J Syst Sci Syst Eng (Jun 2009) 18(2): 203-221 DOI: 10.1007/s11518-009-5095-0
ISSN: 1004-3756 (Paper) 1861-9576 (Online) CN11-2983/N
TOWARD COLLECTIVE INTELLIGENCE OF ONLINE COMMUNITIES: A PRIMITIVE CONCEPTUAL MODEL∗ Shuangling LUO1
Haoxiang XIA2
Taketoshi YOSHIDA1
Zhongtuo WANG2
1
School of Knowledge Science, Japan Advanced Institute of Science and Technology, Ishikawa 923-1292 Japan {slluo,yoshida}@jaist.ac.jp 2
Institute of Systems Engineering, Dalian University of Technology, Dalian, 116024, China
[email protected] ( )
[email protected]
Abstract Inspired by the ideas of Swarm Intelligence and the “global brain”, a concept of “community intelligence” is suggested in the present paper, reflecting that some “intelligent” features may emerge in a Web-mediated online community from interactions and knowledge-transmissions between the community members. This possible research field of community intelligence is then examined under the backgrounds of “community” and “intelligence” researches. Furthermore, a conceptual model of community intelligence is developed from two views. From the structural view, the community intelligent system is modeled as a knowledge supernetwork that is comprised of triple interwoven networks of the media network, the human network, and the knowledge network. Furthermore, based on a dyad of knowledge in two forms of “knowing” and “knoware”, the dynamic view describes the basic mechanics of the formation and evolution of “community intelligence”. A few relevant research issues are shortly discussed on the basis of the proposed conceptual model. Keywords: Community intelligence, online community, knowledge supernetwork, knowing, knoware
1. Introduction
colony, highly-intelligent collective activities of
In recent years, the concept of collective
the whole colony may emerge from the local
intelligence has been widely discussed from
interactions between the individual ants, which
various aspects. One series of related work is
embody very limited intelligence per se.
inspired
Intelligence”
Accordingly, various bio-inspired algorithms
phenomena that can commonly be observed in
have been developed to solve different social
the biological world. For example, in an ant
and technical problems (e.g. Bonabeau et. al.
by
the
“Swarm
∗
This work is supported in part by National Natural Science Foundation of China under Grants No. 70431001, No. 70620140115, and No. 70871016, respectively. H. Xia would also appreciate the financial support from Chinese Scholarship Council to conduct this international collaboration. Part of the paper was presented in the conference of IEEE SMC 2008.
© Systems Engineering Society of China & Springer-Verlag 2009
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
204
1999, Kennedy & Eberhart 2001). Inspired by
Second, some groups like business organizations
such Swarm Intelligence, an interesting question
may be well-structured, while some other groups
may be raised: what the collective activities
like
would be in case that the individuals are smarter
weakly-tied. Therefore, a way of research is to
and more knowledgeable than the individuals of
respectively
those biological swarms. In another word, it
collective intelligences with respect to the
would be curious to study the collective
characteristics of the groups, e.g., to study the
intelligence of human groups.
collective intelligence of a team, of a business
Similar with the Swarm Intelligence concept,
online
discussion study
forums
different
are
more
categories
of
organization, and of a special interest group.
underpinning the collective intelligence of
The study of the team-level collective
human groups is the idea that a human group
intelligence or “team intelligence” (e.g. Akgün
may
of
et al. 2008) can be shown in many research
manifest
higher
capabilities
problem-solving
efforts, basically at the intersection of group
than any individual participant of that group
cognitive psychology and teamwork. At the
does, especially when the participants densely
business level, “business intelligence” has
interact
the
become one of today’s buzzwords since it was
computerized communication channels such as
firstly coined in the 1950s (Luhn 1958). Today
the Internet and the World Wide Web. Various
most of the
related endeavors have already been reported as
intelligence” is focused on developing “smart”
a subject of interdisciplinary research at the
IT applications to support business work, whilst
intersection of sociology, psychology, business,
the
and computer science (e.g. see Levy 1994,
intelligence”
Brown & Lauder 2000, Segaran 2007, Szuba
“organization learning” (Argyris & Schon 1978),
2001, MIT CI Center 2008, Johnson 1998).
“collective mind” (Weick & Roberts 1993), etc.
However, despite all those endeavors, the
To enlarge the research scope, some researchers
collective intelligence of the human groups is
have also studied the global-level collective
much more complex than that of the biological
intelligence or the “Global Brain” (Russell 1983,
swarms, and we are now merely at the beginning
Mayer-Kress and Barczys 1993, Heylighen
stage
1995), making an analogy between the network
information-processing
with
of
the
each
inquiry
and
other
of
through
such
collective
intelligence.
work entitled as “business
human-centered are
studies generally
on
“business
branded
as
of billions of individual human beings and
One critical issue for the study of human
computers in the entire human society and the
collective intelligence is to clarify the universe
network of billions of neurons within one single
of discourse. Generally speaking, all types of
human brain.
human groups or collections can be regarded as
The studies on collective intelligence at the
the subject of research. However, human groups
team level, the business level and the global
may significantly differ from one another. First,
level are no doubt important. Besides, the
the size of a group may range from a small team
collective intelligence within another category
to the entire human society of billions of people.
of human groups is also worthy of investigation,
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
205
i.e. the collective intelligence at the community
theoretical framework from the knowledge and
level, or “community intelligence” as we may
intelligence perspective is still missing in
term. Conceptually, this community intelligence
community studies. There is hence an academic
is located in the middle of a spectrum of
need to construct such a theoretical framework,
collective intelligence from team intelligence
and it may have profound practical implications.
and business intelligence to global intelligence.
