Download Understanding Revenge and Forgiveness: Functional Adaptations in Human Behavior and more Study notes Social Psychology in PDF only on Docsity!
Evolved Mechanisms for Revenge and Forgiveness
Michael E. McCullough Robert Kurzban University of Miami University of Pennsylvania
Benjamin A. Tabak University of Miami
In P. R. Shaver and M. Mikulincer (eds.) (in press). Understanding and reducing aggression,violence, and their consequences. Washington, DC: American Psychological Association.
In this chapter, we describe our efforts to understand the functions of the cognitive systems that underlie humans’ capacities for revenge and forgiveness. A better understanding of these concepts is not simply scientifically interesting; it is socially important as well. In developed nations, the desire for revenge is cited as a causal factor in as many as 20% of homicides (Kubrin & Weitzer, 2003). Roughly 20% of the perpetrators of violent assault and criminal property damage in the United Kingdom cite the desire for revenge as a motive for their behavior (UK Crime and Justice Survey, 2003), and approximately 61% of school shootings in the U. S. between 1974 and June 2000 were vengeance-motivated (Vossekuil, Fein, Reddy,
Borum, & Modzeleski, 2002). The desire for revenge also motivates people to enlist in terrorist organizations (Speckhard & Ahkmedova, 2006). Perhaps because revenge is so closely linked to aggression and violence, it has been fashionable in Western thought since the Stoic (and, later, Christian) philosophers to view revenge as immoral, irrational, or both (Murphy, 2003). As if seeking to re-state this dim view of revenge in more modern, therapeutic terms, social scientists in the past century have also promulgated the idea that the desire for revenge is indicative of psychological dysfunction (Murphy, 2003). Two decades ago, clinical psychologists and therapists began to tacitly endorse the ―revenge as disease‖ conceit as they began the psychological study of forgiveness. The earliest published psychological articles on forgiveness were descriptions of forgiveness-based therapeutic techniques for helping people recover from the effects of traumatic experiences (and residual vengeful feelings) that impaired their psychological and relational functioning (e.g., Hope, 1987). The ―revenge as disease‖ conceit had a predictable effect on how forgiveness came to be understood: If the desire for revenge is a disease, then perhaps forgiveness is the cure. Indeed, many of the earliest empirical studies on forgiveness were related to the use of interventions for promoting forgiveness in therapeutic settings (DiBlasio & Benda, 1991; Hebl & Enright, 1993). These treatments do promote forgiveness—and reduce psychological symptoms of anxiety and depression and boost self-esteem (Lundahl, Stevenson, & Roberts, 2008)—but such facts do not prove that forgiveness is a ―cure‖ for revenge. Recent research casts considerable doubt on the ―disease and cure‖ conceit for conceptualizing the human capacities for revenge and forgiveness. In this chapter, we propose that revenge and forgiveness are distinct behavioral adaptations that solve specific adaptive
associated persons) as a retaliatory measure ‖ (p. 2). Elster (1990) likewise defined revenge as ―the attempt, at some cost or risk to oneself, to impose suffering upon those who have made one suffer, because they have made one suffer‖ (p. 862). Uniacke (2000) also claims that ―revenge is personal and non-instrumental: with revenge we seek to make people suffer because they have made us suffer, not because their actions or values require us to bring them down‖ (p. 62). These definitions, because they are proximate and do not commit to any function, make no obvious predictions about the design features of the psychology of revenge. ―Enjoyment‖ and other proximate explanations (see Govier’s definition) leave a promissory note for an ultimate explanation that must be paid. Why should revenge produce pleasure? For no organism except humans would we accept that an explanation for a behavior is that it brings enjoyment. In short, functional thinking about cognition and behavior reminds us that there is no free lunch. Why would a species such as Homo sapiens engage in costly behavior such as revenge unless it is associated with a benefit in the currency of fitness, or is an incidental by-product of a structure that does yield fitness payoffs (Andrews, Gangestad, & Matthews, 2002)? What could maintain revenge in humans’ behavioral repertoire? The definitions cited above also introduce problems related to intentionality. What does it mean that revenge involves an attempt to impose retaliatory harm on an aggressor? Does the word attempt imply a conscious and deliberate effort to make another individual suffer? Is consciousness necessary? Is deliberation necessary? Or can this striving to harm one’s provoker be automatic and/or unconscious? And is this distinction critical in any case? We think a functional definition of revenge can clarify some of these points. Biologists regularly define behavior functionally, as when Maynard Smith and Harper (2003) defined a signal as ―any act or structure which alters the behaviour of other organisms, which evolved
