Kindness
Kinesthetic Communication
Know-How
Knowledge
Discourse and the Production of Knowledge
Theoretical Background
individuals who want to understand the contents of
multiple information sources dealing with the same
issue from different perspectives must not only con-
struct an integrated mental representation of the situ-
ation described across the sources but also keep the
different sources apart. Thus, both integration and sep-
aration processes are required. According to theKnowledge Acquisition
Complex Declarative LearningAffective and Cognitive Learning in the Online
Classroom
DICK Continuum in Organizational Learning
FrameworkTacit KnowledgeDancing: A Nonverbal Language for Imagining andLearningAltruism and HealthKernel Methods
Learning via Linear OperatorsKN. Seel (ed.), Encyclopedia of the Sciences of Learning, DOI 10.1007/978-1-441# Springer Science+Business Media, LLC 2012The most influential conceptualization of how individ-
uals construct meaning from multiple information
sources is the documents model proposed by Perfetti
et al. (1999). This model builds on and extends cogni-
tive-processing models describing how readers com-
prehend a single text or document. Recently, the
documents model has been further elaborated by
Goldman (2004) and Rouet (2006).
In essence, the documents model explains howKnowledge Acquisition:Constructing Meaning fromMultiple Information Sources
IVAR BRATEN, HELGE I. STRMS
Department of Educational Research, University of
Oslo, Oslo, Norway
SynonymsLearning to construct and integrate; Multiple-
documents literacy; Multiple-texts comprehension;
New literacy
DefinitionConstructing meaning from multiple information
sources denotes a new form of literacy, which involves
locating, evaluating, and using diverse sources of infor-
mation, digital as well as printed, for the purpose of
constructing an integrated, meaningful mental repre-
sentation of a particular issue, topic, or situation.9-1428-6,
1678 K Knowledge Acquisition: Constructing Meaning from Multiple Information Sources
documents model, noting and remembering the
sources of the different perspectives or arguments
(i.e., who said what) are crucial because it allows
readers to achieve overall coherence in their mental
representation of the issue even when contradictory
perspectives or arguments are presented. A person
may, for example, read about the positive effects of
using sunbeds on the website of an organization
representing the industry. However, the same person
may also read a strict warning against sunbeds on
a noncommercial, government-sponsored health por-
tal site. In such instances, noting and remembering
differences between the sources themselves, for exam-
ple, with respect to authors and document types, prob-
ably help the person to accommodate both perspectives
in his or her global understanding of the issue with less
difficulty and confusion.
In addition to author and document type, individ-
uals working with multiple information sources may
note and remember such source characteristics as rhe-
torical goal (intent, audience) and setting (place, date,
cultural setting). These source characteristics are pre-
sumably used when individuals assign weight and posi-
tion to different pieces of information in their overall
representation of the issue. For example, a person may
think that information coming from a particular
source is less trustworthy because it is produced by
a layperson rather than a professional, printed in
a newspaper rather than an encyclopedia, old rather
than new, and intent on selling something rather than
inform. As a consequence, this piece of information is
given less prominence in the overall mental represen-
tation of the issue that the person constructs when
reading the sources than information coming from
sources that the person deems more trustworthy
based on their characteristics. According to Rouet
(2006), source characteristics cannot be ignored when
trying to construct meaning from multiple informa-
tion sources because source information allows the
reader to differentiate documents, and to evaluate the
respective contribution of each document to a global
representation of the situation (p. 68).
Finally, the documents model holds that com-
prehending multiple information sources involves
constructing links that specify similarities and differ-
ences between sources. Such links are supposed to have
the form of predicates which, for example, state thatinformation in a source complements, supports, orcontradicts information in another source. According
to Perfetti et al. (1999), when a set of documents deals
with a controversial issue, such as in normal scientific
and argumentative discourse, predicates specifying
relationships between sources are often dominated by
a solidarity dimension, that is, indicate whether doc-
uments agree/disagree or support/oppose each other.
Whereas sourcesource predicates may sometimes be
explicitly expressed by authors through their use of
references, learners may also have to infer such rela-
tionships based on their prior knowledge of contents
and sources.
Important Scientific Research andOpen QuestionsSystematic research on the construction of meaning
from multiple information sources started in the
1990s. In a landmark study, Wineburg (1991) found
that expert historians who worked through multiple
documents about a particular historical event tried to
piece together a coherent interpretation of the event
described in the documents, at the same time paying
close attention to the different sources on which this
interpretation was based. In this endeavor, the histo-
rians used a strategic approach composed of three
different heuristics. First, they relied on a corroboration
heuristic that involved the systematic comparison of
content across documents to examine potential con-
tradictions or discrepancies among them. Second, they
employed a contextualization heuristic that involved the
use of prior domain knowledge to situate document
content in a broad spatialtemporal context. Third, the
historians relied on a sourcing heuristic, involving the
identification and consideration of source characteris-
tics such as authors, document types, and place and
date of document creation. This information was used
to determine the evidentiary value of each document,
as well as to interpret the documents content. In con-
trast, high school students participating in Wineburgs
study seldom used these three heuristics when working
with the same documents, for example, paying much
less attention to source characteristics and having
difficulty resolving and even noticing discrepancies
among the sources.
Subsequent work, not only in history but also in
other domains (e.g., science and law), has confirmed
the importance of prior knowledge and strategicprocessing to the comprehension and integration of
Knowledge Acquisition: Constructing Meaning from Multiple Information Sources K 1679
Kmultiple documents about a particular issue (Goldman
2004; Rouet 2006). Prior domain knowledge seems
fundamental to drawing inferences needed to link
information and viewpoints across diverse sources,
and also to understanding important sources and
source characteristics in a domain, for example,
which publications or author qualifications may give
credence to a claim. Moreover, active strategic pro-
cessing that involves cross-document elaboration to
compare, contrast, and integrate contents across docu-
ments, as well as the evaluation of sources to decide
which to trust and which to mistrust, seems important
for success in the complex task context of multiple
information sources.
Recently, individuals epistemic beliefs, that is, their
beliefs about knowledge and knowing in a domain,
have also been shown to play an important role in
their comprehension of multiple information sources.
Specifically, believing that knowledge in the domain is
tentative or complex or that knowledge claims need to
be justified by cross-checking of sources is more adap-
tive that believing that knowledge in the domain is
certain and simple and that there is no need or way to
justify knowledge claims (Braten et al. 2011). For exam-
ple, individuals believing knowledge to be complex
may bemore inclined to produce cross-document elab-
orations while working with multiple information
sources, whereas individuals believing knowledge to
consist of an accumulation of simple facts may be
more inclined to rehearse and paraphrase pieces of
factual information from single sources.
Not only individual but also document and task
variables have been shown to play a role in the com-
prehension of multiple information sources. For exam-
ple, integration across documents is easier when the
contents of documents largely overlap, and cross-
document elaboration is facilitated by documents that
explicitly refer to and support or contradict each other
(Goldman 2004). With respect to task variables, some
evidence suggests that comprehension and integration
can be facilitated by argument-centered writing tasks,
that is, by asking students to write argument essays
based on multiple documents (Perfetti et al. 1999;
Rouet 2006). However, research also indicates that
only students with relatively high prior knowledge or
the belief that knowledge is tentative may be able to
take advantage of instructions to construct arguments,whereas students with low prior knowledge or thebelief that knowledge is certain may actually be more
hindered than helped by such task instructions (Braten
et al. 2011).
Although constructing meaning from multiple
information sources generally demands much of the
learner, research demonstrates that reading about
a controversial topic in multiple source documents
rather than a single source presenting the same content
may facilitate deep-level, integrated comprehension
(Britt and Aglinskas 2002). However, spontaneously,
many students at different educational levels have
been found to have great difficulty coping with this
complex activity and, therefore, need explicit instruc-
tion. Britt and Aglinskas (2002) addressed this instruc-
tional challenge and clearly demonstrated that students
can be successfully taught both integration and separa-
tion processes involved in learning from multiple doc-
uments. In their approach, a computer application
called the Sourcers Apprentice was used to present
separate documents about historical controversies, and
students learned to take note of both content and
source characteristics of each document. A series of
evaluation studies with North American high school
students demonstrated that those using the Sourcers
Apprentice improved their ability to integrate informa-
tion across sources and paid more attention to the
sources themselves when compared with students in
control groups (Britt and Aglinskas 2002).
Despite the foundational scientific work referred to
above, much remains to be known about the construc-
tion of meaning from multiple information sources.