Based on the prior considerations, the major
Informally speaking, a “community” refers
focus of the present paper is to incorporate the
to any human group in which the members have
idea of collective intelligence into the studies of
some common characteristics or share some
human communities so as to give an attempt to
interests; for example, it can be formed around
sketch out the basic conceptions of community
the people who have similar hobbies, who share
intelligence. Especially, due to the rapid growth
specific academic interests, or who participate in
of the online communities in which the
a same BBS forum. Such communities are
computer-mediated communications (CMC) are
playing an increasingly important role in our
the prominent means for the community
society.
the
members to interact with each other, in this
explosion of the World Wide Web is the rapid
paper more attention is paid to the intelligence
growth of the different categories of online
of these online communities. With the support of
communities or virtual communities (Rheingold
the
2000), and the collective activities of those
technologies (ICT), the online communities may
virtual
exhibit
Especially,
communities
accompany
often
with
manifest
high
information
and
communication
higher intelligent features than a
problem-solving capabilities. For example, in
traditional community does since ICT firstly
the
a
provides an effective communication channel for
loosely-connected community of programmers
massive exchange of data, information and
who don’t even know each other can collectively
knowledge
develop very complex software products like the
capabilities of the modern ICT may be of great
Linux operating system. In this sense, the
help for the information processing tasks within
exploration of such community intelligence is
the entire community.
Open Source
Software
movement,
and
secondly
the
computation
doubtlessly of practical value. As an academic
This paper is organized as follows. In
response to the increasing role of communities
Section 2 the concept of community intelligence
to human society, a number of relevant research
will be explored under the backgrounds of
efforts have been conducted from various
community
disciplines; and many of the recent efforts on
Thereafter, a theoretical model of community
communities, particularly those around the
intelligence will be described in Section 3,
“community of practice” (Lave & Wenger 1991),
which contains a static or structural view and a
the “community of interest” (Fischer 2001) and
dynamic view. We will shortly discuss the
the like, are closely related to the concept of
corresponding research issues in Section 4 and
community intelligence. However, a clear
conclude the paper in the final section.
and
intelligence
researches.
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
206
2. Community Intelligence under Backgrounds of Community and Intelligence
leads to the increasing interests on the online communities over the Internet and the Web. Third, the growth of communities and the
One major aim of this paper is to make a
development on the supportive information
plea for the attention to the research subject of
technologies greatly stimulate the corresponding
community intelligence or collective intelligence
studies in the pedagogical and managerial fields,
of online communities. The phrase “community
resulting in the extensive discussions on a bunch
intelligence”
of
is
literally
comprised
of
topics
like
“learning
communities”
“community” and “intelligence”, each of which
(Gabelnick et al.1990), “community of practice”
has been a hot research topic for a long history.
(CoP) (Lave & Wenger 1991), “community of
Therefore, in this section we attempt to discuss
interest” (Fischer 2001), and “knowledge-
the essential characteristics of community
building communities” (Scardamalia & Bereiter
intelligence by examining the basic concepts
1994). The latter two aspects have a close
behind “community” and “intelligence”.
conceptual
connection
to
community
intelligence presented in this paper.
2.1 Community Intelligence in the Context of Human Communities
Underlying the endeavors related to learning communities and CoPs is a shift about the focal
Let’s first try to conceive community
point of learning research, from focusing on
intelligence by observing the “communities”. As
individual information-processing and cognitive
informally defined in the prior section, a human
development to emphasizing “situated learning”
community is any human group in which the
(Lave & Wenger 1991) within the context of
members have some common characteristics or
communities and other social groups. Since
share some interests; the participants of the
learning is one key facet of intelligence, the
community generally have some identity to the
concept of community intelligence is actually a
community. Since German sociologist Ferdinand
natural extension of this “situated learning”
Tönnies’ (1887) conception of Gemeinschaft
viewpoint;
(usually translated as “community”), human
community intelligence is that intelligence is
communities have long been a subject of
inherently social rather than something inside an
research in various disciplines such as sociology,
individual’s brain. Associated with this social
anthropology, and psychology. More recently, as
aspect is an “emergence” aspect of intelligence,
a reflection to the increasingly vital role that
which is the key aspect that differentiates our
they play in society, communities have caught
concept of community intelligence from the
wider attention, most significantly in three
well-discussed “situated learning” concepts. In a
aspects. First, with the rising of the complex
community, the participants exchange their
network research, communities have become
opinions and expertise during the collaborative
one key subject of social network analysis (e.g.
learning and problem-solving processes. The
Wellman & Wortley 1990, Girvan & Newman
collective-level intelligence may then emerge
2000). Second, the rapid growth of the Internet
from such knowledge exchange activities,
and
the
central
point
behind
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
207
analogous to the emergence of the collective
“memory” system. Third, incorporating all sorts
intelligence of the biological swarms.
of
Thus, the efficient and effective knowledge
computing
technologies
and
(e.g.
information the
processing
Semantic-Web-based
exchange is critical for the formation of
reasoning tools, Web Services and other
community intelligence. The rapid development
Web-based applications), the Web platform has
of the Internet and the Web may greatly facilitate
obtained some capability of intelligence in its
the inter-personal knowledge exchange within a
own right, and such Web intelligence (Zhong et
community; consequently, the virtual and online
al. 2003) may be furthermore combined with
communities are of particular importance.
participants’
Following Porter (2004), “a virtual community
higher-level community intelligence.
human
intelligence
to
form
of
To sum up, the rapid development of ICT has
individuals or business partners who interact
boosted great research interests on human
around a shared interest, where the interaction is
communities, in particular on virtual and online
at least partially supported and/or mediated by
communities.
Under
this
technology and guided by some protocols or
learning-based
view
on
norms.” The virtual community and the online
furthermore be combined with the emergence
community are usually treated as synonyms.
aspect as is in the case of Swarm Intelligence in
However in this paper we would refer an “online
natural systems; and the communities can hence
community” to the community that uses the
be regarded as a collective intelligent system
computer network (and in particular the Web) as
that is comprised of human beings and the
the
supporting
is
defined
primary
herein
as
an
aggregation
communication
media,
while
ICT
background,
the
communities
can
systems.
