InterJournal Complex Systems, 335
Status: Accepted
Manuscript Number: [335]
Submission Date: 405
The contribution of complexity theory to the study of socio-technical cooperative systems
Author(s): Bernard Pavard ,Julie Dugdale

Subject(s): CX.43, CX.41, CX.17

Category: Article

Abstract:

The aim of this paper is to explore the usefulness of the complexity paradigm in analysing socio-technical cooperative systems. A complex system is intuitively described as a system where it is difficult, if not impossible, to reduce the number of parameters or variables without losing its essential functional properties. We define and analyse four characteristics of complex systems: non-determinism, limited functional decomposability, the distributed nature of information and representation, and emergence and self-organisation. These characteristics are illustrated using examples taken from our work in designing cooperative systems in the domains of air traffic control, emergency control, centres and off-shore deep sea diving. We demonstrate that these four characteristics, which are not treated within the framework of classical analytical approaches, are essential to understand certain functional aspects of cooperative work. For example, from our ergonomic analysis of cooperative work situations we have identified the functional role of the pluri-addressing mechanism. This mechanism is at the heart of the distribution of information between agents in a socio-technical system. By utilising complexity theory we can identify that the mechanism is both non-tracable and non-deterministic. Furthermore, by identifying the distributed nature (in a connectionist sense) of this mechanism we can hypothesise that the robustness of the overall system, i.e. the capacity of the system to handle unforeseen data, is functionally related to the concept of a locally distributed control of information. These mechanisms are principally concerned with local interactions (between the social actors) and are not represented at a central organisational level where certain functional properties (e.g. reliability, robustness of decision making and the occasional abnormal operation of the collective) emerge. This approach and the results would be incomplete if we could not prove them in a productive way, i.e. by simulating the effect of local interactions on the global collective decision during a cooperative scenario. Several simulations are currently being analysed and we briefly report on one simulator that demonstrates the power of this approach. From a general standpoint, we defend the idea of a complementary structural and distributed (also termed 'dynamic') approach both in cognitive science and more generally in social science. These two approaches cover two important dimensions in our understanding of the collective. When used alone the two approaches are necessary, but not sufficient to explain the robustness and dynamic nature of socio-technical systems [Mitchell 99]. The classical analytical reductionist approach is particularly weak in explaining the emergence of functional properties, despite the fact that in socio-technical complex systems, the strength of the collective lies in such properties.

Retrieve Manuscript
Submit referee report/comment


Public Comments: