How I Learned to Stop Worrying and Love the Prisoner’s Dilemma

Abstract

This essay argues that rationality should not be identified with expected utility maximization. It focuses on the Prisoner’s Dilemma and claims that it exhibits the clash between two rival accounts of practical rationality. The orthodox theory is dubbed a best-reply theory, as it counsels that agents choose the utility-maximizing act (or strategy). By contrast, a P-optimal theory would have agents seek to bring about an outcome which is Pareto-optimal. It is evident that in the Prisoner’s Dilemma, the P-optimal outcome is P-superior to the best-reply outcome. The former requires that agents cooperate; this requires that agents realize their objectives (as measured by utility maximization) by acting to bring about some agreed outcome rather than by choosing for oneself on the basis of one’s expectations of others’ choices. The orthodox, best-reply theories insist that any cooperation must be grounded in a deeper level of non-cooperation, a view which seems to verge on incoherence. Persons are concerned to realize their aims; practical reason must surely be understood as conducive to this realization. The essay concludes with the thought that we must P-optimize; there is no going it alone. That may be a practical problem, but only a practical problem. Rightly understood, the Prisoner’s Dilemma shows the way to achieve beneficial interaction—and the way not to achieve it. We should relabel it “the Cooperator’s Opportunity”.