and adjustment.
Left hemisphere: Convergent, concentrates on one
thing at a time, individuality, social, active, no tolerance tendency.
Interest Preference

Specialized Information Processing
Preferences Associated with
Hemisphericity and Styles of Learning
and Thinking Under Different
Dimensions
Learning Style
Verbal

Right hemisphere: Understanding movements of
action, talking while reading or writing, learn best by
instruction which uses visual presentation, likes to
draw more pictures, expression of feelings through
music, dance, and poetry.
Left hemisphere: Understanding verbal explanations, getting things quiet while reading or studying,
learn best of instruction which uses verbal, likes to talk
and write, expression of feelings and thoughts in plain
language (or open mindedness).

Right hemisphere: Invent something new and imaginative; likes to solve complex problems, artistic and
aesthetic, interested in funny things.
Left hemisphere: Improve upon something, likes to
solve simple problems, temporal interest.

Thinking Style
Logic/Fractional

Right hemisphere: Holistic approach, recall faces,
retention and recalling shapes and figures, a good command over total memory and tonal, organizing capacity to show the analogy.
Left hemisphere: Fractional approach, recall names,
retention and recalling numerical figures, analyzing
speech and sound qualities, sequence of ideas analogical relationship.
Divergent/Convergent

Content Preference

Right hemisphere: Interest in soft sciences, open-ended
lessons, likes to learn through main ideas/basic concepts,
writing/likes fiction, learning through exploration.
Left hemisphere: Interest in hard sciences (vocational interest in engineering), structured lessons,
likes to learn through details and specific facts, writing
non-fiction, learning through examining.

Right hemisphere: Deductive learning, independent
thinking, deep thinking while lying down, easily find
directions in strange surroundings, likes to make
guesses.
Left hemisphere: Inductive learning, mentally
receptive and responsive to what is heard and said,
deep thinking while sitting erect, easily find directions
in familiar places, not interested in guesses.

Styles of Learning and Thinking: Hemisphericity Functions

Creative

Right hemisphere: Creative thinking, likes to pre-plan,
intuitive, judgments through feelings and experience,
playful approach in problem-solving.
Left hemisphere: Intellectuality, likes to day dream,
logical approach in judgments, business-like approach.
Problem Solving

Right hemisphere: Absent mindedness, optimistic
view, absence of repression and suppression, passive,
stronger determination and ambition.
Left hemisphere: Never be absentminded, pessimistic view, presence of repression and suppression,
aggressive/short-tempered.
Imagination

Right hemisphere: A strong memory and remembrance
about images, ability to experiment, tactile perception,
imagine, and summarize.
Left hemisphere: Memory of language and pictures,
processes through rational learning and analytical,
lacks haptic or tactile perception, outline, analysis.

Activation of Hemispheric Functions
Different teaching techniques and methodologies can
be adopted to activate and influence the hemispheric
functions of the brain. The teaching techniques in
schools can be undertaken in consonance with the
student’s style of learning and thinking. The teaching
and learning procedures must be organized in such
a way, that they tone up and activate both the hemispheric functions of the brain in students.

S

It has been discovered that imagery is the precursor of creativity in both the arts and music, that
the right hemisphere activity is the precursor of imagery, and that the positive means of the right hemisphere
stimulation through art and music is equally as effective as other incubatory techniques for allaying the left
hemisphere function.
Besides art and music are play. Through play children gradually develop concepts of casual relationship,
the power to discriminate, to make judgments, to analyze and synthesize, to imagine, and to formulate. It
gives an opportunity to use imagination, fantasy, and
creativity which develop the right hemisphere of the
brain. A remarkably effective way of stimulating right
hemisphere thinking is, also, the use of metaphors
(analogies and smiles). They provide awareness of the
relationship between dissimilar objects and situations.

Left Hemispheric Functions
New concepts can be introduced in an analytical manner with verbal emphasis and importance can be given
to the expression of language. Students may be asked to
listen to abstract speeches heard in the radios, televisions, public meetings, and symposium. They may be
given training in analyzing and identifying different
speech sounds and encouraged to give logical reasoning
and examples for unknown activities or functions without experimenting in general. Discussions may be
arranged on general problems, world affairs from the
reading of daily newspapers and magazines. They can be
encouraged in writing non-fiction essays and scientific
explanations in plain language. Games based on verbal
and numerical materials, and events can be encouraged.

Right Hemispheric Functions
The promising strategy by which to cultivate the right
brain is to stimulate the child for a more differentiated
perception and by increasing the proportion of direct
contact he has with phenomena about which he is
learning.
The right hemisphere of the brain is characterized
by numerous functions and no one particular technique of teaching will suffice to develop these functions. These have to be built through different models
of teaching. A combination of models, which shares
many features, is perhaps a useful approach to work out
a system of teaching in this context such that the
inquiry model, inductive, creative training and insight
models.

3229

Important Scientific Research and
Open Questions
Raina’s (1984) work on the right hemisphere shows
that it is capable of processing language if the discriminations are uncomplicated (e.g., a positive from
a negative statement). The analysis of voice intonation,
an integral component of language, appears to be the
function of the right hemisphere. When verbal material
(or material easily coded into language) is presented to
right and left visual fields, most subjects show a rightvisual field superiority presumably due to better right
field connections to the left hemisphere (Kane and
Kaane 1979). During verbal tasks, whether performed
overtly or covertly, increased alpha wave suppression

S

3230

S

Subculture

and a greater evoked response are found over the left
hemisphere. The right-handed people with left hemisphere dominance, ordinarily have the speech centers
in the left hemisphere. The left hemisphere is restively
specialized in verbal functioning and is devoted largely
to semantic abilities and functions. It has been found
that the left hemisphere alone does not process verbal
information. The left hemisphere has also been found
to be involved in comprehension and retention of language (Fitzgerald and Hattie 1983).
The iconic presentation of information (e.g.,
graphic displays, diagrams, flow charts, etc.) greatly
facilitates both the comprehension and the retention
of information. Iconic memory is primarily a function
of the right hemisphere (Raina 1984). The hypothesis
of right hemisphere participation in dreaming implies
that the degree of hemispheric lateralization may contribute to individual differences in dream recall and
content. Divergent thinkers are more likely to report
dreams when awakened from REM sleep than convergent thinkers. These types of cognitive styles are consistent with the assumed specialization of the right and
left hemisphere. High creativity, associated with right
hemisphere functioning, is correlated with more fluent
dreams in terms of number of words and content
elements. With respect to problem solving, the right
hemisphere is able to design thought experiments
which the left hemisphere cannot, because of its rigidity. It is able to hit upon solutions which can then, of
course, be recast into strictly logical terms by the left
hemisphere (Kane and Kaane 1979).
In terms of analytical thinking, the left hemisphere
is far more constrained and it shifts through inputs and
reduces functions to logical-rational forms and acts
more like a digital computer. It requires structure and
order which processes perception and sensory input in
logical and linear modes. It functions easily in acquisition of new habit patterns (Fitzgerald and Hattie 1983).
Researches done by Reynolds and Torrance (1978),
and Fitzgerald and Hattie (1983) indicate that it is
possible to modify a person’s preferred style of learning
and thinking over relatively brief period (6–10 weeks).
It is also possible to control the general direction of the
changes in the style of learning and thinking with the
knowledge of styles of learning and thinking mechanism. It may also be possible to train individuals to
modify their information processing procedures to best
fit their demands of the cognitive tasks.

Cross-References
▶ Adaptation to Learning Styles
▶ Adult Learning Styles
▶ Approaches to Learning and Studying
▶ Cross-Cultural Learning Styles
▶ Jungian Learning Styles
▶ Kolb’s Learning Styles
▶ Learning Style(s)
▶ Neuroeducational Approaches on Learning
▶ Neuropsychology of Learning
▶ Styles of Engagement with Learning

References
Fitzgerald, D., & Hattie, J. A. (1983). An evaluation of your style
of learning and thinking inventory. British Journal of Educational
Psychology, 53, 336–346.
Kane, N., & Kaane, M. (1979). Comparison of right and left hemisphere functions. Gifted Child Quarterly, 23, 157–167.
Raina, M. K. (1984). Education of the left and right: Implications of
hemispheric specialization. Madras: Allied Publishers.
Reynolds, C. R., & Torrance, E. P. (1978). Perceived changes in styles
of learning and thinking (Hemisphericity) through direct and
indirect training. Journal of Creative Behaviour, 12, 245–252.
Venkataraman, D. (1994). Style of learning and thinking. New Delhi:
M/s. Psy-Com Services.

Subculture
▶ Microculture of Learning Environments

Subgoal Learning
RICHARD CATRAMBONE
School of Psychology, Georgia Institute of Technology,
Atlanta, GA, USA

Synonyms
Goals; Hierarchical structure

Definition
A subgoal represents the purpose of a set of steps. When
people learn a procedure, they frequently memorize

Subgoal Learning

a linear series of steps for reaching a solution or
accomplishing a task. The drawback to learning just
a series of steps is that those steps by themselves contain
little or no information regarding how they can be
modified. As a result, learners have difficulty solving
▶ novel problems or adapting old procedures. However, if those steps are connected to subgoals, then the
learner can use the subgoals to guide his or her search
for which steps to alter in order to solve the problem.
Subgoals provide organization and help constrain the
piece(s) of the procedure on which to focus. Thus, new
problems or tasks that share the same subgoals with the
already-learned procedure will be more likely to be
solved by the learner.

Theoretical Background
Learners have difficulty solving problems that involve
more than minor changes to the procedure demonstrated by training problems or examples (e.g.,
Catrambone 1998; Reed et al. 1990). People tend to
form solution procedures that consist of a long series of
steps – which are frequently tied to incidental features
of the problems – rather than more meaningful representations that would enable them to successfully
approach new problems. Thus, if they are given a new
problem that seems similar to an old one – at
a ▶ surface level – they will try to apply a set of steps
from the old problem.
Presumably one of the jobs of education is to equip
people to deal with novel problems and situations,
not just a small recognizable set. Yet it appears that
this job does not get done. A promising approach
would be to help people organize their problemsolving knowledge in some way that ties the steps to
a meaningful hierarchical structure (e.g., Anzai and
Simon 1979; Catrambone 1998; Newell and Simon
1972). This organization is more consistent with the
way that experts solve problems (e.g., Larkin et al.
1980). While an expert’s conception of how to solve
problems does not necessarily tell us how to instruct
novices, it does provide guidance on the types of organizing elements that might be useful for the novice
to form.
The term “subgoal” has been used in at least two
ways in the problem solving and transfer literatures:
(1) something generated by a learner when he or she
reaches an impasse during problem solving (e.g., Newell
and Simon 1972), and (2) a feature of a task structure

S

that can be taught to a learner (e.g., Catrambone 1998).
This latter definition is the focus here.
Subgoals show the breakdown of a problem-solving
procedure into sub-problems (Anzai and Simon 1979).
Problems within a subsection of a domain typically
share the same subgoals. The steps for achieving those
subgoals will differ across problems.
Suppose a learner is attempting to achieve
a particular subgoal in a novel problem and discovers
that the old set of steps, perhaps learned from an
example, will not work. If the learner has learned subgoals, then he or she will have a reduced search space to
consider when changing the procedure: the steps associated with the current subgoal. In contrast, a learner
who has learned a solution procedure consisting of
a single goal reached by a long series of steps will have
a larger space to search, and thus will be less likely to
determine which steps to modify.
Procedures can be analyzed to identify the goals,
subgoals, and steps. Such an analysis has important
implications for training skills because it can help
instructors and instructional designers (e.g., textbook
writers, multimedia learning environment developers)
present material to emphasize the subgoal structure.
The premise is that learning materials that present a
well-developed subgoal organization will lead to more
efficient learning, better retention, and an increased
likelihood that the learner can flexibly apply and generalize the information to new situations.
Consider a student who studies the following
worked example (adapted from Reed et al. 1990):
Tom can mow his lawn in 1.5 h. How long will it
take him to finish mowing his lawn if his son mowed
one fourth of it?
Solution:



1
 h þ :25 ¼ 1 ! ð:67  hÞ þ :25 ¼ 1
1:5
! :67  h ¼ :75 ! h ¼ 1:13hrs;

where h is the number of hours worked.
A student might learn from this that the way such
“work” problems are solved is to take one person’s time
and divide 1 by it, multiply it by the unknown, add the
amount that was already done, and set it all equal to 1.
Such an approach would not be successful for the
following problem (also adapted from Reed et al.
1990):

3231

S

3232

S

Subgoal Learning

Bill can paint a room in 3 h and Fred can paint it in
5 h. How long will it take them if they both work
together?
Solution:

 

1
1
 h þ  h ¼ 1 ! ð:33  hÞ þ ð:20  hÞ ¼ 1
3
5
! :53  h ¼ 1 ! h ¼ 1:89hrs:
These problems cannot be solved using the exact
same set of steps, but they share the same subgoal
structure: represent each worker’s work rate and represent how long each worker worked. These subgoals
might also involve calling on lower-level subgoals to
find a worker’s rate or how long the worker worked.
Thus, these subgoals represent mini-problems in the
context of solving the overall problem.

