Baumard et al. propose that strong social selection based on partner choice in an environment in which mutual helping is highly advantageous led to the evolution of a self-serving sense of fairness in humans. Importantly, fairness preferences do not necessarily lead to equal outcomes but, instead, to payoffs that are distributed according to relative input, where rare abilities may yield particularly high shares. This view supports the market law of supply and demand: Fair payoff distributions mean that switching to another partner will not yield an average higher payoff. An individual with a sense of fairness is both an attractive and a vigilant partner: sharing with cooperating individuals but responding aversively to cheating individuals. While we thoroughly enjoyed the target article, we would like to raise two issues for further discussion.
Our first comment concerns social prestige. The authors argue that their functional approach to morality explains human cooperative behaviours, such as punishment and resource sharing, as a way of securing a good reputation as a cooperator. The underlying psychological mechanism is supposedly a genuine moral sense where cooperative behaviour is seen as intrinsically good rather than as a selfish concern for one's reputation. But how does the vast literature on image scoring and indirect reciprocity fit with this view? Humans are more cooperative if they can gain social prestige (Milinski et al. Reference Milinski, Semmann and Krambeck2002; Wedekind & Milinski Reference Wedekind and Milinski2000) and may even respond cooperatively to the presence of eyes (Bateson et al. Reference Bateson, Nettle and Roberts2006; Haley & Fessler Reference Haley and Fessler2005; but see Fehr & Schneider Reference Fehr and Schneider2010). These findings suggest that human cooperative behaviour is at least partly motivated by strategic concerns about reputation, rather than based solely on a genuine moral sense.
Our second concern is that Baumard et al. do not explicitly address whether a sense of fairness or morality should be unique to humans. Maybe they think that humans are not unique in this respect? At least on the level of outcomes of social interactions we would agree. Animals may indeed achieve rather uniform cooperative and seemingly fair outcomes based on partner selection in repeated games. However, we argue that seemingly fair outcomes need not be based on fairness preferences. While fairness preferences imply that individuals monitor and respond to the relative payoffs accruing to themselves and to a partner, a simpler alternative is that individuals have an internal expectation about payoffs from an interaction and adjust their behaviour (e.g., by switching partners) if these expectations are violated (Chen & Santos Reference Chen and Santos2006). Crucially, responses based on fairness preferences and responses based on self-referent loss aversion can both lead to cooperative and fair outcomes. We illustrate this point with our study system, the marine cleaning mutualism between cleaner wrasse Labroides dimidiatus and their reef-fish “clients.” (Apropos, please note that the biological ecological literature uses the term “mutualism” for cooperation between species, which makes the terminology proposed by the authors confusing when linked to biology.)
In contrast to great ape societies, our cleaning mutualism fulfils the criteria that Baumard et al. stipulate are conducive to the evolution of morality. In brief, the territorial cleaner wrasses are visited by clients at their “cleaning stations.” Both partners strongly benefit from interactions as cleaning is the wrasses' only mode of foraging and parasite removal translates into major health and growth benefits for clients (Ros et al. Reference Ros, Lusa, Meyer, Soares, Oliveira, Brossard and Bshary2011; Waldie et al. Reference Waldie, Blomberg, Cheney, Goldizen and Grutter2011). Conflict arises because cleaners prefer to eat the clients' protective mucus layer over ectoparasites (Grutter & Bshary Reference Grutter and Bshary2003). As individual clients visit cleaners several times per day, the game is clearly repeated. This makes partner switching an efficient partner control mechanism for “visitor” client species, which have access to several cleaning stations: cheating cleaners gain within an interaction but are excluded from future interactions (Bshary & Schäffer Reference Bshary and Schäffer2002). Due to their territoriality, individual cleaners can only win the competition with other cleaners by outbidding rather than by active interference, fitting the assumptions of biological market theory (Noë Reference Noë, Noë, van Hooff and Hammerstein2001). As a consequence, there is very strong convergence between cleaners with respect to the service quality they give to visitors: At 12 observed cleaning stations, visitors jolted on average either 2 or 3 times per 100 sec interaction (Bshary Reference Bshary2002, Fig. 1). Thus, the outcome looks as if based on a hidden contract regarding acceptable levels of cheating, but we consider it most likely that the outcome is due to individual learned optimization of own payoffs by both cleaners and clients.
