Altruistic Punishment Research Articles

Altruism and the Indispensability of Motives

Abstract In this paper we examine Fehr’s notions of “altruism”, “strong reciprocity” and “altruistic punishment” and query his ascription of altruism. We suggest that, pace Fehr, altruism cannot be defined behaviourally because the definition of altruism must refer to the motives of actors. We also advert to certain inconsistencies in Fehr’s usage of his terms and we question his explanation of altruism in terms of ‘social preferences’.

Social Relations Instead of Altruistic Punishment

Abstract Ernst Fehr’s experimental research on altruistic behaviour aims at superseding the classical homo oeconomicus in micro-economic behaviour theory. This essay discusses Fehr’s results from two points of view: first, in regard to the understanding of social action associated with the term “altruism”; second, in regard to the ‘anthropological’ strategy of research that is based on the laboratory method. Against the emphasis on altruism it will be argued that it misleads into providing a distorted description of social acting, and that, due to insufficient clarity about motives for acting, Fehr’s empirical results give evidence not of altruism but rather of phenomena of social recognition. The objection against the anthropological strategy will be that it makes visible only local phenomena within prevailing social conditions and that it thus assumes more than it explains.

Altruistic punishment and the origin of cooperation

How did human cooperation evolve? Recent evidence shows that many people are willing to engage in altruistic punishment, voluntarily paying a cost to punish noncooperators. Although this behavior helps to explain how cooperation can persist, it creates an important puzzle. If altruistic punishment provides benefits to nonpunishers and is costly to punishers, then how could it evolve? Drawing on recent insights from voluntary public goods games, I present a simple evolutionary model in which altruistic punishers can enter and will always come to dominate a population of contributors, defectors, and nonparticipants. The model suggests that the cycle of strategies in voluntary public goods games does not persist in the presence of punishment strategies. It also suggests that punishment can only enforce payoff-improving strategies, contrary to a widely cited "folk theorem" result that suggests that punishment can allow the evolution of any strategy.

The Economics of Altruistic Punishment and the Demise of Cooperation

Explaining the evolution and maintenance of cooperation among unrelated individuals is one of the fundamental problems in biology and the social sciences. Recent experimental evidence suggests that altruistic punishment is an important mechanism to maintain cooperation among humans. In this paper we explore the boundary conditions for altruistic punishment to maintain cooperation by systematically varying the cost and impact of punishment, using a subject pool which extends beyond the standard student population. We find that the economics of altruistic punishment lead to the demise of cooperation when punishment is relatively expensive and/or has low impact. Our results indicate that the 'decision to punish' comes from an amalgam of emotional response and cognitive cost-benefit analysis. Additionally, earnings are lowest when punishment promotes cooperation, suggesting that the scope for altruistic punishment as a means to maintain cooperation is limited.

Egalitarian motive and altruistic punishment

Altruistic punishment is a behaviour in which individuals punish others at a cost to themselves in order to provide a public good. Fehr and Gächter present experimental evidence in humans indicating that negative emotions towards non-cooperators motivate punishment, which, in turn, provokes a high degree of cooperation. Using Fehr and Gächter's original data, we provide an alternative analysis of their experiment that suggests that egalitarian motives are more important than motives for punishing non-cooperative behaviour. This finding is consistent with evidence that humans may have an evolutionary incentive to punish the highest earners in order to promote equality, rather than cooperation.

Egalitarian motive and altruistic punishment (reply)

Fehr and Gachter reply - Fowler et al. raise an important question1. They correctly argue that the desire to reduce inequality may motivate cooperators who altruistically punish free riders in our experiments2. Also, the evolutionary history of humans suggests that egalitarianism shaped many human cultures3 and that egalitarian motives may, therefore, be a powerful force behind the punishment of free riders. In addition, recently developed proximate theories4, which formalize the notion of inequality aversion, also suggest that egalitarian desires may be important. Fowler et al. contrast their egalitarianism hypothesis with our view that negative emotions against free riders drive punishment.

Social relations instead of altruistic punishment. Comments on Ernst Fehr's altruism research

Experimental economists have been trying for some time to discover the laws of behaviour in micro-social situations. Fehr's experimental research on altruistic behaviour attempts to correct the egoistic version of the concept of homo oeconomicus by resorting to the notion of altruistic dispositions. This article discusses Fehr's results from two points of view, namely in regard to the conception of social acting that is associated with altruism, and in regard to the research strategy associated with the laboratory method. The author argues that Fehr's concept of altruism distorts the representation of social acting and that, due to a lack of clarity concerning the motives of action, Fehr's empirical results pertain to phenomena of social recognition rather than to altruism. The charge against the research strategy is that it makes visible only local phenomena within the far wider field of general social conditions. Therefore, this approach presupposes more than it can explain.

