As Guala rightly notes, there is very little evidence that punishment plays a role in the stabilization of cooperation in small-scale societies. On the other hand, as he also notes, it is difficult to totally rule out the strong view of punishment as it is complicated to precisely assess the costs of punishment in the field (Are there really no costs in punishing others? Aren't there many hidden benefits for the individual who punish? etc.). There is, however, another way to disentangle the two views of punishment, namely, the forms that punishments take. Indeed, the two theories – the weak and the strong – make different predictions regarding the logic of punishment.
Group selection theory holds that punishment aims to promote the good of the group by sustaining cooperation and preventing cheating (Boyd et al. Reference Boyd, Gintis, Bowles and Richerson2003; Fehr & Gächter Reference Fehr and Gächter2002; Henrich & Boyd Reference Henrich and Boyd2001). This implies that punishment should be calibrated to deter crimes and render them non-advantageous. Here, group selection parallels the utilitarian doctrine of punishment, which contends that punishment should be used to deter crimes and maximize the good of society (Polinsky & Shavell Reference Polinsky and Shavell2000; Posner Reference Posner1983). The utilitarian theory of punishment holds, for instance, that the detection rate of a given crime and the publicity associated with a given conviction are relevant factors in assigning punishments. If a crime is difficult to detect, the punishment for that crime ought to be made more severe in order to counterbalance the temptation created by the low risk of getting caught. Likewise, if a conviction is likely to get a lot of publicity, a law enforcement system interested in deterrence should take advantage of this circumstance by “making an example” of the convict with a particularly severe punishment, thus getting a maximum of deterrence for its punishment.
By contrast, individual selection predicts a “restorative” or “retributive” logic for punishment (Baumard Reference Baumard2011). Restorative logic holds that punishment aims to restore justice between the criminal and the victim – either by harming the criminal or by compensating the victim. In intuitive terms, people are punished because they “deserve” to be punished, and not because punishing them would be useful for the society at large.
This restorative logic is a direct consequence of the way cooperation has evolved among humans (Baumard Reference Baumard2010a; Trivers Reference Trivers1971). Indeed, human beings belong to a highly cooperative species and get most of their resources from collective actions, solidarity, exchanges, and so forth. (Gurven Reference Gurven2004; Hill & Kaplan Reference Hill and Kaplan1999). In the ancestral environment, individuals were in competition to be recruited for the most fruitful ventures, and it was vital to share the benefits of cooperation in a mutually advantageous manner. If individuals took a bigger share of the benefits, their partners would leave them for more interesting partners. If they took a smaller share, they would be exploited by their partners who would receive more than what they had contributed to produce. This competition to attract cooperative partners is thus likely to have led to selection for a “sense of fairness,” a cognitive device that motivates individuals to share the costs and benefits of social interaction in an impartial way (André & Baumard Reference André and Baumard2011). If cooperation is based on fairness, then crimes create an unfair relationship between the criminal and her victim, and people have the intuition that the criminal ought to compensate the victim or to be punished in order to restore justice.
It is worth mentioning that this theory does not mean that punishment should be absent in human societies. As Guala notes, modern societies have found many institutional ways to reduce the costs of punishments. Although these institutions are absent in smaller societies, justice can still be restored by individuals seeking to retaliate. Retaliation is indeed advantageous from an individual perspective and can indeed be found in many nonhuman species (Clutton-Brock & Parker Reference Clutton-Brock and Parker1995). As Evans-Pritchard noted, in societies where there is no penal system, “self-help, with some backing of public opinion, is the main sanction” (Evans-Pritchard Reference Evans-Pritchard1940/1969, p. 169).
In this kind of situations, selfish and moral motives converge: The victim (or his allies) attacks the criminal to signal his strength and gains a reputation as someone who cannot be attacked without risk; and by doing so, he also punishes the wrongdoer by allowing justice to be done. In line with this idea, people in small-scale societies distinguish between legitimate (and proportionate) retaliation and illegitimate (and disproportionate) retaliation (von Fürer-Hameindorf Reference von Führer-Haimendorf1967; Miller Reference Miller1990). Retaliation is thus clearly limited by moral concerns: within the group, it has to be proportionate to the prejudice. As the Lex Talionis says, “an eye for an eye, a tooth for a tooth,” but no more.
Individual selection thus clearly predicts some kind of punishment, and, more importantly, it predicts that punishments should aim toward a specific goal (restoring fairness) that differs from the utilitarian goal predicted by group selection (preventing wrongdoing). Experimental studies, relying on a variety of methodologies, suggest that punishments fit individual selection more than group selection. Indeed, when people punish harmdoers, they generally respond to factors relevant to a retributive theory of punishment (magnitude of harm, moral intentions) and ignore factors relevant to the group selection theory (likelihood of detection, publicity, likelihood of repeat offending) (Baron et al. Reference Baron, Gowda and Kunreuther1993; Baron & Ritov Reference Baron and Ritov2008; Carlsmith et al. Reference Carlsmith, Darley and Robinson2002; Darley et al. Reference Darley, Carlsmith and Robinson2000; Glaeser & Sacerdote Reference Glaeser and Sacerdote2000; Sunstein et al. Reference Sunstein, Schkade and Kahneman2000).
In line with these results, field observations have extensively demonstrated that, in keeping with the prediction, the level of compensation in stateless societies is directly proportional to the prejudice inflicted to the victim: For example, the wrongdoer owes more to the victim if the wrongdoer has killed a family member or eloped with a wife than if he has stolen animals or destroyed crops (Hoebel Reference Hoebel1954; Howell Reference Howell1954; Malinowski Reference Malinowski1926). To conclude, punishment does not seem to be a group adaptation. It follows the logic of fairness rather than the interests of the group.
