2009 International Conference on Intelligent Networking and Collaborative Systems
Group Intelligence: a distributed cognition perspective
Osama Mansour
School of Mathematics and Systems
Engineering, Växjö University
Växjö, Sweden.
osama.mansour@vxu.se
Abstract—The question of whether intelligence can be
attributed to groups or not has been raised in many scientific
disciplines. In the field of computer-supported collaborative
learning, this question has been examined to understand how
computer-mediated environments can augment human
cognition and learning on a group level. The era of social
computing which represents the emergence of Web 2.0
collaborative technologies and social media has stimulated a
wide discussion about collective intelligence and the global
brain. This paper reviews the theory of distributed cognition
in the light of these concepts in an attempt to analyze and
understand the emergence process of intelligence that takes
place in the context of computer-mediated collaborative and
social media environments. It concludes by showing that the
cognitive organization, which occurs within social interactions
serves as a catalyst for intelligence to emerge on a group level.
Also a process model has been developed to show the process
of collaborative knowledge construction in Wikipedia that
characterizes such cognitive organization.
fields of semantic web, distributed artificial intelligence, and
ambient intelligence [7]. Distributed cognition means cognitive
processes that are distributed across the members of a social group
[9]. These cognitive processes are involved in memory, decisionmaking, inference, reasoning, learning, etc which sum up human
intellectuality [10]. A collaborative system can be viewed as - a
large cognitive system involving multiple people interacting with
each other and a wide range of artifacts to perform an activity
[21]. Hence with the increasing developments in collaborative and
knowledge sharing technologies [25], the focus on group
intelligence is a compelling issue to explore how state-of-the-art
technologies may influence human cognition. Therefore the paper
tries to examine how social media and collaborative technologies
mediate cognitive processes of a group of people and the
influence on the emergent intelligence. A thorough review of the
theory of distributed cognition has been conducted and different
aspects of the theory have been examined to show how cognitive
processes take place within a group activity supported by social
media technologies. This will contribute to understanding how
cognition is distributed and shared among groups and the
influence of the flow of cognitive processes on intelligence.
Keywords: Group Intelligence, Theory of Distributed
Cognition, Social Media, Web 2.0, Collaborative.
I.
II. THEORY OF DISTRIBUTED
COGNITION
INTRODUCTION
Intelligence has always been a controversial term when
discussed in the context of collaborative environments in an
attempt to attribute it to groups. This is because it is habitual to
attribute thoughts and intentions to individuals and to reduce
group phenomenon into actions by the individual members of the
group ([7], [9], [23]). In this paper, group intelligence refers to
the aggregation of individuals’ distributed cognitive processes
over a common platform in the form of socially shared knowledge
and meaning. It is opted that intelligence is a property that
emerges during a group activity and thus it is an attribute on a
group level. The paper uses the term group intelligence rather than
collective, collected intelligence or any other terms based on the
idea that all these terminologies refer to intelligence that takes
place during a social group activity.
The current era of social computing which represents the
emergence of an assortment of social media applications that
support collaboration, knowledge sharing, social networking, and
self-organization ([5], [17], [20], [25], [30]) has stimulated a wide
discussion about concepts like collective intelligence,
crowdsourcing, global brain, etc ([15], [24], [25]). Consequently,
there is a need for a profound understanding and examination of
how intelligence can be achieved by groups supported by social
media technologies. At this respect, distributed cognition (c.f.
section II) has been a focus by researchers where they discussed
different aspects of intelligence that emerge within the social
interactions of group members (e.g. [9], [23], [26]). In this sense,
the use of the distributed cognition approach can be found in the
978-0-7695-3858-7/09 $26.00 © 2009 IEEE
DOI 10.1109/INCOS.2009.59
Edwin Hutchins has developed the theory of distributed
cognition in a simulation experiment of an airline cockpit (c.f.
