The Adaptive Rationality of Interpersonal Commitment

Abstract

Why are people inclined to build friendships and maintain durable, non-reproductive relationships? Previous computational modeling work showed that it can be an efficient survival strategy to choose interaction partners based on relationship length, even if, as a consequence, individuals become unconditionally cooperative in long-term relationships (interpersonal commitment). Such committed individuals can outperform conditional cooperators who play in a fair, reciprocal manner (e.g. tit for tat). However, previous studies did not conduct a sufficiently strict test of the viability of commitment because they did not account for exploiters who specifically take advantage of the tolerance of commitment players. We allow for this by extending previous studies with the possibility of randomly mutating strategies under evolutionary pressures, and thus give a much larger coverage of an infinite strategy space. Our results point to the lack of stable strategies: we find that emerging populations alternate between temporarily stable states. We also show that the viability of strategies increases with increasing levels of interpersonal commitment, and that the effect of interpersonal commitment on viability is larger than the effect of fairness.