The
basic
regarding “virtual community” as a more
characteristics of such community intelligence
general phrase referring to the community that is
can then be discussed from the “intelligence”
mediated via all sorts of telecommunication
point of view.
technologies. The discussion of community intelligence
is
then
focused
on
the
Web-mediated online communities as the Web
2.2 Community Intelligence from the “Intelligence” Aspect
has shown great advantages to become the base
In accordance with the prior description, we
platform to enable “true” intelligence. First,
may furthermore try to characterize community
nowadays the Web has grown to be a pervasive
intelligence by comparing it with some other
platform for wide-area communication and
types of “intelligence”.
collaboration,
facilitating
Intelligence is an everlasting topic of
knowledge (and information) exchange and
philosophical debate and scientific inquiry. It is
fusion (i.e. consensus building) between the
beyond the scope of this paper to give a
community members. Second, the Web itself has
thorough discussion on the Intelligence-related
become a massive distributed inventory of
researches. Instead, we give a simplified
information and knowledge so that it partially
4-part-division for the intelligence research field,
fulfills
as shown in Figure 1, trying to “position”
the
hence
greatly
functionality
of
a
distributed
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
208
community intelligence in the landscape of
intelligence within the intelligence research
“intelligence” researches.
space, as also shown in Figure 1. Therewith, the essential characteristics of
Collective Intelligence Community Intelligence (CI) Biological Intelligence
community intelligence can further be discussed by comparing it with other related categories of
Machine Intelligence
“intelligence”;
and
three
distinctive
characteristics may be identified as to be discussed below.
Individual Intelligence
1) Community Intelligence IS Mixed Collective Intelligence
Figure 1 Community intelligence in the
At the first glance, community intelligence
two-dimension-space of intelligence research
belongs to “Collective Biological Intelligence”, In Figure 1, the intelligence research field is
and this differentiates community intelligence
divided by two axes. The first axis divides the
from Web intelligence and the like, which
field
into
“Biological
Intelligence”
and
belong to the category of “Collective Machine
Intelligence”.
Intelligence
was
Intelligence”. However, community intelligence
the
is not purely biological intelligence. In a
problem-solving capabilities of living creatures,
Web-mediated online community, various ICT
in particular of human beings. We term this type
technologies that may be branded as Web
of intelligence as “Biological Intelligence”, in
intelligence play a critical role to support the
contrast to the “Machine Intelligence” which
collective human intelligence of that community.
was introduced since the blooming of artificial
In this sense, community intelligence is the
intelligence in the 1950s. In the second axis, the
mixed intelligence of “Biological Intelligence”
intelligence research can be classified into
and “Machine Intelligence”.
“Machine
conventionally
Individual
considered
Intelligence
as
and
Collective
Intelligence. This division reflects the shift of the
underlying
assumption
of
intelligence
2) Community Intelligence Is Emergent Intelligence
research from “cognition is something inside an
The second key feature of community
individual’s head” to “mind is social” (see
intelligence is that it is self-organizing and
Preface of Kennedy & Eberhart 2001). Taking
“emergent”; and this may be the distinctive
both axes into account, the intelligence research
feature of community intelligence from “team
field can then be grouped into four parts, i.e.
intelligence” and “business intelligence”.
Individual Biological Intelligence, Individual
Community
intelligence
shares
some
Machine Intelligence, Collective Biological
important characteristics of team intelligence,
Intelligence,
Machine
especially in the sense of shared mental models
Intelligence. Keeping these four parts in mind,
(Cannon-Bowers et al. 1993, Klimoski &
we may give a rough positioning of community
Mohammed
and
Collective
1994).
However,
the
major
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
209
difference is that team intelligence usually
to model the entire human society as a “brain”.
involves a small group of people and thus it
Comparatively, community intelligence is more
lacks the “swarm effect” that the higher-level
investigable for two reasons: first, the size of a
intellectual
community is usually much smaller than that of
capabilities
emerge
from
the
the entire human society so that the evolution of
interactions of massive participants. One critical difference between a human
“community intelligence” is more traceable;
community (particularly an online community)
second, the members of a community usually
and an institutional organization (e.g. a business
have shared interests and this makes the
company or a governmental agency) is that the
knowledge structure at the aggregate level more
community is more open and flexible. In
focused and more utilizable to actual problem
contrast to the fixed structure (commonly
situations. For this reason, it is the authors’
hierarchical) of an institutional organization, the
contention
community boundary is somehow vague and
intelligence might be a first step to reify the bold
people have more freedom of joining and
vision of the “global brain”.
that
the
study
of
community
leaving the community. As a result, fresh new
In general, the studies of team intelligence,
source of ideas and knowledge may then be
business or organization intelligence (namely the
brought in together with the recruitment of the
human
new members; and this continual flowing-in of
community intelligence and the global brain
new ideas and knowledge is beneficial for
may form a full spectrum of human collective
knowledge innovation inside the community. In
intelligence research. Community intelligence is
this sense, the collective intelligence of an
located somewhere in the middle of this
institutional
better
spectrum. Such middle position may be an
designed and structured, while community
appropriate position in order to study the
intelligence is more dynamic, and better fitted
complexity and the emergent properties of
the characteristics of self-organization and
collective intelligence.
organization
might
be
aspect
of
business
intelligence),
“emergence” as being broadly discussed in the context of complex adaptive systems. In another word, community intelligence is more suitable
3) Community Intelligence Is Knowledgebased Intelligence
to be modeled as a “neural network” of the
The “emergence” property is the common
individual participants, analogous to a human
feature between community intelligence and
brain that is a network of the biological neurons.
