Rationality, melioration and law-of-effect models for choice.

Abstract

Economists usually assume that human choice behavior is rational, by which they mean that it maximizes some utility function. Psychologists are more impressed by the evident irrationality of behavior and tend to look for choice mechanisms (which cannot act rationally under all conditions). Richard Herrnstein (1990a) has recently argued that the choices of pigeons and people are dynamically driven by a moment-by- moment tendency to equalize payoff per unit of activity invested, a mechanism he terms melioration. I argue that economic models are not so bad, and melioration is not so good, as Herrnstein contends. The problem with rational choice is not that it is wrong but that it is too flexible. The problems with melioration are that it is poorly defined and refers only to events in the recent past (local events).