Another interesting aspect of the cleaner wrasse mutualism is that cleaners often inspect clients jointly in established male-female couples, where the two cleaners face an iterated prisoner's dilemma-like game (Bshary et al. Reference Bshary, Grutter, Willener and Leimar2008). Couples find a cooperative solution which is based on asymmetric punishment. The larger, dominant males punish females for cheating, whereas females never punish males (Raihani et al. Reference Raihani, Grutter and Bshary2010). Although the pattern fits the definition of third-party punishment (the male punishes the female for biting a client), the punishment is obviously self-serving as it promotes future cooperation and males even fine-tune levels of punishment to their losses (Raihani et al. Reference Raihani, Pinto, Grutter, Wismer and Bshary2012). Intra-specific punishment in cleaner fish serves to restore “fair” outcomes during pairwise cleaning inspections, just as Baumard et al. have suggested that human punishment might. However, male cleaner fish need not attend to females' payoffs to know when to punish. Instead, they could use client departure and the associated reduction in payoffs (relative to expectations) as a cue to punish cheating females. Thus, punishment may be motivated by loss aversion rather than fairness preferences.
In conclusion, we have presented evidence for similarity in the importance of partner choice and punishment as control mechanisms, and similarity in outcomes between cleaning mutualism and human cooperation. Nevertheless, we would not assume a similarity in underlying mechanisms. More generally, there is little evidence that nonhuman animals have evolved fairness preferences, even though other-regarding behaviour is common across a diverse range of taxa (Burkart et al. Reference Burkart, Hrdy and van Schaik2009). The precise ecological conditions that favoured the evolution of a sense of morality which appears to be unique to humans therefore remain to be determined.
You have
Access
Baumard et al. propose that strong social selection based on partner choice in an environment in which mutual helping is highly advantageous led to the evolution of a self-serving sense of fairness in humans. Importantly, fairness preferences do not necessarily lead to equal outcomes but, instead, to payoffs that are distributed according to relative input, where rare abilities may yield particularly high shares. This view supports the market law of supply and demand: Fair payoff distributions mean that switching to another partner will not yield an average higher payoff. An individual with a sense of fairness is both an attractive and a vigilant partner: sharing with cooperating individuals but responding aversively to cheating individuals. While we thoroughly enjoyed the target article, we would like to raise two issues for further discussion.
Our first comment concerns social prestige. The authors argue that their functional approach to morality explains human cooperative behaviours, such as punishment and resource sharing, as a way of securing a good reputation as a cooperator. The underlying psychological mechanism is supposedly a genuine moral sense where cooperative behaviour is seen as intrinsically good rather than as a selfish concern for one's reputation. But how does the vast literature on image scoring and indirect reciprocity fit with this view? Humans are more cooperative if they can gain social prestige (Milinski et al. Reference Milinski, Semmann and Krambeck2002; Wedekind & Milinski Reference Wedekind and Milinski2000) and may even respond cooperatively to the presence of eyes (Bateson et al. Reference Bateson, Nettle and Roberts2006; Haley & Fessler Reference Haley and Fessler2005; but see Fehr & Schneider Reference Fehr and Schneider2010). These findings suggest that human cooperative behaviour is at least partly motivated by strategic concerns about reputation, rather than based solely on a genuine moral sense.
