Putting the dilemma back into Prisoner's Dilemma

Abstract

What is wrong with so much of the experimental work using the Prisoner's Dilemma game? Why does the Prisoner's Dilemma seem so exciting and the experimental work so often misdirected? The main reason is that the dilemma has been ignored. In not allowing or severely restricting the communication that can occur between the players, researchers have treated the Prisoner's Dilemma as a game between two isolated rational (selfinterested) units, and consequently have not examined the social mechanisms that human groups develop to enable cooperation to occur when the members' interests are somewhat opposed in the interesting way that the Prisoner's Dilemma describes. These mechanisms include norms, rules, contracts, and group solidarity and loyalty. In particular, the emphasis on developing indices of propensity to make cooperative choices based on the values in the matrix is misdirected because these indices reflect how much the game resembles one of pure conflict or perfect identity of interests between the players. These two extremes are, in a way, uninteresting because cooperative behavior is unproblematic. What is needed is an elucidation of the concept dilemma and an index which reflects the degree of dilemma in a Prisoner's Dilemma game, rather than how much the game resembles a zero-sum game or a game of perfect identity of interests, so that the ways in which groups solve these dilemmas can be studied. The two-person Prisoner's Dilemma takes the familiar form shown in Table 1.