Important Scientific Research and
Open Questions
Relatively little research has been conducted on how
learners form subgoals; most efforts have been directed
toward predicting transfer by learners assumed to
already possess the subgoals. One attempt to explain
how subgoals might be learned was made by Anzai and
Simon (1979). They recorded the moves and verbal
protocol of a learner as she solved the Tower of Hanoi
problem multiple times. Over trials the learner began
to chunk groups of moves. That is, she would make
a set of moves in quick succession (a “burst”) followed
by a pause before the next set of moves. Anzai and
Simon argued that each burst of moves represented
a chunk. Each chunk may have reflected a subgoal the
learner was achieving by the particular set of steps.
Anzai and Simon (1979) suggested that subgoal
acquisition is greatly aided when the search space for
operators (e.g., possible moves in the Tower of Hanoi
problem) is constrained. When the search space is
constrained, working memory load is reduced. One
hypothesized advantage of a working memory load
reduction is that a learner is better able to notice and
remember a sequence of steps that led to a particular
outcome. In Anzai and Simon’s model, this working
memory load reduction aids subgoal formation
because a subgoal is formed when learners are working
toward a certain goal and notice that a set of steps puts
them in a situation that allows them to ultimately
achieve the goal. Learners will be better able to notice

the result of the first set of steps, and be able to chunk
that sequence of steps, if working memory load has
been reduced (see also Sweller (1988)).
The “subgoal learning model” (Catrambone 1998)
assumes that if a learner is cued that a set of solution
steps belong together, he or she will be more likely to
try to self-explain why those steps belong together, that
is, to determine their purpose. This is similar to one of
the types of self-explanations that Chi and her colleagues have observed as students studied various
texts in physics and biology (e.g., Chi et al. 1989). The
subgoal learning model can be summarized as follows:
1. A cue leads learners to group a set of steps.
2. After grouping the steps, learners are likely to try to
self-explain why those steps go together.
3. The result of the self-explanation process is the
formation of a goal that represents the purpose of
that set of steps.
While most learners can presumably engage in
a self-explanation process with varying degrees of success, good students seem better at determining the
appropriate boundaries between meaningful groups
of steps in a solution procedure (Chi et al. 1989). The
use of a label in examples might serve as a cue to the
boundaries.
A reasonable question to ask at this point is: why
not directly state the subgoals to learners rather than
embedding them in examples? There are two problems
with this approach. First, a variety of studies have
shown that learners exhibit a clear preference for learning from and referring to examples when faced with
new problems. Further, adding examples to training
materials aids learning. Second, although there have
been a small number of successes teaching solution
procedures directly, most attempts have been unsuccessful. These attempts to teach procedures directly
might have failed though because the materials encouraged learners to memorize the steps, instead of forming subgoals that organized the steps. Thus, the
efficacy of direct subgoal instruction remains an open
question.
Research suggests that appropriately designed
examples are an effective way to help learners to form
more generalized solution procedures such as ones
organized around subgoals. The methods learned to
achieve those subgoals will possibly need to be adapted
when the learner encounters new problems. A learner

Subitizing

will have a better chance of adapting a solution procedure that is organized by subgoals, and methods for
achieving those subgoals, compared to a solution procedure that consists only of a long series of steps.
However, systematic research still needs to be
conducted on exploring a variety of ways to aid subgoal
learning.

Cross-References
▶ Advance Organizers
▶ Cognitive Load Theory
▶ Cognitive Models of Learning
▶ Cognitive Skill Acquisition
▶ Effects of Instruction and Modeling on Skill
Learning
▶ Example-Based Learning
▶ Knowledge Organization
▶ Learning by Doing
▶ Learning by Doing Versus Learning by Thinking
▶ Mathematical Learning
▶ Problem Solving
▶ Schema-Based Problem Solving
▶ Worked Example Effect

References
Anzai, Y., & Simon, H. A. (1979). The theory of learning by doing.
Psychological Review, 86(2), 124–140.
Catrambone, R. (1998). The subgoal learning model: creating better
examples so that students can solve novel problems. Journal of
Experimental Psychology: General, 127(4), 355–376.
Chi, M. T. H., Bassok, M., Lewis, R., Reimann, P., & Glaser, R. (1989).
Self-explanations: how students study and use examples in learning to solve problems. Cognitive Science, 13, 145–182.
Larkin, J., McDermott, J., Simon, D. P., & Simon, H. A. (1980). Expert
and novice performance in solving physics problems. Science,
208, 1335–1342.
Newell, A., & Simon, H. A. (1972). Human problem solving. Englewood Cliffs: Prentice-Hall.
Reed, S. K., Ackinclose, C. C., & Voss, A. A. (1990). Selecting analogous problems: similarity versus inclusiveness. Memory & Cognition, 18, 83–98.
Sweller, J. (1988). Cognitive load during problem solving: Effects on
learning. Cognitive Science, 12, 257–285.

Subitization
▶ Subitizing

S

3233

Subitizing
HENRY RAILO1, MINNA M. HANNULA-SORMUNEN2
1
Department of Psychology, University of Turku,
Turku, Finland
2
Department of Teacher Education and Centre for
Learning Research, University of Turku,
Turku, Finland

Synonyms
Subitization

Definition
▶ Subitizing, a term coined by Kaufman and colleagues
(Kaufman et al. 1949), refers to a fast and highly accurate, effortless process by which a small number of
items can be enumerated without counting. Consider,
for example, that you are viewing a pool game and that
at some point you are asked to report how many balls
are on the table. If there are only three balls on the pool
table, you can determine their number with little difficulty and high confidence. However, if the number of
balls is higher, say seven, you probably cannot quantify
their number just at a glance, but you must count
them one by one or in groups. The range of numbers
of items that can be enumerated without counting is
termed subitizing range. Although estimates of the
subitizing range differ between studies, and individual
and developmental differences are known to exist,
subitizing range is nowadays usually considered to
range from one to about three or four objects in adults,
and from one to about two to three in children under
school age.

Theoretical Background
The difference between enumerating small and larger
sets of objects has attracted researchers from the very
beginning of psychological science, and the question is
entangled with a more general issue of capacity-limited
processes of the cognitive system. In 1860, William
Hamilton, following earlier philosophers, wondered
how many objects could be apprehended at once. By
casting a varying number of marbles on the floor and
examining how many objects he could view at once
without confusion, Hamilton concluded that the mind

S

3234

S

Subitizing

is able to ingest up to six items at a glance, reasoning
that “the greater the number of objects among which
the attention of the mind is distributed, the feebler and
less distinct will be its cognisance of each.” (Hamilton
1860, p. 254).
Subsequently, enumeration of small and larger
arrays has been studied in more controlled situations
(e.g., Jevons 1871; Kaufman et al. 1949; Trick and
Pylyshyn 1994). In a typical experiment, participants
are asked to report the number of visually presented
objects as fast as possible. The classical finding is a slope
of response times that reminds of a hockey stick, in
other words, enumeration times do not increase linearly as the number of items to be enumerated increases.
For the first few items, each additional object increases
response times only slightly (60 ms per item), but
after three or four items, every additional object
increases response times more dramatically (300 ms
per item). When participants are asked to enumerate
briefly presented arrays of items, enumeration accuracies and participants’ confidence ratings show a
similar trend: Enumeration of small numbers is highly
accurate and confident, whereas if the number of
objects is about four or higher, enumeration is more
error prone and the participants more uncertain of
their answers. The fact that enumeration is highly
accurate in the subitizing range demonstrates that
subitizing is not merely estimation of number, but
a process through which number can be recognized
accurately.
The view that subitizing and counting are different
processes is supported by neuropsychological patient
data showing that patients who demonstrate clear
deficits in counting may still be able to enumerate
small sets of numbers accurately (Dehaene and
Cohen 1994). In addition, functional brain imaging
studies have demonstrated that enumerating small
and large sets of items rely on partly different brain
circuits. It is important to note that subitizing and
counting are means by which number can be determined, they do not directly tell the mechanism by
which number, or quantity in general, is represented
in the brain. It has been proposed that small numbers
of objects can be represented discretely through separate indexes and, thus, practically without an error,
whereas larger quantities are coded by continuous
and variable magnitude representations. The ability
to individuate small numbers of objects has been

proposed to play a key role in the development of
numerical competence.

Important Scientific Research and
Open Questions
A number of theories have attempted to explain
how enumeration of small sets of items differs from
counting (see Trick and Pylyshyn 1994). Theories have
suggested that subitizing is merely fast counting,
recognition of canonical patterns associated with
number words, or that subitizing would reflect a
capacity limitation of the visual system, attention, or
working memory. Nevertheless, despite decades of
research, the exact processes involved in subitizing
remain elusive.
One of the most influential theories has proposed
that subitizing and counting rely on different attention
mechanisms (Trick and Pylyshyn 1994). According to
this view, subitizing is based on a preattentive parallel
processing stage during which a few items are individuated, although it is acknowledged that subitizing is an
intentional process and, thus, also constitutes an attentional component. Counting was, unsurprisingly,
deemed to be a serial, attention-demanding process:
Items or groups of items are selected one after another,
which is reflected, for example, in the fact that quantification times increase considerably as the number of
counted objects increases. Recent evidence has, however, shown that manipulations that minimize attentional processing compromise, not only counting, but
also subitizing. This demonstrates that also subitizing
is dependent on attentional resources (Railo et al.
2008). Despite the fact that subitizing is an attention
demanding process, the theory that subitizing and
counting recruit different attentional processes remains
viable. Attentional resources might enable the individuation or conscious apprehension of a few objects
at a time, while enumerating a higher number of
objects requires additional attention-dependent processes, such as updating the contents of working
memory.
It is yet unsolved whether subitizing is a strictly
visual process, or if auditory or tactile stimuli could
also be enumerated by subitizing. This is a significant
issue as it might reveal important insights into the
underpinnings of subitizing – whether subitizing
reflects a modality-specific capacity or a general limitation in the ability consciously access perceptual

Subject of Learning

information. Some studies have suggested that enumeration of tactile and auditory stimuli do reveal two
different quantification processes, but other studies
have reached an opposite conclusion. Even if an auditory or a tactile task would reveal the characteristic
hockey-stick function in reaction times and error
rates, conclusions would not necessarily be straightforward. For instance, what if an individual’s subitizing
range, or slope, is clearly different between visual, tactile, and auditory modalities? Could all nonetheless
exhibit subitizing?
It is known that subitizing performance can be
modified not only by changing the experimental procedure (e.g., the presentation time or pattern of the
items), but also by training, both in adults and in
children. Thus, subitizing can be aided and modified
by other processes, although how this exactly happens
is an outstanding question. Likewise, the developmental role of subitizing on later numerical skills needs to
be understood better. How do babies’ capabilities of
small-number discrimination become explicit smallnumber representations linked to number words, and
how do these early verbal number skills develop further
to verbal counting skills?

Cross-References
▶ Attention and the Processing of Visual Scenes
▶ Behavioral Capacity Limits
▶ Numerical Skills in Animals
▶ Working Memory

References
Dehaene, S., & Cohen, L. (1994). Dissociable mechanisms of
subitizing and counting: Neuropsychological evidence from
simultagnostic patients. Journal of Experimental Psychology.
Human Perception and Performance, 20, 958–975.
Hamilton, W. (1860). Consciousness – Attention in general. In L.
Mansel & J. Veitch (Eds.), Lectures on metaphysics and logic
(Vol. I, pp. 246–263).
Jevons, W. S. (1871). The power of numerical discrimination. Nature,
3, 281–282.
Kaufman, E., Lord, M., Reese, T., & Volkmann, J. (1949). The discrimination of visual number. The American Journal of Psychology, 62, 498–525.
Railo, H., Koivisto, M., Revonsuo, A., & Hannula, M. M. (2008). The
role of attention in subitizing. Cognition, 107, 82–104.
Trick, L., & Pylyshyn, Z. (1994). Why are small and large numbers
enumerated differently? A limited capacity preattentive stage in
vision. Psychological Review, 101, 80–102.