because of that effect, and which is effective because the receiver’s response has also evolved‖ (p. 3). By designating a function, it becomes possible to search for evidence of the features – behavioral or physiological – that contribute to accomplishing the putative function. In similar fashion, we would define revenge functionally as behavior resulting from a mechanism designed to deter the imposition of costs on (or the withholding of benefits from) oneself or one’s allies by the imposition of costs following a target’s imposition of costs (or withholding of benefits), where costs and benefits are defined in terms of their effects on lifetime reproductive fitness. That is, revenge is a deterrence system designed to change others’ incentives regarding their behavioral options to protect oneself and one’s kin or allies (see Shaver, Segev, & Mikulincer, Chapter 2, this volume, for a similar functional analysis of what they call the power behavioral system). By imposing costs after harm (or the absence of benefits), revenge signals that subsequent acts will be subject to the same contingent response, altering others’ incentives. Our hypothesis is that humans have developed specific adaptations designed for revenge. This functional definition has several important features. First, it replaces considerations of intentionality (e.g., whether the organism is deliberately or consciously attempting to do something) with considerations of design (e.g., what the system that motivates revenge was designed to do). Moreover, the concept of design makes powerful empirical commitments; adaptation is a strong claim (Williams, 1966), and to the extent the psychological mechanisms do not show features that support a deterrence function, the hypothesis that humans have a psychology of revenge is undermined. We review some relevant evidence below. This definition includes as cases of revenge all retaliatory impositions of costs that are caused by a mechanism designed for this purpose—even acts that are not based on deliberation
retaliation will repeat their injurious actions against the retaliator in the future. Consistent with Clutton-Brock and Parker’s analysis, we hypothesize that selection gave rise to such a system in humans. In this sense, the adaptive consequences of revenge come not from what revenge causes per se, but from what it prevents. For illustrative purposes here, we will distinguish among three types of deterrence. The first two—direct and indirect deterrence of aggression—involve preventing the imposition of costs. The third involves preventing the withholding of benefits. Direct deterrence. By direct deterrence, we mean that revenge discourages aggressors from harming the avenger a second time. The logic of direct deterrence is straightforward: If a potential aggressor has a decision in which he or she can take an action that imposes costs on a potential victim to acquire some benefit, then the potential victim is better off if he or she can change the potential aggressor’s incentives so that the expected value of the cost-imposing action on the potential victim is negative. Revenge can accomplish this transformation of expected value by conveying to an aggressor that the retaliatory infliction of fitness costs will exceed the potential benefits to be gained by aggressing against the potential victim a second time (Sell, Chapter 3, this volume). Nevertheless, direct deterrence gives rise to strategic complications. For example, although revenge at time 1 might predict revenge at time 2, there is nothing that forces this to be true. An organism could be, for example, intermittently vengeful. This leads to well- known problems of signaling that one’s vengeful dispositions are stable over time (Frank, 1988). Experimental evidence in support of revenge’s effectiveness as a direct deterrent comes from experiments involving economic games such as the sequential and iterated prisoner’s dilemma (Axelrod, 1984). In the sequential prisoner’s dilemma game, there is one round of play, but the second mover chooses only after seeing the first player’s choice. In such games, the second player is much more likely to cooperate after a cooperative move than after a defecting
move. More relevant to our present point, defection is almost always met with retaliatory defection (Clark & Sefton, 2001; see Table 6), an observation that holds not only in the United States (Hayashi, Ostrom, Walker, & Yamagishi, 1999). However, because non-cooperation and punishment are the same in the prisoner’s dilemma, such findings must be interpreted with care. In the iterated, prisoner’s dilemma, subjects play multiple rounds of the game with either the same partner or different ones. For the present purpose, key issues are whether people respond to defection with defection—moves plausibly interpretable as revenge—and whether such moves elicit subsequent cooperation from one’s partner. Experiments using large numbers of trials in prisoner’s dilemma games suggest that people do respond to defection with defection (Bixenstine & Wilson, 1963) though the details vary across studies (Rapoport & Chammah, 1965). Reciprocal strategies such as tit-for-tat or variants on it tend to elicit cooperation from experimental subjects (e.g., Wilson, 1971), hinting at their effectiveness in deterring defection. Moreover, in an analysis of data from 5 different laboratory studies of negotiations in dyads in which partners played 250 consecutive trials in which they could either punish, reward, or withhold reward (and punishment) from each other, Molm (1997) found that the frequency with which retaliatory punishment was used (i.e., the infliction of punishment after one’s negotiation partner previously punished the actor) was positively associated with the frequency with which partners rewarded each other. Likewise, the use of punishment following non-reward (i.e., the withholding of benefits) was associated with higher rates of rewarding. These findings suggest that retaliatory infliction of punishments in response to punishments and the withholding of rewards create a relational climate in which the exchange of reward is more frequent. In contrast, Molm reported that the frequency with which dyads punished non-contingently (that is, independently of whether the punishment was a retaliatory response to punishment or the
