THE EVOLUTION OF COOPERATION

Starting from a social-scientific view of human nature which assumes that individuals behave egoistically, U. of Michigan political scientist Axelrod seeks to account for cooperative behavior--through computer games. Axelrod invited eminent game theorists from economics, psychology, and other fields to enter a tournament centered on the now-famous Prisoner's Dilemma: is it in the interests of each of two individuals to cooperate or not cooperate under specific circumstances? (Briefly, it is in the interests of both parties to claim innocence and receive a middling punishment than for one to tell on the other; but it is in the interests of each separately to tell, not knowing what the other will do.) Programs were submitted for this ""iterated"" Prisoner's Dilemma game--i.e., there would be a continuous sequence of moves rather than one single choice--and the winner turned out to be ""TIT FOR TAT"": a solution which said that after the first move, the second player would reciprocate. When pitted against more complicated strategies, the model of reciprocity always won. This leads Axelrod to the conclusion that in a context of repeated interaction, a pattern of behaving ""nicely"" will dominate a pattern of ""meanness,"" and result in cooperative behavior. (Axelrod ran his tournament twice, the second time with many more participants and with all of them informed of the outcome of the first tournament; the TIT FOR TAT entry won again.) If non-game players intuitively respond that this is a strange way to state the obvious, Axelrod's turn to a historical example seen through TIT FOR TAT eyes shows that this is indeed the case. The example is WW I trench warfare and the well-known case of fraternization across the trenches. With the same soldiers facing each other day after day, some individuals on both sides adopted a strategy of not trying to shoot each other; only outsiders, not part of the game, broke the peace (e.g., lobbing artillery shells from a distance). Axelrod thinks his model explains the phenomenon, but it may just as well be said that the phenomenon explains the model. His lame suggestions for ""players"" in diplomacy and business are to get to know each other and to remember that the game will go on for a while. Then act nice, but retaliate immediately if the other side ""defects,"" forcing them to be nice too. Not a breakthrough.