S

3235

Subject of Learning
FRANZ SCHOTT
Department of Psychology, Technical University
Dresden, Dresden, Germany

Synonyms
Learned change of behavior potential, learned competency, learned ability

Definition
The subject of learning is what is learned. Learning is
defined as a change in the behavior potential of a
system (a change in competence or ability) in a given
situation caused by experience and sustained for some
time (e.g., Bower and Hilgrad 1981, p. 11). The system
can be a human being, an organism, or a machine.
In the context of learning, the meaning of “behavior”
is understood in a broad sense, comprising operations such as moving, thinking, wishing, feeling, or
processing information. The behavior potential of a
system describes a certain competency (ability), for
example, knowledge of Canada, or a part of a competency, for example, knowing the capital of Canada. It is
not observable. What is observable is the performance
which refers to a certain competency, for example, a
person says “The capital of Canada is Ottawa.” The
learned change of a behavior potential in a given situation is the subject of learning. It is important to realize
that every subject of learning describes a change.
A behavior potential or a competency itself is not a
subject of learning. Therefore, learning objectives, learning outcomes, and learning content are not subjects of
learning. For example, a person learns something about
Canada: The change from that person’s previous
knowledge to the knowledge he or she possesses when
reaching the learning outcome or the learning objective
is the subject of learning.
The meaning of “subject” as a human being, such as
in the statement: “A subject is a being which has subjective experiences,” is not what is meant here.
There is a large variety of different subjects of learning. They depend on the type of behavior that is
acquired, for example, thoughts, emotions, motivations, observable behavior of human beings, or processes of a computer in a machine, as well as on the

S

3236

S

Subject of Learning

domain of the behavior, for example, mathematics,
various disciplines in science and humanities, music,
painting, sports, cooking, information technology.

Theoretical Background
Because common definitions of learning view the subject of learning as a change in behavior potential (ability, competency), a theoretical analysis of the subject of
learning has to clarify what in this context is meant by
“change,” “behavior,” and “behavior potential.” Let us
begin with the concept of behavior. One general characteristic of any behavior is a change in state. If nothing
happens, there is no behavior. Any change in state can
be broken down into a beginning state and a final state.
For example: Mary’s behavior in unlocking the front
door of her house consists in the beginning state: door
is locked, and the final state: door is unlocked. Many
behaviors can be broken down into parts of behavior.
In our example: taking the key, putting it into the lock,
and so on. Each of Mary’s activities of unlocking the
front door differs from her previous trails of unlocking
it. The set of all these different trails describes a type of
behavior: How Mary unlocks her front door. Because
learning refers to behavior change that is more or less
permanent, the subject of learning is always a change
of a type of behavior, not of one singular behavior. Mary
is able to unlock her front door; in other words: she
has the behavior potential to do it. A behavior potential describes the capability (competency) to execute
a certain behavior (performance) in a given situation.
The manner in which Mary’s husband unlocks the
front door may be different in some ways; it may be
more hasty and error-prone. The abstraction from the
various types of unlocking the front door (by Mary, her
husband, or other persons) describes the set of behaviors necessary for unlocking Mary’s front door. The
abstraction of unlocking all doors by using a key
describes an even more abstract type of behavior. The
more abstract the description of a subject of learning is,
the more problems of misunderstanding may occur.
Take as an example: “learning to behave as a responsible citizen.”
Let us now focus on the concept of “change.” As
mentioned above, learning is defined as a system’s
change of behavior potential in a given situation. Any
behavior describes a change in state. If we call this
change a first-order change, a behavior change describes
a change concerning a certain change in state, which is a

second-order change. In educational settings, a secondorder change of this kind is a change from the inability
to execute a certain task correctly to the ability to do so.
Take as an example a behavior change in fact knowledge:
Mary has to perform a completion test. She reads: “The
capital of Canada is . . .?” This is the beginning state of
the behavior. She fills in the space with “Toronto.” The
final state of this behavior is: “The capital of Canada is
Toronto.” This statement is wrong. Mary has to learn.
After this learning process, the beginning state of Mary’s
factual knowledge in this situation is again: “The capital
of Canada is . . .?,” but the final state will be: “The capital
of Canada is Ottawa.” The states describe the content
aspect of the behavior or task; the change describes the
behavior aspect. In our example, the content aspect is
“The capital of Canada is Toronto” and “The capital of
Canada is Ottawa”; the behavior aspect is completing
“The capital of Canada is . . .?” by filling in the name of
the capital. Learning is indicated here by the substitution of Toronto by Ottawa. This substitution can be an
observable behavior of speaking or writing, but it
is also a thinking process. This substitution as a change
in behavior potential describes a certain subject of
learning.
To summarize: The subject of learning is what is
learned; learning is a change in a behavior potential
(competency, ability) and can be described as a secondorder change.
A third-order change of states can define the type of
change in the learning process.

Important Scientific Research and
Open Questions
The large variety of subjects of learning leads to the
question of whether it is possible to identify universal
characteristics for any subject of learning. In the quest
for parsimonious explanations, some psychological
theories, especially behavioristic ones, reduce all psychological phenomena to two classes of events: stimuli
and responses. In these stimulus–response models the
subject of learning is a change in a certain response to
a given stimulus.
Current cognitive theories in psychology focus
on information processing as the basis of behavior
and behavioral change. When using these theories, we
have to ask the question of whether or not a change in
behavior can be reduced to a change in information or
information processing. We have to take into account

Subject of Learning

two problems: What is the characteristic of information? And is behavior always physical?
The nature of information is not physical. Information is immaterial, but it needs some material to be
realized. When Mary has learned that Ottawa is the
capital of Canada, she can say or write: “Ottawa is the
capital of Canada.” The medium transporting this
information can vary: air when spoken, light when
written; the information itself remains the same. Additionally, although this information is immaterial, it has
a physical basis in Mary’s brain.
A change in information by itself or a change in
information processing by itself is often not a sufficient
condition for a change in a behavior potential. If, for
example, Mary wants to improve her tennis serve, the
realization of this improvement via learning depends not
only on a change in the information process controlling
the behavior. This realization also depends on her fitness,
the individual structure of her limbs, her potential to
strengthen her muscles, etc. In addition, the constraints
of the necessary change in information processing in
Mary’s brain are not only dependent on logical laws
but also on characteristics of her neuronal equipment.
Therefore, the analysis of learning and learning
subjects is not only a matter of information processing,
but also a matter of the characteristics of the system in
which the learning happens.
It is problematic to equate subjects of learning, as
changes of behavior potentials (ability, competency),
with changes in cognitive structures, representations,
or neural information processing as some researchers
do, because this may lead one to confuse the task of
describing the subject of learning with the task of
explaining learning processes. Since the acquisition of
the same subject of learning may be explained by different learning theories, from a methodological point of
view it is useful to separate description and explanation.
Up to now, no general learning theory exists that
can explain all learning phenomena. Accordingly, there
is no general theory concerning the subject of learning.
The reason for this may be the large variety of different
types of subjects of learning. Furthermore, applied
research on learning focuses only on specific types of
subjects of learning such as those found in mathematics
or sports. The development of a general theory of learning requires an analysis of the subject of learning.
A general theory of this kind should answer questions
like the following: Which characteristics of the subject

S

of learning are universal and which are specific,
depending on the domain in focus (for example: what
are typical differences between subjects of learning in the
acquisition of mathematics and those in the acquisition
of a foreign language)? What kind of influence do these
different characteristics have on the processes of learning
and what are good explanations for this influence?
Looking at behavior as a change in state may offer
some possibilities for constructing a general theory of
learning. For example, the analysis of the states may
help one to quantify the content aspect of a certain
behavior by counting the elements and relations
describing the state. And one can also quantify the
behavior aspect by calculating the similarity between
the beginning state and the final state (Schott 1992).
Specific structures of states and changes in state can be
discovered.
A problem in defining subjects of learning is
determining which aspects of a whole situation constitute parts of the subject of learning or of the surrounding situation where the behavior takes place. With
regard to our example of unlocking the front door, we
have to ask: Are the type of lock and environmental
conditions such as rain, darkness, and so on parts of the
description of the behavior? Or are they parts of the
description of the situation where the behavior takes
place? In classical ▶ conditioning learning is often
described as performing the same response to
a different stimulus. But if the stimulus controlling
the response is viewed as part of the beginning state
of the behavior, classical conditioning also causes
a change in a behavior potential. A similar problem is
the following: One may ask the question whether or not
each behavior describes a change in state. For example:
Mary is offered a cigarette and she declines. Is her
refusal a change in state; does anything happen? Since
her refusal is a reaction to a temptation which can be
described as a part of the beginning state of the behavior, it can be interpreted as a change in state. The
answers to these questions lead to problems concerning
the dependence of behavior and situation and to problems concerning the ▶ transfer of learning.

Cross-References
▶ Competency-Based Learning
▶ Conditioning
▶ Learning Objectives
▶ Transfer of Learning

3237

S

3238

S

Subliminal Perception

References

Definition

Bower, G. H., & Hilgrad, E. R. (1981). Theories of learning. Englewood
Cliffs: Prentice-Hall.
Schott, F. (1992). The useful representation of instructional objectives. A task analysis of task analysis. In S. Dijkstra, H. P. M.
Krammer, & J. J. G. Merriënboer (Eds.), Instructional models in
computer-based learning environments (pp. 43–60). New York:
Springer.

Substitution error is the automatic replacement of one
item in a sentence, strategy, word, or phoneme when
the specific information is forgotten or unknown. For
instance, when the subject is requested to repeat nonwords immediately after listening to them, he/she
might produce “zama” after hearing the nonword
“jama.” In this example, the substitution of the phoneme [j] occurred for the sound of [z].

Subliminal Perception
This is the unconscious processing of external stimuli.

Subliminal Perceptual Learning
▶ Task-Irrelevant Perceptual Learning

Submodular Function
▶ Adaptive Proactive Learning with Cost-Reliability
Trade-off

Substantive Rationality
The notion that rationality should be studied by focusing on an abstract analysis of the situation rather than
on decision makers and the processes they use. Sometimes used as a synonym for global rationality.

Substitution Errors in Learning
F. H. SANTOS
Department of Experimental Psychology, UNESP,
São Paulo State University, Assis, São Paulo, Brazil

Synonyms
Error of substitution

Theoretical Background
In Cognitive Psychology, the term is mostly used to
explain errors that occurred on the performance of
paradigms, tests, and tasks related to language, memory, attention, calculation, and executive functions.
Therefore, substitution error is a general concept in
cognition science.
In memory research, substitution errors tap on the
interaction between linguistic and memory fields
(Gathercole et al. 1994). It is well established that in
the memory test for nonwords, a short-term memory
skill measured by instruments such as CNRep or
BCPR, children with typical development will produce a small percentage of errors, whereas children
with specific language impairment will perform significantly poorly than their counterparts in this task.
Phonemic errors (when the identity of the target item
is not present in any position in the repetition)
are classified as substitution, omission, and addition
while order errors (refer to the movement of the target
to nontarget position in the output attempt) are classified as migration. However, the phonotastic rules
(syllabic stress, length, spelling, syllabic construction,
prosody) of the language determine the characterization of the substitution errors. For instance, in English
nonwords the substitution errors mainly occur by the
end of the stimuli, while in Portuguese it might happen in both, by the middle and at the end of the
nonwords. These differences may be discussed in
terms of distinctiveness, clustering, and redintegration hypothesis. Therefore, it is important to obtain
specific stimuli for each idiom (Santos et al. 2006;
Santos and Bueno 2003).
Age effect. The number of phoneme substitution
errors decrease in older children, despite the fact that
the system capacity remains relatively stable during
the development. However, the schooling effect could
be a better determinant of the performance since

Suggestive Accelerative Learning

the phonological awareness improves with the literacy
proficiency.
Length effect. The length effect reflects the subvocal
rehearsal component of the phonological working
memory. Thus, more phoneme substitution errors are
observed in long over short items. This effect is also
observed for Chinese logographemes writing.
Wordlikeness effect. The rate of wordlikeness
between nonword and general words affects the performance in the nonword repetition task; there is a
progressive increase of the substitution error rates
from nonwords with high to low similarity with
words.
Position effect. The pattern of errors in nonword
repetition depends on the phonotactic rules of the
language. For instance, in Portuguese 5-syllablepseudoword errors occur mainly in the middle of
the stimuli (third syllable), before the syllabic stress.
However, substitution errors also appear more in the
end of stimuli (fifth syllable), after the stress. In certain
circumstances, positional errors result from impairment to an abstract ordinal code with graded activation of letter positions from first to last, and this
code is specific to tasks involving orthographic
representations.
Substitution errors can also be investigated by
the context of long-term memory (prospective, gist,
and episodic memory) using recognition tasks, remembering/knowing paradigms, measures of everyday
functioning, and cue detection focused mainly on
semantic substitution errors.
In language research, the term “substitution errors”
is used to describe the pattern of errors production
observed in patients with aphasia, dyslexia, dyspraxia
and children with specific language impairment dueling with reading, comprehension, writing, and repetition tasks. In general, the pattern of errors is similar to
controls but the quantity of error is higher in people
with language impairments. Substitution errors in language studies may be associated with the use of the
third person singular pronouns, anomia, semantic substitution errors (say “arm” when “leg” is intended),
semantic verbs substitution, and substitution of an
orthographically similar word with letters that overlap
the target either in early or late letter positions. In the
language field, substitution errors are investigated
mostly in lexical decision, reading, writing, speech production, and comprehension tasks.

