Rationality and the Myth of the Maximum

Abstract

It is the business of all normative inquiry-ethics, decision theory, welfare economics and the like-to formulate rational decision-making principles. This is sometimes the business of the social sciences too. Now and then a social scientist seeking to explain someone's behavior will look for a rational principle behind his decisions-a reasonable decision-making rule that might have led the subject to act as he did. Rational decision-making principles (DMPs) may be as different as the various decision-making problems for which they are designed. An acceptable method of picking political leaders and a reasonable rule for budgeting household expenses may bear little resemblance to one another. What I want to discuss, though, is not the differences among rational DMPs, but the underlying similarities: What is common to all rational DMPs ? What can we reasonably demand of any DMP? The traditional answer is the Maximization Thesis: a rational DMP specifies something to be maximized. On this view, every decision-making problem is a maximization problem; to optimize is to maximize something; a rational agent (one who is motivated by a rational DMP) is perforce a maximizer. I've argued before that the Maximization Thesis, even when trimmed to its bare bones, is untenable.2 This time the target of my critical fire is not so much the Maximization Thesis itself as