By Martin Dufwenberg , Flora Li , and Alec Smith

We study the effect of communication on beliefs and behavior in a three-stage trust game with punishment. We propose a novel behavioral mechanism, frustrationdependent anger, that links unmet payoff expectations with the willingness to forgo material payoffs to punish others. We conjecture that communication works through this mechanism to raise expectations about the likelihood of belief-dependent costly punishment and to increase trust, cooperation, and efficiency. In an experiment we allow communication in the form of a single pre-play message. We measure beliefs and our design permits the observation of promises and deception. The results are consistent with the theory that costly punishment results from belief-dependent anger and frustration. Promises drive the effect of communication on beliefs and broken promises lead to higher rates of costly punishment.

Preliminary draft. September 13, 2018

By Collin M McCabe

Large-scale cooperation, or the willingness of individuals to incur costs in order to help others, is a defining trait of the human species. However, cooperation poses a theoretical puzzle: since it is individually costly to cooperate, it seems that natural selection should favor non-cooperation (defection). Recently, it has been proposed that coordinated, collective punishment by cooperators of defectors can allow cooperation to invade a population of defectors. Here, we address the fact that in this previous analysis, coordinated punishment was only available to cooperators; defectors had no ability to punish cooperators (i.e. antisocial punishment was not possible). In other models, the inclusion of antisocial punishment has been shown to undermine the ability of punishment to promote cooperation. Thus we examine the effect of allowing coordinated antisocial punishment on the emergence of cooperation. Our results suggest that punishment confers no competitive advantage when it is a strategy available to both cooperators and defectors. While coordinated prosocial punishers can invade a population of non-punishing defectors, they cannot invade a population of coordinated antisocial punishers. These results question the conclusion that coordinated punishment played a central role in the evolution of human cooperation, and highlight the importance of not arbitrarily excluding antisocial punishment strategies from evolutionary models.

Research Gate · May 2014, 20p.