[10]). In this experiment, Hutchins views the cockpit as a large
cognitive system with emergent cognitive properties. These
properties emerge as a result of interactions of the crewmembers,
interactions between the crew and the artifacts, coordination
mechanisms, and the propagation of information in the form of
representational states (i.e. mental state, state of knowledge)
across representational media that is the main concern of the
theory ([8], [10], [21]). The propagation of information in the
form of representational states refers to the transformation of
information during the conduct of a group activity [21].
Representational media is located within individuals (e.g.
memory, knowledge, skills,), within group members (e.g. shared
meanings), and in the physical structure (e.g. tools such as social
media) ([8], [9]). These representational media have different
properties that constraint the required cognitive processes to
communicate representational states [10]. Therefore, the emphasis
is on the propagation of information in the distributed cognition
theory as this process entails group meanings, which form the
emergent properties of the system ([8], [10], [23]). In this sense,
Rogers [21] explained that the distributed cognition approach
seeks to explicate the complex interdependencies between people
and artifacts in their work activities, of which an important part is
identifying the problems, breakdowns and the distributed problem
solving processes that emerge to deal with them.
247
Hutchins and Klausen [10] emphasized the importance of the
distributional characteristics of information, which are essential to
maintain intersubjective understandings and shared meanings.
Intersubjective understanding emerges when individuals’
knowledge gets shared during a group activity to form a shared
resource for their distributed negotiations, interactions, and the
development of shared meanings ([9], [23]). At this respect, a
major aspect of this theory is that it moves beyond understanding
individual cognitive processes to reach an understanding of
cognition that occurs on a system or a group level ([8], [9], [19],
[21], [22]). Therefore, the distributed cognition approach takes the
whole system as a unit of analysis rather than studying the
components of the system ([9], [19], [21]). Stahl [23] reflected on
the issue of the unit of analysis as methodological rather than
ontological where the focus should be the group rather than the
individual member of the group. Three kinds of distribution of
cognitive processes are involved with the theory of distributed
cognition: cognitive processes distributed across the members of a
social group, cognitive processes may involve coordination
between external and internal (material and environmental)
structure, and processes maybe distributed through time ([8], [9]).
To sum up, distributed cognition approach seeks to understand the
organization of large cognitive systems, which could be seen as
sort of cognitive architecture on a group level. To understand
intelligence, we need to consider such larger systems where the
parts get assembled together to achieve tasks and develop shared
meanings [9]. These tasks and shared meanings could not be
achieved by a sole individual and may influence human cognition
on a group or community level ([9], [26]), which therefore impact
their ability to collectively solve problems and make decisions.
this respect, Argyris and Schön [1] mentioned “Learning is a term
applicable to individuals within the context of a group, but when
individuals learn to interact with one another so as to carry out
shared tasks, one can speak of the group itself as learning” (p.
322). In this view of interactional learning, the source of
individual knowledge is the group. Stahl [23] provided a number
of views on learning from the field of computer-supported
collaborative learning (CSCL) that represent diverse sources of
knowledge and meanings both from individuals and groups. These
different views of learning may provide some perspectives on the
propagation of information both internally inside an individual
mind or externally through the interaction among individuals and
with physical artifacts, which is central to the theory of distributed
cognition. In this sense, Smith [22] argued that groups have more
potential to remember information rather than individuals due to
transactive memory creating a larger distributed memory capacity
than an individual can possess. Also, he reflected that the
communication and distribution of information among group
members helps to effectively search for diverse information which
can potentially improve the overall group performance. As a
result, cognition emerges on a group level due to the distribution
and communication of knowledge between group members, which
may result an amplified cognition that is not reducible to the
individual mind. Further, the emergent group cognition consists of
group memory, group problem solving, and group decisionmaking abilities, which form the basis for non-reducible group
intelligence to individuals. However, in order to achieve and
maintain group intelligence, there is a need for an effective
communication and collaboration media so that cognition is
distributed and shared by groups. The next section discusses
different aspects of Web 2.0 and social media technologies, which
may provide platforms for the propagation and communication of
information.
