Towards a Theory of Convention

Some thirty years ago Lewis published his Convention: A philosophical Study (Lewis 1969). Besides exciting the logical community by providing the seminal analysis work on common knowledge, it also laid the foundations for the formal approach to the study of social conventions by means of game theory. Like for the study of common knowledge much has happened in this latter field since then. The theory of convention has been developed and extended so as to include multiple types as well as a basis for the study of social norms. However, classical game theory is currently undergoing severe crisis as a tool for understanding and explaining social phenomena; a crisis emerging from the problem of equilibrium selection around which any theory of convention must revolve. The so-called evolutionary turn in game theory marks a transition from the classical assumptions of rationality and common knowledge of such to evolutionary game theoretical frameworks inspired by the models of (Maynard Smith & Price 1973), (Taylor & Jonker 1978) and (Maynard Smith 1982). By providing an account of equilibrium selection these are thought to work as well-defined metaphors of learning processes upon which a revised theory of convention may be erected. In this article I outline one way this might be done, as well as point to some problems and perspectives that the evolutionary turn leaves in its tracks when brought to serve in a theory of convention.