In her book Rationality and Coordination (Cambridge University Press,

Cristina Bicchieri brings together (and adds to) her own contributions to game theory and the philosophy of economics published in various journals in the period 1987-1992. The book, however, is not a collection of separate articles but rather a homogeneous unit organized around some central themes in the foundations of non-cooperative game theory. Bicchieri's exposition is admirably clear and well organized. Somebody with a good knowledge of game theory would probably benefit mainly from reading the second part of Chapter 3 (from Section 3.6 onward) and Chapter 4. On the other hand, those who have had little exposure to game theory, would certainly benefit from reading the entire book. I shall begin with an overview of the content of the book and then offer some critical comments on what I consider to be the most important part of it. Chapter 1, entitled 'Rationality and predictability', deals with 'the role individual rationality plays in predicting and explaining economic phenomena' (p. 1). Bicchieri distinguishes between two kinds of ration-ality: practical and epistemic. Practical rationality coincides with expected utility maximization, while epistemic rationality means that the agent's beliefs are rational. Bicchieri argues that, with very few exceptions, agents must be credited with both practical and epistemic rationality if individual as well as systemic behavior in interactive contexts is to be explained satisfactorily. In particular, it is necessary to understand the process by which agents form their beliefs about the environment, which includes the actions of the other agents. This is the central theme of the entire book. The chapter ends with a discussion and criticism of the rational expectations hypothesis. Chapter 2, entitled 'Equilibrium', contains a general discussion of games in normal (or strategic) form, introducing the notions of pure and mixed strategy, weakly and strictly dominant strategy, rationalizability and Nash equilibrium. The discussion also covers the concepts of focal point, Pareto dominance, trembling-hand perfect equilibrium and proper equilibrium. As Bicchieri states at the beginning of the chapter (p. 33), 'the general thesis ... is that, in many non-trivial cases, knowledge of the game being played and self-evident principles of rationality are not sufficient to guarantee that an equilibrium will be attained'. For a reader with limited knowledge of game theory Chapter 2 is useful reading. However, Bicchieri's discussion leaves out some new developments in the literature, which are highly relevant to the issues that are central to this book. For example, a …