S

3239

Important Scientific Research and
Open Questions
A search accomplished on March 14, 2010, at medline/
pubmed using the words “substitution and errors
and learning” found 56 articles. After reading their
abstracts, only 42 were kept on the list considering the
specific subject because in the other studies the word
“substitution” was not used as an error type. The
period related to these papers ranged from 1976 to
2009. Substitution errors were reported in a variety of
subject studies: animal response in discrimination or
problem solving tasks (pigeons, rats, monkeys), gender
differences in attention, memory or executive function
tasks, children or adults’ speech production or reading,
cognitive performance of brain-damaged people, psychoactive drug effects, people with dyslexia, apraxia of
speech, dyspraxia and aphasia, and mental disorders
(dementia, depression, melancholia, schizoaffective
disorder, and schizophrenia).

Cross-References
▶ Achievement Deficits of Students with Emotional
and Behavioral Disabilities
▶ Language-Learning-Based Disabilities
▶ Phonological Representation

References
Gathercole, S. E., Willis, C. S., Baddeley, A. D., & Emslie, H. (1994).
The children’s test of nonword repetition: a test of phonological
working memory. Memory, 2, 103–127.
Santos, F. H., & Bueno, O. F. (2003). Validation of the Brazilian
children’s test of pseudoword repetition in Portuguese speakers
aged 4 to 10 years. Brazilian Journal of Medical and Biological
Research, 36(11),1533–1547.
Santos, F. H., Bueno, O. F. A., & Gathercole, S. E. (2006). Errors in
nonword repetition: bridging short- and long-term memory.
Brazilian Journal of Medical and Biological Research, 39, 371–385.

Successive Approximation
Procedure
▶ Insight Learning and Shaping

Suggestive Accelerative
Learning
▶ Superlearning

S

3240

S

Summation

Summation
▶ Cue Summation and Learning

Superlearning
KAZUHIKO HAGIWARA
School of Languages and Linguistics, Griffith
University, Nathan, QLD, Australia

same authors expanded a chapter in the book to publish another book, “Superlearning”.
The method was largely inspired by the Bulgarian
development of Suggestopedia. It borrowed elements
such as the use of Baroque music and the repetitive
cycle in the use of three different types of intonation
from an experimental version of Suggestopedia. It also
added elements such as rhythmical abdominal breathing exercise, and visualization and meditation training
from other yogic therapies such as Transcendental
Meditation and Sophology.

Influences from Yoga and Hypnosis
Synonyms
Affective Learning; Neuro Linguistic Programming;
Speedlearning; Suggestive Accelerative Learning

Definition
“Superlearning” is the title of a book authored by Sheila
Ostrander, Lyn Schroeder, and Nancy Ostrander. It was
first published in 1979 and an updated version was
published in 1997 as “Superlearning 2000.” The term
superlearning also refers to a group of memorization
and self-developing techniques. In a broader sense, the
term superlearning is often used as a synonym for other
learning speedup methods. In a narrow sense, it is
a registered trademark at the US Patent and Trademark
Office (as of 2009).

Theoretical Background
Superlearning, as a learning method, claims that it
establishes a quick and stress-free learning by enhancing learner’s abilities through suggestion and/or autosuggestion that is directed to the relaxed mind and
body. Characteristically, it uses yoga-like breathing
exercise, visualization, and meditation training with
the help of specified background music. A typical program includes a set of learning materials that consist of
a book with the learning content and recorded audio
materials. Superlearning is basically designed as a selflearnable program, and it does not necessarily require
a teacher and a learning group.
Superlearning was created on the basis of journalistic research and interviews conducted in the communist Russia, Bulgaria, and Czechoslovakia in 1967 and
1968 by the authors of the book “Psychic Discoveries
Behind The Iron Curtain” (1969). Ten years later, the

Superlearning is generally positive about hypnosis. It
interprets Suggestopedia as a group of yogic techniques
to obtain hypnotic effects in the waking state (p. 66). As
a result, it puts special importance on the tempo or the
speed of the music as the means to obtain the desirable
state of relaxation. It believes that the speed of 60 beats
per minute in slow Baroque music can assist the ideal
rhythm of breathing to establish the stable appearance
of alpha brain waves. It also believes that a cycle of
certain beat count (8 or 10 s cycle) can assist learners
to obtain a certain hypnotic effect. So-called “New Age
Music” is also used in Superlearning for the same
purpose. Such music pieces are generally characterized
as calm, slow, steady, and “hypnotic” with repetitive
phrases and passages. Special intonation used for the
“concert reading” also reinforces such an effect. Three
types of intonation are repetitively used for each phrase
in the textbook when it is read: “(1) normal (declarative), (2) soft whisper (quiet, ambiguous, misleading
tone), (3) loud, commanding voice (with a domineering tone)” (p. 69).

Use of Suggestions and Relaxation
Superlearning preconditions the brain to maximize the
learners’ abilities by verbal suggestions directed to the
state of physical and mental relaxation. Verbal and
direct suggestions are given to the learners so that
they believe in their abilities to overcome the limitations built into their personalities. Learners are directed
to passively receive the suggestions and the learning
contents. The required state of relaxation for such
a purpose is measured as the maximum frequency of
appearance of alpha brain waves. It claims such relaxation is attained and maintained, without using
a biofeedback device, by abdominal rhythmical

Superlearning

breathing that is paced by the constant “60 beats per
minute” rhythm of the music.

Important Scientific Research and
Open Questions
Needs for Distinction Between
Superlearning and (De)
Suggestopedia
Superlearning is often confused with Suggestopedia
(or Desuggestiopedia) because of its attribution to
Lozanov’s research on yoga and his early success
with hypermnesic experiments with nonclinical
suggestions. However, such an attribution to
Lozanov’s’ research outcomes does not necessarily
secure grounds on which one can conclude that
Superlearning is Suggestopedia. Rather, Superlearning
and Suggestopedia should be considered as two different learning methods as they contradict each other in
major aspects of their key concepts such as: (1) attitude
toward hypnosis, (2) use of suggestion, and (3) required
“relaxation.” For example, whereas Superlearning is
positive about the use of hypnosis and tries to give
learners control over their mind and body through
hypnotic effects, Suggestopedia is consistently negative
about the use of hypnosis and tries to give learners
control by freeing them from a hypnosis-like state

S

3241

induced by the “social suggestive norms.” Accordingly,
whereas Superlearning uses verbal and direct suggestions toward the state of meditative deep relaxation,
Suggestopedia uses nonverbal and peripheral suggestions toward the state of what it calls “psycho-relaxation” that is very close to waking state (Lozanov 1978).
Following table shows other differences in the two
methods (Table 1).
Previous research made on either Superlearning or
Suggestopedia has not been able to confirm the
methods’ superiority to conventional teaching methods,
and criticism has often been directed at Superlearning or
Suggestopedia regarding their lack of scientific consistency (e.g., Scovel 1979; Hagan 2002). However, many
such pieces of research and criticism are based on
a mixture of the two similar but completely different
methods. Each method should be clearly distinguished
from the other to be precisely evaluated.

Is Superlearning Effective? If So, Why?
As mentioned above, Superlearning cannot fully attribute its effectiveness to the success of Suggestopedia
because neither yoga nor hypnosis, the main properties
of Superlearning, has ever been the central issue in
Suggestopedia. It is true that much scientific research
has been conducted on yogi, their brain wave patterns,
depth of relaxation, and their mental potentials, or

Superlearning. Table 1 Differences in Superlearning and (De) Suggestopedia

Theory and Theoretical
aims
background

Superlearning

(De) Suggestopedia

Journalistic study on “New Age”
science:

Study of suggestion (clinical and nonclinical)

Suggestology

Psychology, Psycho physiology

Yoga

Psychiatry, Mental health care

Sophology

Neuro science, Brain physiology

Gestalt psychology
Russian studies on ESP
Biofeedback researches
Brain hemispheric researches
Aims

Self enlightenment

Mental health/Brain health

Quick learning

Liberation of personality

Hypermnesia (super memory)

Promotion of creativity

Physical development

Restoration of inherent learning ability

Development of extra sensory
perceptions (ESP)

S

3242

S

Superlearning

Superlearning. Table 1 (continued)

Practice

Learning forms

Use of hypnosis

Superlearning

(De) Suggestopedia

Self learnable

Not self learnable

A teacher or a learning group not
necessary

A teacher and a learning group required

Positive

Negative

Breathing exercise Rhythmic abdominal breathing

Avoided

8 s cycle recommended
Baroque music and New Age music
with the tempo of 60 beats per
minutes

Symphonies and concertos from Classical to
Romantic period (1st reading session)

Use of visual art

Aid for visualization exercise

A source of “desuggestive suggestions”

Chairs for concert
sessions

Reclining, laying

Normal, upright

Learner attitudes
for concert
sessions

Passive

Pseudo-passive, creative

Suggestion type

Verbal, directive

Nonverbal, peripheral, desuggestive

Required state of
relaxation

Physical and mental relaxation

Concentrative psycho-relaxation (Mental only)

The state of steady alpha wave
established by rhythmic abdominal
breathing exercise

The state of vigorous mental activity established in
the creative personal communication, and
dynamism in the music

Visualization
exercise

Imagery guided with verbal
suggestions while meditation

Avoided

Use of recorded
voice in the
concert sessions

Yes

No

Information
directed to

Subconscious

Conscious and paraconscious

Characteristics of
the course
activities

Hypnotic

Holistic (central and peripheral)

Affective

Awakening, Critical

Music for concert
sessions

Amount of
information

Concertos and organ music from Baroque period
(2nd reading session)

Programming

Spontaneous, Creative

Subliminal

Dynamic, multi-angled

Calm

Changing, fun, surprising Pleasantly emotional

Not considered

Large mass of information stored in the
paraconscious area is important requirement for the
healthy activity of the brain

hypnosis and hypnotized subjects. However, this
research has not necessarily proven the educational
effectiveness of Superlearning. Comparative study in
terms of attainment of proficiency against other nonyoga-hypnotic methods may be desirable.

Is Superlearning Safe?
On one hand, Superlearning is positive about the use
of hypnosis, on the other hand, it seems that care is

not well taken about the negative side of hypnosis.
Lozanov (2009 pp. 47–50) shares his apprehension
with Weitzenhoffer (1989) that the use of hypnosis in
a nonclinical environment can cause serious problems
for the personality of the hypnotized subject. Lozanov
points out that the techniques that Superlearning
heavily relies on, such as breathing exercise and visualization training, can induce unwanted hypnosis.
A longitudinal research study on the relationship

Supervised Learning

between the effects of such activities and the mental
health of the learners is desired. Such research is important for Superlearning to ensure its safety.

