Computer Based Learning
Computer Based Learning
The other method used for measuring calibration is People may directly access their memories to evaluate
the calibration curve or graph (Keren 1991). Actual the status of their knowledge, they may make evalua-
performance is plotted on the y-axis, and judged perfor- tions based on inferences or heuristics about how much
mance is plotted on the x-axis. Perfect calibration is they believe they know about a general domain, they
represented by the 45 line, that is, judgments are exactly may make evaluations based on how self-efficacious
equal to performance. Points that fall below the 45 line they feel about their performance on a particular task,
are interpreted as overconfidence, and points that fall or all of these contributors may come into play. The
above the line are interpreted as underconfidence. accuracy of the calibration judgment will be deter-
Calibration curves are easily interpreted and provide mined by how well all those contributors to the judg-
a graphical means of representing the degree of corre- ment are able to predict performance on a criterion
spondence between perceptions of performance and task. In other words, calibration accuracy depends on
actual performance. the extent to which memories are accessed, the infer-
In the metacognitive literature, an important ences or heuristics are made, or the self-efficaciousness
distinction is made between calibration and discrim- felt conform to the knowledge that is tested on the
ination or resolution (Nelson 1996). Calibration is a criterion task.
measure of absolute accuracy, and discrimination is Accurate calibration is an essential component of
a measure of relative accuracy. Although both con- effective self-regulated learning. In an era of high stakes
structs involve metacognitive monitoring, they repre- tests and accountability, the ability to perform well on
sent different aspects of metacognitive monitoring and tests is essential. Students studying for a test need to be
are measured in different ways. Whereas, calibration accurate in their monitoring of their knowledge reten-
provides estimates of the degree to which a person’s tion if they hope to successfully control further study.
perception of performance corresponds with his or her Students who are overconfident (i.e., a positive bias in
actual performance, relative accuracy provides esti- calibration judgments) may have a false sense of how
mates of the degree to which a person’s judgments well they have mastered the material. They may believe
can predict the likelihood of correct performance of they are prepared when in fact they are at risk for
one item relative to another (Nelson 1996). Calibration failure. Or students could intentionally inflate their
provides estimates of overall memory retrieval, and overconfidence during test preparation as a self-
relative accuracy provides estimates of whether a per- handicapping strategy, excusing or attributing their
son can discriminate between what is known or not poor performance to external causes (Winne 2004).
known. Studies that have compared absolute and rela- Underconfidence (i.e., a negative bias in calibration
tive accuracy have found only small correlations between judgments) also can be detrimental to academic per-
the two, suggesting that the two types of accuracy are formance because students may fail to disengage from
tapping different aspects of metacognitive monitoring studying and misallocate study time if they assume the
(e.g., Hacker et al. 2011). material is not yet mastered. When students demon-
strate strong biases in their calibration judgments, they
Theoretical Background may not take the remedial steps necessary to improve
Calibration is a metacognitive monitoring process. or carefully evaluate their responses during or after an
Monitoring provides information at the metacogntive exam (Hacker et al. 2008). “Learning will be inversely
level about the status of one’s knowledge or strategies at proportional to the degree of calibration bias and
a cognitive level (Nelson 1996). Based on this informa- proportional to calibration accuracy” (Winne 2004,
tion, metacognitive control can be exerted to regulate p. 476).
one’s knowledge or strategies. More specifically, after
a person acquires and hopefully retains a specific chunk Important Scientific Research and
of knowledge, he or she may evaluate the status of that Open Questions
knowledge in memory, that is, to what degree does the There are some consistent findings in the literature
person believe the knowledge has been retained. There related to calibration accuracy. Many studies have indi-
may be many contributing variables to that evaluation. cated that calibration accuracy is linked to achievement
Calibration C 497
level (e.g., Bol et al. 2005; Hacker et al. 2000). More insufficient for improving calibration accuracy. Reflec-
specifically, higher-achieving students tend to be tion and instruction on monitoring and calibration
more accurate but underconfident when compared to were found to be effective, particularly for higher-
their lower-achieving counterparts who are less accu- achieving students. External rewards or incentives
rate and overconfident. Calibration inaccuracy and were found to enhance postdiction accuracy among C
overconfidence among the lower-achieving students lower-achieving students. More recently, group calibra-
has been linked to theories of self-serving bias, attri- tion practice and the provision of guidelines have been
bution theory, self-handicapping strategies, and ego shown to improve calibration accuracy and achieve-
defense (Bol et al. 2005; Hacker et al. 2008). Lower- ment among high school students (Bol et al. 2009).
achieving students seem to anchor their calibration Attempts to further identify consistent patterns
judgments on optimistic yet inaccurate beliefs about of findings across calibration studies are compromised
their own abilities rather than prior performance in an by the lack of common definitions and standard mea-
effort to protect their sense of self-worth. sures. Some researchers refer to calibration as confi-
Another consistent finding is that predictions are dence or self-efficacy; whereas, others refer simply to
almost always less accurate than postdictions. This self-monitoring and not necessarily calibration. Varia-
phenomenon is known as the testing effect or the tions in how calibration has been measured exacerbate
upgrading of prediction accuracy (Pressley and Ghatala the problem (e.g., local or global judgments, confi-
1990). Upgrading makes intuitive sense because a dence ratings, absolute differences). A common termi-
person should be better able to judge how he or nology and standard measurement procedures would
she performed on a task at the completion of the advance this line of inquiry.
task due to familiarity and exposure. Consider the Several other important open questions remain.
context of test-taking. Once students have completed The first centers on ecological validity or the ability to
a test, their predictive judgments of performance generalize findings from laboratory-based studies to
turn from expectations of what may happen to studies conducted in more naturalistic settings such
postdictive judgments of what actually happened. as classrooms. A closely related issue is population
The test itself and students’ performance on it pro- validity or the ability to generalize results beyond col-
vide feedback that informs their postdictions (Hacker lege-age students. The vast majority of calibration
et al. 2000). studies have been conducted with college students, in
However, task difficulty also influences calibration laboratory settings, and employ inauthentic tasks. Fur-
accuracy. In fact, the upgrading of prediction accuracy ther research on effective interventions is warranted.
has been reduced when more complex tasks are required. Studies investigating the effectiveness of interventions
Juslin et al. (2000) have worked with the hard-easy for improving calibration accuracy and confirming
effect in which students tend to be more accurate but its link to achievement are needed. Past research has
underconfident on easy items and less accurate but suggested that the effectiveness of interventions may
overconfident on difficult items. This phenomenon is vary depending on prior achievement, implying that
related to achievement level. Lower-achieving students interventions might be tailored to better meet the
tend to be less accurate and overconfident than their needs of students at risk for failure. Initiating studies
higher-achieving peers on the more difficult items. on the psychological bases of judgments also will be
There is less variability in accuracy on easy items as a productive addition to the literature. What are the
a function of achievement level. significant contributing variables to calibration: Can
Attempts to improve calibration accuracy, or to people directly access memory and accurately judge
debias calibration judgments, in classroom settings the status of memories; are inferences based on
have been met with mixed success (Hacker et al. 2008). domain knowledge or self-efficacy responsible; or do
Repeated calibration practice, across trials, does not people rely on anchoring heuristics? Finally, researchers
seem to enhance accuracy, particularly among lower- might further explore how calibration judgments are
achieving students. Calibration tends to be stable, influenced by social variables. Two avenues for future
suggesting that feedback and practice alone are study include attributional retraining to promote more
498 C Calibration of Comprehension
Chunks are the basic unit of short-term memory, as articulatory suppression). This is because, while
composed not of the smallest atom of information, but both tasks require the temporary maintenance of
instead of an interrelated collection of items (such short-term information, the spatial task relies on the
as a word). Because the interrelations between these visuospatial sketchpad and articulatory suppression
items are stored in long-term memory (e.g., knowledge recruits the phonological loop. On the other hand, C
about words), they do not take up further space in verbal tasks such as reading are impaired by articula-
short-term memory. In fact, because larger chunks tory suppression because both processes rely on the
take up an identical amount of space in short-term phonological loop.
memory, chunks act to extend short-term information Perhaps, the most important contribution of
capacity. Thus, Miller found that while short-term Baddeley and Hitch’s working memory model was
memory is limited to 7 2 chunks, depending on the the explicit pairing of storage with processing via the
content they represent, the chunks themselves can store central executive. Although details of the central exec-
a huge amount of information. utive are left relatively unspecified, its role is to allocate
Though Miller’s highly influential work pioneered attention (i.e., determine what content is placed in the
modern research on the structure of memory, simple temporary stores) and mediate the active processing of
short-term capacity lacked sufficient detail to explain content stored in the slave systems. These characteris-
more complex tasks, such as learning and problem tics of the central executive account for basic expecta-
solving. In an effort to overcome this limitation, tions about memory performance – for example, that
Baddeley and Hitch (1974) proposed working memory, short-term stores are not automatically overwritten by
a model of short-term capacity that focused, in new stimuli in the environment and that complex tasks
particular, on how memory is applied toward complex share a common processing resource, regardless of the
processing goals as opposed to recalling simple lists. In modality of their content.
the original framework, short-term processing is car- Research on the capacity limitations of memory
ried out by a system consisting of three components. and learning has advanced with particular emphasis
The first, active component of this system is the on the active processing perspective. One criticism
central executive. The central executive has no memory of early short-term memory approaches was that,
capacity itself, but instead manipulates content stored while individuals differ in their short-term capacity as
in the other two components of the system – the so- measured by forward span, these differences are only
called slave systems. The slave systems, the phonological weakly related to performance on more complex
loop and visuospatial sketchpad, are limited-capacity processing tasks. However, measures of capacity that
passive stores that hold content specific to a partic- involve both storage and processing, such as Daneman
ular modality. The phonological loop holds acoustic and Carpenter’s (1980) reading span exhibit highly
information, while the visuospatial sketchpad holds robust correlations with performance on complex tasks
visual information. A fourth component has also such as reading comprehension and vocabulary learning.
been added – a multimodal episodic buffer that serves Although initially thought to reflect domain-specific lan-
to bind information from the phonological loop, the guage capacities, reading span also correlates with mea-
visuospatial sketchpad, and long-term memory into sures of “executive control,” such as the ability to filter
a unitary episodic representation. out irrelevant content, maintain task goals, and inhibit
The Baddeley and Hitch model of working memory prepotent responses. This has led to the claim that read-
is valuable because it yields specific predictions about ing span and other complex storage and processing tasks
the nature of online processing. In particular, because tap into a domain-general working memory capacity
the phonological loop and visuospatial sketchpad are (Turner and Engle 1989). Differences in working mem-
dissociable stores, learning and performance scenarios ory capacity relate to performance in a number of com-
that draw on one store are unaffected by demands plex processing and learning activities such as encoding
placed on the second. For example, a spatial task such of new information, memory retrieval, reasoning, rule-
as mental rotation is relatively unaffected by reciting based and logic learning, mathematical performance,
“the” over and over again (a secondary task known following directions, and language comprehension.
500 C Capacity Limitations of Memory and Learning
Important Scientific Research and Thus, while working memory capacity is critical for
Open Questions many learning situations, under certain conditions,
Because working memory capacity has been identified increased capacity may not always be a good thing.
as a highly robust predictor of complex behavior, One final avenue of research has provided evidence
a great deal of capacity research is concerned with that the capacity of working memory may even vary
studying this construct – asking questions about its depending on context. Specifically, scenarios that are
fundamental structure, its specific role in complex highly stressful have been found to disrupt the normal
behavior, and about the conditions that may affect its operation of working memory and thus interfere with
normal operation. One such question is the extent to normal learning and performance. This has been
which working memory capacity represents a stable shown in a variety of stressful situations. For example,
property of the individual or is malleable based on when a math-anxious individual is placed in a math-
experience. While there is clear evidence that working related situation, their ability to allocate working mem-
memory capacity varies across development, how ory toward task-related processes is interfered with by
much of this variation is a function of intellectual their anxiety about the task. This results in worse per-
experience or is predetermined by neural development formance, particularly on those problems and situa-
is not well established. tions that place the most demands on working memory
One tool for resolving this debate resides in train- (Ashcraft and Kirk 2001). This negative relationship
ing studies that expose participants to a regimen of between anxiety and working memory capacity has
demanding working memory tasks. While early results been replicated in situations where high-stakes incen-
in this literature were criticized for their lack of appro- tives (like a standardized test) or negative stereotypes
priate controls, recent work has shown evidence of (e.g., for women, the stereotype that women are bad
effective capacity training under more rigorous condi- at math) lead to performance anxiety (see, Beilock
tions. However, the debate regarding whether or not 2008 for a review). These scenarios carry important
working memory capacity can be enhanced via train- real-world implications for the relationship between
ing is by no means settled. For instance, evidence situational factors and online capacity limitations in
from twin studies suggests that the development and learning and memory.
capacity of working memory does have a genetic com-
ponent. While this finding does not preclude the effi- Cross-References
cacy of training interventions, it does suggest that, ▶ Abilities to Learn: Cognitive Abilities
in the normal population, biological predispositions ▶ Cognitive Load Theory
also play a role in determining capacity limitations of ▶ Individual Differences
the individual. ▶ Intelligence, Learning and Neural Plasticity
The positive association between working memory ▶ Short-Term Memory
capacity and academic performance has become a
pervasive finding in the psychology and education lit- References
eratures. Yet, the full maturation of this resource is Ashcraft, M. H., & Kirk, E. P. (2001). The relationships among
completed only after an individual reaches adulthood, working memory, math anxiety, and performance. Journal of
Experimental Psychology: General, 130(2), 224–237.
long after many critical learning milestones have
Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. Bower
been surpassed. This has lead to the proposal that
(Ed.), The psychology of learning and motivation: Advances in
working memory capacity might actually impede the research and theory (pp. 47–90). New York: Academic.
acquisition of some linguistic and creativity tasks Beilock, S. L. (2008). Math performance in stressful situations. Cur-
(Thompson-Schill et al. 2009). Higher levels of work- rent Directions in Psychological Science, 17, 339–343.
ing memory related to age or natural variation in adults Daneman, M., & Carpenter, P. A. (1980). Individual differences in
working memory and reading. Journal of Verbal Learning and
can actually impede the learning of certain skills
Verbal Behavior, 19, 450–466.
that are best acquired without the guidance of explicit Miller, G. A. (1956). The magical number seven, plus or minus two:
rule-based reasoning processes – processes thought Some limits on our capacity for processing information. Psycho-
to be at the heart of working memory capacity. logical Review, 63(2), 81–97.
Carroll’s Model of School Learning C 501
Thompson-Schill, S. L., Ramscar, M., & Chrysikou, M. (2009). time actually spent for learning. Both variables, in turn,
Cognition without control: When a little frontal lobe goes are dependent on other internal and external variables,
a long way. Current Directions in Psychological Science, 8(5),
such as the learner’s general intelligence and the quality
259–263.
Turner, M. L., & Engle, R. W. (1989). Is working memory capacity of instruction.
task dependent? Journal of Memory and Language, 28, 127–154. C
Theoretical Background
In the 1960s, Carroll developed a conceptual model
of school learning in which the factor time plays
a central role (Carroll 1963). In this model, the achieve-
Care Ethics ment of a student or the degree of learning effectiveness
A theory of prosocial development based on work by is defined as a function of the actual time needed for
Noddings (1984) and Gilligan (1984) that focused on learning and the time actually spent for learning. The
establishing conditions in a particular setting likely to effect of both variables on the degree of learning effec-
encourage goodness. tiveness has been expressed in a functional equation:
Time actually spent for learning
References Degree of learning ¼ f
Actually spent time for learning
Gilligan, C. (1982). In a different voice. Cambridge, MA: Harvard
University Press. Both time variables refer only to active learning and
Noddings, N. (1984). Caring: A feminine approach to ethics and moral are dependent on other variables, such as understand-
education. Berkeley: University of California Press. ing of the task requirements and the student’s aptitude
for a particular task. The understanding of task require-
ments is considered a function of quality of instruction
and the student’s ability to understand instruction that
in turn depends on the student’s general intelligence
Career Interests and verbal aptitudes. The individual task-specific apti-
▶ Stability and Change in Interest Development tude on its part is considered a function of basic apti-
tudes and the time saved by prior learning. All together,
Carroll’s model of school learning can be depicted as in
Fig. 1.
A central educational perspective in Carroll’s model
Carroll’s Model of School is the concept of quality of instruction. He distinguishes
Learning between the substance or content of a learning task and
its communication which plays the most important
NORBERT M. SEEL role for instructional quality. More specifically, quality
Department of Education, University of Freiburg, of instruction contains a clear description of the learn-
Freiburg, Germany ing task, its adequate presentation to the students, as
well as an appropriate sequencing of learning tasks and
a sufficient observance of the students’ characteristics.
Synonyms This distinction allows differentiating between apti-
Conceptual model of school learning tudes and capabilities that are immediately relevant
for learning and abilities which are necessary for under-
Definition standing instruction. In the case that the quality of
Carroll’s model of school learning specifies the distinc- instruction is high there is no high demand for under-
tive roles of generalized abilities and task-specific apti- standing the instruction. The time actually needed for
tudes in determining the effects of instruction on learning can be referred to task-specific aptitudes. In
learning. The degree of learning effectiveness is defined contrast only students with very good task-specific
as a function of the time needed for learning and the aptitudes can understand low-level quality instruction
502 C Carroll’s Model of School Learning
General aptitude
Task-specific
Time saved by aptitude
prior learning
General
intelligence Ability to Understanding
Actually
understand of task
needed time
instruction requirements Degree of
Verbal aptitudes
learning
effective-
ness
Quality of instruction Actually
Perseverance
spent time
Clear description of
task
whereas students with a lack of task-specific aptitudes allowed for learning, perseverance, and required time
need more time for learning. for learning. The time allowed for learning can be
In consequence, Carroll distinguishes between two smaller or bigger than the required time for learning.
kinds of cognitive abilities: The first one refers to com- However, the time actually spent for learning is also
munication and instruction and is general because it constrained by (a) the time a learner is motivated to
applies onto a multitude of learning tasks; the other spend for accomplishing a task, (b) the perseverance,
kind of cognitive abilities are specific with regard to (c) the time needed for accomplishing the task, and
a particular learning task. In addition to the afore- (d) the learner’s aptitudes. Consequently, a learners
mentioned factors, another learner-specific factor stops learning when the time allowed for learning
plays a significant role within Carroll’s model, namely is too short or the motivation for learning is not
perseverance. sufficient.
In addition to the learner-specific factors, the actu-
ally spent time for learning is constrained by the time Important Scientific Research and
allowed for task learning, i.e., the opportunity to learn. Open Questions
This is dependent on the teacher or the curriculum Carroll’s model of school learning was the fundamental
but also on grouping or individualization in order to basis for a number of follow-up attempts to identify
homogenize learning speed and the contents to be and structure the primary variables of effective school
learned. Instructional decisions determine also the learning. Nevertheless, it has been criticized due to
sequencing of learning tasks influencing the student’s some shortcomings. For example, Harnischfeger and
opportunities to learn. The degree of the teacher’s Wiley (1978) criticizes the model as individualistic
adjustment to particular needs and characteristics of inasmuch as it refers only to one learner and one
the learners by means if adequate sequences of learning learning task. It neither incorporates the classroom as
steps are provided is a central part of the quality of sum of individuals nor the sequencing of different
instruction. interrelated learning tasks nor the curriculum as an
In sum: The time actually spent for learning corre- entity. Rather, the quintessence of the model consists
sponds with the smallest of the three factors: time in the fact that the factors aptitude, opportunity for
Case-Based Learning C 503
learning, and perseverance are expressed in terms of Harnischfeger, A., & Wiley, D. E. (1978). Conceptual issues in models
measured time. of school learning. Journal of Curriculum Studies, 10(3), 215–231.
Slavin, R. (2006). Educational psychology: Theory and practice (8th ed.,
A major aspect of Carroll’s argumentation is the
pp. 277–279). Needham Heights: Allyn and Bacon.
precept that the teacher should focus on controllable Squires, D., Huitt, W., & Segars, J. (1983). Effective schools and
variables constituting the quality of instruction. classrooms: A research-based perspective. Alexandria: Association C
Accordingly, some follow-up models of school learning for Supervision and Curriculum Development.
addressed additional classroom and school-level vari-
ables. So, for example, Squires et al. (1983) relabeled
“perseverance” with “involvement” and they added
“coverage” (defined as overlap of the content taught
and content tested) and “success” (defined as degree of
Case-Based e-Learning
achievement in performing academic tasks). ▶ Case-Based Learning on the Web
Another alteration of Carroll’s model is the QAIT
model of Slavin (2006), in which Q denotes the quality
of instruction in Carroll’s sense. A refers to the appro-
priate levels of instruction by redefining student’s
aptitude and ability to understand instruction in Case-Based Learning
terms of teaching behavior variables. I is the abbrevia-
tion of “incentive” and replaces perseverance through a CLAUS ANDREAS FOSS ROSENSTAND
teacher behavior variable. Finally, T stands for time and Department of Communication and Psychology,
corresponds with Carroll’s original variable of oppor- Aalborg University, Aalborg, Denmark
tunity to learn. It is not the place here to justify these
different approaches in more detail. It is sufficient to
show that Carroll’s model was influential enough to Synonyms
evoke follow-up and alternative conceptual models for CBL
school learning.
Most important is the observation that the variables Definition
“time spent” and “perseverance” of Carroll’s model Case-based learning (CBL) is a pedagogical concept,
have been replaced through the concept academic where work method, problem, and discipline are identi-
learning time conceived in general as a combination fied by the learner (or learners) through the learning
of content overlap, involvement, and success. Academic process. Case-based learning is oriented toward a case,
learning time revealed as an appropriate time variable which from different perspectives generates different
for research (Berliner 1978) due to the fact that it is and equally correct problems. Case-based learning is
directly influenced by classroom variables and is the about choosing, deciding priorities, and combining
result of many decisions about how much time is spent different disciplines, and as such is best practiced in
for learning in the classroom. a multidisciplinary context.
solution-based learning, and definition-based learning, Case-Based Learning. Table 1 Cross tabulation of four
was not made until 2008 (Kjærulff et al. 2008). This pedagogical concepts with three pedagogical
time, case-based learning was perceived as an answer to components
the challenge of developing a multidisciplinary study
Pedagogical
across three faculties at Aalborg University, Denmark, component Work
where problem-based learning had been practiced in all Pedagogical concept method Problem Discipline
study programs since the early 1970s.
Definition-Based + + +
Case-based learning takes the preconditions of Learning
problem-based learning a step further. In problem-
Solution-Based – + +
based learning, both work method and problem are
Learning
identified by the learner through the learning process.
In study programs where conventional problem-based Problem-Based – – +
Learning
learning is practiced, the blind spot is the inherent
perspective in the discipline and how they may influ- Case-Based Learning – – –
ence the studies. The pedagogical philosophy of prob-
lem-based learning asserts that a study project will
never be better than the problem investigated. How- In definition-based learning, the learner is provided
ever, it is implicit – though not stated – that this with work method, problem, and discipline. It is a routine
investigation must be within the discipline and related learning situation, where the learning process has a
paradigm of the study. permanent form. Definition-based learning is good
In modern society, common challenges have to for providing the learner with qualifications in situa-
be met with a multidisciplinary approach (Qvortup tions where the ability to produce solutions is essential.
2003). Different perspectives on the same phenomena, In solution-based learning, the learner is provided
or case, result in different problems. Or more accu- with both problem and discipline. However, the learner
rately: The truth is inherent in the perspective, and all has to identify the work method through the learning
perspectives might be equally correct (Rosenstand process. It is a problem-solving learning situation, where
2008). There is not one privileged and valuable truth. the learning process has a solid but not permanent
In practice – both in science, industry, and life in form. When the learner has identified a work method,
general – all cases benefit from being approached the learning situation shifts to a routine learning situ-
from multiple perspectives. In a society where knowl- ation, where the work method is identified and provided
edge is essential and highly valuable, more knowledge is by the learner. The learner might discover that the work
produced by addressing a case from different disci- method is not too wise and turn back to a problem-
plines. Thus, it is essential that pedagogical concepts solving situation conducting solution-based learning
which address multidisciplinarity are introduced and once more. In this way, definition-based learning is
used as part of modern study programs – Case-Based included in solution-based learning, at least as a pre-
Learning is such a concept. condition for producing a solution. Solution-based
In order to address the blind spot of problem-based learning is good for providing the learner with compe-
learning, discipline is added to work method and prob- tences in situations where the ability to choose solution
lem as a pedagogical component that the learner has to methods is essential.
identify through the learning process. A simple cross In problem-based learning only the discipline is
tabulation of pedagogical concepts with pedagogical provided to the learner. As the paradigm behind the
components results in Table 1, including solution- study often is inherent in the study culture, it is
and definition-based learning. rarely questioned as the correct perspective. The learner
A plus (“+”) in Table 1 marks that the learner is has to identify work method and problem through
provided with a pedagogical component as part of the the learning process. It is a problem-oriented learning
learning process, and a minus (“–”) marks that the situation, where the learning process has a loose
learner has to identify the pedagogical component but not unpredictable form. This is often termed
through the learning process. a problem-oriented pedagogy, where the learner has
Case-Based Learning C 505
to orientate himself toward a problem. When the methods that are either too simple or too complex. In
learner has identified a problem, the situation shifts this case, the learner, and the group of which he is
to a solution-based learning situation, where the prob- usually a part, will turn back to an innovative learning
lem is identified and provided by the learner. The situation, where the discipline has to be renegotiated in
learner might find out that the problem is not suffi- order to include new – and perhaps exclude old, per- C
ciently fertile, because the identification of the work spectives. This requires an open and flexible study
method is either too simple, or too complex, in relation culture. As exemplified, case-based learning includes
to the problem – or the identified work method in problem-based learning, which again includes solu-
relation to the problem results in a trivial outcome tion- and definition-based learning. Below, brackets
when shifting to the routine learning situation. In this are used to illustrate how a pedagogical concept
case, the learner can turn back to a problem-oriented includes another pedagogical concept:
situation conducting problem-based learning. In this
● Case-based learning (Problem-based learning
way, problem-based learning includes solution- and
(Solution-based learning (Definition-based
definition-based learning. Problem-based learning is
learning)))
good for providing the learner with creativity in situa-
tions where the ability to produce new solution methods Case-based learning is good for providing the
is essential. learner with culture, where the ability to set up a new
In case-based learning, the learner has to identify framework – a new perspective – is essential. Case-based
work method, problem, and discipline – none of the learning adds an extra dimension to the education of
pedagogical components are provided to the learner. the learner because in order to participate in the
It is an innovative learning situation where the learning required open learning culture, he has to accept that
process has an unpredictable form; in this sense, case- other perspectives than his own can be equally correct,
based learning is a medium in which all the pedagog- even if the different perspectives seem incompatible.
ical components can take different forms. A discipline Table 2 sums up the characteristics of the four
has to be identified through a process in which differ- pedagogical concepts.
ent disciplines and their inherent perspectives are The relationship between the knowledge forms
chosen, prioritized, and combined in an interdisciplin- qualification, competences, creativity, and culture is
ary process, where the different perspectives benefit built on ▶ Greagory Bateson’s “four levels of learning”
from a negotiation in a multidisciplinary context. (Qvortup 2003).
Only geniuses can do this alone – and not always
with all the relevant perspectives. Actually, we do not Important Scientific Research and
know which discipline will prove fertile until we know Open Questions
and understand the very essence of a case, and this Only very few study programs have practiced case-
includes the shift through a problem-oriented, prob- based learning as defined in this entry. As such, it is
lem-solving, and routine learning situation, where a new theory of learning that has yet to be discussed in
problem-, solution-, and definition-based learning is the literature. However, there is comprehensive scien-
conducted, respectively. The identified discipline might tific research in the closely related field of ▶ problem-
turn out to generate or require problems and/or work based learning.
Different organizational study forms have yet to situation – Multidisciplinarity as pedagogical precondition for
be developed and experimented with. How is motiva- innovation]. In J. Stolt & C. Vintergaard (Eds.), Tværfaglighed
& Entrepreneurship [Multidisciplinarity & Entrepreneurship].
tion and talent combined in an interdisciplinary
Copenhagen: IDEA København og Øresund Entrepreneurship
learning culture, where students have to take an inter- Academy.
est in areas other than the one that has their initial Qvortup, L. (2003). The hyper complex society. New York: Peter Lang.
interest?
Where and when is case-based learning a wise
approach? It has been tried out at university level in
different courses, semesters even, with some success; Case-Based Learning on the
but it has not been tried out as the overall pedagogical Web
concept for an entire bachelor or candidate program.
Should such an experiment be conducted, it would LOWELL DEAN TONG, CHRISTIAN BURKE, ANN N. PONCELET
be necessary to include several study programs simul- School of Medicine, University of California San
taneously in order to ensure the multidisciplinary Francisco, San Francisco, CA, USA
preconditions.
Other open questions are: If case-based learning is
conducted, how much weight should this approach Synonyms
carry compared to the other pedagogical approaches Case-based e-learning
in higher education? Will it apply in certain cases
only? When? Could public schools and colleges bene- Definition
fit from case-based learning? How? And how does Case-based learning on the web (CBLW) occupies the
case-based learning apply to an industrial context intersection of case-based learning and online learning.
as a pedagogical approach that supports innovative There is scant research on CBLW specifically; what
processes? exists is mainly in the field of health sciences education,
and medical education in particular. CBLW prepares
Cross-References the learner for authentic and situational performance,
▶ Bateson, Gregory (1904–1980): Anthropology of rather than presenting canonical technical learning
Learning through a straightforward instructional demonstration
▶ Complex Problem Solving video on the web. CBLW is also distinct from case
▶ Cooperative Learning studies of online learning as a pedagogical method.
▶ Creativity, Problem Solving, and Feeling In the medical education literature, CBLW is typi-
▶ Cross-Disciplinary Learning cally implemented through a free-standing educational
▶ Culture of Learning module, or a set of modules, which the learner accesses
▶ Guided Discovery Learning and engages with, via the web through a computer or
▶ Machine Learning of Natural Language similar device, and which is based on a specific and
▶ Problem Solving highly realistic scenario. It typically uses audio and
▶ Problem-Based Learning video components, in addition to text, illustrations,
and other media, and can include assessment of the
learner as well as assessment of the web-based curricu-
References
lum itself. The student may be allowed to experience
Kjærulff, U. B., Rosenstand, C. A. F., Stage, J., & Vetner, M. (2008).
Case-based learning (CBL) – A new pedagogical approach to the module in a free-flowing path, choosing to navigate
multidisciplinary studies. In F. Fink (Ed.), 36th SEFI Annual in any direction, or forced to be linear, with a “one-
Conference 2008 – Quality Assessment, Employability and Innova- way” direction, or even through the use of a selective
tion. Denmark: Sense Publisher. (CD media and Google release mechanism, whereby a student must verify
Scholar).
completion of one section, with perhaps a quiz, before
Kolodner, J. L. (Ed.). (1993). Case-based learning. Boston: Kluwer
Academic Publishers. being allowed to move to the next section. It usually
Rosenstand, C. (2008). Innovation som situation – Flerfaglighed som is designed to push the learner to engage with the
pædagogisk forudsætning for innovation [Innovation as material, making reasoned decisions along the way,
Case-Based Learning on the Web C 507
applying previously learned general principles to the design and collaboration system. It is now used wher-
case at hand, and creating new general hypotheses ever people collaborate to produce media-rich projects
from the case. especially in video cases, interactive media, and
E-learning. It is a design method used in the creative
Theoretical Background arts to shape the content from the learner’s perspective. C
CBLW combines the features of simulation-based It is similar to the architect’s small-scale model of
learning (case-based), with ready access (online learn- a building but adds the dimension of time or flow.
ing). Furthermore, CBLW can standardize learning in A storyboard is both a method for collaborators to
content and quality, facilitate active learning by the communicate about the design and experience its
student, avoid the cost of staging multiple or repeated function, as well as the platform which learners will
simulations, and can allow asynchronous learning for access, navigate, and use. Conceptual storyboards illus-
learners dispersed amongst temporally or geographi- trate on paper what is imagined as the actual computer
cally distributed learning environments, such as med- screen images, combined with exactly how the learner
ical students who are assigned to clinical experiences at interacts with the material: starting, pausing, navigat-
different times and clinical locations. Because it can ing. Each frame or page of the conceptual storyboard
capture and record learner responses, CBLW can assess is visual in nature. Production storyboards document
the learner’s level of competency around elements of the flow of the module and can be either in the form
the case. CBLW need not exclude real-time exchange of pictograms or text, analogous to a program book,
with instructors or fellow learners; there can be desig- table of contents, or sitemap. This leads to effective
nated times for synchronous participation and learn- CBLW, where the student actually experiences the prob-
ing, such as required completion of a CBLW activity lems illustrated by the case, rather than learns about the
prior to a scheduled online discussion. Asynchronous problems of the case (Fig. 1).
or elective use of a CBLW allows for on-demand learn- CBLW is more dynamic than a textbook and allows
ing, either just-in-time learning, or as needed out of for ongoing content updates, providing students with
curiosity, for a repeated experience, or reference. up to date content. This format allows the inclusion of
CBLW may become more essential to formal learn- images, video, audio, and animation. Multimedia helps
ing as virtual learning environments become more target different learning styles and can be tailored to
commonplace in conventional schools and “evening content. Expanded references are instantly available
schools,” and with the rise of new, accredited schools using web links. The web is flexible and becoming
designed to be completely virtual. This modality easier to use with more tools such as Wikis, blogs,
represents a bridge of five unique areas of expertise: and web sites which only require an HTML coder.
(1) educational pedagogy, (2) content and skill exper- The web is searchable and can readily archive learning
tise, e.g., clinical reasoning, (3) video production, objects such as cased-based modules. There are a grow-
(4) case construction, and (5) web technology. Collab- ing number of existing platforms, software building
oration amongst those who together have expertise in blocks, and digital technologies that make the design
all five critical areas is essential for high-quality CBLW. of E-learning tools faster and less costly.
CBLW must be based upon a solid foundation of
curriculum development, and well-established educa- Important Scientific Research and
tion principles must be applied, for example, active Open Questions
learning, activation of prior knowledge, constructiv- There is a paucity of scientific research on CBLW. There
ism, and feedback. is far more research on E-learning in general, emerging
How CBLW differs from conventional learning from reviews of health education studies (Chumley-
methods, such as a lecture or classroom small group, Jones et al. 2002; Ruiz et al. 2006; Cook 2007; Wong
is exemplified by the utility and near necessity of et al. 2010). Their reviews and summaries, based on
storyboards in the curricular design process. The studies that include some CBLW, point out that curric-
storyboarding process, as practiced today, was devel- ula must be well-designed curricula regardless of mode;
oped at Walt Disney studio during the 1930s. The that E-learning is not intrinsically superior in either
Disney approach developed storyboarding as a visual learning outcome or learner satisfaction; that it can be
508 C Case-Based Learning on the Web
ME Quiz
WELCO
1 4
2 5
3 6
cost-saving; and that since the internet is here to stay, learning with peers is lost, and only partly realized if
so is E-learning, including its CBLW variant. Wong the module is used to engage with other students. That
proposes a set of questions related to technology accep- the learning is on the web, as opposed to a classroom
tance and achieving interactive dialogue for educators or typical workplace, increases the possibility of the
to address to maximize effectiveness and perceived learner to be distracted simultaneously by other web-
usefulness. The five questions are: How useful will the based and electronic activities, such as e-mail, web-
prospective learners perceive the Internet technology surfing, music, and other entertainment. On the other
to be? How easy will the prospective learners find hand, the web can also provide access to instant and
this technology to use? How well does this format fit unlimited sources of learner-centered reference mate-
in with what learners are used to and expect? How rials, which may enhance case-specific learning. The
will high-quality human–human (learner–tutor and effect on learning by the increased availability of all
learner–learner) interaction and feedback be achieved? these types of distraction, and subsequent decrease in
How will high-quality human–technical interaction singular focus on the case-based module, is unknown.
and feedback be achieved? The future of the web, including its capacities and
One of the few studies of CBLW specifically is also how it is accessed, will lead to intriguing possibilities
from the medical education literature (Nathoo et al. for the design and use of CBLW. Mobile web access
2005), and its findings include higher levels of stu- via increasingly portable equipment, web-based social
dent engagement and relationships with faculty and learning platforms and culture, and the development
accountability to the learner peer group compared to of artificial intelligence web applications will greatly
the classroom problem-based learning tutorial format. expand the scope of how and for what educators and
CBLW also provided benefits of self-selected student learners can use case simulations on the web. For
pace and more realistic student experiences. Nathoo example, in the health-care education and practice
et al. suggest the need for developing new metrics for arenas, the emergence of the electronic medical record,
measuring level of student collaboration outside the artificial intelligence–derived real-time guidance and
classroom, and evaluation systems that test higher projected health outcomes based on gaming theory,
levels of abstraction, beyond simple recall of factual instant access to patient study results, and virtual com-
information, and that measure authentic challenges munication with patients and other members of the
and competencies that medical practitioners face. health-care team may converge with simulation-based
Applying research on E-learning to CBLW sug- learning. CBLW may someday even transform into
gests that there are important limitations. It is not a personalized, real-time learning that is no longer
replacement for learning through real case-based prac- simulated, but instead a form of web-enabled aug-
tice such as piloting an airplane or working with mented reality.
a patient, but is effective as preparation for real prac-
tice. The actual mentor–learner relationship is lost, Cross-References
though this is mitigated in those cases when the mod- ▶ Active Learning
ule is used for synchronous or asynchronous learn- ▶ Assessment in Learning
ing with the teacher. Similarly, the social context of ▶ Asynchronous Learning
Categorical Learning C 509
▶ Audiovisual Learning
▶ Blended Learning Categorical Learning
▶ Case-Based Learning
▶ Computer-Based Learning SHAWN ELL1, MONICA ZILIOLI2
1
▶ Distributed Learning Environments Department of Psychology, Graduate School of C
▶ e-Learning and Digital Learning Biomedical Sciences, University of Maine, Orono,
▶ Evaluation of Student Progress in Learning ME, USA
2
▶ Learning by Doing Department of Psychology, University of Maine,
▶ Learning from Video Orono, ME, USA
▶ Online Learning
▶ Problem-Based Learning
▶ Simulation-Based Learning Synonyms
▶ Technology-Based Learning Categorization; Category learning; Classification
▶ Twenty-First-Century Skills
▶ Video-Based Learning Definition
▶ Virtual Learning Environments From bacteria categorizing a molecule as nutrient or
poison to humans categorizing individuals as friend
References or foe, categorical learning is a process that is vital
Chumley-Jones, H. S., Dobbie, A., & Alford, C. L. (2002). Web- for the existence of any organism. More formally,
based learning: Sound educational method or hype? A categorical (or category) learning is the process of
review of the evaluation literature. Academic Medicine, 77(10), establishing a memory trace that improves the effi-
S86–S93. ciency of assigning novel objects to contrasting
Cook, D. A. (2007). Web-based learning: Pros, Cons and controver-
groups. In addition to facilitating the categorization
sies. Clinical Medicine, 7, 37–42.
Huang, C. (2005). Designing high-quality interactive multimedia of objects, categorical knowledge also facilitates a vari-
learning modules. Computerized Medical Imaging and Graphics, ety of cognitive processes.
29(2–3), 223–233. In defining categorical learning, it is useful to
Nathoo, A. N., Goldhoff, P., & Quattrochi, J. (2005). Evaluation of an consider what it is not. One important distinction is
interactive case-based online network (ICON) in a problem
between categories and concepts. A category is a col-
based learning environment. Advances in Health Sciences Educa-
tion, 10(3), 215–230.
lection of related objects (from a single or multiple
Ruiz, J. G., Mintzer, M. J., & Leipzig, R. M. (2006). The impact of stimulus modalities). A concept, in contrast, is a col-
E-learning in medical education. Academic Medicine, 81(3), lection of related ideas. Another important distinction
207–212. is between novel and well-learned categories. The rules
Wong, G., Greenhalgh, T., & Pawson, R. (2010). Internet-based med- that govern the learning of novel categories and the
ical education: A realist review of what works, for whom and in
access of information from well-learned categories are
what circumstances. BMC Medical Education, 10, 12–22.
likely quite different. For instance, patients with neu-
rological damage resulting in the loss of a well-learned
category (e.g., tools as in one type of category-specific
visual agnosia) do not lose the ability to learn novel
Case-Based Reasoning categories. Also, patients with neurological damage
resulting in a categorical learning impairment (e.g.,
▶ Analogical Reasoning patients with Parkinson’s disease) do not lose well-
▶ Schema-Based Reasoning learned categories.
Theoretical Background
Since antiquity, categorical learning has been thought
Categorical Analysis to be a central ability underlying cognition. Not sur-
prisingly, categorical learning has been one of the most
▶ Categorical Representation thoroughly studied areas of ▶ cognitive psychology.
510 C Categorical Learning
Theories of categorical learning, however, did not that every exemplar is stored in memory. Given
obtain prominence until the seminal work of Bruner current knowledge of the neural substrates of memory
et al. (1956) at the dawn of the cognitive revolution. formation, this assumption is implausible as a general
The work of Bruner and colleagues can be traced feature of categorical learning.
back to the Greek philosopher Aristotle and postu- At the same time exemplar theory was gaining
lates that categories are represented by their defining prominence, decision-bound theory (based upon gen-
attributes. Defining attributes are the set of singly eral recognition theory, Ashby and Townsend 1986)
necessary and jointly sufficient features for category was also being developed. Decision-bound theory is a
membership. Although, this so-called classical theory multivariate generalization of signal-detection theory.
continues to be highly influential, its shortcomings are It is assumed that, on each trial, the perceived stimulus
widely accepted. For instance, there are many catego- can be represented as a point in a multidimensional
ries for which it is difficult, if not impossible, to list psychological space and that each participant parti-
the necessary and sufficient features (e.g., games). In tions the perceptual space into response regions by
addition, classical theory incorrectly predicts that all constructing a decision bound. The participant deter-
category ▶ exemplars are equally representative cate- mines which region the perceived stimulus is in, and
gory members. then makes the corresponding response. An important
Prototype theory was proposed as an alternative contribution of decision-bound theory is that it sepa-
to classical theory (Rosch and Mervis 1975). rates perceptual and decisional influences on categori-
According to prototype theory, the category ▶ repre- cal learning. Thus, selective attention, for example, can
sentation consists of the most typical member of the affect the perceptual representation of the stimulus as
category (i.e., the ▶ prototype) and the categorization well as how stimulus dimensions are weighted in mak-
of novel exemplars is based upon similarity to the ing categorization decisions.
prototypes of contrasting categories. Prototype theory, Decision bounds can take many different forms
unlike classical theory, predicts that category mem- and, therefore, can mimic other theories of categorical
bership is graded and, as a result, captures the well- learning. One class of decision-bound models assumes
documented finding that some category members that independent decisions are made about all (or
are more typical than others. Although economical, some subset) of the stimulus dimensions. Such models
the assumption that the category representation is are closely related to classical theory and have led to
restricted to only the most prototypical member is the development of so-called rule-based theories of
rather limiting. For instance, information about vari- categorical learning. According to rule-based theories,
ability and correlational structure within a category has logical expressions are used to evaluate category mem-
been lost. bership (e.g., if the stimulus has a value on dimension
Exemplar theory, in contrast, provides a richer cat- X greater than some decision criterion, it belongs
egory representation by assuming that the categoriza- in category A; otherwise it belongs in category B).
tion of novel exemplars is based upon similarity to Thus, rather than a list of defining attributes, the cat-
the stored representations of all previously experi- egory representation is simply the decision criterion.
enced instances of the contrasting categories (Nosofsky Another class of decision-bound models assumes that
1986). The high resolution of the category representa- the decision boundary is midway between the catego-
tion enables exemplar theory not only to predict ries. Such models are generally equivalent to proto-
the phenomena accounted for by prototype theory, type models because the same categorization response
but also to predict effects that are dependent upon would be predicted regardless of whether distance to
within-category variance and correlation such as the the category boundary or distance to the category pro-
influence of category members that are far from the totypes is used to make a decision. Although decision-
prototype. An additional contribution of exemplar bound theory can mimic other theories of categorical
theory was to formally incorporate a mechanism for learning, the fundamental category representation is
selectively weighting some stimulus dimensions over restricted to the decision boundary thereby limiting
others (i.e., ▶ selective attention). One enduring criti- decision-bound theory as a general theory of categor-
cism of exemplar theory is based on the assumption ical learning.
Categorical Learning C 511
Although these divergent theoretical perspectives Prototype-distortion tasks are those in which the cate-
have been hotly debated for more than 50 years, it is gory members are generated by randomly perturbing
difficult for any theory to claim victory. Arguably, the category prototype. Prototype-distortion tasks typ-
exemplar theory has been the most popular of the ically instruct participants to distinguish between cat-
categorization theories. It is important to note, how- egory members and nonmembers (i.e., A-not A tasks), C
ever, that mathematical models derived from prototype but it is not uncommon to use two contrasting catego-
and decision-bound theory have often been shown ries (i.e., A-B tasks). Numerous behavioral dissocia-
to outperform, or perform equivalently to, models tions between these tasks support the utility of this
derived from exemplar theory. For instance, when pro- task-based taxonomy. It is important to note, however,
totype models make the more realistic assumption that that although different categorical learning systems
a category has multiple prototypes (rather than a single may be better suited to learn a particular task, there
prototype) many of the aforementioned criticisms of can be considerable individual differences in how par-
prototype theory are resolved. ticipants learn these tasks.
It is important to stress that all of these theories Cognitive neuroscience research utilizing neuroim-
make an important contribution. Indeed, many aging and neuropsychological methdologies indicate
researchers have embraced the idea that the “correct” that categorical learning in these three types of tasks
theory, or ▶ system, varies depending upon the partic- relies upon different neural circuits. Rule-based tasks
ular categorization task. This is not too surprising given have been shown to depend upon lateral prefrontal
that most studies advocating a particular theoretical cortex and anterior regions of the ▶ basal ganglia.
perspective tend to investigate the same type of cate- Information-integration tasks have been shown to
gorization task. For example, exemplar theory has depend upon a neural circuit linking high-level, sen-
enjoyed considerable success in accounting for data sory cortical areas (e.g., inferotemporal cortex in
from categorical learning tasks where memorization is the case of visual stimuli) to high-level motor areas
plausible given the small number of category exem- (e.g., premotor areas) via posterior regions of the
plars. Similarly, decision-bound theory has enjoyed basal ganglia and the thalamus. A-not A prototype-
considerable success in accounting for data from cate- distortion tasks depend upon extrastriate visual corti-
gorical learning tasks where memorization is implau- cal regions whereas A-B prototype-distortion tasks also
sible given the large number of category exemplars. depend upon prefrontal and parietal cortices.
The idea that distinct learning systems contribute
to categorical learning has been suggested by many Important Scientific Research and
researchers over the last 30 years (e.g., Ashby et al. Open Questions
1998). Multiple systems theorists generally agree that Multiple systems theorists are faced with at least two
one system is rule-based. Differences between alterna- critical challenges. The first centers on characterizing
tive theories center on how best to characterize the categorical learning systems. In pursuit of this task,
other system(s), in particular, issues related to the researchers must carefully define criteria for determin-
nature of the category representation (e.g., exemplar ing whether a putative system is, in fact, a separate
vs. prototype). system or run the risk of system proliferation. Impor-
The argument for multiple categorical learning tantly, characterizing systems not only requires specifi-
systems has been fueled, in large part, by the fields of cation of the cognitive processes, but also the neural
behavioral and ▶ cognitive neuroscience. Such substrates. As systems are characterized, the challenge
research has generally focused on three types of cate- of understanding how they interact becomes para-
gorical learning tasks. Rule-based tasks are those in mount. Current theories assume that categorical learn-
which the categories can be learned by an explicit ing systems operate in parallel and compete with each
reasoning process using logical rules. Information- other, but there is little data to rule out other types of
integration tasks are those in which logical rules have interactions (e.g., cooperation). Even if competition is
limited success and, instead, accuracy is maximized by the correct assumption, there is very little data to guide
combining information from two or more stimulus theorizing on how competition is resolved on a trial-
dimensions prior to making a categorization response. by-trial basis.
512 C Categorical Learning in Pigeons
Historically, cognitive psychology has been insular ▶ Explicit Versus Implicit Learning
in its study of psychology and the study of categorical ▶ Mathematical Models/Theories of Learning
learning has been no exception. Recent research has
embraced classic findings from other disciplines within
References
psychology and is beginning to incorporate these ideas
Ashby, F., Alfonso-Reese, L., Turken, A., & Waldron, E. (1998).
into theorizing on categorical learning. For instance, A neuropsychological theory of multiple systems in category
categorical learning is influenced by an individual’s learning. Psychological Review, 105(3), 442–481.
motivation for performing the task and how these Ashby, F., & Townsend, J. (1986). Varieties of perceptual indepen-
motivations match the task incentives. In addition, it dence. Psychological Review, 93(2), 154–179.
is now apparent that the ▶ social stressors we encoun- Bruner, J. S., Goodnow, J. J., & Austin, G. A. (1956). A study of
thinking. New York: Wiley.
ter in our daily lives can have a profound influence on
Markman, A., & Ross, B. (2003). Category use and category learning.
categorical learning. Whether social stressors impair or Psychological Bulletin, 129(4), 592–613.
enhance categorical learning depends upon the type Nosofsky, R. M. (1986). Attention, similarity, and the identification-
of categorization task. Currently, there is no mecha- categorization relationship. Journal of Experimental Psychology:
nism within purely cognitive theories of categorical General, 115, 39–57.
Rosch, E., & Mervis, C. B. (1975). Family resemblances: Studies in
learning, or cognitive neuroscience theories, to ade-
the internal structure of categories. Cognitive Psychology, 7,
quately explain these data. 573–605.
A thorough understanding of categorical learning
requires an appreciation of differences in training
methodology. The vast majority of studies that have
guided theory development can be classified as
supervised learning studies in which a trial consists of
stimulus presentation, categorization response, and
Categorical Learning in Pigeons
corrective feedback. In contrast, unsupervised learning
FABIAN A. SOTO, EDWARD A. WASSERMAN
studies omit corrective feedback. Another popular
Department of Psychology, Delta Center, University of
methodology requires the participant to use the
Iowa, Iowa City, IA, USA
value of the stimulus on a subset of the dimensions
and the correct category label to infer the value of the
stimulus on a missing dimension. These methodologies
have been used in isolation or hybridized in various
Synonyms
Concept learning in pigeons; Pigeon classification
ways. Importantly, however, the choice of methodology
behavior
can have a profound impact on the category represen-
tation (Markman and Ross 2003). For example, the
category representation resulting from unsupervised Definition
training is restricted to be rule-based. In addition, Categorical Learning in Pigeons refers to the process by
supervised training enhances the representation of which these animals come to treat different stimuli
between-category differences whereas inference train- equivalently, as members of a single class. Evidence of
ing enhances the representation of within-category this learning would require affirmative answers to these
similarities. Together, these training methodologies three questions: can pigeons respond differently to
constitute a powerful set of tools to study categorical members of different classes of stimuli, can pigeons
learning. respond similarly to members of the same class of
stimuli, and can pigeons transfer these behavioral ten-
Cross-References dencies to novel instances of the relevant categories?
▶ Categorical Learning in Pigeons Pigeons can indeed categorize stimuli in these ways
▶ Categorical Representation when class membership is defined by a variety of
▶ Concept Learning criteria, including perceptual resemblance, common
▶ Explicit and Procedural-Learning Based Systems of associations with an event, and abstract properties of
Perceptual Category Learning stimulus collections.
Categorical Learning in Pigeons C 513
training exemplars leads to slower learning, but it also same equivalence class start sharing a common repre-
leads to higher levels of generalization to novel stimuli. sentation. Now, those stimuli become more difficult to
There is important evidence indicating that pigeons discriminate from each other than would otherwise
can spontaneously detect the perceptual cohesiveness have been the case.
of natural categories even if they are not required to do In abstract categorization learning, pigeons come
so in order to earn reinforcement. Thus, when pigeons to sort stimuli into classes on the basis of abstract
learn that members of one category are associated with relations among stimulus elements. Perhaps the sim-
a response, they more robustly generalize this response plest form of discrimination of abstract stimulus fea-
to other members of the same category than to mem- tures is relational learning. Here, pigeons learn to
bers of different categories. Also, pigeons seem to respond to stimuli on the basis of their relative position
exploit this perceptual cohesiveness during categoriza- in a physical dimension (“larger than” or “brighter
tion training, as suggested by their quicker learning of than”) instead of their absolute value along that
tasks in which all members of the same category are dimension. Although traditional demonstrations of
assigned to the same response than their learning of relational learning in pigeons can be explained as
tasks in which the stimuli are randomly assigned to the arising from the interaction of absolute associative
discriminative responses. values, recent evidence has questioned the generality
The behavioral mechanisms of category learning in of that explanation.
pigeons are flexible enough to allow the concurrent Pigeons have also shown the ability to discriminate
classification of the same stimuli into their basic per- collections of items on the basis of their variability, that
ceptual classes and into superordinate classes created by is, the degree to which the items composing an array
the union of two natural categories (e.g., the class of repeat or vary from each other (Wasserman et al. 2004).
“man-made objects,” created by “chairs” and “cars”). So, after training to discriminate arrays of 16 items on
There is also evidence indicating that these behavioral the basis of whether all of the items are identical or
mechanisms can involve pigeons’ reliance on category- nonidentical, pigeons can be tested with novel arrays
relevant features of the stimuli which additively com- involving mixtures of identical and nonidentical items.
bine to support a particular response: for example, Here, the likelihood of pigeons’ responding “noniden-
wheels, a body, and a roof prompt the recognition of tical” to the mixtures increases with the variability in
a car. Finally, there is growing agreement that these the test array. Still other evidence suggests that such
behavioral phenomena are largely governed by the variability discrimination may lie at the root of the twin
same principles that are at work in associative learning concepts of “same” and “different.”
and stimulus generalization, although until recently Finally, there is evidence showing that pigeons can
there has been little effort to support this claim either also learn to match stimulus collections on the basis
theoretically or empirically. of the second-order relations between them. Thus,
In associative category learning, pigeons group pigeons learn that after being shown a 16-item sample
stimuli on the basis of their association with a common display with a particular relation among its elements
response or some other event, in much the same way (either all of the elements are identical or all of the
that we come to call shoes, pants, and hats items of elements are nonidentical), they must choose another
“clothing.” After training of such common associa- display exhibiting the same relation among elements
tions, if pigeons learn to give a new response to some (either identical or nonidentical). Such relational
stimuli from the original training set, then this new matching-to-sample may represent a form of analogy.
response may generalize to all of the other members of Across all of these different categorization tasks,
the class. Such generalization is taken as evidence of a common factor which increases the likelihood of
pigeons’ ability to acquire an equivalence class, where generalization to novel exemplars of the category is
stimuli are treated equally, not on the basis of their the number of trained stimuli. Larger training sets
perceptual resemblance, but on the basis of their com- lead to better evidence of abstract learning. One inter-
mon training contingencies. There is also evidence pretation of this result is that experience with several
suggesting that after being associated with the same exemplars from each category is necessary for detecting
experimental outcome, stimuli belonging to the the abstract properties of the stimuli. Without such
Categorical Representation C 515
rich experience, the pigeons might pay attention to Herrnstein, R. J. (1990). Levels of stimulus control: A functional
more concrete perceptual features, which are irrele- approach. Cognition, 37, 133–166.
Lazareva, O. F., & Wasserman, E. A. (2008). Categories and concepts
vant to task solution. Another possibility is that a
in animals. In J. H. Byrne (Ed.), Learning and Memory:
large number of training stimuli simply increases the A Comprehensive Reference (pp. 197–226). Oxford: Academic.
likelihood of a test stimulus being perceptually similar Mackintosh, N. J. (2000). Abstraction and discrimination. In C
to one or more of the training stimuli. Considerable C. M. Heyes & L. Huber (Eds.), The Evolution of Cognition
work has explored these two possibilities. (pp. 123–141). Cambridge: MIT Press.
Wasserman, E. A., Young, M. E., & Cook, R. G. (2004). Variability
discrimination in humans and animals: Implications for adap-
Important Scientific Research and tive action. The American Psychologist, 59, 879–890.
Open Questions
In the past, considerable research and discussion have
focused on determining whether pigeons and other
nonhuman animals use something like human con-
cepts in mastering categorization tasks. This anthropo- Categorical Perception
centric line of research paid little attention to the
Categorical perception is defined as an “abrupt percep-
possibility that different types of conceptual processing
tual change at the boundary” (Harnad 2005), which
exist in nature. However, empirical results have forced
can be seen in situations where the perceived change in
researchers to consider just such a possibility. Future
some attributes (e.g., color) does not occur gradually
research is likely to focus more on determining what
but as instances of different categories.
kinds of conceptual processes pigeons and other
animals exhibit and to disclosing similarities and dif-
ferences in these processes across diverse species.
References
Harnad, S. (2005). Distributed processes, distributed cognizers
Expanding the scope of research toward studying
and collaborative cognition. Pragmatics and Cognition, 13(3),
more categorization tasks and more species is likely to 501–514.
be crucial to gain a better understanding of the evolu-
tion of conceptual processes.
Current research in the area of pigeon categoriza-
tion is quickly shifting from studies aimed at discover-
ing pigeons’ categorization abilities to studies aimed at Categorical Representation
pinpointing the mechanisms underlying these abilities.
Recent research has prompted several accounts of ARASH SHABAN-NEJAD
pigeons’ categorization behavior, some of them formal- McGill Clinical & Health Informatics, Department of
ized in quantitative models. Future research and theory Epidemiology, Biostatistics and Occupational Health,
will likely move beyond behavioral comparisons across McGill University, Montreal, QC, Canada
species and toward deeper comparisons involving the
mechanisms of categorization.
Synonyms
Cross-References Categorization; Categorical analysis
▶ Animal Learning and Intelligence
▶ Association Learning Definition
▶ Categorical Learning The origin of the term “categories” is the Greek word
▶ Comparative Psychology and Ethology “ΚatZgοrίai” (Katēgoriai), which refers to the manu-
▶ Similarity Learning script written by Aristotle, wherein he defined ten
fundamental modes (categories) of being (things),
References namely substance, quantity, quality, relative (relation),
Critchfield, T. S., Galizio, M., & Zentall, T. R. (Eds.). (2002). Catego- somewhere (location), sometime (when), being-in-
rization and concept learning [Special issue]. Journal of a-position, having (state), acting, or being affected
the Experimental Analysis of Behavior, 78, 237–607. (Ackrill 1975). The word “representation,” as defined
516 C Categorical Representation
by the Oxford English Dictionary, means “the action or e.g., man, horse, tree, small, big. Then new concepts
fact of expressing or denoting [a thing] symbolically.” can be learned through a set of descriptive information
Categorical representation can be described as the pro- (e.g., logical expressions) based on the basic concepts
cess of expressing things in different modes and layers and attributes, e.g., Centaur (man-horse) or Pony
of abstraction based on similarities and differences in (small horse). ▶ Iconic representation and feature dis-
their attributes and relations. Categorical representa- crimination, which lead us toward categorical repre-
tion has been a subject of study in knowledge represen- sentation, contribute to learning by acquaintance, and
tation, mathematics, cognitive science, linguistics, ▶ symbolic representation, which uses category names
philosophy, psychology, art, and so forth. Members of as the atomic symbols, is used for descriptive learning
a category have common attributes and together repre- (Harnad 1987). Together, categorical representation
sent perceptual or conceptual knowledge about a par- and iconic/symbolic representations (Harnad 1996)
ticular domain of interest. enable us to describe and model the real world in
terms of categories and their members, their relation-
Theoretical Background ships, and their attributes.
The human brain has the ability to organize the details A categorical representation of a domain can be
of perceived objects within a series of categories based performed by defining categories at different hierarchi-
on their common features. Similarly, the objects in the cal levels, depending on the level of granularity, using
real world can be processed (i.e., compared, evaluated, different mediums such as Hierarchies, Sets, Lists, and
and remembered) by the brain, based on known attri- so forth. In Artificial Intelligence (AI), ontologies are
butes and past experiences. This allows people to incre- employed based on this ability of humans to find
mentally acquire new knowledge (e.g., discriminating things familiar by using the categorizations in their
between life-threatening situations versus safe ones) brains. Ontologies, as hierarchical organizations of cat-
and communicate with each other through the shared egories from general to specific, are meant to provide
conceptualization of the subject. Therefore, the process a semantic and conceptual basis for sharing knowledge
of categorization is important for decision making. about a domain of interest by defining concepts,
Categorical representations of different types of expres- properties, and axioms. In an educational sense, this
sions (e.g., facial, phonetic, emotional, and mental conceptual model enables humans to apply their expe-
expressions) to distinguish between different concep- riences of the past to similar future situations. For
tual and perceptual behaviors in the human brain have example, the experience of riding a bicycle can be
been widely studied in the literature. Categorization applied to riding different bicycles with different
also plays a crucial rule in human cognitive develop- brands, models, colors, and sizes. Since categories are
ment and is essential to several learning activities, highly dependent on a human’s knowledge about the
including language acquisition, grammar learning, real world, they will evolve (be recategorized) as our
and speech perception. knowledge increases.
Two types of categorization, namely perceptual
(based on perceptual similarities between entities) Important Scientific Research and
and conceptual (based on the functions and interac- Open Questions
tions between entities), can be commonly defined in The idea of categorization is central to many disci-
the human brain, even from the early stages of infancy plines in AI, machine learning, cognitive science, knowl-
(Berg-Cross 2006). Unlike perceptual categorization, edge representation, and so on. Through technological
which is more focused on the appearance of entities, advances, different formalisms and methods can be used
the conceptual model is based on experience-driven to support categorical representations. An example is
patterns and needs a greater degree of maturity in the employing neural networks for iconic and categorical
human’s mental model. As stated by Harnad (1987), representations of different cognitive systems. In lin-
the basic categories are generated through ▶ categori- guistics, the associations between labels (terminologies)
cal perception and specified through a learning process and perceptual categories are considered key factors for
(▶ learning by acquaintance). In this way, one classifies language acquisition studies, perceptual learning, and
the perceived objects (things) and then names them, developing “generative grammars.”
Categorization C 517
Research on categorical representation has been ▶ Explicit and Procedural-Learning Based Systems of
faced with several challenging questions on the nature Perceptual Category Learning
and semantics of categories and types of representa- ▶ Hierarchical-Network Model for Memory and
tions. Categories are derived based on different data Learning
sources (e.g., cognitive, behavioral, and environmental ▶ Knowledge Integration C
data). Categorization is defined in cognitive science ▶ Knowledge Organization
as “the process of dividing the world into categories, ▶ Knowledge Representation
and usually involves constructing concepts that provide ▶ Mental Representations
mental representations of those categories” (Thagard ▶ Ontology and Semantic Web
and Toombs 2005), and can be done for both observ- ▶ Representation, Presentation and Conceptual
able concepts (e.g., humans, limbs) and nonobservable Schemas
concepts (e.g., genes, disease agents, a process such ▶ Similarity Learning
as injection). In the case of categorizations for ▶ Vocabulary Learning
nonobservables, the process also involves creating con- ▶ Word Learning and Lexical Development Across the
cepts for the unambiguous rationalization of the real Lifespan
world (Thagard and Toombs 2005). More formal cate-
gorization is also referred to as “any systematic differ- References
ential interaction between an autonomous, adaptive Ackrill, J. L. (1975). Aristotle: Categories and de interpretatione
sensorimotor system and its world” (Harnad 2005). (Clarendon Aristotle Series). USA: Oxford University Press.
Berg-Cross, G. (2006). Developing knowledge for intelligent agents:
In this definition, the term “systematic” has been used Exploring parallels in ontological analysis and epigenetic robot-
to exclude arbitrary interactions (e.g., the effects of ics. NIST PerMIS conferences 2006.
the wind blowing on the sand) and an “autonomous, Harnad, S. (1987). Category induction and representation. In
adaptive sensorimotor system” means a dynamic sys- S. Harnad (Ed.), Categorical perception: The groundwork of cog-
tem that interacts and changes in time through nition. New York: Cambridge University Press. Chapter 18.
Harnad, S. (1996). The origin of words: A psychophysical hypothesis.
adaptive changes in the states of the system. “Differen-
In B. Velichkovsky & D. Rumbaugh (Eds.), Communicating
tial” implies that the categorization process generates meaning: Evolution and development of language (pp. 27–44).
a different kind of output from a different kind of input New Jersey: Erlbaum.
(Harnad 2005). Harnad, S. (2005). To cognize is to categorize: Cognition is categori-
The categorical perspective in knowledge represen- zation. In H. Cohen & C. Lefebvre (Eds.), Handbook of categori-
tation intends to express universal notions (truths). zation in cognitive science (pp. 19–43). Amsterdam: Elsevier.
Thagard, P., & Toombs, E. (2005). Atoms, categorization and con-
Category theory, with its universal grammar, provides
ceptual change. In H. Cohen & C. Lefebvre (Eds.), Handbook of
an advanced abstract mathematical model that is used categorization in cognitive science (pp. 243–254). Amsterdam:
to represent and analyze the behavior of interacting Elsevier.
objects within categories. The basic notations in cate-
gory theory consist of a class of objects and a class of
morphisms (relations between the objects), an identity
and a composite morphism. The declarative approach Categorization
offered by category theory represents and describes
objects only in terms of their relationships and inter- The ability to group objects or events according to
actions with other objects, without the necessity of a common attribute (or by category). In categorization,
knowing about the internal structure of objects. stimuli are grouped based on complex features,
multimodal properties, or behavioral relevance. Sen-
Cross-References sory similarity alone does not necessarily place stimuli
▶ Categorical Learning in the same category.
▶ Categorical Learning in Pigeons
▶ Classification Learning Cross-References
▶ Classification of Learning Objects ▶ Categorical Learning
▶ Conceptual Clustering ▶ Categorical Representation
518 C Categorization of Variation in Movement
live in the minds of others without knowing it.” The For purpose of discussion, suppose that grief
last two lines became particularly important because involves bodily preparation to cry. Sobbing with tears
Mead and his followers did not follow them up, as would require, at the least, muscular contractions in
indicated below. Cooley was referring to the reflex- order to sob, activation of the tear glands, and some
ive self-consciousness of our experience, how we adrenaline to energize these preparations. The more C
continually monitor our self from the point of view rapidly these preparations are carried through, the
of others. less feeling of sadness. If one cried copiously and
G. H. Mead developed Cooley’s idea in a different instantly, little sadness would be experienced. Sadness
way. He called it “taking the role of the other.” Mead requires delay, just as sexual pleasure can be heightened
pointed out that ordinary discourse is so ambiguous by foreplay. Crying, under certain conditions discussed
that we must be constantly moving in and out of the below, might be the orgasm for grief.
mind of the other person, guessing at the meaning of Embarrassment/shame provides another example.
their discourse by seeing it not only from our own When my students tell the class their most embarrassing
point of view, but also from theirs. Unlike Cooley, moment, many of them are convulsed with laughter
Mead and his followers failed to note how unconscious telling the story. Laughter seems to be the orgasm of
this process becomes. And neither Cooley nor Mead shame. However, it is often difficult to attain enough
realized the relevance of their work to the distancing of distance, especially if one was deeply humiliated. Many
emotions. repetitions of just talk about the incident may be needed
My students experience roller coasters as pleasur- before one can find humor in it.
able, but only if they are sure that the ride is safe. It also needs to be said that just as there is a good cry
They allow themselves to feel fear because they are and a bad one, there is also a good laugh and a bad one.
able, at the same time, to feel safe, rather than becom- A good laugh turns out to be when one is laughing at
ing completely caught up. Levine (1997) refers to this one’s self (“silly me”) or the universe, but not at other
process as pendulating, moving very rapidly in and out people. Laughing at others, as Billig has pointed out
of emotions that would otherwise be painful. We move (2005), usually is ridicule, driven by anger: no help to
so fast that we usually do not realize it. These states can either party. There is also faked laughing, which does
occur not only in the theater but whenever we feel safe not engage any part of the cathartic system, but is more
enough to replay intense emotional experiences, such like a voluntary speech act.
as describing them to another person we trust, or, Esthetic distance is experiencing strong emotions in
occasionally, reliving them alone. a safe environment: theater, film, books, songs, or tell-
ing one’s experience to an empathic person, or even to
Important Scientific Research and one’s self. I once had an intense fear experience in this
Open Questions mode. After an excruciatingly dangerous moment,
Aristotle linked catharsis to clarification or illumina- when I was safe, I realized that I was still tensed up
tion, but he did not explain the connection. In which because of the danger I had encountered. Not knowing
way does catharsis lead to these desirable outcomes? In what to do, I began repeating the phrase “I am afraid.”
order to understand what is taking place in catharsis, After many repetitions my body took over, shaking and
emotions need to be defined. John Dewey (1894/1895) sweating till my clothes were drenched. It was not
proposed that felt emotions are certain bodily prepa- painful, and I felt completely relaxed when it was
rations to act that have been delayed. Since Dewey’s over. Perhaps it was the nearest that I ever came to an
article dealt only with emotions in general, and not illumination. Shaking and sweating would seem to
specific emotions, it had very little influence. It signal the catharsis of fear.
becomes relevant only if we apply it to specific emo- Like many people, when angry I may lash out. But
tions, like grief or shame, anger or fear. These emotions I have had several anger experiences of a quite different
occur when the body is mobilized to act in certain kind. I told the culprit “I am angry at you because. . . . .”
ways, states of bodily arousal in order to complete in an ordinary voice. Since this approach is so undra-
certain acts. What are these acts, and how can they matic, I have had to repeat my complaint several times.
be completed? Then two things happened: the other person started
520 C Causal Attribution
Cross-References
▶ Aristotle (384 BC–322 BC)
▶ Aristotle on Pleasure and Learning Causal Learning
▶ Dewey, John
▶ Psychodynamics of Team Learning AARON P. BLAISDELL1, RALPH R. MILLER2
1
Department of Psychology, University of California,
References Los Angeles, CA, USA
Bergson, H. (1911). Laughter: An essay on the meaning of the comic. 2
SUNY-Binghamton, Binghamton, NY, USA
New York: C. Brereton.
Billig, M. (2005). Laughter and ridicule: towards a social critique of
humor. Thousand Oaks, CA: Sage.
Cooley, C. H. (1922). Human nature and the social order. New York: Synonyms
Scribner’s Sons. Causal induction; Causal inference; Causal reasoning;
Dewey, J. (1894/1895). The theory of emotion. Psychological Review, Contingency learning
1, 553–569, and II, (2), (1895): 13–32.
Freud, S., & Breuer, J. (1895/1966). Studies on hysteria.
New York: Avon. Definition
Goddard, H. (1951). The meaning of Shakespeare. Chicago, IL: Learning the cause–effect relationships or determining
University of Chicago Press. the causal status among a set of two or more events.
Levine, P. A. (1997). Waking the tiger: healing trauma. Berkeley, CA: Learning causal relationships can be characterized as
North Atlantic Books.
a bottom-up process whereby events that share contin-
Mead, G. H. (1936). Mind, self, and society. Chicago, IL: University of
Chicago Press.
gencies become causally related, and/or a top-down
Nussbaum, M. (1986). The fragility of goodness. Cambridge: Cam- process whereby cause–effect relationships may be
bridge University Press. inferred from observation and empirically tested for
Scheff, T. (1979). Catharsis in healing, ritual, and drama. Berkeley, its accuracy.
CA: University of California Press. Re-issued in 2001 by
iUniverse.
Scheff, T. (2007). Catharsis and other heresies: a theory of Theoretical Background
emotion. Journal of Social, Evolutionary and Cultural Psychology, Causal learning has its roots in philosophy. Aristotle
1(3), 98–113. proposed four causes: material (what something is
Scheff, T. (2009). Social science of emotions. http://www.youtube.com/ made of), formal (i.e., structural, how something is
watch?v=DM_MxBizcQk1513cath111march1-10
made, its structure and form), efficient (or moving;
necessary for the effect’s existence), and final (i.e.,
functional, the purpose, an egg is the cause of
a chicken). The British Empiricists (Hume, Lock,
Causal Attribution J. S. Mill, et al.) suggested that cause–effect relation-
ships cannot be observed, but are merely inferred
▶ Attribution Theory of Motivation through statistical regularities between events, often
Causal Learning C 521
captured in associative properties (see e.g., Hume or desirable outcomes) or prevent the goal from
1739). Nativists, such as Kant (1781), argued that the happening (for aversive or undesirable outcomes)
human mind has a priori knowledge of the construct of (Dickinson 2001). In this framework, instrumental
causality. The concept of causation is applied to our actions are suggested to be mediated by causal knowl-
knowledge (both a priori and acquired through expe- edge. Much of the work to support this framework C
rience) to allow us to label events as causal when they comes from research investigating the parallels between
appear so to us. associative learning phenomena in nonhuman animals
Investigation of causal learning in psychology fol- and similar phenomena in human contingency learn-
lows from these philosophical roots. Treatment of ing experiments. The degree to which effects in human
concepts involving causal learning and induction fall contingency learning mirror those found in animal
into three groups: Perception, Associative learning, conditioning experiments establishes the latter as a
and Reasoning. model for the former. This approach has been largely
Belgian psychologist Albert Michotte argued that successful in establishing a connection between these
causality is determined directly through perception. two research paradigms, and few would dispute that
He demonstrated this by describing our perception of this similarity is meaningful. Where the debate centers
causality in how billiard balls move and interact on a is on the interpretation of this similarity between ani-
billiard table. When one billiard ball strikes a second, mal conditioning experiments and human contingency
the first ball transfers its motion to the second. learning experiments. Proponents of the associationist
Michotte (1963) referred to this perception of transfer approach argue that the similarity reflects the role
of movement from one colliding object to the next of the simple, algorithmic-level learning mechanisms
as “ampliation of the movement,” what is now generi- of Pavlovian and instrumental conditioning in causal
cally referred to as the “launching effect.” This gestalt learning in both nonhuman animals and humans. An
approach treats causal knowledge as being derived alternative perspective is that the similarities between
directly from perception rather than acquired through these two research paradigms reflect the operations of
experience of contingency relations between causally rational top-down psychological principles of causal
connected events. Thus, Michotte’s framework – reasoning and induction at least in humans and per-
which still dominates the field of causal perception – haps in nonhuman animals as well.
shares more with Kant’s nativist framework than An alternative theoretical approach to causal learn-
with Hume’s empiricism. ing and reasoning involves the application of rational
The associative learning approach to causal learn- statistical models (also called normative or functional
ing is a direct descendent of the associationist phi- models) to human causality. This approach has also
losophy of David Hume. Proponents of an associative been extended to work with nonhuman animals in
learning approach to causal learning and induction recent years (Penn and Povinelli 2007). According to
argue that the laws of associative learning, such as the normative approach, causal knowledge is acquired
contiguity, contingency, and temporal priority, provide by computing the covariation between candidate
a sufficient account for how humans and other animals causes and effects. The delta-p model is one popular
acquire understanding of cause–effect relationships. generic form of the computation rule for the contin-
Pavlovian conditioning involves pairing an anteced- gency between cause and effect (see Fig. 1; after Allan
ent event (called a conditioned stimulus or CS) with 1980). The indicated conditional probabilities can be
a subsequent, usually motivating, event (called the pieced together into a causal model. A causal model is
unconditioned stimulus or US), thereby establishing a representation containing both a structural frame-
a CS–US association. The CS–US association may be work consisting of links between causes and effects,
represented causally, with the CS as the cause of the US. and the strength of the relationship of each link, also
Instrumental learning, in which changes in behavior referred to as causal power (Cheng 1997). Rational
are driven by their consequences, may also serve as models typically focus on delineating the rules that
a model of causal learning. This case is particularly govern causal structure learning or how causal power
strong for goal-directed learned behavior in which the is computed. An implicit assumption in these models
action is made as if to produce the goal (for appetitive is that causal relationships reflect either a force that
522 C Causal Learning
▶ Human Contingency Learning effect for a given cause and to attribute the most prob-
▶ Inductive Reasoning able cause for the events in their environment. Learning
▶ Inferential Learning and Reasoning causal relationships between the events in our environ-
▶ Normative Reasoning and Learning ment and between our own behavior and those events
▶ Pavlovian Conditioning is critical for survival. From learning what causes fire C
▶ Psychology of Learning (Overview Entry) (so that we could either produce or prevent the occur-
▶ Role of Prior Knowledge in Learning Processes rence of fire at will) to learning what causes rain, what
causes cancer, or what caused that particular silly acci-
References dent that we had with the car a few days ago, both the
Allan, L. G. (1980). A note on measurement of contingency between history of humankind and our individual history are
two binary variables in judgment tasks. Bulletin of the full of examples in which causal learning is crucial. But,
Psychonomic Society, 15, 147–149. as can be said for other forms of learning as well, causal
Cheng, P. C. (1997). From covariation to causation: A causal power learning is not free of errors. Systematic biases and
theory. Psychological Review, 104, 367–405.
errors are known to occur under certain conditions.
Dickinson, A. (2001). The 28th Bartlett memorial lecture causal
learning: An associative analysis. The Quarterly Journal of Exper-
One of such common biases is the illusion of control.
imental Psychology, 54B, 3–25. The illusion of control can be defined as the belief that
Hume, D. (1964). In L. A. Selby-Bigge (Ed.), Treatise of human nature. one’s behavior is the cause of a desired event that is
London: Oxford University Press (first published 1739). actually independent of it. Illusions of control are an
Kant, I. (1965). Critique of pure reason. London: Macmillan. (Original important factor in the development of superstitions.
work published 1781).
Marr, D. (1982). Vision: A computational investigation into the human
For instance, the superstitious belief that by dancing
representation and processing of visual information. San Francisco: one can produce rain, is normally accompanied by the
W. H. Freeman. illusion of controlling rain.
Michotte, M. (1963). The perception of causality (trans: Miles, T. R. &
Miles, E.). New York: Basic Books. Theoretical Background
Penn, D. C., & Povinelli, D. J. (2007). Causal cognition in human and
The origins of research on causal learning can be traced
nonhuman animals: A comparative, critical review. Annual
Review of Psychology, 58, 97–118.
back to the Greek philosopher Aristotle and it has ever
Shanks, D. R., Holyoak, K. J., & Medin, D. L. (1996). Causal learning since interested philosophers, experimental psycholo-
(The psychology of learning and motivation, Vol. 34). San Diego: gists, cognitive scientists and, in general, all scientists
Academic. interested in how humans learn and acquire knowl-
edge. Nowadays, causal learning is generally studied
in the experimental psychology tradition and is nor-
mally considered to be a central aspect of cognition.
However, as it is the confluence of causal learning and
Causal Learning and Illusions the illusion of control research what we are addressing
of Control in this entry, it is interesting to note that this general
cognitive perspective has not been applied to the study
HELENA MATUTE, MIGUEL A. VADILLO of the illusion of control until very recently. The illu-
Departamento de Fundamentos y Métodos de la sion of control has traditionally been regarded as one of
Psicologı́a, Universidad de Deusto, Bilbao, Spain those cases in which the cognitive system fails to work
in an adaptive manner. As such, the study of the illu-
sion of control has been more often linked to Clinical,
Synonyms Health, and Social Psychology than to the Cognitive
Contingency learning; Illusions of causality; Supersti- and Learning Sciences. Today, however, the study of the
tious behavior illusion of control is recovering its place as part of the
Learning Sciences and is being regarded as the normal
Definition consequence of the way the learning system works.
▶ Causal learning is the process by which people and In a typical laboratory experiment on the illusion
animals gradually learn to predict the most probable of control, a given outcome (e.g., getting points in
524 C Causal Learning and Illusions of Control
a computer game) is programmed to occur at certain specifically, a particular case of causal learning), general
intervals, or according to a predetermined sequence, learning theories that can account for causal learning
and the experimental participants are instructed to try can in principle be applied to the illusion of control
to obtain it. The usual result is that, when asked at the as well. These include theories of ▶ associative learn-
end of the experiment about the extent to which they ing, ▶ connectionist learning, ▶ Bayesian learning, or
believe to have controlled the outcome, participants ▶ inferential learning. Despite their differential pro-
normally believe their control to be significantly greater posals, what is common to all these learning theories
than the value of zero which has been programmed by is that all of them would assume that the illusion of
the experimenter (e.g., Alloy and Abramson 1979). The control is the outcome of a much more general cogni-
current use of ▶ web-based control for experiments on tive mechanism. Many theories that explain causal
human learning allows demonstrating that these effects learning as the formation of associations between
occur not only in the laboratory but also in the more causes and effects, or as statistical reasoning or even
noisy and uncontrolled arena of the Internet. This as an inferential process, would agree to predict an
suggests that the illusion of control is a robust phe- illusion of control when both the candidate cause and
nomenon that develops easily in natural settings. the to-be-explained effect occur frequently and do
Ever since the seminal laboratory studies on the coincide frequently by chance. Not surprisingly, these
illusion of control, Ellen Langer (1975) showed that are the conditions where the illusion of control is most
the personal implication of the participant was an often observed.
important factor in producing the illusion. Therefore, An important additional prediction of the learning
a traditional interpretation has been that emotional approach is that, if the illusion is the result of a normal
and motivational factors, such as a need for control process of causal learning, then it should occur regard-
and a need to protect self-esteem, were at the basis of less of whether the potential cause is the participant’s
the effect. Moreover, an association between the illu- own behavior or an external cause. This is not what
sion of control and an absence of depression has been the Social and Clinical Psychology theories of the illu-
repeatedly reported, which has also lead to the sugges- sion would predict. According to these latter views, the
tion that either the illusion protects from depression, illusion occurs to protect self-esteem and whenever
or depression protects from the illusion (Alloy and the potential cause is external there is no need to
Abramson 1979; Taylor and Brown 1988). In line with protect self-esteem. The amount of evidence in the
this, the illusion of control has been described as the area of learning that shows that causal illusions occur
inverse of the learned ▶ helplessness effect that occurs when the potential cause is an external event suggests
when people realize that desired events are uncontrol- that personal involvement is not needed to produce
lable (e.g., Langer 1975; Matute 1996). These findings these illusions. Personal and motivational factors
have also been taken sometimes as supportive of the could perhaps enhance the illusions, but they are not
motivational, self-esteem explanation, though, as we necessary. Both the illusion of causality that occurs
will see, they do not necessarily support this view over when the potential cause is external and the illusion of
the learning approach. control that occurs when the potential cause is the
Even though it seems clear that the illusion of participant’s behavior are enhanced under the same
control can provide beneficial effects on self-esteem as conditions that are predicted to be critical by the many
well as a protection from depression and helplessness, theories of causal learning. Indeed, many ▶ machine
these prophylactic effects, however comfortable they learning algorithms designed to learn according to
may feel, do not provide an explanation for the illusion. the theories of natural learning will necessarily suffer
This is so because, in the first place, protection of self- illusions of causality (and of control) when exposed
esteem could well be a side effect of the illusion rather to those conditions. Such conditions are many, but
than its cause. Secondly, and most important, because perhaps the most relevant can be summarized as
the self-esteem hypothesis does not attempt to explain follows: (a) a high frequency of occurrence of
how our cognitive system produces the illusion: it a desired uncontrollable outcome (or a low frequency
simply postpones the question. Being the illusion of when the outcome is aversive); (b) a high frequency of
control the product of a learning system (and more the potential cause (i.e., our own behavior when we
Causal Learning and Illusions of Control C 525
speak of an illusion of control; any other cause when linked during causal learning will, from time to time,
we speak, more generally, of an illusion of causality); turn out to be causally unrelated. This would be a
and (c) a high number of coincidences of the poten- collateral effect of the causal learning system working
tial cause and the outcome (Alloy and Abramson in a way which will most often be adaptive and
1979; Matute 1996; Matute et al. 2010). It is interest- correct, but sometimes vulnerable (Matute 1996). C
ing to note that the high frequency of the potential In consequence, as we already noted, many artificial
cause is equivalent with a high personal involvement and machine learning algorithms that model learning
when the potential cause is the participant’s behav- according to the predictions of current theories of
ior. It is possibly for this reason that many of those ▶ human causal learning do also suffer the illusion.
results have often seemed to support the self-esteem This does not mean that the algorithms are programmed
explanation. to do so. However, the illusion is a consequence of
their causal learning dynamics. As of natural selection,
Important Scientific Research and a system that detects causal relations that sometimes
Open Questions result illusory might be more adaptive than an alterna-
One of the challenges related to this topic is to find out tive system with such a high threshold for the detection
what the role of personal involvement really is. Does of causal relations that often fails to detect relations that
it really increase the illusion? If so, why? How? Is it do exist (e.g., McKay and Dennett 2009).
because our perceptual and learning abilities are mod- In addition, the illusion of control itself could be
ified when we evaluate the efficacy of our own behav- adaptive on its own (Langer 1975; Matute 1996; McKay
ior? Could it be that we learn causal relationships in the and Dennett 2009; Taylor and Brown 1988). If the
same way regardless of whether it is our own behavior illusion makes us remain active in our trying to obtain
or an external cause what plays the role of the cue, but desired events, such as rain or fire or health, then,
that we then make a different judgment as a function whenever we are uncertain about whether a relation-
of whether the potential cause is our own behavior? ship is really causal, it should be adaptive to maintain
Many questions related to these ones are becoming the illusion that our behavior is being useful so that
really exciting topics of debate right now. The percep- we persist in trying to obtain the desired outcome.
tion of action, of will, of authorship. . . How do we As a source for behavioral persistence, the illusion of
attribute a given outcome to our own behavior or to control could be at the basis of human change and
other sources? How do we decide that we are responsi- adaptation. The alternative option, which would con-
ble for a certain action? Does this depend on the con- sist in realizing that there is no control over important
sequences of the action? These and other related outcomes and that therefore it makes no sense to
questions concentrate a great deal of the research keep on trying, would produce ▶ helplessness, which
being conducted at present (and possibly in the follow- includes behavioral cessation in addition to depression
ing years) on the illusions and perceptions of causality and other problems. In this sense, it appears that
and of personal control. maintaining a high level of activity is possibly an adap-
Another important issue is whether these effects are tive strategy. Sometimes, however, ceasing dancing for
adaptive and should be promoted, or, by contrast, rain, and even going through a transient depression
should be regarded as maladaptive effects to be after realizing that we cannot cause rain, can be adap-
“corrected” in therapy. This question can be under- tive too. It could cause our efforts to be redirected so
stood in various ways. If we look at the evolution of that we can discover better ways to bring water to
our species, we must admit that if superstitions and our land. As we already noted somewhere else (Matute
illusions of control have survived up to our days, this et al. 2010), applying what we know about the illusion
necessarily must mean either that they are adaptive of control to reduce the impact of superstition in
on their own right or that they are an innocuous our society should contribute to a better world. In
collateral effect of an otherwise adaptive learning pro- one way or another, there must be an optimal level
cess. A possible consequence of the normal functioning of the illusion of control (not too low, not too high)
of the learning system could be that those potential which enhances persistence while still allowing room
causes and effects that occur together and become for change.
526 C Causal Perception
Cross-References
▶ Associative Learning Cause-Effect Learning Versus
▶ Bayesian Learning Effect-Cause Learning
▶ Causal Learning
▶ Predictive Versus Diagnostic Causal Learning
▶ Connectionist Theories of Learning
▶ Human Causal Learning
▶ Inferential Learning and Reasoning
▶ Learned Helplessness
▶ Machine Learning CBL
▶ Web-Based Control for Experiments on Human
▶ Case-Based Learning
Learning
References
Alloy, L. B., & Abramson, L. Y. (1979). Judgements of con- Central Bottleneck
tingency in depressed and nondepressed students: Sadder
but wiser? Journal of Experimental Psychology: General, 108, ▶ Capacity Limitations of Memory and Learning
441–485.
Langer, E. J. (1975). The illusion of control. Journal of Personality and
Social Psychology, 32, 311–328.
Matute, H. (1996). Illusion of control: Detecting response-outcome
independence in analytic but not in naturalistic conditions. CER - Conditioned Emotional
Psychological Science, 7, 289–293.
Matute, H., Yarritu, I., & Vadillo, M. A. (2010). Illusions of causality
Response
at the heart of pseudoscience. British Journal of Psychology.
▶ Conditioned Suppression
doi:10.1348/000712610X532210.
McKay, R. T., & Dennett, D. C. (2009). The evolution of misbelief.
Behavioral and Brain Sciences, 32, 493–561.
Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social
psychological perspective on mental health. Psychological Bulle-
tin, 103, 193–210.
Chameleon Effect
▶ Mimicry in Social Interaction: Its Effect on Learning
Definition
According to Robbins and Judge (2010) values repre- customer, community, and corporate. Moreover,
sent basic convictions that “a specific mode of conduct employees are encouraged to believe in three critical
or end-state of existence is personally or socially pref- values: commitment, consistency, and communication.
erable to an opposite or converse mode of conduct or A learning camp which incorporates the concept of
end-state of existence.” Values lay the foundation for constructionism is implemented. Constructionism is
our understanding of people’s attitudes and motivation defined by the organization as a learning method
and influence our perceptions and behavior. Hence, it where learners determine what they want to learn and
can be said that one’s values are determined by one’s how they want to go about it. Learners create new
belief and one’s values determine one’s attitudes which knowledge by building on to their old or current
at the end determine one’s behavior. knowledge. They reflect and share. They learn the con-
tent but, most importantly, they learn how to learn.
Theoretical Background Moreover, the company implements a team-learning
In the study Implementing Change Practice through activity where learners are grouped to work on selected
Learning and Development: A Case Study of Kaeng projects under the guidance of facilitators.
Khoi Cement Plant (SKK), Siam Cement Group, Team Development activities focus on creating a
Thailand (Utsahajit 2009), the company begins its sense of excellence, trust, and collaboration among
change primarily because of external pressure. As the employees. The organization strongly believes changes
business competition becomes more severe and a few become successful challenges when employees embrace
international big players in cement industry have excellent quality, communicate truthfully among one
shown their interest in expanding their current busi- another, and are willing to do everything possible to
ness and investing new businesses in Thailand, the achieve mutual goals. Team Building is one example of
Siam Cement Group has decided to commit in an Team Development activities where both indoor and
extensive change practice to level up their organiza- outdoor learning activities are effectively implemented.
tional performance. Environmental Improvement activities focus on
Change practice can be grouped into three catego- bringing changes into solid, visualized evidence. The
ries (as shown in Fig. 1), namely, Employee Perception, activities entail improvement both in terms of physical
Team Development, and Environmental Improvement. environment and of work atmosphere.
Employee Perception activities focus on aligning Additionally, change practice cannot be made
employees’ perception toward changes in organization. successful without well-designed support strategies.
The activities devote to continuously learning together Goodstein and Burke (2000) suggested methods of
through hands-on experience, both mentally and phys- implementing a change include individual change
ically, creating the readiness for change in employees strategy (e.g., setting up a comprehensive training pro-
by promoting the attitude of accepting changes as gram), technostructural strategy (e.g., modifying the
challenges and pathways to success in three levels: structure, individuals’ jobs, and/or work procedures),
528 C Change of Values Through Learning in Organizations
data-based strategy (e.g., conducting a companywide change must be positive and they must recognize that
survey to assess organizational culture for the purpose change is good, essential and attainable, first and fore-
of using the data to pinpoint required changes), and most. Then when they are open to change and feel
organization development strategy (e.g., collecting ready, people can be put into work group and team
information from organizational members about their building process can begin. They are, therefore, in
views regarding what needs to be changed and acting a stage of readiness to learn and develop. People behave
accordingly). All of these strategies are evident in the and act with trust. They feel comfortable admitting
change practice at Kaeng Khoi Cement Plant. For their ignorance, reflecting and sharing their knowl-
individual change strategy, the company develops a edge and feeling. Finally, the third component of the
comprehensive learning and development scheme change practice can be realized. Improvements are then
using various activities. For technostructural strategy, felt and seen around the plant. Figure 2 shows how
the company restructures the organizational hierarchy employees’ behaviors develop.
to be flatter and less centralized. For data-based strategy, From Fig. 2, as employees’ perceptions toward
the company deploys the organization-wide communi- change become appropriate and healthy, they feel
cation campaign and provides various communication more confident and ready for change. Then they are
channels for information to flow upwardly, down- developed individually and collectively through a
wardly, and laterally. Finally, for organization develop- series of team-learning activities. They become aware
ment strategy, the company undergoes many activities of themselves and others, trust other group members,
to ensure the involvement of people in voicing their and willing to share with and learn from one another.
opinions and valuable ideas regarding change practice. Finally project-based activities are assigned to the
Change practice at Kang Khoi Cement Plant is employees at the right time. Successful results are
a planned program involving the whole system and obtained and each team member feels good about the
relying on many experience-based learning activities, outcomes. This, in return, solidifies the right attitude of
and the focus is on group behavior and team develop- employees toward change, increases their confidence
ment. The company believes that before change can and readiness for change, raises the awareness of the
take place people in the organization have to first importance of working in teams to achieve the ultimate
perceive change. Their value and attitude toward goal, and promotes a trusting learning atmosphere.
Theoretical Background
Character Education The capacity to obtain knowledge from others, rather
than exclusively relying upon what is observed or
A theory of prosocial development focused on the experienced directly, offers important opportunities
teaching of virtue through modeling, direct instruc- for learning. This capacity has long been recognized
tion, and practice. in developmental psychology and has been a major
focus of ▶ sociocultural theory. For decades, cognitive
Cross-References developmental psychologists have searched for the
▶ Moral Learning earliest evidence of children’s ability to make use of
▶ Video Games for Prosocial Learning information they obtain from others. Research in this
area has shown that by age 1, children are capable of
using the emotional responses of caregivers to guide
their judgments about which objects or people are to
Child Development be avoided.
More recently, there has been greater interest in
▶ Infant Learning and Development what happens in the years following infancy. During
this time, children’s language production and com-
prehension skills improve, giving them increased
opportunities to learn from others, and they begin to
develop cognitive skills that allow them to evaluate
Child-Centered Teaching what they have learned more effectively. A central
▶ Learner-Centered Teaching assumption of this research is that because information
that is obtained from others is not always accurate, it is
important for children to critically assess what others
tell them, and that without such an ability they are
vulnerable to being misinformed and manipulated.
Children’s Critical Assessment The primary focus of recent work in this area con-
of the Reliability of Others cerns how children evaluate the credibility of specific
informants. Much of this work has involved showing
GAIL D. HEYMAN children pairs of potential informants who differ on
Department of Psychology, University of California, a key dimension, and measuring which informant chil-
San Diego, La Jolla, CA, USA dren prefer. A standard paradigm involves a training
phase in which young children are presented with
familiar objects such as a ball and a cup, and then
Synonyms observe one informant providing accurate names for
Credibility judgments; Evaluation of testimony; the objects and another informant providing inaccu-
Selective trust; Skepticism rate names. Then, during a test phase, the two infor-
mants identify a series of novel objects using different
Definition novel labels such as “mido” and “loma.” Participants
Children’s critical assessment of the reliability of others are asked which name is most likely to be accurate and
refers to the ability of children to evaluate the extent which informant would be most likely to provide
to which specific individuals are reliable sources of accurate information in the future. Results indicate
Children’s Critical Assessment of the Reliability of Others C 531
that 3–4-year-olds consider informants with a history relating to scientific or supernatural explanations of
of being accurate to be more reliable than those with the natural world. Findings suggest that children are
a history of being inaccurate (Harris 2007). capable of applying different systems of beliefs to
different contexts, and that they often make creative
Important Scientific Research and attempts to merge different frameworks of beliefs in an C
Open Questions effort to main coherence and consistency. For example,
One reason researchers have sought to determine Legare and Gelman (2008) found that South African
how children learn to think critically about the reliabil- children often explained AIDS in ways that integrated
ity of others is that this understanding is thought to be their beliefs about witchcraft with scientific explana-
closely linked to developing conceptions of mental tions about the nature of the disease.
life. Recent research has provided evidence of such It will be important for future researchers to inves-
an association, including direct links between source tigate how children think critically about the informa-
evaluation and tests of mental state understanding tion they obtain from others in real-world contexts that
(Vanderbilt et al. in press). Further evidence comes have significant implications for their well-being, such
from demonstrations that before children reach age 5, as when deciding whether to disclose personal infor-
they are sensitive to a wide range of cues that can serve mation to individuals they meet online. Another key
as indicators of an informant’s knowledge. For exam- area for future research is to understand how children’s
ple, they consider the extent to which informants ability and willingness to engage in critical thinking is
have access to relevant information, and expect indi- influenced by their desires and emotions. Finally, more
viduals who create objects to be more reliable infor- research is needed concerning the types of experiences
mants about the objects than are other individuals (see that foster critical thinking most effectively. This work
Heyman 2008). should lead to insights into how to help children max-
By age 4, children have some understanding that imize the benefits of learning from others, while min-
people are not always motivated to convey what imizing the risks.
they know accurately. However, this does not mean
that they are generally successful at recognizing and
acting upon such motives. For example, 4-year-olds Cross-References
will often accept and act upon the advice of individuals ▶ Belief Formation
who they have repeatedly observed trying to deceive ▶ Collaborative Learning and Critical Thinking
others (Vanderbilt et al. in press). Even 6-year-olds ▶ Scaffolding Learning
have difficulty anticipating the potential effects of ▶ Social-Cognitive Influences on Learning
motives that relate to social desirability (Heyman ▶ Socio-Cultural Learning
2008), and understanding the ways in which judg- ▶ Vygotsky’s Philosophy of Learning
ments can be biased due to personal relationships
(Mills and Keil 2008). References
There has been considerable interest in the effec- Frazier, B. N., Gelman, S. A., & Wellman, H. M. (2009). Preschoolers’
tiveness of children’s efforts to seek out information search for explanatory information within adult-child conversa-
from others. This work has shown that when children tion. Child Development, 80, 1592–1611.
as young as age 4 are actively seeking explanations they Harris, P. L. (2007). Trust. Developmental Science, 10, 135–138.
Heyman, G. D. (2008). Children’s critical thinking when learning
are often able to evaluate the adequacy of the responses from others. Current Directions in Psychological Science, 17,
they obtain, and may repeat their questions or devise 344–347.
their own explanations if the answers they receive are Legare, C. H., & Gelman, S. A. (2008). Bewitchment, biology,
not satisfactory (Frazier et al. 2009). However, children or both: the co-existence of natural and supernatural explana-
of this age often have substantial difficulty with gener- tory frameworks across development. Cognitive Science, 32,
607–642.
ating effective questions.
Mills, C. M., & Keil, F. C. (2008). Children’s developing notions of
Another emerging research area concerns children’s (im)partiality. Cognition, 107, 528–551.
use of information obtained from others to construct Vanderbilt, K. E., Liu, D., & Heyman, G. D. (in press). The develop-
more elaborated systems of beliefs, including those ment of distrust. Child Development.
532 C Children’s Learning from Television
Long-term effects. Huston et al.’s (2001) early learn- instruction 3 years later, when they subsequently
ing model focuses on the long-term effects of educa- entered first or second grade. Moreover, in the lon-
tional media, and how such media might interact with gest-term study to date, even high school students
all of the other influences in children’s lives. Under this who had watched more educational television – and
model, three facets of early development are proposed Sesame Street in particular – as preschoolers had sig- C
as pathways by which long-term effects can result: nificantly higher grades in English, Mathematics, and
(1) learning preacademic skills, particularly related Science in junior high or high school. They also used
to language and literacy, (2) developing motivation books more often, showed higher academic self-
and interest, and (3) acquiring behavioral patterns of esteem, and placed a higher value on academic
attentiveness, concentration, nonaggressiveness, and performance. (See Fisch and Truglio 2001 for a review
absence of restlessness or distractibility. These factors of these and other studies.)
contribute to early success in school, which then plays Beyond this powerful evidence for the educational
a significant role in determining children’s long-term effectiveness of Sesame Street, numerous other studies
academic trajectories (e.g., placement in higher ability show that Sesame Street is not alone in helping children
groups, more attention from teachers, greater motiva- learn. Summative studies on other educational series
tion to do well). In addition, these early successes may for preschool and school-age children have shown that
also affect the types of activities in which children educational television can enhance children’s knowl-
choose to engage; for example, good readers may edge, skills, and attitudes in a wide variety of subject
choose to read more on their own. Each of these out- areas. These include effects of series such as Between the
comes can then result in further success over time. In Lions and The Electric Company on children’s language
this way, the model posits a cascading effect in which and literacy skills; Square One TV and Cyberchase on
early exposure to educational television leads to early children’s use of mathematics and problem solving;
academic success, which in turn, contributes to a long- 3-2-1 Contact and Bill Nye the Science Guy on children’s
term trajectory of success that can endure for years. understanding of science and technology; children’s
news programs on knowledge of current events; and
Important Scientific Research and preschool series such as Blue’s Clues and Barney and
Open Questions Friends on more general school readiness. Many other
Academic effects. Decades of research have demon- examples exist as well. (See Fisch 2004 for a review.)
strated clearly that both preschool and school-age chil- Prosocial effects. Parallel to the academic effects of
dren learn from educational television series. Perhaps educational television, numerous studies have found
the most prominent – and certainly the most exten- that viewing prosocial television programs produces
sively researched – example of an educationally effec- significant positive changes in children’s social behav-
tive television series is Sesame Street. A number of ior. Such effects have been documented as increases in
major summative research studies have examined several domains: “friendliness” and positive interac-
both immediate and long-term effects of Sesame Street tions in general, altruism and cooperation, self-control
on its viewers. Together, these studies demonstrate that and delay of gratification, and reduction of stereotypes.
extended viewing of Sesame Street produces significant Most of this research has been conducted with pre-
immediate effects on a wide range of academic skills school children, so the bulk of the evidence to date
among preschool children (e.g., knowledge of the relates to this age group. However, some research on
alphabet, vocabulary size, letter–word knowledge, stereotypes has been conducted with older children as
math skills, sorting and classification, knowledge of well. (See Mares and Woodard 2001 for a review.)
shapes and body parts, relational terms, time spent Nevertheless, the impact of televised prosocial mes-
reading and in educational activities, telling connected sages is likely to be mediated by lessons learned from
stories when pretending to read). In addition, several family and peers, as well as children’s own life experi-
longitudinal studies have found long-term effects as ences. In some cases, these experiences may work hand-
well; for example, preschool viewers of Sesame Street in-hand with the prosocial lessons shown on-screen. In
were found to be more likely to read storybooks other cases, however, the messages from these various
on their own and less likely to require remedial reading sources may conflict with each other. For example,
534 C Choice Reaction Time and Learning
through the rapid identification and differential hold true and was affected by the specific design of CRT
responding to multiple stimuli. The dependent vari- task used.
ables are the reaction time and accuracy in making Over the last century, not only have CRT tests
a correct choice to different, paired or multiple stimuli, evolved to probe very specific psychological phenom-
which may be presented in either simultaneous or ena that underlie human and animal behavior, but they C
sequential mode. This contrasts with simple reaction have been utilized to examine psychomotor and atten-
time tasks where a single response is made to a single tional function following different manipulations of
stimulus, with no decision or choice required of the normal physiological state. Such manipulations
the subject. Originally designed for people, CRT tasks have included dehydration, stress, and hypoglycaemia,
are now widely used in animal research as probes of as well as examining the psychomotor effects of drugs
focused and spatial attention, vigilance, neglect, and and toxic substances. CRT tests are also widely used to
psychomotor learning, and are used primarily as probe neurological conditions including depression,
probes of forebrain function. The present entry will schizophrenia, Huntington’s disease, attention deficit
focus on the use of these tests in rodents. hyperactivity disorder, Parkinson’s disease and brain
injury, and related treatment strategies, including phar-
Theoretical Background macological and cell-based interventions. In order to
CRT tasks were first developed by Franciscus Cornelis probe the defining deficits of such disorders, CRT tasks
Donders in the nineteenth century to assess psycho- have been developed to assess very precise behavioral
motor function (For review; Smith 1968). Three tasks phenomena and the anatomical pathways and regions
were originally developed to dissect the different psy- of the brain that subserve them.
chological processes involved in responding to a In its most common form, CRT tasks present stim-
specific stimulus in a choice paradigm, based on uli in a series, and the number of stimuli utilized can be
a predefined rule. The first was a simple reaction anything from two choices upward, as can the number
time task – where the participant had to make a of possible responses. In the most basic paradigm, there
response to the appearance of one stimulus. The second are two stimuli and two responses. However, it is pos-
was what we now term a go/no-go task (sometimes sible to have more stimuli than responses or visa versa.
referred to as a recognition reaction time task) – A further adaptation to the CRT paradigm are serial
where two different stimuli were presented indepen- reaction time tasks. In a serial reaction time task trials
dently and the participant had to respond in a set are not presented as discrete trials, but instead admin-
manner to one stimulus but refrain from responding istered as a continuous stream of stimuli, e.g., brief
to the second stimulus. The final was the development pulses of light presented in different locations, and
of a 2-choice reaction time task – where two separate the subjects are required to respond to the correct
stimuli were presented independently, with a different stimulus location by pushing a response button, touch-
response required for each. By comparing the reaction ing a touch-screen, or (for rodents) nose poking into
times achieved on these three tasks, Donders developed a hole, or pressing a lever as rapidly as possible. The
a mathematical procedure termed the subtraction number of locations used can be varied and randomly
method, which worked out the time taken to categorize presented, thus introducing a spatial aspect to what is
a stimulus and select an appropriate response. Follow- essentially a vigilance task, in which participants have
ing Donders work, research focused on theories of to monitor the light array continuously for the appear-
CRT performance using modifications of the two ance of the stimuli, and respond appropriately. This
CRT task; hypothesizing on how stimuli were inter- type of serial reaction time task can further be adapted
nally represented and categorized, and how correct to include sequences, either overt in a sequence learn-
responses were selected in order to perform such tasks ing task, or covert where predictable sequences are
(For review; Smith 1968). One example of this is Hick’s embedded within apparently randomized stimulus
law to determine the speed of CRT when an increasing presentations. This covert use of sequences is designed
number of stimuli are used. Hick (1952) stated that to probe implicit learning in tasks such as the rodent
CRT increased logarithmically with the number of serial implicit learning task (SILT). Another variation
choices of stimuli. However, this law did not always to the basic 2-CRT task, which is also used to examine
536 C Choice Reaction Time and Learning
attention, is the continuous performance task. During with the food magazine. These errors indicate patho-
this task, subjects must respond to just one target logical processes related to specific neural substrates
stimulus within a stream of different stimuli, and all (For review: Robbins 2002).
the other irrelevant stimuli must be ignored. This reac- A different type of CRT is the 2-CRT task for
tion time task is particularly sensitive to perseverative rodents, commonly known as the “Carli” task, which
and disinhibitory changes. was also developed to probe lateralized responding by
Unlike in humans, where the rules for responding rats in the 9-hole operant box (Carli et al. 1985). This
are explicitly explained to the participant before task has been used to assess motor function, sensory
performing the test, animals have to be first trained neglect, and the ability to initiate movements in uni-
on the particular S-R associations to be tested. There- lateral lesion models of neurological disorders. The
fore, in this category of tests, not only is CRTexamined, task is often used to assess unilaterally applied thera-
but also associative learning and habit formation. peutic interventions such as cell or neuroprotective
In rodents, CRT tasks run in the 9-hole box operant gene therapies in lesion models of neurological disor-
chamber or standard 2-lever “Skinner box” operant ders, including the excitotoxic model of Huntington’s
chambers depending on the design employed. In rats disease, the dopamine depletion model of Parkinson’s
and mice, the most utilized of these tasks is the 5-choice disease, or unilateral middle cerebral artery occlusion
serial reaction time task. This task is performed in the as a model of ischemic stroke. In this task, only the
9-hole operant chamber, as developed by Robbins central three holes of the 9-hole box are used. Rodents
and colleagues (For review: Robbins 2002). The animal are trained to make a sustained nose poke into the
must respond to light stimuli presented randomly center hole for a variable duration prior to a brief
across a horizontal 5-hole array, with each correct presentation of the stimulus light in either the left or
response resulting in the presentation of a reward. right hole, to which the animal must respond. The
This test paradigm is regularly employed to assess the dependent variables on this task are reaction time
effects of drugs, or lesions on attentional performance (time to withdraw from the center hole), which is
of animals, and increasingly transgenic animals are a measure of the time required to detect the stimulus
being probed with this and other CRT tasks. The and initiate a motor response, and movement time
5-choice serial reaction time task provides measures (time to move from center hole to response hole)
of reaction time, number of correct, premature, missed as a measure of motor function, as well as task accuracy.
(errors of omission), and incorrect responses (errors There are two versions of the task (“SAME” and
of commission). When using this task, a number of “OPPOSITE”), which require the animal to respond
probes can be introduced to assess attentional function, either in the hole where the light was presented, or in
including randomizing the stimulus lengths, random the unlit hole, respectively. With the two versions of the
intertrial intervals (time between trials) and bursts of task, it is possible to determine whether the deficits
white noise. Alternative versions of the task have been present in a unilateral animal model are due to sensory
designed in which different numbers of stimuli are neglect or deficits in the ability to initiate movement.
used. It is also possible to examine the effects of other Different theoretical explanations of the functional
manipulations, such as brain lesions to uncover the processes underlying correct task performance –
neural basis of attentional processing. In both animal sensory, sensorimotor, or motor – make quite different
and human studies, analysis of the error types provides predictions about the side on which a deficit will be
a detailed description of the functional neuropathology observed in the two tasks in animals with unilateral
of the individual, so whereas a reduction in accuracy lesions. The conventional “Carli” task has been further
may be demonstrated in a particular animal group, adapted to analyze discrimination between different
analysis of the error terms can provide a detailed choice response holes separately in ipsilateral (same
description of why those errors are occurring, for side as the brain lesion) and contralateral (opposite
example, the animal may be simply making the wrong side to brain lesion) space, which then allows analysis
choice, responding prematurely, perseverating in the of whether lateralized deficits are related to egocen-
previously correct hole, or may even become fixated tric (mapped by internal cues) or allocentric space
Choreographies of School Learning C 537
novice and expert and adapts his lessons accordingly. that for every important learning area, sequences of
Like a sports coach, she likewise has to know the the process can be described. The order of sequences
epistemological obstacles to advance in the learning can be normatively fixed. The right organization of this
process. Besides his elaborated knowledge of the deep structure shall be the determining sign for quality
subject, he must know the developmental steps of of learning. A BM describes the learning sequences
learning, the meta-cognitive knowledge about the epis- in regard to certain learning goals in a certain domain.
temology of knowledge and the teaching skills. What is It consists of those concatenations of operations or
the crucial point to be able to progress? The teacher, as groups of operation, which are somehow necessary
a choreographer, is coaching the process from the nov- for every learner and cannot be replaced by anything
ice to the expert. else (Oser 1993). Such learning scripts as concatena-
tions of operations can be viewed in two ways: They can
Theoretical Background be described as phenomena; for instance, when chil-
Oser (Oser and Patry 1990; Oser et al. 1997; Oser and dren construct a concept they proceed in such and
Baeriswyl 2001; Oser 2006) developed the theory of such a way. Or, one can ask how teachers and children
choreographies of teaching. His main hypothesis is subjectively imagine such scripts. Both approaches
that the “very sequence of (school) learning is based complement one another (Oser and Baeriswyl 2001).
on a choreography that binds, on the one side, freedom Twelve BM were developed altogether (Oser and
of method, choice of social form and situated impro- Patry 1990):
visation with, on the other side, the relatively rigor of
1a. Learning through personal experience
the steps that are absolutely necessary in inner learning
1b. Discovery learning
activity (Entwistle 2000; Charness et al. 2005; Hattie
2. Development as an aim of education
2009). Such an hypothesis requires a double operatio-
3. Problem solving
nalization: Firstly, in view of the relationship between
4a. Meaning building
the basis-models and the visible structure, and sec-
4b. Concept building
ondly, in view of the rule-bound character already
5. Contemplative learning
referred to on one hand and freedom to stimulate on
6. Learning of strategies
the other” (Oser and Baeriswyl 2001, p. 1043).
7. Routines and skills
The concept of basis models (BM) is, first of all,
8. Motility
based on the differentiation between surface structure
9. Social learning
and deep structure of teaching and, secondly, based on
10. Construction of values and value identity
the assumptions that the learning process precedes
11. Hypertext learning
goals and is domain specific. If the learning goal is
12. Learning to negotiate
to build up certain values or attitudes, for example,
in law or economics, the learning process must be Each one of these models contains a defined deep
choreographed differently, than if the goal is to build structure of certain elements, which are chained
up conceptual knowledge. The surface structure together.
includes all teaching methods (lectures, project learn- For example, model 1a consists of the following
ing, case studies, problem-based learning, anchored elements:
instruction, etc.), all social forms of learning (e.g.,
1. Anticipating and planning possible actions
individual work, partner work, group work), all
2. Performance of the action
media, and media-based teaching forms. The surface
3. Constructing the meaning for the activity
or visible structure of lessons is directly observable.
4. Generalizing the experience
The surface structure of a lesson is not a major indica-
5. Reflecting similar experiences
tor of learning and teaching quality.
The deep structure refers to the learning process as A situated learning unit contains several basis
a psychological process. It constitutes a construct and is models, which are intercalated.
therefore not directly observable. Oser has assumed Figure 1 shows an example of such an intercalation.
Choreographies of School Learning C 539
the responsibility of the teacher, but the learner has to Oser, F. K., & Baeriswyl, F. J. (2001). Choreographies of teaching:
recognize it as a principle. The goal is that every student Bridging instruction to learning. In V. Richardson (Ed.), Hand-
book of research on teaching (4th ed.). Washington: American
understands his or her learning as a planned act for
Educational Research Association.
which he takes the necessary responsibility. The strict Oser, F., & Patry, J.-L. (1990). Choreographien unterrichtlichen
sequencing and chaining of learning steps, as theorized, Lernens. Basismodelle des Unterrichts. Berichte zur Erziehungs-
is probably not really necessary. The elements are impor- wissenschaft, Nr. 89. Pädagogisches Institut der Universität
tant, and should be present (see also Bereiter and Freiburg (Schweiz).
Oser, F., Patry, J.-L., Elsässer, T., Sarasin, S., & Wagner, B. (1997).
Scardamalia 2006). But the human mind has sufficient
Choreographien unterrichtlichen Lernens. Schlussbericht an
flexibility and does not require a strictly followed den Schweizerischen Nationalfonds zur Förderung der
sequence of learning steps in order to learn successfully. wissenschaftlichen Forschung. Projekt 1113-042353.94/1. Bern.
Resnick, L. B. (1991). Shared cognition: Thinking as social practice.
In L. B. Resnick, J. M. Levin, & S. D. Teasley (Eds.), Perspectives on
Cross-References socially shared cognition (pp. 1–20). Washington, DC: American
▶ Didactics (Didactic Models) and Learning Psychological Association.
▶ Learning Strategies Seidel, T. (2003). Lehr-Lernskripts im Unterricht. Münster: Waxmann.
▶ Teaching Methods Seidel, T., & Shavelson, R. (2007). Teaching effectiveness research in
the past decade: The role of theory and research design in
disentangling meta-analysis results. Review of Educational
References Research, 77(4), 454–499.
Bauch-Schremmer, C. H. (1993). Untersuchungen zu den Wagner, B. (1999). Lernen aus der Sicht der Lernenden. Frankfurt
Choreographien unterrichtlichen Lernens bei Oser – Über die a Main: Peter Lang.
Kombinierbarkeit der Basismodelle. Wissenschaftliche (interne)
Arbeit an der Pädagogischen Hochschule Ludwigsburg.
Bereiter, C., & Scardamalia, M. (2006). Education for the Knowledge
Age: Design-Centered Models of Teaching and Instruction. In
P. A. Alexander & P. H. Winne (Eds.), Handbook of Educational Choreographies of Teaching
Psychology (pp. 695–713). Maywah, NJ: Lawrence Erlbaum
Associates. ▶ Choreographies of School Learning
Charness, N., Tuffiash, M., Krampe, R., Reingold, E. & Vasyukova, E.
(2005). The Role of Deliberate Practice in Chess Expertise.
Applied Cognitive Psychology, 19, 151–165.
Entwistle, N. (2000). Promoting deep learning through teaching and
assessment: conceptual frameworks and educational contexts. CHREST
Paper presented at the TLRP Conference, Leicester, UK.
Retrieved March 17, 2011, from http://www.tlrp.org/acadpub/ A cognitive architecture, developed by Fernand Gobet
Entwistle2000.pdf and Peter Lane, emphasizing a close interaction
Fischler, H., Schröder, H.-J., Tonhäuser, C., & Zedler, P. (2002).
between perception, learning, and memory. It proposes
Unterrichtsskripts und Lehrerexpertise: Bedingungen ihrer
Modifikation. Zeitschrift für Pädagogik, 45, 157–172.
that human cognition is constrained by a number of
Haenni, S. (1996). Das Motilitätsmodell. Dissertation am limitations, such as span of attention and capacity of
Pädagogischen Institut der Universität Freiburg/Schweiz. short-memory. Learning, which to some extent miti-
Teildruck. Freiburg. gates the limits imposed by bounded rationality, is
Hattie, J. A. C. (2009). Visible Learning. A synthesis of over 800 meta- done through the acquisition of chunks and templates.
analyses relating to achievement. New York: Routledge.
CHREST stands for Chunk Hierarchy and REtrieval
Lave, J., & Wenger, E. (1991). Situated learning. Legitimate peripheral
participation. Cambridge: Cambridge University Press. STructures.
Oser, F. (1993). Sichtstruktur und Basismodelle des Unterrichts:
Über den Zusammenhang von Lehren und Lernen unter dem
Gesichtspunkt psychologischer Lernverläufe. Unterlagen zum
gleichnamigen Vortrag auf der 49. Tagung der AEPF in Wien.
Oser, F. (2006). Das Lob der Unvollendetheit: Hans Aebli’s Glaube an Chunks
operative Veränderung. In M. Baer, M. Fuchs, P. Füglister,
K. Reusser, & H. Wyss, (Eds.), Didaktik auf psychologischer A meaningful unit of information built from smaller
Grundlage. Von Hans Aeblis kognitionspsychologischer Didaktik pieces of information. Chunks consist of several items
zur modernen Lehr – und Lernforschung. Bern: h.e.p. of information that have been learned and stored as
Chunking Mechanisms and Learning C 541
a unit in long-term memory such as BMW, KGB, and occurs during perception. Here, we talk about percep-
USA. George Miller proposed that short-term memory tual chunking.
can hold 7 þ/ 2 chunks.
Theoretical Background
Cross-References Chunking as a mechanism was initially proposed by C
De Groot (1946/1978) in his study of chess experts’
▶ Video-Based Learning
perception, memory, and problem solving, to explain
their ability to recall briefly presented positions with
References a high level of precision. It was also a central ingredient
Miller, G. A. (1956). The magical number seven, plus or minus two: of Miller’s (1956) classical article about the limits on
Some limits on our capacity for processing information. Psycho- human information-processing capacity. Miller pro-
logical Review, 63(2), 81–97. posed that chunks are the correct measure for the
information in the human cognitive system, and that
7 2 chunks can be held in short-term memory. Chase
and Simon (1973) proposed a general theory of pro-
cesses underpinning chunking. It is interesting to note
Chunking
that the approaches of De Groot as well as Chase
▶ Deductive Learning and Simon emphasize the implicit nature of chunks,
▶ Restructuring in Learning which are seen as the product of automatic learning
processing sometimes called perceptual chunking.
Miller’s view emphasizes a type of strategic, goal-
oriented chunking, where chunking is essentially re-
coding of the information in a more efficient way. For
Chunking Mechanisms and example, the 9-digit binary number 101000111 can be
Learning re-coded as the 3-digit decimal number 327, making it
easier to process and memorize for humans. The pres-
FERNAND GOBET1, PETER C. R. LANE2 ence of chunks explains how humans, in spite of
1
Department of Psychology, School of Social Sciences, strict cognitive limitations in memory capacity, atten-
Centre for the Study of Expertise, Brunel University, tion, and learning rate, can cope efficiently with the
Uxbridge, Middlesex, UK demands of the environment. Chunking has been
2
School of Computer Science, University of established as one of the key mechanisms of human
Hertfordshire, Hatfield, Hertfordshire, UK cognition and plays an important role in showing
how internal cognitive processes are linked to the exter-
nal environment.
Definition There is considerable empirical evidence supporting
A ▶ chunk is a meaningful unit of information built the notion of a chunk, for example, in our ability to
from smaller pieces of information, and ▶ chunking is perceive words, sentences, or even paragraphs as single
the process of creating a new chunk. Thus, a chunk can units, bypassing their representation as collections of
be seen as a collection of elements that have strong letters or phonemes; this explains, for example, how
associations with one another, but weak associations skilled readers may be insensitive to word repetition or
with elements belonging to other chunks. Chunks, deletion. Particularly strong evidence is found in those
which can be of different sizes, are used by memory studies that use information about the timing of
systems and more generally by the cognitive system. responses to infer the presence of chunks. The use of
Within this broad definition, two further meanings response times assumes that the output of elements
can be differentiated. First, chunking can be seen as within a chunk will be faster than the output of elements
a deliberate, conscious process. Here, we talk about across different chunks. This is because the elements
goal-oriented chunking. Second, chunking can be seen within a chunk belong to the same structure, as well as
as a more automatic and continuous process that sharing a number of relations. There is good empirical
542 C Chunking Mechanisms and Learning
between chunks. For example, the most direct expla- A first implication of chunk-based theories is that
nation for observing a set of actions as a chunk is for acquiring a new chunk has a time cost, and therefore
the actions to be represented internally as a single unit, time at the task is essential, be it in mathematics or
i.e., a chunk, and so retrieved and output together. dancing. As documented by research into ▶ deliberate
However, it is also possible for a subject to plan practice, practice must be tailored to the goal of C
output actions ahead, and so either break long improving performance. Chunk-based theories give
sequences into subparts (e.g., to take a breath when attention a central role – see for example the CHREST
reciting the alphabet) or else compose short sequences model – and such theories are therefore suitable models
into what appear as longer ones (e.g., where a second of deliberate practice. In particular, conceptual knowl-
chunk begins naturally from where the first one fin- edge is built on perceptual skills, which in turn must
ished). Distinguishing between these types is only be anchored on concrete examples. Thus, curricula
possible with the aid of a computational model, where should provide means to acquire perceptual chunks in
the precise items of information known by the subject a given domain.
at a given point in time can be ascertained (Gobet et al. There are different useful ways to direct attention
2001). The advantage of using computer models is and to encourage the acquisition of perceptual chunks:
discussed in more detail in the entry on ▶ Learning in to segment the curriculum into natural components, of
the CHREST Cognitive Architecture, a model based on the right size and difficulty; to present these compo-
the template theory. nents with an optimal ordering and suitable feedback;
Chunk-based theories, such as the chunking and and to highlight the important features of a problem.
template theories, not only provide a powerful explana- If perceptual chunking is an important way of
tion of learning and expert behavior, but also offer useful storing knowledge, then a clear consequence is that
information as to how learning occurs in the classroom transfer will be difficult. Unfortunately for learners,
and how it could be improved (Gobet 2005). We briefly this prediction is correct, both for school knowledge
discuss some of the implications for education (further and more specific skills such as sports and arts. More
principles are listed in Table 1). than 100 years of research have established that trans-
fer is possible from one domain to another only when
the components of the skills required in each domain
Chunking Mechanisms and Learning. Table 1 overlap. Thus, it might be helpful to augment the
Educational principles derived from chunk-based theories teaching of specific knowledge with the teaching of
(After Gobet 2005) metaheuristics – including strategies about how to
● Teach from the simple to the complex learn, how to direct one’s attention, and how to mon-
itor and regulate one’s limited cognitive resources.
● Teach from the known to the unknown
As noted above, an important idea in Chase and
● The elements to be learned should be clearly identified Simon’s (1973) theory is that perceptual chunks can
● Use an “improving spiral,” where you come back to the be used as conditions to actions, thus leading to the
same concepts and ideas and add increasingly more acquisition of productions. Then, an important aspect
complex new information of education is to balance the acquisition of the condi-
● Focus on a limited number of types of standard tion and action parts of productions. Another impor-
problem situations, and teach the various methods in tant aspect of education is to favor the acquisition of
these situations thoroughly
templates (schemata). Templates are created when the
● Repetition is necessary. Go over the same material context offers both constant and variable information.
several times, using varying points of view and a wide As a consequence, and as is well established in the
range of examples
educational literature, it is essential to have variability
● At the beginning, do not encourage students to carry during learning if templates are to be created.
out their own analysis of well-known problem situations,
Finally, chunk-based theories are fairly open to the
as they do not possess the key concepts yet
possibility of large individual differences in people’s
● Encourage students to find a balance between rote
cognitive abilities. In particular, while they postulate
learning and understanding
fixed parameters for short-term memory capacity and
544 C Chunking Theory
learning rates, it is plausible that these parameters Gobet, F., & Simon, H. A. (1996). Templates in chess memory:
vary between individuals. In addition, differences in A mechanism for recalling several boards. Cognitive Psychology,
31, 1–40.
knowledge will lead to individual differences in per-
Miller, G. A. (1956). The magical number seven, plus or minus two:
formance. A clear prediction of chunk-based theories Some limits on our capacity for processing information. Psycho-
is that individual differences play a large role in the logical Review, 63, 81–97.
early stages of learning, as is typical of classroom
instruction, but tend to be less important after large
amounts of knowledge have been acquired through
practice and study. Chunking Theory
Important Scientific Research and Theory developed by Chase and Simon in 1973,
Open Questions explaining how experts circumvent the limitations of
Chunk-based theories have spurred vigorous research cognitive processes through the acquisition of domain-
in several aspects of learning and expertise. A first specific knowledge, in particular, small meaningful
aspect is the acquisition of language, where recent units of interconnected elements (chunks).
research has shown that chunking plays an important
role in the development of vocabulary and syntactic
structures. A second aspect is related to the neurobio-
logical basis of chunking. Recent results indicate that Circumscribed Interests
perceptual chunks are stored in the temporal lobe,
and in particular the parahippocampal gyrus and fusi- Circumscribed interests are a child’s narrow preoccu-
form gyrus. pations or ritualistic activity that is unusually intense in
Other issues being currently researched include the terms of their focus. Circumscribed interests often are
effect of order in learning, and in particular how characterized by difficulty removing the individual
curricula can be designed so that they optimize the from engagement with the interest, high intensity of
transmission of knowledge. A possible avenue for focus, and long duration of fascination and engage-
future research is the design of computer tutors that ment with the interest. Circumscribed interests have
use chunking principles for teaching various materials, been embedded in activities as a basis for promoting
optimizing instruction for the abilities and level of each a child’s participation and use of other behaviors (e.g.,
student by providing personalized curricula, providing social interaction).
judicious feedback, and teaching strategies. Cross-References
▶ Interest-Based Child Participation in Everyday
Cross-References Learning Activities
▶ Bounded Rationality and Learning
▶ Deliberate Practice
▶ Development of Expertise
▶ Learning in the CHREST Cognitive Architecture Civilization, Archaic
▶ Schema
▶ Culture in Second Language Learning
References
Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive
Psychology, 4, 55–81.