As Guala rightly notes, there is very little evidence that punishment plays a role in the stabilization of cooperation in small-scale societies. On the other hand, as he also notes, it is difficult to totally rule out the strong view of punishment as it is complicated to precisely assess the costs of punishment in the field (Are there really no costs in punishing others? Aren't there many hidden benefits for the individual who punish? etc.). There is, however, another way to disentangle the two views of punishment, namely, the forms that punishments take. Indeed, the two theories – the weak and the strong – make different predictions regarding the logic of punishment.
Group selection theory holds that punishment aims to promote the good of the group by sustaining cooperation and preventing cheating (Boyd et al. Reference Boyd, Gintis, Bowles and Richerson2003; Fehr & Gächter Reference Fehr and Gächter2002; Henrich & Boyd Reference Henrich and Boyd2001). This implies that punishment should be calibrated to deter crimes and render them non-advantageous. Here, group selection parallels the utilitarian doctrine of punishment, which contends that punishment should be used to deter crimes and maximize the good of society (Polinsky & Shavell Reference Polinsky and Shavell2000; Posner Reference Posner1983). The utilitarian theory of punishment holds, for instance, that the detection rate of a given crime and the publicity associated with a given conviction are relevant factors in assigning punishments. If a crime is difficult to detect, the punishment for that crime ought to be made more severe in order to counterbalance the temptation created by the low risk of getting caught. Likewise, if a conviction is likely to get a lot of publicity, a law enforcement system interested in deterrence should take advantage of this circumstance by “making an example” of the convict with a particularly severe punishment, thus getting a maximum of deterrence for its punishment.
By contrast, individual selection predicts a “restorative” or “retributive” logic for punishment (Baumard Reference Baumard2011). Restorative logic holds that punishment aims to restore justice between the criminal and the victim – either by harming the criminal or by compensating the victim. In intuitive terms, people are punished because they “deserve” to be punished, and not because punishing them would be useful for the society at large.
This restorative logic is a direct consequence of the way cooperation has evolved among humans (Baumard Reference Baumard2010a; Trivers Reference Trivers1971). Indeed, human beings belong to a highly cooperative species and get most of their resources from collective actions, solidarity, exchanges, and so forth. (Gurven Reference Gurven2004; Hill & Kaplan Reference Hill and Kaplan1999). In the ancestral environment, individuals were in competition to be recruited for the most fruitful ventures, and it was vital to share the benefits of cooperation in a mutually advantageous manner. If individuals took a bigger share of the benefits, their partners would leave them for more interesting partners. If they took a smaller share, they would be exploited by their partners who would receive more than what they had contributed to produce. This competition to attract cooperative partners is thus likely to have led to selection for a “sense of fairness,” a cognitive device that motivates individuals to share the costs and benefits of social interaction in an impartial way (André & Baumard Reference André and Baumard2011). If cooperation is based on fairness, then crimes create an unfair relationship between the criminal and her victim, and people have the intuition that the criminal ought to compensate the victim or to be punished in order to restore justice.
It is worth mentioning that this theory does not mean that punishment should be absent in human societies. As Guala notes, modern societies have found many institutional ways to reduce the costs of punishments. Although these institutions are absent in smaller societies, justice can still be restored by individuals seeking to retaliate. Retaliation is indeed advantageous from an individual perspective and can indeed be found in many nonhuman species (Clutton-Brock & Parker Reference Clutton-Brock and Parker1995). As Evans-Pritchard noted, in societies where there is no penal system, “self-help, with some backing of public opinion, is the main sanction” (Evans-Pritchard Reference Evans-Pritchard1940/1969, p. 169).
In this kind of situations, selfish and moral motives converge: The victim (or his allies) attacks the criminal to signal his strength and gains a reputation as someone who cannot be attacked without risk; and by doing so, he also punishes the wrongdoer by allowing justice to be done. In line with this idea, people in small-scale societies distinguish between legitimate (and proportionate) retaliation and illegitimate (and disproportionate) retaliation (von Fürer-Hameindorf Reference von Führer-Haimendorf1967; Miller Reference Miller1990). Retaliation is thus clearly limited by moral concerns: within the group, it has to be proportionate to the prejudice. As the Lex Talionis says, “an eye for an eye, a tooth for a tooth,” but no more.
Individual selection thus clearly predicts some kind of punishment, and, more importantly, it predicts that punishments should aim toward a specific goal (restoring fairness) that differs from the utilitarian goal predicted by group selection (preventing wrongdoing). Experimental studies, relying on a variety of methodologies, suggest that punishments fit individual selection more than group selection. Indeed, when people punish harmdoers, they generally respond to factors relevant to a retributive theory of punishment (magnitude of harm, moral intentions) and ignore factors relevant to the group selection theory (likelihood of detection, publicity, likelihood of repeat offending) (Baron et al. Reference Baron, Gowda and Kunreuther1993; Baron & Ritov Reference Baron and Ritov2008; Carlsmith et al. Reference Carlsmith, Darley and Robinson2002; Darley et al. Reference Darley, Carlsmith and Robinson2000; Glaeser & Sacerdote Reference Glaeser and Sacerdote2000; Sunstein et al. Reference Sunstein, Schkade and Kahneman2000).
In line with these results, field observations have extensively demonstrated that, in keeping with the prediction, the level of compensation in stateless societies is directly proportional to the prejudice inflicted to the victim: For example, the wrongdoer owes more to the victim if the wrongdoer has killed a family member or eloped with a wife than if he has stolen animals or destroyed crops (Hoebel Reference Hoebel1954; Howell Reference Howell1954; Malinowski Reference Malinowski1926). To conclude, punishment does not seem to be a group adaptation. It follows the logic of fairness rather than the interests of the group.