Happy Pi day + 1 everyone.
I’ve been personally buried in an avalanche of reading, writing, processing, thinking, sensemaking, and job seeking, so there hasn’t been much activity here at toddsuomela.com.
I just finished writing the following introduction to a paper for my complex systems class. It will give you some idea of what I’ve been thinking about over recent days and weeks.
Collective intelligence, group-think, organizational knowledge, distributed cognition, situated action, and the wisdom of crowds are just some of the many different phrases used to describe the similar phenomenon of people collecting, evaluating, and acting on information as a group. There are many situations when the wisdom or intelligence of a group of people may be greater than the wisdom or intelligence of individuals within the group. James Surowecki describes a number of cases in “The Wisdom of Crowds,” including successful examples of guessing the weight of a cow, the location of a submarine, or the outcome of an election.
A student of complex systems should not be surprised by these phenomenon. One of the central ideas in complexity is emergence. From the work of many individuals arises a whole that is different than its parts. On an individual level ants, the cells of a CA algorithm, or even a human being, may be following a set of simple rules that together lead to collective behavior and collective intelligence.
There are some philosophical concerns raised by the idea of collective intelligence. The traditional philosophic view of intelligence is based on the concept of the individual. It is hard enough to determine what makes an individual intelligent without adding in the intelligence of groups. Is the intelligence of a group just the sum of the intelligences of the group members? How can we explain groups that behave irrationally, such as mobs or committees? These are important caveats to the idea of collective intelligence but they will not be addressed in this paper.
Thomas Szuba in “Computing Collective Intelligence” offers the following definition of collective intelligence.
emerges in a social structure of interacting agents or beings, over a certain period, if and only if the weighted sum of problems they can solve together as a social structure is higher during the whole period than the sum of problems weighted in the same way that can be solved by the agents or beings when not interacting.
This operationalizes some of the basic notions of CI: emergence, social structures, increased ability to solve problems. It will be sufficient for the discussion that follows.
In order to narrow the scope of this paper I have chosen to focus on CI at the level of organizations and groups of people. My analysis is based on two different models of groups. The first is proposed by Cohen and Axelrod in “Harnessing Complexity.” The description of groups is borrowed directly from developments in complex adaptive systems theory. The second is taken from “Sensemaking in Organizations” by Karl Weick, which describes the process of sensemaking in organizations. Both theories complement each other very well and offer some interesting suggestions for where a complex systems based approach to CI might go in the future.
Many people have addressed the problem of group information processing by designing software that is frequently called groupware or computer supported cooperative work (CSCW). Grudin identifies eight persistent difficulties encountered by these software projects. The final section of this paper attempts to apply the two models described above, complex adaptive systems (CAS) and sensemaking (SM), to groupware. I hope that this analysis of a concrete problem will improve our understanding of both theories and enable us to act with more wisdom in the future.