Epistemic Foundations of Game Theory

Titelangaben

Pacuit, Eric ; Roy, Olivier:
Epistemic Foundations of Game Theory.
In: Stanford Encyclopedia of Philosophy. - Stanford, Calif. : The Metaphysics Research Lab, Stanford Univ. , 2016

Volltext

Link zum Volltext (externe URL):

Abstract

Foundational work in game theory aims at making explicit theassumptions that underlie the basic concepts of thediscipline. Non-cooperative game theory is the study of individual,rational decision making in situations of strategic interaction. Thisentry presents the epistemic foundations of non-cooperativegame theory (this area of research is called epistemic gametheory)., Epistemic game theory views rational decision making in games assomething not essentially different from rational decision makingunder uncertainty. As in Decision Theory (Peterson 2009), to chooserationally in a game is to select the “best” action inlight of one’s beliefs or information. In a decision problem,the decision maker’s beliefs are about a passive state ofnature, the state of which determines the consequences of heractions. In a game, the consequences of one’s decision depend onthe choices of the other agents involved in the situation(and possibly the state of nature). Recognizing this—i.e., thatone is interacting with other agents who try to choose the best courseof action in the light of their ownbeliefs—brings higher-order information into thepicture. The players’ beliefs are no longer about a passive orexternal environment. They concern the choices and theinformation of the other players. What one expects of one’sopponents depends on what one thinks the others expect from her, andwhat the others expect from a given player depends on what they thinkher expectations about them are., This entry provides an overview of the issues that arise when onetakes this broadly decision-theoretic view on rational decision makingin games. After some general comments about information in games, wepresent the formal tools developed in epistemic game theory andepistemic logic that have been used to understand the role ofhigher-order information in interactive decision making. We then showhow these tools can be used to characterize known “solutionconcepts” of games in terms of rational decision making inspecific informational contexts. Along the way, we highlight a numberof philosophical issues that arise in this area.