The target article by McCullough et al. is a psychological version of reputation models pioneered by biologist Robert Trivers (Reference Trivers1971), who called revenge “moralistic aggression” and the associated emotion “moral outrage,” and economist Robert Frank (Reference Frank1988), who called it “passion within reason.” This model of reputation effects, with similar assumptions, was deepened in the economics literature (Fudenberg et al. Reference Fudenberg, Levine and Maskin1994; Kreps & Wilson Reference Kreps and Wilson1982; Schmidt Reference Schmidt1993).
The psychological dimension to the reputational model enters because the proximate motives for seeking revenge in human societies are generally not to enhance reputation, but rather to obtain satisfaction by harming the offender. Moreover, revenge is common in humans even when its cost is greater than its expected reputational gains, a fact that is difficult to reconcile with the biological and economic reputational models. One example is strong reciprocity which, in a social dilemma context where there is no opportunity for reputation formation, involves being predisposed to cooperate initially and as long as others reciprocate, and to punish non-contributors at personal cost (Bowles & Gintis Reference Bowles and Gintis2011; Fehr & Gachter Reference Fehr, Gachter, Putterman and Ben-Ner1998; Gintis Reference Gintis2000; Gintis et al. Reference Gintis, Bowles, Boyd and Fehr2005). Another example is third party punishment where, even under conditions of anonymity, an individual punishes an agent who has harmed a stranger, or who has committed a social norm violation that does not affect the punisher (Buchholtz et al. Reference Buchholtz, Asplund, Dux, Zald, Gore, Jones and Marois2008; Fehr & Fischbacher Reference Fehr and Fischbacher2004).
McCullough et al. explain the psychology of revenge and its widespread occurrence in situations where deterrence is not involved by arguing that in our hunter-gatherer prehistory, revenge had positive fitness effects by establishing the reputation of the revenge-seeker as an individual who is not to be exploited and who will defend his family and allies. A genetically based human cognitive deterrence system thereby became adaptive. This deterrence system persists in modern life where it is maladaptive because, by contrast with the Pleistocene, contemporary social conditions include many one-shot and anonymous interactions. The absence of one-shot and anonymous interactions in the human hunter-gatherer societies of the Pleistocene explains why evolution gave rise to a cognitive deterrence system that does not condition revenge behavior on the level of expected future returns.
There are three problems with this argument. First, modern humans routinely distinguish between situations in which reputation building is possible and situations in which it is not, and cooperate much more in the former case (Bowles & Gintis Reference Bowles and Gintis2011, Ch. 3). Assuming that this capacity is a cognitive adaptation, there must have been frequent and fitness-relevant non–reputation-building interactions in our evolutionary history.
Second, even in a world of repeated interactions among well-acquainted individuals, anonymous interactions (e.g., hiding a kill from others) are common in contemporary hunter-gatherer societies (Kaplan et al. Reference Kaplan, Hill, Hawkes and Hurtado1984; Hawkes Reference Hawkes1993; D. S. Wilson Reference Wilson1998), and hence doubtless in Pleistocene and early Holocene times as well. Indeed, such behavior is routinely recorded in chimpanzees, and hence is likely an attribute of our most recent common ancestor some eight million years ago (Boehm Reference Boehm2011; de Waal Reference de Waal1997).
Finally, it is not the case that general individuals in prehistoric hunter-gatherer communities were life-long social interactants. The evidence supporting this assertion comes from Late Pleistocene climate records, archaeological records of the causes of death, and genetic evidence bearing on exogamy and migration. Neither the likely size of groups, nor the degree of genetic relatedness within groups, nor the typical demography of foraging bands is favorable to the view that Late Pleistocene human groups sustained cooperation through either kin-based or reciprocal altruism. Rather our ancestors were cosmopolitan, civic-minded, and warlike. They benefited from far-flung coinsurance, trading, mating, and other social networks, as well as from coalitions and, if successful, warfare with other groups (Bowles & Gintis Reference Bowles and Gintis2011, Ch. 6).
I offer the following sketch of an alternative model of revenge-seeking, which treats this motive as a form of moral behavior: Individuals seek revenge not when they have been hurt, but when they have been morally wronged, or when they countenance others violating the social norms of the community.