Our second concern is that Baumard et al. do not explicitly address whether a sense of fairness or morality should be unique to humans. Maybe they think that humans are not unique in this respect? At least on the level of outcomes of social interactions we would agree. Animals may indeed achieve rather uniform cooperative and seemingly fair outcomes based on partner selection in repeated games. However, we argue that seemingly fair outcomes need not be based on fairness preferences. While fairness preferences imply that individuals monitor and respond to the relative payoffs accruing to themselves and to a partner, a simpler alternative is that individuals have an internal expectation about payoffs from an interaction and adjust their behaviour (e.g., by switching partners) if these expectations are violated (Chen & Santos Reference Chen and Santos2006). Crucially, responses based on fairness preferences and responses based on self-referent loss aversion can both lead to cooperative and fair outcomes. We illustrate this point with our study system, the marine cleaning mutualism between cleaner wrasse Labroides dimidiatus and their reef-fish “clients.” (Apropos, please note that the biological ecological literature uses the term “mutualism” for cooperation between species, which makes the terminology proposed by the authors confusing when linked to biology.)
In contrast to great ape societies, our cleaning mutualism fulfils the criteria that Baumard et al. stipulate are conducive to the evolution of morality. In brief, the territorial cleaner wrasses are visited by clients at their “cleaning stations.” Both partners strongly benefit from interactions as cleaning is the wrasses' only mode of foraging and parasite removal translates into major health and growth benefits for clients (Ros et al. Reference Ros, Lusa, Meyer, Soares, Oliveira, Brossard and Bshary2011; Waldie et al. Reference Waldie, Blomberg, Cheney, Goldizen and Grutter2011). Conflict arises because cleaners prefer to eat the clients' protective mucus layer over ectoparasites (Grutter & Bshary Reference Grutter and Bshary2003). As individual clients visit cleaners several times per day, the game is clearly repeated. This makes partner switching an efficient partner control mechanism for “visitor” client species, which have access to several cleaning stations: cheating cleaners gain within an interaction but are excluded from future interactions (Bshary & Schäffer Reference Bshary and Schäffer2002). Due to their territoriality, individual cleaners can only win the competition with other cleaners by outbidding rather than by active interference, fitting the assumptions of biological market theory (Noë Reference Noë, Noë, van Hooff and Hammerstein2001). As a consequence, there is very strong convergence between cleaners with respect to the service quality they give to visitors: At 12 observed cleaning stations, visitors jolted on average either 2 or 3 times per 100 sec interaction (Bshary Reference Bshary2002, Fig. 1). Thus, the outcome looks as if based on a hidden contract regarding acceptable levels of cheating, but we consider it most likely that the outcome is due to individual learned optimization of own payoffs by both cleaners and clients.
Another interesting aspect of the cleaner wrasse mutualism is that cleaners often inspect clients jointly in established male-female couples, where the two cleaners face an iterated prisoner's dilemma-like game (Bshary et al. Reference Bshary, Grutter, Willener and Leimar2008). Couples find a cooperative solution which is based on asymmetric punishment. The larger, dominant males punish females for cheating, whereas females never punish males (Raihani et al. Reference Raihani, Grutter and Bshary2010). Although the pattern fits the definition of third-party punishment (the male punishes the female for biting a client), the punishment is obviously self-serving as it promotes future cooperation and males even fine-tune levels of punishment to their losses (Raihani et al. Reference Raihani, Pinto, Grutter, Wismer and Bshary2012). Intra-specific punishment in cleaner fish serves to restore “fair” outcomes during pairwise cleaning inspections, just as Baumard et al. have suggested that human punishment might. However, male cleaner fish need not attend to females' payoffs to know when to punish. Instead, they could use client departure and the associated reduction in payoffs (relative to expectations) as a cue to punish cheating females. Thus, punishment may be motivated by loss aversion rather than fairness preferences.
In conclusion, we have presented evidence for similarity in the importance of partner choice and punishment as control mechanisms, and similarity in outcomes between cleaning mutualism and human cooperation. Nevertheless, we would not assume a similarity in underlying mechanisms. More generally, there is little evidence that nonhuman animals have evolved fairness preferences, even though other-regarding behaviour is common across a diverse range of taxa (Burkart et al. Reference Burkart, Hrdy and van Schaik2009). The precise ecological conditions that favoured the evolution of a sense of morality which appears to be unique to humans therefore remain to be determined.