from harming the interests of their opponents when they know that the opponents have a strong ability to retaliate (Lawler, Ford, & Blegen, 1988). The lessons of empirical studies on the direct deterrent effects of punishment are not always straightforward, however. For example, Fehr and List (2004) used the trust game, a two- step dyadic game in which an ―investor‖ first entrusts a sum of money to a second person called the ―trustee,‖ which is then multiplied by some constant (often tripled) by the researchers. In the second step, the trustee is given the opportunity to return some amount of money to the investor. Fehr and List permitted investors to indicate a minimum amount of money they required from their trustees in return. If trustees failed to return that minimum amount, that amount was automatically deducted from their payoffs. Return transfers were highest when the punishment option was available but left unused by investors. Nonetheless, the majority of investors used the punishment option. Houser et al. (2008), using a similar design, found that the threat of punishment reduced the fraction of money trustees returned to investors even if the threat of punishment was applied as a result of a random process rather than a decision on the part of investor. Likewise, revenge in some experimental settings increases, rather than deters, non- cooperation. Dreber, Rand, Fudenberg, and Nowak (2008) found that using punishment against non-cooperators reduced players’ gains in an iterated prisoner’s dilemma, possibly because punishment in this experiment imposed very large costs: The cost of punishment and the size of the damage it inflicts clearly influence revenge’s deterrent effects. Deterring third parties. Revenge may also have been selected for its efficacy in deterring would-be aggressors by virtue of its ability to signal an avenger’s aggressive potential. Reputation is important for understanding how third-party deterrence might work. Ancestral humans lived in small, close-knit groups (Boehm, 2008) without the benefit of institutions for
protecting individual rights, so a readiness to retaliate against interpersonal harms might have been an important component of people’s social reputations. Researchers have documented the importance of defense of honor—which is, more or less, the perceived ability to defend one’s interests with violent force when necessary—and the revenge that it stimulates, as a major cause of violence among individuals from many societies (for review see McCullough, 2008, and Cohen, Chapter 7, this volume). Consistent with the idea that revenge is enacted partly out of reputational concerns, laboratory studies show that the psychological mechanisms that cause revenge are sensitive to the presence of third parties. Victims retaliate more strongly against their provokers when an audience has witnessed the provocation—especially if the audience communicates to the victim that he or she looks weak because of the harm suffered, or if the victim knows that the audience is aware that he or she has suffered particularly unjust treatment (Brown, 1968; Kim, Smith, & Brigham, 1998). Also, when two men get into an argument, the mere presence of a third person doubles the likelihood that the argument will escalate to a violent encounter (Felson, 1982). Deterring the withholding of benefits. Finally, we think revenge might have been selected for changing others’ behavior to increase the delivery of benefits (as opposed to only reducing harm). Public goods games are useful for illustrating how revenge can deter the withholding of benefits. In these games, a few (often 4 to 6) participants receive an initial endowment of money and are instructed to choose how to split that endowment between two different pools. One pool is private, and participants simply keep any money they place in it. The other pool is shared— money placed into this pool is multiplied by some amount (greater than 1) and the resultant total is subsequently divided evenly among all the players in the group. Money maximizers keep everything in their private pools; aggregate group wealth is maximized when everyone
