In the target article, McCullough et al. propose a functionalist model of revenge in which retaliatory aggression is neither irrational nor a sign of pathological dysfunction. Rather, such behavior is driven by the implicit calculations of a revenge system that seeks to deter harmful future acts against oneself, kin, and allies. This revenge system operates in tandem with a forgiveness system designed to reassess whether a person is worthy of inclusion in one's ingroups.
The authors' analysis of revenge is in harmony with the person-centered approach to moral judgment (Pizarro & Tannenbaum Reference Pizarro, Tannenbaum, Shaver and Mikulincer2011; Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011; Uhlmann & Zhu, under review; Uhlmann et al., under review a; under review b), which emphasizes the adaptive manner in which social perceivers derive character information from moral acts. In the same spirit as McCullough et al., we argue that what at first appear to be irrational decision-making biases often “make sense” when one considers the adaptive goals individuals must meet as they navigate their social environments. Unlike McCullough et al., we emphasize that people often wish to avenge moral transgressions that not only do not harm them or their kin, but in some cases cause no material harm at all.
In addition to assessing the permissibility of acts, people use behaviors to draw inferences about the moral character of the agents who carry them out. Whereas moral judgments of acts are comparatively more likely to center on the tangible harm caused, judgments of persons focus on whether the behavior signals the presence or absence of positive moral traits (Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011; Uhlmann et al., under review a; under review b). Relatively harmless acts can therefore provoke outrage when they suggest severe deficits in moral character.
Some behaviors are more informative than others regarding an agent's personal character (Nelson Reference Nelson2005; Nelson et al. Reference Nelson, McKenzie, Cottrell and Sejnowski2010; Reeder & Brewer Reference Reeder and Brewer1979). Drawing inferences about character based on such signals is critical to resolving the collective action problems central to McCullough et al.'s analysis of revenge. Because moral traits predict whether a person will cooperate with us or betray us, character-relevant information becomes extremely valuable when navigating social environments. Even acts that are not especially harmful can speak strongly to personal character and therefore prove useful for anticipating more consequential future acts. For the same reason, relatively harmless acts can elicit a desire to castigate the transgressor and exclude him or her from social ingroups.
Evidence for the person-centered approach to moral judgment is provided by act–person dissociations in which an act is seen as less immoral than a comparison act, yet as a clearer indicator of negative moral traits. For example, although beating one's girlfriend is viewed as more morally blameworthy than beating her cat, the latter act is seen as indicating a more coldhearted and sadistic person (Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011). This might seem like a bias in moral judgment until one recalls that animal cruelty predicts antisocial behaviors and an erosion of normal empathic responses (Becker et al. Reference Becker, Steuwig, Herrera and McCloskey2004).
Additional studies demonstrate act-person dissociations in the context of truly harmless acts. For instance, although the use of a racial slur (in private, and with no one overhearing) was seen as a less blameworthy act than physical assault, use of a slur was perceived as providing more negative information about the person's character (Uhlmann et al., under review a). Consistent with the idea that person-centered judgments serve the function of determining who to include in one's social ingroups, participants were more willing to be friends with the target who had been physically aggressive than with the bigot.
Negative gut reactions to harmless-but-disgusting transgressions are frequently cited as a case of moral bias. Indeed, participants can find themselves dumbfounded when asked to justify why they feel eating a dead dog and having sex with a chicken carcass are morally wrong (Haidt Reference Haidt2001; Haidt et al. Reference Haidt, Koller and Dias1993). Part of the reason for strong intuitive responses to such transgressions is that they provide more diagnostic information about the personal character of the agent than do most harm violations. Participants rated eating a dog to be less immoral than stealing a steak, yet more informative of poor moral character (Uhlmann & Zhu, under review).
People are left at a loss to justify their intuitions regarding harmless-but-disgusting acts because they cannot be defended using rational criteria such as the degree of harm caused. In contrast, because of their high informational value regarding underlying traits, there is a clear rational basis for drawing strong character inferences from such behaviors. Although participants were morally dumbfounded when asked whether sex with a chicken was an immoral act, they were not at all dumbfounded when asked whether a person who engaged in sex with a chicken had negative moral traits (Uhlmann & Zhu, under review). This lack of dumbfounding regarding person judgments was driven by the behavior's high informational value.
As further evidence that person-centered judgments are not subjectively irrational (Pizarro & Uhlmann Reference Pizarro and Uhlmann2005), act–person dissociations are observed under conditions of both joint and separate evaluation (Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011; Uhlmann & Zhu, under review). Joint evaluation, in which social targets are evaluated side-by side, promotes logical comparisons and attenuates many decision-making biases (Gaertner & Dovidio Reference Gaertner, Dovidio, Dovidio and Gaertner1986; Hsee et al. Reference Hsee, Loewenstein, Blount and Bazerman1999). This suggests that participants do not view their tendency to judge actions and persons differently as irrational. If they did, they would correct their judgments under conditions of joint evaluation so as not to show any “bias.”
In closing, I agree with McCullough et al. that vengeance against wrongdoers – like moral outrage more generally – has a logic to it. Indeed, their functional analysis of revenge converges with our findings in highlighting the adaptive, reputation-based, and person-centered nature of moral cognition. The desire to exact revenge – and willingness to ultimately forgive – may often be less about the transgression itself than what it says about the agent's moral worth as a person.