De Groot, A. D. (1978). Thought and choice in chess (first Dutch edition
in 1946). The Hague: Mouton.
Classical Conditioning
Gobet, F. (2005). Chunking models of expertise: Implications for
The procedure where an initially neutral stimulus, such
education. Applied Cognitive Psychology, 19, 183–204.
Gobet, F., Lane, P. C. R., Croker, S., Cheng, P. C.-H., Jones, G., as a tone, is repeatedly paired with a biologically sig-
Oliver, I., & Pine, J. M. (2001). Chunking mechanisms in nificant stimulus, such as food. As a consequence, the
human learning. Trends in Cognitive Sciences, 5, 236–243. tone elicits a response that anticipates the food.
Classification of Learning Objects C 545
Three taxonomies are particularly interesting on the didactics aspects related to the LO.The LOs are
within the classification based upon formalized grouped into:
criteria.
● Receptive: The learner is simply the beneficiary of
The first is Wiley’s taxonomy (Wiley 2000), called
the contents. Usually the learner’s activity exploits
“Preliminary Taxonomy of Learning Object Types.” It
LOs of little size.
focuses above all on the structural aspect of the LO. It
● Internally interactive: There is interaction between
counts five kinds of LOs having eight characteristics:
user and computer. The LCMS or the models cre-
● Fundamental (i.e., a video of a hand typing on ated by the teacher guide the learner.
a keyboard) ● Cooperative: Containing brainstorming or prob-
● Combined–closed (i.e., a video of a hand typing on lem-solving sessions which require communicative
a keyboard with a background sound) activities among the students.
● Combined–open (i.e., a web page containing an
The two taxonomies considered so far are respec-
image and a file containing an animation with an
tively based on the relationships among the types and
interactive text)
the characteristics (the former), and the interaction
● Generative-presentation (i.e., a java applet able to
with the user (the latter) (Table 2).
generate an html page lay-out, or to show an
The Osel taxonomy (IJKLO 2006) implies the join-
editor with a correspondent code or to ask ques-
ing of the two taxonomies, whose result is the creation
tions to the learner)
of 15 different classes. Among these many cannot be
● Generative-instructional (i.e., an interface that
considered valid by the research group.
teaches how to play an instrument)
The types of LOs considered admissible in the OSEL
Whereas the characteristics are: Taxonomy are nine:
● Number of elements combined: The number of the
● B-simple: It is the derivation of the classifying
single elements (as video clips, images, etc.) which
combination of fundamental (Wiley) and recep-
constitute the LO.
tive (Redeker). It represents a noninteractive LO,
● Type of objects contained: The kinds of LOs that
made up of a single content constituted by a
can set up a new LO.
single element, or a simple media. Group activities
● Reusable component objects: It indicates if it is
are not allowed. For instance: a JPEG image or
possible to have access to the different components
a text.
of LOs in order to reuse them in other learning
● B-passive: The classifying combination of com-
contexts.
bined–closed (Wiley) and receptive (Redeker). It
● Common function: The basic use of an LO.
represents a noninteractive LO having a single con-
● Extra-object dependence: It indicates if the LO
tent made up of at least two internal elements
needs other information about other LOs (i.e., the
combined between them. Group activities are not
place on the web).
allowed. For instance: a JPEG image with textual
● Type of logic contained in object: It describes the
description.
function of the algorithms and the proceedings
● B-active: The classifying combination of combined–
contained in the LO.
open (Wiley) and receptive (Redeker). A
● Potential for inter-contextual reuse: It indicates the
noninteractive LO constituted by a single content
number of learning contexts in which the LO can be
made up of many internal and external elements
used, i.e., its potential to be reused.
combined among them. Group activities are not
● Potential for intra-contextual reuse: It highlights
allowed. For instance: a textual description connected
the times an LO can be reused within the same
to many JPEG images, among which at least one is on
area or domain (Table 1).
an http out of the platform.
The “Educational Taxonomy for Learning Objects” ● T-simple: The classifying combination of basic
by Redeker (Redeker 2003) is the second taxonomy on (Wiley) and internal interactive (Redeker). An
which the OSEL taxonomy is based. It focuses above all interactive LO constituted by at least two contents
Classification of Learning Objects C 547
Classification of Learning Objects. Table 1 Relationships among types and characteristics in Wiley’s classification
Fundamental Combined– Combined– Generative- Generative-
Learning object learning closed learning open learning presentation instructional learning
characteristic object object object learning object object
Number of elements One Few Many Few – many Few – many C
combined
Type of objects Single Single, All Single, Single, combined–
contained combined–closed combined–closed closed, generative-
presentation
Reusable (Not No Yes Yes/no Yes/No
component objects applicable)
Common function Exhibit, Pre-designed Pre-designed Exhibit, display Computer generated
display instruction or instruction and/ instruction and/or
practice or practice practice
Extra-object No No Yes Yes/No No
dependence
Type of logic (Not None, or answer None, or Domain-specific Domain-independent
contained in object applicable) sheet-based item domain-specific presentation presentation,
scoring instructional strategies instructional and
and assessment assessment strategies
strategies
Potential for inter- High Medium Low High High
contextual reuse
Potential for intra- Low Low Medium High High
contextual reuse
made up of a single element. Group activities are ● T-active: The classifying combination of combined–
not allowed. open (Wiley) and internal interactive (Redeker).
● T-passive: The classifying combination of com- An interactive LO constituted by many internal
bined–closed (Wiley) and internal interactive and external contents having many elements
(Redeker). An interactive LO made up of at least combined among them. Group activities are not
two internal contents made up of at least two ele- allowed.
ments combined between them. Group activities ● W-simple: The classifying combination of basic
are not allowed. (Wiley) and cooperative (Redeker). An interactive
548 C Classification of Levels of Intellectual Behavior in Learning
Cross-References
▶ Cognitive Models of Learning Synonyms
▶ Courseware Learning Classroom discipline
Classroom Management and Motivation C 549
ability to regulate motivation include structuring learning tasks, which in turn enhances the likelihood
opportunities for students to identify and provide that they will initiate and sustain engagement in
their own consequences for behavior; teaching students learning (Anderman & Leake 2005). A success cycle
goal-orientated self-talk strategies; subdividing and is established whereby successful learning leads to
teaching students how to subdivide task into smaller enhanced self-efficacy, which can therefore lead to C
chunks; and teaching students to adopt attributional increased motivation to engage in learning tasks and
control strategies that help them view engagement and therefore more opportunities for successful learning
potential success as within their own personal control (Brooks & Shell 2006). Classroom management prac-
(Brooks & Shell 2006). tices for enhancing these aspects of the motivational
The need for belonging serves as a second category context in a classroom environment include helping
under which to discuss constructs relating to motiva- students set attainable goals; teaching students to
tion (Anderman & Leake 2005). People have a psycho- adaptively attribute their successes and challenges
logical need for belonging or attachment to other with given tasks; providing students with realistic
human beings. From this perspective, motivation and immediate feedback that enhances self-efficacy;
is seen to be enhanced in classroom environments and providing learning opportunities and materials
where the classroom management plan has taken matched to students’ learning styles and strengths
into account specific approaches to building teacher– (Anderman & Leake 2005; Jones & Jones 2010).
student and peer relationships that are mutually
respectful and help learners feel connected to others Important Scientific Research and
in the environment (Anderman & Leake 2005; Jones & Open Questions
Jones 2010). These may include such approaches as There is a wide body of research on the relationships
holding regular class meetings; offering students ways among varying discrete concepts relating to motiva-
to express their opinions and feelings to each other tion, such as interest, goal orientation, self-efficacy,
and privately with the teacher; explicitly teaching and outcome expectancy, attributional orientations,
practicing the social skills necessary for successful cognitive engagement, intrinsic and extrinsic motiva-
learning and social interactions in the classroom; tion, locus of control, task value, and self-efficacy
and implementing systems of behavioral manage- (Anderman & Leake 2005; Keller 2008; Na et al.
ment that focus on engaging students in identifying 2010). However, there has been little systematic
prosocial behaviors for effective learning and reflect- research done that has yielded a comprehensive under-
ing on the outcomes of their own behavioral choices standing of how teachers can foster the development of
(Jones & Jones 2010). Attention to the recognition of motivation among particular students in specific class-
unintended bias, differential expectations, and plan- room environments (Na et al. 2010; Turner & Patrick
ning to ensure equal participation and inclusion also 2008). Turner and Patrick (2008) suggest that in order
support the development of a relational motivational to develop a research agenda that yields findings that
context that promotes engagement in learning tasks are useful to practitioners, the focus of research on
among students (Anderman & Leake 2005; Jones & motivation should be turned toward analysis of stu-
Jones 2010). dents’ participation in groups, in the context of how
Finally, competence is a category under which various groups construe tasks differently. Additionally,
many of the theoretical constructs relating to moti- while much research has focused on ways to increase
vation might be grouped, including those related to learner success through self-regulation, comparatively
expectancy beliefs, goal setting, attributions, self- little has been done with an explicit focus on the regu-
concept, and self-efficacy (Anderman & Leake 2005). lation of motivation (Brooks & Shell 2006). The class-
All human beings have some underlying need to room management structures teachers put in place may
feel that they can be capable and successful with promote or discourage the development among stu-
the tasks they undertake. Research has shown that dents of various self-regulatory tools for managing
when students feel competent, they feel more certain motivation; this is an area in which additional empir-
that they can be successful with a wider range of ical research is needed.
552 C Classroom Teaching and Learning
Cross-References
▶ Learning Motivation of Disadvantaged Students “Clever Hans”: Involuntary and
▶ Motivation Enhancement Unconscious Cueing
▶ Motivation, Volition and Performance
▶ Multifaceted Nature of Intrinsic Motivation NORBERT M. SEEL
▶ School Motivation Department of Education, University of Freiburg,
▶ Self-Regulation and Motivation Strategies Freiburg, Germany
▶ Understanding Intrinsic and Extrinsic Motivation
References Synonyms
Anderman, L. H., & Leake, V. S. (2005). The ABCs of motivation: An
Impulsive cueing; Instinctive cueing
alternative framework for teaching preservice teachers about
motivation. Clearing House: A Journal of Educational Strategies, Definition
Issues and Ideas, 78(5), 192. Cueing has different definitions in different contexts.
Brooks, D. W., & Shell, D. F. (2006). Working memory, motivation, Here, the definition is limited to the context of
and teacher-initiated learning. Journal of Science Education and
responding to externally provided stimuli. Cueing is
Technology, 15(1), 17–30.
Jones, V., & Jones, L. (2010). Comprehensive classroom management: another name for “foldback,” which is a process used
Creating communities of support and solving problems (9th ed.). to return a signal to a performer instantly. Cueing is
Upper Saddle River, NJ: Merrill. achieved via prompts, signals, hints or, more generally,
Keller, J. (2008). An integrative theory of motivation, volition, and cues, which include anything that is connected in some
performance. Technology, Instruction, Cognition & Learning, 6(2),
way to information to be processed and which prompts
79–104.
Na, L., Kang-hao, H., & Chun-hao, C. (2010). A cognitive-situative
its retrieval. This entry refers to the story of “Clever
approach to understand motivation: Implications to technology- Hans,” which can serve as a splendid example of invol-
supported education. US-China Education Review, 7(5), 26–33. untary and unconscious cueing.
Turner, J. C., & Patrick, H. (2008). How does motivation develop and
why does it change? Reframing motivation research. Educational
Psychologist, 43(3), 119–131.
Theoretical Background
Involuntary and unconscious cueing can be illustrated
by referring to the story of Clever Hans from the end
of the nineteenth century. Clever Hans was an Arab
stallion from Russia. His owner, Wilhelm von Osten,
Classroom Teaching and a retired schoolmaster, was convinced that animals
Learning possess an intelligence comparable to that of humans.
After many unsuccessful attempts to teach animals, he
▶ Didactics, Didactic Models and Learning found in Hans a partner for life. Von Osten taught
the horse to respond to questions requiring mathemat-
ical calculations by tapping his hoof. If Hans was
asked, for instance, what the sum of 3 plus 2 is, the
horse would tap his hoof five times. It appeared that the
Classroom-Based Knowledge horse was responding to human language and was
Construction capable of grasping mathematical concepts. In a short
time, Hans was able to work out reasonably complex
▶ Rapid Collaborative Knowledge Improvement
calculations, including some square roots. The horse
could also tell time and name people, but in the liter-
ature the focus is usually on his mathematical skills.
In the 1890s, von Osten began to showing his intel-
Classwide Peer Tutoring ligent horse to the public. Clever Hans and his owner
enjoyed worldwide acclaim, but the scientific commu-
▶ Reciprocal Learning nity remained skeptical. Clever Hans had been tested by
“Clever Hans”: Involuntary and Unconscious Cueing C 553
many people. He mastered each test successfully and ability to answer diminished even further. The second
the observers could not see any trickery. As a conse- major finding was that Hans could only answer cor-
quence, it was claimed that Hans had the intellectual rectly if the questioner also knew the answer to the
ability of a 14-year-old boy (Fig. 1). question. When the questioner did not know the answer
Clever Hans became a real sensation and people to the question, Hans could not find the answer. C
flocked to see his demonstrations when Professor Carl Based on these observations, the “Hans Commis-
Stumpf completed a first scientific testing in 1904 and sion” concluded that Hans was not using intelligence
certified Hans’s ability as genuine (Freund 1904). Other to work out the answers but was responding to visual
scientists, however, remained skeptical. Therefore, cues provided by the questioner or other present
Oskar Pfungst retested Hans in 1907, applying a more persons. Although the people interacting with Hans
rigorous test setting (Pfungst 1911/1998). A group of were not conscious of providing him with cues, the
13 scientists was assembled, known as the “Hans Com- horse was simply responding to muscle tensions, facial
mission.” Pfungst had the idea to separate Hans from expressions, and other involuntary cues produced in
his owner as well as from any other person. While one interacting with Hans. No evidence of cheating was
member of the team wrote down the numbers and left found.
the room, everyone else moved behind the blackboard. Thus, people were cueing Hans unconsciously by
Thus, only the horse could know what was inscribed tensing their muscles until he produced the correct
on the blackboard. Now Clever Hans failed every test. answer. The horse really was clever because he could
Pfungst concluded that when the correct answer was perceive and “interpret” very subtle muscle move-
not known to anyone present in the room, the horse ments. Although Hans could not process human lan-
did not know it either. More specifically, the test dem- guage as his owner maintained, he had an ability of
onstrated effects of involuntary and unconsciousness some kind to respond to involuntary and unconscious
cueing. It became apparent that Hans needed some cues in his environment. People can unconsciously
visual contact with the questioner in order to answer communicate information through subtle move-
correctly. The further away the questioner was, the less ments of muscles, and some animals can perceive
accurate Hans became, and when he was blinkered his these unconscious and involuntary cues.
“Clever Hans”: Involuntary and Unconscious Cueing. Fig. 1 Clever Hans in a test situation
554 C Climate of Learning
enthusiastic and quality instruction, and high expec- be the father of observational learning, believes there
tations for learning-related behavior and academic are four processes necessary for observational learning:
achievement. The climate of learning is a specific com- attention (children are attracted to high status,
ponent of school climate and school culture, which are same-sex models), retention (committing a behavior
a much broader set of factors that may influence stu- to memory), production (imitating the behavior), and C
dent achievement. motivation (the child must be motivated to replicate
the behavior). Bandura’s well-known 1964 Bobo doll
Theoretical Background study demonstrated the effects of observational learn-
In 1924, a group of researchers conducted a study on ing in young children.
the relationship between light intensity and employee Several learning theories exist, and have a direct
productivity at Hawthorne Works, a Western Electric effect on the climate of learning. B.F. Skinner, consid-
plant near Chicago, Illinois. Researchers increased and ered the father of behaviorism, demonstrated that
decreased light intensity and changed other factors of behaviors that are rewarded increase in frequency,
the workday, but the results were inconclusive. After while those that are punished decrease in frequency.
9 years of research and interviews with employees, Teachers shape the learning climate in accordance
investigators discovered that when workers felt valued with the theory or theories they find most compelling.
and understood, their productivity increased. This Cognitive learning theory focuses on how humans
finding is known as the “Hawthorne Effect,” and perceive, store, and remember information. Construc-
it illuminated the social and emotional influence of tivist learning, also known as discovery learning,
climate on human productivity and motivation encourages students to discover concepts and princi-
(Sonnenfeld 1985). ples through personal exploration and activation of
Consistent with the findings in the Hawthorne prior knowledge. Experiential learning, also referred
Study, learning cannot be separated from the social, to as service learning, emphasizes learning through
emotional, and physical factors that surround it. direct experiences. The Montessori approach to learn-
According to Dr. Urie Bronfenbrenner’s bioecological ing places special emphasis on individual development
perspective on human development, “Human beings levels, and encourages children to be self-directed,
create the environments that shape the course of cooperative learners. Students pursue their own aca-
human development (2004, p. 28).” Learning and devel- demic interests and complete work at their own pace.
opment occurs within an interconnected set of sys-
temic levels. The microsystem consists of people and Important Scientific Research and
places with whom the child has the most contact, such Open Questions
as family members at home, and teachers and staff at Since the National Commission on Excellence in
school. In the microsystem, learning experiences are Education published “A Nation at Risk” in 1983, the
bidirectional; both the learner and the teacher shape American Public has become more aware of school
the climate of learning. The intermediate level consists performance and student achievement. As a result of
of indirect influences on the child such as parental this report, more attention was placed on standard-
work environment (e.g., income level, parental work ized test scores that emphasize mathematical and lin-
schedules), and community services. The outermost guistic aptitude to measure student achievement.
level, the macrosystem, consists of global contexts Recently, the No Child Left Behind Act of 2001 has
such as the state and federal economic systems, increased the emphasis on standardized testing and
prevailing cultural norms, and societal laws. Each sys- student achievement to a greater degree. As a result,
temic level is interconnected, and all play a role in most of the time, energy, and resources in schools
shaping development. are channeled toward teaching to the test. Social
Acccording to social learning theory, people learn and emotional facets of education are often usurped
by interacting with others. Learners acquire skills, strat- by the pursuit of greater academic achievement.
egies, and beliefs by observing and modeling others However, research indicates that the climate of learn-
in their environment. Albert Bandura, considered to ing is an important variable that can have direct
556 C Climate of Learning
implications on student achievement. Solid classroom evidence base for social and emotional learning from
management techniques, clear and high expectations, preschool to high school (CASEL 2010). CASEL has
and positive, respectful interactions between students developed curriculum materials for schools across the
and teachers are components of the learning climate. country. The following are the core beliefs of CASEL,
Though they are social and emotional in nature, they obtained from the CASEL Web site:
have a direct impact on student achievement. When
● Adults have a responsibility to help children to
classroom management techniques that minimize dis-
become knowledgeable, responsible, healthy, car-
ruptions to learning are utilized, students spend more
ing, and contributing members of society.
time engaged academically and perform better
● Rigorous science provides an essential foundation
(Freiberg et al. 2009).
for effective educational policies and practices;
Parents send their children to school with the
a core aspect of rigorous science is to ground
expectation that their students will become lifelong
development and testing in real-life settings and
learners and happy, well-adjusted members of society
conditions.
(Cohen 2006). Cohen’s research indicated that parents
● Effective, integrated SEL programming is the most
are more concerned about their children’s social and
promising educational reform to promote the
emotional functioning as adults, as opposed to their
academic success, engaged citizenship, healthy
academic functioning. However, in the American edu-
actions, and well-being of children.
cation curriculum, little emphasis is placed on teaching
● Cross-disciplinary collaboration produces the
students social, emotional, and ethical skills. Yet,
richest insights, biggest impacts, and best outcomes
a strong social, emotional, and ethical curriculum is
in work on behalf of children.
necessary to produce citizens who will actively partic-
● We strive for excellence in all our work. We have
ipate in a democracy. Cohen argues that the lack of
high expectations for ourselves, and we encourage
such a curriculum is not only an injustice to American
and expect the best from others.
schoolchildren, but also a violation of their human
● CASEL leadership, staff, and collaborators must
rights (2006). Even the Founding Fathers indicated
model social and emotional competence and ethical
that all citizens are entitled to “The pursuit of hap-
behavior.
piness.” Children deserve a holistic education that
addresses their academic, social, emotional, and phys- Learning is a lifelong, holistic endeavor and is nei-
ical needs. ther limited to a classroom, nor the first 18 years of life.
Knowles also noted the discrepancy between how Athletes learn from coaches and teammates on the
children are taught and what they need to learn. In athletic field. The resident learns from the practicing
order for children to become healthy adults, they physician. The journeyman teaches and guides the
must become self-directed learners (Knowles 1970). apprentice. A Girl Scout learns financial literacy skills
Traditional pedagogical methods often view students from her dedicated leader. A university student logs on
as sponges, soaking up knowledge with little input to his computer, fulfilling requirements for an online
or experience to draw from and creating dependency course. Wherever learning takes place, a learning cli-
on the teacher. The climate of learning is a crucial mate exists. The social, emotional, and physical impact
element in the maturation process and encouraging of the learning climate profoundly shapes the learning
self-direction. Students must feel respected, accepted, experience.
supported, and physically comfortable in order to Online learning environments are growing in pop-
reach their fullest potential (Knowles 1970). ularity, and accommodate a wide range of lifestyles.
The movement to incorporate social and emotional Many universities, secondary schools, and home
education into school curriculums is growing. In 1994, schooling associations are taking advantage of online
Daniel Goleman, author of ▶ Emotional Intelligence, learning communities. This new learning environment
cofounded CASEL (Collaborative for Academic, Social, creates the need for a solid research base on the social
and Emotional Learning) along with Eileen Rockefeller and emotional effects of online learning climates. Do
Growald. CASEL is a nonprofit organization dedicated blogs and discussion boards provide the same oppor-
to advancing scholarly research and broadening the tunities for comprehension and retention of material as
Coaching and Mentoring C 557
(e.g., “Who was the 35th President?”). ▶ Procedural that engenders all cognitive deficits, domain-specific
knowledge is also spared with aging; skills that have theorists believe that age-related decline is specific
been used throughout the lifetime, such as job-related to individual cognitive areas. For example, Naveh-
skills or musical training, are usually retained into Benjamin (2000) has suggested that older adults are
older age. These types of retained knowledge are con- specifically impaired in their ability to form associa-
sidered to be the basis for ▶ crystallized intelligence tions between pieces of novel information. Such a
and expertise. hypothesis is supported by older adults’ relatively
well-preserved recognition of information that is
Emotion Regulation familiar, coupled with their frequent inability to recol-
The ability to regulate one’s reactions to emotional lect the context in which they learned that information.
stimuli – calming oneself after seeing a snake, for As noted earlier, older adults may easily recognize
example – is preserved as people age. Some researchers a new acquaintance, but they may fail to recollect his
suggest that this ability actually improves with age, name or to recall where they met him.
with older adults being able to direct attention away Much research has also been done to investigate
from negative experiences and to maintain positive how older adults can compensate for these cognitive
experiences more effectively than younger adults. deficits. Yaakov Stern and colleagues (1994) have
Older adults also tend to select more emotionally ful- posited the notion of “cognitive reserve,” where
filling activities to participate in than young people do, environmental factors like advanced education and
perhaps because older adults are more likely than healthy lifestyles can be protective factors against
young adults to prioritize social and emotional well- neurodegeneration and cognitive function. It has also
being (Carstensen et al. 1999). been suggested that healthy older adults recruit addi-
tional brain regions than young adults, to compensate
Important Scientific Research and for neural declines in other regions. Roberto Cabeza
Open Questions and others (2002) have used ▶ neuroimaging methods
One of the open debates about cognitive aging is to show that younger adults often recruit brain struc-
whether age-related cognitive deficits are domain- tures on one side or another for various cognitive tasks
general or domain-specific. Domain-general theorists (for example, recruiting the left, but not right, ▶ hip-
suggest that there is one specific deficit that underlies pocampus when learning new information); healthy
all of the age-related impairments. For example, Lynn older adults, however, will often recruit structures
Hasher and Rose Zacks (1988) have suggested that bilaterally (for example, recruiting the hippocampus
older adults’ cognitive impairments arise from a on both sides when learning new information).
deficiency in inhibitory ability. Inhibition theory
hypothesizes that older adults are less good than
young adults at inhibiting thoughts and actions, and
Cross-References
▶ Emotional Regulation
therefore have more difficulty appropriately deploying
▶ Human Cognitive Architecture
attention (and ignoring extraneous information).
▶ Individual Differences in Learning
Older adults are therefore constantly juggling more
▶ Memory Dynamics
information, which leads to memory and processing
▶ Verbal Learning and Aging
speed deficiencies. Other researchers have suggested
that sensory deficiencies – such as loss of hearing and
vision – underlie older adults’ impairment; if informa- References
tion is harder to process, then it becomes harder to Cabeza, R., Anderson, N. D., Locantore, J. K., & McIntosh, A. R.
select, maintain, update, and remember that informa- (2002). Aging gracefully: Compensatory brain activity in high-
tion. Processing speed, as described earlier, is another performing older adults. NeuroImage, 17, 1394–1402.
Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking
domain-general explanation.
time seriously: A theory of socioemotional selectivity. American
By contrast, other researchers believe that cognitive Psychologist, 54, 165–181.
decline differentially affects specific aspects of cogni- Craik, F. I. M. (1986). A functional account of age differences
tion. Rather than the existence of a common thread in memory. In F. Klix & H. Hagendorf (Eds.), Human
Cognitive and Affective Learning Strategies C 563
memory and cognitive capabilities, mechanisms and performance Learning strategies are to be placed at a medium
(pp. 409–422). New York: Elsevier. level of granularity. They differ from learning styles,
Hasher, L., & Zacks, R. T. (1988). Working memory, comprehension,
that is, general approaches to learning, a widely used
and aging: A review and a new view. The Psychology of Learning
and Motivation, 22, 193–225. distinction being that between deep level and surface
Luchies, C. W., Schiffman, J., Richards, L. G., Thompson, M. R., level approaches to learning. A learning technique or C
Bazuin, D., & DeYoung, A. J. (2002). Effects of age, step direction, tactic, such as underlining a keyword definition while
and reaction condition on the ability to step quickly. The Journals of studying a text, constitutes a smaller unit of thought or
Gerontology, Series A, 57, M246–M249.
behavior than a learning strategy. Hence, learning strat-
Naveh-Benjamin, M. (2000). Adult age differences in memory per-
formance: Tests of an associative deficit hypothesis. Journal of egies can be understood as collections of learning tech-
Experimental Psychology: Learning, Memory, and Cognition, 26, niques orchestrated by the learner.
1170–1187. Learning strategies are central to models of self-
Stern, Y., Gurland, B., Tatemichi, T. K., Tang, M. X., Wilder, D., & regulated learning and some researchers equate skilled
Mayeux, R. (1994). Influence of education and occupation on the execution of learning strategies with self-regulated
incidence of Alzheimer’s disease. The Journal of the American
learning. Self-regulated or strategic learners are assumed
Medical Association, 271, 1004–1010.
to have knowledge of various learning strategies, employ
appropriate strategies in order to attain their learning
goals, and flexibly adapt their choice of strategies to the
task and context they face.
Cognitive and Affective
Learning Strategies Theoretical Background
The role of the learner’s deliberate thought processes and
ISABEL BRAUN, JOHANNES GURLITT, MATTHIAS NÜCKLES strategic behaviors in bringing about learning outcomes
Department of Educational Science, University of began to be investigated in the 1960s as a result of the
Freiburg, Freiburg, Germany fundamental paradigm shift in cognitive and educa-
tional psychology. As the behaviorist view of learning
as a strengthening of responses to stimuli by means
Synonyms of rewards became replaced by the cognitive view of
Learning skills; Self-regulated learning strategies; Self- learning as information processing, researchers turned
regulatory processes; Study skills; Study strategies their attention to basic cognitive learning strategies,
particularly mnemonic strategies. But it was not until
Definition the emergence of the constructivist paradigm in psy-
A college student who prepares for an exam by sum- chology and education that researchers focused on com-
marizing the textbook chapters assigned by the course plex learning strategies. At the heart of constructivism is
instructor, drawing a map of key concepts, monitoring the view of learning as active information processing:
his understanding while drawing the map, setting goals The learner actively selects and organizes to-be-learned
on a daily basis, and checking his progress against the information in working memory and integrates new
goals uses a repertoire of learning strategies. When an information with information stored in long-term
elementary school student practices a poem by repeat- memory. This view of the learner as a sense maker
ing the lines over and over again, researchers also refer entails the assumption that the learner employs cogni-
to this behavior as a learning strategy. The term learn- tive learning strategies that are more complex than, for
ing strategy denotes thoughts and behaviors the learner example, mnemonic strategies (Mayer 1996).
employs with the intention of acquiring knowledge and There are a number of taxonomies or systems for
improving task performance (Weinstein and Mayer categorizing learning strategies. The broad distinc-
1986). In line with this definition, learning strategies tion between primary strategies (aimed at cognitive
can be cognitive and affective. Cognitive learning strat- processing) and support (or affective) strategies is
egies exert a direct influence on knowledge acquisition, widely recognized and accommodates the finding that
whereas affective learning strategies facilitate learning learners who have a repertoire of cognitive learning
via, for example, motivation and volition. strategies may not succeed in achieving certain learning
564 C Cognitive and Affective Learning Strategies
outcomes. Both skill (the ability to select appropriate Support strategies or affective learning strategies
cognitive learning strategies and execute them success- exert an indirect influence on cognitive processing.
fully) and will (the motivational and volitional require- When learners employ affective learning strategies
ments for effective strategy use) are necessary. they aim at setting a positive mood for learning,
In their well-known taxonomy of learning strate- arranging the environment to be suitable for studying,
gies, Weinstein and Mayer (1986) distinguished six managing internal and external resources such as con-
types of cognitive learning strategies: rehearsal strate- centration and time, and coping with anxiety and other
gies for basic and complex learning tasks, elaboration emotions about learning. The focus of research on
strategies for basic and complex learning tasks, and affective learning strategies has not been, however, on
organizational strategies for basic and complex learn- the strategies just described but on strategies targeting
ing tasks. Hence, Weinstein and Mayer considered motivation and volition. Motivational learning strate-
both the specific functions of learning strategies with gies include, for example, goal setting strategies and
regard to information processing as well as differences strategies for sustaining academic self-efficacy. The
in the complexity of learning tasks. Rehearsal strate- learner employs volitional learning strategies to
gies, such as mentally reciting keyword definitions, form intentions and maintain commitment toward
serve the cognitive functions of selection (transfer of learning goals, particularly in the face of competing
new information into working memory) and acquisi- non-academic tasks and activities. The focus on moti-
tion (transfer into long-term memory). The cognitive vational and volitional learning strategies might stem
functions of elaboration strategies are construction from research providing insights into the major diffi-
and integration. According to Weinstein and Mayer, culties experienced by self-regulated learners. One of
the learner constructs connections between pieces these difficulties is procrastination. It has been concep-
of new information or integrates new information tualized as a failure to exert volitional control over one’s
with prior knowledge when generating elaborations, own learning, meaning procrastinators fail to employ
such as mental images or analogies. Today, most appropriate strategies for managing their commitment
researchers define only the latter process, that is, the toward learning goals.
construction of connections between new information
and prior knowledge, as elaboration. Organizational Important Scientific Research and
strategies, the remaining type of cognitive learning Open Questions
strategies in the Weinstein and Mayer taxonomy, are Empirical research on cognitive and affective learning
directed at the construction of internal connections strategies has centered on the development of strategic
within new information. Outlining a textbook chapter learning and on the training of learning strategies.
is an example of an organizational strategy. In addition Research indicates that strategy use increases with age
to cognitive learning strategies, Weinstein and Mayer and that the development of strategic learning follows
included affective strategies and comprehension moni- a trajectory from rudimentary, sporadic strategy use to
toring strategies in their taxonomy of learning strategies. appropriate, consistent strategy use. When trained or
Self-monitoring of comprehension during learn- prompted young children benefit from and can acquire
ing constitutes a metacognitive learning strategy. cognitive and metacognitive learning strategies even
Metacognitive learning strategies also include planning before they enter the elementary grades. In the second-
and reflection. Models of self-regulated learning stress ary grades, learners enlarge their repertoire of learning
the importance of metacognitive activities during strategies and eventually have sophisticated cognitive
learning and there is ample evidence that these and metacognitive learning strategies at their disposal
activities indeed contribute to learning. However, cog- (e.g., Zimmerman and Martinez-Pons 1990). However,
nitive and metacognitive learning strategies are very several deficiencies regarding the acquisition and use
much intertwined as metacognitive strategies operate of learning strategies have been identified and some
on domain knowledge. Therefore, the effectiveness of of them have been linked to developmental processes.
employing metacognitive learning strategies when they Young children have been demonstrated to have a
do not form part of a well-orchestrated repertoire of mediation deficiency (Flavell et al. 1966), which refers
cognitive learning strategies has to be considered small. to executing a cognitive or metacognitive learning
Cognitive and Affective Learning Strategies C 565
strategy in an inadequate or incomplete way and problem solving, or foreign language learning. One
hence failing to benefit from it. Learners who show of the best known direct interventions is reciprocal
a mediation deficiency are at a cognitive developmental teaching (Palinscar and Brown 1984). Reciprocal teach-
level that does not permit them to construct the cogni- ing was designed to facilitate the acquisition and trans-
tive “mediators” required for adequate and full execu- fer of reading comprehension strategies in the regular C
tion of the learning strategy. In other cases, learners classroom. During reciprocal teaching, students read
show production deficiencies (Flavell et al. 1966): a text passage by passage and take turns in executing
Although they may spontaneously produce appropri- a sequence of cognitive and metacognitive learning
ate cognitive and metacognitive learning strategies, strategies: asking questions, summarizing, seeking clar-
they typically fail to do so. In yet other cases, learners ification, and making predictions. The teacher models
at the secondary and college levels fail to benefit and scaffolds the use of each strategy, fading his/her
from their use of cognitive and metacognitive learn- support as students gain proficiency in executing the
ing strategies although they spontaneously produce strategies. The effectiveness of reciprocal teaching has
appropriate strategies. It has been suggested that been established in field and laboratory studies. Recip-
utilization deficiencies (Miller 1994) underlie the rocal teaching produces short-term and long-term
failure to benefit from the use of cognitive and effects on measures of comprehension, strategy knowl-
metacognitive learning strategies once the learner has edge and skills, and transfer to novel tasks.
completed the developmental trajectory for a specific The implementation of learning journals as an
strategy. Which factors contribute to the occurrence adjunct to classroom instruction at the secondary level
of utilization deficiencies is not yet completely under- or as a supplement to traditional college coursework
stood. Among the causal mechanisms discussed are forms an example of an indirect learning strategy inter-
high cognitive load during initial strategy use, insuffi- vention. The learning journal constitutes a specific writ-
cient metacognitive self-regulation, low perceived ing task that requires students to organize, elaborate,
self-efficacy and cognitive developmental factors, and reflect on learning contents, typically over an
particularly working-memory capacity limitations dur- extended period of time (Nückles et al. 2009). It has
ing childhood. been demonstrated to produce learning gains and to
At the most general level, learning strategy inter- promote the acquisition of cognitive and metacognitive
ventions differ in how they promote effective strategy learning strategies. For the potential benefits of learn-
use. Learning strategies can be trained directly through ing journals to unfold, however, instructional support
explicit instruction on the cognitive, metacognitive, appears necessary. Such support can be provided, for
and affective components of strategic learning. But example, through strategy prompts embedded in the
they can also be trained indirectly in learning environ- writing instruction. The ineffectiveness of learning
ments that facilitate or require strategic learning. Some journals written without instructional support, which
researchers argue that it may be most productive to emerged in early studies, seemed to confirm concerns
combine elements of direct and indirect interventions about indirect learning strategy interventions that had
to promote the acquisition of learning strategies. How- been raised since the early decades of research on learn-
ever, the appropriate balance of explicit instruction and ing strategies (e.g., Weinstein and Mayer 1986). How-
implicit facilitation has yet to be established through ever, powerful evidence of the effectiveness of guided
empirical studies. journal writing has weakened these concerns.
Numerous studies on direct interventions are At the conceptual level, a central issue of debate
reported in the learning strategies literature. Altogether, among researchers is the nature of strategic learning.
these studies show that direct interventions are effective Underlying the taxonomy of learning strategies, the
when they are carried out over an extended period of developmental model and the training approaches
time, provide the learner with information about when outlined above is the assumption that skill and will,
to use which learning strategy, and include instruction that is, relatively stable learner characteristics, underlie
on metacognitive strategies, that is, planning, self- the use of learning strategies. Several researchers have
monitoring, and control strategies. Most direct argued, however, that the nature of strategic learning
interventions target reading, writing, mathematical has to be conceptualized differently. Two alternative
566 C Cognitive and Noncognitive Processes
Cross-References
▶ Elaboration Effects on Learning
▶ Elaboration Strategies and Human Resources
Development Cognitive Apprenticeship
▶ Learning Styles Learning
▶ Metacognition and Learning
▶ Self-Regulated Learning JOANNA K. GARNER
▶ Self-Regulation and Motivation Strategies Berks College Department of Psychology,
The Pennsylvania State University, Reading, PA, USA
References
Flavell, J. H., Beach, D. R., & Chinsky, J. M. (1966). Spontaneous
verbal rehearsal in a memory task as a function of age. Child
Development, 37(2), 283–299. Synonyms
Mayer, R. E. (1996). Learning strategies for making sense out of Reciprocal teaching; Scaffolding; Situated cognition;
expository text: The SOI model for guiding three cognitive Situated learning
processes in knowledge construction. Educational Psychology
Review, 8(4), 357–371.
Miller, P. H. (1994). Individual differences in children’s strategic Definition
behaviors: Utilization deficiencies. Learning and Individual Cognitive apprenticeship learning is situated within
Differences, 6(3), 285–307. social constructivist approaches to instruction. It pri-
Nückles, M., Hübner, S., & Renkl, A. (2009). Enhancing self-regulated oritizes the use of authentic tasks and situations, and
learning by writing learning protocols. Learning and Instruction,
the role of interactions between more and less skilled
19, 259–271.
Palincsar, A. S., & Brown, A. L. (1984). Reciprocal teaching of
individuals in order to foster the development of
comprehension-fostering and comprehension-monitoring activ- metacognitive strategies and domain-specific problem-
ities. Cognition and Instruction, 1(2), 117–175. solving skills. A focus on cognitive rather than physical
Weinstein, C. E., & Mayer, R. E. (1986). The teaching of learning skill development and the use of planned rather than
strategies. In M. C. Wittrock (Ed.), Handbook of research in entirely naturalistic opportunities for skill development
teaching (Vol. 3, pp. 315–327). New York: Macmillan.
and practice differentiate cognitive apprenticeship from
Zimmerman, B. J., & Martinez-Pons, M. (1990). Student differences
in self-regulated learning: Relating grade, sex, and giftedness to more traditional models of craft apprenticeship.
self-efficacy and strategy use. Journal of Educational Psychology, A key process goal of cognitive apprenticeships is
82(1), 51–59. to make otherwise tacit cognitive and metacognitive
Cognitive Apprenticeship Learning C 567
processes explicitly available during the performance of the type of contextually embedded practice of authen-
complex tasks. This is done via the instructor serving tic skills that supports the development of expertise
as an expert and coach who models and verbalizes and, instead, limits the development of content mas-
their thought process as well as supporting increasingly tery and undermines intrinsic motivation. Apprentice-
independent and reflective practice of these same pro- ships offer opportunities for the learner to engage in C
cesses by the learner. Specifically, according to Collins meaningful and contextualized practice of transferable
et al. (1987), the instructor uses the timely implemen- knowledge and skills (Collins et al. 1987). Accordingly,
tation of three instructional methods designed to facil- not only is domain knowledge viewed as epistemolog-
itate the acquisition of skills (modeling, coaching, and ically inseparable from the context in which it will be
scaffolding), two methods designed to improve the used, but on a practical level, it must be acquired within
clarity of observation and self-observation (articula- contextualized instructional experiences in order to
tion and reflection), and one method designed to pro- be available during problems requiring generalization
mote learner autonomy (exploration). These methods and transfer.
are grounded in social learning theory because learners Cognitive apprenticeship differs from traditional
observe, enact, and respond to feedback. Cognitive instruction in two key ways. First, its social construc-
apprenticeship techniques make use of scaffolding, tivist foundation incorporates a view of meaning as
which allows students to perform at a higher level something that is negotiated and developed among
than they would otherwise be able to if acting alone individuals who reside within a community. Dialogic
because of the instructor’s accuracy in monitoring and interactions between teachers and students are thus
diagnosing each student’s current ability level. These essential to the learning process. All are expected to
methods broadly align with research on the contextu- be actively engaged in discovering, articulating, model-
ally situated nature of expertise, in part because encul- ing, and refining conceptions of the content as well
turation into an expert community is adopted as an as the conditions under which that content can be
implicit instructional goal. Finally, these methods draw meaningfully used and how such conclusions came
upon stage-based descriptions of skill acquisition in about. Second, cognitive apprenticeship learning
which the goal is for the learner to be able to execute involves the practice of authentic tasks instead of iso-
their skills automatically and in a wide variety of appro- lated component skills. The role of the teacher thus
priate contexts. becomes one of coach and facilitator, whose job is to
assist students as they interact with complex and mean-
Theoretical Background ingful problems – first through modeling, then scaf-
Within a professional domain, experts display the folding, then prompted reflection. Thus, learners move
ability to identify and solve problems because they toward more expert-like performance because they
possess a substantial body of interconnected con- have the chance to observe, discuss, and receive feed-
ceptual knowledge and accessible procedural heuris- back on their use of strategies. They acquire flexible,
tics or decision-making strategies (Collins et al. task-oriented problem-solving strategies, and become
1987). Experts also have a wealth of cognitive and more articulate about their strategy use because of an
metacognitive knowledge and strategies at their dis- increased ability to reflect meaningfully on the learning
posal, but the seemingly effortless execution of these process itself.
strategies often renders the complex constituents of Cognitive apprenticeship is possible because of sev-
these processes invisible to the novice learner (Mayer eral key psychological concepts. One, articulated by
1991). Proponents of cognitive apprenticeship learning Bandura (1977) through his social learning theory, is
have argued that traditional schooling reduces the the idea that humans have a tremendous capacity to
opportunity to observe and emulate authentic problem learn through the actions and verbalizations of another
solving, isolates the presentation and use of informa- person. Bandura called this person a model, and thus,
tion from the contexts in which it will be relevant, and the term “modeling” was adopted to refer to cognitive
masks expert-like thought processes from novice and behavioral displays meant to teach another
learners (Brown et al. 1989). Thus, it does not foster person. In cognitive apprenticeship, the teacher serves
568 C Cognitive Apprenticeship Learning
as the initial model, but as students progress, they are 1. Modeling: In modeling, the expert carries out a task
encouraged to adopt this role for one another. A second or solves a problem. This is done in such a way that
key concept is the process of scaffolding (Vygotsky students can observe the required steps, but also can
1978). In scaffolding, a more skilled person provides listen to the control processes and decisions that the
assistance to allow a less skilled individual the oppor- expert uses along the way.
tunity to perform at a level that he or she could not do 2. Coaching: In coaching, students are provided with
alone. The facilitator ensures the maintenance of a zone prompts, feedback, and other reminders pertinent
of proximal development, a conceptual space through to the successful completion of a specific task.
which the learner progresses as skills are developed. 3. Scaffolding: In scaffolding, teachers provide physi-
Scaffolding processes include verbal and physical cal and verbal prompts and support but only to the
prompts such as questions during expository text read- degree that the teacher completes parts of the task
ing, or cue cards for presented during composition which the students cannot autonomously attempt.
processes. Scaffolding also includes the joint comple- As skill levels increase, supports are removed
tion of task components which cannot be achieved through the process of fading.
independently. In combining these concepts of model- 4. Articulation: In articulation, teachers prompt stu-
ing and scaffolding, cognitive apprenticeship calls dents to explicitly state their approaches and strat-
upon the teacher to simultaneously model expert-like egies, and to characterize their beliefs about the
skills and provide appropriate scaffolds during each domain or skill.
stage of the learning process. A third important concept 5. Reflection: In reflection, a reflection on process is
is the important role played by metacognitive strategies permitted via analysis of recent performance. This
in domain expertise. In addition to declarative, pro- can take place via verbal review or by reviewing
cedural, and conditional knowledge, experts possess a recording.
a significant capacity to recognize patterns, connect 6. Exploration: In exploration, students are encour-
problem states with solution paths, and execute strat- aged to seek and define new problems within the
egies that yield solutions (Mayer 1991). The novice domain, in order to practice using and transferring
learner therefore benefits from metacognitive reflection skills from one context to another.
by the expert who is modeling the skill in order to learn
the conditions under which particular strategies are Important Scientific Research and
used and how troubleshooting may occur. This allows Open Questions
the novice to understand how and why problems are Initially, the cognitive apprenticeship approach was
framed, approached, and solved. Thus, the expert identified in programs designed to improve reading
thinks out loud and uses appropriate prompting tech- comprehension, writing, and mathematics skills.
niques within a collaborative dialogue in which options Collins et al. (1987) describe these techniques in detail.
for solving a problem or completing a task are explic- They are reciprocal teaching (Palinscar and Brown
itly verbalized. In application, this may mean that the 1984), procedural facilitation of writing (Bereiter and
teacher models how to generate appropriate “why” Scardamalia 1987), and Schoenfeld’s (1994) approach
questions while reading, how to revise the first draft to teaching mathematical problem solving. Of these,
of an essay, and how to dissect math problems into reciprocal teaching has garnered the most empirical
givens and unknowns. attention. Employed in elementary, middle, and high
Alongside this argument for the benefits of school settings, reciprocal teaching is a reading com-
a situated learning approach, Collins et al. (1987) prehension strategy instruction method where teachers
presented a framework for cognitive apprenticeship and students alternate between reading and then
learning that included six processes teachers use to discussing the content and the metacognitive processes
promote student learning. During learning episodes required to comprehend the text. Working in small
which are carefully sequenced for increasing com- groups, teachers and students take turns to lead the
plexity and diversity of required skill, the following discussion. Scaffolding and prompting takes place as
take place: necessary, and over time, the goal is to fade the usage of
Cognitive Apprenticeship Learning C 569
such prompts due to the increasingly expert-like nature ubiquitous presence of the type of social interaction
of the students’ reading strategies. required for successful apprenticeship-based lessons.
In their seminal research, Palinscar and Brown Cognitive apprenticeship models often assume that
(1984, study 1) used reciprocal teaching techniques in learners are able and willing to engage in extensive
a group of seventh grade students with poor compre- collaborative discussion in which their inner thought C
hension skills. Comparison groups received instruction processes can be articulated. Järvelä identified the need
on how to locate information during testing, or to investigate the process of individualizing instruc-
participated in the testing schedule but received no tion, since scaffolding and modeling often need to be
additional instruction. After the intervention, compre- adjusted to suit the needs of linguistically, motivation-
hension accuracy increased from 30% to 80% for the ally, and socially diverse learners.
experimental group, who also performed at or above Along similar lines, relatively little emphasis has been
grade level on measures of skill generalization and placed on gathering data to measure the moment-by-
transfer. Dialogue analyses showed that as time went moment interactions among students and between
on, more expert-like questioning and summarization students and teachers. It is important to understand
strategies were used by the experimental group. These how the process of apprenticeship, as revealed through
results have been replicated – in a review of 16 quanti- small group processes, leads to the achievements
tative studies of reciprocal teaching, Rosenshine and documented in many studies. This may require mixed
Meister (1994) reported a median effect size of 0.32 methods research that can connect intra-individual,
and 0.88 for standardized and experimenter-developed inter-individual, and contextual variables with learning
comprehension tests, respectively. However, in a qual- outcomes. In addition, research is needed to document
itative analysis assessing the success of adopting recip- the long-term feasibility and success of classrooms that
rocal teaching methods, Hacker and Tenent (2002) adopt cognitive apprenticeship methods, whether these
reported that teachers encountered difficulties in are defined from a teacher, learner, or administrative
ensuring that groups of students stayed on-task, used perspective. Finally, questions have been raised about
strategies effectively, and generated sufficiently inter- the legitimacy of the claim that tasks used to teach
rogative questions. Enduring challenges of adopting within apprenticeship-based lessons are truly authen-
cognitive apprenticeship models may therefore be tic, as well as the context and conditions in which cogni-
how to balance classroom logistics and developmental tive apprenticeship techniques are most appropriate and
needs in the absence of low student-to-teacher ratios, feasible. Apprenticeship approaches seems to be most
and how to effectively train teachers to become flexible successful in content domains where a metacognitive or
in their implementation of apprenticeship-liked teach- self-regulatory process lies at the heart of what needs to
ing strategies. be learned, and where problem solving is integral to
Other examples of cognitive apprenticeship tech- the desired skill. This is the case even when the tasks
niques can be found in areas such as science and scien- presented cannot always be considered to be entirely
tific inquiry (Roth and Bowen 1995), instructional authentic. But when students must undergo radical cog-
technology, computer programming, teacher profes- nitive restructuring as well as domain-related skill devel-
sional development, medicine, and psychology (Järvelä opment, such as in the case of conceptual change in
1996). Each emphasizes qualities of situated learning science, apprenticeship models may prove to be less
that map onto cognitive apprenticeship, such as learn- successful (Vosniadou 2007).
ing in a social context through the joint considera-
tion of ill-defined problems, and the importance of Cross-References
drawing upon the knowledge of peers or instructors ▶ Apprenticeship-Based Learning in Production
who demonstrate higher levels of expertise (Roth and Schools
Bowen 1995). However, in a review, Järvelä (1996) ▶ Scaffolding
identified several areas that would benefit from addi- ▶ Self-regulated Learning
tional research. She criticized proponents of cognitive ▶ Situated Learning
apprenticeship for making assumptions about the ▶ Socio-constructivist Models of Learning
570 C Cognitive Apprenticeship Modeling
References
Bandura, A. (1977). Social learning theory. New York: General Cognitive Architecture
Learning.
Bereiter, C., & Scardamalia, M. (1987). Fostering evaluative, diagnos-
A theory, expressed as a suite of computer programs
tic and remedial capabilities. In C. Bereiter & M. Scardamalia that provides specification for structures and related
(Eds.), The psychology of written composition (pp. 265–298). processes of the cognitive system. Models derived from
Hillsdale: Erlbaum. the architecture are typically used to explain phenom-
Brown, J. S., Collins, A., & Duguid, P. (1989). Situated cognition and ena in several domains.
the culture of learning. Educational Researcher, 18, 32–42.
Collins, A., Brown, J. S., & Neuman, (1987). Cognitive apprentice-
ship: teaching the craft of reading, writing, and mathematics. Cross-References
Technical Report No. 403. Center for the Study of Reading. ERIC
▶ Schema-Based Architectures of Machine Learning
Document 284181.
Hacker, D. J., & Tenent, A. (2002). Implementing reciprocal teaching
in the classroom: Overcoming obstacles and making modifica-
tions. Journal of Educational Psychology, 94, 699–718.
Järvelä, S. (1996). New models of teacher-student interaction:
A critical review. European Journal of Psychology of Education,
11, 249–268. Cognitive Artifacts and
Mayer, R. E. (1991). Thinking, problem solving, cognition. New York: Developmental Learning in
Worth.
Palinscar, A. S., & Brown, A. L. (1984). Reciprocal teaching of com-
a Humanoid Robot
prehension fostering and comprehension monitoring activities.
Cognition and Instruction, 1, 117–175. ARTUR ARSÉNIO
Rosenshine, B., & Meister, C. (1994). Reciprocal teaching: A review of Instituto Superior Técnico, Technical University of
the research. Review of Educational Research, 64, 479–530. Lisbon, Lisbon, Portugal
Roth, W.-M., & Bowen, G. M. (1995). Knowing and Interacting:
A study of culture, practices, and resources in a grade 8 open-
inquiry science classroom guided by a cognitive apprenticeship
metaphor. Cognition and Instruction, 13, 73–128. Synonyms
Schoenfeld, A. (1994). Reflections on doing and teaching mathemat- Artificial intelligence; Educating robots, teaching
ics. In A. Schoenfeld (Ed.), Mathematical thinking and problem robots as humans; Human-aided machine learning;
solving (pp. 53–69). Hillsdale: Erlbaum. Human-robot social interactions
Vosniadou, S. (2007). The cognitive-situative divide and the
problem of conceptual change. Educational Psychologist, 42,
55–66. Definition
Vygotsky, L. (1978). Mind in society: The development of higher psy-
Development Learning in a humanoid robot defines
chological processes. Cambridge, MA: Harvard University Press.
an incremental, staged methodology for machine
learning based on similar principles that guide chil-
dren’s development. It is therefore strongly tied to
developmental psychology, according to the epigenetic
principle: as each stage progresses, it establishes the
Cognitive Apprenticeship foundation for the next stages. Cognitive Artifacts are
Modeling a humanoid robot’s (and a child’s) learning aids –
▶ Expert Cognitive Modeling and Problem-Based or cognitive enhancers – such as books, toys, puzzles,
Learning drawing boards, or construction bricks, employed by
a caregiver in order to guide development learning
of a humanoid robot. They are an important tool
to achieve socially intelligent humanoid robots
(Arsenio 2004a, b) – introducing robots into our soci-
Cognitive Approach ety and treating them as us – using child development
as a metaphor for developmental learning of a human-
▶ Situated Cognition oid robot.
Cognitive Artifacts and Developmental Learning in a Humanoid Robot C 571
students (see, e.g., Reif 2008). However, for Richard when he felt he did not fully understand quantum
Feynman (1918–1988), a famous physicist and Nobel mechanics. As Davies wrote (1997, pp. 420–421):
Prize winner, “subjects like philosophy and psychology
" If teachers continue to give the impression that they do
are hard, but physics is easy and that’s precisely why we
have a better basic understanding of such fundamen-
know so much about it.” But if physics is “easy,” why is
tals than their students, the students will see their own
C
it difficult to teach and learn? Certainly, there are many
perplexity and uncertainty as a negative reflection on
reasons for that. One, surely not the least important, is
their own capabilities. Even in this group today there
that teachers soon face the harsh reality of how deep
will be some of you who will remember the relief you
and extensive their students’ difficulties are and how
felt when you could use some equation, and your
naı̈ve it is to assume that kids are just as enthusiastic
mathematics, to answer a problem, rather than stay
about the curriculum as they are.
with your uncertainties regarding the concepts
Besides the many social-cultural problems teachers
involved. We learn and teach others to use mathemat-
face in their teaching, it can be argued that learning
ics to manipulate the symbols associated with myster-
science, and physics in particular, is like learning a new
ies. This does not mean that we or they have a grasp of
language – a language that uses many of the same words
the mysteries themselves.
as ordinary language but with altered and far more
precise meanings. Physics deals with conceptual objects such as force,
The essence of the problem of learning the language velocity, energy, radiation, etc. These are all words that
of physics is learning to make conceptual distinctions are in common use in everyday language. In fact, many
among related but distinct concepts. It is, essentially, people will use words like momentum, force, and energy
a matter of familiarization with the lexicon of the interchangeably in casual conversation. Nonetheless, in
language and its proper use in specific contexts. Famil- physics these concepts and the words we use to name
iarization is an important issue when learning science these concepts are quite distinct. Force, in the language
(and mathematics). And, for some eminent scientists, of physics, is the “rate of change of momentum.”
becoming familiar with is so important to the success of Energy or work can be related to force as can momen-
scientific ideas that new ideas only become triumphant tum, but neither force nor momentum is conceptually
because supporters of old ideas die, as Planck wrote in the same as energy. Power is yet a different concept.
his autobiography (Planck 1950, pp. 33–34): “A new Another important issue in learning such abstract
scientific truth does not triumph by converting its concepts, and one that is intimately related to familiar-
opponents and making them see the light, but rather ization, is the issue of reification, that is, of concretiza-
because its opponents eventually die, and a new gener- tion of abstract objects. According to Wright and Wright
ation grows up that is familiar with it.” Scientists fre- (1998, p. 128), “Reification is a central goal [. . . of
quently say that they do not understand some of the learning science and mathematics]; it essentially
most fundamental concepts or theories in their own defines scientific literacy. It is the foundation for com-
field. For example, Feynman confessed that he did not mon sense about how the world works (. . .).”
really understand quantum mechanics. If we take the position that reification and familiar-
Experienced physics teachers also alert us to the ization are essential aspects of learning physics and
fragile nature of our understanding. For example, mathematics, we are led to ask how can such learning
many of the useful concepts of physics are, for teen- be improved with technology and, specifically, with
agers, mysterious and difficult to grasp; the nature of an computers? Hebenstreit, writing about the role of
electrostatic charge, of a magnetic field, of electromag- computers in education, coined a term that provides
netic wave propagation in vacuum, or of charm and an important insight into understanding how com-
color of quarks are examples. There is no absolute puters can help in the reification of knowledge. For
understanding or knowledge of the nature of these Hebenstreit, computers allow us to manipulate a new
entities, yet any young adult will certainly wonder type of object; a kind of object that he calls a concrete–
about their nature. In physics education, there is surely abstract object (Hebenstreit 1987). Concrete in the
the need for the kind of humility shown by Feynman, sense that they can be manipulated on the screen and
574 C Cognitive Artifacts, Technology, and Physics Learning
react as “real objects” and abstract because they might representations of non-concrete formal objects. This
be physical or mathematical constructs such as vectors, ability accounts, at least in part, for the increasing
equations, fields, etc. importance of computer visualization and simulation
Teachers tend to teach what they can teach, not in science in general and in physics in particular.
necessarily what they think it would be useful to Galison (1997), for example, wrote about the new
teach. This is what some authors call technological “epistemic position” of computers and simulations in
determinism. For example, most of the practical and the production of physics knowledge.
theoretical teaching is dependent on the limited math- Nickerson (1995) pointed out that researchers had
ematics that students (and also teachers) can use: sim- not focused on students as authors of simulations.
ple analytical tools that often need complex algebraic He argued that “it is only difficult, not impossible,
manipulation. But with computer tools, one can use and the work that goes into the successful building of
numerical approaches that can turn out to be simpler a microworld is likely to deepen one’s understanding
than analytical ones and lead to improved familiariza- of whatever the microworld is intended to simulate”
tion and reification of physics by students. (p. 16). To build simulations, one can use program-
A characteristic feature of using a computer as ming languages, but these often require technical
a cognitive artifact is that the emphasis is on meaning knowledge and skill outside of the domain being sim-
and semiquantitative reasoning instead of formulaic ulated. This is the reason why Nickerson proposed
solving of well-classified problem types. A good exam- the development of specific tools that can be used by
ple of what is semiquantitative reasoning can be done people without that knowledge (p. 16): “For student-
with a mathematical object such as dx/dt = 4 t. (For developed simulations to be practical for educational
the sake of concreteness, assume that x is a distance and purposes, it will probably be necessary to develop tools
t is a time – in that case the 4 represents an accelera- that are designated to facilitate the building of simula-
tion.) What does this tell us? First, the rate of change of tions by people without such language facility and
x is proportional to t and that means that the larger the programming experience.” Such tools have been devel-
value of t the larger the rate of change of x. More oped in the last decade (see, e.g., Teodoro 2003), and its
precisely, when t is 5 time units, for example, the rate impact has been assessed (see, e.g., Teodoro and Neves
of change will be, at that instant, 4 5 = 20 velocity 2011). These tools have been used extensively as an
units. If t is 10 time units, then the rate of change will be “integral part” of new curricula, such as Advancing
4 10 = 40 velocity units. That is, if t doubles, the rate Physics, developed by the Institute of Physics in the
of change of distance doubles. Moreover, x always UK (2000/2008).
increases for positive values of t. Consider another In the early 1980s, it was not clear how important
example: dx/dt = 4 x. (NB in this case the 4 represents and ubiquitous computers would become in our soci-
1 divided by a time or a frequency.) Now we have a rate ety. Now computers have changed working practices
of change of x that is proportional to x at any instant of and leisure activities, and everybody agrees that
time. For example, if x is zero, then the rate is also zero. using computers is part of literacy and numeracy. The
For a positive value of x, at any instant of time, the rate impact on science is so profound that, for the National
of change is positive and so x increases. Experts can Research Council (USA), scientific computation can
readily do this kind of semiquantitative reasoning even be considered a third fundamental methodology of
if they do not know the specific details of a calculation. science – parallel to the experimental and theoretical
Physics is a science in which visualization plays an approaches.
important role, even when visualization is only used to It has been pointed out that computers, like all
show mathematical objects, such as vectors or field technological innovations in schools, tend to follow
lines. It therefore seems reasonable to suppose that a cycle of high expectations, rhetoric about the need
computer visualization can help learners create mean- to innovate, oriented policy and finally limited use.
ing from manipulations of abstract objects. This capa- There have been many promises of radical change in
bility of the computer has been used extensively in education from technology enthusiasts. After intelli-
many contexts and is stressed by many authors, who gent tutoring systems, multimedia, Internet, etc., edu-
pointed out to the capability of making dynamic cators have become cautious of what can really make
Cognitive Artifacts, Technology, and Physics Learning C 575
a difference. Educators increasingly tend to focus on into lectures, practical and laboratory work, what is the
supportive systems, on coaching and scaffolding. best way to integrate computer cognitive tools? How do
Groups such as the group that worked with the Educa- these tools relate with interactive digital documents?
tion Technology Center in Harvard between 1985 and We are beyond the point of needing short-term
1995 have initiated this perspective. The Harvard per- programs that assume that innovation is guaranteed C
spective was based on four principles (Harvard Educa- because it has proven to work with enthusiastic
tional Technology Center 1988): adopters. We need programs that encourage cumulative
improvement committed to ongoing slow but clear
Goals: Focus on key concepts and on the overall nature
change. Computer tools and computer networks have
of knowledge, evidence, and inquiry in a discipline.
an enormous potential impact in learning, and it will
Teaching Approaches: Help students develop a deep
increase as technology advances. But, as Seymour Papert
understanding of the subjects they study by taking
pointed out 30 years ago, there is a world of difference
into account their prior theories and by integrating
between what computers can do and what society will
teacher-directed instruction with opportunities
choose to do with them. We all face the challenge of using
and challenges for critical inquiry.
technology to empower learning (as well as other
Technology: Use technologies selectivity to make a dis-
human activities), and not to create a kind of Aldous
tinct contribution to teaching and learning, for
Huxley Brave New World where machines control
example, to present dynamic models of key ideas
everything, dehumanizing schools and learning.
or to enable students to participate in disciplined
inquiry. Cross-References
Implementation: Design technology-enhanced teaching ▶ Cognitive Artifacts and Developmental Learning in
modules and approaches that can be gradually and a Humanoid Robot
gracefully integrated into existing curriculum and ▶ Learning Through Artifacts in Engineering
practice. Education
As we can see in these statements, technology is not ▶ Models and Modeling in Science Learning
a goal in itself but a selective contribution “to make
a distinct contribution to teaching and learning.” And References
it is the teacher that really can make the difference Davies, B. (1997). Physics like you’ve never had before. Physics Edu-
cation, 32(6), 418–421.
in creating powerful educational environments with Galison, P. (1997). Image & logic, a material culture of microphysics.
technology. Chicago: The University of Chicago Press.
Harvard Educational Technology Center. (1988). Making sense of
Important Scientific Research and the future. Cambridge, MA: Harvard Graduate School of
Open Questions Education.
Hebenstreit, J. (1987). Simulation et pédagogie: Une rencontre du
Embedding the use of computers as information deliv-
troisième type. Gif Sur Yvette: École Superieure d’Electricité.
ery tools has been done in schools in the last decade. This Institute of Physics & Ogborn, J. (2000/2008). Advancing physics AS /
use usually adds nothing fundamentally different from A2. London: The Institute; OCR.
previous tools of delivering information. But integrating Nickerson, R. S. (1995). Can technology help teach for understand-
computers as powerful cognitive tools in the physics ing? In D. N. Perkins, J. L. Schwartz, M. M. West, & M. S. Wiske
curriculum (as well as in mathematics and other scien- (Eds.), Software goes to school. NY: Oxford University Press.
Norman, D. A. (1991). Cognitive artefacts. In J. M. Carroll (Ed.),
tific subjects) is a much more difficult endeavor. It needs Designing interaction: Psychology at the human-computer inter-
a coherent view of the role of cognitive tools, a culture of face. Cambridge: Cambridge University Press.
teaching and learning close to the way science is done, Planck, M. (1950). Scientific autobiography and other papers. London:
and reasonable organizational conditions. Williams & Norgate.
A properly balanced integration of computer cog- Reif, F. (2008). Applying cognitive science to education: Thinking and
learning in scientific and other complex domains. Cambridge, MA:
nitive tools in the curriculum remains to be found.
MIT Press.
Important open questions left for future research are, Teodoro, V. D. (2003). Modellus: Learning physics with mathematical
for example: Is there an optimal set of tools that min- modelling. PhD Thesis, Lisboa: Universidade Nova de Lisboa.
imizes cognitive opacity? If a course is organized Retrieved from http://hdl.handle.net/10362/407.
576 C Cognitive Aspects of Deception
Teodoro, V. D., & Neves, R. G. (2011). Mathematical modelling in about and/or understanding behavioral maneuvers or
science and mathematics education. Computer Physics Commu- even attribution of mental states.
nications (182), 8–10. doi:10.1016/j.cpc.2010.05.021.
In practice, it is difficult to distinguish between
Wright, J. C., & Wright, C. S. (1998). A commentary on the profound
changes envisioned by the national science standards. Teachers different orders of intentionality, as acts carried out to
College Record, 100(1), 122–143. affect the beliefs of others do not look any different
from acts that shall affect merely the others’ behavior.
To date, there is little evidence that nonhuman animals
are capable of full mental state attribution (theory
of mind), i.e., to understand that others have beliefs
Cognitive Aspects of Deception and desires, but there are some persuasive examples
of precursor elements like visual perspective taking.
THOMAS BUGNYAR Recently, attempts have been brought forward to spec-
Department of Cognitive Biology, Konrad Lorenz ify cognitive building blocks of deception, which may
Research Station & Department of Neurobiology and underlie the transition from different orders of inten-
Cognition Research, University of Vienna, Vienna, tionality. The most promising among them are the
Austria ability to flexibly inhibit normal behaviors and the
understanding that conspecifics can be manipulated.
(from mammals to birds and fish), who also live in conceal information and/or provide false information
a complex social environment, have received increased to naive conspecifics in food competition contests.
attention. Long-tailed macaques Macaca fascicularis, in contrast,
do not seem to be capable of actively concealing infor-
Important Scentific Research and mation from a competitive human experimenter in the C
Open Questions foraging context, although they frequently hide from
dominant conspecifics during sexual intercourse.
Experimental Studies on Intentional For testing how animals respond to deception by
Deception others, studies usually involve the use of human exper-
Studies concerning the cognitive underpinning of imenters, who either give false information in choice
deceptive tactics may follow different experimental studies (i.e., they point out the incorrect location of
lines but generally make use of (experimentally hidden food) or they do not share the reward after
induced or naturally occurring) variation in informa- having relied on the behavior of the test subject to find
tion about desired objects (i.e., food). The focus is it. In most of these studies, nonhuman animals like apes,
either on how informed subjects act to prevent others monkeys, and dogs Canis familiaris learn to provide no
from gaining these objects or on how naı̈ve subjects cues and/or to give wrong cues to the experimenter but
respond to receiving false information. only as a result of intensive training. Reversing the
In the “informed forager” paradigm, a particular roles between experimenter and test subject leads to
individual gets informed about the location of food, a drop in performance, which supports the interpre-
usually by allowing her visual access to the hiding tation that the animals base their deception on the
procedure. The subject is then allowed to retrieve the others’ behavior (which they have to learn anew) and
food together with other group members, which may not on an understanding of the others’ intention.
be dominant and/or uninformed about the food
location. In chimpanzees Pan troglodytes, mangabeys Open Questions and Future Research
Cercocebus torquatus, and ravens Corvus corax, some Complex social life involves various ways of commu-
subjects start withholding the correct information nication, cooperation, and competition, offering a
from naı̈ve dominants (who are likely to steal the range of opportunities in which deception would pay
food) and learn to mislead them to false locations. off. Hence, tactical deception may be a widespread
In chimpanzees, dominant subjects may even develop phenomenon that is primarily constrained by social
counter tactics to avoid being cheated by subordinates. structure (e.g., risk of detection, punishment) rather
Other species, in contrast, seem to have problems in than by phylogeny. Empirical evidence for this assump-
learning that others can be deceived (e.g., ring-tailed tion is still scarce but reports on mammals, birds, and
lemurs Lemur catta) or readily adopt alternative strat- recently also on fish point in this very direction. How-
egies to outwit others when misleading attempts are ever, the occurrence of tactical deception does not
not successful (e.g., domestic pigs Sus scrofa). allow inferring the underlying cognitive mechanism.
Knower-guesser studies, originally designed for Notably, tactical deception may reflect intentional
testing mental attribution, feature aspects of the behavior on side of the deceiver but does not need to
informed foraging paradigm and frequently involve reflect attribution of mental states. The critical points
deceptive maneuvers on side of the informed subjects for future research are thus to study the acquisition and
(knowers) against the uninformed guessers. Corvids flexibility of deceptive tactics and to tease apart differ-
like ravens and Western scrub jays Aphelocoma ent levels of intentionality.
californica spontaneously hide from others when they
cache food, and thus withhold information from pos- Cross-References
sible competitors that could subsequently pilfer the ▶ Complex Learning
caches. Ravens also actively distract others from cache ▶ Intelligent Communication in Animals
sites and do false caches, indicating naturally occurring ▶ Observational Learning: The Sound of Silence
forms of misleading behavior. Likewise, chimpanzees ▶ Social Cognitive Learning
and tufted capuchins Cebus apella may spontaneously ▶ Social Construction of Learning
578 C Cognitive Aspects of Natural Communication in Primates
Research with natural populations has been partic- shown to convey details about external events. For
ularly useful in addressing some of these questions, example, chimpanzee screams during agonistic inter-
recently also including great apes. It is now clear that actions reflect the nature of the event, the role of the
nonhuman primates are able to produce messages that caller, the severity of the attack, and whether high-
convey not only their inner states but also something ranking group members are nearby. Similarly, chim- C
about their external world, that they use various com- panzees produce acoustically variable “rough grunts,”
munication strategies to this end, and that they can which covary with the perceived quality of the encoun-
develop a fairly complex understanding of the social tered food (Zuberbühler et al. 2009). Yet a number of
consequences of their signals. Some profound differ- basic questions are still unsolved. What evolutionary
ences between human and nonhuman communication processes can explain the acoustic structure of the
have equally emerged, as summarized in the following. different call types within a species’ repertoire? Habitat
structure, caller physiology, and receiver psychology
Signal and Sequence-Based Meaning are likely candidates, but the details are not well under-
A central problem in primate communication research stood. How widespread are meaningful acoustic vari-
concerns the psychological states underlying and driv- ants of basic call types in primate communication, and
ing signal production. A widely held belief is that which call types are especially prone to acoustic vari-
nonhuman primates only experience different degrees ants? How much control do callers have during call
of arousal, which act as the main agent of signal pro- production, and how do they acquire them? Why did
duction. Another version of the arousal model is that humans evolve so much greater control over their
different events trigger different kinds of arousal, which vocalizations than all other primates?
are then linked to signals. At the same time, there is A second mechanism by which nonhuman pri-
good evidence that nonhuman primates possess mental mates can increase their small repertoire is by combin-
representations and organize their world along mental ing different call types into sequences. This has been
concepts (Tomasello and Call 1997), and it seems found in a number of primates, from Old World
unreasonable to assume that this should not also affect monkeys to gibbons and great apes. There is good
their communication. In the end, psychological states evidence that receivers can discriminate the different
are private, but there are no empirical grounds to favor sequences, i.e., they are semantically meaningful to
arousal-based vocal production over concept-based them. Numerous questions remain open such as: Are
models. Whatever psychological states involved in sig- meaningful call sequences a general feature of primate
nal production, primates often behave as if their signals communication? Are there population differences in
convey meaning by referring to the external events or call sequences? How much control do primates have
inner states that trigger the signals. over call sequences, and what is the role of learning?
Most primates are forest dwellers, a habitat in which A more difficult issue is whether sequential signaling
vocal communication is especially important. Neverthe- is relevant for understanding the origins of human
less, primate vocal repertoires tend to be small, with syntax. One notion of human syntax is that its basic
a finite number of basic call types tightly linked to units (e.g., words) have their own stable and indepen-
specific biological functions. However, sometimes indi- dent meanings, something that has not been shown in
viduals produce meaningful acoustic variants within the primate examples. Also, primates do not make
some of the basic call types. For example, female much advantage of the generative power of call com-
Campbell’s monkeys regularly exchange contact calls, binations, suggesting that they have very little active
which help individuals to stay with the group in the control or cognitive understanding of these vocal
dense rainforest habitat. The calls are exchanged products.
according to some social rules and are individually dis- One key issue in vocal production is the role of
tinct. Some acoustic convergence effects have been learning. A wealth of data has shown primate vocal
reported in the structure of calls of closely affiliated repertoires are very species-specific with little acoustic
group members (Zuberbühler et al. 2009). In other variation between populations and rigid developmen-
research, acoustic variation within call types has tal patterns, suggesting that learning does not play an
580 C Cognitive Aspects of Natural Communication in Primates
important role during ontogeny, at least at the level of example, terrestrial alarm calls often do not have nar-
call morphology. Of course, the same is also true for row or fixed referents but are typically given to an array
nonlinguistic human vocalizations, although humans of events that do not always have an obvious shared
are able to mimic calls fairly accurately and at will. conceptual structure. Context then plays a key role
Another key difference is that human infants go and primates often respond very differently to the
through a babbling phase and then gradually gain con- same calls, depending on the circumstances, suggesting
trol over their vocal apparatus and learn to produce basic inferential reasoning. In one experiment, Diana
speech signals. In nonhuman primates, learning does monkeys responded differently to guinea fowl ground
play a role in developing call comprehension and call predator alarms, depending on whether the birds’
use. Individuals are generally very attentive to their alarm calls were caused by a leopard or a human, two
own and other species’ vocal signals, and appear predators that require different antipredator responses.
to have a sophisticated understanding of the meaning Similarly, baboons attend to entire exchanges of calls
of these calls. In terms of call use, young primates begin between group members, a pattern also found in chim-
by generating the different call types in the appropriate panzees. In human communication, transmission of
larger context, but then learn how to fine-tune call meaning also depends largely on the context in which
production in more detail. The classic example is the utterance is produced.
young vervet monkeys, who discriminate from the However, human communication goes beyond
beginning between aerial and terrestrial dangers but context contributing to the meaning of an utterance.
require experience to produce the alarm calls to the Human speakers seek to establish common ground
few relevant predator classes. However, very little sys- with their partners, by taking into account common
tematic research has been done, so it is not clear what knowledge, beliefs, and assumptions, while listeners
the general pattern is like. also make assumptions about the speaker’s intent.
Research on gestural communication has gener- These skills develop gradually from early childhood
ated a somewhat different picture, by showing that and become first visible in joint attentional episodes
there is much variation in the gestural repertoire during which both individuals are aware of each other’s
within different species, especially the apes. While attention to an external object (Tomasello and Carpen-
some gestures appear to be almost universal, others ter 2007). During subsequent stages of development,
can appear and disappear over time. Learning seems humans begin to monitor whether their communica-
to play a role, with signalers and receivers converging tive intentions are properly received and understood.
on what looks like ontogenetically ritualized gestural There is currently no good evidence that nonhuman
conventions (Call and Tomasello 2007). In general, primates possess the same cognitive capacities to
gestural signals are used much more flexibly than take shared knowledge and speaker intention into
vocalizations, but they are also more restricted to account during acts of communication. Yet, some
some contexts, especially play. Many gestures do not key precursor abilities are in place, such as a general
carry much meaning apart from acting as enhancers awareness of the audience and the likely consequences
of ongoing social interactions. One open problem in of producing signals. Audience awareness is particu-
gestural research is also what exactly counts as a ges- larly obvious in the gestural domain. Chimpanzees,
ture, i.e., whether a behavior in question has proper for instance, will not produce visual gestures before
signal character, either by design or intention. Other having established visual contact with the receiver.
communicative modes, especially olfactory communi- Bonobos are capable of engaging in joint activities
cation, are poorly researched and only little is known with human caregivers, in which both partners play
about the underlying cognition involved in production complementary roles, and gesture to their (human)
and perception. partners if they interrupt or are reluctant to pursue
the joint activity. In the vocal domain, chimpanzees
Inferential and Intentional Processes are aware of the composition of the audience and
Another important cognitive process in communica- the potential implications of their calls, as shown by
tion is that primates take the ongoing context into several studies (Zuberbühler et al. 2009). Whether or
account when responding to the signals of others. For not apes are willing to actively inform others about
Cognitive Aspects of Prosocial Behavior in Nonhuman Primates C 581
choice task. In the prosocial choice task, subjects are survival of the offspring. Some propose that ancestral
presented with a choice between a prosocial option that hominids were cooperative breeders, that modern
provides a single reward (often food) to himself or human minds are adapted for a cooperatively breeding
herself and to the recipient (referred to as the “1/1” environment, and that one of the ways the cooperative
option to denote that one reward is received by the breeding environment influenced our psychology was
actor and one reward is received by the recipient) and to predispose individuals to behave prosocially (e.g.,
another option which provides a reward for the actor Burkart et al. 2009).
only (the “1/0” option). The effort required of the actor Therefore, the cooperative breeding hypothesis pre-
is the same for both choices; the choices differ only dicts that prosocial preferences would be expressed not
by whether or not the recipient also receives a reward. by our closest living primate relatives the chimpanzees,
The proportion of trials on which actors choose the but instead by cooperative breeders. In the primate
prosocial option is compared with a control condition order, cooperative breeding occurs in the taxonomic
in which no recipient is present (a nonsocial control). family Callithrichidae, the marmosets, and tamarins.
Evidence of prosocial behavior is assumed if the actor Empirical support for the cooperative breeding
chooses the prosocial option more often when a recip- hypothesis was generated by presenting marmosets
ient is present to receive the reward than when there is and tamarins with the same prosocial choice task
no recipient present. that was utilized with chimpanzees. Unlike chimpan-
The resurgence of interest in nonhuman primate zees, marmosets and tamarins demonstrated prosocial
prosociality was sparked by findings indicating that preferences (e.g., Burkart et al. 2009; Cronin et al.
chimpanzees did not demonstrate prosocial behavior 2010). These findings support the hypothesis that
on this task. In fact, chimpanzees across multiple cap- there are psychological adaptations associated with
tive populations chose randomly between the two cooperative breeding that positively influence prosocial
choices, showing no increase in the prosocial response preferences.
when a partner was present compared to absent However, positive results from the prosocial choice
(e.g., Silk et al. 2005). These findings provided initial task are emerging from primate species that are not
support for the hypothesis that prosocial preferences cooperative breeders, indicating that cooperative
are uniquely human and emerged in the human breeding is not necessary for prosocial behavior (e.g.,
lineage after our ancestors diverged from the other Massen et al. 2010). Furthermore, under some experi-
great apes, or within the last six million years of evolu- mental conditions, cooperative breeders do not show
tion (hypothesis 1, above). prosocial preferences on the prosocial choice task
Positive results from additional primate species (Cronin et al. 2009). These mixed results suggest that
soon followed that suggested prosocial preferences the expression of prosocial behavior will not be
are not uniquely human and may in fact be a charac- explained by social systems or evolutionary history
teristic shared by humans and cooperative breeding alone and that prosocial behavior is dependent upon
species (hypothesis 2, above). Across primate species, a myriad of ultimate and proximate influences. Along
breeding systems can be arranged along a continuum these lines, de Waal and colleagues have proposed that
defined by which individuals bear responsibility for the proximate mechanism that elicits prosocial behav-
offspring care. At one end of the continuum are inde- ior among nonhuman primate species is empathy, or
pendent breeders. In independently breeding species, the sharing of an emotional state with another (hypoth-
care is provided nearly exclusively by the mother. esis 3, above). de Waal argues that some basic form of
This is the breeding system of most primate species, empathy is present throughout the primate order. The
including chimpanzees. However, at the other end likelihood of expressing prosocial behavior among pri-
of the continuum are cooperative breeders in which mates therefore depends on the ability to match the
many group members are actively involved in infant emotional state of the potential recipient, an ability
care, including the father, siblings, aunts, uncles and that will be affected by social factors such as the degree
sometimes unrelated individuals. Helpful behaviors of social closeness with that individual (de Waal and
by the nonbreeding individuals are essential to the Suchak 2010).
Cognitive Automatisms and Routinized Learning C 583
transition from declarative knowledge to procedural possible for a person to become aware of his or her
knowledge remains a delicate operation because the actions and, as in the case of bad habits, attempt to
automatic process can lock some know-how into tight change those behavior patterns. This question of how
procedures that are not as subject to dynamic alterna- automatic and conscious motivations interact when in
tion as circumstances may actually require. Human conflict is one of practical as well theoretical impor-
judgment is, thus, necessary to update these proce- tance, and we are now investigating parameters of this
dures, but this may occur only after a mistaken appli- interaction” (Bargh 1997, p. 52).
cation of an inflexibly automated procedure.
Bargh (1997) integrates principles of motivations as Important Scientific Research and
described in the self-determination theory. He observes Open Questions
to what extent the emotional, cognitive, and motiva- The question of consciousness in mental processes has
tional conditions that characterize an environment can always been a thorny one. Recent studies converge on
serve as the basis for a preconscious psychological state the fact that the consciousness vs. automaticity oppo-
that can generate an automatic response – automatic in sition is a dichotomy that is not clear because it seems
that it escapes the individual’s awareness and direct that consciousness and automation coexist and influence
consciousness. The underlying idea is that the routin- each other, sometimes in nonconscious ways. Psycholo-
ization of certain procedures helps an individual focus gists agree that both processes evolve together. Acknowl-
his/her attention on essential, new, and creative tasks. edging the role of consciousness in memorization
What is new here is the manner in which Bargh analyses implies recognizing that chance and the environment
motivation. Indeed, nothing happens by accident. First have a limited role. In terms of memorization, this boils
of all, before walking may become an automatic pro- down to no longer focusing all attention on the mech-
cess, we have learnt how to walk; and second of all we anisms of procedural knowledge learning, and to
intend to walk. Bargh (1997) introduces an automotive acknowledging the fact that declarative knowledge is
model to explain to what extent mental representations essential. In other words the transition from representa-
are essential to the development of cognitive mecha- tion to action is a mechanism that needs to be explained
nisms (see Fig. 1). if we are to understand how our procedural knowledge
The interactions between cognition and motivation evolves and why there is a gap between what an individ-
are therefore essential and must be taken into account. ual thinks he/she does and what he/she actually does.
Consciousness initiates the process of skill acquisition Modification of our forms of memorization must be
with possible tensions during this learning stage: “But considered in relation to changes. Individuals, as well as
even in the case of these automatic motivations, it is organizations, must learn to manage them, and to chan-
nel the emotions generated by modifications in the col-
lective representations.
One may note that the debate on routines and
Evaluative automatisms has always had a more or less positive
system
connotation because in everyday language, a routine is
regarded as automatic behavior, in contrast to
designed and implemented strategic plans. This is
the reason why Langer (1989) emphasized the notion
Environmental Motivational
Behavior of mindfulness to highlight individuals’ attention
features system
inside cognitive automatisms. In this perspective,
individuals should make sense of what they do and
perceive, by increasing their acuity so as to be able to
Perceptual integrate new information, to continuously update and
system
refine their mental categories. Indeed, the notion of
mindfulness emphasizes the necessity of focusing not
Cognitive Automatisms and Routinized Learning. so much on simple quantitative questions of data
Fig. 1 Motivation and behavior according to Bargh (1997) storing, but on the quality of the memorization.
Cognitive Conflict and Learning C 585
Experimental studies show that working groups that Langer, E. (1989). Mindfulness. Reading: Addison-Wesley.
apply this principle memorize what they learn better Shiffrin, R. M., & Schneider, W. (1977). Controlled and auto-
matic human information processing. Perceptual learning
and are more creative (Langer 1989). This principle has
automatic attending and a general theory. Psychological Review,
also been implemented in complex technological envi- 84, 127–190.
ronments so as to reduce the risk of accidents and Weick, K. E., & Sutcliffe, K. M. (2006). Mindfulness and the quality C
prevent major technological disasters (Weick and of organizational attention. Organization Science, 17(4),
Sutcliffe 2006). Potential change in routines should 514–524.
not be seen as a fateful coincidence related to external Further Reading
and disruptive factors, but as a crucial ingredient to the Johnson, E. J., Bellman, S., & Lohse, G. L. (2003). Cognitive
revitalization of individuals and organizations. This lock in and the power law of practice. Journal of Marketing,
67(2), 62–75.
leads us to reconsider the very meaning of the term
“routine” and to focus on individual and collective
memorization processes. The involvement of individ-
uals in the development of new procedural knowledge
is a delicate exercise because deliberate reasoning and Cognitive Change
mindfulness attitude, at the individual level, is a con-
▶ Mediated Learning Experience (MLE) and Cognitive
trolled, effortful, process, whereas other cognitive activ-
Modifiability
ities such as reasoning or intuition appear to be effortless
and to involve a level of automaticity. This also explains
why learning may appear to be costly at an individual
level and why the motivational dimension may play
a critical role in going beyond preexisting cognitive skills Cognitive Conflict and Learning
that are deeply entrenched in the habitual skills. Indeed
skill-based habits acquired through a trial and error MATTHEW WAXER, J. BRUCE MORTON
learning process may become increasingly automated Department of Psychology, The University of Western
as a function of the amount experience with it, creating Ontario, London, ON, Canada
some “cognitive lock-in” resisting to changes. The orga-
nizational context may provide (or not) opportunities
to go beyond these cognitive lock-in with the creation of Synonyms
systems that may facilitate learning. Cognitive dissonance; Conceptual conflict; Disequilib-
rium; Socio-cognitive conflict
Cross-References
▶ Automaticity in Memory Definition
▶ Human Cognition and Learning Cognitive conflict is a psychological state involving
▶ Individual Learning a discrepancy between cognitive structures and experi-
▶ Memory Dynamics ence, or between various cognitive structures (i.e.,
▶ Mental Effort mental representations that organize knowledge,
▶ Mindfulness and Meditation beliefs, values, motives, and needs). This discrepancy
▶ Motivation and Learning: Modern Theories occurs when simultaneously active, mutually incom-
▶ Rote Memorization patible representations compete for a single response.
▶ Routinization of Learning The detection of cognitive conflict is thought to trigger
compensatory adjustments in executive control pro-
cesses, which serve to reduce and prevent subsequent
References
instances of similar cognitive conflict.
Anderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.
Bargh, J. (1997). The automaticity of everyday life. In J. A. Bargh & Theoretical Background
R. S. Wyer Jr. (Eds.), The automaticity of everyday life (pp. 1–61). Cognitive conflict is a part of many different psycho-
Mahwah: Lawrence Erlbaum. logical theories, and has often been regarded as more
586 C Cognitive Conflict and Learning
deleterious than beneficial. For example, Freud (1901/ of the containers. This alternation provides the neces-
1953) viewed distortions of rational thinking and neu- sary conditions for the fourth step, which is simulta-
roses as the result of conflict between basic drives. neous attention to both height and width and their
Similarly, early learning-theoretic investigations of coordination into a mutually compensating system. It
conflict focused on different types of response compe- is at this point that the child recognizes the two con-
tition that lead to negative outcomes (Miller 1944). tainers contain the same amount of liquid (i.e., con-
However, other theorists such as Piaget (1977) and serve quantity).
Festinger (1957) viewed the effects of cognitive conflict Festinger’s cognitive dissonance theory shares
as playing a beneficial role in rational thinking and many similarities with Piaget’s theory of equilibration.
intellectual development, insofar as conflict drives pos- Festinger (1957) suggested that the perception of
itive cognitive adaptation. inconsistency between two simultaneously held ideas
Piaget viewed cognitive development as involving generates a state of psychological discomfort or cogni-
the attainment of successively higher states of equilib- tive dissonance. The theory of cognitive dissonance
rium or balance. Piaget proposed that the mechanism holds that individuals have a motivational drive to
of transition from one state of equilibrium to another resolve dissonance by either changing their beliefs,
was the process of equilibration. According to Piaget, attitudes and behaviors, or rationalizing their beliefs,
this process is fueled by conflict or “disequilibrium,” attitudes, and behaviors. For example, it is widely
either between cognitive structures and experience or accepted that smoking is associated with a greater
between various cognitive structures. Disequilibrium probability of developing lung cancer. At the same
then motivates an individual to resolve the conflict time, most individuals desire to live a healthy life. On
and attain a new state of equilibrium. this account, the desire to live a healthy life is dissonant
One example used to illustrate the processes of with engaging in activities that will most certainly
equilibration is the acquisition of conservation of con- shorten one’s life. The conflict produced by simulta-
tinuous quantity. A child is presented with two identi- neously holding these contradictory ideas may be
cal beakers that have been filled to exactly the same level reduced by quitting smoking, or rationalizing one’s
with juice; one is identified as belonging to the child smoking.
and the other to the experimenter. After the child has The ability to recognize and learn from instances of
acknowledged that the amount of juice is the same in cognitive conflict is an important evolutionary adapta-
each beaker, the experimenter pours the contents of tion, and as such, understanding the biological systems
one jar into a short, broad container and that of the that underpin this ability remains an important line of
other into a tall, thin one. The experimenter then asks research. Recent theoretical advances in cognitive neu-
the child if the containers contain different amounts of roscience have started to shed light on the underlying
liquid or the same amount. If the answer is the “same neural mechanisms of cognitive conflict and its resol-
amount,” the participant is said to have “conserved” ution. One theory that has garnered a considerable
the substance of the liquid; and with respect to this amount of attention is the conflict monitoring theory
problem, the child’s thinking has reached a new state of (for review see Botvinick et al. 2004). On this account,
equilibrium. specific subsystems of the human brain detect instances
According to Piaget, all equilibration processes go of conflict in information processing, particularly
through four steps. In the Step 1, the child attends to response competition, and then engage other executive
only one dimension (e.g., the height of the container), brain regions to diminishing conflict in succeeding
and judges the tall drink to contain more liquid (i.e., time intervals. The anterior cingulate cortex (ACC) is
fails to conserve quantity). With repeated experience thought to be the monitoring center that is responsi-
on similar problems, in Step 2 the child then focuses ble for the online detection of response conflict. The
on the opposite dimension (e.g., the width of the conflict signal that is detected by the ACC is then
container), and judges the broad container to contain transmitted to other brain regions, such as the dorso-
more liquid. The third step may be viewed as a mixture lateral prefrontal cortex (DLPFC), to increase the level
of the first two steps. More specifically, the child will of cognitive control and reduce the amount of cogni-
now alternate responses between the two dimensions tive conflict.
Cognitive Developmental Robotics C 587
Important Scientific Research and these two bodies of literature begin to emerge when
Open Questions the findings of single cell recording studies are taken
Many empirical investigations of the effects of cognitive into consideration. Recording studies in nonhuman
conflict in human participants have shown that when primates have failed to find any evidence of conflict-
conflict arises between behavioral responses in experi- related signals in the ACC (Mansouri et al. 2009). C
mental tasks, performance is adversely affected in terms Reconciling these differences remains an important
of speed and accuracy. For example, in the Stroop task, challenge for future research.
participants are presented with the name of a color Many psychological theories, including develop-
printed in colored ink. The participant’s task is to mental, social, clinical, and cognitive neuroscientific,
identify the color of the ink as quickly and accurately have emphasized the importance of cognitive conflict.
as possible. On high-conflict trials, when the color’s Despite the importance of cognitive conflict in many
name differs from the ink color, participants are slower different psychological theories, the development of
and less accurate than on low-conflict trials, in which a unifying theoretical framework remains an important
the color name and ink match one another, or than on challenge for researchers.
neutral trials, in which the word is not color-related.
A large corpus of neuroimaging studies in humans
Cross-References
▶ Cognitive Dissonance in the Learning Processes
using event-related potential (ERP) recordings, and
▶ Metacognition and Learning
functional magnetic resonance imaging (fMRI) have
reported activation of the ACC to be greater in high- References
conflict conditions relative to low-conflict or neutral Botvinick, M., Braver, T. S., Yeung, N., Ullsperger, M., Carter, C. S., &
conditions during the performance of different tasks Cohen, J. D. (2004). Conflict monitoring: Computational and
designed to elicit conflict (cf. Botvinick et al. 2004; empirical studies. In M. I. Posner (Ed.), Cognitive neuroscience of
Mansouri et al. 2009). attention (pp. 91–104). New York: Guilford.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford:
Although the behavioral effects of conflict have
Stanford University Press.
been typically associated with decrements in speed Freud, S. (1953). The psychopathology of everyday life. In J. Strachey
and accuracy, this relationship is dynamically modu- (Ed.), The standard edition of the complete psychological works of
lated by previous experience with conflict. More spe- Sigmund Freud (Vol. 6). London: Hogarth Press [Original work
cifically, response latencies on high-conflict trials that published in 1901.].
Mansouri, F. A., Tanaka, K., & Buckley, M. J. (2009). Conflict-induced
are immediately preceded by high-conflict trials are
behavioural adjustment: A clue to the executive functions of the
shorter than those on high-conflict trials that are prefrontal cortex. Nature Reviews Neuroscience, 10, 141–152.
immediately preceded by low-conflict trials. Addi- Miller, N. E. (1944). Experimental studies of conflict. In J. McV Hunt
tionally, conflict-related ACC activation has been (Ed.), Personality and the behavior disorders (Vol. 1). New York:
shown to be modulated by preceding conflict, with Ronald.
greater ACC activation observed on high-conflict Piaget, J. (1977). The development of thought: Equilibration of cognitive
structures. New York: Viking.
trials that were preceded by low-conflict trials rela-
tive to high-conflict trials that were preceded by high-
conflict trials. The facilitative effect of previous
experience with conflict has been referred to as the
“conflict adaptation effect”; and has been observed
Cognitive Curiosity
across a wide range of conflict tasks (cf. Botvinick ▶ Curiosity and Exploration
et al. 2004; Mansouri et al. 2009). ▶ Epistemic Curiosity
Cognitive conflict appears to be a ubiquitous phe-
nomenon that can also be observed in non-human
primates and other animals. For example, studies
using nonhuman primates tested on analogs of con- Cognitive Developmental
flict tasks used in human research have shown Robotics
similar behavioral responses to conflict as humans (cf.
Mansouri et al. 2009). However, discrepancies between ▶ Developmental Robotics
588 C Cognitive Disequilibrium
learning resulting in the construction of new knowl- discrepancy and completing the procedure. When
edge structures. As an added benefit to the learning using impasse-driven learning to design learning envi-
process, the motivational aspects of resolving cognitive ronments, the quality of the information used by the
dissonance create an environment where learners are learners to continue with the procedure is of critical
continually exposed to content-relevant information importance. Inaccurate information used to repair C
facilitating deeper processing. an impasse will result in internalized misconceptions
of procedural knowledge. Historically, these miscon-
Applying Cognitive Dissonance to ceptions are difficult to repair and can result in persis-
Learning Environments tent misunderstandings.
As the process of knowledge acquisition involves inte-
grating new knowledge with existing schema, allowing Important Scientific Research and
learners to be in a state of cognitive dissonance is ideal Open Questions
for new learning. Theories of cognitive dissonance can The concepts surrounding cognitive dissonance are
be applied to both problem-based learning and proce- a foundational element in learning processes and
dural learning. can occur no matter what the knowledge level of a
Probably the most natural instructional environ- particular learner. An awareness that this phenomenon
ments in which to study the phenomenon of cognitive exists and the processes used by learners to resolve these
dissonance are ones that employ problem solving. In discrepancies is critical to understanding learning pro-
problem-solving exercises, learners are presented with cesses. Designers interested in using the positive effects
information and are asked to use their knowledge to of cognitive dissonance should focus on the creation of
extract the correct information and solve the problem. situations where learners can satisfy their internal need
As soon as learners are presented with the components to resolve dissonant information thereby increasing
of a problem (problem state, goals, operators), they their deep processing of the content.
begin the process of resolving conflicting information, With the abundance of research in advanced learn-
selecting relevant information needed to solve the prob- ing technologies, adaptive systems, simulation, and
lem, and constructing a procedure to solve the problem. game-based environments that require instructional
The intrinsic human need to move from disequilibrium approaches through problem solving, research in the
to equilibrium creates a constant process of examining effects of cognitive dissonance on learning processes is
and reexamining information until a satisfactory solu- ongoing. Design-based research (Barab and Squire
tion is reached. This trial-and-error process leading to 2004) is a methodological approach that proposes the
insight is a cornerstone for the design of game-based design of environments to specifically verify theories of
learning environments (Van Eck 2007). One key consid- learning and the effects of instructional design on the
eration in the design of these environments is to under- learning process. Because cognitive dissonance is
stand the relationship between the level of cognitive closely related to problem solving, the design and
dissonance and the motivation to solve problems. evaluation of problem-based learning environments
Learners are quickly bored with a level of dissonance provides a perfect opportunity to test and validate
that is too easily resolved but on the other hand can be assumptions about cognitive dissonance and the pro-
frustrated with a level of dissonance that is too high. cesses of learning.
Cognitive dissonance can also be used to promote Several specific questions can be addressed through
procedural learning. The impasse-driven learning the- design-based research to verify the effects of cognitive
ory (van Lehn 1988) is defined as a point in which dissonance on learning and motivation to learn. For
learners are presented with a procedural step that can- example, one might assess whether the level of cogni-
not be accomplished due to a discrepancy in their tive dissonance has positive or negative motivational
knowledge base. This theory has been used to propose effects on learners. Variables such as level of challenge
designs for expert systems in procedural domains such and affordances can be manipulated to increase or
as mathematics. After reaching an impasse, learners decrease levels of cognitive dissonance contributing to
go through a repair-and-reflect cycle replacing the a deeper understanding of motivational issues such as
590 C Cognitive Dissonances
Cross-References
▶ Complex Problem Solving Synonyms
▶ Designing Learning Environments Mental efficiency; Optimal thinking; Problem-solving
▶ Development of Team Schemas efficiency
▶ Emotional Learning
Definition
▶ Emotions in Cognitive Conflicts
Cognitive efficiency (CE) is a multifaceted construct
▶ Games-Based Learning
that describes the ability to reach learning, problem
▶ Motivation and Learning
solving, or instructional goals through optimal use of
▶ Problem Solving
mental resources. CE can be defined as optimal effort
▶ Procedural Learning
needed to perform a task, optimal performance on
▶ Schema Development
a task, or as the relationship between maximum per-
▶ Schema-Based Learning
formances on a task while exerting minimum effort
▶ Schema-Based Problem Solving
(Hoffman and Schraw 2010).
References In general, all views construe CE as the tradeoff
Barab, S., & Squire, K. (2004). Design-based research: Putting a stake between benefits such as increases in the rate, amount,
in the ground. Journal of the Learning Sciences, 13(1), 1–14. or conceptual clarity of knowledge versus costs such as
Elliot, A. J., & Devine, P. G. (1994). On the motivational nature of time, effort, or the cognitive resources expended to
cognitive dissonance: Dissonance as a psychological discomfort. complete a task. Three main criteria influence the
Journal of Personality and Social Psychology, 67(3), 382–394.
understanding and utility of CE: the discipline of appli-
Festinger, L. (1957). A Theory of cognitive dissonance. Stanford, CA:
Stanford University Press. cation, measurement of the construct, and individual
Piaget, J. (1975). The equilibration of cognitive structure. Chicago, IL: differences among learners.
University of Chicago Press.
Van Eck, R. (2007). Six ideas in search of a discipline. In B. Shelton & Theoretical Background
D. Wiley (Eds.), The design and use of simulation computer games Beginning in the late nineteenth century, experimental
in education. Rotterdam, The Netherlands: Sense Publishing. and behavioral psychologists such as Ebbinghaus,
van Lehn, K. (1988). Toward a theory of impasse-driven learning. Hull, and Thorndike conducted research using tasks
In H. Mandl & A. Lesgold (Eds.), Learning issues for intelligent
such as maze learning, the memorization of nonsense
tutoring systems. Berlin: Springer.
Zanna, M., & Cooper, J. (1974). Dissonance and the attribution
symbols, and learning word lists in an attempt to
process. In J. Harvey, W. Ickes, & R. Kidd (Eds.), New directions explain individual differences in efficient cognition.
in attribution research (pp. 199–217). Hillsdale, NJ: Erlbaum. These researchers concluded that CE was based on the
amount of time needed to complete learning, and sig-
nificant within-person variability on tasks determined
the relative efficiency of learning conditions. Concur-
Cognitive Dissonances rently, efficiency research in diverse disciplines includ-
ing industry, economics, and management examined
▶ Discontinuities for Mental Models what methods and conditions fostered productive
Cognitive Efficiency C 591
outcomes while minimizing waste. Combined, these (Neubauer and Fink 2009). Additionally, the influence
findings have lead to a focus on CE research in three of gender is related to greater neural efficiency with
primary disciplines: philosophy, neurobiology, and female dominance on verbal tasks, and male superior-
education. ity on spatial tasks.
C
Philosophy Education/Psychology
Spawned by the “efficiency movement” in the early CE research in education and psychology is focused in
1900s, and popularized by Frederick Taylor’s work several diverse areas including problem-solving effi-
(1911) in scientific management, the philosophical ciency in mathematics, verbal efficiency in reading,
view of CE combined psychological and sociological and strategy efficiency across domains. Problem-
perspectives. This systemic approach stated that individ- solving efficiency is the ability to arrive at accurate
ual competence cannot be achieved without efficiency, problem solutions with minimal effort or time (Schraw
and productive cultures are based upon the moral obli- and Hoffman 2010). Verbal efficiency emphasizes the
gation of citizens to maximize effort and avoid wasting ability to quickly decode text and enhance reading
human resources. According to this view, maximal pro- comprehension, while strategy efficiency focuses on
ductivity influences all aspects of life, including applica- how effectively learners apply strategies to solve prob-
tion of efficiency principles to the science of education. lems quickly and accurately.
Research grounded in the discipline of philosophy places Much of the CE research in education and psychol-
strong emphasis on measuring teacher competence and ogy has investigated the effectiveness of instructional
attempts to quantify educational efficiency by determin- design and pedagogy, or focused on determining
ing optimal teaching methods. what factors influence information processing during
learning and problem solving. The characteristics of
Neuroscience instructional materials such as the complexity of the
Neurologically, CE is assessed by the frequency, speed, information and the presentation modality affect CE.
and location of prefrontal cortical activity as measured Grounded in cognitive load theory (Sweller et al. 1998)
by brain imaging technology such as functional mag- cognitive efficiency is constrained due to the limited
netic resonance imaging (fMRI) and positron emission capacity of working memory to process and store
tomography (PET), which detects changes in cerebral information simultaneously. Information that is
blood flow or neural activation. Individuals with faster intrinsically complex and presentation modalities that
and more localized brain connectivity are deemed require learners to engage in extraneous processing
neurologically more efficient when fewer cognitive (e.g., embedding descriptive labels for a diagram in
resources are used and less energy is expended to cor- text rather than near the diagram) reduce CE because
rectly solve cognitive processing tasks such as digit- they create excessive burdens on processing resources
symbol substitution or spatial reasoning tasks, for and can interfere with learning.
example, the Raven Progressive Matrices test. Slower
processing and greater neurological activity is deemed Measurement of CE
less efficient and typically associated with lower intel- Three primary methods are used to measure CE, each
ligence and diminished performance (Neubauer and with different computational formulas (Hoffman and
Fink 2009). Schraw 2010). Studies investigating instructional effi-
Individuals who complete tasks faster and with ciency typically measure differences between perfor-
greater accuracy have lower brain activation and higher mance and effort. These studies convert raw effort
brain alpha levels, meaning they achieve superior task and performance scores obtained when completing
performance with less cognitive effort. However, many tasks to standardized scores and measure the difference
empirical ambiguities exist with the typical negative between control and experimental groups similar to
relations between brain activity and performance calculating effect sizes. For example, according to this
attenuated for complex tasks. Other variables including method of measurement, if two individuals have the
structured practice and adaptive strategy use mediate same test score, the individual that spent less time or
brain activation and increase neurological efficiency effort is deemed to be more cognitively efficient.
592 C Cognitive Efficiency
The second method is processing efficiency, which is memory. Other person variables that influence CE
a measure of the ratio of performance (i.e., accuracy or include metacognitive awareness and motivation.
number of errors) divided by cost (i.e., time or effort) Metacognitive awareness involves knowing what
between participants in different groups (e.g., experi- strategies to use, and how and when to use them.
mental and control groups). The primary focus is on Furthermore, learners’ motivation to use strategies
rate of change, or change relative to the amount of influences CE. So domain knowledge, metacogni-
effort or time that was needed to achieve accuracy. tive awareness, and motivation can help students
A student could either complete a task with greater become more cognitively efficient, even if they have
accuracy, or with less time or effort, and be considered lower WMC.
cognitively efficient. This method differs from the first Instructional variables, including the quality of
method as the construct of interest is the rate of change, instructional materials and the presentation format of
not the difference in change. to-be-learned information, also influence CE. Mate-
A third approach holds a factor constant (e.g., rials enhance CE when individuals can expend less
background knowledge) and uses existing CE measures effort and achieve relatively higher performance out-
to predict future outcomes of efficiency, similar to comes. For example, when solving problems using
using a covariate for statistical control. For example, worked examples, a modeled sample problem, learners
in a situation where an athlete is running an obstacle are more successful than when merely asked to solve
course, the athlete’s efficiency using the ratio of time to a problem without aids. Overly complex or poorly
distance may be the same or even worse than the designed materials lower CE because learners waste
previous run. Using previous information can be help- valuable working memory resources deciphering mate-
ful to determine differences in CE over time by calcu- rials and thus devote less attention to learning.
lating the relative gain after additional instruction or CE can be inhibited even when learners possess
practice. This method of measurement differs from necessary domain knowledge and WMC. Individuals
those previously described as it calculates the condi- lacking in awareness or the motivation to use auto-
tional rate of change from an existing level of CE to mated strategies may forego efficient problem solving
measure relative gain, when some relevant prior mea- in favor of more time-consuming methods such as
sure of practice or learning is considered. calculation. Similarly, effort can influence the potency
of cognitive resources dedicated toward a task, with
Important Scientific Research and greater effort associated with more complex tasks and
Open Questions a reduction in CE. The extra effort strains WMC
Person variables, such as working memory resources resulting in longer problem-solving time, thus reducing
and domain knowledge, influence CE. Learners have efficiency. Although employing more effort usually
limited working memory resources, which means they impedes CE, overconfidence in problem-solving success
can concurrently process and store limited amounts of can result in withholding effort. Individuals anticipat-
information at any given time. Thus, the availability of ing success may not try as hard as usual, miserly appro-
working memory resources influences how much priating effort, resulting in lower performance, and
information learners can process, how quickly they ultimately reducing CE.
can process it, and the strategies they use to process it The mode of presentation, context of learning, and
(Hambrick and Engle 2003). For instance, individuals the nature of pedagogy may influence CE. Some mate-
with greater working memory capacity (WMC) typi- rials are more suitable to the schemata of experts than
cally solve problems more accurately and more effi- novices and instructional methods such as discovery
ciently than individuals with lesser WMC (Mayer and learning can be counterproductive (Kirschner et al.
Wittrock 2006). 2006). These facets of CE assume knowledge acquisi-
Domain knowledge influences CE. When individ- tion and problem-solving ability of a more seasoned
uals have knowledge that is deep, well-structured, learner can be encumbered by information which is
and schematically well-organized, they think more redundant or unnecessary. Materials or methods pro-
efficiently, use strategies judiciously, and are better viding information ancillary to learning can create
able to retrieve information from long-term cognitive congestion, lowering CE.
Cognitive Instruction C 593
RICHARD E. MAYER
Department of Psychology, University of California, Theoretical Background
Santa Barbara, CA, USA The science of learning is the scientific study of how
people learn (Mayer 2011). Over the past 120 years,
researchers have developed three conceptions of how
Synonyms learning works – response strengthening, information
Knowledge change; Learning acquisition, and knowledge construction. According
to the response-strengthening view, learning involves
the strengthening or weakening of an association
Definition
between a stimulus and a response, in which responses
Cognitive learning is a change in knowledge attributable
that are followed by satisfaction are strengthened and
to experience (Mayer 2011). This definition has three
responses that are followed by dissatisfaction are
components: (1) learning involves a change, (2) the
weakened. The instructor is a dispenser of rewards
change is in the learner’s knowledge, and (3) the cause
and punishments whereas the learner is a passive re-
of the change is the learner’s experience. An example of
cipient of rewards and punishments. The response-
cognitive learning includes being able to give the defini-
strengthening view reached prominence in the first
tion of cognitive learning after reading this entry.
half of the twentieth century, and is reflected in classic
Cognitive learning can be distinguished from
research by Thorndike (1911/1965) on trial and error
behavioral learning on the basis that cognitive learning
learning by cats.
involves a change in the learner’s knowledge whereas
According to the information acquisition view,
behavioral learning involves a change in the learner’s
learning involves adding new information to memory,
behavior. However, a change in knowledge (i.e., cogni-
in which the amount of practice or time spent studying
tive change) must be inferred from the learner’s behav-
is related to the amount of information learned. The
ior (i.e., behavioral change), so cognitive learning is
instructor is a dispenser of information and the learner
closely related to behavioral learning.
is a passive recipient of information. The information
Knowledge change is at the heart of cognitive learn-
acquisition view reached prominence in the 1960s and
ing; so it is useful to distinguish among five kinds of
1970s in conjunction with the information-processing
knowledge (Mayer 2011):
revolution in cognitive psychology, and has its roots in
Facts – factual knowledge about the characteristics classic research by Hermann Ebbinghaus (1885/1964)
of things, such as knowing that the numeral “5” on the role of practice in memorizing lists of nonsense
corresponds to the word “five” syllables.
Cognitive Learning C 595
According to the knowledge construction view, memories have unlimited capacity to hold sensory
learning is an active process of sense making in which representations for very brief periods (i.e., less than
the learner constructs a mental representation by 1 s). The second row represents working memory in
selecting relevant incoming information, mentally which selected aspects of incoming sounds and images
organizing it into a coherent structure, and integrating from sensory memory are mentally organized into C
it with appropriate prior knowledge. The instructor is coherent cognitive verbal and pictorial representations,
a cognitive guide who helps the learner engage in respectively. However, working memory capacity is
appropriate cognitive processing during learning, and severely limited; so only a small amount of cognitive
the learner is an active sense maker. The knowledge processing can take place within each channel at any
construction view has been prominent since the 1980s, one time, and information that is not processed decays
and has its roots in classic research by Frederic Bartlett quickly (i.e., in less than 1 min). Finally, the third row
(1932) on learning and memory as constructive activ- represents long-term memory, which is the learner’s
ities that depend on the learner’s existing knowledge. storehouse of knowledge – a memory store for knowl-
Although all three conceptions of learning are still edge with nearly unlimited capacity and long duration.
influential today, they may be most relevant for differ- These distinctions are consistent with the limited-
ent kinds of learning situations. capacity principle, which holds that learners can process
only a small amount material at any one time in work-
Important Scientific Research and ing memory.
Open Questions Concerning cognitive processes, the arrows repre-
Figure 1 presents a framework for cognitive learning, sent the three major kinds of cognitive processing
which consists of two channels, three memory stores, required for cognitive learning – selecting, organizing,
and three cognitive processes (Mayer 2009). Concer- and integrating. Selecting occurs when learners attend
ning channels, the top row represents the auditory/ to aspects of the incoming information in sensory
verbal channel whereas the bottom row represents the memory for further processing in working memory,
visual/pictorial channel. This distinction is consistent as indicated by the arrows from sensory memory to
with the dual-channel principle, which holds that working memory (i.e., selecting words and selecting
learners have separate channels for processing verbal images). Organizing occurs when learners mentally
and visual material. arrange verbal elements into a coherent verbal repre-
Concerning memory stores, the first row represents sentation (indicated by the organizing words arrow)
sensory memory in which incoming spoken words and mentally arrange pictorial elements into a coherent
impinge on the ears and are held in acoustic form for pictorial representation (indicated by the organizing
a very brief time within auditory sensory memory images arrow). Integrating occurs when learners acti-
whereas incoming pictures and printed words impinge vate relevant knowledge from long-term memory
on the eyes and are held in visual form for a very brief and connect it with incoming information in work-
time within visual sensory memory. These sensory ing memory (as indicated by the integrating arrow).
students encounter difficulties when learning from and Robinson 1972). Numerous principles have been
combinations of different verbal and pictorial repre- identified on how to design texts in a manner, which
sentations. Such combinations not only offer various support students’ learning. These principles address
learning opportunities for students, but also place issues of content as well as structure and layout. No
increased demands on the students. For instance, stu- one assumes, however, that texts designed according to C
dents need to understand (1) how information is these principles guarantee that students will learn suc-
encoded in each single representation, (2) how each cessfully. Rather, students are taught – from the ele-
representation is related to the subject domain, and mentary to the university level – reading and learning
(3) how information in one representation can be strategies which take the specific characteristics of texts
related to information in another representation (cf. into account. These strategies involve both internal
Ainsworth 2006). Thus, students not only have to learn learning activities (e.g., paraphrasing text segments)
how to identify the relevant components of verbal and and external learning activities (e.g., highlighting text
pictorial representations, but how to relate them to segments). Thus, after many years of education, the
each other as well. If the representations are dynamic, students have acquired and exercised a number of
students must also learn how to identify and relate internal and external techniques which help them to
both spatially and temporally separated compo- systematically approach particularly complex and dif-
nents. Interactive representations place even more ficult texts.
demands on the students in that they need to plan, If empirically evaluated strategies for learning from
to monitor, and to evaluate their interactions with the texts are available, but there exist almost no strategies
representations. for learning from other external representations, then
One approach to support learning from different one obvious approach to conceptualizing strategies for
static and dynamic representations is the principled learning from other representations is to draw upon the
design of digital media. Based on theories and models strategies for learning from texts. However, strategies
of human learning, such as Richard Mayer’s (2001) for learning from texts cannot be directly mapped onto
theory of multimedia learning, this approach essen- strategies for learning from other representations.
tially aims at designing digital media in a way that Because each external representation has its own char-
make the identification and selection, as well as the acteristics and places its own demands on learners,
organization and integration of information as easy as a conceptual model that mediates such a mapping is
possible for students. Examples of important design needed. Theories on multimedia learning create
principles are the multimedia principle, the split- a promising starting point for formulating the required
attention principle, and the modality principle (cf. conceptual model.
Mayer 2005). Several studies have demonstrated that Mayer’s (2001) theory of multimedia learning
the principled design of digital media facilitates learn- emphasizes four different kinds of cognitive processes:
ing. Over the past 10 years, research on learning from selection, organization, transformation, and integra-
digital media has focused on this approach. tion of information. Selected textual and pictorial
Another approach to improve learning from exter- information is initially processed in separate channels.
nal representations is the principled design of learn- Subsequently, the selected information is then orga-
ing strategies. Also based on theories and models of nized into two separate models: one model for verbal
human learning, it aims at empowering students to information and one model for pictorial information.
initiate, plan, organize, monitor, and regulate their During information processing, verbal representations
own learning and to competently deal with challeng- may be transformed into pictorial representations (e.g.,
ing learning material. With respect to learning from by constructing mental images) and vice versa (e.g., by
digital media, this approach has been largely neglected internally verbalizing images). In order to make mul-
up until now. timedia learning successful, both models need to be
One example in which research on the design of integrated and related to prior knowledge.
external representations has been successfully coupled If strategies for learning from multimedia are to be
with research on the design of cognitive learning strate- based on strategies for learning from text, the models
gies is when applied to learning from texts (e.g., Thomas can support this conceptualization in two different
598 C Cognitive Learning Strategies for Digital Media
ways. First, learning techniques used in strategies for investigated in order to analyze the learning effective-
learning from text can be categorized according to the ness of the strategy. One group of students learned
cognitive processes which they aim to induce. Subse- without the strategy whereas another group of students
quently, analogous techniques for learning from mul- learned with the strategy. It was demonstrated that the
timedia have to be constructed in such a way that they students who employed the strategy attained signifi-
stimulate the same cognitive processes. In this case, the cantly better learning results with medium to large
model serves as a synthetic aid for “mapping” tech- effect sizes.
niques, which have been designed for learning from The strategies proposed by Kombartzky et al.
one representational system to those techniques (2010) and Schlag and Ploetzner (in press) are two
designed for learning from another representational examples of cognitive strategies for learning from dif-
system. Secondly, once a learning strategy is available, ferent combinations of digital media. Additional exam-
the learning techniques employed within the strategy ples of such strategies are self-explaining while learning
can be categorized, as described above, in order to from text and pictures (e.g., Ainsworth and Loizou
determine whether each of the cognitive processes is 2003) and guided discovery learning while learning
promoted by a corresponding learning technique. In from interactive simulations (e.g., de Jong and van
this case, the model serves as an analytic aid in order to Joolingen 1998). However, there is much potential for
verify that all four kinds of cognitive processes are further research on strategies for learning from digital
supported by the strategy. media. For instance, a learning strategy can be provided
to the students in many different ways. The complete
Important Scientific Research and strategy can either be presented to the students at once
Open Questions on a worksheet or the students can be prompted incre-
Based on the conceptual model described above, mentally and adaptively for single learning techniques
Kombartzky et al. (2010) proposed a cognitive strategy when they are working on specific parts of the learning
for learning from animations and spoken text. Two material. Currently, we do not know which possibility
different experimental studies were conducted in is more beneficial to learning.
order to evaluate the strategy. In the first study, one In the long run, one also needs to investigate whether
group of students learned from an animation without the learning strategies can be taught to students in such
the strategy, whereas a second group of students was a way that the students internalize the strategies step by
encouraged to make use of the proposed strategy dur- step and then automatically apply them to new learning
ing learning. The use of the strategy was not monitored. situations. This commonly requires the training of
The students who were encouraged to take advantage learning strategies over a longer period of time. Research
of the strategy learned significantly better than the indicates that the use of a newly acquired, but not yet
students who were not asked to do so. In the second automatized learning strategy demands a great deal of
study, three groups of students were investigated. The mental effort and might therefore – temporarily – even
first group learned from an animation without the impede learning. Only after a longer period of training
strategy. The second group was encouraged to make does it become easier to apply the strategies, hence
use of the strategy during learning but use of the strat- learning improves.
egy was not monitored. The third group was also There might also be potential for optimizing the
encouraged to make use of the strategy during learning proposed strategies. On the one hand, we need to better
and their use of the strategy was monitored. The results understand how the learning techniques employed in
of the second study replicated the findings of the first the strategies contribute to learning success. Are the
study. Furthermore, learning was most successful when learning techniques of equal importance or could some
the students’ use of the learning strategy was moni- of the learning techniques be neglected? On the other
tored. The effect sizes are medium to large. hand, only processes at the cognitive level are currently
On the basis of the same conceptual model, Schlag induced by means of the strategies. Various studies,
and Ploetzner (in press) developed a cognitive learning however, indicate that learning might be even more
strategy in order to support learning from written text successful if processes at the metacognitive level were
and static pictures. Two groups of students were also taken into account. It could therefore be of interest
Cognitive Load Measurement C599
to further investigate whether or not it is beneficial to
complement the proposed learning techniques at the Cognitive Load
cognitive level with learning techniques at the ▶ Mental Effort
metacognitive level.
C
Cross-References
▶ Animation and Learning
▶ Audio-Visual Learning Cognitive Load Measurement
▶ Cognitive and Affective Learning Strategies
▶ Learning Strategies TAMARA VAN GOG1, FRED PAAS1,2
1
▶ Multimedia Learning Institute of Psychology, Erasmus University
▶ Representational Learning Rotterdam, Rotterdam, The Netherlands
2
▶ Strategic Learning University of Wollongong, Wollongong, Australia
References Synonyms
Ainsworth, S. (2006). DeFT: A conceptual framework for considering Measurement of working memory load; Workload
learning with multiple representations. Learning and Instruction,
measurement
16, 183–198.
Ainsworth, S., & Loizou, A. (2003). The effects of self-explaining
when learning with text or diagrams. Cognitive Science, 27, Definition
669–681. Cognitive load can be defined as the load imposed on
de Jong, T., & van Joolingen, W. R. (1998). Scientific discovery an individual’s working memory by a particular (learn-
learning with computer simulations of conceptual domains.
ing) task. It can be measured using various techniques.
Review of Educational Research, 68(2), 179–201.
Kombartzky, U., Ploetzner, R., Schlag, S., & Metz, B. (2010). Devel-
oping and evaluating a strategy for learning from animation. Theoretical Background
Learning and Instruction, 20(5), 424–433. Cognitive load theory (CLT) is discussed extensively
Mayer, R. E. (2001). Multimedia learning. New York: Cambridge
elsewhere in this Encyclopedia, and therefore not
University Press.
Mayer, R. E. (Ed.). (2005). The Cambridge handbook of multimedia
repeated in detail here. What is important to note for
learning. New York: Cambridge University Press. cognitive load measurement, though, is that the intrin-
Schlag, S., & Ploetzner, R. (in press). Supporting learning from illus- sic load imposed by a learning task results from both
trated texts: Conceptualizing and evaluating a learning strategy. task and learner characteristics. The higher the number
Instructional Science. of novel interacting information elements a task con-
Streblow, L., & Schiefele, U. (2006). Lernstrategien im Studium.
tains, the higher the intrinsic cognitive load it imposes
[Learning strategies in academic studies]. In H. Mandl &
H. F. Friedrich (Eds.), Handbuch Lernstrategien [Handbook on working memory. With increasing practice, ele-
of learning strategies] (pp. 352–364). Göttingen: Hogrefe Verlag. ments are combined or chunked into a schema, which
Thomas, E. L., & Robinson, H. A. (1972). Improving reading is stored in long-term memory and can be retrieved
in every class: A source-book for teachers. Boston: Allyn and and handled in working memory as a single informa-
Bacon.
tion element. Because schemata can be handled as
a single element, the same task imposes less cognitive
load for people who have had more practice than for
people who are unfamiliar with the task, that is, their
Cognitive Learning Strategy performance is more efficient. Therefore, measuring
cognitive load next to the more traditional perfor-
Goal-directed mental activities aimed at enhancing mance measures (e.g., accuracy, number, or type of
one’s knowledge and skill. Examples of cognitive learn- errors), before, during, or after a learning phase, can
ing strategies include summarizing, outlining, concept provide additional information on the level of exper-
mapping, creating analogies, generating elaborations, tise of a learner or group of learners relative to that of
sub-goaling, self-questioning, etc. other learners.
