Dangerous but contagious altruism: recruitment of group members and reform of cooperation style through altruism in two modified versions of Hammond and Axelrod's simulation of ethnocentric cooperation

Abstract

This paper explores the possibility of recruiting other agents to one's own kind and changing their cooperation style to one's own through altruistic generosity in a simulation. This is done by modifying Ross Hammond and Robert Axelrod's well-known simulation, which shows that ethnocentric behavior is the most successful strategy in a spatialized game of cooperation where color is the only attribute that agents can act on. Their simulation is altered in two steps. First, the simulation is changed so that each agent can “recruit” neighboring agents to its own color with a low probability R through out-group altruism toward non-cooperative out-group members. If recruitment is successful, there is also a probability M that the converted agent “morally reforms,” that is, changes cooperation style to the cooperative agent's style. The result is that the strategy of altruism can successfully compete with ethnocentric strategies if it leads to recruitment around 1–2% of the time, and a change of cooperation style to altruism around 0.4–0.5% of the time. Second, the simulation is altered so that only agents of one color, green, can recruit the other colors through altruism. The most obvious result of this simulation is that agents of this group become increasingly dominant with an increasing probability of recruitment. More interesting is that the overall proportion of altruists decreases and ethnocentrism becomes more dominant as a cooperation strategy compared to the first modified simulation, although not as dominant as in the original simulation by Hammond and Axelrod. That is, green can dominate the board through recruitment of other colors even though only some of the green agents are “proselytizing” altruists. The simulations have bearing on (1) how altruistic behavior spreads through cultural transmission in a population and (2) the historical problem of why apparently self-destructive behaviors of generosity and non-self-defense toward out-group individuals seems to have contributed to the expansion of religious movements in certain historical periods.