Network of Excellence in Professional
Learning
PROLEARN
European Commission Sixth Framework Project (IST 507310)
Deliverable
5.3
Common Paper on Conceptual Learning Management
Editor
Work Package
Ambjörn Naeve
WP5
Status
Version 1.0
Date
2005-06-30
The PROLEARN Consortium
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
Universität Hannover, Learning Lab Lower Saxony (L3S), Germany
Deutsches Forschungszentrum für Künstliche Intelligenz GmbH (DFKI), Germany
Open University (OU), UK
Katholieke Universiteit Leuven (K.U.Leuven) / ARIADNE Foundation, Belgium
Fraunhofer-Gesellschaft zur Förderung der angewandten Forschung e.V. (FHG)
Wirtschaftsuniversität Wien (WUW), Austria
Universität für Bodenkultur, Zentrum für Soziale Innovation (CSI), Austria
École Polytechnique Fédérale de Lausanne (EPFL), Switzerland
Eigenössische Technische Hochschule Zürich (ETHZ), Switzerland
Politecnico di Milano (POLIMI), Italy
Jo ef Stefan Institute (JSI), Slovenia
Universidad Polictécnica de Madrid (UPM), Spain
Kungl. Tekniska Högskolan (KTH), Sweden
National Centre for Scientific Research “Demokritos” (NCSR), Greece
Institut National des Télécommunications (INT), France
Hautes Etudes Commerciales (HEC), France
Technische Universiteit Eindhoven (TU/e), Netherlands
Rheinisch-Westfälische Technische Hochschule Aachen (RWTH), Germany
Helsinki University of Technology (HUT), Finland
Document Control
Title: A Conceptual Modeling Approach to Studying the
Learning Process
Author/Editor:
Ambjörn Naeve
E-mail:
amb@nada.kth.se
AMENDMENT HISTORY
Version
Date
Author/Editor
Description/Comments
0.3
2005-01-07
Ambjörn Naeve
Draft status (final version due month 18)
0.6
2005-05-22
Ambjörn Naeve
Draft status (final version due month 18)
0.7
2005-05-31
Ambjörn Naeve
Draft status (final version due month 18)
0.8
2005-06-05
Ambjörn Naeve
Draft status (final version due month 18)
0.9
2005-06-16
Ambjörn Naeve
Draft status (final version due month 18)
1.0
2005-06-30
Ambjörn Naeve
Final version
Contributors
Name
Institution
Ambjörn Naeve
Royal Institute of Technology (KTH)
Mikael Nilsson
Royal Institute of Technology (KTH)
Matthias Palmér
Royal Institute of Technology (KTH)
Richard Wessblad
Royal Institute of Technology (KTH)
Mia Lindegren
Uppsala Learning Lab / Uppsala University
Miltiadis Lytras
Computer Science Dept / University of Patras
Nikos Korfiatis
Computer Science Dept / University of Patras
Bernd Simon
Wirtschaftsuniversität Wien
Fridolin Wild
Wirtschaftsuniversität Wien
Barbara Kieslinger
Universität für Bodenkultur
Pertti Yli-Louma
University of Oulu
Milos Kravcik
Fraunhofer FIT, Sankt Augustin
Vana Kamtsiou
NCSR Demokritos
Dimitra Pappa
NCSR Demokritos
Legal Notices
The information in this document is subject to change without notice.
The Members of the PROLEARN Consortium make no warranty of any kind with
regard to this document, including, but not limited to, the implied warranties of
merchantability and fitness for a particular purpose. The Members of the PROLEARN
Consortium shall not be held liable for errors contained herein or direct, indirect,
special, incidental or consequential damages in connection with the furnishing,
performance, or use of this material.
A CONCEPTUAL MODELING APPROACH
TO STUDYING THE LEARNING PROCESS
Ambjörn Naeve, Pertti Yli-Luoma, Milos Kravcik, Miltiadis Lytras, Bernd Simon, Mia
Lindegren, Mikael Nilsson, Matthias Palmér, Nikos Korfiatis, Fridolin Wild, Richard
Wessblad, Vana Kamtsiou, Dimitra Pappa, Barbara Kieslinger
Version 1.0 (2005-06-30)
1
TABLE OF CONTENT
1.
INTRODUCTION ................................................................................................................................................ 3
1.1.
1.2.
2.
SCOPE OF THE PAPER ........................................................................................................................................ 3
STRUCTURE OF THE PAPER ............................................................................................................................... 3
SOME PERSPECTIVES ON KNOWLEDGE AND LEARNING............................................................... 4
2.1.
2.2.
2.3.
3.
KNOWLEDGE TRANSMISSION VERSUS KNOWLEDGE CREATION ..................................................................... 6
KNOWLEDGE PUSHING VERSUS KNOWLEDGE PULLING .................................................................................. 6
FORMAL VERSUS INFORMAL LEARNING PROCESSES ....................................................................................... 7
ASSEMBLY LINE MODELING OF THE LEARNING PROCESS .......................................................... 8
3.1.
3.2.
3.3.
4.
THE PROCESS/PEDAGOGY/TOOLS ABSTRACT MODEL ...................................................................................... 8
INSTANTIATING THE PROCESS/PEDAGOGY/TOOLS ABSTRACT MODEL ........................................................... 8
REFINING THE PROCESS PART OF THE MODEL ............................................................................................... 10
THE SECI KNOWLEDGE CREATION PROCESS ................................................................................... 11
4.1.
4.2.
4.3.
4.4.
THE SECI MODES OF KNOWLEDGE CONVERSION ......................................................................................... 11
BA – A PLACE FOR INTERACTIVE KNOWLEDGE CREATION ........................................................................... 12
KNOWLEDGE ASSETS ..................................................................................................................................... 13
LEADING THE SECI KNOWLEDGE-CREATING PROCESS ................................................................................ 14
5.
THE SECI PROCESS FRAMEWORK.......................................................................................................... 14
6.
EMPIRICALLY VALIDATED PEDAGOGICAL SUPPORT .................................................................. 16
6.1.
6.2.
6.3.
6.4.
6.5.
7.
INTRODUCTION AND OVERVIEW .................................................................................................................... 16
SOCIALIZATION PROCESS ............................................................................................................................... 17
EXTERNALIZATION PROCESS ......................................................................................................................... 18
COMBINATION PROCESS ................................................................................................................................. 18
INTERNALIZATION PROCESS .......................................................................................................................... 19
CONCLUSIONS AND FUTURE WORK ...................................................................................................... 20
7.1.
7.2.
APPLYING THE FRAMEWORK MODELS TO PROLEARN................................................................................... 20
FUTURE RESEARCH ISSUES............................................................................................................................. 21
8.
ACKNOWLEDGEMENTS .............................................................................................................................. 22
9.
REFERENCES ................................................................................................................................................... 22
APPENDIX
10.
CONCEPTUAL MODELING........................................................................................................................ 28
10.1.
10.2.
10.3.
10.4.
10.5.
10.6.
THE CONCEPT ’ CONCEPT’ ............................................................................................................................ 28
PROPERTIES OF THE CONCEPT ‘CONCEPT’................................................................................................... 28
UML – A GLOBAL MODELING LANGUAGE .................................................................................................. 28
THE UNIFIED LANGUAGE MODELING TECHNIQUE ..................................................................................... 28
THE MODELING THEORY OF DAVID HESTENES ........................................................................................... 29
EDUCATIONAL MODELING LANGUAGES..................................................................................................... 30
11.
PROCESS MODELING: THE GOAP APPROACH................................................................................. 31
12.
THE ONSCAIL NETWORK ......................................................................................................................... 32
12.1.
12.2.
13.
ENABLING SEMANTIC COOPERATION AMONG CONTENT PROVIDERS ......................................................... 32
MODELING THE CFL COURSE DEVELOPMENT PROCESS ............................................................................. 33
THE OPEN-LMS ASSESSMENT PROJECT............................................................................................. 35
13.1.
13.2.
13.3.
13.4.
OBJECTIVES .................................................................................................................................................. 35
SCOPE ........................................................................................................................................................... 35
PROCESS AND METHOD ................................................................................................................................ 35
PRACTICAL USER TESTING ........................................................................................................................... 36
2
1.
INTRODUCTION
In our emerging knowledge society, a firm understanding of the interplay between the management of
knowledge and learning is of strategic importance in order to create and maintain effective learning processes in
a large variety of non-traditional learning situations (Lytras, Naeve and Pouloudi [36]). For example, as
described by Grace and Butler in [18], Zuboff [87] argues that learning, integration and communication become
key to leveraging employee knowledge. Accordingly, managers must “switch from being drivers of people to
being drivers of learning”. Argyris and Schön [3] point out that “there is a virtual consensus that we are all
subject to a learning imperative, and in the academic as well as the practical world, organizational learning has
become an idea in good currency.”
However, as Grace and Butler observe, “learning is a complex phenomenon and the concept of learning within
organizations has numerous dimensions, making it even more complicated than individual learning. Multiple
levels of learning have been distinguished, including individual learning, group learning and organizational
level learning and several processes of learning within an organizational context have been differentiated.
Among these are knowledge creation, knowledge acquisition, information distribution, information
interpretation and organizational memory (Fiol and Lyles [16], Levitt and March [33], Nonaka [53]).”
1.1. Scope of the paper
In this paper we present a conceptual approach to the study and analysis of learning processes with emphasis on
the knowledge-creating types of learning processes that often occur in workplace learning.1 We present a
framework (abstract model) for categorizing knowledge-creating learning processes based on process modeling
and Nonaka’s (SECI spiral) theory of knowledge creation ([53], [54], [55]). Our framework makes use of
assembly-line-style process modeling in order to show how different parts of a learning process are supported by
different pedagogical aspects and tools. It also uses the Unified Language Modeling (ULM) technique [43] in
order to improve the conceptual overview and increase the visibility and clarity of the structures involved.
By bringing together the three aspects of SECI knowledge creation, process modeling and ULM, the paper in
fact describes the present “state of the art” of concept-based e-learning within the Prolearn consortium. Although
the validity of many aspects of our SECI process framework is empirically supported by pedagogical research
(section 6), we must stress that the framework in its entirety has not yet been applied to the study and
classification of learning processes. However, we plan to initiate such studies within the near future, with a
special focus on learning processes in the workplace. In fact, our framework suggests a method for describing
such learning processes and how they are supported by different pedagogical aspects and tools. Applying this
method to the description of concrete learning processes will provide an empirical basis from which we can find
out what works well and why it works well, which will lead to more effective ways of creating and managing
learning processes in general.
This focus reflects our response to the main criticism from the first Prolearn review in March 2005,
which stated that the consortium should strengthen its involvement with learning at the workplace. Because of
this critique, the originally planned focus of this paper on LMS brokerage and interoperability issues has been
changed, and the corresponding discussion (the openLMS project) has been placed in an appendix.
1.2. Structure of the paper
In section 2 we present a general discussion on knowledge and learning and introduce a definition of knowledge
as consisting of efficient fantasies. In section 3 we apply assembly-line process modeling [14] in order to outline
a learning process framework (abstract model) that indicates how the different parts of a learning process are
motivated by different pedagogical aspects and supported by different tools. Instantiating the abstract model at
the specific level, we should be able to figure out how each part of a specific learning process is motivated by
specific pedagogical or didactical aspects, and how these aspects are supported by the specific tools that are used
in the corresponding part of the learning process. At the abstract level, our model considers a learning process as
orchestrated by pedagogical/didactical aspects, which are supported by various tools. As an illustration, we
1
In contrast to e.g., traditional academic learning (courses), where the knowledge exists prior to the execution of the learning process.
3
instantiate this abstract process/pedagogy/tools model with a specific example of pedagogical aspects and
supporting tools. This example is based on the work of the KMR group [88] at KTH as presented in [46].
In section 4 we review (part of) the unified model of dynamic knowledge creation as presented by Nonaka,
Toyama and Konno in [55], and apply both process modeling and UML-style conceptual modeling [43] to it in
order to construct an abstract model of the most important parts of the knowledge creation process: Four
different types of knowledge-conversions (SECI) supported by four different types of ba (‘interaction spaces’)
and resulting in four different types of knowledge assets.
In section 5 we combine process modeling (section 3) with SECI knowledge conversion (section 4) in order to
formulate the SECI process framework, which is an abstract model for the study and classification of knowledgecreating learning processes.
