Knowledge, Awareness and Probabilistic Beliefs

Bayesian game theory investigates strategic interaction of players with full awareness but incomplete information about their environment. We extend the analysis to players with incomplete awareness, who might not be able to reason about all contingencies in the first place. We develop three logical systems for knowledge, probabilistic beliefs and awareness, and characterize their axiom systems. Bayesian equilibrium is extended to games with incomplete awareness and we show that it is consistent with common prior and speculative trade, when common knowledge of rationality is violated.

COMMON REASONING IN GAMES: A LEWISIAN ANALYSIS OF COMMON KNOWLEDGE OF RATIONALITY

We present a new class of models of players’ reasoning in non-cooperative games, inspired by David Lewis's account of common knowledge. We argue that the models in this class formalize common knowledge of rationality in a way that is distinctive, in virtue of modelling steps of reasoning; and attractive, in virtue of being able to represent coherently common knowledge of any consistent standard of individual decision-theoretic rationality. We contrast our approach with that of Robert Aumann (1987), arguing that the former avoids and diagnoses certain paradoxes to which the latter may give rise when extended in particular ways.

The Logical Antinomies of Knowledge

The critique of common knowledge of rationality (CKR) developed in the preceding chapters should convince researchers interested in explaining social reality to simply avoid the concept. The actual cost of abandoning CKR in terms of explaining social behavior is minimal because the Nash equilibrium concept itself is problematic when the recursive nature of interagent beliefs is important and the correlated equilibrium is by far the more cogent equilibrium concept. Nevertheless, it may seem curious that we must reject CKR even in situations where all players are in fact rational. What, after all, is the problem with assuming agents know something that is in fact true? This chapter discusses the pitfalls of naïve epistemic logic, the common knowledge of logicality paradox, the Surprise Examination problem, the modal logic of knowledge, and a solution to the Surprise Examination conundrum.

Rationalizability and Common Knowledge of Rationality

This chapter deals with the implications of rationality in normal form games. It first explores the ramifications of the rationalizability assumption and shows that in many cases rational individuals will not play rationalizable strategies. It argues that the informal reasoning supporting rationalizability must be replaced by a more rigorous analytical framework. This framework is known as epistemic game theory. Using epistemic game theory, it presents the argument that not rationality, but rather common knowledge of rationality, implies that players will only use rationalizable strategies. The chapter concludes by showing that there is no justification of the common knowledge of rationality assumption, and hence there is no reason to believe that in general rational players will choose rationalizable strategies. It strengthens this conclusion by showing that even assuming common knowledge of rationality, there is no reason for a rational player to conform to the iterated elimination of strongly dominated strategies.

Extensive Form Rationalizability

The extensive form of a game is informationally richer than the normal form since players gather information that allows them to update their subjective priors as the game progresses. For this reason, the study of rationalizability in extensive form games is more complex than the corresponding study in normal form games. There are two ways to use the added information to eliminate strategies that would not be chosen by a rational agent: backward induction and forward induction. The latter is relatively exotic (although more defensible). Backward induction, by far the most popular technique, employs the iterated elimination of weakly dominated strategies, arriving at the subgame perfect Nash equilibria—the equilibria that remain Nash equilibria in all subgames. An extensive form game is considered generic if it has a unique subgame perfect Nash equilibrium. This chapter develops the tools of modal logic and presents Robert Aumann's famous proof that common knowledge of rationality (CKR) implies backward induction. It concludes that Aumann is perfectly correct, and the real culprit is CKR itself. CKR is in fact self-contradictory when applied to extensive form games.

Interpreting Knowledge in the Backward Induction Problem

Robert Aumann argues that common knowledge of rationality implies backward induction in finite games of perfect information. I have argued that it does not. A literature now exists in which various formal arguments are offered in support of both positions. This paper argues that Aumann's claim can be justified if knowledge is suitably reinterpreted.