600 C Cognitive Load Measurement
Next to intrinsic cognitive load, the way in which or (an adapted version of) the 9-point symmetrical
the task was designed or presented to the learner category mental effort rating scale developed by Paas
may affect cognitive load. In this case, measuring cog- (for reviews, see Paas et al. 2003; Van Gog and Paas
nitive load in combination with performance can – 2008). This mental effort rating scale asks students
at least when the level of intrinsic load is kept to indicate “how much mental effort did you invest
constant – provide information on the effects of differ- in solving this problem?” (or “. . .in studying this
ent task designs relative to each other. For example, example,” or “. . .in completing this task”), with answer
when we know that two groups of learners (A and B), of options ranging from (1) very very low mental effort to
equal levels of expertise (i.e., materials will impose (9) very very high mental effort. Mental effort is
a comparable intrinsic load), both experience the defined as “the aspect of cognitive load that refers to
same level of cognitive load during learning with two the cognitive capacity that is actually allocated to
different instructional formats, say A (Group A) and accommodate the demands imposed by the task; thus,
B (Group B), we do not know very much. However, it can be considered to reflect the actual cognitive load”
if we know that the learning outcomes of Group (Paas et al. 2003, p. 64). To illustrate the difference
B were higher than those of Group A, we can conclude between “objective” cognitive load (e.g., as defined by
that the cognitive load they experienced must have the number of interacting information elements) and
resulted from different cognitive processes: The load actual cognitive load as measured by mental effort: if
experienced by Group B was imposed by cognitive a task is very high in intrinsic load, but the learner does
processes that were more effective for learning than not allocate any cognitive capacity to the task (i.e., does
those in Group A. Or alternatively, if Groups A and not engage in it, which can be the case, e.g., when
B had obtained the same test scores, but Group learners perceive a task as being too difficult), the task
A experienced more cognitive load during learning will not actually impose any cognitive load on the
than Group B, the learning process of Group B was learner’s working memory. Subjective ratings are usu-
more efficient (Van Gog and Paas 2008). ally collected immediately after each task, in which case
In sum, cognitive load is the load imposed on they do not give insight into fluctuations in load over
working memory by the cognitive processes that a time. They can also be applied repeatedly during the
(learning) task evokes. It can be measured at different task, in which case, some information on fluctuations
levels. Xie and Salvendy (2000) distinguish between in load is available.
instantaneous load, peak load, average load, accumu- A more objective way of measuring cognitive load
lated load, and overall load. Instantaneous load reflects is the use of secondary-task procedures, in which the
the dynamics of cognitive load, which fluctuates every amount of load imposed by the primary (learning)
moment during execution of the (learning) task. Peak task is measured by the performance or response time
load is the maximum value of instantaneous load while on a secondary task: the higher the load imposed by
working on the task. Accumulated load is the total the primary task, the less cognitive capacity is available
amount of load that the learner experiences during for attending to the secondary task, and as a conse-
a task. Average load represents the mean intensity of quence, response to the secondary task will be ham-
load during the performance of a task. The average pered/slower (for a review, see Brünken et al. 2003). For
value of instantaneous load equals the accumulated example, learners could be asked to respond to a color
load per time unit. Finally, overall load is the experi- change of a letter placed above the multimedia mate-
enced load based on the whole working procedure (see rials they are studying as soon as possible (see Brünken
also Paas et al. 2003). et al. 2003). The slower their response to the color
change, the more cognitive capacity was being devoted
Important Scientific Research and at that moment to the multimedia materials. Note that
Open Questions in order for the secondary task to be sensitive to vari-
Cognitive load can be measured with different tech- ations in cognitive load, it should draw on the same
niques. Most CLT research applies subjective rating working memory resources as the primary task. More-
scales to assess cognitive load, such as an adapted over, if learners decide to devote more cognitive capac-
version of the NASA-Task Load Index (NASA-TLX) ity to the secondary task, this might hamper their
Cognitive Load Theory C 601
than biologically primary knowledge (Geary 2008). We is provided by the borrowing and reorganizing principle.
have evolved to acquire primary knowledge such as Information is borrowed (and reorganized) from the
listening to and speaking our first language, recog- long-term memories of other people by imitating what
nizing faces, engaging in routine social relations, and they do, listening to what they say, and reading what
using general problem-solving strategies, over many they write.
generations. Each of these skills is modular and not While information can be borrowed from other
closely related to other primary skills. Primary knowl- people, that information must be created in the first
edge can be acquired effortlessly, unconsciously, and instance. Information is created during problem-
without explicit instruction by immersion in a human solving by the randomness as genesis principle, using
society. a random generate and test for effectiveness proce-
Secondary knowledge is cultural. We have not dure. Random generation of moves can result in an
required specific examples of secondary knowledge unmanageable number of possible moves. Knowledge
until relatively recently and so have not evolved to held in long-term memory is used to reduce the range
acquire any particular form of such knowledge. For of possible moves.
example, we have not specifically evolved to read and When dealing with novel information, knowledge
write in the way we have evolved to listen and speak. may be unavailable to sufficiently limit the range of
Similarly, it is plausible to argue that we have not possible moves. Instead of using knowledge to reduce
evolved to acquire the content of any subject com- the range of moves, the narrow limits of change prin-
monly taught in educational institutions. In contrast ciple is used. Our limited capacity, limited duration
to primary knowledge, the acquisition of secondary working memory prevents us from attempting to gen-
knowledge requires a general cognitive architecture erate a large number of complex moves.
applicable to a wide variety of areas, rather than mod- Lastly, the environmental linking and organizing
ular systems specific to a particular area. Furthermore, principle uses information from long-term memory
acquiring secondary knowledge tends to be effortful, to alter the characteristics of working memory. Indef-
conscious, and enhanced by explicit instruction. The inite quantities of organized information held in long-
cognitive architecture used by cognitive load theory term memory can be transferred to working memory
applies to secondary rather than primary knowledge for indefinite periods. As a consequence, information
and is central to cognitive load theory. When dealing held in long-term memory transforms a working mem-
with secondary knowledge, human cognition can be ory that is limited in capacity and duration into a
considered a natural information processing system working memory with no known capacity or duration
whose evolution has been driven by an analogous limits. That information from long-term memory
natural information processing system, evolution by determines how we interact with our environment.
natural selection (Sweller and Sweller 2006). The char- This cognitive architecture is used by cognitive load
acteristics of natural information processing systems as theory to generate instructional procedures. The aim
applied to human cognition will be described using five of instruction, based on the information store principle,
basic principles. is to accumulate knowledge in long-term memory.
The information store principle deals with the stor- Once stored, the environmental linking and organizing
age of information in human long-term memory. principle allows us to use the information to function
All learning requires information to be stored in long- in our environment. Knowledge stored in long-term
term memory. If nothing is stored in long-term memory is most easily acquired from other people
memory, nothing has been learned. In biologically sec- using the borrowing and reorganizing principle. If
ondary areas, massive amounts of domain-specific knowledge held by others is unavailable to us, it can
information in schematic form are stored in long- be created using the randomness as genesis principle. In
term memory. The primary goal of instruction is to both cases, the narrow limits of change principle indi-
assist learners to store that information. cates that instruction needs to minimize an unneces-
Because the amount of information stored in long- sary working memory load.
term memory is so large, an efficient procedure for The cognitive load (or working memory load)
acquiring that information is required. That procedure imposed by instructional material depends on the
Cognitive Load Theory C 603
number of elements (or schemas) with which learners separate form. For example, an explanation associated
must simultaneously deal. If elements interact, they with a diagram may be presented next to the diagram
must be dealt with simultaneously by working memory. rather than at appropriate points on the diagram.
Interacting elements that are intrinsic to the instructional Learners must mentally integrate a text and diagram
material impose an intrinsic cognitive load that cannot that are physically separate and mental integration C
be reduced other than by changing the nature of the task requires working memory resources that consequently
or by learning to group elements together into a higher- are unavailable for learning, imposing an extraneous
order schema that acts as a single element. Interacting cognitive load. Placing text at appropriate points on
elements that are extraneous to the instructional area a diagram allows working memory resources to be
impose an extraneous cognitive load that should be used for learning instead of relating the two sources
reduced by altering instructional procedures. Working of information. The split-attention effect occurs when
memory resources devoted to dealing with intrinsic cog- physical integration is superior to mental integration.
nitive load are germane to the task at hand and are The effect requires the two sources of information to
sometimes referred to as germane cognitive load. Effec- be unintelligible in isolation. If, for example, text
tive instruction maximizes working memory resources merely redescribes a diagram, the split-attention effect
dealing with intrinsic cognitive load that is germane will not be obtained (see the redundancy effect below).
to the task at hand and minimizes working memory The modality effect is demonstrated by comparing
resources dealing with extraneous cognitive load. information presented in both visual (e.g., a diagram)
Cognitive load theory has been used to generate and spoken (text) modes to information presented in
many instructional effects using randomized, con- a visual mode only with written text. The effect occurs
trolled experiments comparing various instructional when a dual, audio-visual mode of presentation is
procedures. These are described in the next section. superior to a single, visual only (with written text)
mode of presentation. The modality effect is related
Important Scientific Research and to the split-attention effect in that both effects require
Open Questions one or more sources of information to be unintelligible
The worked example effect is demonstrated when study- in isolation. The effect occurs because working mem-
ing worked examples increases problem-solving skill ory capacity and learning can be increased by using
more than solving the equivalent problems. Searching both auditory and visual processes. The effect will not
for problem solutions using the randomness as genesis be obtained if the information includes long textual
principle imposes a heavy, extraneous cognitive load passages. These passages must be presented in written
that reduces learning. In contrast, studying worked form because it may not be possible to process them
examples makes use of the borrowing and reorganizing appropriately in working memory.
principle. Skilled problem solvers have learned to rec- The redundancy effect occurs when multiple sources
ognize problem states and the best move for each state. of information are unnecessary for understanding,
Worked examples are ideally suited to indicate which unlike the split-attention and modality effects that
moves are best for particular problem conditions. only occur when each source of information is essential.
The problem completion effect is related to the Unnecessary information must be processed in working
worked example effect. Instead of learners being memory and so imposes an extraneous cognitive load
presented with fully worked examples, they are presented that is eliminated by eliminating the redundant informa-
with partially completed worked examples that they tion. The effect is obtained when learning is enhanced
must complete themselves. Characteristically, learners by the elimination of redundant information.
who complete partially completed problems learn more The expertise reversal effect is obtained when in-
and perform better on subsequent tests than learners structional procedures that facilitate learning by
who solve full problems, demonstrating the problem novices reduce in their relative effectiveness as levels
completion effect. of expertise increase. Instructional procedure A may
The split-attention effect occurs when learners must result in more learning that procedure B for novices but
split their attention between multiple sources of infor- for more knowledgeable learners, B may be superior to
mation that are unnecessarily presented in physically A. This effect is an outcome of the redundancy effect.
604 C Cognitive Load Theory
Information critical for novices may be redundant for extraneous cognitive load (all of the above effects),
experts and so impose an extraneous cognitive load. intrinsic cognitive load must be high. If intrinsic cogni-
There are many versions of the expertise reversal effect tive load is low due to low intrinsic element interactivity,
depending on relations between the categories of infor- reducing a high extraneous cognitive load may not mat-
mation. One version is particularly important and is ter because total cognitive load may be below working
discussed next. memory limits. Cognitive load effects require complex
The guidance fading effect is an example of the information.
expertise reversal effect that is dependent on the worked The isolated-interacting elements effect can be
example and completion effects. Worked examples only obtained if element interactivity due to intrinsic cog-
are effective in comparison to solving problems for nitive load is too high for working memory to process
novice learners. With increasing expertise, the relative the information. Element interactivity and its atten-
effectiveness of worked examples decreases and eventu- dant working memory load can be reduced by initially
ally reverses. For more expert learners in an area, study- presenting the interacting elements as though they are
ing worked examples is redundant and learning may be isolated without reference to the interactions between
facilitated if worked examples are replaced by comple- them before presenting them in fully interacting
tion problems. As expertise increases further, even form. Presenting information in isolated followed by
completion problems may be redundant and should be interacting form facilitates learning compared to only
replaced by full problems. In this manner, the informa- presenting the information with all interactions between
tion provided to learners is faded from worked examples elements emphasized.
to completion problems and finally, to full problems as The variability effect also depends on variations in
relevant information is stored in long-term memory and intrinsic cognitive load. If learners are presented new
so becomes redundant if provided during instruction. material with examples that vary in many surface char-
The imagination effect occurs when learners who acteristics, they must not only learn a new concept or
imagine concepts or procedures learn more than procedure, they also must learn to extract the concept
learners who study those concepts or procedures. or procedure from the surface structure in which it
Imagining requires rehearsal of concepts or proce- is embedded. Intrinsic cognitive load is likely to be
dures in working memory, a procedure that can better high. It can be lowered by reducing the surface vari-
transfer information to long-term memory than sim- ability but then learners no longer learn to distinguish
ply studying. The imagination effect is obtained when between different surface variations. Providing there
learners asked to imagine concepts or procedures learn is sufficient working memory capacity to handle the
more than learners asked to study the same concepts or increased element interactivity, high variability exam-
procedures. ples will result in more learning and transfer than low
The goal-free effect is obtained when learners are variability examples.
presented the givens of a problem without the goal These effects, generated by cognitive load theory,
and asked to make as many problem moves as they can indicate instructional procedures that can facilitate
without reference to a goal. Conventional problems with learning. The theory emphasizes the storage of large
a conventional goal require problem solvers to consider amounts of biologically secondary information in
their current problem state, the goal state, differences long-term memory after processing in a limited
between the two, and possible moves to reduce those working memory. Stored information governs expert
differences. Under goal-free conditions, problem solvers performance. Novel information is best obtained
only need to consider whether any move can be made. from other people. Cognitive load theory assumes
The reduced working memory load enhances learning that during instruction, learners do not acquire very
compared to solving conventional problems. This tech- general cognitive strategies because general strategies
nique only is likely to be effective using problems for are biologically primary and so learned easily and
which the number of moves that can be generated from automatically. Rather, the function of instruction is
the givens without a goal is very limited. to assist in the acquisition of a large number of
The element interactivity effect depends on intrinsic domain-specific, biologically secondary knowledge
cognitive load. For effects dependent on reducing structures.
Cognitive Modeling with Multiagent Systems C 605
Cross-References
▶ Cognitive Load Measurement Cognitive Modeling with
▶ Goal-Free Effect Multiagent Systems
▶ Guidance-Fading Effect
▶ Imagination Effect ANGELO CANGELOSI C
▶ Modality Effect on Learning Centre for Robotics and Neural Systems, University of
▶ Redundancy Effect on Learning Plymouth, Plymouth, UK
▶ Split-Attention Effect on Learning
▶ Worked-Example Effect
Synonyms
References Adaptive agents; Artificial life; Evolutionary robotics;
Geary, D. (2008). An evolutionarily informed education science.
Embodied agents
Educational Psychologist, 43, 179–195.
Sweller, J. (2003). Evolution of human cognitive architecture. In
B. Ross (Ed.), The psychology of learning and motivation Definition
(Vol. 43, pp. 215–266). San Diego: Academic Press. The computational modeling of cognition highly ben-
Sweller, J., & Sweller, S. (2006). Natural information processing sys- efits from the use of computer models of the learning
tems. Evolutionary Psychology, 4, 434–458. of behavioral and cognitive capabilities in simulated
Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory.
agents, such as for language development and evolu-
New York: Springer.
tion, or for the development of sensorimotor skills.
Through the simulation of the dynamics and interac-
tions in groups of agents it is possible to investigate
Cognitive Lock-In the role of social and group-based processes contrib-
uting to the development of cognition. In addition,
▶ Cognitive Automatisms and Routinized Learning multiagent systems can be used to investigate phyloge-
netic processes affecting the evolution of cognitive
capabilities. Examples of multiagent systems method-
ologies used for the study of cognition are artificial life
Cognitive Map models and evolutionary robotics. The main areas of
investigation in this field are language learning and
Internal representation of spatial relationship between
sensorimotor strategy development.
cues within the environment.
Theoretical Background
Cognitive modeling through agent-based systems per-
Cognitive Mapping mits the development and testing of specific hypotheses
on the ontogenetic and evolutionary acquisition of
▶ Spatial Learning behavioral and cognitive capabilities. Through the use
of multiagent systems it is possible to investigate, with
computer simulations, the role of social and group-
based processes in the development of cognition. In
Cognitive Model particular, researchers can adopt a synthetic modeling
Representation strategy (Cangelosi and Parisi 2002; Langton 1997),
which is quite different from classical scientific meth-
▶ Mental Models
odologies based on the analytic approach. For example,
in the natural sciences such as biology, a top-down
approach is often used by assuming the analysis, i.e.,
Cognitive Modeling division, of the global biological system into its main
component (e.g., the body is analyzed as a set of func-
▶ Adaptive Game-Based Learning tional systems and organs). In linguistics, language is
606 C Cognitive Modeling with Multiagent Systems
analyzed, i.e., decomposed, into syntax, words, pho- well as the interaction between evolutionary dynamics
nemes. On the contrary, a synthetic modeling approach and ontogenetic learning processes.
to cognition uses a bottom-up strategy to reconstruct Evolutionary robotics regards the autonomous de-
the behavioral and cognitive system. The researcher can sign of the controllers of (simulated or physical) robots
define the basic components of a cognitive agent, the through the use of genetic algorithms (Nolfi and
rules by which these components interact, and the Floreano 2000). This robotics approach can be consid-
environment in which the agent and its components ered as a subset of the artificial life methodologies,
interact with each other. The computer program will with an additional focus on the role of embodiment
then simulate the interactions among the components in cognition due to the simulation of the robot’s sen-
to observe the emergence of the various higher-level sorimotor system. Although great part of the early
capabilities. With a simulation model, the feasibility work in evolutionary robotics focused on low-level
and validity of assumptions regarding the components, sensorimotor capabilities (navigation, object avoid-
their interaction rules, and the environment can be ance, foraging), more recent work has extended the
tested. For example, wrong, incomplete, or inadequate use of this methodology for higher-order motor and
assumptions will make it impossible to observe the cognitive capabilities such as object manipulation and
emergence of higher-level entities or of entities that language learning. In addition, evolutionary robotics
do not have realistic properties and do not exhibit has been recently applied to more complex models
realistic phenomena. The bottom-up approach of syn- of robotic platforms, moving from the use of simple
thetic simulations of multiagent cognitive systems also wheeled robots to humanoid robot platforms. In addi-
permits the study of problems and phenomena that are tion, evolutionary robotics has also been used to develop
analytically intractable, such as those of complex and models of the evolution of morphology of both the body
nonlinear systems, as it is the case of the phylogenetic and the brain of the agents.
and ontogenetic development of cognition. If we consider the main areas of cognition that
Within the field of cognitive agent modeling, two have been investigated through both approaches in
main synthetic modeling methodologies have been multiagent systems, we can identify five main behav-
employed: (1) artificial life models and (2) evolutionary ioral and cognitive capabilities where important scien-
robotics. Artificial life refers to the synthetic modeling tific insights have been produced:
of natural and artificial life-like systems, an innovative
approach developed by Langton and collaborators in ● Navigation, exploration, and foraging strategies. This
the late 1980s at the interface between biology and is the area where evolutionary robotics, as well as
computer science (Langton 1997). Although part of early artificial life models, has contributed most
the initial efforts within artificial life focused mostly (Nolfi and Floreano 2000). These studies typically
on the modeling of plant systems (e.g., Lindenmayer used wheeled robots (khepera, e-pucks) to investi-
systems) and low-level biochemical interactions (e.g., gate the evolutionary emergence of flexible, adap-
protein binding), artificial life systems have been exten- tive strategies for optimal exploration strategies and
sively used for the modeling of behavioral and cogni- foraging. Models demonstrated the strict coupling
tive capabilities in multiagent systems (Cangelosi and between the agent’s own sensorimotor system and
Parisi 2002; Steels and Belpaeme 2005). These agent- of the constraints of their environment.
based artificial life models typically consist of the sim- ● Categorization. The adaptive interaction with the
ulation of a group of agents that have to survive by environment requires the capability to categorize
adapting to the social and physical requirements of the objects and entities in the world, consistently
the environment and have to reproduce through with the agent’s own internal needs and social
genetic algorithms. Common artificial life tasks regard context. For example, Steels and Belpaeme (2005)
navigation and exploration of the environment, social analyzed which mechanisms a population of auton-
cooperation, and communication. The behavioral and omous agents benefits from to arrive at a repertoire
cognitive capabilities of each agent are controlled using of perceptually grounded color categories. They
a variety of methods, such as artificial neural networks, compared three main approaches to human cate-
that permit the modeling of learning mechanisms, as gorization: nativism, empiricism, and culturalism.
Cognitive Modeling with Multiagent Systems C 607
Multiagent simulations showed that the collective agents (e.g., model of humanoid robots) and carry out
choice of a shared repertoire must integrate multi- extensive simulation experiments (e.g., on the evolu-
ple constraints, including constraints coming from tion of brain and body morphology). In addition, the
communication. increasing empirical evidence in neuroscience and psy-
● Biological and cultural evolution of language. Lan- chology on the embodiment bases of cognition opens C
guage is definitely one of the main areas where new challenges for the understanding of the interaction
synthetic multiagent systems have produced signif- between sensorimotor knowledge and other cognitive
icant impact and scientific explanations (Cangelosi capabilities.
and Parisi 2002). In particular, numerous models of Such significant technological and scientific advances
the biological and cultural evolution of language have opened up a series of new challenges in cognitive
have shed light on the crucial factors favoring the modeling through multiagent systems. Here we list a few
evolutionary emergence of languages, such as social of the key research questions for future research:
learning phenomena and internal representation
– How can more complex embodiment systems, such
capabilities. More recently, such computer models
as simulation models of humanoid robots, be used
have been put in relationship with empirical data
to explain the fine mechanisms of the grounding of
on human languages (Vogt 2009).
cognition (e.g., microaffordance effects of action–
● Development and grounding of cognition in embodi-
vision links, action-compatibility effects in lan-
ment systems. Language again has been used as
guage processing)?
a test case for investigating the role of embodiment
– What are the evolutionary and developmental mech-
in cognition. For example, Cangelosi (2010) uses
anisms that supported the coevolution of brain and
a variety of multiagent systems to examine the
behavior?
grounding of language into the agent’s own action
– How can multiagent systems be used to investigate
repertoire, both in simulation agents and in human-
the effects of different social interaction protocols
oid robots. These models are consistent with increas-
in the establishment and maintenance of social
ing empirical evidence from neuroscience and
structures?
cognitive psychology on embodied cognition.
– How can the current minimal cognitive models
● Social coordination. Synthetic multiagent models
used in multiagent systems be scaled up to investi-
have been utilized to study social coordination
gate higher-order cognitive capabilities?
(both competition and cooperative interactions)
– What are the interaction dynamics between genetic
amongst groups of cognitive agents. For example,
evolution and cultural evolution in the emergence
coevolutionary simulations on prey–predator com-
of language?
petition experiments demonstrated an “arms race”
– What is the role of evolutionary and cognitive fac-
phenomenon where increase in complexity in one
tors in the emergence of syntax?
population, e.g., escape strategies of the prey, can
cause the emergence of complex strategies in the
Cross-References
coevolving predator species (Nolfi and Floreano
▶ Agent-Based Modeling
2000).
▶ Cognitive Models of Learning
▶ Cognitive Robotics
Important Scientific Research and ▶ Learning Agents and Agent-Based Modeling
Open Questions
The field of synthetic multiagent systems has received
an important boost in the last few years because of References
technological progress on computer simulation sys- Cangelosi, A. (2010). Grounding language in action and perception:
tems and robotic agent modeling. Thanks to advances From cognitive agents to humanoid robots. Physics of Life
Reviews, 7(2), 139–151.
in computationally intensive simulation tools for evo-
Cangelosi, A., & Parisi, D. (Eds.). (2002). Simulating the evolution of
lutionary and multiagent systems and to the availability language. London: Springer.
of open-source physics engines, it is now possible to Langton, G. C. (1997). Artificial life: An overview. Cambridge, MA:
build more detailed and accurate models of cognitive MIT Press/Bradford Books.
608 C Cognitive Models of Learning
Nolfi, S., & Floreano, D. (2000). Evolutionary robotics: The for just about any imaginable context. A cognitive
biology, intelligence, and technology of self-organizing machines. model for a given domain or problem solving task
Cambridge, MA: MIT Press/Bradford Books.
typically represents an expert’s knowledge, which can
Steels, L., & Belpaeme, T. (2005). Coordinating perceptually
grounded categories through language: A case study for colour. sometimes take years (or even a decade) to form in the
The Behavioral and Brain Sciences, 28(4), 469–489. mind of that expert. For a learner seeking to become
Vogt, P. (2009). Modelling interactions between language evolution an expert in that domain, the developmental path to
and demography. Human Biology, 81(2–3), 237–258. that desirable end state can be just as complex, if not
more, than the domain knowledge itself. The tools of
cognitive science can also be used to describe the
processes learners engage to acquire knowledge and
expertise in a given domain. To construct such cogni-
Cognitive Models of Learning tive models of learning, a variety of approaches are
used to collect relevant data while students are
H. CHAD LANE engaged in learning. These include think-aloud pro-
Institute for Creative Technologies, University of tocols, problem solving traces, diagnostic tests, and
Southern California, Los Angeles, CA, USA even neurological analyses of brain activity. Because
learning can occur in different ways, in different
contexts, and for different knowledge types, a variety
Synonyms of models that account for learning have emerged.
Cognitive skill acquisition; Computational models of Further, cognitive models of learning can take a des-
learning; Conceptual change criptive form reporting empirical observations and
strategies revealed from learner thinkalouds to a
Definition more formal, computational form suitable for simu-
A cognitive model is a descriptive account or compu- lation on a computer (Ohlsson 2008).
tational representation of human thinking about a Acquisition of cognitive skills is a common focus of
given concept, skill, or domain. Here, the focus is on cognitive models of learning. Here, learning is focused
cognitive knowledge and skills, as opposed to sensori- on solving problems in a given domain. Substantial
motor skills, and can include declarative, procedural, empirical evidence exists showing that cognitive skill
and strategic knowledge. A cognitive model of learn- acquisition progresses in three stages: (1) cognitive
ing, then, is an account of how humans gain accurate stage: learners develop a declarative encoding of the
and complete knowledge. This is closely related to domain knowledge, (2) associative stage: through prac-
metacognitive reasoning and can come about as a result tice, errors in knowledge are identified and repaired,
of (1) revising (i.e., correcting) existing knowledge, and (3) autonomous stage: continued practice increases
(2) acquiring and encoding new knowledge from speed and accuracy during execution of the cognitive
instruction or experience, and (3) combining existing skill. Models of cognitive skill acquisition generally
components to infer and deduce new knowledge. strive to follow the same pattern, and deal with the
A cognitive model of learning should explain or simu- complexities that learners also face. They track learning
late these mental processes and show how they produce of individual rules, or knowledge components, to mul-
relatively permanent changes in the long-term memory tiple interacting pieces of knowledge at once, and
of learners. It is also common to consider impoverished finally, on to the final stages when practice produces
cognitive models of learning which can be useful for autonomy (VanLehn 1996).
diagnosis of learner errors and misconceptions, and Cognitive models of learning are tied closely to
in many cases, prescribing appropriate instructional metacognition, which can informally be understood
interventions. as “thinking about thinking.” Metacognitive thinking
represents an essential aspect to cognitive models of
Theoretical Background learning because they define control mechanisms the
Cognitive modeling is a basic tool for the field of learner must apply in order to actually acquire new
cognitive science used to account for human thinking knowledge. That is, to reach the end state of possessing
Cognitive Models of Learning C 609
usable and accessible new knowledge in long-term Important Scientific Research and
memory, learners must actively regulate their own cog- Open Questions
nitive processes, decide where to direct their attention, Cognitive models represent an important class of tools
self-assess to decide if they understand, self-explain in the study of human cognition and learning. To date,
in order to establish connections between domain researchers have made incredible strides in studying C
principles and the object of study, decide if they will and modeling complex human learning (Ohlsson
seek help, and so on. For example, learners who study 2008; VanLehn 1999). However, any model of human
worked-out examples learn more effectively if they learning is almost by definition, incomplete. It is always
choose to frequently stop to check their own under- necessary to restrict a cognitive model of learning in
standing and identify underlying principles that pro- some way, whether it be the domain it operates on
vide justification for problem solving steps (Chi et al. or the kinds of reasoning of which it is capable.
1989). A good example of a computational model of Nowhere is this more evident than in recent efforts to
these activities, along with other learning mechanisms, integrate affective and emotional processes into models
is captured in the computational cognitive model of of learning (Kort 2009). Here, researchers are focused
learning, Cascade (VanLehn 1999). The model simu- on understanding the interplay between emotion and
lates learning from worked-out examples as well as learning to answer basic questions such as when
from problem solving and produces cognitive changes instruction is most effective, at what point do learners
on the impasse–repair–reflect cycle, a model derived respond positively to challenge, and when does frus-
from empirical studies of human learners (Chi et al. tration hinder or impede learning. These questions
1989). During learning, if Cascade finds that its current represent key open questions in both the psychological
domain knowledge is insufficient to move forward in literature on human learning, as well as in the cognitive
reading or problem solving (i.e., it is at an impasse), modeling literature. Ohlsson (2008) points out that an
this triggers a learning event. The system seeks to mod- assumption made by many computational models of
ify its existing knowledge or add a new rule that will learning is that learning mechanisms are tested inde-
allow it to overcome the impasse (i.e., a repair). Finally, pendently (p. 384). This suggests that as more models
reflection is achieved via explanation-based reasoning are tested for validity and completeness, they should be
on the proposed solution to determine correctness. In done so in complex learning contexts that involve mul-
Cascade, the approach is to leverage commonsense tiple learning mechanisms. It is the interaction between
knowledge in conjunction with existing knowledge to learning mechanisms that may pose a hidden threat to
construct new rules for future use (VanLehn 1999, the success of existing computational models of learn-
pp. 86–87). ing. In addition, research on emotions in learning pro-
Broadening the perspective beyond cognitive skill cesses can be viewed as a positive step because they are
acquisition, researchers have also investigated cog- inherently contextual (i.e., learning is never focused on
nitive models of conceptual change during learning sitting down to simply experience an emotion – it always
and development. Here, models deal directly with the involves a cognitive target). Finally, very few cognitive
fact that learners enter into learning situations with models of learning have integrated findings from cogni-
preconceived and naı̈ve conceptions and misconcep- tive neuroscience, and so this represents a key open area
tions about the world. Recent research on conceptual of future research. To date, researchers have determined
change has shifted focus to the learner by introduc- areas of the brain that are involved in learning, emotion,
ing intentional conceptual change, defined as “goal- and automaticity. This empirical data may shed light on
directed and conscious initiation and regulation of cognitive models of learning by providing evidence for
cognitive, metacognitive, and motivational processes setting of parameters (e.g., rate of learning or memory
to bring about a change in knowledge” (Sinatra and decay) and testing of underlying assumptions.
Pintrich 2003, p. 6). These approaches therefore over-
lap significantly with metacognitive models of learning, Cross-References
but with substantially more of a focus on developmen- ▶ ACT
tal and repair activities necessary for long-term con- ▶ Cognitive Dissonance in the Learning Processes
ceptual understanding. ▶ Cognitive Learning
610 C Cognitive Neuroscience
Definition
Cognitive psychology of music learning is the study of
Cognitive Plasticity the perceptive and generative processes involved in
listening to, performing, analyzing, improvising, and
▶ Mediated Learning Experience (MLE) and Cognitive composing music.
Modifiability
Theoretical Background
and remember music they hear. Development of Music Learning and Cognition
musical abilities follows a developmental trajectory. Learning Theories
Both genetic predispositions and early instruction Theories of the development of musical learning
experiences affect the development of the musical include both a thinking component – cognition – and
brain. Contextual influences, such as parent sup- a learning component. The learning component is C
port, as well as intrinsic motivation factors deter- experiential in nature. Musicians grow in both knowl-
mine the extent to which an individual becomes edge and experience. The next sections briefly describe
a musician. the essential general learning theories, in the context of
Musicians who began receiving music instruction at music learning.
an early age have different brain structures when com- Behaviorism. Adherents to this theory of musical
pared to individuals who never received instruction. learning believe in the power of classical conditioning.
Moreover, musicians exhibit different electrophysio- Classical conditioning, developed by Pavlov in his
logical brain responses while performing different experiments with dogs, revealed that a neutral stimulus
tasks. Regardless of differences in brain structure, the will elicit a response after repeatedly being paired with
underlying processes that govern how we perceive, another stimulus that already elicits that response.
process, and respond to musical stimuli is the same Stimulus-response chains can then be developed that
for musicians and nonmusicians. will lead to predictable, generalizable behaviors. Exam-
Musical activity encompasses every part of our ples of behaviorism in music teaching and learning can
brain. Different parts of the brain process different be found in the area of traditional instrumental music
elements of music, such as timbre, beat, rhythmic pat- education. Wind bands practice for festivals, where
terns, tonalities, harmonies, song lyrics, and so on. they receive ratings that reinforce or inhibit their
Different parts of the brain are interconnected and behaviors. On a smaller level, instrumentalists’ musical
dependent on each other when attempting to catego- practice habits can be reduced to a series of behaviors
rize incoming musical stimuli, using memory, reason- that researchers who espouse to this theory of learning
ing, and evaluation. A variety of factors determine how can describe and measure quantitatively. Examples
brain mechanisms work in the case of evaluating and of behaviorism in music therapy include using music
responding to music. as a contingency for modifying behavior or as a cue
We evaluate music based on our previous experi- for teaching new skills. Jayne M. Standley and her
ences. Through musical experiences our brain learns research with premature infants, Clifford K. Madsen,
to associate different sounds as pleasant, soothing, and Alice-Ann Darrow are examples of researchers
calming, or arousing. As we listen to music, our brains firmly grounded in Behaviorism.
work quickly to categorize and impose structure over Cognitive Psychology. Cognitive psychology, specif-
the incoming stimuli. This process is ongoing because ically, in music learning theory was a shift from the
the feedback we receive from our brain subsequently focus on observable measurable behaviors to a focus on
becomes a factor determining how we respond and examining internal processes of cognition and intellec-
evaluate future musical stimuli. Gradually our brain tual growth. Constructivism is a branch of cognitive
develops a complex system of expectations that helps psychology grounded on the principle that all individ-
us understand musical genres, harmonies, and rhyth- uals are born with certain cognitive functions, and
mic or tonal patterns. that these cognitive functions develop over time. So,
Musical compositions are designed around vali- each cognitive function builds on previous age-based
dating or violating our expectations. Our familiarity and/or experience-based versions of that particular
with specific musical genres determines whether function. Jean Piaget is probably the most influential
our brain will interpret a piece as simple or com- figure in this line of research. Within the music learn-
plex. An individual who has heard jazz music ing area, researchers such as Mary Louise Serafine and
throughout his or her life will have a different evalu- her Generative Processes theory, and Edwin Gordon
ating response when hearing a jazz composition in and his Music Learning Theory are examples of
comparison to an individual who has only heard jazz researchers in music education firmly grounded in
music sporadically. Cognitive Psychological theory. Kenneth Aigen is
612 C Cognitive Psychology of Music Learning
a music therapy researcher who has applied Serafine’s research. Performance practice has been considered
theory to explain client responses in music therapy. almost exclusively in terms of Western Classical
Sociohistorical Theory. Proponents of the sociohis- music-making. Future research might examine prac-
torical theory of learning emphasize the importance tice in terms of non-Western Classical music-making:
of context and history in the development of all man- ethnic ensembles, popular music ensembles, and new-
ifestations of learning, including both cognitive and music ensembles. Improvisation – constraints imposed
experiential. Through this theoretical lens, researchers on the process of improvisation has been a topic of
such as Vygotsky have proposed that learning does not research. The measure of musical ability and impro-
center entirely on the solitary actions of individuals, as visation has been examined most notably in jazz.
the behaviorists would imply, or on the interaction Researchers have examined group improvisation as
between the individual and his or her environment as a social construct, one where individual identities are
the constructionist would imply. Rather, sociohistor- shaped by participation in the group. Future work in
ical theorists see all human learning as occurring this area could include examining how an identity as an
within particular cultures, with particular histories. improviser in a group is different than an identity as
Vygotsky’s work in sociohistorical theory can be seen a performer. Composition – work in the area of compo-
in the music education literature in the work of sition learning has focused on processes, products, and
Patricia Shehen Campbell and others. In music ther- the meaning of composition to individuals. There is
apy, Mercédès Pavlicevic, Gary Ansdell, and Brynjulf a focus currently on understanding composition learn-
Stige are prominent researchers influenced by socio- ing in particular teaching and learning contexts. Future
historical theory. work could probe qualitatively the value and meaning
Connectionism. Connectionist theorists use innova- of composition learning to students.
tive technology such as electroencephalography (EEG), Perception of Musical Sounds. A number of different
electromagnetic-encephalography (EMG), event- areas have been examined in the area of music percep-
related potential (ERP), magnetic resonance imaging tion. Researchers have focused their efforts on under-
(MRI), computer tomography (CT), and positron standing the perception of pitch, tonal cognition,
emission tomography (PET) to measure activity in musical timbre, musical time (meter and rhythm),
the brain when individuals are engaged in musical and musical memory. This line of research has
activities. Neural networks provide the basis for an blossomed alongside the multiple technological inno-
individual’s musical representations. Learning within vations that have made measurement in this area
this theory is then related to physiological conditions of more feasible.
the brain. Notable pioneer researchers within this area Music Cognition and Psychobiology. Empirical stud-
of music learning area are Donald Hodges, John Flohr, ies examining how music cognition and the aesthetic
Daniel Miller, and Diane Persellin. In music therapy, qualities of music affect cognitive, affective, sensory,
Michael H. Thaut is a prominent researcher who devel- and motor human responses are continuing to emerge
oped Neurologic Music Therapy, a scientific model of in the music therapy literature. Such findings continue
examining the therapeutic uses of music in neurologic to inform clinicians who use biomedical applications of
rehabilitation, neoropediatric therapy, nerogeriatric music in therapy or music as therapy.
therapy, development, and adaptation.
Cross-References
Important Scientific Research and ▶ Human Cognitive Architecture
Open Questions ▶ Shared Cognition
Measurement of Musical Abilities. The measurement of ▶ Situated Cognition
various musical abilities have been examined in
research, including: performance, improvisation, and
References
composition. Performance – practice habits and moti-
Gruhn, W., & Rauscher, F. (2002). The neurobiology of music cogni-
vation have been explored recently in the area of musi- tion and learning. In R. Colwell & C. Richardson (Eds.), The new
cal performance. The function of family support to the handbook of research on music teaching and learning. New York:
practicing musician has been a component of that Oxford University Press.
Cognitive Robotics C 613
Hallam, S., Cross, I., & Thaut, M. (2009). The oxford handbook of neuroscience. In addition to the technological aim of
music psychology. New York: Oxford University Press. designing autonomous robots, cognitive robots are
Hodges, D. A. (1996). Handbook of music psychology (2nd ed.).
also widely used as embodied computational models
San Antonio, TX: IMR Press.
McPherson, G. (2006). The child as musician: A handbook of musical investigating the organization of learning and cogni-
development. New York: Oxford University Press. tion within the cognitive and neural sciences. A C
National Association for Music Education. (2000). Music makes growing field of cognitive robotics has taken a devel-
the difference: Music, brain development, and learning. Reston, opmental (i.e., ontogenesis) flavor in recognition of the
VA: MENC.
fundamental role of learning in the final performance
of biological cognitive systems.
Theoretical Background
In the fields of cognition, neuroscience, and robotics
Cognitive Restructuring there is growing theoretical and empirical evidence on
▶ Cognitive-Behavioral Family Therapy the role of embodiment, situated learning, and the
grounding of cognitive capabilities in sensorimotor
knowledge in natural and artificial cognitive systems
(Pfeifer and Bongard 2006). Recent advances in cogni-
tive psychology, neuroscience, cognitive linguistics,
Cognitive Robotics and developmental psychology support an embodied
view of cognition, i.e., the fact that cognitive functions
GIORGIO METTA1, ANGELO CANGELOSI2 (e.g., perception, categorization, reasoning, and lan-
1
Department of Robotics, Brain and Cognitive guage in particular) are strictly intertwined with sen-
Sciences, Italian Institute of Technology, Genoa, Italy sorimotor and emotional processes (Rizzolati and
2
Centre for Robotics and Neural Systems, University of Craighero 2004). This is particularly evident in numer-
Plymouth, Plymouth, UK ous experimental psychology studies on the grounding
of language, and other cognitive capabilities, in action
and perception.
Synonyms Such evidence is consistent with cognitive robotics
Cognitive systems; Developmental robotics; Epigenetic research. This uses knowledge from neural and cogni-
robotics; Humanoid robots; Neuro-robotics tive sciences to derive bio-inspired design principles
for cognitive development that are then tested in
Definition robotic platforms. The training of a robot to acquire
Cognitive robotics, also known as artificial cognitive sensorimotor, cognitive, and social capabilities implies
systems research, regards the use of bio-inspired that these skills are developed through dynamic inter-
methods for the design of sensorimotor, cognitive, actions between the entire cognitive system and its
and social capabilities in autonomous robots. Other environment. As such, most studies in cognitive robot-
designations have been proposed in the short history ics require the simultaneous learning of several cog-
of cognitive robotics which spans approximately the nitive skills, although a certain progression can be
last 15 years, as for example, Epigenetic Robotics, identified by studying human cognitive development
Autonomous Mental Development (AMD), or Cog- (von Hofsten 2004).
nitive Developmental Robotics (CDR). Robots are Within the field of cognitive robotics, in fact, the
required to learn suchcapabilities (e.g., attention and developmental (epigenetic) robotics approach focuses
perception, object manipulation, linguistic communi- on the autonomous mental development of cognition
cation, social interaction) through interaction with through incremental and maturational stages (Weng
their environment and via incremental develop- et al. 2001; Lungarella et al. 2003). Such an approach
mental stages. The biological- and cognitively inspired is directly inspired by ontogenetic stages studied in
methods and design principles are derived from stud- developmental psychology, as in Piaget’s epigenetic
ies in cognitive and developmental psychology, and psychology. Development adds an important aspect
614 C Cognitive Robotics
to the study of cognitive robotics by considering the grounding. A simulated robot is first trained to
possibility that cognitive skills arise only through a learn, by imitation, a set of action primitives,
process of maturation rather than being fixed and and a corresponding set of action words describing
hand-coded a priori by a human designer. Typically, these motor categories. Subsequently, the robot is
developmental robotics attempts at identifying a small taught linguistic combinations of the names of
number of early behaviors (the inductive bias) and the actions to describe compositional, higher-order
rules of development that transform the early behaviors actions (e.g., “grab” as a result of the simultaneous
in new skills via interaction of the cognitive agent with use of the left and right arms). Through a symbol
the environment (including social interaction). grounding mechanism, implemented in the robot’s
The main areas of research in cognitive and devel- own neural architecture, the robot is then able to
opmental robotics regard the following topics: transfer the grounding of basic action words to
higher-order compositional actions. This simula-
● Curiosity, attention, vision. The development of
tion model is currently being extended to language
humans is driven by motives that can be social
learning experiments with the iCub robot.
(interaction) or even motoric (it seems that exercis-
● Social interaction, imitation, and cooperation.
ing the motor system is a strong motive by itself).
Great part of early work on cognitive humanoid
This is important since cognition develops at the
robotics has centered on imitation and social
interface between the brain and action but requires
learning (Schaal 1999). This is also explained by
goals and a motivated subject (von Hofsten 2004).
developmental psychology focus on learning by
Attention and more in general vision clearly shape
imitation from parents and peers and its impor-
profoundly the acquisition of cognitive skills. In
tance for social development. Social learning and
robotics, many of these skills and their developmen-
imitation studies have proposed models of learning
tal counterparts have been modeled and this repre-
by imitation (e.g., imitation of motor behavior
sents one of the main trends in cognitive robotics.
from a teacher or demonstrator) as well as social
For a review of the relevant literature, the interested
learning for higher-order cognitive capabilities
reader is redirected to Lungarella et al. (2003) and
(e.g., perspective-taking).
Vernon et al. (2007).
● Locomotion. There is consistent developmental
● Manipulation. Tantalizing results from neurosci-
literature that locomotion in humans opens up
ence have shed light into the intricacy of the con-
the doors of spatial understanding. Numerous
trol of manipulation in the brain (Rizzolati and
experiments show that certain perceptual judg-
Craighero 2004). Many examples of the cognitive
ments develop in tight synchrony with the develop-
control of manipulation (comparing this to more
ment of crawling (or more in general with the ability
traditional model-based manipulation) have been
to move in the environment). Robotic research in
proposed, often at the boundary of imitation and
this direction concentrated though mostly in the
social interaction as an attempt to explain not only
technical skills (motor control) required for stand-
the how (that is the realm of neuroscience) but also
ing and walking (e.g., Asimo) rather than in the
the whys of certain brain circuits. One pivotal dis-
cognitive aspects connected with walking. Further-
covery is clearly that of mirror neurons (Rizzolati
more, most of this same research does not consider
and Craighero 2004 for a review) which has gener-
a developmental progression (Thelen and Smith
ated consistent interest in the cognitive robotics
1994) and rather addresses the problem of the gen-
community (see Arbib et al. 2008).
eration of suitable trajectories and feedback stabi-
● Communication and language. Language learning
lizing controllers.
is one of the key research topics in cognitive robot-
ics as it provides a prototypical example of how
higher-order cognitive skills (semantics, syntax) Cognitive Robotics Platforms
are directly grounded on sensorimotor knowl- In the literature there is a variety of robotics plat-
edge. For example, Cangelosi and Riga (2006) forms, using different actuators configurations (mobile
developed an epigenetic robotics model of language robots, arm manipulators, humanoid), that have been
Cognitive Robotics C 615
employed for cognitive modeling research. However, invariant aspects of the cognitive system and those
humanoid robots provide a more general and suitable that are independent from the task. Provided with
test platform for cognitive robotics as they permit knowledge, the cognitive architecture was theoreti-
the investigation of complex sensorimotor capabilities cally capable of performing a given task. Conversely,
(e.g., object manipulation) and realistic human–robot for embodied and developmental systems the defi- C
interaction (HRI) scenarios. The humanoid platforms nition of a Cognitive Architecture is less clear. One
most commonly used in cognitive robotics are the iCub attempt of a definition as proposed in the above
(RobotCub Consortium), Qrio (Sony Corp.), AIBO mentioned paper by Vernon et al. identifies the
(Sony Corp.), Asimo (Honda), and NAO (Aldebaran Cognitive Architecture of a developmental system
robotics). as its phylogeny. In this respect, the Cognitive
The humanoid robot iCub (Sandini et al. 2007) is Architecture contains the initial skills of the system
one of the platforms gaining significant impact in cog- together with its developmental rules.
nitive and developmental robotics. The iCub has been ● Interaction between development, maturation, and
developed as part of the RobotCub EU project (IST FP6 phylogeny. Within cognitive robotics, most of the
004370) with the explicit goal of providing a complex focus has been on incremental (i.e., developmental,
platform for cognitive systems research. With this in ontogenetic) learning. On the other end, other
mind, the iCub was designed with complex hands approaches such as evolutionary robotics mostly
for manipulation (9 degrees of freedom each), facial focus on phylogenetic changes. Future research
expressions (for interaction), and locomotion abilities should look at the interaction between such
(crawling). Sensors are also important and in this phenomena (as in the Baldwin effect) and the
respect, the iCub sports cameras, microphones, gyro- interaction with neural and morphological matura-
scopes, accelerometers, position sensors of various tional mechanism, known to affect learning and
types, and a sensorized skin. The platform is distrib- development.
uted as Open Source following a GPL license in an ● Robustness in unstructured environment. One of
attempt to make it the platform of choice for research the main challenges that cognitive robotics aims
in cognitive systems. About 20 iCubs have been built to address, in comparison with classical robotics
as part of this endeavor. This allowed the creation of approaches such as industrial automation, is the
a community of users and the possibility of sharing capability of robots to adapt to dynamic and
results or building on each other’s success. unpredictable environments. This is the case for
One important aspect of the availability of such example of humanoid robots that have to operate in
complex platforms at many locations is the possibility open and unstructured environments (e.g., walking
of benchmarking and experimental validation. Experi- in home, table-top manipulation tasks).
ments and models can be now truly tested on the ● HRI and social acceptance. The increasing availabil-
same hardware and results compared quantitatively. ity of humanoid and mobile robots in service robot-
In a sense, the dependence on the platform becomes ics, such as companions for elderly, has important
less important since many share the same platform implications for defining users’ acceptability criteria
(the iCub). to facilitate human–robot interaction.
Cangelosi, A., & Riga, T. (2006). An embodied model for sensorimo- which actions are relevant, under which circumstances
tor grounding and grounding transfer: Experiments with epige- those actions should be performed, how they relate
netic robots. Cognitive Science, 30(4), 673–689.
to the person’s goal, and what their effects are likely
Lungarella, M., Metta, G., Pfeifer, R., & Sandini, G. (2003). Develop-
mental robotics: A survey. Connection Science, 15(4), 151–190. to be. This type of knowledge is variously referred to as
Newell, A., & Simon, H. A. (1976). Computer science as empirical “competence,” “expertise,” “know-how,” “practical
inquiry: Symbols and search. Communications of the Association knowledge,” “procedural knowledge,” and “skill knowl-
for Computing Machinery, 19, 113–126. Tenth Turing award lec- edge.” No single term is standard; practical knowledge
ture, ACM, 1975, March 1976.
will serve.
Pfeifer, R., & Bongard, J. (2006). How the body shapes the way we
think: A new view of intelligence. Cambridge, MA: MIT Press.
Practical knowledge is intrinsically related to goals
Rizzolati, G., & Craighero, L. (2004). The mirror-neuron system. and actions, while declarative knowledge consists of
Annual Review of Neuroscience, 27, 169–192. facts, episodes, and generalities that are true or false
Sandini, G., Metta, G., & Vernon, D. (2007). The iCub cognitive independent of a person’s intentions or behavior (e.g.,
humanoid robot: An open-system research platform for enactive the Earth is round). Practical knowledge is primarily
cognition. In M. Lungarella, F. Iida, J. Bongard, & R. Pfeifer,
acquired via practice, while declarative knowledge is
(Eds.), 50 years of artificial intelligence. Essays dedicated to the
50th Anniversary of artificial intelligence series: Lecture notes in primarily acquired via observation and discourse.
computer science (Vol. 4850). Heidelberg: Springer. A popular belief holds that the two types of knowledge
Schaal, S. (1999). Is imitation learning the route to humanoid robots? follow different forgetting curves, with declarative
Trends in Cognitive Sciences, 3(6), 233–242. knowledge (e.g., the content of a text) decaying faster
Thelen, E., & Smith, L. (1994). A dynamic systems approach to the
than practical knowledge (e.g., the skill of riding a
development of cognition and action. Cambridge, MA: MIT Press.
Vernon, D., Metta, G., & Sandini, G. (2007). A survey of cognition
bicycle), but this belief is not grounded in research.
and cognitive architectures: Implications for the autonomous Cognitive skills are exemplified by symbolic activi-
development of mental capabilities in computational systems. ties like chess and mathematics and by professional ac-
IEEE Transactions on Evolutionary Computation, 11(2). Special tivities like medical diagnosis, computer programming,
issue on AMD, April 2007. and ship navigation. Successful performance depends
von Hofsten, C. (2004). An action perspective on motor develop-
primarily on the processing of conceptual information.
ment. Trends in Cognitive Science, 8, 266–272.
Weng, J., McClelland, J., Pentland, A., Sporns, O., Stockman, I., In contrast, motor skills (a.k.a. “perceptual-motor
Sur, M., & Thelen, E. (2001). Autonomous mental development skills” and “sensori-motor skills”) are exemplified by
by robots and animals. Science, 291, 599–600. tasks such as baseball, dance, and juggling. Successful
performance depends primarily on the physical char-
acteristics of the person’s movements: acceleration,
amplitude, direction, force, speed, timing, and so
on. The boundary between the two types of skill is
Cognitive Skill Acquisition not sharp.
psychologists worked with natural scientists and tech- internalization of task instructions, if any (a.k.a.
nologists who developed the first information tech- “proceduralization” and “knowledge compila-
nologies, including feedback systems. Information tion”); the use of analogies to already mastered
processing concepts revolutionized cognitive psychol- tasks; the study of solved examples and demonstra-
ogy, but they were initially applied to other problems tions, if available; reasoning from prior declarative C
than learning. In the 1960s, applied psychologists like knowledge; and capturing positive outcomes of
P. Fitts, R. Gagné, and A.T. Welford developed the heuristic search (a.k.a. “trial and error”).
enduring notions of learning curves, phases of practice, 2. The mastery phase lasts from the first complete
and multiple modes of learning, The modern study of performance until the correct performance can be
cognitive skill acquisition began with a 1979 article by reliably produced. The cognitive mechanisms that
Y. Anzai and H.A. Simon that reported a computer are likely to be important during this phase include
simulation of the acquisition of a problem solving learning from the feedback (a.k.a. “knowledge of
skill in a single subject. results”) provided by the task environment (see
The essence of practice is to attempt to perform below).
a target task that one has not yet mastered, with the 3. The optimization phase begins when the task has
intent to master it. Each attempt at performing the task been mastered and lasts as along as the learner
is a training trial, or trial for short. The learner’s behav- continues to perform the task. The cognitive mech-
ior changes gradually over trials: The learner makes anisms that are likely to operate during this phase
fewer erroneous or unnecessary steps, hesitates less, include the discovery of new, qualitatively different
and executes the appropriate actions faster. These strategies; the identification of redundancies and
changes can be represented by a learning curve (a.k.a. shortcuts; the optimization and speedup of actions
“practice curve”): If performance, measured, e.g., by and cognitive operations; and the replacement of
the time for task completion, is plotted as a function of multistep processes with retrieval from memory of
the amount of experience with the target task, mea- repeatedly produced answers.
sured, e.g., in terms of number of trials, the result is
The observable effects of practice – fewer errors,
invariably a negatively accelerated curve. That is, the
faster performance – are cumulative effects of the inter-
rate of improvement is fastest in the beginning, slows
actions among the multiple learning mechanisms. The
down as practice progresses, and eventually approaches
three phases should not be seen as sharply bounded.
an asymptote that represents the best possible perfor-
They represent gradual shifts in the relative importance
mance. There is disagreement about the best mathe-
of different mechanisms as practice progresses.
matical description of such curves, but the negatively
accelerated shape of empirical learning curves is one of
the most thoroughly documented regularities in the Important Scientific Research and
study of learning. An accurate theory must account Open Questions
for this phenomenon. However, it turns out that neg- 1. Feedback. The term “feedback” is imported from
atively accelerated learning curves can be derived from engineering. In the study of cognitive skill acquisi-
several different theoretical assumptions, so this test is tion, positive feedback (a.k.a. positive reinforcement)
less stringent than it first appears. is information to the effect that an action taken by
A variety of cognitive mechanisms have been pro- the learner was appropriate, correct, or useful, while
posed to explain the basic practice effects. It is useful to negative feedback (a.k.a. negative reinforcement) is
organize these by the phase during practice when they information to the effect that the learner’s action
are most likely to be active: was inappropriate, incorrect, or unhelpful. Feedback
is sometimes intrinsic to a task environment (e.g.,
1. The initial phase lasts from the first encounter with error messages in computer software), but a coach,
the target task until the task has been completed supervisor, teacher, trainer, or tutor can support skill
for the first time. Cognitive mechanisms that are acquisition by delivering additional feedback in the
likely to operate during this phase include the course of practice. The two central questions are
618 C Cognitive Skill Acquisition
when, under which circumstances, a tutor should strategies, and the encoding into memory of a
intervene, and what information should be included large number of particular cases have all been pro-
in a feedback message. posed as possible sources of flexibility. It is widely
Immediate feedback is more helpful than believed that varied problem solving experience is
delayed feedback. Other aspects of feedback have more likely to foster transferable skills than drill on
turned out to be less straightforward. If feedback very similar practice tasks. It is also widely believed
helps, it seems to follow that more feedback should that conceptual understanding of why a particular
help more. In some studies, increasing the fre- strategy works facilitates application of a strategy
quency of feedback resulted in more effective learn- to unfamiliar situations. Neither belief is strongly
ing. But in others, higher feedback density appears supported by research. Due to the intrinsic contra-
to impair learning. With respect to content, some diction between effectiveness and generality, it
studies show that bare bones feedback (“yes/no” or is unlikely that the transfer problem has a princi-
“right/wrong”) is less effective than feedback with pled solution.
explanations (“this answer is wrong, because. . .”), The problematic consequence for school learn-
while other studies have found no advantage for ing is that there is no way to ensure that skills
the explanatory content. Both positive and negative learned in a classroom will be applied outside
feedback can be helpful, but if the learner interprets school. In vocational and professional training sit-
negative feedback as punishment, it is likely to have uations, the standard solution to this problem is to
an adverse effect on motivation. Even when feed- trade off generality for effectiveness and conduct
back is purely informational, results vary. Some training in so-called high-fidelity training environ-
studies show strong effects of negative feedback, ments. These are designed to be as similar to the
while others do not. Common sense suggests that future application environment as possible. Exam-
negative and positive feedback in combination is ples include flight simulators for airline pilots and
more helpful than either in isolation, but this simulated battlegrounds for the military. Virtual
assumption has no extensive research support. reality technology makes this training strategy
The problem of feedback is central to the design more widely applicable.
of intelligent tutoring systems, educational software 3. Long-term practice effects. As practice progresses
systems that use artificial intelligence techniques over long time, the consequences depend on the
to compute on line the feedback to be delivered to type of skill and the characteristics of the training.
the learner. Tutoring systems are more helpful than A simple skill performed over and over again with
independent practice or lectures but less helpful little variation – a.k.a. drill – becomes automatized.
than human tutors. The effort to design more help- Automatized skills (a) are triggered when appropri-
ful tutoring systems would benefit from more deci- ate even without deliberate decision making, (b) are
sive research on the effects of different feedback rigid in their execution, and (c) impose low levels of
variables. The possibility of accessing tutoring sys- cognitive load. Automaticity can require thousands
tems via the Internet suggests that their importance of training trials.
will grow over time. Complex skill sets applied in varied situations
2. Transfer. The finding that cognitive skills become exhibit a different type of long-term outcome com-
more adapted to the particulars of the target task monly referred to as expertise. This is the outcome
during practice raises the question of how a mas- sought in professional training and other practice
tered skill can be applied (transferred) to tasks that scenarios. Expertise is characterized by fast but
differ in their details from the training task. Effec- flexible decision making and superior memory for
tive performance requires high specificity, while area-related information. Experts engage in deliber-
broad application requires abstraction. Yet, people ate practice, i.e., they intentionally vary already
tend to be both effective and flexible in their mastered performances to explore possibilities for
everyday behavior. Abstract declarative knowledge, improvement. The amount of practice required to
anticipation of the future situations in which a achieve top-level performance is approximately
skill is to be applied, the hierarchical structure of 10 years, if the learner practices 4 h a day, 6 days
Cognitive Tasks and Learning C 619
References
Ohlsson, S. (2008). Computational models of skill acquisition. In Cognitive Tasks and Learning
R. Sun (Ed.), The Cambridge handbook of computational psychol-
ogy (pp. 359–395). Cambridge: Cambridge University Press. LIESBETH KESTER, PAUL A. KIRSCHNER
VanLehn, K. (1996). Cognitive skill acquisition. Annual Review of
Centre for Learning Sciences and Technologies,
Psychology, 47, 513–539.
Open University in the Netherlands, Heerlen,
The Netherlands
Definition
Cognitive Strategies Cognitive tasks are those undertakings that require
a person to mentally process new information (i.e.,
▶ Learning Strategies acquire and organize knowledge/learn) and allow
620 C Cognitive Tasks and Learning
them to recall, retrieve that information from memory or similar situations – in other words to perform well
and to use that information at a later time in the same on cognitive tasks – one must first possess the necessary
or similar situation (i.e., transfer). individual cognitive skills for schema acquisition/
schema construction. Then, one must be able to coor-
Theoretical Background dinate the separate skills that constitute the task. In
addition, these skills must be integrated with prior
Cognitivism knowledge and existing attitudes. Finally, successful
The roots of cognitive psychology and the role of performance of cognitive tasks requires differentiation
cognitive tasks lie with David Ausubel’s Psychology by recognizing qualitative differences among the task
of Meaningful Verbal Learning (Ausubel 1963) and characteristics that influence the constituent skills that
Robert Gagné’s Conditions of Learning (Gagné 1977). have to be applied.
According to Gagné, cognitive tasks aim at the acqui-
sition of intellectual skills and consist of eight hier- Cognitive Tasks and Learning
archically organized cognitive processes: stimulus Performing cognitive tasks taxes the learner’s
recognition, response generation, procedure following, limited working memory (cf. Sweller 1988). In other
use of terminology, discriminations, concept forma- words, it induces significant cognitive load. Because
tion, rule application, and problem-solving. Gagné of this, effective learning can only commence if
identified five major categories of learning (verbal the specific instructions within a cognitive task are
information, intellectual skills, cognitive strategies, properly aligned with cognitive architecture (Van
motor skills, and attitudes), each requiring different Merriënboer and Kirschner 2007). In their Ten Steps
internal and external conditions for it to occur. to Complex Learning Van Merriënboer and Kirschner
Cognitivism was a response to behaviorism which outline an instructional design model based upon
saw learning as a simple response to environmental a whole-task approach and provide strategies to
stimuli. Ausubel, in response to behaviorism, believed align instruction to human cognitive architecture and
that understanding concepts, principles, and ideas is help people learn how to perform the complex cogni-
achieved through deductive reasoning requiring active tive tasks.
participation in of a learner whose actions are a conse- Part-task models of skill acquisition dominated
quence of thinking. He called this meaningful learning; the field of instructional design until the late 1980s. In
as opposed to rote memorization. that approach, one aspect of a skill was learned and
practiced until mastery, at which time a new – often
Schema Theory of Learning related aspect of the skill – was then learnt and mas-
That which is meaningfully learned is organized in tered, etc., until the “whole” skill was considered to be
schemata. The schema theory of learning (Anderson mastered. Van Merriënboer and Kirschner (2007)
1977) views organized knowledge as an elaborate stress the use of a whole-task model of learning since
network of abstract mental structures which represent part-task models have three major shortcomings,
how one understands the world. Schemata (1) are namely, they lead to compartmentalization (i.e., teach-
constructed by the learner, (2) are meaningfully ing knowledge, skills, and attitudes separately, thus
organized, (3) are added to and refined as an indi- hindering complex learning and competence develop-
vidual gains experience (Piaget: assimilation), (4) are ment), fragmentation (i.e., analyzing a complex learn-
reorganized when incoming data make this necessary ing domain in small pieces corresponding with
(Piaget: accomodation), and (5) are embedded in other specific learning objectives, and then teaching it
schemata and contain sub-schemata. In other words, piece-by-piece without paying attention to the rela-
learning can be seen as change in a learner’s schemata. tionships between pieces), and limit transfer (i.e.,
transfer paradox: using instructional methods that are
Cognitive Tasks highly efficient to reach specific learning objectives, but
To mentally process new information effectively, that are not suitable to reach transfer of learning.). Due
retrieve it from memory, and then use it in the same to this, there has been a growing interest in whole-task
Cognitive Tasks and Learning C 621
models of learning and instructional design. In dealing theories, etc.), and a way of reasoning which a scientist
with the learning of cognitive tasks, whole-task models uses to support or refute the claim. They must eval-
provide an alternative to part-task models. Whole- uate the quality of the argumentation and the informa-
task models, in contrast, analyze tasks as a coherent, tion used.
interconnected whole and then teach them from very A reverse task presents a goal state and an accept- C
simple, yet meaningful wholes that are representative able solution, but the learners have to trace the impli-
for the whole task to increasingly more complex cations for different claims (i.e., predict the given). In
wholes, fostering coordination, integration, and trans- the context of troubleshooting, for example, learners
fer of learning. might be told that a particular component is faulted
Whole meaningful tasks, thus, are seen as the driv- or has failed and predict the behavior of the system
ing force for learning. Easy-to-difficult sequencing based on this (i.e., what they should observe in order to
techniques and learner support and guidance, which reach a correct diagnosis themselves). Like case studies,
may be faded as learners acquire more expertise (i.e., reverse tasks focus learners’ attention on useful solu-
scaffolding), are studied as methods to deal with task tions and require them to relate solution steps to given
complexity. Second, there is a focus on the develop- situations.
ment of the whole person (i.e., learner-centered) rather An imitation task presents a conventional task in
than the acquisition of isolated pieces of knowledge, combination with a case study of an analogous task.
and the learner is co-responsible for a process of The solution presented in the case study provides
competency development. Third, there is a renewed a blueprint for approaching the new task, focusing
interest in the study of instructional methods that attention on possibly useful solution steps. Imitation
explicitly aim at transfer of learning. Methods that tasks are quite authentic, because experts often rely on
work the best for reaching isolated, specific objectives their knowledge of specific cases to guide their prob-
are often not the methods that work best for reach- lem-solving behavior on new problems – a process
ing integrated objectives and increasing transfer known in the field of cognitive science as case-
of learning (Van Merriënboer et al. 2006). A whole- based reasoning.
task approach takes this paradox into account and A nonspecific goal task stimulates the exploration of
is always directed toward reaching multiple, inte- relationships between solutions and the goals that can
grated objectives that go beyond a limited list of highly be reached by those solutions. It invites learners to
specific objectives. Therefore, whole-task approaches move forward from the givens and to explore the
are characterized by the use of mathemagenic instruc- problem space, which helps them construct cognitive
tional methods that give rise to meaningful learning schemas. This is in contrast to traditional, goal-specific
and transfer. problems that force learners to work backward from
the goal. For novice learners, working backward is
What This Means a cumbersome process that may hinder schema con-
Van Merriënboer and Kirschner (2007) present a series struction (Sweller 1988).
of cognitive task types which are well suited to the A completion task provides a given state, criteria
learner’s cognitive architecture and which are also for an acceptable goal state, and a partial solution.
aimed at carrying out and learning from whole tasks. Learners must complete the partial solution by deter-
Different types of cognitive learning tasks can be mining and adding the missing steps, either at the end
constructed by manipulating the information given to of the solution or at one or more places in the middle of
the learner, the goal state to be achieved by the learner, the solution. A particularly strong point of such tasks is
and/or the solution that the learner is required to come that learners must carefully study the partial solution
up with. Here, for the field of problems in the natural provided to them, because they will otherwise not
sciences for example, explanations and examples of the be able to come up with the complete solution. Well-
different types are presented: designed completion tasks ensure that learners can
In a case-study, learners receive a media claim, understand the partial solution and still have to per-
relevant articles, and/or information (i.e., facts, form a nontrivial completion.
622 C Cognitive Underpinnings of Primate Communication
survival and is as highly influential to cognitive schemas therapy. An emphasis needs to focus on examining
in the processing of information. the application with different types of family problems
The combination of the cognitive-behavioral and also cultural variations (Dattilio 1998b). It would
approach with families is equally effective with the also be interesting to examine the various character-
behavioral conditions, although cognitively focused istics of family members and determine what consti-
interventions tend to produce more cognitive change, tutes differential responses to treatment, as well as
while behavioral interventions modify behavioral optimal sequences of behavior and the restructuring
interactions. of schemas. Comparative studies, if conducted, would
help to isolate the specific characteristics that render
Important Scientific Research and cognitive-behavior family therapy effective and also
Open Questions discover which components are most advantageous
Cognitive-behavior therapy has been subjected to for integrative purposes with other modalities.
more controlled outcome studies than any other
therapeutic modality in existence (Dattilio 2001). Cross-References
There is substantial empirical evidence from treatment ▶ Application of Family Therapy on Complex Social
outcome studies, using cognitive-behavior therapy to Issues
indicate the effectiveness with relationships, although ▶ Behavior Modification, Behavior Therapy, Applied
most studies have primarily focused on the behavioral Behavior Analysis and Learning
interventions of communication training, problem- ▶ Behavior Therapy
solving training, and behavior contracts, with only
a handful of studies examining the impact of cognitive
restructuring procedures (see Baucom et al. 1998, for
References
Barrowclough, C., & Tarrier, N. (1992). Families of schizophrenic
a review that employs stringent criteria for efficacy).
patients: Cognitive-behavioral interventions. London: Chapman
Baucom et al. (1998) review of outcome studies indi- & Hall.
cate that cognitive-behavior therapy is efficacious in Baucom, D. H., Shoham, V., Mueser, K. T., Daiuta, A. D., &
reducing relationship distress. Cognitive-behavioral Stickle, T. R. (1998). Empirically supported couples and family
approaches gained popularity and respect among cli- therapy for adult problems. Journal of Consulting and Clinical
nicians, including couple and family therapists (Bitter Psychology, 66, 53–88.
Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive
2009; Dattilio 1998a; Dattilio and Epstein 2003; Epstein
Therapy of Depression. New York: Guilford.
and Baucom 2002; Davis and Piercy 2007; Nichols and Bitter, J. M. (2009). Theory and practice of family therapy and counsel-
Schwartz 2008). ing. Belmont: Brooks/Cole, Cengage Learning.
Epstein (2001) has produced an excellent overview Dattilio, F. M. (Ed.). (1998a). Case studies in couple and family
of the empirical status of cognitive-behavior therapy therapy: Systemic and cognitive perspectives. New York: Guilford
Press.
with relationships. More recently, Dattilio and Epstein
Dattilio, F. M. (1998b). Finding the fit between cognitive-
(2003) and Dattilio (1998a) published an overview of behavioral and family therapy. The Family Therapy Networker,
both couples and family therapy with additional empha- 22(4), 63–73.
sis on family schema. A more comprehensive text by Dattilio, F. M. (2001). Cognitive- behavioral family therapy: Con-
Dattilio (2010) outlines all of the research literature up temporary myths and misconceptions. Contemporary Family
to date. Sadly, the area of cognitive-behavior therapy Therapy, 23(1), 3–18.
Dattilio, F. M. (2010). Comprehensive cognitive-behavior therapy with
with couples has substantially more quantitative studies
couples and families. New York: Guilford Publications.
than that of family therapy (Baucom et al. 1998; Dattilio Dattilio, F. M., & Epstein, N. B. (2003). Cognitive-behavioral couple
and Epstein 2003; Epstein 2001; Dattilio 2010). The and family therapy. In G. Weeks, T. Sexton, & M. Robbins (Eds.),
most recent of the family therapy studies include Handbook of family therapy: Theory research and practice
the treatment of schizophrenia in the early 1980s, as (pp. 147–173). New York: Routledge.
Davis, S. D., & Piercy, E. P. (2007). What clients of couple
well as those studies conducted by Barrowclough and
therapy model developers and their former students say
Tarrier (1992). about change: Part 1. Model dependent common factors
An open question remains the need for future across three models. Journal of Marital and Family Therapy,
research with family cognitive-behavioral family 33(3), 318–343.
Cognitive-Code Learning C 625
Epstein, N. B. (2001). Cognitive-behavioral therapy with couples: of linguistic theories and the findings of psycholinguis-
Empirical status. Journal of Cognitive Psychotherapy, 15(2), tic research, cognitive psychologists and applied lin-
299–310.
guists, such as John B. Carroll and Kenneth Chastain,
Epstein, N. B., & Baucom, D. H. (2002). Enhanced cognitive therapy
for couples: A contextual approach. New York: Guilford Press. advocated the cognitive-code approach to the study of
Nichols, M. P., & Schwartz, R. C. (2008). Family therapy: Concepts and a second language as an alternative to the audio-lingual C
methods (8th ed.). Boston: Allyn & Bacon. method prevalent at the time. Cognitive-code learning
theory (Chastain 1971) proposes that learning a second
language requires explicit instruction and a study of
the language as a complex and rule-governed system
Cognitive-Behavioral Therapy (Carroll 1964). This approach took the view of a con-
scious study of the language structure as central and
▶ A Tripartite Learning Conceptualization of placed a great deal less emphasis on the development
Psychotherapy of a second language as a combination of skills. In
the current perspective on second language learning,
cognitive-code theory is largely seen as an updated
variety of the traditional grammar-translation method,
Cognitive-Code Approach with an attendant goal of overcoming the shortfalls
of the audio-lingual approach. At its core, cognitive-
▶ Cognitive-Code Learning code learning represents a theoretical, rather than a
pedagogical approach. In part due to the fact that this
theoretical proposal met with debate and skepticism,
its tangible outcomes in the form of curricula, methods,
Cognitive-Code Learning or teaching techniques did not materialize.
Providing learners opportunities for a great deal
ELI HINKEL of meaningful practice in a second language con-
Department of Anthropology, Seattle University, stitutes the central precept of the cognitive-code
Seattle, WA, USA approach. The main emphasis on meaningful practice
underscored the need for the learner first to under-
stand the language rules and then apply them in the
Synonyms context of practical language use. Thus, the explicit
Code-cognition approach; Cognitive-code approach; study of language rules, such as in grammar and
Cognitive-code learning theory vocabulary, was not only expected, but strongly
encouraged. In the context of structural linguistics
Definition and behavioral psychology, cognitive-code learning
Cognitive-code learning refers to a theory of second envisions practice to be meaningful when learners
language teaching and learning rooted in cognitivist clearly understand and are able to apply language
psychology and structural applied linguistics developed rules in practice. The essential difference between
in the 1960s. The theory emphasizes the central role the audio-lingual approach and the cognitive-code
of cognition in the conscious and explicit learning of approach is that in the former, structural learning
the rules of a language as a code. The cognitive-code without an explanation and pattern drills are seen as
approach to learning a second language sees it as a leading to modifications in the learners’ language
study of language as a complex system with the goal behavior, while in the latter, students need to
of gaining conscious control of the grammatical, lexical understand the linguistic rules before these can be
(vocabulary), and auditory patterns. implemented in practice. According to Carroll (1966,
p. 102), “the theory attaches more importance to the
Theoretical Background learner’s understanding of the structure of the foreign
Cognitive-code learning theory was proposed and language than to the facility in using that structure,
widely debated in the 1960s. Based on the foundations since it is believed that provided the student has
626 C Cognitive-Code Learning Theory
a smooth continuum of wavelengths, humans break distract, but some features prove critical. We are begin-
that continuum into a small set of labeled regions. ning to develop an understanding of these issues in
According to Berlin and Kay (1969), humans break terms of generalization, sample complexity, and com-
the color spectrum into 11 basic color categories, putational learning theory, but many questions remain,
although cultures differ in whether they have basic such as these: How can we measure the importance of C
color terms for all 11. We judge color differences as features? When does the discriminating power of a new
smaller if they come from the same category, and larger feature justify the added expense of increasing com-
if they come from different categories, even when the plexity? What are the best ways to recognize the impor-
wavelength difference is the same. The same effect tant features as we learn a task from scratch? How
characterizes perception of speech sounds such as the do our categorical assumptions change our view of
stop-consonant categories /ba/, /da/, and /ga/ (Harnad the world – and what is their effect on learning? It is
1987, p. 2). difficult to answer these questions in the general sense
Other evidence indicates that perceptual categories because relevance and value depend upon the task at
are the result of experience and learning. For example, hand, the agent, and the relative costs of computation
infants that grow up in a particular language environ- time, mistakes, and risk.
ment, say, English, appear to lose the ability to discrim- Choosing an appropriate representation often is the
inate speech sounds absent from that environment most critical step of solving a problem – as will be
during the first year of life (Werker and Tees 1984). apparent to anyone who has tried to multiply numbers
These changes suggest that human perception develops represented as Roman numerals. Cognitive economy
category distinctions that give us a functional advan- assumptions affect every field involving decision-
tage for interactions with our environment. making or skilled performance. Examples include the
Human reasoning and problem solving also appear following.
to benefit from our innate cognitive-economy assump- Writing: Effective technical writing provides the
tions. Herbert Simon (1957) used the term bounded reader with an appropriate level of detail. Too much
rationality to describe our limited cognitive capacity – detail will confuse the reader.
much too small to produce objectively rational behav- Teaching: Students need to develop effective repre-
ior in the real world. Therefore, we construct a simpli- sentational constructs for the topic at hand, and the
fied model of reality that allows us to discard details teacher must communicate concepts in terms the stu-
that appear tangential to our task. dents can understand.
Simon’s analysis has been born out in studies of Athletic performance: Keep your eye on the ball!
experts and novices. Experts appear to represent the Learn how to recognize an opponent’s intentions and
relevant details of their tasks much more efficiently pending actions.
than novices. The experts have learned “what to look Design: Objects may present perceived affordances
for,” and have learned to disregard spurious features. that enable the user to recognize how the object may be
For example, experiments with chess players have used. For example, some doors have a brass push plate
shown that a key difference between master players on one side and a pull handle on the opposite side,
and lesser players is that the masters are able to making usage obvious. Design of human–machine
immediately recognize the important attributes of interfaces (e.g., computer technology) can either make
a chess position (de Groot 1965). But this ability only human use error-prone and frustrating (and thus expen-
extends to chess positions from actual games. When sive), or intuitive and empowering. In automobiles and
presented with random chess boards, the experts had in airplane cockpits, cognitive economy assumptions
no advantage. may have life-and-death consequences.
Machine learning: Feature extraction and feature
Important Scientific Research and selection are both areas of active research at major
Open Questions conferences such as AAAI (the annual conference of
Cognitive economy is based on the common sense idea the Association for the Advancement of Artificial
expressed so aptly by Albert Einstein: “Make everything Intelligence) and ICML (the International Conference
as simple as possible, but not simpler.” Irrelevant details on Machine Learning). In order for researchers to
628 C Cognitivism
Cross-References
▶ Affordances Co-learning
▶ Categorical Representation
▶ Cognitive Efficiency ▶ Learning by Eliminating
▶ Expertise
▶ Judgment of Similarity
▶ Knowledge Organization
▶ Knowledge Representation
▶ Mathematical Models Theories of Learning Collaboration
▶ Mental Representation From Latin com- þ laborare to work jointly with others
▶ Role of Prior Knowledge in Learning Processes or together especially in an intellectual endeavor. A
synergic relationship among participants sharing their
References knowledge or skills, engaged in a specific context using
Berlin, B., & Kay, P. (1969). Basic color terms. Berkeley, CA: University implicit or explicit interaction rules to achieve one or
of California Press.
more valuable and situated outcomes.
de Groot, A. D. (1965). Thought and choice in chess. The Hague:
Mouton & Co.
Harnad, S. (1987). Introduction: Psychophysical and cognitive Cross-References
aspects of categorical perception: A critical overview. In
▶ Learning Through Social Media
S. Harnad (Ed.), Categorical perception (pp. 1–25). Cambridge:
Cambridge University Press.
Rescher, N. (1989). Cognitive economy: The economic dimension of the
theory of knowledge. Pittsburgh: University of Pittsburgh Press.
Rosch, E. (1978). Principles of categorization. In E. Rosch &
B. B. Lloyd (Eds.), Cognition and categorization (pp. 27–48). Collaboration Scripts
Hillsdale, NJ: Lawrence Erlbaum.
Simon, H. A. (1957). Models of man: social and rational. New York:
Wiley.
KATI MÄKITALO-SIEGL1, INGO KOLLAR2
1
Werker, J. F., & Tees, R. C. (1984). Cross-language speech perception: Finnish Institute for Educational Research, University
Evidence for perceptual reorganization during the first year of of Jyväskylä, University of Jyväskylä, Finland
life. Infant Behavior and Development, 7, 49–63. 2
Department of Psychology, University of Munich,
Munich, Germany
Synonyms
Cognitivism Cooperation scripts; Instructional support; Scaffold-
A theory of learning which considers that learning can ing; Scripted cooperation
be compared to the way a computer processes infor-
mation. The learner gathers information and con- Definition
structs an internal symbolic database of real world Collaboration scripts are a specific type of scaffolds
objects and experiences. Information may come from for collaborative learning that is characterized by its
a perceived expert (the teacher). Learning focuses on focus on supporting learning through direct manipu-
structured schemas. lation of collaboration processes rather than through
Collaboration Scripts C 629
offering content-specific support. A collaboration a text, then summarizing it, then making predictions).
script is a set of instructions which aims to guide and Fourth, activities are often clustered to roles that are
support two or more learners to interact and behave distributed (e.g., an explainer and a listener) and
during collaborative learning in a way that all learn- may be switched among the learning partners. Finally,
ing partners benefit from collaboration. The aim of scripts can vary with respect to their type of representa- C
the collaboration script is to enhance learning of tion, that is, they may be presented to the learners as
group members by engaging them in cognitive (e.g., oral instructions, but also as instructional texts or they
explaining, questioning, summarizing), metacognitive may be embedded in the communication interface in
(e.g., monitoring, regulating, formulating arguments), a CSCL environment.
and social activities (e.g., taking turns, listening, Two prototypical realizations of the collaboration
playing specific roles, etc.) related to individual script approach are “Scripted Cooperation” (O’Donnell
knowledge and skill acquisition. Collaboration scripts 1999) and “ArgueGraph” (Dillenbourg and Jermann
specify, sequence, and distribute these activities among 2007). The Scripted Cooperation approach supports
the learning partners of a group and often attach groups of two learners in learning from text. The text
them to specific collaboration roles. Thereby, they can is split up into paragraphs by the learners or the teacher
vary with respect to how much structure they provide: before learners read the first passage individually. After
so-called macro-scripts bring specific collaboration that, learners put the text aside and one learner has
phases in a certain order, but do not give further the role of the recaller, whose aim is to recall the text
instructions on how learners should act in these information as completely as possible. Simultaneously,
phases. So-called micro-scripts, in contrast, offer the learning partner is in the role of the listener, who
more specific instructions on how to perform certain tries to catch misconceptions and identify omissions.
activities, for example, by prompting one learner to After this, learners jointly elaborate the text content to
give a constructive critique on a learning partner’s make it more memorable. Then the next passage of
contribution (e.g., “What I did not understand in the text is read and the procedure of the activities is
your contribution was. . .”). repeated as in the first round except that learners switch
their roles. The learning objectives for the learners are
Theoretical Background acquiring the domain-specific content knowledge and
There is plenty of evidence both in face-to-face and domain-general text-learning strategies. In order to
▶ computer-supported collaborative learning (CSCL) reach these goals, learners engage in cognitive activities,
situations showing that unstructured collaboration such as explaining, and metacognitive activities, such
usually does not lead to high achievements with as monitoring. Activities are sequenced in a fixed order,
respect to learning. Therefore, different approaches according to which learners need to read a text, then to
for instructing learners’ collaborative activities, called summarize it and identify misconceptions and omis-
collaboration scripts, have been created both for sions, and jointly elaborate the text content. Learners
face-to-face and CSCL situations (Fischer et al. 2007). are assigned to play roles, such as the recaller and the
Research on collaboration scripts includes a vast variety listener, and these roles are switched several times dur-
of different script variants. Despite this diversity, five ing the learning process. The script instructions are
central conceptual components of collaboration scripts usually presented by the teacher and practiced prior
can be identified, which are the following (Kollar et al. to collaboration.
2006): First, collaboration scripts are directed towards Unlike Scripted Cooperation, which offers instruc-
specific learning objectives, such as the acquisition of tional support for collaborative processes occurring
domain-specific knowledge or domain-general compe- within dyads of learners, ArgueGraph aims to integrate
tences. Second, they try to engage learners in parti- small group, individual, and whole-class activities in
cular activities that are functional with respect to the computer-supported classroom situation. First,
reaching these objectives (e.g., explaining, argumenta- learners are asked individually to fill in a computer-
tion, questioning). Third, these activities are typically based multiple-choice questionnaire (e.g., on the topic
to be shown in a particular sequence (e.g., first reading “Theoretically driven courseware design”) and give an
630 C Collaboration Scripts
argument for each choice in an open text window. orchestration of learning processes within a complex
Once each student has completed the questionnaire, social system by integrating individual, small group,
the software system creates a graph which shows and whole-class activities, it is a proponent of more
learners’ positions on different predefined dimensions macro-level scripting which leaves learners consider-
(e.g., permissiveness) compared to other learners based able freedom to interact in a way they want and play
on their answers. Second, the graph is looked through different roles. Scripted Cooperation, in contrast, aims
and discussed with the whole class in the plenary to support small group activities in a more fine-grained
session. After this, the system automatically builds manner and requires both learners to engage in partic-
pairs by selecting the learners who have the most ular predefined activities and play two roles during
contrasting positions in the graph to work together. collaboration. Therefore, Scripted Cooperation repre-
Third, the dyads’ task is to answer the same question- sents more micro-level scripting. However, activities
naire together, select one answer per question, and and roles may be even more prestructured by using
finally give a joint argument for their selection. During prompts or sentence starters (e.g., “What does. . ..
that process, significant argumentation activities are mean?,” “Tell me more about. . .”), which has been
expected to happen. After this phase has been com- realized especially in purely computer-based collabora-
pleted, the system shows a new graph based on the tion scripts that can be found in the literature (e.g.,
answers of individuals and pairs as well as an aggre- Weinberger et al. 2005).
gation of the arguments. Fourth, there is a plenary
phase in which the teacher discusses with the learners Important Scientific Research and
about their arguments, asking explanations, organizing Open Questions
their arguments into theories, clarifying definitions, The research conducted with collaboration scripts
etc. Finally, each learner writes a synthesis of arguments shows positive effects with respect to domain-specific
from a specific question. Learning objectives are to knowledge and domain-general competence. However,
acquire domain-specific knowledge (e.g., courseware there are variations in learners’ outcomes. Even highly
design and learning theories) and domain-general coercive collaboration scripts cannot be expected to
competences (e.g., argumentation). Activities students completely determine the success of collaborative
engage in are, for example, elaborating, explaining, and learning. Rather, the students also bring “internal”
formulating arguments and counterarguments. The collaboration scripts with them which guide their
activities are sequenced both with respect to their behavior during collaboration. These “internal” collab-
type and the social level (individual, small group, ple- oration scripts have been acquired through repeated
nary) on which they are supposed to occur. For exam- experience in collaborative situations and are highly
ple, in pairs, learners are not only asked to formulate resistant to change (Schank 1999). Therefore, learning
arguments, but also counterarguments, when trying can be interrupted if learners’ internal scripts are
to reach a joint position with their fellow learners. inconsistent with a given external script or if the exter-
Learners are not explicitly assigned to the roles, but nal script overestimates learners’ skills and ability. Fad-
they are, for example, taking the roles of the opponent, ing might be a solution for reconciling external and
the defender, or the explainer. Switching roles takes internal scripts in an adaptive way in order to avoid
place when learners are engaging in different activities. over- and under-scripting problems. This approach
The script’s representation is located in both the means that less and more support can be provided
teacher’s instruction and the particular design of the during activities depending on an individual learner’s
computer-based learning environment which provides or group’s needs. In order to increase or decrease sup-
instruction on what is supposed to happen during the port in the right moment means that the group
particular learning phase. processes need to be observed in real time. Yet, moni-
Scripted Cooperation and ArgueGraph represent toring multiple groups on a more detailed level at the
a considerable amount of diversity that also becomes same time is an impossible task for one teacher in
apparent when more collaboration scripts are consid- a classroom, but could be done by technology. First
ered in the literature. As ArgueGraph focuses on an computer-supported analysis software tools are within
Collaborative Learning C 631
Important Scientific Research and time of this study a report by the American Association
Open Questions for the Advancement of Science advised that the work
Research at all levels of schooling has indicated that of professionals in the sciences is not done in isolation,
students learn and retain more when they have agency but collaboration is necessary at all levels. Current
in the process and have opportunities to speak, listen, instructional methods that focused on traditional C
share, interact, reflect, and be active. Over 750 studies teaching rather than student learning raised concerns
have focused on the positive aspects of collaborative that professionals were ill prepared to solve real world
learning and its underlying learning theory (Johnson problems in cooperative ways. Consequently, frequent
and Johnson 2005). Two studies will be highlighted group activity in the classroom and experience sharing
here to illustrate the influence that both simple and responsibility for learning was called for at a national
more complex forms of collaboration can have on level. In the meta-analysis conducted of the 39 highest
student outcomes in comparison to traditional teach- quality studies, the use of small group learning for
ing practices. undergraduates in STEM classes showed significant
In a well-known study Ruhl et al. (1987) examined and positive effects on ▶ achievement, persistence in
what happens when students are given opportunities to courses, and favorable attitudes toward courses when
share understanding of classroom content at key points compared to students who did not work collabora-
in a lecture sequence. Two groups of university students tively. On these three broadly defined outcome mea-
received the same instruction in two different ways. In sures the effect size for all three variables was about
the experimental group, an instructor paused for 2 min 0.50. Achievement differences which included grades
on three occasions (intervals between pauses were and test scores showed an effect size of 0.51. This
approximately 15 min) during each of five lectures. difference would move a student from the 50th per-
During the pauses, while students simply worked in centile to the 70th percentile in a course. An effect size
pairs to discuss and rework their notes, no interaction of 0.46 was identified in the area of persistence, indi-
occurred between instructor and students. At the cating that by using small group and collaborative
end of each lecture, students were given 3 min to learning methods classes and institutions would
write down everything they could remember from the reduce their attrition by 22%. Student attitudes
lesson. Twelve days after the last lecture students were about their own competence and the subject matter
also given a multiple-choice test to measure long term were also positively affected by their exposure to
retention. A control group received the same lectures as small-group instruction. The effect size on this mea-
those in the “pause procedure” group and was similarly sure was 0.55 for attitudes in small group settings
tested. In two separate courses repeated over two versus an average effect of 0.28 for other classroom
semesters, the results were consistent. Students who interventions.
experienced more student-to-student interaction and
were more involved in the learning process did signif- Cross-References
icantly better on the daily assessments and on the final ▶ Action-Based Learning
multiple-choice test. The magnitude of the difference ▶ Active Learning
in mean scores between the two groups was large ▶ Collaborative Learning and Critical Thinking
enough to make a difference of two letter grades. This ▶ Collaborative Learning Strategies
study suggests that if teachers talk less (even 6 min less ▶ Collaborative Learning Supported by Digital Media
as noted in this study) and brief pauses for collabora- ▶ Group Cognition and Collaborative Learning
tion are engineered, students can learn more effectively. ▶ Group Learning
In a highly regarded study funded by the National ▶ Learning in the Social Context
Science Foundation, Springer et al. (1999) reviewed ▶ Multimodal Learning Through Media
hundreds of studies to conduct a metanalysis of the ▶ Social Construction of Learning
effect of small-group/collaborative instruction on stu- ▶ Social Interactions and Learning
dent outcomes in university level science, technology, ▶ Social Learning Theories
engineering, and mathematics (STEM) classes. At the ▶ Trust into e-Learning
634 C Collaborative Learning and Critical Thinking
two to four students while a group of three to five activities of individual learners. A psychoanalysis of the
students may be appropriate for a semester-long pro- group discussions reveals useful information. The goal
ject (Slavin 1995). Thus, a group size of three to four is is to enhance the probability that interactions in a
optimum and promotes positive interdependence, yet group are educational and result in enhancing higher-
provides sufficient diversity of opinions and back- order thinking skills. C
grounds, which is influenced by group composition. Thinking is often casual and informal but critical
Much of the literature emphasizes that groups thinking calls for persistent effort to apply theoretical
should be heterogeneous when possible; heterogeneity constructs to understanding the problem, consider evi-
among group members refers to general differences like dence, and evaluate methods or techniques for forming
age, gender, race, ethnicity, and performance in school, a judgment. The cognitive skills of analysis, interpreta-
or task-specific differences like proficiency in the sub- tion, inference, explanation, evaluation, and of moni-
ject-matter. Studies indicate that some difference of toring and correcting one’s own reasoning are at the
viewpoints is required to trigger interactions but within heart of critical thinking (APA 1990). Critical thinking
the boundaries of mutual interest and intelligibility. not only mimics the process of scientific investigation –
Studies indicate that grouping learners with even distri- identifying a question, formulating a hypothesis, gath-
bution of abilities results in better learning when com- ering and analyzing relevant data, using it to test and
pared to learning outcomes of randomly mixed groups eventually accepting or rejecting the hypothesis, and
with varied student abilities. There is no clear way to finally drawing conclusions – but executes it repeatedly.
maximize group diversity and prevent individual isola- Collaborative learning facilitates the expression of
tion. An advantageous compromise is to cluster at least the thought processes in a non-stressful environment
two students of the same kind, say two women or two and provides opportunities to examine and reexamine
students of common ethnicity, or two students of beliefs and conceptions of the subject-matter in the
same ability, in a group of four (Cooper et al. 1990). light of evidence that may or may not support them.
A consciously designed group permits a healthy balance When students are confronted with different interpre-
of homogeneity and heterogeneity among its members. tations of the same situation, the peer support system
For collaborative learning to be effective, the makes it possible for the learner to internalize new
instructor must view teaching as a process of devel- knowledge and convert that into tools for intellectual
oping and enhancing students’ ability to learn. The functioning. The medium provides students with
instructor’s role is not to transmit information, but to opportunities to analyze, synthesize, and evaluate
serve as a facilitator for learning. This involves creating ideas cooperatively. The informal setting facilitates dis-
and managing meaningful learning experiences and cussion and interaction. This group interaction helps
stimulating students’ thinking through real-world students to learn from each other’s scholarship, skills,
problems. Yet, the task must be clearly defined and be and experiences.
guided by specific objectives. There is no reason to When collaboration is structured, group diversity
expect that unstructured collaboration will result in in terms of age, gender, ethnicity, and knowledge and
the expected learning outcomes so this predicament experience contributes positively to the learning process.
has been tackled by the use of scripts. Students are asked to go beyond mere statements of
Scripts structure collaborative learning by creating opinion by giving reasons for their judgments and
roles and mediating interactions while allowing for reflecting upon the criteria employed in making these
flexibility in dialog and activities (Kollar et al. 2006). judgments. Thus, each opinion is subject to careful
Scripting is a compromise between the constraints scrutiny, and the ability to admit that one’s initial opin-
usually induced by instructional design and the free- ion may have been incorrect or partially flawed is valued.
dom of collaborative learning. There are two broad Proponents of collaborative learning claim that the
types of scripts: macro-scripts and micro-scripts. active exchange of ideas within small groups not only
Macro-scripts aim at creating situations within which increases interest among the participants but also
desired interactions will occur by describing groups, promotes critical thinking. According to Johnson and
roles, and phases while micro-scripts emphasize the Johnson (1989), there is persuasive evidence that coop-
communication process students must engage in and erative teams achieve at higher levels of thought and
636 C Collaborative Learning Environment
retain information longer than students who work virtualization for enhancing critical thinking. A psy-
quietly as individuals. The shared learning gives stu- choanalysis of online group discussions could reveal
dents an opportunity to engage in discussion, take useful information.
responsibility for their own learning, and thus become
Cross-References
critical thinkers. Gokhale (1995) found that students
▶ Collaborative Learning
who participated in collaborative learning performed
▶ Collaborative Learning Strategies
significantly better on a critical-thinking test than stu-
▶ Cooperative Learning
dents who studied individually, while both groups did
▶ Creative Inquiry
equally well on a drill-and-practice test. Students are
▶ Creativity and Learning Resources
capable of performing at higher intellectual levels when
▶ Critical Learning Incidents
asked to work in collaborative situations than when
asked to work individually. References
The development and enhancement of critical- American Philosophical Association. (1990). Critical thinking:
thinking skills is one of the primary learning goals in A statement of expert consensus for purposes of educational
technical disciplines. Educational research investigates assessment and instruction. ERIC document ED (pp. 315–423).
Cooper, J., Prescott, S., Cook, L., Smith, L., Mueck, R., & Cuseo, J.
effective methodologies for nurturing higher-order
(1990). Cooperative learning and college instruction: Effective use
thinking skills and preparing students to deal with of student learning teams. Long Beach: California State University
increasingly complex workplace problems. Researchers Foundation.
report that students working in small groups tend to Gokhale, A. A. (1995). Collaborative learning enhances critical think-
learn more of what is taught and retain it longer than ing. Journal of Technology Education, 7(1), 22–30.
Johnson, D. W., & Johnson, R. T. (1989). Cooperation and competi-
when the same content is presented in other instruc-
tion: Theory and research. Edina: Interaction Book Company.
tional formats. Additionally, students learn how to Kollar, I., Fischer, F., & Hesse, F. (2006). Collaboration scripts –
communicate effectively, provide leadership, and prac- A conceptual analysis. Educational Psychology Review, 18(2),
tice social skills. 159–185.
Slavin, R. E. (1995). Cooperative learning: Theory, research, and prac-
Important Scientific Research and tice (2nd ed.). Boston: Allyn & Bacon.
Open Questions
The explosion of knowledge and information technol-
ogy has altered the characteristics of the learning envi-
ronment; higher education continues to adapt to the Collaborative Learning
digital culture and changes in student learning styles. Environment
Today, it is even more imperative that students acquire
critical thinking skills to manage information overload. ▶ Online Collaborative Learning
So the question is, how do we investigate the effective-
ness of collaborative learning to enhance critical think-
ing skills in digital environments? How relevant is
heterogeneity among group members when avatars Collaborative Learning
are taking the place of real people? How does an instruc- Strategies
tor provide structure and effectively intervene in asyn-
chronous communications? What is the difference in ALICE UDVARI-SOLNER
dynamics of face-to-face and online communications? Department of Curriculum and Instruction, University
The corporate culture is changing too with virtual of Wisconsin-Madison, Madison, WI, USA
meetings, remote access, and globally spread-out teams
becoming a reality. Critical thinking is part of a life-
long learning process and collaboration fosters its Synonyms
development through discussion, clarification of Active learning strategies; Cooperative learning strate-
ideas, and evaluation of others’ ideas. Future research gies; Small group learning strategies; Team learning
studies need to investigate the implications of strategies
Collaborative Learning Strategies C 637
12 = 120 and 15–13 = 2. The teachers could differ- interactions that take place in the context of this
entiate instruction easily by creating cards (and collaborative learning strategy require and promote
ultimately equations) that ranged in difficulty level. positive interdependence. To engage in the activity
Some student groups could form algebraic equa- and ultimately be successful students cannot function
tions, others fractions, and still others could repre- in isolation. By sharing their knowledge and finding
sent the process of addition or subtraction. relationships between what is represented on their card
● In a university level pharmacology course for phy- and others’, the individual is ultimately promoting the
sician’s assistants, drug types identified by their group’s achievement of joint goals (i.e., to constitute
generic names were placed on cards. Students were a meaningful category that integrates each individual’s
asked to classify themselves in multiple ways by the contribution).
use of the drug, the side effects of the drug, and the Representation of social justice principles: In this
populations of patients who benefit from its use. example students are placed in an empowering and
“knowing” position at the outset of the learning expe-
Representation of social constructivism: A critical
rience. Rather than assuming the students are not
tenet of this theory is that knowledge or the way
knowledgeable and must be taught what is relevant
humans understand their experiences and reality is
from an instructor, students individually and then col-
not simply constructed, it is co-constructed through
lectively must use their existing knowledge to discover
the frameworks of language and culture in relation-
and make personal meaning from the content. They are
ships among individuals. In this example social dis-
not passive recipients of instruction that is dictated by
course is not only encouraged but required to make
others but have significant agency in their own learn-
meaning of the academic content. The individual must
ing. The teaching/learning relationship is reoriented to
seek out others to make deliberate comparisons, judg-
one that becomes a dialogue first among students and
ments, and analyses. In doing so, each interaction with
then with the teacher who is informed by the students’
another class member reveals new perceptions and
discovery of new patterns and conceptions. Students
interpretations, consequently shaping new knowledge
have an equitable role in conveying relevant concepts
that has been developed collaboratively within a unique
alongside the teacher. This process promotes greater
social context. In addition, learners have multiple
spontaneity in instruction and assures instructional
opportunities to interact with and learn from more
time is not spent directly teaching what students
competent peers during the interchange of information
already know or could discover.
representing the zone of proximal development defined
by Vygotsky (1978). Important Scientific Research and
Representation of social learning theory: Social Open Questions
learning theory emphasizes that by observing others Research regarding collaborative learning strategies is
and engaging in reciprocal social and academic inter- generally subsumed under broader investigations of
actions the individual develops new and more complex collaborative learning. If collaborative learning strate-
behavioral and intellectual repertoires. The strategy gies are held distinct from cooperative learning, it is
Classify, Categorize, and Organize establishes an arena difficult to find studies that have extensively investi-
for individuals to observe the language and behavior of gated the use of one particular strategy. Overall in
other group members while problem solving. Models reviews of research, outcome measures of achievement,
are present as exemplars for appropriate attitudes, reac- reasoning, frequency of new idea generation, and trans-
tions to questions, and higher level thinking skills. fer of content learned from one situation to another
Since students must integrate their knowledge and were found to be superior in collaborative learning
information and then convey it to the rest of the par- approaches as compared to competitive or individual-
ticipants, there is opportunity to rehearse or practice istic structures (Barkley et al. 2005; Johnson and John-
new behavior. son 2005). Additionally, research examining the use of
Representation of social interdependence theory: collaborative learning strategies as forms of differenti-
Social interdependence exists when the outcomes of ation in instruction and vehicles to promote the inclu-
individuals are affected by the actions of others. The sion of students with disabilities in general education
Collaborative Learning Supported by Digital Media C 639
environments are being established (Udvari-Solner structure over another but to create a balance of expe-
and Kluth 2007). It seems the critical research question riences in the classroom that serves the needs of mul-
is no longer: Should collaborative learning be installed tiple learners well.
in our elementary, secondary, and higher education
classrooms? Instead the pertinent question arises: Cross-References C
When should collaborative learning be selected over ▶ Active Learning
other approaches? ▶ Collaborative Learning
Collaborative learning is not a panacea for all ▶ Collaborative Learning and Critical Thinking
instructional purposes. It is also wrought with poten- ▶ Collaborative Learning Supported by Digital Media
tial downfalls in implementation, the learning process, ▶ Cooperative Learning
and group dynamics. Ineffectively sharing information ▶ Group Cognition and Collaborative Learning
held by individuals to the group, social loafing, limita- ▶ Group Learning
tions in information processing and conflict resolution, ▶ Social Construction of Learning
and the ability to rectify failing projects have been ▶ Social Learning Theories
documented problems (Kirschner et al. 2009). Conse-
quently, an instructor must make conscious choices
regarding the most appropriate instructional strategy
References
Barkley, E. F., Cross, K. P., & Major, C. H. (2005). Collaborative
to match the demands of an academic task. Kirschner learning techniques: A handbook for college faculty. San Francisco:
et al. (2009) use cognitive load to theory to propose Jossey-Bass.
that the complexity of the task should influence the Johnson, D., & Johnson, R. (2005). New developments in social
decision whether to use individualistic or collabora- interdependence theory. Genetic, Social, and General Psychology
tive approaches. On a basic level, cognitive load theory Monographs, 131(4), 285–358.
Kirschner, F., Paas, F., & Kirschner, P. A. (2009). A cognitive load
relates to the perceived mental effort expended by the
approach to collaborative learning: United brains for complex
individual under specific instructional conditions. tasks. Educational Psychology Review, 21, 31–42.
"
Roschelle, J., & Teasley, S. D. (1995). The construction of shared
It is, therefore, hypothesized that the more complex
knowledge in collaborative problem solving. In C. O’Malley
the learning task (i.e., the higher the intrinsic cognitive
(Ed.), Computer supported collaborative learning (pp. 69–97).
load), the more efficient and effective it will be for Berlin: Springer.
individuals to collaborate with other individuals in Udvari-Solner, A., & Kluth, P. (2007). Joyful learning: Active and
a manner that reduces this load. By contrast, less com- collaborative learning in inclusive classrooms. Thousand Oaks:
plex tasks that can easily be solved by a single individ- Corwin Press.
Vygotsky, L. S. (1978). Mind and society: The development of higher
ual will lead to less efficient learning in groups than in
psychological processes. Cambridge, MA: Harvard University
individuals alone, because the required group commu- Press.
nication and coordination process (i.e., transaction
costs) impose an additional cognitive load upon the
group members, regardless of whether this communi-
cation and coordination is beneficial to learning or not
(Kirschner et al. 2009, p. 37). Collaborative Learning
The critical message for teachers as they design Supported by Digital Media
assessment, curriculum, and instruction is that the
learning activities designated for collaborative interac- CHRISTIANE METZNER, RICARDO A. CATTAFI
tions should be complex enough in nature that they Universidad Central de Venezuela, Caracas, Venezuela
cannot be easily carried out by individuals. In addition,
based on the replete research history, collaboration in
learning carries with it opportunities to build valued
academic skills concurrently with essential social skills Synonyms
that are required for complex human relationships. The Collaborative e-learning; Computer-supported collab-
key is not to make exclusive choices for one learning orative work; Computer-supported cooperative work
640 C Collaborative Learning Supported by Digital Media
Extranets, blogs, online forums, wikis, podcasts, tools and techniques but also its social, organizational,
lifestreams, social bookmarking, Web communities, and psychological impact. Groupware applications
social networking, and avatar-based virtual reality. integrate concurrent activities of users working on
They are used in a wide spectrum of problem domains, a single project connected on a network or Internet
including business computing. Teleconferencing tools (Wilson 1995). It is worthwhile noting that “collabora- C
enable interactive information sharing and each par- tive” and “cooperative” are sometimes used indistinctly
ticipant can set and access data in a shared blackboard; in the definition of CSCW; however, some authors
video and audio are used for information exchange, consider a semantic difference between these two con-
forums for asynchronous virtual discussions, and chat cepts (Dillenbourg and Schneider 1995). Dillenbourg
rooms are platforms for synchronous virtual discus- and Schneider argue the difference lies in how the
sions. Online conferencing and email are two technol- tasks to be accomplished are decomposed: cooperative
ogies available and easy to use which has made them means tasks are decomposed hierarchically into
a tool of choice for collaborative courses (e.g., Jonassen independent subtasks; collaborative means tasks are
et al. 1995; Warschauer 1997). decomposed hetero-hierarchically into interchange-
The key to implementing successfully these strate- able layers (Cattafi and Metzner 2007).
gies lies in the analogy between mental structures Coordination is a process used to exchange infor-
and processes and the associative structure and mation among people using a common system of
hyperlinking processes of the Web. The challenge is to symbols, signs, and behavior, requiring a dependency
construct an instructional environment accurately management between activities and support of inter-
reflecting the instructor’s knowledge structure (Miller dependencies among participants (Bordeau and Wasson
and Miller 1999). 1997). Cooperation requires coordination when the
Several guidelines for the successful application of a results are to be integrated while collaboration is
digitally supported teaching and learning strategy are usually a synchronous process. Therefore, communi-
proposed: establishing a highly structured, positive cation is essential in any coordination or collaboration
learning environment that encourages individual activity.
responsibility and creates high expectations, to teach In a wider context, Computer-Supported Collabo-
collaboratively; creating spaces for learner collabora- rative Work, Groupware, Computer-Supported Coop-
tion and peer review; redefining the instructor’s role; erative Work, and Collaborative Learning Supported by
building a community; and exploiting time (Kuriloff Digital Media are considered teaching/learning strate-
2005). gies supporting Web education or e-learning, where all
educational activities use digital media and software
Important Scientific Research and tools, and Internet is the communication platform.
Open Questions Computer-mediated communication involves the use
Computer-Supported Collaborative Work, Group- of computer communication technologies to connect
ware, Computer-Supported Cooperative Work, and learners. It can be used in various forms in a teaching/
Collaborative Learning Supported by Digital Media learning process for content publication, support of
are frequently used as synonymous; however, this administrative tasks, increasing availability via elec-
point of view is at least theoretically incorrect (Cattafi tronic tutorials and promoting collaborative learning
and Metzner 2007). by enhancing communication between learners and
Computer-Supported Collaborative Work (CSCW) instructors. In Web education or e-learning, strategies
is a multidisciplinary research field dealing with the can be collaborative or cooperative and encompass
development of tools and techniques providing sup- cognitive procedures for self-control and self-
port to people performing shared tasks on a network or regulation applied by participants on attention,
a distributed platform (Greif and Cashman 1988). memory, and comprehension; in a collaborative form,
Although the terms “CSCW” and “Groupware” are activities are performed in the classroom – virtual or
generally used indistinctly, some authors point out real – by small groups of learners after explanations of
that CSCW focuses not only on studying groupware the instructor.
642 C Collaborative Learning Supported by Digital Media
Information technologies supporting Web educa- and meaning can also impact the process of collabora-
tion or e-learning should have the following properties: tive learning and tools. Additionally, collaborative
accessibility, multiplatform, multimedia format for learning could affect power relations in participants
displaying information, graphical interfaces, group and influence performance. These power relations can
restricted access, hypertext structure of information be studied by allowing role changes of the participants.
and content, interpersonal communication, learner
follow-up, tools for collaborative work, learner Cross-References
management and control, creation of evaluation and ▶ Collaborative Learning
self-evaluation assignments, access to information ▶ Computer-Based Learning Environments
and content on the Web, and interactions among ▶ e-Learning and Digital Learning
group members. In e-learning the learning process is
viewed in terms of the increasingly skilled participation References
of members in a knowledge-based community. Bordeau, J., & Wasson, B. (1997). Orchestrating collaboration in
When developing a collaborative/cooperative pro- collaborative telelearning. In Proceedings of the eighth world con-
ject some degree of structuring groups is necessary. A ference on artificial intelligence in education (pp. 565–567).
Amsterdam: IOS Press.
structure is the result of an organization viewed as a
Cattafi, R., & Metzner, C. (2007). A didactic experience in collabora-
social or administrative entity (Chiavenato 1999). As
tive learning supported by digital media. Journal of Issues in
social entities, organizations consist of people having Informing Science and Information Technology, 4, 15–28. Retrieved
to reach some goals. Every goal requires work to be March 15, 2010, from http://proceedings.informingscience.org/
accomplished. The work is decomposed into tasks and InSITE2007/IISITv4p015-028Catt351.pdf.
assigned to members of the group. Formal organizations Chiavenato, I. (1999). Introducción a la teorı́a general de la
administración (5th ed.). Mexico: Mc Graw Hill.
are based on a structure rigorously defined in official
Dillenbourg, P., & Schneider, D. (1995). Collaborative learning
documents and on a rational division of tasks to be and the Internet. In Proceedings of ICCAI 1995. Retrieved
accomplished, specializing the functions and entities March 20, 2010, from http://tecfa.unige.ch/tecfa/tecfa-research/
by activities. In informal organizations the structure CMC/colla/iccai95_1.html
emerges on the fly as an outcome of human relations Greif, I., & Cashman, P. (1988). CSCW: A book of readings. San Mateo:
Morgan Kaufmann.
based on friendship or antagonism between individuals
Holzinger, A. (2002). Learning. Cognitive fundamentals of
playing specific roles in the underlying formal organiza-
multimedial information systems. Multimedia Basics, 2, 55.
tion; hence, groups emerge that are not considered in the Retrieved April 12, 2010, from http://www.basiswissen-
official documents (Cattafi and Metzner 2007). multimedia.at.
As administrative entities, organizations are respon- Jonassen, D., Davidson, M., Collins, M., Campbell, J., & Haag, B.
sible for planning, integrating, structuring resources, cre- (1995). Constructivism and computer- mediated communica-
tion in distance education. The American Journal of Distance
ating entities, and assigning activities. According to the
Education, 9(2), 7–26.
degree of formality of the relationships between the Kuriloff, P. (2005). Breaking the barriers of time and space:
members of an organization, different kinds of organiza- More effective teaching using e-pedagogy. Innovate Journal of
tional structures can be identified: bureaucratic, team, Online Education, 2(1). Retrieved March 13, 2010, from http://
and spontaneous cooperation based on teams and col- Innovateonline.info
Miller, M. S., & Miller, L. K. (1999). Using instructional theory to
laborations. Whether communication strategies consider
facilitate communication in web-based courses. Educational
time elements (i.e., synchronous vs. asynchronous)
Technology & Society, 2(3), 106–114. Retrieved April 3, 2010,
allows classifying organizations as hierarchical, dynamic, from http://digitalebookden.com/using-instructional-theory-
and virtual (Whittaker et al. 2001). Each kind of orga- to-facilitate-communication-in-web.html
nization shows different degrees of collaboration Schneider, D. (1994). Teaching and learning with internet tools:
between its members (individuals or groups) when A position paper. In Proceedings of the first international confer-
ence on the world-wide web. Geneva: University of Geneva.
accomplishing their tasks (Cattafi and Metzner 2007).
Retrieved April 7, 2010, from http://tecfa.unige.ch/edu-comp/
As in any human endeavor, specific social or task edu-ws94/contrib/schneider/schneide.fm.html
factors can impact collaborative learning, and to be Schunk, D. (1997). Teorı́as del aprendizaje (2nd ed.). México:
useful these factors should be measurable. Language Prentice Hall.
Collective Development and the Learning Paradox C 643
Wagner, E. D. (1997). Interactivity: From agents to Outcomes. something genuinely novel is therefore impossible and
In T. E. Cyrs (Ed.), Teaching and learning at a distance: What that all essential structures must be present at birth. It
it takes to effectively design, deliver and evaluate programs
appeared difficult to counter these objections and con-
(New directions for teaching and learning, Vol. 71, pp. 19–26).
San Francisco: Jossey-Bass. sequently this extreme Nativist position had a devastat-
Warschauer, M. (1997). Computer-mediated collaborative learning: ing impact on the popularity of theories of learning and C
Theory and practice. The Modern Language Journal, 81(4), development (e.g., Piaget’s theory).
470–480. Collective learning refers to a conceptualization of
Whittaker, P., MacKinnon, J., & White, D. (2001). Virtual desire
learning that takes the structures and processes of social
and virtual reality. A case study highlighting the reality of
becoming a virtual organization. In Proceedings 2001 BITWORLD cooperation into account as a “reality sui generis”
conference, Cairo. (Miller 1987). It is an alternative conceptualization
Willis, J. (2006). Creating a working model for technology integration of learning that promises to avoid the learning paradox.
through a lesson planning WebQuest. Electronic Journal for the Collectively accepted knowledge is knowledge that
Integration of Technology in Education, 5, 25–33. Retrieved April cannot be denied and yet is not necessarily completely
1, 2010, from http://ejite.isu.edu/Volume5No1/.
comprehended. It thus creates the possibility of exp-
Wilson, B. G. (1995). Metaphors for instruction: Why we talk about
learning environments. Educational Technology, 35(5), 25–30. eriencing contradictions without reference already to
the subsequent level of understanding. The suggestion
is that if we explain, in this social way, how a group of
peers who seriously try to solve a dispute can under-
stand a disturbance and can learn something genuinely
Collaborative Learning with novel, we are not invoking the contradictions alluded
Emerging Technologies to in the learning paradox in our explanations.
▶ Interactive Learning Environments
Theoretical Background
The novelty problem was articulated by Fodor some
30 years ago. He provided a modern formulation of the
ancient (Plato) learning paradox, making it relevant
Collective Development and to learning and the conception of stage development
the Learning Paradox as entertained by Piaget. He concluded that it is impos-
sible to learn something fundamentally new. Novel
JAN BOOM knowledge cannot be derived completely from old
Department of Developmental Psychology, University knowledge or it would not be new. Yet the new tran-
of Utrecht, Utrecht, The Netherlands scendent element of it cannot be wholly new either,
because then it could never be understood. In Plato’s
“Meno” dialogue, the paradox is brought forward by
Synonyms Meno but the arguments underlying it originated with
Innateness controversy; Novelty; Sociological model of the Sophists. They used the paradox to argue against
learning the view that learning is an activity of learning persons
themselves. They tried to convince their opponents
Definition that learning is completely dependent upon instruc-
The learning paradox refers to a set of arguments tion. For if it were true that learning depended on
that, in the 1980s, questioned the received way of con- asking and searching, learning would not be possible
ceptualizing learning. The core of the arguments was – asking for something means that you already know
that novel knowledge cannot be derived completely what you are looking for, in which case you do not need
from old knowledge, or it would not be new. Yet the to learn it anymore. However, if you do not know it yet,
new transcending part of it cannot be completely you cannot learn it either, because then you cannot
new either, for then it could never be understood. In know what it is that you are looking for. Plato did not
particular, Fodor (1980) maintained that learning agree with the conclusion that learning is exclusively
644 C Collective Development and the Learning Paradox
dependent on instruction. Although he admitted that Notice that all of the three steps or conditions are
some knowledge must be presupposed, he maintained necessary for the paradox to occur. If one is omitted,
that this knowledge could be dormant. The immortal there is no paradox. For example, if condition (c) is not
soul already knows everything before being born; met, it is perfectly possible that something is learned.
learning is a matter of recollection. More recently Fodor admits that complex concepts might be learned
the same controversies have reemerged. Fodor (1980) because they can be represented initially by other (i.e.,
maintains that learning something genuinely novel is primitive) concepts (Fodor 1981, p. 271). Also, relative
impossible and therefore that all essential structures learning is possible because input and output are of
must be present at birth. Fodor is primarily concerned a different order. The initially available cognitive struc-
with the issue of concept learning, which he believes to tures are powerful enough to generate new hypotheses
be a confused notion. He claims that all learning theo- (new content), as long as these hypotheses do not
ries are based on inductive extrapolation, and therefore transcend the boundaries of the present framework.
must acknowledge hypothesis formation and confir- As we will see, Miller’s solution of the Meno paradox
mation among the processes involved in learning. He is also based on a definition of the input as of a different
then shows that given such premises, there can be no order than the output, such that condition (c) does not
such thing as concept learning, or achieving a new stage apply. If there existed a form of learning in which
in development as Piaget would have it. conditions (a) or (b) were not required, the paradox
The line of argument entertained in such learning likewise would not follow. If a test and confirmation are
theories (specifically within the empiricist tradition), not required to learn something truly new, a represen-
and Fodor’s objections to it, can be reconstructed in tation is not necessary prior to the acquisition of the
three steps: (a) First, a subject has to have an idea of new knowledge. In the absence of condition (b), the
what he or she wants to learn. A representation of it emergence of a representation of something novel and
(e.g., a hypothesis specifying a general rule) must be learning something novel are the same phenomenon.
present: the input. (b) Second, the subject should The attainment of a mathematical insight might be an
test whether these ideas conform with experience. For illustration of such a learning step. It may be unclear
example, the hypothesis must be put to test. That is how such a step can occur, but there no longer exists
why it has to be representable in the first place. Fodor’s a paradox! Note, therefore, that Fodor’s (1980) basic
main examples involve concept learning. Testing would argument that it is impossible to represent a richer
amount to verifying whether the concept is used cor- logic in terms of a weaker logic, while being true, is
rectly after the inference of a rule that specifies the not sufficient to allow the conclusion that learning
right use. Correction, in this case, is carried out by a richer logic is impossible. Only if learning is defined
other competent speakers. The predicate learned in such a way that step (b) is indispensable is this
(novel knowledge) is only justified after confirmation conclusion warranted. Fodor is very clear about this
of the hypothesis, so something is learned if and only if because he adds: “if what you mean by learning is
this step has been completed: the output of the learning hypothesis formation and confirmation” (p. 148). Of
process. (c) A problem of circularity will occur in the course, giving up the notion of confirming the new
special case when the input and output are of the same insight is a heavy price to pay to avoid the paradox.
kind. In this case, the learning process presupposes as But it is logically possible that a form of learning exists
input that which is only available as output. Fodor that involves neither hypothesis formation nor confir-
points out an instance of circularity. A problem arises mation. Fodor (1981) does not subscribe to the empir-
when we want to learn a “primitive” concept (concepts icists’ account of learning. Instead, he maintains that all
having no further internal structure and hence not primitive concepts (and fundamental structures) must
representable in terms of other concepts). It is impos- be innate, although he admits that experience plays
sible (by definition) to form a hypothesis about them a role by triggering the concepts. Triggering is consid-
without the use of the primitive concept itself. But since ered to be a causal process and for that reason not
this is exactly the concept that is to be attained, the a form of (constructive) learning. In this way Fodor
paradox follows (Fodor 1981). avoids the paradox.
Collective Development and the Learning Paradox C 645
We now turn to another account of learning, based collective argument are much more restrictive than
on the idea of internalization as advanced by social those governing individual thinking. An isolated indi-
interactionists. Chapman (1992) proposed that joint vidual could easily ignore conflicting information.
activity in which subjects come to share the knowledge However, in a collective argument this is not accept-
that each alone possesses can lead to the construction able as long as the goal – developing a joint argu- C
of new knowledge neither individual possessed before. ment that gives an answer to a disputed question –
Such an idea has been worked out originally and fairly is retained.
elaborated by Miller (1987). Assuming these principles are indeed in operation,
Miller claims that it is possible to experience dis- it is conceivable that one participant in the argument
turbances in a relevant and meaningful way without asserts proposition A while another participant asserts
reference to the subsequent level of understanding. proposition B, with both statements mutually exclusive
However, this is only possible, according to him, by and both traceable to the same shared base of collec-
means of discussion by a group of peers who seriously tively accepted knowledge. Consider the well-known
try to solve a dispute. Miller maintains that cognitive balance scale task. If two or more children address
development can be explained adequately only if this problem, one child may claim that the one arm is
the structures and processes of social cooperation are heavier due to a greater number of weights, while
taken into account as a “reality sui generis” (Miller another child maintains that the other arm is heavier
1987). In collective argumentation – which is the because of the greater distance of the weights from the
model for all argumentation – the primary goal is to fulcrum. Since both children are at a stage in which
develop a joint argument that answers a disputed ques- they acknowledge only one of the variables, they must
tion by relating it to collectively accepted knowledge. in principle be able to understand each other’s reason-
Of greatest relevance is a discussion between peers ing (albeit with difficulty). What they were unable to
sharing the same developmental level. On the basis do is to coordinate both points of view and to see their
of theoretical considerations as well as empirical interconnectedness. The conclusion that Miller draws
research, Miller states that under such circumstances from this example is that a child can no longer simply
a disturbance can be understood and something novel ignore what is going on and is bound to experience
can be learned. He claims that such collective argu- some form of contradiction. At least he or she will be
ments are regulated by a very specific set of rules and made aware of the fact that his or her current knowl-
more specifically three cooperation principles. These edge is not sufficient to reach a consensus (Miller
three basic cooperation principles of argumentation 1986). Collectively accepted knowledge is knowledge
can be in operation (in some form) between very that cannot be denied and yet is not necessarily
young subjects. They function as a coordinating device completely comprehended. It thus creates the possibil-
that determines the processes of argumentation in such ity of experiencing contradictions without reference to
a way that, in principle, a set of collectively valid state- the subsequent stage.
ments can be found and agreed upon. The principle of
generalizability specifies that a statement is justified if Important Scientific Research and
(a) it is either immediately acceptable (belongs to the Open Questions
collectively valid) or (b) if it can be converted to the The learning paradox posed a huge problem for devel-
collectively acceptable. The principle of objectivity opmental psychology and learning theories. An alter-
states that if a statement cannot be denied (i.e., its native to individualistical, psychological theories of
denial cannot be converted into a collectively valid learning was felt to be needed by many. Yet, approaches
statement), it belongs to the realm of the collectively just stressing the social character of learning often
valid, whether it confirms or falsifies the point of view do no more than shift the problem of novelty to the
of some participants. The principle of consistency for- sociocultural plane. In that case, either all novelty is
bids that contradictions enter into – or (once they have denied or novelty remains unaccounted for. For exam-
been discovered) remain in – the realm of the collec- ple, if children learn new ways of thinking from their
tively valid (Miller 1987). These conditions governing parents and their parents have learned them from
646 C Collective Knowledge
their parents, and so forth, we get an infinite regress. Miller, M. (1987). Argumentation and cognition. In M. Hickmann
Miller (1986) offered a far more interesting version of (Ed.), Social and functional approaches to language and thought
(pp. 225–250). New York: Academic.
a Vygotskian social approach based on collective learn-
Molenaar, P. C. M. (1986). On the impossibility of acquiring more
ing principles. powerful structures: A neglected alternative. Human Develop-
However, despite the huge impact on developmen- ment, 29, 245–251.
tal psychology and learning theories the learning
paradox is not a hot topic any more. Interactionism
has become more fashionable, partly due to nonlinear
dynamic systems theory (Molenaar 1986). It is
accepted nowadays that interactions in a dynamic sys- Collective Knowledge
tem may lead to the emergence of new structures and ▶ Collective Learning
sudden reorganizations. That reorganization can take
place quite suddenly and have rather severe conse-
quences is not only possible, but even plausible for
systems as complex as the human mind. Although
Fodor’s strict functionalism has lost much of its cred- Collective Learning
ibility and interactionism is nowadays conceived of as
much broader than just collective argumentation – for THOMAS N. GARAVAN, RONAN CARBERY
example, from the neuronal to the social level – the Department of Personnel and Employment Relations,
fundamental questions involved in the learning para- Kemmy Business School, University of Limerick,
dox should not be ignored because otherwise they will Limerick, Ireland
undoubtedly return in some new form.
Cross-References Synonyms
▶ Bootstrapping: How Not to Learn Collective knowledge; Learning networks
▶ Can Children Learn by Bootstrapping?
▶ Collaborative Learning Definition
▶ Collaborative Learning and Critical Thinking Collective learning is a complex concept that is vari-
▶ Collective Learning ously defined. It is generally conceptualized as a
▶ Conceptual Change dynamic and cumulative process that results in the
▶ Cooperative Learning production of knowledge. Such knowledge is institu-
▶ Piaget, Jean tionalized in the form of structures, rules, routines,
▶ Plato norms, discourse, and strategies that guide future
▶ Social Construction of Learning action. Learning emerges because of interactive mech-
▶ Socio-Constructivist Models of Learning anisms where individual knowledge is shared, dissem-
inated, diffused, and further developed through
relational and belonging synergies. Collective learning
References can therefore be conceived as an evolutionary process
Boom, J. (1991). Collective development and the learning paradox. of perfecting collective knowledge.
Human Development, 34, 273–287.
Chapman, M. (1992). Equilibration and the dialectics of organiza-
tion. In H. Beilin & P. B. Pufall (Eds.), Piaget’s theory: Prospects Theoretical Background
and possibilities (pp. 39–59). Hillsdale: Lawrence Erlbaum. The concept of collective learning draws on a wide
Fodor, J. (1980). Fixation of belief and concept acquisition. In body of theory related to learning, organization the-
M. Piattelli-Palmarini (Ed.), Language and learning: The debate ory, sociology, and psychology. It recognizes the role
between Jean Piaget and Noam Chomsky. Cambridge, MA:
of social interactions in the construction of values
Harvard University Press.
Fodor, J. (1981). The present status of the innateness controversy. and identity. Collective learning may result in a com-
In J. Fodor (Ed.), Representations (pp. 257–316). Brighton: munal language, in which collective approaches and
Harvester Press. knowledge are expressed and cultivated. Garavan and
Collective Learning C 647
McCarthy (2008) highlight a multiplicity of concepts the collective process differentiates collective learning
that fall within the rubric of collective learning, includ- from individual learning.
ing organizational learning; the learning organization; Different types of collective learning are highlighted
team learning; communities of practice; collective in the literature:
knowledge and memory; and collaborative learning.
● Aggregate learning is conceptualized as the aggre-
C
Collective learning, therefore, represents a macro con-
gation of learning gained though trial and error
cept that addresses learning at the levels of the team,
at the individual level. The emphasis is on individ-
the organization, and society. An important distinction
ual learning processes rather than any collective
is made between individual learning and collective
perspective. Aggregate learning may give rise to
learning. Individual learning tends to be conceptual-
fragmentation and individualization rather than
ized as an information system where learning is
inclusion and collectivity.
interpreted, retained, and retrieved by individuals. Col-
● Group learning focuses on the processes that a
lective learning is viewed as a more macro-level concept
group uses to acquire new skills, knowledge, ways
that emphasizes the synergy and advantages of the
of interacting, change patterns between group
collective element.
members, standard operating procedures, and
Collective learning has been defined in a variety
behavioral routines.
of ways. It is possible to view it as an aggregate of
● Institutional learning is conceptualized as learning
individual learning. According to this perspective,
that institutions undergo in order to meet their
collective learning occurs when individuals create,
public brief or mission. It is a form of learning
acquire, and share unique knowledge and informa-
that enables the institution to function effectively;
tion. A second perspective suggests that collective
however, it may lead to institutionalized practices
learning is assumed to occur when a collective
such as hierarchy, paternalism, and authority.
engages in behavior such as asking questions, seeking
● Associational learning is conceptualized by high
feedback, experimenting, reflecting, and discussing
symbolic complexity, but low levels of structural
options and errors. Another view suggests that collec-
openness. It focuses on the coordination and syn-
tive learning is a dynamic process in which learning
thesis of cognitive structures of associated individ-
process and the behavior of the collective change as the
uals and groups. The emphasis in associational
collective learns.
learning is on collective identity.
This third view considers the collective to be an
● Double contingency learning is conceptualized as
open, living system that continuously interacts with
a process of social or discursive construction that
its environment. Many collectives are structured; how-
delineates a field of experiences. It involves the
ever, others are unstructured, yet they take on charac-
erection of boundaries and the exclusion of others.
teristics of complex, living entities. Collectives are
It may result in situations where consensus is
essentially self-organizing through their interactions
expected, disagreements are avoided, and, in some
with the environment. Collectives can be both closed
cases, it leads to a form of fundamentalism.
and open. Some components do not change whereas
● Triple contingency learning is characterized by both
others are transformed.
structural openness and symbolic complexity. This
Central to collective learning is the notion that the
learning occurs due to the emergence of discourses,
collective is enhanced in three ways: (a) it achieves the
cognitive forms, and the capacity to observe, chal-
capacity to restructure and to meet changing condi-
lenge, evaluate, and form opinions. It has strong
tions; (b) it can add and use skills, knowledge, and
self-constituted and self-organized characteristics.
behaviors; and (c) it becomes highly sophisticated in
its capability to deal with feedback and reflect on its Collective learning is considered valuable for indi-
actions. Evolutionary theory defines learning as viduals, organizations, and societies. The outcomes of
a process of cumulative knowledge, taking place in collective learning may be both individual and collec-
firms where common and shared rules exist which tive. Commentators such as Simons and Ruijters
allow individuals to coordinate their action in search (2001) consider collective learning to be collective in
for problem solutions. The social element embedded in the sense of process and outcomes. Their restrictive
648 C Collective Learning
definition has, however, been broadened by other knowledge is central to the competitive success of the
researchers to accommodate individual learning pro- organization and while the existence of linkages and
cess with collective outcomes. emulation is important for this to occur, linkages or
emulation do not simply transfer knowledge directly.
Important Scientific Research and Instead, they are part of the social context in which
Open Questions learning occurs and new knowledge is generated within
From a theoretical point of view, there is much to be the organization. Therefore, particular phenomena
done to further develop the concept of collective learn- such as localized interfirm networks and spin-offs and
ing. Collective learning processes potentially include intra-regional labor mobility become crucial focuses of
a variety of perspectives such as a cognitive and/or attention and indicators of the possibility of innovation
behavioral focus; whether learning is individual learn- and learning.
ing within the collective or genuine collective learning. The conditions that facilitate collective learning are
The factors that facilitate collective learning are not yet largely hypothetical and primarily focused on analogies
fully understood. The role of learning networks, for to individual learning and on experience. Nonaka and
example, are highlighted as important because they Takeuchi (1995) refer to a number of conditions that
provide physical or virtual platforms for human inter- stimulate collective learning, including the presence of
actions (Fu et al. 2006). a vision which directs the processes of knowledge cre-
Camagni (1991) suggests that collective learning is ation, an avoidance of information, and a creative focus
not simply the acquisition of information, and that the which stimulates interaction with the environment.
availability of information is not a central issue. Recent theories on innovation, mainly from cultural–
Instead, it is the process by which available information individual perspectives, focus on supportive conditions
becomes useable knowledge that is the main focus. for collective learning including learning skills, learning
Organizations within the innovative environments motivation, and collective foreknowledge.
seek to cope with the problem of uncertainty by devel- The distinction between individual learning and
oping a “transcoding function” that translates external collective learning requires more detailed investiga-
information into a language that the organization can tion. Both concepts share elements of continuity and
understand. Crucial to this process is the emergence of dynamic synergies; however, they differ in terms of the
a common language and culture that act as precondi- social nature of the latter process. Commentators high-
tions to enable this transcoding to take place. Further light the public dimension of collective learning. The
research may be required to understand the shared mechanisms for transfer of learning focus on ratings
cognitions that facilitate collective learning. The term and behaviors. The operationalization of collective
cognitive consensus has been used in relation to shared learning is, therefore, problematic and researchers are
cognition and collective learning. Cognitive consensus faced with significant problems concerning how best to
refers to the degree of similarity among the mental measure it or identify and label it as a social construct.
models by members of the collective. This consensus
increases over time depending on the level of interaction. Cross-References
The role of trust is also important in terms of the extent ▶ Collective Development and the Learning Paradox
of social interaction and the extent to which a shared ▶ Communities of Practice
cognition will emerge expediently (Capello 1999). ▶ Cooperative Learning
Camagni (1991) distinguishes between “links- ▶ Group Learning
based” and “non-links-based” mechanisms by which ▶ Organizational Learning
this common language or shared cognitions emerges. ▶ The Learning Organization
In the context of organizations and firms, of particular
importance on the links-based side are supply chain
References
linkages or links established via the movement of
Camagni, R. (1991). Local “milieu”, uncertainty and innovation net-
labor between firms. Non-links-based forms of learn- works: towards a new dynamic theory of economic space.
ing include imitation, emulation, and reverse engineer- In R. Camagni (Ed.), Innovation networks: spatial perspectives
ing. This perspective proposes the belief that, while (pp. 121–42). London: Belhaven.
Communication and Learning in the Context of Instructional Design C 649
Capello, R. (1999). Spatial transfer of knowledge in high technology ▶ Discourse and the Production of Knowledge
milieux: learning versus collective learning processes. Regional ▶ Social Interaction Learning Styles
Studies, 34(4), 353–365.
Fu, W., Lo, H., & Drew, D. S. (2006). Collective learning, collective
knowledge and learning networks in construction. Construction
Management and Economics, 24, 1019–1028. C
Garavan, T. N., & McCarthy, A. (2008). Collective learning processes
and human resource development. Advances in Developing Communication and Learning
Human Resources, 10(4), 451–471. in the Context of Instructional
Nonaka, I., & Takeuchi, H. (1995). The knowledge creating company:
how Japanese companies create the dynamics of innovation. New
Design
York: Oxford University Press.
Simons, R. J., & Ruijters, M. (2001). Learning professionals: PHILIP GRISÉ
Towards an integrated model. Paper presented at the bian- The College of Communication and Information
nual conference of the European Association for Research Studies, Florida State University, Tallahassee, FL, USA
on learning and instruction, Aug, 26 – Sept, Fribourg,
Switzerland.
Synonyms
Listening; Reasoning; Thinking
Commitment Definition
Learning may take place through self-discovery, by
▶ Motivational Variables in Learning accident, through observation, by reading, or through
communication with another individual. This descrip-
tion focuses on purposeful teaching/training of an
individual or group, using principles of instructional
Commitment for Learning design. Learning is the purposeful adoption of an
Goals organism’s behavior to its environment based upon
insight gained from encounters with the environment
▶ Volition for Learning directly or communication that provided the insight.
Somewhere between the two fields of communica-
tion and learning lies a blend, an opportunity for the
creative, entertaining stimuli of communication theory
Common Understanding to merge with teaching/training materials to offer to
learners instructional tools that are at once both rigor-
▶ Concept Similarity in Multidisciplinary Learning ous and pleasurable. The result is a measurable out-
come where one knows that learning/training took
place effectively and efficiently, while at the same time
the learner/trainee comes away from the event with
a warm positive feeling, ready to expand ones learning
Communal Learning and tackle even more difficult scenarios.
▶ Group Learning
Theoretical Background
Beginning in the 1960s, elements of what today is
called “instructional design” coalesced from compo-
Communication nents in education, experimental psychology, educa-
tional learning theory, and industrial psychology. Early
▶ DICK Continuum in Organizational Learning work by Robert Gagné and Leslie Briggs laid the
Framework groundwork for others who followed at Florida State
▶ Discourse University. Because these researchers were rooted in
650 C Communication Anxiety
hard scientific backgrounds, and used that perspective Communication and learning needs to pay heed
to develop their teaching/learning paradigms, some of to ongoing research with nonhumans. Current studies
the softer, humanistic aspects of interpersonal interac- with dolphin (Viegas 2010) demonstrate that mammal
tion may have been accidentally slighted. Among the sea creatures have astonishing levels of intellect and
casualties might have been communication and its creativity, and are readily able to understand, prob-
relationship to learning. Some of the fundamental lem-solve, empathize, and otherwise demonstrate
resources that provide extensive background on the “human” characteristics.
interrelationships between communication and instruc-
tional design include works by (Briggs 1979); (Dick and Cross-References
Carey 1990); (Gagné 1985); (Kaufman and Grisé 1995), ▶ Communication Theory
and (Keller 1987). ▶ Discourse
Building on the notion of system thinking (Capra ▶ Discourse and the Production of Knowledge
1996; Wheatley 2006; Senge 2004) a deeper under-
standing of the holistic nature of learning and commu-
References
nication begins to pull together. All pieces are connected
Briggs, L. J. (1979). Handbook of the procedures for the design of
and impact all others. It is this communication between instruction (Monograph #4). Pittsburgh: American Institutes
beings that enables learning to arise. As we intensify our for Research.
study by turning to one another, clarity and wisdom Capra, F. (1996). The Web of Life. New York: Doubleday.
may bring about understanding and perhaps even Dick, W., & Cary, L. (1990). The systematic design of instruction
harmony. (3rd ed.). New York: Harper Collins.
Gagné, R. M. (1985). Conditions of learning. New York: Holt, Rinehart
& Winston.
Important Scientific Research and Kaufman, R., & Grisé, P. J. (1995). Auditing your educational
Open Questions strategic plan. Thousand Oaks: Corwin Press. http://grise.
Leslie Briggs once remarked that the essence of instruc- wordpress.com.
tion (for purposes of the teaching/learning paradigm) Keller, J. M. (1987). Development and use of the ARCS model of
motivational design. Journal of Instructional Development, 10(3),
was to “Tell the learners about what you are going
2–10. http://www.arcsmodel.com/.
to teach them, then teach them, then tell them what Senge, P. (2004). Presence: Human purpose and the field of the future.
you taught them” (personal conversation 1969). This Cambridge, MA: The Society for Organizational Learning.
is a more straightforward way of describing David http://www.solonline.org/PeterSenge/bio/.
Ausubel’s concept of advanced organizers, wherein Viegas, J. (2010, Jan 10). Dolphins: Second smartest animals? (http://
www.ibtimes.co.in/articles/20100108/dolphins-deserve-human-
the learner is stimulated to become aware that he/she
status-say-scientists.htm).
was about to be taught something which would consist Wheatley, M. (2006). Leadership and the New Science: Discovering
of specific elements for a specific purpose. Bob Gagné Order in a Chaotic World. San Francisco, CA: Berrett-Koehler
(1965–1992) also carefully dissected the teaching/ Publishers.
learning process into the necessary condition of learn-
ing to ensure that instruction would take place.
Sadly, all of these works have an element of sterility
to them. The emphasis of experimental design on the
process cannot be mistaken. Conversely, marketing and Communication Anxiety
advertising domains – created to persuade people to act
▶ Apprehension and Communication
in certain predictable ways – do not integrate the rigor
of instructional design but rely more on group dynam-
ics and interpersonal communication. The inclusion of
system thinking with its outcome-based orientation
has done much to move teaching/learning theory Communication Aversion
toward a humane application of principles that are
valuable and functional for learners. ▶ Apprehension and Communication
Communication Theory C 651
encoder with their most accurate translation of what and can exhibit empathy. Following are three of an end-
they perceive was transmitted. This feedback enables less array of examples of creatures other than humans
a series of successive approximations to close in on communicating with one another, engaging higher order
a harmonious understanding of what the message was reasoning processes that move the interchange from
supposed to be. In reality, such feedback looping rarely a pure S-R pattern to a communication scenario.
occurs, leaving less than an ideal communication event.
1. A dog is hit by a vehicle on a busy highway. He is
Additional communication hurdles are presented
crippled. Another dog observes the cries for help
by both internal interference and external interference,
and comes to his assistance. He pulls his “friend”
in every situation. These interference features are
out of harm’s way. See YouTube, December 3, 2008.
sometimes referred to as noise. Internal interference
http://www.youtube.com/watch?v=m2qSakxWt54
arises when the decoder is physically or emotionally
2. Animal psychologist Joyce Poole, from Cornell
distracted. This may arise from a headache, pressing
University, and others are conducting elephant
matters of one’s schedule, concern for an ill family
listening projects in Kenya, Africa, as well as central
member, or other distractions. Additionally, a value
and western Africa. We have learned that elephants
judgment may be made by the decoder as to whether
routinely emit subsonic (to humans) utterances
he/she wishes to believe/accept/appreciate the message
that can be distinguished by other elephants more
being transmitted. This can be because the encoder is
than a mile away. These vocalizations are in addi-
a person of another gender, another race, another cul-
tion to the sonic vibrations that elephants make by
ture, a working subordinate, a lost love, or a host of
stomping the ground and in turn detecting those
other rejections. By external interference we mean
vibrations through their feet at a considerable dis-
confusion caused by the external environment – quite
tance. See CBS Television “60 Minutes, January 3,
literally, noise.
2010,” and National Geographic (2003). http://
news.nationalgeographic.com/news/2003/02/0221_
Important Scientific Research and 030221_elephantvocal1.html
Open Questions 3. A pet, be it a cat, a dog, or another creature, rou-
Communication holds a special place within psychol-
tinely responds to its owner’s requests, not always in
ogy, education, and other social sciences. It is a prac-
the manner desired (the same can be said of chil-
tical application of a variety of disciplines, melded
dren). There is no question on the part of the
together to be a functional device for living organisms.
human “master” that the animal is not simply
Because communication is such an inherent element of
performing an S-R conditioned response pattern
society and culture, its roots and complexity are often
such as Pavlov’s dog salivation experiments
overlooked. Now in the twenty-first century, the defi-
(English publication 1927). The communication
nition demands leaping beyond all-too-frequent per-
that exists is at a much higher level than S-R pat-
spectives such as a process of conveying ideas, and
terns. Pets truly can conduct nonverbal dialog with
thoughts, and feelings between people.
their owners and others. This interplay should be
B.F. Skinner’s operant conditioning methodology
considered a form of communication, as we shall
requires an observable change in behavior caused by
see through the expanded definition.
an event. The stimulus-response (S-R) observation
may be covert and not readily seen – such as change Although many other examples of animal commu-
in blood pressure, galvanic skin response, retinal dila- nication can be cited (e.g., dolphin and primate
tion, and so on. The “skinner box” was often used research), for simplicity’s sake, the definition of com-
within experimental psychology to trigger responses munication shall be described by actions and reactions
in an organism. This is not communication. between people.
Let the loop be closed here so that the definition
expands beyond humans to include other animals for- Functional Definition of
merly referred to as “dumb” animals. Situational obser- Communication
vation and experimentation routinely demonstrate, for It is time to take a new look at the functional defi-
example, that animals possess problem-solving skills, nition of communication. It is time to embrace
Communication Training for Health Professionals C 653
communication features that are very real and take are unobstructedly listening to each other, not
place daily in human to human contact, in human to assuming, second guessing, ignoring, but earnestly
other species contact, and in other species to other working on maintaining a focus between one another.
species contact. Age, gender, culture, attitude, etc.: life gets in the way.
One last mention should be made regarding the Communication requires work. C
functionality of external sounds within a communica- The more one studies communication, the more
tion setting. Often unbeknownst to a film audience, awareness is gained regarding how imprecise and acci-
a movie director is employing long-understood princi- dental communication between two individuals really
ples of the psychological impact film music can have on is. Through improved communication skills, people
an audience. “Film music” refers to more than the can and must do a better job sharing experiences and
melody carried by music, but also the background working in harmony, rather than remaining antago-
effects that are brought to play which might further nistic based upon misunderstandings.
the message of the film. One example is the application From the 1967 Paul Newman film Cool Hand Luke,
of subsonic or very low pulsation softly in the back- the phrase “What we’ve got here, is a failure to
ground to raise the tension level in an audience. communicate” became a symbol for culture clashes.
A rather sublime scene can move viewers to a feeling Communication across racial, religious, cultural, and
of uneasiness, not by what is seen on the screen, or even ethnic lines is essential. Retired US Senator Bob
by the words or melodic track, but by the foreboding Graham, chair of the United States Commission on
rumble that strikes up an innate fear response. Another the Prevention of Weapons of Mass Destruction Pro-
classic example of communication impact within liferation and Terrorism noted that failure of commu-
a movie, without the use of words is music itself. Alfred nication between various US intelligence agencies
Hitchcock’s classic film “Psycho” (1960) rivets the enabled the terror of 9/11 to take place. Had better
audience in fear as violins strike up a screeching noise communication existed between cultures, perhaps the
as the slasher cuts through the shower curtain, killing War on Terror would find itself to be unnecessary.
the bathing Janet Leigh. Proof of the subliminal impact
communicated is easily demonstrated by replaying the
same scene without audio. A much more innocuous
Cross-References
▶ Communication and Learning in the Context of
murder takes place.
Instructional Design
▶ Discourse
To Recap and Add a Touch of
▶ Discourse and the Production of Knowledge
Philosophy
A message that intends to convey meaning is commu-
nicated by an encoder using one or more channels to References
transmit a message. A decoder absorbs the elements Andrews, P. H., & Baird, J. E. (2005). Communication for business and
presented as best as possible, interpreting through the professions (8th ed.). Long Grove: Waveland.
the filters and experiences of his/her own frame of Hamilton, C. (2008). Communicating for results (8th ed.). Belmont:
Wadsworth/Thompson Learning.
reference, paying attention to language, paralanguage,
Shockley-Zalabak, P. S. (2006). Fundamentals of organizational com-
and nonverbal cues. Ideally, the decoder will respond munication (6th ed.). Boston, MA: Pearson Education.
to the encoder with their interpretation of the mean- Wood, J. T. (2004). Communication theories in action (3rd ed.).
ing, seeking validation, or obtaining redirection/ Belmont: Wadsworth/Thompson Learning.
correction. Additionally, within a communication
setting, internal and external stimuli often create dis-
traction/noise.
Woven throughout the definition of the com-
munication process is an essential element – Communication Training for
listening! For all the encoding, decoding, and feed- Health Professionals
back loops to function, success of the communication
depends upon a situation wherein both parties ▶ Empowering Health Learning for the Elderly (EHLE)
654 C Communities of Practice
experiencing high levels of either construct alone (Lee implementation of SLCs found that SLCs were most
and Smith 1999; Wasley et al. 2000; Cotton 2001; Marks often implemented as career or freshmen transition
2000; Shouse 1996). academies and were moderately implemented based
Most research on SLCs, and small schools in gen- on criteria such as: common planning time for
eral, suggests that school size is not the proximate cause teachers, autonomy over program policies and staffing
of the improved student outcomes. Instead, school size decisions, the availability of course offerings related to
is said to facilitate the development of school charac- a given theme, and career related graduation require-
teristics associated with positive student outcomes. ments. There are several factors commonly identified as
Cotton’s 2001 comprehensive analysis of SLCs identi- facilitating SLC implementation and sustainability
fied several characteristics of successful SLCs grouped including: strong school leadership, supportive central
into five categories: administration, high levels of staff buy-in, and suffi-
cient space to make SLCs separate and distinctive.
● Self-determination: The ability to make decisions
However, scheduling and logistical issues, lack of
regarding building usage, scheduling, budget, cur-
physical space to separate SLC programs, and lack of
riculum, instruction, and personnel in order to
qualified staff to accommodate smaller class sizes are
establish a distinct identity.
commonly described as having a negative influence on
● Identity: A common sense of vision and mission
SLC implementation and sustainability.
around enhanced student learning as well as an in-
Research aimed at associating enhanced student
depth focus on a particular theme that distinguishes
outcomes with SLC conversions have focused on atten-
an SLC from the larger building in which it is housed.
dance, graduation rates, student engagement, and aca-
● Personalization: Focus on the social relationships
demic achievement. In his examination of extant
developed among teachers and students, as well
research on the impact and challenges of SLCs, Levin
as substantive efforts to involve parents and the
(2010) examined findings from four separate SLC eval-
community.
uative studies. Three of the four studies examined
● Support for teaching: Includes a variety of strate-
identified significant increases in attendance rates of
gies that enhance the role and decision making
SLC schools contrasted to comparison schools. In the
authority of teachers, including bottom up deci-
studies that had at least one cohort of students reach
sion-making structures, job-embedded ongoing
graduation, there was evidence suggesting SLCs had
professional development, teaching teams, and the
higher graduation rates than comparison schools.
implementation of integrated curricula.
However, there is not yet a critical mass of research
● Functional accountability: Incorporating authentic
examining SLCs with cohorts having reached gradua-
performance assessments that measure what students
tion and, thus, conclusions about whether they impact
can do as well as what they know, in addition to
graduation rates are tentative at best.
standardized tests for gaging student learning. Addi-
Most research at the present cannot provide suffi-
tionally, the use of non-traditional accountability
cient evidence either to support or refute SLCs as
measures such as measures of teacher efficacy, com-
a means for improving academic achievement. Three
mitment, and collective responsibility for student
of the four studies Levin (2010) examined did not
learning that provide time for the SLCs structures
identify statistically significant differences between
and strategies to take effect in measureable ways.
SLC schools and comparison schools on academic
achievement defined by performance on standardized
Important Scientific Research and math and reading tests. However, there was modest,
Open Questions but statistically significant improvement in middle and
Early research on SLCs has focused on the extent to high schools’ achievement in one study where SLCs had
which SLCs are successfully implemented, issues facil- been implemented for 8 years, the longest of any of the
itating and inhibiting implementation of SLCs, as well studies examined.
as on the nature of the various structural methods and The existing research examining the impacts of
strategic reforms SLC schools employ. A study commis- SLCs has several limitations that should be noted.
sioned by the US Department of Education examining First, most studies do not include comparison groups
Community of Learners C 657
or base line data to contrast with SLC outcomes. Addi- SLCs are often situated in low performing schools
tionally, often SLCs are implemented in conjunction using the existing building and employing the same
with other reform efforts, and thus it is difficult to faculties and staff. Thus, school norms, patterns of
isolate the SLC structure as the cause for any improve- relationships, and community expectations of the
ment identified. school are difficult to change. Levin (2010) suggests C
As SLCs mature as a school reform measure, several that stakeholders explicitly discuss the challenges and
challenges to its promise to promote the educational pitfalls of history so that SLCs may be better able to
success of at-risk populations have emerged. Propo- attain a level of change that includes altering patterns of
nents claim that SLCs allow teachers time to collaborate student learning.
on instructional improvements and relationships with Research that demonstrates the impact of SLCs on
students; however, collaboration among teachers has students over longer time frames is needed. Addition-
mostly focused on addressing SLC logistical issues or ally, research examining how SLCs successfully navi-
data analysis of standardized test scores. Teachers spend gate pressures from the division, state, and national
much less time reforming curriculum and instruction levels and sustain curriculum and instructional
in meaningful ways such as team teaching and curric- reform in the face of high stakes testing and account-
ulum integration than on student behavior manage- ability should be undertaken. Finally, research com-
ment. Despite being granted autonomy, SLCs often are paring the effectiveness of the various strategies and
still subject to pressures from the district, state and structures SLCs employ incorporating baseline data
national level regarding testing and accountability or comparison groups should be conducted so that
implications. The result is often a lack of substantive what works about SLCs is more clearly elucidated
change in the approach to teaching and learning; leav- and communicated.
ing SLCs as small versions of the large schools they are
designed to replace. Cross-References
Proponents of SLCs suggest that the structure can ▶ At-Risk Learners
better match individual students’ interests, learning ▶ Interests and Learning
styles and career ambitions. SLCs can tailor the curric- ▶ Learner-Centered Teaching
ulum, instructional approach, and school culture, to ▶ School Climate and Learning
the specific interests of the student population. How- ▶ Small Group Learning
ever, the variety of SLCs, in terms of themes and ▶ Student-Centered Learning
instructional focus, may also pose challenges for ensur- ▶ Workplace Learning
ing consistent levels of rigor across SLCs. When stu-
dents are given a choice of SLCs, their selection is
usually based on, “the extent to which they were willing References
Barker, R., & Gump, P. (1964). Big school, small school: High school size
to let high school make demands on their time and
and student behavior. Stanford: Stanford University Press.
effort” (Ref., p. 121), and thus resulted in stratifying Cotton, K. (2001). New small learning communities: Findings from
high and low performing students based upon SLCs’ recent literature (70 pp.). Portland, OR: Northwest Regional
reputations (Lee and Ready 2007). The allocation of Educational Lab
teaching staff by teacher preference can have a similar Kahne, J. E., Sporte, S. E., de la Torre, M., & Easton, J. Q. (2008). Small
impact. For example in one reorganized SLC, teachers high schools on a larger scale: The impact of school conversions
in Chicago. Educational Evaluation and Policy Analysis, 30(3),
who previously worked in an International Baccalau- 281–315.
reate program all chose to work in the same SLC and Levin, T. (2010). What research tells us about the impact and
thus attracted academically high achieving students challenges of smaller learning communities. Peabody Journal of
while another SLC chosen by coaches and athletic Education, 85, 276–289.
staff attracted student athletes. Lee, V. E., & Ready, D. D. (2007). Schools within schools: Possibilities
and pitfalls of high school reform. New York: Teachers College
Newly formed SLCs must also transcend school
Press.
history. Stand-alone small schools are often founded Lee, V., & Smith, J. B. (1999). Social support and achievement for
in new buildings with new faculties and create new young adolescents in Chicago: The role of school academic press.
norms rather than challenging old ones. However, American Educational Research Journal, 36(4), 907–945.
658 C Comparative Cognition
Marks, H. M. (2000). Student engagement in instructional Romanes, a friend and student of Charles Darwin.
activity: Patterns in the elementary, middle, and high school Flourens’ book title represented the first use of the
years. American Educational Research Journal, 37(1), 153–184.
term, comparative psychology (Psychologie Comparée
Shouse, R. (1996). Academic Press and sense of community: conflict
and congruence in American High Schools. In A. M. Pallas (Ed.), 1864) and predated Romanes’ Animal Intelligence
Research in the sociology of education and socialization. Green- (1882). Both proposed a science that would compare
wich: JAI Press. animal and human behavior, Romanes postulating
U.S. Department of Education, Office of Planning, Evaluation and the existence of a gradient of mental processes and
Policy Development, Policy and Program Studies Service
intelligence from the simplest animals to man – the
(2008) Implementation study of smaller learning communities,
final report. Washington, DC.
comparative approach much in use today. Romanes
Wasley, P. A., Fine, M., Gladden, M., Holland, N. E., King, S. P., strengthened his proposal by a vast collection of anec-
Mosak, E., & Powell, L. C. (2000). Small schools: Great strides. dotal accounts of clever behavior in dozens of animal
A study of new small schools in Chicago. New York: Bank Street species. Though perhaps best known today for the
College of Education. fallacies of his anecdotal method and for his easy
assignment of human mental faculties to animals –
anthropomorphism – Romanes nevertheless succeeded
in establishing his idea of a gradient of mental processes
Comparative Cognition across the animal kingdom as a basic premise of early
comparative psychology. Ethology too has a mixed
▶ A Salience Theory of Learning
parentage. Isidore Geoffroy-Saint-Hillaire first used
▶ Learning in Invertebrates
the term in 1859, though Oskar Heinroth, a late nine-
▶ Linguistic and Cognitive Capacities of Apes
teenth century German biologist, was one of the first to
apply the methods of comparative morphology to ani-
mal behavior; he is thus considered to be one of the
founders of ethology.
Comparative Music Education Both disciplines had many adherents in the early
and middle parts of twentieth century: Comparative
▶ International Perspectives in Music Instruction and
Psychology in the USA under the influence of the
Learning
learning psychologists (e.g., Ivan Pavlov and Edward
Thorndike), the behaviorists (e.g., Zing-Yang Kuo,
John Watson, and B. F. Skinner), and the epigeneticists
(e.g., T. C. Schneirla, Daniel Lehrman, Ethel Tobach,
Comparative Psychology and and Gilbert Gottlieb), while Ethology became firmly
Ethology established after World War II in Europe under the
influence of biologists such as William Thorpe, Nikko
GARY GREENBERG Tinbergen, and Konrad Lorenz. The latter two, in fact,
Department of Psychology, Wichita State University, were awarded the Nobel Prize in medicine (there is no
Wichita, KS, USA separate prize for behavioral research) in 1972 for their
animal behavior studies (they shared this prize with
Karl von Frisch, an early twentieth century biologist).
Definition
Comparative psychology and ethology are both sciences Theoretical Background
which study animal behavior, typically nonhuman Given the biological roots of both comparative psy-
behavior, though both have often studied humans. chology and ethology, evolution was seen to play an
Comparative psychology is a subdiscipline of psychol- important role in behavioral origins by both disci-
ogy and ethology of biology. Both can trace their roots to plines, though in different ways. Comparative psychol-
the late nineteenth century. Depending on which history ogy, strongly influenced by early twentieth century
one reads, the first comparative psychologist was Pierre Functionalists (e.g., William James, John Dewey),
Flourens, a protégé of Baron Cuvier, or George John believed behavior allowed organisms to adapt to their
Comparative Psychology and Ethology C 659
environments (i.e., Darwinism); behavior itself was not it is in biology, though many in both camps understand
an evolved phenomenon, though the organism was. behavior to be a biopsychosocial phenomenon. The
Thus, as organisms changed through evolution, new significance of both psychological and biological devel-
or different behavioral potentials arose. Ethologists, on opment, long ignored, is now seen to be crucial to
the other hand, understood behavior itself to be an a full understanding of behavioral origins. While focus- C
evolved process, the route being genes –> instincts, or ing primarily on issues of comparative psychology,
inherited behaviors. In later years, this one-way route, the many open questions still confronting the study
from genes to behavior, became to be known as the of animal behavior are reviewed in a recent textbook
central dogma of molecular biology. Additionally, (Greenberg and Haraway 2002). For example, though
while comparative psychology tended to engage pri- studied now for well over 100 years, there are still new
marily in laboratory research, ethology emphasized the developments to be found in the area of learning.
significance and importance of studying behavior out-
side the laboratory, in natural settings. Current Status
These two fundamentally different approaches to While comparative psychology grew in America,
the study of behavior lead to a serious intellectual and ethology remained somewhat stagnant in Europe.
theoretical “war” around the 1950s. Ethology advo- Many still identified with the discipline, though it
cated the position that behavior was a biological phe- was clear that they had abandoned the hard-nosed
nomenon, determined, and not merely influenced by biological determinism of the classical ethologists.
the organism’s genotype; much animal behavior was Beginning in 1944 with the initiation of the American
thus believed to be instinctive. Indeed, Lorenz, whose Psychological Association’s divisional structure, com-
mentor was Oskar Heniroth, and Tinbergen spelled out parative psychology had a home in Division 6, Physi-
the full meaning of what instinctive behavior was. The ological Psychology and Comparative Psychology. In
clearest statement of this is found in Tinbergen’s book, the 1990s, in an effort to attract new members, the
The Study of Instinct (1951). Comparative psycholo- division entered into discussion of a name change –
gists, on the other hand, tended to take an epigenetic the important point for the present discussion was the
approach, stressing the importance of development, retention of “comparative psychology” in the new
experience, and other psychological processes. The dif- name adopted at the 1995 APA meeting, Behavioral
ferences were summarized in an important paper by Neuroscience and Comparative Psychology. While
Daniel Lehrman (1953), which today still represents membership in Division 6 was falling, comparative
one of the best critiques of instinct theory. While psychology as a field of study remained healthy as
healthy, the ensuing debates settled little. It was an illustrated by the appearance of several comparative
important 1966 book by Robert Hinde (Animal behav- psychology societies in the closing years of the twenti-
iour: A synthesis of ethology and comparative psychology) eth century: The Southwestern Comparative Psychol-
that seemed to resolve the differences between these ogy Association (founded in 1983 by Michael Domjan,
two opposing views. Indeed, a later 1981 book by the Del Thiessen, Steve Davis, and Gary Greenberg); the
ethologist S. A. Barnett (Modern ethology: The science of Comparative Cognition Society (founded in 1994 by
animal behavior) was able to discuss the discipline Ron Weisman, Mark Bouton, Marcia Spetch, and Ed
without resorting to instinct explanations. Wasserman); and the International Society for Com-
parative Psychology (founded in 1983 by Ethel Tobach
Important Scientific Research and and Gary Greenberg). An even earlier group, the Inter-
Open Questions national Society for Developmental Psychobiology, was
The two disciplines historically sparred over the nature– founded in 1967 by George Collier, Norman Spear,
nurture issue: Was behavior a biological or a psycholog- Bryon Campbell, John Paul Scott, and others. The
ical phenomenon? Endless debates over this issue have annual and biennial meetings of these societies attract
yet to see it formally resolved. Contemporary reports of animal behavior researchers from several disciplines
the discovery of a gene for a behavior are routinely across the globe; their membership is also interna-
retracted following failures to replicate such findings – tional. There are, of course, several other such societies
but the search continues. This is as true in psychology as in countries around the world.
660 C Comparative Psychology and Ethology
The picture was not so rosy for ethology which scientist in 2011 can object to the significance of evo-
seemed to languish in the same period. This was likely lution to psychology?
because, “The simple truth is that ethology never did There has also been new life breathed into ethology
deliver as a science of comparative behavior. . .” and sociobiology. The sociobiological idea of the
(Plotkin 2004, p. 105). Indeed, in 1989 ethology was genetic basis of human altruism has recently been
declared: somewhat retracted by one of its earliest proponents,
E. O. Wilson. While this is comforting news to many
" dead, or at least senescent. That is, if you think of
non-reductionistic comparative psychologists and other
ethology in the narrow sense – the study of animal
animal behaviorists, it does not sit well with all students
behavior as elaborated by Konrad Lorenz, Nikolas Tin-
of behavior (Marshall 2010), attesting to the staying
bergen, and Karl von Frisch. It has been quiescent for
power of the classical ideas of ethology. In a recent
some time. No exciting ideas were emerging, and data
analysis, Salzen (2010) makes a case for interpreting
gathering on key issues had lost its direction. (Barlow
the ideas of ethology in modern neuroscientific terms.
1989, p. 2)
There is in fact a discipline known as “neuroethology,”
However, the biological study of animal behavior has which describes animal behavior in terms of how the
thrived well into the twenty-first century. Ethology was nervous system works. As a comparative psychologist,
reborn in the early 1970s as a new science, that of I take comfort in the staying power of my discipline. Its
sociobiology (Wilson 1975), the goal of which was to history has been long, though not nearly as tumultuous
biologicize the social sciences. But this blatant attempt as that of ethology.
at understanding animal and human behavior as a
purely biological phenomenon was met with scathing Cross-References
criticism (Hull 1988; Lustig et al. 2004) from numerous ▶ Animal Culture
quarters. The main point of contention centered ▶ Biological and Evolutionary Constraints of Learning
around the continuing nature–nurture issue and the ▶ Developmental Cognitive Neuroscience and
question of whether behavior, especially human behav- Learning
ior, was the result of genetic and biological determin- ▶ Evolution of Learning
ism. To many opponents of sociobiology, psychology
was not a biological science at all, but a uniquely psy-
References
Barlow, G. W. (1989). Has sociobiology killed ethology or revital-
chological science (e.g., Greenberg 2007).
ized it? In P. P. G. Bateson & P. H. Klopfer (Eds.), Perspectives
The intellectual sparks flew for years, well into the in ethology (Whither ethology? Vol. 8, pp. 1–45). New York:
end of the twentieth century, which witnessed the Plenum.
appearance of a still new iteration of ethology, evolu- Buss, D. M. (Ed.). (2005). The handbook of evolutionary psychology.
tionary psychology. This approach focuses primarily Hoboken: Wiley.
Greenberg, G. (2007). Why psychology is not a biological science:
on human behavior and posited that we owe our uni-
Gilbert Gottlieb and probabilistic epigenesis. European Journal of
versal nature to evolutionary adaptations faced by our Developmental Science, 1, 111–121.
Pleistocene ancestors that we have inherited in our Greenberg, G., & Haraway, M. M. (2002). Principles of comparative
genomes. A good source for reviewing the tenets and psychology. Boston: Allyn and Bacon.
the research conducted in this field is The Handbook of Hull, D. (1988). Science as a process. Chicago: University of Chicago
evolutionary psychology (Buss 2005). With evolutionary Press.
Lehrman, D. S. (1953). A critique of Konrad Lorenz’s theory of
psychology, instincts are once again in vogue. As with
instinct. Quarterly Review of Biology, 28, 337–363.
ethology and sociobiology, evolutionary psychology is Lickliter, R., & Honeycutt, H. (2003). Developmental dynamics:
not without its critics (e.g., Lickliter and Honeycutt toward a biologically plausible evolutionary psychology. Psycho-
2003). It is not the application of evolution to behavior logical bulletin, 129, 819–835.
that is at question, but the manner in which it is Lustig, A., Richards, R. J., & Ruse, M. (Eds.). (2004). Darwinian
heresies. Cambridge: Cambridge University Press.
understood to apply to behavioral origins. Evolution-
Marshall, M. (2010). Sparks fly over origin of altruism. New Scientist,
ary psychology, though seen by many to be seriously 2780, 8–9.
flawed, is a rather popular orientation in the contem- Plotkin, H. (2004). Evolutionary thought in psychology: A brief history.
porary behavioral sciences. After all, what serious Maiden, MA: Blackwell.
Comparator Hypothesis of Associative Learning C 661
Salzen, E. (2010). Whatever happened to ethology? The case for the reflects the degree to which a cue activates a neural
fixed action pattern in psychology. History and Philosophy of representation of the outcome. However, there are
Psychology, 12.
a number of observations that challenge that simple
Tinbergen, N. (1951). The study of instinct. Oxford: Oxford
University Press. assumption. Most notably, studies of contingency
Wilson, E. O. (1975). Sociobiology: The new synthesis. Cambridge: found that behavioral control by a cue depends not C
Harvard University Press. only on the probability of the cue being followed
by the outcome (p[outcome|cue]), but also on the
probability of the outcome in the absence of the
cue, that is the context (p[outcome|context alone]).
Thus, behavioral control by a cue seemingly reflected
Comparative Psychology of p(outcome|cue) – p(outcome|context alone). Initially,
Learning it was unclear whether the critical context was that of
▶ Evolution of Learning training or test, and whether the computation occurred
after each training trial or at the beginning of each test
trial. But subsequent research determined that the crit-
ical context was that of training and that this compu-
tation occurred at the time of each test trial.
Comparator Hypothesis of Miller and Matzel (1988) used these two findings to
Associative Learning formulate, in associative terminology (as opposed to
conditional probabilities), the original Comparator
RALPH R. MILLER1, JAMES E. WITNAUER2 Hypothesis, which went well beyond contingency the-
1 ory by allowing nontarget cues that were present dur-
State University of New York at Binghamton,
Binghamton, NY, USA ing target training (not only the training context) to
2 serve as the basis of comparison (i.e., as comparator
State University of New York at Brockport,
Brockport, NY, USA stimuli). This provided a new account of cue competi-
tion (e.g., overshadowing and blocking) as well as the
contingency phenomena on which the model was
Synonyms based. Prior accounts of cue competition assumed
Comparator theory; Performance-focused model; that cue competition is caused by a failure to acquire
Response rules; Retrieval-focused model the target cue-outcome association. When the pairings
occurred in the presence of another cue, the most
Definition common account asserted that the two cues competed
The central tenet of the Comparator Hypothesis is that for a limited amount of available associative strength
responding to a cue requires that the cue signal a that could be supported by the outcome. The Compar-
change in reinforcement. That is, given prior cue- ator Hypothesis instead assumes that each cue acquires
outcome pairings, responding to the cue is not an association with the outcome independent of the
a direct function of the strength of the outcome repre- presence of the other cue. The impaired behavioral
sentation activated by the cue. Instead, responding control of the target cue after it is trained in compound
depends on the degree to which the cue predicts an with a nontarget [comparator] cue is a consequence of
increase (or decrease) in the likelihood of the outcome a comparison between the target cue-outcome and
relative to the likelihood of the outcome in the training comparator stimulus-outcome associations; each
context (which might differ from the test context) in serves as the context of learning for the other. However,
the absence of the cue. as testing of the target can occur in the absence of the
comparator stimulus, activation of the comparator-
Theoretical Background outcome association must be mediated by activation
Both early theorizing and prevailing contemporary of the target cue-comparator stimulus association (see
models of learning (e.g., Rescorla and Wagner 1972) Fig. 1). Thus, the Comparator Hypothesis states that
posit that responding to a cue in a Pavlovian situation behavioral control by a target cue is a direct function of
662 C Comparator Hypothesis of Associative Learning
X-US
Presentation Association
Directly activated
of target CS X US representation
at test Link 1
X-Comparator
stimulus Link 2 Comparator Response
association to the CS
Comparator Hypothesis of Associative Learning. Fig. 1 The Original Comparator Hypothesis. Learning by contiguity
(not total error reduction); Responding by modulation, which is the basis of stimulus interaction
the target-outcome association (Link 1) and an inverse Link 2 also explains the context specificity of these
function of the product of the target-comparator stim- effects. The reduced behavioral control by the target
ulus association (Link 2) and the nontarget cue- cue seen as a result of CS-alone presentations (i.e.,
outcome association (Link 3). Critically, this account CS-pretraining exposure, partial reinforcement, and
views cue competition as something that influences extinction) is viewed as a consequence of strengthening
expression rather than acquisition of associations. of Link 2, with the training context serving as the
Hence, the Comparator Hypothesis is centrally a comparator stimulus.
response rule, with acquisition governed by a simple The Comparator Hypothesis not only anticipates
local error reduction rule, that is, a learning mechanism excitatory responding to the target cue when Link 1 is
that reduces the predictive error of each cue separately, strong compared to the product of links 2 and 3. It also
rather than the overall predictive error of all cues pre- anticipates behavior indicative of condition inhibition
sent on a given trial. when Link 1 is weak relative to the product of Links 2
Great flexibility was obtained by allowing compar- and 3. In contrast with traditional associative models,
ator stimuli to be either a punctuate companion cue the Comparator Hypothesis does not posit negatively
or a protracted training context. The Comparator valued associations or associations between cues and
Hypothesis so framed readily accounts for cue compe- no-outcome representations. Rather, all associations
tition effects, all of which depend on strong associa- are positive (i.e., excitatory), and behavior indicative
tions both between the target cue and the competing of conditioned inhibition arises from an interaction
cue (Link 2, with the competing cue serving as the among positive associations. This is a strength of the
comparator stimulus) and between the competing cue Comparator Hypothesis, as it obviates perplexing
and the outcome (Link 3). Additionally, the reduced issues concerning encoding of information that sup-
behavioral control by the target cue seen as a result of ports behavior indicative of conditioned inhibition.
presentations of the outcome alone during target cue The Comparator Hypothesis, unlike prior models
training and the outcome-preexposure effect are con- of learning, avoided using a learning mechanism
sequences of a strong training context-outcome asso- dependent on total error reduction (i.e., a discrepancy
ciation; the requirement that this be the training between the outcome that occurs on a trial and the
context (not the test context) in order to establish expectation of the outcome based on all cues present on
Comparator Hypothesis of Associative Learning C 663
that trial). Because cue competition is viewed as mod- (the association between this third cue and the [first-
ulation of performance, rather than modulation of order] comparator stimulus), and Link 3 (now
acquisition that is governed by total error reduction, Link 3.1) was similarly down modulated by the product
the Comparator Hypothesis was the first model to of Link 3.2 (the association between the [first-order]
account for retrospective revaluation. Retrospective comparator stimulus and a third cue) and Link 3.3 (the C
revaluation (as an empirical phenomenon) refers to association between this third cue and the outcome).
a change in responding to a cue as a consequence of a Although these changes seemingly complicate the
change in the associative status of another cue that was Comparator Hypothesis, they actually simplify it by
previously paired with the target cue. Such demonstra- eliminating the arbitrary assumptions that there can
tions are challenging for many models of learning be only one comparator stimulus and that Link 1
because they assume that a cue must be present for was special in being potentially down modulated,
a change to occur in its behavior control. Most dem- whereas Links 2 and 3 were immune to this process.
onstrations of retrospective revaluation consist of In the Extended Comparator Hypothesis, all stimuli
decreasing the associative status of the target’s compar- and associations are treated equally. The consequence
ator stimulus (i.e., the cue with which the target was of potential higher-order comparator stimuli is that
trained) and observing an increase in responding to the a second-order comparator stimulus can reduce the
target. The best known example of retrospective reval- effectiveness of a first-order comparator stimulus,
uation is recovery from overshadowing. Following just as a first-order comparator stimulus can reduce
overshadowing of a target cue by a nontarget cue (by responding to a target cue. Thus, a post-target training
reinforcing them in compound, which results in change in the associative status of a second-order com-
reduced behavioral control by the target relative to its parator stimulus should produce a change in behav-
being reinforced by itself), extinction of the nontarget ioral control by the target cue in the same direction as
cue increases behavioral control by the target. the second-order comparator. This contrasts with
Although the Comparator Hypothesis provided the changes in the associative status of a first-order com-
first coherent account of retrospective revaluation, parator stimulus, which ordinarily induce a change in
alternative accounts (e.g., Van Hamme and Wasserman behavioral control by the target cue in the opposite
1994) were soon developed that viewed retrospective direction.
revaluation as the consequence of changes in the Stout and Miller (2007) provided a mathematical
value of the target-outcome association during the implementation of the Extended Comparator Hypoth-
retrospective revaluation trial despite the absence of esis. In addition to formalizing the Extended Compar-
the target cue. The Extended Comparator Hypothesis ator Hypothesis, this implementation added a feature.
(Denniston et al. 2001), which elaborated the Compar- Both the original Comparator Hypothesis and its
ator Hypothesis, made predictions that differentiated extension assumed that the product of Links 2 and 3
its approach from that of the new acquisition-focused are always subtracted from Link 1 yielding so-called
models. The changes in the Extended Comparator negative mediation (e.g., cue competition). However,
Hypothesis relative to the original Comparator phenomena like second-order conditioning and sen-
Hypothesis were twofold. First, it allowed multiple sory preconditioning suggest that, under some circum-
comparator stimuli to summate in down modulating stances, the indirect pathway from the target cue to the
responding to a target cue, whereas the original Com- outcome (i.e., Link 2 and Link 3) adds to the direct
parator Hypothesis assumed that only the companion pathway (i.e., Link 1) yielding so-called positive medi-
cue with the strongest association to the target cue ation. The mathematical implementation assumes
would serve as a comparator stimulus. Second, the that the determinant of the type of mediation is
Extended Comparator Hypothesis not only assumed whether the organism has had sufficient opportunity
that Link 1 was down modulated by the product of to discriminate between the directly and indirectly
Links 2 and 3 as in the original Comparator Hypothe- activated representations of the outcome. With few
sis, but that Link 2 (now Link 2.1) was down modu- training trials, the discrimination is difficult, so the two
lated by the product of Link 2.2 (the association outcome representations summate. When there have
between the target cue and a third cue) and Link 2.3 been sufficient trials to facilitate the discrimination,
664 C Comparator Hypothesis of Associative Learning
the product of Links 2 and 3 is subtracted from Link 1 Seemingly, animals are relatively resistant to losing
rather than added. previously acquired behavioral control, a conservative
The Extended Comparator Hypothesis makes pre- evolutionary strategy. To circumvent this problem,
dictions that differentiate it from acquisition-focused studies have been performed in which the target cue
models of retrospective revaluation. Notably, when is not made biologically significant until after infla-
a target cue has two (as opposed to one) comparator tion. In this situation, retrospective revaluation is
stimuli that are themselves associated, each compar- seen to result from both posttraining associative
ator stimulus can act as a first-order comparator for deflation and inflation of comparator stimuli. This
the other comparator, thereby reducing the effect of confirms the basic prediction of the Comparator
the other comparator on the target cue. Thus, two Hypothesis, but does not integrate into the model
treatments, each of which independently decreases proper an account of why posttraining inflation of
responding to a target, collectively can result in more comparator stimuli does not work well in first-order
behavioral control by the target than with either treat- conditioning.
ment alone. For example, a target cue having two The Comparator Hypothesis is designed to explain
blocking cues as comparator stimuli can evoke elemental learning and interactions between cues (and
stronger responding than the same target with only outcomes) trained in compound. However, the Com-
one blocking cue. These so-called counteractions are parator Hypothesis does not explain stimulus interfer-
widely seen. Often the context serves as one of the ence, that is, interactions between stimuli (outcomes or
competing cues. For instance, degraded contingency cues) trained apart.
and overshadowing counteract; that is, context- The Comparator Hypothesis is a trial-wise model
outcome pairings, relative to context-alone trials, that assumes information processing necessary for
when interspersed among target–outcome pairings, responding occurs at the beginning of each test trial
reduce responding to the target, and compound cue and information processing necessary for new learn-
trials reduce responding relative to elemental cue ing occurs at the end of each trial. That is, it is not
trials. But compound cue trials interspersed with a real-time model. Hence, it is unable to account for
context-outcome trials result in stronger responding a number of timing effects.
to the target than either response reducing treatment
alone. Counteraction has also been reported between Cross-References
cue-preexposure and overshadowing, trial massing ▶ Associative Learning
and overshadowing, and long duration cues and ▶ Context Conditioning
overshadowing. Moreover, counteraction has been ▶ Pavlovian Conditioning
reported between treatments that enhance excitatory
behavior control such as second-order conditioning
References
supported by a context as the first-order cue and
Denniston, J. C., Savastano, H. I., & Miller, R. R. (2001). The
second-order conditioning supported by a punctuate extended comparator hypothesis: Learning by contiguity,
stimulus as the first-order cue. Counteraction has responding by relative strength. In R. R. Mowrer & S. B. Klein
also been reported between two inhibitory treatments (Eds.), Handbook of contemporary learning theories (pp. 65–117).
(Pavlovian conditioning inhibition training and differ- Hillsdale: Erlbaum.
ential inhibition training, Urcelay and Miller (2008)). Miller, R. R., & Matzel, L. D. (1988). The comparator hypothesis:
A response rule for the expression of associations. In G. H. Bower
(Ed.), The psychology of learning and motivation (Vol. 22,
Important Scientific Research and pp. 51–92). Orlando: Academic.
Open Questions Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian
The Comparator Hypothesis anticipates changes in conditioning: Variations in the effectiveness of reinforcement
behavioral control of a target cue to occur as a result and nonreinforcement. In A. H. Black & W. F. Prokasy (Eds.),
Classical conditioning II: Current research and theory (pp. 64–99).
of both posttraining associative deflation and inflation
New York: Appleton.
of its comparator stimulus. Deflation is readily seen to Stout, S. C., & Miller, R. R. (2007). Sometimes competing retrieval
increase responding to the target, whereas inflation’s (SOCR): A formalization of the extended comparator hypothe-
decreasing responding to the target is more elusive. sis. Psychological Review, 114, 759–783.
Compartmentalization in Learning C 665
Urcelay, G. P., & Miller, R. R. (2008). Counteraction between two phenomenon that occurs when an individual has
kinds of conditioned inhibition training. Psychonomic Bulletin & two or more different, potentially conflicting conceptual
Review, 15, 103–107.
schemas concerning a particular domain in his/her
Van Hamme, L. J., & Wasserman, E. A. (1994). Cue competition in
causality judgments: The role of nonpresentation of compound cognitive structure. Certain situations trigger one
stimulus elements. Learning and Motivation, 25, 127–151. schema and other situations stimulate another. Com- C
partmentalization is evident, by and large, when a given
situation does not activate the schema that is most
relevant to the specific situation and, instead, activates
another – plausibly less relevant. One way to under-
Comparator Theory stand the notion of conceptual schema is, following
▶ Comparator Hypothesis of Associative Learning Seel in this volume, as representing a particular way
of organization of the generic and abstract knowledge
a person has acquired in the course of numerous
individual experiences with objects, people, situa-
tions, and events.
Comparison Task
Definition 2 (Based on
A comparison task requires the observer to mentally
Representations)
imagine whether two objects could be rotated into con-
This definition of compartmentalization is used more
gruence with each other. The mental rotation can take
extensively in the field of mathematics education,
place in the three dimensions, and the experimenter can
since mathematical concepts are accessed, processed,
include distractor shapes by scrambling the shape in
and transmitted only through semiotic representa-
a different way or creating a mirror image of the item.
tions. Applied to representations, the phenomenon of
The participant must therefore exclude the plausible
compartmentalization reveals the cognitive difficulty
alternatives and detect the correct choice.
that arises from the need to accomplish flexible and
competent conversions back and forth between differ-
ent types of mathematical representations of the
same concept. These cognitive difficulties reveal defi-
Compartmentalization in ciencies in representational flexibility, which indicate
Learning a fragmentary mathematical understanding of the rel-
evant concept.
ATHANASIOS GAGATSIS
Department of Education, University of Cyprus, Definition 3 (Experimental-
Nicosia, Cyprus Operational)
From a statistical perspective and based on the idea that
compartmentalization refers to the splitting up of an
Synonyms idea or concept into (sometimes more or less incon-
Inconsistency; Inflexibility gruent) parts, compartmentalization is the phenome-
non of the establishment of two distinct clusters that
Definitions correspond to different mathematical conceptualiza-
There are three relevant definitions for the term tions or different representations of, or different cog-
compartmentalization depending upon the content in nitive processes related to, the same concept which have
which it is used. a weak statistical relation (correlation, implication,
similarity) between them.
Definition 1 (Based on the Concept of
Conceptual Schema) Theoretical Background
The term “compartmentalization” is used in the sci- Psychologists, cognitive scientists, and philosophers
ences of learning and cognition to designate the have all contributed toward theories on the architecture
666 C Compartmentalization in Learning
of mind. Some key constitutive notions of these theo- There is a possible parallel here between the
ries are modules and modularity of mind. Fodor (1983) modular theory of mind and compartmentalization.
examines the modularity of mind and the extent to Compartmentalization is not indicated only by the
which the nativist thesis and the alleged domain spec- inconsistency of one’s behavior due to the activation
ificity of cognition are relevant to constraints on the of different schemas. Sometimes, in a given situation
architecture of the mind. He argues that the input the schema that is the most relevant to the specific
systems or perceptual modules as well as the system situation is not activated and instead, another plausibly
for processing language are domain-specific, encapsu- less one is. Compartmentalization represents the act of
lated, mandatory, fast, hardwired in the organism, and partitioning an idea or concept into (sometimes more
have a fixed neural architecture. As Fodor states infor- or less incongruent) distinct components and, in an
mational encapsulation is at the heart of modularity. attempt to simplify things, trying to impose thinking
However, Fodor forcibly argues that the brain apart processes that eventually impede attempts to allow
from its input systems is not modular either in struc- these components to connect again. Thus, several
ture or in function since processing in certain domains authors describe compartmentalization as the implicit
is not informationally encapsulated from information or explicit knowledge that is automatically activated in
in other domains. The processes in the higher cognitive everyday life and operates independently of other
centers, that is, the cognitive areas minus the input forms of knowledge. This phenomenon is described
systems, are holistic in the sense that the knowledge as knowledge compartmentalization (Schoenfeld 1986).
stored in the system can affect all sorts of processing. A distinction is made between at least five types of
All beliefs in the system are formed within the back- knowledge compartmentalization that differ with
ground of the total body of knowledge stored in the regard to their effects on further learning and knowl-
brain. In other words, there are no higher cognitive edge application. These are discussed below.
functions that are not affected by cognitive functions
elsewhere in the brain and, thus, there are no The Compartmentalization of Correct
compartmentalized areas of knowledge; the mind is and Incorrect Concepts
not modular and cognition is massively abductive – In this case instruction does not replace the incorrect
abductive inferences are inferences to the best ideas by the correct concepts, but just provides addi-
explanation. Equivalently, there are no higher cognitive tional pieces of knowledge; correct and incorrect
systems whose function relies only on information knowledge coexist. The major deficiency resulting
stored in their proprietary data-bases. Instead, these from this kind of knowledge compartmentalization is
functions depend on information stored everywhere that in situations where only the use of the correct
in the brain. concept enables problem solving, the problem solver
Raftopoulos (2009) claims that even in the case of often depends on the old inadequate misconceptions
the perceptual system the distinction between percep- and not on the scientific concepts he/she has recently
tion and cognition is not as clear cut as Fodor thinks in developed which would be more appropriate (Mandl
so far as locations in the brain that participate in et al. 1993).
perceptual tasks also participate in cognitive tasks,
although they perform different functions in each case. The Compartmentalization of Several
Karmiloff-Smith (1992) has examined the ques- Correct Concepts
tion “Is the initial architecture of the infant mind Different concepts that are closely associated are
modular?” and criticizes Fodor’s ideas. In her model acquired as separate pieces of knowledge and are
of representational redescription (RR) she describes the stored in different compartments. This causes over-
way procedural knowledge is initially represented and simplifications on the application of these knowledge
processed and then modularized and again becomes structures because their complicated interconnec-
explicit and non-modular following her four Phases tions are not reachable. This kind of compartmen-
of Modularization. Karmiloff-Smith’s theory is that talization results in limited understanding and
modularity is the result of ontogenetic and not phylo- oversimplification in knowledge application (Mandl
genetic processes. et al. 1993).
Compartmentalization in Learning C 667
References Cross-References
Elia, I., & Gagatsis, A. (2008). A comparison between the hierarchical ▶ Activity- and Taxonomy-Based Knowledge
clustering of variables, implicative statistical analysis and confir-
Representation
matory factor analysis. In R. Gras, E. Suzuki, F. Guillet, &
F. Spagnolo (Eds.), Statistical implicative analysis: Theory and
applications (pp. 131–163). Heidelberg: Springer-Verlag.
Fodor, J. (1983). The modularity of mind. Cambridge, MA: The MIT
Press.
Karmiloff-Smith, A. (1992). Beyond modularity. Cambridge, MA: Competency-Based Learning
The MIT Press.
Mandl, H., Gruber, H., & Renkl, A. (1993). Misconceptions and SHAHRAM AZIZI GHANBARI
knowledge compartmentalization. In G. Strube & K. F. Wender
Department of University Teaching and E-Learning,
(Eds.), The cognitive psychology of knowledge (pp. 161–176).
Amsterdam: Elsevier.
International University Institute Zittau, Zittau,
Raftopoulos, A. (2009). Perception and cognition: How do psychology Germany
and the neural sciences inform philosophy. Cambridge, MA:
The MIT Press.
Schoenfeld, A. H. (1986). On having and using geometrical knowl- Synonyms
edge. In J. Hiebert (Ed.), Conceptual and procedural knowledge:
Ability-based; Learning; Learning object; Performance;
The case of mathematics (pp. 225–264). Hillsdale, NJ: Erlbaum.
Prescription
Definition
Competence is a skill which is acquired. It is described
Compassion by a certain set of tasks which can be executed if one has
the relevant expertise. This set of tasks can contain
▶ Altruism and Health
subsets of different types of tasks. “Task,” in fact, does
not refer to the colloquial use of the word, but rather
the exact description of a particular action scheme
(e.g., for adding). Competence consists of one or
Competence more degrees of competence which indicate how well
these tasks can be performed and describes a skill with
▶ Learner Characteristics and Online Learning a certain degree of sustainability, i.e., it should – if it is
a characteristic of a person – survive for an extended
period of time.
However, despite broad use of the concepts “com-
Competence Development petence” and “competency” in educational literature,
the terms are not explicitly defined and, therefore, the
▶ Comprehensive Learning question of how to differentiate these terms still
remains topical.
Theoretical Background
Competence-Based Knowledge There is neither a standard method for describing com-
Space Theory petencies nor a universal definition of “competency.”
We (e.g., Schott and Azizi Ghanbari 2008) propose
A mathematical psychological framework for domain a definition of competency which is – in our opinion
and learner knowledge representation that is applied – appropriate for educational research. To this end, we
Competency-Based Learning C 669
turn to Schott’s (1992). “Appropriate” means that the 8. Behavioral aspect (the measure of executing the
competencies to be determined should be as desirable task): The behavioral aspect of a task, its operator,
as possible for the educational system, as well as suffi- can be inferred from the change from the initial to
ciently describable, conveyable, and verifiable. the final state of the task.
If – as in the definition above – “competency” 9. Contextuality: The context of each execution of a C
consists of a set of tasks and competence levels, an task may vary. A distinction is made between an
accurate description of competency is a problem of internal variation of the scheme of change describ-
appropriate task analyses. According to Schott’s work ing a task and an external variation, which is
on task analysis (e.g. Schott 1992), it is useful to split related to the situational circumstances in which
each task into states and changes and to take into the scheme of change is implemented.
account that the concept of “task” includes the follow- 10. Universality: There are no range limitations
ing elements: concerning which operations can describe tasks.
Tasks are not limited only to cognitive tasks.
1. Task name (what the task is called, e.g., “adding,”
“first aid”). The lack of a generally accepted operational defini-
2. Task objective (what the task is, e.g., “applying first tion of competence/competency is generally acknowl-
aid to a health problem,” “summing up several edged. Some authors simply accept this fact and
numbers by means of a rule of calculation”): support a pragmatic approach. Stoof et al. (2002)
The task objective describes the deeper structure label the search for an overarching definition of the
of the task. term an objectivist approach in which the “criterion
3. Task representation (the manner in which a task is for a competence definition is not whether the defini-
represented, e.g., “73 + 25 = 98” or “if you add tion is true but the extent to which the constructed
LXXIII and XXV, you obtain LXXXXVIII”): The definition has proved to be adequate in the context in
task representation describes the surface structure which it is used (i.e., viability)” (p. 347).
of the task. In the literature, many definitions of competence/
4. Basic formal structure as scheme of change: Every competency can be found – almost as may there are
task describes an operation as a change: It can be authors writing on competence-related matters. In
broken down into an initial state (the question), the following, a selection of definitions by various
a final state (the answer/solution), and an opera- authors will be compared (for a detailed discussion
tor which transforms the initial state into the see Kouwenhoven 2003).
final state. For a more detailed description, inter- The basic structure of competence requires the fol-
mediate states may be specified. Given that such lowing distinction:
an operation of change can never be repeated
● The distinction between competence and performance.
in exactly the same way, it is always a scheme
Competence is the ability of a person to carry
of change.
out a certain task (e.g., to have command of the
5. Proficiency (how well a problem is solved): A
German language). Performance is the implemen-
degree of competence can be specified in quanti-
tation of a concrete subset of the task (e.g., to speak
tative terms (e.g., 90% of the solutions are correct)
or to write German). A person’s competence can be
or qualitative terms (e.g., certain facts have to
diagnosed only through his or her performance.
occur).
● The distinction between competence as prescription
6. Degree of resolution (the level of detail of the task
and as ability. Competence as prescription relates to
description): Tasks can be broken down into sub-
a code or directive. Competence as ability describes
tasks or combined to form higher-level, complex
what a particular person can actually do.
tasks: The former increase the degree of resolution
Educational goals describe competencies as pre-
of the description, whereas the latter decrease it.
scription or “prescriptive skills.” Learning controls
7. Content aspect (the subject of the task execution):
describe individual abilities or “personal skills.”
The content aspect of a task may be inferred from
Opponents view the movement toward
the initial state and the final state of the task.
competency-based systems as reductionistic and
670 C Competency-Based Learning
prescriptive, especially in general education areas semantic and rational task analysis is, however, limited,
(Betts and Smith 1998). since they are determined by the requirements of spe-
● The distinction between subject-specific skills and cific subject content.
mental ability. When one regards mental abilities Instructional psychology has a long tradition of
as people’s psychological dispositions to live, act, assigning teaching materials to certain forms of learn-
and behave (e.g., the ability to perceive, remember, ing, which shall not be discussed here in detail. The
or feel), then a professional role is not necessarily a empirical task analysis finally determines what a person
mental ability. Mental abilities are relatively inde- is actually doing when solving the tasks which represent
pendent of the requirements of a particular techni- the competence to solve a given problem (e.g., Schott
cal field. Subject-specific skills, however, are mainly and Azizi Ghanbari 2008, p. 62).
determined by the context and professional recom-
mendations of the technical field in question. Con-
siderations regarding the skills needed in a particular
Important Scientific Research and
field describe prescriptive but not personal skills.
Open Questions
The educational term competency-based learning is not
Thus, a psychological investigation is not sufficient
the result of a fashion of introducing new words and
for the determination of a subject-specific skill or for
concepts, but an objective phenomenon in education
the development of a theoretical competency model.
motivated by social and economic, political, and educa-
tional conditions. First of all, it is professional education’s
The clarification of competence in education using
reaction to changes in social and economic demands and
semantic, rational, psychological, and empirical task
to the innovative processes which have appeared together
analysis.
with the global market economy (Lobanova and Shunin
If one accepts the definition above, which states that
2008). At the international level, work in the field of
each competency is precisely described by specifying
competencies began in 1990 under the aegis of the
a set of tasks and the corresponding degrees of compe-
Organization of Economic Cooperation and Develop-
tence, then the determination of specialist skills is
ment (OECD) with the international interdisciplinary
a problem of task analysis. It is useful to distinguish
program DeSeCo (Definition and Selection of Compe-
four types of task analysis: “semantic,” “rational,” “psy-
tencies: theoretical and conceptual foundations).
chological,” and “empirical” task analysis.
The quality of subject-specific skills taught at all
These four kinds of task analysis for determining
levels of education – secondary, higher, and continuing
competence have different functions and relate to each
– plays a decisive role in establishing individual and
other. As the first step of determining competence, the
national well-being. Good education depends inter alia
semantic task analysis describes what is meant by a
on the quality of teaching itself, i.e., the methods used
specific competency.
to convey the relevant educational material.
The rational task analysis and the psychological task
The two most important quality criteria for practi-
analysis contain requirements of a process which is
cal interventions both in education and in any other
necessary for solving the tasks that describe the skills.
area of application are:
As far as rational task analysis is concerned, these
requirements are provided from the perspective of the 1. “the relevance of the objective of the respective
respective fields. Psychological task analysis refers to intervention, its justification and desirability
additional assumptions of a specific person’s psycho- 2. the efficiency of the intervention. A[n] intervention
logical processes. is effective if:
For reasons of field orientation, rational task anal- a. the desired effect or goal of the measure (i.e.,
ysis precedes psychological task analysis. During a step- output) is achieved in a sufficient manner. In
wise clarification of competence, the results of the education, this is mainly characterized by the
individual task analyses may reveal repercussions on intended learning results; is resource friendly.
the previous task analyses. For example, the result of That is, the desired effect is achieved at a good
an empirical task analysis may result in a correction of cost-benefit ratio without harmful side effects.”
the psychological task analysis. The revision of the (Schott and Azizi Ghanbari 2010) (p. 481).
Competitive Learning C 671
Efficiency is of fundamental importance to the Schott, F., & Ghanbari, S.A. (2010). Zur Theorie und Praxis kompeten-
(successful) teaching of specialized skills. A theoretical zorientierten Lehrens und Lernens Probleme und Lösungsmö-
glichkeiten). Zeitschrift für Report Psychologie (RP). Report
competence model that requires an unrealistic amount
Fachwissenschaftlicher Teil., 474–487.
of time is without practical value. Relevance and effi- Stoof, A., et al. (2002). The boundary approach of competence:
ciency of teaching measures are referred to as “quality A constructivist aid for understanding and using the concept C
of intervention.” of Competence. Human Resource Development Review, 1(3),
The pathways of learning no longer lead automat- 345–365.
ically to traditional institutions of higher education.
Instead, they lead most directly to learning opportuni-
ties in which competencies are defined explicitly and
delivery options are multiple. This new paradigm will
ultimately redefine the roles of faculties, institutions,
Competitive Learning
and accreditation authorities.
PITOYO HARTONO
Although cognitive skills and abilities gained
Department of Mechanics and Information
through traditional higher education programs are
Technology, Chukyo University, Toyota, Aichi, Japan
the decisive results of education, the choice of compe-
tencies can still hardly be reduced to these frameworks
only. This is just one aspect of the difficulty to be con-
sidered. As the theory and practice of hiring procedures
Definition
Competitive learning is a learning mechanism where
for young specialists demonstrates, noncognitive aspects
the components of the learning systems compete for
play an important role, such as practical skills, attitudes,
motivation, value preferences, and ethics, which are not the executions of the learning procedures. As opposed
to the noncompetitive learning algorithms, where in
necessarily achieved and developed in the field of
each learning step all of the components of the learning
formal education. Furthermore, terms like competence,
system take part in the learning procedure, in compet-
competency, key competences, and skills are often used
itive learning algorithm only a part of the components
ambiguously.
that fulfill a predefined criterion win the right to
execute the learning procedure. The competition
Cross-References between the components of the learning system usually
▶ Cognitive Learning
results in the clear division of the training data or
▶ Cognitive Tasks and Learning
underlying dynamics of the learning target among the
▶ Subject of Learning
components.
X
K X
N
E¼ aij jjCi xj jj2 ð1Þ 2.5
i¼1 j¼1
4.2
4
3.8
3.6
C
3.4
3.2
3
2.8
2.6
4.2
3.8
3.6
3.4
3.2
2.8
2.6
2.4
2.2
4 4.5 5 5.5 6 6.5 7 7.5 8
competitive layer of SOM contains Nx Ny neurons most similar prototype input is then designated as the
that are aligned in a two-dimensional grid. The i-th winner according to (5):
neuron in the competitive layer represents a prototype
vector Ci 2 Rd. Similar to K-means algorithm, this wðxðtÞÞ ¼ arg min jjxðtÞ Ci ðtÞjj2 ð5Þ
i
prototype Ci should be the reference for input vectors
x that are similar to it. SOM ensures the topological The winner and the neurons in its neighborhood
correctness of the map by assigning similar prototypes are allowed to modify their prototypes as follows:
to the neighboring neurons in the map.
This topological-correctness is obtained through Ci ðt þ 1Þ ¼ Ci ðtÞ þ ðtÞdist ði; wðxðtÞÞ
ð6Þ
a competitive learning process, in which, when an ðxðtÞ Ci ðtÞÞ
input vector x is presented at time t, the neurons in
the competitive layer compete to be the reference for Here, (t) is a constantly decreasing function and
this input by measuring the distance between their dist(i,w) is the distance between the i-th neuron and
current prototype vectors with this input where. The the winner.
674 C Competitive Learning
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Competitive Layer are connected with light color, while dissimilar pro-
totypes are connected by dark color. From this figure,
we can learn that most of the neighboring neurons
represent similar prototypes and it is also obvious
that the map is roughly divided into two parts by
a string of dark areas, which can be regarded as a kind
of border in the data space.
The simple example shows that the ability to visu-
Input Layer
alize the multivariate data helps us in intuitively under-
standing the structure of the data.
Input Vector
Learning Vector Quantization
Unlike K-means algorithm and SOM that quantize
Competitive Learning. Fig. 5 Structure of SOM
unlabeled data into a specified number of prototype
vectors, learning vector quantization (LVQ) quantizes
To give a clearer understanding on the topology- labeled data into a predefined number of labeled pro-
preserving mapping characteristics, SOM with 10 10 totype vectors. Similar to K-means and SOM, the pro-
neurons is trained with the original 150 points of the totypes are generated through a competitive learning
four-dimensional Iris data (Fisher 1936) which natu- process, but taking the labels of the training data into
rally cannot be visualized on their original data space. account. After the learning process, the collections
The result is shown in Fig. 6, where the hexagons are of the prototype vectors can be used for classifying
the neurons in the map. The gray area of a hexagon is unlabeled vector.
proportional to the number of inputs that refer the The training process in LVQ is started by initially
corresponding neuron as their prototype (also shown setting K prototypes, usually by choosing K vectors
with a number inside the hexagon). Figure 7 explains from the labeled data. In the competitive training pro-
the topological characteristics of this map. In this fig- cess for each presentation of a labeled vector x(t),
ure, neurons are represented as gray hexagons, while a winner prototype Cw(x) is chosen as K-means algo-
the colors of the areas connecting these hexagons rithm in (2). However, in LVQ the labels of the given
indicate the similarities of the prototypes of the vector x(t) and the winning prototype Cw(x) play an
corresponding neurons, in which similar prototypes important part in modifying the prototype as follows:
Competitive Learning C 675
Hits
8
4 1 2 1 3 1 2 1 0 1
7 1 1 1 1 1 2 1 2 0 1 C
6 1 1 2 1 1 1 1 1 5 3
1 1 2 1 2 0 1 0 3 3
5
3 0 1 1 0 5 0 1 1 1
4
1 3 0 2 1 1 1 0 3 3
3
1 2 1 2 2 2 0 3 2 1
2
2 1 2 2 1 4 0 1 2 5
1 1 1 1 2 1 0 1 1 3 2
0 2 3 2 2 0 3 0 1 1 1
−1
0 2 4 6 8 10
−1
0 2 4 6 8 10
CwðxÞ ðtÞ þ ðxðtÞ CwðxÞ ðtÞÞ lðCwðxÞ ðtÞÞ ¼lðxðtÞÞ vector. In this case, when an unlabeled vector is
CwðxÞ ðt þ 1Þ ¼
CwðxÞ ðtÞ ðxðtÞ CwðxÞ ðtÞÞ lðCwðxÞ ðtÞÞ 6¼ lðxðtÞÞ presented, the vector is assigned the label of the most
ð7Þ similar prototype. The competitive learning mecha-
nism of LVQ is improved in LVQ2 and LVQ3 where
It is obvious that with the modification rule in (7), the modifications of the prototype vectors are only
a winner prototype with the same label as the input executed when they are in the vicinity of the borders
vector is pulled toward the vector while a winner between different classes, which will generate better
prototype with different label is repelled away from dividing hyperplane between different classes.
the input vector. The non-winner prototype vectors
remain unmodified. An example of LVQ is given with Important Scientific Research and
a simplified Iris data similar to that of K-means, except Open Questions
that the data are labeled. Originally the Iris data are Here three of the most utilized competitive learning
labeled with three classes; however in this example for algorithms are explained. However, over the recent
simplicity, the labels of data belonging to one class were decades, there are many interesting competitive learn-
kept, while the two other classes were merged and ing mechanisms with various objectives and properties.
labeled as one new class. The distribution of the two- Neural gas (NG) (Martinetz et al. 1993) is a kind of
class data is shown in Figs. 8 and 9, where data belong- self-organizing algorithm similar to SOM; however the
ing to one class are expressed with △s and the data prototype vectors in NG are not bounded in grid
from the other class are shown with ○s. In these figures neighborhoods. In NG, the similarity between neigh-
the prototypes are shown with large △s and ○s. bors is decided using the ranking of the Euclidean
Figures 8 and 9 show the distribution of the prototypes distances between the input vector and the prototype
when their numbers are three and six, respectively. It is vectors. The most significant difference between SOM
clear that the prototypes are well positioned to quantize and NG is that in modifying the prototype vectors
the data. NG minimizes a global cost function which is not
After the termination of the learning process, LVQ available for SOM. Modular network SOM (mnSOM)
can be used for deciding the label of an unlabeled (Tokunaga and Furukawa 2009) was proposed to
4.5
3.5
2.5
2
4 4.5 5 5.5 6 6.5 7 7.5 8
4.5
4
C
3.5
2.5
2
4 4.5 5 5.5 6 6.5 7 7.5 8
Cross-References
▶ Hebbian Learning Complex Communication
▶ Learning in Artificial Neural Networks
▶ Self-organized Learning ▶ Intelligent Communication in Animals
▶ Supervised Learning
▶ Unsupervised Learning
References
Fisher, R. A. (1936). The use of multiple measurements in taxonomics Complex Declarative Learning
problems. Annals of Eugenics, 7, 179–188.
Forgy, E. (1965). Cluster analysis of multivariate data: Efficiency vs. STELLAN OHLSSON
interpretability of classifications. Biometrics, 21, 768–780. Department of Psychology, University of Illinois at
Kohonen, T. (1982). Self-organized formation of topologically cor-
Chicago, Chicago, IL, USA
rect feature maps. Biological Cybernetics, 43, 59–69.
MacQueen, J. (1967). Some methods for classification and analysis of
multivariate observation. In Proc. of the Fifth Berkeley Sympo-
sium, Vol. 1, (pp. 281–297). University of California Press, Los Synonyms
Angeles. Knowledge acquisition
Martinetz, T., Berkovich, S., & Schulten, K. (1993). “Neural-gas”
network for vector quantization and its application to time-
series prediction. IEEE Trans. on Neural Networks, 4(4), 558–569.
Definition
Tokunaga, K., & Furukawa, T. (2009). Modular network SOM. Neural Declarative knowledge is knowledge about what the
Networks, 22(1), 82–90. world is like. Examples include specific facts, e.g., that
678 C Complex Declarative Learning
bananas grow on trees; general principles, e.g., that bound on the size of the declarative knowledge
spring follows winter; and episodic information, e.g., base of an adult must hence be on the order of a
that such-and-such a person was absent (or present) million knowledge units. There are no estimates of
on a particular occasion. In epistemology, the term the upper bound.
“knowledge” is used normatively to refer to assertions There are three main models of LTM. In the propo-
that are, in fact, true, but in the learning sciences, the sitional model, the unit of knowledge is the proposition,
term is used to refer to whatever assertions a person which is approximately the meaning of a declarative
believes to be true. sentence. Propositions are linked by logical relations
Declarative knowledge contrasts with practical (e.g., follows from, instance of ) and form intuitive
knowledge (also known as “competence,” “expertise,” theories (also known as belief systems). In the schema
“know-how,” “procedural knowledge,” and “skill”). model, the unit of declarative knowledge is instead the
Practical knowledge is knowledge about how to per- schema, which consists of slots (also known as “roles”),
form tasks such as tying one’s shoelaces, using an which are linked by semantic relations (e.g., instrument
electronic device, or proving an algebraic theorem. for, recipient of ). For example, a schema for a birthday
Practical knowledge is intrinsically related to goals party has slots for, at least, the person whose birthday
and actions, while declarative knowledge (e.g., the it is, the host, the presents, the cake, and the guests.
Earth is round) is neutral with respect to purpose. To create a memory of a birthday party, the slots are
Practical knowledge is primarily acquired via practice, filled with the details of the particular event. Schemas
while declarative knowledge is primarily acquired via are interconnected because a schema can fill a slot in
observation and discourse. A popular belief holds that another schema. In the network model, every concept
the two types of knowledge follow different forgetting (node) is linked to other concepts, and the links repre-
curves, with declarative knowledge (e.g., the content of sent adjacency in time or space, causal relations, or
a text) fading faster than practical knowledge (e.g., the semantic similarity. The propositional model empha-
skill of riding a bicycle), but this belief is not grounded sizes the organization of declarative knowledge by
in research. topic, the schema concept highlights the importance
It is useful to distinguish between episodic knowl- of abstraction, and the network model captures the
edge, i.e., knowledge of particular events, and concep- interrelatedness of all declarative knowledge. Neither
tual knowledge, i.e., knowledge of concepts, facts, and model explains all relevant phenomena.
principles. Many episodic memories are autobiographi- Learning declarative knowledge involves at least
cal, i.e., they are memories of a person’s own experi- three types of processes. First, the knowledge must be
ences. Memories for events in stories are episodic but acquired. That is, it must be encoded into LTM; meta-
not autobiographical, because the events happened to phorically, it is said to be stored in LTM. The acquisi-
the protagonist of the story. tion process constructs new knowledge units. Second,
the retention of knowledge in LTM is not perfect.
Theoretical Background Subjective experience suggests that knowledge decays
Declarative knowledge resides in long-term memory over time, but there is less evidence for this than
(LTM). There is no way to measure the total capacity for interference among memory units. Third, using
of LTM directly. But the average educated adult in stored knowledge requires retrieval. Metaphorically,
a Western nation has been estimated to know approx- the retrieval process moves information from LTM
imately 50,000 words, and hence approximately that into working memory (WM). The latter holds those
many concepts. Each concept enters into more than knowledge units that are currently attended. There is
one piece of knowledge. Furthermore, estimates of consensus that the storage metaphor, albeit convenient
the number of knowledge units required for expert and widely used, is misleading. It is more accurate to
performance in a cognitive domain fall in the 10,000– think of retrieval as the activation of a subset of LTM.
100,000 range. Competent but not expert perfor- Failure to recall can be due to failure to encode, imper-
mance is likely to require fewer knowledge units, but fect retention, or failure to retrieve. When the learned
an individual is typically competent in more than one knowledge is complex, these processes become com-
domain (cooking, driving, gardening, etc.). The lower plex as well.
Complex Declarative Learning C 679
Important Scientific Research and tends to be better retained than the first two. Each
Open Questions transformation draws upon the reader’s prior knowl-
edge. Differences in prior knowledge between author
Acquisition and reader probably accounts for a significant propor-
A significant proportion of the content of LTM con- tion of failures to learn from text. C
sists of autobiographical information acquired in the The acquisition of declarative knowledge becomes
course of everyday experience. The creation of autobio- even more complex when a discourse directly contra-
graphical memories requires no intentional effort, but dicts the learner’s prior knowledge. This case is studied
subjectively the process appears selective: Some expe- under the label cognitive consistency (also known as
riences are remembered well and others poorly or not “cognitive dissonance”) in social psychology and con-
at all. A popular hypothesis holds that the probability ceptual change in the learning sciences. Resistance
of encoding a particular experience is proportional to to contradictory information is proverbial and easily
how closely the person pays attention. The explanatory observed in public discourse, but it also operates in
power of this principle is limited by the lack of a theory reading and declarative learning generally. A variety of
of degrees of attention. Another popular hypothesis cognitive mechanisms have been proposed to explain
holds that the probability of encoding is proportional resistance. These include doubting the veracity of the
to the emotional quality and intensity of the experi- source, creating exceptions, and introducing new
ence. The research evidence for the latter principle is assumptions (also known as “abductive reasoning”).
mixed: Some studies have found better recall for emo- Evidence from both social psychology and the history
tionally intense events (also known as “flash bulb of science supports the intuition that the degree
memories”), while others have not. The intriguing of resistance is a function of the centrality of the
but implausible hypothesis that all experiences are contradicted belief. Resistance processes might cause
stored in LTM is proposed from time to time but new information to be distorted or misunderstood.
difficult to test. There is no widely accepted theory of how resis-
Conceptual knowledge is typically acquired via dis- tance to contradictory information is overcome. The
course. Knowledge about abstractions (e.g., the gross idea that resistance can be removed by undermining
national product, the square root of 2), the past prior conceptions with anomalies – counterarguments
(e.g., World War II), theoretical entities (e.g., chemical and demonstrations – has not been shown to improve
atoms), and other matters with which we have no the effectiveness of school learning. An alternative
firsthand experience is necessarily learned via some hypothesis holds that a misunderstanding is due to
type of communication (discussing, listening, reading, a misclassification of some phenomenon under the
watching, etc.). wrong ontological category, so successful acquisition
The essence of the acquisition process is to relate the requires an ontological category shift. A related proposal
new information to previously acquired knowledge. is that a phenomenon can be understood differently by
Research on discourse comprehension, specifically, has re-subsuming it under a different intuitive theory. It is
revealed multiple knowledge-based processes, including likely that there are multiple paths to new knowledge,
lexical disambiguation (identifying the intended mean- each involving different processes.
ings for ambiguous words), parsing (identifying the
relations between parts of a sentence), and bridging Retention
inferences that link the sentences in a text into a coherent Everyday experience shows that the longer the time
whole. There is also evidence that the mental represen- since acquisition, the lower the probability of successful
tation of a text undergoes successive transformations in recall. Experimental studies have revealed that forget-
the course of reading: The initial perception of the words ting follows a negatively accelerate curve, i.e., forgetting
and sentences – the surface representation – is transient is rapid immediately after acquisition but the rate of
and rapidly replaced by a representation of the meaning forgetting decreases over time. The mechanism that
of the sentences and their relations to each other – the produces this regularity is not fully understood.
text base – which in turn generates a representation of Complex declarative knowledge is also affected by
the referent of the text – the situation model. The latter processes that alter the content of memory. F. Bartlett
680 C Complex Declarative Learning
proposed in the 1930s that the memory of a complex transfer, i.e., the application of knowledge acquired in
structure like a story undergoes a process of abstrac- one context (e.g., a classroom) in another, possibly
tion and compacting that leaves only the gist, typi- dissimilar context (e.g., everyday life). Cognitive psy-
cally embellished with a few striking details. Later, chologists find less transfer than they expect in labora-
D. Ausubel proposed the related principle that memory tory experiments, and educators lament that students’
for expository text looses in specificity over time and knowledge is “inert,” i.e., not retrieved when needed.
becomes absorbed into its overarching abstraction The possibility of retrieval, given an application con-
(“obliterative subsumption”). Also, there is strong evi- text, depends on how the knowledge was encoded
dence for both proactive and retroactive interference initially (encoding specificity). One way to increase the
between successive acquisition processes. Researchers probability of retrieval is therefore to anticipate the
have found evidence that declarative knowledge is future use of knowledge while it is acquired and encode
affected by repeated efforts to recall and use it. Each recall it accordingly (transfer appropriate processing). When
involves a certain amount of reconstruction to fill gaps in future use cannot be anticipated, transfer can be facil-
the stored information. The reconstructions are them- itated by encoding the information in multiple ways
selves stored and become part of the memory. Future (encoding variability).
retrievals may or may not distinguish between the orig-
inal information and the subsequent reconstructions. Related Areas
The underlying cause of such effects might be lack of Research on the acquisition of declarative knowledge
source monitoring, which causes pieces of information has generated novel instructional techniques. Research
from different sources to be fused in memory. on knowledge distortion has proven useful in the eval-
uation of eyewitness reports and other topics in law
Retrieval and psychology. Social research on prejudice and ste-
Retrieval requires a probe (also known as “cue”) that reotypes is closely related to, but not well integrated
specifies the needed information. The source of the with research on knowledge acquisition in the learning
probe can be a question asked by someone else, a sciences.
deliberate attempt to recall events in the environment
or implicit task demands. The retrieval probe guides Cross-References
the search through LTM. Successful retrieval requires ▶ Abductive Learning
that the cognitive system makes contact between the ▶ Advance Organizers
probe and the sought after knowledge structure. Due to ▶ Analogical Reasoning
the size of the knowledge base and the probabilistic ▶ Ausubel, David P. (1918–2008)
character of the retrieval process, the latter might fail ▶ Belief Formation
to access a piece of knowledge even though it is, in fact, ▶ Categorical Learning
available (i.e., present) in LTM. ▶ Classification Learning
The probability of successful retrieval, given that ▶ Cognitive Dissonance in the Learning Process
the target information has been encoded and retained, ▶ Conceptual Change
is a function of multiple factors. These include the ▶ Discourse and the Production of Knowledge
number of times a knowledge unit has been retrieved ▶ Dogmatism and Learning
in the past and the time since the last retrieval. Coher- ▶ Episodic Learning
ent and interconnected knowledge structures provide ▶ Fact Learning
more support for retrieval than isolated fragments. ▶ Meaningful Verbal Learning
Consequently, individuals with large amounts of well- ▶ Schema-Based Learning
organized knowledge that is used frequently – experts – ▶ Verbal Learning
exhibit superior memory for knowledge that is relevant
Further Reading
to their area of expertise.
Chi, M. T. H., & Ohlsson, S. (2005). Complex declarative learning. In
The greater the similarity between the probe and K. J. Holyoak & R. G. Morrison (Eds.), The Cambridge handbook
the target representation in memory, the higher the of thinking and reasoning (pp. 371–399). Cambridge: Cambridge
probability of retrieval. This creates the problem of University Press.
Complex Learning C 681
Chinn, C. A., & Brewer, W. F. (1993). The role of anomalous data in approach in education reduces complex contents and
knowledge acquisition: A theoretical framework and implica- tasks into simpler elements, until a level where the
tions for science instruction. Review of Educational Research,
distinct elements can be transferred to learners through
63, 1–49.
Dole, J. A., & Sinatra, G. M. (1998). Reconceptualizing change in presentation and/or practice. The elements are thus
the cognitive construction of knowledge. Educational Psycholo- taught as readymade pieces, which correspond to C
gist, 33(2/3), 109–128. specific, single objectives. This approach works well if
there are few interactions between the elements or
associated objectives, but, according to the holistic
perspective, it does not work well if objectives are
Complex Learning interrelated to each other. For such integrative objec-
tives, the whole is more than the sum of its parts.
JEROEN J. G. VAN MERRIËNBOER Holistic approaches basically try to deal with complex-
FHML, Department of Educational Development and ity without losing sight of the relationships between
Research, Maastricht University, Maastricht, elements. They do so by teaching from simple to
The Netherlands complex wholes. Right from the start, learners are
confronted with the most important relationships
between the elements of complex tasks or complex
Synonyms information.
Integrative goals Another characteristic of the atomistic approach in
education is that skills, knowledge, and attitudes are
Definition often taught separately. For example, knowledge is
A common complaint of students is that they experi- taught in lectures, skills are taught in a skills lab, and
ence the curriculum as a disconnected set of topics and attitudes are taught in role plays. This approach makes
courses, with implicit relationships between them and it difficult if not impossible for learners to integrate
unclear relevance to their future profession. This com- objectives from different domains of learning. Charac-
plaint prompted the initial interest in complex learn- teristic of complex learning is that integrative objectives
ing. The term was introduced in the 1990s to refer to are assumed to be rooted in different domains of learn-
forms of learning aimed at ▶ integrative goals (Gagné ing, including the declarative or conceptual domain,
and Merrill 1990). Learning goals that require the inte- the procedural or skills domain (including perceptual
gration of multiple objectives are frequently encoun- and psychomotor skills), and the affective or attitudes
tered when instruction must reach beyond a single domain. It thus refers to the simultaneous occurrence
lesson or course, for example, when professional com- of knowledge construction, skill acquisition, and atti-
petencies or complex skills are taught. Complex learn- tude formation.
ing takes a holistic rather than atomistic perspective on
learning and teaching processes (van Merriënboer Important Scientific Research and
2007). First, complex contents and tasks are not Open Questions
reduced into simpler elements up to a level where the The main research question is how complex learning
single elements can be transferred to learners through could best be evoked and supported. Most educational
presentation and/or practice, but they are taught from theories assume that complex learning occurs in situa-
simple-to-complex wholes in such a way that relation- tions where learning is driven by rich, meaningful
ships between elements are retained. Second, complex tasks, which are typically based on real-life, profes-
contents and tasks are not divided over different sional tasks. Such tasks are called learning tasks (van
domains of learning, but knowledge, skills, and atti- Merriënboer and Kirschner 2007), enterprises (Gagné
tudes are developed simultaneously. and Merrill 1990), scenarios, projects, or problems.
Well-designed learning tasks explicitly aim at integra-
Theoretical Background tive objectives, by forcing learners both to coordinate
The concept of complex learning is rooted in holism different aspects of task performance and to integrate
(van Merriënboer 2007). The traditional atomistic knowledge, skills, and attitudes. Guidance is necessary
682 C Complex Problem Solving
to help learners deal with the complexity of tasks, Reiser, B. J. (2004). Scaffolding complex learning: The mechanisms of
that is, to provide supports that enable them to deal structuring and problematizing student work. Journal of the
Learning Sciences, 13(3), 273–304.
with more complex content and skill demands than
Van Merriënboer, J. J. G. (2007). Alternate models of instructional
they could otherwise handle. Moreover, provided design: Holistic design approaches and complex learning. In
guidance and support should gradually decrease in R. A. Reiser & J. V. Dempsey (Eds.), Trends and issues in instruc-
a process of “scaffolding,” as learners gain more exper- tional design and technology (pp. 72–81). Upper Saddle River:
tise (e.g., Reiser 2004). ▶ Cognitive load theory (van Pearson/Merrill Prentice Hall.
Van Merriënboer, J. J. G., & Kirschner, P. A. (2007). Ten steps to
Merriënboer and Sweller 2005) explicitly studies
complex learning. Mahwah: Lawrence Erlbaum/Taylor & Francis.
methods that might help to reduce the high cognitive Van Merriënboer, J. J. G., & Sweller, J. (2005). Cognitive load theory
load that is imposed by rich learning tasks. Van and complex learning: recent developments and future direc-
Merriënboer et al. (2003), for example, describe on tions. Educational Psychology Review, 17, 147–177.
the basis of ▶ four-component instructional design Van Merriënboer, J. J. G., Kirschner, P. A., & Kester, L. (2003). Taking
methods that might help reduce high cognitive load: the load of a learners’ mind: instructional design for complex
learning. Educational Psychologist, 38(1), 5–13.
(a) simple-to-complex sequencing of classes of equally
difficult whole tasks, (b) working from worked exam-
ples to conventional problems, (c) just-in-time presen-
tation of helpful information, and (d) provision of
part-task practice for routine aspects of tasks. Complex Problem Solving
With regard to learning outcomes, complex learn-
ing explicitly aims at ▶ transfer of learning, that is, the JOACHIM FUNKE
ability to apply what has been learned to unfamiliar Department of Psychology, Heidelberg University,
problems and/or in new situations. The main assump- Heidelberg, Germany
tion is that complex learning yields a highly integra-
ted knowledge base, organized in cognitive schemas,
which facilitates transfer (Gagné and Merrill 1990). Synonyms
On the one hand, particular types of learning tasks Dealing with uncertainty; Dynamic decision making;
(e.g., goal-free problems, worked examples, comple- Problem solving in dynamic microworlds
tion tasks), which are carefully tuned to the current
level of expertise of learners, contribute to the devel- Definition
opment of an integrated knowledge base and subse- Complex problem solving takes place for reducing the
quent transfer performance; on the other hand, barrier between a given start state and an intended goal
▶ variability of practice should ensure that the whole state with the help of cognitive activities and behavior.
set of learning tasks varies on all dimensions on which Start state, intended goal state, and barriers prove
tasks also differ from each other in the real world, complexity, change dynamically over time, and can be
including surface features and structural features, to partially intransparent. In contrast to solving simple
reach transfer (for an overview, see van Merriënboer problems, with complex problems at the beginning of
and Sweller 2005). a problem solution the exact features of the start state,
of the intended goal state, and of the barriers are
Cross-References unknown. Complex problem solving expects the effi-
▶ Cognitive Load Theory cient interaction between the problem-solving person
▶ Four-Component Instructional Design and situational conditions that depend on the task.
▶ Transfer of Learning It demands the use of cognitive, emotional, and
▶ Variability of Practice social resources as well as knowledge (see Frensch and
Funke 1995).
References
Gagné, R. M., & Merrill, M. D. (1990). Integrative goals for instruc- Theoretical Background
tional design. Educational Technology Research and Development, Since 1975 there has been started a new movement in
38(1), 23–30. the psychology of thinking that is engaged in complex
Complex Problem Solving C 683
problems in contrast to simple problems. Essential more reliable assertions. Complexity demands
impulses for this development came from external, from the problem solver a simplification through
shocking events like the oil crisis or the first analyses reduction to the essential.
of the “Club of Rome” at that time, which showed the 2. Connectivity between involved variables. Needless to
constraints of growth and which made humanity- say, it is not the pure number of variables that is C
threatening problem fields visible. Besides that, the decisive for the workload on the problem-solving
dissatisfaction about the nonpredictability of relevant person, but the connectivity between these. Assum-
characteristics like professional, economical, or politi- ing that in a system of 100 variables every variable is
cal success based on classical intelligence tests led to connected to only exactly one other, the connectiv-
a search of alternative measurements for the assessment ity is lower than in a system in which all variables
of the way humans deal with complex situations, are connected to each other. For making mutual
a search for “operative intelligence,” as it was coined dependencies understandable, a model of the con-
by Dietrich Dörner. nectivity is required from the problem solver.
As an alternative, the use of computer-simulated 3. Dynamics of the situation. This feature explains the
scenarios was proposed. Such “microworlds” allow fact that interventions into a complex, networked
experimental research of complex problems under con- system might activate processes whose impact was
trolled conditions (Brehmer and Dörner 1993). For possibly not intended. A unique variant is the own
example, the scenario “Lohhausen” (Dörner 1997) (intern) dynamic (“eigen-dynamics”). It signifies
simulated the events in a fictitious village. The subject that in a lot of cases the problem does not wait for
had to act as the mayor of a small city for simulated the problem-solving person and his/her decisions,
10 years (essentially reduced to nearly 10 h of gaming but the situation changes itself over time. Dynamic
time) and had to care about the well-being of requires from the problem solver the consideration
the community and its financial wealth. For this task, of the factor “time.”
the fictitious mayor could control the events and shape 4. Intransparency concerning the variables involved
the town according to her or his visions. Based on the and concerning the definition of the goal. In an
data from successful and less successful subjects in this intransparent situation, not all required informa-
scenario, interesting hypotheses about the conditions tion about variables and possible goals are given.
of success and failure in dealing with uncertainty and Intransparency requires from the problem solver
complexity have been formulated. the active acquisition of information.
Since that early start of this research program with 5. Polytely. In a complex situation, reaching goals can
“Lohhausen” in the mid-1970s, numerous scenarios be complicated. Usually there is more than one
with varying extent and from different domains (e.g., goal in a complex situation that has to be consid-
economy, ecology, policy, technology) have been ered. Conflicts due to antagonistic goals require
developed and applied in both basic and applied the forming of compromises and the definition
research. In the following sections, I will outline char- of priorities.
acteristics of complex problems, describe tendencies in
Two approaches concerning research with complex
research, illustrate empirical results, and discuss prob-
problems differentiate with respect to procedures and
lems and perspectives of this approach.
to goals:
Characteristics of complex problems considerably
differ from requirements of simple problems. Five ● The experimental approach: “Systematic manip-
features have been differentiated traditionally (Funke ulation of scenarios.” Essential features of this
2003): approach are the experimental manipulation of
the stimuli (the complex systems) and its condition
1. Complexity of the problem situation. Traditionally, of presentation. Particularly the systematic manip-
complexity is defined based on the number of ulation of scenarios (or system features) became
variables in the given system. Surely, this is only a characteristic of this approach: degree of connec-
a first orientation for the estimation of problem tivity, presence or absence of eigen-dynamics, or the
difficulty, but additional characteristics permit degree of time delays show influences on knowledge
684 C Complex Problem Solving
requirements in different cultures. Cultural com- tasks set by the different scenarios. Thereby, one
parison does not mean changing between nations would get from blanket description to precise testi-
or continents, but could happen simply on the level monies. Scenarios have to be analyzed in form and
of “subcultures.” Assessing how variations in con- content. It has to be explained properly what is
text lead to variation in strategies and subjectively measured. C
constructed problem spaces within the process of ● Characteristics of the problem-solving process. Once
problem solving might be an important task of the requirements are known, cognitive processes
future research. within the acting person can be focused in detail.
● Training and the question of domain specificity or Particularly the differentiation between implicit
generalizability. The question of domain specificity and explicit processes and their relation to the dis-
of problem-solving activities is closely related to the tinction between novice and expert problem solving
issue of context sensitivity. In case of research in could be of peculiar interest. Based on this research,
complex problem solving, the question is one of training procedures could be designed. Existing
transfer of knowledge and strategies between spe- dynamic scenarios contributed to this purpose
cific scenarios. It is generally accepted that confron- already because of their differentiation between
tation with different scenarios leads to an extension different forms of knowledge, of strategies, and of
of the realm of experience – however, there are metacognition.
no empirical evidences. The simple repetition ● Heuristics. It seems promising to transfer our
of processing the same scenario leads to learning knowledge about heuristics found in research on
effects, but training itself means more: the acquisi- decision making to the field of complex problem
tion of strategic competences universally applicable. solving. Possibly simple heuristics control the
Finding rules for unpredictable situations could be processing of complex problems, an idea which
the squaring of a circle. Concerning application would be helpful for finding a global theory.
aspects, there is a huge challenge of psychological
research in problem solving. Cross-References
● Missing theory. The major problem of current ▶ Complex Problem Solving
research is the lack of a firm theory about dealing ▶ Learning and Thinking
with complex problems. It is not even clear if there ▶ Problem Solving
is a need for another theory besides a theory for ▶ Simulation and Learning: The Role of Mental
solving simple problems. Indeed a global theory of Models
cognition that describes and explains dealing with ▶ Simulation-Based Learning
all forms of problems is needed. But such a “unified
theory of cognition” (Alan Newell) does not seem
References
Brehmer, B., & Dörner, D. (1993). Experiments with computer-
to appear on the horizon.
simulated microworlds: Escaping both the narrow straits of the
Perspectives. Within the major area called “psychol- laboratory and the deep blue sea of the field study. Computers in
Human Behavior, 9, 171–184.
ogy of thinking and reasoning,” the exploration of
Dörner, D. (1997). The logic of failure. Recognizing and avoiding error
complex problems represents a question that is of in complex situations. New York: Basic Books.
great significance beyond our discipline. Thereby, a Frensch, P. A., & Funke, J. (Eds.). (1995). Complex problem solving:
chance appears to devote psychology on a basis of The European perspective. Hillsdale: Lawrence Erlbaum
verified findings to a field of application within areas Associates.
like politics and business consulting (“give psychology Funke, J. (2003). Problemlösendes Denken. Stuttgart: Kohlhammer.
Osman, M. (2010). Controlling uncertainty: A review of human
a-way”). For this reason, more intensive data pooling
behavior in complex dynamic environments. Psychological Bul-
and the refinement of appropriate theoretical approaches letin, 136, 65–86.
are needed. Interesting developments could be expected Wenke, D., Frensch, P. A., & Funke, J. (2005). Complex problem
in following areas: solving and intelligence: Empirical relation and causal direction.
In R. J. Sternberg & J. E. Pretz (Eds.), Cognition and intelligence:
● Task and requirement analysis. It seems profitable to Identifying the mechanisms of the mind (pp. 160–187). New York:
undergo an analysis of requirements concerning the Cambridge University Press.
686 C Complex Psychology and Learning
CLINT RANDLES1, JOHN KRATUS2, PAMELA BURNARD3 Important Scientific Research and
1
Center for Music Education Research, School of Open Questions
Music, University of South Florida, Tampa, FL, USA Research in the area of music teaching and learning has
2
College of Music, Michigan State University, East been focused primarily on music performance for
Lansing, MI, USA much of the twentieth century. From about 1950
3
Faculty of Education, University of Cambridge, on, small pockets of researchers began the study of
Cambridge, UK composition learning in the context of music educa-
tion. While some countries – most notably England,
Australia, and Finland – have adopted composition as
Synonyms a regular part of the music curriculum, composition as
Creative Thinking in Music; Creativity; Creativity in a major facet of the teaching and learning of music
Music; Music Creation in the United States has been far less common.
The following sections describe the most notable
Definition research efforts.
Composition learning in music education refers to the
result of creative thinking in music that takes shape in What Children Compose
a process of bringing a musical product into existence Pioneering research on children’s original music
by an individual or group of composers. Composition appeared in a series of monographs published in the
learning has specific meaning for the composer. 1940s by Gladys Moorhead and Donald Pond. These
Composition work takes the form of either notated studies examined the musical characteristics of vocal
music and/or audio recording. Composition learning and instrumental music created by children in an
in music education has traditionally held a secondary unstructured setting. Their research found that even
status to performance learning in music education the youngest school-aged children make use of simple
curricula around the world. melodic and rhythmic patterns in their compositions.
Later research by John Kratus looked at the character-
Theoretical Background istics of music composed by children aged 5–13 in
J. Paul Guilford is known for being a pioneer in the a more structured context. He found that between the
study of general creativity. His speech to the American ages of 5 and 11 there is steadily increasing use of
Psychological Association in 1950 marks the beginning melodic development, rhythmic patterns, tonal orga-
of the study of general creativity in the United States. nization, and metric organization, suggesting a greater
Composition Learning in Music Education C 687
one of the missing links to composition’s inclusion as Moorhead, G. E., & Pond, D. (1941-1951/1978). Music of young
a major curricular area within music education. children. Santa Barbara: Pillsbury Foundation for the Advance-
ment of Music Education.
Randles, C. (2010). Creative identity in music teaching and learning
Assessment (Doctoral dissertation, Michigan State University, 2010). Disser-
Researchers have considered how to assess composition tation Abstracts International, __, ____A.
work. Generally speaking, the study of the assessment Swanwick, K., & Tillman, J. (1986). The sequence of musical devel-
of creativity started in the 1950s with J.P. Guilford. opment: A study of children’s composition. British Journal of
Music Education, 3(3), 306–339.
E. Paul Torrance built on this work by devising
Webster, P. (2009). Children as creative thinkers in music: Focus
standardized tests of creativity. Peter Webster then on composition. In S. Hallam et al. (Eds.), The oxford
took the knowledge gained from the work of these handbook of music psychology (pp. 421–428). New York: Oxford
researchers and developed the Measurement of Creative University Press.
Thinking in Music. The MCTM is likely the most
widely used assessment tool for measuring creative
thinking in music.
Regarding composition learning in music educa-
tion specifically, assessment has been a more frequent
Composition of Groups
topic of discussion in the United Kingdom than in the ▶ Cooperative Learning Groups and Streaming
United States. Researchers in England have been work-
ing on sophisticated rubrics to help teachers rate the
composition work of students at all of the key stages.
Researchers such as Teresa Amabile and Maud Hickey
have taken up the task of developing ways of assessing Composition of Learning
student musical compositions. Groups
Cross-References BIEKE DE FRAINE, BARBARA BELFI, JAN VAN DAMME
▶ Cognitive Psychology of Music Learning The Education and Training Research Group,
▶ Developmental Psychology of Music K.U. Leuven, Leuven, Belgium
References
Amabile, T. (1983). The social psychology of creativity. New York: Synonyms
Springer. Average group level; Group configuration; Group
Barrett, M. (2003). Freedoms and constraints: Constructing musical homogeneity; Group heterogeneity; Group mix
worlds through the dialogue of composition. In M. Hickey (Ed.),
Why and how to teach music composition: A new horizon for music
education. Reston: MENC.
Definition
Burnard, P. (2006). The individual and social worlds of children’s The composition of a learning group refers to how
musical creativity. In G. McPherson (Ed.), The child as musician: a group of learners is composed. The learning group
A handbook of musical development. New York: Oxford University can refer to the school (school student body), a class
Press. group (class composition), or to the more flexible
Hickey, M., & Lipscomb, S. (2006). How different is good? How good
grouping of students within a class (within class group-
is different? The assessment of children’s creative musical think-
ing. In I. Deliege & G. A. Wiggins (Eds.), Musical creativity: ing). The term group composition is used in the sciences
Multidisciplinary research in theory and practice. New York: of learning and cognition to refer to the characteristics
Psychology Press. of the group in terms of ability, achievement level,
Kratus, J. (1985). Rhythm, melody, motive, and phrase characteristics gender, ethnicity, age, etc. The group can be described
of original songs by children aged five to thirteen. (Doctoral
from two main perspectives: the average level of the
dissertation, Northwestern University, 1985).
Kratus, J. (1989). A time analysis of the compositional processes used
group and the heterogeneity (mix) of the group. The
by children ages 7 to 11. Journal of Research in Music Education, group-level variables are calculated by aggregating
37, 5–20. the background characteristics of all individuals in the
Composition of Learning Groups C 689
group to an average (level) or a measure of heteroge- average level. This also means that heterogeneous
neity (e.g., standard deviation). When both perspec- groups are generally beneficial for weak students’
tives are combined, three types of groups emerge: achievement and that homogeneous groups are best
homogeneous weak groups, homogeneous strong for strong students. However, heterogeneous grouping
groups, and heterogeneous groups. The gender com- is generally considered as the best grouping practice in C
position of learning groups, for example, has three most cases, since the benefits for weaker students tend
main categories: two types of single-sex groups (all to be larger than the disadvantages for the stronger
boys’ groups, all girls’ groups) and coeducational students. High-ability students tend to do well in
groups. Studies on tracking and ability grouping address either type of group. However, the range of abilities
the effects of grouping and mixing students by ability. within the group should not be too wide, to increase
Studies on multigrade and multiage classes address the productive interaction in cooperative small groups
effects of grouping and mixing students by age. (Wilkinson and Fung 2002).
The group composition effect (sometimes called con- With regard to the students’ academic self-concept,
textual effect) refers to the effect of the group-level the grouping advice is exactly the opposite (see
variable (level or heterogeneity) on learners’ outcomes BFLPE). High-ability students have a higher academic
over and above the effect of the individual-level vari- self-concept in heterogeneous classes; while for low-
able. Group composition influences both academic and ability students, it is better to be grouped in homoge-
nonacademic outcomes. neous classes.
One of the best-known group composition effects is Two main explanations have been put forward with
the big-fish-little-pond effect (BFLPE) (Marsh et al. regard to group composition effects: sociopsychologi-
2001). This effect indicates that equally able students cal processes and instructional processes.
have a lower academic self-concept when placed in The sociopsychological processes of group learning
a group with a higher average achievement level. It is (peer influences) refer to the normative and compara-
beneficial for students’ academic self-concept to be part tive processes in and between groups. The composi-
of a group with a low average achievement level. tion affects the group’s norms about effort and
investment in learning. The group rewards or pun-
Theoretical Background ishes individuals for conformity or deviant behavior.
It has been widely established that the learning out- Classes with an advantaged group composition develop
comes of an individual are not only affected by his/her a pro-academic culture in which academic achievement
individual background characteristics (age, gender, is highly valued, thereby stimulating everyone in the
general ability, socioeconomic status, achievement group to achieve. In disadvantaged groups, nonconfor-
level, etc.), but also by the composition of the group mity with academic objectives and alienation from
in which the learning takes place. school are often rewarded.
Class composition practices originated as an answer The group can also be a comparative reference
to the diversity in students’ instructional needs. In group, constituting a frame of reference against which
homogeneous classes, teachers can better adjust their the student evaluates his/her own accomplishments.
materials, level, and pace of instruction to the needs According to his/her perceived position, the student
and interests of individual students (Hattie 2002). develops feelings of relative deprivation or gratification
However, opponents claim that homogeneous group- that may affect his/her feelings and behavior. This is
ing denies students to learn from peers of other ability, also called the “frog–pond effect”: the student com-
sex, ethnicity, social class, and/or age. Furthermore, pares himself (size of the frog) to his/her fellow
lower tracks often get stigmatized, which leads to students (size of the pond). There is empirical evidence
teachers not wanting to teach lower-ability tracks and for comparative effects on the self-image, but not
lower-track students feeling discouraged (Hattie 2002). on achievement.
The majority of the studies on group composition Students compare themselves not only to the others
effects have found that it is generally beneficial for all in their group, but comparisons are also made across
students’ achievement to be part of a group with a high groups. Interclass comparisons produce labels, and these
690 C Composition Writing
collective labels influence expectations of teachers, peers, variable and the individual variable. Such an interac-
and parents. This process may stigmatize groups with an tion indicates that the group composition might have
unfavorable group composition and activate a self- another effect on different types of students. For exam-
fulfilling prophecy of failure. Through group identifica- ple, girls show higher math achievement in single-sex
tion and assimilation, the labels also affect students’ self- classes, whereas boys show higher achievement in coed-
concept and expectations. ucational classes.
The second main explanation of the group compo-
sition effect is the instructional process. Advantaged Cross-References
groups tend to show less disciplinary problems, more ▶ Ability Grouping (and Effects) on Learning
higher-order questions, a broader curriculum, etc. In
lower-ability groups, there is a more limited academic References
focus and a reduced opportunity to learn. Talented and Hattie, J. A. C. (2002). Class composition and peer effects. Interna-
motivated teachers are often teaching advantaged tional Journal of Educational Research, 37, 449–481.
groups while low-ability classes are assigned to the Marsh, H. W., Köller, O., & Baumert, J. (2001). Reunification of East
least well-prepared teachers. However, these are corre- and West German school systems: longitudinal multilevel
modeling study of the big-fish-little-pond effect on
lations between group composition and instructional
academic self-concept. American Journal of Educational Research,
practices, making it difficult to disentangle composi- 38, 321–350.
tion and instruction effects. They can have separate and Slavin, R. (1990). Achievement effects of ability grouping in second-
joint effects on student outcomes. ary schools: a best evidence synthesis. Review of Educational
Moreover, Wilkinson and Fung (2002) argue that Research, 60, 471–499.
Wilkinson, I. A. G., & Fung, I. Y. Y. (2002). Small-group composition
peer influences interact with instructional processes to
and peer effects. International Journal of Educational Research, 37,
mediate the effects of group composition on students’ 425–447.
learning.
children were asked to listen to them with the expec- Difficult text also could contribute to inaccurate judg-
tation that if they were monitoring their com- ments. Difficult text usually requires a diligent reader
prehension, the inconsistencies would be detected. to reread, and judgments of comprehension may be
Unfortunately, children proved to be quite poor at more strongly influenced by the amount of rereading
monitoring their comprehension, and the inconsis- than by actual comprehension. The amount of text that
tencies went largely undetected. Although Markman’s can be recalled after reading could affect the accuracy of
work ignited a great deal of research in the area of judgments. If readers are unable to recall verbatim
comprehension monitoring that mostly involved the much of what they have read, the assumption might
reading of text, questions about the use of the error- be made that the text was not understood, even though
detection paradigm arose concerning whether the the overall gist of the text was well remembered. Finally,
kinds of reading children used in the research were the kind of text could affect judgments of comprehen-
similar to the kinds of reading in which people nor- sion. Typically, readers view expository text to be more
mally engage, namely, reading considerate text that is difficult to understand than narrative text, and the
largely error free. Failures to detect inconsistencies may accuracy of judgments of comprehension will vary as
not necessarily indicate a failure to monitor compre- a function of perceived difficulty.
hension, but rather, may indicate that the reader is There has been much debate and theorizing over
monitoring for purposes unrelated to the implanted the issue of how metacomprehension judgments
errors (Hacker 1994). are made. People’s retrospective judgments of com-
Glenberg and Epstein (1985) and Maki and Berry prehension made after reading a text and their pro-
(1984) introduced an alternative paradigm. After spective judgments of future performance on a test
reading error-free texts, readers were asked to make about that text are likely tapping into unique but
metamemory judgments about whether they had overlapping psychological processes (Maki et al.
comprehended text material well enough to perform 2005). Understanding those processes is something a
accurately on a criterion task, such as judging infer- comprehensive theory of comprehension monitoring
ences based on the texts or answering questions about or metacomprehension will provide. Such a theory
the text. In subsequent research on calibration moni- has yet to be proposed. Dunlosky et al. (2002) have
toring or metacomprehension, people were typically suggested that an integration of theories of text com-
asked to read a text, make a judgment of comprehen- prehension with theories of metacognitive monitoring
sion of the text, and then asked to make a prediction of may lead to productive research that could contribute
how well they will perform on a criterion task designed to such a comprehensive theory.
to measure comprehension of the text. Most findings in
this literature have corroborated Markman’s findings: Important Scientific Research and
People are typically poor at monitoring their reading Open Questions
comprehension. In spite of the evidence that the accuracy of calibra-
Maki has added significantly to our understanding tion monitoring or metacomprehension is low, there
of metacomprehension. Across 25 studies from her lab, remains optimism that accurate monitoring and effec-
she reported that the mean correlation between ratings tive control of comprehension (i.e., self-regulated
of comprehension and test performance was only 27 comprehension) holds great promise in educational
(Dunlosky and Lipko 2007). Dunlosky, across 36 stud- contexts. This optimism has been fueled, at least in
ies of metacomprehension, also has reported similar part, by evidence from differing approaches showing
low correlations (Dunlosky & Lipko). Maki and asso- that improved accuracy of monitoring does correspond
ciates (2005) identified several factors that could with an increase in learning (e.g., Dunlosky et al. 2005).
account for such poor monitoring of comprehension. Moreover, the ability to exercise self-regulation was
When readers are unfamiliar with the domain being shown to be an important factor, such that “the efficacy
addressed in a text, their judgments of comprehension of monitoring to enhance learning was undermined
may be poorly gaged because the judgments are when the task did not afford self-regulation” (Dunlosky
influenced more by their unfamiliarity with the et al. 2005, p. 9). This evidence highlights the impor-
domain than by their comprehension of the text. tance of self-regulation in educational settings and the
Comprehensive Learning C 693
need to improve students’ ability to monitor and con- Dunlosky, J., Rawson, K. A., & Hacker, D. J. (2002).
trol their comprehension. Metacomprehension of science text: Investigating the levels-of-
disruption hypothesis. In J. Otero, J. León, & A. C. Graesser
New lines of research have shown promise in this
(Eds.), The psychology of science text comprehension (pp. 255–
endeavor. Huff and Nietfield (2009) improved fifth 279). Mahwah: Lawrence Erlbaum Associates.
grade students’ monitoring accuracy by explicitly Dunlosky, J., Hertzog, C., Kennedy, M. R., & Thiede, K. W. (2005). C
teaching comprehension monitoring strategies over The self-monitoring approach for effective learning. Cognitive
a 12-day period. Rawson, Dunlosky, and Thiede Technology, 10, 4–11.
Glenberg, A. M., & Epstein, W. (1985). Calibration of comprehen-
(2000, as cited in Dunlosky and Lipko 2007) doubled
sion. Journal of Experimental Psychology. Learning, Memory, and
metacomprehension accuracy simply by having partic- Cognition, 4, 702–718.
ipants reread passages twice, a strategy also known to Hacker, D. J. (1994). Comprehension monitoring as a writing pro-
improve reading comprehension. Thiede, Dunlosky, cess. In E. C. Butterfield (Ed.), Children’s writing: Toward a
Griffin, and Wiley (2005, as cited in Dunlosky and process theory of the development of skilled writing. Greenwich:
Lipko 2007) also nearly doubled accuracy by asking JAI Press.
Hacker, D. J. (1998). Metacognition: Definitions and empirical foun-
participants to summarize texts after a short delay.
dations. In D. J. Hacker, J. Dunlosky, & A. C. Graesser (Eds.),
They also showed that summarizing texts after reading Metacognition in educational theory and practice (pp. 1–23).
could be reduced to simply generating five key terms Mahwah: Erlbaum.
that captured the essence of a text, and still accuracy Huff, J. D., & Nietfield, J. L. (2009). Using strategy instruction and
was improved. confidence judgments to improve metacognitive monitoring.
Metacognition and Learning, 4, 161–176.
With respect to future research, any setting that
Maki, R. H., & Berry, S. (1984). Metacomprehension of text material.
involves self-regulated comprehension would benefit Journal of Experimental Psychology: Learning, Memory, and Cog-
from new methods that improve the accuracy of com- nition, 10, 663–679.
prehension monitoring. Although individual differ- Maki, R. H., Shields, M., Wheeler, A. E., & Zacchilli, T. L. (2005). Indi-
ences such as verbal ability or test performance have vidual differences in absolute and relative metacomprehension
been examined in some detail (Maki et al. 2005), the accuracy. Journal of Educational Psychology, 97(4), 723–731.
Markman, E. M. (1977). Realizing that you don’t understand:
complex nature of individual differences leaves a
A preliminary investigation. Child Development, 48, 986–992.
great deal of potential factors remaining for explora- Nelson, T. O., & Narens, L. (1990). A theoretical framework and new
tion. In addition, studies in comprehension monitor- findings. The Psychology of Learning and Motivation, 26, 125–141.
ing have focused on the learning of text material,
but there may likely be other fields of education includ-
ing verbal communication that would be relevant to
and benefit from similar research. Finally, due to
the complex nature of comprehension monitoring, Comprehensive Learning
researchers should strive to be clear about the types of
metacognitive judgments that may be included in any KNUD ILLERIS
study of it. Department of Learning, The Danish University
School of Education, Aarhus University, Copenhagen
Cross-References NV, Denmark
▶ Calibration
▶ Metacognition and Learning
▶ Metacognitive Control Synonyms
▶ Reading and Learning Competence development; Everyday learning;
▶ Self-managed Learning Qualification
▶ Self-regulated Learning
Definition
References All normal learning includes the three dimensions of
Dunlosky, J., & Lipko, A. R. (2007). Metacomprehension: A brief content, incentive, and interaction (or the cognitive,
history and how to improve its accuracy. Current Directions in the emotional, and the social) (Illeris 2002, 2007).
Psychology Science, 16, 228–232. However, as the immediate understanding of learning
694 C Comprehensive Learning
is very often narrowly focused on the acquisition of research has estimated that people averagely spend
knowledge and skills and the emotional and social about 20% of energy on mental processes such as
dimensions are more or less neglected, it becomes thinking, remembering, and learning. The strength
important to emphasize that even when these dimen- and nature of this mobilization depend on what is
sions are not considered they are always involved and usually described as motivation, which has to do with
influence both the learning process and the learning emotion, interest, need, inclination, desire, volition,
result. Basically, this is due to the way the human brain duty – or with a general term incentive. A strong incen-
is working (Damasio 1994; Goldberg 2001) and to the tive favors a differentiated and durable learning result,
fact that people are social beings – and all this is what which can be activated in a broad range of different
the concept of comprehensive learning is referring to. situations, whereas a weak incentive will instead lead
to a learning result which is superficial, difficult to
Theoretical Background remember, and only turns up in situations which
The most fundamental understanding of how human strongly resemble or relates to the learning situation.
learning takes place is that all learning involves two very In Fig. 1, the interaction process of learning is
different processes. The one process is the interaction depicted as a vertical double arrow between the indi-
between the learner and his or her environment. In vidual and the environment, and the acquisition pro-
principle, this process is ongoing all the time when cess of learning is depicted at the individual level as
individuals are not asleep. Sometimes it is very vivid, a horizontal double arrow between the elements of
sometimes it is almost fading out. But whenever it content and incentive.
contains something which is new or different in rela- When the two double arrows of Fig. 1 are framed by
tion to what people have already learned, they have a triangle it gives an illustration of the three dimen-
a possibility to learn from it. sions, which are involved in all learning. Furthermore,
However, learning only takes place if people also by adding a circle around the triangle, indicating that
involve themselves in a mental process of acquisition. all learning is situated in and influenced by the envi-
In this process the new information, which learners ronment of a society, Fig. 2 shows the main elements
have perceived from the interaction by their senses, is and structure involved in human learning or what may
related to whatever prior learning learners subjectively be called a model of comprehensive learning.
and often unconsciously find relevant, and through this The claim of this model is that all learning involves
encounter the learning result is developed. Conse- the elements shown and, consequently, that no learning
quently, this result depends on both the nature of the process or learning situation can be fully understood,
new input information and the nature of what is
already developed in the mind, and this is why different
persons learn different things from the same input
information.
Content Incentive
Further, the acquisition process always contains two Acquisition
elements. The one is the learning content. This is, as Individual
mentioned, usually conceived of as knowledge or skills,
but in a comprehensive understanding of learning it
may also be opinions, insights, meanings, attitudes,
Interaction
Society
into an already existing scheme. This is, for example,
what traditional school teaching is generally aiming at
and over the years people learn a tremendous lot of
things in this way. Assimilative learning is not very
energy demanding, the durability of the learning results C
Content Incentive
depends on how often there was a need to use them,
and they can be recalled in all situations when people
are mentally oriented toward the schema(s) to which
they are subjectively related.