In section 6 we introduce some support for our learning process model from pedagogical research. The empirical
results of Yli-Luoma and others are discussed and related to the SECI process framework, showing how they
underscore and validate the different parts of the model.
In section 7 (conclusions and future work) we indicate how the SECI process framework can be applied to the
knowledge creating processes of the Prolearn NoE – especially roadmapping and knowledge work management.
The section concludes with a discussion on some general research issues that are raised by the models developed
in this paper, and which we will attempt to address in the future.
In order to keep a clear focus, a few sections have been placed in an appendix. Sections 10 and 11 describe
some basic ideas of conceptual modeling and process modeling, whereas sections 12 and 13 describe related
ongoing work within two Swedish projects, the Open Network for Semantic Collaboration Around Informal
Learning (ONSCAIL) and the openLMS assessment project of open source based L(C)MSes that is carried out
for the Swedish Netuniversity under the coordination of Uppsala Learning Lab [99].
2.
SOME PERSPECTIVES ON KNOWLEDGE AND LEARNING
Since the time of ancient Greece, the philosophical discussions and debates on the nature of knowledge and
learning have been recurrent, and several schools of thought have made substantial contributions. As pointed out
by Nonaka, Toyama and Konno in [55], in traditional Western epistemology 2 truthfulness is the essential
attribute of knowledge. It is the absolute, static and non-human view of knowledge, and it fails to address its
relative, dynamic and humanistic dimensions.
According to Sperber and Wilson “Human beings are efficient information-processing devices. This is their most
obvious aspect as a species” ([71], p. 45) . This quote from one of the classics of cognitive psychology provides a
good example of the Western emphasis on explicit knowledge, as opposed to tacit knowledge, a term which was
introduced by Michel Polanyi in [62]. The term tacit knowledge refers to the implicit and silent (pre-logical)
knowledge that we all carry within ourselves, and which Polanyi expressed as “we can know more than we can
tell ([62], p. 4).
In his dialogue seminars, Bo Göranzon [19] has introduced the following different types of knowledge and
described useful methods for their exploration:
• Explicit knowledge consists of statements, which can be explored through standardized surveys (quantitative
studies).
• Implicit knowledge consists of statements that are harder to directly recall, and which require more of reflection
and introspection. Common ways of exploring implicit knowledge is by deep interviews and ethnographic
methods, which are both qualitative in nature and therefore require substantial elements of interpretation.
• Silent knowledge consists of knowledge that (for logical reasons) is not available in the form of statements, but
which is primarily expressed in the form of practical actions. It can also be studied through deep interviews and
ethnographical methods.
• Sub-conscious knowledge – or feelings – that can be explored with psychological methods.
In his famous taxonomy of learning, Bloom [8] identifies 6 different levels of knowledge in the cognitive
domain. They are shown in Figure 1 (slightly revised by Anderson and Kratwohl [1]). The truncated pyramid
2
The theory of knowledge.
4
indicates that each level builds on the ones below it. A similar analysis for the affective domain3 has been carried
out by Kratwohl, Bloom and Marsia [32] and in the psycho-motor domain by Dave [12]. These domains all
represent important dimensions of learning, which need to be taken into account in a full analysis, but they will
not concern us further here.
Figure 1.
A taxonomy of learning in the cognitive, affective and psycho-motor domains.
4
Different perspectives on learning will be taken up in section 6, where we will discuss some contributions from
e.g., Vygotsky, Kolb, Piaget, Ravenscroft, Keeves, Sweller and Hestenes.
In [43], Naeve defines (mental) knowledge as consisting of efficient fantasies5 and describes (mental) learning as
based on inspiring fantasies. Each fantasy has a context, a purpose and a target group, and it is only when we
have described how we are going to measure the efficiency of our fantasies - within the given context, with the
given purpose, and against the given target group - that we can speak of knowledge in a way that can be
validated.
Figure 2.
Learning and Knowledge Management perspectives of the Learning Process:
Transforming inspiring fantasies into efficient fantasies.
From this perspective, management of the learning process is concerned with exposing the learner to inspiring
fantasies and assisting her/him in transforming them into efficient fantasies. This involves two complementary
aspects, learning management, which is people-oriented and focuses on learning as a process, and knowledge
management, which (traditionally) is technology-oriented and focuses on knowledge as a resource. See e.g., [18]
for a more thorough discussion on the attempts of modern LMSes to bridge this traditional gap and
accommodate both of these important perspectives.
3
which is concerned with the perception of value issues.
The figure is based on Atherton [4].
5
As opposed to muscular knowledge, which he defines as “efficient reflexes”. The word “fantasy” is used instead of the synonymous word
“conceptualisation” in order to emphasise that the conceptual structures are constructed from within.
4
5
2.1. Knowledge transmission versus knowledge creation
Here we will introduce a distinction between knowledge-transmitting and knowledge-creating learning
processes, a distinction that separates formal and informal learning. In a knowledge-transmitting type of learning
process, the desired knowledge (as expressed e.g., in the curriculum of a traditional course) exists prior to the
execution of the learning process, whereas in an informal type of learning process, (substantial parts of) the
desired knowledge is often created during the execution of the learning process itself.
Note that our choice of terms does not imply that we believe that transmitted knowledge can be received as
knowledge6. In contrast, we share the constructivist belief that knowledge has to be constructed by each separate
individual, preferably within a collaborative learning process that involves interacting with others. Hence, we are
well aware that knowledge creation occurs during the execution of any type of learning process. However, in the
knowledge transmission type of learning process, this knowledge creation takes place only among the learners7,
since it is driven by a fixed curriculum, which exists prior to the course.
Figure 3.
Kowledge-transmitting versus knowledge-creating learning processes.
As depicted in Figure 3, a knowledge-transmitting type of learning process leads to a knowledge-simulating type
of behaviour, where the learners are trying to figure out the right answers, whereas a knowledge-creating
learning type of learning process leads to knowledge-stimulating type of behaviour, where the learners are trying
to figure out the right questions. The reader is referred to [39] and [40] for further discussions on these matters.
2.2. Knowledge pushing versus knowledge pulling
In [43] Naeve discusses the demands for flexible and personalizable learning in terms of the distinction between
knowledge-pushing and knowledge pulling types of learning processes. The traditional learning processes are
based on teacher-centric, curriculum-oriented, knowledge-push. The new demands on learning are largely
concerned with a shift along all of these dimensions in order to support more learner-centric, interest-oriented,
and knowledge pulling types of learning processes. In [46], it is shown that the infrastructure, frameworks and
tools of the KMR group are designed to encourage and support the latter type of learning processes. Since we
will make use of these contributions in our modeling examples, we briefly introduce them here for the
convenience of the reader.
Over the last few years, members of the WGLN [94] and Prolearn [96] networks have made numerous
contributions towards a Public Knowledge and Learning Management Environment ([46], [48], [49], [79]) based
on open source, open ICT standards and the technology for the emerging next generation Internet – the so-called
Semantic Web [107]. In this paper, we will instantiate our process/people/tools framework (meta-model) with
examples from this work.
The PKLME is structured in the form of a Knowledge Manifold, which is an information architecture that
consists of a number of linked conceptual information landscapes (context-maps), whose concepts can be filled
with content, and where one can navigate, search for, annotate and present all kinds of electronically stored
information [39], [40], [41]. The PKLME also includes:
6
This is the ghist of the so called “transmission theory” of knowledge, which no one seems to believe in these days.
And not among the other stakeholders, such as e.g., teachers or administrators. Of course, when we are dealing with changing the structure
of the learning process itself - e.g. by changing a course - new knowledge is created by all stakeholders.
7
6
• The Edutella infrastructure: A democratic (peer-to-peer) network infrastructure for search and retrieval of
information about learning resources [48], [52].
• The SCAM framework: (Standardized Contextualized Access to Metadata): A framework that helps
applications to store and share information about learning resources [59].
• The SHAME framework (Standardized Hyper-Adaptable Metadata Editor): An editor framework that supports
an evolving annotation process of learning resources in a way that enables the growth of an “ecosystem” of
quality metadata [51], [91].
• The Formulator (or SHAMEditorEditor): A tool for editing metadata editors that is built on top of the SHAME
framework.
• The Confolio network: A network of conceptual electronic portfolios (built on top of SCAM, SHAME and
Edutella) that supports collaborative and reflective learning techniques.
• The Conzilla concept browser: A knowledge management tool that supports the construction, navigation,
annotation and presentation of the information in a knowledge manifold [40], [42].
• The VWE composer: An environment for composing learning resources and building customized learning
modules.
The knowledge roles of a Knowledge Manifold
The KM architecture supports the following seven different knowledge roles [41]:
• the knowledge cartographer, who constructs and maintains context-maps.
• the knowledge librarian, who fills context maps with content-components.
• the knowledge composer, who constructs customized learning modules.
• the knowledge coach, who cultivates questions.
• the knowledge preacher, who provides live answers.
• the knowledge plumber, who directs questions to appropriate preachers.
• the knowledge mentor, who is a role model and supports self-reflection.
2.3. Formal versus informal learning processes
Of course, the knowledge-dimensions of creating ---- transmitting and pushing ---- pulling are not independent
of each other, but in fact highly correlated. For example, it is obvious that a knowledge-transmitting learning
process must push the knowledge items on its curriculum in order to be successful. As shown in Figure 4, this is
characteristic of formal learning processes, such as e.g., traditional academic courses. This type of learning
process leads to an imitative learning behaviour, where the learners are rewarded for figuring out the right
answers.
Figure 4.
Formal versus informal learning
In contrast, a knowledge-creating learning process creates its own curriculum (more or less) at runtime, i.e.,
during the time that it executes, which requires more of a knowledge-pulling strategy in order to be effective.
This is characteristic of informal learning, which occurs e.g., in academic research as well as in many forms of
workplace learning - especially among the knowledge-workers of companies that produce knowledge-intensive
products and/or services. This type of learning process leads to an explorative learning behaviour, where the
learners are rewarded for figuring out fruitful questions (and, of course, also for providing answers to them).
In section 4 we will explain how knowledge-creating learning processes can be effectively described by making
use of Nonaka’s dynamic theory of knowledge creation [53] and the unified model of Nonaka, Toyama and
Konno [55]. By combining this theory with process modeling we will arrive at an abstract model that can serve
as a basis for the analysis and classification of many professional learning processes.
7
3.
ASSEMBLY LINE MODELING OF THE LEARNING PROCESS
Process modeling presents powerful ways of describing dynamic interactions – ways that seem to have found
little use for educational modeling within the TEL community [65].8 Here we will make use of a type of process
modeling described e.g., in [14], where (horizontal) assembly lines and (vertical) support arrows indicate how
the various parts of a process are supported by different kinds of resources. The GOAP9 approach to process
modeling is described in the appendix (section 11).
3.1. The process/pedagogy/tools abstract model
In Figure 5 the learning process has been divided into the sub-processes Analyze, Develop, Perform and
Evaluate, which are high-level abstract descriptions of the different stages of an overall learning process. The
vertical arrows show which tools that support which pedagogical aspects in which parts of the process.
Figure 5.
The process/pedagogy/tools abstract model. The learning process is supported by pedagogical aspects,
which, in turn, are supported by tools.
In Figure 5 the black dots indicate that the pedagogical aspect B and the tool E support the ‘Analyze’ part of the
process, while the pedagogical aspect A and the tool D support the ‘Develop’ part of the process, etc. The arrow
from ‘Evaluate’ back to ‘Analyze’ indicate a feed-back loop that is characteristic for a never-ending (life-long)
type of learning process.
3.2. Instantiating the process/pedagogy/tools abstract model
Here we will present a simple example of how the abstract process/pedagogy/tools model can be instantiated to
show the interplay between processes, pedagogical aspects and tools in a more specific example. In this example,
which is presented in Figure 6, we make use of the knowledge manifold educational architecture (described in
section 2.2) where the pedagogical aspects correspond to the knowledge roles, and the tools are divided up into
infrastructure, frameworks10 and tools.
In Figure 6 the learning process is described as the following sequence of sub-processes:
1) learning needs analysis, which produces a description of the knowledge gap (= the difference between present
and desired knowledge).
2) learning preparation & content development, which produces a set of learning offerings.
3) learning process execution (= LP-performance = LP-instantiation), which produces (some measure of :-)
understanding.