Cross-References
▶ Attention, Memory, and Meditation
▶ Desuggestopedia
▶ Learning and Recall Under Hypnosis
▶ Music Therapy

References
Hagan, K. (2002). Pseudoscience at 30,000 feet: Suggestology,
suggestopedia and accelerated language learning Skeptic, Fall.
Lozanov, G. (1978). Suggestology and outlines of suggestopedy. New
York: Gordon and Breach.
Lozanov, G. (2009). Suggestopedia/reservopedia theory and practice of
the liberting-stimulating pedagogy on the level of the hidden
reserves of the human mind. Sofia: Sofia University Publishing
House.
Ostrander, S., & Schroeder, L. (1970). Psychic discoveries behind the
iron curtain. Englewood Cliffs: Prentice-Hall.
Ostrander, S., Schroeder, L., & Ostrander, N. (1982). Superlearning.
New York: Laurel.
Scovel, T. (1979). Book review of Lozanov’s “Suggestology and Outlines of Suggestopedy”. TESOL Quarterly, 13(2), 255–266.
Weitzenhoffer, A. M. (1989). The practice of hypnotism (Vol. I and II).
New York: Wiley.

Superstitious Behavior
▶ Causal Learning and Illusions of Control

Supervised Learning
QIONG LIU1, YING WU2
1
FX Palo Alto Laboratory, Palo Alto, CA, USA
2
EECS Department, Northwestern University,
Chicago, IL, USA

Synonyms
Active learning; Classification; Inductive machine
learning; Learning from labeled data; Learning with
a teacher; Regression; Semi-supervised learning;
Supervised machine learning

S

3243

Definition
Supervised Learning is a machine learning paradigm
for acquiring the input-output relationship information of a system based on a given set of paired inputoutput training samples. As the output is regarded as
the label of the input data or the supervision, an inputoutput training sample is also called labeled training
data, or supervised data. Occasionally, it is also referred
to as Learning with a Teacher (Haykin 1998), Learning
from Labeled Data, or Inductive Machine Learning
(Kotsiantis 2007). The goal of supervised learning is
to build an artificial system that can learn the mapping
between the input and the output, and can predict the
output of the system given new inputs. If the output
takes a finite set of discrete values that indicate the class
labels of the input, the learned mapping leads to the
classification of the input data. If the output takes
continuous values, it leads to a regression of the
input. The input-output relationship information is
frequently represented with learning-model parameters. When these parameters are not directly available
from training samples, a learning system needs to go
through an estimation process to obtain these parameters. Different from Unsupervised Learning, the training data for Supervised Learning need supervised
or labeled information, while the training data for
unsupervised learning are unsupervised as they are
not labeled (i.e., merely the inputs). If an algorithm
uses both supervised and unsupervised training data, it
is called a Semi-supervised Learning algorithm. If an
algorithm actively queries a user/teacher for labels in
the training process, the iterative supervised learning is
called Active Learning.

Theoretical Background
Figure 1 shows a block diagram that illustrates the form
of Supervised Learning. In this diagram, (xi, yi) is a
supervised training sample, where “x” represents system input, “y” represents the system output (i.e., the
supervision or labeling of the input x), and “i” is the
index of the training sample. During a Supervised
Learning process, a training input xi is fed to the Learning System, and the Learning System generates an output ỹi. The Learning System output ỹi is then compared
with the ground truth labeling yi by an arbitrator that
computes the difference between them. The difference,
termed Error Signal in this diagram, is then sent to the
Learning System for adjusting the parameters of the

S

3244

S

Supervised Learning

yi

Training Data Set
{(x1, y1), ...,
(xn, yn)}

xi

Learning System

∼
yi

Abitrator (−)

Error Signal

Supervised Learning. Fig. 1 Block diagram that illustrates the form of Supervised Learning

learner. The goal of this learning process is to obtain a
set of optimal Learning System parameters that can
minimize the differences between ỹi and yi for all i,
that is, minimizing the total error over the entire training data set.
A notable phenomenon is that a minimum training
error does not necessarily indicate a good performance
in testing. Training is referred to as the learning process
that estimates the parameters of the learner based on
the ground truth supervised data seen, while testing is
to evaluate the predictions of the learner for the data
unseen, that is, the data used in testing have not been
included in the training process. Therefore, even if
a learner achieves a minimum error on the set of
training data, it does not guarantee to perform well to
the data unseen. The reason for this is mainly due to the
possible overfitting to the training data, that is, the
learner has an unnecessary order of complexity in
learning the mapping. This issue is referred to as the
generalizability. A good learning algorithm must have
a good generalizability. To take into consideration of
the generalizability in designing the learner, a learning
algorithm needs to balance the objective of minimizing
the training error and the complexity of the learner
(e.g., the structure and the order of the learner).
For example, in the Support Vector Machine, the
generalizablity of the learner is characterized by the
margin of the learned discrimination boundary.
The larger the margin, the better the generalization.
The support vector machine learns a maximum margin
classifier over the training set, and thus it naturally leads
to a good generalization performance (Vapnik 1995).
The Supervised Learning paradigm does not restrict
the sources of the input or output data. The input or
output may belong to a vector space or a set of discrete
values. The learning paradigm does not have special

restrictions on the arbitrator either. If yi is drawn from
a continuous space, the error signal is usually computed via yi  ỹi. If yi belongs to a set of discrete values,
the arbitrator usually outputs the error signal based on
the equality between yi and ỹi. For example, the arbitrator may output 0 for equaled yi and ỹi and output 1
for different yi and ỹi.
There are different approaches to the design of
the Learning System in Supervised Learning. Some
well-known approaches include the logic-based
approach, the multi-layer perceptron approach, the
statistical learning approach, the instance-based learning approach, the Support Vector Machines, and
Boosting.

Important Scientific Research and
Open Questions
Advantages and Disadvantages
The foremost advantage of Supervised Learning is that
all classes or analog outputs manipulated by the algorithm of this paradigm are meaningful to humans.
And it can be easily used for discriminative pattern
classification and for data regression. But it also has
several disadvantages. The first one is caused by the
difficulty of collecting supervision or labels. When
there is a huge volume of input data, it is prohibitively
expensive, if not impossible, to label all of them. For
example, it is not a trivial task to label a huge set of
images for image classification. Second, as not everything in the real world has a distinctive label, there are
uncertainties and ambiguities in the supervision or
labels. For example, the margin for separating the
two concepts of “hot” and “cold” is not distinct; and
it is difficult to name an object that is a cross between
a love seat and a bed. These difficulties may limit the

Supervised Learning in Spiking Neural Networks

applications of the Supervised Learning paradigm in
some scenarios.
To overcome these limitations in practice, other
learning paradigms, such as Unsupervised Learning,
Semi-supervised Learning, Reinforcement Learning,
Active Learning, or some mixed learning approaches
can be considered.

S

3245

Supervised Learning in Spiking
Neural Networks
RĂZVAN V. FLORIAN
Center for Cognitive and Neural Studies (Coneural),
Romanian Institute of Science and Technology,
Cluj-Napoca, Romania

Applications
Supervised Learning enables a machine to learn the
human behavior or object behavior in certain tasks.
The learned knowledge can then be used by the
machine to perform similar actions on these tasks.
Since the computing machinery may perform some
input-output mappings much faster and more persistent than the human, machines equipped with a good
supervised learner can perform certain tasks much
faster and accurate than the human. On the other
hand, because of the limitation in hardware, software,
and algorithm designs, existing Supervised Learning
algorithms still cannot match human’s learning ability
on many complicated tasks.
Supervised Learning have been successfully used
in areas such as Information Retrieval, Data Mining,
Computer Vision, Speech Recognition, Spam Detection, Bioinformatics, Cheminformatics, and Market
Analysis (Wikipedia 2010).

Synonyms
Spiking neural networks; Supervised learning

Definition
In ▶ supervised learning, the learner is presented,
for a set of inputs, the desired (target) outputs
corresponding to the given inputs. The learner should
then learn to predict the correct output for any valid
input. Spiking neural networks are neural models,
where neurons transmit information through each
other by firing action potentials, or spikes, as real
neurons do. Supervised learning in spiking neural networks refers to how spiking neurons modify their
parameters in order to be able to reproduce and generalize the input–output associations that they were
taught in the past.

Theoretical Background
Cross-References
▶ Adaptation and Unsupervised Learning
▶ Classification of Learning Objects
▶ Feature Selection (Unsupervised Learning)
▶ Interactive Learning
▶ Machine Learning from Pairwise Relationships
▶ Statistical Learning Theory and Induction
▶ Supervised Learning in Spiking Neural Networks

References
Haykin, S. (1998). Neural networks: A comprehensive foundation
(2nd ed.). Upper Saddle River: Prentice Hall. ISBN 0-13273350-1.
Kotsiantis, S. (2007). Supervised machine learning: A review of classification techniques. Informatica Journal, 31, 249–268.
Vapnik, V. (1995). The nature of statistical learning theory. New York:
Springer. ISBN 0-387-98780-0.
Wikipedia. (2010). Supervised learning. http://en.wikipedia.org/
wiki/Supervised_learning

Humans and animals learn through coordinated
changes in the properties of their neural systems. In
neural models, this is simulated by changes of the
parameters of these models, such as synaptic efficacies. The study of ▶ learning in artificial neural networks focuses on the rules that govern these changes
such that they allow the networks to process and
memorize information. Learning rules for neural
models are studied analytically and in computer simulations, and this often sheds light on the function of
synaptic and other forms of plasticity observed in
the brain.
In the brain, information is exchanged between
neurons through action potentials, or spikes. Neurons
integrate the spikes received through their input synapses and, when a threshold is reached, they fire spikes
at their turn, which are transmitted to other neurons.
The effects on a neuron of the spikes received through

S

3246

S

Supervised Learning in Spiking Neural Networks

a particular synapse depend on the synaptic efficacy of
that synapse.
In artificial neural networks where information
is represented by the continuous activation of
model neurons, which models the firing rate of real
neurons, supervised learning has been thoroughly
studied. The famous backpropagation learning rule
for this kind of networks is a cornerstone of ▶ connectionist theories of learning. However, the study
of supervised learning in spiking neural networks
has been hampered, until recently, by the discontinuities generated by spikes, which create difficulties
in applying the typical methods for deriving supervised
learning rules in artificial neural networks.

Important Scientific Research and
Open Questions
The first supervised learning method for spiking neurons was SpikeProp (Bohte et al. 2002), a method
inspired by the backpropagation algorithm used
for training classical, non-spiking artificial neural networks. SpikeProp works by minimizing the difference between the timing of an output spike and the
desired timing. The method is designed for adjusting
just the timing of a single (first) spike per output
neuron and assumes that the synapses are such
that each output neuron fires at least one spike for
the given inputs. The method is not suitable for
adjusting the number of output spikes nor for training
a network to fire given output spike patterns that
extend in time.
Pfister et al. (2006) have derived supervised learning rules for probabilistic neurons. The learning
method is based on gradient ascent in the space
of synaptic efficacies, which maximizes the likelihood
of having a trained neuron firing at the desired
moments. Because of the probabilistic framework,
the learning rules do not involve the actual timing of
the output spikes, but the probability of having a
particular output spike train given a particular input
spike train.
The tempotron (Gütig and Sompolinsky 2006)
implements supervised learning for a particular task,
where an output neuron either fires one spike or does
not fire, when presented with an input spike train. The
approach assumes that after the neuron emits a spike in
response to an input pattern, all other incoming spikes
have no effect on the neuron (are shunted), which is

artificial. The method requires information that is
nonlocal in time, needing to monitor the maximum
of the output, and information that is not available
to the neuron, such as the maximum of the membrane potential that would have been reached if the
neuron would have not fired. The timing of the output
spike cannot be controlled with this method. The
tempotron has a binary response and, thus, its output
cannot distinguish between more than two input
categories. All these constraints undermine its biological plausibility and its applicability to more general
problems.
ReSuMe (Ponulak and Kasiński 2010) is a general
supervised learning method for spiking neurons that
allows learning of arbitrary output spike trains. However, this learning rule has been conjectured without an
analytical justification, by analogy to the Widrow–Hoff
rule for analog neurons. To date, it has been shown
analytically that ReSuMe will converge to an optimal
solution only for the case of one input spike and one
target output spike. Simulations have shown that not
all the terms of the conjectured learning rule are needed
for learning.
Florian (2011) introduced two general supervised
learning rules for spiking neurons that encode information in the precise timings of spikes, for both input
and output (chronotrons). The first rule, E-learning, is
analytically derived and much more efficient than
ReSuMe, while the second rule, I-learning, has a high
degree of biological plausibility. Both learning rules
allow neurons to fire spikes at the desired timings,
with sub-millisecond precision. Using these rules,
chronotrons can learn to classify their inputs, by firing
identical, temporally coded spike trains for different
inputs belonging to the same class. When the input is
noisy, the classification also leads to noise reduction.
These rules allowed the computation of the memory
capacity of single spiking neurons.
I-learning implies that synaptic changes are proportional to the corresponding synaptic currents, being
thus biologically plausible. Postsynaptic spikes lead to
synaptic depression similar to anti-Hebbian spiketiming-dependent plasticity, while the timings of target
postsynaptic spikes trigger potentiation. The depression and potentiation should balance each other when
actual spikes occur at the target timings. These target
timings could be indicated by spikes coming from other,
teacher neurons. The trained neurons’ firing should

Supplantation Effect on Learning

then become increasingly correlated to the one of
teacher neurons, eventually mimicking their firing
with a lag corresponding to the delay of the arrival of
the teaching spikes. The existence of similar learning
rules in the brain remains to be established. Similar
mechanisms might be responsible for neural synchronization that modulates neural interactions, such as the
synchronization of thalamic neurons needed for driving
the cortex through weak synapses; for encoding of information through synchronization; or for the fine temporal tuning of excitation relative to inhibition that
contributes to stimulus selectivity in rat somatosensory
cortex.
Supervised learning rules for spiking neural networks were reviewed by Kasiński and Ponulak (2006).