Swarm Intelligence in the biological world. The
This feature is generally absent in the current
major difference of these two types of collective
discussions about the organizational intelligence.
intelligence is twofold. First, the participants or
As being mentioned in the introduction
individuals within a human community are
section of this paper, the neural-network
humans, who are knowledgeable and intelligent
metaphor is also adopted in the vision of the
by themselves, while the elemental entities of
“global brain”. However, the study of the global
Swarm Intelligence are usually with very limited
brain may encounter great practical difficulties
intellectual capabilities. Thus, the community
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
210
intelligent system is essentially a system of
intelligence is based on the shared knowledge
massive intelligent systems (i.e. human beings);
structure of the entire community. Accordingly,
on the contrary the Swarm Intelligence system is
our modeling effort in this paper basically
an emergent intelligent system of massive
involves two parts. The macroscopic structure of
unintelligent entities.
community intelligence is firstly discussed.
Second, the interaction modes in a biological
Then, we try to describe how such macroscopic
swarm are different from the modes in a human
intelligent system forms and evolves through the
community.
microscopic
The
emergence
of
Swarm
Intelligence is usually from the stigmergy-based
knowledge-based
interactions
between the participants.
interactions between the individuals (Balis et al. 1967, Bonabeau et al. 1999). In contrast, the interactions in a community intelligent system
3.1 The Structural Model of Community Intelligence
are knowledge-centric; and the emergence of
Regarding the community intelligent system
community intelligence is essentially due to the
as a knowledge-based intelligent system, our
knowledge exchange and knowledge fusion
contention is that a Web-mediated online
activities. Therefore, as the key research theme
community may self-grow to have at least three
of Swarm Intelligence is to study how the
essential characteristics. First, the community
micro-level stigmergy-based interactions give
should, generally in a distributed fashion,
rise to the emergence of intelligent activities at
contain a memory system that stores information
the macro-level, a fundamental theme of
and knowledge, analogous to the memory
community intelligence is how the community-
system
level intelligence may generate from the
community should have the capability of
knowledge-related activities of the participants
“intelligent” problem-solving, i.e. the capability
or the community members.
of utilizing the stored knowledge to solve
in
a
human
brain.
Second,
the
problems; and the community should commonly
3. A Knowledge-based Model of Community Intelligence
exhibit higher-level intelligent capability than any community member per se. Third, as we
A bunch of related research issues may be
will explain in Section 3.2, this memory system
raised from the prior discussions on community
dynamically evolves, analogous to the cognitive
intelligence. One key issue is to develop a
development of an individual human mind.
theoretical model to answer what the community
Comparing to the human brain which is
intelligence is and how it develops. An attempt
comprised of the interconnected neurons, the
is given in this section to build such a model.
community intelligent system can also be
As discussed in the previous section, our
viewed as a neural network in which the
basic view is that community intelligence is the
“neurons” are the community members (i.e.
emergent collective intelligence based on the
human participants) together with the underlying
local knowledge exchange activities of the
computer systems. Furthermore, this “neural
participants; and the emergent community
network” is not just a network, but a network of
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
211
networks or a “supernetwork” (Nagurney &
support human communications and interactions.
Wakolbinger
2005).
This Web-based media network is interwoven
supernetwork
basically
This
knowledge triple
with the human network in that it provides a
interwoven networks, namely a technological
cyberspace to promote human communications
network or
consists
of
supports
and collaborations. The nodes of this media
information and knowledge transfer (in a
network (i.e. the Websites) are created by
Web-mediated online network, the Web is the
humans (i.e. the nodes of the human network)
major component of this technological network),
and accessed by other nodes of the human
a human network of community members, and a
network (i.e. other community members).
content network of knowledge and information
Through the Web-mediated communications,
(called the “knowledge network”) which is
knowledge creation, sharing and fusion happen
hosted in humans and computer systems. The
within the human network or the human
structure of such knowledge supernetwork is
community. The collective activities of the
illustrated in Figure 2.
community member would then give rise to the
media
network
that
Knowledge Network
Human Network
emergence of the third network, i.e. the “knowledge network”. The knowledge
network embodies the
collective knowledge of the community. As having been widely discussed in Knowledge Engineering, knowledge may essentially be represented as a networked structure, for example, a semantic network of concepts and Media Network (primarily the Web)
predicative relations, a linked structure of a set of reasoning rules, or elements interconnected
Figure 2 Illustration of the supernetowrk of
by a cognitive schema or a mental model. Thus
community intelligence
the overall knowledge content of a community can naturally be viewed as a network of
In this supernetwork, the media network
“knowledge elements”, or the “knowledge
itself is a two-tier supernetwork comprised of
network”. This knowledge network is an abstract
the physical network of the computers (i.e. the
network with its components (or fragments)
Internet) and the logical network of Websites (i.e.
being actually embedded in the human brains
the Web). To simplify the discussion, here we
(i.e. the nodes of the human network) and the
just consider the logical level structure of this
Websites (i.e. the nodes of the media network).
technological network, i.e. the Web (or more
The overall knowledge network is structured by
strictly a fraction of the Web that is connected to
the conceptual and logical connections of these
the specific human community) and we call it as
knowledge components.
media network since it is the primary media to convey knowledge and information and to
Some components or fragments of the knowledge
network,
which
are
basically
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
212
classified as codified knowledge, are contained
Individual with Personal Knowledge
in the nodes of the technological network (i.e. the Websites). One Website may connect to
Web-mediated Interactions Knowledge network
multiple nodes of the knowledge network; and one node of the knowledge network can vice versa be stored in multiple Websites. The connections between the Websites and the
Mapping
fragments of the knowledge network exhibit the basic means to link the technological network and the knowledge network. The knowledge network is also connected to the human network
Figure 3 Virtual “memory” structure of the community
in that the largest proportion of the overall
may emerge by interconnecting of the personal
knowledge of the community is still stored in
knowledge of the members. In general, there
human brains although the media network may
exist two basic ways to interconnect the personal
help
codified
knowledge. First, through communications and
knowledge. A single member of the human
interactions, one community member may get
community connects to a sub-network of the
the “know-who” knowledge of some other
entire knowledge network, which actually maps
members; thus he or she may direct seek help
to the personal knowledge structure residing in
from them when encountering problems. By
the member’s brain. The whole knowledge
such
network is then the union of all these
knowledge links may be established between
sub-networks, indicating that the knowledge
these members. The second way of knowledge
network as a whole serves as the “memory” of
interconnection is an indirect way. The members
the intelligent system of the community, which
can “externalize” their personal knowledge to
can furthermore be regarded as the collective
the media network by, for example, posting
“memory”
Such
articles in some Web-space. By reading the
“memory” structure can further be illustrated by
articles, some other members may assimilate the
Figure 3.
embedded knowledge into their own knowledge
contain
of
some
all
amount
the
of
participants.
