Rationality of Belief Or : Why Savage ' s

Economic theory reduces the concept of rationality to internal consistency. As far as beliefs are concerned, rationality is equated with having a prior belief over a “Grand State Space”, describing all possible sources of uncertainties. We argue that this notion is too weak in some senses and too strong in others. It is too weak because it does not distinguish between rational and irrational beliefs. Relatedly, the Bayesian approach, when applied to the Grand State Space, is inherently incapable of describing the formation of prior beliefs. On the other hand, this notion of rationality is too strong because there are many situations in which there is not sufficient information for an individual to generate a Bayesian prior. It follows that the Bayesian approach is neither sufficient not necessary for the rationality of beliefs. 1. Rationality of Belief and Belief Formation Economic theory is both the birthplace and the prime application of the rational choice paradigm. Throughout the 20 century, economics has relied on rationality, and refined the definition of rational choice, offering concepts such as subjective expected utility maximization and Nash equilibrium, which have proved useful in several other disciplines. ∗ Our thinking on these issues was greatly influenced by discussions with many people. In particular, we wish to thank Edi Karni, Marion Ledwig, Dan Levin, Stephen Morris, Peter Wakker, and two anonymous referees for comments and references. 1 Tel-Aviv University, and HEC, Paris. [email protected] 2 University of Pennsylvania. [email protected] 3 Tel-Aviv University and Ohio State University. [email protected]