The University of Sussex

On simulating the evolution of communication

Jason Noble, Dave Cliff

The prospects for modelling the evolution of communication are considered, including the problem of intentional explanation, and the possibility of grounding simulation work in theoretical biology. The seminal work of MacLennan and Burghardt (1994) on the evolution of cooperative communication is described, and their experiment replicated. Our results were broadly similar, in that evolved communication was observed, but specific differences are discussed. MacLennan and Burghardt's work is extended and their methodology critiqued in detail. Their experiment remains a useful demonstration, but artefactual features make their results difficult to interpret. Furthermore, we argue that too many factors are simultaneously investigated for any general principles to be extracted, and suggest an alternative program of narrowly-focused simulations.

Download compressed postscript file