First, more online processing data are needed on how
individuals work with multiple documents to compre-
hend a particular issue, topic, or situation. Useful
approaches in this regard may be the collection and
analysis of concurrent verbal reports (i.e., think-
alouds), traces in the learning materials (e.g., soft-
ware-logged activities), and eye movements during
task completion. Second, assessing the learning out-
comes of working with multiple information sources
in reliable and valid ways remains a vital issue. Building
on preliminary work employing both questionnaires
and essay tasks (Braten et al. 2011), an important goal
is to define standard procedures for assessing individ-
uals competence that can be adapted to different edu-
cational levels and subject areas. Finally, given that
only a small set of intervention studies have been car-ried out, more research on how individuals can be
sources is greatly needed.
Strategic Learning
Text Relevance
1680 K Knowledge and Learning in Natural LanguageReferencesBraten, I., Gil, L., & Strms, H. I. (2011). The role of different task
instructions and reader characteristics when learning from mul-
tiple expository texts. In M. T. McCrudden, J. P. Magliano, &
G. Schraw (Eds.), Text relevance and learning from text. Green-
wich: Information Age.
Britt, M. A., & Aglinskas, C. (2002). Improving students ability to
identify and use source information. Cognition and Instruction,
20, 485522.
Goldman, S. R. (2004). Cognitive aspects of constructing meaning
through and across multiple texts. In N. Shuart-Faris & D.
Bloome (Eds.), Uses of intertextuality in classroom and educa-
tional research (pp. 317351). Greenwich: Information Age.
Perfetti, C. A., Rouet, J. F., & Britt, M. A. (1999). Toward a theory of
documents representation. In H. Van Oostendorp & S. R.
Goldman (Eds.), The construction of mental representation during
reading (pp. 99122). Mahwah: Erlbaum.
Rouet, J. F. (2006). The skills of document use: From text comprehension
to web-based learning. Mahwah: Erlbaum.
Wineburg, S. (1991). Historical problem solving: A study of the
cognitive processes used in the evaluation of documentary and
pictorial evidence. Journal of Educational Psychology, 83, 7387.
Knowledge and Learning inNatural Language
CHARLES YANG
Department of Linguistics & Computer Science,
University of Pennsylvania, Philadelphia, PA, USA
DefinitionKnowledge and Learning in Natural Language (KLNL)Cross-ReferencesBeliefs about Learning Learning Strategies for Digital Media
Learning with Multiple Goals and Representations Literacy and LearningNaturalistic Epistemology
Online Learningeffectively taught knowledge, beliefs, and strategies that
can promote their learning from multiple information(Yang 2002) is a synthesis of computational learningand theoretical linguistics to provide explanations for
child language development. It was a first effort to
integrate probabilistic learning mechanisms with the
theory of Universal Grammar, and has provided
a quantitative link between the statistical properties of
the input data to the learner and the developmental
patterns of the learners grammar.
Theoretical BackgroundThe Principles and Parameter framework (Chomsky
1981) is a response to the challenges posed by language
learnability. By attributing the totality of languagevariation to a finite set of parameters, the learners
hypothesis spacemay be effectively constrained to facil-
itate language acquisition. Even though the parameter-
based approach is most closely identified with the
Chomskyan approach to language, most modern lin-
guistic theories similarly admit only a finite range of
possible grammars; the learners task is to select those
used in his or her linguistic environment.
Thus, the acquisition problem becomes one of
parameter setting. The dominant approach follows
the conception of triggering (Chomsky 1981; Gibson
and Wexler 1994). The learner is identified with
a grammar (i.e., a string of parameter values) and
makes changes to that grammar (i.e., changing param-
eter values) as input data are processed. Aside from
learnability problems with this approach, the assump-
tion of one grammar at a time fails to account for the
fact that childs language during acquisition generally
cannot be identified with a single adult-like grammar.
Moreover, child grammar typically develops gradually,
challenging the notion of triggering where the learner
makes categorical changes to his or her grammar
(Valian 1991).
KLNL makes a break with the tradition of trig-
gering and much work in the tradition of Universal
Grammar. The learner is modeled as a population of
grammars, whose probabilistic distribution changes
in response to the input data in a Darwinian selec-
tionist fashion. The learner nondeterministically
selects a string of parameter values with their associ-
ated probabilities, and rewards or punishes these
probabilities based on the success or failure of ana-
lyzing an input utterance (Bush and Mosteller 1951).
The rise of the target grammar is smooth, which pro-
vides an explanation for the gradualistic developmentof child language.
possibility that the mechanisms of language learning
are domain-general while still operating within
domain-specific constraints of language. On the one
Unlearning (The Nature of. . .)
Knowledge Change K 1681
KImportant Scientific Research andOpen QuestionsPerhaps the most attractive feature of the KLNL learn-
ing model lies in its ability to connect the input data to
child grammar development. Under the competition-
based scheme, parameter values which are expressed
more frequently in the input data will be established
sooner than those expressed less frequently. This allows
the researcher to correlate the statistical properties of
the input data, which is available through collections of
child-directed speech, with the longitudinal develop-
ment of child grammar. In addition, under the KLNL
model nontarget grammars (before their demise) can
be probabilistically accessed by the learner. This main-
tains the fruitful tradition in acquisition research that
the errors in child language may depart from the target
language but nevertheless is systematic and linguisti-
cally possible. Similar approaches have been applied to
the acquisition of morphology and phonology. Lin-
guistic rules can be associated with probabilities
thereby capturing well-known developmental findings
such as frequency effects that previously had no place in
the traditional approach to child language under gen-
erative grammar.
An important direction for research is to further
clarify the formal properties of probabilistic learning
models similar to KLNL. Convergence proof has been
obtained, though the efficiency of learning depends on
the landscape of the parameter space and, given the
probabilistic nature of the learning model, the distri-
bution of the input data. Parameters are designed to tie
together wide-ranging generalizations to linguistic
expressions that are readily available in the input.
Recent work exploring large and linguistically realistic
samples reveals that the space of parameters may be
sufficiently smooth to facilitate plausible language
acquisition. An open question here concerns the
scope and limit of parameters in linguistic description
and acquisition: every language has specific rules and
constructions that are results of history and cannot be
plausibly attributed to the innate endowment. Current
efforts are devoted to the development of an integrated
model that carries out probabilistic selection among
innate parametric choices while at the same time con-
structs language-particular rules and determines their
range of applications.
A final and broader consideration that emergesfrom the KLNL approach lies in the connectionReferencesBush, R., & Mosteller, F. (1951). A mathematical model for simple
learning. Psychological Review, 68, 313323.
Chomsky, N. (1981). Lectures on government and binding.
Dordrectht: Foris.
Chomsky, N. (2005). Three factors in language design. Linguistic
Inquiry, 36(1), 122.
Gibson, E., & Wexler, K. (1994). Triggers. Linguistic Inquiry, 25,
355407.
Valian, V. (1991). Syntactic subjects in the early speech of American
and Italian children. Cognition, 40, 2182.
Yang, C. (2002). Knowledge and learning in natural language.
New York: Oxford University Press.
Knowledge Changehand, this suggests that individual variation in lan-
guage acquisition, including certain clinical cases,
may not be the result of deficiencies in the affected
individuals linguistic ability, but in the ability to pro-
cess and analyze the input data that may be reflected in
other cognitive and perceptual tasks. On the other
hand, the probabilistic learning mechanisms in param-
eter setting appear to be evolutionarily ancient (Bush
and Mosteller 1951), which raises interesting questions
for the evolution of language: specifically, how the ways
we learn might have shaped the organization of lan-
guage in its attested form (e.g., parameters).
Cross-References Formal Learning Theory Infant Learning and Development Language Acquisition and Development
Learnability Learning and Instinct
Mathematical Models/Theories of LearningReinforcement Learning in Animals
Statistical Learning in Perception Stochastic Models of Learningbetween the language faculty and other cognitive sys-
tems (Chomsky 2005). The KLNL model opens up theCognitive Learning
Definition
practices, and developing expertise. It has been pro-
posed that this view of learning is a third main meta-
Theoretical Background
1682 K Knowledge ClaimThe knowledge creation metaphor is a claim that there
is an emerging trend of theories about human learning
and cognition which aim at understanding how people
organize their work and learning for developing andphor of learning, which is becoming more and more
important in modern society (in contrast to the acqui-
sition metaphor of learning and the participation met-
aphor of learning).An overall term for such theories and views of learning
which emphasize learning and human cognition as
processes of developing and pursuing certain novelties
(artifacts, products, practices, concepts, activities, pro-
cesses) collaboratively and with distributed means
where individuals initiative is embedded in fertile
social and institutional practices and processes. The
focus is on advancing knowledge, transforming socialKnowledge Claim
Belief Formation
Knowledge Compilation
Restructuring in Learning
Knowledge Creation Metaphor,The
SAMI PAAVOLA
Institute of Behavioral Sciences, University of Helsinki,
Helsinki, Finland
SynonymsCollaborative knowledge creation; The knowledge
creation approach; The knowledge creation metaphor
of learningcreating things together. This metaphor is a sequel toAnna Sfards (1988) famous distinction between acqui-
sition and participation metaphors of learning. Sche-
matically described, the acquisition metaphor focuses
on processes of adopting or constructing subject-
matter knowledge or conceptual knowledge within an
individuals mind, whereas theories representing the
participation metaphor focus on processes of socializing
in social communities and in social interaction and
practices. The acquisition metaphor emphasizes usu-
ally logically oriented epistemology, whereas the par-
ticipation metaphor emphasizes such things as
communities, social identities, cultural mediation,
and the situatedness of human cognition. It can be
maintained that this distinction is a very apt character-
ization of basic theories of learning as such, but in
order to understand the emerging new phenomena
related to collaborative creativity and learning, a third
basic metaphor of learning should be defined where
change, an organized aim of developing something
new, and the role of mediating artifacts are emphasized.
According to Paavola et al. (2004); see also Hakkarainen
et al. (2004), influential representatives of the knowl-
edge creation metaphor in learning sciences are the
knowledge building approach (Bereiter 2002), and
expansive learning as well as cultural-historicalactivity theory (Engestrom and Sannino 2010), and,
within organizational sciences, Nonaka and Takeuchis
theory of knowledge-creating companies. The term
knowledge creation owes a lot to the theory of orga-
nizational knowledge creation (Nonaka and Takeuchi
1995), but nowadays it is used often in various fields of
research. Theories representing the knowledge creation
metaphor are quite different from each other. In spite
of clear and in some sense fundamental differences
among these theories, they have many features in com-
mon, that is, focuses that include (1) the pursuit of
newness; (2) processes of mediation to avoid Cartesian
dualisms (such as mind vs matter, or concepts vs mate-
rial objects); (3) social processes, while also emphasiz-
ing the role of individual subjects in knowledge
creation; (4) going beyond propositional and concep-
tual knowledge as a sole locus of learning, while recog-
nizing conceptualizations and conceptual artifacts as
central for knowledge creation; and (5) the interaction
around and through shared objects.
It has been maintained that various changes in
modern society form a basis for the knowledge crea-tion metaphor of learning, such as: (1) the rapid
Knowledge Creation Metaphor, The K 1683
Kdevelopment of new technology, which has formed and
is all the time forming qualitatively new possibilities for
distributed interaction and collaboration; (2) the pres-
sure to create and learn deliberately to create new
knowledge and transform existing practices in various
areas of life; and (3) the complexity of modern society,
which requires people to combine their expertise to
solve emerging and often unforeseen complex problems.
Important Scientific Research andOpen QuestionsThe concept of the knowledge creation metaphor
(KCM) of learning was first developed in relation to
computer-supported collaborative learning (CSCL)in order to broaden or better understand the epistemo-
logical basis of technology-mediated learning and how
processes of innovative knowledge advancement and
discovery are taken into account (Paavola et al. 2002).
Partially because of this history, this concept has
been developed and used more in relation to CSCL
research and technology-mediated collaborative learn-
ing and inquiry learning. There are new methodologi-
cal developments in how to research knowledge
creation processes and how to advance and support
collaborative knowledge creation with technology and
related knowledge practices. Quite often the KCM is
interpreted as a continuation or even a synonym for the
knowledge building approach because both of them
emphasize a new kind of approach needed for answer-
ing the challenges of modern knowledge work. The
original idea of the KCM has, however, been that it
includes different kinds of approaches where some of
them emphasize collaborative idea development (such
as knowledge building) and some collaborative practice
transformations (such as expansive learning). In the
CSCL context, it is maintained that the KCM provides
a basis for a trialogical approach to learning where an
emphasis on meaning making and dialogs is
complemented with a focus on jointly constructed
knowledge practices and knowledge artifacts.
The KCMhas also been used inmany contexts other
than CSCL, such as when reviewing various approaches
to workplace learning and organizational learning, or
doing research on teacher education and teacher com-
munities, networked learning, and networked exper-
tise, while defining a new digital epistemology for
schools and teacher education, or analyzing the natureof open source communities. The KCM should notstand processes of collaborative knowledge creation. It
remains an open question as to what extent the partic-
ipation metaphor differs from the knowledge creation
metaphor, or if the participatory approaches can be
developed so that they include knowledge creation
processes. Another basic criticism against the KCM is
that it aims at seeing commonalities between theories
which are epistemologically and methodologically very
different, and leads to eclecticism. According to this
view, better alternatives to the acquisition and the
participation approaches would be such metaphors as
expansion (Engestrom and Sannino 2010) or
knowledge building, but not both together. One
answer to these criticisms is that the KCM is not
meant to be used for overruling differences between
various approaches, but aims at pointing out some
important phenomena in areas related to collaborative
learning which can be used for enriching methodolog-
ical and theoretical development and discourse in these
areas of research.
Cross-References 21st Century Skills
Collaborative Knowledge BuildingComputer-Supported Collaborative LearningContradictions in Expansive Learning
Cultural-Historical Theory of Development Learning Metaphors
ReferencesBereiter, C. (2002). Education and mind in the knowledge age.
Hillsdale: Erlbaum.
Engestrom, Y., & Sannino, A. (2010). Studies of expansive learning:
Foundations, findings and future challenges. Educationalonly be seen in relation to new digital technology
because similar mediated processes which are orga-
nized for producing things together can be done with-
out the use of any special technology. Very often the
KCM is, however, related to new technology that pro-
vides new means for collaboration, such as new collab-
orative learning environments, or Web 2.0 technology
in general, wikis, podcasts, and so on.
The KCM is proposed as a supplement or an alter-
native to theories belonging to the participation meta-
phor of learning with the idea that theories
representing the participation metaphor (or the acqui-
sition metaphor) are not enough if the aim is to under-Research Review, 5(1), 124.
Nonaka, I., & Takeuchi, H. (1995). The knowledge-creating company:
How Japanese companies create the dynamics of innovation.
munity, proceedings of: CSCL 2002 (January 711, 2002, Boulder,
Colorado, USA). Hilldale: Lawrence Erlbaum.
choosing just one. Educational Researcher, 27(2), 413.
cess into ones physical bearing and actions, e.g., the
lying rule or a scientific concept. Because they lack this
viously unrelated pieces of knowledge together in their
1684 K Knowledge EmbodimentKnowledge Gaps
Cognitive Dissonance in the Learning Processes
Knowledge Generation
Generative Learning
Knowledge Improvement
A process of learning in which learners are engaged inbodily understanding that people bring to religious
worship or a trade.
Knowledge Encoding
Routinization of LearningKnowledge Embodiment
Process of incorporating knowledge of a cultural pro-Paavola, S., Lipponen, L., & Hakkarainen, K. (2004). Models
of innovative knowledge communities and three metaphors of
learning. Review of Educational Research, 74(4), 557576.
Sfard, A. (1998). On two metaphors for learning and the dangers ofNew York: Oxford University Press.
Paavola, S., Lipponen, L., & Hakkarainen, K. (2002). Epistemological
foundations for CSCL: A comparison of three models of inno-
vative knowledge communities. In G. Stahl (Ed.), Computer
support for collaborative learning: Foundations for a CSCL com-Hakkarainen, K., Palonen, T., Paavola, S., & Lehtinen, E. (2004).
Communities of networked expertise: Professional and educational
perspectives (Advances in learning and instruction series).
Amsterdam: Elsevier.constructively critiquing and improving each othersmemory, often as a result of instruction.
Theoretical BackgroundResearch on knowledge integration is closely relatedabstract background knowledge, novices tend to focus
on the superficial differences between pieces of knowl-
edge they acquired in seemingly unrelated situations
(cf. Chi et al. 1981). They, then, store these pieces
independent of each other in their long-term memory,
which results in fragmented knowledge. Knowledge
integration takes place whenever learners connect pre-ideas. It may involve the generation of a groups idea
that is co-constituted by the interactions between mem-
bers of a group discussing their own individual ideas.
Knowledge Integration
MICHAEL SCHNEIDER
Institute for Behavioral Sciences, ETH Zurich,
Zurich, Switzerland
SynonymsIntegrating knowledge; Merging knowledge structures
DefinitionKnowledge integration refers to the process of merging
two or more originally unrelated knowledge structures
into a single structure. In the most general sense, it can
encompass the complexities of how two digital data-
bases can be merged together or how two companies
can effectively combine the knowledge of their workers.
In the learning sciences, however, the term usually
refers to knowledge integration within persons mem-
ory. Learners pick up pieces of knowledge (e.g., expe-
riences, observations, ideas, hypotheses, explanations)
in many different situations, for example, everyday life
observations, conversations with friends, the Internet,
and school instruction. Novices in a domain often do
not see which of these newly acquired pieces of knowl-
edge relate to each other and why they should be related
at all. Recognizing relations usually depends on rele-
vant prior knowledge, for instance, knowing an under-to research on conceptual change. Both investigate
Knowledge Integration K 1685
Khow learners acquire complex knowledge structures
in conceptually rich domains (diSessa 2006). Some
conceptual change researchers argue that conceptual
knowledge mostly takes the form of subjective naive
theories which are rather coherent and integrated. This
is the knowledge as theory perspective, from which
the problem of knowledge integration is negligible.
Other researchers argue that conceptual knowledge is
frequently fragmented, particularly for learners with
little expertise in a domain. This is the knowledge as
elements perspective on conceptual change. From this
perspective knowledge integration is an important
learning mechanism which should be activated and
strengthened by instruction.
One of the most widely recognized knowledge as
elements theories has been formulated by Andrea
diSessa and colleagues (diSessa et al. 2004). According
to this view, students acquire knowledge in pieces.
These pieces of knowledge are abstractions of common
experiences, for example, everyday life observations,
which diSessa calls phenomenological primitives or
p-prims. P-prims may or may not be conscious. They
are initially not related to each other, because the
learner does not have the necessary knowledge to relate
them to each other. Over time, that is, with increasing
expertise and knowledge in a domain, learners inte-
grate p-prims into more organized knowledge struc-
tures. DiSessa emphasizes the importance of what
he calls coordination classes for this process. Coordina-
tion classes are systematic collections of strategies
for reading related information out from the world.
Coordination classes eventually lead to a different view
of the world, a view in which previously unconnected
p-prims are seen as aspects of the same more general
concept.
A related theoretical approach (Schneider and Stern
2009) emphasizes the roles of working memory and
long-term memory in knowledge fragmentation and
integration. Learners must create a complex, well-
structured, and well-integrated network of knowledge
in long-term memory in order to acquire expertise in
a domain. Long-term memory has a virtually unlim-
ited capacity. However, it does not allow for the active
transformation and integration of knowledge. This
integration is possible only in working memory,
which can hold only a few pieces (also called chunks)
of knowledge at a time. Thus, building up well-integrated knowledge structures in long-term memoryrequires many cycles of loading some pieces of knowl-
edge into working memory, integrating them when
appropriate, and storing the results back in long-term
memory. This process is complicated by the fact that
novices in a domain often do not understand which
pieces of knowledge are plausible candidates for knowl-
edge integration and should be loaded into working
memory simultaneously.
Finally, Baroody (2003) and other researchers on
mathematics learning recommend that in addition to
integrating pieces of conceptual knowledge, learners
should also integrate their concepts and their prob-
lem-solving procedures. Conceptual knowledge can
help with the construction of new procedures, the
modification of existing ones, the transfer of proce-
dures from well-known to new problem types, adap-
tive choices between alternative procedures, and
monitoring of the execution of a procedure. Proce-
dural knowledge, in turn, is necessary to quickly and
efficiently apply conceptual knowledge to the solution
of problems. However, this is possible only when
learners realize how concepts and procedures relate
to each other.
Important Scientific Research andOpen QuestionsResearch on knowledge integration is a young field
which has emerged over the last 20 years from studies
on science learning (diSessa 2006; Linn 2006) and
mathematics learning (Baroody 2003; Schneider and
Stern 2009). It still needs to be generalized to other
content domains.
Marcia C. Linn (2006) and her colleagues con-
ducted over 40 empirical case studies on the way
knowledge integration occurs and how it can be facil-
itated by learning environments. They found that the
processes of knowledge fragmentation and integration
are common in science learning. Successful instruc-
tional interventions which foster knowledge integra-
tion usually include some or all of the following
four activities. First, successful instruction stimulates
students to elicit their current ideas. Students profit
most when they are prompted to consider ideas from
many different contexts, for example, school, home,
recreation, and museum. Eliciting ideas in group set-
tings helps learners to also consider the ideas of their
peers and to integrate knowledge by reflecting on theseideas. Second, adding new, normative ideas can have
representation of physics problems by experts and novices. Cog-
ing and instruction. In R. K. Sawyer (Ed.), The Cambridge hand-
book of the learning sciences (pp. 243264). Cambridge:
1686 K Knowledge Management
a positive effect on students. Learners sometimes prefer
learning something new over reflecting on their (par-
tially incorrect) prior knowledge. The acquisition of
new ideas can increase knowledge integration, when
these ideas stimulate the reconsideration of existing
views, for example, by demonstrating the connection
between two everyday life experiences or by illustrating
an abstract idea. Third, instructions which foster
knowledge integration should help students to develop
criteria for the evaluation of their own ideas. Interven-
tions helping students to acquire this metacognitive
competence will encourage students to judge ideas
based on their validity and plausibility. This is an
important competence, because in everyday life
learners encounter a mix of valid ideas and bogus
stories, for example, on the Internet. Looking for sys-
tematic connections between claims, empirical evi-
dence, sources of information, persons perspectives
and intents can help to sort out valid ideas by seeing
them as parts of a bigger system of knowledge. Finally,
the integration of knowledge can require sorting out
ideas that are incorrect and contradict other parts of
the learners knowledge base. This sorting process is
based on the abovementioned processes, because it
requires students to already have acquired some nor-
matively correct ideas and criteria for distinguishing
between more and less adequate ideas. The process of
sorting out ideas leads to a knowledge base that is more
focused and thus more easily comprehended. In sum-
mation, empirical research shows that knowledge inte-
gration is a challenging and time-consuming process
which is facilitated by well-prepared learning environ-
ments. Ideally, these are adapted to the learners specific
prior knowledge, their everyday life, and the broader
sociocultural environment so that knowledge from all
of these sources is integrated.
Cross-ReferencesComplex Declarative LearningConceptual Change
Constructivist LearningKnowledge Acquisition: Constructing Meaning from
Multiple Information Sources
Knowledge Organization
Learning and UnderstandingModel-Based Learning
Role of Prior Knowledge in Learning Processes
Science, Art and Language ExperiencesKnowledge Maps
Concept Maps
Knowledge Organization
GABI REINMANN
Learning and Teaching with Media, Universitat der
Bundeswehr Munich, Germany
SynonymsInformation classification; Information organization;
Knowledge representation; Knowledge structuring
DefinitionThe term knowledge organization has at least twoKnowledge Management
Acquiring Organizational Learning NormsCambridge University Press.
Schneider,M., & Stern, E. (2009). The inverse relation of addition and
subtraction: A knowledge integration perspective. Mathematical
Thinking and Learning, 11(1), 92101.nitive Science, 5(2), 121152.
diSessa, A. A. (2006). A history of conceptual change research. In
R. K. Sawyer (Ed.), The Cambridge handbook of the learning
sciences (pp. 265280). Cambridge: Cambridge University Press.
diSessa, A. A., Gillespie, N. M., & Esterly, J. B. (2004). Coherence
versus fragmentation in the development of the concept of force.
Cognitive Science, 28, 843900.
Linn, M. C. (2006). The knowledge integration perspective on learn-ReferencesBaroody, A. J. (2003). The development of adaptive expertise and
flexibility: The integration of conceptual and procedural knowl-
edge. In A. J. Baroody & A. Dowker (Eds.), The development of
arithmetic concepts and skills: Constructing adaptive expertise
(pp. 133). Mahwah: Erlbaum.
Chi, M. T. H., Feltovich, P. J., & Glaser, R. (1981). Categorization andmeanings depending on the concept of knowledge.