III. GROUP INTELLIGENCE
The question of how cognition could be examined on a group
level has been raised in many scientific disciplines such as social
psychology and cognitive science ([7], [10], [23]). The traditional
conceptions that view cognition as limited to the individual mind
and that learning and thinking only occurs on an individual level,
have led to difficulties in understanding intelligence on a group
level ([1], [7], [9], [23]). However, there is an increasing number
of approaches proposing that human cognition is not solely found
inside an individual brain but also involves interactions with other
minds ([7], [8], [21], [22]). These approaches are focused on the
cognition of larger systems that emerge on a group level. Theories
of distributed cognition, extended mind, and situatedness are
focusing on the extension of individual cognition to involve
interconnections with other minds and the ability to form patterncompleting and pattern-creating actions [6]. In this sense,
distributed cognition is a confluence with collective or group
intelligence and cognition could be understood as collective
intelligence that is cognitive processes and structures that emerge
at the social level [7].
In contrast to traditional conceptions of learning and thinking,
many researchers viewed knowledge as a result from social
construction that is achieved by individuals rather merely
individual observation ([1], [7], [22], [23]). This has led to the
idea that individuals learn due to social interactions with other
individuals in the course of a group activity [23]. A simple
example to clarify this is about a newborn learning child. In the
early childhood, a child starts interacting with adults and artifacts
around him. Over time, the child starts to learn and develop new
functional skills that would allow him to function in the absence
of adults. Therefore, learning could be seen as interactional. At
IV. WEB 2.0 AND SOCIAL MEDIA
It might be said that the vision of the global brain has been
reached in this current era of social computing. The global brain is
an intelligent network formed by the people of this planet together
with the knowledge and communication technologies that connect
them together [7]. This vision is manifested in the application of a
variety of Web 2.0 and social media technologies. Web 2.0 is a
generic term that refers to a variety of social computing
applications such as Wikis, blogs, peer-to-peer downloading,
social networking, etc ([17], [25]). The underlying concepts of
Web 2.0 technologies involve the architecture of participation,
collaborative knowledge construction, network as a platform,
consuming and remixing data from multiple sources, etc ([2], [6],
[11], [14], [17], [20], [25]). The significant impact of Web 2.0
technologies lies in connecting large numbers of people and in
their flexibility that allows people to appropriate their features in a
way that enable them to link and share their individual tacit
knowledge, which represent individual cognitive processes, and
align this knowledge with the social context of producing group
knowledge [13]. This knowledge is not reducible to individuals as
it contains multiple ideas and interpretations that have been
processed and manipulated by other people. Many researchers
have reflected on the capabilities of Web 2.0 or social web and the
potential for augmented social cognition, collective intelligence,
and wisdom of crowds ([2], [6], [20], [24], [25], [27]). Web 2.0
environments reflect the concept of networks of practice.
Electronic and networked communities of practice represent large
applications of Web 2.0 and social media technologies with their
rich knowledge exchange and social interactions. Wasko and
248
Faraj [28] defined electronic networks of practice as “a selforganizing, open activity system focused on a shared practice that
exists primarily through computer-mediated communication.” (p.
3). In this sense, Web 2.0 environments support large number of
individuals who voluntarily choose to be part of online social
communities and engage in open activities where they share their
knowledge to help solve problems of common interests. The
properties of these technologies reflect a major notion, which is
the propagation and distribution of shared knowledge within
social interactions. This notion is implied in the theory of
distributed cognition where group cognition takes the form of
shared knowledge. In this sense, Touzet [26] explained, “…we
can just about to conceive distributed cognition as sum of
contextual intelligences operating as a whole structure.” Based on
the conceptions of Web 2.0, a Web 2.0 collaborative system could
be seen as a single and large structure integrating multiple groups
and communities with common interests where they communicate
and share their knowledge. This results cognitive processes of
these groups to be distributed and are manifested in collaborative
knowledge. Collaborative knowledge is rich of multiple ideas,
meanings, and interpretations, which externalize the contents of
thought, and is embodied in the coordination of our identities in a
social and collaborative context [22]. Consequently, Web 2.0
environments mediate the distribution of our cognitive processes
and support our social interactions in a sense that the knowledge
transformed across members of the groups, from a distributed
cognition perspective, can be viewed as a form of group
intelligence. Such group or social Intelligence can result from
social learning that is interactional, which in turn illuminate
augmented social cognitive processes.