participants could punish members of their own groups or members of other people’s groups. In the key treatment, the ―one-way TPP‖ (third party punishment) condition, almost no one punished. The fact that one-way third-party punishment was so minimal when directed toward non-cooperators in groups to which the subjects themselves did not belong strongly suggests that absent the possibility of revenge, people tend not to punish. The Evolution of Forgiveness Forgiveness: A Functional Definition Natural selection gave rise to the psychological system that produces revenge, we posit, by virtue of the fitness payoffs associated with direct deterrence, third party deterrence, and, possibly, coerced delivery of benefits. However, avengers trade off the potential benefits lost by virtue of any damage that revenge does to relations with the harm doer, and they incur the (probabilistic) costs associated with any counter-revenge that might ensue as the result of their revenge. We therefore presume that the revenge system is designed to adjust its operation in response to the potential costs and benefits associated with revenge in any particular instance. When the costs of revenge are too high relative to its expected deterrence benefits, an organism might pursue an alternative course of behavior—forgiveness being one of the more likely ones. Over the past decade, the first author’s research group has defined forgiveness as a set of motivational changes whereby an organism becomes (a) decreasingly motivated to retaliate against an offending relationship partner; (b) decreasingly motivated to avoid the offender; and (c) increasingly motivated by goodwill for, and a desire to reconcile with, the offender, despite the offender’s harmful actions (McCullough, 2008; McCullough, Worthington, & Rachal, 1997). Here, we refine this definition by adding a functional addendum: that a ―forgiveness system‖
produces these motivational changes because of their efficacy during evolution in promoting the restoration of beneficial relationships in the aftermath of interpersonal harms. This newly ―functionalized‖ definition of forgiveness permits all of the important conceptual distinctions that other theorists (e.g., Enright & Coyle, 1998; Worthington, 2005) consider important (e.g., that forgiveness is different from forgetting an offense, denying its reality, condoning it, or attempting to minimize its significance), and it enables a tighter conceptual link between forgiveness and reconciliation than has previously been recognized. Many theorists have been careful to distinguish forgiveness from reconciliation, with the latter concept indicating a restoration of the relationship between offender and victim (Worthington, 2005). In light of the functional definition of forgiveness that we propose, it might be possible to forgive a harm doer (i.e., experience motivational changes by which one becomes less vengeful, less avoidant, and more benevolently disposed toward him or her) without reconciling (i.e., restoring the relationship). Nevertheless, we reason that modern humans are capable of forgiving because ancestral humans who deployed this strategy enjoyed the fitness benefits that came from restoring potentially valuable relationships. Forgiveness, like revenge, involves costs. Forgiveness prepares a victim to re-enter constructive relations with a harm doer based on the prospect of capturing benefits from that relationship, but the motivational change that the forgiveness system creates entails foregoing revenge and its deterrent effects. Forgiveness, therefore, involves a loss of gains from changing the harm doer’s incentives—potentially inviting recidivism (e.g., see Gordon, Burton, & Porter,
- and attacks from those who see the opportunity to exploit the forgiver. In short, forgiveness undermines the function of the revenge system by undermining deterrence. Thus, the forgiveness system, like the revenge system, should be sensitive to costs and benefits, and these
―echo effect‖ (Axelrod, 1984), whereby individuals who are cooperatively disposed nevertheless become locked in costly cycles of retaliation when initial unintended defections occur due to noise. Second, individuals who can forgive their reciprocal altruism partners following defections would have been able to preserve their access to benefits that their partners would have been able to provide them, and would have spared themselves the costs associated with establishing new relationships with new individuals whose social dispositions would be unknown (Hruschka & Henrich, 2006). On average, it may simply be less costly to forgive some number of defections from a well-established relationship partner than to retaliate, or to withdraw from the relationship entirely, following an isolated defection. Indeed, in computer simulations of the evolution of reciprocal altruism—especially when the possibility of noise is assumed—evolutionarily stable strategies tend to be more forgiving than ―tit for tat,‖ which responds to defection with defection and to cooperation with cooperation (Frean, 1994; Hauert & Schuster, 1998; Wu & Axelrod, 1995). This is especially true when one models reciprocal altruism as occurring largely among small networks of individuals (e.g., friendship groups or individuals within small living groups) who focus their cooperative efforts on other individuals within the network, and limit their cooperation with individuals outside of the network (Levine & Kurzban, 2006): Under such circumstances, agents are expected to forgive up to 80% of other network members’ defections (Hruschka & Henrich, 2006). Other types of relationships generate still other types of benefits that redound to lifetime reproductive fitness. The benefits that might accrue from forgiving a mate are different from the benefits that might accrue by forgiving a friend, which in turn are different from the benefits that a forgiver might receive by forgiving an ally. Because the fitness-enhancing properties come in different currencies, the psychological system that produces forgiveness is likely set up to
