Evolution of contingent altruism when cooperation is expensive

Ross A Hammond,Robert Axelrod
DOI: https://doi.org/10.1016/j.tpb.2005.12.002
Abstract:The ubiquity of cooperation has motivated a major research program over the last 50 years to discover ever more minimal conditions for the evolution of altruism. One important line of work is based on favoritism toward those who appear to be close relatives. Another important line is based on continuing interactions, whether between individuals (e.g., reciprocity) or between lines of descent in a viscous population. Here, we use an agent-based model to demonstrate a new mechanism that combines both lines of work to show when and how favoritism toward apparently similar others can evolve in the first place. The mechanism is the joint operation of viscosity and of tags (heritable, observable, and initially arbitrary characteristics), which serve as weak and potentially deceptive indicators of relatedness. Although tags are insufficient to support cooperation alone, we show that this joint mechanism vastly increases the range of environments in which contingent altruism can evolve in viscous populations. Even though our model is quite simple, the subtle dynamics underlying our results are not tractable using formal analytic tools (such as analysis of evolutionarily stable strategies), but are amenable to agent-based simulation.
What problem does this paper attempt to address?