Salience, inductive reasoning and the emergence of conventions

This paper develops Lewis's theory of conventions to show that the spontaneous emergence of conventions depends on shared conceptions of salience. It offers a reconstruction of a mode of reasoning that is compatible with the emergence of conventions, and argues that such reasoning is pragmatically rational. This is a form of non-Bayesian inductive reasoning in which an individual's private and subjective conceptions of salience can influence the inferences she makes. This mode of reasoning is then shown to be pragmatically rational in a more general sense, relevant to problems of induction discussed in the philosophy of science.