Knowledge Organization K 1687
KIf you concentrate on personal knowledge to which only
the individual has direct access, knowledge organiza-
tion indicates amental and therefore internal process of
structuring or transforming (e.g., visualizing) knowl-
edge. If you concentrate on public knowledge (syno-
nym: information) which is materialized in documents
(text, image, audio, video) so that principally all per-
sons can access it, knowledge organization refers to the
technical and external processes of describing and clas-
sifying information. A third definition takes a socio-
logical perspective and indicates such social entities as
knowledge organizations whose work primarily relies
on knowledge (like universities); this latter meaning
will not be included in this entry.
Theoretical BackgroundThe theoretical background as well as the history of the
term knowledge organization has two disciplinary
roots: (1) learning and cognitive psychology and
(2) information and computer sciences. The preferred
perspective depends on the underlying conception of
knowledge. Therefore, the term knowledge organiza-
tion can be applied only when considering the concept
of knowledge. There are many understandings of
knowledge which differentiate its special characteristics
and forms. Among others, Jean Piaget (18971980) has
had great influence on the conception of knowledge.
Piaget represents a theory of genetics of cognitive
structures giving the individual a crucial and active role
in constructing knowledge. But the theory also con-
siders the strong interdependence between individual
constructions and external stimuli of the external world
(and their experience). These two complementary pro-
cesses which balance this interdependence are called
accommodation and assimilation.
Against this background knowledge is a very per-
sonal and mental phenomenon to which only the indi-
vidual has direct access. Under this perspective you can
speak about personal knowledge (Seiler 2004). Personal
knowledge can be of different specification (especially
conceptual, visual, and enactive) and is therefore more
or less easy to articulate and to make explicit. Never-
theless, persons can communicate and collaborate so
that one must assume that there is not only implicit or
idiosyncratic but also shared knowledge mostly con-
ventionalized through language and often materialized
in written, spoken, or visualized documents like text,audio, image, or video. Documented knowledge likethis is of principally public access. Under this perspec-
tive you can speak of public knowledge or information
(Seiler 2004).
To distinguish between personal and public knowl-
edge is the only one alternative of many to structure the
domain of knowledge. But exactly this possibility can
aid one to make the decision to choose either the psy-
chological or technical discipline as the leading one for
research of knowledge organization. Learning and cog-
nitive psychology provide theoretical and empirical
insights referring to personal knowledge organization
which is always an internal process. Computer and
information sciences in contrast deliver the scientific
basis to organize public knowledge or rather informa-
tion which always results in an external representation.
Internal knowledge organization: The growth of per-
sonal knowledge is part of human development and
does not happen randomly but in some organized
manner. There are three representational constructs
describing this internal knowledge organization:
semantic networks, theories, and schemas (see Chi
and Ohlsson 2005).
Knowledge can be represented in semantic networkswhich consist of concepts (called nodes) and rela-
tions (called links). The organization of knowledge
in semantic networks indicates that everything is
related to everything. The quality of this organiza-
tion is determined by the number and character of
relations (e.g., hierarchical, temporal, or causal)
between concepts as well as by content similarity
of concepts. So cognitive psychologists assume that
knowledge in semantic networks is grouped by
domain.
Domain-specific knowledge can also be representedas theories: Theories in this sense can be well-
articulated structures with core knowledge ele-
ments (big ideas) in the center and peripheral
concepts around them. This attribute (i.e., different
importance of elements) characterizes theories in
expert knowledge as well as those in novice knowl-
edge. The latter, however, are rather intuitive theo-
ries lacking the depth of scientific theories of
experts.
The representation of knowledge in schemas resultsin the assumption that humans construct patterns
of experience. A schema is a set of attributes (calledslots) which can take on different values referring to
1688 K Knowledge Organization
phenomena in the external world. Schemas are typ-
ically abstract and organize knowledge about spe-
cific stimulus domains. They are retrieved as units
and are used to organize learning, thinking, and
acting.
To a certain extent internal representations like
semantic networks, theories, and schemas are the
results of natural processes of personal knowledge
organization which do not have to be externally
instructed and controlled. Nevertheless, there are
some possibilities to foster and optimize the internal
knowledge organization through educational support.
Nearly all proposals which do this stem from the
research of (cognitive) learning strategies and are
based upon the principle of producing external repre-
sentations. These external representations can be ver-
bal, pictorial, or something in between, for example,
maps and other logical pictures.
You can foster internal knowledge organization byarticulating knowledge verbally or in written form.
To think out loud can help to structure knowledge
because transforming implicit knowledge in
conventionalized words fosters the construction or
retrieval of schemas. Furthermore, specific pro-
grams have been developed to support knowledge
generation and organization through writing and
text production, respectively.
A widely used strategy in organizing knowledgerecurs to visualization. Psychological research on
memory shows that there are spontaneous produc-
tions of internal imageries probably influencing the
internal knowledge organization. So you can use
external imagery strategies to foster these processes.
But there is greater evidence in the surplus of logical
forms of visualizations like diagrams and maps.
Because of the structural similarity to assumed
semantic networks (as internal knowledge repre-
sentations) concept mapping is a widespreadstrategy or cognitive tool to produce effective exter-
nal representations.
Processes and methods of internal knowledge orga-
nization are often integrated in educational settings (as
learning strategies), but they are also part of so-called
personal knowledge management strategies through
which adults should optimize their personal growth aswell as work performance (Holsapple 2004).External knowledge organization: The growth of
public knowledge, or rather information, is part of
the sociocultural development and needs sophisti-
cated forms of organization because of the increasing
amount of scientific knowledge focused in this context.
There are traditional and technology-based forms for
external knowledge organization in order to gather,
describe, index, classify, store, and find documents
(Hjrland 2008).
Information can manually be indexed and classifiedlike in former libraries or archives. This procedure is
slow but brings the advantage that persons (librar-
ians) work withmeaning in practice. However, today
all information institutions are using computer sys-
tems for archiving, identifying, and retrieving infor-
mation relevant to specific purposes.
So information also can automatically be structuredin very different ways using insights of linguistics,
logic, mathematics, and philosophy. Information
scientists know a lot of methods for organizing
information, for example, from the general to the
specific, facet-analytical, bibliometric, user-ori-
ented, and so forth.
Like personal knowledge information can be visu-ally represented using new information technolo-
gies: Visualization presupposes well-structured
information and greatly facilitates the retrieval of
information which most notably is an advantage
with huge amounts of information.
For a short time even technology-based informa-tion can also be socially organized, for example,
through social tagging or social bookmarking.
Instead of expert taxonomies leading the external
knowledge organization, so-called folksonomiesare structuring the public knowledge on the
Internet.
Processes and methods of external knowledge orga-
nization are not only used in libraries and archives,
but also in knowledge-based organizations engaging
in organizational knowledge management whereknowledge organization covers not all but an impor-
tant part of management tasks (Holsapple 2004).
Important Scientific Research andOpen QuestionsKnowledge organization is an interdisciplinary chal-lenge combining several disciplines which do not use
many contexts (e.g., education and work place).
behavior is associated with the existence of applicable
to perform a complex process, task, or activity.
Modern cognitive science sees cognition and learn-
Knowledge Representation K 1689
Khand and computer and information science on the
other hand. (b) External knowledge organization in
contrast presupposes objectives and criteria which are
possible only on the basis of internal representations of
persons engaging in articulation, visualization, or other
forms of externalization. Perhaps the field of visualiza-
tion is a seminal research object which connects differ-
ent disciplines working on knowledge organization
(see Tergan and Keller 2005). Information and knowl-
edge visualization share the common goal to organize
information and knowledge for better access, search,
and understanding and use comparable techniques and
methods. Joint research can be noticed with concept
mapping as a method of graphical representation fos-
tering internal and external forms of knowledge orga-
nization for different purposes.
Cross-ReferencesKnowledge Integration
Knowledge Representation Learning Strategies
Organizational Change and Learning Schema(s)
ReferencesChi, M. T. H., & Ohlsson, S. (2005). Complex declarative learning. In
K. J. Holyak & R. G.Morrison (Eds.),The Cambridge handbook of
thinking and reasoning (pp. 371400). Cambridge: Cambridge
University Press.