Cross-References
▶ Learning in Artificial Neural Networks
▶ Reinforcement Learning in Spiking Neural Networks
▶ Supervised Learning

References
Bohte, S. M., Poutré, H. L., & Kok, J. N. (2002). SpikeProp: errorbackpropagation for networks of spiking neurons.
Neurocomputing, 48, 17–37.
Florian, R. V. (2011). The chronotron: a neuron that learns to fire
temporally-precise spike patterns. Nature Precedings. http://precedings.nature.com/documents/5190/version/3.
Gütig, R., & Sompolinsky, H. (2006). The tempotron: a neuron that
learns spike timing-based decisions. Nature Neuroscience, 9(3),
420–428.
Kasiński, A., & Ponulak, F. (2006). Comparison of supervised learning methods for spike time coding in spiking neural networks.
International Journal of Applied Mathematics and Computer Science, 16(1), 101–113.
Pfister, J.-P., Toyoizumi, T., Barber, D., & Gerstner, W. (2006). Optimal spike timing-dependent plasticity for precise action potential firing in supervised learning. Neural Computation, 18(6),
1318–1348.
Ponulak, F., & Kasiński, A. (2010). Supervised learning in
spiking neural networks with ReSuMe: sequence learning,
classification, and spike shifting. Neural Computation, 22(2),
467–510.

Supervised Machine Learning
▶ Supervised Learning

S

3247

Supplantation Effect on
Learning
JOERG ZUMBACH, BIRGIT REISENHOFER
Department of Science Education and Teacher
Training, University of Salzburg, Salzburg, Austria

Synonyms
Media-based observational learning; Script and/or
schema acquisition through external representations

Definition
The supplantation effect describes a function of media
that conducts a mental operation by an external representation. If a learner is not able to accomplish a
mental, covert operation internally and an external
representation is able to replace this process, supplantation occurs. External representation supporting supplantation can either be static representations like
graphs or dynamic representations like animations or
film. The basic function of the external representation
is to show a functional link between objects that cannot
be constructed by a learner on his or her own. Learners
can internalize this link and, thus, reconstruct a
relationship between internally represented objects
without greater mental effort. Nevertheless, an active
cognitive processing is required.

Theoretical Background
The term and the basic concept of supplantation
are derived from the work of Salomon and colleagues
in the late 1970s and 1980s. Based on socioconstructivism theories, Salomon (1970, 1979)
assumes that symbol systems are not only communicational devices but rather tools for thinking. A basic
assumption of supplantation is that media as external
representations are related and interact with internal
mental operations and representations. The mapping
process between external and internal representation
depends on the learners’ prior knowledge, because
learners have to understand the coding system as well as
the objects and their relationships. This does not include
that the relation between objects is already known but
rather that learners are able to understand this relation.
Vogel et al. (2007) describe this process as follows (see Fig. 1): A learner might have an internal

S

3248

S

Supplantation Effect on Learning

External representation

Object 1

Transformation1

Subjects’ mind

Object 1

Surface

OperationExternal

Transformation1

OperationSurface

Object 2

External

External
representations

Internal

Transformation2

Transformation1

Object 2

T2

Object 2

External

T2

OperationInternal

Object 2

Surface

Internal

Surface
representations

Internal
representations

Supplantation Effect on Learning. Fig. 1 Supplantation of operations

representation of two objects (Object 1 and Object 2)
but does not have a representation of the operation that
links these two objects. The external representation
shows this operation. At a first step (Transformation1)
a surface representation is built in the learner’s mind. In
a second step, an internal representation (Transformation2) is built. This process is not a unidirectional mode
of information processing but rather an interactive
mapping and schema construction process between
external, surface, and internal representations. While
Fig. 1 presents a complete process, Salomon (1972) as
well as Vogel et al. (2007, p. 1289) single out that several
degrees of supplantation are possible:
1. Including the necessary mental activities by showing nothing but the beginning state of a process
2. Presenting the initial situation and its final
modification
3. Presenting the initial situation and its transformation but not the final modification
4. Maximal supplantation: presenting the initial situation, its transformation, and the final modification
The degree of supplantation varies depending on
learners’ aptitudes, especially regarding their prior
knowledge and mental skills (i.e., supplantation is
a process that underlies an Aptitude-Treatment-Interaction, ATI). For example, Salomon (1979) shows that
prior knowledge is necessary for understanding the
objects and the transformation that should be
supplanted but – on the other hand – might also inhibit

the supplantation process; with already fair skill mastery
it is more effective to just activate this skill instead of
providing complete skill supplantation, because interference processes are likely to occur. With poor initial
mastery, learners benefit rather from supplanting codes
but not from codes that only aim to activate skills.
In general, the process of supplantation remains in
the tradition of observational learning transferred to
media and, meanwhile, multimedia-based learning
environments. Supplantation is compatible with most
of recent models of processing information from pictures and text as well as multimedia learning environments like Mayer’s S-O-I-model of multimedia
learning or the model of text and picture comprehension by Schnotz and Bannert (cf. Vogel et al. 2007).

Important Scientific Research and
Open Questions
Support for the supplantation hypothesis is provided
within several studies. Salomon (1972) used film as
media in order to show the effect of supplantation. In
several studies, he used different versions of films.
Within a supplantation version, an operation between
two states was presented by zooming in and out (the
material was about the overall impression and details of
Breughel’s paintings). In control group versions these
operations were shortened by only showing the initial
and the final state of an operation. Both of these conditions were compared to an activation condition
where learners were presented only the initial state of

Surface Approaches to Learning

an operation and were required to activate the appropriate mediating operation from their own mental
repertoire. Results reveal that the supplantation version
and the activation version led to superior results in
selecting details out of the presented paintings. There
was also a strong ATI-effect showing that poor scorers
on cue-attendance (important for singling out details)
and verbal reasoning benefited far more from filmic
supplantation than learners with high scores in these
aptitudes. A subsequent study using a different task
(laying out solid objects and folding them back again)
compared supplantative vs. short-circuited film and
replicated the findings from the prior study.
Nevertheless, not only films can be used to achieve
supplantation effects. In a study by Seel and Dörr
(1994) sets of images integrated in a computer program
were used to show different objects as well as their
transformations in a supplantation condition (the
learning domain was to watch a three-dimensional
object and then to imagine the corresponding orthogonal projection and vice versa). This was compared
with an imagery condition without the visualization
of the transformation. Authors found strong evidence
for the supplantation hypothesis stating that showing
the operations within both types of visualizations leads
to superior task performance than just presenting single object states without the transformation process.
Similar results by using a computer program for
interpreting data and graphs are reported by Vogel
et al. (2007).
In a study by Zumbach et al. (2008) authors
compared animations – within the domain of eutrophication of lakes – to static versions about the same
topic presented as text with pictures. The animations
showing dynamically the transition between different
states of a lake during the year proved to be more
effective in building up a mental model about the
process and underlying parameters and their relationship than text and pictures. This also supports the
supplantation hypothesis showing that media itself
can overtake cognitive functions by supporting knowledge construction through internalizing supplanted
operations.

S

References
Salomon, G. (1970). What does it do to Johnny? Viewpoints: Bulletin
of the School of Education, 46(5), 33–62.
Salomon, G. (1972). Can we affect cognitive skills through visual
media? AV Communication Review, 20(4), 401–422.
Salomon, G. (1979). Media and symbol systems as related to cognition and learning. Journal of Educational Psychology, 71(2),
131–148.
Seel, N., & Dörr, G. (1994). The supplantation of mental images
through graphics: Instructional effects on spatial visualization
skills of adults. In W. Schnotz & R. W. Kulhavy (Eds.), Comprehension of graphics (pp. 271–290). Oxford: Elsevier.
Vogel, M., Girwidz, R., & Engel, J. (2007). Supplantation of mental
operations on graphs. Computers & Education, 49, 1287–1298.
Zumbach, J., Reisenhofer, B., Czermak, S., Emberger, P., Landerer, C.,
& Schrangl, G. (2008). The role of attribution, modality, and
supplantation in multimedia learning. In J. Zumbach, N.
Schwartz, T. Seufert, & L. Kester (Eds.), Beyond knowledge: The
legacy of competence (pp. 237–246). Dordrecht: Springer.

Supplementary Questions
▶ Adjunct Questions: Effects on Learning

Supporting Student-Centered
Learning Tasks/Environments
▶ Scaffolding for Learning

Supportive Questions
▶ Adjunct Questions: Effects on Learning

Suppositions
▶ Beliefs About Language Learning

Cross-References
▶ Aptitude-Treatment-Interaction
▶ Children’s Learning from TV
▶ Multimedia Learning

3249

Surface Approaches to Learning
▶ Deep Approaches to Learning in Higher Education

S

3250

S

Surprise and Anticipation in Learning

Surprise and Anticipation in
Learning
LUIS MACEDO1, RAINER REISENZEIN2, AMILCAR CARDOSO1
1
University of Coimbra, Coimbra, Portugal
2
University of Greifswald, Greifswald, Germany

Synonyms
Anticipatory behavior; Curiosity and learning; Expectation-based behavior; Role of unexpectedness vs predictability in learning; Surprise-based learning

Definition
Surprise. Common-sense psychology conceptualizes
surprise as a peculiar state of mind, usually of brief
duration, caused by unexpected events of all kinds.
Subjectively, surprise manifests itself centrally in a phenomenal experience or “feeling” with a characteristic
quality, that can vary in intensity (e.g., one can feel
slightly, moderately, or strongly surprised). In addition, the surprised person is often aware of a variety
of surprise-related mental and behavioral events: she
realizes that something is different from usual or other
than expected; she notices that her ongoing mental
processes and actions are being interrupted and that
her attention is drawn to the unexpected event; she may
feel curiosity about the nature and causes of this event;
and she may notice the occurrence of spontaneous
epistemic search processes. Objectively (i.e., from the
perspective of the outside observer), surprise may
reveal itself – depending on circumstances – in any of
a number of behavioral indicators, including: interruption or delay of ongoing motor activities; orienting of
the sense organs to the surprising event; investigative
activities such as visual search and questioning others;
spontaneous exclamations (“Oh!”) and explicit verbal
proclamations of being surprised; and a characteristic
facial expression consisting, in full-blown form, of eyebrow-raising, eye-widening, and mouth-opening/jaw
drop. Furthermore, psychophysiological studies suggest that surprising events may elicit a variety of bodily
changes, commonly subsumed under the so-called
orienting response, such as a temporary slowing of
heart rate and an increased activity of the eccrine
sweat glands. It must be emphasized, however, that
the behavioral manifestations of surprise occur by no

means in all situations and are in general only loosely
associated with one another.
Anticipation, surprise, and learning. Many natural,
as well as some artificial agents are not (just) reactive
but anticipatory in that they are able to generate predictions or expectations of possible future situations
and to use these models of the future to optimize their
current decisions. To be adaptive, it is essential that an
agent’s models of the current and future situations are
approximately correct, and this means that these
models must, when necessary, be updated to reflect
newly acquired information. It is in this context that
the surprise mechanism plays an important role.
Although surprise is not confined to explicit learning situations, it occurs frequently in such situations,
and it plays an important role for both explicit and
incidental learning. Surprise occurs in learning situations, for example, when the anticipated effects of an
observed event, or an own action, are inconsistent with
the feedback from reality. The surprising effects may
lead to learning or schema-update (e.g., in changed
expectations about the effects of an action). In addition, the agent may actively select those parts of the
environment that it expects to be most relevant for
learning (see below).

