Social Evolution in Imperfect World

Agent based simulation is a powerful tool for the research of complex systems such as human society. The important key word of this kind of research is “bounded rationality ” of the agents. From the stand-point, we adopt following two hypotheses as the basis of our analysis: (1) each agent interacts only with his neighbors, and (2) each agent behaves by mimicking its neighbors. How the society gropes its way towards equilibrium in this imperfect world? The purpose of this research is to investigate the role of mimicry in so-cial evolution. In this paper, we examined the growth of conventions in the society where individuals rely on hearsay to determine their action. Specific conditions as to which conventions are most likely to emerge from an amorphous state are clarified with computer simu-lations.