4) learning assessment & certification, which produces quality certificates.
8
For example, IMS Learning Design [105], a leading international standardization effort that deals with the description of learning
processes, does not seem to make use of process modeling in a systematic way.
9
Goals, Obstacles, Actions, Prerequisites.
10
Here, the word ‘framework’ denotes a code library, which could be regarded as a programmer’s form of abstract model, which s/he
instantiates by writing a computer program that makes use of the library.
8
In Figure 6, below the description of the learning process there is listed the set of knowledge roles for a
knowledge manifold educational architecture.11 These roles represent different types of human involvement in
the learning process. Below these roles the figure lists the infrastructure (Edutella), frameworks (SCAM and
SHAME) and tools (Formulator, Confolio, Conzilla, VWE) of a knowledge manifold, as well as the Flashmeeting tool. 12
Figure 6.
Instantiated pedagogical aspects and tools for a Knowledge Manifold with Flash meeting support.
The information in Figure 6 should be interpreted in the following way:
• During the learning needs analysis stage, the cartographer makes use of:
• Conzilla in order to map out the present and desired competence of a learner
and describe the corresponding knowledge gap.
• During the learning preparation and content development stage, the cartographer makes use of:
• Conzilla in order to create context-maps of the relevant knowledge areas.
• Formulator in order to create suitable metadata editors to describe the various concepts involved.
• During the learning preparation and content development stage, the librarian makes use of:
• Conzilla in order to fill the context-maps created by the cartographer with information about content.
• Edutella in order to search for and locate information about this content on the Semantic Web.
• Formulator in order to create suitable metadata editors for the description of this information.
• Confolio in order to store this information (and sometimes also the content itself).
• During the learning preparation and content development stage, the composer makes use of:
• Conzilla in order to locate relevant material for a certain learning module
from the context-maps created by the cartographer.
• Confolio in order to locate relevant material for a certain learning module
from the material gathered by the librarian.
• VWE in order to assemble the located material into a customized learning module.
• During the learning preparation and content development stage, the developer makes use of:
• SCAM in order to develop relevant material in the form of computer programs.
• SHAME in order to develop relevant material in the form of computer programs.
11
In order to illustrate the connection with the frameworks (code libraries) SCAM and SHAME, the knowledge-mentor has been substituted
for the knowledge-developer, which is not part of the original seven knowledge roles for a knowledge manifold. The latter role could also be
called LearningObject-developer, or Content-developer.
12
Flash meeting is developed by the Knowledge Media Institute of the Open University under the coordination of Peter Scott.
9
• During the learning process execution stage, the coach makes use of:
• VWE in order to run (execute) the learning process in the learning module created by the composer.
• Conzilla in order to let the learners map their own knowledge as it develops over time.
• Confolio in order to let the learners collect and present their own material, as well as to reflect and
comment on the material of others.
• During the learning process execution stage, the preacher makes use of:
• Flash meeting in order to “preach on request” and present live answers to learner questions,
and record them for future access and storage in the Conzilla-based knowledge archive.
• During the learning process execution stage, the plumber makes use of:
• Conzilla in order to browse different context-maps looking for relevant preachers to answer a question.
• Confolio in order to browse different content archives looking for relevant preachers to answer a question.
• Edutella in order to search the Semantic Web looking for relevant preachers to answer a question.
• During the learning assessment and certification stage, someone13 makes use of:
• Confolio in order to let the learners presents the knowledge they have gained during the learning process.
The scenario described above is related to question-based learning as described in [39] and [40].
3.3. Refining the process part of the model
Observe that the description of the learning process still remains very abstract and general. In fact, process
modeling is typically performed top-down, where each sub-process is divided into different parts and described
in more detail, until satisfactory level of concretion is reached. For example, two of the sub-processes in Figure 6
are named in an aggregated way that immediately suggests subdivision into parts. These two sub-processes are
“learning preparation and content development” and “learning assessment and certification.” In Figure 7 we
show the refinement of the model with respect to the former of these sub-processes.
Figure 7.
Refining the process part of the framework
The names of the other three sub-processes of Figure 6 do not directly suggest how to break them up into parts,
and the way to do this will in fact be a characteristic of the actual learning process under study. Hence, by
modeling the way that different learning processes refine the abstract process model, we can in fact create an
empirical basis for their classification.
13
Here a supporting tool is introduced without specifying who (= which knowledge role) is making use of it.
10
Figure 8.
Breaking learning process execution into smaller parts (sub-processes)
To give a brief indication of how this works, in Figure 8 the learning process execution process has been broken
up into the three sub-processes collecting, weeding and reflecting, and in Figure 9 the reflecting process has been
broken up into the three sub-processes constructing, testing and refactoring.14
Figure 9.
4.
Breaking the reflecting process into smaller parts (sub-processes)
THE SECI KNOWLEDGE CREATION PROCESS
In their award-winning book from 1995 The Knowledge Creating Company [54], Nonaka and Takeuchi
introduce their theory of organizational knowledge creation, first put forward by Nonaka [53] in 1994.
Although it is well known within the knowledge management community, it seems to have had relatively little
influence within the learning management community, which could be taken as an indicator of the strong
traditional separation between these two communities.
According to Nonaka and Takeuchi, the Cartesian split between subject and object, the knower and the known,
has given birth to a western view of an organization as a mechanism for information processing. While this view
has proven to be effective in explaining how an organization functions, it does not really explain the concepts of
innovation and knowledge creation. In the Nonaka-Takeuchi theory of knowledge creation, the cornerstone is the
distinction between tacit and explicit knowledge. The dominant form of knowledge in the West is explicit
knowledge, which can be easily transmitted across individuals - formally and systematically. In contrast, the
Japanese view knowledge as primarily tacit – something that is not easily visible and expressible, but which is
deeply rooted in an individual’s actions and experiences.
4.1. The SECI modes of knowledge conversion
According to Nonaka [53] the key to knowledge creation lies in the following four (SECI) modes of knowledge
conversion, that occur when tacit and explicit knowledge interact with each other:
• Socialization, which is the process of sharing experiences (tacit knowledge), thereby creating new tacit
knowledge.
• Externalization, which is the process of articulation and conversion of tacit knowledge into explicit knowledge.
• Combination, which is the process of restructuring and aggregating explicit knowledge into new explicit
knowledge.
• Internalization, which is the process of reflecting on and embodying explicit knowledge into tacit knowledge.
14
Naturally, in Figures 4, 5, and 6 there are also feedback loops with ‘breakout criteria’ for the different process chains, but for reasons of
simplicity they are not shown in these figures.
11
Figure 10.
The SECI spiral of knowledge creation
As illustrated in Figure 10, which is based on Naeve [43], a knowledge-creating spiral occurs when these modes
of interaction between tacit and explicit knowledge are elevated from the individual, to the group and
organizational levels. Organizational knowledge creation, therefore, should be understood as a spiraling process
that organizationally amplifies the knowledge created by individuals and crystallizes it as a part of the
knowledge network of the organization. This process takes place within an expanding “community of
interaction” which crosses intra- and inter-organizational levels and boundaries ([74], p. 51).
4.2. Ba – a place for interactive knowledge creation
Nonaka and Takeuchi emphasize that, on the organizational level, the spiral of knowledge creation is guided by
dialectical thinking15 and driven by organizational intention, i.e. an organization’s aspiration to achieve its goals.
Moreover, they introduce the Japanese concept of ba (which roughly means “place for interactions”) as a crucial
enabler for effective knowledge creation. The Japanese word ‘ba’ is a concept that unifies physical space (such
as e.g., an office space), virtual space (such as e.g., e-mail), and mental space (such as e.g., shared ideas). Within
an organizational context, it is the role of middle managers to maintain the necessary manifestations of such ba
in order to support the knowledge creation spiral and make it efficient for the purposes of the organization.
Figure 11.
A conceptual model of the SECI knowledge creation process.
There are four types of ba that support the four different modes of knowledge conversion: originating ba,
dialoguing ba, systemizing ba and exercising ba [55]. Each ba offers a context for a specific step in the
knowledge-creating process. Building, maintaining and utilising ba is important to facilitate organizational
knowledge creation.
15
Which tries to transcend paradox by achiving a Hegelian synthesis between thesis and anti-thesis.
12
• Originating ba provides a context for socialization. It is a place where individuals
transcend the boundaries between self and others by sympathising or empathising with others and sharing tacit
knowledge in the form of experiences, feelings, emotions and mental models. From originating ba emerge care,
love, trust and commitment, which form the basis for knowledge conversion among individuals.
• Dialoguing ba provides a context for externalisation. Tacit knowledge is shared and articulated through
dialogues amongst participants. Dialoguing ba is the place where individuals' mental models and skills are
shared, converted into common terms, and articulated as concepts. The articulated knowledge is also brought
back into each individual, and further articulation occurs through self-reflection.
• Systemising ba provides a context for the combination of existing explicit knowledge into new forms.
Information technology, through such things as on-line networks, groupware, electronic mailing lists, news
groups, databases, etc., offers a virtual collaborative environment for the creation of systemizing ba.
• Exercising ba provides a context for internalisation. Here, individuals embody explicit knowledge that is
communicated through virtual media, such as written manuals or simulation programs. Exercising ba synthesises
the transcendence and reflection through action, while dialoguing ba achieves this through thought.
Ba exists at many levels that may be connected to form a greater ba. Individuals form the ba of groups/teams,
which in turn form the ba of organisation. Then, the market environment becomes the ba for the organisation. Ba
is a concept that transcends the boundary between micro and macro, and the organic interactions amongst these
different levels of ba can amplify the knowledge-creating process.
4.3. Knowledge assets
At the base of knowledge-creating processes are knowledge assets. Nonaka, Toyama, and Konno [55] define
knowledge assets as “firm-specific resources that are indispensable to create values for the firm”. According to
them “knowledge assets are the inputs, outputs and moderating factors of the knowledge-creating process. For
example, trust amongst organizational members is created as an output of the knowledge-creating process, and at
the same time it moderates how ba functions as a platform for the knowledge-creating process.”
Knowledge assets must be built and used internally in order for their full value to be realized, as they cannot be
readily bought and sold. To understand how knowledge assets are created, acquired and exploited, Nonaka,
Toyama, and Konno propose to categorize knowledge assets into four types – corresponding to the four (SECI)
modes of knowledge conversion: experiential knowledge assets, conceptual knowledge assets, systemic
knowledge assets, and routine knowledge assets (see Figure 11). They give the following characterization of
these four types ([55], p. 21-22):
• Experiential knowledge assets consist of the shared tacit knowledge that is built through shared hands-on
experience amongst the members of the organisation, and between the members of the organisation and its
customers, suppliers and affiliated firms. Skills and know-how that are acquired and accumulated by individuals
through experiences at work are examples of experiential knowledge assets. Other examples of such knowledge
assets include emotional knowledge, such as care, love and trust, physical knowledge such as facial expressions
and gestures, energetic knowledge such as senses of existence, enthusiasm and tension, and rhythmic knowledge
such as improvisation and entrainment.
• Conceptual knowledge assets consist of explicit knowledge articulated through images, symbols and language.
They are the assets based on the concepts held by customers and members of the organization. Brand equity,
which is perceived by customers, and concepts or designs, which are perceived by the members of the
organization, are examples of conceptual knowledge assets.
• Systemic knowledge assets consist of systematized and packaged explicit knowledge, such as explicitly stated
technologies, product specifications, manuals, and documented and packaged information about customers and
suppliers. A characteristic of systemic knowledge assets is that they can be transferred relatively easily. This is
the most visible type of knowledge asset, and current knowledge management focuses primarily on managing
systemic knowledge assets, such as intellectual property rights.
• Routine knowledge assets consist of the tacit knowledge that is ‘routinized’ and embedded in the actions and
practices of the organisation. Know-how, organisational culture and organizational routines for carrying out the
13
day-to-day business of the organisation are examples of routine knowledge assets. A characteristic of routine
knowledge assets is that they are practical.
4.4. Leading the SECI knowledge-creating process
As mentioned above, the SECI process is guided by dialectical thinking, which focuses on transcending paradox
by creating a synthesis between opposing forces, such as between order and chaos, micro and macro, tacit and
explicit, body and mind, emotion and logic, and action and cognition. As pointed out by Nonaka, Toyama and
Konno in [55], the SECI process cannot be managed in the traditional sense of management, which centers on
controlling the flow of information. In contrast, top and middle management take a leadership role by “reading
the situation”, as well as leading it, working on all three elements16 of the knowledge-creating process. Leaders
provide the knowledge vision, develop and promote the sharing of knowledge assets, create and energize ba, and
enable and promote the continuous spiral of knowledge creation. This overall organizational knowledge creation
process is modeled in Figure 12.17
Especially crucial to the SECI process is the role of knowledge producers, i.e. middle managers who actively
interact with others to create knowledge by participating in and leading ba. Nonaka, Toyama and Konno
emphasize that in order to create knowledge dynamically and continuously, an organization needs a vision that
synchronizes it. It is the role of top management to articulate the knowledge vision and communicate it
throughout (and outside) the company. The knowledge vision defines what kind of knowledge the company
should create and in what domain. In short, it determines how the organization and its knowledge base evolve
over the long term. The knowledge vision also defines the value system that evaluates, justifies and determines
the quality of the knowledge the company creates.
Figure 12.
5.
The overall organizational knowledge creation process
THE SECI PROCESS FRAMEWORK
By combining learning process modeling (section 3) with the SECI theory of knowledge creation (section 4), we
can create a SECI process framework (abstract model) for the description and classification of knowledgecreating learning processes. In Figure 13 we have introduced the four different kinds of ba, as well as their
corresponding tools of support. Socialization occurs in originating ba, where experiencing and empathizing
activities are supported by community building tools. Externalization occurs in dialoguing ba, where articulating
and conceptualizing activities are promoted by discussion supporting tools. Combination occurs in systemizing
ba, where connecting and deducing activities are supported by conceptual modeling tools. Internalization occurs
in exercising ba, where reflecting and embodying activities are supported by reflective analysis tools.
16
17
SECI, ba and knowledge assets.
which is a process model version of Figure 8 from [55].
14
Figure 13.
The SECI process framework: increasing understanding
through experiencing, articulating, deducing and reflecting.
In each of the four SECI knowledge conversion stages a learning process takes place. As shown in Figure 13,
sharing experiences in the socialization process, with input from visions, challenges and activities, produces new
individual understanding of the issues at stake. This new individual understanding is then externalized and
articulated into new collective understanding of the same issues. Then the combination process deductively
produces increased collective understanding, which is then internalized by reflection and embodied into
increased individual understanding.
Figure 14.
The SECI process framework with a process model within each ba.
In Figure 14 the different knowledge conversions have been modeled as processes. During the socialization
process we respond to challenges and activities by collecting inspiring experiences. During the externalization
process they form the input for discussions, which produce articulated concepts. During the combination
15
process, these articulated concepts are connected and combined into conceptual models, and during the
internalization process, these conceptual models are reflected upon, which results in increased understanding of
the issues involved.
In Figure 14 we think of each process as described by the kind of process/pedagogy/tools model that was
introduced in section 3. This is difficult to draw in the overall diagram, but in the Conzilla-based version, a
double-click on the top diagram within each ba would open up the corresponding process/pedagogy/tools
abstract model. In order to describe a concrete professional learning process we can then “drill down” of the
processes in each ba and perform a top-down construction of their corresponding process/pedagogy/tools model.
By mapping out concrete learning processes in this way, we lay the empirical foundation for their future
classification. Hence the SECI process framework provides a methodology for researching the structure of
knowledge-creating learning processes.
6.
EMPIRICALLY VALIDATED PEDAGOGICAL SUPPORT
6.1. Introduction and overview
We will now present some previous research into the development and testing of pedagogical ideas related to the
SECI process framework (Figure 16). The corresponding learning model involves four latent variables, namely
Socialization, Externalization, Combination, and Internalization, which will be reviewed from a pedagogical
perspective. These latent variables consist of the measurable processes: Collecting experiences, Discussing
experiences, Modeling articulated concepts, and Reflecting on the models (Figure 14). These processes take
place within the corresponding ba (space of interaction, which has been marked with a dashed rectangle and a
descriptive name in Figure 14. In front of every process variable we have inserted a descriptive variable of the
available tools. The Socialization variable in the learning model consists of Collecting experiences, which will
take place by interacting with other learners in Originating ba. In this ‘space’ a collaborative learning group will
be formed with Inspiring Experiences as an end product. When forming the learner group Community building
tools will be used.
Community building tools include a support process (at least in academic learning level), which is partly covered
by an Interaction process, which might activate the Exploratory learning behaviour. The exploratory behaviour
is activated only if the interaction process is good enough. Self-esteem is the first endogenous variable, which is
predicted by the quality of the interaction process. On the other hand, the intrinsic motivation is further predicted
by learner’s self-esteem. First when the intrinsic motivation is activated, the exploratory behaviour or the process
of Collecting experiences under Socialization is activated. This part has been empirically tested by Yli-Luoma
using the LISREL method ([83], p. 211; [84], pp. 15-28). Here he shows that the Collecting experiences process
is emotionally ([82], p. 106; [83], p. 211) and socially ([77], p. 163) loaded. It should be observed that the
quality of the emotional network of the social group would seem to increase the self-esteem of the group, which
further activates the Collecting experiences process.
Within the Externalization phase, the Discussing experiences process is still emotional. However, cognitive
dimensions are needed ([22], p. 114), where creativity is also activated ([10], p. 152). Kolb argues further that
this process becomes effective in the form of teamwork (see further [28]).
When the Externalization phase goes over to the Combination phase, the Modeling process is activated, and
Hypothetical-Deductive thinking abilities are needed for the modeling approach ([25], p. 209; [27], p. 117; [26],
pp. 388-390; [18], p. 175). It should be observed here that the Constructivistic learning process only covers the
Socialization and Externalization phases of the learning process ([27], p. 175). In any type of learning that
involves deductive or abductive reasoning, such as e.g. the learning of science and mathematics, a modeling
approach is needed ([15], p. 1; Hestenes [20]).
The learning process requires further that students engage in seeking to understand and explain the conceptual
models developed in the Combination process ([81], pp. 92-93; [61], p. 1), which means that the process of
Internalization is activated. This is a process, where learners reflect on the new structures (models) by using
critical thinking abilities when testing or applying them.
16
6.2. Socialization process
In the Socialization process, the social interaction between students and their teachers is included. Yli-Luoma
([83], p. 175; [84], p. 27) has observed that when this interaction is good enough, and when it covers three
special dimensions: emotional attachment, cognitive support, and moral values, it will advance the internal
working models, which include intrinsic motivation. The intrinsic motivation, however, is mainly activated by
strong self-esteem, which is a product of the interaction process. So the best support would seem to be the
advancement of strong self-esteem among the learners in order to activate their learning processes. Bowlby ([9],
p. 238) argues that secure emotional attachment activates exploratory behaviour, which is best conceived as
mediated by a set of behavioural systems evolved for the special function of extracting information from the
environment. Activation results from novelty and termination from familiarity.
The activation process was tested empirically by using a LISREL model. The model (Figure 15) was run with
one exogenous variable (Interaction) and three endogenous variables (Self-Esteem, Motivation, and Learning).
The measures used are shown in the boxes above the latent variables. The beta-coefficient (between self-esteem
and motivation) and the gamma-coefficient (between teacher-student interaction and self-esteem) were both
statistically very significant (γ21 = 0.54, p < 0.001; β32 = 0.44, p < 0.001), and the beta-coefficient between
motivation and learning (β43 = 0.52, p < 0.001) (Figure 15) was also very significant.
The Interaction variable was measured by three dimensions: Emotional, Supporting, and Moral. The Self-Esteem
variable was measured with Social and Academic dimensions - according to Shavelson et al. ([68], p. 407). The
latent variable Motivation has the dimensions of Intrinsic motivation and Extrinsic motivation as a measurement
model. The learning process was measured with two types of learning: Imitative and Explorative.
Figure 15.
The LISREL model of the interaction process
Vygotsky ([76], [77]) claims that the social context has a significant impact on the learning process. He argues
further that it takes place on two levels, the social and the psychological level. The social interaction process is
observed by inter-personal relationships. The psychological process takes place on the intra-psychological level,
which means that the learners construct new information with their thinking abilities. This type of approach has
made a contribution to social constructivism, which was developed by Berger and Luckman ([6]).
The interaction process above refers to synchronous face-to-face learning. What about synchronous or
asynchronous distance learning? How do we activate the exploratory behaviour (i.e., how do we motivate) at a
distance and asynchronously? Interaction design is the art of effectively creating interesting and compelling
experiences for others ([70]).
The distance in time and place seems to impede the process of bonding (attachment) and of building cohesion in
a group. Cohesiveness in a group is positively reinforced if the group goals match the members’ personal goals,
if the group interacts effectively and harmoniously, and if the members are attracted to each other ([67]). To
build trust and create a feeling of cohesion, intensive personal attention and presence is required, which is
difficult to achieve via Internet-based communication. Bonding (social attachment) is much easier to advance if
members have met face-to-face first.
The social interaction among the online learners is crucial not only for knowledge construction and mutual
support, but for the reduction of isolation and anxiety during the independent learning process (compare
Vygotsky's psychological level).
Comparing face-to-face learning and online learning, the social context might be the one dimension, where most
differences can be found. The social context, however, is one of the cornerstones in the learning process. How
then can online learning be arranged in order to take place in such a way that the participants maintain mutual
caring and understanding through the interactions, which can be offered online? A good arrangement would
17
mean that the online learners would be able to develop a sense of belonging, social-emotional bonds or
attachment, and supportive relationships.
Collecting experiences is positioned in the Socialization knowledge conversion process, which is emotionally
and/or affectively loaded. If the learner does not like the subject, s/he would not be interested to collect any new
information or experiences either. The Kolbian approach replaces these two aspects together ([28], [22]).
Moreover, brain research has demonstrated that learning is based on collecting experiences ([9]). In his study of
Kolbian learning styles, Yli-Luoma ([85]) has observed that collecting experiences is one of the basic learning
styles, but if it remains the preferred style of the learner, then the learning process would seem to remain
qualitatively quite week. It would further seem to haven a negative prediction ability for college and university
performance. This would mean that the students with this learning style only would not seem to perform well in
their studies.
6.3. Externalization process
The Socialization process is needed in order to activate a collaborative discussing phase between the online
learners ([22], [28]). Ravenscroft ([64]) argues further that a socio-cultural framework is needed for cognitive
change. According to the argumentation of Vygotsky ([76]), the higher cognitive processes provide a basis and
motivation for collaborative, argumentative and reflective discourse.
Bransford et al. ([10]) suggest further that the collaborative discussing phase includes creativity. Zohar ([86])
argues that the creative thinking process demands that we can break old rules or are able even for a shift of
paradigms. Some brain researchers argue that this kind of thinking is placed in human brains within the same
area as motivation, vision, value and meaning.
According to Keeves ([27]), the constructivist approach still works in this phase. Students construct the
information and experiences towards their new knowledge using the Piagetian cognitive developmental stage at
the Concrete Operational Stages, but they do not need to go beyond these stages. Keeves argues further that at
least in the fields of mathematics and science, the basic principles of constructivism are incomplete and
inadequate for both learning and teaching these fields. This argument is strongly supported by the modeling
theory of Hestenes ([46], p. 355-356), which he has applied to the education of high-school physics teachers for
almost two decades ([20]).
Sweller ([73]) questions strongly the efficacy of so-called 'constructivist based' learning and argues that evidence
for the effectiveness of these learning procedures is almost totally missing with a lack of systematic and
controlled experimentation. The experimentation, however, should be re-positioned after the modeling process
(or the Combination process as it is called in the present study).
We stress the relevancy of cognitivism in the Externalization phase. One of the learning strategies that support
cognitivism is concept mapping (Novak [57]), which is a technique for the expression and visualization of
domain concepts and their relationships. Concept maps are tools for organizing and representing knowledge.
They include concepts and propositions. Concepts are defined as a perceived regularity in events or objects, or
records of events or objects, designated by a label (Novak [58]). Propositions are statements about some object
or event in the universe, either naturally occurring or constructed. Propositions contain two or more concepts
connected with other words to form a meaningful statement.
6.4. Combination process
The Modeling approach (modeling articulated concepts) takes place in Systemizing ba after Externalization
(Figure 14). Here the learners need Conceptual modeling tools in order to advance the articulated concepts
towards forming Conceptual models. For these kinds of processes, higher thinking abilities with advanced
modeling tools are needed. The Piagetian Formal Operations Stage would seem to fulfill this demand. At the
formal operational stage, students are able to formulate and test a single hypothesis - they are able to go beyond
the data. When the problem is more complex - and several hypotheses are needed - a model approach would
seem to be more suitable. Kaplan ([25], p. 117) argues that the term 'model' is useful when the symbolic system
it refers to is significant as a structure - a system that allows for exact deductions and explicit correspondences.
The value of the model lies in the deductive fertility of the model, so that the unexpected consequences can be
predicted and then tested by observation and experiment. Evers ([15]) has presented a connectionist model of
artificial neural networks in an educational situation. Penner's article ([61]), titled 'Cognition, Computers, and
Synthetic Science: Building Knowledge and Meaning through Modeling', laid the foundations for a shift towards
what he recognizes to be a modeling approach. However, Penner fails to recognize that a model must be tested
18
for adequacy. While he considers practical work in the traditional teaching of science, he does not see clearly its
role in a modeling approach. Keeves ([27]), however, argues very clearly for a modeling approach, which has to
satisfy the following requirements:
• a model should lead to a prediction of consequences,
• a model should contain both associative and structural relationships,
• a model should reveal a causal direction leading to explanations, and
• a model should give rise to new concepts and new relationships.
Keeves has identified several types of models:
• analogue models,
• semantic models,
• schematic models,
• mathematical models, and
• causal models.
These modeling processes might well be described as construction processes, and the term 'constructionism'
could be employed. Nevertheless, the term 'constructionism' has already been used in association with social
constructivism and in this context it has a very different meaning. Since most often a simple construction with
the characteristics of a model is not being built through social constructionism, the term 'constructionism' is best
avoided and an alternative word should be sought. That is why the term 'modeling' could be adopted. In the
present learning model, the term Combination (Figure 14) is also used for this process. The Combination
process in the present study is closely related to the Kolbian Learning Style of the Theorist. Its predictive value
in the academic learning process is the highest possible ([85]). In the very same comparative study of Kolbian
learning styles, which uses a new measurement model advanced by Yli-Luoma ([85]), it is shown that the
Kolbian theorist (closely related to our Systemizing ba) is very rare as the preferred learning style among
European polytechnic students (Finland 10.0%, France 23.8%, Italy 3.3%, and Spain 13.0%).
The modeling approach has already been used in Bloom’s taxonomy ([8]). For example, at the Apply level (see
Figure 1) the learner should be able to construct a model of the phenomenon under study and demonstrate how it
will work. At the Analyze level the learner should be able to make a flow chart to show the critical stages of
knowledge or construct a graph to illustrate the selected information. Moreover, Novak ([56]) has turned
modeling into a real cognitive tool in his conceptual mapping procedure. While constructing good concept maps
([110]) the learner is modeling e.g., a laboratory activity, or a particular problem or question that s/he is trying to
understand.
During the Combination phase, case-based reasoning (CBR) is also important, as an approach to learning and
problem solving based on previous experiences (Kolodner [29]). Past experiences are stored in the form of
solved problems (‘cases’) in a so-called case base. A new problem is solved based on adapting solutions of
known similar problems to this new problem. This kind of inference is necessary for addressing ill-defined or
complex problems. Key to such reasoning is a memory that can access the right experiences (cases) at the times
they are needed.
6.5. Internalization process
The Internalization process consists of Reflection on the models, which takes place in Exercising ba (Figure 14).
This would mean that the learners should already have conceptual models of the knowledge (theory) they are
articulating. They should now advance experiments, laborations etc. in order to test the conceptual models they
have developed. This process should increase their understanding.
Yli-Luoma's ([81], p. 92) comparative study among pre-university students of physics learning reveals the
importance of experimental and testing processes. He had access to data from seven different countries of which
three made use of an experimental approach (Exercising ba) and the remaining four did not. The results expose
how pre-university students understand physics without having evolved their understanding in an experimental
context. In those countries in which the students were involved with an experimental approach, the thinking
abilities and understanding of physics were much better developed than in the countries where the experimental
approach was missing. The Test on Understanding Physics for those using an experimental approach
(Exercising ba) the score was µ = 38.3, σ = 9.0 and for those not using an experimental approach (no
Excersising ba) µ = 11.5, σ = 4.8 and for these two groups (Exercising ba and no Exercising ba) the t-test
19
value was calculated t = 36.9, p < 0.001. This would seem to give a very strong evidence for advancing a well
working Excersising ba.
From the above it can be concluded that the theoretical approach in a learning process is not enough, but an
experimental learning approach (Exercising ba), with testing of knowledge, will lead to a better quality of
learning.
How is the experimental learning process implemented in online learning? Simulations might be useful as
laboration tools in an experimental approach. Nakajima ([47]) tested it in physics-learning, using chat-forum as a
reflection tool. His experiment would seem to confirm the idea of using simulations as a part of the experimental
approach.
7.
CONCLUSIONS AND FUTURE WORK
In this paper we have presented a conceptual approach to the studying of learning processes. We have introduced
the process/pedagogy/tools model and shown how its assembly-line style of process modeling can be used to
describe which pedagogical aspects and which tools that support which parts of a specific learning process.
Moreover we have introduced the distinction between knowledge-transmitting and knowledge-creating learning
processes, a distinction that separates formal learning from informal learning, as well as (traditional) courses
from research. Finally, we have presented the SECI process framework for the study and analysis of knowledgecreating learning processes, and we have shown how the different SECI modes of knowledge conversion are
empirically supported by pedagogical research.
Naturally, both knowledge-transmitting and knowledge-creating learning processes have to be supported in
workplace learning. As the percentage of “knowledge workers” is rapidly increasing and 50% of all employee
skills become outdated in three to five years (Moe and Blodgett [38]), re-qualification plays an important role.
Since re-qualification is often based on learning already existing knowledge and skills, knowledge transmission
is typically required in such situation. On the other hand, companies also need to collect and analyze the
feedback from customers and their own employees, investigate the market, compare their products with those of
the competition, and design and develop innovations. In such situations new knowledge has to be created, and
this is also the critical demand of the present “knowledge age”. Hence our deliverable focuses on a type of
learning process – knowledge creation - that is crucial for workplace learning, and which in the past has not been
investigated as much as knowledge transmission.
7.1. Applying the framework models to Prolearn
The major aim of this paper has been to formulate abstract conceptual models that could provide a foundation for
useful ways to characterize and classify work place learning processes. We believe that both the
process/pedagogy/tools model (section 3.1) and the SECI process framework (section 5) are very relevant for
studying learning at the workplace. When applied to the knowledge-creation of the Prolearn NoE, the SECI
spiral goes from Core partners to Associate partners and further outwards to the Scientific Community &
Industry, as illustrated in Figure 16. In fact, this framework has already been adopted as the basis for the
Prolearn roadmapping process (Kamtsiou et al. [24]), and we also plan to apply it to the study of work place
learning within the Prolearn workpackage on knowledge work management.18 In fact, this work will be part of
the deliverable D 7.5 (Antecedents for Effective Learning Management at the Workplace), which is due on
month 24. As discussed earlier, work place learning tends to be much more knowledge-creating than traditional
course-based learning, which is why we have chosen the SECI theory of knowledge creation as a basis for this
conceptual approach.
18
In fact, the Prolearn Virtual Competence Centre provides an excellent ba for this type of research process.
20
Figure 16.
Applying the SECI process framework to Knowledge work management could give a classification framework
for knowledge-creating types of workplace learning processes.
7.2. Future research issues
How should we design effective learning processes in the workplace? In the present study we have introduced
several different types of knowledge. Göranzon’s [20] four types of knowledge have been discussed, as well as
Bloom’s six levels of cognitive knowledge [8] (slightly revised by Anderson and Kratwohl [1]) ranging from
simple remembering of facts at the lowest level through more complex and abstract mental levels to the highest
one, classified as Creation in Figure 1. Moreover, Naeve [43] defines a new dimension of knowledge, namely
efficient fantasies. We need here a synthesis of the different types or dimensions of knowledge, which could be
tested by confirmatory factor analysis.
Also, knowledge transmission or creation and pushing and pulling concepts were presented as two important
dimensions that distinguish formal and informal learning processes. These two types of learning could be tested
by a comparative approach, which is one strong feature in the LISREL –method (LInear Structural
RELationships [80]).
Another interesting comparison of knowledge-transmitting and knowledge-creating types of learning processes
recognizes knowledge-simulating and knowledge-stimulating type of behaviour (Figure 3). Here the following
question arises: how can we distinguish real and simulated knowledge? One possibility might be by means of the
Bloom taxonomy – which has knowledge simulation (or imitation) as its lowest level. In the knowledge-creating
learning process learners are trying to figure out the right questions. This corresponds with the revised concept of
intelligence as specified by R. C. Schank ([11]). The easier is it to get information the lower is its value. But the
value of good questions increases. In the future, intelligence will mean ability to reach the boundaries of the
knowledge base.
The SECI knowledge creation process consisted of four different dimensions, namely Socialization,
Externalization, Combination, and Internalization. It was also concluded that these four different types of
knowledge creation processes take place in four different types of interaction spaces (ba). These spaces are of
great interest especially for INTeL –project group (INTeractive e-Learning), which will be collecting the data to
be analyzed. And the data collecting procedure should be carefully planned to cover all the processes and
theoretical features, which are included in the measurement model
The main aim of the present study has been to theoretically investigate different types of knowledge, knowledge
transmission or creation, knowledge creation processes, and different types of interaction spaces (ba).
In the previous chapters a hypothetical model of learning was advanced and compared with previous research.
However, a major aim for the future is to test its validity when applied in online learning. The INTeL –project19
19
Managed by the University of Oulu. which is one of the Prolearn associate partners.
21
will setup an empirical testing process for the present hypothetical model (see Figure 16Figure 16). The first aim
should be to collect all the used concepts in the four stages in the hypothetical process model, which should
cover the useful theoretical features in every stage to be used. A structural model will be the aim product. After
that a measurement model will be advanced to be able to collect correct empirical data for testing purposes. The
LISREL-method could be applied as an analysis method for the following purposes:
• a structural model can be tested,
• all the hypotheses can be confirmed in one model run,
• in LISREL a hypothetical model can also be modified manually or automatically, and
• a simultaneous comparative analysis in several populations can also be undertaken.
8.
ACKNOWLEDGEMENTS
The authors acknowledge the support of the Prolearn Network of Excellence, which has made it possible for
them to collaborate.
REFERENCES20
9.
Papers and Books
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
[17]
20
Anderson, L. W., Kratwohl, D. R. (eds.) (2001) A Taxonomy for Learning, Teaching and Assessing: A
Revision of Bloom’s Taxonomy of Educational Objectives, New York: Longman.
Argyris, C., Schön, D. A. (1978) Organizational Learning, Reading, Massachusetts: Addison-Wesley
Argyris, C., Schön, D. A. (1996) Organizational Learning II: Theory, Method and Practice, Reading,
Massachusetts: Addison-Wesley.
Atherton, J. S. (2004) Teaching and Learning: Bloom’s taxonomy, Online UK,
http://www.learningandteaching.info/learning/bloomtax.htm
Alexander, S. (2001) E-Learning developments and experiences, Education + Training, Vol. 43, Number
4/5, pp 240-248, MCB Univ. Press, http://www.emerald-library.com/ft
Berger P. and Luckman, T. 1969. The Social Construction of Reality: A Treatise in the Sociology of
Knowledge, Garden City, NY: Doubleday.
Blomqvist, U., Handberg, L., Naeve, A. (2003) New Methods for focusing on Students Learning Process
and Reflection in Higher Education. Proc. of the 28th IUT (Improving University Teaching) Conference.
Växjö, Sweden: May 2003.
Bloom, B. S. (ed.) (1956) Taxonomy of Educational Objectives, the classification of educational goals,
Handbook 1: Cognitive Domain, New York: Longman.
Bowlby, J. (1987) A Secure Base. Parent-Child Attachment and Healthy Human Development, New
York: Basic Books Inc.
Bransford, J., Brown, A., Cocking, R. (2002) How People Learn, Washington D.C.: National Research
Council.
Brockman, J. (ed.) (2002) The Next Fifty Years: Science in the First Half of the Twenty-First Century,
Barnes & Noble publishers.
