Mutual Commitment, Norm Formation and Indirect Regulation of Agent Society

Norm formation is one of the most basic processes of social interaction. To clarify this process, we introduce mutual commitment and meta commitment among agents. R. Axelrod has analyzed the commitment structure of norm formation and the collapsing process of the norm as his norm and meta norm games [1] . He analyzes the games with genetic algorithm. In this paper we generalize the games as mutual commitment processes of agents [6] . We also introduce the commitment by central authorities such as a government that enforces legal policies. We deal with the model from theoretical point of view. For the purpose of describing a social learning dynamics, we introduce replicator dynamics (RD). RD is usually used in evolutionary game theory. We induce RD from stochastic learning process. At first we reformulate the norm game as a coupling model between the replicator dynamics of alternatives {C,D} and the dynamics of normative attitude. We also extend this model and introduce several types of centralized and decentralized commitment mechanisms on agent societies.