Individuals are reluctant to cooperate for public goods, especially for risky ones that involve violent confrontations where they may get hurt while others freeride on their efforts. Cooperation can be fostered by norms, incentives, reputations, and training for risky situations, which take time to develop and depend on monitoring and reliable information transmission through a group's network.
... [Show full abstract] Nevertheless, untrained amateurs engage in impromptu street fights, rescue victims at disasters, and participate in protests against oppressive regimes, largely without these mechanisms. An Ising model that reflects asymmetric payoffs of cooperation and defection demonstrates that under high uncertainty, agitation by situational turmoil such as aggressing opponents is sufficient to get cooperation started. At a critical level of turmoil, just a few accidental cooperators entail a burst of cooperation among the rest. If the proportion of unconditional defectors is above a critical threshold, however, bursts do not occur, which is confirmed by video data of street violence. The model also explains why in larger groups, violent subgroups are typically small whereas larger groups have a better chance to win. In the end we discuss how the model can be generalized by incorporating the usual mechanisms to explain cooperation without turmoil.