As shown in Figure 3, the community
direct
knowledge
exchange,
virtual
structure. In this way, the media network then
members have their own knowledge residing in
intermediates
their personal memory systems. Furthermore,
structures of the different community members.
collective knowledge creation, transmission, and
By these direct and indirect connections, the
fusion
take
place
through
to
connect
the
knowledge
Web-mediated
knowledge network of the whole community,
interactions amongst the community members.
which is actually a virtual network embedded in
Through these collective knowledge activities,
the human network and the media network, may
on one hand the community members update
form from the individual knowledge structures
their personal knowledge; on the other hand, a
of all the community members. The structure of
virtual knowledge network at the collective level
this knowledge network is the logical union of
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
213
the knowledge structures of all the community
and transferred between the individual and the
members.
organization levels. Despite this salient work
The
previously-described
networked
and many further developments thereafter, e.g.
“memory” structure of the online community
Wierzbicki and Nakamori’s (2005) “trefoil”
also implies the utilization of knowledge in the
knowledge creation model, the very basic
solving of the real-world problems. Analogous
notions behind knowledge and knowledge-
to
integrating
related activities and processes are still under
different knowledge sources in human memory,
human
problem-solving
debate in academia. Cook and Brown (1999)
the basic process of “intelligent” problem-
criticize that in the models proposed by Nonaka
solving in the community is accomplished by
as well as other organization-learning advocators
integrating different knowledge sources from the
what is taken into account is “knowledge” being
collective “memory” of the community. Such
regarded as something that people “possess”
knowledge integration and utilization is further
whilst the actions people do during epistemic
accomplished
by
direct
by
indirect
work are largely neglected. Therefore, regarding
communications between the participants or
and
“knowing” as “the epistemic work that is done
community members.
as part of action or practice”, they draw a distinction between knowledge and knowing,
3.2 The Dynamic View of Community Intelligence Based on the prior structural view of
and argue that the actual organizational learning process is a “generative dance” between knowledge and knowing.
“community intelligence”, we can go a further
We agree with Cook and Brown’s ideas with
step to a more dynamic view, to study the
reservations. One of our reservations is that their
evolution
knowledge
concept of “knowing” is somehow confusing. If
network, or in other words to study the cognitive
“knowing” were just a process or “action itself”
development of the community as an intelligent
as they argued, it would be of no necessity to
“living creature”.
conceptually
of
the
community’s
distinguish
“knowing”
from
The development of the knowledge network
knowledge, since it has been a commonsense
is essentially based on the creation, transmission
that knowledge is achieved and used in the
and fusion of knowledge within the community.
process of “knowing”. What makes “knowing”
This subject has been
in the
comparable with “knowledge” is that “knowing”
knowledge management arena for years. For
can also be regarded as some “entity” in its own
example, a well-noted contribution is the
right. From an epistemological view, all an
“knowledge conversion model” or the SECI
individual knows is from his or her experiences
model proposed by Nonaka (1994) in the
or practices; in return, any action of learning and
organizational context. In that model, a learning
doing of an individual is based on what he or she
spiral is presented in which knowledge is
knows. Thus, to an individual, the action of
transformed between the tacit and explicit forms
knowing and the “entity” of what he or she
discussed
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
214
knows cannot be separated. Furthermore, what
external knowledge sources into an individual’s
an individual knows is intrinsically dynamic
personal “knowing”. Following the pattern of
because of the continuing action of knowing of
the words “hardware” and “software” being
that individual. In this sense we can term what
created, we use a term “knoware” to refer to this
an individual knows as “knowing”, which can
external form of knowledge, in hope to stress
not be apart from the individual’s action of
that
is
some
ontological
“substance”
learning and doing. Another characteristic of this
independent
from
any
individual’s
“knowing” is that it cannot be apart from the
consciousness.
individual who holds it; in another word, here “knowing”
refers
to
personal
knowledge
it
single
This dyad of “knowing” and “knoware” of knowledge implies
a
generic
process
of
endogenously embedded in one’s consciousness.
inter-personal knowledge transfer, as shown in
Thus the development of personal cognition
Figure 4.
essentially means the development of personal “knowing” in the process of learning and doing. The previous understanding of “knowing” implies
that
“knowing” is one
form of
knowledge. Our second contention is that there is another form of knowledge, which is
Personal knowing
Knoware
Externalize Knowledge sender
something “out there” or something existing independent of any individual’s mind. For
Personal knowing
Internalize Knowledge Media receiver space
Figure 4 Generic knowledge transfer process
example, we usually acknowledge that books this
The argument behind the knowledge transfer
“knowledge” contained in a book is very much
process depicted in Figure 4 is that there is no
different from the above-mentioned “knowing”
direct channel to transfer one individual’s
residing in one’s mind. If one individual dies,
personal knowledge (i.e. the personal “knowing”)
the “knowing” of his or her disappears
to the other, since the “personal knowing”
accordingly. Nevertheless, if he or she writes a
adheres to its owner. Instead, the knowledge
book, the “knowledge” contained in the book
sender has to externalize a part of his or her
would remain even after the author’s death. Thus
“knowing” toward some media space (e.g. audio,
the “knowledge” contained in the book can be
video, or texts) to generate some sort of
viewed as some “being” out there. In addition,
“knoware”. Such “knoware” can then be
there is no direct mapping from the knowledge
acquired by another individual; and through a
in the book to an individual’s mind. A book
series of internal cognitive activities, the second
about the superstring theory would mean
individual
nothing to an individual without the knowledge
“knoware” to his or her internal “knowing”; and
background of modern physics. That means,
such assimilation process from knoware to
cognitive endeavors are required to absorb
knowing can also be termed as “internalization”.
contain
“knowledge”.