Dave, R. H. (1975) in Developing and Writing Behavioural Objectives (R. J. Armstrong, ed.),
Educational Innovators Press.
Decker, S. & Manning, C., Naeve, A., Nejdl, W., Risch, T., Studer, R. (2001) Edutella - An Infrastructure
for the exchange of Educational Media, Part of the PADLR proposal [49] to WGLN [94], granted June
2001,
Eriksson, H. E., Penker, M. (2000) Business Modelling with UML – Business patterns at Work, Wiley &
Sons Inc., New York, 2000, ISBN 0-471-29551-5.
Evers, C. (2000) Connectionist modelling and education, Australian Journal of Education 44(33) pp. 209225.
Fiol, C. M., Lyles, M. A. (1985) Organizational learning, Academy of Management Review, Vol. 10,
No. 4, pp. 284-295.
Godet, Mitchel (1989) Effective Strategic management: The Prospective Approach, Technology Analysis
& Strategic Management I.
All listed URLs have been accessed on 2005-xx-xx.
22
[18]
[19]
[20]
[21]
[22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
Grace, A., Butler, T. (2005) Learning Management Systems: A new beginning in the management of
learning and knowledge, International Journal of Knowledge and Learning (IJKL), Vol. 1, Nos. 1-2,
Inderscience publishers.
Göranzon, B. (1990) Det Praktiska Intellektet (The Practical Intellect), Carlssons förlag.
Hestenes, D. (1995) Modeling Software for Learning and Doing Physics, in: Thinking Physics for
Teaching., Ed. Carlo Bernardini et al, pp. 25-65, http://modelingnts.la.asu.edu/pdf/ModelingSoftware.pdf
Hestenes, D. (2002) Reforming the Mathematical Language of Physics, Oersted Medal Lecture 2002 – in
response to his reception of the Oersted medal award,.
http://modelingnts.la.asu.edu/pdf/OerstedMedalLecture.pdf
Honey, P., Mumford, A. (1982) Manual of Learning Styles, London: P.Honey.
Hummel, H., Manderveld, J., Tattersall, C., Koper, R. (2004) Educational modelling language and
learning design: new opportunities for instructional reusability and personalized learning, International
Journal of Learning Technology (IJLT), Vol. 1, No. 1, pp. 111-126, 2004.
Kamtsiou, V., et al. (2005) The Prolearn Roadmapping Process, Deliverable 12.1.2.
Kaplan, A., (1997) Scientific Methods in Educational Research, In Keeves (ed.) Educational Research,
Methodology, and Measurement: An International Handbook, Oxford: Pergamon.
Keeves, J. (1997) Models and Model Building, In Keeves (ed.) Educational Research, Methodology, and
Measurement: An International Handbook, Oxford: Pergamon
Keeves, J. (2002) Learning in Schools: A Modelling Approach. In Suortti & Heikkinen (eds.) Mind,
Science, and Technology. Kajaani: University of Oulu Publications.
Kolb, D. (1984) Experiential Learning: Experience as the Source of Learning and Development. Upper
Saddle River (NJ): Prentice Hall.
Kolodner, J. (1993) Case-Based Reasoning, San Mateo: Morgan Kauffman Publishers.
Koper, E. J. R. (2001) Modelling Units of Study from a Pedagogical Perspective: the Pedagogical
Metamodel Behind EML, http://eml.ou.nl/introduction/docs/ped-metamodel.pdf
Kraan, W. (2003) Using SHAME to fill your SCAM, Center for Educational Technology Interoperability
Standards, November 5, 2003, http://www.cetis.ac.uk/content2/20031105152216
Kratwohl, D. R., Bloom, B. S., Masia, B. B. (1964) Taxonomy of Educational Objectives, the
classification of educational goals, Handbook II: Affective Domain, New York: Longman.
Levitt, B., March, J. G. (1988) Organizational learning, Annual Review of Sociology, Vol. 14, pp. 319340.
Lytras, M., Pouloudi, A., Poulymenakou, A. (2002) Knowledge Management Convergence: Expanding
Learning Frontiers. Journal of Knowledge Management 6 (1) pp. 40-51.
Lytras, M., Naeve, A., (eds) (2005), Intelligent Learning Infrastructures for Knowledge Intensive
Organizations: A Semantic Web Perspective, IDEA publishing group, 2005.
Lytras, M., Naeve, A., Pouloudi, A. (2005) Knowledge Management as a Reference Theory for ELearning: A Conceptual and Technological Perspective, International Journal of Distance Education
Technologies, Vol. 3, No. 2, pp. 1-12, (Special Issue on Knowledge management Technologies for ELearning), April-June 2005.
Lytras, M., Naeve, A., Pouloudi, A. (2005) A Knowledge Management Roadmap for E-Learning: The
Way Ahead, International Journal of Distance Education Technologies, Vol. 3, No. 2, pp. 68-75, (Special
Issue on Knowledge management Technologies for E-Learning), April-June 2005.
Moe, M., Blodgett, H. (2000) The knowledge web. Merrill Lynch & Co., Global Securities Research &
Economics Group, Global Fundamental Equity Research Department, Gartner Group and International
Data Corporation.
Naeve, A. (1997) The Garden of Knowledge as a Knowledge Manifold - A Conceptual Framework for
Computer Supported Subjective Education, CID-17, TRITA-NA-D9708, KTH, Stockholm,
http://kmr.nada.kth.se/papers/KnowledgeManifolds/cid_17.pdf
Naeve, A. (1999) Conceptual Navigation and Multiple Scale Narration in a Knowledge Manifold, CID52, TRITA-NA-D9910, KTH, Stockholm, http://kmr.nada.kth.se/papers/ConceptualBrowsing/cid_52.pdf
Naeve, A. (2001a) The Knowledge Manifold - an educational architecture that Supports Inquiry-Based
Customizable Forms of E-learning, Proc. of the 2nd European Web-Based Learning Environments
conference, pp. 200-212, Lund, Sweden, October 24-26, 2001,
http://kmr.nada.kth.se/papers/KnowledgeManifolds/KnowledgeManifold.pdf
Naeve, A. (2001b) The Concept Browser - a New Form of Knowledge Management Tool, Proc. of the 2nd
European Web-Based Learning Environment conference, pp. 151-161, Lund, Sweden, October 24-26,
2001, http://kmr.nada.kth.se/papers/ConceptualBrowsing/ConceptBrowser.pdf
Naeve, A. (2005), The Human Semantic Web – Shifting from Knowledge Push to Knowledge Pull,
International Journal of Semantic Web and Information Systems (IJSWIS), Vol. 1, No. 3, pp. 1-30, JulySeptember 2005.
23
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]
[58]
[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
Naeve, A., Nilsson, M., Palmér, M. (2001) The Conceptual Web - our research vision, Proceedings of the
first Semantic Web Working Symposium, Stanford, July 2001,
www.semanticweb.org/SWWS/program/position/soi-nilsson.pdf
Naeve, A., Knudsen, C., Nilsson, M., Palmér, M., Paulsson, F., Petersson, A., Müller, E., Pargman, D.,
Blomqvist, U. (2002) En publik e-lärandeplattform byggd på kunskapsmångfalder, öppen källkod och
öppna IT-standarder, (A public e-learning platform built on knowledge manifolds, open source and open,
international ICT standards. Report to the Swedish Netuniversity. Nov. 2002,
http://kmr.nada.kth.se/papers/SemanticWeb/Natuniv-KMR.pdf
Naeve, A., Nilsson, M., Palmér, M., Paulsson, F. (2005) Contributions to a Public e-Learning Platform –
Infrastructure, Architecture, Frameworks and Tools, International Journal of Learning Technology
(IJLT), Vol. 1, No 3, 2005, pp. 352-381. http://kmr.nada.kth.se/papers/SemanticWeb/Contrib-to-PeLP.pdf
Nakajima, Y. 2002. The simulation program and chat in the e-Learning course in physics (mechanics).
http://www.aace.org/dl/files/ELEARN2002/paper _3009_2617.pdf
Nejdl. W., Wolf, B., Qu, C., Decker, S., Sintek, M., Naeve, A., Nilsson, M., Palmér, M., Risch, T. (2002)
Edutella: A P2P Networking Infrastructure Based on RDF, Proc. of the 11th World Wide Web
Conference (WWW2002), Hawaii, May 7-11, 2002. http://kmr.nada.kth.se/papers/SemanticWeb/p597nejdl.pdf
Nejdl, W., et al. (2001) PADLR-1 (Personalized Access to Distributed Learning Repositories) proposal to
WGLN, granted in March 2001, www.learninglab.de/pdf/L3S_padlr_17.pdf
Nilsson, M. (2001) The Semantic Web: How RDF will change learning technology standards, Feature
article, Centre for Educational Technology Interoperability Standards (CETIS), October 2001,
http://kmr.nada.kth.se/papers/SemanticWeb/CETISSemWeb.pdf
Nilsson, M., Palmér, M., Naeve, A. (2002) Semantic Web Metadata for e-Learning - Some Architectural
Guidelines, Proceedings of the 11th World Wide Web Conference, Hawaii, May 7-11, 2002.
http://kmr.nada.kth.se/papers/SemanticWeb/p744-nilsson.pdf
Nilsson, M., Naeve, A., Palmér, M. (2004) The Edutella P2P Network - Supporting Democratic ELearning and Communities of Practice, in McGreal, R. (ed.), Online Education Using Learning Objects,
Routledge-Falmer, New York, 2004, ISBN 0-415-33512-4,
http://kmr.nada.kth.se/papers/SemanticWeb/Edutella-chapter.pdf
Nonaka, I. (1994) A dynamic theory of organizational knowledge creation, Organization Science, Vol. 5,
No. 1, pp. 14-37.
Nonaka, I., Takeuchi, H. (1995) The knowledge-creating company: How Japanese companies create the
dynamics of innovation, New York: Oxford University Press.
Nonaka, I., Toyama, R., Konno, N. (2000) SECI, Ba and Leadership: a Unified Model of Dynamic
Knowledge Creation, Long Range Planning, Vol 33 (2000), Elsevier Science Ltd.
Novak, J. D. (1990) Concept maps and Vee diagrams: Two metacognitive tools for science and
mathematics education, Instructional Science, 19, 29-52.
Novak, J. D. (1991) Clarity with concept maps, The science teacher 58(7), pp. 45-49.
Novak, J. D. The Theory Underlying Concept Maps and How To Construct Them,
http://cmap.coginst.uwf.edu/info/printer.html
Palmér, M., Naeve, A., Paulsson, F. (2004) The SCAM-framework – helping applications to store and
access metadata on the semantic web, Proceedings of the First European Semantic Web Symposium
(ESWC 2004), Heraklion, Greece, May, 2004, Springer, ISBN 3-540-21999-4,
http://kmr.nada.kth.se/papers/SemanticWeb/SCAM-ESWS.pdf.
Palmér, M., Naeve, A. (2005) Conzilla – a Conceptual Interface to the Semantic Web, Invited paper at the
13:th International Conference on Conceptual Structures, Kassel, July 18-22, 2005. To be published in the
series of Springer Lecture Notes on Computer Science.
Penner, D. (2001) Cognition, Computers, and Synthetic Science: Building Knowledge and Meaning
Through Modeling, In Secada (ed.) Review of Research in Education 25 (1-36), Washington D.C.
Polanyi, M. (1967) The Tacit Dimension, New York, Anchor books (based on the 1962 Terry lectures).
Powell, A., Nilsson, M., Naeve, A., Johnston, P. (2004) DCMI – Abstract Model, white paper, working
draft, 2004, http://dublincore.org/documents/abstract-model
Ravenscroft, Andrew. (2002) Communities, Communication and Cognitive Change: Social Processes and
Designing Engaging e-learning Discourse, http://www.aace.org/dl/
Rawlings, R., van Rosmalen, P., Koper, R., Rodriguez-Artacho, M., Lefrere, P. (2002) Survey of
Educational Modeling Languages (EMLs), Version 1, CEN/ISSS WS/LT Learning Technologies
Workshop, Sept 19, 2002.
Rumbaugh, J., Jacobson, I., Booch, G. (1999) The Unified Modeling Language Reference Manual,
Addison Wesley Longman Inc.
Sears, D.O., Peplau, L.A., Taylor, S.E. (1991) Social psychology. Englewood Cliffs, NJ: Prentice-Hall.
