Thursday, May 29, 2008

On the emergence of punishment

source: Dreber et al (2008). Winners don't punish. nature 452:348-350.

Cooperation in a team ensures higher profit with lower risk in both ecological and sociological contexts. But how do we maintain cooperation? Through two basic concepts: reward and punishment. In biological terms reward is defined as a cost for the donor which translates into a profit for the receiver; however, in case of punishment, this manifests as a larger cost for the other end.

In a game theory set-up based on prisoner dilemma, the authors have tried to model punishment in a repeated setting. In this model, each individual can cooperate, defect or punish in every trial. The results from many rounds of pairwise competitions
(i.e. between different behavioral strategies) show that high ranking individuals tend not to punish and individuals with costly punishment behaviors end up in the lowest ranks.

Based on these models, the authors have concluded that costly punishment is not adaptive in cooperation games and the fact that they exist suggests reasons beyond the promotion of cooperation.

No comments: