The Cultural Evolution of Indiscriminate Altruism in a Large Randomly Matching Population

Abstract

Altruism is hard to explain because altruistic acts are costly to the individuals who perform them. Although past work has identified conditions under which altruism can evolve, there is wide agreement among evolutionary theorists in the social sciences and biology that indiscriminate altruism cannot evolve in a large randomly matching population. Building on earlier work (Mark, 2002), I show that cultural transmission can create a cultural evolutionary force toward indiscriminate altruism in a large randomly matching population. The cultural transmission of a behavior (altruistic or selfish) disproportionately exposes those who acquire that behavior to that behavior prior to its acquisition. That is, individuals who have acquired an altruistic behavior through cultural transmission (i.e., who learned to behave altruistically from other people) were disproportionately exposed to the altruistic acts of others. Likewise, individuals who have acquired a selfish behavior through cultural transmission were disproportionately exposed to the selfish acts of others. Because of this disproportionate prior exposure, altruists have disproportionately benefitted from the altruistic acts of others, and selfish individuals have disproportionately been hurt by the selfish acts of others. If the benefits of being the target of altruistic acts increase one's attractiveness as a behavioral model, then a cultural evolutionary force toward altruism results.

[An appendix to the article is featured as an online supplement at the publisher's website.]

Keywords:

• altruism
• culture
• evolution

Acknowledgments

I thank Jonathan Bendor, Karen Cook, James Fearon, Herbert Gintis, Mark Granovetter, Michael Hannan, Douglas Heckathorn, Michael Macy, James March, Miller McPherson, Mark Mizruchi, Arthur Stinchcombe, Nobuyuki Takahashi, and Toshio Yamagishi for comments and discussion.

Notes

¹Although scholars do adopt alternative definitions (Piliavin and Charng, 1990), this definition is widely accepted among scholars who approach the topic of altruism from an evolutionary perspective (Hamilton, 1964; Wilson, 1975; Simon, 1990; Gintis, Bowles, Boyd, and Fehr, 2003; Axelrod, Hammond, and Grafen, 2004; Hammond and Axelrod, 2006).

²“Altruism is not … a kind of pleasant ornament of our social life, but one that will always be its fundamental basis” (Durkheim, 1893/1984, p. 173).

³The significance of this puzzle is emphasized by Boorman and Levitt's (1980) “principle of the failure of indiscriminate altruism” (p. 59). Axelrod and Hamilton (1981) also note that “if the payoffs are in terms of fitness, and the interactions between pairs of individuals are random and not repeated, then any population with a mixture of heritable strategies evolves to a state where all individuals are defectors” (pp. 1391–1392).

⁴The prisoner's dilemma has been closely associated with reciprocity which can evolve when the game is iterated (Trivers, 1971; Axelrod, 1984) and which does not meet the definition of altruism because a reciprocal relationship provides net benefits to each of the individuals in the relationship.

⁵These generations can be thought of as a very simplified representation of generations of individuals moving through a formal organization.

⁶Social dilemmas are situations that discourage individuals from behaving in ways that increase collective welfare. Yamagishi and Cook (1993) define a social dilemma as a situation in which many individuals each have the opportunity to cooperate or defect and two conditions are met: (a) for each individual, given any set of decisions by each of the other individuals to cooperate or defect, the individual's payoff will be higher if she defects than if she cooperates; and (b) the collective payoff is highest if everyone cooperates and lowest if no one does.

Focusing on cases where opportunities to give create social dilemmas means I am focusing on cases where the amount of altruism people practice positively affects collective welfare.

⁷The parameters a and c represent resources actors possess on entering the system. c is the portion of these resources the actor will lose if she gives. The portion a stays with every actor through the actor's second round. As I discuss below, the parameter a determines the minimum fitness an actor can possess and is the probability that an actor with the lowest fitness level is imitated by her young partner. Thus, inclusion of the parameter a in the model rests on the noncontroversial assumption that this probability might be greater than zero. Past work on the evolution of altruism toward kin (Hamilton, 1964), reciprocal altruism (Boorman and Levitt, 1980), and limited war (Maynard Smith, 1982) has employed analogous assumptions.

⁸An individual's genetic reproductive fitness is the individual's expected number of genetic offspring surviving to reproductive maturity (Boorman and Levitt, 1980).

⁹Actors of these age–behavior combinations entered the system with a + c, and they were given b by their first partner (a giver). Keepers do not give up their own benefit c in giving to someone else, and although the young givers will, they have not yet had the opportunity.