24
[68]
[69]
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
[80]
[81]
[82]
[83]
[84]
[85]
[86]
[87]
Shavelson, R. & Hunber, J., Stanton, G. (1976) Self Concept: Validation of Construct Interpretations,
Review of Educational Research 46, 407-441.
Sjöberg, L. (1997) Studieintresse och Studiemotivation – en analys av de grundläggande faktorerna.
Institutet för individanpassad skola, SAF, ISBN 91 7152 777X.
Shedroff, N. (1999) Interaction Design Course Syllabus, http://www.nathan.com/thoughts/course.html
Sperber, D., Wilson, D. (1995) Relevance, Blackwell Publishing (2:nd ed.).
Strong, M. (1996) The Habit of Thought: From Socratic Seminars to Socratic Practice, Chapel Hill, NC,
New View, 1996.
Sweller, John. (1999) Instructional Design in Technical Areas, Camberwell: ACER.
Takeuchi, H., Nonaka, I., (eds) (2004) Hitotsubashi on Knowledge Management, Wiley & Sons, 2004.
Vygotsky, L. (1962) Thought and Language, Cambridge, MA: MIT Press.
Vygotsky, L. (1978) Mind and Society, Cambridge, MA: MIT Press.
Vygotsky, L. (1986) Thought and Language, Cambridge, MA: MIT Press.
Wilson, S. (2001) The Next Wave: CETIS interviews Mikael Nilsson about the Edutella project, Feature
article, Centre for Educational Technology Interoperability Standards (CETIS), September 2001,
http://kmr.nada.kth.se/papers/SemanticWeb/TheNextWave.pdf
Wolpers, M., Grohman, G. (2005) PROLEARN: technology-enhanced learning and knowledge
distribution for the corporate world, International Journal of Knowledge and Learning (IJKL), Vol. 1,
Nos. 1-2, 2005, pp. 44-61.
Yli-Luoma, P. (1990) Predictors of Moral Reasoning. Stockholm: Almqvist & Wiksell International.
Yli-Luoma, P. (1992) Predictors of Critical Thinking Abilities – A Rasch Model Approach. Research
Bulletin 81. Department of Education, University of Helsinki.
Yli-Luoma, P. (1995) Hemmets inverkan på inlärning – Bindningsteorins tolkningsförmåga för elevers
socialisation och skolframgång. Helsingfors: Helsingfors Universitet, Pedagogiska Institutionen, Nr 144.
Yli-Luoma, P. (1996) Attachment theoretical predictors of school behaviour. In Yli-Luoma, P. (ed.) From
Metascience to Educational Policy. Helsiniki: Research Bulletin 94. Department of Education, University
of Helsinki.
Yli-Luoma, P. (2003) Hyvä opettaja (A Good Teacher) Helsinki: International Multimedia & Distance
Learning Oy Ltd.
Yli-Luoma, P. (2005) Kolbian Learning Styles as Predictors of Student Achievement A MetaAnalytical Study, To appear in The Journal of Educational Psychology.
Zohar, D. (1997) Rewiring the Corporate Brain: Using the New Science to Rethink How We Structure
and Lead, San Francisco, CA:Berrett-Koehler Publishers.
Zuboff, S. (1988) In the Age of the Smart Machine: The Future of Work and Power, New York: Basic
Books.
Web sites
KMR (Knowledge Management Research) group: http://kmr.nada.kth.se
CID (Centre for user oriented IT Design): http://cid.nada.kth.se/en
SCAM: http://scam.sourceforge.net
SHAME: http://kmr.nada.kth.se/shame
Conzilla: www.conzilla.org
VWE (Virtual Workspace Environment): www.vwe.nu
WGLN (Wallenberg Global Learning Network): www.wgln.org
PADLR: http://www.learninglab.de/padlr/index.html
Prolearn: http://www.prolearn-project.org
L3S (Learning Lab Lower Saxony): http://www.learninglab.de
Edutella: http://edutella.jxta.org
ULL (Uppsala Learning Lab): www.ull.uu.se
UDBL (Uppsala Data Base Laboratory): http://user.it.uu.se/~udbl
Soft Infrastructure for IT in education (MSU): http://mjukis.skolutveckling.se
Digital Media Library (UR): http://www.ur.se/mb
National Centre for Flexible Learning (CFL): http://www.cfl.se
Dublin Core Metadata Initiative: http://dublincore.org
IMS-LD (2003) Learning Design v. 1.0 (Final Specification),
available from http://www.imsglobal.org/learningdesign/index.cfm
[106] ISO/IEC JTC1 SC36: http://jtc1sc36.org
[107] Semantic Web: www.SemanticWeb.org
[108] RDF (Resource Description Framework): www.w3.org/RDF
[88]
[89]
[90]
[91]
[92]
[93]
[94]
[95]
[96]
[97]
[98]
[99]
[100]
[101]
[102]
[103]
[104]
[105]
25
UML (Unified Modeling Language): http://www.uml.org
Concept maps: http://www.graphic.org/concept.html
Topicmaps: http://www.topicmaps.org
e-Europe: http://europa.eu.int/information_society/eeurope/2005/text_en.htm
League of World Universities, Large City Universities and the Public School System, Rectors
Conference, New York University, April 19-20, 1993, www.nyu.edu/rectors/conf41993A.html
[114] Learn in Freedom, Books Critiquing the School System, http://learninfreedom.org/system.html
[115] Hestenes modeling website, http://modelingnts.la.asu.edu/html/Modeling.html
[109]
[110]
[111]
[112]
[113]
26
//////////////////////////
APPENDIX
27
///////////////////////////////
10.
CONCEPTUAL MODELING
10.1. The concept ’concept’
Def: A concept is a representation of something that we have experienced or that we can imagine, and which we
can apply to the objects that we are aware of.
Def: A description of the most important concepts - and their relations - within a specific problem domain is
called a conceptual model of the domain.
Def: The definition of a concept describes its intention, i.e. what qualities it aims to express and delimit with
respect to its surroundings.
Def: The set of objects that exemplify a concept is called its extension.
Def: Each member element of the extension-set is called an example = object = instance of the concept.
Def: The concept, whose extension consists of a set of instances, and whose intention describes their common
structure, is called the type or the class of these examples.
Def: To identify a concept by observing similarities and differences within a group of examples is called to
classify the examples.
Def: We say that a concept can be applied to a specific example, if this example fulfills the intention of the
concept.21
10.2. Properties of the concept ‘concept’
• A concept must always be defined by making use of other concepts.
• A concept can be denoted by one or several names (= symbols).
• A concept is always idealized, because it contains simplification that focus on some aspects and disregard
others.
• The definition of a concept always depends on the context within which it will be used. The aim is always to
disregard what is inessential and focus on what is essential within that context.
10.3. UML – a global modeling language
UML (Unified Modeling Language) [66], [109], is a language for specifying, visualizing and documenting
conceptual models within many different domains.
UML was developed during 1993-1997 within the object-oriented software industry as an attempt to unify the
250 different modeling languages that were in use by the mid 1990s.It represents a collection of practically
tested modeling techniques that have proven to be effective in the description of large and complex systems.
UML provides a “standardized visual language” where you can draw draw the most important concepts (and
their relations) within a specific problem domain. You get a visible background against which you can discuss
and where it is clear how you have been thinking up til now. This facilitates further development of the
conceptual model and increases the possibilities to “calibrate the model” and reach consensus with respect to
what should be regarded as important.
10.4. The Unified Language Modeling technique
Unified Language Modeling [39], [40] is a context-mapping technique, which has been developed by Ambjörn
Naeve during the past decade. It is designed to visually represent a verbal description of a subject domain in a
coherent way. Today, the ULM technique is based on the Unified Modeling Language [66], [109], which is a de
facto industry standard for systems modeling.
21
i.e. the conditions of its definition.
28
Figure 17.
The basic verbal/visual correspondence of Unified Language Modeling.
In ULM the resulting context-maps have a clearly defined and verbally coherent visual semantics, which makes
it easy to cognitively integrate the conceptual relations and achieve a clear overview of the context. Moreover,
making the context visually explicit provides important support for the conceptual calibration activities that form
an integral part of the learning process. The ULM verbal-to-visual contextual representation technique has a
crucial advantage in comparison with similar techniques - such as concept maps [110] or topic maps [111] which have to rely on purely verbal semantics in order to convey their conceptual relationships.
10.5. The modeling theory of David Hestenes
Here we will briefly describe the basic ideas that underlie the modeling theory of David Hestenes – a wellknown physicist and physics education researcher. Although he has applied his modeling theory mainly to
physics education [115] – and over the past two decades achieved striking results within this field - we share his
belief that his theory is applicable to learning in general.
Hestenes uses the term “conceptual learning” for the type of learning that is the opposite of “rote learning”. Here
follows a brief presentation 22 of Hestenes’ five general principles of conceptual learning that he has incorporated
into his instructional theory and applied repeatedly in the design of instruction.
• Conceptual learning is a creative act. This is the crux of the so-called constructivist revolution in education,
most succinctly captured in Piaget’s maxim: “To understand is to invent!” Its meaning is best conveyed by an
example: For a student to learn Newtonian physics is a creative act comparable to Newton’s original invention.
The main difference is that the student has stronger hints than Newton did.23
• Conceptual learning is systemic. This means that concepts derive their meaning from their place in a coherent
conceptual system. For example, the Newtonian concept of force is a multidimensional concept that derives its
meaning from the whole Newtonian system. Consequently, instruction that promotes coordinated use of
Newton’s laws should be more effective than a piecemeal approach that concentrates on teaching each of
Newton’s laws separately.
• Conceptual learning depends on context. This includes social and intellectual context. It follows that a central
problem in the design of instruction is to create a learning environment that optimizes the learner’s opportunities
for systemic learning of targeted concepts. The context for scientific research is equally important, and it is
relevant to the organization and management of research teams and institutes.
• The quality of learning depends on the conceptual tools. The quality of learning is critically dependent on
conceptual tools at the learner’s command. The design of tools to optimize learning is therefore an important
subject for educational research [20].
22
23
This presentation is condensed from [21].
Hestenes attaches the following warning to this (Constructivist) Learning Principle: ”There are many brands of constructivism, differing
in the theoretical context afforded to the constructivist principle. An extreme brand called “radical constructivism” asserts that constructed
knowledge is peculiar to an individual’s experience, so it denies the possibility of objective knowledge. This has radicalized the
constructivist revolution in many circles and drawn severe criticism from scientists. I see the crux of the issue in the fact that the
constructivist principle does not specify how knowledge is constructed. When this gap is closed with the other learning principles and
scientific standards for evidence and inference, we have a brand that I call scientific constructivism.”
29
• Expert learning requires critical feedback. Expert learning requires deliberate practice with critical feedback.
There is substantial evidence that practice does not significantly improve intellectual performance unless it is
guided by critical feedback and deliberate attempts to improve. Students waste an enormous amount of time in
rote study that does not satisfy this principle.
Figure 18.
A ULM context-map of Hestenes’ five learning principles.
The textual interpretation of Figure 18 is the following: Conceptual Learning is a Creative Act. Conceptual
Learning is (a) Systemic. Conceptual Learning depends on Social Context and Intellectual Context. Conceptual
Learning depends on Learning Tools, especially Modeling Tools. Conceptual Learning depends on Practice,
which consists of (= has) Deliberate Improvement Attempts and Critical Feedback.
In his response [21] to the Oersted medal reward in 2002, Hestenes writes:
”I believe that all five principles are essential to effective learning and instructional design, though they are
seldom invoked explicitly, and many efforts at educational reform founder because of insufficient attention to
one or more of them.”
”I see the five Learning Principles as equally applicable to the conduct of research and to the design of
instruction. They support the popular goal of teaching the student to think like a scientist.”
10.6. Educational Modeling Languages
In [65] an Educational Modeling Language (EML) is defined as: “A semantic information model and binding,
describing the content and process within a ‘unit of learning’ from a pedagogical perspective in order to support
reuse and interoperability.”
According to [30], p. 5, the need for an EML is based on the fact that the prevailing learning object models (as
expressed e.g. by IMS Simple Sequencing and IMS Content Packaging) express a common overall structure of
objects within the context of a unit of study, but they do not provide a model to express the semantic relationship
between the different types of objects in the context of use in an educational setting.