However,
may
assimilate
the
external
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
215
With the mediation of “knoware”, in this process
network, or personal “knowing” through the
it would look like that the sender’s “knowing” is
personal cognitive processes. What’s more, the
transferred to the receiver. To conclude, every
collective “cognitive” processes may emerge
interpersonal knowledge transfer process is
from
inherently
interactions amongst the community members.
a
knowing-knoware-knowing
all
sorts
of
communications
and
In the collective cognitive processes, knowledge
transformation process. transfer
is transmitted between the community members
process is then the elementary micro-process
through different types of “knoware”, which is
that eventually leads to the formation and
conveyed by the “media network”. Figure 3
evolution of the collective intelligence at the
illustrates the general picture how the parallel
community
and
This
interpersonal
level.
aforementioned
knowledge
In
other
supernetwork
words,
the
structure
of
community intelligence generates and evolves
repetitive
“knowing-knoware-knowing”
transformation processes lead to the formation and evolution of community intelligence.
through the parallel and repetitive actions of
In Figure 5, the human network and the
such micro-level “knowing-knoware-knowing”
media network is explicit; but the knowledge
transformation processes.
network is virtual. This virtual knowledge network manifests the memory system of the
Virtual knowledge network of knowing and “knoware” Personal knowing
Human Network
Generates Personal knowing
Communications and interactions Virtual space of “knoware”
“intelligent” community, which is comprised of the personal knowing of the community members and the external “knoware” conveyed by the media network. In the case of a Web-mediated online community, much of such “knoware” resides in the Web-space. The evolution of the knowledge network or the “memory” of the community is the result of the
Internalization
Externalization
“knowing-knoware-knowing”
transformation
and the consequent knowledge transmission, Supports
Media network as human-network-supporter and “knoware”-conveyor
Figure 5 The overall process of community intelligence development
integration (fusion) and creation. Following the dynamic processes of the knowledge network evolution, the community’s collective “memory” is more than the summation of the personal memories of the community members by themselves. Analogous to personal cognition which is built on schemas (Marshall 1995) and mental models (Johnson-Laird 1981), it would
At the individual level, each community
be natural to expect that the formation of the
member develops his or her personal knowledge
collective “schemas” and “mental models”
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
216
the
inevitably lead to the study of the dynamic
community intelligent system. The emergent
processes of a supernetwork. Such study can be
collective “mental models” may reside in the
respectively carried out at the micro- and
member’s personal minds, as being discussed by
macro-levels following a generative research
various authors in the context of teamwork
fashion (Epstein 1999). At the micro-level, one
(Klimoski & Mohammed 1994). This idea that
key research issue is the deeper analyses of the
community intelligence is generated upon the
knowledge transfer model between human
shared mental models reflects the conceptual
beings,
connection between community intelligence and
“knowing-knoware-knowing”
team intelligence, as we have discussed earlier.
Further
The community “mental models” may embody
analyses, which may be rooted on social
as the shared understandings which basically
psychological and communicational studies, are
reside in the community members’ minds and
necessary to pursue better understandings of
which
actual knowledge-transfer mechanics through
would
be
an
often
emergent
attribute
property
some
of
degree
of
intangibility or tacitness; but they can also be exhibited in more tangible forms such as the
based
on the
conceptual,
currently proposed
logical
framework. and
empirical
human communications and collaborations. Based on the exploration of the microprocess of interpersonal knowledge transfer, the
written norms and regulations.
macro-level community intelligence may be
4. Some Relevant Research Issues
investigated. We suspect that the Swarm-
In the preceding section we try to develop a
Intelligence-like dynamics may take effect so
conceptual model of “community intelligence”.
that the community would evolve to be an
However, this model is still primitive at the
“intelligent being” from local interpersonal
current stage. Especially, the elaborate dynamics
interactions. However, comparing with any
of how the micro-level knowledge activities lead
process in which Swarm Intelligence emerges,
to the macro-level evolution of community
additional complexity has to be handled to
intelligence has not been well-addressed in the
model the formation and evolution of the
proposed
collective intelligence in human communities.
model
and
extensive
theoretical
explorations are deserved.
Two reasons can instantly be identified. One
Essentially, the generation and evolution of
reason is the inherent structural complexity of
community intelligence rely heavily on the
the involved system. We have to handle the
development
“knowledge
complexity of the structural development of a
network” of the community, as being discussed
supernetwork which contains heterogeneous
in the structural modeling part in the preceding
relationships, as at the center of community
section. This knowledge network is furthermore
intelligence is the idea of a knowledge
dependent on the associating human network
supernetwork of triple interwoven networks. The
and media network; therefore, the study of the
second reason is the complexity of knowledge
knowledge
itself. To explore the dynamics of the formation
of
the
network
virtual
development
would
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
217
and evolution of community intelligence, we
alleviate the difficulties of “knowing-knoware”
must set the focus on the knowledge processes
transformation. The “knowledge organization”
such as knowledge transmission, integration and
(Hjørland 2003) and “knowledge visualization”
creation,
intangible.
