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