The Evolution of Communication and Cooperation

Institute For Advanced Study, Princeton NJ

Investigators

Abstract

Nowak 0079000 The investigator and his colleagues develop new theories for the emergence of communication and cooperation among individuals. They focus on theories to explain the evolutionary transitions from simple indexical signals through to iconic and referential signs and finally to symbols and syntax in human language. These transitions are related to increases in fitness brought about through natural selection acting on communication strategies. They approach these problems using evolutionary game theory and population dynamics. The technical objective is to discover evolutionarily stable solutions to language games, such that when all individuals adopt these solutions, no mutant behavior can invade the population under the influence of natural selection. The challenge is to characterize the dynamics of the language game by specifying a realistic set of strategies and payoffs, and then to analyze the stability properties and accessibility of the equilibrium solutions. The mathematical equivalence of the game-theoretical replicator equation to the Lotka-Volterra equation offers a large tool box of techniques from nonlinear dynamics with which to explore the games. They also aim to relate these results to findings in information theory, in particular the formulation due to Shannon and Weaver (1949). Theoretical studies of language therefore provide a set of fascinating conceptual and technical challenges, as the strategies described with the models must be rich enough to capture the symbolic and open-ended properties of signals. While cooperation and animal communication are both areas in which much important theoretical work has been conducted, few attempts have been made to bring them together. The study of language provides an important focus for such a synthesis. Theoretical studies of the evolution of animal communication have stressed extravagant and costly ornamentation, using Darwin's theory of sexual selection, whereas evolutionary approaches to language tend to be represented by studies in historical linguistics within human societies. It is the objective of this project to occupy the center ground and provide a theoretical bridge between these two disciplines. Thus the work is neither a study of animal signals and cooperation nor a mathematical investigation of human syntax and grammar, but rather an attempt to reconstruct the continuous series of stages through which evolution might have proceeded from one to the other. The project does not not attempt to explain the glorious intricacies of human grammar studied by linguists, such as in X-bar theory. Instead it focusses on simple design features of animal communication and human language and attempts to describe their evolution in mathematical terms. These probably antedated grammar, phrase structure, anaphora and related linguistic constructs, and thereby provide some biological foundations upon which these later concepts might rest.