In the target article, McCullough et al. propose a functionalist model of revenge in which retaliatory aggression is neither irrational nor a sign of pathological dysfunction. Rather, such behavior is driven by the implicit calculations of a revenge system that seeks to deter harmful future acts against oneself, kin, and allies. This revenge system operates in tandem with a forgiveness system designed to reassess whether a person is worthy of inclusion in one's ingroups.
The authors' analysis of revenge is in harmony with the person-centered approach to moral judgment (Pizarro & Tannenbaum Reference Pizarro, Tannenbaum, Shaver and Mikulincer2011; Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011; Uhlmann & Zhu, under review; Uhlmann et al., under review a; under review b), which emphasizes the adaptive manner in which social perceivers derive character information from moral acts. In the same spirit as McCullough et al., we argue that what at first appear to be irrational decision-making biases often “make sense” when one considers the adaptive goals individuals must meet as they navigate their social environments. Unlike McCullough et al., we emphasize that people often wish to avenge moral transgressions that not only do not harm them or their kin, but in some cases cause no material harm at all.
In addition to assessing the permissibility of acts, people use behaviors to draw inferences about the moral character of the agents who carry them out. Whereas moral judgments of acts are comparatively more likely to center on the tangible harm caused, judgments of persons focus on whether the behavior signals the presence or absence of positive moral traits (Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011; Uhlmann et al., under review a; under review b). Relatively harmless acts can therefore provoke outrage when they suggest severe deficits in moral character.
Some behaviors are more informative than others regarding an agent's personal character (Nelson Reference Nelson2005; Nelson et al. Reference Nelson, McKenzie, Cottrell and Sejnowski2010; Reeder & Brewer Reference Reeder and Brewer1979). Drawing inferences about character based on such signals is critical to resolving the collective action problems central to McCullough et al.'s analysis of revenge. Because moral traits predict whether a person will cooperate with us or betray us, character-relevant information becomes extremely valuable when navigating social environments. Even acts that are not especially harmful can speak strongly to personal character and therefore prove useful for anticipating more consequential future acts. For the same reason, relatively harmless acts can elicit a desire to castigate the transgressor and exclude him or her from social ingroups.
Evidence for the person-centered approach to moral judgment is provided by act–person dissociations in which an act is seen as less immoral than a comparison act, yet as a clearer indicator of negative moral traits. For example, although beating one's girlfriend is viewed as more morally blameworthy than beating her cat, the latter act is seen as indicating a more coldhearted and sadistic person (Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011). This might seem like a bias in moral judgment until one recalls that animal cruelty predicts antisocial behaviors and an erosion of normal empathic responses (Becker et al. Reference Becker, Steuwig, Herrera and McCloskey2004).
Additional studies demonstrate act-person dissociations in the context of truly harmless acts. For instance, although the use of a racial slur (in private, and with no one overhearing) was seen as a less blameworthy act than physical assault, use of a slur was perceived as providing more negative information about the person's character (Uhlmann et al., under review a). Consistent with the idea that person-centered judgments serve the function of determining who to include in one's social ingroups, participants were more willing to be friends with the target who had been physically aggressive than with the bigot.
Negative gut reactions to harmless-but-disgusting transgressions are frequently cited as a case of moral bias. Indeed, participants can find themselves dumbfounded when asked to justify why they feel eating a dead dog and having sex with a chicken carcass are morally wrong (Haidt Reference Haidt2001; Haidt et al. Reference Haidt, Koller and Dias1993). Part of the reason for strong intuitive responses to such transgressions is that they provide more diagnostic information about the personal character of the agent than do most harm violations. Participants rated eating a dog to be less immoral than stealing a steak, yet more informative of poor moral character (Uhlmann & Zhu, under review).
People are left at a loss to justify their intuitions regarding harmless-but-disgusting acts because they cannot be defended using rational criteria such as the degree of harm caused. In contrast, because of their high informational value regarding underlying traits, there is a clear rational basis for drawing strong character inferences from such behaviors. Although participants were morally dumbfounded when asked whether sex with a chicken was an immoral act, they were not at all dumbfounded when asked whether a person who engaged in sex with a chicken had negative moral traits (Uhlmann & Zhu, under review). This lack of dumbfounding regarding person judgments was driven by the behavior's high informational value.
As further evidence that person-centered judgments are not subjectively irrational (Pizarro & Uhlmann Reference Pizarro and Uhlmann2005), act–person dissociations are observed under conditions of both joint and separate evaluation (Tannenbaum et al. Reference Tannenbaum, Uhlmann and Diermeier2011; Uhlmann & Zhu, under review). Joint evaluation, in which social targets are evaluated side-by side, promotes logical comparisons and attenuates many decision-making biases (Gaertner & Dovidio Reference Gaertner, Dovidio, Dovidio and Gaertner1986; Hsee et al. Reference Hsee, Loewenstein, Blount and Bazerman1999). This suggests that participants do not view their tendency to judge actions and persons differently as irrational. If they did, they would correct their judgments under conditions of joint evaluation so as not to show any “bias.”
In closing, I agree with McCullough et al. that vengeance against wrongdoers – like moral outrage more generally – has a logic to it. Indeed, their functional analysis of revenge converges with our findings in highlighting the adaptive, reputation-based, and person-centered nature of moral cognition. The desire to exact revenge – and willingness to ultimately forgive – may often be less about the transgression itself than what it says about the agent's moral worth as a person.