The next section discusses Wikipedia, which is based on the
Wiki technology and has been rising as the largest online
encyclopedia ([4], [12], [15]) to explain the processes of largescale collaborative knowledge construction and decision-making.
1) and every aspect is cognitive-ingrained. Figure 1 above
visualizes theoretical aspects of the theory of distributed cognition
that represent sequential and iterative processes of collaborative
knowledge construction in Wikipedia. The process model
describes processes that take place during a cognitive activity,
which is continuously executed by a group of people for
collaborative knowledge construction. These processes entail
propagation of multiple states of people. People are performing
these processes iteratively. Each process is associated and altered
by cognitive abilities of everyone resulting cognitive processes to
be distributed across the group. Such distribution characterizes a
kind of cognitive organization, which is created and maintained
through social interactions and connectivity that are represented
by arrows in the process model. The iteration and recurring efforts
in the process of collaborative knowledge construction in
Wikipedia create a cognitive feedback loop leading to some sort
of organization. According to Robert (1964 cited in [9], p. 2) that
“such social organization could be seen as a sort of architecture of
cognition at the community level”. This iterative process and
cognitive organization in Wikipedia implies a kind of - social
capital, which is embedded within networks of social structure
([18], [28]). Social capital is a term used in community studies,
which highlights the central importance of networks of
relationships that provide a basis for trust, cooperation, and
collective action (ibid). Nahapiet and Ghoshal [18] argued that
social capital facilitates the creation of intellectual capital based
on the idea that networks of relationships constitute a resource for
the social conduct and provide the members of the groups with
collectivity-owned capital (Buordieu, 1986, p. 249 cited in [18]).
In this sense, they referred to intellectual capital as “…the
knowledge and knowing capability of a social collectivity, such as
an organization, intellectual community, or professional practice.”
(ibid, p. 5). The community of Wikipedians is supported by the
Wiki technology that constitutes the structure or the platform for
their social interactions and collective actions. This helps to
maintain the growth of social capital through enabling knowledge
exchange and intellect leverage that is represented and combined
in large numbers of articles. The cognitive characteristics of the
community represent its memory capacity and the ability to
manage and retrieve information, which can be determined by
looking at what information is there, where it is located and how it
moves in the community (ibid). In this sense, figure 1 represents a
sequence of representational states leading to movement of
information in Wikipedia. Hutchins [9] explained that such
movement of information has an impact on the organizing
behavior and coordination of the group and also depends on their
assessment of own states of knowledge and others. Also, repeated
patterns of activity may lead to the consolidation of functional
assemblages and the result can be individual learning or
organizational learning [9]. He further added that an important
property of aggregate systems is that they may give rise to forms
of organization that cannot develop in the component parts. Hence
one may argue that the cognitive organization is by itself an
intelligent effort enabled by Web 2.0 environments and also
serves as a catalyst for intelligence to emerge on a group level.
A. Wikipedia: A Process Model
The Wiki technology refers to a simple website that enables
everyone to collaboratively create, edit, tag, and link content
([11], [12], [16], [25]).
Figure 1. Process model describing the processes of collaborative
knowledge construction in Wikipedia.
In this sense, articles in Wikipedia are collaboratively created and
content is refined and filtered in an iterative manner. Everyone
can contribute knowledge to a particular subject ([12], [16]).