Hjrland, B. (2008). What is knowledge organization (KO)? Knowl-
edge Organization, 35(2/3), 86101 (International Journal
devoted to Concept Theory, Classification, Indexing and Knowl-Beneath this practical argument there are strong theo-
retical reasons for collaboration between scientists of
different origin: (a) Internal knowledge organization
often needs external representations as well as technical
tools to stimulate and support them. Further research
has to analyze the interrelation between methods based
upon cognitive and learning psychology on the onethe same concepts of knowledge or information. Com-
munication between these disciplines is rare and diffi-
cult because there is little grounding referring to the
underlying core concept. As a consequence, different
scientific camps which widely ignore each other are
developing although psychological and technical pro-
cesses of knowledge organization should correspond inedge Representation).ing as a complex process with many facets, including
symbolic representations of objects and events which
are not immediately present but exist only in imagina-
tion. Accordingly, most cognitive scientists agree on the
basic assumption that cognition and learning take place
in the use of mental representations, in which individ-
uals organize symbols of experience or thought in suchknowledge. By relating intelligence and knowledge, the
systems behavior becomes more or less reconstructible
and predictable. The most discussed distinction is
between declarative (knowing that) and procedural
(knowing how) knowledge (see Anderson 1983).
Declarative knowledge is defined as factual
knowledge, whereas procedural knowledge is defined
as the knowledge of specific functions and proceduresHolsapple, C. W. (Ed.). (2004). Handbook of knowledge management.
Berlin: Springer.
Seiler, T. B. (2004). The human foundation of knowledge manage-
ment. In J. Gadner, R. Buber, & L. Richards (Eds.), Organising
knowledge. Methods and case studies (pp. 4359). Hampshire:
Palgrave, Macmillan.
Tergan, S.-O., & Keller, T. (Eds.). (2005). Knowledge and information
visualization. Searching for synergies. Berlin: Springer.
Knowledge Representation
PABLO PIRNAY-DUMMER, DIRK IFENTHALER,
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany
SynonymsInternal representation
DefinitionKnowledge representation is a key concept in cogni-tive science and psychology. To understand this
theoretical term one has to distinguish between
knowledge and its representation. Intelligent
behaviors of a system, natural or artificial, are usually
explained by referring to the systems knowledge.
In other words: The capability of performing intelligenta way that they effect a systematic representation of this
experience or thought as means of understanding it or
of explaining it to others (Seel 1991). This author
describes the function of knowledge representation by
distinguishing three zones: the object zoneW as part of
the physical world, the knowledge zoneK, and the zone
of knowledge representation R (Seel 1991, p. 17).
The interplay between these zones can be depicted
as in Fig. 1.
As shown in Fig. 1, there exist two classes of func-
tions: (1) fin as the function for the internal represen-
tation of the objects of the world (internalization), and
(2) fout as the function for the external re-representa-
tion back to the physical world (externalization).
In cognitive science usually a distinction has been
made between representation schemas which refer to
1690 K Knowledge Representationdeclarative or procedural knowledge. Conceptual graphs
and semantic networks are the most popular schemas
for representing declarative knowledge, whereas proce-
dural knowledge is often represented by means of pro-
duction systems. Both kinds of representational schemas
are discussed in particular entries of this encyclopedia.
Furthermore, frames must be mentioned as a repre-
sentation schema which can be seen as a compound of
semantic networks and productions.
Theoretical BackgroundIn Piagets tradition of semiotic functions (see the entry
on Semiotics and Learning), learning and thinkingare seen as a process of using and operating with sys-
tems of signs (gestures, images, language, or symbols).
These systems enable people to visualize and express
Representation R
Knowledge K
World W
Rep
rese
ntat
ion
Inte
rnaliz
atio
n: f in
Re-
Rep
rese
ntat
ion
Exte
rnaliz
atio
n: f ou
t
r1 r2
k1 k2
w1 w2
Knowledge Representation. Fig. 1 The interplay
between knowledge, its representation, and the physicalworldtheir subjective experiences, ideas, thoughts, and feel-
ings. Consequently, the idea of mental representations
advanced to one of the most significant concepts of
cognitive science.
Types of Knowledge RepresentationAll work on knowledge representation is based on the
hypothesis that people memorize knowledge by means
of specific propositions about the conceptual informa-
tion that is inherent in the issue to be remembered.
Alternatively, there is the conception that information
also can be memorized by means of images which
correspond as analogues to the original perception.
This basic alteration corresponds with semiotics
where, however, a third form of knowledge representa-
tion can be found. Actually, semiologists differentiate
between signs depending on whether they are used as an
index (e.g., smoke as a sign of fire, red as a sign of
danger), as an icon or pictorial sign (e.g., the line draw-
ing of a face, a figure, a vase), or as a symbol (e.g., letters,
numerals, musical notes, mathematical symbols). Cor-
respondingly, cognitive psychologists often differenti-
ate basically between enactive (i.e., activity-based),
iconic, and symbolic types of knowledge representation
(Bruner 1964).
In a process which corresponds to the biological
development of the central nervous system, people
first develop the functions for perception, motor skills,
and coordination. Bruner has thus referred to enactive
representation as a mode of representing past events
through appropriate motor response . . . Such segments
of our environment bicycle riding, tying knots,
aspects of driving get represented in our muscles, so
to speak (Bruner 1964, p. 2). This anchoring of
enactive representation in the neurobiologically oldest
functions of the central nervous system allows this
representation format to maintain its central signifi-
cance as a sign function over the entire lifespan.
The idea of an iconic representation of knowledge
corresponds to the idea of encoding-specificity, which
posits that content frommemory is usually retrieved in
the same format in which it was coded when learned.
Thus, many psychologists (e.g., Larkin and Simon 1987)
use the old adage a picture speaks 1,000 words to
underline the significance of iconic representation and
highlight the fact that it can be influenced by graphics
and images. On the other hand, there are also cognitivescientists (e.g., Pylyshyn 1984) who take up arguments
Knowledge Representation K 1691
Kof the Wurzburg School (from the beginning of the
twentieth century) and reject the idea of representing
knowledge by means of visualization. In the literature,
this controversial argumentation became popular as
imagery debate which dominated the discussion
about knowledge representation in the 1980s. The
iconic representation of knowledge is discussed in par-
ticular entries on imagery.
Finally, knowledge representation occurs by means
of symbols. Psychologists often speak of language as
a special form of symbolic representation. Indeed,
human language is symbolic because it is impossible
to infer the signified (a building in which people live)
from the signifier (e.g., the sounds or letters H + O +
U + S + E). As one and the same signified can be
expressed by any number of different signifiers (e.g.,
maison, Haus, casa), this type of sign is referred
to as arbitrary. By using linguistic means of expression,
humans can visualize ideas, thoughts, and feelings and
describe things that they can only express to a limited
extent in actions or ideational images. It is a distinctive
characteristic of humans that they can deal with their
own experiences through language. Linguistic expres-
sions can trigger ideas (even very vivid ones) which can
even reconstruct facts which cannot be experienced
directly, complex connections, and abstract thought
constructions (e.g., scientific theories). Accordingly,
concepts are probably the most useful form of symbolic
knowledge representation. Concepts are categoriza-
tions as forms of generalized abstractions and require
the application of cognitive operators for class forma-
tion and abstraction. Other symbols which are used for
knowledge representation aremusical notes, digits, and
conventionalized signs. Based on concepts specific rep-
resentation schemas have been developed for
representing declarative knowledge.
Representation Schemas forDeclarative KnowledgeThe idea of declarative knowledge representation
grounds on the assumption that one can consider the
issue of representation largely independent on the
methods of applying knowledge. Declarative knowledge
is considered as a set of facts that can be represented as
data structure. Mylopoulos and Levesque (1984) pro-
vide a taxonomy of representation schemes which
ground on the assumption that the world can be con-sidered as a collection of individuals among whichmanifold relationships exist. The collection of all indi-
viduals and relationships at any one time in any world
constitutes a state, and there can be state transforma-
tions that cause the creation of individuals or that can
change the relationship between them. Depending on
whether the starting point for a representation scheme
is individuals and relationships, or assertions about
states, a distinction can be made between semantic
networks and logical representation schemas.
