In the target article, Guala argues that although costly punishment occurs in lab settings, there is little evidence of costly punishment “in the wild.” Thus, he questions whether lab studies replicate the processes and conditions that support cooperation in the real world. We believe that Guala has defined “cost” too narrowly. Costs should include not only reduced material resources, but also decreased social status and psychological well-being. We argue that punishers often experience a number of social and psychological costs, and that such costs should be considered when searching for evidence of costly punishment “in the wild.”

Punishers can experience social costs (e.g., reduced status), especially when observers question whether the punishment they enacted was proportionate to the seriousness of the transgression (Trevino Reference Trevino1992). For example, when employees observe managers punishing fellow employees in their organization, they sometimes report becoming less trusting, less respecting, and more fearful of those managers (Atwater et al. Reference Atwater, Waldman, Carey and Cartier2001). Feelings of distrust likely translate into less commitment to the manager (see Kramer & Cook [Reference Kramer and Cook2004] for a review), which could lead to poorer outcomes for the punitive manager over time. Moreover, the fact that managers experience these social costs – even when punishing is somewhat legitimated by their social role – suggests that institutionalized punishment may not always mitigate the social costs of punishing. Research on the fundamental attribution error has demonstrated that people are prone to locating the cause of behavior within actors while failing to adequately account for the situational forces (e.g., roles) that may have caused the behavior (Ross Reference Ross and Berkowitz1977). As a result, people may still perceive punishers negatively (i.e., as aggressive or untrustworthy) even when punishment is encouraged by the punisher's role.

Enacting less harsh forms of punishment (such as ostracism, gossip, and verbal reproach) may also entail social or psychological costs. Although Guala argues that ostracism is not very costly, research suggests it is cognitively taxing and ego-depleting for the ostracizer. Participants who ostracized a confederate performed worse on subsequent tests of their physical and mental capacities (e.g., they solved fewer anagrams) then participants who did not ostracize a confederate (Ciarocco et al. Reference Ciarocco, Sommer and Baumeister2001). Even a seemingly innocuous form of punishment, gossip, is not without social costs: People who gossip negatively about others are less trusted and are more prone to negative reputations than those who do not gossip, even when controlling for the frequency of gossip (Turner et al. Reference Turner, Mazur, Wendel and Winslow2003). Finally, verbal reproach is also socially costly. Whistle-blowers who speak up against illegal behaviors perpetrated by employees of their organizations are susceptible to retaliation (e.g., negative performance evaluations, ostracism, dismissal) from members of the organization (Miceli et al. Reference Miceli, Near and Dworkin2008; Near & Miceli Reference Near and Miceli1995; Rothschild & Miethe Reference Rothschild and Miethe1999). Indeed, the prevalence of retaliation against whistle-blowers has led to the passage of legislation in the United States and other countries to attempt to protect whistle-blowers (see Miceli et al. [Reference Carlsmith and Darley2008] for a review). Other examples of the costs of verbal reproach abound: Whites such as Viola Liuzzo who protested racial discrimination and segregation during the Civil Rights Movement in the United States suffered physical harm, reputational and material costs, and even death (Stanton Reference Stanton2000). In sum, ostracism, gossip, and verbal reproach can all be psychologically or socially costly forms of punishment. Although many factors likely influence whether these costs are experienced in any given situation, we simply highlight that punishers sometimes incur such costs.

Beyond these psychological and social costs, there is also anecdotal evidence of material costs associated with punishing “in the wild,” such as when individuals or groups choose to boycott an organization. For example, the Dean and faculty at Vermont Law School denied military recruiters access to their campus facilities for many years because they opposed the military's “Don't Ask, Don't Tell” policy that prevents those who are openly gay, lesbian, or bisexual from serving in the military. As a result, the military had a difficult (but not impossible) time recruiting Vermont Law students, and the school forfeited approximately $500,000 in federal funding annually (Sanchez Reference Sanchez2005).

Given the various types of costs we have reviewed, it is worth noting that empirical evidence supports Guala's speculation that people's emotions or motivations might lead them to punish even when it is against their immediate self-interest. Psychological research demonstrates that people's desires to punish are driven primarily by retribution, such that people punish to see the offenders suffer in a manner proportionate to their wrongdoing, even if the punishment will not effectively deter future transgressions (see Carlsmith & Darley [2008] for a review). In other words, people may punish to satisfy their retributive desires, even when it is costly to do so.

In conclusion, Guala dismisses non-material costs by claiming that they are not very costly or that they are not relevant to arguments of group fitness. In contrast, we argue that broadening the definition of costs to include social and psychological costs can help to inform the debate about whether there is evidence of costly punishment “in the wild.”