Several EMLs have been developed in order to remedy this deficiency [65]. Prominent among these is the
OUNL-EML developed by the Open University of the Netherlands [23], [30], which now forms the basis for the
standardization efforts of IMS Learning Design [105].
Although EMLs are important in several respects, they will not concern us much in this paper. This is mainly
due to the follwing reason: An EML descibes a unit of study in a formal way, so that automatic processing
becomes possible. This is an important general characteristic of information modeling, which aims to prepare for
the construction of an automated information system that should support some ongoing human activities at a
higher (informal) level.
In contrast, the models that are presented in this paper have no such aim. Instead they aim to facilitate effective
communication at the human level by highlighting important concepts of the activities that humans are engaged
in. Since our domain of interest is the management of the learning process, these activities involve the interplay
of processes, human (pedagogical) roles and tools.
30
Moreover, the EMLs presented and compared in [65] do not contain any support for dynamic (e.g. process or
activity) modeling 24. They introduce dynamic concepts (such as e.g., Learning Activity), but these concepts are
described only in terms of static UML models (class diagrams). This makes EMLs unsuitable for expressing the
kinds of process-related concepts that we are interested in exploring here.
11.
PROCESS MODELING: THE GOAP APPROACH
Here we will briefly describe the GOAP (Goal, Obstacles, Actions, Prerequisites) approach to process modeling.
It is often used in connection with business modeling25, but since our present ‘business’ is the learning process,
we can just as well apply it to the modeling of learning processes. This approach was not explicitly followed in
the CFL modeling work described above, but will be applied by CFL in the future in order to validate its course
development process model.
The GOAP approach starts out by modeling the goals of the business or organization. A specific goal in the
business is described as an object of a goal type. Goals can be quantitative or qualitative, and both of these goal
types have a goal description as an attribute. A quantitative goal also has a goal value.
The relationships between goals are dependencies and associations. A dependency is represented in UML by a
dashed line from the larger (dependent) goal to the smaller (partial) goal ending with an open arrow. The
dependency should be interpreted as stating that the fulfillment of the smaller (partial) goal contributes towards
the fulfillment of the larger (dependent) goal. If a larger goal can be completely broken down into smaller
(partial) goals, a dashed line is drawn across the dependencies and a constraint is written next to the line in this
form: {complete}. If the goal cannot be completely broken down into partial goals, {incomplete} is written (this
is also the default if nothing at all is written).
A goal that has been completely broken down into partial goals indicates that the goal will automatically be
fulfilled if all of the partial goals are met. This is what could be seen as a logical AND condition between the
partial goals. The <<contradictory>> stereotype can be used to denote mutually exclusive goals. Typical
examples of contradictory goals are “high quality” and “low cost”. Contradictory goals normally cannot both be
fulfilled, but must both be taken into consideration.
In connection with describing the goals we also describe the obstacles that stand in their way. An obstacle is a
problem that hinders the achievement of a goal. By finding the problem, new goals or partial goals are
discovered that attempt to eliminate the problem. An obstacle is therefore always linked to a goal. Similar to a
goal, an obstacle can also be broken down into partial obstacles. By modeling the connections between the goals
and the obstacles, we construct a goal/obstacle model, which can be expressed in a goal/obstacle diagram.
An obstacle can be informally specified in a note with the stereotype <<problem>> attached to its goal object.
An obstacle can be a temporary problem that can be solved once and for all, or it can be a continuous problem
that requires continuous action in order to prevent the problem from reoccurring.
Obstacles are eliminated (overcome) by actions. An action plan can be formulated from the goal/obstacle model,
where temporary obstacles are resolved as soon as possible, and the goals linked to the continuous obstacles are
allocated to processes in the business.
The action plan should contain:
1) a list of the obstacles,
2) the cause of each obstacle,
3) the appropriate action for each obstacle,
4) the prerequisites for each action, and finally,
5) the resource or process responsible for overcoming it.
This can also be shown visually in the goal/obstacle model through the use of stereotyped notes. The stereotypes
<<problem>>, <<cause>>, <<action>>, and <<prerequisite>> specify the purpose of the note. In UML all of
these optional concepts are defined through the use of informal notes, because they are normally described in
24
25
as described e.g. in [14].
A more detailed description of this use can be found e.g. in [14].
31
simple text and cannot be formally defined. The problem-note is attached to the goal, the cause note to the
problem-note and so on. The example presented in the figure below is taken from [14]:
As described in [14], a technique for identifying goals is to ask these questions:
• Why should we achieve this goal?
• How should we achieve this goal?
The answers to the why question will identify higher goals (goals to which the current goal is a partial goal), and
the answers to the how question will identify partial goals. Answering these two questions makes it possible to
identify new goals from existing goals, and to reveal additional goals that might not have been discovered by the
people in the business.
All the primary goals of the business can be summarized in a diagram of their own, where any conflicts between
the goals (contradictory goals) are shown. Each of the primary goals can then be described in a diagram of its
own, with its corresponding partial goals. It is important to realize that the goal/obstacle diagram should not be
over-formalized or described in too computational terms. The purpose of the goal/obstacle diagram is to identify
and structure the different goals of the business and to break down the goal descriptions to a level at which the
corresponding goals can be allocated to individual processes.
12.
THE ONSCAIL NETWORK
The ONSCAIL network involves a number of stakeholders (shown in the figure below) that have teamed up and
are now jointly contributing to a Public e-Learning Platform (PeLP) 26. These stakeholders include the Swedish
Educational Broadcasting Company (UR), the Swedish National Agency for School Improvement (MSU), the
Swedish National Agency for Education (SV), the Swedish National Centre for Flexible Learning (CFL), the
Swedish Terminology Centre (TNC), the Swedish Museum Window (SMW), and the Stockholm Public
Education Management Agency (UTBF). The Soft infrastructure for IT in education project of MSU [101], the
Digital Media Library of UR [102], and the Learning Resource Centre of CFL [103] are three of the important
stakeholders projects in the PeLP.
The present stakeholder roles of the ONSCAIL network are listed below, together with the organizations that are
implementing these roles:
• Content provider: UR, SV, UTBF, Museum Window
• Service provider: MSU, UR, Ateles.
• Course developer: CFL.
• Terminology integrator: TNC
• Technology Integrator: DataDoktorn, Uppsala Learning Lab, Uppsala Database Laboratory, Swedish
Netuniversity.
The overall idea of the ONSCAIL network is that each participating organization should work out a conceptual
model (ontology) that describes its corresponding role(s). This will create a set of open role-interfaces that
describe how to play each one of the roles within the network. In this way, any organization that wants to
participate, will be provided with an explicit interface for the role that it wants to play, and by implementing this
interface, the organization can take on the corresponding role and interoperate with the other stakeholders of the
network.
12.1. Enabling semantic cooperation among content providers
Here we will describe a metadata interoperability modeling effort, which has been performed within the
ONSCAIL network in order to lay the foundations for future semantic collaboration among content providers.
This modeling work was carried out within a Swedish standardization group for learning technology called
26
By an e-learning platform we mean a system that contains the functionalities of an LMS (Learning Management System) and/or an LCMS
(Learning Content Management System). An LMS typically contain features for administration, assessment, course management, possibly
content management and authoring. An LCMS typically emphasizes content management/authoring and includes many features of an LMS.
In generic terms, we will describe an e-learning platform in terms of its infrastructures, architectures, frameworks and tools.
32
TK450, which is part of ISO-JEC/SC36. During the spring of 2004, Ambjörn Naeve mediated a series of
modeling sessions involving UR, SV, CFL, and MF, who were represented by their respective metadata experts.
The aim was to find out the similarities and differences of the respective description structures that were used
among them, as well as to decide on some common terms for things that were perceived (thought of) in the same
way. The documentation of this modeling process is available (in Swedish) at
http://knowgate.nada.kth.se:8080/portfolio/main?cmd=open&manifest=amb&uri=scam%3A%2F%2Famb%2F1
776c
12.2. Modeling the CFL course development process
During the spring of 2004, Ambjörn Naeve conducted a distance course in conceptual modeling with
participators from CFL, which is located in two different Swedish cities, Norrköping and Härnösand. The
process modeling part of the course worked with the example of modeling the CFL course development process
– as described in a number of internal text documents. As the modeling processed, several discrepancies and
inconsistencies of these textual descriptions were uncovered, and the participants became enthusiastic about the
conceptual clarity that grew out of the visual models. This kind of process modeling is gaining increased
attention within CFL and is rapidly becoming established as a working method within the departments of CFL
that participated in the course.
Although the modeling was carried out in Swedish, the resulting CFL course development process model has
been translated into English and uploaded to the CFL confolio27 at
http://knowgate.nada.kth.se:8080/portfolio/main?cmd=open&manifest=CFL&uri=urn%3Axsmac.nada.kth.se%3ACFL%3A12
Screenshots of the model have been included below, but in order to be able to read the text properly, the online
slides at the URL above are recommended.
27
To view the slides properly, press "PP-view" (= PowerPoint view), which lets you go through the slides without "jumping in and out" of
the portfolio folder.
33
34
A notational novelty in these process diagrams is the introduction of a ‘musical-note’ inspired system for
describing the participation of different competencies (competency roles) in the various parts of the process. An
unfilled note denotes participation, and a filled note denotes participation and responsibility. More precisely, if
the corresponding vertical arrow (connecting the competencies with the information objects and the subprocesses) is pointing upwards, then the filled note means that the corresponding competency role is responsible
for initiating the info object or the sub-process, and if the vertical arrow is pointing downwards, the filled note
means that the corresponding competency role is responsible for approving the info object or the results of the
sub-process.
13.
THE OPEN-LMS ASSESSMENT PROJECT
The openLMS assessment project of open source based L(C)MSes is carried out for the Swedish Netuniversity
under the coordination of Uppsala Learning Lab [99].
13.1. Objectives
The openLMS project is performing a survey-style assessment of open source based systems for technologyenhanced learning (here collectively referred to as Learning Management Systems). The overall objective of the
project is to facilitate the exchange of experiences of such systems in order to enable cost-effective and qualityenhancing collaborative development of such systems and promote their use within Swedish universities and
university colleges. Hence the survey focuses on assessing the “potential for interoperability” of open source
based LMSes from a technical perspective.
13.2. Scope
The study focuses exclusively on open source LMS projects where the software technology and the development
framework used is founded on open source code, open standards, open development tools and open operating
systems with no limitations on distribution and use – e.g. GPL, LGPL type of license. Consequently commercial
products are explicitly excluded from the assessment. For a LMS project to be considered for this evaluation, it
must show results in the form of code, functional LMS application(s) and documentation. Also, the project must
be focused on learning management – and not just content management.
13.3. Process and method
The study has been carried out in accordance with the following process:
• Find the initial technology candidates that fit the defined scope.
35
• Identify criteria/metrics and use cases useful in evaluating open source LMSes.
• Assess the selected technology candidates using software technology criteria.
• Select the top third of these candidates as quality candidates for quality evaluation.
• Further assess these candidates using ISO9126 software quality criteria.
• Select the top third of these candidates as UC candidates for testing based on Use Cases.
13.4. Practical user testing
The UC candidates were hands-on tested through demo installations available on the respective project’s web
site and building, installing and configuring the software from open source code repositories. User accounts with
admin/teacher and student roles were created, inspection of the online help, user and instructor documentation,
and commentary of the user community were preformed in order to assess the UC candidates.
A group of teachers and students then tested the basic functionality of the UC candidates and answered a set of
questions based on groups of use cases. The questions were organized in a “question repository” consisting of
seven different groups - based on use cases that reflect the basic functionalities of an LMS. These groups were
Registration, Scheduling, Delivery, Tracking, Assessment, Content handling, and Activity handling. From this
question repository, two different questionnaires were constructed, one for administrators/teachers and one for
students. The questions were constructed around a common template formulated as “It is easy to carry out X”,
where X refers to some functionality of an LMS. The answers were given in the form of check-boxed
“percentage of agreement”: I agree to 0%, 20%, 40%, 60%, 80%, 100% and “not applicable”. The testers were
also encouraged to write down their qualitative impressions in the form of test-notes. At the time of writing these
user tests are still being evaluated.
Figure 19.
The Sakai LMS technical profile
36
Figure 20.
An evolutionary model for user-centered query shaping of LMS tests.
37
View publication stats