The Reasons for Being of Two Minds: A Comment on Schelling's “Enforcing Rules on Oneself”

Abstract

T.C. Schelling's subject is our frequent desire to use our present judgment to influence our own future behavior because we know or suspect that our future judgment will be different; we will be of two minds about what to do. His paper is intended to suggest a structure, a hierarchy, of the rules we use to exercise that influence and to evaluate the characteristics of those rules that make for their success or failure. At the outset he compares these issues of self-control with those interpersonal conflicts on contract performance in which society, through the courts, is willing to intervene. The subject is rules. Not rules as routinized, repetitive behavior-habituation in the manner of Scitovsky or Nelson and Winter-that economizes on our bounded rationality, leaving more of it available for unresolved issues. Here rules are standards of acceptable behavior, set up to guide our future selves. Professor Schelling's discussion of the dimensions of those behaviors is characteristically rich and nicely illustrated, distinguishing between sins of omission and commission; the periodicity or constancy of the need for selfcontrol; the ease or difficulty of monitoring the desired behavior; its ambiguity or visibility; target levels as zero or average values, discrete or continuous; differences in the time lags between moral lapse and the unwanted behavior; internal-thought behavior as against external-physical behavior. It is a useful catalogue, essential to any full understanding of this sort of behavior. But the sense of structure-of organizing these complicated behaviors-that Schelling promises does not clearly emerge. Part of the problem is complexity: there is a lot going on here with different characteristics of behavior interacting with different incentives over different time periods. So, for instance, with considerable insight Schelling suggests that piecemeal rules may work better to induce desired actions (taking things one at a time) while all-or-nothing rules may work best to induce desired abstinence (all must be lost with that first drink). But taking it the next step, if