(Eppler & Burkard 2004) technologies may be
Modeling of such knowledge movements in the
of great help in this respect. More computing
social and technological environments would be
tools
a big challenge. Despite the difficulties, we are
comprehensions of group and social cognitions
not pessimistic in conducting researches in this
would also be worthy of pursuing. At the macro-
direction. Various researches on knowledge
level, the problem is how the Internet/Web
diffusion, especially those on agent-based
technologies to support the knowledge activities
modeling of knowledge diffusion patterns, for
in a community, for example in a community of
example what has been done by Cowan and his
e-learning groups. Due to the evolutionary
collaborators (Cowan & Jonard 2004), may give
nature of community intelligence, the emphasis
implications to the study of the knowledge
should be paid on using computing tools to
network evolution in our settings.
promote participation and to facilitate more
which
are
essentially
The second category of research efforts that
that
dynamic
are
and
developed
under
“democratic”
deeper
knowledge
catches our attention is stemmed from a
dissemination, integration, and creation. From
practical and technological point of view. As
this aspect, various technologies under the
discussed
and
umbrella of “Web 2.0” (O’Reilly 2005) and
communication technologies are critical for the
social computing (Wang et al. 2007) may set the
formation
technological cornerstones for further research
earlier, of
the
information
community
intelligence.
A
subsequent question would then be how the
and development.
community and community intelligence can be supported
by
the
information
and
5. Conclusion
communication technologies. More specifically,
In literature, “intelligence” has become a
in our Internet age, how can the Internet and
widely-discussed metaphor to describe various
Web technologies be adopted to facilitate
social entities. The discussions are generally
smoothening
focused on the team and organization levels,
knowledge
transmission,
integration and creation in an online community?
aiming at developing models of the “collective
Again this question may be respectively
mind” (Weick & Roberts 1993), the “knowledge
answered at the micro- and macro-levels. At the
creating
micro-level, the problem is to exploit IT
“organizational learning” (Argyris & Schon
technologies to bridge the knowledge gaps
1978), etc. Another stream of inquiry is on the
between individuals. Conceptually, knowledge is
vision of the “global brain” (Russell 1983). The
transferred from one person to another through a
attention of the present work, nevertheless, is to
“knowing-knoware-knowing” procedure, thus
some extent between the prior two research
the key to smoothen knowledge transfer is to
streams. In comparison with the organizational
company”
(Nonaka
1994),
and
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
218
level researches, the research on “community
Reading MA
intelligence” would basically focus on larger-
[2] Akgüna, A.E., Dayanb, M. & Benedettoc,
scale social communities with more openness
A.D. (2008). New product development team
and flexibility; on the other hand, the research
intelligence: antecedents and consequences.
on community intelligence would technically be
Information & Management, 45 (4): 221-226
more traceable than the “global brain” research.
[3] Balis, M.E., Krakoff, I.H., Berman, P.H. &
With
the
observation
that
various
social
Dancis, J. (1967). Urinary metabolites in
communities have become the vehicle of
congenital hyperuricosuria. Science, 156
knowledge transmission and innovation, we
(3778): 1123
contend that there is a need to study the phenomena of “community intelligence”.
(1999). Swarm Intelligence: from Natural to
From this aspect, in this paper we call for attention to the possible academic field of community intelligence, by examining it from the
backgrounds
of
[4] Bonabeau, E., Dorigo, M. & Theraulaz, G.
“community”
and
Artificial Systems. Oxford University Press, OUP USA Inc., New York [5] Brown, P. & Lauder, H. (2000). Collective Intelligence. In: Baron, S., Field, J., Schuller,
“intelligence” researches A rough conceptual
T.
model of community intelligence is
Perspectives, pp. 230. Oxford University
then
presented, based on the idea of knowledge
(eds.),
Social
Capital:
Critical
Press, New York
supernetworks and the view of knowledge
[6] Cannon-Bowers, J.A., Salas, E., & Converse,
movements upon the distinction of “knowing”
S. (1993). Shared mental models in expert
and “knoware”. A few related research issues are
team decision making. In: Castellan, N.J. Jr.
briefly discussed as a starting point for the
(eds.),
possible future research. However, we have to
Making:
admit that the ideas presented here is fairly
Hillsdale, NJ: LEA
primitive, and we hope that this immature work would foster more thoughtful endeavors.
Individual
between
the
Group
decision
pp.
221-246.
generative
organizational
organizational
The authors are grateful for the suggestions
issues,
[7] Cook, S.D. & Brown, J.S. (1999). Bridging epistemologies:
Acknowledgment
and
Current
knowing.
knowledge
dance and
Organization
Science, 10 (4): 381-400
from the anonymous referees, which are of great
[8] Cowan, R. & Jonard, N. (2004). Network
help to improve the quality of this paper. The
structure and the difusion of knowledge.
hard work of the Journal editor, Prof. Jian Chen,
Journal of Economic Dynamics and Control,
is cordially appreciated.
28 (8): 1557- 1575 [9] Epstein,
References [1] Argyris,
J.
(1999).
Agent-based
computational models and generative social C.
&
Schon,
D.A.
(1978).
Organizational Learning. Addison-Wesley,
science. Complexity, 4 (5): 41-60 [10] Eppler, M.J. & Burkard, R.A. (2004).
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
219
Knowledge visualization: towards a new
[17] Johnson-Laird, P.N. (1981). Mental models
discipline and its fields of application. ICA
in cognitive science. In: Norman, D.A. (eds.),
Working Paper #2/2004, University of
Perspectives on Cognitive Science, pp.
Lugano, Lugano
147-191. Norwood, NJ: Ablex; and Hillsdale,
[11] Fischer,
G.
(2001).
Communities of
Interest: learning through the interaction of multiple knowledge systems. In: Bjornestad, S. et al. (eds.), Proceedings of the 24th
NJ: Erlbaum [18] Kennedy, J. & Eberhart, R. (2001). Swarm Intelligence.
Morgan
Kaufmann,
San
Francisco
Information Systems Research Seminar in
[19] Klimoski, R. & Mohammed, S. (1994).
Scandinavia, 1-14, Ulvik, Norway, August
Team mental model: construct or metaphor?
11-14,
Journal of Management, 20 (2): 403-437
2001,
Bergen:
Department
of
Information Science, University of Bergen
[20] Lave, J. & Wenger, E. (1991). Situated
[12] Gabelnick, F., MacGregor, J., Matthews,
Learning: Legitimate Peripheral Participation.