False friends are worse than bitter enemies: “Altruistic” punishment of in-group members

One of the most critical features of human society is the pervasiveness of cooperation in social and economic exchanges. Moreover, social scientists have found overwhelming evidence that such cooperative behavior is likely to be directed toward in-group members. We propose that the group-based nature of cooperation includes punishment behavior. Punishment behavior is used to maintain cooperation within systems of social exchange and, thus, is directed towards members of an exchange system. Because social exchanges often take place within groups, we predict that punishment behavior is used to maintain cooperation in the punisher's group. Specifically, punishment behavior is directed toward in-group members who are found to be noncooperators. To examine this, we conducted a gift-giving game experiment with third-party punishment. The results of the experiment ( N=90) support the following hypothesis: Participants who are cooperative in a gift-giving game punish noncooperative in-group members more severely than they punish noncooperative out-group members.

NETWORK DEPENDENCE OF STRONG RECIPROCITY

Experimental evidence suggests that human decisions involve a mixture of self-interest and internalized social norms which cannot be accounted for by the Nash equilibrium behavior of Homo Economicus. This led to the notion of strong reciprocity (or altruistic punishment) to capture the human trait leading an individual to punish norm violators at a cost to himself. For a population with small autonomous groups with collective monitoring, the interplay of intra- and intergroup dynamics shows this to be an adaptive trait, although not fully invasive of a selfish population. However, the absence of collective monitoring in a larger society changes the evolution dynamics. Clustering seems to be the network parameter that controls maintenance and evolution of the reciprocator trait.

The neural basis of altruistic punishment.

Many people voluntarily incur costs to punish violations of social norms. Evolutionary models and empirical evidence indicate that such altruistic punishment has been a decisive force in the evolution of human cooperation. We used H2 15O positron emission tomography to examine the neural basis for altruistic punishment of defectors in an economic exchange. Subjects could punish defection either symbolically or effectively. Symbolic punishment did not reduce the defector's economic payoff, whereas effective punishment did reduce the payoff. We scanned the subjects' brains while they learned about the defector's abuse of trust and determined the punishment. Effective punishment, as compared with symbolic punishment, activated the dorsal striatum, which has been implicated in the processing of rewards that accrue as a result of goal-directed actions. Moreover, subjects with stronger activations in the dorsal striatum were willing to incur greater costs in order to punish. Our findings support the hypothesis that people derive satisfaction from punishing norm violations and that the activation in the dorsal striatum reflects the anticipated satisfaction from punishing defectors.

The Dynamics of Neighborhood Watch and Norm Enforcement

We analyze the dynamics of neighborhood watch programs in a local interaction framework. Agents can watch their neighbors' houses and, thus, deter burglars from breaking in. At the same time, agents also try to recruit their neighbors to join the neighborhood watch program. The probability of an agent joining the neighborhood watch program depends on the success of the program, i.e., whether burglaries continue to occur. We show that the punishment of burglars plays a dual role in this context. On the one hand, punishment deters burglaries if the level of punishment is sufficiently high. On the other hand, it also affects the probability of an agent joining the neighborhood watch program. In particular, we show that if recruitment is harder when burglaries do not occur, a legal policy attempting to improve deterrence using more severe punishment is suboptimal. In a second part, we extend our model to the study of norm enforcement in public goods dilemmas and show that our results remain valid if agents can punish each other (instead of burglars) for not contributing to the public good. Our paper, thus, provides a first analysis of the evolution of altruistic punishment in large populations with local interaction.

Third-party punishment and social norms

We examine the characteristics and relative strength of third-party sanctions in a series of experiments. We hypothesize that egalitarian distribution norms and cooperation norms apply in our experiments, and that third parties, whose economic payoff is unaffected by the norm violation, may be willing to enforce these norms although the enforcement is costly for them. Almost two-thirds of the third parties indeed punished the violation of the distribution norm and their punishment increased the more the norm was violated. Likewise, up to roughly 60% of the third parties punished violations of the cooperation norm. Thus, our results show that the notion of strong reciprocity extends to the sanctioning behavior of “unaffected” third parties. In addition, these experiments suggest that third-party punishment games are powerful tools for studying the characteristics and the content of social norms. Further experiments indicate that second parties, whose economic payoff is reduced by the norm violation, punish the violation much more strongly than do third parties.