Accommodative learning or accommodation is the
other main type of learning which people practice daily,
although certainly not as often as assimilation, because
Interaction it is much more energy demanding. People engage in
accommodative learning in situations in which they
cannot immediately understand or interpret what is
happening, but have a strong incentive to do so. In
such cases, learners have the possibility of breaking
Comprehensive Learning. Fig. 2 The elements of down parts of one or more existing schemas and recon-
comprehensive learning (© Illeris 2007, p. 26) struct them in a way so that the new impulse can be
included. This is typically experienced as something
people come to realize by a sort of break-through,
they suddenly understand a structure or a connec-
analyzed, planned, or in other ways dealt with if all of tion, a light is dawned on us, or the like. So it is by
these elements are not taken into account. accommodative learning that most important new
In addition to this, the theory of comprehensive insights are gained or people take a qualitative step
learning as presented by the Danish learning researcher in some direction. The learning result has precisely the
and theorist Knud Illeris (2002, 2007) expounds that nature of understanding, it will usually be remem-
the acquisition process can take place in four different bered until it may be changed by new accommodative
ways, the four fundamental types of learning. These learning, and it can be recalled in all kinds of relevant
are defined in relation to how the learning input is situations.
connected and incorporated into already developed Finally, transformative learning or transformation
learning schemas or schemata. is the most complex and demanding type of learn-
In the case of cumulative learning, the learning ing, in which several schemes are reconstructed involv-
results from the start of a new schema, that is, there is ing a change in the organization of the self or the
no existing schema to which it can be related. This identity. People engage in this type of learning only
happens frequently in the first years of human life, when they very much need or wish to do so as it is
but after a couple of years only in very few situations strongly personally demanding and often experi-
with the character of rote learning. One example from enced like the overcoming of some kind of crises.
later life could be the learning of a new pin-code (but The learning result becomes part of the self, it remains
even in this case people often try to invent some kind of with persons and can only be obliterated by a new
system, reference, or mnemonic rule, which actually transformation or by being completely irrelevant
implies that they try to relate the code to some already under changed life conditions. Earlier transformative
existing schema). The results of cumulative learning learning was closely related to psychotherapy, but
can be characterized as rigid and they can only be in modern life conditions are for many people
recalled in situations which are subjectively narrowly changing so often and so radically that this type of
related to the learning situation. learning becomes actualized, and today it is often
Assimilative learning or assimilation is the everyday related to schooling and education in youth and
type of learning in which a new element is integrated adulthood.
696 C Compressed Curriculum
formed to think within the principles of a given system, skills for everyday life and the performance of simple
even when the system is defined as free. The rationale of tasks, selects those who will be given further educa-
the dominant groups tends to give shape to a particular tion and training for higher technical and academic
system. These groups presume to have found the best vocations and professions, and determines the collec-
way to live and think and assume their right to enforce tive culture and values education should promote
it. Even high-quality compulsory education in most to enhance the integrity and unity of the nation. The
democratic states would be regarded by some critics basic compulsory curriculum is expected to act as
as mind-control par excellence. a harmonizing tool, both for the society and for
There are two parts to an education system: the the market. In addition to the compulsory curriculum,
connection of its structure and organization to the the secular state is also expected to make provisions
general system of the country, and its curricula. for the right to religion and freedom of conscience of
National education is connected to the demands of its citizens, to allow them to learn a religion through
the production system of the country by the state and is alternative means and freely practice it. Accordingly,
given the philosophy of the regime as its guiding prin- the public and private schools should not interfere with
ciple. The human model the system adheres to should the religious practices of the parochial schools (see for
be suited to these basic requirements (O’Keeffe 2004). an overview: Rotberg 2004).
Equivalent compulsory education is also possible in Obviously, the modern state arrogated to itself the
private schools and parochial schools. The basic struc- power to oversee the education of all of its citizens and
tural and organizational criteria all of these types of the people within its domain according to its interests.
schools need to meet for the compulsory years are the The individuals in charge of the state can keep design-
same: The law determines the rules for establishing ing the educational system and maintain their domi-
schools, organizing the stages of education, determin- nance as long as the circumstances are favorable.
ing the number of required years, the school model, the Primary education is the primary means by which
division of labor among the school types, etc., while the improvements in new generations can be introduced at
program structure and balance, hours per day, and days an early age, thereby benefitting the individual and soci-
per year the students are to be engaged in learning are ety. Traditionally, stages of education were, as much as
determined by the central authorities. The student child pedagogy permits, designed in keeping with the
assessment standards, the economics of vertical and technological levels in a country. Accordingly, as the
horizontal movement of students within the education utilization of technology rose, many countries intro-
system, diploma and certification equivalency, pas- duced compulsory education through at least the pri-
sages to life, and acceptance in the labor market are mary stage, often extending to the lower-middle level
all regulated. For these reasons, the system is criticized and some to the level of secondary education. Yet, it is
by some thinkers for supporting an economics of com- pivotal to understand that the content and length of
pulsion (Kanpol 1997). compulsory education are directly related to the demand
In compulsory education every nation has a unique for employment in an economy’s markets, which is in
agenda which makes its curricula national and serves to turn related to the development level of the country.
project the functional content for the formation of the Pushing for longer durations of compulsory education
mind, mold the citizen, and train the work force. Again, and starting vocational training at later ages creates
the soundness of the curricula must be approved; the unfair competition in economies that are not ready for
education and certification of teachers and the quality it. Compulsory schooling is not an area where the deci-
of teaching must be regulated according to the stan- sion of extending the length of education can be made by
dards set by the central authorities of the state. looking at others. The length of compulsory schooling
The compulsory curriculum should have certain must correspond to the demand for employment in the
characteristics: The state sets the official language of production sector. The system is meant to provide pas-
teaching, mandates civics courses for various age sages to life and offer further stages of education for
groups for the entire population of young children, those whose work is needed at a higher level.
offers courses in mathematics, science, and technology However, the universalization of modern education
appropriate for different age groups, includes necessary did not take place as it did in the Western world, even
Compulsory Education and Learning C 699
though the promoters were Western. A few of the early and the government are exerting more pressure to
nation-states that extended their realm of activity increase the duration of compulsory formal education
to the colonies carried over the structure and organi- at both ends of the scale to include the kindergarten
zation of their education systems and schools but never years and secondary school and are vying to encompass
the content of their curricula. Colonial regimes could the whole globe with their ideology. C
not provide compulsory education to all of the young
natives but rather only educated a select few who would Important Scientific Research and
serve the masters with the colonizers’ language and Open Questions
their ways of handling work. If missionaries were able Four milieus of research are vital for functional solu-
to reach those remote places, the rest of the young tions for compulsory education and learning (Chappell
natives were left to their hands, who were preaching 2010):
proselytism. Another dilemma of the modern secular Firstly, decisions on the duration of compulsory
nation-state is that it found itself supporting religious education are generally propagated by international
indoctrination in the colonies. In the face of the impos- organizations, which rely on evidence from the level
ing superiority of the colonizer, compulsory national of technology and the market needs of the developed
education never came to the foreground in most coun- countries. This is evident in the ILO and UNESCO
tries until the second half of the twentieth century definitions of the child, UN and WB indexes, OECD
(Lauder et al. 2006). statistics, and WTO decisions, which end up making
In the course of its development, compulsory edu- the same recommendations to most countries. The
cation started under the authority of the state as the question of the duration of formal and compulsory
duty of the citizen and treated the child as a creature education has to be studied in its pedagogical, social,
of the state. As democratic thought took root in the and economic aspects and in relation to the individual
independent countries, education came to be per- needs of both developed and developing countries
ceived as the right of the individual and the duty of (Lauder et al. 2006).
the state. Compulsory education at the primary level Secondly, although there is research, the problem of
was affirmed as a human right by the 1948 Universal inclusion with regard to geographical, physical, social,
Declaration of Human Rights. Although they are still economic, and psychological differences will need to
far from being fully realized, the principles of social be solved at the compulsory primary education level,
justice and equal opportunity in education took prior- which is becoming more complicated as mobility,
ity in the governments’ policies. In the interest of a alienation, and bilingualism increase.
humanization of education, steps were taken to make Thirdly, there is abundant research on how to bring
physical, mental, social, and economic provisions for flexibility to compulsory learning and teacher prepara-
groups with special needs, and the demand for educa- tion for this purpose, and much of it is being put to use.
tional choices brought the individual into focus. Examples include individual learning strategies, the
In brief, in the developed countries the compulsory multiple-intelligence approach, constructivist learning,
national education and learning policies of the nation- student-based learning, minimal invasive education,
state have served their purposes. Production increased, freedom from unnecessary guidance, self-organizing
wealth accumulated, and their citizens enjoyed welfare group learning, and so on. Future research will need
and democracy. But this was not necessarily the case for to emphasize the education of the multi-dimensional
many countries. individual, take into account the free economic and
With the advent of a new era of globalization during political awareness needs of the child, and decrease
the last quarter of the twentieth century, the content the dose of ideology from a powerful center (Kanpol
and style of education came under debate, because 1997).
education now had to take into account new needs Lastly, the libertarians’ choice of educating their
of continuity and change. The liberal voices in the children on their own has attracted considerable
Western societies demand actual freedom for the family research activity (O’Keeffe 2004). Home schooling
and young individuals and a loosening of compulsory and school voucher systems are being applied, both
education. On the other hand, the market economy of which have problematic aspects for the parents.
700 C Computational Emotions
Cross-References
▶ Aligning the Curriculum to Promote Learning Computational Models of
▶ Curriculum and Learning Classical Conditioning
▶ Formal Learning
▶ Twenty-First Century Skills NESTOR A. SCHMAJUK
Department of Psychology and Neuroscience,
References Duke University, Durham, NC, USA
Chappell, C. (2010). Changing pedagogy: Contemporary vocational
learning. Adelaide: National Centre for Vocational Education
Research.
Green, A. (1990). Education and state formation: The rise of education
Synonyms
systems in England, France and the USA. London: McMillan. Associative learning; Mathematical models; Neural
Kanpol, B. (1997). Issues and trends in critical pedagogy. New Jersey: networks; Pavlovian conditioning
Hampton Press.
Lauder, H., Brown, P., Dillabough, J., & Halsey, A. H. (2006). Educa-
tion, globalization and social change. London: Oxford University
Definition
Press. Computational models of classical conditioning are
O’Keeffe, D. (2004). Compulsory education: An oxymoron of moder- mathematical models – including neural network
nity. London: Libertarian alliance. Retrieved September 17, 2010, models – that describe associative learning in terms of
from http://www.libertarian.co.uk/lapubs/educn/educn036.htm. the computation of different intervening variables, such
Rotberg, I. C. (Ed.). (2004). Balancing change and tradition in global
as attention, associations, predictions, and responses.
educational reform. Lanham: Scarecrow Education.
Most times, the models require the use of computer
simulations because they are formulated as nonlinear
systems for which analytical solutions are unknown
or difficult to obtain. The models can reproduce and
Computational Emotions predict experimental results under different condi-
tions. Explanations for the observed behaviors can be
▶ Emotion-Based Machine Learning derived from the observation of the model variables in
a given simulated experiment.
Theoretical Background
During classical (or Pavlovian) ▶ conditioning,
Computational Intelligence (CI) humans and animals change their behavior as a result
▶ Mathematical Models/Theories of Learning (TL) of their experience with different possible relationships
between the conditioned stimulus (CS) and the uncon-
ditioned stimulus (US). Although apparently simple,
many models were proposed to account for the numer-
ous experimental results – described at the end of this
Computational Learning entry – regarding classical conditioning. Here, we intro-
Theory duce some of the most significant models, briefly explain
their mechanisms, and show how they address some
▶ Mathematical Linguistics and Learning Theory important experimental results.
Computational Models of Classical Conditioning C 701
Competition to Gain Associations representations of the CS, the US, the interstimulus
interval (ISI) and the intertrial interval (ITI), direct
Competition Between CSs to Gain CS-US associations, and indirect CS-US associations
Association with a US through configural stimuli. Configural stimuli are cre-
Bush and Mosteller (1955) offered a differential equa- ated by combining the internal representations of sim- C
tion describing how the CS-US association increased ple CSs. Configural stimuli are maximally active when
whenever the CS was presented with the US and some specific CSs are present and others are absent.
decreased when the CS was presented by itself. Rescorla The model was the first model of classical condition-
and Wagner (1972) modified the Bush–Mosteller ing to include an individual error term (Blough 1975,
(Bush and Mosteller 1955) equation to reflect the p. 20) to limit the associations gained by a single CS.
assumption that CSs compete to gain association In addition to the results explained by the Rescorla–
with the US. The ▶ Rescorla–Wagner Model can Wagner model, the SD model also describes – among
describe acquisition, partial reinforcement, generaliza- other paradigms – conditioning with different CS dura-
tion, extinction by increasing the US strength, US- tions, rapid reacquisition, learning to learn, compound
preexposure effect, forward blocking, unblocking, conditioning, negative and positive patterning, ISI
supernormal overshadowing, conditioned inhibition, effects, ITI effects, serial feature-positive (FP) and fea-
conditioning, overexpectation, and simultaneous fea- ture-negative (FN) discriminations, and biconditional
ture-positive and feature-negative discriminations. discrimination. Schmajuk et al. (1998) extended the SD
Sutton and Barto (1981) introduced a version of the model to describe how the conditioned response (CR)
Rescorla–Wagner model that describes learning as is determined by both the US and the CS, an important
a moment-to-moment (“real time”) process. Van issue in occasion setting.
Hamme and Wasserman (1994) described a modified
version of the Rescorla and Wagner (1972) model. They Competition, Timing, and
proposed that the association of a CS with the US Configurations
decreases when the CS is absent, instead of staying Desmond and Moore (1988) offered a neural network
constant as in the original model. In addition to the that describes adaptive timing in classical condition-
paradigms listed above, the modified model can ing. Grossberg and Schmajuk (GS) (Grossberg and
explain paradigms recovery from overshadowing and Schmajuk 1989) presented a model that assumes that
blocking, and backward blocking. a CS generates multiple temporal representations and
can describe training with multiple USs. Buhusi and
Competition and Configurations Schmajuk (1999) combined the SD and the GS models
Kehoe (1988) offered a layered network model of asso- to describe timing of the peak CR, training with mul-
ciative learning in which the CS inputs, using a com- tiple USs, the temporal specificity of blocking, and
petitive rule as the previous models, learn to activate temporal specificity in serial FP discriminations.
configural hidden units when the US is presented. In
turn, the hidden units can become associated with the Competition Without Configurations
US. In addition to most of the results explained by the Some models incorporate competitive rules but do
original Rescorla–Wagner model, the model is able to not use configural representations to solve nonlinear
address rapid reacquisition, learning to learn, com- problems. For instance, McLaren and Mackintosh
pound conditioning, and negative and positive pat- (2000) developed an elemental associative theory
terning. Gluck and Myers (1993) also introduced which assumes that all stimuli activate a set of common
a model that also incorporates a competitive rule and elemental units which provide a solution to negative
configural stimuli. patterning and biconditional discriminations. The
Schmajuk and DiCarlo (SD) (Schmajuk and model is also able to describe latent inhibition and
DiCarlo 1992) presented a “generalized” version of perceptual learning. Similarly, Harris (2006) proposed
the Rescorla–Wagner (Rescorla and Wagner 1972) a model in which a limited-capacity attentional mech-
rule into a model that also included temporal anism boosts the activation of elements that enter an
702 C Computational Models of Classical Conditioning
attention buffer. Therefore, individual elements lose validity, conditioned inhibition as a slave process,
activation when a stimulus is part of a compound. inhibitory sensory preconditioning, and counteraction
The model explains negative patterning because more between overshadowing and latent inhibition.
elements enter the buffer when A or B are presented
separately on A+ or B+ trials, than when they are Attentional Models
presented together on AB trials. On AB trials,
inhibitory associations are formed between the stron- Attention Increases When the CS Is
ger elements of each CS (referred to as A and B) and the a Good Predictor of the US
weaker elements of the other CS (referred to as b and a), Mackintosh (1975) suggested that ▶ attention to a
because A and B are in the attention buffer and a and given CS increases when that CS is the best predictor
b are outside the buffer. However, this mechanism does of the US, and decreases otherwise. The model can be
not allow the model to describe learning to learn and applied to forward blocking, overshadowing, and
occasion setting. Among other paradigms, the model latent inhibition. At the same time, Grossberg (1975)
also explains positive patterning, biconditional dis- offered a ▶ neural network in which CSs compete to
criminations, latent inhibition, and the results of com- activate their input nodes in proportion to their
pound conditioning of an excitor and an inhibitor. salience and association with the US. Interestingly,
the network implements Mackintosh’s (1975) atten-
Pure Configurations tional rule. Along the similar lines, Moore and Stickney
Pearce (1987) proposed a purely configural model acti- (1982), Schmajuk and Moore (1989), and Schmajuk
vated by the whole pattern of stimulation. For instance, and DiCarlo (1991) presented real-time versions of
presentation of A activates node A and presentation of Mackintosh’s (1975) rule and Grossberg’s (1975) net-
the compound AB activates a different node AB, which works. Both Moore and Stickney (1982) and Schmajuk
allows the model to readily solve negative patterning and Moore (1989) incorporate simultaneous excitatory
and biconditional discriminations. and inhibitory associations.
simultaneous excitatory and inhibitory associations, mediated acquisition, mediated extinction, and medi-
explains latent inhibition, the Hall–Pearce effect, and ated attentional changes.
unblocking by decreasing the US. The model describes, among other paradigms, US-
preexposure effect, forward blocking, unblocking by
Attention Increases When the CS Is increasing the US, unblocking by decreasing the C
a Poor Predictor of the US or Any US, overshadowing, conditioned inhibition, super-
Other CS, and When the CS Is Poorly normal conditioning, overexpectation, recovery from
Predicted by Other CSs or the CX overshadowing, recovery from forward blocking, back-
Schmajuk, Lam, and Gray (SLG) (Schmajuk et al. 1996; ward blocking, recovery from backward blocking, con-
Schmajuk and Larrauri 2006; Larrauri and Schmajuk ditioning with different CS durations, ISI and ITI
2008; Schmajuk 2009) proposed a neural network effects, latent inhibition, recovery from LI, counterac-
model of classical conditioning. Figure 1 shows a tion and synergy between overshadowing and latent
block diagram of the network which includes (1) a inhibition, external disinhibition, spontaneous recovery,
short-term memory and feedback system, (2) an atten- renewal, reinstatement, rapid or slow reacquisition,
tion system, (3) an association system, and (4) a nov- extinction of conditioned inhibition, conditioned inhi-
elty system. The SLG includes equations that portray bition as a slave process, second-order conditioning,
behavior on a moment-to-moment basis, attentional excitatory and inhibitory sensory preconditioning, and
control of the formation and retrieval of CS-US learned irrelevance (Schmajuk 2010).
and CS-CS associations, competition among CSs to
become associated with the US or other CSs, and reen- Combined Architectures
trant feedback of predictions of the CSs. Attention to Buhusi and Schmajuk (1996) combined the mecha-
the CS is controlled by the CS-US associations, by nisms of the SLG and the SD models into a model
context-CS (CX-CS) associations, and by CS-CS asso- that explains all the results previously addressed by
ciations. The feedback system allows the model to each model. Along a similar line, Le Pelley (2004)
describe inferences and cognitive mapping, as well as presented a model that included excitatory and
Novelty’
CSs, US
Associations
VCS1–CS1
Computational Models of Classical Conditioning. Fig. 1 Block diagram of the Schmajuk–Lam–Gray (SLG) (Schmajuk
et al. 1996) network. CS conditioned stimulus, US unconditioned stimulus, tCS short-term memory trace of the CS, BCS
prediction of the CS, zCS attentional memory, XCS internal representation of the CS, VCS1-CS1, VCS1-CS2, . . .,VCS1-US associations
CS1-CS1, CS1-CS2, . . ., CS1-US, CR conditioned response
704 C Computational Models of Classical Conditioning
inhibitory associations like Schmajuk and Moore (1989) 9. Conditioned ▶ inhibition. Stimulus CS2 acquires
model, an individual error term like Schmajuk and inhibitory conditioning with CS1 reinforced
DiCarlo (1992), an “attentional” associability similar trials interspersed with CS1-CS2 nonreinforced
to Mackintosh’s (1975), and a “salience” associability trials.
defined as in the Pearce–Hall (1980) model. 10. Supernormal conditioning. Reinforced CS1-CS2
presentations, following inhibitory conditioning
The Evolution of Computational of CS1, increase CS2 excitatory strength compared
Models of Conditioning with the case when it is trained in the absence
This entry presents a number of models that describe
of CS1.
many features of classical conditioning in terms of dif-
11. Overexpectation. Reinforced CS1-CS2 presenta-
ferent computations carried out on the conditioned
tions following independent reinforced CS1 and
and unconditioned stimuli. It is clear that the models
CS2 presentations result in a decrement in their
evolved – and are still evolving – from a few, relatively
initial associative strength.
simple equations to the present complex models able to
12. Simultaneous feature-positive discrimination.
account for many experimental results. The computa-
Reinforced simultaneous CS1-CS2 presentations,
tional complexity of these models puts our understand-
alternated with nonreinforced presentations of
ing of their workings beyond the ability of our intuitive
CS2, result in stronger responding to CS1-CS2
thinking and makes computer simulations irreplaceable.
than to CS2 alone. In this case, CS1 gains a strong
Interestingly, the complexity of the models frequently
excitatory association with the US.
results in function redundancy, a natural property of
13. Simultaneous feature-negative discrimination.
biologically evolved systems that is much desired in
Non-reinforced simultaneous CS1-CS2 presenta-
technologically designed products.
tions, alternated with reinforced presentations of
Some Important Classical CS2, result in weaker responding to CS1-CS2 than
Conditioning Results to CS2 alone. In this case, CS1 gains a strong inhib-
1. Acquisition. After a number of CS-US pairings, itory association with the US.
the CS elicits a conditioned response (CR) that 14. Recovery from overshadowing. Extinction of
increases in magnitude and frequency. the CS1 results in increased responding to the
2. ▶ Partial reinforcement. The US follows the CS overshadowed CS2.
only on some trials. 15. Recovery from forward blocking. Extinction of the
3. Generalization. A CS2 elicits a CR when it shares blocker CS1 results in increased responding to the
some characteristics with a CS1 that has been blocked CS2.
paired with the US. 16. Backward blocking. Conditioning to CS1 following
4. Extinction. When CS-US pairings are followed by conditioning to CS1-CS2 results in a weaker con-
presentations of the CS alone or by unpaired CS ditioning to CS2 than that attained with CS2-US
and US presentations, the CR decreases. pairings.
5. US-Preexposure effect. Presentation of the US in 17. Conditioning with different CS durations. Condi-
a training context prior to CS-US pairings retards tioning first increases and then decreases with
production of the CR. increasing CS durations when the US is presented
6. Forward blocking. Conditioning to CS1-CS2 fol- at the end of the CS.
lowing conditioning to CS1 results in a weaker 18. Rapid reacquisition. CS-US presentations follow-
conditioning to CS2 than that attained with CS2- ing extinction result in faster reacquisition.
US pairings. 19. Learning to learn. Learning a CS1-US association
7. Unblocking. Increasing the US increases facilitates the subsequent learning of a CS2-US
responding to the blocked CS2. association.
8. Overshadowing. Conditioning to CS1-CS2 results 20. Compound conditioning. Reinforced CS1-CS2
in a weaker conditioning to CS2 than that attained results in stronger responding to the compound
with CS2-US pairings. than to the components.
Computational Models of Classical Conditioning C 705
21. Positive patterning. Reinforced CS1-CS2 presenta- 33. Latent inhibition. Preexposure to a CS followed
tions intermixed with nonreinforced CS1 and CS2 by CS-US pairings retard the generation of the CR.
presentations result in stronger responding to 34. Perceptual learning. Preexposure to a couple of CSs
CS1-CS2 than to the sum of the individual facilitates the acquisition of a discrimination
responses to CS1 and CS2. between them. C
22. Negative patterning. Nonreinforced CS1-CS2 pre- 35. Simultaneous excitatory and inhibitory associa-
sentations intermixed with reinforced CS1 and CS2 tions. A CS can simultaneously act as excitor and
presentations result in weaker responding to CS1- inhibitor of the CR.
CS2 than to the sum of the individual responses to 36. Backward conditioning. Excitatory conditioning is
CS1 and CS2. obtained when the US precedes the CS by a short
23. Interstimulus interval (ISI) effects. Conditioning is interval and inhibitory conditioning when the
maximal at an optimal ISI and gradually decreases interval is long.
with increasing ISIs. 37. Conditioned diminution or facilitation of the
24. Intertrial interval (ITI) effects. Conditioning to the unconditioned response (UR). A reduction in the
CS increases with longer ITIs. amplitude of the UR that immediately follows
25. Serial feature-positive discrimination. Reinforced a previously reinforced CS.
successive CS1-CS2 presentations, alternated with 38. Pretrial CS. Presentation of a CS before CS-US
nonreinforced presentations of CS2, result in stron- pairings decreases conditioning for short CS-CS
ger responding to CS1-CS2 than to CS2 alone. In intervals and increases conditioning for long CS-
this case, CS1 acts as an occasion setter. CS intervals.
26. Serial feature-negative discrimination. 39. Pretrial US. Presentation of a US before CS-US
Nonreinforced successive CS1-CS2 presentations, pairings decreases conditioning.
alternated with reinforced presentations of CS2, 40. Recovery from LI. Presentation of the US in the
result in weaker responding to CS1-CS2 than to context of preexposure and conditioning results in
CS2 alone. In this case, CS1 acts as an occasion renewed responding to the preexposed CS.
setter. 41. Unblocking by decreasing the US. Decreasing the
27. Biconditional discrimination. Four stimuli are US in the second phase of forward blocking can
paired in four different combinations, two that increase responding to CS2.
are reinforced (AB+ and CD+), and two that are 42. Hall–Pearce negative transfer effect. CS-US associ-
not (AC and BD). ations with a weak US slow down subsequent CS-
28. CR is determined by both the US and the CS. The US associations with a strong US.
nature of the CR is determined not only by the US 43. Counteraction between overshadowing and latent
but also by the CS. inhibition. The combined effect of latent inhibi-
29. Timing of the peak CR. The CR peaks at the time of tion and overshadowing results in stronger
the US presentation during training (equivalent to responding than that individually obtained with
responding at the ISI). each procedure.
30. Training with multiple USs. A CS trained with a US 44. Synergy between overshadowing and latent inhi-
presented at different ISIs will present peaks cen- bition. The combined effect of latent inhibition
tered at those ISIs. and overshadowing results in weaker respond-
31. Temporal specificity of the competition between CSs ing than that individually obtained with each
in blocking. Blocking is observed when the blocked procedure.
CS is paired in the same temporal relationship 45. External desinhibition. Presenting a novel stimulus
with the US as the blocking CS. immediately before a previously extinguished CS
32. Temporal specificity in serial FP discriminations. might produce renewed responding.
A serial feature-positive discrimination is best 46. Spontaneous recovery. Presentation of the CS after
when the feature-target interval during testing some time after the subject stopped responding
matches the training interval. might yield renewed responding.
706 C Computational Models of Classical Conditioning
47. Renewal. Presentation of the CS in a novel context Bush, R. R., & Mosteller, F. (1955). Stochastic models for learning.
might yield renewed responding. New York: Wiley.
Denniston, J. C., Savastano, H., & Miller, R. R. (2001). The extended
48. Reinstatement. Presentation of the US in the con-
comparator hypothesis: learning by contiguity, responding by
text of extinction and testing might yield renewed relative strength. In R. R. Mowrer & S. B. Klein (Eds.), Handbook
responding. of contemporary learning (pp. 65–117). Mahwah: Lawrence
49. Rapid or slower reacquisition. Based on the length Erlbaum.
of the extinction phase, CS-US presentations fol- Desmond, J. E., & Moore, J. W. (1988). Adaptive timing in neural
models: The conditioned response. Biological Cybernetics, 58,
lowing extinction might result in faster or slower
405–415.
reacquisition. Dickinson, A., & Burke, J. (1996). Within-compound associations
50. Extinction of conditioned inhibition. Inhibitory mediate the retrospective revaluation of causality judgments.
conditioning is extinguished by CS2-US presenta- Quarterly Journal of Experimental Psychology, 49B, 60–80.
tions, but not by presentations of CS2 alone. Gelperin, A., Hopfield, J. J., & Tank, D. W. (1985). The logic of Limax
51. Conditioned inhibition as a slave process. After CS1- learning. In A. Selverston (Ed.), Model neural networks and
behavior (pp. 237–261). New York: Plenum.
US and CS1-CS2 presentation, extinction of the
Gluck, M. A., & Myers, C. E. (1993). Hippocampal mediation of
CS1-US association results in the elimination of stimulus representation: A computational theory. Hippocampus,
the retardation in conditioning the conditioned 3, 491–516.
inhibitor CS2. Grossberg, S. (1975). A neural model of attention, reinforcement, and
52. Second-order conditioning. When CS1-US pairings discrimination learning. International Review of Neurobiology,
18, 263–327.
are followed by CS1-CS2 pairings, presentation of
Grossberg, S., & Schmajuk, N. A. (1989). Neural dynamics of adap-
CS2 generates a CR. tive timing and temporal discrimination during associative
53. Sensory preconditioning (Excitatory). When CS1- learning. Neural Networks, 2, 79–102.
CS2 pairings are followed by CS1-US pairings, Harris, J. A. (2006). Elemental representations of stimuli in associa-
presentation of CS2 generates a CR. tive learning. Psychological Review, 113, 584–605.
54. Sensory preconditioning (Inhibitory). When CS1-X/ Kehoe, E. J. (1988). A layered network model of associative learning:
Learning to learn and configuration. Psychological Review, 95,
CS2-X alternated presentations are followed by 411–433.
CS1-US pairings, CS2 becomes inhibitory. Le Pelley, M. E. (2004). The role of associative history in models of
55. Learned irrelevance. Random exposure to the CS associative learning: A selective review and a hybrid model.
and the US retards conditioning even more than Quarterly Journal of Experimental Psychology, 57B, 193–243.
combined latent inhibition and US preexposure. Mackintosh, N. J. (1975). A theory of attention: Variations in the
associability of stimuli with reinforcement. Psychological Review,
56. Recovery from backward blocking. Extinction of the
82, 276–298.
blocker CS1 results in increased responding to the McLaren, I. P. L., & Mackintosh, N. J. (2000). An elemental model of
blocked CS2. associative learning: I. Latent inhibition and perceptual learning.
Animal Learning & Behavior, 28, 211–246.
Cross-References Miller, R. R., Schachtman, T., & Spear, N. E. (1985). Conditioning
▶ Cognitive Models of Learning context as an associative baseline: Implications for response
generation and the nature of conditioned inhibition. In
▶ Computational Models of Human Learning R. R. Miller (Ed.), Information processing in animals: Conditioned
▶ Conditioning inhibition (pp. 51–88). Hillsdale: Lawrence Erlbaum.
▶ Fear Conditioning Moore, J. W., & Stickney, K. J. (1980). Formation of attentional-
▶ The Role of Attention in Pavlovian Conditioning associative networks in real time: Role of the hippocampus
and implications for conditioning. Physiological Psychology, 8,
References 207–217.
Blough, D. S. (1975). Steady state data and a quantitative model of Pearce, J. M. (1987). A model for stimulus generalization in Pavlovian
operant generalization and discrimination. Journal of Experi- conditioning. Psychological Review, 94, 61–73.
mental Psychology: Animal Behavior Processes, 104, 3–21. Pearce, J. M., & Hall, G. (1980). A model for Pavlovian learning:
Buhusi, C. V., & Schmajuk, N. A. (1996). Attention, configuration, Variations in the effectiveness of conditioned but not of uncon-
and hippocampal function. Hippocampus, 6, 621–642. ditioned stimuli. Psychological Review, 87, 532–552.
Buhusi, C. V., & Schmajuk, N. A. (1999). Timing in simple condi- Rescorla, R. A., & Wagner, A. R. (1972). A theory of Pavlovian
tioning and occasion setting: A neural network approach. Behav- conditioning: Variation in the effectiveness of reinforcement
ioral Processes, 45, 33–57. and non-reinforcement. In A. H. Black & W. F. Prokasy (Eds.),
Computational Models of Human Learning C 707
Input units Output units To make this simple model more powerful,
researchers have extended it to include hidden units
CS1 (see Fig. 2). In this way, the computational power of the
US1 model greatly increased. As a simple example, deter-
mining the parity (odd/even) of the number of input
CS2 units that is active is impossible without hidden units,
but it becomes possible when hidden units are added.
US2 This type of model is applied to domains where com-
CS3
plex input–output transformations are required such
as reading aloud (orthography-phonology mapping),
sensory coordinate transformations (e.g., from eye-
to head-centered object representations), and mental
arithmetic. To train a model containing hidden units,
Teaching signal a learning rule is specified in which the training signal
given at the output level is passed backward in the
Computational Models of Human Learning. Fig. 1 network (from output to hidden units, a process called
Learning procedure in supervised model back propagation; see Fig. 2). Unfortunately, this pro-
cess is even less biologically plausible than the Rescorla-
Wagner learning rule.
These two indices are often in a trade-off relation (see Unsupervised models do not receive teacher signals,
below). as the name suggests, and are therefore more biologi-
Historically, two types of model have been influen- cally plausible. These models trace back to the influen-
tial: supervised and unsupervised models. In super- tial proposal by Donald Hebb that if two units (cells)
vised models, an external teaching signal is injected in are active together, their connection weight will be
the computational system and used in the model’s increased. Learning rules based on this principle are
learning rule (Fig. 1). This is then used to adapt the called Hebbian learning rules. Much data attest to the
configuration of the system. A seminal model from this biological plausibility of this learning rule. However,
class is the Rescorla-Wagner model to account for models based (only) on Hebbian learning typically
aspects of classical conditioning. One of the main have much less computational power than supervised
attractions of this model was that it was able to account learning models and are in this sense less behaviorally
for the phenomenon of blocking, observed in human plausible.
and nonhuman organisms, which holds that CS–US A third class of models is reinforcement learning
relations are learned only when the US is not predict- models. They strike a middle ground between the
able. This attests to the behavioral plausibility of the two traditional classes of model (supervised and
Rescorla-Wagner model, and made it extremely popu- unsupervised models) and may be called weakly super-
lar as a model of relatively simple learning tasks. On the vised. They have recently become very popular because
other hand, there is little direct biological evidence for of their biological plausibility. In such models, there is
the existence of this type of learning rule in the human no feedback to the system about what activation level
brain, except in very specific areas such as the cerebel- each of the units should have (as is the case in super-
lum (Gluck and Myers 2001). vised models), but instead a broad reinforcement signal
Researchers in cognitive science have applied and is provided which informs the system whether its per-
extended this model to human learning in domains formance was “good” or “bad” (Fig. 3). One reason for
well beyond conditioning. For example, Rumelhart the recent interest in reinforcement learning models is
and McClelland (1986) applied it to learning to gener- the remarkable convergence on similar concepts in two
ate the past tense of English verbs. Attesting to its traditionally separated research streams, computer sci-
behavioral plausibility, it was observed that during ence and neurophysiology. In computer science, so-
learning the model exhibited similarities to the error called temporal difference models of reinforcement
patterns of children. learning learn from a signal expressing the difference
Computational Models of Human Learning C 709
(intermediate
(e.g., orthographic features) (e.g., phonological features)
computations)
Backpropagated
teaching signal
Teaching signal
Computational Models of Human Learning. Fig. 2 Learning procedure in supervised model with hidden units
(back propagation)
reasoning is at this time not easily accommodated by Roelfsema, P. R., & van Ooyen, A. (2005). Attention-gated reinforce-
the type of models described here. To address this, ment learning of internal representations for classification. Neu-
ral Computation, 17, 2176–2214.
recent models endow biologically plausible models
Rumelhart, D., & McClelland, J. (1986). Parallel Distributed
with extra computational power. One interesting devel- Processing. Cambridge, MA: MIT Press.
opment is to add hierarchical representations to rein- Sutton, R. S., & Barto, A. G. (1998). Reinforcement learning: An
forcement learning models (Botvinick et al. 2009) introduction. Cambridge, MA: MIT Press.
which can be used for hierarchical planning. Another
is to add randomly connected neurons that have dif-
ferent activation states at different time points, which
can be used for precisely timed action sequences (e.g., Computational Models of
dancing; Buonomano and Maas 2009).
Learning
As mentioned above, Hebbian learning in itself
is not very powerful; however, it can be used as ▶ Cognitive Models of Learning
a building block for more powerful rules. For example,
a reinforcement learning rule can be constructed
by modulation of a Hebbian learning rule. Hebbian
learning in a particular cortical region could be Computational Natural
increased whenever a dopaminergic reinforcement sig-
nal arrives there. This even allows construction of
Language Learning
learning rules with the same computational power as ▶ Machine Learning of Natural Language
backpropagation learning (Roelfsema and van Ooyen
2005). In recent years, other neurotransmitters have
been proposed to provide important learning signals
for the cortex (e.g., serotonin, noradrenalin). Also
Computer Adaptive Testing
these neurotransmitters could operate by modulating
Hebbian learning processes. Recent research investi- ▶ Adaptive Evaluation Systems
gates how these neurotransmitters interact with corti-
cal areas to obtain powerful devices for learning in the
human brain (Doya 2008).
Computer Simulation Model
Cross-References
▶ Computational Models of Classical Conditioning DIRK IFENTHALER
▶ Connectionist Theories of Learning Institut für Erziehungswissenschaft, Albert-Ludwigs-
▶ Reinforcement Learning University Freiburg, Freiburg, BW, Germany
References
Botvinick, M. M., Niv, Y., & Barto, A. C. (2009). Hierarchically Synonyms
organized behavior and its neural foundation: A reinforcement
Simulation model; Simulator model; System simula-
learning perspective. Cognition, 113, 262–280.
Buonomano, D. V., & Maas, W. (2009). State-dependent computa- tion model
tions: Spatiotemporal processing in cortical networks. Nature
Reviews Neuroscience, 10, 113–125. Definition
Doya, K. (2008). Modulators of decision making. Nature Neurosci- A computer simulation model is a computer program
ence, 11, 410–416. or algorithm which simulates changes of a modeled
Frank, M. J., Seeberger, L. C., & O’Reilly, R. C. (2004). By carrot or by
system in response to input signals.
stick: Cognitive reinforcement learning in Parkinsonism. Science,
306, 1940–1943.
Gluck, M. A., & Myers, C. E. (2001). Gateway to memory: An intro- Theoretical Background
duction to neural network modeling of the hippocampus and learn- Simulation has become a widely used tool for training
ing. Cambridge, MA: MIT Press. and for research on human interaction with complex
Computer Simulation Model C 711
work and learning environments. A simulation is Computer simulation models include three major
defined as a working representation of reality (Jones determinants: Time, behavior, and data (see Fig. 1).
1980). Further, a simulation may be an abstracted, The factor time defines a static or dynamic com-
simplified, or accelerated model of a process or system puter simulation model. A static computer simulation
which allows exploration where reality is too expensive, model includes variables and parameters which are not C
complex, dangerous, fast, or slow (de Freitas 2006). time dependent. Dynamic computer simulation
The richness in the use of simulations to support models include variables and parameters which model
learning is clearly documented in the literature and time-varying states of the simulated phenomenon.
numerous research projects. Training programs using The factor behavior defines a stochastic or determin-
simulations have been successfully applied in the fields istic computer simulation model. Stochastic computer
of flight training, health care education, dental educa- simulation models are characterized through their
tion, command and control training of large incidents, indeterminacy in future evolutions which are described
team-based decision making, simulations for the train- by probability distributions. Deterministic computer
ing of firefighters, teacher education, and many other simulation models include no randomization in the
domains (see Ifenthaler 2009). development of future events of the simulation.
Technically, computer simulations which model A special deterministic model is a chaotic computer
some specific domain of reality allow users to change simulation model whose behavior cannot be entirely
input variables by manipulating objects or entering predicted. The factor data defines a discrete or contin-
data. The results of the simulation are represented as uous computer simulation model. Discrete computer
dynamically generated graphs, numeric displays, and simulation models include variables which change only
texts (de Jong and van Joolingen 2008). Hence, three at specific points in time at which an event occurs.
major components of simulations can be identified: Continuous computer simulation models include vari-
(1) the simulation model, (2) the execution of the ables which change in a continuous way including
simulation model, and (3) the analysis of the executed infinite number of states.
simulation model. First, the simulation model may be The development of adequate computer simulation
based on declarative, conceptual, or functional under- models requires the definition of the three major
standing of a specific phenomenon to be simulated. determinants; time, behavior, and data. Besides these
Second, the execution of the simulation model is factors, numerous design decisions must be taken into
defined through specific algorithms, e.g., serial execu- account, e.g., the application area, the programming
tion, parallel execution, or fuzzy execution algorithms. framework, the user interface, the system support, the
Finally, the analysis approach of the executed simula- simulation engine, and so forth (for details see
tion model may focus on the input–output processes, Fishwick 1998; Sulistio et al. 2004).
the verification of results, the visualization of output
data, and the validation of the simulated output. How-
ever, a sufficient simulation requires a well-founded Behavior
model of the simulated phenomenon – the computer • Stochastic
simulation model. • Deterministic
Initially, a simulation model has been realized as Time Data
a mathematical model. These mathematical models • Static • Discrete
have become a useful tool in physics, chemistry, biol- • Dynamic • Continuous
ogy, economics, engineering, and social sciences. How-
ever, the ad hoc manipulation of variables from outside Computer
a predefined mathematical model is not possible. simulation
Accordingly, the application of computer programs or model
algorithms enables higher variability and stronger indi-
vidualized simulation runs. This is primarily realized
through the change of parameters of the computer Computer Simulation Model. Fig. 1 Determinants of
simulation model. computer simulation models
712 C Computer Simulation Model
e1
x1
p12 p12 e1
x1 y1
p21
y1
Computer Simulation Model. Fig. 2 Simple path model Y2
p11
including one independent variable
e2
Important Scientific Research and
Open Questions Computer Simulation Model. Fig. 3 Expanded path
From a cognitive science perspective, learning initi- model
ated by simulation involves explorative thinking and
inductive and analogical reasoning. This places high
cognitive and metacognitive demands on the learner, The visualization of these two equations results in
who must generate hypotheses and test them by the following path model (see Fig. 3):
accomplishing learning tasks as well as performing The path coefficients can be used to decompose the
experiments in the simulated environment. Accord- correlations in the path model into direct and indirect
ingly, simulations of complex processes and systems effects (the total causal effect of variable i on variable j is
often require complex problem solving. Complex the sum of the values of all the paths from i to j). The
problem solving requires iterative steps of hypothesis total causal effect on y1 is the sum of all direct and
testing as well as increased time for constructing appro- indirect effects (see (4)):
priate mental models. Mental models are constructed
ef f total ¼ p12 p11 þ p21 ð4Þ
in order to hypothesize and understand the structure of
the simulation process or system and to simulate trans- Applying the mathematical assumptions of a path
formations of these processes and systems mentally. model and the related path coefficients, we are able to
Currently, research focuses on the development of transform a path model into equations for realizing the
adequate computer simulation models for the social necessary computer simulation model. All equations of
and cognitive sciences. Ifenthaler (2009) suggests the direct and indirect effects from the path model are
application of path models for the development of included in the computer simulation model.
computer simulation models. Path models include Future research should address a formal and struc-
path coefficients, which are standardized regression tural comparison of available computer simulation
coefficients showing the direct effect of an independent models as well as a meta-analysis of simulation models
variable on a dependent variable. Additionally, regres- and their effects in the field of learning and instruction.
sion residuals are considered in the equations. A simple
path model including one descriptive variable x1 is Cross-References
shown in (1): ▶ Modeling and Simulation
▶ Simulation and Learning: The Role of Mental
y 1 ¼ p12 x1 þ e1 ð1Þ
Models
e1 indicates the residual of the path model. ▶ Simulation-Based Learning
The path model described in (1) can be visualized as
follows (see Fig. 2): References
If the path model in Fig. 2 is expanded with an de Freitas, S. I. (2006). Using games and simulations for supporting
additional descriptive variable, it can be specified by learning. Learning, Media and Technology, 31(4), 343–358.
the following path equations (see (2), (3)): de Jong, T., & van Joolingen, W. R. (2008). Model-facilitated learning.
In J. M. Spector, M. D. Merrill, J. van Merrienboer, &
y 1 ¼ p11 y 2 þ p12 x1 þ e1 ð2Þ M. P. Driscoll (Eds.), Handbook of research on educational com-
munications and technology (pp. 457–468). New York: Taylor &
y 1 ¼ p21 x1 þ e2 ð3Þ Francis.
Computer-Based Learning C 713
Definition
In computer-based learning (CBL), the computer is
used for instructional purposes whereas the computer
Computer-Assisted Instruction hard- and software as well as the peripherals and input
(CAI) devices are key components of the educational envi-
▶ Interactive Learning ronment. CBL assists individuals in learning using
multiple representations of information for a specific
educational purpose. Common innovative realizations
of CBL to improve teaching and learning are hypertext,
simulations, and microworlds.
Computer-Assisted Learning
(CAL) Theoretical Background
The development of the first integrated circuit by
▶ Interactive Learning Noyce and Kilby in the late 1950s marked the dawn
of the role of computer technology in education.
In the following years the microcomputer was de-
veloped featuring audio, colors, peripherals, and
input devices, as well as a graphical user interface
Computer-Assisted Training (Ifenthaler 2010). In CBL, the computer is regarded
(CBT) as the key component of the educational environment.
▶ Interactive Learning Individuals are assisted in learning from multiple rep-
resentations of information for a specific educational
purpose. CBL provides promising opportunities for
fostering meaningful learning (Lajoie 2000). Common
innovative realizations of CBL to improve teaching
Computer-Based Collaborative and learning are hypertext (text that links to other
Learning information), simulations (characteristics of a system
can be influenced through change of underlying vari-
▶ Online Collaborative Learning ables), and microworlds (environment where individ-
uals explore information and construct or change the
environment).
In the 1960s and 1970s, PLATO (Programmed
Computer-Based Education Logic for Automated Teaching Operations) was the
(CBE) first computer system which was used for programmed
instruction (Lockee et al. 2008). Programmed ins-
▶ Interactive Learning truction has influenced modern instructional design
714 C Computer-Based Learning
processes and laid the foundations for computer- interests (Seel and Casey 2003), and a basic education in
mediated instruction (e.g., Glaser 1965; Hartley 1974; information technology became a real hit in these years
Stolurow 1961). (Altermann-Köster et al. 1990). Educators tried just
Especially the 1980s and 1990s produced a huge about everything they could to teach their students
range of CBL, e.g., computer-assisted learning how to use computers. More important than these
(McDougall 1985), multimedia learning environments changes in the classroom, however, was the fact that
(Mayer 2001), hypermedia environments (Dillon and information and communication technology were
Jobst 2005), or simulations, games, and microworlds increasingly becoming a part of the daily lives of chil-
(Reiber 2005). However, since the early days of CBL it dren, teenagers, and adults.
has been subject of scrutiny and debate with arguments Like it or not, the general proliferation of com-
being advanced both in support of and against the use puter-based information and communication tech-
of computers for learning and instruction. nologies is irreversible, and computers now play an
A quarter of a century ago Greenfield (1984) took important role in human learning in everyday life as
up the topic of new media and communication tech- well as at educational institutions (Ifenthaler 2010).
nologies and discussed their possible effects on the
learning and behavior of children. The topic was
approached from a fundamentally positive, albeit crit-
Important Scientific Research and
Open Questions
ical perspective. New technologies were understood
Today, there is widespread agreement among educa-
as cultural artifacts that demand complex cognitive
tional theorists on the point that educational applica-
skills for their use which are not learned or taught in
tions of modern information and communication
school, but rather only through active manipulation
technologies can be made more effective when they
and practice in everyday life. However, the discussion
are embedded in multimedia learning environments
in the 1980s was dominated more by critical voices.
created to enable productive learning. CBL environ-
Günther (1986), for instance, warned vehemently
ments should be designed to enable learners to
against an overly hasty introduction of computers in
explore them with various amounts of guidance and
schools and was only prepared to accept it if the schools
construct knowledge and develop problem-solving
also offered regular outdoor excursions and field trips.
methods independently (Ifenthaler 2009; Seel et al.
The well-known proponent of educational reform von
2009). The key to success is seen not so much in how
Hentig (1987) recommended waiting as long as possi-
the information is presented as in how well the learners
ble to offer computer courses to school students.
can manipulate the different tools available in the
Haft (1988) commented on this discussion by
CBL environment on their own. However, empirical
pointing out that every technological advancement in
research also shows that students often struggle while
history has led to a perceived loss of immediacy, belief,
confronted with a CBL environment (Lajoie and
and confidence in one’s own experiences but that
Azevedo 2006). Extensive use of a computer as a tool
in most cases the pessimistic predictions concerning
for solving problems can help learners to concentrate
the proliferation of new technologies has turned out to
on understanding and solving problems rather than the
be ungrounded. Whereas ardent educational reformers
finished product or the acquisition of declarative
warned of the dangers of the computer, parents and
knowledge and can awaken their curiosity and creativ-
children were quick to see the potential of the computer,
ity. Several characteristics of the new technologies con-
and the PC made its way rapidly into children’s bed-
tribute to this effect:
rooms – more rapidly, at any rate, than into schools.
Schools began reacting to this challenge in the ● The new information and communication technol-
1990s and made systematic efforts to improve the ogies are interactive systems.
information technology competence of their students. ● The learners themselves are placed in control of
Computer literacy, the ability to work competently what and how they learn.
and effectively with computer technologies and pro- ● The computer can model real situations and com-
grams, advanced increasingly to the fore of pedagogical plex systems and simulate their behavior.
Computer-Based Learning C 715
● The learners can receive immediate feedback on can be supported effectively is sometimes left out of
their activities. the picture.
● In many cases the computer can also execute com- Much of what we discussed above is already dated
plex operations (e.g., simulations of dangerous sit- in a technological as well as a pedagogical sense and will
uations) which cannot be executed as well or at all in a few years be hardly more than a historical footnote C
by other media (Seel and Dijkstra 2004). like the Jasper Woodbury Series (Cognition and Tech-
nology Group at Vanderbilt 1997) or the goal-based
Indeed, when one considers that modern com- scenarios (Schank et al. 1994). We believe that the days
puters can represent all forms of information and of pre-programmed online courses are numbered, in
knowledge needed for learning and problem solving, which the learner – as in the classical paradigm of
the current state of computer technology seems to programmed instruction – is viewed more as an audi-
make the tedious process of integrating traditional ence than as an active constructor. In the near future,
media (such as texts, graphics, video) technically learners will be the constructors of their own environ-
superfluous and obsolete. Moreover, recent develop- ments and create the structures of the content units on
ments in the area of interactive software provide their own.
unique possibilities for creating virtual learning envi-
ronments and modeling complex systems without Cross-References
professional guidance. The options for independent ▶ Blended Learning
development of interactive environments are manifold, ▶ Learning Management Systems
and the graphical capabilities of new software programs ▶ Model-Based Learning with System Dynamics
include exciting animations and simulations of highly ▶ Programmed Instruction
complex processes. Last but not least, everything is
comparatively inexpensive and thus readily available
to the broader public (Ifenthaler 2010).
References
Altermann-Köster, M., Holtappels, H. G., Günther, H., Kanders, M.,
However, the advantages of CBL lie not only in the Pfeiffer, H., & de Witt, C. (1990). Bildung über computer?
area of education, but also in administrative, financial, Informationstechnische Grundbildung in der Schule. München:
and social domains. The main educational advantages Juventa.
may be summed up as follows: Cognition and Technology Group at Vanderbilt. (1997). The Jasper
project. Lessons in curriculum, instruction, assessment, and profes-
● The independence of learning and teaching from the sional development. Hillsdale, NJ: Lawrence Erlbaum.
constraints of time and space: Learners (e.g., college Dillon, A., & Jobst, J. (2005). Multimedia learning with hypermedia.
In R. E. Mayer (Ed.), The Cambridge handbook of multimedia
students) can follow a course from any point on the
learning (pp. 569–588). New York: Cambridge University Press.
earth and at any point in time, and the courses can Glaser, R. (Ed.). (1965). Teaching machines and programmed
be offered worldwide. instruction (Vol. II). Washington, DC: National Education
● The individuality of learning: Courses can be Association.
adapted to the needs of each individual learner Greenfield, P. M. (1984). Mind and media. The effects of television,
computers, and video games. Cambridge, MA: Harvard University
and course materials can be reused and rearranged
Press.
as often as one likes (provided that they are orga- Günther, H. (1986). Jugend und computer. Auswertung einer
nized in modules). empirischen Untersuchung. Pädagogische Rundschau, 40, 669–686.
Haft, H. (1988). Einführung: Neue Medien und Sozialisation – Die
Although these advantages are actually all beyond Technik rennt, die Forschung humpelt. Unterrichtswissenschaft,
question, the discussion on the educational use of 16(4), 2–4.
learning in the digital age often suffers from being Hartley, J. (1974). Programmed instruction. Programmed Learning
limited to the technological potential of information and Educational Technology, 11, 278–291.
Ifenthaler, D. (2009). Using a causal model for the design and devel-
and communication technologies (Seel and Ifenthaler
opment of a simulation game for teacher education. Technology,
2009). The technological possibilities for designing Instruction, Cognition and Learning, 6(3), 193–212.
CBL environments are doubtlessly great, but the Ifenthaler, D. (2010). Learning and instruction in the digital age.
pedagogically significant question as to how learning In J. M. Spector, D. Ifenthaler, P. Isaı́as, Kinshuk, &
716 C Computer-Based Learning Environment
theories and empirical research about what leads to context in which learning takes place. Cultural and
effective learning experiences. CBLEs can be designed societal issues are also considered in learning theories
to help students during thinking, problem solving, and that consider communities of learners and communi-
learning by providing them with opportunities to use ties of practice as a factor in learning. CBLEs can
their knowledge in complex contexts and meaningful provide situated learning experiences, where learners C
activities or situations. interact with complex problem-solving situations,
There is a long history of CBLEs and consequent using multiple media (e.g., text, video, animations,
theories underlying their design. Behaviorist stimulus- and diagrams). Theories of intelligence and aptitude
response theorists such as Skinner (1957) influenced tell us that learners differ in how they learn, for example
the use of computer-assisted instruction where multi- some learners respond better to verbal material and
ple-choice questions could be administered to students others respond better to visual material. There is not
automatically. The computer was seen as a “teaching one best way to teach individuals given these individual
machine” whose key benefit was its ability to provide differences and consequently CBLEs that use multiple
immediate feedback to reinforce correct responses or to representations can help to meet a variety of individual
correct incorrect responses with predetermined solu- learning needs (Moreno and Mayer 2007). Further-
tions. Questions would increase in difficulty as learners more, CBLEs can scaffold learners in the context of
demonstrated mastery at a particular level. Once stu- their learning by providing adaptive technological
dents mastered one set of problems they would then assistance in the form of computer tutoring (e.g., intel-
move on to the next level of difficulty. ligent tutoring systems) or pedagogical agents or with
The influence of developments in the information human assistance of those more proficient. The social
sciences (e.g., mathematical theory of communication, context of learning and collaboration using CBLEs is
computer programming, systems analysis) contributed a field in itself where complex methodologies docu-
to the development of cognitive theories that viewed ment how human dialogue by peers and mentors leads
learning as a form of information processing. Instead of to better understanding.
just looking at learning outcomes, problems could be
decomposed into the individual cognitive processes Important Scientific Research and
needed to solve them. The development of domain- Open Questions
specific cognitive models made it possible to identify As technology becomes part of our everyday lives,
and remediate errors that learners might make in the educators need to incorporate such changes in their
context of a particular problem-solving situation. classrooms. Researchers can support educators by
CBLEs could be designed to use complex production demonstrating the effectiveness of CBLEs and by
rules to detect and correct student misconceptions designing more interactive and engaging environ-
(Anderson 1996) and provide appropriate feedback ments. Technology can respond to individuals through
based on the identification of learning impasses. The- its actions be they text-based, verbal, or reactions of
ories of expertise led to the identification of complex personal/pedagogical agents or avatars. The prevailing
models of competency that could be used to help the view is that the more natural the interaction with
less proficient become more proficient more efficiently. computer-based learning environments, the less awk-
CBLEs could use such models as exemplars for novices ward and more realistic the learning situation. Identi-
to observe, as well as benchmarks for dynamic forms fying the optimal level of realism to promote effective
of assessment of individual learners, to determine the engagement and learning is an open question, though
type of feedback learners would need in the context many researchers strive toward passing Turing’s (1950)
of learning. test of machine intelligence, whereby a reasonable per-
Situated learning theories describe how human son would not be able to distinguish between a human
thought and action are best supported in contexts and computer response to his or her actions. Given that
that provide opportunities for learners to integrate situating learning in authentic, meaningful, and engag-
their information from multiple sources (Greeno ing settings is the goal of current CBLEs, it is very likely
1989). Learning theories are now looking at the inter- that we need to keep moving forward in pursuit of
section between cognition, motivation, and the social artificial intelligence techniques applied to education
718 C Computer-Based Music Instruction (CBMI)
practice. One particular area of current research is in in project-based activities where students may create
the use of natural language techniques and dialogue. their own content. This type of CBLE is a challenge
For example, Graesser has developed AutoTutor (www. to traditional modes of schooling and is often best
autotutor.org), a system that engages in dialogue with assessed in the context of design experiments and
students learning about Newtonian physics and adap- other participatory methodologies.
tively responds using a combination of explanations,
prompts, and feedback on errors. Cross-References
A second approach to enhancing engagement is to ▶ Agent-Based Modeling
detect and respond to changes in students’ emotions ▶ Computer-Based Learning
and levels of motivation as they use a CBLE. This builds ▶ Computer-Supported Collaborative Learning
on the research of Lepper, Malone, and others that ▶ Human-Computer-Interaction and Learning
shows that successful human tutors are able to main- ▶ Interactive Learning Environments
tain and direct continuous attention to both cognitive/ ▶ Technology-Enhanced Learning Environments
informational and motivational/affective factors, and ▶ Virtual Reality Learning Environments
formulate specific goals to maintain students’ confi-
dence, challenge, curiosity, and control. Sensor tech- References
nology can be used to detect emotion through a Anderson, J. R. (1996). ACT: A simple theory of complex cognition.
combination of physiological measures (e.g., EEGs, American Psychologist, 51, 355–365.
seat position, eye gaze, facial expression, skin conduc- Greeno, J. G. (1989). A perspective on thinking. American Psycholo-
gist, 44, 134–141.
tance). A relatively new area of research is investigating
Lajoie, S. P., & Azevedo, R. (2006). Teaching and learning in
how these physiological data can be used concurrently technology-rich environments. In P. A. Alexander &
with observational data, self-report, and outcome data P. H. Winne (Eds.), Handbook of educational psychology
to create motivating learning circumstances using (2nd ed., pp. 803–821). Mahwah, NJ: Erlbaum.
CBLEs. Building on previous work in gesture and Moreno, R., & Mayer, R. E. (2007). Interactive multimodal learning
environments. Educational Psychology Review, 19, 309–326.
face recognition, Lester has devised computational
Shute, V. J., Ventura, M., Bauer, M. I., & Zapata-Rivera, D. (2009).
models of affect recognition (automatically recogniz- Melding the power of serious games and embedded assessment
ing students’ affective states) and affect expression (that to monitor and foster learning: Flow and grow. In U. Ritterfeld,
automatically recognize and classify students’ affective M. Cody, & P. Vorderer (Eds.), Serious games: Mechanisms
states). Lester, Moreno, Azevedo, and VanLehn are each and effects (pp. 295–321). Mahwah, NJ: Routledge/Taylor and
currently examining how pedagogical agents (intelligent Francis.
Skinner, B. F. (1957). Verbal behavior. Englewood Cliffs, NJ:
virtual tutors) can employ language, facial expressions,
Prentice-Hall.
and gestures to engage learners and create effective Turing, A. (1950). Computing machinery and intelligence. Mind,
learning experiences. 236, 433–460.
Engagement is a necessary, but insufficient, condi-
tion for learning. New and innovative assessments need
to be created concurrently with new CBLEs to ensure
that we are collecting evidence of learning in these new
contexts. This can be a challenge given that different Computer-Based Music
domain-specific competencies are assessed in each Instruction (CBMI)
CBLE. One innovative approach to this challenge is
the use of stealth assessment, a process by which learner ▶ Technology in Music Instruction and Learning
performance data is continuously gathered during
the course of playing/learning. Stored in dynamic,
learner models, inferences are continuously drawn
about student competencies (Shute et al. 2009). Fur-
thermore, many of the more inquiry-based CBLEs pro-
Computer-Based Training
vide a significant amount of learner control to students ▶ Computer-Based Learning
Computer-Supported Collaborative Learning C 719
Computer-Supported Collaborative Learning, and the Four major empirical strands can be discerned
Springer (formerly Kluwer) Computer-Supported Col- as influential in CSCL. The experimental paradigm,
laborative Learning book series, among other venues. which typically compares an intervention to a control
The research field has been characterized by one if its condition by carefully manipulating variables, has
founders, Timothy Koschmann, as “a field centrally roots in cognitive and educational psychology. Exper-
concerned with meaning and practices of meaning- imentalism has been critiqued for failing to examine
making in the context of joint activity and the ways in learning in specific cases of interaction (most analyses
which these practices are mediated through designed aggregate the behavior of multiple individuals), and
artifacts” (Stahl et al. 2006). Understood in this way, for weak ecological validity due to the contrived situa-
CSCL is not merely a specialization of collaborative tions needed to control variables. The iterative design
learning within educational psychology, but rather is tradition continuously improves artifacts intended
relevant to any field of inquiry concerned with inter- to mediate learning and collaboration, with changes
subjective meaning-making (Suthers 2006). at each iteration driven by theory, observation, and
engagement of stakeholders. This tradition derives
Theoretical Background from CSCL’s roots in computer science and human–
Work undertaken in CSCL is based on several alter- computer interaction. Traditions of interaction analysis
native theoretical views of how social settings bear in CSCL are influenced by conversation analysis and
upon learning (Suthers 2006). Some theories treat the ethnomethodology, and examine how learning is
individual as the locus of learning. Research under accomplished in practice. These traditions privilege
a knowledge-communication epistemology examines participants’ own behavior and accounts rather than
how to more effectively present knowledge in some prior theoretical accounts, and typically focus on
medium, or how to otherwise communicate in ways short episodes of interaction (Stahl et al. 2006). Such
that cause or support learners’ acquisition of the methods are well suited to existentially quantified
desired knowledge. CSCL has moved decidedly away claims, yet are less developed for making predictive
from views of learning as transfer of knowledge, and generalizations. Finally, sociocultural analysis examines
toward more constructivist and interactional views. how institutional, cultural, and historical processes,
Constructivist epistemologies emphasize the agency of structures, and tools bear upon learning, identifying
the learner in constructing knowledge based on her how infrastructures produced at meso- and macro-
efforts to make sense of her experiences. These may scales influence learning in specific settings (Jones
include social experiences in which new ideas are et al. 2006).
encountered, some of which may conflict with one’s
own ideas, and the expectation to defend one’s own Important Scientific Research and
ideas. Some interactionalist epistemologies emphasize Open Questions
learners’ efforts to find “common ground” and share Some relevant findings in CSCL derive from or overlap
information with others. Other interactional episte- with the field of cooperative learning in education,
mologies, such as group cognition, treat learning as which has studied the conditions that affect whether
a process in which new ideas are jointly created through groups are beneficial for learning (e.g., group compo-
interaction. Here the agent of learning is the group sition, reward structures, task characteristics, role spe-
rather than the individual, and learning itself is not cialization, various forms process guidance). Due to
just a product of interaction but actually consists of space limitations, this article provides a sampling of
interaction. Participatory epistemologies bring the important trends within the field of CSCL itself and
agency of learning to the community level: becoming associated open questions. See Stahl et al. (2006) for
a member of a community of practice is not merely a brief history of CSCL and pertinent references.
a matter of an individual internalizing the knowledge A sampling of earlier research in CSCL may be found
and practices of that community, but also a process of in Koschmann et al. (2001).
the community’s own self-replication and growth as it A common strategy in CSCL is to identify interac-
takes on new members. tions that lead to learning and then try to get students
Computer-Supported Collaborative Learning C 721
to engage in these kinds of interactions. Based on socio- knowledge differences in heterogeneous groups,
cognitive conflict theory and research showing the and how scripts can drive software agents participat-
beneficial effects of attempting to articulate and justify ing in the collaboration. Critical issues include the
one’s own ideas, a major thrust of work in CSCL has coerciveness of scripting and the danger of denying
sought to engage learners in argumentation with each participants’ agency in learning to direct their own C
other (Andriessen et al. 2003). Here, “argumentation” learning.
is not used as synonymous with verbal conflict, but Technology-centric work in CSCL is in a delicate
rather to include cooperative interactions in which position, requiring an understanding of the concept
participants take a critical stance to ideas and their of affordances. Affordances are relationships between
justifications, exposing them to tests and comparing agents and their environments, relationships that offer
alternative points of view in an effort to reach greater potentials for action. Because human beings are cul-
understanding. Interventions explored include ICT- tural agents, our use of technologies is not deter-
supported role-playing, sentence-opener prompts that mined by their properties. Affordances are enacted
make different argumentative moves explicit, and through the meaning-making activities of learners.
representational notations and tools that support argu- Yet, affordances are not purely socially constructed or
mentation by making ideas and their interrelations and entirely relativistic: the properties of technologies make
evidence visible. The effectiveness of different com- some kinds of practices more available than others.
puter-mediated communication tools for supporting Consequently, designers of technologies for CSCL can-
argumentation has also been studied. Argumentation not treat their designs as directly controlling or deter-
scripts lead us to the next major area of research mining learning. Rather, an indirect approach is called
in CSCL. for in which designers offer potentials for desirable
Learners do not spontaneously engage in practices practices and examine how these potentials are actually
that lead to effective collaborative learning, such as taken up (Jones et al. 2006). Open questions lie in the
coordinating their joint efforts, referencing each others’ design and study of fundamentally social technologies
contributions, and building and evaluating grounded that are informed by the affordances and limitations of
arguments. Furthermore, they may be distracted from those technologies for mediating intersubjective mean-
such practices when attention must be allocated to ing-making (Suthers 2006).
managing the ICT and their group processes. For An advantage of studying learning in small groups
these reasons, collaboration scripts are studied as ways is that participants will display their understanding
to make learners’ interactions more productive for to other participants in ways that are also accessible
learning (see Fischer et al. 2007, on which this para- to educators and researchers (Stahl et al. 2006). Small
graph is based). Scripts are understood in psychology groups are also of interest because they lie at the
to refer to memory structures that guide people in boundary of and mediate between individuals and
understanding and participating in social action a community: the knowledge building that takes place
sequences, in computer science as formal structures within small groups becomes “internalized by their
that may be visualized or used to drive computational members as individual learning and externalized in
processes, and in education as practical means for their communities as certifiable knowledge” (Stahl
organizing learning activities. Scripts may apply at et al. 2006). Yet there has been insufficient research
a “macro” level in advance of a session by organizing that actually makes connections between these levels
who is collaborating on what task in what roles; and of analysis: most work examines either individual
at a “micro” level, by specifying the processes by learning outcomes or group processes, and does not
which learners conduct their activities. Research exam- trace connections between these levels. Also, the ways
ines issues such as the most effective ways to struc- in which institutions select and implement the infra-
ture interaction (e.g., scripting collaboration versus structures of CSCL that influence local interaction need
scripting reasoning), the conditions under which to be made visible (Jones et al. 2006). Hence, some
collaboration scripts are internalized so that external CSCL researchers are examining ways to bridge
support can be removed, the use of scripts to bridge between levels of analysis.
722 C Computer-Supported Collaborative Work
The development of the Internet and Web into Koschmann, T., Hall, R., & Miyake, N. (Eds.). (2001). CSCL II.
technological infrastructures for networked individ- Carrying forward the conversation. Mahwah, NJ: Lawrence
Erlbaum.
ualism and sociability has led to new challenges.
Stahl, G., Koschmann, T., & Suthers, D. D. (2006). Computer-
CSCL research has traditionally focused on strong supported collaborative learning: An historical perspective. In
relationships of cooperation and collaboration, but R. K. Sawyer (Ed.), Cambridge handbook of the learning sciences
is now faced with the question of whether to also (pp. 409–426). Cambridge: Cambridge University Press.
embrace proliferating “weak ties” of the new Suthers, D. D. (2006). Technology affordances for intersubjective
meaning-making: A research agenda for CSCL. International
networked society, or instead to offer a critical voice
Journal of Computer-Supported Collaborative Learning, 1(3),
in favor of strong relationships (Jones et al. 2006). At 315–337.
the community level, CSCL has also focused on cohe-
sive groups who share an enterprise and repertoire,
raising the question of whether “communities of prac-
tice” or “networked learning” based on weak ties is
more productive with respect to the learning of the Computer-Supported
individual participant (Jones et al. 2006). Promising Collaborative Work
topics for research in the networked society include
▶ Collaborative Learning Supported by Digital Media
identifying how the mutability and mobility of digital
artifacts can serve to recruit participants in new
social arrangements that make new forms of learning
possible, the conditions for productive entanglement
of multiple individual trajectories of participation, and Computer-Supported
how the social affordances of technologies operate over Cooperative Learning
larger spans of time and larger collections of actors
(Suthers 2006). ▶ Computer-Supported Collaborative Learning
Cross-References
▶ Asynchronous Learning Networks
▶ Collaboration Scripts Computer-Supported
▶ Collaborative Knowledge Building Cooperative Work
▶ Collaborative Learning
▶ Collaborative Learning Supported by Digital Media
▶ Collaborative Learning and Critical Thinking
▶ Online Collaborative Learning
▶ Collaborative Learning Strategies
▶ Collaborative Learning Supported by Digital Media
▶ Collective Learning
▶ Communication and Learning
▶ Online Collaborative Learning Computer-Supported
Intentional Learning
References Environment
Andriessen, J., Baker, M., & Suthers, D. (Eds.). (2003). Arguing to
▶ Online Collaborative Learning
learn: Confronting cognitions in computer-supported collaborative
learning environments. Boston: Kluwer.
Fischer, F., Kollar, I., Mandl, H., & Jaake, J. M. (2007). Scripting
computer-supported collaborative learning: Cognitive, computa-
tional and educational perspectives. New York: Springer.
Jones, C., Dirckinck-Holmfeld, L., & Lindstrom, B. (2006). Computer-Supported Learning
A relational, indirect, meso-level approach to CSCL design in
the next decade. Computer-Supported Collaborative Learning, ▶ Situated Prompts in Authentic Learning
1(1), 35–56. Environments
Concept Formation: Characteristics and Functions C 723
same planet exists both as “morning star” and as “eve- objects and events designated and denoted by words;
ning star,” and the same person exists both as “Josef they also contain a personal component which includes
Ratzinger” and as “Pope Benedict XVI.” Two linguistic emotional judgments based on subjective experiences
expressions may thus have the same meaning but and feelings. It is known that even strong emotions can
a different sense. A general outline of this conception be associated with concepts. This connotative meaning
is provided in Fig. 1. of concepts changes in the course of an individual’s
Words allow people to communicate with their development in dependence on learning experiences
surroundings and to have a conception of things and and communication with others. Parallel to the con-
occurrences in the world. However, this presupposes tinuous development of cognitive operations and semi-
that words are considered as objects of knowledge. otic functions, humans learn concepts which are more
If people use words as a means of communicating and more abstract (e.g., SOCIAL WELFARE, TOLERANCE).
their knowledge they must be able to retrieve both the
signs and meanings of the words they wish to use. The Attributes of Concepts
German philosopher Lorenz (1987) therefore empha- Concept formation begins with a determination of the
sized the contrast between factual knowledge, which is common characteristics or attributes of things, quali-
independent of concrete speech acts, and linguistic ties, and events which can then be united to form
knowledge, which is knowledge about objects as well a semantic category on the basis of these similar attri-
as about possibilities of expressing those objects in butes. However, the attributes used to form these cat-
language and communicating them. egories can vary in quantity and quality (relevance,
distinguishability). A quadrilateral, for example, has
" A single word I say
four relevant attributes: a closed shape, a plane figure,
It’s only words,
And words are all four angles, and four sides. The same four attributes
I have to take your heart away. plus two additional ones – right angles and equal sides
The Bee Gees: Words, 1968. – are used to define a square. Thus, the attributes “right
angles” and “equal sides” are relevant for the concept of
From a psychological point of view, Bruner et al. (1956)
SQUARE but not for that of QUADRILATERAL.
determined that the meaning of a concept is the result
Cognitive psychology differentiates between pri-
of the association of perceived and learned characteris-
mary and secondary attributes of concepts, depending
tics of an object with attributes stored in memory. In on whether they ascribe to objects concrete character-
other words, the psychological meaning of a concept is
istics (e.g., form, location, color, and size), functional
determined by an individual’s existing knowledge of characteristics, or characteristics based on opinions
the world. It is formed by associating information
(e.g., characterizing an object as “beautiful,” “good to
about things, qualities, and events with attributes
sit on,” etc.). Accordingly, a distinction can be made
defined in memory. These attributes do not only con-
between sensory and categorical concepts. Clearly, sen-
tain information on the qualities and characteristics of
sory concepts classify objects on the basis of concrete
attributes and are represented in memory primarily by
Thing means of these attributes, whereas categorical concepts
Denotes
are formed on the basis of non-concrete and functional
Attribute
characteristics. The formation of categorical concepts
Signifies
Relation extends to abstract concepts, which result from cog-
Concept Word nitive processing and must not correspond to any
Individual concrete object or occurrence. Nevertheless, even
Class exceedingly abstract concepts like ETERNITY or ENDLESS-
Means
NESS still may have a residual concreteness for many
Structure
people. Abstract concepts encompass not only many
Concept Formation: Characteristics and Functions. attributes of the underlying class of concepts; they are
Fig. 1 The relationship between concept and word in also often related to other concepts in the same subject
terms of three-dimensional semantics domain. This led Klix (1984) to the conclusion that the
Concept Formation: Characteristics and Functions C 725
more abstract a concept is the more relations it will them. More specifically, a subordinate concept is char-
have to other concepts. acterized by all attributes of its superordinate concept
The formation of concrete concepts and many cat- plus the attributes which characterize it and distinguish
egorical concepts is grounded in the assumption that it from the other concepts on the same level of the
attributes are separable, that is, easy to distinguish from concept hierarchy. Take for instance the concepts BIRD C
one another. However, in many cases this condition is and MAMMAL: Birds have warm blood like mammals,
not fulfilled, causing the semantic contour of a concept but they do not have mammary glands (or udders) and
to become indistinct and “blurred.” Actually, many of their offspring are not born live. But despite these
the concepts humans operate on in daily life are vague, differences, both birds and mammals possess common
and in consequence the boundaries between these con- attributes and are thus both classified as belonging to
cepts are not only indistinct but also variable. However, the TETRAPODA (vertebrates with four legs or limbs),
the less possible it is to differentiate between attributes a class which also includes reptiles and amphibians.
the more difficult concept formation becomes. The hierarchical organization of attribute con-
Another important structural feature of concept cepts correlates to a great extent with the degree of
formation has to do with the relations within a concreteness of the attributes. This also has conse-
concept. The first step in the process of associating the quences for the assignment of things, qualities, and
attributes of a concept is to establish the common and events to concepts on various hierarchical levels. Hoff-
distinguishing attributes of the objects of a domain. mann (1986) and others have shown that the first
Objects with common attributes can then be combined concept to be identified as such is the one which is
to form a class. Thus, all attribute concepts are based on characterized by both the smallest and most compre-
the one-attribute relation which ascribes certain char- hensive set of concrete attributes. This concept, which
acteristics to the objects. Examples are SUGAR – sweet, represents the “lowest common multiple” of the con-
JAGUAR – spotted, FROG – croak, DOG – bark, etc. The crete attributes of objects, is referred to as a primary
attribute relation is also used to construct semantic concept (e.g., Hoffmann 1986) and is the point of
categories by testing whether things, qualities, and departure for addressing the significant attribute classes
events can be combined to form a class on the basis of in memory.
common characteristics. This, however, presupposes From an extensional standpoint, a concept may be
that the attributes are separable. Then it is possible to defined by a class containing an undefined amount of
distinguish between several “relations within a con- objects. Many semantic classes, however, comprise only
cept,” for example (a) the contrastive characterization a single object – the MOON, the EARTH, the PRESIDENT OF
of two concepts with reference to a certain attribute THE UNITED STATES, FRANCE. Other concepts comprise
(e.g., HIGH – LOW, GIANT – DWARF, MOUNTAIN – VALLEY) and many objects (e.g., SONGBIRD = [nightingale, lark,
the comparative characterization of two concepts (e.g., robin, titmouse]), and others even comprise an infinite
SICK – INFIRM, WIND – STORM, JOG – RUN). amount of objects (e.g., RATIONAL NUMBER). Many con-
A far-reaching assumption of semantics and psy- cepts cam be characterized by examples and coun-
chology states that concepts are hierarchically orga- terexamples. Any rectangle is a good example of
nized. Two complementary aspects of concept QUADRILATERAL, but a bad example of TRIANGLE; and it is
hierarchies are emphasized in the literature: the inher- also a bad example of ANIMAL and all other concepts
itance of attributes and the intensification of attributes. which do not signify geometric forms. Finally, the
The inheritance principle, which states that a subordi- examples of a concept also vary in how open they are
nate concept always includes the attributes of its super- to sensory perception. On the one hand, there are
ordinate concept (as a more comprehensive class), can examples that one can see, hear, smell, or feel, but on
be understood as a cognitive operation of specializa- the other hand, there are concepts whose examples are
tion. The complementary operation consists in gener- not perceptible and thus also difficult to represent (e.g.,
alizing abstractions, which result in an intensification of ATOM, GENETIC CODE).
attributes. This is because superordinate concepts are Generally, a concept is defined by the attributes
formed on the basis of the conjunctive association of which all members of the semantic category it sig-
the common attributes of the concepts subordinate to nifies have in common. But Wittgenstein (1953)
726 C Concept Formation: Characteristics and Functions
demonstrated with the example of the concept GAME understanding and solving a problem; (4) concepts
that not all members of a category share all of the facilitate the learning of other concepts. The first two
same attributes. Some games, like chess or checkers, usages have to do with the classification or, as Novak
require a board, others require cards, balls, or paddles, (1998) says, the assimilation of concepts, the last two
and some (like hide-and-go-seek or guessing games) with the transfer of concepts.
do not require any equipment at all. Many games are A central aspect of the transferability of concepts is
competitions, some are not. Some people who play that humans are capable of learning how to learn con-
games do it for fun, whereas others treat them like cepts. In fact, they acquire new concepts throughout
a sport and complain about stress and pressure. This their lives and learn with time the principles of acquir-
example makes it easy to understand why Wittgenstein ing new concepts. Referring to Piaget’s seminal work on
chose to work with the term family resemblance and to the development of the concepts of numbers, room,
assume that objects are combined to form a class and time, Aebli (1987) described concept formation as
because they resemble each another and not because a structural process which can be stimulated effectively
they possess all or even most of the same attributes. by external influence (e.g., instructional methods).
Whereas Wittgenstein argued along logical principles,
some decades later Rosch was able to demonstrate in Important Scientific Research and
numerous individual experiments that many natural Open Questions
categories include members which are judged to be As said in the introductory part of this entry, concept
more typical for a category than the rest. In one exper- formation has been a central issue of theoretical and
iment, for instance, Rosch (1975) presented to subjects practical consideration since ancient times. Ros (1989/
the names of members of everyday categories (e.g., 1990) described in full detail the history of the philo-
vegetables, furniture) and asked them to rate the sophical consideration of “rationale and concept” from
items in a list according to their value for the category. Socrates to Wittgenstein. In his description of the
The results revealed that carrots, for instance, are modern concept of “concept,” Ros centers on the def-
judged to be more typical for the category VEGETABLES inition of concepts as directly accessible subjects of self-
than pumpkins but less typical than peas. Typical items consciousness that are created autonomously by the
share many attributes with the other members of the human mind. This corresponds to Locke’s understand-
same category, but only few with members of other ing of concepts (or general ideas) as templates of
categories. Correspondingly, atypical representatives of existing mental images and Leibniz’s idea of concepts
a category have only little in common with other items as capabilities to imagine “forms” as well as to Kant’s
in the same category, whereas they may have more or definition of concepts as capabilities to produce many
less attributes in common with items from other cate- optional mental representations of concrete objects in
gories. Typicality is strongly dependent on the degree of compliance with a rule. This understanding of con-
family resemblance. In psychology, the most typical cepts, which has since been modified and revised by
member of a semantic category is referred to as the Wittgenstein, is clearly the fundamental basis of mod-
prototype (see Eckes 1991). It is assumed that the pro- ern philosophy of language and psychology. Actually,
totype is at the center of a category, whereas atypical apart from behaviorism all new movements of twenti-
members are at its margins. The prototype serves as eth century psychology referred more or less explicitly
a point of reference for the classification of objects in to Kant and his followers in discussing the formation of
a category. concepts. This can be demonstrated by the example of
the Würzburg school of psychology (e.g., Ach 1921)
Transferability of Concepts and its focus on the so-called imageless thoughts (i.e.,
Once individuals have learned a concept, there are conscious sets, awarenesses, and thoughts). Addition-
several ways in which they can use it in other situations: ally, developmental psychology has focused on concept
(1) New things, qualities, or events can be assigned to formation since Piaget’s seminal work on the formation
a concept; (2) concepts can be ordered in a hierarchy of the concepts of number, space, time, etc., in children
through the identification of superordinate or subor- (Wetzel 1980). Actually, Piaget’s epistemology and
dinate relations; (3) a concept can be used as an aid in early research on cognitive development initiated an
Concept Formation: Characteristics and Functions C 727
abundance of empirical studies on childhood concept learning and artificial intelligence but in particular also
learning and the origin and evolution of everyday con- on educational approaches to concept learning. As
cepts (see, e.g., Novak 1998). a consequence, there have been instructional princi-
In addition, Piaget also had a strong influence on ples pertaining to concepts in the literature of educa-
the ascent of cognitive psychology in the 1950s. For tional psychology for decades (see, e.g., Aebli 1987; C
instance, concept formation as semantic classification Klausmeier and Ripple 1971). However, probably the
was at the core of Bruner’s work (e.g., Bruner et al. most successful instructional application of research
1956) whereas Ausubel and others focused on the hier- on concept formation was the idea of visualizing
archical organization of concepts (see the entries on relations within and between concepts by means of
“▶ Assimilation Theory of Learning” and “▶ Meaning- maps and graphs (Novak 1998). There are hundreds of
ful Verbal Learning”). Since the paradigm shift known studies that demonstrate the effectiveness of concept
as the “cognitive revolution,” concept formation has mapping as a tool for structuring and assessing
become an important research topic throughout the domain-specific knowledge as well as for learning
world, and especially again in Europe. This can be new concepts.
illustrated by the research of Aebli (1980), a Swiss psy-
chologist and student of Piaget, and East German psy- Cross-References
chologists such as Hoffmann (1986) and Klix (1984). ▶ Abstract Concept Learning in Animals
Altogether, it can be said that concept formation is ▶ Bruner, Jerome S.
probably the most important branch of cognitive psy- ▶ Categorical Learning/Category Learning
chology ever. In consequence, it has also become an ▶ Categorical Representation
important topic for cognitive science and informatics ▶ Concept Maps
with its emphasis on machine learning and artificial ▶ Concept Similarity in Multidisciplinary Learning
intelligence (cf. Brodie et al. 1984; Sowa 1984). ▶ Conceptual Change
Machine learning refers to cognitive psychology, ▶ Conceptual Clustering
often especially to Bruner et al. (1956), and focuses ▶ Language Acquisition and Development
on the development of computational approaches to ▶ Meaningful Verbal Learning
concept formation and learning. Machine learning ▶ Prototype Learning Systems
may apply different approaches depending on how ▶ Psycholinguistics and Learning
concept formation is to be modeled. Discriminative ▶ Word Learning
approaches do not entail an explicit model of a
concept but only a procedure for discriminating References
between members and nonmembers of mutually Ach, N. (1921). Über die Begriffsbildung [On concept formation].
exclusive contrasting categories, whereas distribu- Königsberg: Buchner.
Aebli, H. (1980). Denken: das Ordnen des Tuns. Band I: Kognitive
tional approaches operate with a model of a concept
Aspekte der Handlungstheorie. Stuttgart: Klett-Cotta.
as a probability distribution and classify new instances Aebli, H. (1987). Zwölf Grundformen des Lehrens. Eine Allgemeine
as members of a category if their estimated probability Didaktik auf psychologischer Grundlage (3. Aufl.). Stuttgart:
of family resemblance exceeds a threshold. Distribu- Klett Cotta.
tional approaches are regularly based on Bayesian Brodie, M. L., Mylopoulos, J., & Schmidt, J. W. (Eds.). (1984). On
learning and include “novelty detection” techniques conceptual modelling. Perspectives from artificial intelligence, data-
bases, and programming languages. New York: Springer.
which operate not only with positive examples but Bruner, J. A., Goodnow, J. S., & Austin, G. J. (1956). A study of
also with negative examples of principled generaliza- thinking. New York: Wiley.
tion. The correspondences between approaches of Chater, N., Tenenbaum, J. B., & Yuille, A. (2006). Probabilistic models
cognitive psychology and machine learning are obvi- of cognition: Conceptual foundations. Trends in Cognitive Sci-
ous, and it can be said that computational approaches ence, 10(7), 287–291.
Eckes, T. (1991). Psychologie der Begriffe. Strukturen des Wissens und
attempt to close the gap between human and machine
Prozesse der Kategorisierung. Göttingen: Hogrefe (Psychology of
concept learning (Chater et al. 2006). concepts).
Cognitive psychology and its research on concept Frege, G. (1892/1980). On sense and meaning (M. Black, Trans.). In
learning not only had a strong influence on machine P. Geach & M. Black (Eds.), Translations from the philosophical
728 C Concept Learning
writings of Gottlob Frege (3rd ed., pp. 56–78). Oxford: Blackwell. concept horse refers to a largely domesticated mammal
(Original work published 1892). species with distinctive and well-recognized head and
Hoffmann, J. (1986). Die Welt der Begriffe. Psychologische
body shapes). Other concepts are more abstract and
Untersuchungen zur Organisation des menschlichen Wissens.
Berlin: VEB Deutscher Verlag der Wissenschaften (The world difficult to pin down in terms of specific, observable
of concepts). characteristics (e.g., the concept cousin is defined by
Klausmeier, H. J., & Ripple, R. E. (1971). Learning and human abilities. a particular familial relationship; the concept freedom is
Educational psychology (3rd ed.). New York: Harper & Row. defined by the lack of physical and social constraints on
Klix, F. (Ed.). (1984). Gedächtnis, Wissen, Wissensnutzung. Berlin:
one’s behavior). On average, concrete concepts are
VEB Deutscher Verlag der Wissenschaften (Memory, knowledge,
knowledge use).
learned more quickly and easily than abstract ones.
Lorenz, K. (1987). Weltwissen und Sprachwissen. Ihre Rekonstruktion
in Dialogsituationen. In J. Engelkamp, K. Lorenz, & B. Sandig Theoretical Background
(Hrsg.), Wissensrepräsentation und Wissensaustausch (pp. 35–45). Some early behaviorists attempted to explain concept
St. Ingbert: Röhrig Verlag (World knowledge and language learning in terms of the strengthening of certain of S–R
knowledge).
associations; for example, children will form the con-
Novak, J. D. (1998). Learning, creating, and using knowledge: Concept
maps as facilitative tools in schools and corporations. Mahwah: cept red if they are consistently reinforced for saying
Lawrence Erlbaum. “red” in response to red objects. Although this expla-
Ros, A. (1989/1990). Begründung und Begriff (3 Vols.). Felix Meiner: nation might at least partly explain concept learning
Hamburg. in very young children and nonhuman species, it did
Rosch, E. (1975). Cognitive representations of semantic categories.
not hold up to close scrutiny in laboratory research
Journal of Experimental Psychology: General, 104, 192–253.
Sowa, J. F. (1984). Conceptual structures. Information processing in
with older children and adult humans, who appear to
mind and machine. Reading: Addison-Wesley. mediate their overt responses to particular classes of
Wetzel, F. G. (1980). Kognitive Psychologie. Eine Einführung in objects with internal, mental responses (e.g., Kendler
die Psychologie der kognitiven Strukturen von Jean Piaget. et al. 1962).
Weinheim: Beltz. Cognitive psychologists have offered several alter-
Wittgenstein, L. (1953). Philosophical investigations. (G. E. M.
native explanations regarding the nature of concepts
Anscombe, Trans.). Oxford: Basil Blackwell.
and concept learning, at least for human learners.
Further Reading Perhaps the first prominent theory grounded in a cog-
Rosch, E. (1973). Natural categories. Cognitive Psychology, 7, 532–547.
nitively oriented framework was one involving hypoth-
esis testing (Bruner et al. 1956). In particular, when
confronted with a label that is believed to represent
an unknown class of objects, a learner forms and tests
Concept Learning a series of hypotheses regarding features that might
possibly define the concept (e.g., color, shape) either
JEANNE ELLIS ORMROD singly or in combination. However, laboratory studies
School of Psychological Sciences (Emerita), University supporting this perspective were highly contrived
of Northern Colorado, Greeley, CO, USA and unrepresentative of real-world concept-learning
situations.
Several other cognitively oriented theories do appear
Synonyms to have some relevance to real-world concept-learning
Categorization; Classification; Concept formation situations. For example, Eleanor Rosch (e.g., Rosch
1978) has proposed that many concepts are formed,
Definition at least in part, by acquiring mental prototypes that
A ▶ concept is a mental representation of a class of capture the features of a typical, average member of
objects or events that share one or more common a concept (e.g., a sparrow-like creature might be a good
properties. Some concepts are fairly concrete, in that prototype of the concept bird; a penguin or ostrich
the objects or events they encompass share certain would be less representative of birds in general).
easily detectable physical features (e.g., the concept Other theorists (e.g., Ross and Spalding 1994) have
red refers to a certain range of light wavelengths, the suggested that mental representations of many
Concept Learning in Pigeons C 729
concepts may be based on a variety of examples, or Several factors have been found to facilitate concept
exemplars, that reflect the variability that concept learning in instructional settings. Explicit definitions
members may show (e.g., the concept fruit might be that identify critical features of concept members are
mentally represented by such diverse exemplars as helpful, as are visual or other modality-specific repre-
apples, bananas, and grapes) and can include atypical sentations that highlight those features. Illustrative C
concept members (e.g., although most mammals give examples are beneficial as well, but it is also important
birth to live young, platypuses and a few other mammal to show non-examples that are “near misses” to cate-
species lay eggs). gory membership (e.g., a spider is not an insect because
When concepts are not easily represented by pro- it has eight legs instead of six).
totypes or exemplars, a mental feature list of category Much of the existing research on concept learning
members may be involved (e.g., Ward et al. 1990). In has involved studies with adults (or in some cases
particular, learning a concept may involve learning the nonhuman animals) learning artificial concepts in lab-
one or more features that characterize many or all oratory settings. Such research is helpful in illuminat-
instances of the concept, along with probability esti- ing cognitive processes that might underlie concept
mates for each feature. Identifying an object or event as learning. However, the extent to which the principles
an example of a particular concept, then, is a matter of derived from such research can be generalized to more
determining whether the object or event includes natural concept-learning phenomena has yet to be
enough of these features to qualify. determined.
The various theoretical explanations just described
are not necessarily mutually exclusive (Ormrod 2008). Cross-References
Quite possibly, mental representations of concepts ▶ Abstract Concept Learning in Animals
include (1) prototypes that capture a typical, average ▶ Categorical Learning
concept member; (2) exemplars that reflect variability ▶ Categorical Representation
among concept members; (3) a set of features that ▶ Concept Formation: Characteristics and Functions
facilitate identification of new examples; and (4) one ▶ Meaningful Verbal Learning
or more automatic responses to concept members. ▶ Prototype Learning Systems
Hypothesis testing may come into play in situations ▶ Schema(s)
where a learner is given a concept label and a set of References
examples and non-examples but no explicit definition. Bruner, J. S., Goodnow, J., & Austin, G. (1956). A study of thinking.
New York: Wiley.
Important Scientific Research and Kendler, T. S., Kendler, H. H., & Learnard, B. (1962). Mediated
Open Questions responses to size and brightness as a function of age. American
Journal of Psychology, 75, 571–586.
Researchers have observed the ability – and, some
Ormrod, J. E. (2008). Human learning (5th ed.). Upper Saddle River,
might say, a natural tendency – to categorize objects NJ: Pearson.
and events in human infants as young as three months Rosch, E. H. (1978). Principles of categorization. In E. Rosch &
old, and also in several other mammal species. Acqui- B. Lloyd (Eds.), Cognition and categorization (pp. 27–48).
sition of any particular concept may occur over Hillsdale, NJ: Erlbaum.
a period of time, with learners sometimes initially Ross, B. H., & Spalding, T. L. (1994). Concepts and categories. In
R. J. Sternberg (Ed.), Handbook of perception and cognition (Vol.
showing under-generalization (i.e., they fail to recog-
12, pp. 119–148). New York: Academic.
nize all concept members) or over-generalization (i.e., Ward, T. B., Vela, E., & Haas, S. D. (1990). Children and adults learn
they mistakenly include nonmembers as being exam- family-resemblance categories analytically. Child Development,
ples of the concept) before fully mastering the concept. 61, 593–605.
Also, learners may sometimes mistakenly identify the
essential features of concept members; for example,
many young children restrict their understanding of
the concept animal to creatures with four legs and Concept Learning in Pigeons
a lot of fur, thus disqualifying fish, insects, and people
as animals. ▶ Categorical Learning in Pigeons
730 C Concept Map
Theoretical Background
Concepts and relations can be conceptualized as key-
Concept Maps constructs of knowledge and thought. Epistemological
foundations of concept maps can be found within the
JOHANNES GURLITT realm of graphical knowledge representation that are
Department of Educational Science, University of based on logic and the study of ontology. While logic
Freiburg, Freiburg, Germany provides the formal structure and rules of inference,
ontology deals with questions about entities relevant
for the respective domain and how such entities can be
Synonyms grouped, related within a hierarchy, and subdivided
Conceptual maps; Knowledge maps according to similarities and differences (Sowa 2000;
see also ▶ Ontology and Semantic Web in this encyclo-
Definition pedia). The earliest known semantic network appeared
Concepts can be defined as objects, events, situations, in a commentary on Aristotle’s categories, by the
or properties that possess common critical attributes philosopher Porphyry in the third century A.D. (see
and are represented by icons or symbols, such as key Sowa 2000).
words (Ausubel 2000). Concept maps are external From an educational point of view, concept map-
network structures that allow two-dimensional, spatial ping is based on the assimilation theory of David
processing along preconstructed or to-be-constructed Ausubel (see Novak and Gowin 1984; Ausubel 2000).
connecting lines. In its simplest form, a concept map In short, assimilation theory points out that all new
would consist out of two concepts and a linking word information is linked to relevant, preexisting aspects of
for example “cats – are ! mammals.” Although orig- the learner’s cognitive structure and that both, the
inally conceptualized as hierarchical structures (Novak newly acquired and the preexisting structure are mod-
and Gowin 1984), current conceptualizations use a ified in the process. The assimilation of new informa-
broader scope that is the basis for the following defini- tion includes establishing relations between same-level
tion: Concept maps provide an external network-like concepts (combinatorial learning), generalization pro-
representation of knowledge structures. They consist of cesses creating new subsumers (superordinate learn-
spatially grouped nodes with key words representing ing), and anchoring a new idea below a higher-level
Concept Maps C 731
anchoring idea (subsumption learning). According to learners being overwhelmed by the unfamiliar repre-
Ausubels hierarchical view of knowledge, these pro- sentation or confused by the tasks to be carried out.
cesses of concept assimilation are perceived as the A second limitation is ambiguity: Concepts are usually
major learning activities of school children and adults represented by one or two key words only. Hence, the
(Ausubel 2000). Related to these processes described by justifications for certain connections may not be explicit. C
Ausubel, hypotheses about knowledge representation This limits external judgments, such as scoring or grad-
distinguished between inter-concept relations and ing procedures that are used to assess the learners’ prior
intra-concept relations (e.g., Klix 1980). Inter-concept knowledge. Negative effects of these limitations may
relations are relations between concepts and events be softened or overcome through self-assessment and
that have been directly observed and experienced for prestructuring: In self-assessments learners realize which
example “the boat is in the water.” Intra-concept rela- concepts or relationships they know or do not know yet.
tions are based on common or distinguishable features The “lost in mapping space” phenomenon may be
within the concepts that are not directly extractable reduced when the task is prestructured to a substantial
from experience or observed but have to be inferred, degree. However, prestructuring the task too much bares
for example, by comparative processes such as “high is the risk of superficial processing. When provided with
the opposite of low” or inferences such as “a hammer is a completely worked-out map, learners may not engage
tool.” These considerations lead to the still-debated in meaningful learning and rather process the mate-
question whether and which relations are pre-stored rial in a superficial mode, which may lead to rote
in semantic memory or have to be computed dynam- learning. Therefore it seems appropriate to design
ically. Based on empirical research, Klix hypothesized mapping tasks that leave certain achievable but chal-
that, in general, inter-concept relations are stored lenging tasks, targeted at deep-level cognitive and
directly in memory, while intra-concept relations are metacognitive processes.
not stored directly in memory but are derived or
generated dynamically depending on the respective Important Scientific Research and
task demand. Thus, although a direct relation between Open Questions
external and internal representations is naı̈ve, these O’Donnell et al. (2002) summarize that concept maps
considerations about internal processes and human facilitate the recall of central ideas, benefit especially
memory lead to the question how specific affordances those learners with low verbal abilities and low prior
of concept maps may trigger or facilitate internal knowledge, and facilitate cooperative learning. Further-
processes. more, they are more effective for learning when struc-
Concept maps focus on the visualization of key tured according to Gestalt principles (e.g., the use of
concepts and key relationships which makes them color and shapes to show similarity or groupings to
potentially valuable tools for planning, learning, show proximity). The meta-analysis from Nesbit and
and (self-)assessment. When used for planning activi- Adesope (2006) showed a small effect in favor of study-
ties, concept maps allow an overview and the detection ing maps compared to studying text, a small effect in
of the “red line” running through different topics, favor of studying maps compared to studying outlines
steps, or key concepts. In learning settings, concept or lists, and a small effect in favor of constructing maps
mapping can facilitate organization and elaboration compared to constructing text or outlines.
processes leading eventually to the construction of The benefits of concept maps outlined above raise
high-level schemas. For assessment, concept maps pro- the questions how concept maps should be used
vide the possibility to tap into a learner’s cognitive for learning and whether elicited cognitive and
structure and externalize, for both, the learner and metacognitive processes are different for different map-
the teacher, what the learner already knows and does ping tasks. With respect to the task, concept maps may
not know. However, it is important to keep in mind be created entirely by the student, or instructors can
two limitations of concept maps. First, many learners prepare incomplete maps that require learners to
are initially not familiar with this representation and perform specific activities, such as filling in some
therefore experience a “lost in the mapping space” nodes or labeling links. Empirical research indicates
phenomenon. This can be described as a feeling of that different mapping tasks lead to different cognitive
732 C Concept Similarity
Cross-References Definition
▶ Advance Organizer A common understanding can be regarded as the reflec-
▶ Learning Strategies tion of a shared knowledge structure of a team (Novak
▶ Ontology and Semantic Web and Gowin 1984). A knowledge structure on the other
hand can be characterized as an elaborated and highly
interconnected framework of related concepts (Mintzes
References et al. 1997). Although there are many different attempts
Ausubel, D. P. (2000). The acquisition and retention of knowledge: to capture knowledge structures, concept mapping is
A cognitive view. Boston: Kluwer.
regarded as a more direct approach and thus has been
Gurlitt, J., & Renkl, A. (2010). Prior knowledge activation: how
different concept mapping tasks lead to substantial differences often used to elicit and represent knowledge structures
in cognitive processes, learning outcomes, and perceived self- (Ruiz-Primo 2004).
efficacy. Instructional Science, 38, 417–433. Concept maps (CMAPs), as a collection of con-
Klix, F. (1980). On structure and function of semantic memory. In cepts and interconnections among concepts, make
F. Klix & J. Hoffmann (Eds.), Cognition and memory (pp. 11–25).
knowledge structures assessable. A CMAP is a graph
Amsterdam: North-Holland.
Nesbit, J. C., & Adesope, O. O. (2006). Learning with concept and
consisting of nodes representing concepts and labeled
knowledge maps: A meta-analysis. Review of Educational lines denoting relationships between a pair of nodes.
Research, 76, 413–448. One important characteristic of CMAPs is the expres-
Novak, J. D., & Gowin, D. B. (1984). Learning how to learn. New York: sion of propositions, which is represented by using
Cambridge University Press. two or more concepts connected by linking words
O’Donnell, A. M., Dansereau, D. F., & Hall, R. H. (2002). Knowledge
or phrases to convey meaning. Thus, a CMAP can be
maps as scaffolds for cognitive processing. Educational Psychol-
ogy Review, 14, 71–86. described as a set of concepts and a set of propositions;
Sowa, J. F. (2000). Knowledge representation: logical, philosophical, and accordingly, the similarity of concepts can be deter-
computational foundations. Boston: MIT Press. mined as a function of the propositions.
Concept Similarity in Multidisciplinary Learning C 733
where cij = jth concept in the ith CMAP; Fij = set of Whether concepts in Sk are identical, similar, or
propositions associated with cij; mi = number of con- dissimilar is determined by the propositions in
cepts in CMAPi; and Fij is defined as the shared, overlapping, or distinctive set. If both
the overlapping and distinctive sets are empty and the
Fij ¼ fðpijl Þj1 i n; 1 j mi ; l
1g ð3Þ
shared set is not, then the concepts are identical. Con-
where pijl = lth proposition associated with the jth cepts are dissimilar if both the shared and overlapping
concept in the ith CMAP. sets are empty and the distinctive set is not. For all
To compare concepts from different CMAPs, this other situations, concepts are considered similar.
study defines a set, Sk, which is a collection of con- In the following, formal definitions are given.
cepts, as well as their associated propositions, selected Definition 1: A set of concepts in Sk is identical, if
from different CMAPs for similarity comparisons, and only if, for any Fijk
i.e., no two concepts in Sk are from the same CMAP. \
n
In this way, the comparison of a set of CMAPs is Fijk ¼ Fijk ð9Þ
transformed into the similarity analysis of a set of Sk, i¼1
where
w = total number of concept sets to be compared Definition 3: Concepts in Sk are similar, if and only
or S . if, (1) there exists at least one Fijk (1 i n, 1 j mi),
An Sk merely regroups the concepts and proposi- such that
tions contained by CMAP. Thus
\
n
Thus, the similarity measure, sim(Sk), can be defined immediately associated with concepts to be compared.
as follows: Sometimes, propositions that are not directly associ-
ated with the concepts may also have an impact on
1. Condition 1: when distinctive = Ø and overlapping =
similarity analyses. Such an impact is not considered
Ø, sim(Sk) = 1
2. Condition 2: when shared = Ø and overlapping =
in the proposed similarity measure. C
Ø, sim(Sk) = 0 and
3. Condition 3: when overlapping = Ø or
Cross-References
▶ Concept Formation
overlapping = Ø but Distinctive 6¼ Ø and
▶ Concept Mapping
shared 6¼ Ø
▶ Knowledge Representation
Yn ^
i Ni
simðSk Þ ¼ i ð13Þ References
i¼1
n Nk
Cooke, N. J., Kiekel, P. A., Salas, E., & Sout, R. (2003). Measuring
team knowledge: A window to the cognitive underpinnings of
Where n = number of concepts, CMAPs, or stu-
team performance. Group Dynamics: Theory, Research and Prac-
dents; i = number of concepts that a proposition tice, 7(3), 179–199.
belongs to; Ni = number of propositions that are shared Fruchter, R. (1999). A/E/C teamwork: A collaborative design and
by i concepts (2 i n 1); and Nk = total number of learning space. Journal of Computational Civic Engineering,
propositions associated with Sk. 13(4), 261–269.
Fruchter, R., and Luth, G. P. (2004). ThinkTank – A web-based
Once the similarity of concepts is obtained, this
collaboration tool. Proceedings of the ASCE Structures Conference.
study uses an average method to measure the similarity Reston, VA: ASCE.
of CMAPs by aggregating the results of concept simi- Mintzes, J. J., Wandersee, J. H., & Novak, J. D. (1997). Meaningful
larity analyses. In other words, the similarity of CMAPs learning in science: The human constructivist perspective. Hand-
is proportional to the similarity of concepts in CMAPs. book of academic learning (Series in Educational Psychology,
After a concept analysis, a similarity value is derived for pp. 405–447). Orlando, FL: Academic Press.
Novak, J. D., & Gowin, D. B. (1984). Learning how to learn. Cam-
each Sk; thus the similarity of CMAPs is
bridge: Cambridge University Press.
structures (diSessa 2006). Empirical evidence to date is from the more familiar domain to the other. This could
inconclusive with respect to which of the two views lead to conceptual restructuring in the target domain.
is more adequate. The two perspectives agree on the Closely related to analogies, the role of providing
importance of prior knowledge for subsequent learning adequate models or external representations has been
and on conceptual understanding as based on a com- shown as important in promoting conceptual change, C
plex system of knowledge rather than on single and because these models and representation can be used to
unitary ideas. clarify aspects of a scientific explanation that are not
In addition to describing conceptual change in apparent in other models.
learners, research on conceptual change also investi- Yet another implication of conceptual change
gated how these processes can be influenced by means research relates to the curriculum and the way in
of instructional interventions (for an overview, see which it is organized. If certain concepts in science
Mason 2001). One central idea is that to achieve con- and mathematics are particularly difficult and give
ceptual change, a cognitive conflict should be elicited in rise to misconceptions, it may be more profitable
learners, by confronting them with information that to focus more deeply on a limited number of topics
contradicts their current state of knowledge. However, rather than superficially dealing with many topics.
several conditions have to be met for a cognitive con- Moreover, the order in which concepts are dealt with
flict to be meaningful. Learners have to be motivated should be carefully considered in order to avoid certain
to process the anomalous information, they need misconceptions.
sufficient prior knowledge to understand the anoma-
lous character of the new information, their episte- Important Scientific Research and
mological beliefs about the subject matter or about Open Questions
learning and teaching may hinder a revision, and they In order to analyze the content and structure of
need adequate reasoning abilities to detect conflict learners’ conceptual knowledge in scientific domains
and revise existing knowledge. Despite all these poten- and how these structures change over time, conceptual
tial difficulties, cognitive conflict is still acknowledged change researchers typically conduct interviews. For
as an important condition which can lead to concep- example, Vosniadou and Brewer (1992) asked various
tual change. first-, third-, and fifth-graders questions about the
There are several ways in which cognitive conflict shape of the earth, such as “Can you draw a picture of
can be achieved through instruction. One way is to use the earth?” or “If you walked for many days in a straight
refutational texts, which directly explain common mis- line, where would you end up?” They categorized
conceptions and why they are wrong. Another way is children’s answers as indicating one of six alternative
through peer collaboration and discussion, since social mental models of the earth. There was a clear age trend
interaction with peers may promote learners’ aware- leading away from more naı̈ve conceptions (e.g., the
ness of their own beliefs and therefore of a possible earth as a flat square) over several conceptions mixing
conflict with new information. The common denomi- naı̈ve and scientific ideas, toward more scientifically
nator is that these approaches attempt to develop correct concepts (the earth as a sphere) with increasing
learners’ metaconceptual awareness: Learners are not age. The answer patterns could be interpreted in terms
always aware of their presuppositions and beliefs that of conflicts between new information about the earth
constrain their learning, and when they are, they do not as a sphere and children’s prior knowledge that the
always understand their theoretical or contradictory ground they stand on appears to be flat and that objects
nature, or that they are open to falsification. fall from underside of a sphere in everyday life. This
A further important approach is the use of analo- study is paradigmatic for many subsequent studies with
gies. Analogical reasoning, in particular, cross-domain a similar methodology, that is, interviews with children
mapping, has been shown to play a major role in of different ages about their physics concepts.
restructuring learners’ existing knowledge. This is Although interview methods are still used in
because the comparison between two domains may most studies on conceptual change, there is a growing
highlight their common features and reveal unnoticed awareness of the importance of complementing them
commonalities, and foster the projection of inferences by alternative approaches. Interviews yield only very
738 C Conceptual Clustering
Conceptual Configurations
Conceptual Framework
▶ Representations, Presentations, and Conceptual
Schemas ▶ Advance Organizer
F T F F T T
not-red square incorrect not-red square correct
Conditional Reasoning by F F F F F T
Nonhuman Animals not-red not-square incorrect not-red not-square correct
Conditional Reasoning by Nonhuman Animals. Fig. 1 Correct and incorrect assignments according to conjunctive,
disjunctive, conditional, and biconditional rules when Red and square are focal attributes
partitions are red objects that are not-square; no con- author (Bourne 1970) relied partly on the subjects’
ditions are specified for being incorrect when p is verbal explanations to confirm how they had reasoned.
not-red. In Bourne’s (1970) research, subjects had to It is hoped that one result of the present article will be
infer which truth-table was applicable based on exper- to prevent future researchers from misinterpreting or
imenter feedback, such as, saying “correct” or “incor- misrepresenting, either inadvertently or intentionally,
rect” according to whether the discriminanda were the results of typical conditional-discrimination, rule-
being partitioned consistently with a given truth- learning research using nonhuman animals.
table’s contingencies. The illustration also shows
how discriminanda must be partitioned according Important Scientific Research and
to conjunctive, disjunctive, conditional, or bicondi- Open Questions
tional truth-tables when red and square are the focal The typical conditional learning task used with
attributes. nonhuman animals involves two successively presented
There is an extensive history of investigating “con- discriminanda, represented here as A and B, only one of
ditional discrimination learning,” “conditional rule which is presented on a given trial, and two simulta-
learning,” “if–then rule learning,” etc., by nonhuman neously presented discriminanda, represented here as
animals using various procedures, and often it is stated X and Y, which appear on every trial. A or B serves as an
or implied that the animals had demonstrated condi- associative cue to select either X or Y. It is tempting to
tional reasoning corresponding to forms such as, “if p, describe and conceptualize such tasks, as many inves-
then q.” However, this article questions whether there tigators have done, as embodying conditional reason-
has ever been a valid demonstration of conditional ing such as: “If A, then X and if B, then Y.”
reasoning by nonhuman animals. Typically, relatively few discriminanda are used and
Previous investigators used methods that either they are presented more than once. Repeated presenta-
(a) confounded conditional reasoning with the possi- tions make it likely that the relatively few specific con-
bility of rote-memorization or (b) confounded the figurations afforded by the discriminanda might be
possibility of conditional reasoning with conjunctive learned by rote-memorization. As others have noted,
reasoning. The only nonverbal procedure of which I am such configuration learning is confounded with the
aware that might be used to show unequivocal condi- possibility that the animals used conditional reasoning.
tional reasoning by an animal was developed for use However, such confounding prevents such studies
with humans. However, that experiment appears to be from providing conclusive evidence for conditional
impractically difficult for nonhuman animals, and its reasoning by animals. Even if specific configuration
744 C Conditional Reasoning by Nonhuman Animals
learning is precluded, there remains a fundamental object-member of the difference-pair that was closest
problem that all known experiments using animals to the center door.
have confounded the possibility of conjunctive with In the final stage of training, the center door was
conditional reasoning. raised to expose either a triangle or a heptagon; then,
There are three basic ways to avoid specific config- it was closed to cover the triangle or heptagon before
uration learning: (a) use exemplars from conceptual the outer doors were raised to expose the same and
categories for the successive discriminanda, (b) use different pairs of objects. Intervals between closing the
exemplars from conceptual categories for the simul- center door and concurrently raising the outer doors
taneous discriminanda, or (c) use exemplars from were increased systematically. The best performing
conceptual categories for both the successive and monkey met a stringent criterion of correct responding
simultaneous discriminanda. Burdyn and Thomas’s (13 of 15 correct on 15 triangle-same trials and 13 of
(1984) investigation will illustrate both the use of con- 15 correct on 15 heptagon-different trials within a 30-
ceptual categories as discriminanda and how conjunc- trials session) with a 16 s. interval. Therefore, when the
tive and conditional reasoning are confounded. successive cues were visually absent, “triangularity”
Burdyn and Thomas (1984) used exemplars of the and “heptagonality” had to be retained symbolically
conceptual categories “same” and “different” as the in working memory as cues for “same” and “different,”
simultaneous discriminanda; an exemplar of “same” respectively.
was an identical pair of objects and an exemplar of It is tempting to conceptualize the monkeys’ suc-
“different” was a nonidentical pair of objects. New cessful performances as conditional reasoning which
pairs of objects were used on each trial in the con- might be expressed as “if triangle, then same” and “if
ceptual category phases of the testing which pre- heptagon, then different.” However, Burdyn and
cluded the monkeys from memorizing specific Thomas realized that they could not conclude that
discriminanda and reinforcement associations. The unequivocally, because it was also possible that the
successive discriminanda involved the conceptual cat- monkeys were reasoning conjunctively such as “triangle
egories “triangularity” and “heptagonality” which were and same” and “heptagon and different.” This general
represented by using 120 discriminable triangles and interpretational problem appears to have affected all
120 discriminable heptagons. Such a large number of other so-called conditional rule-learning studies in ani-
discriminanda together with trial-unique exemplars of mals. It should be noted also that most animal studies
“same” and “different” made it unlikely that the mon- have not used conceptual-category discriminanda
keys memorized and associated specific triangles and which means their subjects might have memorized
heptagons with same and different. the specific configurations associated with the
An apparatus with three guillotine doors was used. discriminanda-reinforcement contingencies.
During most of the training, all three doors were raised Bourne (1970) also realized that his subjects might
and lowered concurrently. On a given trial, (a) either have performed on some basis other than implementing
a triangle or a heptagon appeared as the center door the requirements of the appropriate truth-table, but he
was raised, (b) a pair of identical objects appeared as was able to determine through a series of transfer exper-
a result of raising one of the outer doors, and (c) a pair iments that his subjects had learned the rules. Some of
of nonidentical objects appeared as a result of raising the transfer experiments involved the experimenter and
the other outer door; the choice of triangle or heptagon the subjects discussing the applicable rule. It is unlikely
and the left-right locations of the same and different that such verbal validation will be available to animal
pairs were determined quasi-randomly for each trial. researchers, and it remains to be seen whether animals
When a triangle was presented, the correct response will show the kind of perfect or near-perfect transfer of
was to displace the object-member of the same-pair training that is necessary otherwise to confirm that the
that was closest to the center door; doing so revealed subject reasoned conditionally. By “near-perfect,” it is
a food well with a bit of fruit reinforcement beneath meant that there must be so few mistakes that the subject
the object. When a heptagon was presented, the likely could not have memorized specific discriminanda
correct response, similarly reinforced, was to the and reinforcement relationships.
Conditioned Avoidance C 745
A minimum of four trials is necessary merely reasoning based on methods that embody truth-
to present the minimal information to show which functional logic.
rule is operating, namely, one trial each to manifest
each row contingency in a given truth-table. After Cross-References
being trained on a succession of problems based on ▶ Abstract Concept Learning in Animals C
the same logical operation, Bourne’s human subjects ▶ Animal Intelligence
learned to use the four informational trials to attain ▶ Associative Learning
thereafter perfect or near-perfect performances on new ▶ Categorical Learning
problems. Presumably, this could be done only if the ▶ Complex Learning
subjects had inferred correctly and followed the appro- ▶ Complex Problem Solving
priate truth-table. ▶ Concept Learning
Future animal research on conditional reasoning ▶ Conditional Reasoning
can and must be improved by precluding the possibility ▶ Conditions of Learning
of rote-memorization of the discriminanda or con- ▶ Discrimination Learning Model
figurations of the discriminanda. This is best done by ▶ Evolution of Learning
using conceptual-category discriminanda. Response ▶ Inductive Reasoning
contingencies that allow the subject to affirm or negate ▶ Laboratory Learning
exemplars might be helpful. If animal experiments ▶ Logical Reasoning and Learning
are based on Bourne’s procedure, they would involve ▶ Measures of Similarity
reinforcing an animal’s responses that correctly affirmed ▶ Nature of Creativity
or negated each discriminandum in accordance with the ▶ Problem Solving
applicable truth-table. A series of problems should be ▶ Rote Memorization
administered according to a single operation, until,
following the administration of the four mandatory, References
Bailey, A. M., McDaniel, W. F., & Thomas, R. K. (2007). Approaches
informational trials on new problems, the animal con-
to the study of higher cognitive functions related to creativity in
tinued with perfect or near-perfect performances,
nonhuman animals. Methods, 42, 3–14.
or until it seemed unlikely that the animal would be Bourne, L. E., Jr. (1970). Knowing and using concepts. Psychological
able to attain such performances. If perfect or near- Review, 77, 546–556.
perfect performances were seen on new problems, it Braine, M. D. S., & O’ Brien, D. P. (Eds.). (1998). Mental logic.
should be reasonable to attribute the use of the condi- Mahwah: Lawrence Erlbaum.
Burdyn, L. E., Jr., & Thomas, R. K. (1984). Conditional discrimina-
tional reasoning to the animal (or conjunctive rea-
tion with conceptual simultaneous and successive cues in the
soning, etc., depending upon which truth-table was squirrel monkey (Saimiri sciureus). Journal of Comparative Psy-
being applied). chology, 98, 405–413.
This article would be incomplete without acknowl- Thomas, R. K. (1980). Evolution of intelligence: An approach to its
edging that some scholars have tried to reconcile stan- assessment. Brain, Behavior and Evolution, 17, 452–474.
dard logic with what some refer to as “natural” or
“mental logic” (e.g., Braine and O’ Brien 1998). Such
logic is said to apply to cases of reasoning that reflect
genuine, “if–then” conditional reasoning without Conditional Rule Learning
using procedures that fulfill the requirements of the
truth-table for the conditional. However, consideration ▶ Conditional Reasoning by Nonhuman Animals
of natural versus standard logic has not revealed how
the methods associated with natural logic will enable
us to design experiments to distinguish how animals
may have reasoned. Thus, it appears that the most
conservative and justifiable approach is to continue to
Conditioned Avoidance
attempt to investigate animals’ use of the conditional ▶ Aversive Learning in Drosophila melanogaster
746 C Conditioned Inhibition
The proper assessment of conditioned inhibition for responding, both tests create a nonzero response
has long been a contentious issue. Some of this contro- baseline which allows a negative tendency to be distin-
versy revolves around what treatments should serve guished from simply not responding at all. The intro-
as a control for experience with the CS and US, and duction of the two-test strategy by Rescorla (1969)
the relationship between them. It is both convenient provided a solid foundation for a rich set of later exp- C
and theoretically meaningful to regard the signaling erimental findings in an area largely dormant since
power of a CS as falling somewhere on a scale from Pavlov’s original observations.
+1.0 (excitor) to 1.0 (inhibitor) with a zero neutral One of these new findings is that conditioned inhi-
point. The best control treatment then would theoret- bition is mediated by multiple mechanisms. Some con-
ically leave the control CS with zero signaling power. ditioned inhibitors seem to cancel a US expectation
One possibility might be to schedule the control CS that evoked a specific CS (negative occasion setter)
at random times during the experimental session. but not all other CSs (standard conditioned inhibi-
Unfortunately, chance forward pairings of the control tion). Conditioned inhibitors in the former class reduce
CS with the US can sometimes lead to elevated the behavioral effects of the specific excitor they previ-
responding, leaving the neutrality of the control CS in ously accompanied, but less so a new excitor in
question. Any alternative procedure in which one of a summation test. The negative occasion-setting mech-
the two main players in the relationship is omitted, anism seems to be favored when the inhibitory CS on
the CS or the US, is difficult to defend. This state of the trial terminates shortly before its excitatory partner
affairs has led to the tailoring of control procedures is nonreinforced. Situational or apparatus cues are also
to suit the experiment at hand. For example, if arbi- more likely to act as negative occasion setters than
trary letters are used to stand for which of several conditioned inhibitors. A short list of other generally
possible conditioned stimuli in the experiment are agreed characteristics of conditioned inhibition are:
actually present on a given trial (A, B, C. . .), and the
presence (“+”) or absence (“”) of a subsequent US is ● Learned expectations must be acquired before they
indicated, the compound trial method can be denoted can be suppressed by conditioned inhibition.
as A+, AB. Here, B is the conditioned inhibitor. ● Conditioned inhibition is less well retained than
Accordingly, we might schedule A+, AB, and C conditioned excitation over a retention interval.
trials in the experimental group, and A, AB, and C ● A CS will not lose its inhibitory power when
+ trials in the control group. In theory, B should be unreinforced in isolation outside of the original
more inhibitory in the experimental group than in the unreinforced compound.
control group because it clearly signals the omission ● Instrumental actions can serve as conditioned
of an otherwise expected US (evoked by A). However, inhibitors in avoidance learning. Here, the action
it is possible the intermixing of trials in the A, AB, signaling that a potentially aversive event will not
C+ control arrangement could also lead to some inhi- occur serves as the inhibitory stimulus.
bition as AB is differentiated from C+. This has led ● Conditioned inhibition develops whenever an
to the adoption of multiple control procedures in expectation of the forthcoming US is greater than
some experiments, none of which on its own is the the US actually obtained on the trial; hence,
single best. two excitors previously trained on separate trials
Historically, the most troublesome aspect of the will lose associative strength if they occur in
assessment question is how best to distinguish between compound and are actually followed by the US
a subject curtailing an otherwise likely response and (overexpectation).
simply not responding. The general approach has been ● A neutral stimulus reinforced in the presence of
to insist the inhibitor show the ability to suppress a conditioned inhibitor will gain strength extraor-
responding evoked by another known excitor (summa- dinarily quickly because of the larger than normal
tion test), as well as to show the inhibitor is not easily discrepancy between the subject’s “negative” expec-
converted into a signal for the presence of the US tation and the delivery of the US on the trial (super-
(retardation test). By providing an alternative impetus normal conditioning).
748 C Conditioned Response
● An excitatory CS can be protected from extinction Information processing in animals: Conditioned inhibition
if it is accompanied by a conditioned inhibitor (pp. 1–49). Hillsdale: Erlbaum.
Pavlov, I. P. (1927). Conditioned reflexes. New York: Dover.
when nonreinforced.
Rescorla, R. A. (1969). Pavlovian conditioned inhibition. Psycholog-
● Conditioned inhibitors sometimes, but not always, ical Bulletin, 72, 77–94.
convey information about the omission of a partic- Tobler, P. N., Dickinson, A., & Schultz, W. (2003). Coding of
ular identifiable event, such as the absence of sugar- predicted reward omission by dopamine neurons in
water but not the absence of food. a conditioned inhibition paradigm. Journal of Neuroscience, 23,
10402–10410.
Urcelay, G. P., & Miller, R. R. (2006). A comparator view of Pavlovian
Important Scientific Research and and differential inhibition. Journal of Experimental Psychology:
Open Questions Animal Behavior Processes, 32, 271–283.
Some of the properties of conditioned inhibitors Wagner, A. R., & Rescorla, R. A. (1972). Inhibition in Pavlovian
remain open to question. Researchers have reported conditioning: Application of a theory. In R. A. Boakes &
that extensive extinction of the particular CS+ associ- M. S. Halliday (Eds.), Inhibition and learning (pp. 301–336).
New York: Academic.
ated with the CS can eliminate conditioned inhibi-
tion. The conditions under which such deactivation
occurs continue to be studied (Urcelay and Miller
2006). Another question of continuing interest is to
what degree inhibitory conditioning plays a role in
human causal inference. Is a preventative cause just
Conditioned Response
an inhibitory signal that an expected effect will not When the pairing of one stimulus with another
occur? Much applied research has been directed at results in some specific change in response to either
enhancing the effectiveness of extinction (conditioned stimulus, then that change can be identified as
inhibition) to dampen traumatic memories. Can phar- having a conditioned basis. The oft-cited example of
macological agents be used to both quicken the process a conditioned response is Pavlov’s serendipitous
of extinction and make it stick in new contexts? Alter- observation that hungry dogs will come to salivate
natively, does fast extinction simply create the condi- to a bell that has previously signaled the delivery
tions for response recovery? Finally, do unreinforced of food.
trials given in the short temporal window after acqui-
sition prevent consolidation of a freshly learned CS-US
association?