Like other organisms, humans have a preference ordering over states of affairs, and they act to best achieve their desired states of affairs, given the material and informational resources available to them. These preferences are strongly influenced by genetic predispositions, but they are affected by group culture. Culture for humans is not merely a set of learned techniques, but also a set of moral values that are internalized by group members (Parsons Reference Parsons1967). The capacity to internalize values through socialization is an evolved adaptation of humans (Durkheim Reference Durkheim1902/1967; Gintis Reference Gintis2003; Simon Reference Simon1990), and accounts both for cultural diversity across societies and (limited) cultural stability across generations.
Human social cooperation is governed not by genes alone, but by social norms that are widely embraced by social actors, and act as moral values present in individual preference functions. We term these other-regarding preferences (Gintis Reference Gintis2009). Individuals incorporate moral values in their actions by trading off the costly attainment of other-regarding goals against self-regarding goals. The ubiquity of altruistic cooperation and altruistic punishment around the world suggests that these values are strong genetic predispositions, although the evidence indicates that their expression is strongly modulated by local cultural values (Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr and Gintis2004; Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath, Alvard, Barr, Ensminger, Henrich, Hill, Gil-White, Gurven, Marlowe, Patton and Tracer2005; Reference Henrich, McElreath, Barr, Ensminger, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2006). There are several plausible models of the evolution of these predispositions (Bowles & Gintis Reference Bowles and Gintis2011; Boyd et al. Reference Boyd, Gintis and Bowles2010; Gintis Reference Gintis2000). There are also plausible evolutionary models of the internalization of norms, the mechanism by which moral values become represented in the individual's preference ordering (Boehm Reference Boehm2011; Bowles & Gintis Reference Bowles and Gintis2011; Gintis Reference Gintis2003).
In this alternative framework, revenge and forgiveness can be self-regarding acts aimed at deterring malefactors and warning others of the cost of aggression. But revenge can also be an other-regarding act carried out to redress wrongs on a purely moral level. This explains why individuals punish not only those who hurt themselves, their families, and their allies, but more generally those who violate social norms. It also explains why individuals will seek vengeance against aggressors even when there is no possible deterrent effect.
The target article by McCullough et al. is a psychological version of reputation models pioneered by biologist Robert Trivers (Reference Trivers1971), who called revenge “moralistic aggression” and the associated emotion “moral outrage,” and economist Robert Frank (Reference Frank1988), who called it “passion within reason.” This model of reputation effects, with similar assumptions, was deepened in the economics literature (Fudenberg et al. Reference Fudenberg, Levine and Maskin1994; Kreps & Wilson Reference Kreps and Wilson1982; Schmidt Reference Schmidt1993).
The psychological dimension to the reputational model enters because the proximate motives for seeking revenge in human societies are generally not to enhance reputation, but rather to obtain satisfaction by harming the offender. Moreover, revenge is common in humans even when its cost is greater than its expected reputational gains, a fact that is difficult to reconcile with the biological and economic reputational models. One example is strong reciprocity which, in a social dilemma context where there is no opportunity for reputation formation, involves being predisposed to cooperate initially and as long as others reciprocate, and to punish non-contributors at personal cost (Bowles & Gintis Reference Bowles and Gintis2011; Fehr & Gachter Reference Fehr, Gachter, Putterman and Ben-Ner1998; Gintis Reference Gintis2000; Gintis et al. Reference Gintis, Bowles, Boyd and Fehr2005). Another example is third party punishment where, even under conditions of anonymity, an individual punishes an agent who has harmed a stranger, or who has committed a social norm violation that does not affect the punisher (Buchholtz et al. Reference Buchholtz, Asplund, Dux, Zald, Gore, Jones and Marois2008; Fehr & Fischbacher Reference Fehr and Fischbacher2004).
McCullough et al. explain the psychology of revenge and its widespread occurrence in situations where deterrence is not involved by arguing that in our hunter-gatherer prehistory, revenge had positive fitness effects by establishing the reputation of the revenge-seeker as an individual who is not to be exploited and who will defend his family and allies. A genetically based human cognitive deterrence system thereby became adaptive. This deterrence system persists in modern life where it is maladaptive because, by contrast with the Pleistocene, contemporary social conditions include many one-shot and anonymous interactions. The absence of one-shot and anonymous interactions in the human hunter-gatherer societies of the Pleistocene explains why evolution gave rise to a cognitive deterrence system that does not condition revenge behavior on the level of expected future returns.