Altruism, Altruistic Punishment and Social Investment

The concept of altruism is used in very different forms by computer scientists,economists, philosophers, social scientists, psychologists and biologists. Yet, in order to be useful in social simulations, the concept "altruism" requires a more precise meaning. A quantitative formulation is proposed here, based on the cost/benefit analysis of the altruist and of society at large. This formulation is applied in the analysis of the social dynamic working of behaviors that have been called "altruistic punishments", using the agent based computer model Sociodynamica. The simulations suggest that "altruistic punishment" on its own cannot maintain altruistic behaviors. "Altruistic behavior" is sustainable in the long term only if these behaviors trigger synergetic forces in society that eventually make them produce benefits to most individuals. The simulations suggest however that "altruistic punishment" may work as a "social investment", and is thus better called "decentralized social punishment". This behavior is very efficient in enforcing social norms. The efficiency of decentralized social punishment in enforcing norms was dependent on the type of labor structured of the virtual society. I conclude that what is called "altruistic punishment" emerges as a type of social investment that can evolve either through individual and/or group selection, as a successful device for changing or enforcing norms in a society. Social simulations will help us in better understanding the underlying dynamic working of such devices.

Altruism, evolution, and welfare economics

Henrich clearly presents the convincing evidence that the evolution of prosocial preferences—altruism and altruistic punishment—is both theoretically possible and empirically present in human populations. Henrich convinces us using only the most restrictive arguments. If one goes beyond the assumptions of one-gene, one-trait, or recognizes the existence of “social cognition” (Caporael, 1997), the case for the existence of prosocial behavior is overwhelming. Furthermore, while Henrich is probably correct in saying that large-scale social cooperation is unique to the human species, mounting evidence suggests that prosocial behavior is also present in such diverse non-human species as lions, meerkats, fire ants, and Arabian babblers (Clutton-Brock, 2002). So why is the reaction against the notion of pure altruism so strong in both economics and biology? A likely reason is that it goes against the dominant cultural belief in progress through competition among individuals. Both evolutionary biology and political economy matured in Victorian England where Darwin’s careful and cautious argument for evolution by natural selection was recast as “survival of the fittest”, a metaphor which fit the dog-eat-dog world of commercial society. Rugged individualism is reflected in the gene-based reductionism of biology as well as the agent-based reductionism of welfare economics (Bergh and van den Gowdy, 2003). In his contribution, Henrich establishes the theoretical basis for cultural group selection based on large-scale cooperation. He shows that the existence of prosocial behavior cannot be explained by individual-based characteristic alone such as kin selection, reciprocal altruism or costly signaling. This does not contradict Darwinian natural selection which can operate on any statistically reliable pattern. Patterns of culture may vary widely in human societies based on the different kinds of cultural transmission including conformist transmission (a short cut to reliable information), prestige-based transmission, punishment of non-conformists, and normative conformity by which people want their behavior to match that of the group. The interplay between these kinds of cultural transmission and the in-

The evolution of strong reciprocity: cooperation in heterogeneous populations

How do human groups maintain a high level of cooperation despite a low level of genetic relatedness among group members? We suggest that many humans have a predisposition to punish those who violate group-beneficial norms, even when this imposes a fitness cost on the punisher. Such altruistic punishment is widely observed to sustain high levels of cooperation in behavioral experiments and in natural settings. We offer a model of cooperation and punishment that we call strong reciprocity: where members of a group benefit from mutual adherence to a social norm, strong reciprocators obey the norm and punish its violators, even though as a result they receive lower payoffs than other group members, such as selfish agents who violate the norm and do not punish, and pure cooperators who adhere to the norm but free-ride by never punishing. Our agent-based simulations show that, under assumptions approximating likely human environments over the 100,000 years prior to the domestication of animals and plants, the proliferation of strong reciprocators when initially rare is highly likely, and that substantial frequencies of all three behavioral types can be sustained in a population. As a result, high levels of cooperation are sustained. Our results do not require that group members be related or that group extinctions occur.