Theoretical Background
Many natural, as well as some recent artificial cognitive
agents are not (purely) reactive but anticipatory. This
means that the behaviors of these agents are guided by
a model of the environment that includes the possible
future development of the current situation. Such
a model may also be used to decide which information
to focus on, for example, in learning situations.
Typically, the states of the world relevant for decision and action are only partly known to agents. Due to
the imperfection of their perceptual and sense-making
processes, and because it is impossible or too costly to
obtain all potentially relevant pieces of information
about the world, agents never have complete information about the environment in which they act.
According to several psychologists, ethologists, and
cognitive scientists, the evolutionary solution to this
problem in humans and other animals consisted of the
emergence of information-processing mechanisms
which generate assumptions or expectations that fill in
gaps in the available knowledge (e.g., Piaget 1952; cited
in Macedo et al. 2009). Although many of these

Surprise and Anticipation in Learning

assumptions or expectations concern unknown current
states of the world, those most important for action
concern possible future situations. The efficacy of an
agent’s behavior can therefore be greatly improved if its
model of the world includes a good model of the future
states of the world. Because future worlds cannot be
perceived in any direct sense, natural agents have
evolved mechanisms (inference procedures) that
allow them to make predictions about future states of
the world on the basis of their knowledge or beliefs
of the current state of the world. However, relying
on assumptions and expectations rather than (only)
on perceptions of the current situation comes at a
cost: The assumptions and expectations may turn out
to be false.
Although anticipatory agents cannot completely
prevent such prediction errors, they can learn from
them by correcting their model of the world, i.e., by
updating their schemas or beliefs. To be able to do this,
they need to have a mechanism or mechanisms that
detect inconsistencies or conflicts between their
assumptions and expectations (either of the present
state of the world or of future states of the world) and
the actual state of the world, as soon as the latter
becomes known to them. One such mechanism is surprise (Ortony and Partridge 1987; Meyer et al. 1997).
Partly because of its important role in the process of
knowledge updating, surprise has long been of interest
to philosophers and psychologists, with first theories of
surprise dating back as far as Aristotle (about 350 B.C.)
(for more details on the history of research on surprise,
see Macedo et al. 2009). Within contemporary experimental psychology, aspects of surprise first came to be
discussed in the 1960s under the headings of “orienting
reaction” and “curiosity and exploration” (Berlyne
1960; cited in Macedo et al. 2009). Surprise as an

Perception/
cognition of
event

Beliefs/
schemas
concerning
event

S

independent phenomenon came under study in the
1970s, when evolutionary emotion theorists, referring
back to Charles Darwin, proposed that surprise is
a basic emotion that serves essential biological functions. One of these functions – surprise as an instigator
of epistemic (specifically causal) search and a precondition for learning and cognitive development – came
to be particularly emphasized by developmental psychologists (see Charlesworth 1969). In the 1970s and
1980s, this suggestion was taken up by social psychologists interested in everyday causal explanations, who
emphasized unexpectedness as a main instigator of
causal search (e.g., Weiner 1985; cited in Macedo
et al. 2009).
Meyer et al. (1997) have tried to integrate the
modal views of previous surprise theorists with attributional analyses of reactions to unexpected events
within the framework of schema theory (Rumelhart
1984; cited in Macedo et al. 2009). The resulting cognitive-evolutionary model of surprise (Fig. 1) assumes that
(ultimately) surprise-eliciting events initiate a fourstep sequence of processes. The first step in this
sequence consists of (1) the appraisal of an event as
schema-discrepant, or unexpected. If the degree of
schema-discrepancy
(unexpectedness)
exceeds
a certain threshold, then (2) ongoing mental processes
are interrupted, attention is shifted to the unexpected
event, and surprise is experienced. This second step
serves to enable and prepare (3) the analysis and evaluation of the unexpected event plus – if this analysis
suggests so – (4) immediate reactions to the unexpected
event and/or an updating, extension, or revision of
the schema or schemas that gave rise to the discrepancy.
Ideally, successful schema change (belief update)
enables the person to predict and, if possible, to control
future occurrences of the schema-discrepant event, to

Feeling of surprise
about event
Appraisal of
event as
unexpected
(schema-discrepant)

Analysis and
evaluation of
event

Schema
update

Interruption of
processing/
reallocation of
processing resources

Surprise and Anticipation in Learning. Fig. 1 The cognitive-evolutionary model of surprise (Meyer et al. 1997)

3251

S

3252

S

Surprise and Anticipation in Learning

avoid the event if it is negative and uncontrollable, or to
ignore the event if it is irrelevant for action.
The surprise mechanism is assumed to consist at its
core of a device that continuously compares, at an
unconscious level of processing, the currently activated
cognitive schemas (which may be regarded as constituting the person’s working-memory model of her
present situation) with newly acquired information
(beliefs). As long as this mechanism registers congruence between schema and input – as long as events
conform to expectations – the person’s informal theories are supported by the evidence, and there is hence
no need to revise them. Rather, the interpretation of
events and the control of action take place largely
automatically and without effort. In contrast, if
a discrepancy between schema and input is detected,
a surprise reaction is elicited: ongoing information
processing is interrupted, processing resources are
reallocated to the unexpected event, surprise is experienced, and cognitive processes (as well as, possibly,
overt actions) aimed at the analysis and evaluation of
the unexpected event are initiated. The function of
these processes is, on the one hand, to enable and
motivate immediate adaptive actions directed at the
surprising event (short-term adaptation); and on the
other hand, to promote the appropriate revision of
the disconfirmed schemas and thereby, future adaptive
actions (long-term adaptation).
Although the model proposed by Meyer et al. concerns surprise in humans, this or a functionally similar
surprise mechanism can be regarded as a crucial component of general intelligence, and hence as an essential
component of any anticipatory agent that, like humans,
is resource-bounded and operates in an imperfectly
known and changing environment. The function of the
surprise mechanism in such an agent is the same as in
humans: to promote the short- and long-term adaptation to unexpected events (Meyer et al. 1997). And as in
humans, this function of surprise entails a close connection of surprise to curiosity and exploration (Berlyne
1960; cited in Macedo et al. 2009), as well as to belief
revision and learning (e.g., Charlesworth 1969).

Important Scientific Research and
Open Questions
Beginning in the 1960s, when surprise first came into
the focus of experimental psychology, research on surprise has steadily increased and is carried out today by

investigators in different subfields of psychology, as
well as in computer science. Topics addressed by recent
psychological research on surprise are, for example: the
relation between surprise intensity and the strength of
cognitive schemas; the role of surprise in spontaneous
attention capture; the effects of surprise on the hindsight bias; the spontaneous facial expression of surprise; and the role of surprise in advertising (see
Macedo et al. 2009).
In recent years, several computational models of
surprise, including concrete computer implementations,
have been developed (see Macedo et al. 2009). The aim
of these computational models of surprise – which are in
part based on psychological theories and findings on the
subject – is on the one hand to simulate surprise in order
to advance the understanding of surprise in humans,
and on the other hand to provide artificial agents
(softbots or robots) with the benefits of a surprise mechanism. This second goal is motivated by the belief that
surprise is as relevant for artificial agents as it is for
humans (Ortony and Partridge 1987). Of particular
interest to these models are the details of the computation of surprise. Although current models already do
a reasonably good job in modeling surprise in humans,
research on this subject is continuing.
This research also has relevance for machine learning. To perform well, current supervised machine
learning systems must be trained on hundreds or even
thousands of labeled examples. For many learning
tasks, these learning instances are difficult, time-consuming, or expensive to obtain. To overcome these
problems, so-called active learning paradigms allow
the learning algorithm to choose the data from which
it learns (Settles 2009). The aim is to select training data
that have high information value, allowing to improve
performance with less training. This raises the question
of how the information value of data should be gauged.
One measure of information value frequently used by
active learning algorithms is the Shannon entropy.
Alternatively, surprise could be used as an index of the
relevance of information for learning (Macedo et al.
2009). In this case, the goal is to select data that maximize the intensity of surprise or more precisely, the
intensity of expected surprise. For example, consider
conducting experiments as learning experiences. If the
possible outcomes of an experiment are expected to be
unsurprising, the principle of maximizing expected
surprise dictates that the experiment should not be

Sutherland, Edwin H. (1883–1950)

conducted. In contrast, if one anticipates surprise from
an experiment, one should conduct it to learn from it.
Recent years have also seen an increased interest in
the topic of anticipation in artificial systems. Pezzulo
et al. (2008) present an overview of this research field
that focuses on both the understanding of anticipation
in natural agents and its implementation in artificial
systems.

S

Sutherland, Edwin H.
(1883–1950)
THOMAS ANTWI BOSIAKOH
Department of Sociology, University of Ghana, Legon,
Accra, Ghana

Cross-References

Life Dates

▶ Active Learning
▶ Adaptation and Anticipation: Learning from
Experience
▶ Adaptability and Learning
▶ Anticipation and Learning
▶ Anticipatory Learning Mechanisms (in Autonomous
Agents)
▶ Belief-Based Learning Models
▶ Curiosity and Exploration
▶ Emotions: Functions and Effects on Learning
▶ Machine Learning
▶ Schema(s)

Edwin H. Sutherland, an American sociologist and
criminologist, was born at Gibbon, Nebraska, in the
USA on 13th August, 1883. He was the third of seven
children born to George Sutherland, a clergy/college
professor/college president and Lizzie (Pickett) Sutherland. Sutherland’s father had been raised in intensely
religious and sober family and he committed to bring
his children up in similar ways. The family lived in
a rural area, and in that setting, they succeeded in
inculcating the sense of moral implications involved
in carrying out one’s duties. Snodgrass (1972) has
observed that, Sutherland’s father was a religious fundamentalist and followed all austere and strict practices
of his Baptist faith in bringing up his children.
In 1893 when Sutherland was 10 years, his family
moved from Gibbon Nebraska to Grand Island,
Nebraska, where young Sutherland attended Grand
Island College. Originally, Sutherland studied history,
and graduated with bachelor’s in 1904 from Grand
Island College where his clergy father, George Sutherland was president (Goff 1982). In 1905, he enrolled in
a correspondence course in introductory sociology
offered by the home study department of the University
of Chicago. At this point in his life, he was planning to
do graduate studies in history at the University of
Chicago, and a course in sociology was a requirement.
It was this university requirement that set Sutherland
on the path of reading sociology, and eventually
becoming a reputable sociologist and helping to
develop one key subfield in sociology – criminology.
In 1906, Sutherland began graduate study at the
department of sociology, University of Chicago where
he changed his major from history to sociology. He was
persuaded by his correspondence course instructor to
take a course with Henderson who eventually became
his graduate studies advisor. Sutherland completed his
Ph.D. in Sociology in 1913 with specialization in political economy (Gaylord and Galliher 1988).

References
Charlesworth, W. R. (1969). The role of surprise in cognitive development. In D. Elkind & J. H. Flavell (Eds.), Studies in cognitive
development (pp. 257–314). Oxford: Oxford University Press.
Macedo, L., Cardoso, A., Reisenzein, R., Lorini, E., & Castelfranchi, C.
(2009). Artificial surprise. In J. Vallverdú & D. Casacuberta
(Eds.), Handbook of research on synthetic emotions and sociable
robotics: New applications in affective computing and artificial
intelligence (pp. 267–291). Hershey: IGI Global.
Meyer, W. U., Reisenzein, R., & Schützwohl, A. (1997). Towards
a process analysis of emotions: The case of surprise. Motivation
and Emotion, 21, 251–274.
Ortony, A., & Partridge, D. (1987). Surprisingness and expectation
failure: What’s the difference?. Proceedings of the 10th international joint conference on artificial intelligence (pp. 106–108).
Milan: Morgan Kaufmann.
Pezzulo, G., Butz, M. V., Castelfranchi, C., & Falcone, R. (Eds.).
(2008). The challenge of anticipation: A Unifying Framework for
the Analysis and Design of Artificial Cognitive Systems. LNAI 5225.
Berlin: Springer.
Settles, B. (2009). Active learning literature survey (Technical report
1648). Madison: University of Wisconsin.