R.S. & Smith, B.L. (1990). Learning
University of Cambridge Press, Cambridge,
Communities: Creating Connections among
UK
Students, Faculty, and Disciplines (New
[21] Levy, P. (1994). Collective Intelligence:
Directions for Teaching and Learning #41).
Mankind’s Emerging World in Cyberspace.
Jossey-Bass, John Wiley & Sons, Inc.
Basic Books
[13] Girvan, M. & Newman, M.E.J. (2002). Community structure in social and biological networks. Proceedings of National Academy
[22] Luhn, H.P. (1958). A business intelligence system. IBM Journal, October, 314-319 [23] Marshall, S.P. (1995). Schemas in Problem Solving. University of Cambridge Press,
of Sciences, 99 (12): 7821-7826 [14] Heylighen, F. (2005). Conceptions of a global brain: an historical review. Available
Cambridge, UK [24] Mayer-Kress, G. & Barczys, C. (1995).
DIALOG.
The global brain as an emergent structure
http://pespmc1.vub.ac.be/Papers/GBconcepti
from the Worldwide computing network, and
ons.pdf. Cited 2005
its
via
[15] Hjørland, B. (2003). Fundamentals of knowledge
organization.
Knowledge
implications
[16] Johnson, N., Rasmussen, S., Joslyn, C.,
modeling.
Information Society, 11 (1): 1-27 In:
intelligence.
Handbook Available
of
collective
via
DIALOG.
Rocha, L., Smith, S. & Kantor, M. (1998).
http://scripts.mit.edu/~cci/HCI/index.php
Symbiotic
Cited Nov 18th, 2008
intelligence:
self-organizing
The
[25] MIT Center for Collective Intelligence (2008).
Organization, 30 (2): 87-111
for
knowledge on distributed networks driven by
[26] Nagurney, A. & Wakolbinger, T. (2005).
human interactions. In: Adami, C. et al.
Supernetworks: An introduction to the
(eds.), 6th Artificial Life Conference '98,
concept and its applications with a specific
MIT Press
focus
on
knowledge
supernetworks.
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
220
International Journal of Knowledge, Culture
[36] Wang, F., Carley, K., Zeng, D. & Mao, W.
and Change Management, 4 [27] Nonaka, I. (1994). A dynamic theory of organizational
2002
knowledge
creation.
(2007).
Social computing:
from social
informatics to social intelligence. IEEE Intelligent Systems, 22 (2): 79-83
Organization Science, 5 (1): 14-37 [28] O’Reilly, T. (2005). What is Web 2.0:
[37] Weick, K.E. & Roberts, K.H. (1993).
design patterns and business models for the
Collective mind in organizations: heedful
next generation of software. Available via
interrelating on flight decks. Administrative
DIALOG.
Science Quarterly, 38: 357-381
http://www.oreillynet.com/pub/a/oreilly/tim/
[38] Wellman, B. & Wortley, S. (1990).
news/2005/09/30/what-is-web-20.html. Cited
Different
2005
community ties and social support. American
[29] Porter, C.E. (2004). A typology of virtual communities: a multi-disciplinary foundation for
future
research.
Journal
of
strokes
from
different
folks:
Journal of Sociology, 96 (3): 558-588 [39] Wierzbicki, A.P. & Nakamori, Y. (2005). Creative
Space:
Models
of
Creative
Computer-mediated Communication, 10 (1):
Processes for the Knowledge Civilization
Article 3
Age. Springer, Berlin
[30] Rheingold,
H.
(2000).
The
Virtual
Community: Homesteading on the Electronic Frontier. MIT Press, London [31] Russell, P. (1983). The Global Brain: Speculations on the Evolutionary Leap to Planetary Consciousness. Houghton Mifflin, Boston, MA [32] Segaran,
T.
(2007).
Programming
Collective Intelligence: Building Smart Web 2.0 Applications. O’Reilly Press
[40] Zhong, N., Liu, J.M. & Yao, Y.Y. (2003). Web Intelligence. Springer, Berlin Shuangling Luo Ph.D. Candidate at Japan Advanced Institute of Science and Technology, Nomi, Ishikawa, Japan. She received her M.Sc. from Dalian University of Technology in 2006. Her
research
interest
is
in
knowledge
communication and visualization. Haoxiang Xia Associate professor at Dalian
[33] Scardamalia, M. & Bereiter, C. (1994).
University of Technology, Dalian, China. He
Computer support for knowledge-building
obtained his Ph.D. degree from Institute of
communities.
Systems Engineering, Dalian University of
Journal
of
the
Learning
Sciences, 3 (3): 265-283 [34] Szuba,
T.
(2001).
Technology (DUT) in 1998. Before working at Computational
DUT since 2000, he was a postdoctoral fellow at
Collective Intelligence. Wiley, New York
Institute of Systems Science, the Chinese
[35] Tönnies, F. (1887). Gemeinschaft und
Academy of Sciences from 1998 to 2000. His
Gesellschaft, Leipzig: Fues's Verlag. English
main research interests include Web-based
translation by Loomis, P.C. as Community
information systems, swarm intelligence and
and Society. Courier Dover Publications,
complex networks.
Luo et al.: Toward Collective Intelligence of Online Communities: a Primitive Conceptual Model J Syst Sci Syst Eng
221
Taketoshi Yoshida Professor at Japan Advanced
Zhongtuo Wang Professor at Dalian University
Institute of Science and Technology, Nomi,
of Technology, Dalian, China. He obtained B.S.
Ishikawa, Japan. He received his Ph.D. degree
from the department of Electrical Engineering,
from the department of Systems Engineering,
Tsinghua
Case Western Reserve University in 1984. He
Academician of the Chinese Academy of
worked for IBM Japan from 1985 to 1997. His
Engineering. His current research interests are in
research interests are in systems science and
knowledge management and knowledge systems
knowledge-handling information systems.
engineering.
University in 1951. He
is
an