Other people may change or edit the content according to their
experiences in each subject. Continuously, this iterative process of
refinement embodies social knowledge that is communicated and
shared with self-organized communities of Wikipedians involving
decision-making processes especially when topics are
controversial. Topics that are controversial require collaborative
assessment and consensus, which may involve argumentations
and discussions over the Wiki until a peer-reviewed article is
reached. Any changes to the content can be seen by group
members and are aware of them. Such collaborative process of
creating and communicating content in Wikipedia embodies
different aspects of the theory of distributed cognition (see Figure
VI. CONCLUSIONS
In this paper, the theory of distributed cognition has been
examined in the context of computer-mediated environments in an
attempt to address the question of how intelligence can emerge on
a group level. The theory provides some perspectives on the
emergence process of intelligence through examining the
249
distribution of cognitive processes across the members of social
groups. Particularly, the theory has been examined in Web 2.0
environments, which entail rich social interactions and cognitiveingrained collaborative processes. The cognitive organization that
is achieved through the propagation of knowledge, division of
labor, and ongoing social interactions across the group results
augmented social cognition and therefore serves as a catalyst for
intelligence to emerge on a group level. This cognitive
organization results due to the distribution of cognitive processes
and is manifested in collaborative knowledge. The resulted
collaborative knowledge is rich of multiple ideas, meanings, and
interpretations of everyone within the group, which helps to
enhance the ability of making decisions and solve problems and is
not reducible to any of the members of the group. Furthermore, in
order to provide an example of cognitive organization, a process
model has been developed that describes the different processes
involved in collaborative knowledge construction in Wikipedia.
The model visualizes different theoretical aspects of the theory of
distributed cognition. It also shows how different cognitiveingrained states propagate throughout the whole process creating a
cognitive feedback loop leading to architecture of cognition at the
community level.
[9] Hutchins, E. (2000) Distributed Cognition. IESBS.
[10] Hutchins, E. & Klausen, T. (1996). ‘Distributed Cognition in
an Airline Cockpit’. In Middleton, D, Engeström, Y.
Communication and Cognition at Work. Cambridge: Cambridge
University Press, 15-34.
[11] Kane, G., Fichman, R. (2009). The Shoemaker's Children:
Using Wikis for Information Systems Teaching, Research, and
Publication. MIS Quarterly, Vol. 33, pp. 1-22.
[12] Lee, M., Lan Y. (2007). From Web 2.0 to conversational
Knowledge Management: Towards Collaborative Intelligence.
Journal of Entrepreneurship Research. Vol. 2, No. 2, pp. 47-62.
[13] Leonardi, P. (2007). Activating the informational capabilities
of information technology in organizations. Organization Science,
Vol. 18, pp. 813-831.
[14] Li, Q., C-Y Lu, S. (2008). Collaborative tagging applications
and approaches. IEEE Multimedia, Vol. 15, pp. 14-21.
[15] Libert, B., Spector, J. (2007). We Are Smarter Than Me.
How to Unleash the Power of Crowds in Your Business. 1st ed.,
Wharton School publishing.
[16] Mader, S. (2007). Wikipatterns. Weily Publishing, Inc.,
Indianapolis, Indiana.
[17] McAfee, A. (2006). Enterprise 2.0: The Dawn of Emergent
Collaboration. MITSloan Management review. Vol. 47, pp. 21-28.
[18] Nahapiet, J., Ghoshal, S. (1998). Social Capital, Intellectual
Capital, and the Organizational Advantage. The Academy of
Management Review, Vol. 23, No. 2 (Apr., 1998), pp. 242-266.
[19] Nardi, B. (1992). Studying context: A comparison of activity
theory, situated action models and distributed cognition.
Proceedings East-West Conference on Human-Computer
Interaction. August 4–8, St. Petersburg, Russia. pp. 352–359.
[20] O’Reilly, T. (2005). Web 2.0: Compact definition – O’Reilly
Radar.
[Online].