Cultural group selection, coevolutionary processes and large-scale cooperation

In constructing improved models of human behavior, both experimental and behavioral economists have increasingly turned to evolutionary theory for insights into human psychology and preferences. Unfortunately, the existing genetic evolutionary approaches can explain neither the degree of prosociality (altruism and altruistic punishment) observed in humans, nor the patterns of variation in these behaviors across different behavioral domains and social groups. Ongoing misunderstandings about why certain models work, what they predict, and what the place is of “group selection” in evolutionary theory have hampered the use of insights from biology and anthropology. This paper clarifies some of these issues and proposes an approach to the evolution of prosociality rooted in the interaction between cultural and genetic transmission. I explain how, in contrast to non-cultural species, the details of our evolved cultural learning capacities (e.g., imitative abilities) create the conditions for the cultural evolution of prosociality. By producing multiple behavioral equilibria, including group-beneficial equilibria, cultural evolution endogenously generates a mechanism of equilibrium selection that can favor prosociality. Finally, in the novel social environments left in the wake of these cultural evolutionary processes, natural selection is likely to favor prosocial genes that would not be expected in a purely genetic approach.

The evolution of altruistic punishment.

Both laboratory and field data suggest that people punish noncooperators even in one-shot interactions. Although such "altruistic punishment" may explain the high levels of cooperation in human societies, it creates an evolutionary puzzle: existing models suggest that altruistic cooperation among nonrelatives is evolutionarily stable only in small groups. Thus, applying such models to the evolution of altruistic punishment leads to the prediction that people will not incur costs to punish others to provide benefits to large groups of nonrelatives. However, here we show that an important asymmetry between altruistic cooperation and altruistic punishment allows altruistic punishment to evolve in populations engaged in one-time, anonymous interactions. This process allows both altruistic punishment and altruistic cooperation to be maintained even when groups are large and other parameter values approximate conditions that characterize cultural evolution in the small-scale societies in which humans lived for most of our prehistory.

The Evolution of Strong Reciprocity: Cooperation in Heterogeneous Populations

How do human groups maintain a high level of cooperation despite a low level of genetic relatedness among group members? We suggest that many humans have a predisposition to punish those who violate group-beneficial norms, even when this imposes a fitness cost on the punisher. Such altruistic punishment is widely observed to sustain high levels of cooperation in behavioral experiments and in natural settings. We offer a model of cooperation and punishment that we call strong reciprocity: where members of a group benefit from mutual adherence to a social norm, strong reciprocators obey the norm and punish its violators, even though as a result they receive lower payoffs than other group members, such as selfish agents who violate the norm and do not punish, and pure cooperators who adhere to the norm but free-ride by never punishing. Our agent-based simulations show that, under assumptions approximating likely human environments over the 100,000 years prior to the domestication of animals and plants, the proliferation of strong reciprocators when initially rare is highly likely, and that substantial frequencies of all three behavioral types can be sustained in a population. As a result, high levels of cooperation are sustained. Our results do not require that group members be related or that group extinctions occur.

Altruistic punishment in humans.

Human cooperation is an evolutionary puzzle. Unlike other creatures, people frequently cooperate with genetically unrelated strangers, often in large groups, with people they will never meet again, and when reputation gains are small or absent. These patterns of cooperation cannot be explained by the nepotistic motives associated with the evolutionary theory of kin selection and the selfish motives associated with signalling theory or the theory of reciprocal altruism. Here we show experimentally that the altruistic punishment of defectors is a key motive for the explanation of cooperation. Altruistic punishment means that individuals punish, although the punishment is costly for them and yields no material gain. We show that cooperation flourishes if altruistic punishment is possible, and breaks down if it is ruled out. The evidence indicates that negative emotions towards defectors are the proximate mechanism behind altruistic punishment. These results suggest that future study of the evolution of human cooperation should include a strong focus on explaining altruistic punishment.

Spatial evolutionary game theory: Hawks and Doves revisited

We consider a spatial generalization of evolutionary game theory in which strategies are distributed over a spatial array of sites. We assume that the strategy corresponding to a given site has local interactions with the strategies sitting on neighbouring sites, and that the strategies change if neighbouring strategies are doing better. After briefly setting the stage with a formal definition of spatial evolutionary game theory, we consider the spatial extension of the Hawk-Dove game, and we show that the results are qualitatively different from those obtained from classical evolutionary game theory. For example, the proportion of Hawks in the population is in general lower in the spatial game than in the classical one. We also consider spatial generalizations of the extensions of the Hawk-Dove game obtained by including strategies such as Retaliator and Bully. Here, too, the results from the spatial game are very different from the classical results. In particular, with space Retaliator is a much more successful strategy than one would expect from classical considerations. This suggests that, in general, spatial structure may facilitate the evolution of strategies such as Retaliator, which do not necessarily prosper classically, and which are reminiscent of the \`nice', \`provokable' and `forgiving' strategies which seem to play a central role in the evolution of cooperation. The results indicate that including spatial structure in evolutionary game theory is a fruitful extension.