We compare Colman's proposed “psychological game theory” with the existing literature on psychological games (Geanakoplos et al. 1989), in which beliefs and intentions assume a prominent role. We also discuss experimental evidence on intentions, with a particular emphasis on reciprocal behavior, as well as recent efforts to show that such behavior is consistent with social evolution.

This comment focuses on the informational distinction Brian Skyrms makes between rational choice theories of the social contract and theories based on evolutionary dynamics. The basic point is that to dismiss the rational choice method because of the restrictive informational assumptions may discount interesting work done in the area of bounded rationality. Further, the comment argues that combining the best elements of both approaches into an evolutionary theory of boundedly rational agents can improve the power of social contract theories. To (...) illustrate the point, we work through an example of analysing the data from a bargaining experiment. (shrink)