Available
at:
http://radar.oreilly.com/archives/2005/10/web-20
compactdefinition.html [Accessed 15 January 2009]
[21] Rogers, Y. (2006) Distributed Cognition and
Communication. In The Encyclopedia of Language and
Linguistics 2nd Edition. Oxford, pp. 181-202.
[22] Smith, E. (2008). Social relationships and groups: New
insights on embodied and distributed cognition. Cognitive
Systems Research, Vol. 9, pp. 24-32.
[23] Stahl, G. (2005). Group cognition in computer-assisted
collaborative learning. Journal of Computer Assisted Learning,
Vol. 21, No. 2. (April 2005), pp. 79-90.
[24] Surowiecki, J. (2005). The Wisdom of Crowds: Why the
Many Are Smarter Than the Few. Abacus, London, UK.
[25] Tapscott, D., Williams, A. (2006). Wikinomics: How Mass
Collaboration Changes Everything. Penguin Group, Inc., New
York, N. Y.
[26] Touzet, J. (n.d.). Distributed cognition or intelligence?
Patterns of group thought within online graduate-level
coursework. Paper submitted for OISE/UT CTL1608F06
[27] Wang2, F., Zeng, D., Carley, K., Mao, W. (2007). Social
computing: from social informatics to social intelligence. IEEE
Computer Society, Vol. 22, pp. 79-83.
[28] Wasko, M., Faraj, S. (2005). Why should I Share?
Examining Social Capital and Knowledge Contribution in
Electronic Networks of Practice. MIS Quarterly, Vol. 29 No. 1,
pp. 35-57/March 2005.
[29] Wood, L. (2005). Blogs & Wikis: Technologies for
enterprise applications?. The Glibane Report. Vol. 12, 2-9.
[30] Yukihiro, K. (2007). In-house use of Web 2.0: Enterprise
2.0. NEC Technical Journal. Vol. 2, p. 46-49.
VII. FURTHER RESEARCH
Theories of social psychology and cognitive science are of
great benefit and importance for the communities of CSCW and
CSCL. Such theories provide diverse and enlightening
perspectives of group dynamics, adaptive behavior, networks of
social relationships, and studies of mind. These perspectives are
essential to enhance our understanding of collaborative and group
processes especially with increasing developments of computing
technologies that facilitate social interactions and group work.
Therefore, further research stemming from these theories is
needed to further increase our understanding of group dynamics.
In particular, these theories are useful to explore more about
group and collective intelligence, which are prevailing concepts in
the era of social computing.
References
[1] Argyris, C., & Schön, D. (1978) Organizational learning: A
theory of action perspective, Reading, Mass: Addison Wesley.
[2] Chi, E. (2008). The social web: research and opportunities.
IEEE Computer. Vol. 41, pp. 88-91.
[3] Flor, N.V., Hutchins, E. (1992). Analyzing Distributed
Cognition in Software Teams: a Case Study of Collaborative
Programming During Adaptive Software Maintenance. In
Empirical Studies of Programmers: Fourth Workshop, eds. J.
Koenemann-Belliveau, T.
[4] Gloor, P., Copper, S. (2007). The new principles of swarm
business. MITSloan Management Review. Vol. 48, pp. 81-84.
[5] Grossman, M., McCarthy, R. (2007). Web 2.0: is the
Enterprise ready for the adventure? Issues in Information Systems.
Vol. VIII. No. 2, pp. 180-185.
[6] Gruber, T. (2008). Collective knowledge systems: where the
social web meets the semantic web. Web Semantics: Science,
Services and Agents on the World Wide Web. Vol. 6, pp. 4–13.
[7] Heylighen, F., Heath, M., Van Overwalle, F. (2004) The
emergence of distributed cognition: a conceptual framework.
Proceeding of Collective Intentionality IV, Seina (Italy).
[8] Hollan, J., Hutchins, E., Kirsch, D. (2000). Distributed
Cognition: Toward a New Foundation for Human-Computer
Interaction Research ACM Transactions on HCI, Vol. 7, pp. 174196.
250