Commitment and Weakness of Will in Game Theory and Neoclassical Economics

A decision-maker may be said to be rational if his decisions fulfill two conditions: first, they are consistent with his preferences in the light of his beliefs, and second, his beliefs are consistent with the available evidence. Either of these conditions may be fulfilled exactly or only approximately. Selten writes (1975 p. 320) “ … game theory is concerned with the behavior of absolutely rational decision makers whose capabilities of reasoning and memorizing are unlimited …. ” The same is true of neoclassical economics. But it is not an open question whether this “concern ” is descriptive of actual human decision-makers. There is ample evidence to reject it. No rational decision-maker, choosing among theories, can choose one that is based on absolute rationality. Real human rationality is bounded1, and although boundedly rational decisions may approximate those that would be made by one of Selten’s decision-makers, theories based on the postulate of absolute rationality can be quite misleading. (Akerlof and Yellen 1985, Akerlof 2001) Nevertheless, absolute rationality retains some intellectual interest. On the one hand, the “possible world ” inhabited by absolutely rational beings is a possible world tha