Good examples for logical representation schemas
are knowledge bases that are constructed on the basis
of first-order logic. Statements about the world domain
to be represented are translated into formulas which
permit conclusions. This form of a representation
schema clearly presupposes appropriate inference
rules which must also be available. Possibly more pop-
ular are semantic networks as form of declarative
knowledge representation. In its most basic form
a semantic network represents knowledge in terms of
a collection of objects (nodes) and binary associations
(directed labeled edges); the former standing for indi-
viduals (or concepts of some sort), and the latter stand-
ing for binary relations over these. According to this
view, a knowledge base (or structure) is a collection of
objects and relations defined over them, and modifica-
tions to the knowledge base occur through the inser-
tion or deletion of objects and the manipulation of
relations (more detailed information can be found in
the entry on Semantic Networks).Another form of declarative knowledge representa-
tion is the Conceptual Dependency Structure from
Schank (1975) that centers on conceptualizations
which attribute cases to actions. Finally, also frames
must be mentioned in the context of declarative knowl-
edge representation although frames aim at a com-
bination of declarative and procedural knowledge
representation.
Representation Schemas forProcedural KnowledgeDepending onwhether the starting point for a represen-
tation schema is state transformations the application
of logical representation schemas and semantic net-
works is limited and demands for another type of
representation. Clearly, the most popular procedural
representation schemas operate with productions and
production systems. They are described in detail in theentry on Production Systems and Operator Schemas.
Brachman, R. J., & Levesque, H. L. (Eds.). (1985). Readings in knowl-
Ifenthaler, D., Pirnay-Dummer, P., & Seel, N. M. (Eds.). (2010).
Computer-based diagnostics and systematic analysis of knowledge.
P. Winston (Ed.), The psychology of computer vision. New York:
Pylyshyn, Z. (1984). Computation and cognition: Toward a foundation
for cognitive science. Cambridge, MA: MIT Press.
1692 K Knowledge Representation and Reasoning and Learning
Important Scientific Research andOpen QuestionsKnowledge representation is not only a central problem
in cognitive science but also in Artificial Intelligence
due to the fact that intelligent systems need the
availability of expert knowledge along with associated
knowledge handling facilities. Since the emergence of
cognitive science countless studies have focused on the
use of representational schemas by intelligent systems,
natural or artificial. Actually, the problem of knowledge
representation is at the center of most publications in
the field of cognitive psychology (e.g., Markman 1998)
and Artificial Intelligence (e.g., Brachman and
Levesque 1985; Davis et al. 1993).
With regard to cognitive psychology and its
focus on human knowledge representation the most
important result of research consists in the observa-
tion that people are able to use different forms
of representation of memorized information. People
can either recall an appropriate form of representa-
tion from memory or transform memorized informa-
tion in an appropriate form of representation in
dependence on situational demands. However,
because it is not possible to assess directly internal
representations of knowledge one of the most impor-
tant issues of research on knowledge representation is
concerned with reliable and valid measurements of
declarative and procedural knowledge (see for an
overview Ifenthaler et al. 2010). Here, it is argued
that different types of knowledge require different
types of representations.
Cross-ReferencesACT (Adaptice Control of Thought)Activity- and Taxonomy-Based Knowledge
Representation
Conceptual Clustering
Imagery and LearningKnowledge OrganizationMental Imagery
Mental Representation Pictorial Representation and Learning
Production Systems and Operator Schemas inProcedural Learning
Representation, Presentation and ConceptualSchemas
Semantic Networks
Wurzburg SchoolSchank, R. (1975). Conceptual information processing. New York:
Elsevier.
Seel, N. M. (1991). Weltwissen und mentale Modelle. Gottingen:
Hogrefe [World knowledge and mental models].
Knowledge Representation andReasoning and Learning
Semantic Technologies and Learning
Knowledge Restructuring
Conceptual Change
Knowledge Storage
Routinization of Learning
Knowledge StructureMcGraw-Hill.
Mylopoulos, J., & Levesque, H. J. (1984). An overview of knowledge
representation. In M. L. Brodie, J. Mylopoulos, & J. W. Schmidt
(Eds.), On conceptual modelling. Perspectives from artificial intel-
ligence, databases, and programming languages (pp. 317).
New York: Springer.New York: Springer.
Larkin, J. H., & Simon, H. A. (1987). Why a diagram is (sometimes)
worth ten thousand words. Cognitive Science, 11, 6599.
Markman, A. B. (1998).Knowledge representation. Mahwah: Erlbaum.
Minsky, M. (1975). A framework for representing knowledge. Inedge representation. Los Altos: Kaufmann.
Bruner, J. S. (1964). The course of cognitive growth. The American
Psychologist, 19, 116.
Davis, R., Shrobe, H., & Szolovits, P. (1993). What is a knowledge
representation? Artificial Intelligence Magazine, 14(1), 1733.ReferencesAnderson, J. R. (1983). The architecture of cognition. Cambridge, MA:
Harvard University Press.Concept Similarity in Multidisciplinary Learning
Koffka, Kurt (18871941)
Life Dates
tion with Max Wertheimer and Wolfgang Kohler in
developing the foundation of Gestalt psychology. In
ception but also in many philosophical issues as well
LearningKoffka believed that most of the childs early learning is
Kohlberg, Lawrence (19271987) K 1693
K1924, Koffka came to the Unites States in order to
serve as visiting professor at Cornell University and
the University of Wisconsin. Then he accepted a posi-
tion at Smith College in Northampton, Massachusetts,Kurt Koffka was born in Berlin in 1886. In 19041905,
he studied at the University of Edinburgh in Scotland,
where he met British scholars, and studied in English.
He completed a Ph.D. at the University of Berlin in
1908 under the supervision of Carl Stumpf. Koffka
served as research assistant at the universities of
Wurzburg and Frankfurt before moving to the Univer-
sity of Giessen, where he served as professor until 1924.
When Koffka was in Frankfurt he began his collabora-NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, GermanyKnowledge-Based Learning
Analytic LearningKnowledge Structuring
Knowledge Organization
Knowledge Transfer
Process of using knowledge from a learning situation in
a situation outside of the initial learning, e.g., applying
school mathematics to shopping tasks.
Cross-ReferencesTransfer of Learningwhere he remained until his death in 1941.sensorimotor learning which occurs after a consequence.
For instance, a child who touches a hot stove will learn
not to touch it again. Furthermore, Koffka believed that a
lot of learning also occurs by imitation, which he saw as
a natural occurrence. However, the highest type of learn-
ing is ideational learning, which makes use of language.
Koffka notes that an important time in childrens devel-
opment is when they understand that objects have names.
Cross-ReferencesHistory of the Sciences of Learning
Kohler, WolfgangWertheimer, Max
ReferencesKoffka, K. (1924). The growth of the mind. London: Routledge.
Reprinted 1999.
Kohlberg, Lawrence(19271987)
DAMIAN GRACE, MICHAEL JACKSON
Department of Government and International
Relations, The University of Sydney, Sydney, Australia
Life DatesLawrence Kohlberg (October 25, 1927January 19,
1987) was an American psychologist whose theory of
the development of moral reasoning made him one ofas in development, learning and thinking. Thus, he
published not only an important book on child psy-
chology in 1921, but also a complete and systematic
overview on the principles of Gestalt in 1935. In con-
sequence, Koffka focused on developmental aspects of
learning in accordance with Gestalt principles.
Contribution(s) to the Field ofTheoretical BackgroundKoffka was one of the founders of Gestalt theory (along
with Wertheimer and Kohler). Like Wertheimer and
Kohler, he was not only interested in research on per-the most significant psychological researchers of the
1694 K Kohlberg, Lawrence (19271987)
twentieth century. Kohlberg was born in New York
City, attended a private school, and worked for a time
in the merchant marine. In this capacity he served the
Zionist cause by smuggling Jewish refugees into Pales-
tine. In 1948 he was admitted to The University of
Chicago, which awarded him a bachelors degree after
just 1 year. He enrolled in graduate studies at Chicago
and began his academic career at Yale University
(19561961). From 1962 to 1968, he was a professor
at Chicago and thereafter at Harvard. In 1971,
Kohlberg contracted a debilitating parasitic disease
while working in Belize. For the rest of his life he
suffered from its effects, including depression. In
1987, he committed suicide by drowning at Winthrop,
just outside Boston (Walsh 2000).
Contribution(s) to the Field ofLearningKohlberg began investigating themoral development of
children in his doctoral dissertation (1958), and con-
tinued this line of research for the next 30 years. His
many publications include Essays on Moral Develop-
ment: Vol. 1. The Philosophy of Moral Development
(vol. 1, 1981; vol. 2, 1984) and Child Psychology and
Childhood Education: A Cognitive-Developmental View
(1987). A Web of Knowledge search in 2010 returned
65 publications in journals, including the American
Journal of Education, American Journal of Mental Defi-
ciency, American Journal of Orthospyschiatry, American
Sociological Review, Child development, Human Devel-
opment, and Zygon. A parallel search returned more
than 750 citations of these works. In short, his work
was widely disseminated and has been very influential.