There are three problems with this argument. First, modern humans routinely distinguish between situations in which reputation building is possible and situations in which it is not, and cooperate much more in the former case (Bowles & Gintis Reference Bowles and Gintis2011, Ch. 3). Assuming that this capacity is a cognitive adaptation, there must have been frequent and fitness-relevant non–reputation-building interactions in our evolutionary history.
Second, even in a world of repeated interactions among well-acquainted individuals, anonymous interactions (e.g., hiding a kill from others) are common in contemporary hunter-gatherer societies (Kaplan et al. Reference Kaplan, Hill, Hawkes and Hurtado1984; Hawkes Reference Hawkes1993; D. S. Wilson Reference Wilson1998), and hence doubtless in Pleistocene and early Holocene times as well. Indeed, such behavior is routinely recorded in chimpanzees, and hence is likely an attribute of our most recent common ancestor some eight million years ago (Boehm Reference Boehm2011; de Waal Reference de Waal1997).
Finally, it is not the case that general individuals in prehistoric hunter-gatherer communities were life-long social interactants. The evidence supporting this assertion comes from Late Pleistocene climate records, archaeological records of the causes of death, and genetic evidence bearing on exogamy and migration. Neither the likely size of groups, nor the degree of genetic relatedness within groups, nor the typical demography of foraging bands is favorable to the view that Late Pleistocene human groups sustained cooperation through either kin-based or reciprocal altruism. Rather our ancestors were cosmopolitan, civic-minded, and warlike. They benefited from far-flung coinsurance, trading, mating, and other social networks, as well as from coalitions and, if successful, warfare with other groups (Bowles & Gintis Reference Bowles and Gintis2011, Ch. 6).
I offer the following sketch of an alternative model of revenge-seeking, which treats this motive as a form of moral behavior: Individuals seek revenge not when they have been hurt, but when they have been morally wronged, or when they countenance others violating the social norms of the community.
Like other organisms, humans have a preference ordering over states of affairs, and they act to best achieve their desired states of affairs, given the material and informational resources available to them. These preferences are strongly influenced by genetic predispositions, but they are affected by group culture. Culture for humans is not merely a set of learned techniques, but also a set of moral values that are internalized by group members (Parsons Reference Parsons1967). The capacity to internalize values through socialization is an evolved adaptation of humans (Durkheim Reference Durkheim1902/1967; Gintis Reference Gintis2003; Simon Reference Simon1990), and accounts both for cultural diversity across societies and (limited) cultural stability across generations.
Human social cooperation is governed not by genes alone, but by social norms that are widely embraced by social actors, and act as moral values present in individual preference functions. We term these other-regarding preferences (Gintis Reference Gintis2009). Individuals incorporate moral values in their actions by trading off the costly attainment of other-regarding goals against self-regarding goals. The ubiquity of altruistic cooperation and altruistic punishment around the world suggests that these values are strong genetic predispositions, although the evidence indicates that their expression is strongly modulated by local cultural values (Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr and Gintis2004; Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath, Alvard, Barr, Ensminger, Henrich, Hill, Gil-White, Gurven, Marlowe, Patton and Tracer2005; Reference Henrich, McElreath, Barr, Ensminger, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2006). There are several plausible models of the evolution of these predispositions (Bowles & Gintis Reference Bowles and Gintis2011; Boyd et al. Reference Boyd, Gintis and Bowles2010; Gintis Reference Gintis2000). There are also plausible evolutionary models of the internalization of norms, the mechanism by which moral values become represented in the individual's preference ordering (Boehm Reference Boehm2011; Bowles & Gintis Reference Bowles and Gintis2011; Gintis Reference Gintis2003).
In this alternative framework, revenge and forgiveness can be self-regarding acts aimed at deterring malefactors and warning others of the cost of aggression. But revenge can also be an other-regarding act carried out to redress wrongs on a purely moral level. This explains why individuals punish not only those who hurt themselves, their families, and their allies, but more generally those who violate social norms. It also explains why individuals will seek vengeance against aggressors even when there is no possible deterrent effect.