An individual-level model that may help to explain why humans punish selfish behaviours that do not affect them personally is presented in Nature. The study validates the model in human adults, and shows that third-party punishment (TPP) sends a signal of the punisher's trustworthiness and provides reputational benefits that outweigh the initial costs of punishing.
TPP seems to be universal across cultures and is thought to have evolved through group selection, because TPP promotes cooperation, and selection may therefore favour groups with third-party punishers. However, whether TPP confers any individual-level benefits on the punisher has remained unclear.
David Rand, Jillian Jordan and colleagues developed a theoretic model of TPP and validated their model in economic game experiments with human adults. They show that individuals can evolve to punish selfish others to signal that they are not selfish themselves. They demonstrate that TPP is a strong signal of trustworthiness when it is the only signal available, but it weakens when the punisher has the option to instead engage in costly helping.The authors note that although TPP seems to be altruistic, their study shows that it can actually serve the punisher's own self-interests by boosting the perception of the punisher's trustworthiness.