CSs or USs and promote extinction of suppression to suppression to each was equal. They then conducted
the context. The experimenter can see if the ongoing a series of assays designed to see if learning was weaker
behavior rate is adequate for testing the CS, and, if to light than to tone, but that did not appear to be so.
so, can measure suppression to it in a subsequent ses- Finally, they arranged for suppression to be weaker to
sion. Off-line conditioning procedures also allow the tone than to light but still found more freezing to tone.
use of short CSs (except on test trials) if that is desired. The results suggest that conditioned suppression may
CS durations of 1–15 s. are common. In contrast, with be a more sensitive measure of fear than directly
on-line techniques, the CS must be long enough to observed freezing (at least as freezing was defined).
permit a reliable measure of the ongoing behavior in Importantly, the fact that the degree of freezing and
its presence. So in the on-line procedure, CSs are typ- suppression are correlated supports the idea that con-
ically 1–3 min. in duration. ditioned suppression in rats does indeed reflect fear
A danger in using off-line techniques arises when the and therefore should be a good animal model for the
experiment requires complex procedures and multiple study of fear acquisition and elimination in humans.
phases. When these phases are conducted off-line before Frequently, rats will show strong suppression to
an ultimate test trial, the experiment can resemble a novel CSs, particularly when they are brief. Looking
magic act. The audience sees the magician pull the rabbit only at suppression of the measured baseline
out of the hat on the test trial but can only guess how the responding, one might be tempted to believe that the
rabbit got there. In contrast, a similar experiment novel CS is frightening. Direct observation, however,
conducted on-line is fully transparent (for a discussion reveals otherwise (e.g., Ayres et al. 1987). Novel CSs
of this issue, see Rauhut et al. 2000, pp. 106–107). tend to evoke a great deal of activity, including a lot of
rearing (standing up on the hind legs). This behavior
Important Scientific Research and seems to reflect an orienting or investigatory response
Open Questions rather than a conditioned response. Thus, asking what
A criticism of conditioned suppression is that it tells us the rat actually does during the CS can help to deter-
what the subject is not doing during the CS (it is not mine whether CS-evoked suppression does or does not
engaging in the ongoing or baseline behavior), but it depend upon CS-US pairings.
does not tell us what the subject is doing. If Estes and
Skinner were correct in assuming that conditioned sup- Cross-References
pression reflects fear, then we should be able to predict ▶ Associative Learning
what the animal actually does during the CS. A basis for ▶ Comparator Hypothesis and Learning
such a prediction is an observation by Fanselow and ▶ Computational Models of Classical Conditioning
Lester (1988), who noted that fear restricts an animal’s ▶ Conditioned Inhibition
behavior to a small number that have an evolutionary ▶ Conditioning
history of thwarting predation. They stated that in the ▶ Context Conditioning
laboratory rat, the most dominant of these behaviors ▶ Contingency in Learning
seems to be freezing or defensive immobility. Usually, ▶ Emotional Learning
but not always, freezing occurs in a crouched position. ▶ Fear Conditioning in Animals and Humans
A number of studies have directly observed behav- ▶ Pavlovian Conditioning
ior during conditioned suppression and have system-
atically measured freezing (e.g., see Kim et al. 1996 and References
citations therein). They have found that freezing does Ayres, J. J. B., Haddad, C., & Albert, M. (1987). One-trial excitatory
indeed occur during the CS and that the degree of backward conditioning as assessed by conditioned suppression
suppression is correlated with the degree of freezing. of licking in rats: Concurrent observations of lick suppres-
(Freezing was defined as the absence of any movement sion and defensive behaviors. Animal Learning & Behavior, 15,
212–217.
save that of the sides needed for breathing.) Interest-
Estes, W. K., & Skinner, B. F. (1941). Some quantitative properties of
ingly, Kim et al. found that freezing was not the whole anxiety. Journal of Experimental Psychology, 29, 390–400.
story of conditioned suppression, because subjects Fanselow, M. S., & Lester, L. S. (1988). A functional behavioristic
froze more to tone CSs than to light CSs even though approach to aversively motivated behavior: Predatory
Conditions of Learning C 751
Theoretical Background
Conditioned Tolerance In order to understand the conditions of learning, one
first must have an understanding of the five categories
▶ Drug Conditioning of learning outcomes that Gagné identified. These five
categories are verbal information, intellectual skills,
cognitive strategies, attitudes, and motor skills. Each
of these five categories, along with the external instruc-
Conditioning tional conditions that support learning within that
category, is described below.
▶ Associative Learning
▶ Learning in Honeybees: Associative Processes
▶ Psychology of Learning (Overview Article)
Verbal Information
Gagné (1985) indicates that verbal information in-
volves the ability to state, tell, or describe facts, names,
labels, and/or principles, either as individual entities
or as interrelated elements, also known as bodies of
Conditioning Applications knowledge (such as the names of all the capital cities in
a particular region of the world). A person is said to
▶ Behavior Modification, Behavior Therapy, Applied have acquired, or learned, some verbal information
Behavior Analysis and Learning when that person is able to state, tell, or describe that
information in sentence form. Gagné points out that an
essential characteristic of verbal information learning is
that the learner states that information in essentially
Conditioning Therapies the same form in which it was presented, simply as
a fact, name, label, and so on. In other words, the
▶ Behavior Modification, Behavior Therapy, Applied learner is said to have acquired that information simply
Behavior Analysis and Learning by being able to restate it; the learner need not have to
752 C Conditions of Learning
apply that information in order to demonstrate that he recall that information on multiple occasions over
or she has learned it. time, is likely to aid retention.
cases a learner must be able to identify independent beginning the instruction by demonstrating the entire
clauses and coordinating conjunctions (concept learn- process to learners so that they get a preview of the
ing) before the learner can correctly insert commas into whole task.
sentences that contain those elements. Thus, before Providing learners with opportunity to practice
being taught the desired new skill, learners should be applying the rules they are being taught is another C
asked to recall the component skills, if they have already crucial condition of learning Gagné and Driscoll dis-
learned them, or should be taught those skills if they cuss. In doing so, they emphasize that simply because
have not as yet acquired them. a learner can state a rule does not mean he or she can
In instances where the desired learning outcome apply it; thus the need to have the learner practice
involves concept learning, it is important that instruc- application of the rule. Moreover, Gagné and Driscoll
tion direct a learner’s attention to the distinctive fea- indicate that spaced practice, practice of the same rule
tures of the concept to be learned. Moreover, if learners on multiple occasions over an extended period of time,
are likely to have difficulty distinguishing between two will greatly facilitate a learner’s ability to retain the skill
closely related concepts, Gagné and Driscoll indicate he or she has learned. Gagné and Driscoll also point to
that it is important to direct learner attention to the the value of having learners practice applying a skill in
features that serve to differentiate the two. Thus, for a variety of situations and contexts, thus promoting
example, if the goal is to teach learners to identify transfer of that skill.
proper fractions, the instruction should not only Finally, Gagné and Driscoll point to the importance
include definitions and examples of proper fractions, of feedback during rule using. They discuss the value of
but should also provide examples of fractions in which reinforcing correct responses and point to the impor-
the numerator is larger than, or equal to, the denomi- tance of corrective feedback when learners are having
nator and an accompanying explanation as to why such difficulty performing a rule-using task properly.
fractions are not proper fractions.
Inasmuch as rule learning often involves per- Cognitive Strategies
forming a series of steps, Gagné and Driscoll indicate According to Gagné, cognitive strategies are the means
that one of the key instructional strategies for teaching via which learners guide their own remembering,
such skills is to provide learners with cues that will thinking, and learning. For example, a learner might
help learners recall the sequence of steps, or a particular use a mnemonic device in order to recall the names of
step, in the process. For example, a verbal cue that is the planets in our solar system.
likely to help a learner recall how to divide by fractions
would be “invert the divisor and multiply.” Obviously, Conditions of Learning for Cognitive
in most cases stating this cue will just serve as one of Strategies
many instructional events that will be employed as Cognitive strategies are often developed by learners
a learner is being taught the desired skill. Nonetheless, independently as they engage in some learning activity.
this cue is quite likely to be a crucial one, one that will Nonetheless, Gagné and Driscoll (1988) indicate that
help the learner recall the necessary procedure for there are at least three categories of instructional activ-
dividing by fractions. ities that can be employed in order to help learners
Gagné and Driscoll also suggest that rules that acquire and use cognitive strategies. First, cognitive
involve a large number of steps should be taught in strategies may be demonstrated and/or described to
chunks. That is, if a rule-using task involves more than the learners. For example, when learners are being
one or two steps (e.g., balancing a checkbook), the taught how to solve complex problems, a strategy for
learner might first be provided with instruction and identifying the essential and irrelevant ideas presented
practice on the one or two steps in the process before in the problem situation can be described and dem-
the learner is presented with instruction and practice onstrated to the learners. As additional problems of
on any of the other steps. It is important to point out this nature are presented, demonstrations of how
that the employment of this approach, which has been to apply the strategy may be faded and replaced by
labeled by some as a “part-task approach” (e.g., van simple instructions reminding the learners to apply
Merrienboer 2007) does not preclude the possibility of the strategy.
754 C Conditions of Learning
Second, Gagné and Driscoll discuss the importance Moreover, Gagné and Driscoll indicate that in some
of providing learners with frequent opportunities to cases it may be valuable to have learners engage in
practice employing cognitive strategies. They suggest mental practice of physical skills, indicating that
that providing students with a variety of novel prob- learners may benefit from forming mental images of
lems within a particular content area will facilitate their how to perform the skill.
ability to apply a particular cognitive strategy or set of As Gagné and Driscoll indicate, when learners are
strategies to other novel problems within the same area. engaged in the physical practice of a motor skill, it is
Third, Gagné and Driscoll point to value of providing also very important to provide learners with feedback
learners with informative feedback as they are learning regarding their performance. The authors pay particu-
cognitive strategies. They indicate that this type of lar attention to two characteristics of the feedback that
feedback does not simply inform the learner as to should be provided to learners when they are engaging
whether his or her proposed solution to the problem in physical practice of motor skills. First, they discuss
was correct; in addition it might indicate the extent to the importance of immediate feedback, indicating that
which the process the learner employed in arriving at if feedback is not immediate, learners may get into
the solution was original, creative, or inventive (this the habit of performing a skill incorrectly, making it
assumes that the strategies that were employed are that much harder to teach them the proper execution
observable). In addition, one might presume that of the skill. Second, they discuss the need to provide
such feedback could also focus on the efficiency of the informative feedback, namely feedback that indicates
strategies the learner used. Moreover, in those cases in to the learner what aspect of performance was faulty
which strategies did not meet particular criteria, it and describing or demonstrating the correct manner of
would be useful to provide feedback recommending performing that action.
alternative techniques.
Attitudes
Motor Skills Attitudes, according to Gagné (1985), are the internal
Gagné indicates that motor skills usually involve feelings or beliefs that influence the choice of personal
a sequence of physical movements that “constitute actions an individual takes. For example, a person’s
a total action that is smooth, regular, and precisely beliefs about the value of wearing a seat belt while
timed” (Gagné 1985, p. 62). Examples include a wide driving is likely to influence his or her decision as to
array of physical activities, such as serving a tennis ball, whether to use one.
driving a car, printing the letters of the alphabet,
performing a type of dance, and innumerable other Conditions of Learning for Attitudes
physical actions. Gagné and Driscoll (1988) describe a variety of learning
conditions that can be employed to promote learner
Conditions of Learning for Motor Skills acquisition of particular attitudes. One involves the use
According to Gagné and Driscoll (1988), one of the of human modeling. They suggest that learners may be
important steps in teaching learners how to perform influenced to adopt a particular attitude if they are
a particular motor skill is to describe and demonstrate shown examples of a positive role model, someone
the various physical procedures (also called the execu- they admire or respect, displaying that attitude. The
tive subroutines) which constitute that skill. They also authors point out that usually it is not sufficient to
suggest that for complex skills, in addition to demon- simply have the model talk about the value of adopting
strating the skill as a whole, it is valuable to divide the a particular attitude, it is important that the model
skill into parts and describe and demonstrate each part actually display the action that reflects that attitude.
separately. For example, rather than simply stating it is important
The authors also point to the importance of pro- that drivers bring their cars to a complete stop at stop
viding learners with many opportunities to engage signs, a model should also display that behavior.
in the physical practice of a motor skill so that learners Another strategy involves establishing an expec-
can not only learn how to perform the skill, but can tancy for success on the part of the learner. Gagné and
have opportunities to fine-tune that performance. Driscoll indicate that if learners are rewarded or
Conditions of Learning C 755
experience some form of success after engaging in some centered around Gagné’s views regarding a particular
action based on personal choice, they are more likely to event, such as providing learning guidance (Alutu
continue to engage in such actions. In other words, the 2006) or presenting instructional cues (Tomic 1980).
learners will be more likely to adopt the attitude that In a much larger number of cases, researchers have
led to that choice of action. For example, an individual examined how various levels of a particular instruc- C
who receives praise for recycling an item is more likely tional event, such as modeling behaviors (West and
to continue to engage in recycling. In a similar vein, Graham 2007) or providing feedback (Ifenthaler 2011),
Gagné and Driscoll indicate that when a role model affect learning. However, most of these studies have
engages in some attitudinal behavior, it is important to not specifically focused on Gagné’s views regarding
demonstrate how the role model is rewarded or receives these events.
some satisfaction from taking that action. For example, While some studies related to Gagné’s work have
a role model might discuss the satisfaction he or she focused on a single instructional event or condition,
received from doing some voluntary service activity several studies have focused on lessons or materials
(such as the satisfaction I have received from preparing that incorporate several of the events of instruction
this entry for this encyclopedia!). described by Gagné and have examined how the pres-
ence versus absence of one or more of these events
Important Scientific Research and affected student learning and attitudes. In several
Open Questions such studies (Martin et al. 2007; Martin and Klein
Many of the studies that have examined Gagné’s views 2008), providing learners the opportunity to practice
regarding the conditions of learning have focused on desired skills proved to be the instructional event
a particular instructional event or condition that he that had the greatest effect on learning. These studies
postulated as facilitating a particular type of learning call for further examination of how a combination of
outcome. For example, as noted earlier, Gagné (1985) Gagné’s events of instruction affect student learning
indicated that intellectual skills that involve a large and attitudes.
number of steps should be taught in small chunks, Other studies have focused on the degree to which
with the learner receiving instruction and practice on Gagné’s events of instruction are employed during a
a few steps at a time. In recent years, several authors lesson, and have examined how various levels of use
(e.g., van Merrienboer 2007; Merrill 2009) have raised are correlated with student learning and attitudes. For
questions about this viewpoint. As a result, several example, in a study examining the instructional activ-
researchers have compared instructional approaches ities employed in 37 sections of undergraduate com-
that employ this “part-task approach” with a “whole- puter science and chemistry courses, Hampton and
task approach” in which, from the outset of a lesson, Reiser (2004) found that student learning and motiva-
the practice activities presented to learners require tion were positively correlated with the degree to which
them to perform all the skills or steps that constitute Gagné’s events of instruction were employed.
the whole task, starting with a simple version of the Many models of teaching that are often character-
whole task and, over time, progressing to more com- ized as “direct instruction” (Magliaro et al. 2005) pre-
plex versions of the task. Results of a recent study scribe the use of a set of instructional activities similar
revealed that skill acquisition and transfer was greater to Gagné’s events of instruction. Given the current
among students in a whole-task condition than among debate about the appropriate degree of instructional
students in a part-task group (Lim et al. 2009). How- guidance that should be provided to learners (Clark
ever, the researchers clearly indicated that a great deal and Hannafin 2012), there is a need for additional
of additional research is necessary in order to get research examining the effects of direct instructional
a clearer picture of the relative merits of these two approaches such as those proposed by Gagné.
approaches across a wide variety of cognitive skills
and learners. Cross-References
Research has also been conducted on many of the ▶ Abilities to Learn: Cognitive Abilities
other instructional conditions or events of instruction ▶ Attitudes: Formation and Change
described by Gagné. In a few instances, research has ▶ Chunking Mechanisms and Learning
756 C Cone of Experience
References
Alutu, A. N. G. (2006). The guidance role of the instructor in the
teaching and learning process. Journal of Instructional Psychology,
33(1), 44–49. Cone of Experience
Ausubel, D. P. (1978). In defense of advance organizers: a reply to the
critics. Review of Educational Research, 48, 251–257. ▶ Multimodal Learning Through Media
Clark, R. E., & Hannafin, M. J. (2012). Debate about the benefits
of different levels of instructional guidance. In R. A. Reiser &
J. V. Dempsey (Eds.), Trends and issues in instructional design and
technology (3rd ed.). Boston: Pearson.
Gagné, R. M. (1965). The conditions of learning (1st ed.). New York: Cone of Learning
Holt, Rinehart & Winston.
Gagné, R. M. (1985). The conditions of learning (4th ed.). New York: ▶ Multimodal Learning Through Media
Holt, Rinehart & Winston.
Gagné, R. M., & Drsicoll, M. P. (1988). Essentials of learning for
instruction (2nd ed.). Englewood Cliffs: Prentice Hall.
Gagné, R. M., & Medsker, K. L. (1996). The conditions of learning:
training applications. Fort Worth: Harcourt Brace. Confidence in Retrieval
Hampton, S. E., & Reiser, R. A. (2004). Effects of a theory-based
feedback and consultation process on instruction and learning in ▶ Calibration
college classrooms. Research in Higher Education, 45, 497–527.
Ifenthaler, D. (2011). Bridging the gap between expert-novice differ-
ences: the model-based feedback approach. Journal of Research
on Technology in Education, 43(2), 103–117.
Lim, J., Reiser, R. A., & Olina, Z. (2009). The effects of part-task and
whole-task instructional approaches on acquisition and transfer Confidence Judgments in
of a complex cognitive skill. Educational Technology Research and Learning
Development, 57(1), 61–77.
Magliaro, S. G., Lockee, B. B., & Burton, J. K. (2005). Direct instruc-
CHRISTOPH MENGELKAMP1, MARIA BANNERT2
tion revisited: a key model for instructional technology. Educa- 1
tional Technology Research and Development, 53(4), 41–55.
Department of General and Educational Psychology,
Martin, F., & Klein, J. (2008). Effects of objectives, practice, and University of Koblenz-Landau, Landau, Germany
2
review in multimedia instruction. Journal of Multimedia and Instructional Media, University of Wuerzburg,
Hypermedia, 17(2), 171–189. Wuerzburg, Germany
Martin, F., Klein, J. D., & Sullivan, H. (2007). The impact of instruc-
tional elements in computer-based instruction. British Journal of
Educational Technology, 38(4), 623–636.
Merrill, M. D. (2009). First principles of instruction. In Synonyms
C. M. Reigeluth & A. Carr (Eds.), Instructional design theories Metacognitive judgments
and models: building a common knowledge base (Vol. III).
New York: Routledge Publishers. Definition
Pressley, M., Levin, J. R., & Delaney, H. D. (1982). The mnemonic
In research about ▶ metacognition (for an overview see
keyword method. Review of Educational Research, 52, 61–91.
Tomic, W. (1980). The concept of instructional cues. Twente educa- Dunlosky and Metcalfe 2009) confidence judgments
tional memorandum number 24. Twente, Netherlands: University or metacognitive judgments are defined as assessments
of Twente of the current state of knowledge. Thus, referring to
Confidence Judgments in Learning C 757
▶ cognitive learning students assess their own knowl- text. Then she judges how well she has understood the
edge themselves at certain points of time during their text and how good she will be at the examination, that
learning. In this regard three kinds of judgments are of is, she judges her learning (JOL). As she feels that she
particular interest: ease of learning (EOL) judgments has not got everything right, she decides to reread some
that are taken before the learning begins, judgments of parts of the text and to have a look at the boxes that she C
learning (JOL) that are taken after learning but before neglected during her first reading. Coming to the end
a performance test, and retrospective confidence (RC) of the text again, Jennifer fills in the practice test at the
judgments that are taken after testing. end of the chapter. Because she feels very confident that
her answers in the practice test are correct (RC judg-
Theoretical Background ments) she decides not to read anymore about steam
In the framework of metacognition proposed by engines. From an educational perspective the core
Nelson and Narens (1992), cognitions are split into question is: How well will Jennifer do at her examina-
two levels, namely, the object-level and the meta-level. tion? The answer to this question partly depends on the
All cognitions about the content to be learned are accuracy of her judgments, because her decisions dur-
located at the object-level, for example, activities like ing learning are conductively only if the judgments
reading or elaborating, and representations like the reflect the actual knowledge at that point of time in
definitions of terms or mental images of pictures. Men- the learning process. Otherwise she will allocate study-
tal representations about these cognitions are located at time toward texts that she already knows, reread text
the meta-level, for example, the belief that a certain parts superfluously, and she will not invest additional
learning strategy will be efficient, the plan to reach effort for learning following an illusion of knowing.
a learning goal, the belief that one has comprehended Thus, confidence judgments are a central component
the content, etc. Thus, the learner constructs a mental of ▶ self-regulated learning and affect the learning pro-
model at the meta-level that maps the cognitions at the cess and the learning outcome. But how are judgments
object-level, and this model may be altered as the learn- and their accuracy obtained? We will give a brief intro-
ing process continues, that is, the model at the meta- duction into the methods in the next section.
level changes. One of the processes that potentially To obtain confidence judgments learners are
alter the mental model is monitoring that is defined as usually asked questions like “How well will you be
assessing the cognitions at the object-level. Confidence able to complete a test over this material?” The answer
judgments are an important part of this monitoring as is often given using a scale from 0% to 100%.
they assess the current state of knowledge about the If a multiple choice test is used the lower limit is
learning content. After the knowledge has been judged adjusted to guessing, for example, 20% for an answer
and the mental model has been updated, the cognitions format using five alternatives. As the actual parameter
on the object-level may be controlled. Thus, besides for guessing can differ from this value depending on
monitoring control is the second important process in the used distracters – see 3-PL models in item response
the framework by Nelson and Narens (1992). During theory – we suggest using open answer formats or other
learning such control may lead to rereading, selection formats with a guessing nearby 0% whenever possible.
of another study-strategy, etc. Judgments can be obtained as local or global judg-
We will elaborate on the role of judgments in ments. Local EOL judgments or JOLs are obtained for
learning more deeply using an example. Assume that each content or text section to be learned; local RC
Jennifer reads a text about the functioning of a steam judgments are obtained after each item of a test. In
engine as a part of her preparation for an exam. Before contrast global judgments are made for the whole
she begins to read she asks herself how difficult it will be learning material or the whole test. As we have argued
to comprehend this text, that is, she judges her EOL. above, the accuracy of judgments is crucial for self-
Depending on her judgment she will reserve time for regulated learning and for learning outcome. There
reading about steam engines and allocate less or more are mainly two different kinds of accuracy measures
study-time to other themes. Afterward she begins read- calculated: Absolute accuracy (aka calibration) is based
ing, but she decides not to read some of the additional on the difference or the absolute difference between the
material that is printed in boxes throughout the main judgment and the performance indicating how much
758 C Confidence Judgments in Learning
learners’ judgments derivate from their performance. learning (see, e.g., Dunlosky and Metcalfe 2009).
Absolute accuracy can be visualized plotting calibra- Study-time allocation was mainly studied in laborato-
tion curves (see, e.g., Fig. 1) that are indicating under- ries using word pairs or short sentences as learning
and ▶ overconfidence at each level of confidence. A material. One result is that JOLs predict the allocation
prerequisite for calculating absolute accuracy is that of study-time in a following learning phase. There are
judgments either are made on the same scale as the two hypotheses explaining this result. The discrepancy-
performances (e.g., “I will solve 12 items correctly”) or reduction hypothesis states that study-time is allocated
on a percentage scale. Relative accuracy (aka resolution) to those items that are judged as least known, whereas
indicates to what extend learners discriminate bet- the region-of-proximal-learning hypothesis states that
ween test performances on correct vs. incorrect items. study-time is allocated to those items that are judged
Relative accuracy is calculated as the within-person as not yet known, but among these items the easiest
correlation between the judgments and performances, ones will be chosen first. There is evidence for both
often using the nonparametric gamma correlation. hypotheses found in the literature, and maybe both
A prerequisite for calculating relative accuracy is that mechanisms are used by learners. However, each of
local judgments have been used. the hypotheses explains why the accuracy of judgments
is related via the mediator study-time allocation to the
Important Scientific Research and learning outcome. And indeed, there is evidence for
Open Questions JOLs’ relative accuracy being correlated to the learning
In this section we will briefly review core research about outcome in text comprehension (Maki and McGuire
confidence judgments in learning and stress major 2002). Beyond that experiments have shown that
research questions. Four topics are addressed: (1) judg- enhancing JOLs’ accuracy causes an increase in learn-
ments, study-time allocation, and learning outcome; ing outcome. Moreover researchers have shown that
(2) enhancing JOLs’ accuracy; (3) stability and gener- the relative accuracy of RC judgments is correlated with
ality of judgments and their accuracy; and (4) calibra- comprehension in reading too. In sum this body of
tion in classroom studies. research supports the claim that JOLs and RC judg-
Researchers have been interested in study-time ments are important for the control of learning pro-
allocation as a part of metacognitive control during cesses, and that self-regulation is effective only if the
accuracy of these judgments differs from zero.
Based on the function of confidence judgments for
Intermediate test
learning there has been considerable effort made to
100
enhance the accuracy of judgments (see, e.g., Dunlosky
and Metcalfe 2009; Dunlosky and Lipko 2007). There
(698)
are several interventions found to be effective in push-
Percentage correct
80
ing the relative accuracy of JOLs: (a) JOLs are more
(84) (372)
accurate if they are obtained delayed rather than imme-
60 (458) diately after learning. One explanation is that delayed
judgments include processes of activation from ▶ long-
(648)
40 term memory only whereas immediate judgments
additionally rely on ▶ working memory. In contrast
performance in a knowledge test is based on long-
term memory and therefore the delayed judgments
40 60 80 100
reflect the knowledge more accurately than immediate
Judgments (%)
judgments. (b) Deeper understanding of texts facili-
Confidence Judgments in Learning. Fig. 1 Example of tates the relative accuracy of judgments, for example,
a calibration curve using RC judgments. Frequencies of throughout rereading, writing summaries, and gener-
judgments are given in brackets. The dotted line ating key terms before taking the JOLs. But how is this
indicates perfect calibration. Cited from Mengelkamp effect explained? Firstly, it is known from the ▶ levels
and Bannert (2010) of processing approach that deep processing leads to
Configural Cues in Associative Learning C 759
better comprehension of texts. Secondly, techniques was trained explicitly and incentives for accurate judg-
like generating key terms are indicative for the depth ments were given, a positive effect on calibration and
of processing. As JOLs reflect the depth of processing performance in the final test was achieved.
and the performance reflect the depth, too, the relative Confidence judgments are important for learning.
accuracy of the judgments increases. Nevertheless it is often ignored that judging ones C
From an interindividual perspective, the stability knowledge during learning is a highly complex process,
and generality of judgments and their accuracy was and in order to improve learning significantly theories
investigated (see, e.g., Mengelkamp and Bannert of SRL and metacognition need to be integrated.
2010). There is evidence that the judgments themselves
Cross-References
are considerably stable over the time at least within
▶ Comprehension Monitoring
each kind of judgment, and this corresponds to evi-
▶ Metacognition and Learning
dence from research using RC judgments in test-taking.
▶ Metacognitive Strategies
This stability may reflect stable characteristics of per-
▶ Self-regulated Learning
sons like beliefs about one’s own ability or self-confi-
dence. Further, it was questioned if there is a stable and References
general metacognitive ability; thus the stability and Dunlosky, J., & Lipko, A. R. (2007). Metacomprehension: A brief
generality of the judgments’ accuracy is of interest. history and how to improve its accuracy. Current Directions in
Psychological Science, 16(4), 228–232. doi:10.1111/j.1467-
Results indicate that relative accuracy is not stable at
8721.2007.00509.x.
all and generalizes not across different domains. In
Dunlosky, J., & Metcalfe, J. (2009). Metacognition. Thousand Oaks:
contrast absolute accuracy seems to be moderately Sage.
stable and generalizable. But as absolute accuracy is Hacker, D. J., Bol, L., & Keener, M. C. (2008). Metacognition in
not mathematically independent from the magnitude education: A focus on calibration. In J. Dunlosky & R. A. Bjork
of judgments and the magnitude of performances, the (Eds.), Handbook of metamemory and memory (pp. 429–456).
New York: Psychology.
latter result potentially is an artifact. To sum up this
Maki, R. H., & McGuire, M. J. (2002). Metacognition for text:
section, relative accuracy of judgments seems not to be Findings and implications for education. In T. J. Perfect &
much of a trait but it is sensitive to characteristics of the B. L. Schwartz (Eds.), Applied metacognition (pp. 68–92).
learning situation and thus open for interventions. Cambridge, UK: University Press.
The research presented so far was mainly conducted Mengelkamp, C., & Bannert, M. (2010). Accuracy of confidence
judgments: Stability and generality in the learning process and
in laboratories using relative measures of accuracy.
predictive validity for learning outcome. Memory & Cognition,
Since the late 1990s, there is a growing body of research
38, 441–451. doi:10.3758/MC.38.4.441.
that has been conducted in classrooms using absolute Nelson, T. O., & Narens, L. (1992). Metamemory: A theoretical
measures of accuracy (see, e.g., Hacker et al. 2008). framework and new findings. In T. O. Nelson (Ed.), Metacogni-
One result is the replication of the “unskilled but tion: Core readings (pp. 117–130). Needham Heights: Allyn
unaware” effect in educational settings, that is, low- and Bacon.
achieving persons overestimate themselves whereas
high-achieving persons are quite well calibrated or
show slight underconfidence. Further, almost all stud-
ies found a correlation between the absolute accuracy Configural Cues in Associative
of JOLs or RC judgments and the final test perfor- Learning
mance. Therefore, one aim of research in the class-
room is to get students better calibrated in order to STEVEN GLAUTIER
improve the learning outcome. There are some studies School of Psychology, Southampton University,
addressing this aim using RC judgments. One obvious Southampton, UK
way is to give practice tests toward the students in order
to increase their accuracy of judgments. First results
show that giving the students practice tests together Synonyms
with judgments and feedback is not increasing the Conjunction; Part; Pattern; Stimulus configuration;
absolute accuracy of their judgments. But if monitoring Whole
760 C Configural Cues in Associative Learning
Definition
Configural cues are stimuli provided by the juxtaposi-
tion, in time or space, of individual stimulus elements.
Associative learning is said to have taken place when
presentation of a stimulus elicits a new response as
a result of a history of pairing with another stimulus.
The outcomes of many associative learning experi-
ments demonstrate that organisms use configural cues
during learning.
Theoretical Background
As long ago as 350 BC, Aristotle considered the distinc-
tion between individual elements and their assembly:
“In the case of all things which have several parts and in
which the totality is not, as it were, a mere heap, but the
whole is something beside the parts . . .” (Aristotle,
350 BC, Book VIII, Part 6; as cited by Ross (2009)).
More recently the Gestalt psychologists are commonly
associated with the expression “the whole is greater Configural Cues in Associative Learning. Fig. 1 Stimuli
than the sum of its parts” for which Aristotle’s words exhibiting “configuration“ effects. Left-hand-side two
are one precursor. The general principle referred to arrangements of four corners, and right-hand-side, the
is straightforward and there are many nice visual exam- Thatcher Illusion
ples from the perception literature, such as the one
shown on the left of Fig. 1. The arrangement of
four corners in the top left gives the impression of between features and thus facilitates rapid identifica-
a square shape, which is quite different to the impres- tion of anomalies such as feature inversion.
sion formed when the same elements are rearranged, as Early demonstrations of configural effects in asso-
shown in the bottom left of the figure. Discussions of ciative learning appeared in the western literature
stimulus configuration effects are not confined to the during the 1930s. Following-up on the work of other
perception literature. Other examples arise in different Russian investigators, including Pavlov, Gregory
areas of psychology including attention, face percep- Razran documented faster learning in humans when
tion, and associative learning. In studies of divided compound stimuli were used as conditioned stimuli
attention it has been argued that response times when than when individual elements were used. Salivation
two signals (e.g., tone and letter) are presented together responses were acquired more rapidly to an alternating
are faster than could be expected if the signals are pattern of red and green lights which signaled food
processed separately on independent channels. Instead, than to a single red or green light of the same duration.
it has been suggested that attention is allocated to A little later, Charles Woodbury was working with dogs
a third signal internally generated from the combina- and found that the dogs could learn a negative pat-
tion of the two experimenter defined signals. The terning discrimination. This discrimination involves
Thatcher illusion has been proposed as evidence in reinforced presentations of two stimuli presented indi-
support of the view that configurations of facial fea- vidually but non-reinforcement of the two stimuli
tures play an important part in face perception. The presented in compound, a procedure which can be
face of Margaret Thatcher, shown on the right of Fig. 1, summarized as a series of intermixed A+, B+, and
does not appear dramatically unusual at first glance. AB trials. The fact that conditioned responding
However, when viewed after a rotation of 180 the can be lower to a compound than to either of the
corruption of the image is readily apparent. Upright elements presented alone shows that the animals
viewing uses configural cues coding the relationships were responding to something that was unique to the
Configural Cues in Associative Learning C 761
compound, distinguishing it from the elements from Important Scientific Research and
which it was composed. Other discriminations, such as Open Questions
a biconditional discrimination (AB+, BC, CD+, and Although it is clear that humans and animals can
AD trials), which can be readily solved by humans represent and learn about stimulus configurations
and animals, also suggest that stimulus configurations the best characterization of those representations is C
are attended to during associative learning. In the not firmly established. Wagner recently proposed an
biconditional discrimination, all stimulus elements elaborated version of the configural cue hypothesis,
are reinforced and non-reinforced equally often; there- the Replaced Elements Model. In the Replaced Ele-
fore, alone, they cannot inform the subject of the ments Model stimulus compounds not only generate
response requirement. Instead, to respond appropri- new configural cues but also, simultaneously, these
ately in this type of task it is necessary to processes configural cues inhibit components of the original ele-
stimulus conjunctions such as “A and B.” ments. So, a compound of A and B generates a repre-
Of course, because these configural cues are sentation of the conjunction, c, and at the same time
unobservable theoretical entities, inferred as a means produces some changes in A and B so the final charac-
to understand behavior in particular experimental sit- terization of the compound is A0 B0 c (see Wagner 2008).
uations, there has been considerable speculation about One of the advantages of such a model is that it natu-
their underlying nature. It is commonly thought that rally predicts asymmetrical generalization gradients
configural cues are processed in the same way as ele- from adding and removing elements. According to
mental cues and, following Clark Hull’s lead, that they this model, after conditioning of A, if a test trial
arise as the result of a perceptual interaction between of an AB compound is presented then the loss of
simultaneously or successively presented stimuli. So, it responding will be proportional to the difference
is supposed, for example, that a tone sounds slightly between A and A0 (the novel B0 and c components of
different when presented in compound or soon after the compound are not assumed to affect responding).
a buzzer (and the sound of the buzzer is changed in On the other hand, after conditioning of an AB
a similar way) and the “unique elements” generated by compound, should a test trial of A be presented then
the compound code the conjunction. Hull coined the the loss of responding will be proportional to the
phrase “afferent neural interaction” to capture this difference between A0 B0 c and A. In most situations
idea. However, one challenge to this notion has arisen it would be expected that the difference between
from the suggestion that the unit of analysis should A0 B0 c and A would be greater than the difference
be the pattern itself, rather than the elements. Pearce between A and A0 . Studies with humans and animals
(1994) argued that all of the elements present in a have confirmed this result (Brandon et al. 2000;
stimulus are represented as a single configural unit, or Glautier 2004). Moreover, this “Replaced Elements
pattern, and it is this pattern which is processed by the Model” is significant in that it allows for a degree of
learning mechanism. Thus, during biconditional dis- flexibility in the extent to which cue conjunctions are
crimination four different patterns (AB, BC, CD, and encoded. This idea is consistent with a developing view
AD) gain associative strength. Although reinforcement in the literature that different experimental prepara-
is distributed evenly across the elements (A, B, C, and tions can result in a more or less configural represen-
D) rendering a simple elemental account of learning of tation of the stimulus compounds which are presented
this discrimination inadequate, there is differential (Melchers et al. 2008).
reinforcement at the level of pattern. The fact that the
patterns in a biconditional discrimination share some Cross-References
elements, and hence have nonzero similarities, means ▶ Associative Learning
that this discrimination should be more difficult to ▶ Computational Models of Classical Conditioning
learn than one which just involves A+, B, C+, and ▶ Conditioning
D trials. The fact that biconditional discrimination is ▶ Discrimination Learning
learned more slowly than a simple discrimination is ▶ Formal Learning Theory
consistent with this analysis. ▶ Representation Changes in Learning
762 C Confirmation Bias
References
Brandon, S., Vogel, E. H., & Wagner, A. R. (2000). A componential Confucian Educational
view of configural cues in generalization and discrimination in Philosophy and Its Implication
Pavlovian conditioning. Behavioural Brain Research, 110, 67–72.
Glautier, S. (2004). Asymmetry of generalization decrement in causal
for Lifelong Learning
learning. The Quarterly Journal of Experimental Psychology, 57B,
315–329. QI SUN
Melchers, K. G., Shanks, D. R., & Lachnit, H. (2008). Stimulus coding Adult and Post Secondary Education, Department of
in human associative learning: Flexible representation of parts Professional Studies, College of Education, University
and wholes. Behavioural Processes, 77, 413–427. of Wyoming, Laramie, WY, USA
Pearce, J. M. (1994). Similarity and discrimination: A selective review
and a connectionist model. Psychological Review, 101, 587–607.
Ross, W. D. (2009). Metaphysics by Aristotle book VIII.
http://classics.mit.edu/Aristotle/metaphysics.8.viii.html. Accessed Synonyms
20 Dec 2010. Confucian educational philosophy; Confucianism;
Wagner, A. R. (2008). Evolution of an elemental theory of Pavlovian Confucius; Lifelong education; Lifelong learning
conditioning. Learning & Behavior, 36, 253–265.
Definitions
Confucius (551 B.C.E. – 479 B.C.E.) was the greatest
educator, philosopher, and eminent figure in the his-
Confirmation Bias tory of China. Confucius is a latinized name for “Kong
Fu Zi” by the Italian Jesuit Matteo Ricci (1552 C.E. –
The tendency to search and select information
1610 C.E.), when Confucian works started to be intro-
confirming personal hypotheses and beliefs, ignoring
duced to the Western world during the late sixteenth
contrary evidence.
century. His last name was Kong, people generally
called him “Kong Fu Zi.” “Fu Zi” added following a
surname was an honorific title back then, which meant
Master. “Kong Fu Zi,” translated as Confucius, thus has
Conformist Bias been known to the world.
▶ Theory of Conformist Social Learning Confucianism started from the thoughts of Confu-
cius (Huang 2006; Zhang 2009), developed, enriched,
and joined by the thoughts of Mencius (372 B.C.E. –
289 B.C.E.), Xun Zi (298 B.C.E. – 238 B.C.E.), and
other followers. Historically, Confucianism has gone
Conformist Transmission through many stages, such as the “original Confucian-
▶ Theory of Conformist Social Learning ism,” the “Han Confucianism (206 B.C.E. – 220 C.E.),”
and the “Neo-Confucianism.” For over 2,500 years,
Confucian philosophy has exerted a profound influ-
ence on almost every aspect of Chinese society, partic-
Conformity ularly, in the education arena.
Lifelong learning is a broad concept. It generally
▶ Reproductive Learning denotes that learning, either for personal, profes-
sional, institutional/organizational, and/or societal
purposes, continues via all kinds of learning activities
throughout life span, whether formal, nonformal,
Confucian Educational and/or informal. Lifelong learning has been inter-
Philosophy changeably utilized with concepts such as lifelong edu-
cation, continuing education, continuing professional
▶ Confucian Educational Philosophy and Its Implica- education, recurrent education, adult learning, and
tion for Lifelong Learning adult education.
Confucian Educational Philosophy and Its Implication for Lifelong Learning C 763
Today, the emergence of the knowledge economy perspective, Ren is the knowledge of morality and
in an era of globalization evidently leads to an increasing humanity. Confucius believed Ren is gained through
economic competitiveness, thus makes lifelong learning lifelong self-cultivation. In fact, the processes of lifelong
primarily a governmental instrument for the promotion learning deepen the facets of developing morality and
of a knowledge society, skills’ formation, transferable humanity toward a multidimensional world of which
skills, multiskilling and careership, which all reflect humans are a part.
human capital theory (Green 2006; Olssen 2006). Sage: It is the Confucian ideal human model, who
As can be seen, from a humanistic tradition, the has realized Ren. “Sage is one who has reached the
concept understands individual growth and develop- highest realm and become (1) the undivided “I” with
ment. Embedded in the pragmatism, it sees the need for the Universe; (2) the unity of “I” with other humans and
societal and organizational improvement, for social other beings; and (3) the wholeness of “I” with self” (Sun
development or transformation. Then, influenced by 2004, p. 79). With this nature, sage, being at the most
the utilitarianism and human capital theory, it centers perfect stage, can fully develop his or her own nature.
on economic effectiveness as opposed to a social polit- More importantly, he or she can fully develop the nature
ical perspective advocating citizenship. of his or her fellows and all other things. In fact, he or she
We would all agree that each argument embraces is able to assist the transforming and nourishing powers
valid points. However, from the Confucian perspective, of Heaven and Earth, and ultimately form a triad with the
all seems to exclude other equally important dimen- Universe. Confucius positively advocated that through
sions of human development via lifelong learning. It is lifelong learning and practice every human being was
in this context that the Confucian educational philos- capable of gaining access to Ren, to be a Sage. This
ophy presents us significant implication for lifelong ideal human model framed the realistic purpose of
learning and lifelong education. the Confucian learning, which is to educate Jun Zi.
The core of Confucian educational philosophy is Jun Zi: A term Confucius adopted and replenished
how we learn to be human. Human beings are the ends the meaning to refer to a person who is willing to learn
not the means. Confucius believed that the ultimate and practice Ren via lifelong learning and cultivation.
end of learning was to realize the true nature of human Jun Zi also stands for the Confucian educated, an
beings – become fully human. We are all human beings. exemplary and model characterized by outstanding
However, we are not born fully human. Each of us must knowledge, courage, and multiple skills. Ideally, Jun
still consciously learn to be so. To Confucius, lifelong Zi learns to realize and manifest his or her true nature
learning enables human beings to realize and practice toward the universe, the natural world and other
their true nature and live happily in and with the beings, the social world and other humans, and the
worlds of different kinds: universe, nature, society/ inner world of self. Toward the universe, Jun Zi respects
other human beings, and inner self (Sun 2004). Thus, the Tao of Heaven and understands each human’s fate
the Confucian educational philosophy has a fuller per- (Lun Yu, [the Analects], XVI, 8; XX, 3). Toward society,
spective on purposes and functions of lifelong learn- Jun Zi has strong social accountability. Toward other
ing. Several conceptions Ren, Sage, and Jun Zi reviewed beings, Jun Zi follows the precept “do not impose on
by Sun (2004, 2008) are critical in understanding the others what you do not desire (Lun Yu, XII, 2; XV. 24).”
Confucian multidimensional learning. Toward self, Jun Zi never ceases self-strengthening.
Ren: Generally translated as humanity, morality, Jun Zi holistically develops the wholeness and
and righteousness, it is the backbone of Confucian becomes a multidimensional model of learning and
philosophy, which has manifestly influenced Confu- doing, to be harmonious in both internal and external
cian educational thought and practice. From an axio- qualities. In the social sphere, for instance, he or she not
logical perspective, Ren is the utmost virtue of the only cultivates the self but also establishes others. He or
Universe. It is the totality of morals. It leads human she is not only a self-directed learner but also an edu-
beings to manifest their true nature. Confucius under- cator of others. He or she is not only knowledgeable but
stood that humans are potentially moral. Yet, the also action oriented.
potentials need to be cultivated and developed through Hence, the Confucian educational philosophy
lifelong learning and practice. From an epistemological presents significant implications for the practice of
Confucian Educational Philosophy and Its Implication for Lifelong Learning C 765
today’s lifelong learning. Comparing with the Western importantly, could lifelong learning resolve all kinds of
approaches to lifelong learning, Confucian educational social ills and economic problems, without looking at
philosophy through the ideal human model of Sage the center that is our human beings? These fundamental
and realistic educational end Jun Zi, presents holistic yet overlooked questions invite us to critically ponder
and multidimensional goals and functions for human and move beyond our own paradigms of thoughts. C
beings to learn lifelong. As can be seen, the Confucian Although Western thought and way of thinking
lifelong learning includes but also goes beyond the have become dominated through its “scientific”
purposes advocated by each stage of the development research, more and more scholars (Merriam and Asso-
of lifelong learning during past several decades, ciates 2007) acknowledge that there are truly huge
whether humanistic for personal development, or missing parts of non-Western outlooks that could ben-
pragmatic and utilitarian for economic crisis. For Con- efit the whole human beings’ learning. Optimistically,
fucius, human beings live in and interact with different Confucian philosophy and his educational practice
worlds: universe, nature, society, and self. Thus, human may provide us alternative lens and perspectives to
beings need to learn for and from, live with, and live find possible solutions if we are willing to explore and
in each reality. All help realize the true and complete open to perspectives other than our own (Sun 2008).
nature of being human. In other words, any single
aspect of development no matter how successful and Cross-References
full will only lead to an incomplete of a human being. ▶ Experiencing Wisdom Across the Lifespan
▶ Learning and Fluid Intelligence Across the Lifespan
Important Scientific Research and ▶ Lifelong and Worklife Learning
Open Questions ▶ Mental Models and Life-Long Learning
Since 1970s, the functions and goals of lifelong learning ▶ Values in Education and Life-Long Learning
have been changed from one set to another to practi-
cally meet needs under new contexts. These modifi- References
cations also reflect efforts of scholars’ debates and Dehmel, A. (2006). Making a European area of lifelong learning
criticism on the narrowly focused purposes associated a reality? Some critical reflections on the European union’s
with certain philosophical beliefs (Dehmel 2006; lifelong learning policies. Comparative Education, 42(1), 49–62.
Green, A. (2006). Models of lifelong learning and the ‘knowledge
Holford and Jarvid 2000; Schuetze 2006). Researchers
society’. Compare, 36(3), 307–325.
have called for an overarching conceptual framework, Hinchliffe, G. (2006). Re-thinking lifelong learning. Studies in Phi-
“one that describes the basic dimensions, relates central losophy and Education, 25, 93–109.
elements and points to strategic issues and consider- Holdford, J., & Jarvis, P. (2000). The learning society. In A. L. Wilson
ations relevant for policy and practice” (Tuijnman and & E. R. Hayes (Eds.), Handbook of adult and continuing education
Boström 2002: 105). In a changing society, lifelong (new edn., pp. 643–659). San Francisco: Jossey-Bass.
Huang, Z. (2006). Confucian moral education theory review. Wuhan:
learning must be on the condition of human beings
Wuhan University.
(Lindgren 2002). However, we are experiencing, “the Lindgren, A. (2002). Lifelong learning in a changing world. In
key driver of the development of lifelong learning is K. Harney, A. Heikkinen, S. Rahn, & M. Schemmann (Eds.),
thus the emergence of the knowledge economy in an Lifelong learning: One focus, different systems (pp. 55–72).
era of globalization” (Hinchliffe 2006, p. 94). Evidently, New York: Peter Lang.
Merriam, S. B., & Associates. (2007). Non-Western perspectives on
the current practice of lifelong learning leaves some of
learning and knowing. Malabar: Krieger.
the other dimensions of human existence out that they OECD. (1996). Lifelong learning for all: Meeting of the educational
have vanished from sight. What do we miss? Should we committee at ministerial level. Paris: OECD.
revisit the centuries’ old question: What is the purpose Olssen, M. (2006). Understanding the mechanisms of neoliberal
of learning and education? Are human beings the ends control: Lifelong learning, flexibility and knowledge capitalism.
or the means of lifelong learning? Do we now have The International Journal of Lifelong Education, 25(3), 213–230.
Schuetze, H. G. (2006). International concepts and agendas of life-
different understandings of ends and means than did
long learning. Compare, 36(3), 289–306.
ancient sages such as Confucius? What can we learn Sun, Q. (2004). To be Ren and Jun Zi: A confucian perspective of
from the Eastern ancient educational philosophy for the practice of contemporary education. The Journal of Thought,
the modern practice of lifelong learning? More 39(2), 77–91.
766 C Confucianism
Sun, Q. (2008). Confucian educational philosophy and its implica- and Cohen 2003). Confusion indicates an uncertainty
tion for lifelong learning and lifelong education. International about what to do next or how to act (Graesser et al.
Journal of Lifelong Education, 27(5), 559–578.
2007). Thus, confusion accompanies cognitive disequi-
The analects of Confucius Bao, S. (Trans. into Modern Chinese)
Lao, A. (Trans. into English), (1992), (Shandong, Ji Nan: librium which plays an important role in comprehen-
Shandong Friendship Press). sion and learning processes (Piaget 1952; Graesser and
Tuijnman, A., & Bostrom, A. (2002). Changing notions of lifelong Olde 2003).
education and lifelong learning. International Review of Educa- Under this theory, people start in a state of equilib-
tion, 48(1/2), 93–110.
rium. In this phase, they are processing information
Wain, K. (2004). The learning society in a postmodern world: The
educational crisis. NewYork: Peter Lang.
from the world around them, but not always at a deep
Zhang, X. (2009). Review and interpretation of the phenomenon of level. Deep comprehension occurs when learners con-
Confucius: Ritual and music in life and philosophy. Shanghai: front contradictions, anomalous events, obstacles to
China Eastern Normal University. goals, salient contrasts, perturbations, surprises, equiv-
alent alternatives, and other stimuli or experiences that
fail to match expectations (Mandler 1976; Schank
1986). At this point, the person moves into a state of
Confucianism cognitive disequilibrium where a misunderstanding
is realized and attempts are started to reconcile the
▶ Confucian Educational Philosophy and Its Implica- conflicting internal and external sources of informa-
tion for Lifelong Learning tion. Cognitive disequilibrium has a high likelihood of
activating conscious, effortful cognitive deliberation,
questions, and inquiry that aim to restore cognitive
equilibrium. The affective state of confusion is diag-
Confucius nostic of cognitive disequilibrium (Graesser and Olde
2003; Graesser et al. 2007) and the resolution of the
▶ Confucian Educational Philosophy and Its Implica-
confusion is essential to restoring equilibrium.
tion for Lifelong Learning
Important Scientific Research and
Open Questions
Recent, empirical evidence substantiates the predic-
Confusion’s Impact on Learning tions that confusion is related to learning. Craig et al.
(2004) conducted an observational study in which
SCOTTY D. CRAIG confusion among other affective states (i.e., boredom,
Department of Psychology/IIS, The University of eureka, flow, frustration, and neutral) were observed
Memphis, Memphis, TN, USA during a learning session with an intelligent tutoring
system. Of the observed affective states, confusion
and flow displayed significant positive correlates with
Synonyms learning. Boredom was observed to be negatively cor-
Uncertainty related. Of the affective states, only confusion was
observed to predict learning, accounting for 27% of
Definition the variance. Further, when learner’s performance for
Confusion is a cognitive-affective state that occurs sessions with and without confusion present was com-
when a person is aware of an inconsistency between pared, significant differences were revealed. Partici-
their knowledge and observed information and is pants in confusion-present sessions exhibited a 46%
actively seeking to reconcile the discrepancy. increase in learning (Cohen’s d = .64) over participants
with confusion-absent sessions.
Theoretical Background However, the presence of confusion might not
Recent empirical research has pointed to confusion as always produce ideal learning. Since confusion occurs
an important affective state for scientific study (Rozin during cognitive disequilibrium, the learner could
Connectionism C 767
either resolve the confusion and move back into a Graesser, A. C., & Olde, B. (2003). How does one know whether
state of equilibrium or fail to resolve it. D’Mello and a person understands a device? The quality of the questions the
person asks when the device breaks down. Journal of Educational
Graesser (2010) demonstrated this empirically using
Psychology, 95, 524–536.
an offline self-report methodology. Under this meth- Mandler, G. (1976). Mind and emotion. New York: Wiley.
odology, learners viewed a video of their interaction Piaget, J. (1952). The origins of intelligence. New York: International C
with an intelligent tutor and indicated their affective University Press.
states from a fixed list of affective states: Confusion, Rozin, P., & Cohen, A. B. (2003). High frequency of facial expressions
corresponding to confusion concentration, and worry in an
Boredom, Flow, Frustration, Delight, Surprise, and
analysis of naturally occurring facial expressions of Americans.
Neutral. They found that confusion states associated Emotion, 3, 68–75.
with more learning showed conflict resolution and Schank, R. C. (1986). Explanation patterns: Understanding mechani-
thus a return to cognitive equilibrium of neutral or cally and creatively. Hillsdale: Lawrence Erlbaum Associates.
flow states. However, when participants exhibited con-
fusion states that were not resolved they tended to
transition to frustration and boredom and decreased
learning.
When the learner is confused, there might be a
Congruence
variety of potential paths to pursue. The learner could Congruence, realness, or genuineness is a most basic
be allowed to continue being confused during the attitude for the facilitation of learning. It means that
cognitive disequilibrium (and the affiliated increased the feelings the facilitator is experiencing are available
physiological arousal that accompanies all affective to his or her awareness, that he or she is able to live
states). The learner’s self-regulated thoughts might these feelings, to be them, and able to communicate
hopefully restore equilibrium when feedback to learner them if appropriate (Rogers 1983).
errors is delayed. Alternatively, after some period of
time waiting for the learner to progress, indirect hints
could be provided to nudge the learner into more References
productive trajectories of thought. However, the opti- Rogers, C. R. (1983). Klientenzentrierte Psychotherapie. In
J. R. Corsini (Ed.), Handbuch der Psychotherapie (S. 471–512).
mal paths have yet to be determined.
Weinheim: Beltz.
Cross-References
▶ Affective and Cognitive Learning in the Online
Classroom
▶ Boredom of Learning
Conjunction
▶ Emotions: Functions and Effects on Learning ▶ Configural Cues in Associative Learning
▶ Flow Experience and Learning
References
Craig, S. D., Graesser, A. C., Sullins, J., & Gholson, B. (2004). Affect Connectionism
and learning: An exploratory look into the role of affect in
learning. Journal of Educational Media, 29, 241–250.
THEMIS N. KARAMINIS, MICHAEL S. C. THOMAS
D’Mello, S., & Graesser, A. (2010). Modeling cognitive-affective
dynamics with hidden Markov models. In R. Catrambone & Department of Psychological Sciences, Birkbeck
S. Ohlsson(Eds.), Proceedings of the 32nd Annual Cognitive College, University of London, London, UK
Science Society (pp. 2721–2726). Austin: Cognitive Science
Society.
Graesser, A. C., Chipman, P., King, B., McDaniel, B., & D’Mello, S.
(2007). Emotions and Learning with AutoTutor. In R. Luckin, K.
Synonyms
R. Koedinger, & J. Greer (Eds.), Artificial intelligence in education: Artificial Neural network modeling; Connectionist
Building technology rich learning contexts that work (AIED07) modeling; Neural nets; Parallel Distributed Processing
(pp. 554–556). Washington, DC: IOS Press. (PDP)
768 C Connectionism
X1 W1
input Σ XiWi
activity from W2 output to
other weighted threshold other
X2 y
neurons sum of input activation neurons
activity function
W3
X3
decoding
‘Wickelfeature’
representation of past
tense
‘Wickelfeature’
representation of verb
root
encoding
Connectionism. Fig. 3 The Rumelhart and McClelland (1986) model for the learning of the English past tense. The core of
the model is a two-layered feed-forward network (pattern associator) which learns mappings between coarse-coded
distributed representations (Wickelfeature representations) of verb roots and past tense forms
In the connectionist framework an artificial neural The environment is presented as pairs of input patterns
network model depicts cognition when it is able to and desired output patterns (or targets), where the
respond to its environment with meaningful activa- target is provided by an external system (the notional
tion patterns. This can be achieved by modifications “supervisor”). The network is trained on the task of
of the values of the connection weights, so as to regulate producing the corresponding targets in the output
the activation patterns in the network appropriately. when an input pattern is presented.
Therefore, connectionism suggests that learning involves The Backpropagation of Error algorithm (Rumelhart
the shaping of the connection weights. A learning algo- et al. 1986b) as proposed for training such networks.
rithm is necessary to determine the changes in the Backpropagation is an error-driven algorithm. The aim
weight values by which the network can acquire of the weight changes is the minimization of the output
domain-appropriate input-output mappings. error of the network. The Backpropagation algorithm
The idea that learning in artificial neural networks is based on the delta rule:
should entail changes in the weight values was based on
DWij ¼ ðti ai Þaj ð2Þ
observations of neuropsychologist Donald Hebb on
biological neural systems. Hebb (1949) proposed his The delta rule is a modification of the Hebbian
cell assembly theory also known as Hebb’s rule or Hebb’s learning rule (Eq. 1) for neurons that learn with super-
postulate: vised learning. In the delta rule, the weight change
"
(Dwij) is proportional to the difference between the
When an axon of cell A is near enough to excite a cell
target output (ti ) and the output activation of the
B and repeatedly or persistently takes part in firing it,
receiving neuron (ai ), and the output activation of
some growth process or metabolic change takes place
the sending neuron (aj ).
in one or both cells such that A’s efficiency, as one of
Backpropagation generalizes the delta rule in net-
the cells firing B, is increased. (1949, p.62)
works with hidden layers, as a target activation value is
Hebb’s rule suggested that connections between not available for the neurons on these internal layers.
neurons which present correlated activity should be Internal layers are necessary to improve the computa-
strengthened. This type of learning was also termed tional power of the learning system. In a forward pass,
correlational or associative learning. the Backpropagation algorithm calculates the activations
A simple mathematical formulation of the Hebbian of the units of the network. Next, in a backward pass the
learning rule is:
DWij ¼ ai aj ð1Þ
Target activity values
The change of the weight (Dwij) from a sending unit
Output patterns
j to a receiving unit i should be equal to the constant
multiplied by the product of output activation values Output
units
(ai and aj) of the units. The constant is known as
learning rate.
algorithm iteratively computes error signals (delta terms) O’Reilly and Munakata (2000) proposed the
for the units of the deeper layers of the network. The LEABRA (Local, Error-driven and Associative, Biolog-
error signals express the contribution of each unit to ically Realistic Algorithm) algorithm. This algorithm
the overall error of the network. They are computed combines error-driven and Hebbian Learning,
based on the derivatives of the error function. Error exploiting bidirectional connectivity to allow the C
signals determine changes in the weights which mini- propagation of error signals in a biologically plausible
mize the overall network error. The generalized delta fashion.
rule is used for this purpose: The supervised learning algorithms assume a very
detailed error signal telling each output how it should
DWij ¼ di aj ð3Þ
be responding. Other algorithms have been developed
According to this rule, weight changes equal to the that assume less detailed information. These approaches
learning rate times the product of the output activation are referred to as reinforcement learning.
of the sending unit (aj) and the delta term of the Another class of neural networks is trained with
receiving unit (dii ). unsupervised learning. In this type of learning, the
Although the Backpropagation algorithm has been network is presented with different input patterns.
widely used, it employs features which are biologically The aim of the network is to form its own internal
implausible. For example, it is implausible that error representations which reflect regularities in the input
signals are calculated and transmitted between the neu- patterns.
rons. However, it has been argued that since forward The Self-Organizing Map (SOM; Kohonen 1984)
projections between neurons are often matched by is an example of a neural network architecture that is
backward projections permitting bidirectional signal- trained with unsupervised learning. As shown in Fig. 2,
ing, the backward projections may allow the imple- a SOM consists of an array of neurons or nodes. Each
mentation of the abstract idea of the backpropagation node has coordinates on the map and is associated with
of error. a weight vector, of the same dimensionality as the input
Pursuing this idea, other learning algorithms patterns. For example, if there are three dimensions
have been proposed to implement error-driven learn- in the input, there will be three input units, and each
ing in a more biologically plausible way. The Contras-
tive Hebbian Learning algorithm (Hinton 1989) is
a learning algorithm for bidirectional connected net-
x1 x2 x3
works. This algorithm considers two phases of training
in each presentation of an input pattern. In the first
Input vector
one, known as the minus phase or anti-Hebbian update,
the network is allowed to settle as an input pattern is
wij
presented to the network while the output units are free
to adopt any activation state. These activations serve
as noise. In the second phase (plus phase or Hebbian
update), the network settles as the input is presented
while the output units are clamped to the target out-
puts. These activations serve as signal. The weight
change is proportional to the difference between the Pattern class 1
products of the activations of the sending and the
receiving units in the two phases, so that the changes
reinforce signal and reduce noise: Pattern class 2
Array of nodes
DWij ¼ ai þ aj þ ai aj ð4Þ (output layer)
Learning is based on contrasting the two phases, Connectionist Theories of Learning. Fig. 2 Unsupervised
hence the term Contrastive Hebbian Learning. learning in a simple self-organizing map (SOM)
774 C Conscientiousness
output unit will have a vector of three weights microstructure of cognition. Volume 1: Foundations (pp. 45–76).
connected to those input units. Cambridge, MA: MIT Press.
Rumelhart, D. E., Hinton, G. E., & Williams, R. J. (1986b).
The aim of the SOM learning algorithm is to pro-
Learning internal representations by error propagation. In
duce a topographic map that reflects regularities in the D. E. Rumelhart, J. L. McClelland, & The PDP Research Group
set of input patterns. When an input pattern is (Eds.), Parallel distributed processing: Explorations in the micro-
presented to the network, the SOM training algorithm structure of cognition. Volume 1: Foundations (pp. 318–362).
computes the Euclidean distance between the weight Cambridge, MA: MIT Press.
vector and the input pattern for each node. The node
that presents the least Euclidean distance (winning node
or best matching unit [BMU]) is associated with the
input pattern. Next, the weights vectors of the neigh- Conscientiousness
boring nodes are changed so as to become more similar
to the weights vector of the winning node. The extent One of the big five personality factors. Individuals
of the weight changes for each of the neighboring scoring high on this dimension tend to be organized
nodes is determined by its location on the map using and mindful of details.
a neighborhood function. In effect, regions of the output
layer compete to represent the input patterns, and
regional organization is enforced by short-range excit-
atory and long range inhibitory connections within Consciousness and Emotion:
the output layer. SOMs are thought to capture aspects Attentive vs. Pre-attentive
of the organization of sensory input in the cerebral Elaboration of Face Processing
cortex. Hebbian learning to associate sensory and
motor topographic maps then provides the basis MICHELA BALCONI
for a system that learns to generate adaptive behavior Department of Psychology, Catholic University of
in an environment. Milan, Milan, Italy
Cross-References
▶ Adaptive Learning Systems Synonyms
▶ Bayesian Learning Aware; Facial expression; Implicit elaboration;
▶ Computational Models of Human Learning Unaware; Unconscious processing
▶ Connectionism
▶ Hebbian Learning Definition
▶ Learning in Artificial Neural Networks Emotional facial expressions represent facial displays of
▶ Reinforcement Learning in Spiking Neural Networks emotions which determine different patterns of mus-
▶ Self-Organized Learning cular correlates, cognitive responses, and brain activa-
tion. Autonomic and central nervous systems cooperate
References in order to provide a coherent pattern of mimic
Hebb, D. O. (1949). The organization of behavior: A neuropsychological responses to specific contextual cues. Positive (i.e., hap-
approach. NewYork: Wiley. piness) vs. negative (i.e., anger) facial expressions are
Hinton, G. E. (1989). Deterministic Boltzmann learning performs
produced respectively in consequences to aversive or
steepest descent in weightspace. Neural Computation, 1, 143–150.
Kohonen, T. (1984). Self-organization and associative memory. Berlin: appetitive contexts. People are able to consciously pro-
Springer-Verlag. duce and comprehend facial expressions, but in many
O’Reilly, R. C., & Munakata, Y. (2000). Computational explorations in cases, they may obtain emotional information from
cognitive neuroscience: Understanding the mind by simulating the face by using an unconscious processing (pre-attentive
brain. Cambridge, MA: MIT Press.
processing). Typically, tasks that can be performed in less
Rumelhart, D. E., Hinton, G. E., & McClelland, J. L. (1986a).
A general framework for parallel distributed processing. In than 200 ms are considered pre-attentive. Simple fea-
D. E. Rumelhart, J. L. McClelland, & The PDP Research Group tures are extracted from the visual display in the pre-
(Eds.), Parallel distributed processing: Explorations in the attentive system and later joined in the focused attention
Consciousness and Emotion: Attentive vs. Pre-attentive Elaboration of Face Processing C 775
system into coherent objects. Pre-attentive processing is predisposition to respond to emotional situations.
done quickly, effortlessly and in parallel. Taking advan- This fact is in line with previous research that have
tage of pre-attentive processing can greatly improve used autonomic (skin conductance measures or
intuitiveness of representations yielding in a faster and cardiovascular indexes) measures or ▶ conditioned
more natural way of acquiring information. Contrarily, responses, that pointed out unconscious affective C
conscious production and comprehension of facial stimuli may have effect for the appraisal of conscious
expressions allow a more detailed and complete way of stimuli. It seems that the information presented under
processing information, it being a serial and effortful way pre-attentive conditions may be perceived and cogni-
of take information. tively processed. For this reason, facial expressions of
emotion are considered unique in their ability to orient
Theoretical Background the subjective cognitive resources, even if people are
Rapid detection of emotional information is highly unable to process information consciously.
adaptive, since it provides critical element on environ- Also the responses to unconscious stimulation
ment and on the attitude of the other people (Darwin showed to be sensitive to the emotional content of the
1872). Specifically, motivational significance of emotions facial stimuli, as revealed by different behavioral and
has an effect on subjects’ responses, since it was found physiological measures. That is, it was hypothesized
that emotionally salient stimuli (unpleasant compared that subjects are able to assign a semantic value to the
to neutral; more arousing compared to less arousing) emotional content of faces even in an unaware condi-
generate greater magnitude cognitive, cortical, and tion. Unconsciously processing for facial stimuli can
autonomic system responses. Thus, significance of also be demonstrated in clinical context, such as in
emotional facial expressions in terms of the relevance case of prosopagnosia. In most cases, prosopagnosics
of the emotional patterns for the subjective safeguard appear to recognize familiar faces even though they fail
influences the degree of attentional resources allocated. to identify the persons verbally. Therefore, the patients
It was showed that each emotion and its facial showed an unconscious recognition that cannot be
expression represents a specific response to a particular accessed consciously (Tranel and Damasio 1985).
kind of significant event, that is it is evaluated by the Similarities in processing between attentive vs. pre-
subject in line with its motivational significance. This attentive stimulation can also be assessed from neural
▶ appraisal process is regulated by two main criteria: point of view: Consistent analogies in the aware and
the arousing power of the stimulus (high or low); unaware processing structure were well-founded,
the ▶ valence of the emotional stimulus (positive or suggesting that similar neural activity is involved. In
negative) (Russell 2003). Thus, the entire emotional humans, evidence for the unconscious perception of
universe is representable by the two axes, and the “sig- emotional face has been revealed in terms of subjective
nificance” attributed to the emotional expressions may reports, autonomic reactions, brain imaging measure,
have a direct effect on the cognitive level and the degree as well as ERPs (event-related potentials) and brain
of attention allocated. oscillations. Brain areas generally involved in evalua-
Moreover, significant affective processes happen tion of the emotional and motivational significance of
outside consciousness (LeDoux 1996). It has been facial expressions appear to be mediated by the amyg-
shown that the affective information contained in facial dala and orbitofrontal cortex, while structures such as
expression is perceived involuntarily, and is able to the anterior cingulate, prefrontal cortex, and somato-
constrict automatically the focus of attention. Infor- sensory areas are linked to the conscious representation
mation presented under the conscious threshold may of emotional facial expression for the strategic control
be processed on a high level even if the subject is of thought and action (Adolphs 2003).
unaware of this information, since pre-attentive
response to emotional faces is effective in eliciting Important Scientific Research and
coherent subjective responses. Two main factors seem Open Questions
to be relevant in orienting subject’s response to the Numerous studies have sought to demonstrate that
emotional cues in case of unconscious stimulation: emotional information can be perceived without aware-
the content of the stimulus and the subjective ness. The conclusion that emotional facial expressions
776 C Consciousness and Emotion: Attentive vs. Pre-attentive Elaboration of Face Processing
can be perceived without consciousness is not surpris- comprehension, such as ERP. Specifically, ERP mea-
ing given the importance of emotional information sures are very useful tools to examine the time course
for human survival. Nevertheless, although the exis- of the conscious vs. unconscious stimulus elaboration
tence of unconscious affect elaboration was accepted, at a very high temporal resolution (Balconi and Mazza
the question concerning its relevance for emotional 2009). ERPs furnish a valid measure of the qualitative
decoding is still open. Specifically, what remains nature of the emotional mechanisms, checking the
unclear was the specific semantic value this perception resemblance of the underlying processes for attentive
has and in what measure the subject can elaborate the and not attentive emotional elaboration. For this rea-
unconscious emotional stimuli. son, it is interesting to compare ERP profiles in con-
Only a limited number of studies have explored scious vs. unconscious condition, in order to verify the
the significance of conscious vs. unconscious face com- similarity of the comprehension processes. Therefore,
prehension, based on ▶ priming effect or subliminal the main topics to be explored in the future about
stimulation. The short stimulus presentation in pre- attentive vs. pre-attentive processing are related to the
attentive condition prevents the subjects to have a semantic significance of unconsciously processed emo-
clear cognition of the stimulus. Generally, an objective tional stimulus. Specifically:
threshold for pre-attentive condition is provided. It
– The resemblance of the two processes, attentive and
was defined by an identification procedure, the case
pre-attentive, from a qualitative point of view
where stimulus is perceived by the subject no more
– The existence of some differences in terms of the
than in 50% of the times. According to signal detection
type and the amount of cognitive resources required
theory (SDT), when detection threshold sensitivity is at
to activate the two processes
chance (d0 = 0), it is unlikely that there is conscious
– The temporal course of these mechanisms, with
awareness of the stimulus.
a delayed or anticipated effect for unconscious
Moreover, another useful measure to analyze con-
elaboration
scious and unconscious perception of faces is the
– The resemblance of attentive vs. pre-attentive
masking procedure. By low intensity and brief expo-
process in response to different emotional facial
sure, a target stimulus can be made unrecognized
expressions
when another stimulus is presented simultaneously,
shortly before (forward masking), or shortly after
(backward masking). This paradigm is used to investi-
Cross-References
▶ Attention and Implicit Learning
gate below awareness response to emotional perception
▶ Emotional Mental Models
in which facial expressions are followed immediately
▶ Emotional Schema(s)
by a masking face. Evidence for the unconscious per-
▶ Explicit Versus Implicit Learning
ception of masked faces has been revealed in terms of
subjective reports, autonomic activity, and functional References
brain imaging measures. Nevertheless, the effect of this Adolphs, R. (2003). Cognitive neuroscience of human social behav-
masking technique was not largely used for the emo- iour. Nature Reviews. Neuroscience, 4, 165–178.
tional face detection, and there is no precise knowledge Balconi, M., & Mazza, G. (2009). Brain oscillations and BIS/BAS
of the actual effect of masked emotional stimulus on (behavioral inhibition/activation system) effects on processing
the elaboration of the target one. masked emotional cues. ERS/ERD and coherence measures
of alpha band. International Journal of Psychophysiology, 74,
Most of the recent research on the detection and
158–165.
analysis of emotionally significant information from Darwin, C. (1872). The expression of the emotions in man and animals.
face have used fMRI (functional Magnetic Resonance London: John Murray.
Imaging) measures, which are based on relatively LeDoux, J. E. (1996). The emotional brain: The mysterious underpin-
slow hemodynamic brain responses, and the studies ning of emotional life. New York: Simon and Schuste.
Russell, J. A. (2003). Core affect and the psychological construction of
on the time course of emotional processing have
emotion. Psychology Review, 110, 145–172.
been relatively scarce. Thus, these methods need to be Tranel, D., & Damasio, A. R. (1985). Knowledge without awareness:
completed with measures that provide insights An automatic index of facial recognition by prosopoagnosics.
into temporal parameters of unconscious emotional Science, 228, 1453–1454.
Constraint Satisfaction for Learning Hypotheses in Inductive Logic Programming C 777
Definition
Consensus Learning Inductive logic programming is a subfield of machine
▶ Brainstorming and Learning learning which uses first-order logic as a uniform rep-
resentation of examples, background knowledge, and
hypotheses. In many works, it is assumed that examples C
are clauses and the goal is to find a consistent hypoth-
esis H, that is, a clause entailing all positive examples
Consolidation and no negative example. Entailment is frequently
checked using y-subsumption which is a decidable
The time-dependent strengthening of a memory after restriction of logical entailment. Given a clause T
a training trial (or trials) that results from biological called a template, the template consistency problem
processes in the brain. deals with finding a substitution s such that H=Ts
is a consistent hypothesis. Both entailment checking
Cross-References and template consistency problems are combinatorial
▶ Covert Reorganization / Spatial Learning problems that can be solved using constraint satisfac-
tion techniques.
Theoretical Background
In the core form, Inductive Logic Programming (ILP)
Constraint Networks deals with the problem of finding a hypothesis covering
all positive examples and excluding negative examples
A constraint network is a set of variables and con- (Muggleton and De Raedt 1994). For the sake of com-
straints that interrelate and define the valid values plexity analysis, a formalization of core ILP tasks was
for the variables. Constraint networks have proven proposed by the seminal paper (Gottlob et al. 1999).
to be a useful mechanism for modeling computa- Gottlob defines two basic ILP problems: the bounded
tionally intensive tasks in artificial intelligence. They consistency problem and the template consistency
operate by expressing a problem as a set of variables, problem. In both, it is assumed that examples are
variable domains, and constraints and define a clauses and the goal is to find a consistent hypothesis
search procedure that tries to satisfy the constraints H, that is, a clause entailing all positive examples and
by assigning values to variables from their specified no negative example. Entailment is checked using
domains. y-subsumption (Plotkin 1970) which is a decidable
restriction of logical entailment. For simplicity of nota-
tion, we can assume clauses to be expressed as sets
of literals, and, without loss of generality, we can only
work with positive literals, that is, non-negated atoms.
Constraint Satisfaction for All terms in the learning examples (hypotheses, respec-
Learning Hypotheses in tively) are constants (variables) written in the lower
Inductive Logic Programming (upper) cases. For instance, E = {arc(a,b), arc(b,c),
arc(c,a)} is an example and H = {arc(X,Y), arc(Y,Z)}
ROMAN BARTÁK1, FILIP ŽELEZNÝ2, ONDŘEJ KUŽELKA2 is a hypothesis. Hypothesis H subsumes example E, if
1
Faculty of Mathematics and Physics, there exists a substitution y of variables such that Hy
Charles University in Prague, Prague, Czech Republic E. In the above example, substitution y={X/a, Y/b, Z/c}
2
Faculty of Electrical Engineering, Czech Technical implies that H subsumes E.
University, Prague, Czech Republic In the bounded consistency formulation, the num-
ber of literals in H is polynomially bounded by the
number of examples. In the template consistency for-
Synonyms mulation, it is instead required that H = Ts for
Template consistency problem some substitution s, where T is a given clause called
778 C Constraint Satisfaction for Learning Hypotheses in Inductive Logic Programming
a template. Since all terms in H are supposed to be constraints). The constraint is arc consistent if all
variables, the task lies in determining which subsets values in the domains of constrained variables have
of variables in T should be unified. For generality, we some support in the constraint; in particular, the
assume that all variables in T are mutually different, value is part of a tuple satisfying the constraint. The
that is, each variable occurs exactly once in T, as in T= values without a support are removed from the vari-
{arc(X1,X2), arc(X3,X4)}. An exemplary hypothesis ables’ domains. For example, the constraint A=B,
H may be obtained from this T by applying unification where the domain of A is {1,2} and the domain of B is
X2 = X3 (and then suitably renaming the variables). {1,2,3}, is made arc consistent by removing value 3
Clearly, if no unification is applied and the template from the domain of B. If any domain becomes empty
consists only of the predicates in the example (arc/2 then the problem has no solution. If all the constraints
in our case) then the hypothesis subsumes that exam- are arc consistent and the domains are still not single-
ple. The reason for introducing unifications is thus ton then some variable is selected, its domain is
to prevent H from subsuming negative examples. In split into two (or more) disjoint subsets, and the
our case, hypothesis obtained by applying unifications obtained subproblems are solved using the same tech-
X2 = X3 and X1 = X4 to T does not subsume the above nique. This domain splitting forms a choice point in
example E. the search procedure. Other branching strategies exist,
Gottlob shows that both bounded consistency and for example, taking some constraint in the form of
template consistency problems are equivalent in terms exclusive disjunction such as C∨ C 0 and adding C to
of computational complexity and belong among S2P the problem in one search branch and C 0 to the prob-
complete problems. In both cases, the complexity arises lem in the second search branch. An optimization
from two sources: version of a CSP called a Constraint Optimization
Problem (COP) adds a objective function that evalu-
1. “The subsumption test for checking whether
ates the solutions. The task is to find a solution to a CSP
a clause subsumes an example”
that minimizes (or maximizes) the value of the objec-
2. “The choice of the positions of variables in the
tive function.
atoms (of the clause)”
Informally, (2) corresponds to the task of searching Important Scientific Research and
the space of admissible clauses, and (1) corresponds to Open Questions
evaluating an explored clause. Both task (1) and (2) can Maloberti and Sebag (2004) used constraint satisfac-
be solved using constraint satisfaction techniques. tion techniques to address the above complexity source
Constraint satisfaction problem (CSP) is a triple (1). In particular, they proposed a y-subsumption
(X, D, C), where X is a finite set of decision variables, algorithm called Django that is based on reformulation
for each xi 2 X, Di 2 D is a finite set of possible values of y-subsumption as a binary constraint satisfaction
for the variable xi (the domain), and C is a finite set of problem. Thanks to powerful CSP heuristics, Django
constraints (Dechter 2003). A constraint is a relation brought dramatic speed-up for y-subsumption and
over a subset of variables (its scope) that restricts the consequently for the entire ILP system. The constraint
possible combinations of values to be assigned to the model for each example looks as follows. First, for each
variables in its scope. The constraints can be expressed predicate symbol p with arity k we collect all k-tuples of
in extension using a set of compatible value tuples. values from atoms of this predicate in the example.
A solution to a CSP is a complete instantiation of vari- These value tuples define in extension a k-ary con-
ables such that the values are taken from respective straint cp. Now, for each atom of predicate p with vari-
domains and all constraints are satisfied. Constraint ables {Y1,. . ., Yk} in the hypothesis we post constraint cp
satisfaction techniques are frequently based on the over these variables. Clearly, based on instantiation
combination of inference techniques and search. The of variables {Y1,. . ., Yk} we can find an atom in the
most widely used inference technique is arc consistency example to which a given atom from the hypothesis
(the name goes from the graph representation of a CSP, is mapped to. Let {arc(a,b), arc(b,c), arc(c,a)} be all
where nodes describe the variables and arcs specify the atoms of predicate arc/2 in the example. Then binary
Constraint Satisfaction for Learning Hypotheses in Inductive Logic Programming C 779
constraint carc is defined in extension by a set of value where n is the total number of variables (in constraint
pairs {(a,b), (b,c), (c,a)}. Atom arc(X,Y) from the element(X,List,Y) X and Y are variables and L is list of
hypothesis is represented by constraint carc(X,Y) and variables; the constraint describes a relation that Y
instantiation X=a, Y=b means that that this atom is equals to the X-th element of List.). In other words, if
mapped to arc(a,b) in the example. In summary, variable Xi is mapped to Xj (Ii = j) then Xj is not C
any solution to a CSP defined by constraints cp mapped to any preceding variable (Ij = j, i.e., IIi = Ii).
describes a substitution y such that Hy E. The fol- For example, [1,1,2] is not a valid list of indices (it
lowing example demonstrates the constraint model represents X1 = X2 and X2 = X3), the correct represen-
for the subsumption problem (let us call it a subsump- tation of this unification should be [1,1,1] (X1 = X2
tion model): and X1 = X3). The element constraints thus ensure that
each set of unifications is represented by a single list of
Example: indices. The following example demonstrates the base
arc(a,b), arc(b,c), arc(c,a), red(a), red(c) unification model:
Hypothesis:
Template:
arc(Y1,Y2), arc(Y2,Y3), red(Y2)
arc(X1,X2), arc(X3,X4), arc(X5,X6), red(X7), red(X8), red(X9),
Subsumption model: green(X10)
Variables Y1, Y2, Y3 Unification model:
Domains {a,b,c} Variables I1, . . ., I10
Constraints carc(Y1,Y2), carc(Y2,Y3), cred(Y2) Domains Di = {1,. . .,i} 8i = 1,. . .,10
Solutions {Y1 =c, Y2 =a, Y3 =b}, {Y1 =b, Y2 =c, Y3 =a} Constraints element(Ii, [I1,.., I10], Ii) 8i = 1,. . .,10
To address the above complexity source (2) Barták The unification model needs to be connected to the
(2010) proposed a constraint model and dedicated subsumption models for individual examples. This can
search strategy for specifying which variables in the be done again via the element constraints in the fol-
template should be unified to obtain a consistent lowing way. Assume that n is the number of variables in
hypothesis. Recall that each variable appears exactly the template. Then for each example Ej we plug a set
once in the template so one can order the variables. Xj,1,. . ., Xj,n of fresh variables into H, where these vari-
Indices in the following example show this ordering T ables participate in the subsumption model for prob-
={arc(X1,X2), arc(X3,X4), arc(X5,X6)}. The model is lem Ej. Note that each example may define different
based on the observation that if a set of variables is compatible tuples for the constraints and hence
unified then we can select the variable with the smallest a different solution (subsumption) yj. The constraints
index to represent this set and all other variables in the element(Ii, [Xj,1,.., Xj,n], Xj,i) ensure that the variables in
set are mapped to this variable. For example, unifica- the subsumption models are properly unified based on
tion X2 =X3 can be represented by mapping X3 to X2. the decision about unifications done in the unification
The constraint model uses index variable Ii for each model. These constraints can be posted in advance all
variable Xi in the template to describe the mapping. together as we require all positive examples to be sub-
The domain of Ii is {1,. . .,i} (variable Xi can only be sumed. However, for the negative example, we require
mapped to itself or to some preceding variable). To that the corresponding CSP has no solution. This can
express that variables Xi and Xj are unified we simply be ensured by trying to solve the CSP for the negative
post a constraint Ii = Ij (both variables are mapped to example (while respecting the so far decided unifica-
an identical variable). To ensure that each variable is tions) and if there is any solution found, this solution is
mapped to the first variable in the set of unified vari- broken by additional unification introduced to the
ables we use a constraint unification model (which is immediately propagated
to the constraint models for positive examples to
8i ¼ 1; . . . ; n elementðIi ; ½I1 ; . . . ; In ; Ii Þ; validate whether it does not conflict the positive
780 C Constraints on Learning
concept. This incremental process involves the three- textual databases as sources of observations, and use
step concept formation cycle, that is, the abstraction, propositionalization, a form of abstraction, and aggre-
integration (generalization), and assigning a concept gation techniques to transform representation space
definition/name, which introduces a new concept (Kietz and Morik 1994; Perlich and Provost 2006).
(descriptor).
A fundamental role in the constructive induction Important Scientific Research and
process plays the representation space, as it provides Open Questions
context for learned concept definitions. The target Independent of the strategy for generating new descrip-
concept never changes. Its definition does change tors is control over conceptual vocabulary. Machine
according to essential characteristics found in the space. learning methods can surely benefit from human expe-
In the formation of any concept, the capability of making rience in this regard, where philosophy is the founda-
comparisons among observations is critical. In this con- tion of science, and epistemology is the foundation of
text, “similarity is the relationship between two or more philosophy. The requirements of cognition determine
observations that possess the same characteristics but in the objective criteria of conceptualization. They can be
different measure or degree” (Rand 1990). summed up best in the form of an epistemological
The process of abstraction can be realized by many “razor”: concepts are not to be multiplied beyond neces-
methods capable of conceptually clustering (grouping, sity – the corollary of which is; nor are they to be
agglomerating, sorting) observations. The need and integrated in disregard of necessity (Rand 1990).
complexity for abstracting depends on the current Another challenge for constructive induction is
state of the representation space: starting with empty naming of constructed concepts. The challenge is not
representation space (with no observations and no in assigning a name, as such can easily be generated by
descriptors) to fully developed, that is, having examples a computer program, rather assigning a meaningful
described in terms of relevant descriptors, where it is name, understandable in human language. This might
a matter of selecting terms for building consistent and require some human–computer interaction to allow
complete descriptions with regard to the learning task. understanding of concept definition or the factors
Constructive induction term was first used in that contributed to differentiation of concept examples
machine learning, specifically in the context of con- from other examples and then scope of generalization
cept learning from examples (Michalski 1978). Con- of the selected concept examples.
structive induction systems may employ different The underlying premise included in the above
strategies for generating new descriptors. Based on challenges is the need for maintaining synergy in
the primary strategy employed, the systems can be human–computer interaction to preserve consistency
divided into four categories: data-driven, hypothesis- in modeling real-world problems.
driven, knowledge-driven, and multistrategy (Wnek
Cross-References
and Michalski 1994). Data-driven constructive induc-
▶ Concept Learning
tion systems analyze and explore the observations,
▶ Conceptual Clustering
including the interrelationships among their descrip-
▶ Feature Selection (Unsupervised Learning)
tors, and on that basis introduce changes in the
▶ Inferential Theory of Learning
representation space. Hypothesis-driven constructive ▶ Machine Learning
induction systems incrementally transform the repre-
▶ Multistrategy Learning
sentation space by analyzing inductive hypotheses
generated in one iteration and then using detected References
patterns as attributes for the next iteration. Knowl- Kietz, J.-U., & Morik, K. (1994). A polynomial approach to the
edge-driven constructive induction systems apply constructive induction of structural knowledge. Machine Learn-
expert-provided domain knowledge to construct new ing, 14, 193–217.
Michalski, R. S. (1978). Pattern recognition as knowledge-guided com-
terms. Multistrategy constructive induction systems
puter induction (Tech. Rep. No. 927). Urbana-Champaign: Uni-
combine different approaches and methods for versity of Illinois, Department of Computer Science.
constructing new terms. Real-world applications of Michalski, R. S. (1983). A theory and methodology of inductive
constructive induction systems utilize relational or learning. Artificial Intelligence, 20, 111–161.
Constructivist Learning C 783
Theoretical Background
Constructivism: Sociocultural Constructivist learning has emerged as a prominent
Approaches approach to learning and teaching on the basis of the
work by John Dewey (1858–1952), Jean Piaget (1896–
▶ Meaning Development in Child Language: A
1980), Lev Vygotsky (1896–1934), Jerame Bruner
Constructivist Approach
(1915–), Maria Montessori (1870–1952), and Ernst
von Glasersfeld (1917–), who, among others provide
historical precedents for constructivist learning theory.
Constructivist learning claims that learners do not just
Constructivist Agents absorb information. Instead, learners construct infor-
mation by actively trying to organize and make sense of
▶ Anticipatory Learning Mechanisms it in unique ways.
In the literature of constructivism in education,
there are many types of paradigms including
cognitive/personal constructivism, social constructiv-
Constructivist Learning ism, radical constructivism, critical constructivism,
cultural constructivism, genetic epistemology,
AYTAC GOGUS constructionism, information-processing constructiv-
Center for Individual and Academic Development, ism, interactive constructivism, cybernetic systems,
Sabanci University, Istanbul, Turkey and sociocultural approaches to mediated action.
Although there are many paradigms with different
emphases, they share many common perspectives
Synonyms about teaching and learning. These common perspec-
Effective learning; Meaningful learning tives provide the basis for constructivist learning.
784 C Constructivist Learning
Constructivist learning is articulated in contrast to The major foundation for cognitive constructivist
objectivist learning. Jonassen (1999) compares objectiv- approaches to teaching and learning is Piaget’s theory of
ist conceptions of learning by constructivist concep- cognitive development (1973), which describes how
tions of learning: children develop cognitive abilities. Piaget’s theory of
cognitive development (1973) has two major parts:
" Objectivist conceptions of learning assume that knowl-
ages and stages. According to Piaget there are four
edge can be transferred from teachers or transmitted
stages through birth to 12 years: the sensorimotor
by technologies and acquired by learners. Objectivist
period (birth to 2 years), preoperational thought (2–
conceptions of instructional design include the analy-
6/7 years), concrete operations (6/7–11/12 years), and
sis, representations, and resequencing of content and
formal operations (11/12 to adult). According to Piaget
tasks in order to make them more predictably and
(1973), learners must construct their knowledge
reliably transmissible.
through experiences by relying on ▶ cognitive struc-
Constructivist conceptions of learning, on the
ture (i.e., schemas and mental models). These cognitive
other hand, assume that knowledge is individually
structures are changed and enlarged through three
constructed and socially coconstructed by learners
complementary processes of ▶ assimilation, ▶ accom-
based on their interpretations of experiences in the
modation, and correction. Within Piagets’s theory
world. Since knowledge cannot be transmitted, instruc-
(1973), the basis of learning is discovery: “To under-
tion should consist of experiences that facilitate knowl-
stand is to discover, or reconstruct by rediscovery, and
edge construction. (p. 217)
such conditions must be complied with if in the future
Objectivist learning expects that teachers trans- individuals are to be formed who are capable of
mit knowledge and learners replicate the presented production and creativity and not simply repetition”
content and gain the same understanding as the (Piaget 1973, p. 20).
teacher (Jonassen 1999). Objectivist learning approach Social constructivism suggests that reality takes on
assumes that learner can gain the same understanding meaning which is formed and reformed through the
when systematic rules are used for logical conclusion. social process. Vygotsky’s constructivist theory (1978),
Therefore, objectivist learning does not provide appro- which is called social constructivism, emphasizes the
priate training for creative thinking, higher-order importance of culture and social context for cognitive
problem solving, transferring and applying knowledge development. Vygotsky’s zone of proximal development
to concrete experiences. On the other hand, construc- (1978) concept argues that learners can master con-
tivist learning provides an opportunity for reflection cepts, which they cannot understand on their own,
and critical thinking to make sense of the world and with help from instructors and peers. Vygotsky (1978)
create understanding, not just the memorizing of right divides the child’s language development into three
answers (Brooks and Brooks 1999). Learning is a search stages (at age 2, 3, and 7). In each stage, the child
for meaning, which requires understanding of the learns through observing and interacting with his/her
whole content as well as its parts, so the learning pro- immediate social environment. According to Vygotsky
cess focuses on individual understanding, not isolated (1978), the culture provides the cognitive tools to the
facts (Brooks and Brooks 1999). child for development such as cultural history, social
Two major types of the constructivist learning context, language, and technology. Adults such as
perspectives are cognitive constructivism and social con- instructors and parents guide learning by means of
structivism. While Piaget (1973) developed the cogni- these cognitive tools. The type and quality of
tive constructivism view of learning, Vygotsky (1978) these tools play an important role on learning and
developed the social constructivism view of learning. development.
These two constructivist view of learning are different Like Piaget and Vygotsky, Bruner studied cognition
in emphasis, but there is also a great deal of overlap and language learning in young children and defined
between them. Vygotsky shares many of Piaget’s learning as an active process in which a learner
assumptions about how children learn, but Vygotsky constructs new ideas or concepts based upon his/her
puts more emphasis on the social context of learning. both current and past knowledge by selecting and
Constructivist Learning C 785
transforming information, constructing hypotheses, world, so people create their own mental models to
and making decisions through representing individual make sense of their experiences. Also, constructivist
experiences in a cognitive structure (Brooks and learning emphasizes that the social and cultural context
Brooks 1999). In the constructivist classroom, the has a huge impact on learning. Therefore, learning is
instruction should be surrounded by an active dialog defined as a social process in which learners share, C
between the instructor and student while the instructor compare, and reformulate ideas to restructure new
tries to encourage students to discover principles by understandings. If an instruction or training allows
themselves. Therefore, the role of the instructor is to learners to exchange their personal views and test
translate information to be learned into a format them with others’, learners can build their own under-
appropriate to the learner’s level of understanding so standings with empirical evidence through activities
that the student constantly builds upon what he/she and observations. Since learners’ level of potential
has already learned. development has a critical impact on understanding,
Dewey (1966) is a reformer in educational policy. learners’ cognitive maturity, their interests, previous
He emphasizes that schools should not focus on repet- experiences should be considered in instructional
itive, rote memorization and that they should be design besides their social, cultural, and other contex-
engaged in real-world, practical training to be able to tual characteristics. Therefore, constructivist learning is
demonstrate their knowledge through creativity and defined as both an individual and a social process.
collaboration. Dewey’s book “Democracy and Educa-
tion” (1966) states that processes of instruction should
Important Scientific Research and
focus on the production of good habits of thinking so
that students should have opportunities to think them-
Open Questions
Constructivist learning views learning as a social activ-
selves and articulate their thoughts. According to
ity. Learning is influenced by social interaction and the
Dewey, students should be involved in meaningful
language that the learner uses. Besides social interac-
activities and apply the concepts they are trying to
tion and language, other major factors that influence
learn. Dewey (1966) uses term active learner by
learning is learner’s pervious knowledge, learner’s
stressing that learning is an active process in which
motivation, and learner’s characteristics such as beliefs,
the learners construct their own meaning. In other
prejudices, and fears. These factors are associated with
words, learning is not a passive acceptance of presented
individual, social, and cultural aspects of learning.
knowledge by teachers, but is constructing meaning.
Constructivist learning requires educators to think
Constructing meaning happens in the mind; therefore,
about epistemology and pedagogy to be able to allow
educators should design both hands-on activities and
learners construct knowledge individually and socially.
mental activities. Dewey (1966) emphasizes that learn-
In order to teach well, educators must understand the
ing happens through reflective activities as a product of
mental models to support and challenge the learner’s
critical thinking. Learners should reflect on what they
thinking (Brooks and Brooks 1999). Brooks and
understand.
Brooks (1999) list 12 characteristics for teaching by
According to von Glasersfeld (1996), the human
implementing constructivist learning theory into class-
mind can only know what the human mind has
room instruction:
made. In radical constructivist approach, there is an
important point that how we know is more essential ● Encourage and accept student autonomy and
than what we know. There are two main principles of initiative.
radical constructivism (von Glasersfeld 1996): ● Use raw data and primary sources along with
(1) knowledge is not passively received but actively manipulation, interaction, and physical materials.
built up by subject; (2) cognition is adaptive and serves ● Use cognitive terminology such as “classify,” “ana-
the organization of the experiential world, not the lyze,” “predict,” and “create” when assigning tasks
discovery of ontological reality. to the students.
In summary, constructivist learning emphasizes ● Allow student responses to drive lessons, shift,
that people construct their own understanding of the instructional strategies, and alter content.
786 C Constructivist Learning Environments
● Inquire students’ understanding of concepts before their experiences. The teachers’ understanding of the
sharing their own understanding of these concepts. approach is limited to their personal experiences.
● Encourage students to engage in a dialog both with Therefore, teachers should gain a proper understand-
the teacher and with one another. ing of constructivist philosophy and approaches in
● Encourage student critical thinking and inquiry order to create effective instructional activities that
by asking thoughtful, open-ended questions, and are reflective of constructivist orientation.
encourage students to ask questions to each other.
● Seek elaboration of student’s initial response. Cross-References
● Engage students in experiences that might engender ▶ Active Learning
contradictions to their initial hypotheses and then ▶ Bruner, Jerame (1915–)
encourage discussion. ▶ Constructive Learning
● Allow wait time after posing questions. ▶ Dewey, John (1858–1952)
● Provide time for students to construct relationships ▶ Piaget, Jean (1896–1980)
and create metaphors. ▶ Piaget’s Learning Theory
● Nurture students’ natural curiosity through fre- ▶ Project-Based Learning
quent use of the learning cycle models. ▶ Social Construction of Learning
Constructivist learning allows students to take ▶ Social Learning Theories
responsibility for their own learning and establish con- ▶ Vygotsky’s Philosophy of Learning
nections between ideas and thus to analyze, evaluate,
and defend their ideas (Brooks and Brooks 1999). References
Jonassen (1999) describes constructivist learning Brooks, J. G., & Brooks, M. G. (1999). In search of understanding: The
environment (CLE) as having eight characteristics: case for constructivist classrooms. Alexandria: ASCD - Association
for Supervision and Curriculum Development.
active/manipulative, constructive, collaborative, conver-
Dewey, J. (1966). Democracy and education. New York: Free Press.
sational, reflective, contextualized, complex, and inten- Jonassen, D. (1999). Designing constructivist learning environments.
tional. Construction of knowledge by learners should In C. Reigeluth (Ed.), Instructional design theories and models:
have these eight qualities. In constructivist learning A new paradigm of instructional theory (Vol. II, pp. 215–239).
environment, the student as an active learner mediates Mahwah: Lawrence Erlbaum Associates.
Piaget, J. (1973). To understand is to invent. New York: Grossman.
and controls learning by engaging in meaningful
von Glasersfeld, E. (1996). Introduction: Aspects of constructivism.
social interaction with other students and teacher. In C. Fosnot (Ed.), Constructivism: Theory, perspectives, and
The teacher as a moderator provides students with practice (pp. 3–7). New York: Teachers College Press.
variety of activities that promote collaboration, inter- Vygotsky, L. (1978). Mind in society: The development of higher psy-
action, reflection, experimentation, interpretation, and chological processes. MA: Harvard University Press.
construction (Brooks and Brooks 1999). The challenge
for educators is to design instructional strategies to
actively engage learners in knowledge construction,
being able to negotiate meaning, and solving complex
problems (Jonassen 1999). Constructivist Learning
Educators at all levels have tried to improve their Environments
instructional practices through experimenting with
▶ Open Instruction and Learning
constructivist learning principles. This is because con-
structivism focuses on how people learn and it sug-
gests that learning occurs through active engagement in
problem solving, and not simply from taking in infor-
mation, replicating the information. The challenge Constructivist Learning
in teaching is to create experiences that engage stu-
Principles
dents in learning activities and support their own
explanation, evaluation, and communication about ▶ Cybernetic Principles of Learning
Content-Area Learning C 787
Definition
Constructivist Learning Theory Content area learning is closely associated with the
Constructivist learning theory takes on several forms – academic skills and instructional pedagogies needed
individual, social, cognitive, postmodern – but all to succeed within various core academic courses. Each
emphasize that learners construct knowledge using content area has its own traditions, knowledge base, C
their own activities, and that they interpret concepts and pedagogies, including strategies for teaching and
and principles in terms of the schemata that they have learning. Content area learning is typically driven and
already developed. The verbs used in constructive defined by the unique learning traditions of the four
alignment (above) are the “construction tools” that major specialty areas taught in secondary schools (i.e.,
students use to meet the learning outcomes intended. social studies, mathematics, science, English Language
Arts); although content area learning can also refer to
learning that takes place in other courses (e.g., art
history, business-related courses). Furthermore, con-
tent area teachers play a substantial role in shaping
Contemplation learning goals and instructional pedagogies within
▶ Mindfulness and Meditation their courses via individually constructed understand-
ings of the discipline’s knowledge and learning tradi-
tions. Content area specialists in secondary (middle
and high) schools typically demonstrate proficiency
Contemplative Science in their area through completion of a full degree or
other closely related coursework. Several nations main-
▶ Mindfulness and Meditation tain the tradition of content area teachers completing
an adjoining degree in teaching, which leads to national
or local professional licensure.
Within each content area, there is voluminous
Content Area Literacy information to transfer to students; teachers typically
organize content within units of study guided by
▶ Content-Area Learning national or local standards and other curricular guide-
lines. A common understanding of content area learn-
ing is that each discipline’s academic demands revolve
around students’ cognitive and academic capacity (e.g.,
Content-Addressable Memory reading comprehension, writing skills) to efficiently
navigate voluminous background knowledge and liter-
▶ Associative Memory and Learning acy demands. Rich background knowledge helps facil-
itate student constructions of conceptual knowledge
within a particular discipline; however, this cognitive
construction process is often enabled or restricted
Content-Area Learning by students’ literacy skills. Students need strong liter-
acy skills to successfully interact with the substantial
DONALD D. DESHLER, BELINDA B. MITCHELL, demands generated by academic coursework, but simul-
MICHAEL J. KENNEDY, LESLIE NOVOSEL, FRANCES IHLE taneously need strong literacy skills to demonstrate
Department of Special Education, University of understanding and proficiency on assessments and
Kansas, Lawrence, KS, USA other course requirements.
Embedded within the cognitive construction of
content area material is (a) a requirement to master
Synonyms significant quantities of vocabulary terms, including
Adolescent Literacy; Content Area Literacy; Disciplin- complex concepts through a combination of reading
ary Literacy and in-class instruction; (b) the need for students
788 C Content-Area Learning
to develop and use metacognitive strategies for in secondary-level content courses, and contain signif-
interacting with the discipline’s content; and (c) the icant quantities of vocabulary terms and concepts,
teacher’s roles in providing explicit instruction that and many are written with the assumption that readers
structures readiness to process, comprehend, and possess vast background knowledge about the
critique discipline-specific texts. Furthermore, as stu- corresponding content area (Kamil et al. 2008). The
dents progress within their respective academic pro- mixture of significant quantities of new vocabulary
grams, content area learning demands, especially those terms and “newness” of content without existing cog-
related to literacy, are augmented in significant quan- nitive schemas to activate during learning can hinder
tities (Carnegie Council on Advancing Adolescent Lit- the comprehension of all students, but especially those
eracy 2010). In summary, content area learning is a with reading or other learning challenges.
multifaceted, generative process shaped by the knowl- Content area courses employ discipline-specific
edge construction traditions unique to each content texts as well as primary source documents (e.g., journal
area, but is largely dependent on learners’ basic literacy entries, laboratory notes, letters, and policy docu-
and higher-order thinking skills and processes (Kamil ments). Primary source documents often contain
et al. 2008). wide disparities in terms of text structure, purpose,
use of jargon, and levels of reading difficulty. In short,
Theoretical Background these documents, while rich in terms of building stu-
Foundational Literacy Skills. The foundation for success dents’ conceptual understanding of persons, events,
in content area learning is strong basic literacy skills. and processes, can be very difficult to read with the
To succeed in content-specific courses, especially at the efficiency and effectiveness needed for success in class-
secondary level, a strong base of literacy skills, includ- rooms. Content area teachers need to provide students
ing comprehension (and its component parts), writing, with explicit instruction in order to successfully engage
and capacity to participate in discourse, is required challenging discipline-specific texts and related docu-
(Kamil et al. 2008). The demands of content area ments. Promoting and nurturing student metacogni-
courses, especially at the secondary level, frequently tion is a critical element of content area learning that
include substantial demands of discipline-specific must be carefully addressed by teachers.
texts in terms of reading levels and overall accessibility Another commonality across the content areas is
of document(s). Policymakers, researchers, practi- the quantity of information that is transmitted from
tioners, and other stakeholders have brought signifi- the teacher (and text and other materials) to students.
cant attention to the issue of literacy learning within Content area courses are built from and organized
the content areas across the past 30 years (Carnegie around standards derived from national and/or local
Council for Advancing Adolescent Literacy 2010). education agencies. In the current age of accountabil-
Despite this attention, effective interventions for pro- ity, the demands for content coverage have been
moting content area learning for all subpopulations augmented, while available instructional time has
have not been widespread. Thus, for many practi- often not kept pace. This has resulted in what is
tioners and teacher preparation programs, the empha- sometimes called a “pedagogy of telling” which
sis for content area teachers revolves around the results in teachers being compelled to cover a wide
content and methods for teaching, while readiness to breadth of material by way of sacrificing depth of
promote literacy instruction is marginalized (Moje understanding. For many students, including those
2007). This area is currently receiving substantial with learning challenges, the pedagogy of telling is
attention in the professional literature, and will con- a mismatch for their learning needs and preferences,
tinue to be a source of new research and innovation and frequently results in academic struggles and fail-
in the future. ures. A critical area of research in the field is the
Commonalities Across Content Area Learning. There design and validation of instructional methods for
are several characteristics of content area learning addressing the compounding demands of state and
that remain constant despite preexisting differences in local standards and curricula while not ignoring best
subject matter. The first hallmark of content area pedagological practices within each respective disci-
learning is the use of texts. Textbooks are a mainstay pline area (Moje 2007).
Content-Area Learning C 789
Teacher Preparation. Shulman (1987) pedagological corroboration, helps teachers explicitly teach students
content knowledge (PCK) is a well-known construct to think and act as historians do, which promotes
for organizing and understanding how content area deeper engagement with content. In order to partici-
teachers make sense of content and select/design pate in the active discourse of content it is necessary for
appropriate methods for teaching students. Methods students to possess adequate background knowledge, C
for organizing this content to efficiently and effectively along with sufficient literacy skills to engage complex
convey content to students (pedagogy) differ from text-based documents or materials. In science courses,
content area to content area. For teacher educators students are often taught to participate in active
and practitioners, understanding how PCK informs inquiry activities, which require significant know-how
teaching in the respective content areas is essential. on the part of students that has been transmitted from
That said, students’ basic literacy skills are essential their teachers.
for proficient content area learning; hence, teachers Shulman’s PCK and Moje’s DLP frameworks help
have the dual responsibility of promoting literacy inform the work of content area teachers by specifying
learning in the service of enabling advanced content and promoting the norms and practices of experts in
mastery (Kamil et al. 2008). the respective fields. These norms and practices guide
Teacher preparation programs for content special- orchestration of hands-on learning made possible
ists feature coursework on discipline-specific teach- through underlying background knowledge and liter-
ing methods, but frequently are limited in terms of acy tools needed to promote active and independent
disciplinary literacy pedagogy (DLP) (Moje 2007). student learning. This constructivist approach con-
Moje’s construct of DLP is an expansion of Shulman’s trasts significantly from traditional “stand and deliver”
PCK in that it reflects the need to specifically highlight and pedagogy of telling methods that rely on passive
and implement discipline-specific knowledge creation learning on the part of students.
traditions from the respective content areas. Highlight-
ing and explicitly teaching students the epistemo- Important Scientific Research and
logical practices of the discipline is a substantial Open Questions
departure from many traditional approaches to con- Research in the field of content area learning can be
tent area learning. In addition, a DLP framework characterized as the study of methods for helping
focuses on infusing specific literacy instruction into teachers organize and deconstruct content (graphic
teaching, so that students develop capacity within the organizers, vocabulary instruction) in order to help
respective content areas to meaningfully interact with students access subject matter. Research that is needed
course texts. This epistemological approach is essential in this area includes furthering understanding of
for understanding content area learning, as the capacity how discipline experts can capture the metacognition
to understand the respective content area traditions involved in knowledge construction and critique that
and construction methods holds the key for higher- translates into instructional materials relevant for nov-
order thinking and learning on the part of students. ice learners. In addition, significant attention in the
That said, publicizing the discipline’s respective episte- professional literature has been dedicated to strategy
mologies and translating those traditions into struc- instruction. Students can become cognitively active
tured lessons for novice learners is an extraordinarily when they are explicitly taught how to learn and are
complex undertaking for many practitioners, and con- given authentic opportunities to engage in the learn-
stitutes an area of significant need within the field of ing activities of experts in the respective disciplines
content area learning in terms of future research and while receiving ongoing feedback from teachers. Fur-
innovation (Moje 2007). ther research is needed in order to create teacher
Given the unique demands of each content area, preparatory programs and practices that lead to the
teachers must make decisions about how to deliver preparation of content area teachers to implement a
their content in the most efficient and effective disciplinary literacy framework in their classroom.
manner. In social studies courses, for example, Research in strategy instruction as it relates to
Weinberg’s (1991) framework for knowledge con- content area learning has focused on cognitive and
struction, (a) sourcing, (b) contextualization, and (c) metacognitive processes that learners must employ in
790 C Content-Extending Reasoning
important requirement for modern information and discussed with regard to speech behavior (see, e.g., the
communication systems where context sensitivity context-sensitive associate theory of Wickelgren 1969)
of applications refers to the adaptivity to the situ- and word processing (Schvanefeldt and McDonald
ations in which a system needs to act. This enables 1981); and in the area of machine learning context-
more efficient and robust functioning in dynamic sensitive learning methods for text categorization are C
environments. comparable (e.g., Cohen and Singer 1999).
In contrast with schema-based argumentations,
Theoretical Background researchers in the field of mental models argue that
A basic assumption of constructivist approaches of
context sensitivity occurs consciously and inten-
learning is that learners respond sensitively to charac-
tionally. Among others, Anzai and Yokoyama (1984)
teristics of the environment, “such as the availability of
assume that learners encode information on a problem
specific information at a given moment, the duration
in a mental model as soon as they begin working on it
of that availability, the way the information is struc-
in order to gain a basic understanding of the situation
tured” (and presented), “and the ease with which it
and its demands. This initial experiential model can –
can be searched” (Kozma 1991, p. 180). However, this
and the learner is generally aware of this – be false or
seems dependent on the learning strategies which
insufficient for accurately representing the subject
students use in a more or less consistent manner.
domain in question. However, it is semantically sensi-
Entwistle (1981), for example, assumes that some
tive toward key stimuli in the learning environment
learners are more consistent in their use of strategies
and can thus be transformed into a new model through
while others behave more opportunistically or with
accurate processing and interpretation of these key
more sensitivity to the requirements of their immediate
stimuli. The results of the experimental study of
situation. This argumentation corresponds, to a large
Anzai and Yokoyama (1984) as well as those of other
extent, with the concept of a biological sensitivity to
studies (e.g., Ifenthaler and Seel 2005; Seel and Dinter
context as discussed in the area of psychopathology
1995) indicate the following characteristics of contex-
where biological reactivity to environmental stressors
tual semantic sensitivity in the learning-dependent
is widely discussed (see, e.g., Boyce and Ellis 2005).
progression of mental models:
Furthermore, what Kozma says also corresponds with
the idea of contextual cueing within the realm of cogni- ● If the learner’s initial mental model is strongly
tive psychology, where powerful and sophisticated dependent on previous knowledge from experience
selection mechanisms exist to spontaneously focus rather than on acknowledged principles (for instance
on aspects of a complex scene that are of significant of physics) and if specific key stimuli from the learn-
relevance for information processing. For example, ing environment capture the learner’s attention, the
in complex visual search tasks, the global context initial experiential model is semantically sensitive
may direct attention toward specific elements involved toward these key stimuli and can be changed into
in the scene. This contextual guidance of visual a more correct model.
attention reflects context sensitivity to meaningful reg- ● Semantic sensitivity requires for the key stimuli to
ularities and covariances between objects within a be related to the knowledge on which the initial
scene (cf. Treisman and Gelade 1980). In accordance model is based. Key stimuli which capture the
with schema-theoretical approaches of information learner’s attention but are not related to the knowl-
processing, it has been argued (e.g., Chun and Jiang edge on which the “experiential model” is based are
1998) that relevant contextual knowledge is mainly ineffective for changing this mental model.
acquired through implicit learning processes which ● Which key stimuli in the environment are taken
occur without intention or awareness. Incidentally into account for the further development of a
acquired contextual knowledge forms a highly robust, model depends primarily on the learner’s domain-
instance-based, implicit memory for context and con- specific knowledge. The mental model created at
stitutes the fundamental basis of contextual cueing the beginning of a problem-solving process is only
as a form of schema-based automaticity. Similarly, sensitive toward key stimuli in the environment if
in cognitive psychology context sensitivity has been the learners are able to recognize the principles
792 C Context and Semantic Sensitivity in Learning
which their knowledge indicates to being relevant The experimental condition for the second problem
for mastering the given situation. was designed in the same way.
The results of both experiments demonstrate the
We can summarize that context and semantic
effectiveness of the analogues on display in the exper-
sensitivity is widely accepted in various disciplines
imental room and thus also the context sensitivity of
which are concerned with learning and information
the learners. However, the learners’ context sensitivity
processing. Not only cognitive and educational psy-
was only effective when they were given enough time to
chology operates with this theoretical term but also
solve the problems.
linguistics, machine learning, and artificial learning
Similarly, some research on mental models focused
research.
on the effects of semantic sensitivity on learning and
problem solving. A prominent example has been pro-
Important Scientific Research and vided by Anzai and Yokoyama (1984) who distinguish
Open Questions between the stage of initial model construction and
Context sensitivity is fundamental to intelligent behav-
the subsequent process of model development which is
ior. It is the context of the learners that determines
dependent on the learner’s semantic sensitivity toward
which stimuli will be perceived, how interpretations
relevant key stimuli in the learning environment. The
are placed on incoming information, and how the
starting point of their study was the observation that
learner responds to the stimuli. By paying attention to
many students have only fragmentary knowledge of
the context, an intelligent agent can spontaneously
physics. Although they acquire a good deal of formal
select appropriate responses to stimuli, especially of
knowledge in school, they are unable to apply this
unanticipated events.
knowledge to new physics problems. Rather, they
From an educational perspective, an excellent
tend to devote their attention to, and to let themselves
example for illustrating the relevance of context sensi-
be distracted by, surface attributes of the problems
tivity for learning is an experimental study of Dreistadt
and end up forming naive internal representations
in 1969 (Dreistadt 1969). In this study, adults had to
on the basis of these attributes. However, as Anzai
solve two well-defined problems: In the first problem,
and Yokoyama could show that students sensitively
the subjects had to separate the area of a farm into four
adjusted their mental models to particular information
parts of equal size and shape (see Fig. 1), and in the
provided by the learning environment. This observa-
other problem they had to plant 20 trees in five straight
tion could be replicated in experimental studies done
rows of four trees each. For the experimental condition,
by Seel and Dinter (1995) and Ifenthaler and Seel
Dreistadt provided pictures of various objects in the
(2005).
experimental room which supplied analogues to the
Alternatively to this research on mental models,
given problems and indicated an idea for a solution.
various schema-theoretical approaches demonstrate
The first experiment provided the subjects with a
also the effectiveness of context sensitivity, for instance,
map of the USA on which Texas and several flight
in visual processing as well as verbal processing. Actu-
paths were highlighted, a diagram with curves, and a
ally, many researchers have used concepts such as “con-
clock on a dresser half covered by a radio (see Fig. 2).
text” or “typicality” in order to explain the influence of
knowledge structures on processing visual and verbal
stimuli (see, e.g., Antes et al. 1981; Schvanefeldt and
McDonald 1981; Treiman, Kessler and Bick 2002).
When we view a visual scene, we are able to determine
rapidly and effortlessly the scene’s constituent objects,
spatial relations, and to what semantic class the scene
belongs. This corresponds largely with the schema
hypothesis, according to which a visual scene is rapidly
Context and Semantic Sensitivity in Learning. Fig. 1 The identified as a member of a semantic category, and
farm problem in Dreistadt’s (1969) study on the use of contextually sensitive predictions are then used for
analogies subsequent object identification (Henderson 1992).
Context and Semantic Sensitivity in Learning C 793
11 12 1
10 2
9 3
8
7
Context and Semantic Sensitivity in Learning. Fig. 2 The provided analogies for solving the farm problem
(Dreistadt 1969)
According to this hypothesis, schemas function as contextual associations between an object and other
a framework which promotes context-bound under- objects with which it typically appears. In addition to
standing and coherence when we process visual object-based facilitation, a context-based mechanism is
information. At the same time, schemas regulate the proposed to trigger top-down facilitation through con-
attention we devote to information depending on textual associations between objects in scenes. Fenske
whether it is related to a schema or not. Everyday et al. point out that object- and context-bound top-
experiences and observations by psychologists indicate down processes operate together in promoting efficient
that information which is atypical for a schema attracts recognition by framing early information about a
more attention and is thus more likely to be retained. visual scene within the constraints provided by a life-
Context sensitivity does not only play an important time of experience with contextual associations.
role in psychological research on visual and verbal
processing but rather also in the field of machine learn- Cross-References
ing and Artificial Intelligence where schema-based ▶ Anticipatory Schema(s)
approaches of context-sensitive reasoning are popular ▶ Schema(s)
since the 1990s (see, e.g., Cohen and Singer 1999; ▶ Schema-Based Reasoning
Turner 1994; Turney 1996). ▶ Visual Perception Learning
Schema-based approaches of context sensitivity ▶ Word Learning
operate basically with a top-down mechanism in suc-
cessful recognition as discussed in recent neuropsy- References
Antes, J. R., Penland, J. G., & Metzger, R. L. (1981). Processing global
chological models and research findings (e.g., Fenske
information in briefly presented pictures. Psychological Research,
et al. 2006). Actually, there is sufficient evidence for 43(3), 277–292.
top-down facilitation of recognition that is triggered Anzai, Y., & Yokoyama, T. (1984). Internal models in physics problem
by early information about an object, as well as by solving. Cognition and Instruction, 1, 397–450.
794 C Context Awareness
of food would be a discrete association between the of context conditioning may involve different pro-
food and the illness. However, returning to the restau- cesses. Background context conditioning may require
rant in the absence of the offensive food may also be more attention or vigilance in order to form a strong
sufficient to evoke nausea. This would be because the context association because the discrete CS may be in
context of the restaurant has become associated with competition with the context for cognitive resources. In C
the illness. As stated, in context conditioning, the indi- support of the idea that background and foreground
vidual stimuli that compose the context are bound conditioning involve different processes, inhibition of
together as a gestalt such that an individual stimulus protein synthesis immediately after training disrupted
from the environment may not be sufficient to evoke foreground but not background contextual fear condi-
a response, but when the contextual stimuli are pre- tioning (Stiedl et al. 1999). Because foreground and
sented as a whole, a strong response is evoked. background context conditioning may involve differ-
Research examining the neural substrates of classi- ent processes, experimental variables may not have the
cal conditioning suggests that the classical conditioning same effects on each type of conditioning. This fact
of discrete stimuli such as a tone and the classical should be considered when designing and interpreting
conditioning of contextual stimuli may involve differ- experiments.
ent systems. In classical fear conditioning, lesions of the Other forms of context conditioning exist in addi-
amygdala disrupt both the conditioning of a discrete tion to contextual fear conditioning. Conditioned place
stimulus and the conditioning of the training context; aversion is a type of context conditioning in which
however, lesions of the dorsal hippocampus only dis- subjects are exposed to different contexts that are sep-
rupt context conditioning, leaving conditioning to a arated by an opaque Plexiglas wall. One context is
discrete stimulus intact. The hippocampus is involved repeatedly paired with a control substance such as
in processing contextual and spatial stimuli and is saline and the other side is repeated paired with
thought to play a role in binding stimuli together. a potentially noxious stimulus such as an aversive
Further suggestion that context conditioning and con- dose of a drug. After multiple trials, the Plexiglas
ditioning of discrete stimuli involve separate processes divider is removed and the time subjects spend in
comes from pharmacological studies demonstrating each context is measured. If the subjects have learned
that a drug can selectively affect one type of condition- to associate the noxious stimulus with the context in
ing without affecting the other. For example, nicotine which it was administered, they should spend less
administration enhances contextual fear conditioning time in that context. Just as a context can be associated
but not fear conditioning using a discrete auditory with aversive stimuli, context conditioning can occur
stimulus as the CS (Kenney and Gould 2008). If con- with appetitive stimuli. One example of this is condi-
text conditioning and conditioning with a discrete CS tioned place preference. The training of conditioned
involved the same processes, then they should be sim- place preference is similar to the training previously
ilarly affected by pharmacological manipulations or by described for conditioned place aversion except that
inactivation of brain regions. instead of pairing a noxious stimulus, an appetitive or
Just as there are multiple types of classical condi- rewarding stimulus is paired with one context. If the
tioning, there are multiple types of context condition- subjects form an association between the context and
ing. One distinction is whether the context is the that stimulus, they should spend more time in that
primary CS or a secondary CS. Using classical fear context at testing. The paradigm is often used to exam-
conditioning again as an example, when a discrete ine how drugs of abuse become associated with con-
auditory CS is paired with a mild shock US, the context textual information and the processes that support this
is a secondary CS. The context conditioning in this case type of learning.
would be background context conditioning (Odling- Another context association that can occur with
Smee 1975). However, if no discrete CS is paired with drugs of abuse is context conditioned tolerance. With
the US, the context becomes the primary CS; this is repeated administration of a drug, the same dose of the
called foreground context conditioning. The distinc- drug may come to elicit less of a response; this is known
tion between foreground and background context con- as tolerance. When a drug is repeatedly administered in
ditioning is an important one because these two types the same context, the context can become associated
796 C Context Conditioning
with the drug administration and this context-specific pharmacological classes (e.g., stimulant versus halluci-
association can lead to the expression of tolerance. nogen) do not substitute their control of the condi-
However, because the context is controlling the expres- tioned response unless they share a common effect in
sion of tolerance, administration of the same dose in the nervous system. Unfortunately, there is very little
a novel context can result in an overdose. For example, research in this area with humans and it potential
in a study that examined conditioned tolerance, rats import. The limited research with nonhuman animals
were injected with doses of heroin that escalated over suggests that it could be quite important in such areas
time; injections occurred in one of two contexts and the as drug addiction and eating disorders.
paring of injection condition with context remained
stable throughout the experiment. Rats were then given Important Scientific Research and
a test dose that was nearly twice as high as the last dose Open Questions
of heroin administered. Rats given the test dose of One important issue for understanding context condi-
heroin in the environment in which heroin was previ- tioning is clarifying whether context conditioning is
ously administered were less likely to show signs of one learning process where the context becomes asso-
overdose than rats given the same dose in an environ- ciated with a stimulus or two different learning pro-
ment that was heroin naı̈ve (Siegel et al. 1982). In cesses where the context is learned as one process and
addition to the context being able to elicit tolerance, the representation of the context is then associated with
contextual stimuli can also elicit cravings. Environ- a US. In contextual fear conditioning, the context and
ments associated with self-administration will evoke the US are presented during the same trial, which
drug-seeking behaviors in rodents and reports of drug makes it difficult to determine if learning a context is
craving in humans. This ability of the context to con- different from context conditioning. This issue has
trol the expression of tolerance and cravings has serious been clarified through a series of experiments that
implications for understanding and treating addiction. demonstrates that context learning and context condi-
As just one example, treating a patient for substance tioning can occur as separate processes. If a rodent is
abuse in a clinic and then returning them to the envi- put into a conditioning chamber, immediately admin-
ronment where they consumed the drugs may greatly istered the unconditioned stimulus, and then removed;
increase the likelihood of relapse. the rodent does not show robust context conditioning,
Drugs such as the heroin discussed in previous par- though conditioning can occur and changes in exper-
agraph have perceptible interoceptive effects. These imental design can change this outcome. However, if
perceptible effects can serve as an internal contextual the previous experiment is repeated except this time
stimulus much like the exteroceptive cues that compose the naı̈ve subject is also allowed to passively explore
the room where the addict takes drug, or the chamber the training context on the day before the immediate
where experiments are conducted. Like exteroceptive conditioning, context conditioning results. This dem-
context stimuli, the internal context induced by a drug onstrates that for context conditioning to occur, the
can acquire control of approach or avoidance-related context must be first learned and then entered into an
conditioned responses when the drug state is paired association with the unconditioned and suggests that
with an appetitive or aversive stimulus, respectively the context learning and the context conditioning may
(Bevins and Murray 2011). As an example, rats can be separate processes (Fanselow 2000).
receive daily nicotine sessions intermixed with daily Contextual associations play an important role in
saline sessions. On nicotine sessions, sucrose is avail- several types of mental illness. One example in anx-
able intermittently; no sucrose is available on saline iety disorders includes posttraumatic stress disorder.
days. The internal context induced by the nicotine Contextual stimuli can become associated with a
comes to control an anticipatory approach and search stressful or anxiogenic event. Reexposure to these con-
in the area where sucrose had been previously given. textual stimuli can result in reexperiencing stress and
Research in this area has indicated that the internal anxiety. This becomes problematic if the repeated
context is specific to the neurobiological process exposure does not lead to a decrease or extinction of
underlying the drug. Thus, drugs within and across the stress and anxiety responses. In addition, further
Context Fear Learning C 797
Context fear learning is believed to be a two- information via its three-layered laminar structure to
stage process. First, through active exploration of the form a multi-modal spatial representation, or cognitive
experimental apparatus, the animal must integrate the map, of the environment. It is this cognitive map that
multimodal stimuli into a unified “contextual repre- serves as the contextual representation in context fear
sentation” that can be used as a CS. Second, this con- conditioning. The site of the context-shock associa-
textual representation is then associated with the tion is believed to be in the amygdala where hippocam-
aversive US. The context-shock association then sub- pal inputs and shock-related information converge.
sequently drives the fear CR. A number of phenomena Strengthening of the hippocampal-amygdala synapses
in contextual fear conditioning have led to this view. via Hebbian long-term potentiation allows subse-
Most important among these is the immediate shock quent activation of these inputs to drive amygdala
deficit. If the aversive US is presented immediately after activity. Amygdala activation by the contextual CS
the subject is placed in the context it will acquire no then activates downstream structures, such as the
contextual fear, and it thus exhibits the immediate peri-aquedectual gray (PAG), which coordinate the
shock deficit. Extensive experimentation has demon- fear response. Thus, when comparing context fear learn-
strated that this deficit occurs because the formation of ing with learning fear of a simple discrete cue such as
contextual representation has not yet occurred prior to a sound, the hippocampus is involved in context but not
the immediate shock and therefore there is no CS to cued fear. However, the amygdala is equally important
associate the shock with. Pre-exposure to the condi- for both types of fear.
tioning chamber prior to the immediate shock rescues There are a number of important caveats to this
the immediate shock deficit. This pre-exposure rescue view. The first is that context fear learning can readily
indicates that if the subject has already formed the occur in the absence of the hippocampus as long as
contextual representation it can retrieve this represen- more than one shock is presented during training.
tation prior to the immediate shock and thereby form Thus, animals with lesions or pharmacological inacti-
the context-shock association. The length of time vation of the hippocampus prior to training with
between placement in the conditioning chamber and two or more shocks can acquire normal levels of
presentation of the shock is referred to as the “place- contextual fear. Lesions after training or pharmacolog-
ment to shock interval” or PSI. Short PSIs produce ical inactivation prior to testing, however, consistently
little to no conditioning, as just described. As the PSI produce context fear deficits, even when multiple
is increased the level of conditioning increases, up to shocks are used. This discrepancy between pre- versus
approximately 3 min when the level of conditioning post-training manipulations has been interpreted in
becomes asymptotic. This placement to shock interval the following way: When hippocampal function is
function indicates that the formation of the contextual compromised during training, alternate structures are
representation occurs very rapidly, but is clearly not able to compensate and generate a contextual represen-
instantaneous. It requires integration of multi-modal tation that is sufficient to support conditioning. Post-
sensory experience over time. training lesions are more effective because normally the
hippocampus actively inhibits and/or outcompetes these
Important Scientific Research and alternate structures. Therefore when training occurs with
Open Questions an intact hippocampus, the alternate structures are not
A major area of current research is focused on deter- recruited. In addition, these alternate structures are less
mining the underlying neural mechanisms of context efficient, which is why more training is required in the
fear learning. The current view is that formation of the absence of the hippocampus. Furthermore, they are less
contextual CS occurs in the hippocampus, a region that accurate, resulting in inappropriate fear responses to
is critical in many forms of learning and memory. It other contexts that were not paired with shock. The
receives highly processed multi-modal sensory infor- exact site of these alternate structures is still a matter of
mation from the lateral entorhinal cortex and precise debate; however, the slower learning rate and reduced
spatial information from the medial entorhinal cortex. specificity are consistent with theoretical predictions of
It is believed to further process and integrate this learning in cortical structures.
Context-Based Learning C 799
best capture an overtly practical engagement with the However, as indicated, it is the actual practice
learning environment. The approach is still, in many of context-based learning that, for the most part,
ways, an immanent response to the day in day out reveals its commitments and implicit assumptions.
process of learning and transmitting knowledge, The method redefines the roles of both learner and
whereby new techniques have been developed bot- teacher: the former is to be actively involved in the
tom-up rather than inspired and determined by learning process and the latter is to facilitate the
preexisting theoretical commitments. And nowhere is learners’ taking possession of the knowledge for him
such a pragmatic approach more appropriate than in a or herself. The learning process is not about rote learn-
pedagogical methodology that seeks to integrate the ing of facts, but is interest governed in the sense that
interests of the would-be knower with the body of the learner perceives that there is something at stake
knowledge before him or her. in the learning rather than the mere propensity to pass
Of course, theoretical precedents have been sought an examination or gain credit. Learning is no longer
in the educational literature, and two thinkers above all seen as something happening to one, but an activity
others seem to stand out: John Dewey, the American in which one is engaged. Consequently, the teacher
pragmatist, and Lev Vygotsky, the Russian psycholo- becomes a facilitator or a supervisor of tyro researchers;
gist. From the former, context-based learning derives he or she is no longer a dispenser of facts and theories,
an overtly pragmatic commitment: learning is an activ- but an organizer of a social community of equal
ity bound up with human interests. Just as the standard learners.
of truth and knowledge in Dewey’s deeper philosophical As the learner becomes the center of their own
theory is their utility, a statement is true if it is useful, so educational experience situated within a communal
must learning prepare the subject for social engagement. group, so he or she reflects upon the first axis of the
In response to his theoretical commitments, Dewey context (the social environment), the object (the inter-
believed that education ought to be dominated by real- section of the knowledge with empirical reality), and
life tasks and challenges and that theory and facts were to the experience of learning. The advantage is obvious:
be learned through activity, rather than the standard through learning, the leaner is also learning to learn and
model of a passive student receiving knowledge from progresses from a dependent student to an indepen-
an expert or superior. The latter thinker, Vygotsky, dent subject. The second axis of the context concerns
asserted that culture and the learner’s immediate envi- the engagement with real-life learning challenges.
ronment determine both how he or she thinks, that is, These activities would ideally involve both intellectual
the processes of reasoning, and also the content of his or and physical activity: the movement of the students, the
her thinking, that is, the elements of knowledge that are seeking out of data, the measuring of objects, and so
combined and used in the thinking process. One’s suc- on. The learning context must be both a concrete real-
cess in learning is dependent upon the environment of ity and the site of an investment of abstract ideas and
learning and the activity is best facilitated through epistemological mores. So, one approach would be to
a process of problem solving in collaboration with invest academic knowledge of philosophy, physics, lit-
peers, relations, or teachers. Intellectual development erature, and so on into a real context such as health
depends greatly on the social situation of learning and care, art galleries, the music industry, political events,
how interactions with teachers, relations, and peers and so on, reflecting the interests of the learners them-
around the learner occur. So, the context in which learn- selves. A second approach would be to frame the
ing is based is a dual axis: on the one hand, the context pursuit of academic knowledge in terms of real-life
is the social situation of learning whereby knowledge is challenges: the objective of making river water safe
acquired, processed, and produced through collabora- to drink (chemistry) or the understanding of why
tion and use rather than direct dissemination; on the public consensus is so outraged when cadaver’s
other hand, the context must be an engagement with organs are used without consent (philosophy). In
a real-life task whereby knowledge interfaces with an both approaches, it is obvious that the epistemological
actual, empirical reality. Both axes instigate a move base of the discipline is broadened (in the chemistry
away from the hierarchical model of passive-learning example, there must be an explicit discussion of why
in the traditional lecture hall or classroom situation. we would want safe water and where the technology
Context-Based Learning C 801
would be useful) and that the acquisition of knowledge (Anthony et al. 1998; Hansman 2001; Rose 2009). The
crosses disciplines (in the philosophy example, data on majority of this empirical research is concerned with
the medical use of organs would have to be compiled). the effectiveness of imparting skills and relating aca-
The process of learning should involve distinct demic knowledge to real-life challenges. In these stud-
phases. One, the learner begins with empirical engage- ies, there is an attempt to compare the acquisition of C
ment with the site or interaction of the knowledge that knowledge concepts through traditional means and
relies on facts and theories already belonging to the through context-based approaches. The hypothesis
learner as well as knowledge shared with peers and that underpins much of the research is that if a learner
the encounter with new knowledge in situ. Two, the can understand why they are learning what they are
learner then conceptualizes that reality in terms of learning, it will somehow be of significance to them
concepts and theories is drawn from one or more and hence retained. The empirical case needs to show
academic disciplines. The knowledge acquired from that this is more than a mere truism and the theoretical
a traditional discipline is perceived as useful to the work has to articulate a framework that explains why
completion of a task or in the satisfaction of the this is the case.
learner’s self-directed interest. Three, the concepts There are, however, also some theoretical assump-
and theories of the discipline are applied or used and tions that deserve more attention. Most obviously,
thus engage reality and a concrete problem or object, so the reduction of epistemology to simple pragmatism
that the student sees and commands them in action. whereby discipline-specific knowledge is only of value
The learner takes possession of the knowledge in order if it can be utilized or applied to tasks and social
to satisfy an interest. Four, the results and conclusions integration seems to prioritize skill learning over
acquired and generated are disseminated in a variety of facts. Moreover, context-based learning ought to per-
ways, determined by appropriateness: presentations, haps be conditional and not a universal theory. It may
reports, theses, web pages, and so on. The approach perhaps be more suited to specific disciplines and the
encourages higher-order thinking alongside the passive balance between student-led and facilitator-governed
acquisition of discipline-based knowledge and involves learning will differ from discipline to discipline and
the learner in the social construction of knowledge that perhaps also from learner to learner. A core dissemi-
interfaces with a concrete reality. nation of knowledge is required at the dependent
The advantages of the approach are that the learn- stage and should not be discounted: independence
ing environment facilitates the internalization of develops from dependence and the use of context-
knowledge and facts because they are connected to based approaches should be attentive to these consid-
the reality of learners’ lived experience. Learners are erations. Empirical investigations are required to
involved in the production of the knowledge in a tyro measure the effectiveness of the technique and to dis-
researcher role whereby a hands-on experience makes cern the correct balance between the dependence and
learning into a doing and not just a happening. Further- independence at the various stages of an education.
more, learners are motivated to acquire the knowledge
and see it as valuable because it solves a specific prob- Cross-References
lem or engages a distinct reality. The motivation to ▶ Bottom-Up and Top-Down Learning
engage in learning is interest rather than punitively ▶ Collaborative Learning
driven. As a pedagogical method, it implicitly builds ▶ Dewey, John
upon the knowledge that learners already possess and ▶ Interests and Learning
so increases confidence and independence through ▶ Learner-Centered Teaching
active involvement and social collaboration. ▶ Personalized Learning
▶ Problem-Based Learning
Important Scientific Research and ▶ Project-Based Learning
Open Questions
Much of the research into context-based learning References
is rather appropriately an active engagement with Anthony, S., Mernitz, H., Spencer, B., Gutwill, J., Kegley, S., &
new pedagogical techniques in specific disciplines Molinaro, M. (1998). The ChemLinks and ModularCHEM
802 C Contextual / Context Stimuli
Contingency
Contextual / Context Stimuli
▶ Contingency in Learning
Stimuli in the background whenever learning and
remembering occur. These stimuli can be external
(e.g., room cues) or internal (e.g., drug or emotional
states).
Contingency in Learning
CHARLES R. GALLISTEL
Contextual Conditioning Cognitive Science and Behavioral Neuroscience,
Rutgers University, New Brunswick, NJ, USA
▶ Context Conditioning
Synonyms
Assignment of credit; Association; Contingency;
Contextual Control Correlation; Dependence; Prediction; Retrodiction
a b
1 1
p(US |CS)=.1
0.8 p(US |CS)=.2 0.8
Strength of CR p(US |CS)=.4
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.1 0.2 0.3 0.4 0 0.2 0.4 0.6
p(US|~CS) UC Contingency
c d
1 1
0.8 0.8
Strength of CR
0.6 0.6
0.4 0.4
0.2 0.2
0 0
0 0.2 0.4 0.6 0 0.2 0.4 0.6
f Contingency ΔP Contingency
Contingency in Learning. Fig. 1 (a). The strength of the CR on first test trial as a function of p(US|CS) and p(US|
CS) in
Rescorla’s (1968) experiment on the role of CS–US contingency as against temporal pairing. Although in each of the three
conditions, the temporal pairing of US and CS [hence p(US|CS)] was held constant, the strength of the CR declined to zero
as the contingency was degraded by increasing p(US|
CS). (b) Performance data in a plotted against the uncertainty
coefficient (UC) measure of contingency. (c) Performance data in a plotted against the f measure of contingency.
(d) Performance data in a plotted against the DP measure of contingency
Contingency in Learning. Table 1 22 contingency Table-based measures are, however, problematic
table when applied to instrumental and Pavlovian condi-
#US #~US Row totals
tioning experiments, which do not reliably have a
definable trial structure (Gallistel and Gibbon 2000).
#CS: a b a+b
This is apparent when one considers how to construct
#~CS: c d c+d the contingency table for Rescorla’s experiment. In that
Col totals: a+c b+d experiment, the CS always lasted 2 min. The interval
between CSs varied around an average of 10 min.
There is no doubt about how many CSs and USs
there were, so the first cell (a in Table 1) is readily
is recommended by Gibbon et al. (1974), while the determined. All the other cells are problematic, because
difference in the conditional probabilities of the US, there is no objectively justifiable answer to the ques-
tion: How many not-USs and how many not-CSs
a c
DP ¼ pðUS jCS Þ pðUS j
CS Þ ¼ were there? The values of contingency underlying
aþb cþd
Fig. 1 were obtained by following the common practice
has been used extensively in studies of human contin- of assuming that the intervals between CS presenta-
gency and causality judgment (see, e.g., Allan et al. tions are composed of “trials” of 2-min durations
2008). each, during which a US either occurs or does not.
Contingency in Learning C 805
The number of
CSs is taken to be the number of such the surprisal. Intuitively, the less probable the event, the
arbitrary subdivisions. The number of
USs is the total more unexpected or surprising it is, the more we are
number of 2-min intervals, including those when the informed by its occurrence – but, by the same token,
CS was present, minus the number in which a US the less often we are so informed. As may be seen from
occurred. However, the 2-min “trials” during the inter- Eq. 2, the entropy, H, of a distribution is simply the C
vals between CSs are a fiction, as is the number of not- average surprisal, that is, the amount of information
USs. Absent objectively defined trials, not-USs, and provided by each of the possible events weighted by its
not-CSs have no objectively definable relative fre- relative frequency.
quency, so one cannot construct a contingency table. Entropy is the technical term for the amount of
This problem is acute in the instrumental conditioning uncertainty in a probability distribution, which is the
case, because there are no trials in those protocols. same as the amount of information available from
A second problem with measures based on a con- that distribution, because information reduces uncer-
tingency table, and with the correlation coefficients as tainty. The mutual information between two events
well, is that they take no account of time. The contin- with observed or experimenter-defined probability
gencies of ordinary experience are defined over time, distributions is:
and the temporal intervals between the events are
I ðCS; US Þ ¼ H ðCS Þ þ H ðUS Þ H ðCS; US Þ;
centrally relevant to the psychological perception of
contingency and causality. The importance of “close” where H ðUS; CS Þ is the entropy of the joint CS–US
temporal pairing – of response and reinforcer, or of distribution. In the case where a contingency table can
CS and US – has always been stressed in the condi- be constructed, the US distribution is given by the
tioning literature. However, attempts to specify what normalized column totals, that is, the column totals
constitutes “close” have never succeeded. Clearly, a in Table 1 divided by N; the CS distribution is given
psychologically useful measure of contingency must by the normalized row totals; and the joint distribution
take time into account. is given by the normalized cell values (a/N, b/N,
A measure that does this is the uncertainty coeffi- c/N, d/N). For each distribution, the entropy is:
P
cient, also known as the entropy coefficient. It is the H¼ pi lnð1=pi Þ.
percent reduction in uncertainty about when (or The UC measure applies to temporal uncertainty as
whether) a predicted event (US) will occur gained well (Balsam and Gallistel 2009). If USs (or reinforcers)
from knowledge of the times at which (or trials on occur at random times, then the uncertainty regarding
which) the predictor event (CS) occurred: when the next US will occur is the entropy of an
exponential distribution, which depends only on the
UC ¼ I ðCS; US Þ=H ðUS Þ: ð1Þ
average US–US interval (the reciprocal of the base
I ðCS; US Þ is the mutual information between CS and rate). This entropy is the basal uncertainty about
US. H ðUS Þ is the entropy of the US distribution. It is when the next US will occur. It is the amount of avail-
also called the amount of “available” or “source” infor- able information. If a CS always precedes a US and
mation. It is the information-theoretic measure of the always tells us exactly when to expect the US, then
uncertainty regarding when and/or whether a US will there is no residual objective uncertainty about when
occur. In the case of atemporal dichotomous variables, the next US will occur once the CS has occurred. In
where there are objectively definable trials, hence objec- that case, the UC is 1, that is, the CS reduces the
tive probabilities for the failure of a US to occur, uncertainty about when the next US will occur by
X 100%. However, humans and other common labora-
HðUSÞ ¼ pi Inð1=pi Þ ¼ pðUS Þ lnð1=pðUS ÞÞ tory animals can only estimate the duration of an
ð2Þ
þ pð
US Þ lnð1=pð
US ÞÞ: elapsing interval with about +/ 15% accuracy. To be
useful, the CS must precede the US by some interval.
The lnð1=pi Þ is the amount of information pro- Our residual uncertainty about when exactly to expect
vided by the occurrence of the ith event in the set of the US is then determined by our imprecision in esti-
possible events over which a probability distribution is mating when the remembered CS–US interval has
defined (e.g., the US and
US events). It is also called elapsed. Thus, the effective percent reduction in our
806 C Contingency Learning
Cross-References
▶ Association Learning
▶ Associationism Continuous Assessment
▶ Bayesian Learning
▶ Communication Theory ▶ Formative Assessment and Improving Learning
▶ Connectionist Theories of Learning
▶ Formal Learning Theory
▶ Human Contingency Learning
▶ Law of Effect Continuous Improvement
▶ Reinforcement Learning
▶ Temporal Learning in Humans and Other Animals ▶ Learning Cycles
Contradictions in Expansive Learning C 807
of a new (“high-tech”) mode of production, they pose activity, which largely neglects the level of subject/
new questions with respect to the theorization of subjectivity and therefore ignores the kind of power
contradictions. Similar to Foucault’s governmentality effects which Foucault, for example, addressed as
approach, Langemeyer maintains that, under these “subjectivation” (assujettisement).
conditions, self-responsibility and self-management By contrast, Holzkamp’s “subject-science” of learn-
are not only aspects of self-determination, but para- ing discusses “internalized constraints” and the “expro-
doxically also a means of adaptation and subordina- priation of expansive learning,” for instance when one’s
tion. Due to this shift from “disciplinary power” own interests and those of others are so “intertwined”
toward “technologies of the self,” self-dependent that “power is not acting on the subjects from the
forms of learning are marked by new contradictions, outside but through them, through their subjectivities”
or more precisely, by power relations that act increas- (Holzkamp 1993, p. 523, my translation). However,
ingly “through subjectivity.” Instead of encouraging this problem is insufficiently reflected in Holzkamp’s
and enhancing collective learning, these contradictions use of the analytical categories “defensive learning” and
enforce tendencies of individualization and thereby “expansive learning.” In particular, the concept of
impair the potential of (expansive) learning. “defensive learning” is tailor-made for the problems
of schooling (e.g., the resistance of pupils against edu-
Important Scientific Research and cation), whereas “expansive learning” seems to be only
Open Questions its positive counterpart, but still conceived within the
Contradictions as both the motive and the object of same paradigm. Holzkamp exemplifies his vision of a
collective learning activities at various workplaces (e.g., self-determined education with some of his own indi-
in the Finnish health care sector) were investigated vidual experiences – of learning something “for its own
empirically by Engeström and collaborators at the sake.” Against this background, expansive learning
University of Helsinki (Engeström 2001). These studies becomes associated with a practice free from restric-
are centered on the idea of “developmental work tions, disturbances, or contradictions. Yet this interpre-
research” which harnesses workers’ active involvement tation would be misleading with regard to challenges of
in improving their cooperative work activities and their self-responsibility under flexibilized and precarious
working conditions (cf. Toikka et al. 1985; Engeström working conditions.
2005). Developed on the basis of psychological thought An empirical study on workplace learning of IT
(Vygotsky, Leont’ev, and others), this approach is specialists by Langemeyer shows a way of combining
not only highly regarded internationally, it has also insights from Engeström’s and Holzkamp’s approaches
exerted broad influence on various other disciplines, in view of a new societal problem. To explain the
such as the sociology of work and organizations, new type of contradictions, Langemeyer (2005) argues
human resources management, communication sci- that motivation to embrace the challenges of self-
ences and the media, software design, and science and responsibility depends on how a person makes sense
technology studies (cf. Roth and Lee 2007, p. 188). of them. The subjective meaning ascribed to one’s
However, despite this wide recognition, several cri- own living conditions is analyzed as a reflection of
tiques have emerged concerning Engeström’s theoreti- one’s vital needs and interests and situated knowledge
cal framework as well as his methodological basis (cf. Holzkamp 1993). Moreover, subjectivity is seen as
(e.g., Toomela 2008; Langemeyer 2006; Langemeyer immersed in social relations of everyday life, shaped by
and Roth 2006; Avis 2007). Among other objections, social processes of interpretation and negotiation, and
Engeström’s conceptualization of transformation and thus as susceptible to narrow-mindedness and ideolo-
change was ultimately seen as “conservative” (Avis gies. Consequently, the capacity or competence for self-
2007), as adjustable to capitalist needs of revolutioniz- regulated learning and self-management, Langemeyer
ing the mode of production, and thus as incompatible argues, does not “reside” as a stable character trait
with Vygotsky’s engagement for transformative social “inside” a person. She thus contradicts approaches
practice and dialectical thinking (cf. Stetsenko 2008) which assume that this competence would exist prior
Furthermore, these critics rejected Engeström’s adap- to specific learning or work activities. Instead, she
tation of functionalist and systemic views on human explains that it develops with raising awareness of the
Contradictions in Expansive Learning C 809
matters of work or life in general, how they emerged, itself. The study of contradictions in expansive learning
why they are at stake, and in what ways they can be is therefore a constant challenge “to generate – each
changed. This awareness is seen as a result of collective time anew – critical perspectives on these societal prac-
learning, which is envisaged, following Engeström, as tices in which we participate, and on our own social-
both a theoretical and practical intervention. In other individual basis to act and to reflect on the problems C
words, the desired competence coevolves with the and conflicts to be resolved” (Langemeyer and Roth
learning or work activity. 2006, p. 40).
The new type of contradictions is then illuminated
as follows: On the one hand, the desired effects of
self-responsibility essentially depend on the growth of
Cross-References
learners’ personal sense and self-will (Eigen-Sinn), that ▶ Activity Theories of Learning
▶ Apprenticeship Learning in Production Schools
is, on a specific kind of personality development.
▶ Collaborative Learning and Critical Thinking
On the other hand, this personal sense and self-will is
▶ Collective Learning
often pervaded by work relations in which resources
▶ Communities of Practice
are limited and objectives are shortsighted or even
▶ Cultural-Historical Theory of Development
contradictory. This kind of contradiction can be stud-
▶ Independent Learning
ied best by focusing on learning trajectories rather than
▶ Learning Activity
subjective reasons to learn as Holzkamp suggests.
▶ Lifelong and Worklife Learning
Given, for example, the subordination of workplace
▶ Self-determination of Learning
learning to work routines or management structures,
▶ Self-organized Learning
learning trajectories are often constrained by inade-
quate forms of participation and cooperation. ▶ Self-regulated Learning
▶ Sociocultural Research on Learning
Although at one moment in time, a number of aspects
▶ Socio-technological Change of Learning Conditions
of expansive learning may be prevailing (motivation
▶ Trajectories of Participation; Temporality and
for learning, engagement for problem-solving, and self-
Learning
responsibility may be high at the beginning of a course
▶ Workplace Learning
or a training program), in the long run they may be
gradually overshadowed by a discrepancy between the
learner’s desired and actual performance and between References
planned achievements and shortcomings. Holzkamp’s Avis, J. (2007). Engeström’s version of activity theory – a conservative
focus on subjective reasons and on learning as individ- praxis? Journal of Education and Work, 20(3), 161–177.
ual action proves to be too narrow to address this Engeström, Y. (1987). Learning by expanding: An activity-theoretical
problem, while Engeström’s notion of contradiction approach to developmental research. Helsinki: Orienta-Konsultit.
Engeström, Y. (2001). Expansive learning at work: toward an activity
as systemic dysfunction and as the driving force of
theoretical reconceptualization. Journal of Education and Work,
development is too broad. By contrast, Langemeyer’s 14(1), 133–156.
notion of “contradictions in expansive learning” aims Engeström, Y. (2005). Developmental work research (ICHS –
at theorizing the dynamics of expansive learning. Her Schriftenreihe). Berlin: Lehmann’s Media.
empirical approach consists of a three-dimensional Foucault, M. (1977). Discipline and punish: The birth of the prison.
analysis of forms of cooperation, modes of participa- London: Allen Lane.
Holzkamp, K. (1983). Grundlegung der Psychologie. Frankfurt/M:
tion, and changing aspects of a person’s situatedness. Campus.
It understands power relations that prevent the learn- Holzkamp, K. (1993). Lernen. Subjektwissenschaftliche Grundlegung.
ing or working subjects from exerting influence and Frankfurt/M: Campus.
gaining the power to act as the crucial point of societal Langemeyer, I. (2005). Kompetenzentwicklung zwischen Selbst- und
contradictions. In so doing, Langemeyer does not Fremdbestimmung. Arbeitsprozessintegriertes Lernen in der
Fachinformatik. Eine Fallstudie. Münster: Waxmann.
expect contradictions to be the driving force for devel-
Langemeyer, I. (2006). Contradictions in expansive learning –
opment or an obstruction for learning per se. She towards a critical analysis of self-dependent forms of learning
reasons that any engagement for changing and enhanc- in relation to the contemporary socio-technological change.
ing activities must be seen as a contradictory practice Forum Qualitative Social Research, 7(1), Art. 12 [43 paragraphs].
810 C Contrast
Langemeyer, I., & Roth, W. M. (2006). Is cultural-historical activity to be initiated by the subject. It is considered to be
theory threatened to fall short of its own principles and possi- limited, slow, serial, effortful, and used for unskilled
bilities in empirical research? Outlines. Critical Social Studies,
tasks. It is initiated intentionally and shows benefit
8(2), 20–42.
Roth, W. M., & Lee, Y. J. (2007). ‘Vygotsky’s neglected legacy’: cul- from practice. Performance will change from controlled
tural-historical activity theory. Review of Educational Research, to automatic after extensive training under the precisely
77(2), 186–232. the same conditions. Automatic processing is considered
Stetsenko, A. (2008). From relational ontology to transformative to be the opposite process to controlled processing.
activist stance: expanding Vygotsky’s (CHAT) project. Cultural
Studies of Science Education, 3(2), 465–485.
Toikka, K., Engeström, Y., & Norros, L. (1985). Entwickelnde
Theoretical Background
Arbeitsforschung. Theoretische und methodologische Elemente. During the 1950s, the cognitive psychology focused
Forum Kritische Psychologie, 15, 5–41. on the capacity limits of human information pro-
Toomela, A. (2008). Activity theory is a dead end for methodological cessing (HIP), such as how the brain treats incoming
thinking in cultural psychology too. Culture & Psychology, 14(3), information (stimuli). The British psychologist
289–303.
Broadbent introduced a significant model of informa-
tion processing in 1958 and was one of the first to
draw a distinction between automatic and controlled
processes. Further work by Posner and Snyder (1975)
Contrast implicated the automatic process to be an unconscious
▶ Simultaneous Discrimination Learning in Animals and unintentional process, whereas the controlled pro-
cess requires conscious intention. This view was
redefined by Schneider and Shiffrin in 1977 and has
since then been supported by convincing evidence
Control Processes and thereby kept its relevance during the decades.
In the “dual-process” information processing model
Control processes in the Atkinson–Shiffrin model are of Schneider and Shiffrin, a distinction between “auto-
strategies for managing learning such as deciding how matic detection” and “controlled search” emphases
to encode the material (verbal repetition, semantic two fundamentally different human information
associations, visual images) and subsequently retrieve processing operations. According to this view, auto-
it from memory. For example, attempting to recall the matic processing is parallel, fast, and a result of
names of all 50 states in the USA could be organized repeated training on a task, whereas controlled
either by alphabetical order or by geographical regions. processing is slow, serial, limited, and effortful. A new
skill requires controlled information processing and,
increasingly, as the skill is mastered, it becomes more
automatically processed. For example, learning how to
Controlled Information read is initially effortful and requires extensive cogni-
Processing tive capacity and gradually, reading training will
change the information processing to a more auto-
ÅSA HAMMAR matic process. A novice reader needs more time and
Department of Biological and Medical Psychology, has more errors compared to a skilled reader. Another
Division of Cognitive Neuroscience, University of example is when first learning how to drive a car and
Bergen, Bergen, Norway becoming an experienced driver, where information
processing transfers from operations which requires
controlled processing to more automatic operations.
Synonyms
Effortful information processing
Important Scientific Research and
Definition Open Questions
Controlled information processing is a mental process Numerous behavioral studies have shown that exten-
that requires attention and cognitive capacity and has sive training on precisely the same task increase the speed
Convergent Thinking and Learning C 811
References
Maturana, H. R., & Valera, F. (1987). Distributed processes, distrib-
uted cognizers and collaborative cognition. Pragmatics and Cog-
Conversation nition, 13(3), 501–514.
▶ Communication Theory
▶ Discourse
Cooperation
▶ Altruistic Behavior and Cognitive Specialization in
Conversation Analysis Animal Communities
▶ Multi-robot Concurrent Learning
Also known as CA, is a special type of discourse analysis
specially designed for the study of everyday verbal
and nonverbal communication. The aim of CA is to
describe structure and patterns of casual conversation
and of institutional talk (e.g., in school, surgery, or Cooperation Scripts
court). Developed in the late 1960s and early 1970s ▶ Collaboration Scripts
principally by the sociologist Harvey Sacks, Emanuel
Schegloff, and Gail Jefferson, CA is grounded in foun-
dational assumptions of ethnomethodology, a branch
of sociology that focuses on the question of how people
produce the mutually shared social order in which they
Cooperative Learning
live (ethnomethodology was founded by Harold
RIM RAZZOUK, TRISTAN E. JOHNSON
Garfinkel and Erving Goffman).
Learning Systems Institute & Department of
Educational Psychology and Learning Systems, College
of Education, Florida State University, Tallahassee,
FL, USA
Convex Relaxations
▶ Relaxations for Learning
Synonyms
Collaborative learning; Group learning; Small group
learning; Team learning
Co-occurrence Definition
Cooperative learning is the instructional use of small
▶ Measures of Association groups through which students work together to
Cooperative Learning C 813
maximize their own and each other’s learning (Johnson cooperative learning: (1) social interdependence, (2)
et al. 1994). It is related to collaborative learning, which cognitive-developmental or constructivism, and (3)
emphasizes that learning occurs as an effect of com- behavioral learning theories. The first theory, social
munity (Johson and Johnson 1999). It is, however, interdependence theory views cooperation as resulting
contrasted with individualistic and competitive learn- from positive interdependence among individuals’ C
ing in which students work by themselves to accom- goals. Groups are seen as dynamic wholes in which
plish learning goals that are not related to others, and a change in the state of any member changes the state
compete with each other for grades (Johnson et al. of other members. According to Johnson et al. (1998),
1998). There are three types of cooperative learning. the basic premise of social interdependence theory is
The first type is formal cooperative learning which con- that the way social interdependence is structured deter-
sists of students working together, for one class period mines how individuals interact, which in turn deter-
or several weeks, to achieve a joint learning goals and mines the individual and group outcomes. Positive
complete tasks assigned. The second type is informal interdependence (cooperation) results from promotive
cooperative learning which includes students working interaction as individuals encourage and facilitate
together to achieve shared learning goals in temporary, each other’s efforts to learn. In the absence of a func-
ad-hoc groups that last from a few minutes to one class tional interdependence (that is, individualism) there
period. The third type is cooperative base groups that is no interaction as individuals work independently
are long term, heterogeneous cooperative learning without interchange with each other. The second the-
groups where members give support, encouragement, ory that guides cooperative learning research is cogni-
and assistance needed to accomplish the shared goal tive-developmental theory that is grounded on the work
and succeed academically (Johnson et al. 1994; Johson of Piaget and Vygotsky. Piaget’s work is based on the
and Johnson 1999). For an activity to be cooper- premise that when individuals cooperate in the envi-
ative, it should have five basic elements: positive ronment, socio-cognitive conflict occurs that creates
interdependence, individual accountability, face-to-face cognitive disequilibrium, which in turn stimulates cog-
promotive interaction, social skills, and group processing nitive development. Vygotsky’s work is based on the
(Johson and Johnson 1999). Positive interdependence premise that knowledge is social, constructed from
means that students feel committed to one another cooperative efforts to learn, understand, and solve
and the success of one member is dependent on the problems. The third theory, behavioral learning theory
other group mates. Individual accountability requires focuses on the impact of group reinforcers and rewards
each group member to be responsible for contributing on learning.
a fair share of the work within the group. Face-to-face
promotive interaction in where students promote each Important Scientific Research and
other’s success by sharing resources, helping, and prais- Open Questions
ing each other’s success. Interpersonal and social skills Many studies have shown that when correctly
include leadership, decision making, and communica- implemented, cooperative learning improves informa-
tion skills. Finally, group processing requires group tion acquisition and retention, higher-level thinking
members to communicate not only how well they are skills (i.e., reasoning skills), interpersonal and com-
achieving but to coordinate their efforts (Johnson and munication skills, and self-confidence (Johnson et al.
Johnson 1999). 1998). These multiple outcomes that have been studied
can be classified into three major categories: achieve-
Theoretical Background ment, positive relationships, and psychological
There are several theoretical perspectives that have health. The research clearly indicates that coopera-
guided cooperative learning. Cooperative learning is tion, compared with competitive and individualistic
based on a variety of theories in anthropology, sociol- efforts, typically results in (a) higher achievement and
ogy, economics, political science, psychology, and greater productivity, (b) more caring, supportive, and
other social sciences. In psychology, however, where committed relationships, and (c) greater psycho-
cooperation has received the most intense study, logical health, social competence, and self-esteem
three major theories have guided the research on (Johnson et al. 1998). Findings from a meta-analysis
814 C Cooperative Learning
(Johnson et al. 2000) supported the effectiveness benefited from homogeneous grouping while low
of cooperative learning on students’ achievement/ achievers did equally well in either a homogeneous or
outcomes (e.g., grades). A total of 158 empirical studies heterogeneous group.
were included in the meta-analysis. Results revealed Although there has been a growing body of litera-
that cooperation promotes higher achievement than ture and empirical studies in the area of cooperative
do competitive (Cohen’s d = 0.82) or individualistic learning; many of the studies conducted looking at
efforts (Cohen’s d = 1.03). Cooperative learning also the impact of cooperative learning methods on
promotes higher achievement as compared to compet- achievement have methodological shortcomings and,
itive or individualistic efforts (Cohen’s d = 0.59 and therefore, any differences found could be the result
0.91 respectively). The authors concluded that it is of methodological flaws rather than the cooperative
reasonable to hypothesize that the effective use of the learning method (Johnson et al. 2000). In the future,
cooperative learning method will likely promote learn- researchers should concentrate on conducting highly
ing and other achievement-related outcomes. controlled (experimental design) studies that add to
As another example, Felder et al. (1998) conducted the confidence with which their conclusions will be
a longitudinal study to examine engineering students’ received. Future research studies need to investigate
achievement and attitudes in a cooperative learning the effect of different variables in the cooperative
environment versus students’ achievement and atti- learning process such as, group composition (hetero-
tudes in traditionally taught classes (i.e., lecture). The geneous versus homogeneous), group selection and
authors found that students in cooperative learning size, structure of cooperative learning, amount of
outperformed students in traditional context. Students teacher intervention in the group learning process,
in the cooperative learning environment had higher differences in preference for cooperative learning
scores and better attitudes toward instruction than associated with gender and ethnicity, and differences
did students in the traditional context. In addition to in preference and possibly effectiveness due to differ-
its effect on learning outcomes and attitudes, coopera- ent learning styles or self-regulation strategies, in
tive learning showed positive effects on retention, crit- addition to any mediating, moderating, or inter-
ical thinking skills (i.e., analysis and synthesis), and action variables that may affect the cooperative learn-
peer interaction. Cooperative learning caused higher ing process.
students’ retention rates, development of critical think-
ing skills, and higher peer interaction as compared to
traditional lecture. Felder et al. (1998) suggested that Cross-References
the more cooperative learning features that instructors ▶ Academic Learning
implement, the greater the learning improvements they ▶ Action-Based Learning
can expect. ▶ Altruistic Learning
Even though there has been many experimental ▶ Collaborative Knowledge Building
studies that examined the effect of cooperative learn- ▶ Collaborative Learning and Critical Thinking
ing on students’ learning outcomes, some researchers ▶ Collaborative Learning Strategies
have further studied the grouping effect, team com- ▶ Collective Learning
position based on achievement scores, (i.e., homo- ▶ Communities of Practice
geneous versus heterogeneous) within cooperative ▶ Engagement in Learning
learning environment on students outcomes. For ▶ Interactive Learning Environments
example, Baer (2003) compared heterogeneous coop- ▶ Knowledge Integration
erative learning groups with homogeneous cooperative ▶ Learner-Centered Learning
learning groups who were formed based on their ▶ Learning in the Social Context
first test scores. The results indicated that, overall, ▶ Participatory Learning
homogeneously grouped students significantly out- ▶ Peer influences on Learning
performed heterogeneously grouped students on the ▶ Peer-Learning
final exam. Particularly, high- or average-achievers ▶ Social Learning
Cooperative Learning Groups and Streaming C 815
tend to reflect social class and ethnic norms, thereby teaching of specific skills. Ability grouping should
perpetuating traditional class and ethnic distinctions. reflect different ability levels within a subject matter
Researchers have noted different teaching prac- area rather than to a general ability measure.
tices in homogeneous classrooms. Instruction for the ● Grouping must be followed by changes in teacher
highest performing students tends to be characterized behavior. Teachers must modify the pace and level
by teaching strategies that require deep and meaningful of instruction when ability groups are formed.
content manipulation and negotiation. Instruction for
those in lower ability groups tends to focus on memo- Small Group Ability Grouping
rization and the application of rules and algorithms. Although many forms of small group learning exist,
Schools that adopt ability grouping tend to employ most fall into three categories: peer tutoring; informal
the most traditional teaching methods. In the United groups; and cooperative learning groups. Peer tutoring
Kingdom, performance discrepancies between high- occurs when a more able peer teaches or mentors a less
and low-performing students were greatest in schools able peer. In informal learning groups, students work
using whole class teaching. together on a common task for a relatively brief time-
Ability grouping, which is closely related to period (ranging from a few minutes to an entire class
achievement in secondary education, impacts students’ period), but often with little structure or guidance
self-concept. Academic self-concept, which is formed defining how group members should collaborate. In
through social comparison, is diminished for students formal learning groups, often termed cooperative
in low-ability groups and is associated with negative learning groups, students collaborate according to
attitudes toward future learning experiences. The extent some form of systematic activity or script that guides
to which ability grouping is practiced within a school participants’ behavior. For example, in Learning
further impacts academic self-concept. Self-concept is Together (Johnson and Johnson 1998), team members
highest in schools with the least ability grouping and work on a common goal that is structured around five
lowest among students attending the most highly strat- themes: positive interdependence; individual account-
ified schools (Ireson and Hallam 2009). ability; effective interaction; communication skills; and
Whole-class grouping continues in many elemen- group processing.
tary and secondary schools although little research Numerous studies have examined the effectiveness
supports the practice. Slavin examined the effects of of learning in small groups. Results suggest that within
ability grouping on achievement in elementary and k-12 schools, small group learning effectiveness
secondary schools. Using an approach known as a increases as group structure increases. Thus, coopera-
best-evidence synthesis (which uses results from tive learning tends to be more effective than other
meta-analytic research and literature reviews), he forms of small group learning in k-12 schools. Even
reported an overall effect size of 0 indicating no benefit in less rigorous studies, cooperative learning is at least
to the practice of between-class grouping (Slavin 1987). as effective as other forms of large and small group
However, some grouping benefits were reported in work and meta-analyses indicate an effect size in excess
elementary schools for subjects that are inherently of .6. At the college level, all forms of small group
hierarchical: Cross-grade grouping benefitted reading learning appear to be effective in Science, Technology,
instruction and within-class ability grouping benefitted Engineering, and Mathematics (STEM) classes. At the
mathematics instruction. Slavin cautioned that when college level, small group learning is associated with
ability grouping is employed, the following guidelines improved academic achievement, persistence, and atti-
should be applied: tudes and the overall effect sizes for achievement are
approximately .5 standard deviations.
● Grouping plans must be flexible. Students tend to Although the superiority of group versus individual
remain in a group once an initial assignment has learning does not appear in question, the issue of group
been made. To be effective, ability grouping must ability composition has not been resolved. Homoge-
allow students to change groups as ability changes. neous grouping proponents claim that high-ability
● Grouping must be specific to content. Assigning students benefit from being academically stimulated
students to ability groups should be limited to the and challenged by similar ability partners. Critics
Cooperative Learning Groups and Streaming C 817
argue that mixed-ability grouping provides accesses to Important Scientific Research and
diverse opinions and resources and better prepares Open Questions
students for life in a diverse world. Additionally, het- According to Slavin, sufficient research has been
erogeneous grouping is said to benefit high-ability stu- conducted on the effects of whole class ability grouping
dents when they explain lesson content to their less able among students through 9th grade, but research is still C
peers, and low-ability students who receive help and needed to examine ability grouping effects in grades
explanations from more able partners. Although dis- 10–12. Research is also needed to examine changes in
agreement exists concerning the effects of ability teaching practices under different grouping plans and
grouping for cooperative learning, one result appears the development of more reliable assessments to sup-
to be consistent: Low-ability students perform less well port accurate data measurement.
when grouped with similar ability than higher ability Research is needed to examine how productive
partners. learning groups can be formed among participants
working at a distance. For example, although coopera-
Gender Grouping tive learning has been validated across cultural con-
The issue of gender grouping has both pedagogical and texts, research is needed to understand how cross-
political implications. To some, single-sex schooling cultural collaboration should be managed to promote
provides females and males with educational settings effective group work.
in which they can thrive: Students can study in envi- Research is also needed to examine the design and
ronments free from the social pressures induced by the effectiveness of computer-based tools that enhance the
opposite sex. Yet to others, single-sex schooling is per- basic elements that make cooperation work. Many of
ceived as a barrier to effective socialization. the studies examining group composition in small
Lee and Bryk (1986) concluded that single-sex edu- groups occurred before the evolution of the Internet.
cation is particularly beneficial to female students in Recent development of the so-called Web 2.0 technol-
secondary schools. They found that single-sex schools ogy creates new opportunities for collaboration in
deliver advantages to their students including increased small groups. Researchers have become interested in
academic achievement, enhanced attitudes and motiva- whether online learning and virtual communities of
tion, and improved academic behavior. Similar results practice can be fostered through collaboration. Hence,
were found in the UK, where single-sex schools were research is needed to determine whether embedded
particularly beneficial to high-performing 16-year-old scripts, computer-tutors or pedagogical agents, or
girls who outperformed males attending single-sex other forms of design can foster collaboration.
schools and were more likely to explore non-gender The composition of online groups could be re-
stereotypic subject matter. Moreover, the rate of high- considered and perhaps reframed in terms of the pro-
performing girls attending single-sex schools was three cesses underlying successful group collaboration. Since
times the rate for those attending coeducational schools young children can be considered domain novices with
(Sullivan et al. 2010). relatively uniform and unspecialized domain knowledge,
Kinzie et al. (2007) found that women benefitted cooperative learning groups in schools can be considered
from the types and frequency of “purposeful activities” homogenous in terms of domain knowledge. Indeed,
and the personal progress made in diverse educational such groups probably also bring well-established social
outcomes that occur at women’s colleges. Students norms (from playground and classroom interaction) to
experience higher personal expectations from faculty the group setting. However, for online groups, the learn-
than do their counterparts at coeducational colleges. ing environment may be heterogeneous in ways that have
They interact more frequently with faculty (who tend not been previously considered. Due to the sparse com-
to be more accessible) and meet faculty outside of class munication and other inherent potentially limiting
more frequently than do women at coeducational insti- features of the setting, it is unclear whether people work-
tutions. Similarly at women’s colleges, student leader- ing in heterogeneous ability groups can communicate
ship opportunities are greater and more students enroll productively at a distance. Similarly, other group com-
in traditionally male-dominated math, science, and position findings may not transfer from face-to-face to
engineering classes. online settings. If so, considerable research is needed to
818 C Cooperative Learning Strategies
deal with externally imposed events and demands that evaluation of a particular encounter with the environ-
the individual perceives as unpleasant or potentially ment, reflecting the person’s evaluation of the situa-
harmful. To most coping theorists, the coping process tion as relevant. They categorize appraisal as irrelevant,
consists of efforts to reduce perceived stress through benign-positive, and stressful. Stress appraisals are fur-
a wide range of thoughts, emotions, and actions ther divided into harm–loss, threat, and challenge sub- C
directed at both external stressors and internal categories, a framework that has received extensive
demands and needs. attention by researchers in the extant general and
Coping has been categorized as both a dispositional sport psychology literature. Threat appraisals are
and a situational construct. As a dispositional con- those in which the perception of danger exceeds the
struct, coping may be considered a person’s style, or perception of abilities or resources to cope with the
disposition. As a situational construct, coping consists stressor. Challenge appraisals, in contrast, are those in
of the conscious use of strategies for the purpose of which the perception of danger does not exceed the
either improving one’s internal resources (e.g., confi- perception of resources or abilities to cope. Thus,
dence, resourcefulness, hardiness, mental toughness) because the absence of a stress appraisal begins with
or managing external demands. Thus, coping style, the athlete’s perception of an event that is appraised as
also called dispositional or higher order coping, stressful, making non-stressful appraisals (e.g., posi-
is defined as a person’s disposition, or orientation, tive, harmless) requires no coping (Anshel et al.
toward the preferred use of selected types, or categories, 2001). An appraisal labeled stressful can reflect negative
of coping strategies (Anshel et al. 2001). Coping style feelings, such as threat or worry, or relatively positive
has been traditionally defined as “methods of coping feelings, such as challenge or heightened arousal.
that characterize the person’s reactions to stress Tomaka et al. correctly acknowledge that harm–loss
either across different situations or over time within appraisals occur after stressful situations abate, while
a given situation.” These coping “methods” are used threat and challenge appraisals occur before or in antic-
consistently in dealing with stressors across time ipation of stressful situations. Threat appraisals, there-
and in various situations. Coping strategies, on the fore, are accompanied by feelings of worry that nothing
other hand, is the situational use of a technique to will be gained from the stressful situation. Challenge
reduce external demands or improve internal resour- appraisals, on the other hand, provide hope that there
ces in dealing with an event perceived as stressful or will be something gained by the situation, and envision
unpleasant. positive incentives or avoidance of an unpleasant event.
Persons who feel inadequate or overwhelmed to
Theoretical Background deal with the stressful situation or view their coping
The process of coping with stress has a rich theoretical skills as inadequate are likely to make threat appraisals.
framework. The coping literature is replete with coping On the other hand, individuals who perceive them-
theory and models that reflect the coping process. Most selves as prepared to handle the stressful event possess
of these models can be represented by the following proper coping skills and feel confident in the outcome
structures and processes commonly referred to as the of the situation are more likely to make challenge
coping process. This section is divided into the coping appraisals. Threat appraisals are more strongly associ-
process and the primary theoretical frameworks that ated with negative emotional reactions than challenge
explain coping. appraisals (Lazarus and Folkman 1984).
Another appraisal conceptual framework is called
The Coping Process: Appraisals and perceived control, or controllability. Perceived control
Coping Strategies refers to the extent to which a person believes that the
The coping process begins with an event or stimulus that outcome of an event can be attributed to internal
is appraised as stressful. Appraisal is the person’s deter- (personal) sources, external (situational/environmen-
mination whether a particular environmental encoun- tal) sources, or to the cause or predictability of an
ter is relevant to his or her well-being and, if so, in what event. Perceived control, therefore, is the person’s belief
way. More specifically, Lazarus and Folkman (1984) that the individual can determine one’s own internal
contend that appraisal consists of the individual’s state and behavior, influence one’s environment, or to
820 C Coping with Stress
bring about a desired outcome from the stressful event, or potential unpleasant event or peacefully confronting
either by producing desirable events or preventing the source of stress by obtaining additional informa-
undesirable events. tion are examples. Emotion-focused coping concerns the
As the case with all types of cognitive appraisal individual’s conscious decision to deal with the stressor
constructs, perceived controllability influences the by regulating his or her emotions, or maintaining
individual’s coping response. The major determinants emotional control. Taking a deep breath and relaxing
of coping responses are the individual’s appraisal of the after a stressful event or discounting the importance of
stressor (Anshel et al. 2001). In the general psychology the stress source are examples. Both coping strategies
literature, personal (e.g., dispositions) and situational are useful and effective, as needed, given the demands
factors (e.g., source and/or intensity of the stressor) and characteristics of the situation (Lazarus and
influence the ongoing appraisal of threats and resources Folkman 1984).
in responding effectively to those threats. Finally, whether the coping effort was successful –
coping effectiveness – is the last segment of the coping
Coping Styles and Strategies process. Authors in the coping literature have desig-
The next step in the coping process is that the person nated nine outcomes of effective coping: (1) to reduce
initiates a coping strategy, which is situational, that psychological distress; (2) to obtain accurate informa-
often, although not always, reflects the person’s coping tion about environmental demands; (3) maintain
style. It is important to note that coping strategies proper internal mechanisms (e.g., attentional focusing,
reflect situational ways of dealing with stress, whereas proper vigilance and arousal level, rapid and accurate
coping style is dispositional and more predictable than decision-making procedures) to process incoming
a strategy. Thus, the person’s coping style should pre- information, and to know when and how to react
dict the type, or category, of coping strategy the person properly to stressful events; (4) reduce or manage
will enact following a stressful appraisal. physiological reactions (e.g., heart rate, muscle ten-
Coping styles, or the coping strategies that reflect sion) that may result in negative emotion and impair
them, have been categorized different among various performance; (5) improve mental well-being and
researchers and theorists. One popular framework is a positive self-image; (6) maximize the likelihood of
approach and avoidance (Anshel et al. 2001; Krohne returning to prestress activities; (7) create a stable
1993). Approach coping (styles and strategies) reflects psychological and emotional status that successfully
the person’s intensified intake and processing of directs energy and intentional behavior to meet exter-
unpleasant or threatening information. If one’s safety nal demands; (8) reduce and, if possible, eliminate
or welfare is at stake, for instance, the person must harmful environmental conditions; and (9) resolve
remain vigilant toward the stressor until the situation the stressful situation by producing a desirable affec-
has been resolved. Avoidance coping, on the other hand, tive or performance outcome. Taken together, there
reflects the person’s conscious attempt at physically or is general agreement that coping is a function of
mentally turning away from the stressful source. For several cognitive processes that are influenced by a
example, because coping consumes energy and atten- series of personal and situational factors. See Zeidner
tional resources, a person may want to be distracted by and Endler (1996) for a more extensive review of this
the stressor or psychologically distance oneself from the literature.
stressful source, similar to understanding the reasons
the explain behavior patterns of an unpleasant person, Coping Theories
or reducing the importance of an unpleasant situation. The theoretical frameworks that help explain coping
A common framework for examining coping strat- include the trait/dispositional model, the contextual/
egies more than coping styles is problem-focused and situational model, and the transactional model. The
emotion-focused coping. Problem-focused coping con- trait/dispositional model posits that a person’s use of
cerns the individual’s attempt to reduce or manage coping strategies is stable and cross-situational; coping
stress by directly dealing with the problem that is caus- is a unidimensional personality variable. It is assumed,
ing the distress, that is, an attempt to manage or control therefore, that a person’s coping thoughts or actions
a stressful situation. Removing oneself from an actual can be predicted from the person’s score on a coping
Coping with Stress C 821
inventory. Researchers have tended to not find exten- Rice 2000). The use of coping strategies have often
sive support for the trait theory of coping because the been used, tested, and reported interchangeably with
coping process has been viewed as multidimensional. coping styles. Relatively little research has been devoted
While trait measures are generally inadequate in to understanding the relationship between the use of
describing the complexity of the coping process, pro- coping strategies and the effectiveness of those strate- C
ponents of the trait model contend that personality gies. That is, coping effectiveness has received relatively
plays an important role in an individual’s persistent scant attention. In addition, coping has been measured
application of their personal coping style following inconsistently across studies, and a person’s self-report
stressful events, such as “the approacher” or “the of their coping strategies has consisted primarily of
avoider.” recalling events that may have occurred years before,
The contextual, or situational, model posits that or are responding to hypothetical situations. Finally,
coping is assessed in relation to specific stressful con- there is a deficiency of psychometrically validated cop-
ditions or situations. It is assumed in this model that ing inventories that were constructed for the sample
coping cognitions and behaviors are influenced by the currently being studied. These research issues have
relationship between the person and the environment clouded conclusions in our understanding of the cop-
following a particular event that is appraised as stress- ing process and the most valid means of measuring this
ful. In this model, then, coping consists of changing process. More experimental new research is needed to
thoughts and behaviors used by the person to manage determine the effect of coping skills training on selected
external demands and/or internal resources (e.g., con- cognitive and behavioral outcomes. Specific questions
fidence, anxiety, arousal, hardiness). To many propo- include the following:
nents of the contextual model, coping consists of
1. What conceptual model of cognitive appraisal,
managing the problem (i.e., problem-focused coping)
which is a mediating variable of coping, best pre-
and regulating emotions (emotion-focused coping).
dicts a person’s use of coping strategies?
The transactional model describes the individual
2. To what extent do coping strategies reflect a
and the environment in a continuous, bidirectional
person’s coping style? Similarly, does coping style
relationship. Transactional theory is designed to be
explain a person’s typical use of selected types of
used in reference to a specific stressful experience,
coping strategies?
rather than explaining the use of coping strategies –
3. How can we best measure coping effectiveness?
both problem-focused and emotion-focused – across
How should “effectiveness” be operationally
situations. In addition, this theory refers to what
defined in our attempts to measure the proper use
a person actually thinks or does (i.e., the use of strate-
of coping strategies?
gies), rather than what they usually do (i.e., reflecting
4. To what extent do moderating variables (e.g., age,
coping style) or what they think they should do. Finally,
gender, culture, stress intensity, coping style, per-
the theory reflects general coping strategies, which
sonality, situational factors) influence a person’s use
apply to a variety of stressful encounters or in various
of coping strategies and the effectiveness of those
stages of a single stressful encounter.
strategies?
5. Which of the primary coping models most strongly
Important Scientific Research and describe, explain, and predict a person’s coping
Open Questions skills?
The coping process is complicated and multi-
dimensional. Each dimension of coping includes dif-
ferent conceptual frameworks and structures. For Cross-References
example, some studies have examined coping in ▶ Stress and Learning
response to chronic stress, while other studies have ▶ Stress Management
examined acute (situational) stress. In addition,
cognitive appraisal has been conceptualized as per- References
ceived controllability or as a function of harm–loss, Aldwin, C. M. (2007). Stress, coping, and development: An integrative
threat, and challenge (Lazarus and Folkman 1984; perspective (2nd ed.). New York: Guildford.
822 C Copying
Anshel, M. H., Kim, K.-W., Kim, B.-H., Chang, K.-J., & Eom, H.-J.
(2001). A model for coping with stressful events in sport: Theory, Corporate Elearning
application, and future directions. International Journal of Sport
Psychology, 32, 43–75. ▶ Advanced Distributed Learning
Krohne, H. W. (1993). Attention and avoidance. Bern: Hogrefe &
Huber.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping.
New York: Springer Publishing Co.
Rice, V. H. (Ed.). (2000). Handbook of stress, coping, and health:
Implications for nursing research, theory, and practice. Thousand
Corpulent
Oaks: Sage.
▶ Obesity Stigma, Evolution, and Development
Zeidner, M., & Endler, N. S. (Eds.). (1996). Handbook of coping:
Theory, research, applications. New York: Wiley.
Corpus Callosum
Copying The corpus callosum is the largest fiber tract in the
▶ Imitation: Definitions, Evidence, and Mechanisms brain. It is the thick, white band of nerves that connects
▶ Imitative Learning in Humans and Animals the two hemispheres of the brain and allows both
halves to communicate sensory, motor, and higher
order information to coordinate activity.
Course of Study
Cortico-spinal Entrainment
▶ Curriculum and Learning
▶ Learning-Related Changes of b-Activity in Motor
Areas
Courseware Learning
Cost Complexity JAE MU LEE
Department of Computer Education, University of
▶ Adaptive Proactive Learning with Cost-Reliability Busan National University of Education, Busan,
Trade-off South Korea
Synonyms
CAI; Educational software; E-learning; ICT education
Cost-Noise Trade-Off
Definition
▶ Adaptive Proactive Learning with Cost-Reliability Courseware is a term that combines the words “course”
Trade-off with “software.” It is software containing educational
content, instruction, and instructional strategies. Its
meaning originally was used to describe additional
educational material intended as kits for teachers or
trainers or as tutorials for students, usually packaged
Counseling for use with a computer. Courseware learning is the
▶ A Tripartite Learning Conceptualization of process of learning through Courseware. CAI and
Psychotherapy educational software are terms that are also used to
describe Courseware. CAI stands for computer assisted
instruction or computer aided instruction. CAI is
a program that contains instruction contents and assis-
tance to instruction using a computer. It is difficult to
Counseling Outcomes distinguish between CAI and courseware. Sometimes
Courseware and CAI were used as the same concepts in
▶ Learning from Counseling
reference to a sort of educational software which refers
to all types of software for education. Educational
software is classified as instructional software, learning
software, and education management software. Classi-
Counting fied instructional software supports group learning
in a classroom for teachers, and learning software
▶ Accounting and Arithmetic Competence in Animals supports individual learning for students. Education
824 C Courseware Learning
management software assists management of educa- Khan 2001). WBI supports learning to overcome time
tion. Some examples of the assistance that education and space limitations. It provides not only the interac-
management software provides are evaluation, educa- tion between contents and learner, but provides the
tional material management, and instructor manage- interaction between instructor and learner, and inter-
ment. This education management software is called action among learners. WBI provides newly updated
CMI (Computer Managed Instruction). material and various types of information through the
Courseware is educational software that can be Internet. Due to Internet development, it made the
categorized as instructional software or learning soft- move from off-line learning to on-line learning and
ware. The instructor should use Courseware in accor- finally leads to e-learning. The essential and most impor-
dance to its advantages and characteristics in learning. tant part of e-learning is the quality of the e-learning
Courseware should not be overused. The important content. E-learning content is the same as Courseware;
thing is that Courseware learning does not substitute; therefore, the success of e-learning is highly dependent
rather it should assist in traditional learning. There- on the quality of the Courseware.
fore, Courseware can be just an auxiliary or subsidiary In 2000, the Internet environment was developed
media to support learning. to support wireless networks so various mobile media
such as cellular phones, ▶ smart phones, and PDA
Theoretical Background (Personal Digital Assistants) appeared. And the mobile
Courseware learning has been continuously changing
media tried to include educational contents and
as a result of the constant development of Courseware
Courseware. In the near feature, we will learn using
format, and Courseware format has changed in accor-
Courseware through new devices supporting ▶ ubiq-
dance with the development of computer environments.
uitous computing environments.
The main concept of the computer and educational
Courseware learning supports individual learning
software was the subsequent change of the Mainframe
with consideration of the difference of individual.
to the personal computer, CD-ROM, Internet, e-learn-
Courseware provides Learner initiated learning. While
ing, Mobile learning, and Ubiquitous learning.
the instructor chooses the contents and gives a lecture
Courseware was operated on the mainframe com-
to learners on a massive scale through his own inten-
puter until 1970. The PLATO (Programmed Logic for
tions in classical learning, the learner can choose the
Automated Teaching Operations, 1960) project was the
contents and can study at his learning pace, at his level,
first developed Courseware in the University of Illinois
and according to his interest in Courseware learning.
that was based on the mainframe computer (Smith and
Courseware provides self-directed, self-paced learning,
Sherwood 1976). It then appeared on personal com-
and learning based on various media.
puters such as the Altair8000 in 1975. It made the
Courseware has characteristics such as interaction,
transfer from the mainframe computer to personal
individual learning, and motivation. Price (1991) men-
computers. By the personal computer emergency, the
tioned the following:
computer can be used increasingly at home and at
school, and Courseware was distributed widely and 1. Courseware supports individual learning and
populated (Ceruzzi 2003). allows for self-directed learning.
When the CD-ROM appeared in 1990, the personal 2. It supports interaction and active learning. The
computer could process the large volume of data and learner can get an immediate response as a result
could process multimedia data such as voice, image, of using Courseware.
and video data. Therefore, Courseware extended from 3. It supports variety. The learner can develop an
text to multimedia data. Courseware supporting mul- interest through the graphics, sound, dynamic ani-
timedia contents are called MBI (Multimedia Based mation, and various feedbacks.
Instruction) (Alessi and Trollip 2001). By the end of 4. It supports record keeping. Through the record
1990, the Internet was developing rapidly and the edu- keeping of the computer, by saving learning history
cational software and Courseware were running based and recording learning accomplishments, Course-
on the Internet. Courseware that is running on the ware can make a diagnosis and provide suitable
Internet is called WBI (Web Based Instruction, learning.
Courseware Learning C 825
5. It supports flexibility. After evaluating the learner, ICAI requires the system to be able to diagnose stu-
we can increase or decrease learning volume. dent’s performance and provide the optimal student
6. It supports timeliness and responds instantly to the modeling process. A branch study within ICAI refers to
learner’s actions. An impatient learner and those adaptive learning system (e.g., iWeiver system) which
with a lack of curiosity in the subject can be capti- attempts to find the most suitable learning strategy C
vated by instant results of Courseware. considering each learner’s learning styles, learning
history, and learning goals. Adaptive learning system
Also, Courseware learning has the following advan-
is an implementation of ▶ Aptitude-treatment interac-
tages: Firstly, through simulation learning, it supports
tion (ATI). Most adaptive learning system studies are
learning activities that cannot be accessed in classical
developed considering learning styles (e.g., Tangow
learning due to time limitations, cost, and danger.
system). Other studies of Courseware development
Secondly, it allows the learner to repeatedly practice
are ▶ authoring tools, educational games and are
the things that he learns. In classical learning the
represented in the instructional model, interaction,
teacher has limitations in giving a lecture repeatedly,
learning motivation, evaluation and feedback design,
but Courseware learning allows the learner to repeat
interface design, and screen design in Courseware.
lectures and practices as many times as is necessary.
The studies for the effective use of Courseware learn-
Thirdly, it makes learning interesting by combining the
ing are mainly concerned with the analysis of learning
contents with games, and activities. This is especially
results or effects in Courseware learning (Kuilk and
useful for children.
Kuilk 1991). Courseware produces positive effects in
However, Courseware has some disadvantages:
logical learning such as mathematics and science sub-
Firstly, it is initiated not by the instructor but by the
jects. Multimedia learning is effective by supporting
learner. There is potential for ineffective results for
multisensory learning (Heinich et al. 1996). Learning
learners who lack learning motivation or intention.
is best facilitated through a combination of comple-
Secondly, it is not easy to build Courseware that is
mentary visual and auditory information (Mayer
assured of quality.
1997). Multimedia learning produces positive results
Courseware designers must consider general peda-
in foreign language learning. Also, Gleason (1981)
gogical issues such as the appropriateness of the com-
applied educational software to general students and
puter, methodology, student practices, lesson length,
students in a controlled group. He reported more effec-
and mastery level. Courseware design should adapt to
tiveness in the controlled environment.
the learner’s skill and knowledge.
The blended learning that was integrated in on-line
learning (e-learning) and off-line learning (face to face
education) is one of the approaches to improving
Important Scientific Research and learning accomplishments.
Open Questions Generally, if Courseware is properly used, Course-
The study of Courseware learning is classified into two
ware can improve learning effectiveness and efficiency.
parts: building effective learning Courseware, and effec-
On the other hand, there are several research studies
tively applying the Courseware in learning activities.
that suggest that Courseware is not meaningful and
Therefore, one is the development of effective Course-
effective because of the insufficient quality of the con-
ware in the computer science field, and the other is
tent (Alessi and Trollip 2001).
effectively using Courseware in learning as instructional
methods and the instructional technology field. Cross-References
The study of Courseware development is combin- ▶ Adaptive Learning System
ing Courseware and ▶ Artificial Intelligence (AI), ▶ Web Based Instruction
Wenger et al. 1987). This is called ICAI (Intelligent
Computer Assisted Instruction, Kearsley 1987) or ITS References
(Intelligent Tutoring System, Sleeman and Brown Alessi, S. M., & Trollip, S. R. (2001). Multimedia for learning method
1982). ICAI (e.g., GUIDON) was developed to improve and development. Needham Heights: A Pearson Education
upon the limitations of traditional CAI. The Study of Company.
826 C Covariation Learning
It is easier to study a specialized use of covert that rhyme, they are confused with one another so
rehearsal to memorize information. Researchers have that the usual benefit of rehearsal does not accrue.
found that series of items typically can be remembered The poorer memory for items that sound similar is
well, with a minimum of effort, if one can covertly termed the phonological similarity effect.
rehearse them (Gathercole and Baddeley 1993). Flavell et al. (1966) studied how children learn to C
Among the vast amount of information stored in the use covert rehearsal to remember sets of pictures. They
human mind, it is possible to think of only a very small made use of the fact that when the list is difficult,
amount at one time (this being the part that is called children often move their lips while rehearsing; this
the current contents of working memory) but often that activity might be considered partly covert (in that not
is all the information one needs. For example, going much sound is being made) and partly overt (in that
into a grocery store you may need to remember to the lips are moving). In order to allow children to be
buy, bread, eggs, cheese, and a spatula. Covert rehearsal less self-conscious, they wore a helmet and the visor
helps one to memorize the list or keep it in working was brought down, obscuring the child’s vision while
memory for a sufficient time. It also can help in the he or she tried to remember the pictures. The visor did
memorization of a new word, such as a person’s name. not cover the mouth, however, so it was possible for
Without covert rehearsal, new information tends to the investigators to see whether the child’s lips moved.
fade in a matter of seconds or is quite vulnerable to The finding was that rehearsal seemed to occur in only
interference from subsequent speech information. 10% of the 5-year-old children, increasing steadily to
60% of the 7-year-olds and 85% of the 10-year-olds.
Important Scientific Research and This rehearsal also went along with better memory for
Open Questions the pictures.
A convenient finding that makes it easier to study Cowan et al. (1987) made use of the phonological
covert rehearsal is that it appears to take place at similarity effect to examine the benefit of rehearsal in
about the same speed as overt rehearsal (Landauer adults. Participants were to remember and then repeat
1962). You can test this yourself with a stopwatch. series of words that sounded dissimilar (brick, spoon,
Ask a friend to count to 20 aloud as fast as possible cat, etc.) or series of words that sounded similar (mat,
while articulating each of the numbers, starting when bat, cat, etc.). To be counted correct on a trial, the
you say “go,” and to knock on the desk as soon as he or serial order of words in the series had to be reproduced
she finishes the last number. Now do the same test correctly. The index of memory was the length of lists
again but ask the participant to count silently instead that could be correctly repeated, or memory span.
of aloud. Do this a few times each way. You probably Memory span in adults displays a strong phonological
will find that the amount of time taken to speak aloud similarity effect: span for phonologically similar words
or silently is remarkably similar. is much lower than for dissimilar words. This effect is
It seems clear that people use covert speech as a thought to occur partly because of the confusion in
means to retain verbal information. For example, in memory between similar words when they are recalled,
the 1960s and early 1970s, R. Conrad published but partly when they are covertly rehearsed. For exam-
research on peoples’ immediate recall of a series of ple, the adult may try to remember the words by
letters. Researchers have referred to this work widely covertly rehearsing them in a cumulative manner,
and have followed up on it (e.g., Cowan et al. 1987; in order, as they are presented. Encountering brick,
Gathercole and Baddeley 1993). Conrad found that the participant rehearses brick; then encountering
even when lists were printed instead of spoken, recall spoon, the participant rehearses brick, spoon; and so
of these lists was impeded most when the letters on. For lists of phonologically similar words, the
sounded similar, not when they looked similar. For order is somewhat likely to be incorrectly changed
example, it is relatively easy to remember the series c, during rehearsal. When adults’ ability to carry out
f, q, p, o, r, y with the letters in order, and much more rehearsal was suppressed by requiring that participants
difficult to remember the series b, t, v, p, c, z, d with the quietly recite the alphabet while hearing the list, both
letters in order, because the letters rhyme in the latter the magnitude of the phonological similarity effect and
case. It is thought that when one tries to rehearse items the overall level of performance (especially on lists of
828 C Covert Rehearsal
its spatial orientation. For example, spatial memory is upon a spatial memory system whose description has
required when we have to navigate in a familiar envi- been greatly improved by rodent studies showing that
ronment or when we have to learn how to go from one so-called hippocampal place cells selectively fire when
point to another in a novel environment. Memory the animal occupies a specific location in its environ-
consolidation refers to the time-dependent process by ment, allowing the creation of a spatial map. Note- C
which recently acquired information is gradually inte- worthy, several studies have disclosed the reactivation
grated into long-term memory stores, a process whose of neuronal ensembles during sleep and wakefulness
duration ranges from hours to years according to immediately following exposure to spatial environ-
theoretical models, neurobiological and neuropsy- ments. Thus, studies conducted in rodents revealed
chological observations. Consolidation of spatial, that firing activity of hippocampal place cells active
hippocampus-dependent memories benefits from during spatial exploration behavior was increased dur-
sleep. However, functional neuroimaging studies have ing subsequent sleep states. Using large ensemble
revealed that this process of consolidation takes place recordings of place cells in the CA1 field of rodents’
by means of a covert reorganization of brain patterns hippocampus, it has been further showed that those
underlying memory performance, which is not neces- cells that fire in a synchronous manner when the ani-
sarily accompanied by overt changes in behavior. mal occupies particular locations in its environment
exhibited an increased tendency to fire together again
Theoretical Background during subsequent non-REM (NREM) sleep, as well
Animal and human studies have demonstrated a pri- as during the immediate post-training wakefulness
mary role for hippocampal areas in spatial learning, period. Synchronous cellular activity during NREM
supporting allocentric representation of the environ- sleep actually reproduced the discharge patterns
ment and encoding of the relationships between observed during task performance, eventually leading
environmental clues. However, spatial navigation in a to the neuronal replay hypothesis, positing that infor-
well-known environment may also be supported by mation acquired during active behavior is reexpressed
activity in the striatum through stimulus–response during sleep, a phenomenon that may represent a
associations. Indeed, whereas a hippocampus- neurophysiological substrate for memory consolida-
dependent strategy is applied in the early phase of tion processes. Although neuronal reactivations have
training, a strategic shifting toward striatum-dependent been repeatedly observed during NREM sleep in
responses may take place after repeated practice (Iaria rodents, similar phenomena have been also disclosed
et al. 2003). Noticeably, active reshaping of brain activ- during REM sleep (see Peigneux et al. 2001 for review),
ity is not necessarily accompanied by overt, detectable suggesting that all sleep stages may support the pro-
change in behavior. For instance, rodent and human cesses of memory consolidation. Additionally, it was
studies have yielded evidence for a covert reorganiza- found that the temporal sequence of neuronal dis-
tion of spatial memory traces during sleep, a state charges observed in hippocampal CA1 neurons during
known to be beneficial for memory consolidation pro- spatial exploration is repeated – recapitulated – during
cesses. In these studies indeed, the neural basis of per- NREM sleep on a similar or faster timescale. Neuronal
formance at retrieval was modified with intervening replay after spatial experience is not restricted to CA1
sleep and/or time after learning, whereas performance hippocampal neurons, since reexpression of firing pat-
levels per se were unchanged. In this section, we review terns during sleep has been observed also in posterior
those animal and human studies having evidenced parietal, visual, and prefrontal cortices (Peigneux et al.
covert reorganization of cerebral activity in spatial 2001). Finally, temporal correlations during NREM
learning, especially in relation to sleep. sleep between hippocampal ripples (high frequency
waves at 140–200 Hz) and spindle activity (phasic
Important Scientific Research and bursts in the 12–16 Hz frequency range) recorded
Open Questions in the prefrontal cortex were observed, reflecting
Route retrieval and way finding in a previously learned coactivation of hippocampal and neocortical pathways.
environment are critical prerequisites to successfully Taken together, offline replay of hippocampal activity
carry out most daily activities. These abilities rely together with coactivation of neocortical areas during
830 C Covert Reorganization / Spatial Learning
sleep probably represents important components in negative at delayed retrieval in the sleep group but
memory consolidation processes, allowing a gradual positive in the sleep-deprived. As a whole, these data
transfer of recently acquired spatial memory traces suggested that brain activity is reorganized during
from short-term hippocampus-based to long-term post-training sleep in such a way that navigation,
neocortical stores. Still, it should be noticed that initially based on a hippocampus-dependent spatial
those animal studies have not evidenced behavioral strategy, becomes progressively contingent on a
changes following the post-training sleep, putatively response-based strategy mediated by the striatum
consolidating period. (Iaria et al. 2003). In other words, sleep favored the
Likewise in humans, post-training reactivation of automation of spatial navigation. These results addi-
spatial navigation-related activity has been reported tionally demonstrated that covert (not directly observ-
during slow wave sleep (SWS; i.e., the deepest compo- able) reorganization of brain activity underlying
nent of NREM) using positron emission tomography navigation after sleep is not necessarily accompanied
(PET) (Peigneux et al. 2004). Furthermore, it was by overt (observable) behavioral changes.
found that overnight gains in task performance were It should be kept in mind that spatial navigation
correlated to hippocampal activity levels during SWS, is not in itself a pure process, but rather involves
suggesting a close association between spatial memory many cognitive operations and different memory com-
consolidation and hippocampal reactivation during ponents including spatial and contextual representa-
sleep (Peigneux et al. 2004). In a follow-up study, tions. More precisely delineated, a spatial memory
Orban et al. (2006) investigated using fMRI the sleep- representation involves the creation of and/or the
and time-dependent reorganization of spatial memory access to a cognitive map of the environment where
traces within the brain using a navigation learning task the spatial relationships between the streets are speci-
in a complex virtual town with a high degree of details fied independently of the salient features of the envi-
(walls, ground textures, objects,. . .). In this experi- ronment. For instance, when attempting to reach the
ment, subjects were scanned during route-finding hospital from the supermarket, one can keep in mind
tasks immediately after learning (consisting in a free an “abstract” map-like representation indicating the
30 min exploration period) and 3 days later. Then, half appropriate direction to follow at each crossroad, inde-
of the subjects were allowed regular sleep, whereas the pendently of specific environmental cues along the
other half was totally sleep-deprived during the first way. Besides this “streets configuration” component
post-learning night. Surprisingly, results showed however, a second, complementary process can be
a striking dissociation between equivalent performance used, which refers to a contextual memory representa-
and distinct neural bases for route retrieval at delayed tion (or “landmarks memory”) in which specific asso-
testing in sleep and sleep-deprived participants, ciations between salient landmark objects and their
suggesting sleep-dependent processes for reorganiza- milieu are stored. For instance, one may remember
tion of learning-related cerebral activity, not paralleled that from school to library, there is a right turn just
by overt changes in behavior. Indeed, whereas route after the post office and then a left turn in front of the
finding elicited increased activity in a well-known church. Thus, a further study wondered whether sleep
navigation-related hippocampo-neocortical network globally promotes consolidation of all memory com-
(Maguire et al. 1998) at immediate and delayed ponents embedded in virtual navigation, or rather
retrieval testing both in sleep and sleep-deprived par- favors the development of specific representations
ticipants (Fig. 1), activity in routine behavior-related (Rauchs et al. 2008). Using the same experimental
striatal areas was associated with delayed retrieval design than in Orban et al. (2006), participants were
activity only in participants allowed to sleep after administered four memory tasks (see Fig. 2) specifically
training. Furthermore, correlations between striatum tapping either the spatial memory component
activity and navigation accuracy were positive in the (“Impoverished” and “Alternate” conditions) or the
sleep group (higher activity in the striatum associ- contextual memory component (“Recognition” condi-
ated with higher navigation accuracy) but negative in tion) or both (“Natural” condition). In the Natural,
sleep-deprived participants (Fig. 1). Likewise, connec- Impoverished, and Alternate conditions, subjects had
tivity between hippocampus and striatum regions was to retrieve the route between two locations in the
Covert Reorganization / Spatial Learning C 831
Effect size
2
–2
–4
c
Covert Reorganization / Spatial Learning. Fig. 1 Navigation accuracy and sleep-dependent modulation of brain activity.
(a) Navigation accuracy, estimated as the distance traveled toward the target location, at the immediate (left) and delayed
(right) retrieval sessions for the sleep (blue) and sleep-deprived (red) groups. (b) Between-group regression analyses of the
average session performance on cerebral activity at delayed retrieval (sagittal and coronal sections). The blue crosshair
indicates the right caudate nucleus. The scatter plot shows that brain response in this area was correlated positively
with performance in the sleep group (blue; r = 0.41) but negatively in sleep-deprived participants (red; r = 0.80).
(c) Psychophysiological interaction analysis using the right caudate nucleus (green crosshair) as seed area. The coupling of
activity between the caudate nucleus and the left hippocampus (coronal section) was negative in the sleep group (blue)
but positive in sleep-deprived participants (red). The blue crosshair indicates the left hippocampus. Blue and red plots
show the size of effect for each group. Error bars are standard deviations
learned environment. In the Impoverished condition, representations were probed. Subjects had then to
the environment was plainly deprived of any wall/ determine whether environmental changes were made
ground feature and objects, promoting the use of spa- as compared to the exploration period (Fig. 2). Again,
tial representations to successfully perform the task. In behavioral performance did not differ between partic-
the Natural and Alternate conditions, the environment ipants allowed regular sleep during the post-learning
was the same as during the training period (one hour of night and those who were sleep-deprived, neither in a
free exploration of the environment performed outside natural setting that engages both spatial and contextual
the scanner), allowing the use of both contextual and memory processes nor when looking more specifically
spatial memory representations. In the Alternate con- at each of these memory components (Rauchs et al.
dition however, direct pathways between starting and 2008). At the neuronal level however, analyses focused
target points were blocked to promote alternative on contextual memory revealed distinct correlations
route-finding strategies that rely more on spatial rep- between performance and neuronal activity. In sleep
resentations. In the Recognition condition, subjects participants, recognition performance was correlated
had to pay attention to the environmental features of with activity in frontal regions, suggesting that
the town while following dots marking the path recollection processes were in use, whereas perfor-
between two locations, thus the spatial requirements mance was associated with parahippocampal activity
of the task were minimized while the contextual in sleep-deprived subjects, suggesting the involvement
832 C Covert Reorganization / Spatial Learning
J
D
I C
Bin G
F B
Station
Lee
A
Recognition condition
Wrong image ?
Impoverished condition
1 2
None
3 4
Covert Reorganization / Spatial Learning. Fig. 2 Virtual environment and navigation tasks. The map depicts an aerial
view of the color 3D virtual town in which subjects navigated at the ground level using a keypad. The ten possible starting
points are represented by letters (from A to J) with associated symbols and colors indicating the location to reach, out of
three possible targets (Bin, Lee and Station). The four snapshots display samples of the environment as seen by the
participant in the Natural, Recognition, Alternate or Impoverished conditions. For the Recognition task, subjects first
navigated in the environment following color dots on the ground (left panel). They were instructed to determine whether
and where environmental changes had been made as compared to the town explored during the training period. At the
end of each walk, a four-choice panel composed of three pictures taken from the path (one of them representing a change
made in the environment), and a white square was presented. Subjects had to respond by selecting the modified image or
the white square if they thought that no modification had been made (right panel)
of familiarity processes (Fig. 3). Likewise, efficient phenomenon is not unique to spatial learning since
spatial memory was associated with posterior cortical lack of overt changes in behavior paralleled with covert
activity after sleep whereas it was correlated with modulations of brain activity following sleep has been
parahippocampal/medial temporal activity after sleep reported also for memory consolidation of emotional
deprivation. Finally, variations in place-finding effi- material. Although further studies are needed to fully
ciency in a natural setting encompassing spatial and understand the functional significance of covert reorga-
contextual elements were associated with caudate nizations, available data suggest that cerebral reshaping
activity after post-training sleep, replicating our prior may precede overt expression of behavioral changes.
study (Orban et al. 2006), suggesting that sleep favors
automation in navigation. Cross-References
To sum up, available data indicate that even in the ▶ Human Cognition and Learning
absence of overt, measurable behavioral modifications ▶ Memory Codes (and Neural Plasticity in Learning)
following time or sleep after spatial learning, post- ▶ Memory Persistence
training sleep covertly reorganizes the neural substrates ▶ Place Learning and Spatial Navigation
of both spatial and contextual memories. Still, the ▶ Spatial Learning
Creative Inquiry C 833
4
3.5 C
3 3
2.5
2 2
1.5
1 1
0.5
0 0
Covert Reorganization / Spatial Learning. Fig. 3 Sleep-dependent modulation of correlation between brain activity and
performance in the Recognition condition. Contrasts are displayed at p < 0.001 (uncorrected) superimposed on the
average T1-weighted MR scan. Correlations were computed at the within-subject level (i.e., between brain activity and
individual variations in trial-to-trial performance). Left panel: higher correlations in sleep than in sleep-deprived
participants in the left frontal gyrus. Right panel: higher correlations in sleep-deprived than in sleep participants in the right
para-hippocampal gyrus
References
Iaria, G., Petrides, M., Dagher, A., Pike, B., & Bohbot, V. D. (2003). Creative Inquiry
Cognitive strategies dependent on the hippocampus and caudate
nucleus in human navigation: variability and change with prac- ALFONSO MONTUORI
tice. Journal of Neuroscience, 23, 5945–5952. Department of Transformative Inquiry,
Maguire, E. A., Burgess, N., Donnett, J. G., Frackowiak, R. S., Frith,
California Institute of Integral Studies, San Francisco,
C. D., & O’Keefe, J. (1998). Knowing where and getting there:
a human navigation network. Science, 280, 921–924.
CA, USA
Orban, P., Rauchs, G., Balteau, E., Degueldre, C., Luxen, A., Maquet,
P., & Peigneux, P. (2006). Sleep after spatial learning promotes
covert reorganization of brain activity. Proceedings of the Synonyms
National Academy of Sciences of the United States of America, Creativity; Passion; Self-inquiry; Transformative Edu-
103, 7124–7129.
cation; Transformative Learning
Peigneux, P., Laureys, S., Delbeuck, X., & Maquet, P. (2001). Sleeping
brain, learning brain: the role of sleep for memory systems.
Neuroreport, 12, A111–124.
Definition
Peigneux, P., Laureys, S., Fuchs, S., Collette, F., Perrin, F., Reggers, J.,
Phillips, C., Degueldre, C., Del Fiore, G., Aerts, J., Luxen, A., &
Creative Inquiry frames education as a larger manifes-
Maquet, P. (2004). Are spatial memories strengthened in tation of the creative impulse rather than as the funda-
the human hippocampus during slow-wave sleep? Neuron, 44, mentally instrumental acquisition, retention, and
535–545. reproduction of information, or Reproductive Learn-
Rauchs, G., Orban, P., Schmidt, C., Albouy, G., Balteau, E., ing (Montuori 1989, 1998, 2005, 2006, 2008). It stresses
Degueldre, C., Schnackers, C., Sterpenich, V., Tinguely, G.,
the role of ongoing inquiry, and the active creative
Luxen, A., Maquet, P., & Peigneux, P. (2008). Sleep modulates
the neural substrates of both spatial and contextual memory process of bringing forth meaning, knowledge, self,
consolidation. PLoS One, 3(8), e2949. and engagement with the world. Creative Inquiry
critiques Reproductive Learning, where the student is
an empty vessel to be filled by the instructor, and
Narcissistic Learning, which places the individual’s
Covert Speech largely unreflective and decontextualized opinions,
likes and dislikes, at the center of a subjectivist, relativ-
▶ Covert Pronunciation and Rehearsal istic world.
834 C Creative Inquiry
correct knowledge from the instructor. This knowledge discipline, a foundation of skills, and immersion in
must be reproduced to the instructor’s satisfaction. the field, in the same way that a creative musician
Creative Inquiry starts from an attitude of “not- must practice scales and learn music theory. But these
knowing,” a willingness to accept the illusion of famil- are not antithetical to creativity. On the contrary, the
iarity that covers the vast mystery of existence, examine foundation in scholarship is essential in order for the C
one’s positions in the process of inquiry, and challenge creativity to emerge (Montuori 2006; Montuori and
fundamental and underlying assumptions that shape Purser 1995).
inquiry. The goal is not to conclude the process by Creative Inquiry (CI) stresses the importance of
having the correct answer, but to encourage a more immersion and active participation in an ecology of
expansive, spacious approach to inquiry that actually ideas, in the existing discourse, literature, and research
generates more potential inquiry rather than stopping (Montuori 2005). It also recognizes that embodied and
at the one “correct” answer, and illuminates the crea- embedded knowing is grounded in existing cultural,
tion of knowledge. As in a jazz group, “band members” social, and historical assumptions, theories, facts, and
are invited to make contributions that will make the beliefs, and that any action in the world is based on, and
overall sound of the band the most interesting and in fact cannot occur, without interpretations of the
surprising. The point of contributions is not to pro- world and specific situations. This knowledge is neces-
vide “the” answer, and thereby to stop the conversa- sary for participation in both discourse and practice.
tion. In the same way that band members can push For Creative Inquiry this knowledge, in the form of
a soloist to greater heights with a series of well-placed paradigms, theories, etc., shared by communities of
chords or percussive accents, or simply verbal encour- inquiry (fields, disciplines, research methods, and
agement, the object of these contributions is to push agendas), and the inquirer’s own implicit assumptions
the dialogue to greater heights and to keep it going and theories, is itself constantly the subject of inquiry,
(Montuori 2003). offering an opportunity to explore and understand the
Creative Inquiry recognizes the limitations of creation of knowledge, perspectives, positions, beliefs,
knowledge and the opportunities for different perspec- theories, for purposes of wise and creative action.
tives, frames, and approaches. This involves an attitude
of epistemological humility and fallibility that recog- Important Scientific Research and
nizes humanity’s always partial and limited under- Open Questions
standing of the world (Bernstein 1983, 2005). Even Culturally and philosophically, Creative Inquiry
more importantly, it also recognizes that not-knowing emerges as an effort to address the opposition between
is a fundamental starting point for creativity. The will- Objectivism and Relativism (Bernstein 1983, 2005).
ingness to be open to the possibility that all knowers With (objectivist) Reproductive Learning, the deter-
have a fallible interpretation of the world allows for the ministic assumption is that the environment, “objec-
emergence of multiple alternative perspectives rather tive reality,” creates the learner. In (subjectivist)
than the assumption of a fixed “given” world. Creative Narcissistic Learning, this assumption is reversed, and
Inquiry encourages constant exploration and self- captured in the popular New Age dictum “I create my
examination for attachment to positions, obsession own reality.” Creative Inquiry proposes a recursive rela-
with certainty and power, and a constant awareness of tionship where “I create a world which creates me.”
the threats of dogma and/or habituation. Above all, an Creative Inquiry is an ongoing creative process in
attitude of not-knowing allows for the space and open- which the inquirer is engaged in self-eco-creation
ness for novelty to emerge. (Montuori 2003; Morin 2008a). Creating not just
Creative Inquiry does not accept the common himself or herself, but creating a relational being
binary opposition between creativity and rigorous whose actions have an impact in an interconnected,
scholarship suggested by the Romantic mythology of interdependent social and natural context. This is
creativity. This mythology’s assumption of “genius a crucial difference with Reproductive Learning,
without learning,” so popular in the West, became where the learner is treated like an isolated cog, to be
Narcissistic Learning. Understood in a wider perspec- molded by the educational process, so as to fit in
tive, the creative process requires and includes a larger machine.
836 C Creative Inquiry
Much important research still needs to be done in The process of self-creation through Creative
the application of creativity, complex thought, and Inquiry is not relativistic, self-centered Narcissistic
co-constructivist epistemologies to education, building Learning, revolving around the learner’s subjective
on the works of Morin, Kegan, Kincheloe, Varela, and likes and dislikes, agreements and disagreements, but
others. Central to this research will be the role of the an integration and embodiment of the inquiry process
inquirer in inquiry and the strong parallels between in a practice of phronesis, defined in this context as wise
Formal Thinking (Reproductive Learning) and Post- action informed by a (self-) reflection on values, beliefs,
Formal Thinking (Creative Inquiry). Creative Inquiry’s and implicit theories. Given the assumption that crea-
improvisational dimension is also akin to the concept tivity is not an exceptional talent confined to a gifted
of expertise from Dreyfus and Dreyfus’s research few but rather the essential condition of all human
(Montuori 2003). beings, the question becomes how that creativity will
be utilized and for what purposes. Self-creation in CI,
Inquiry and/as Self-Inquiry therefore, means taking responsibility for creativity and
Creative Inquiry invites inquirers to explore what they addressing central questions pertaining to the “who,”
are passionate about, and to ground their work in this “why,” and “to what end” of inquiry. Inquiry is not
passion. This passion itself becomes a topic for inquiry a dispassionate, purely “objective” process any more
and self-reflection as inquiry becomes an opportunity but engagement, participation, and responsibility for
for developing self-knowledge. The inquirer is not creation. It is an action in the world, and as such has
a spectator to the world, but embodied and embed- repercussions in the world and ethical consequences, as
ded, an active participant in knowledge-creation and well as being motivated by human passions and social,
praxis. Particular attention is paid to espoused theory political, and economic dimensions.
and theory-in-use, to dialogue between the inquirer’s
views and the research literature, and through dialogue Cross-References
with the perspectives of other co-inquirers. Every ▶ Creativity and its Nature
inquiry becomes self-inquiry in an ongoing process of ▶ Narcissistic Learning
unearthing one’s own implicit theories and assump- ▶ Reproductive Learning
tions, and in turn how they may be related to one’s own
personal history, sense of identity, attachments to References
beliefs and ideologies, and so on. Bernstein, R. (1983). Beyond objectivism and relativism. Science,
A central dimension of Creative Inquiry is the hermeneutics, and practice. Philadelphia: University of
self-reflection on this creative process of knowledge- Pennsylvania Press.
making and knowledge-embodying. Knowledge and Bernstein, R. (2005). The abuse of evil: Politics and religion after 9/11.
Malden: Polity.
concepts are viewed as creative products of the
Bocchi, G., & Ceruti, M. (2002). The narrative universe. Cresskill:
human mind (Deleuze and Guattari 1994) that can be Hampton.
challenged and opened up to reveal underlying Davies, P. (1989). The cosmic blueprint. New discoveries in nature’s
assumptions and the way they define, organize, and creative ability to order the Universe. New York: Simon and
determine knowledge. Theories, frameworks, and so Schuster.
Deleuze, G., & Guattari, F. (1994). What is philosophy? New York:
forth illuminate some dimensions of the world and
Columbia University Press.
obscure or ignore others, and are inevitably limited
Giroux, H. A. (2007). The university in chains: Confronting the
and partial. CI views concepts as creative products. It military-industrial-academic complex. Boulder: Paradigm.
frames inquiry into concepts (theories, paradigms, Giroux, H. A. (2010). Education and the crisis of public values. New
beliefs, etc.) and actions (as embodiments of theories, York: Peter Lang.
paradigms, etc.) as inquiries into the creative process of Kauffman, S. A. (2008). Reinventing the sacred. A new view of science,
reason, and the sacred. New York: Basic Books.
concept-creation. CI is radical in the sense that it
Kaufman, G. D. (2004). In the beginning. . .creativity. Minneapolis:
addresses the underlying roots and matrices from Augsburg Fortress Publishers.
which knowledge emerges, as well as the organization Kincheloe, J. (1993). Toward a critical politics of teacher thinking.
of knowledge and knowledge of organization. Mapping the postmodern. Westport: Bergin & Gray.
Creativity and Its Nature C 837
changing world. Driven by a combination of post- the arts or science to consider herself creative or to be
materialist conditions in many technologically advanced engaged in an enterprise that was labeled as creative by
countries and the explosion of the discourse of self-help others. This meant that creativity could only “exist”
and personal growth, there is also an increasing desire for in a limited number of human activities. In the West
self-expression and self-creation as individuals break out and many other parts of the world, women were tra-
of traditionally established careers and life-paths. It is ditionally not given extensive access to these activities.
also often the case that with rapid technological and For example, in the arts, no musical performances in
economic change, many new professions are emerging public, no study with nude models, and in the sci-
as old ones become obsolete and fall by the wayside. ences, limited access to education and explicit exclu-
Individuals and communities therefore have to reinvent sion from many areas. Women were therefore not in
themselves. Self-creation has become a major societal a position to be considered creative because they sim-
process where creativity takes center stage (Bauman ply could not participate in the activities that were
2008). societally labeled as creative (Eisler and Montuori
The concept of creativity emerged in the West in the 2007). This characterization of creativity therefore
Renaissance, along with individualism, and blossomed made it a very unusual, subjective, contingent phenom-
with the Genius myth of Romanticism. Until the 1980s, enon that was limited to very few individuals during
research on creativity in the West focused primarily on rare moments of inspiration in a closely circumscribed
the three Ps: Person, Process, and Product (Runco set of human endeavors.
2007). In the romantic mythology underlying this Creativity was a puzzling phenomenon in Moder-
atomistic view, the creative person was mostly a lone, nity. The Modern scientific worldview was based on
eccentric genius. The “Who” of creativity could there- a machine or clockwork metaphor in which the world
fore only be an individual person. Groups, organiza- was fundamentally Objective, Rational, and Orderly.
tions, cultures, and relationships were representatives Creativity on the other hand was either associated
of conformity and compliance, and were mostly viewed with subjective experience, the irrationality of mystical
as potential obstacles. insight or a breakdown in Order and hence with Dis-
The “How” of creativity consequently occurred order, whether socially or personally (mental illness,
exclusively “inside” the individual. The classic image of revolution). Creativity was viewed as essentially con-
the creative process involved a light bulb going on over tingent and subjective, rather than a lawful, orderly,
the creator’s head during the Eureka moment. The crea- and objective phenomenon. Science itself could there-
tive process was viewed as a solitary process, initially with fore not account for creativity. The creativity of scien-
mystical or divine sources, and then also increasingly tists did not begin to be systematically addressed until
associated with mental unbalance or even psychopathol- the 1950s as part of the larger emergence of systematic
ogy. The focus of the How was on the generation of creativity research. In his important work The Logic of
the idea, not the process leading up to the idea or how Scientific Discovery, philosopher of science Karl Popper
the idea would become a reality. The “What” or creative stressed the context of justification, and did not in fact
product was associated with “big bang,” earthshaking discuss discovery itself, which was, because of it sub-
insights (Montuori and Purser 1999; Runco 2004, jectivity and contingency, not considered amenable to
2007). Educational institutions and educators were scientific inquiry. By leaving the context of discovery to
not meant to cultivate the insights of genius, but merely psychologists, he was essentially dismissing it as a wor-
to reproduce a certain foundational knowledge base thy subject for science and philosophy, and hence seri-
and social system. The “Where” of creativity was ous inquiry (Popper 2002).
almost exclusively the arts and sciences, and in the Mainstream education mostly did not address cre-
latter mostly physics (Montuori 2011). ativity, because it was considered a gift of unique indi-
If having the Creative Person as the unit of analysis viduals rather than a quality or characteristic that could
by definition ruled out creativity as a possibility for be cultivated, and also because the social and political
educational settings, the Where of creativity by defini- purpose of education was to create good law-abiding
tion made it virtually impossible for somebody not in citizens and workers, not independent thinkers. When
Creativity and Its Nature C 839
the systematic and scientific study of creativity by The new, contextual and collaborative approach
psychologists was ushered in by J.P. Guildford’s Presi- to creativity by the younger generation is matched in
dential address at the American Psychological Associa- the research by a new research interest in the social
tion meeting in 1950, this was part of a larger Cold dimensions of creativity (Montuori 2011; Montuori &
War climate. The main concern was to reestablish Purser 1999). There is a move away from an essentialist C
American scientific supremacy. No effort was made to view of creativity to one that is relational and contex-
foster creativity in all students. Greater attention was tual. The emphasis on these dimensions of creativity
paid to creativity by finding the “best and the brightest” may be significant for education. Traditionally foster-
so they could be given special attention and their gifts ing creativity meant removing exceptional students
nurtured. from their educational context. Their exceptional
Despite the now truly substantial research literature nature was the starting point, but essentially the result
on creativity (Runco 2004, 2007), its impact on educa- of contingency and individual characteristics, and not
tion has been slim. At the beginning of the twenty-first replicable. Historically, there has been little research on
century, numerous critiques of education across all the creation of environments that foster creativity
levels bemoan the lack of creativity, and the focus on across the board for all students (Amabile 1996). The
Reproductive Learning that stresses memorization, focus on the social dimensions of creativity is showing
test-taking, and conformity (Robinson 2001). In the that creativity is also a function of certain kinds of
USA, the Ph.D. dissertation is supposed to be an orig- environments.
inal contribution to one’s field, but tellingly originality Creativity has been consistently mythologized and
and creativity are barely ever discussed during the misunderstood. Educational attempts to go beyond
educational process, unless it is in the context of pla- traditional Reproductive Learning and foster creativity
giarism (Montuori 2010). Research on the difficulties have at times veered perilously into Narcissistic Learn-
American doctoral students have completing their ing, valorizing the subjective, the unusual, and self-
degree found that in large part, the educational system expression at the expense of traditional competencies.
simply does not prepare students to be independent Typical was the left brain/right brain fad of the 70s
researchers (Lovitts 2005). and 80s. It seemed to suggest that the “right brain”
(the non-dominant hemisphere) was all that was
Important Scientific Research and needed for creativity, and the “left brain” was simply
Open Questions a hindrance. Research conclusively shows that creativ-
There are strong indications that in the twenty-first ity involves both hemispheres. Yet it is the simplicity
century, the discourse and practices of creativity itself of the right brain explanation that is so appealing and
may be changing. From the Modern individualistic also so misleading. The underlying dichotomizing is
focus oriented to “eminent” or uncontroversial crea- the same kind of thinking that leads to Narcissistic
tives producing exceptional products (Einstein, Learning and the promotion of a trivial creativity
Picasso, etc.), there has been a shift toward a more that is exclusively self-expressive but not contextually
collaborative, “everyday,” ecological creativity. The appropriate. Indeed when creativity is viewed through
focus is on generative interactions in a variety of a binary logic and decontextualized, it is trivialized and
mundane contexts, rather than the individual lone mutilated.
genius. Millennial college students associate creativity The emerging research on and practices of creativ-
with everyday activities, and with social interaction. ity can be summarized as proposing that:
Whereas for Baby Boomers, creativity came from “emi-
1. Creativity is the fundamental nature of the Universe,
nent creatives” in the form of the guitar of Jimi Hendrix
the process of creation itself, rather the spark of an
or the pens of Herman Hesse or Thomas Pynchon, in
occasional (C)creator, and is therefore a basic “every-
today’s “participatory” culture (Jenkins 2009), the
day, everyone, everywhere” human capacity.
focus is not so much “eminent creatives,” but partici-
2. Creativity is a networked, ecological, and relational
patory processes in video games like Beaterator, and the
process rather than an isolated phenomenon.
Garageband music application.
840 C Creativity and Learning Resources
mechanism. For instance, a single creative tech- metacognition in the acquisition of new competencies
nique like brainstorming (Osborn 1957) – one of has to be highlighted.
the best known creativity techniques, focused on
the free, abundant production of bizarre ideas in
Important Scientific Research and
order to promote innovation – could be used as a
Open Questions C
In order to produce in trainees a stable aptitude to
general tool for developing creative ideas and skills.
think and behave creatively in extra-training contexts,
2. Trainees are like a tabula rasa, that is, before being
it seems that educational tools should:
instructed, they know virtually nothing about how
to be creative; they have no idea or opinion about 1. Develop an integrated structure of various mental
creative strategies and are not able to control them. mechanisms, each playing a role in a particular kind
All this has to be “imprinted” into their allegedly of situation or in a particular phase of the creative
empty minds. process
3. Even though trainees are instructed with non- 2. Use materials that mimic real-life situations or, at
ecologically valid materials (such as puzzles, rid- least, help trainees to recognize the relationship
dles, and so on), the training programs can succeed between the training tasks and such situations
in prompting the subsequent spontaneous transfer 3. Consider individuals’ spontaneous beliefs and ten-
of creative strategies to everyday situations. dencies toward creative thinking and start teaching
4. The development of creative thinking can be in- from their naı̈ve creative competencies, with the
duced by simply asking trainees to perform a specific hope of changing spontaneous beliefs, tendencies,
mental operation a given number of times. In other and strategies by means of an internal restructuring
words, getting some practice in executing an opera- process
tion should be sufficient to allow people to learn it. 4. Show a metacognitive sensibility, that is, train
5. Creativity is only a matter of cognitive processes; learners not only to execute creative strategies, but
therefore, trainees must be taught only to activate also to control their execution (for instance, to
particular kinds of cognitive operations, without select the strategy to be applied and to monitor its
any reference to the complex interaction of these application)
operations with other cognitive processes, emotion, 5. Encourage a creative attitude, e.g., encourage
motivation, and the context. learners to accept the risks and discomforts that
6. Creativity can be promoted as a general ability, creativity involves, to avoid the tendency to stick
without making reference to specific domains. to familiar responses and to induce students to look
for novelty
Given these assumptions, it is not surprising that
the traditional programs designed to stimulate creativ- Various attempts to integrate cognitive, emotional,
ity often failed to reach their goals. In fact, ordinary and personality aspects of thinking have been made.
situations where creative thinking is needed are usually A constructivist point of view – aimed at substituting
complex situations that involve multiple mental oper- the spontaneous beliefs and tendencies of an individual
ations. Furthermore, in everyday life explicit hints to with new and evolved strategies by means of an internal
employ the relevant strategy are seldom given, so that restructuring process – is shared by many contempo-
individuals need to be able to identify the specific rary creativity programs. The features of current train-
features of the situation in question and choose the ing materials are in agreement with the issues discussed
appropriate way to deal with it. Finally, individuals previously. First, they induce individuals to learn a set
must not only know how to think creatively, but also of reasoning strategies that can result in a creative way
must want, that is, be inclined or motivated, to process of thinking. Further, they make people aware of the
situations creatively. These remarks stressed the need strategies they employ, of their relevance, of their ben-
for a different approach to promote creativity. More efits and costs. In other words, the programs stimulate
precisely, various components have to be identified in a metacognitive attitude. They also try to encourage
creativity; more attention to common reasoning and autonomy in the management of thinking strategies.
to complex real-life situations is required; the role of Moreover, the critical situations where learners are
842 C Creativity in Music
been Shaw’s (1989) model of the “Eureka process,” many orientations (Sternberg 2005) requiring a range
Finke et al.’s (1992) “Geneplore” model, Amabile’s of methodological approaches (Mumford 2003) and
(1996) “Componential” model, and Cropley’s (2001) incorporating different disciplinary perspectives.
“Holistic” model to mention but a few. Some of these Reporting on a series of investigations conducted into
process models expand upon the number of stages the nature of scientific creativity more than half a
given in the classic model, while others collapse them century ago, Taylor (1988, pp. 99) concluded that “cre-
into broader categories preferring instead to describe ativity is a very complex human performance” involv-
a wider range of substages or processes. However of ing “all aspects of a person’s response repertoire.” Such
these process models mentioned, only two of them (viz: a response repertoire, must by definition involve both
Shaw and Cropley) highlight the role of affect in the cognitive (thinking) and noncognitive (feeling) com-
creative process. Interestingly, Shaw identified a series ponents. Traditionally however, the field of cognitive
of feedback loops arising between each phase of the psychology has focused solely on the cognitive pro-
classic model and linked them to a set of affect states cesses. This begs the question “What makes creative
both positive and negative in the creative process. problem solving creative?”
Each loop was named after the creativity researcher At a microscopic level of inquiry, recent advances
responsible for theorizing its presence. While Shaw in neuroscience have shed new light on the role of
hypothesized the presence of five such loops (including noncognitive processes in human reasoning and con-
the “Rossman loop” that feeds back from all previous sciousness, revolutionizing thinking concerning the
stages in the model), he speculated the presence of role of feeling and intuition in solving novel prob-
many more. The proposed existence of multiple feed- lems. Working with brain-damaged patients, Damasio
back loops, operating simultaneously and successively, (1994) found individuals, who having presented with
both consciously and non-consciously, over n parallel normal IQ, language ability, and learning capacity,
paths, is consistent with neural network models of the being unable to solve problems, due to impairment of
brain. A diagram showing Shaw’s feedback loops the feeling function within the brain. Damasio goes on
superimposed onto the classic model of creative prob- to describe three kinds of feeling, notably feelings of
lem solving is given in Fig. 1. basic universal emotions, feelings of subtle universal
emotions, as well as background feelings. These feelings
Important Scientific Research and arising from the complex interplay of the brain core
Open Questions (viz: hindbrain, midbrain, and limbic systems) and the
However, despite the rhetoric, or perhaps because of it, cerebral cortex, provide a picture of the body’s internal
the question still remains as to why creativity defies state justa-positioned with information received about
complete explanation and why its nature continues the external one. Such feeling is essential to human
to remain elusive. Working at a macroscopic level survival and consciousness. According to Damasio,
of analysis, recent research would seem to indicate feelings are just as cognitive as other precepts and are
that a confluence of components is needed involving essential for being able to move through a decision
Solving a
Novel Problem Rossman Loop
Rossman Loop
Rossman Loop
Areti Loop Vinacke Loop Lalas Loop Communication
Loop
Intimation
“A ha”
Playing experiencing Verifying
Preparing Incubating Illuminating Elaborating Outcome
Creativity, Problem Solving, and Feeling. Fig. 1 Diagram of the classic model of creative problem solving superimposed
with Shaw’s feedback loops
Crime C 845
making space. Thus, while the traditional view may concert to express a whole raft of mental processes be
have been that feelings interfere with an individual’s they cognitive or not.
ability to solve problems, this old adage failed to point
out that in the absence of feeling an individual is Cross-References
unlikely to solve the problem at all. ▶ Complex Problem Solving C
Evidence of individuals attending to a feeling of ▶ Consciousness and Emotion
cognition in solving novel problems is to be found ▶ Problem Solving
in the historical accounts of notable scientists and ▶ Nature of Creativity
mathematicians. Henri Poincaré, for example, des-
cribes an inner aesthetic feeling guiding his response References
to a new intellectual order, Albert Einstein describes Aldous, C. R. (2009). The genesis of new ideas: Models, feeling and
a feeling of direction sensed visually, going toward solutions. In B. Matthews & T. Gibbons (Eds.), The Process of
something concrete, while Nobel Prize winner Barbara research in education: A festschrift in honour of John P. Keeves
(pp. 338–366). Adelaide, SA: Shannon Research Press.
McClintock, describes a feeling of affinity guiding
Amabile, T. M. (1996). Creativity in context: Update to the social
observation into the making of new “insights.” In psychology of creativity. Boulder, CO: Westview.
each of these examples, attention to feeling is integral Cropley, A. J. (2001). Creativity in education and learning a guide for
to the creative problem-solving process and the devel- teachers and educators. London: Kogan Page.
opment of new ideas. Damasio, A. R. (1994). Descartes’ error: Emotion reason and the
human brain. London: Papermac.
In the light of evidence such as this, it is perhaps
Finke, R. A., Ward, T. B., & Smith, S. M. (1992). Creative cognition:
interesting to note the findings of a recent large- Theory, research and application. Cambridge, MA: MIT Press.
scale study of creative problem solving indicating Guilford, J. P. (1967). Creativity: Yesterday, today and tomorrow. The
that students who attended to a feeling approach to Journal of Creative Behavior, 1(1), 3–14.
reasoning were more likely to be successful in solv- Horstman, J. (2010). The scientific American brave new brain.
ing a novel mathematics problem than those who San Francisco: Jossey-Bass.
Mumford, M. D. (2003). Where have we been, where are we going?
did not. In this study, it was inferred that feelings
taking stock in creativity research. Creativity Research Journal, 15
of cognition served to assist the successful novel (2 & 3), 107–120.
problem solver through the problem-solving space Shaw, M. P. (1989). The Eureka process: A structure for the creative
(Aldous 2009). experience in Science and Engineering. Creativity Research Jour-
Any discussion about the origin of ideas and the nal, 2, 286–298.
Sternberg, R. J. (2005). Creativity or creativities. International Journal
solving of novel problems, however, would not be com-
of Human Computer Studies, 63, 170–382.
plete without making mention of the debate concerning Taylor, C. W. (1988). Various approaches to and definitions of crea-
which process arises first, feeling or thinking. One group tivity. In R. J. Sternberg (Ed.), The nature of creativity: Contem-
of proponents contend that noncognitive (feeling) and porary psychological perspectives (pp. 99–121). New York, NY:
cognitive (thinking) processing operate as independent Cambridge University Press.
systems and that decisions can be made instantaneously Wallas, C. (1926). The art of thought. New York: Harcourt brace.
References
Criminal Autistic Psychopathy Fairclough, N. (1992). Discourse and social change. Cambridge:
Polity Press.
▶ Diagnosis of Asperger’s Syndrome Fairclough, N. (2000). Discourse, social theory and social research:
the discourse of welfare reform. Journal of Sociolinguistics, 4,
163–195.
Crisis Incubation
▶ Barriers to Organizational Learning
Critical Events in Learning
▶ Critical Learning Incidents
Criterion-Referenced
Assessment
When a student’s performance is assessed according to
Critical Learning Incidents
how well the performance meets certain preset stan-
HANNU SOINI
dards or criteria. This is as opposed to norm-referenced
Department of Educational Sciences and Teacher
assessment (NRA) that assesses a student’s perfor-
Education, University of Oulu, Oulu, Finland
mance accordingly to how well it compares with those
of other students.
Cross-References Synonyms
Critical events in learning
▶ Learning Criteria, Learning Outcomes, and Assess-
ment Criteria
Definition
Critical learning incidents are learning situations which
learners have experienced as effective, exceptional, or
personally meaningful. Critical learning incidents may
Critical Discourse lead to educationally significant learning and personal
growth. The term critical refers to the fact that the
The confirmation by the learner of a best judgment by circumstances described in the incident play an impor-
discussing assumptions, realizations, and solutions tant role in determining the outcome of learning. Typ-
with other adults. ical of these experiences is that critical characters of an
incident are described by the learners themselves. This
means that incidents include a multitude of different
kinds of activity and that incidents can only become
critical afterward.
Critical Discourse Analysis
Also known as CDA, is a form of discourse analysis that Theoretical Background
focuses on the ways in which discourses serve as means The study of critical incidents has a long history in
of social and political domination. Developed in the psychology (Butterfield et al. 2005). Flanagan (1954)
last decades by Norman Fairclough, CDA is an inter- developed the critical incident technique (CIT) during
disciplinary approach unified by foundational assump- World War II as a means to gain understanding of the
tion about the links between language and power rather causes of airplane crashes. In 1954, Flanagan published
than by a well-defined set of analytic techniques. an article on the critical incident technique, describing
Critical Learning Incidents C 847
the origins of the method and a flexible set of principles the emphasis is on specific situations and incidents.
which must be followed in order to capture a detailed Instead of writing about abstract concepts, respon-
description of the incident. According to Flanagan, a dents concentrate on describing particular happenings,
critical incident is any activity that is sufficiently com- which are much easier to report than are general
plete in itself to permit predictions to be made about definitions or underlying assumptions. An additional C
the person performing the act (Flanagan 1954, p. 335). advantage of the critical incident technique is that
To be critical, an incident must occur in a situation subjects are talking about themselves without being
where the purpose or intent seems fairly clear to the consciously aware of it. While students are not being
observer. The critical incident technique presumes that asked directly to articulate their ideas or conceptions
participants’ general assumptions are embedded in, of learning, the choice of examples really reveals essen-
and can be inferred from, their specific descriptions tial features about their own ideas and experiences
of particular incidents. of learning.
Recently, the study of critical learning incidents has Woods (1993) has reported several benefits that
been based on the assumption that in order to under- critical incidents possess for the understanding of
stand human learning, we should better take into the nature of student learning. In critical incidents,
account both personal experience and social context learning is integrated in the self, because it is based
as the most essential factors of the learning process. on students’ personal needs and goals. Through per-
Hofer and Pintrich (1997) have assumed that the tra- sonal experiences, students have a real possibility to
ditional research methods in learning might predeter- construct their own view about reality. Learners also
mine approaches of the learning study to focus on the have a large amount of control over their own behavior
dimensions important for the researcher and exclude in learning settings. In other words, students are the
more personally salient perspectives. For example, owners of the products of the learning process.
when studying learning from the learner’s point of The benefit of critical incidents in the analysis of
view, students’ short stories might better describe learners’ personal views on learning is twofold. Firstly,
their ideas about learning, rather than just asking they give insight into learners’ everyday practices. Sec-
them to define the concept of learning. In stories, ondly, describing specific situations, events, and people
students have to locate their learning experience in is much less demanding or threatening for students
everyday situations and to describe it from their per- than being asked to define their general assumptions
sonal point of view. When learning is described in the about or abstract definitions of learning. Brookfield
form of a story, it is constituted as a changing, contex- (1994) believes that the critical incident technique is
tual, and personally meaningful sequence of events. especially appropriate for teachers or other people who
In the study of critical incidents, narrative metaphor are interested in developing the learning of others.
may be used from the retrospective perspective. That
is, students construct their experiences about learning Important Scientific Research and
afterward into the shape of a story. Learning itself may Open Questions
be a chaotic or unconscious process, but through tell- The benefit of critical incident studies rests on the
ing their experiences, students give logical form to their assumption that concrete learning experiences offer
idea of the process of learning. an adequate way to understand learning from the
learner’s point of view. However, the investigation of
Contribution to the Field of Learning critical learning incidents has many methodological
In recent years, the educational power and usefulness of challenges. According to Butterfield et al. (2005) the
critical learning incidents has become evident for many future of critical incident studies is “rooted in the
researchers. Critical incident studies have been used as past, which entails striking the right balance between
the basis for curriculum design in many areas of health respecting technique’s method as articulated by
sciences, teacher education, and the service industry. Flanagan(1954), and embracing its inherent flexibility
According to Brookfield (1994), the advantage of the that allows researchers to adapt it for use across myriad
critical incident method in the field of learning is that disciplines and research questions” (p. 489).
848 C Critical Reflection
Cross-References
▶ Beliefs About Learning Cronbach, Lee J. (1916–2001)
▶ Critical Learning and Thinking
▶ Everyday Learning NORBERT M. SEEL
▶ Experiential/Significant Learning (C. Rogers) Department of Education, University of Freiburg,
▶ Flow Experience and Learning Freiburg, Germany
▶ Learning in Practice and by Experience
▶ Transformational Learning
Life Dates
Lee Joseph Cronbach was born in Fresno, California,
References on April 22, 1916. He received a master’s degree from
Brookfield, S. (1994). Using critical incidents to explore learners’
the University of California at Berkeley. Thurstone’s
assumptions. In J. Mezirow (Ed.), Fostering critical reflection in
adulthood (pp. 177–193). San Francisco: Jossey-Bass. work on the measurement of attitudes had a strong
Butterfield, L. D., Borgen, W. A., Amundson, N. E., & Maglio, A. T. influence on him, and accordingly he studied psychol-
(2005). Fifty years of the critical incident technique: 1954–2004 ogy at the University of Chicago. In 1940, he received
and beyond. Qualitative Research, 5(4), 475–497. his PhD in educational psychology from the University
Flanagan, J. C. (1954). The critical incident technique. Psychological
of Chicago, where he met Ralph Tyler and became
Bulletin, 51, 327–358.
Hofer, B., & Pintrich, P. R. (1997). The development of epistemolog-
his research assistant on the Eight-Year Study – one of
ical theories: Beliefs about knowledge and knowing and the most influential studies in education of that time.
their relation to learning. Review of Educational Research, 67(1), Cronbach’s lifelong interest in education likely had its
88–140. origins in this collaboration. In 1940, Cronbach accepted
Woods, P. (1993). Critical events in teaching and learning. London: an assistant professorship in psychology at Washington
The Falmer.
State University. Toward the end of World War II,
he served as a military psychologist in San Diego and
became increasingly engaged in instructional psychol-
ogy. After the war, he returned to Chicago, then he
moved to the University of Illinois in 1948, and finally
Critical Reflection to Stanford University in 1964, where he served as Vida
The examination of the influences around oneself that Jacks Professor of Education until his retirement in
contribute to a worldview change. 1980. Cronbach died of congestive heart failure in Palo
Alto on October 1, 2001.
As an educational psychologist he made significant
contributions to psychological testing and measure-
ment as well as to instructional psychology (Shavelson
Critical Self-Reflection 2009).
instruction (Snow 1980). Later he expanded the under- Snow, R. (1980). Aptitude processes. In R. Snow, P. Frederico, &
lying information processing model of learning by W. Montague (Eds.), Aptitude, learning and instruction, conative
and affective process analyses (Vol. 1, pp. 27–60). Hillsdale:
introducing cognitive-conative-affective intersections.
Erlbaum Associates.
The objective was to integrate more realistic aspects Snow, R. (1989). Aptitude-treatment interaction as a framework for
of mental life, such as mood, emotion, impulse, research on individual differences in learning. In P. Ackerman,
desire, volition, and purposive striving into instruc- R. J. Sternberg, & R. Glaser (Eds.), Learning and individual
tional models. differences. New York: W.H. Freeman.
Sternberg, R. J. (1996). Matching abilities, instruction, and assess-
The collaboration between Cronbach and Snow
ment: Reawakening the sleeping giant of ATI. In I. Dennis &
set the stage for learning orientation research, which P. Tapsfield (Eds.), Human abilities: Their nature and measure-
attempts to reveal the dominant power of emotions ment (pp. 167–181). Mahwah: Lawrence Erlbaum.
and intentions on the guidance and management of
cognitive processes. In its basic understanding of the
structure and nature of the complex relationships
between learning orientations and interactions, this Cross Talk Between Stored
line of research can easily be traced back to Cronbach’s
original hypothesis that we should find the treatment
Memories
to which each individual can most easily adapt. If synapses are shared by different stored memories, the
Although the ATI concept disappeared gradually as retrieval of one particular memory can be contami-
a research topic after the 1980s, the idea of matching nated by the undesired recall of other memories.
abilities, instruction, and assessment is still at the core Typically, synapses are shared if memory and query
of instructional research today – ATI is the “sleeping patterns are distributed; i.e., each pattern contains
giant” of learning orientation research (Sternberg many active neurons. The strength of cross talk will
1996). increase with the number of stored patterns.
Cross-References
▶ Adaptive Instruction System(s) and Learning
▶ Adaptive Learning Through Variation and Selection
▶ Aptitude–Treatment Interaction
Cross-Cultural Approaches to
Learning and Studying
References
▶ Cross-Cultural Learning Styles
Cronbach, L. J. (1949). Essentials of psychological testing. New York:
Harper and Row.
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of
tests. Psychometrika, 16, 297–334.
Cronbach, L. J. (1954). Educational psychology. New York: Harcourt.
Cronbach, L. (1957). The two disciplines of scientific psychology. The
Cross-Cultural Factors in
American Psychologist, 12, 671–684. Learning and Motivation
Cronbach, L. J. (1977). Educational Psychology (3rd ed.). New York:
Harcourt Brace Jovanovich. JULIAN GEORGE ELLIOTT1, WILMA C. M. RESING2
Cronbach, L. J. (1989). Lee J. Cronbach. In G. Lindzey (Ed.), History 1
School of Education, Durham University,
of psychology in autobiography (Vol. 8, pp. 64–93). Stanford:
Durham, UK
Stanford University Press. 2
Cronbach, L. J., & Snow, R. E. (1977). Aptitudes and instructional Department of Psychology, Department of
methods: A handbook for research on aptitude-treatment interac- Developmental and Educational Psychology, Leiden
tions. New York: Irvington. University, Leiden, Netherlands
Flammer, A. (1975). Individuelle Unterschiede im Lernen. Beltz:
Weinheim.
Seel, N. M. (1979). Wertungen im Geschichtsunterricht. München:
Minerva. Synonyms
Shavelson, R. J. (2009). Lee J. Cronbach (1916–2001): A biographical Cultural factors in learning and motivation; Culture
memoir. Washington, DC: National Academy of Sciences. and learning and motivation
Cross-Cultural Factors in Learning and Motivation C 851
academic achievement. This operated, not merely 3. A powerful and influential level of family sup-
within the classroom, or even the school, but was port. With this comes recognition that family
also reflected by parental views and those of the hardship may be a necessary price to pay to achieve
wider communities from which the students origi- the highest levels of success. Parental obligation
nated. In contrast, Russian classrooms were tradi- involves ensuring that their children learn well.
tionally seen by members of that society as a setting In turn, children feel obligation to honor their
for hard work and maximum engagement, and there parents’ sacrifice by means of their academic
were widely agreed understandings that arduous achievement.
study would also be undertaken after school hours. 4. A belief in discipline and the importance of
2. Powerful peer influences that maximized student demonstrating effort. Traditional virtues include
commitment. In contrast, in the USA and UK, diligence, endurance of hardship, humility, concen-
peer influences appeared to undermine academic tration, and perseverance. These tend to persist
engagement and achievement. While effortless suc- even when families relocate to Western societies.
cess in the classroom was generally acceptable in 5. A strong sense of group identity in which the desires
these latter contexts, it was often considered to be of the individual are subordinated to the needs of
socially undesirable to display heightened interest the class group.
in, or commitment to, one’s studies. Allied to 6. A supportive pro-learning peer culture and the
these social constraints were significant differences employment of high-achieving peers as important
between the Russian and the Western children in role models.
respect of general classroom behavior and accep- 7. Respect for the authority and knowledge of parents
tance of teacher authority. These differences were and teachers.
similarly found in other contemporary cross- 8. Recognition that education is often a demanding
cultural investigations involving these countries. and arduous process and does not need to always be
3. A strong belief in the importance of education as fun or intrinsically appealing.
a vehicle for personal improvement. To be an edu-
cated (“cultured”) person was to be someone who Important Scientific Research and
was generally respected and admired. Thus, what- Open Questions
ever one’s abilities and goals, education was seen as Despite pride in their international standing, several
being an important means of self-improvement. In nations scoring highly on international tests are now
the USA and UK, however, education is widely per- questioning whether their traditional values and
ceived in highly instrumental terms. While this, in approaches adequately equip their students to thrive
itself, did not appear to be a motivational problem in a global marketplace. However, it is recognized that
for those who believed that education could help the introduction of reforms brings associated risks
them achieve their goals, such an orientation proved because many of the personal characteristics that are
to be highly problematic for a significantly high deemed to be valuable for economic success have the
proportion of underachievers who believed that potential to threaten traditional relationships and
they could never achieve such success. power structures. An interesting dilemma is whether
it is possible, in such societies, to maintain the highly
Such a list maps closely onto those cultural factors
disciplined and focused educational orientation of
that have been widely identified in the literature to
young people while simultaneously increasing their
explain the high levels of educational performance of
capacity for autonomy, creativity, risk-taking, indepen-
South East Asian students (i.e., those from cultures
dence, spontaneity, problem-solving, assertiveness, and
primarily underpinned by the Confucian tradition)
perhaps most controversially, their willingness to ques-
(Stevenson and Stigler 1992). Key amongst these are
tion and challenge.
1. Highly positive attitudes to learning and However, this may be a moot issue as powerful
scholarship. social and economic forces are not easily controlled
2. Very high standards and expectations in relation to by government agencies. A breaking away from tradi-
educational achievement. tional attachments and identities, a strong emphasis
Cross-Cultural Learning Styles C 853
upon individualism, and the seeming inability of Elliott, J. G., Hufton, N., Willis, W., & Illushin, L. (2005). Motivation,
existing institutions to guide young people, appear to engagement and educational performance: International perspec-
tives on contexts for learning. London: Palgrave.
be features of all late-modern or postmodern societies
Elliott, J. G., & Tudge, J. (2007). The impact of the West on post-
(Inglehart and Welzel 2005). Thus, a tradition of def- Soviet Russian education: Change and resistance to change.
erence to adult authority, a strong emphasis upon Comparative Education, 43(1), 93–112. C
self-discipline, a readiness to forego social and leisure Inglehart, R., & Welzel, C. (2005). Modernization, cultural change and
pursuits in favor of narrow academic success, and democracy: The human development sequence. Cambridge:
Cambridge University Press.
a willingness to engage with unappealing academic
Reynolds, D., & Farrell, S. (1996). Worlds apart? A review of interna-
material – factors all highly associated with high- tional surveys of educational achievement involving England.
scoring countries – may all ultimately be undermined London: H.M.S.O.
by globalizing influences, irrespective of any national Stevenson, H. W., & Stigler, J. W. (1992). The learning gap: Why our
desire or legislative action. schools are failing and what we can learn from Japanese and
Such phenomena were evidenced in the post-Soviet Chinese education. New York: Summit Books.
Cross-References
▶ Motivation and Learning Synonyms
▶ Motivation to Learn Cross-cultural approaches to learning and studying;
▶ Motivational Variables in Learning Cross-cultural learning approaches; Cross-cultural
▶ Social Learning learning types
▶ Socio-emotional Aspects of Learning
Definition
References Cross-cultural learning styles refer to variations in the
Bronfenbrenner, U. (1970). The two worlds of childhood. New York: cognitive, affective, and physiological traits that are
Russell Sage. relatively stable, self-consistent, and characteristic
854 C Cross-Cultural Learning Styles
indicators of how learners from different cultures per- (Richardson 1994), our knowledge of the relationship
ceive, interact with, and respond to the learning envi- between cross-cultural differences and learning styles is
ronment, including, but not limited to, the processing still rather limited in its scope and its results.
of information. In a more applied manner, cross-
cultural learning styles can also be referred to as the Important Scientific Research and
degree to which the concept that individuals differ in Open Questions
regard to what mode of instruction or study is most The majority of the existing research on learning styles
effective for them varies across cultures. comes from only a limited number of countries leaving
available instruments largely untested in cross-cultural
Theoretical Background settings. By and large, the research on culture’s impact
The notion that culture has an all-pervading influence on learning styles falls into one of two categories –
on all aspects of human life has led to an inquiry into its studies of the learning behavior of students in certain
relationship with learning styles over the past three national cultures, and comparative studies. Among the
decades. The discourse on cross-cultural learning styles first group, a sizable number of studies have been
is deeply rooted in the larger topics of cognitive style conducted on learners in the Anglo-Saxon cultures of
and learning style. Cognitive styles are usually referred Australia, the United Kingdom, and the United States
to as self-consistencies in processing information, solv- of America. Also, with the number of students seeking
ing problems, and making decisions that develop in a degree outside of their home country on the rise,
characteristic and habitual ways around personality several studies have been carried out on populations
trends. The term learning style is both broader and of students studying in foreign host countries, most
narrower than cognitive style at the same time. On the notably on Asian students in English-speaking coun-
one hand, it is narrower as its application is specifically tries. In the second group of studies, we find a number
limited to the context of a learning environment. On of comparative studies focusing on direct comparisons
the other hand, it is broader as it goes beyond the between two or more cultures, as well as a smaller
cognitive by including affective and physiological number of studies on the multicultural classroom.
behaviors. Wide parts of the extant literature still use Unfortunately, as the existing research is far from not
the terms synonymously thereby creating confusion only a consensus about learning styles instruments, but
and incoherence. Some of the better known conceptu- also concerning the measurement of culture, the results
alizations of learning styles are Marton and Saljö’s deep of studies on culture’s influence on learning styles are
vs. surface learning dichotomy, and Kolb’s Learning hardly comparable and largely inconclusive. Some
Style Inventory (Apfelthaler et al. 2007). In the decades studies confirm the influence of culture on learning
after their introduction, a plethora of alternative con- styles and see, for instance, a culture-biased distribu-
ceptualizations and instruments whose psychometric tion of different types of learners across cultures, while
properties vary greatly (Coffield et al. 2004) have been other studies do not. Some authors even attribute
developed. Most notable, among those, are the Learn- greater explanatory power to other factors such as
ing and Studying Questionnaire (LSQ), the Revised discipline, gender, or institutional factors when it
Study Process Questionnaire (R-SPQ), the Approaches comes to variations in learning styles. Similarly, com-
and Study Skills Inventory for Students (ASSIST), the mon national cultural stereotypes such as, for instance,
Revised Approaches to Studying Inventory (RASI), the the Asian learner as rote-learner have both been con-
Learning Styles Questionnaire (LSQ), the Index of firmed and challenged by existing research on the topic
Learning Styles (ILS), the Study Process Questionnaire (Watkins and Biggs 1996).