Surprise-Based Learning
▶ Surprise and Anticipation in Learning

3253

S

3254

S

Sutherland, Edwin H. (1883–1950)

Sutherland had 2 years of predoctoral teaching
career and 36 years or so of postdoctoral teaching/
research career. After graduating with bachelor’s degree
from Grand Island College in 1904, he was employed by
Sioux Falls College in South Dakota where he taught
Latin, Greek, history, and shorthand for 2 years.
Between 1906 and 1926, Sutherland passed through
five academic institutions, namely, the University of
Chicago (1906–1913, as a graduate student), William
Jewell College (1913–1919, as a professor), University
of Kansas (1918, as a visiting scholar of sociology),
University of Illinois (1919–26, as an assistant professor
of sociology, and later as an associate professor of
sociology), and Northwestern University (1922, as
a visiting professor of sociology). In 1926, Sutherland
was appointed a full professor of sociology at the University of Minnesota, a position he occupied until 1929.
Between 1929 and 1930, he worked as a researcher for
the New York Bureau of Social Hygiene.
From 1930 to 1935, Sutherland was a professor of
sociology at the University of Chicago, his alma mater.
From the University of Chicago, he moved to the Indiana University to serve as the head of the department of
sociology from 1935 to 1949. In 1939, Sutherland
was elected the 29th president of the American Sociological Society, and in 1940, president of the Sociological Research Association. Sutherland also became the
president of Indiana University’s Institute of Criminal
Law and Criminology, president of the American
Prison Association, and president of the Chicago Academy of Criminology. He also served as visiting professor of sociology at the University of Washington in
1942, and founded the Bloomington Institute of Criminal Law and Criminology.

Contribution(s) to the Field of
Learning
Sutherland’s contribution to the field of learning is in
the area of criminal learning, i.e., how criminals
come to acquire the criminal skills with which they
operate. This however was not part of his training. In
studying for his bachelor’s, Sutherland specialized in
history, and in his doctorate in sociology, specialized
in political economy. In the early period of his teaching/research career therefore, Sutherland focused
on these specialties – history at Sioux Falls College,
and political economy after his doctorate until the
early 1920s.

It was during his tenure at the University of Illinois
between 1919 and1926 that Sutherland’s work in the
area of crime became evident and eventually his contribution to the field of learning science. In 1921, Sutherland received a proposal from the head of sociology
department at the University of Illinois to write a textbook on criminology. At that time, this proposal seemed
odd, but turned out to be a fortuitous endeavor. Sutherland had only taught criminology occasionally and he
had no publications in criminology. However, his book,
Criminology, was duly published in 1924 and retitled
Principles of Criminology in 1934. In 1932, Sutherland
met Broadway Jones, a Chicago-based renowned grifter
who regaled Sutherland with narratives of his remarkable exploits in the criminal underworld. Excited by
these stories, Sutherland and Jones started compiling
these stories with the hope of publishing them one day.
Eventually in 1937, Sutherland and Jones published
these narratives as the Professional Thief (in which
Jones used the pseudonym “Chic Conwell”).
From the stories of Jones, Sutherland became convinced that professional criminals learn the techniques
and attitudes involved in professional criminal work
from close association with other professional criminals. This simple notion compelled Sutherland to start
theorizing in the third edition of his Criminology (1939,
pp. 4–8) that crime is a function of differential association, i.e., the different interactions/associations and
patterns of learning that take place in groups such as
juvenile gangs. In 1947 when the fourth edition of
Sutherland’s Criminology (the Principles of Criminology) appeared, he had outlined clearly differential association theory, his most renowned export into the
criminological world. This theory contained nine principles suggesting that criminal behavior is learned in
a process of communication in intimate personal
groups. These groups, according to the theory, teach
skills, motivations, attitudes, and rationalizations, as
well as definitions either favorable or unfavorable to
the violation of the criminal law.
The principles of differential association theory are
as follows:
1. Criminal behavior is learned; it is not inherited.
2. Criminal behavior is learned in interaction with
other people in the process of communication.
3. The principal part of learning criminal behavior
occurs in intimate groups.

Syllogistic Reasoning

4. When criminal behavior is learned, the learning
includes (a) techniques, sometimes complicated,
and sometimes very simple and (b) the specific
direction of motives, drives, rationalizations, and
attitudes.
5. The specific direction of motives and drives are
learned from definitions of legal codes as favorable
or unfavorable.
6. A person becomes criminal because of excess of
definitions favorable to the violation of law over
definitions unfavorable to the violation of law.
7. Differential association varies in frequency, duration, priority, and intensity.
8. The process of learning criminal behavior involves
the same mechanisms involved in any other learning process.
9. Both criminal and noncriminal behaviors are
expressions of the same needs and values.
The basic premise of the differential association
theory is that criminal behavior results when one is
exposed to an excess of definitions favorable to the
violation of the law over those unfavorable through
interactions with others. With this theory, Sutherland
rejected the notions of criminal behavior being caused
by psychopathological or economic factors, the leading
explanations of crime causation before the 1930s.
Sutherland’s Criminology (Principles of Criminology) was published in four editions – 1924, 1934, 1939,
and 1947 before his death in 1950, and has since been
kept in print, first by Donald Cressey and later by David
Luckenbill. The Eleventh edition of the book appeared
in 1992. Sutherland also has contributed more than 50
articles to journals and books dealing with different
aspects of the learning process for the acquisition of
criminal skills.

Important Scientific Research and
Open Questions
At the time Sutherland died in 1950, in Bloomington,
Indiana, it was said that, it had become a colloquialism
to refer to him as the dean of American criminology. He
is considered one of the most influential criminologists
of the twentieth century and is most remembered for
the conceptual and theoretical contributions he made
to the field of criminal studies, particularly the learning
process in his differential association theory. His legacy
straddles across teaching and research. A number of

S

3255

Sutherland’s graduate students (including Donald
Cressey, Lloyd Ohlin, Albert Cohen, and Karl
Schuessler) also became equally influential personalities in the field of criminology helping to shed more
light on Sutherland’s work. Recognizing the contribution of Sutherland to the field of criminology and to the
learning aspect of criminological science, the American
Society of Criminology in 1960, established an annual
award that carries Sutherland’s name – The Sutherland
Award. This award, the most prestigious recognition
for an American criminologist, is given to scholars
recognized to have made outstanding contributions
to theory or research in criminology, to the etiology
of criminal and deviance behavior, and to the criminal
justice system, corrections, law, or justice. Daniel Glaser, Albert Cohen, Robert Merton, Charles Tittle, and
Lloyd Ohlin are among the leading criminologists who
have won the Sutherland award. He has also been the
subject of several Ph.D. dissertations including
Snodgrass in 1972 and Goff in 1982.

Cross-References
▶ Differential Association Theory
▶ Social Learning Theory
▶ Value Learning

References
Gaylord, M. S., & Galliher, J. F. (1988). The criminology of Edwin
Sutherland. New Brunswick: Rutgers–The State University.
Goff, C. (1982). Edwin H. Sutherland and white-collar crime.
Unpublished Ph.D. dissertation, University of California, Irvine.
Snodgrass, J. (1972). The American criminological tradition: Portraits
of men and ideology in a discipline. Unpublished Ph.D. dissertation, University of Pennsylvania.
Sutherland, E. H. (1937). The professional thief. Chicago: University of
Chicago Press.
Sutherland, E. H. (1939). Principles of criminology (3rd ed.). Philadelphia: J. B. Lippincott.
Sutherland, E. H. (1947). Principles of criminology (4th ed.). Philadelphia: J. B. Lippincott.

Syllogistic Reasoning
Syllogistic reasoning is concerned with using syllogisms
to draw conclusions from premises. A syllogism
(Greek: sullοgismός – “conclusion,” “inference”) or
logical appeal is a kind of logical argument in which

S

3256

S

Symbol Systems

one proposition (the conclusion) is inferred from two
others (the premises) of a certain form. Although there
are infinitely many possible syllogisms, there are only
a finite number of logically distinct types.

Symbol Systems
The form of symbolic expression used to communicate
ideas such as spoken language, printed text, music,
numbers, images, graphs, photographs, and so forth.

Symbols
Arbitrary signs that have acquired a conventional
significance.

Synaesthesia
This is a neurologically based condition in which stimulation of one sensory or cognitive pathway leads to
automatic, involuntary experiences in a second sensory
or cognitive pathway. In one common form, known as
grapheme-color synaesthesia, letters or numbers are
perceived as inherently colored.

Symbolic Clustering
▶ Conceptual Clustering

Synaptic Consolidation
▶ Linking Fear Learning to Memory Consolidation

Symbolic Expression
▶ Learning with Multiple Representations

Synaptic Efficacy Enhancers
▶ Pharmacological Enhancement of Synaptic Efficacy,
Spatial Learning, and Memory

Symbolic Representation
Symbolic representation is the representation of something in association with something else. In other
words, symbolic representation means representing
new categories grounded on the older previously
known categories.

Symbolic Therapy
▶ Metaphor Therapy

Symbolism
▶ Semiotics and Learning

Synchronous Learning
Any learning event where interaction happens simultaneously in real time.

Synthetic Learning
Environment
PATRICK BLUMSCHEIN
Department of School Education, University of
Education, Freiburg, Germany

Synthetic learning environments can be characterized
in terms of a particular technology (e.g., a simulation

Systemic Therapy

or game), subject matter, learner characteristics, and
some guiding pedagogical principles (Cannon-Bowers
and Bowers 2008). In many cases, synthetic learning
environments involve a computer simulation as
a central component and thus serve as simulationbased training.
As Cannon-Bowers and Bowers (2008) point out,
synthetic learning environments can be developed in
countless ways and many variables must be taken into
account. Therefore, engineering a synthetic learning
environment places great demands on the instructional
design. Cannon-Bowers and Bowers focus on several
instructional features and strategies, such as authenticity and fidelity of learning experiences, model-based
reasoning, the design of cases or scenarios that provide
the context for instruction, and collaborative and social
learning as well as various motivational factors (e.g.,
goal setting, engagement, reinforcement).

References
Cannon-Bowers, J. A., & Bowers, C. A. (2008). Synthetic learning
environments. In J. M. Spector, M. D. Merrill, J. van
Merrienboer, & M. P. Driscoll (Eds.), Handbook of research
on educational communications and technology (3rd ed.,
pp. 317–327). Mahwah: Lawrence Erlbaum.

S

3257

System Dynamics
System dynamics is an approach to modeling the
dynamics of systems which interact strongly with each
other. System dynamics was founded in the early 1960s
by Jay W. Forrester of the MIT Sloan School of Management. System dynamics refers to studying complex systems by applying feedback loops. Stocks and flows are
the basic building blocks of system dynamic models.
They help describe how a system is connected by feedback loops which create the nonlinearity found so frequently in modern day problems. Computer software is
used to simulate a system dynamics model of the situation being studied. Running “what if” simulations to
test certain policies on such a model can greatly aid in
understanding how the system changes over time.

System Simulation Model
▶ Computer Simulation Model

Systematic Repetition
System
In the strongest case, a system refers to an information
processor that can be differentiated from other information processors on the basis of psychological process, the computational algorithms, and the underlying
neural substrates.
The constituents of a system are a set of elements
which are interconnected to aid in driving toward
a desired goal. This definition has two implications:
First, a system consists of elements (parts) that are
interconnected through relationships or functions.
The number and quality of these relationships determine that a system can range from simple to complex.
The central goal of creating a “valid” model of a system
is to capture the most important elements and especially the nature of the relationships. Second, the
goal orientation of a system implies its continuous
improvement. Feedback from the system must be
used to measure the performance of the system against
its desired goal.

▶ Drill and Practice in Learning (and Beyond)

Systemic Functional Linguistics
Also known as SFL, is a branch of linguistic focusing on
language in use and looking on portions of discourse
larger than the sentence. Founded by Michael Halliday
in the 1960s, SFL is grounded in the view of language as
“a network of systems, or interrelated sets of options
for making meaning.” The term “functional” indicates
that the approach is concerned with the contextualized,
practical uses to which language is put.

Systemic Therapy
▶ Application of Family Therapy on Complex Social
Issues

S

3258

S

Systemic-Structural Theory of Activity (SSAT)

Systemic-Structural Theory of
Activity (SSAT)
▶ Learning and Training: Activity Approach

Systems Psychology
▶ Application of Family Therapy on Complex Social
Issues

Systems Theory
Systems Consolidation
▶ Linking Fear Learning to Memory Consolidation

Systems Dynamics
▶ Dynamic Modeling and Analogies

▶ Application of Family Therapy on Complex Social
Issues

Systems Thinking
▶ Metapatterns for Research into Complex Systems of
Learning