A search of dissertation titles and abstracts identified
more than 900 theses that refer to Kohlberg in the title
or abstract.
Kohlbergs initial research was to interview 72 white
working and middle class Chicago boys aged 10, 13,
and 16, and present them with this moral dilemma.
A man named Heinz cannot afford to buy a uniquely
efficacious drug to save his dying wife. The chemist
who discovered the drug demands an extortionate
price for it. Heinz cannot even borrow so large a sum,
so he steals the drug. Did he do the right or wrong thing
in stealing from the chemist?
Based on his subjects responses to this and other
dilemmas, Kohlberg built a classification of stages ofmoral development. The characteristics of these stagesare, first that each is qualitatively different from the
others. Secondly, that they are structured wholes,
that is, responses to one dilemma will be consistent
with responses to others, reflecting patterns of thought
peculiar to each stage. Thirdly, children progress
through the stages in an invariant sequence, neither
skipping stages nor regressing. Fourthly, the stages are
integrated hierarchically, that is, learning at earlier stages
is not lost as subjects move into the later ones. Finally,
the stages are cross-cultural universals, found in all
societies. Kohlberg means by this not that all cultures
have the same beliefs but that the development of moral
reasoning occurs in all cultures according to his stages.
Development might not progress through all the stages,
for stages 5 and 6 demonstrate quite a high degree of
detachment and moral self-direction. Indeed, by 1976,
Kohlberg had stopped using stage 6, first because he
doubted the efficacy of his dilemmas in distinguishing
between stages 5 and 6; and secondly because he
doubted that many subjects reasoned consistently at
Stage 6 (Crain 2005, pp. 158, 165; Cortese 1990 p. 21).
Kohlberg placed his six stages into three levels. Level
1 contains two stages of Pre-conventional Morality. At
this level, children do not yet appreciate the social
nature of morality. At Stage 1, children regard morality
as a set of rules laid down by authority. That is, morality
is an externally imposed set of rules and deviation from
them brings punishment. In the Heinz case, the rules
dictate that it is wrong to steal and children asked
about whether Heinz would be right to steal the drug
typically respond, Its against the law, and youll get
punished. At this stage, egocentrism prevails and other
perspectives are not part of the childs way of thinking
(Crain 2005, p. 154). At Stage 2, moral reciprocity
emerges. Right is what parties reciprocally agree to be
fair, and morality is to this extent relativized. Children
can appreciate that Heinz might have sufficient reason
to steal the drug, but they can also see the chemists side
of things. Self-interest has emerged as a criterion of
right, though tempered by a sense of the chemists
unfairness in demanding too high a price for his drug.
At Level 2, Conventional Morality, children are
entering adolescence and have a grasp of morality as
necessary for social cohesion. The centrality of rules
gives way to the consideration of character and a con-
cern for the welfare of others. Stage 3 reasoning reflects
empathic relationships with family and friends and theobligations these entail. Children at this stage will
act on real issues, rather than just theorize about hypo-
thetical dilemmas. The Cluster Schools did not survive
Kohlberg, but they did inspire others to embark on
similar innovative programs (Walsh 2000).
Criticisms of the TheoryKohlbergs research looked at reasoning rather than
behavior, and this presents a difficulty. Someone
could offer a convincing discussion of ethical conduct
while behaving unethically. Equally, an ethical actor
might not be able or willing to articulate the reasoning
that explains and justifies the conduct.
Probably the most famous critique of Kohlberg was
that of Carol Gilligan, his colleague at Harvard. Her
book, In a Different Voice (1982) had wide influence in
education, gender studies, philosophy, and political
science. Gilligans argument was that Kohlberg had
Kohlberg, Lawrence (19271987) K 1695
Kconsider motives, such as the chemists selfishness and
greed, and Heinzs feelings of love and concern in
stealing the drug. At Stage 4, the moral stage is broad-
ened beyond family and friends to include society. Now
there is a recognition that if everyone acted as Heinz did,
society would fall apart. There is recognition that Heinz
has good motives, but that this does not override the
obligation to consider society as a whole and its reliance
on obedience to the law. Responses at this stage resemble
those at Stage 1, but children at Stage 4 can give reasons
for their views, such as harm to society. At Stage, 1
stealing is just wrong (Crain 2005, pp. 155157).
At Level 3, Post-conventional Morality, Stage 5,
a person moves beyond concern for social cohesion to
questions about the nature of society and the principles
underlying its conventions. What might a good society,
irrespective of its setting, look like? It turns out to look
something like John Rawlss just contractarian political
association enshrining basic rights and using democratic
procedures (Crain 2005, pp. 157158). Stage 5 respon-
dents take a critical position on social conventions,
distinguishing law and morality in the Heinz case.
A good society would protect the chemists property
rights but be sensitive to Heinzs wifes right to live.
Heinz would be justified morally in saving his wife,
and legally a judge should take account of this and
punish him but lightly (Crain 2005, p. 157). Stage 6
reasoners would exhibit a Rawlsian appreciation of the
principled requirements for social agreement in
a democracy and especially the centrality of justice con-
ceived as fairness. This last stage follows logically from
Stage 5, but as indicated above, it lacks the empirical
support of the earlier stages.
Table 1 indicates Kohlbergs view of moral growth,
but note that ages for each stage are correlations.
Kohlbergs stages of moral development made
a considerable impact. In addition to his own publica-
tions and their influence on other researchers, the
stages also entered into the discourse of teaching in
psychology, education, and applied ethics. Publications
referring to Kohlberg in 2010 examined reasoning
across a range of fields beyond education, including
business ethics, nursing ethics, and more.
Important Scientific Research andOpen QuestionsKohlberg recognized that his theory had practicalimplications. If the attainment of the higher stages ofmoral development is promoted by appropriate social
interactions, how could such interactions be facilitated?
His response was the so-called Cluster Schools that
afforded pupils the opportunity to discuss moral ques-
tions with more mature thinkers and to reflect on their
experience. Kohlbergs involvement in the Cluster
School movement began in 1974, when he was involved
in planning the first, and he then applied his vision to
other schools and a womens prison. He envisaged
small, democratic communities even schools
within schools where students and teachers partici-
pated equally in establishing norms. This participation
offered developing moral reasoners, the opportunity to
Kohlberg, Lawrence (19271987). Table 1 Kohlbergs
stages of moral development
Level Age Stage Description
Level 3: Post-conventional
Maturity Stage 6 Kantianuniversality
Maturity Stage 5 Social contract
Level 2:Conventional
Adulthood Stage 4 Obedience toauthority
Teens Stage 3 Good boy/girl
Level 1: Pre-conventional
512 Stage2 Instrumental
05 years Stage 1 Punishmentavoidance
While most individuals develop over the years to sat 4 above, only
a few continue on the stages 5 and even fewer to 6.studied only boys and that aspects of the reasoning of
Kohlbergs highest level of moral development empha-
Kohlbergs stages 5 and 6, and represented a different,
equally important moral voice.
1696 K Kohler, Wolfgang (18871967)Kohler, Wolfgang (18871967)
NORBERT M. SEEL
Department of Education, University of Freiburg,
Freiburg, Germany
Life DataWolfgang Kohler was born in 1887 in Reval, Estonia,
and grew up in Wolfenbuttel, Germany. Kohler studied
philosophy, science, and psychology at the Universities
of Tubingen, Bonn, and Berlin. One of his teachers inCross-ReferencesMoral Learning
Piagets Learning Theory Piaget, Jean (18961980)
ReferencesCortese, A. (1990). Ethnic ethics: The restructuring of moral theory.
Albany: SUNY Press.
Crain, W. C. (2005). Theories of development (5th ed.). Upper Saddle
River: Pearson Prentice-Hall.
Gilligan, C. (1982). In a different voice. Cambridge: Harvard Univer-
sity Press.
Walsh, C. (2000). Reconstructing Larry: Assessing the legacy of
Lawrence Kohlberg. Ed. Magazine, (Harvard Graduate
School of Education). http://www.gse.harvard.edu/news/features/
larry10012000_page1.html. Accessed 24 Oct 2010.sized rules, principles, rights, duties, and the like. In
contrast, Gilligan found that women faced with very