Original article
When in Rome, do as the Romans do: the coevolution of altruistic punishment, conformist learning, and cooperation

https://doi.org/10.1016/j.evolhumbehav.2006.08.002Get rights and content

Abstract

We model the coevolution of behavioral strategies and social learning rules in the context of a cooperative dilemma, a situation in which individuals must decide whether or not to subordinate their own interests to those of the group. There are two learning rules in our model, conformism and payoff-dependent imitation, which evolve by natural selection, and three behavioral strategies, cooperate, defect, and cooperate, plus punish defectors, which evolve under the influence of the prevailing learning rules. Group and individual level selective pressures drive evolution.

We also simulate our model for conditions that approximate those in which early hominids lived. We find that conformism can evolve when the only problem that individuals face is a cooperative dilemma, in which prosocial behavior is always costly to the individual. Furthermore, the presence of conformists dramatically increases the group size for which cooperation can be sustained. The results of our model are robust: they hold even when migration rates are high, and when conflict among groups is infrequent.

Introduction

We are a cooperative species. Experimental evidence and field data show that humans often sacrifice resources in order to benefit nonrelatives, even when those who benefit are not expected to return the favor (Gintis, Bowles, Boyd, & Fehr, 2003). People sometimes use “altruistic punishment” to enforce cooperation, whereby they pay a cost in order to punish noncooperators whom they will never meet again (Fehr & Gaechter, 2000, Fehr & Gaechter, 2002, Ostrom et al., 1992). The combination of unrequited cooperation between nonrelatives and altruistic punishment is known as “strong reciprocity” (Gintis, 2000). Both of these components of strong reciprocity pose a puzzle for the standard evolutionary theories of cooperation: kin selection (Hamilton, 1964) and reciprocal altruism (Axelrod & Hamilton, 1981, Trivers, 1971).

Some authors argue that human cooperation may be explained by the selection of cultural traits at the group level (Bowles et al., 2003, Boyd & Richerson, 1985, Cavalli-Sforza & Feldman, 1981, Sober & Wilson, 1994). Assuming that cooperative groups out-compete less cooperative ones in the struggle for survival, then it may be possible for group level selective pressure to outweigh the maladaptive nature of altruism at the individual level. For this to occur, either noncooperative individuals must invade cooperative groups infrequently or else the amount of intergroup conflict must be very high.

Analytical models suggest that two factors play a crucial role in the emergence of cooperation: altruistic punishment and conformism (i.e., the tendency of individuals to imitate the most common form of behavior; see Boyd & Richerson, 1985, and Henrich & Boyd, 1998). Gintis (2000) proves that, when a group faces the threat of extinction, a small number of altruistic punishers may induce selfish individuals to behave cooperatively. Henrich and Boyd (2001) show that the presence of conformists may permit altruistic punishment to persist and thereby facilitate the emergence and survival of cooperation. Boyd, Gintis, Bowles, and Richerson (2003) report simulations that mimic the environment in which early hominids lived. They show that altruistic punishment enhances cooperative behavior when social learning takes the form of payoff-dependent imitation (i.e., when individuals imitate the most successful forms of behavior). However, this mixture of group selection and punishment cannot sustain cooperation in large groups if the migration rate between groups is high and conflict between groups is low.

Boyd and Richerson (2005) argue that cultural group selection is especially strong in human populations due to the fact that variation among human groups is maintained by an unusual combination of strong reciprocity and conformist social learning. Following their lead, this article uses a group selection approach to explore the coevolution of behavioral strategies and learning rules in the context of a “cooperative dilemma.” By cooperative dilemma we mean a situation in which an individual must choose whether or not to behave cooperatively, and benefit the group, or uncooperatively, and benefit himself. In our model, there are two social learning rules, conformism and payoff-dependent imitation, which evolve by natural selection, and three behavioral strategies, cooperate, defect, and cooperate, plus punish defectors, which evolve under the influence of the prevailing learning rules.

To the extent that our analysis is concerned with competing learning rules, it relates to the literature on endogenous learning. There is, however, one important difference. This literature is primarily concerned with social and individual learning as alternative ways to acquire information about the natural environment. Within such a framework, Boyd and Richerson (1985) demonstrate how the balance between social and individual learning depends on the accuracy of learning and the variability of the environment. Feldman, Aoki, and Kumm (1996) show that social learning can evolve if there is a fixed fitness cost to learning errors, while Henrich and Boyd (1998) show that social learning can evolve as long as the environment is not too variable.

The aims of this article are as follows: firstly, to determine if conformist transmission can evolve within the context of a cooperative dilemma, and secondly, to explore the impact of conformism on cooperation. Henrich and Boyd (2001) and Henrich (2004) observe that conformism to norms that are costly to the individual is most likely to evolve in tandem with individually beneficial conformism. Individuals may find it very difficult to distinguish between actions that are eventually costly to them and those that are eventually beneficial. Under these conditions, it may be best to conform blindly to the prevailing norm, even though this may sometimes involve taking actions that harm oneself. The alternative of doing it alone or seeking to be more selective may be worse. Henrich and Boyd (2001) and Henrich (2004) also observe, without elaboration, that costly conformism might evolve on its own through natural selection. In this article, we show the second observation is correct. We also show that the presence of conformists dramatically increases the group size for which cooperation can be sustained.

Section snippets

Model

We shall now develop a model in which evolution determines both the learning rules that individuals adopt and the behavioral strategies which they follow. The learning rules evolve at the biological level and the strategies chosen by individuals at any time are based on these rules. Our model builds on the work of Boyd et al. (2003), but departs from it by allowing conformist learning, and by making learning rules endogenous.

There are G groups, each of which has N members. Following Boyd et al.

Baseline scenario

Following Boyd et al. (2003), we simulate the model of the previous section for conditions that approximate those in which early hominids lived. Each simulation spans 2000 years of model time. Baseline parameters are given in Table 1. Most of these parameters are taken from Boyd et al. (2003) and we do not justify them here. Our model introduces two new parameters: the death rate and the mutation rate. We set the death rate at q=0.1, which implies a reproductive life of 10 years. The mutation

Conclusion

We have shown that conformism can evolve when the only problem that individuals face is a cooperative dilemma. There is no need to assume that costly conformism is a spin-off from individually beneficial conformism. We have also shown that conformism and altruistic punishment coevolve, allowing groups of greater size to sustain cooperation. This occurs because conformism preserves between-group variation and stabilizes punishment, and because punishment protects groups from the spread of

References (19)

There are more references available in the full text version of this article.

Cited by (0)

View full text