The University of Sussex

An investigation into the evolution of communicative behaviors

Ezequiel A. Di Paolo

Communication is a phenomenon with many different aspects, and it has attracted the attention of a variety of scientific disciplines, Biology being perhaps the one with the best chance of providing a good theoretical backbone by addressing the unifying theme that underlies the different views on the subject. However, recent work in the evolution of communication have tended to evade rather than embrace this task. I provide a critical analysis of the reasons for this situation, which are, for the most part, methodological and conceptual, and manifest themselves in the way biologists characterize the phenomenon, as well as in the tools they use to research it. I present an alternative characterization in terms of autopoietic theory, and show that not only is it possible to work with it, but that it also addresses issues of interest to other disciplines. By choosing as my object of study a game of interactions, I intend to provide some continuity with traditional approaches and the view of communication presented here. Traditional tools, such as game theory, are not blindly discarded, but are extended in order to go beyond equilibrium studies into the nature of the evolutionary dynamics. Further extension involves the use of a computational model, so some of the methodological issues that arise by its use are discussed. Within this model, communication evolves in a society of artificial agents even in the presence of costs \emph{against} it, and this is explained in terms of selective mechanisms acting within the constraints provided by other factors such as spatial organization. A complex network of mechanisms is explored by studying the phenomenology of emergent self-regulating unities in the spatial distribution of agents. Dialogic communication also evolves non-trivially in a similar game in which agents share all the relevant environmental information and, by coordinating their actions, they are able to perform tasks beyond their individual cognitive capabilities, showing that the concept of information has to be used with care, and providing a metaphor for the evolution of cognition as rooted in social activity. Conclusions are drawn both on the general subject of explaining complex processes with many interacting causal factors, and on the relation of these results to the evolution of natural communication.

Download compressed postscript file