identify the types of benefits that a particular type of relationship is likely to confer (and to weight them appropriately with respect to the probability of capturing those benefits, the time horizon at which they will be realized, etc.), and weigh those benefits against the deterrent value of revenge, which the organism would trade off if it chose to forgive instead of seeking revenge. Choosing Forgiveness over Vengeance: Proximate Causation If, as we hypothesize, the forgiveness system is sensitive to tradeoffs associated with sacrificing the deterrence benefits of revenge for the relationship-restoration benefits of forgiveness, then the forgiveness system should be acutely sensitive to variables that influence the value of each option. These variables include, but are not necessarily limited to characteristics of the offender, the transgression itself, and cues that predict the probabilities of future attacks and/or the potential future value of the restored relationship. Put simply, we predict that the motive to forgive will be generated by systems designed to compute and compare the cost of forgone revenge and the benefits that are expected to accrue from a restored relationship. Va lue of deterrence. The value of revenge diminishes to the extent that it does not change behavior that would otherwise occur. In the limiting case, suppose that after an offense, the transgressor could persuasively signal that he or she would never–or could never–again inflict costs. In such a case, revenge would yield no benefit (except through third-party deterrence). Information relevant to inferring intent can come from various sources. For instance, a transgressor’s apology, expression of sympathy for a victim’s suffering, and declaration of his or her intention to behave better in the future could indicate a low likelihood of trying to harm the victim in the future (McCullough et al., 1997). Verbal declarations such as these are susceptible to strategic manipulation, of course. Nonverbal displays such as blushing, which facilitate forgiveness after some transgressions (de Jong, Peters, & de Cremer, 2003), also contain
we expect that cues of kinship will facilitate forgiveness, just as they facilitate the restraint of vengeance (Lieberman & Linke, 2007). In similar fashion, those with whom one has a close history of association, shared interests, and many opportunities for mutually beneficial transactions are good candidates for forgiveness because of the possibility of continued gains. Indeed, priming people with the names of close others (e.g., via subliminal presentation) leads to increased judgments of forgiveness, increased accessibility of the concept of forgiveness, and reduced deliberation about whether forgiveness is an appropriate course of action (Karremans & Aarts, 2007). Karremans and Aarts’s results complement those from several previous studies showing that people are more inclined to forgive individuals with whom they feel close and committed (Finkel, Rusbult, Kumashiro, & Hannon, 2002; McCullough et al., 1998; Slotter & Finkel, Chapter 2, this volume). The reason for these associations (and experimentally induced) effects of closeness or commitment on forgiveness, we would argue, is that relationship closeness and/or commitment acts as a cue of relationship value in many types of interpersonal relations. We think the importance of relationship value can also explain why people tend to want some form of compensation prior to forgiving (Boehm, 1987; Bottom, Gibson, Daniels, & Murnighan, 2002): Compensation may serve as a cue of (among other things) a transgressor’s ability or willingness to be a valuable relationship partner in the future. In computations of expected value, empathy may play a special role. Empathy for transgressors, which is a sympathy- or pity-like response to the plight of another person, appears to be a reliable facilitator of forgiveness (McCullough, et al., 1997; Zechmeister & Romero, 2002)—perhaps as a result of empathy’s long phylogenetic history as a motivator of care for valuable relationship partners (Preston & de Waal, 2002). Whether the empathy-forgiveness link
is part of the design of the forgiveness system, or merely an incidental effect that empathy can exert within the existing forgiveness system, however, is currently difficult to know. Summary The desire for revenge and the ability to forgive seem to be universal psychological endowments of humans (Boehm, 2008; Daly & Wilson, 1988; McCullough, 2008). Species- typical traits call out for explanations in terms of the mind’s evolved mental structures—either as direct products or as by-products of what those structures were designed to do (Andrews et al., 2002). Here, we have taken an adaptationist stance and posited that revenge and forgiveness result from computational mechanisms designed to produce them. Once one has moved into a functional framework, we think it becomes easier to see what should qualify as revenge (or forgiveness) and what should not, and what the important questions are if one wants to understand what revenge and forgiveness are really all about. By outlining the selection pressures that likely gave rise to humans’ penchants for revenge and forgiveness, we have also tried here to identify the types of information that the structures that produce revenge and forgiveness should be designed to process. We hope that introducing this sort of thinking into mainstream research on revenge and forgiveness can help investigators prioritize their research efforts in the future.
References Anderson, C., & Putterman, L. (2006). Do non-strategic sanctions obey the law of demand? The demand for punishment in the voluntary contribution mechanism. Games and Economic Behavior, 54 , 1-24.