(SPQ), the Inventory of Learning Styles (ILS), or the What is surprising is that in the entire literature
Approaches to Studying Inventory (ASI) (Apfelthaler on culture’s implications for learning styles, there are
et al. 2007). Unfortunately, only very few of these only very few references to the vast amount of existing
instruments have been tested or used in more than publications on cross-cultural differences that have
one cultural environment. Therefore, despite decades otherwise attracted considerable attention, such as
of research on learning styles, and a general agreement the works of Dutch researcher Geert Hofstede
that learning styles may vary from one culture to another (e.g., Hofstede 1986). According to Hofstede, cultures
Cross-cultural Training C 855
vary across four dimensions that he calls power dis- Hofstede, G. (1986). Cultural differences in teaching and
tance, individualism, uncertainty avoidance, and mascu- learning. International Journal of Intercultural Relations, 10(3),
301–320.
linity. In some of his earlier works, Hofstede made
Richardson, J. T. E. (1994). Cultural specificity of approaches to
assumptions concerning the consequences of these studying in higher education: A literature survey. Higher Educa-
dimensions for the learning behavior of students, tion, 27(4), 449–468. C
including differences in profiles of cognitive abilities Watkins, D. A., & Biggs, J. B. (1996). The Chinese learner. Cultural,
between the populations from which teachers and stu- psychological and contextual influences. Hong Kong: Comparative
Education Research Center.
dents are drawn and differences in expected patterns
Yamazaki, Y. (2005). Learning styles and typologies of cultural differ-
of teacher/student and student/student interactions. ences: A theoretical and empirical comparison. International
Based on his own research in 40 different countries, Journal of Intercultural Relations, 29(5), 521–548.
Hofstede predicts that students from certain Asian
countries that score low on individualism and high
on power distance, will have a strong preference for
traditional student–teacher relationships that are based
on hierarchy, respect, harmony, and formal instruction Cross-Cultural Learning Types
(Hofstede 1986). It is somewhat surprising that, except
for a few notable recent contributions (e.g., Apfelthaler ▶ Cross-Cultural Learning Styles
et al. 2007; Yamazaki 2005), learning styles research
has not embraced Hofstede’s work on cultural differ-
ences to a greater extent. Based on the existing litera-
ture on cross-cultural learning styles, a number of open
questions and directions for future research can be Cross-cultural Training
identified. These include (1) the testing of existing
learning styles instruments in cultures other than TATIANA STEFANENKO1, ALEKSANDRA KUPAVSKAYA2
1
those in which they have been developed and, if Department of Social Psychology, Moscow State
necessary, their revision; (2) the development and University, Moscow, Russia
2
test of a conceptual model of how culture influences LITE College, London, UK
learning styles; and (3) studies comparing two or more
cultures using those learning styles instruments that
show strong psychometric properties across different Synonyms
cultures. Intercultural training
Cross-References Definition
▶ Cross-Cultural Factors in Learning and Motivation Training is one of the methods of interactive education,
▶ Cross-Cultural Training specifically organized short-term group work, and
▶ Culture of Learning based on the assumption that learners derive knowl-
▶ Learning Styles edge, skills, and competencies from personal – direct or
▶ Social Interaction Learning Styles simulated – experience. Metaphorically, training as a
method can be described as a process of intense social-
References ization, and in the case of cross-cultural training –
Apfelthaler, G., Hansen, K., Keuchel, S., Mueller, C., Neubauer, M., intense enculturation (the realization an individual
Ong, S. H., & Tapachai, N. (2007). Cross-cultural differences in achieves about his own culture) and intense accultura-
learning and education: Stereotypes, myths and realities. In D. tion (the realization an individual achieves about
Palfreyman & D. L. McBride (Eds.), Learning and teaching across a different culture). This relatively new field represents
cultures in higher education (pp. 15–35). Houndmills and New
an interdisciplinary focus of cultural anthropology,
York: Palgrave.
Coffield, F., Mosely, D., Hall, E., & Ecclestone, K. (2004). Learning
cross-cultural psychology, sociology, sociolinguistics,
styles and pedagogy in post-16 learning: A systematic and critical intercultural communication, and multicultural edu-
review. London: Learning and Skills Research Centre. cation (Bennett et al. 2004).
856 C Cross-cultural Training
Diverse programs of cross-cultural training are could be counterproductive as it may call into question
focused on direct interaction with other cultures and the very possibility of successful interaction and under-
are designed to teach individuals to cope with situa- standing between different cultures.
tions of cultural variety, help them effectively deal with Implementing the principle of cultural universal-
the inevitable stress that accompanies the cross-cultural ism requires a high level of professionalism from the
experience, and be efficient in a multicultural environ- cross-cultural trainer at every stage of the program’s
ment. On the macro-level there are two main aims of design – from the methodological development to
cross-cultural training: (1) to bring about change in assessing its effectiveness. The model of an intercultural
a social or cultural situation such as decreasing racism, trainer’s competences by M. Paige consists of four main
chauvinism, and other forms of prejudice and discrim- categories: knowledge, skills, personal attributes, and
ination existing in society, and (2) to resolve conflicts ethics (Paige 1996). An intercultural trainer should
and promote more harmonious intercultural relations possess the following skills: an ability to determine
(Paige 1996). participants needs, to design the training course
(set the goals, objectives, content, and selection of
Theoretical Background methods), and to implement and assess the program.
Any program of cross-cultural training is trying to The trainer also needs to have deep cross-cultural
answer the question “How?”: how an individual can knowledge, cross-cultural self-awareness, familiarity
establish interpersonal contact with other cultures, how with the developmental models of ethnic identity, an
he/she can acquire its values, norms, role structure, understanding of the concept of culture shock, adapta-
etc. This kind of training is supposed to cause a change tion and acculturation, as well as an intercultural edu-
in the learner’s attitudes by developing sensitivity to cation in general. Ideally, he/she should be a mediator
intercultural differences and cross-cultural competence. between cultures. According to various sources, the
Even though any cross-cultural training aims to following personality traits, values, and attitudes are
develop or improve the awareness, emotions, and required: tolerance to uncertainty, flexibility in cognitive
behavior of trainees, the program itself might put style and behavior, possession of a clear idea of his/her
different emphasis on the particular field where key own ethnic identity and universal values, openness to
results need to be achieved – cognitive, emotional, or a variety of views, interest in others, empathy, and the
behavioral. The cognitive approach focuses on giving tendency to lean toward cooperation during conflict.
students information about cultures and cultural dif- Finally, an effective intercultural trainer strictly obeys
ferences and helps learners to understand how stereo- ethical and “do no harm” principles.
types and prejudice affect their interaction with members
of other cultures. Therefore, its objectives are grounded Important Scientific Research and
in knowledge and social representations. The emotional Open Questions
approach focuses on transforming attitudes related to The end of the World War II marks the beginning of the
intercultural interaction by changing feelings toward cross-cultural training field, when Edward T. Hall drew
“others” (from prejudice to tolerance), and teaches attention to the lack of adequate training materials.
learners how to manage emotional reactions (such as This continued to be the case until the mid-1970s
anxiety, fear, or anger) during contact with other cul- when the first handbooks on intercultural communi-
tures. The behavioral approach is designed to develop cation were published. The beginning of research and
skills which are necessary for effective interaction with experimentation produced many of the training tech-
other cultures (Bhawuk and Brislin 2000). niques commonly used today. Familiar methods such
Regardless of the methodological approach chosen, as role plays, critical incidents, case studies, and simu-
cross-cultural training should be built on the princi- lations provided a point of entry for engaging in the
ple of cultural universalism. Any case where an research and theory building that would produce strat-
intercultural trainer suggests some customs, values, egies to prepare people to function interculturally.
and norms of any culture are “right” or “wrong” In the late 1980s, cross-cultural training became widely
could draw the student back to the ethnocentric posi- demanded, therefore programs appeared that were
tion. On the other hand, extreme cultural relativism more sophisticated and targeted to specific audiences.
Cross-cultural Training C 857
As the importance of customizing approaches and 2000). Researchers have noted the positive effect of
activities were taken into account for an extensive commonly used programs in following five phenom-
range of cultural variations, it also gave rise to new ena: the personal growth of trainees, a positive change
models (Pusch 2004). in the perception of other groups and relationships
A typical training program combines didactic and with representatives, adaptation, and achievement in C
experiential methods with either a culture-general work and study. Positive effects of the cross-cultural
or a culture-specific approach (Cushner and Brislin training have also been claimed in some recent research
1997). Each type works with a specific set of methods: for a wide variety of measures, such as subjective
(a) Experiential culture-general training focuses on experience of the training, interpersonal relations,
a trainees’ cultural self-awareness (the model of cul- intercultural sensitivity, and intercultural adjustment
tural identity). Such an approach is implemented in (Van de Vijver and Breugelmans 2008).
cross-cultural workshops by exploring how trainees’ However, the problem of assessing cross-cultural
own socialization has influenced their perception, effectiveness is still very far from being solved. Up to
attitudes, stereotypes, and behavior. Another method today, the effectiveness of very few procedures has been
is culture-general simulation which is usually explicitly demonstrated. For most training programs,
constructed as a “meeting of two cultures” and gives validity data remains absent, which means there are no
trainees an opportunity to gain an experience of methodological foundations for many cross-cultural
belonging to an imaginary culture with its norms, procedures (Van de Vijver and Breugelmans 2008). The
values, and behaviors that are different from the main reasons for that are: (a) problems with identifying
trainees’ native ones. (b) Experiential culture-specific or designing adequate tools to assess cross-cultural
training uses culture-specific simulations and role competence; (b) problems with establishing causality
plays, which attempt to help trainees learn how to in studies of cross-cultural research in general; (c)
interact effectively with members of a specific group. problems of obtaining adequate samples due to the
Cultural assimilators are collections of critical incidents specificity of the subjects; and (d) problems relating
that relate to the experiences of people from two or to interpreting the results of a controlled experiment
more cultures who face the problem of resolving some such as training.
task. Another powerful program is intergroup dialogs, Despite identified problems, demand for cross-
which could be used in situations of disagreement and cultural training continues to grow, and its methods
conflict between different ethnic communities and pro- are being adapted and implemented for many spe-
vide an opportunity to make a mutual step to find a cific audiences in education, health care, hostage nego-
common ground. In many cases, the lack of understand- tiation, dispute resolution, law enforcement, the media,
ing between people from different cultures occurs on the politics, and even cyberspace (Pusch 2004). In dealing
level of interpretation and has no basis on a behavioral with cross-cultural training it is important to take into
level. Therefore attributive culture-specific training consideration that according to many experts in the
focuses on the way people from different cultures inter- field, “intercultural training is both an art, which is
pret the reasons for different behavior. Attributive train- appropriately passed on by experienced teachers, and
ing helps to make expectations about the possible a science, which is appropriately winnowed by empir-
behavior of an individual from different cultures more ical research” (Bennett et al. 2004, p. 8).
accurate and contributes to the development of isomor-
phic attributions. (c) Didactic culture-general training Cross-References
is mainly based on cognitive approaches such as lec- ▶ Competency-Based Learning
tures, films, videos, and culture-general assimilators. ▶ Cross-Cultural Factors in Learning and Motivation
(d) Didactic culture-specific training provides opportu- ▶ Cross-Cultural Learning Styles
nities for trainees to gain information about specific ▶ Developing Cross-cultural Competence
cultures. Training methods include culture-specific ▶ Enculturation and Acculturation
briefings, culture-specific assimilators, and readings. ▶ Learning and Training: Activity Approach
The most popular programs of cross-cultural train- ▶ Social Influence and the Emergence of Cultural
ing were verified in the 1990s (Bhawuk and Brislin Norms
858 C Cross-Disciplinary Education
can be seen as one specific type of transfer of learning The connection between CLI and transfer of
restricted to language-related knowledge being applied learning is also apparent in overlap in factors that
in situations involving the use of another language. influence CLI and factors that influence transfer of
learning. One of the main factors linked to CLI is
Theoretical Background perceived cross-linguistic similarity: CLI is more C
CLI has been a central topic in research and theory likely when an individual perceives similarity between
on second language acquisition (SLA). In a seminal two languages, and less likely when an individual
publication, Odlin (1989) traced the origin of scholarly perceives difference between two languages. Along
work on CLI back to nineteenth century debate about the same lines, individuals’ perceptions of similarity
the effects of language contact and mixing on language and difference (i.e., between tasks and contexts) are
classification and change, for example in the study seen as a major influence on transfer of learning. Other
of pidgins and creoles. In the twentieth century, influ- factors that have been linked to both CLI and transfer of
ential early publications on CLI include Weinreich’s learning are (a) knowledge base (e.g., language profi-
(1953) examination of CLI in the phonetic, grammat- ciency), (b) amount and type of practice, (c) attitudes
ical, and lexical systems of bilinguals, and Lado’s (1957) and motivation, and (d) sociocultural context.
manual on contrastive analysis, which included claims In addition, although transfer of learning is a
that a systematic comparison of two languages could broader construct than CLI, transfer of learning
be used to predict where second language learning research has, like CLI research, examined language-
difficulties would occur. In the 1960s and 1970s, SLA related knowledge, for example vocabulary, grammar,
research expanded in scope and emphasis was placed and pronunciation, as well as reading and writing skills
on factors other than CLI that influence second lan- and strategies. For instance, in an investigation of the
guage learning, such as factors that influence first development of reading skills among elementary
language learning. However, CLI continues to receive school students, Martin-Chang et al. (2007) examined
a great deal of research attention (e.g., in influential techniques for learning new words; this study explored
journals such as Studies in Second Language Acquisition, whether students’ knowledge of new words learned
Language Learning, and Applied Linguistics), and the through different techniques (e.g., individually on
body of scholarly work on this phenomenon shows in flashcards; in the context of a story) transferred to
an increasingly diverse collection of contexts the con- novel reading tasks. Similarly, Williams et al. (2005)
sistency with which CLI plays a role in SLA. studied transfer of learning from a reading comprehen-
In the extensive body of literature on CLI, connec- sion instruction program; this study examined whether
tions between CLI and transfer of learning are appar- training students to use strategies to analyze the struc-
ent. Odlin (1989) pointed out that CLI is seen by ture of a text had any influence on their subsequent
many SLA scholars as a construct that was appropriated performance reading other texts. These and other
from psychology research and theory on transfer of transfer of learning studies that investigated individ-
learning in the mid-twentieth century. Furthermore, uals’ first language knowledge are different from CLI
SLA scholars have suggested that CLI involves more research that by definition examines the interface of
general cognitive processes: Ringbom (1986), for two languages; however, such transfer of learning
example, suggested that it is beneficial to view the research does examine transfer of language-related
source of CLI (e.g., a first language) as only part of knowledge, which is a characteristic it shares with
the knowledge base an individual has that can be trans- CLI research.
ferred, while Faerch and Kasper (1986) suggested that Finally, although CLI research might be seen as
CLI can be seen as a case of a learner extending existing focusing on relatively deep learning (i.e., an individ-
knowledge to new contexts. Also, in a more recent ual’s first language, which has typically been learned
overview of scholarly work on CLI, Jarvis and Pavlenko over an extensive period of time) compared to transfer
(2008) argued that CLI involves not only traditional of learning research (i.e., which might involve studying
categories of language (e.g., phonology, syntax) but transfer after only short periods of training), CLI
also higher-level cognitive concepts (e.g., the way research also can involve more shallow learning. CLI
objects are categorized). has been viewed specifically as the influence of an
860 C Cross-Linguistic Influence and Transfer of Learning
individual’s first language on second language learning context was more accurate if the words had been
and use; however, current definitions of CLI tend to be learned in context (i.e., in a story). From a CLI
broader and also include the influence of a second perspective also, it is worth asking if the influence of
language on the learning and use of a third language first language word knowledge on second language
(or a fourth language, etc.), as well as the influence of learning and use would also be constrained by the
a second language on first language use. Since individ- congruence between learning tasks and transfer tasks.
uals learn second languages in a variety of ways, there is Similarly, Williams et al. (2005) found that training
tremendous variety in the depth of learning associated students to use several strategies to analyze the struc-
with a second language. This variation means that the ture of a compare/contrast expository text had a posi-
source of CLI may not always be as deeply learned as tive influence on their subsequent performance with
a first language system. For example, Kecskes and texts with a similar structure, but not with texts with
Papp (2000) examined CLI among secondary school a different structure. From a CLI perspective also, it
students whose first language was Hungarian and who is worth asking if the use of first language reading
were studying English, French, or Russian as a second strategies to read second language texts would be
language; the findings showed that some types of constrained by similarity in genre between texts used
second language instruction had an influence on the in learning tasks and transfer tasks. Similar questions
students’ use of their first language (e.g., use of subor- can be generated from the numerous other transfer
dinate clauses when writing in Hungarian). of learning studies that have examined language-
related knowledge.
Important Scientific Research and Finally, it is important to ask if and how findings
Open Questions from other transfer of learning research (i.e., research
From a CLI perspective, the connection between CLI that involved knowledge less directly related to lan-
and transfer of learning raises important questions. guage) may also be relevant to CLI. For example,
First, can theories related to transfer of learning help transfer of learning research has focused on the influ-
shed light on unanswered questions about CLI? For ence of factors such as attitudes and motivation, and
instance, one of the central unanswered questions sociocultural context (Haskell 2001). SLA researchers
about CLI is how individuals’ perceptions of similarity have pointed to such factors as relevant to CLI as well.
between languages are triggered. From a transfer of A study by Kecskes and Papp (2000) examined CLI
learning perspective, theories have been offered for among secondary school students who spoke Hungar-
how the human brain identifies similarities in incom- ian as a first language and were studying English,
ing information, for example by being hardwired with French, or Russian as a second language, and CLI
a kind of harmonic structure that facilitates recognition was reportedly influenced by learning contexts/tasks
of similar relationships in different situations (e.g., (i.e., CLI from second language learning to first lan-
relationships between notes in a song that one is famil- guage use varied with the kind of second language
iar with but that is being played in a key higher or lower instruction students received) and student motivation.
than before) (Haskell 2001). Such accounts may also be Both learning tasks/contexts and motivation were also
relevant to CLI. highlighted as a potentially important nonlinguistic
In addition, which findings from transfer of learn- influence in Weinreich’s (1953) influential study of
ing research that has involved language-related knowl- CLI. Finally, the influence of sociocultural context is
edge might apply to CLI? For example, Martin-Chang reflected in the way individuals may adjust their lan-
et al. (2007) found that elementary school students’ guage use patterns – in a way that involves more or less
transfer of first language word knowledge in first lan- CLI – depending on the perceived identity of the per-
guage use situations depended on the congruence son with whom they are speaking (Jarvis and Pavlenko
between the learning technique and the kind of transfer 2008). However, CLI research on these particular fac-
task: Performance on a transfer task involving reading tors is limited. Relevant questions therefore include the
isolated words was more accurate if the words had been following: How do motivation and sociocultural con-
learned in isolation (i.e., on flash cards), and perfor- text influence CLI? Might CLI be more likely in some
mance on a transfer task involving reading words in cases if an individual feels motivated to make use of
Cross-Modal Learning C 861
modality, or the output of several modalities can be feature vectors still reside in the individual modalities,
combined to achieve better performance or robustness. but we construct several intermediate classifiers, which
One example of this is speech recognition. In situated are no longer independent, and combine them. This
dialogue, recognizing sequences of words in an audio requires a close-coupling of semi-synchronous learning
signal can be greatly improved by information about processes, based on interconnected representations, and
the situated context (what is there to be seen, what is leading up to the formation of cross-modal concepts.
there to be done, what have we talked before), and We can also identify a third type of cross-modal
through observation of the speaker. Context and obser- learning that is performed on a higher level of abstrac-
vations aid disambiguation during processing of the tion. Here, a model is acquired that connects modal
auditory signal, possibly also correcting misheard conceptual structures from different modalities by
words, or filling in (grammatically) missing words. learning associations between them. For example, let
For example, lip reading can greatly aid recognition. us suppose that we want to recognize a cup of coffee.
In a continuous learning process, successfully recog- A cup can be recognized visually. Yet, to recognize what
nized lip poses can supervise learning of audio-based is inside the cup we need another sense – smell. We
speech recognition ability. The other way round, cor- need to combine information from both modalities
rectly recognized audio input can provide labels to aid to determine that there is a cup of coffee on the table
the learning of lip reading. This process of coupled and not a cup of a black tea. The learning of required
supervision during learning is also known as co- concepts could be performed largely independently, in
learning. In the end we obtain two classifiers, one in each modality individually. At some point though we
each modality, that can be used individually, or they need to learn to combine the concepts of the cup and
can be combined to further increase the success of the coffee into a concept of a “cup of coffee.” The final
speech recognition. This type of cross-modal learning representation therefore consists of representations
is thus based on a weakly coupled interaction of data from several modalities.
from different modalities, which is done on a rather high Cross-modal learning is related to principles of
level of abstraction. In the case above we assumed that fusion of data from different sensors (Clark and Yuille
both modalities mutually drive the learning in the other 1990), also known as multisensory processing in natu-
one. This process can in principle be unidirectional. If ral cognitive systems (Stein and Meredith 1993). Dif-
the information in one modality is much more reliable, ferent processes interact in a cognitive system to form
it can drive the learning in another modality. a coherent interpretation of experience, based on the
In closely coupled cross-modal learning, learning combination of information obtained through several
processes are more intertwined. A model is learnt by modalities. The process of learning how to combine
combining information from different modalities into this information is a kind of cross-modal learning.
a common level of representation, and then using this As already mentioned, we can consider the term
level as a starting point to build a common cross-modal modality in its wider sense. This includes derived
classifier or predictor. As a result, inference with the modalities. In this case, the type of information that
acquired model requires information coming from characterizes a modality is not attached directly to
several modalities, and cannot be achieved within a sensor, but to a process which interprets the sensorial
a single modality only. This approach is often used in data. For example, suppose that we have a place recog-
sensorimotor learning. Here, low-level features from nition approach that is based on both visual images,
a visual modality and motor (or proprioceptive, or and 3D point clouds representing geometrical struc-
haptic) modalities are merged. Based on the obtained ture. The images may be obtained using a camera. The
cross-modal features, higher-level sensorimotor con- 3D point clouds are obtained using a laser range finder,
cepts are learned. For example, from low-level visual or, alternatively, both the images and the 3D data can
features describing objects and low-level features param- be obtained using a stereo rig. In both cases we can
etrizing actions that could be performed, a model is conceive of the learning of representations of places as
learned that predicts what happens with a particular a kind of cross-modal learning, although in the second
object if a particular action is applied (through classifi- case we have one sensor only. In computer vision, it is
cation or regression). Another example is when the very often favorable to extract several visual cues (such
Cross-Modal Learning C 863
as color, texture, borders, shape, motion), and combine A fundamental aspect of embodied cognition is that
them in order to obtain better classifiers. We can look at understanding is ultimately based in how a cognitive
the learning of such combined classifiers as at a kind of system experiences the world. Since the cross-modal
cross-modal learning as well. learning is based on processing and relating informa-
The relevance of cross-modal learning is alike for tion from several (sensory) modalities it may play an C
natural and artificial cognitive systems (Christensen important role in bringing about grounded forms of
et al. 2010). Both continuously learn to extend their cognitionde (de Sa and Ballard 1998).
knowledge of acting in dynamic environments. The We also have to address the terminological issues,
ability to connect possibly asynchronously devel- since the terms related to cross-modal learning are not
oped models across different modalities provides consistently used in the literature. Sometimes, the term
an important basis for a grounded form of self- cross-modal learning is used only to refer to strongly
understanding. The possibility to interconnect and coupled types of cross-modal learning. Also, the term
thus form an interpretation that is coherent across modality is sometimes used in its narrower sense, con-
multiple modalities indicates what is known relative sidering sensory modalities only. Here, we adopted the
to some experience. Failure to do so may indicate broader meaning of both terms. There is also another
a knowledge gap, and can function as a trigger for term in the literature that is often used to describe
self-aware learning. a similar phenomenon, the term multimodal learning.
One meaning of this term refers to (human) learning
Important Scientific Research and based on different multimedia material involving dif-
Open Questions ferent human senses that facilitate learning. The second
There are arguments for learning to be based on meaning of this term is very close to the meaning of
association, and for learning to be mediated by a cross-modal learning as defined above. Sometimes this
(developing) categorical system. Very often, the inter- term relates to forms of weakly coupled cross-modal
connectivity between modalities is mediated by cate- learning, while very often cross-modal and multimodal
gorical structure. Effectively this establishes a triadic learning are used interchangeably with the same mean-
relation between modalities. The conceptual structures ing (synonyms).
in the modalities can be connected because they can
be understood as related by virtue of their reference
Cross-References
to a shared categorical ground. The arguments for this
▶ Active Learning
type of learning, based on the formation of a mediating
▶ Adaptation and Learning
categorical structure, arise from, for example, child-
▶ Cognitive Models of Learning
hood cognitive development. In word learning it is
▶ Cognitive Robotics
shown that a purely associative, unmediated account
▶ Embodied Cognition
(“child-as-data-analyst,” Sloutsky (2003)) cannot
▶ Learning and Understanding
appropriately account for categorical generalizations
a child is able to make (“child-as-theorist,” Waxman
and Gelman (2009)). The use of mediating categories References
both helps generalization of sensory input beyond Barsalou, L. W. (1999). Perceptual symbol systems. Behavioral and
actual experience, and allows for representations to be Brain Sciences, 22, 577–609.
Christensen, H. I., Kruijff, G. J. M., & Wyatt, J. L. (2010). Cognitive
ultimately grounded in, and influenced by, the embodi- systems (COSMOS 8). Berlin: Springer.
ment of the system (Lakoff and Johnson 1999; Barsalou Clark, J. J., & Yuille, A. L. (1990). Data fusion for sensory information
1999; Glenberg 1997). On the other hand, in many processing systems. Norwell: Kluwer Academic.
cases the modalities interact on a much lower level, de Sa, V. R., & Ballard, D. (1998). Category learning through
like in the case of sensorymotor learning. It is still an multi-modality sensing. Neural Computation, 10(5), 1097–1117.
Glenberg, A. M. (1997). What memory is for. Behavioral and Brain
open question what roles do these different forms of
Sciences, 20(1), 1–55.
learning play in specific types of cross-modal learning, Lakoff, G., & Johnson, M. (1999). Philosophy in the flesh: The
whether in natural or artificial cognitive systems embodied mind and its challenge to western thought. New York:
(Philipona et al. 2003). Basic Books.
864 C Cross-Sectional Research
Philipona, D., O’Regan, K., & Nadal, J. P. (2003). Is there something Theoretical Background
out there? Inferring space from sensorimotor dependencies. During language acquisition, children learn a lexicon
Neural Computation, 15(9), 2029–2049.
containing many thousands of associations between
Stein, B. E., & Meredith, M. A. (1993). The merging of the senses.
Cambridge, MA: MIT Press. words and their meanings, at the rate of around ten
Sloutsky, V. M. (2003). The role of similarity in the development of new words a day. Children accomplish this task rapidly
categorization. Trends in Cognitive Sciences, 7, 246–251. and remarkably successfully, overcoming potentially
Waxman, S. R., & Gelman, S. A. (2009). Early word-learning entails unlimited uncertainty about the meaning of every
reference, not merely associations. Trends in Cognitive Sciences,
new word they encounter, and identifying some
13(6), 258–263.
aspects of word meaning after only a very few expo-
sures through so-called fast mapping. Quine (1960)
famously illustrated the problem of referential uncer-
tainty through the story of an imaginary anthropolo-
gist working with a speaker of an unfamiliar language:
Cross-Sectional Research when a rabbit runs past, the speaker shouts “gavagai,”
and the anthropologist tentatively notes that this new
▶ Longitudinal Learning Research on Changes in
word means “rabbit.” Quine’s insight was to point out,
Learning of University Students
however, that the anthropologist can never be sure that
“gavagai” means “rabbit,” no matter how many future
clarificatory tests are carried out; it could, after all, have
an infinite number of possible meanings of varying
plausibility, including “animal,” “white,” “undetached
Cross-Situational Learning rabbit parts,” “dinner,” or “it will rain.”
Yet despite the philosophical problem of unlimited
ANDREW D. M. SMITH1, KENNY SMITH2 referential uncertainty, children clearly do learn large
1
Literature and Languages, School of Arts and lexicons, and the focus of much research into word
Humanities, University of Stirling, Stirling, UK learning has been on providing explanations for this.
2
Linguistics and English Language, School of The dominant approach has been to identify mecha-
Philosophy, Psychology and Language Sciences, nisms which allow the learner to exclude from consid-
University of Edinburgh, Edinburgh, Scotland, UK eration many meanings which are theoretically possible
but in reality spurious, thus reducing the level of refer-
ential uncertainty in the input to a more manageable
Synonyms level, and simplifying the task of determining the
Associative learning word’s true meaning. A number of heuristics have
been put forward: interpreting behavioral cues in
Definition order to identify the speaker’s focus of attention;
Cross-situational learning is a technique for learning assuming that novel words are more likely to refer to
the meanings of words across multiple exposures, whole objects rather than their parts or properties;
despite uncertainty as to the word’s meaning on each building on existing knowledge about the meanings
individual exposure. The cross-situational learner of other words and assuming that new words will
encounters a word in a number of different situations, have different meanings; making use of the syntactic
each of which provides a set of multiple candidate context in which the new word is presented to infer
meanings; the learner determines the word’s meaning aspects of its meaning (see Bloom 2000, for review).
by selecting from those meanings which reliably recur While quantifying the impact of such heuristics is
across situations. Cross-situational learning is less cog- problematic, it is clear that some referential uncertainty
nitively demanding than many other models of word is likely to remain even after the application of many or
learning, because it does not require a learner to unam- all of them; the utility of cross-situational learning
biguously identify a word’s meaning on a single stems from the fact that it allows words to be learnt
exposure. despite the existence of residual uncertainty.
Cross-Situational Learning C 865
Cross-situational learning works by amalgamating by the environment, it will be sifted out of the set of
information about the meaning of a word from across possible meanings; in homonymous or polysemous
the various different situations in which that word situations where the word has more than one intended
occurs. Each separate context in which the word is meaning (e.g., the English word “bank”), none of the
used yields a (possibly infinite) set of possible candi- intended meanings will appear in all exposures, and C
date meanings, which is potentially reduced through thus the set of possible meanings will be empty (i.e.,
the application of word-learning heuristics such as situations in which “bank” is used as a verb denoting
those described above to a finite set of candidate mean- turning will probably not feature financial institutions
ings (including the true meaning); the same word in their set of likely meanings; likewise, situations in
uttered in a different context may of course yield which it is used as a noun will not reliably feature the
a different set of candidate meanings. Candidate mean- act of turning).
ing sets from different contexts can be combined, These vulnerabilities stem from the pure cross-
enabling the learner to identify the most likely correct situational learner maintaining the maximal amount
meaning, for instance, by identifying the meaning of cross-situational information, namely, an accurate
which lies at the intersection of the sets, as shown in set of candidate meanings which always occur with
Fig. 1. Although each exposure to a new word may the word. At the other end of the spectrum, a learner
provide a large number of possible meanings, and could make minimal use of cross-situational informa-
thus a large degree of referential uncertainty, successive tion by simply remembering a single one of the mean-
exposures in different contexts will gradually reduce ings suggested in a previous exposure, and maintaining
the uncertainty, eventually eliminating it completely this as their preferred meaning so long as it is also
by winnowing the set of possible meanings down to suggested by the current context. Between these two
the true meaning alone. extremes lie an infinite number of potential cross-
This eliminative approach to cross-situational situational learning strategies, much more resilient to
learning illustrated in Fig. 1, however, is vulnerable noise, yet less powerful than pure cross-situational
to failure in a number of real-world circumstances learning (Blythe et al. 2010). In particular, a frequentist
(see Gleitman et al. 2005, for discussion): In noisy strategy, where learners track the frequency with which
situations where the intended meaning is not suggested candidate meanings co-occur with the target word,
appears to match well the data from experimental
tests of cross-situational learning (Yu and Smith 2007;
Target word: “horse” Smith et al. 2011).
demonstrated that his cross-situational learning pro- timescale. Quantifying this critical point, however, is
cedures could be specified so that the algorithm could still problematic, not only because of the difficulties in
recover from errors originating from environmental accurately quantifying the referential uncertainty of
noise and homonymy. More recent formal models naturalistic data, but also because the experimental
(e.g., Yu et al. 2005) have successfully demonstrated evidence for when and how people shift learning strat-
that cross-situational learning can be used to infer the egies is currently rather minimal. Furthermore, existing
meanings of words from increasingly complex and research into different variants of cross-situational
realistic, though still small, corpora of natural language learning has primarily been carried out on adults, pos-
use. Mathematical models, meanwhile, have shown ing the question of whether children shift strategies in
that cross-situational learning is viable not just with response to task demands in the same way as adults, or
small corpora, but also scales up to the learning of whether they even use the same cross-situational learn-
large, human-size vocabularies within reasonable time- ing strategies at all.
scales (Blythe et al. 2010). Despite significant levels of
referential uncertainty at each exposure, the relative Cross-References
learning speed disadvantage of cross-situational learn- ▶ Associative Learning
ing compared to an idealized fast-mapping learner is ▶ Embodied Cognition
surprisingly small. There is, therefore, no necessary link ▶ Heuristics and Problem Solving
between the ability to learn individual words rapidly ▶ Matching
and the ability to acquire large vocabularies. ▶ Meaningful Verbal Learning
A body of research has demonstrated that both
References
adults and infants can effectively exploit cross-
Akhtar, N., & Montague, L. (1999). Early lexical acquisition: The role
situational learning information when learning small of cross-situational learning. First Language, 19(57), 347–358.
numbers of words (e.g., Akhtar and Montague 1999; Yu Bloom, P. (2000). How children learn the meanings of words.
and Smith 2007), using both naturalistic and more Cambridge, MA: MIT Press.
controlled (and therefore quantifiable) stimuli. The Blythe, R. A., Smith, K., & Smith, A. D. M. (2010). Learning times for
effectiveness of cross-situational learning in humans large lexicons through cross-situational learning. Cognitive Sci-
ence, 34(4), 620–642.
is affected by the degree of referential uncertainty: Gleitman, L. R., Cassidy, K., Nappa, R., Papafragou, A., & Trueswell,
performance (in terms of number of words learnt) J. C. (2005). Hard words. Language Learning and Development,
decreases as referential uncertainty increases. Further- 1(1), 23–64.
more, increasing referential uncertainty appears to Quine, W. V. O. (1960). Word and object. Cambridge, MA: MIT Press.
change the mechanism by which cross-situational Siskind, J. M. (1996). A computational study of cross-situational
techniques for learning word-to-meaning mappings. Cognition,
learning takes place, with increased referential uncer-
61, 39–91.
tainty prompting a shift from a pure eliminative strat- Smith, K., Smith, A. D. M., & Blythe, R. A. (2011). Cross-situational
egy to a less demanding, more nuanced, frequentist learning: An experimental study of word-learning mechanisms.
equivalent (Smith et al. 2011). Cognitive Science, 35(3), 480–498.
This recent emphasis on examining when alter- Yu, C., & Smith, L. B. (2007). Rapid word learning under uncertainty
native cross-situational strategies are employed by via cross-situational statistics. Psychological Science, 18(5),
414–420.
learners leads to a number of currently open questions. Yu, C., Ballard, D. H., & Aslin, R. N. (2005). The role of embodied
Increasing referential uncertainty naturally increases intention in early lexical acquisition. Cognitive Science, 29(6),
the time it takes to learn a lexicon, yet weaker forms 961–1005.
of cross-situational learning (those which make less
efficient use of cross-situational statistics) are dispro-
portionately affected by increases in referential uncer-
tainty than stronger forms.
Crosswise Research
At some point, therefore, increasing referential
uncertainty will make a human-size lexicon impossible ▶ Longitudinal Learning Research on Changes in
to learn by cross-situational learning in a reasonable Learning of University Students
Cue Summation and Learning C 867
Delivery Method
Learning content delivered using methods such as
printed handouts, digital text, computer-based graph- Cued Recall
ical presentation, face-to-face instruction, audio
video–based instructions, and computer-based inter- PHILIP A. HIGHAM, MEHMET A. GUZEL
active applications may influence the learning process School of Psychology, University of Southampton,
(Laurillard 2002). Southampton, UK
Delivery Platform
The delivery may influence how learners are involved Synonyms
in the learning process (e.g., whether individuals work Primed recall; Prompted recall
at their own pace or with a group, learning from home,
learning on the move, learning via different technolo- Definition
gies like computers and mobile phones). Cued recall refers to retrieving information from
long-term memory using aids or cues. Cues can be
Learning Motivation external stimuli, such as words, sentences, incomplete
Learning motivation can be defined as learners’ action pictures, letters within a word, and so on, as long as
or behavior behind engaging in the learning process, they have some kind of connection to the to-be-
which reflects on the success of the learning process. remembered (target) information. That connection
might be a semantic or associative relationship, tem-
Cross-References poral co-occurrence of a cue and target, or the cue
▶ Audiovisual Learning could actually be the target presented in an incomplete
▶ Auto-associative Memory and Learning form. For example, recall of the target TIGER might be
▶ Communication Theory cued with lion, A ____ has stripes, an incomplete
Cued Recall C 869
drawing of a tiger, or TI___. Cues can also be internal. when cued with stripes. Recognition failure is problem-
For example, people may be asked to think about what atic for generate–recognize theory because phase-4
they were thinking about, their mood, or the spatio- recall was limited by two bottlenecks (i.e., the target
temporal context at the time that they encountered had to be both generated and then recognized),
a target in an attempt to cue retrieval of it. Cued recall whereas phase-3 recognition was only limited by one C
is often contrasted to free recall, which mainly refers to (i.e., the target item had already been “generated” dur-
the process of retrieving information from long-term ing phase-2). Thus, if recall really involves successive
memory without the provision of any explicit cues. stages of generation followed by recognition, then
Along with recognition memory (discrimination of pre- Tulving and Thomson reasoned that recognition fail-
viously-encountered stimuli from novel ones), free ure of recallable words could not have occurred.
recall and cued recall constitute the most common Tulving and Thomson (1973) argued, instead, that
explicit tests of memory. their results were explained by the encoding specificity
principle. This principle asserts that the effectiveness of
Theoretical Background retrieval cues is largely determined by whether they
Most theories of cued recall posit that retrieval of target were present at the time the target information was
information can be achieved via two routes. The first encoded (i.e., encoded specifically with it). Thus, rec-
is an efficient direct route, involving recollection or ognition failure occurred in their experiment because
ecphory of the sought-for details. The second route to the cues that were encoded specifically with the targets
recall if the direct route fails involves a process of were not available during phase-3, whereas they were
generating plausible candidates, and then attempting available during phase-4. Consequently, phase-4 recall
to recognize the target from amongst the candidates. was superior to phase-3 recognition.
Support for the generate–recognize route to recall The encoding specificity principle and related ideas
has been garnered from studies demonstrating that such as transfer appropriate processing highlight the
cues that were strongly associated to a target were importance of matching the conditions of encoding
more effective than cues that were not (e.g., Bahrick and retrieval to elicit good cued-recall performance.
1970). However, generate–recognize models were crit- Similarity between the conditions of encoding and
icized in the early 1970s by Tulving and colleagues (e.g., retrieval was important because both external and
Tulving and Thomson 1973). The main source of their internal retrieval cues are embedded in those condi-
criticism was the phenomenon of recognition failure of tions. Examples of external retrieval cues include envi-
recallable words, which was demonstrated in an exper- ronmental cues (e.g., the particular room in which
iment involving four stages. First, weakly associated target information was learned/retrieved) or the cue
cue-target pairs were studied for a later memory test words that were presented along with target informa-
(e.g., stripes–TIGER). Next there was a phase intended tion, as in the Tulving and Thomson’s (1973) experi-
to encourage generation of targets, during which par- ments. However, retrieval cues are also embedded in
ticipants produced any words that came to mind in internal contexts, such as the type of mood learners
response to strong associates of the targets (e.g., lion). had or the type of processing that they engaged in
In a third recognition phase, there was an attempt to during encoding (e.g., whether target information
recognize targets from amongst the generated items, was processed for meaning versus in terms of how it
which was often unsuccessful. Finally, the weak cues sounded). In all these cases, if the encoding and
that were presented with the targets during the initial retrieval conditions are matched, then the likelihood
study phase were presented again to cue recall of targets is increased that there will be retrieval cues available at
(e.g., stripes–?). Recognition failure of recallable words test that were specifically encoded with the target infor-
was revealed in that targets not recognized during the mation during learning, thus enhancing recall.
recognition phase of the experiment (phase-3) were
often recalled later in the presence of weak cues Important Scientific Research and
(phase-4). In other words, TIGER might have been Open Questions
generated to lion, but not recognized as a target, even Tulving and Thomson’s (1973) important work on the
though it was successfully recalled in the final phase encoding specificity principle demonstrated that not all
870 C Cued Recall
cues are equally effective at retrieving target informa- context, metacognition refers to people’s assessment of
tion from memory. Although it is generally accepted their own memory system and how well it works. For
that cued recall is superior to free recall, exactly example, people may be asked to make a judgment of
which aspects of cues that cause this superiority has learning (JOL) about how well they will remember
been the source of considerable research. The encoding a target from a cue-target pair (e.g., ocean–TREE)
specificity principle states that cues encoded with when presented with the cue later on (e.g., ocean–?).
the target information are the most effective, but Nelson and Dunlosky (1991) showed that if this judg-
are there extralist cues (cues not presented during ment is delayed for a short period, JOLs are more
encoding) that are effective for retrieval? By comparing accurate than if they are made immediately after study-
free recall with recall aided with different types of cues, ing the pair. One way of explaining this effect is in
it is possible to answer this question. However, it is terms of the types of information that people rely on
important to consider the fact that, when making this when making JOLs. Immediately after studying the
comparison, free recall is free only of externally pro- cue-target pair, both items are in short-term memory
vided cues, such as words, letters, or sentence stems. In making it difficult to use a covert retrieval attempt of
free recall, people are likely to strategically prompt the target using the cue as a basis for the judgment.
themselves with covert cues. For example, a professor However, if some time passes, and the items are no
attempting to remember who attended a recent meet- longer in short-term memory, JOLs may indeed be
ing might covertly “walk around the department,” gen- based more on the success of a covert retrieval attempt,
erating colleagues’ names from memory for the spatial which is a reasonably good predictor of later success.
location of their offices, and then attempting to recog- Other research on metacognitive processes in
nize those names from the meeting. Hence, even so- recall involves trying to separate the front-end compo-
called free recall is likely to be cued in some sense. The nents of the retrieval process from the post-retrieval
fact that cued recall is generally better than free recall editing processes. Higham and Tam (2005) likened
indicates that the cues people use spontaneously are these processes to the generation and recognition pro-
not as effective as those provided in most cued-recall cesses incorporated in the original generate–recognize
situations. models, discussed above. They argued that there are
There are rare cases in which free recall is superior influences on both these early and late selection com-
to cued recall, and these cases are diagnostic of the ponents of recall and that sometimes the same variable
encoding and retrieval processes involved in recall. could have opposite effects on each. For example, the
For example, Reffel (1998) found that college students higher the strength of the associative relationship
in the USA were able to freely recall about 41 of the between cue-target pairs, the more likely the target
50 states (82%). However, when recall of the states was will be generated by the cue. However, strong associa-
cued with a blank map of the United States, recall tive relationships make recognizing the target from
dropped to approximately 32 states (64%). In contrast amongst candidates more difficult because of high
to the professor in the example above, clearly those who inter-candidate similarity. Thus, despite Tulving and
were engaging in free recall of American states were not Thomson’s (1973) influential attack on early generate–
cuing themselves by spatially “moving around” a men- recognize theory, the usefulness of the distinction
tally-created blank map, otherwise they should have between the early- and late-selection processes that it
done at least as well as those actually provided with incorporates has stood the test of time.
such a map. Instead, Reffel suggested that the memorial
representations formed when students learned the Cross-References
50 states were verbal rather than visual and so the visual ▶ Autobiographical Memory
cues provided by the map were not as effective as other ▶ Cueing
verbal cues that people may have spontaneously used in ▶ Human Cognitive Architecture
free recall. ▶ Memory Dynamics
Modern research has been focused on the ▶ Recall and Effect of Repetition on Recall
metacognitive processes involved in recall. In this ▶ Retrieval Cues and Learning
Cueing C 871
principle of encoding specificity applies. That is, the That is, the targets may be falsely remembered as hav-
most effective cues are those that were learned along ing come from the wrong list because very few cues are
with target information. However, Nairne (2002) has available to correctly attribute the source of the targets
argued that the encoding specificity principle is only as one list versus the other.
part of the story; what is more critical than encoding- Cues can also be used to prompt semantic memo-
retrieval match per se is whether a retrieval cue dis- ries, that is, generic memories for definitions of words
criminates the target memory from other candidate or concepts. A common way to do this is priming,
memories. Usually, the greater the match, the more whereby the activation of target information in seman-
a cue specifies a particular target, but this need not be tic memory is facilitated by simultaneous or prior
the case. For example, Nairne discusses the case in processing of another item (prime). For example, in
which people are asked to memorize a list of homo- semantic priming, the presentation of the prime nurse
phones (words that sound the same but have different shortly before (or along with) presentation of the target
meanings; e.g., write, right, rite, write, rite, right. . .) and doctor is likely to shorten the pronunciation latency
then asked to retrieve the third item in the list. Suppose (the time to pronounce) of the target (e.g., Meyer and
further that there was an increase in the match between Schvaneveldt 1971). In this case, retrieval is not char-
the encoding and retrieval conditions because informa- acterized by accessing an event in a person’s personal
tion about how the target information sounded was past, as with retrieval from episodic memory, but
included in the retrieval cue (“the target you are looking rather by temporary activation of a particular concept
for is pronounced \’rı̄t\ or rLIt”). Although compared or entry in a lexicon. Unlike cueing of episodic mem-
to free recall the match between the conditions of ories, cueing (or activating) semantic memories is
encoding and retrieval is increased with the provision not associated with the experience of “reliving” a past
of veridical sound information, and although the event. However, partial cueing of both semantic and
detail added to the cue was encoded specifically with episodic memories can give rise to the tip-of-the-
the target information, retrieval would be unlikely to tongue state (Brown and McNeill 1966), in which peo-
improve. The sound of the target is shared with all ple report the frustrating experience of feeling that the
other candidate memories and so it is not diagnostic sought-for information is near consciousness, but just
of which particular item is being sought. Thus, out of reach. Sometimes, people can accurately report
although it is generally true that the greater the match characteristics of the target information, such as the
between the encoding and retrieval conditions, the first letters of the word, or what it rhymes with, without
better is memory performance, the match per se is being able to actually name the item. In this case,
not sufficient for good retrieval. It is discriminability the subjective experience of cueing is similar despite
of the cues – how well they specify a particular candi- the difference in the types of memories being cued.
date memory – that is really the important factor that
determines the efficacy of cues. Cross-References
In many cases, there may be a combination of ▶ Autobiographical Memory
internal and external cues that uniquely identify spe- ▶ Cued Recall
cific memories, allowing retrieval to occur without a ▶ Human Cognitive Architecture
problem. Other times, however, several of the cues in ▶ Memory Dynamics
the cue set may point to more than one memory, ▶ Recall and Effect of Repetition on Recall
causing errors in retrieval. For example, if two different
targets are encountered in two different sources where
there is a lot of cue redundancy (e.g., the targets TIGER References
Brown, R., & McNeill, D. (1966). The “tip-of-the-tongue” phenom-
and TABLE are studied in the same general experimen-
enon. Journal of Verbal Learning and Verbal Behavior, 5, 325–337.
tal context but in different lists) then source-monitoring Higham, P. A., & Tam, H. (2005). Generation failure: Estimating
errors (e.g., Higham et al. 2011) may occur despite metacognition in cued recall. Journal of Memory and Language,
a close match between the encoding-retrieval contexts. 52, 595–617.
Cultural Influences on Human Learning C 873
Cue-to-Consequence Learning
▶ Selective Associations
Cultural Historical Activity
Theory
Cultural Activities ▶ Cultural-Historical Theory of Development
recommended to use relatively simple sentences for ▶ Cross-cultural Studies on Learning and Motivation
nonnative-speaking students. ▶ e-Learning and Personalization
● Learning style preferences. The learning style ▶ Personalized Learning
preferences among Eastern students and Western ▶ Personalized Learning Systems
students were not statistically significantly different. C
However, students have different learning style
References
Boondao, R., Hurst, A. J., & Sheard, J. I. (2009). Understanding cultural
preferences in each culture group. Instructors and
influences: Principle for personalized E-learning systems. Interna-
course designers need to provide course material tional Journal of Human and Social Sciences, 4(9), 691–695.
that takes into consideration students’ individual World Academy of Science, Engineering and Technology.
learning style preferences. Kim, K. J., & Bonk, C. J. (2002). Cross-cultural comparisons of
online collaboration. Journal of Computer-Mediated Communi-
cations, 8 (1), http://www3.interscience.wiley.com/cgi-bin/fulltext/
Important Scientific Research and 120837864/HTMLSTART
Open Questions Levine, A., Oded, B., Connor, U., & Asons, I. (2002). Variation in
In the context of cultural influences on learning, one EFL-ESL peer response. Teaching English as a Second or Foreign
active line of research is the cultural awareness in Language (TESL-EJ).
e-learning. The research on cultural awareness in Liu, Y. (2007). Designing quality online education to promote
e-learning is made to ensure that culture is taken into cross-cultural understanding. In A. Edmundson (Ed.), Global-
ized E-learning cultural challenges (pp. 35–59). London:
account when developing online courses. Cultural
Information Science Publishing.
awareness is important in communication with people Schiaffino, S., Amandi, A., Gasparini, I., & Pimenta, M. S. (2008).
from other cultures. In the internationalized learn- Personalization in e–learning: the adaptive system vs. the intelli-
ing environment, students and lecturers come from gent agent approaches. Proceedings of the VIII Brazilian Sympo-
multicultural education backgrounds. This can influ- sium on Human Factors in Computing Systems (pp. 186–195).
ACM International Conference Proceeding Series. Brazil:
ence their ways of understanding education, curricu-
Sociedade Brasileira de Computação Porto Alegre.
lum, learning methods, expectations, duties, and other Welzer, T., Druzovec, M., Cafnik, P., & Venuti, M. Z. (2010). Aware-
activities in the educational process. To avoid lack ness of Culture in e-learning. Proceedings of the 9th International
of understanding, culture has to become a built-in Conference on Information Technology Based Higher Education
knowledge of each curriculum (Welzer et al. 2010). and Training (pp. 312–315). IEEE Xplore.
Another lively area of research concerns the person-
alization systems development. Research into person-
alization in e-learning can be divided into two main
directions: adaptive educational systems and intelligent Cultural Issues in Music
tutors. In an adaptive educational system the presenta- Education
tion of content and the order in which the content
is presented to the student is adapted to a student’s ▶ Multicultural Issues in Music Instruction and
model. Adaption is achieved by methods such as adap- Learning
tive ordering, hiding or removing links, and adaptive
link annotation. An intelligent tutor is like a computer
program that learns what interests a student, his or her
preferences and study habits, and provides proactive, Cultural Learning
personalized tutoring (Schiaffino et al. 2008).
DARA CURRAN
Acknowledgments Cork Constraint Computation Centre (4C), Computer
The authors would like to thank T.J. King for his assis- Science Department, University College Cork, Cork,
tance in language editing. Ireland
Cross-References
▶ Cross-cultural Factors in Learning and Motivation Synonyms
▶ Cross-cultural Learning Styles Cultural transmission
876 C Cultural Learning
Theoretical Background
Cultural learning describes the process of information Belew Experiments
transfer between individuals in a population through The experiments undertaken by Belew (1990) were
nongenetic means. Typically this is achieved through among the first to examine the effects of cultural
communication or the creation of artifacts available to transmission on a population of agents employing a
all members of a population for the purposes of cul- similar model to that employed by Hinton and
tural transmission. Nowlan (1987). A mechanism dubbed cultural advan-
Cultural learning is a model which combines popu- tage allowed fit agents to vertically transmit informa-
lation learning with a modified version of lifetime learn- tion to their offspring.
ing that allows populations to pass on knowledge to the The transmission comes in the form of a learning
next generation through nongenetic means through a bias, where agents are given a higher chance of selecting
process of communication or artifact creation, often the same locus (a “1” or a “0”) as their parents to
achieved through imitation. Much research has been replace their “?” locus. Belew found that the inclusion
conducted in the field of imitation, particularly with of this cultural advantage mechanism caused the
respect to robotics and symbol grounding in animals population to converge more quickly to the problem
and artifacts, and a number of models have been solution. Furthermore, he noted that genomes began to
developed to examine the interaction of culture and contain a much higher proportion of “?” loci than the
evolution. In addition, the simulation of culture in Hinton and Nowlan experiment indicating that
populations of artificial organisms has been the focus the addition of culture reduces selective pressure to
of much research (Yanco and Stein 1993; Denaro and find the optimal genome. Finally, oblique cultural
Parisi 1996; Hutchins and Hazlehurst 1991; Borenstein transmission was also examined, where fit agents were
and Ruppin 2003). allowed to share information with both their offspring
Cultural transmission is the transmission of infor- and a number of peers from the same generation.
mation resulting from social interactions across Results obtained from the experiment showed that
and within generations. As such, it is distinct from oblique transmission was effectively equivalent to ver-
genetic evolution although, as described further, it tical transmission given a population of sufficient size.
can interact with the evolutionary process. Transmis-
sion can occur in a number of ways, according to the Best Experiments
social interactions in question. Vertical transmission Another variation on the Hinton and Nowlan model is
describes the transfer of information from a parent described in the experiments undertaken by Best
generation to the next, such as occurs between parents (1999). The work extended the Hinton and Nowlan
and offspring. Horizontal transmission is the trans- model by allowing agents to acquire information
fer of information within a single generation, as socially, not through individual learning. In other
occurs when peers acquire behavior through imitation words, the stated aim of the work was to observe
or learning. Finally, oblique or diagonal cultural whether the evolutionary process could be influenced
Cultural Learning C 877
by cultural learning alone, without the presence of any that examined the evolution of a shared lexicon
individual learning mechanism. through repeated cultural interactions (Hutchins and
The cultural learning mechanism is different from Hazlehurst 1991). Each agent in the experiment is
Belew’s cultural advantage in that it is based on imita- represented by a neural network with a fixed structure.
tion rather than learning bias. Agents communicate The neural network possesses the standard input and C
through a horizontal cultural transmission mechanism output layers as well as two hidden layers. However, the
where agents imitate their peers. second hidden layer (closest to the output layer) is
Agents are evaluated using an external cultural fit- dubbed the “verbal input/output” layer and is used
ness function and social models are chosen accordingly for agent communication.
from the population. A learner is selected randomly Agents are randomly assigned as speakers or listeners,
from the population and for each “?” locus in its where speakers emit signals from their verbal input/
genome, the agent imitates the model agent’s value. output layer and listeners perceive and attempt to rep-
The process of learning is non-Lamarckian – imitated licate the signals using back-propagation. By iteratively
values are not passed on to the next generation. The performing these interactions, the researchers were able
experiments demonstrated that horizontal cultural to show the emergence of a shared lexicon of symbols
transmission can guide the evolutionary process and representing the agent’s environment.
that social learning may be superior to conventional
lifetime learning, leading to faster convergence. Borenstein and Ruppin Experiments
Borenstein and Ruppin suggest that imitative learning
Indexed Memory and Cultural can be harnessed in a similar manner to more explicitly
Artifacts cultural means of information transfer (Borenstein and
It has been proposed that instead of agents communi- Ruppin 2003). Like cultural learning, learning by imi-
cating directly with one another, sometimes in a tation is an alternative to supervised learning and dis-
seemingly random fashion, it may be useful to have penses with the requirement for explicit sources of
them share information through a specified medium training data such as external oracles. Much research
(Spector 1994). This medium is more easily observed has been conducted in the field of imitation, particu-
by the experimenter and direct effects can be produced larly with respect to robotics and symbol grounding in
by modifying its properties. animals and artifacts.
A population may share its information through The model employed by Borenstein and Ruppin
a centralized memory repository where individuals consists of a population of evolving agents possessing
can write and read information about their perceived a genetic encoding of a neural network and assigned
environment. One disadvantage of the approach is a number of benchmark tasks. A number of teachers
that since any agent may write to the shared memory, are selected from the population according to their
there is a risk that agents not well suited to their fitness and agents imitate their behavior using back-
environment could disrupt others by sharing errone- propagation. In this sense, the work is similar to
ous information. research on teacher/pupil interactions as undertaken
by Billard (Billard and Dautenhahn 1999). However, in
Lexicon Evolution this set of experiments, the researchers limit the behav-
The study of communication in artificial populations iors that can be learned by imitation to those acquired
has led some researchers to include fixed lexicons as innately by the teachers. In other words, pupils may not
a part of their experiment (Cangelosi and Parisi 1998). imitate behavior that the teachers have themselves
While this has provided a useful starting point, acquired through learning.
others argue that the use of a fixed lexicon is not
representative of real world language development. Noise as a Source of Diversity
Much research has been done focusing on a dynamic The success of genetic algorithm approaches to func-
lexicon in a population of communicating organisms. tion optimization problems is due in part to the algo-
Hutchins and Hazlehurst developed an experiment rithm’s capability for novelty arising from mutations.
878 C Cultural Mentoring
To investigate whether a similar scheme could be pro- Denaro, D., & Parisi, D. (1996). Cultural evolution in a population of
vided for cultural transmission an experiment was neural networks. Proceedings of the 8th Italian Workshop on
Neural Nets, Trento, Italy (pp. 100–111). New York: Springer.
conducted (Denaro and Parisi 1996) where a popula-
Hinton, G. E., & Nowlan, S. J. (1987). How learning guides evolution.
tion of agents underwent a process of cultural imitation Complex Systems, 1, 495–502.
using the teacher/pupil scenario. The teacher’s output Hutchins, E., & Hazlehurst, B. (1991). Learning in the cultural
to a given situation became the pupil’s input to allow process. In Artificial life II (pp. 689–706). Cambridge, MA:
the pupil to associate a situation in its environment MIT Press.
Spector, L. (1994). Genetic programming and AI planning systems.
with a given signal. The experiment used a purely cul-
In Proceedings of Twelfth National Conference on Artificial Intel-
tural evolution scheme, so no genetic information was ligence (pp. 1329–1334). Seattle, Washington: AAAI Press/MIT
passed on to further generations. Press.
It was found that if a population taught the succes- Yanco, H., & Stein, L. (1993). An adaptive communication protocol
sor generation in the fashion described above, the for cooperating mobile robots. In From Animals to Animats 2.
cultural information passed on would dissipate over Proceedings of the second International Conference on Simulation
of Adaptive Behavior (pp. 478–485). Cambridge MA: MIT Press.
generations. This could be reduced by applying a selec-
tive process to the choice of teacher, but this only seemed
to delay dissipation which was in the end inevitable.
It was suggested that this may have been because of
the lack of novelty in the cultural transmission and that Cultural Mentoring
an equivalent to the genetic algorithm’s mutation
operator could be the addition of noise in the signal ▶ Developing Cross-cultural Competence
from teacher to pupil. The results showed that the
populations were able to sustain communication sys-
tems over successive generations with the inclusion of
random noise.
Cultural Models
Cross-References ▶ Anthropology of Learning and Cognition
▶ Enculturation and Acculturation
▶ Imitative Learning in Humans and Animals
▶ Learning: A Process of Enculturation
▶ Lifelong Learning
▶ Population Learning Cultural Schema–Based
▶ Repeated Learning and Cultural Evolution Expectations
▶ Anthropology of Learning and Cognition
References
Belew, R. K. (1990). Evolution, learning and culture: Computational
metaphors for adaptive algorithms. Complex Systems, 4, 11–49.
Best, M. L. (1999). How culture can guide evolution: An inquiry into
gene/meme enhancement and opposition. Adaptive Behavior,
7(3/4), 289–306.
Cultural Transmission
Billard, A., & Dautenhahn, K. (1999). Experiments in learning by ▶ Cultural Learning
imitation: Grounding and use of communication in robotic
agents. Adaptive Behaviour, 7(3/4), 411–434.
Borenstein, E., & Ruppin, E. (2003). Enhancing autonomous agents
evolution with learning by imitation. Interdisciplinary Journal
of Artificial Intelligence and the Simulation of Behaviour, 1(4),
335–348. Cultural-Historical Activity
Cangelosi, A., & Parisi, D. (1998). The emergence of a language in an Theory
evolving population of neural networks. Connection Science,
10(2), 83–97. ▶ Activity Theories of Learning
Cultural-Historical Theory of Development C 879
Cross-References
▶ Activity Theories of Learning Culture and Learning
▶ Development and Learning
▶ Internalization ▶ Identity and Learning
882 C Culture and Learning and Motivation
In the 1980s and 1990s, educational and linguistic subcultures and thus have different notions on what it
studies investigated manifestations of culture in lan- means to be polite and how politeness should be real-
guage teaching and learning and concerned primarily ized in speech and behavior.
the effects of body language, eye contact, and other The second strand of research in anthropology,
overt communicative behaviors. Comparisons of cul- ethnography, and applied linguistics also includes C
turally defined behaviors focused on such common studies of specific cultures, such as Brazilian, Chinese,
anthropological constructs as hand and head move- Japanese, or Korean. Such studies examine and describe
ment, eye contact, lexical references to broad-range ways of doing, speaking, and behaving in specific
tangible and abstract entities (e.g., measures of dis- cultural and language communities, without necessar-
tance, shapes, colors, and time), forms of address, or ily undertaking to identify commonalities and differ-
terms of kinship and personal relationships that do not ences among various cultures. Both research into
exist outside the specific societies in which they are culture in general and specific cultures can be useful
used. In the 1980s and 1990s, language teaching meth- to language teachers and learners who seek to raise
odologies began to include various techniques for ana- their awareness of the inextricable relationships between
lyzing and teaching cultural behaviors together with the culture of the community and the language usage
instruction on second language skills. Many such of its speakers.
teaching techniques associated with culture learning,
however, encompassed primarily the anthropological Important Scientific Research and
views of culture and only briefly touched on underlying Open Questions
cultural assumptions, beliefs, and values (e.g., meta- In second language pedagogy, a dominant perspective
phors or conversational norms) that are invariably has emerged that language usage and the culture of its
reflected in language uses and interaction. speakers are closely bound up, and, together, they
At present, two parallel strands of research have constitute a unified domain of sociolinguistic experi-
evolved to identify the role of culture in society and ence. Many researchers in language learning and meth-
its influence on human behavior and language use. The odologists in language teaching currently hold the view
first strand includes studies of culture as it applies to that it is simplistic to imply that culture can be exam-
▶ social norms, ▶ beliefs, assumptions, and ▶ value ined, taught, and learned through exercises on reading
systems that affect practically all human activities and news media reports and advertisements. Few believe
is prevalent in the domains of anthropology, sociol- that folklore, festivals, facts, and foods (the 4-F
ogy, ethnography, and intercultural communication. approach to teaching culture) are directly relevant to
Research in these disciplines examines culture as it the impact of culture on learners’ linguistic production
applies to the structure of human societies and organi- and interactive behaviors.
zations, as well as the differences and similarities that A substantial body of research has demonstrated
exist in ▶ social worldviews. Applied linguistics, and convincingly that various aspects of second language
sociolinguistics in particular, undertakes the study of learning are affected by the interpretive principles and
the interconnections between language and ▶ sociocul- paradigms in learners’ natal cultures. Specifically,
tural norms and societal frameworks. Specifically, the language learners’ understanding of conceptualizations
subdisciplines of sociolinguistics and pragmatics have and constructs in second culture is crucially affected by
the goal of analyzing how members of particular cul- their culturally defined assumptions, presuppositions,
tures use language to refer to, describe, or function beliefs, and worldviews. For example, for learners
within social organizations. For example, politeness is socialized in the cultures with a strong tradition of
considered to be a universal feature of language use deference to elders, more egalitarian terms of address,
in social organizations, but its pragmatic, linguistic, such as the use of a first name, may seem somewhat
social, intentional, and conceptual realizations vary inappropriate at best.
substantially among different languages and cultures. The teaching and learning of sociocultural and
Even speakers of the same language, such as Chinese or linguistic norms implicitly or explicitly pervades the
Spanish, or different dialects, e.g., American, British, teaching of conversational discourse, social interac-
or Indian English, may belong to different cultures or tion, and the spoken and written language typically
884 C Culture in Second Language Learning
employed in a language community. Second language leaves open the question of whether adult learners can
learners inescapably become learners of the second be fully socialized in a second culture. Learners’ aware-
culture because a language cannot be learned without ness of sociocultural norms and frameworks and the
considering the cultural context in which it is used concepts they acquire as a part of their socialization
(Hinkel 1999). into assumptions, beliefs, and behaviors remain pre-
In the current understanding of the place of culture dominantly first culture-bound even in the case of
in second language pedagogy and learning, the work of advanced and proficient second language users. As
Michael Byram has played a prominent role. Byram many researchers have noted, language learners cannot
(1989, p. 1) noted that culture represents a “hidden” simply shed their own cultural identity and fully adopt
curriculum in second language teaching. That is, lan- another because their natal culture is a part of them-
guage teaching can rarely take place without implicitly selves, and their socialization processes have formed
teaching the culture of its speakers because language and created them as social individuals (Byram and
invariably refers to their common and shared knowl- Morgan 1994).
edge and perceptions of the world, as well as the Without an understanding of the manifestations
concepts of culture, and cultural learning. Currently, and outcomes of sociocultural values, norms, and
many researchers and language teaching methodologists concepts on speech and behavior in language use, it
largely assume that, in real terms, communicative com- may not be possible to become fully linguistically
petence involves socially and culturally appropriate lan- competent in another language. Being aware of the
guage use, which is almost invariably culture specific. sociocultural frameworks does not mean, however,
Unlike the foundational language skills, such as that learners have to become “native-like,” but an
speaking, reading, or writing, second culture does not awareness of the second cultural norms can allow
represent a separate domain of language instruction. learners to make their own informed choices of what
Rather, the learning of the second culture makes to say and how to say it. Because language use reflects
learners better – and more competent – communica- the culture of its speakers in a myriad of ways, teach-
tors. In language learning, the foundational sociocul- ing the second culture together with the essential
tural principles that determine the norms of linguistic skills more adequately represents the con-
appropriate language use and behavior within the nections between language and culture than teaching
social networks and paradigms represent the invisible second language linguistic skills – or culture – in
culture (Hinkel 2001). As Stewart (1972, p. 16) com- isolation.
ments, “[t]he typical person has a strong sense of what
the world is really like, so that it is with surprise that he Cross-References
discovers that ‘reality’ is built up out of certain assump- ▶ Cultural Learning
tions commonly shared among members of the same ▶ Language Acquisition and Development
culture. Cultural assumptions may be defined as ▶ Second Language Learning
abstract, organized, and general concepts which pervade ▶ Social Influence and the Emergence of Cultural
a person’s outlook and behavior.” To members of Norms
a particular community and culture, these fundamental ▶ Value Learning
assumptions usually appear to be self-evident and
axiomatic. On the other hand, they are not always shared
References
by members of other language communities and cul-
Byram, M. (1989). Cultural studies in foreign language education.
tures whose values are similarly based on unquestioned Clevedon: Multilingual Matters.
fundamental assumptions and concepts. It is also impor- Byram, M., & Morgan, C. (1994). Teaching-and-learning language-
tant to acknowledge that ways of using language (e.g., and-culture. Clevedon: Multilingual Matters.
speaking, listening, reading, and writing) and sociocul- Hinkel, E. (Ed.). (1999). Culture in second language teaching and
learning. Cambridge: Cambridge University Press.
tural frameworks in different communities may conflict
Hinkel, E. (2001). Building awareness and practical skills for cross-
to varying extents (Hinkel 1999). cultural communication in ESL/EFL. In M. Celce-Murcia (Ed.),
The conceptualization of culture as inextricable Teaching english as a second or foreign language (3rd ed.,
from ethnolinguistic and personal identity, however, pp. 443–458). Boston: Heinle & Heinle.
Culture of Learning C 885
Hymes, D. (1972). Models of interaction of language and social life. facilities they have both at home and at school but also
In J. Gumperz & D. Hymes (Eds.), Directions in sociolinguistics some abstract factors such as the attitudes of teachers,
(pp. 35–71). New York: Holt, Rinehart and Winston.
parents, and students toward learning (Kumpulainen
Stewart, E. (1972). American cultural patterns: A cross-cultural per-
spective. Yarmouth: Intercultural Press. and Renshaw 2007).
C
Theoretical Background
Embedded in the definition of culture is that culture
has its start with mankind and end with mankind.
Culture of Learning Likewise we should acknowledge the fact that learning
is not limited with time and space. Culture of learning
ERDOGAN BADA1, YONCA OZKAN1, BILAL GENC2 thus involves enabling students to grasp this fact by
1
Department of English Language Teaching, heart and it implies a substantial change in the attitudes
Cukurova University, Adana, Turkey of students toward learning. Given the fact that in the
2
Department of English Language Teaching, history of teaching much importance has been put on
Inonu University, Malatya, Turkey the role of parents and teachers, we believe, “culture of
learning” will reveal to us the extent to which working
with peers in team environments and students’ moti-
Synonyms vating each other thus enhancing learning.
Attitudes toward learning; Reinforcing the value of Within the concept of culture of learning, learning
learning is considered as a competitive process “through which
cultural resources are distributed within specific local
Definition groups of learners and more broadly throughout a
It is by no means an exaggeration to state that there is society” (Kumpulainen and Renshaw 2007, p. 111).
no other word whose definition proves to be that prob- Thus learning is not seen simply as a technical matter
lematic and incomplete as the definition of the word of effectiveness and efficiency but it is always a norma-
culture for both scholars and the laymen. According to tive and ethical endeavor. The individuals of a society
The Oxford Companion to English Language the first are either promoted to or denied from value-laden
meaning of the term, tillage of the soil, still is in usage resources that affect the level and kinds of participa-
in the sense of raising plants and animals and in the tion that individuals might achieve in a community
scientific “culturing” of microorganisms and tissues. (Kumpulainen and Renshaw 2007). Learning extends
The second meaning which denotes a sense of training our lives into new dimensions. It is cumulative. Instead
(body, mind, ideas, tastes, or manners) underlies such of diminishing in time, like health and strength, its
phrases as physical culture and a cultured manner. returns go on increasing.
A third sense, as in the twentieth-century Western A deep understanding of culture and learning is
culture, refers to a social condition, level of civilization, important for all educators, though the subject must
or way of life (McArthur 1992). Culture of learning be addressed carefully. The relationship of the values of
then refers to the social conditions under which learn- the culture in which a child is currently living, or from
ing takes place. While in broader terms culture of which a child has roots, and the learning expectations
learning denotes the attitudes of people from different and experiences in the classroom are directly related to
culture toward learning, in a narrower sense it denotes the child’s school academic, social, and emotional suc-
the physical and psychological conditions in a school cess. If a classroom teacher is to facilitate successful
environment efficient in leveraging learning through learning opportunities for all learners, he or she must
reinforcing the value of learning. Likewise, in broader “know” the learner. This includes knowing about the
terms culture is not considered as an entity apart from individual’s personality, learning styles and prefer-
learning, but as a closely related factor with learning ences, as well as the acquired cultural values that affect
and in narrower terms students’ achievement is behavior.
expected due to not only concrete factors such the Students whose families value collaboration are
time and energy students assign to their studies, the known to be independent, and exercise self-control
886 C Culture of Learning
due to the value given to spontaneity. And, students discussion and spontaneous and detailed comments
who are rewarded in their families for being social are are encouraged (Samovar and Porter 2003) while in
known to work quietly and alone when needed. Every an eastern culture, attentive listening and brief com-
child of every culture, race, ethnicity, socioeconomic ments after contemplation are expected. When faced
status, gender, age, ability, and talent deserves to have with a problem, westerners value confrontation which
an equal opportunity to be successful in academic and involves reporting one’s feelings honestly and expecting
social life (Martin and Nakayama 2008). Teachers who reciprocal honesty (Stewart and Bennett 1991). This
understand learning and cultural differences will strive contrasts with an eastern culture where harmony is of
for intentional variety in instruction, curriculum, class- prime value (Nakayama 1994) and where confrontation
room management, and assessment, which in turn should be avoided as much as possible.
reflects on the learner’s learning behavior. Administra- In eastern cultures – particularly Asian – influenced
tors who believe in learning styles and the variety of by Confucianism where students are expected to respect
cultural characteristics actively value differences in and not challenge their teachers, many students hesitate
teaching styles. Curriculum designers who practice to voice obvious objections whereas western students are
a learning styles approach encourage diverse programs less likely to be so inhibited.
in classrooms, schools, and the district. The second important difference between east and
They can increase awareness of individual learning west is competition that is the primary method among
styles and cultural differences through encouraging westerners is to motivate members of groups (Stewart
and supporting appropriate professional development and Bennett 1991) while eastern people in general value
experiences for all levels of school personnel, includ- cooperative attitudes (Nakayama 1994). These charac-
ing their own. For instance, characteristics concerning teristics may prove very constructive for both types of
eastern and western cultures should be taken into learners in a class setting where teachers could on one
account by any curriculum designer, administrator, hand encourage competition leading to excellence, and
and teacher while creating a learning environment. on the other, promote cooperation among individuals
Such characteristics may roughly be summarized in leading to solidarity.
the table below: Efficacy differs between two cultures. While western
culture values pragmatism where the focus is on getting
Dominant themes in an Dominant themes in things done (Stewart and Bennett 1991), eastern cul-
eastern culture a western culture ture generally prioritizes other people’s feelings. Thus,
Relaxation; acceptance; Activity; assertiveness; while a class may evolve more in a process-oriented
contemplativeness; part of confrontation; diligence; setting regarding students belonging to an eastern cul-
nature; silence; meditation; coexistence with nature; ture, in a western society, the focus may be on product
consideration of others’ speech; articulation; self-
feelings; content with less assuredness; attempting to
rather than process. Therefore, teachers should be
material assets; love of life; get more of everything; aware of such culture-specific learning values.
austerity; cherishing success; achievement; Another different aspect is that eastern students
wisdom of years; cherishing vitality of youth; have a culture that is geared much more to academic
retirement to be with retirement to enjoy rewards education than those are in a western culture. As
family; teachers and of work; communicating
a result of such a cultural emphasis on eastern students
textbooks; coordination of and learning; teachers as
group support; social and organizers, mentors, to excel academically, they inevitably spend more time
moral learning; guides; horizontal relations studying and doing homework than their western
hierarchical relations counterparts. This may be due to the fact that western
teachers themselves do not stress the importance of
homework relative to other educational activities. In
Important Scientific Research and other words, western students may not regard it as
Open Questions important as eastern students do whether the home-
Students’ attitudes in a formal learning setting display work itself is depicted as important or not. This may be
differences in terms of learning viewed by these a plausible reason as to why some western students
two cultures. In western culture, active initiation of choose to fly through their homework without trying
Cumulative Learning C 887
to really understand the concepts or why their parents Martin, N. J., & Nakayama, T. K. (2008). Thinking dialectically
allow them to do homework while watching television. about culture and communication. In M. K. Asante, Y. Miike,
& J. Yin (Eds.), The global intercultural communication reader
Students’ concentration level is inherently abated if
(pp. 73–92). London: Routledge.
they perceive the diminished value of doing home- McArthur, T. (Ed.). (1992). The Oxford companion to the English
work, not to mention the intrinsic value, to be rather language. Oxford: Oxford University Press. C
poor at best. Nakayama, T. K. (1994). Show/down time: “Race”, gender, sexuality,
We make meaning based on our experiences and on and popular culture. Critical Studies in Mass Communication, 11,
162–179.
the information and ideas we encounter. Based on
Samovar, L. A., & Porter, R. E. (2003). Communication between
these experiences, we interact with other people. In cultures. Belmont: Wadsworth.
order to enable our interaction to be effective, we first Stewart, E. C., & Bennett, M. J. (1991). American cultural patterns:
should recognize other people’s differing viewpoints A cross cultural perspective. Yarmouth: Intercultural Press.
and interpretations. In that sense, teachers or curricu-
lum designers should be aware of the fact that their
interpretation may differ depending on students cul-
tural backgrounds. Recognizing each other’s view-
points and interpretations will prevent us from Culture-Bearer
imposing our own views on others and contribute to
A resource person who is a member of the culture being
reforming each other’s social identities.
represented or studied.
Learning should aim not only at helping students
learn the content of the class itself but also under-
stand its cultural background. Therefore, it is impor-
tant to learn with each other what the expectations
in another culture are in particular situations. Culture-Within-a-Culture
One culture may give priority to verbalizing what
people think spontaneously and clearly while another ▶ Microculture of Learning Environments
may prioritize considering others’ feelings before
expressing whatever occurs in the mind. By under-
standing each other’s different viewpoints, we will
gradually be able to acquire shared meanings indis-
pensable for smooth communication. It is up to the Cumulative Learning
teacher to create opportunities so that effective educa-
tion can be realized by motivating students, most JUNGMI LEE
of which depending on our understanding and appre- Faculty of Economics and Behavioral Sciences,
ciation of our students’ different characteristics, needs, Department of Educational Science, University of
styles, preferences, beliefs, and attitudes. It is only Freiburg, Freiburg, Germany
then we can ensure the maintenance of such motiva-
tion leading to the emergence of a healthy learning
environment. Synonyms
Incremental learning; Layered learning
Cross-References
▶ Cross-Cultural Factors in Learning and Motivation Definition
▶ Cross-Cultural Learning Styles Intelligent systems, human or artificial, accumulate
▶ Learning-Related Motives and the Motivational knowledge and abilities that serve as building blocks
Quality of the Learning Environment for subsequent cognitive development. Cumulative
learning (CL) deals with the gradual development of
References knowledge and skills that improve over time. In both
Kumpulainen, K., & Renshaw, P. (2007). Cultures of learning. Inter- educational psychology and artificial intelligence, such
national Journal of Educational Research, 46(3–4), 109–115. layered or sequential learning is considered to be an
888 C Cumulative Learning
prior knowledge through extensive practice. Further- learns the process of classification, differentiation,
more, these highly automated pieces of knowledge and recall, retention, and transfer of learning due to its
skills are continuously tuned over time through con- continuing cumulative effect. Thus, Gagné assumed
siderable amounts of “deliberate practice” (Ericsson a cumulative organization of learning events based on
et al. 1993, p. 363), helping the learner to gain various pre-existing relationships among learned behaviors. C
levels of expertise. Consequently, learning situations Hence, he posited that instruction should provide
should focus on the important synthetic structural a set of component tasks and a sequence of those
links and association to ensure the gradual develop- tasks to ensure the learners’ mastery of each component
ment of knowledge and skills and hence optimize task and the optimal transfer of the final task. He tried
performance. to empirically demonstrate the effect of CL using
Shuell (1986) states that cognitive psychology has Piaget’s classical conservation task. Though he was
influenced learning theory and research in several ways, criticized for his incorrect analysis of the conservation
including (a) the view of learning as an active, con- task example because he used an ambiguous combina-
structive process as formulated by Ausubel (1960), who tion of “nonmetric judgment of volume” and “conser-
believes a learner’s present cognitive structure consti- vation of identity” in his analysis whereas his specific
tutes the principal factor influencing whether or not he task example presents only the latter, this can hardly
or she will be able to acquire and retain particular degrade the implication of this example in terms of
information; and (b) the cumulative nature of learning demonstrating cognitive development process (see
and the corresponding role played by prior knowledge. Furby 1972).
Shuell, in particular, stated that learning is cumulative Many authors have emphasized the cumulative
in nature and that everything has meaning or is learned nature of learning. Bruner et al. (1956) defined concept
in connection. Furthermore, cognitive conceptions of learning, also known as category learning and concept
learning have considerable importance through prior attainment, as “the search for and testing of attributes
knowledge. This argumentation corresponds to a great that can be used to distinguish exemplars from non-
extent with Gagné’s idea of cumulative learning. exemplars of various categories” (in the reprinted ver-
sion: 1986, p. 233). Consequently, concept learning
Cumulative Learning in Educational requires a learner to compare and contrast categories
Psychology based on the relevance of their features. Bruner
Gagné (1968) coined the term “cumulative learning (1960) viewed learning as an active process in which
theory” on the basis of the assumption that intellectual learners construct new concepts based upon their pre-
skills can be broken down into simpler skills. He sent and prior knowledge by means of selecting and
believed that “learning is cumulative” (Gagné 1965, transforming information, constructing hypotheses,
1968, 1970) and that human intellectual development and making decisions based on a cognitive structure
consists in building up increasingly complex structures, (i.e., schemas and mental models). He hypothesized
the learning of higher-level skills such as rules and that “any subject can be taught effectively in some
principles depending primarily upon the prior mastery intellectually honest form to any child at any stage of
of subordinate skills or concepts. He also viewed the development” (1960, p. 33). Accordingly, he empha-
process of learning as an accumulation of increasingly sized the importance of learning the underlying prin-
complex interacting structures of learned capabilities. ciples of different concepts of children’s learning in
These structures are capable of interacting with each a way that allows the subsequent transfer to be made
other in patterns of great complexity, thereby cumula- and consequently expands their knowledge. Bruner
tively generating an ever-increasing competency level. also noted that arranging information in a spiral
Therefore, any learned capability at any stage of the fashion helps children to organize knowledge into
learning sequence can operate to mediate learning that a structure that makes it increasingly usable in other
was not otherwise deliberately taught. Consequentially, areas beyond the current learning situation. Thus, he
generalization or transfer of knowledge to new tasks is emphasizes that the learning situation should help
an effect of CL. Learning occurs not only in the acqui- learners to actively reorganize the new information,
sition of new associations but also because the learner allowing them to build on existing knowledge in
890 C Cumulative Learning
a meaningful way and use the newly gained knowledge Michalski (1994) views learning as “a goal-guided pro-
effectively in future tasks. In other words, in the course cess of modifying the learner’s knowledge by exploring
of learning situations, repeatedly presented informa- the learner’s experience” (p. 3) in his Inferential Theory
tion is organized from the simple to the more complex, of Learning (ITL). In ITL, the learning process consists
from the general to the specific, and is examined of the input facts, the background knowledge, and the
in association with other information. Accordingly, types of inferences (i.e., induction, deduction, and
Bruner believed that as children grow, the curriculum analogy) a learner makes to generate new knowledge.
should repeatedly represent previously learned infor- Thus, the changes in the knowledge content, its orga-
mation and expand on it until the child understands nization, and its certainty are all seen as bringing about
the information and its relations more completely. a total change in the learner’s knowledge in the course
Ausubel argued that human cognitive structure is of learning. Successfully learned knowledge is assimi-
organized hierarchically from greater to lesser inclu- lated into the learner’s background knowledge and can
siveness (Ausubel and Robinson 1969). New informa- be used in subsequent learning processes.
tion is subsumed into an existing cognitive structure. Zhou (1990) introduced the CSM (classifier system
In other words, new incoming information can be with memory), an extension of the classifier system
connected to relevant existing cognitive structure model that includes mechanisms for analogical and
through the process of subsumption (Ausubel 1963). cumulative learning, and tested it in the domains of
Learners activate their pre-existing knowledge so that robot navigation and letter extrapolation. The CSM
it can be assimilated, tuned, and restructured into new was designed in response to the problems of conven-
schemas by meaningful learning (Ausubel 1968). tional expert systems, namely, the fact that they do not
Hence, meaningful learning can help learners to acti- acquire automated knowledge, update it to any sub-
vate pre-existing schemas more effectively and allow stantial extent (i.e., adding and removing knowledge),
them to use their old knowledge in the new learning or function intelligently beyond their current knowl-
situation more effectively. edge. It can preserve problem-solving expertise, recall
Aebli (1973) defined concepts as the “basic blocks similar solutions by searching its long-term memory,
of any discipline,” and once these basic concepts are construct solutions to similar new situations using
mastered or acquired, the student is ready to take his analogy (i.e., recognizing the similarities between
knowledge to a higher level. These accumulated con- two problems), and adapt them to fit new situations.
cepts lead to the higher level of knowledge. According Rules created by information exchange are stored in
to Aebli, knowledge is acquired as a learner actively a temporary knowledge base. When a set of detectors
constructs and transforms it by integrating newly relays external information to the system, eligible rules
gained information into an earlier knowledge structure may be triggered, which in turn generate new messages
through a process of modification and reinterpretation and then perform actions. The system’s behavior can
of existing knowledge in light of the newly gained be changed through the deletion, modification, and
information. Moreover, such integration promotes creation of rules (i.e., tuning). While the short-term
the activation of other closely related areas of knowl- memory (STM) stores previously accumulated active
edge in that it cumulatively disposes the learner to knowledge, valuable inactive information is also
learn more. maintained separately in the long-term memory
(LTM), thus preventing rapid forgetfulness over time
Cumulative Learning in Artificial and preserving the information for future use. In other
Intelligence words, when the system has accumulated sufficient
In the field of machine learning, information is com- knowledge, it categorizes and generalizes a set of suc-
pared and put into a framework to be used for up- cessful task-independent rules by extracting the com-
coming processes or problem-solving tasks. Pfeffer mon features from a set of relevant rules and then
(2000) defines a cumulative learning agent as one that transfers them from STM to LTM, where it stores
learns and reasons as it interacts with the world by them as chunked building blocks, organized and
using its accumulated knowledge and its observations. indexed hierarchically (from specific to general) for
Cumulative Learning C 891
future problem-solving situations. The CSM shows In the field of artificial intelligence (AI), the agent
that “with the benefit of the prior experience and perceives and models or formulates concepts from its
accumulation of problem solving expertise, it con- environment and generates an appropriate decision by
structs its knowledge base incrementally through aggregating different pieces of them in order to con-
interaction with its environment and improves its struct a solution. Swarup et al. (2006) outlined an C
problem-solving ability over time” (p. 404). This is extracting ontology from the process of cumulative
consistent with the principle of CL. learning to solve future related problems. Each agent
Regarding the cognitive processes which are inher- aggregates and accumulates a packet of knowledge that
ent to CL, a distinction can be made between (a) the is extracted from solutions to multiple tasks, and these
aggregation of knowledge, (b) the abstraction of knowl- packets impact new learning tasks through analogy (i.e.,
edge, and (c) a combination of aggregation and abstrac- the agent recognizes and applies similarities between the
tion of knowledge. While aggregation of knowledge tasks). Thus, the agents can guide each other’s learning
means that multiple units of packets of knowledge are process by grounding symbols of the aggregated cumu-
at the same level of knowledge structure, abstraction of lative knowledge, thereby improving learning perfor-
knowledge means that they are at higher level on the mance. As the agent aggregates experiences and builds
hierarchy of knowledge structure. up its cumulative knowledge to find solutions to new
problems, it is expected that more cumulative knowl-
Aggregation edge will be found (see “▶ Reinforcement Learning”).
In the process of aggregation, a learner extracts and Murphy and Medin (1985) argued that a concept is
identifies information and knowledge and then puts hierarchically related with other concepts as well as
that knowledge into a coherent knowledge structure. structurally related, thus enabling inferences. Accord-
Bruner (1960) terms this process categorization. A ingly, to ensure an effective and efficient CL, each
learner categorizes knowledge on the basis of its sur- aggregated concept should be constructively related
face and structural similarities. The process can also be in the complex knowledge network in order for transfer
called “structural mapping” (Gentner 1983), which to be successful in a future learning situation. The
means that a learner maps the knowledge from model packet of collected knowledge should be stored as
A to model B by analogy. Structural mapping presup- a composition; otherwise the aggregates would waste
poses a mental model in that a learner only compares limited memory space.
his or her mental models of the two structures but Easterlin (1986) divided the process of concept
never compares their facts themselves. If the learner formation in machine learning into the three following
cannot compare the mental models, then he or she components:
induces a schema by analogy. In this process, the
learner might not be able to define the shared features ● Aggregation, in which important instances of expe-
of the two structures but will definitely be able to riences are grouped into a set of aggregates. Here,
aggregate pieces of information or knowledge. To experiences are aggregated by the learning system
categorize the knowledge, a learner needs to define itself for further use based on their contribution
perceptual and surface-structural as well as deep- to a successful problem solution and to system
structural similarities in a long line of aggregation. performance.
For example, in the field of geophysics, the known ● Characterization, in which a description of the
structure (A) can be mapped to the new structure (B) essential information for an aggregate of experiences
through the analogical reasoning of structural map- is generated or constructed in terms of characteristics
ping. Most people rely only on the surface structure that are useful to the system based on individual
of the information rather than its deep structure and descriptions of each member of the aggregate.
thus fail to reach a correct analogical conclusion. Accre- ● Utilization, in which the concept description is
tion and tuning might also be involved in this process, integrated with the performance element of the
but a learner does not reorganize or restructure the system and the important aspects of the aggregate
knowledge at this stage. are captured.
892 C Cumulative Learning
This process leads to the formation of concepts with the concrete aspects of the unknown domain”
containing functional information. Then, the agent (Dörner 1982, p. 140). If one accepts Dörner’s further
applies its operators to the subsequent task by mapping argument that this process of abstraction presupposes
objects and operators of current states to consequent the existence of a comprehensive system of abstract
states, thus producing a generalized schema with its concepts which allow transitions from one concrete
background knowledge. Again, aggregated experiences, aspect to another, the schema-theoretical argumenta-
instances, and entities should be characterized and uti- tion is complete, for which Dörner (1976, pp. 82ff) has
lized structurally (as stated above) to meet the essential defined the following steps:
goal of CL.
(a) Abstraction of certain attributes of the phenome-
How can structural aggregation be accomplished?
non in question (esp. as regards content);
In Pfeffer’s (2000) Integrated Bayesian Agent Language
(b) The search for a model, i.e., for a second phenom-
(IBAL), a learning agent can modify its models based
enon, which constitutes another concretization of
on its collected observations and use them in future
the abstract phenomenon;
situations. He stresses that “a representation language
(c) Transfer of (structural) attributes of the model
must be modular and extensible” (p. 52) so that the
back to the original phenomenon; and
knowledge base can be structurally extended and accu-
(d) A test as to whether the hypothesized attributes are
mulated. Swarup et al. (2006) suggest starting with
actually present in the phenomenon.
small problems with very few easy-to-find solutions
to find networks which have subgraphs that can be In an extension of this argument, Seel (1991)
reused to solve other problems. It seems that chunking describes the function of mental models in analogical
the useful aggregated knowledge into a module which reasoning: A person makes propositions or predictions
can be extended, modified, and dynamically updated is for a certain phenomenon (target domain) by falling
an appropriate means of ensuring the rich process of back on his or her knowledge about similar phenomena
aggregation for successful CL. (base domain) and creating a mental model for both.
On the basis of the structural similarities this person
Abstraction finds between the models of the base and target
In the process of abstraction, the learner extracts domains, he or she reaches a conclusion by analogy,
commonalities from superficial features as well as the integrates both models into a unified solution model
underlying structure of knowledge and then defines under the assumption that they are similar, and
a super-concept of knowledge, identifies underlying tests whether it is possible to create an alternative
relationships, modifies/creates mental model(s), and solution model.
constructs a new schema by analogy. This new schema
can then be applied to the different domains. For Important Scientific Research and
example, Edward Hargraves found gold in Australia Open Questions
in 1851 by comparing and finding similarities between This cumulative effect of learning that results from
the geological characteristics of California and Austra- classification, discrimination, retention, and transfer
lia. This abstraction process, however, presupposes over a period of time has to be explored extensively in
a substantial (re)organization of learned knowledge. human learning. In the field of human learning, the
Dörner (1982) has described this aspect of cumulative term CL has not been used explicitly since Gagné
learning as the “condensation” of an abstract schema (1965). In the fields of machine learning and robotics,
by means of a conclusion by analogy. This process on the other hand, CL has been used and studied
involves a search for a known reality domain (as the explicitly in more recent times, and the idea plays an
base of an analogy) which is in certain respects analo- important role. In these fields, a close relationship
gous to the unknown target domain. It is in essence between CL and schema induction can be found (see
a process in which “the concrete aspects of the known Pfeffer 2000). Assuming that CL is the essential concept
stock of knowledge [. . .] are in a sense ‘boiled away’ or mechanism in the process of meaningful transferable
and the remaining ‘pure’ structure [. . .] is filled learning, bridging the gap between the two fields and
Cumulative Learning C 893
learning from the experiences of both seems to be Überblick in Einzeldarstellungen (pp. 134–148). München:
a promising endeavor. The achievement of learning in Urban & Schwarzenberg. [Learning of knowledge and compe-
tence acquisition].
machines might help us understand how humans learn.
Easterlin, J. D. (1986). Functional properties and concept formation.
Furthermore, CL and its practical implications for In T. M. Mitchel (Ed.), Machine learning: A guide to current
human learning and machine learning should be research (pp. 59–66). Norwell: Academic. C
empirically proven with detailed longitudinal learning Ericsson, K. A., Krampe, R. T., & Tesch-Roemer, C. (1993). The role of
research and cross-sectional studies with general learn- deliberate practice in the acquisition of expert performance.
Psychological Review, 100, 363–406.
ing sequences of large data rather than with peculiar
Furby, L. (1972). Cumulative learning and cognitive development.
small parts of learning sequences. In that way, it would Human Development, 15, 265–286.
be possible to explore how learning systems can acquire Gagné, R. (1965). The conditions of learning and theory of instruction
deeply cumulative knowledge in a long sequence of (1st ed.). New York: Holt, Rinehart, & Winston.
learning. Gagné, R. (1968). Learning hierarchies. Educational Psychologist,
6(1), 1–6.
Cross-References Gagné, R., & Gropper, G. (1965). Individual differences in learning
▶ Abstraction in Mathematics Learning from visual and verbal presentation. Pittsburgh: American Insti-
tutes for Research.
▶ Assimilation Theory of Learning Gagné, R. M. (1970). The conditions of learning (2nd ed.). New York:
▶ Comprehensive Learning Holt, Rinehart & Winston.
▶ Concept Learning of Humans Gentner, D. (1983). Structure-mapping: A theoretical framework for
▶ Concept Mapping analogy. Cognitive Science, 7, 155–170.
▶ Inferential Learning and Reasoning Langley, P. (1995). Elements of machine learning. Palo Alto: Morgan
Kaufmann.
▶ Inferential Theory of Learning
Michalski, R. S. (1994). Inferential learning theory: Developing foun-
▶ Machine Learning dations for multistrategy learning. In R. S. Michalski & G. Tecuci
▶ Meaningful Verbal Learning (Eds.), Machine learning – A multistrategy approach (pp. 3–61).
▶ Reinforcement Learning San Mateo: Morgan Kaufmann.
▶ Schema-Based Learning Murphy, G. L., & Medin, D. L. (1985). The role of theories in
▶ Schema-Based Reasoning conceptual coherence. Psychological Review, 92, 289–316.
Pfeffer, A. (2000). A bayesian language for cumulative learning. In
▶ Schema Development
L. Getoor & D. Jensen (Eds.), Working notes of the AAAI-2000
▶ Schema(s) workshop on learning statistical models from relational data
(SRL-2000), Technical report WS-00–06 (pp. 57–62). Austin:
References AAAI Press.
Aebli, H. (1973). Didactica psihologică. Bucuresti: Editura Didactică si Roske, R. J. (1963). The world impact of the California gold rush
Pedagogică. 1849–1857. Arizona and the West, 5(3), 187–232.
Ausubel, D. P. (1960). The use of advance organizers in the learning Rumelhart, D. E., & Norman, D. A. (1978). Accretion, tuning, and
and retention of meaningful verbal materials. Journal of Educa- restructuring: Three models of learning. In J. U. Cotton &
tional Psychology, 51(5), 267–272. R. L. Klatzky (Eds.), Semantic facts in cognition (pp. 37–54).
Ausubel, D. P. (1963). The psychology of meaningful verbal learning. Hillsdale: Erlbaum.
New York: Grune and Stratton. Seel, N. M. (1991). Weltwissen und mentale Modelle. Göttingen:
Ausubel, D. P. (1968). Educational psychology: A cognitive view. New Hogrefe. [World knowledge and mental models].
York: Holt, Rinehart, & Winston. Shuell, T. (1986). Cognitive conceptions of learning. Review of Edu-
Ausubel, D. P., & Robinson, F. G. (1969). School learning: An intro- cational Research, 56, 411–436.
duction to educational psychology. New York: Holt, Rinehart, & Swarup, S., Lakkaraju, K., Ray, S. R., & Gasser, L. (2006). Symbol
Winston. grounding through cumulative learning. In P. Vogt et al. (Eds.),
Bruner, J. S. (1960). The process of education. Cambridge: Harvard Symbol grounding and beyond. Proceedings of the third interna-
University Press. tional workshop on the emergence and evolution of linguistic
Bruner, J. S., Goodnow, J. S., & Austin, G. A. (1956). A study of communication (pp. 180–191). Berlin Heidelberg: Springer-
thinking. New York: Wiley. Verlag.
Dörner, D. (1976). Problemlösen als Informationsverarbeitung. Stutt- Wittrock, M. C. (1991). Generative teaching of comprehension. The
gart: Kohlhammer. [Problem solving as information processing]. Elementary School Journal, 92(2), 169–184.
Dörner, D. (1982). Lernen des Wissens- und Kompetenzerwerbs. In Zhou, H. H. (1990). CSM. A computational model of cumulative
B. Treiber & F. E. Weinert (Eds.), Lehr-Lern-Forschung. Ein learning. Machine Learning, 5(4), 383–406.
894 C Curiosity
Theoretical Background
Cross-References Curiosity and exploration have both been linked to
▶ Interest-Based Child Participation in Everyday
animals and humans (Berlyne 1966). Much research
Learning Activities
has demonstrated that primates, raccoons, rats,
▶ Play, Exploration, and Learning
birds, dogs, and many other animals seek new stimuli,
explore their environment or behave so as to maximize
knowledge. As for humans, researchers have associated
curiosity and exploration also with cognitive, social,
emotional, and spiritual development across the
Curiosity About People
lifespan.
▶ Interpersonal Curiosity Attachment theorists, for example, acknowledge
the link between attachment, curiosity, exploration,
and learning. Secure attachment style patterns where
an individual feels a positive emotional bond with
a parent, teacher, intimate friend, or spouse support
Curiosity and Exploration safely being curious and exploration of the environ-
ment and subsequent learning. Those with anxious-
THOMAS G. REIO, JR. ambivalent and avoidant attachment styles (insecure
Department of Leadership and Professional Studies, attachment) are far less likely to be curious and explor-
Florida International University, Miami, FL, USA atory, particularly when presented with stressful situa-
tions, as they are less likely to interact well, explore their
environment, and demonstrate emotional resiliency.
Synonyms Therefore, the exploratory and attachment systems
Cognitive curiosity; Diversive exploration; Intrinsic work in tandem to promote socioemotional function-
motivation; Sensory curiosity; Specific exploration ing, learning, and development (Voss and Keller 1983).
Psychosocial theorists (e.g., Erikson 1968) and
Definition researchers associate curiosity and exploration with
Curiosity is a state of increased arousal response pro- identity formation and learning. Identity versus role
moted by a stimulus high in uncertainty and lacking in confusion normatively occurs during adolescence (i.e.,
information. When compared to existing knowledge, it is the central task of adolescence), but is not confined
the novel, uncertain, conflicting, or complex properties to that period. For adolescents, the confluence of indi-
of external stimuli create a conceptual conflict that vidual internal forces of development (biological),
arouses the internal state of arousal called curiosity one’s unique interests, feelings, and needs, and societal
(Berlyne 1966). Once curiosity has been aroused, the demands to find suitable ways to enter adult roles
organism engages in a process of exploration to reduce fosters identity formation. For optimal identity devel-
the state of arousal. There are two basic types of curi- opment, however, adolescents must be curious about
osity: cognitive and sensory. Cognitive curiosity is the and explore social roles in the greater community to
Curiosity and Exploration C 895
learn what activities might provide the best fit for one’s appropriately and embracing both as means to develop
biological and psychological aptitudes and interests. the joy of learning in their children. On children’s
Thus, adolescents should be willing to be proactively playgrounds, curiosity and exploration are closely
curious and explore the possibilities afforded by their linked to both free, imaginative play and games with
environments to learn how to locate themselves suc- rules. Through being curious and exploratory during C
cessfully in the larger social order. free play, children learn how to productively solve
Cognitive theorists and researchers, too, embrace a diverse array of problems related to activities such
curiosity and exploration in their writings. Piaget, for as building a fort, settling an argument without adult
instance, proposed that being proactively curious supervision, and soothing an imaginary crying baby
is a prerequisite for the construction of knowledge when playing house. In games with rules, children
because it both stimulates the acquisition of new infor- need to be proactively curious and exploratory to
mation and the seeking of new stimuli. In essence, he is learn how to play the game in the first place and to
referring to both the cognitive and sensory types of find ways to get peers to invite them to join the game.
curiosity. The information and experiences secured All of these behaviors are conducted in the context of
through interaction of the cognitive and sensory types being relatively stress-free where the children learn and
of curiosity and exploration then promote cognitive develop best cognitively, socially, and emotionally.
development through the construction of new knowl-
edge (Reio et al. 2006). Piagetian theory also highlights Important Scientific Research and
the role of disequilibration in cognitive development. Open Questions
When an individual is faced with discrepant informa- Significant theoretical and empirical advances have
tion, a state of disequilibration is aroused (curiosity) been made in working out how curiosity and explora-
that motivates exploration of the environment for the tion influence learning and healthy human function-
sake of regaining a state of equilibration. Through this ing throughout the lifespan. Open questions remain
process, new information is acquired, creating condi- however about the brain systems that support each.
tions for optimal learning and cognitive development. A promising “SEEKING,” neuroemotional system that
To be sure, there is a dark side to being curious and encompasses the basic need to seek, investigate, and
exploratory. Cognitive curiosity can be associated with understand the environment has been receiving
asking too many questions, being meddlesome, and increased scholarly attention (Panksepp 1999, p. 145).
nosy (forms of specific exploration). Sensory curiosity, The SEEKING system runs through the medial fore-
on the other hand, motivated by the desire for new brain bundle of the lateral hypothalamus where dopa-
sensations and thrills despite the risk, is linked to mine in particular seems to be an essential ingredient in
diversive exploration such as drug experimentation, allowing brain circuitry to operate efficiently. This
engaging in unprotected sex, and driving while intox- appetitive motivational system drives and energizes
icated, to name a few maladaptive behaviors. Being the persistent feelings of curiosity, interest, sensation
appropriately curious and exploratory is regulated seeking, the search for higher meaning, and learning
somewhat by societal and group norms, but when the beyond the simple promise of rewards. More research is
associated risks outweigh possible gains, it can have needed to better understand the mechanisms that
profound negative individual, group, and societal enable curiosity and exploration through this impor-
implications. tant motivational system.
Curiosity and exploration are vital concerns in Pressures continue to mount for increased aca-
educational contexts (Flum and Kaplan 2006; Reio demic performance at our schools. High-stakes testing
et al. 2006). Educators who can arouse their learners’ has evolved as a means for assessing, evaluating, and
curiosity will improve learner attention, promote improving the success of school-related activities
greater breadth and depth of exploration of informa- designed to improve student learning and perfor-
tion related to solving a problem, increase time on task, mance. While arguably laudable, little research has
and boost the likelihood that the learners will want addressed the potential fallout from such activities
to learn more for the sake of learning. Parents also on student motivation, especially intrinsic motivation
foster curiosity and exploration through modeling it like curiosity, exploration, interest, and the like.
896 C Curiosity and Learning
Theoretical Background
Curiosity and Learning
Introduction
▶ Surprise and Anticipation in Learning Learning, in the definition of this encyclopedia, takes
places in numerous settings. Many of those settings
are informal and others are formal. Formal education
settings are, for example, primary and secondary
Curious Learning schooling, tertiary education, but also some kind of
professional development and off-the-job training
▶ Anticipatory Learning activities. In formal settings, the provision of learning
is usually steered by a plan.
In the educational system of formal school educa-
tion, a curriculum operates at various levels at which
one or more different plans for learning and other
Current Concerns curriculum products are in use. At each level, different
▶ Goals and Goalsetting: Prevention and Treatment of curriculum and organizational processes take place.
Depression A useful distinction of the different operational levels
and respective plans for learning and curriculum
products are:
● The nano level of the individual student where the
Curriculum processes of individual learning occur and that are
steered and supported by personal learning plans
▶ Alignment of Learning, Teaching, and Assessment and individualized learning courses
Curriculum and Learning C 897
● The micro level of the classroom providing the Curriculum and Learning. Table 1 Typology of curricu-
environment and activities for learning, where lum representations
teaching plans, textbooks, and other teaching mate- Intended Ideal curriculum: referring to the
rials are used and applied curriculum original vision underlying
● The meso level of schools and educational institu- a curriculum, to its basic philosophy C
tions, where educational and school programs are Formal curriculum: refering to
developed and learning environments and learning curriculum documents and to
activities are organized and administered curriculum materials such as
● The macro level of educational systems and national textbooks, teacher guides, and
student materials, also referred to as
policies, responsible for the development of core written curriculum
objectives, attainment levels, and examination pro-
Implemented Perceived curriculum: referring to the
grams of national curricula
curriculum curriculum as interpreted by its
● The supra level of international policies. The supra various users
level influences national curricula and curriculum
Enacted curriculum: referring to the
reform initiatives through multinational compara- actual instructional processes in
tive studies, for example, PISA and TIMMS, and via classrooms; also referred to as
international frameworks, for example, the Euro- operational curriculum, or curriculum
pean Reference for Languages. In the context of in action
curriculum development in developing countries, Attained Experiental curriculum: referring to
the policies and practices of multinational donors, curriculum the actual learning experiences of
donor countries, and aid agencies often have impli- students
cations on all operational levels of the education Learned curriculum: referring to the
system within the recipient country. resulting learning outcomes of
students
Curriculum Representations
Further understanding of a curriculum is given by the
typology of three so-called curriculum representations ● The technical-professional perspective, concerning
that can be detailed into six forms (van den Akker 2003, the methods and practices of curriculum develop-
adapting a distinction of Goodlad 1979). The curricu- ment to successfully translate the intentions in
lum representations emphasize the different layers of curriculum products to be used for teaching and
the curriculum (Table 1). learning
The so-called hidden curriculum is another termi- ● The sociopolitical perspective, referring to the
nology in use. The terminology does not originate from influences in the decision-making processes and
the previously introduced concept of curriculum rep- the different values and interest of different
resentations. The hidden curriculum refers to the stakeholders
underlying assumptions and beliefs that may not be
The distinctions are useful for analysis. In practice,
formally articulated or clearly expressed but that in
all three perspectives come together and are of
actual reality are of influence for educational practice.
importance.
In addition to the three perspectives of Goodlad,
Curriculum Perspectives
the critical perspective has established itself in recent
Of the many ways to look at a curriculum, the three
decades as an independent variety of curriculum
perspectives of Goodlad (1994) depict classical angles
theorizing (Walker 2003). Curriculum criticism is
on curricular issues:
concerned with issues of domination, exploitation,
● The substantive perspective, focussing on the arche- resistance, and what legitimates the knowledge to be
typical question about what knowledge is of most taught in the curriculum (Marsh and Willis 1999).
worth for inclusion in teaching and learning Although curriculum critical proponents are not
898 C Curriculum and Learning
a homogeneous group, Walker (2003) writes that two stakeholders such as pupils, parents, teachers, aca-
characteristics unite most of the disparate work. The demics, administrators, economy, labor unions, reli-
first is the commitment to view every curriculum as gious groups, social organizations, and policy makers.
a social construction, and second, the determination to Each person has its own educational experience
analyze the construction critically to find out how it influencing perceptions how education should be
contributes to the evils of our days. shaped; each group has its own experiences, concerns
about qualifications and subject-matter insights, and
The Core Curriculum Question of its own consideration and preferences to pedagogical
What to Learn views, political issues, or how society should look like
Spanning across all four of the previous perspectives is (van den Akker 2003).
the core question underlying all curricula and of all Given the complexity of curriculum influences, the
curriculum making: “what do we want students to selection and out-selection of curriculum content
learn, and why”? Or, more elaborated: what are the remains an inherently difficult process. Curriculum
desirable aims and content of education to equip stu- literature mentions three main criteria for selection
dents for their role in today’s and tomorrow’s society and prioritization of aims and content. The three
and, intrinsically, how should students learn to acquire criteria are sometimes referred to as the “three S”:
the competencies that are identified as necessary? Three
● Subject, or knowledge: the academic and cultural
main factors influence the answers to these questions.
heritage essential for learning and future
Firstly, new subject specific findings call for being
development
added in the curriculum. In recent decades, this refers
● Student, or personal development: elements of vital
especially to the developments in the sciences and in
importance from the personal and educational
technology. Secondly, to respond to changes in society
needs of students
and emerging challenges, all kinds of societal groups
● Society, or social preparation: issues relevant for
articulate their demands and expectations about what
inclusion from the perspective of societal trends
students have to learn, pressing for new content to the
and needs.
included in the overall curriculum. Thirdly, over the
past two decades, new understanding about learning
has been gained, for example, from developmental
Quality Criteria for the Learning
psychology as initiated by the work of Piaget and
Potential of Selected Curriculum
Vygotsky, from cognitive sciences, and also because
Content
The learning potential within a curriculum depends on
the overall amount of available knowledge has signifi-
its vertical and horizontal consistency. Vertical curric-
cantly increased and continues to increase. They have
ulum consistency, for example through longitudinal
contributed to the scientific development of theories of
learning trajectories, refers to the buildup and sequenc-
learning regarding the aims of learning and even more
ing of learning activities (subject matter, tasks) to link
about the how of learning; the attention to diversity in
prior learning of students to desired learning and to
learning styles; and the changes in student and learning
future learning. Depending on content or learning
environment outside and inside school.
tasks, the sequencing may be linear, stepped, or spiral.
As a result of developments, the question about the
Horizontal curriculum consistency refers to the
what and how of learning continues to occur for cur-
coherence between related subject content and methods
riculum making. Taking up new findings demands for
at one educational level. A horizontally consistent cur-
changes in curriculum content and for inclusion of new
riculum allows for synergies in teaching and learning
content. Yet, a curriculum is by its very nature finite.
activities to supplement each other across subjects and
The decision about subject content and learning
subject combinations.
methods to be included, intrinsically also meaning
eliminating other to provide time and space for
new ones, is a highly contested area, which is often Curriculum Development
likened as a battlefield. There are many interests vested Curriculum development is a process focused on the
in a curriculum and they are addressed by many improvement and innovation of education. Given the
Curriculum and Learning C 899
definition of this chapter, it relates to curriculum prod- findings that successful innovations occurred when
ucts of any operational levels (see Introduction). planned curricula were not highly specified in advance
Historically, the creation of curriculum develop- but were mutually adapted by users within specific
ment and the early decades of curriculum development institutional settings. Michael Fullan and Alan Pomfret
as a field of study were strongly influenced by educators published in 1977 another study. They reviewed 15 C
and scientists in the USA. The three most well-known major studies in Canada, the United States, and the
classical approaches and stepping stones to contem- United Kingdom and found widespread variation in
porary curriculum development models are Tyler’s whether or not innovations were being put into prac-
rational-linear approach, 1949; Walker’s naturalistic tice. Fullan’s and Pomfret’s use of the two terms fidelity
model, 1971; and Eisner’s artistic approach, 1979. The perspective and process perspective was among the
so-called pragmatic approach has more recently come study’s greatest influence on researchers. The two stud-
into existence and can be seen as a fourth curriculum ies of Berman and McLaughlin and Fullan and Pomfret
development approach. Overarching all approaches have since become highly influential and are widely
are five basic development activities: analysis, design, cited in professional literature on curriculum imple-
development, implementation, and evaluation, often mentation. They have been followed by numerous
abbreviated as ADDIE. In the rational-linear approach, studies and research activities of which many con-
these activities are seen as a linear sequence; in the cluded that in educational reality, the intended learning
pragmatic approach the activities are cyclical. of curriculum innovations has rarely taken place. Given
Each of the four curriculum development those many research findings, the huge gaps between
approaches comprises valuable elements and insights the intentions of a curriculum and what is really
for curriculum development. None of the approaches learned has become an often-repeated global common
should be seen as the ultimate model for curriculum theme for which overarching the following major rea-
construction. None claims to be. This is even more sons are given:
as the tasks and scope of curriculum development at
operational levels differ considerably and the number ● Time frames for implementation were much too
and composition of development teams vary. short (too much was asked in too little time).
● The scope of ambitions was not matched by re-
Important Scientific Research and sources, materials, and organizational and institu-
Open Questions tional support provided, for example, regarding
corresponding professional development of teachers
The Challenge of Large-Scale and the time and room for it.
Curriculum Implementation ● Ownership of the educational change is critical for
Ideally, the methods and practices of curriculum devel- success, but often lacking. There are many stake-
opment are to bridge the intended, implemented, holders involved in bringing a curriculum into
and attained curriculum. Professional literature on appropriate learning activities to achieve the desired
curriculum implementation exists, at least, as early as learning. From a systemic perspective, the transport
the 1950s. The term implementation became much of a national curriculum from the macro level to the
more in use when several scholars, for example John nano level faces many challenges. McLaughlin (1998)
Goodlad, Neal Gross, and Seymor Sarason, highlighted states that teachers’ perspectives on teaching and
around 1970 that curriculum innovations in the USA learning are often rooted in fundamentally different
of the 1960s had one fatal flaw, namely, that ideas premises of action, if not different goals, than those
were not finding their way into classrooms. During of policy makers.
the 1970s, numerous studies on the implementation ● A strict so-called fidelity of implementation
of innovations were undertaken. The Rand study was at approach does not work. It cannot be assumed
that time possibly the most comprehensive research on that top-down introduced curriculum, prescribed
implementation ever done. It encompassed 293 pro- by macro level, will be readily accepted and
jects in school districts in different regions of USA. Paul implemented by schools and teachers and students.
Berman and Milbrey McLaughlin concluded on the Firstly, there are as many interpretations of any
900 C Curriculum Development
curriculum as there are stakeholders (Fullan 2001), countries, firstly, to the supra and macro level and the
and secondly, the transformation of a written reciprocal relationship between education on one side
curriculum into learning requires substantial and sociopolitical and socioeconomic problems on
adaptations in teacher practices, attitudes, and the other. Context also relates, secondly, to the more
understanding. educational nano, micro, and meso level of learners,
● The meaning of a curriculum was not shared classrooms, and schools. Over the past decades,
between those who initiated and developed research has increased understanding about the diver-
a curriculum and those who were expected to sup- sity of educational contexts and its influence on teach-
port it and finally put it into classroom practice. ing and learning. Within the broad range of research on
context, many studies are concerned with questions
Two major conclusions derive from the findings
about the fit between context and curriculum content
and have furthered the thinking about curriculum
(subject matter, methods) as well as about the imple-
and learning. Firstly, curriculum innovations must fit
mentation requirements for particular learning and
existing classroom contingencies and ownership of
teaching contexts.
innovations must be shared. Innovations benefit from
an analysis about existing issues and requirements of
teaching and learning in classrooms and schools. Cross-References
Next to the flow of information from the macro to ▶ Competency-Based Learning
the micro level, there must also be a reverse commu- ▶ Life-Long Learning
nication from micro to macro to address concerns and ▶ Piaget, Jean (1896–1980)
practical implications of schools and teachers. Sec- ▶ Piaget’s Learning Theory
ondly, curriculum development requires room for ▶ School Climate and Learning
more adaptive implementation to provide for the ▶ Vygostky’s Philosophy of Learning
different contexts and realities of teaching and learn-
ing. Formative evaluations at early implementation References
stages may detect implementation obstacles and Fullan, M. (2001). The new meaning of educational change (3rd ed.).
thus allow for interventions before hindrances have New York: Teachers College Press.
added up. Goodlad, J., & Associates (Eds.). (1979). Curriculum inquiry: The
study of curriculum practice. New York: McGraw-Hill.
Goodlad, J. (1994). Curriculum as a filed of study. In T. Husén &
Open Questions and Research T. Postlethwaite (Eds.), The international encyclopedia of educa-
tion (pp. 1262–1276). Oxford: Pergamon Press.
Given the complexity of curriculum and learning, there
Marsh, C., & Willis, G. (1999). Curriculum: Alternative approaches,
are numerous open questions and research interests. ongoing issues. Upper Saddle River: Merrill.
A major focus of recent and ongoing research activi- McLaughlin, M. (1998). Listening and learning from the field: Tales of
ties is about how to best enable “educational change” policy implementation and situated practice. In A. Hargreaves,
realistically occurring in classrooms and schools. A. Liebermann, M. Fullan, & D. Hopkins (Eds.), International
handbook of educational change (Vol. 5, pp. 70–84). Dordrecht:
Based on the increased awareness that schools are
Kluwer Academic Publishers.
the place where curricula are transported in learning Van den Akker, J. (2003). Curriculum perspectives: An introduction.
activities, the question of school development has In J. van den Akker, W. Kuiper, & U. Hameyer (Eds.), Curriculum
become a prominent theme of educational change in landscapes and trends (pp. 1–10). Dordercht: Kluwer Academic
recent years, also because in many countries educa- Publishers.
tional reforms provide schools with more tasks and Walker, D. (2003). Fundamentals of curriculum. Mahwah: Lawrence
Erlbaum.
responsibility. Subtopics of school development are
organization and management at school level, profes-
sional development of teachers, and school-based cur-
riculum development.
Another theme of educational change relates to Curriculum Development
context. The term context refers, especially but not
exclusively for reform initiatives in developing ▶ Didactics, Didactic Models and Learning
Cybernetic Principles of Learning C 901
but is now recognized simply as cybernetics with In his collected writings from the journal Cybernet-
Wiener’s original meaning becoming a subset of the ics and Human Knowing, Glanville (2009) recognizes
expanded meaning. Ashby as one of a handful of scholars who wrote exten-
Unlike previous theories of machines Wiener’s sively and seriously about the Black Box. Glanville
breakthrough was, according to Ashby (1957), to shift himself is another. He extends the use of the Black
the question from one about the nature of mechanical Box as a way of understanding how we learn about
things (what is this?) to one about ways of behaving the world. We cannot know what is inside (i.e., what
(what does this do?). Wiener’s science of control and is “real”), but we build a model of how the world works
communication saw the transmission of an unam- as we interact with the Box and observe the “output.”
biguous message as an engineering problem (as did However, no matter how well our model stands up, we
Shannon’s 1949 Information Theory) where feedback can never say that it is a true representation, because we
governed changes in communication, which changed can never look inside the Box. And no matter how
behavior, which changed the communication and so on objective we try to be in our observations, any obser-
in a circular feedback loop that enabled a system to vation is a transaction between Black Box and observer.
maintain a desired state. Hence, cybernetics originally This is a radical constructivist position. Constructivism
had a close association with physics. However, because is based on the understanding that knowledge is
it deals with all forms of behavior, it is no longer constructed by the learner as s/he interacts with the
confined to that field of study. In fact, Ashby (1957) world. Von Glasersfeld’s radical constructivism says
saw it as providing a common language and set of that we can neither confirm nor reject an external
concepts capable of illuminating the behavior of com- absolute reality. We operate on hypotheses and only
plex systems wherever they occur including machines, revise them when they no longer fit with our experience
brains, and society. He suggests that the system should (von Glasersfeld 2007). In this respect, cybernetics has
be treated as though it were a black box with the strong links with the work of Piaget (1897–1980) who
experimenter or observer seeking patterns that link was originally a zoologist. His work describes how
the black box inputs to its outputs. By ignoring the children build up a picture of the world through mak-
contents of the box (what is this thing) and instead ing their own sense of inputs to the black box of the
concentrating on acting on the box in some way mind, converting experience to a personal understand-
(input) and observing what happens (output), the ing of concrete objects. For Piaget knowledge thus built
experimenter has no need to open the box. Instead, did not provide a “true” picture of an absolute reality
the observer’s role is to discern patterns in behavior that but provided a working model that contributed toward
link inputs to outputs. the organism’s equilibrium.
This focus on behavior and the use of the black box In this view of how we develop understanding, no
device has linked cybernetics to Skinner’s behaviorist one can be the controller sitting outside the system
psychology (e.g., Skinner 1989). Skinner saw the brain pulling levers and knowing exactly what the message
as a black box with inputs in the shape of rewards or (output) is supposed to be. The message is not trans-
punishments and outputs conditioned by those inputs. ferred unambiguously through some machine (or
Thus, according to Skinner, behavior could be modi- Black Box, which could be a mind/brain), but instead
fied and controlled and “learned behavior” could be meanings are constantly being negotiated. This is the
objectively studied. Although this may be close to the experimenter, or in the words of von Foerster (1992),
original engineering approach to cybernetics, most the observer, being both an observing system (observing
cyberneticians have for a long time seen cybernetics as the input to and output from the Black Box) and also
far from a linear input–output model. In a major part of a larger system (observer plus box) observed by
departure from Skinner’s mechanistic idea of behavior, another which in turn is observed by another and so on
Ashby includes the role of the experimenter in the with always yet another system engulfing the observed
experiment. This, as Glanville (2009) points out, indi- system and the observer. This is the observing of
cates that Ashby already understood what became observing the cybernetics of cybernetics.
known later as second-order cybernetics – the meta- The term cybernetics of cybernetics (or second-order
study of box plus investigator studying the box. cybernetics) was first used by Margaret Mead in a paper
Cybernetic Principles of Learning C 903
written in 1968. However, Glanville argues that Mead in one trigger changes in the other. Maturana and Varela
and Bateson (1972) had probably always understood (1987) say that we bring forth the world by living in it.
cybernetics in these second-order terms. He bases this In this respect, they diverge from radical constructiv-
insight on their input into the Macy conferences, ism, which says that there may or may not be a reality
chaired by psychiatrist Warren McCulloch and held in “out there” but if there is we can only ever know what it C
New York between 1946 and 1953, which brought is not, revising our hypotheses whenever we bump up
together the anthropological observations of Bateson against it and our current hypotheses do not fit.
and Mead with for example, the fields of digital com-
puting and neurophysiology. These conferences grap- Important Scientific Research and
pled with the notions of feedback and circular systems. Open Questions
In 1949, when the physicist von Foerster joined the In his musings on the current state of cybernetics,
group, the problem of the observer was raised sparking Glanville (2009) says that it was at one time considered
a debate about whether meaning is an intrinsic com- to be the new super-science. It generated great excite-
ponent of information or something attributed to the ment and was seen as relevant, for example, to physics,
observer and thus different for everyone. This shifted biology, social science, engineering as well as philoso-
the focus of cybernetics from an engineering problem phy, anthropology, and neurophysiology. It was appro-
to one of accounting for behavior in biological and priated into the world of artificial intelligence and
social systems. bionics from where it reappeared in the form of systems
Biologists Humberto Maturana and Francisco Varela theory, by which time, the excitement around cyber-
(1987) continued this shift. They argued that knowledge netics itself had died thus leaving systems theory to take
cannot reflect an ontological world but instead should be center stage. Glanville puts this fading of cybernetics as
judged by effective action in the experiential world – an area of study down to its success in providing valu-
effective, that is, from the perspective of the actor who able tools for use in such a huge range of other disci-
makes decisions on such effectiveness out of personal plines. He suggests that cybernetics itself has practically
experience and survival needs. Hence, the central notion disappeared and the origin of these tools has been
of circularity is here applied to living systems that can forgotten. Heinz von Foerster, who died in 2002 and
only know the world out of their own construction of to whom volume 10 (no. 3–4) of Cybernetics and
meaning from their experiences – which knowing then Human Knowing was dedicated, saw this differently.
becomes part of who they are, constituting the self that He believed that far from dead, cybernetics acts implic-
constructs meaning from experience and so on. In this itly and powerfully across many facets of life today,
case, “knowing” rather than “knowledge” is the impor- particularly its central notion of feedback, which is
tant word, denoting a process rather than some idea of widely known and used.
a fixed and knowable world. However, there are areas in which the field of cyber-
Maturana and Varela (1987) provide a biological netics continues to grow. Within its framework, several
explanation for this process of knowing. They equate areas of research have opened up. Glanville continues
learning with change and say that learning happens to to write extensively on the Black Box and its useful-
us as we adapt and survive. The particular change/ ness in understanding learning. Sociocybernetics has
learning depends on who we are (which depends on become an established field of study with a dedicated
the body we were born with and its history of interac- journal that explores systems science in the social sci-
tions) and the environment with which we interact ences and combines systems theory and cybernetics.
through communication, which they call ▶ languaging And cybersemiotics has been established through
and emotioning. Thus, learning cannot be stored as the work of Søren Brier, who in 1992 founded the
knowledge in the brain but is a process of going on journal Cybernetics and Human Knowing, which now
living in an environment changing and being changed covers not only second-order cybernetics but also
by it. To go on living is to go on learning; learning is ▶ autopoiesis and cybersemiotics. Cybersemiotics is
surviving – a reciprocal dynamic process between liv- a transdisciplinary framework that brings together
ing system and environment (which includes other semiotics and second-order cybernetics with cogni-
living systems and all communication) in which changes tive semantics, language game theory, and Niklas
904 C Cybernetic Principles of Learning