Baumard et al. claim that I and my coauthors, in our work on human strategic choice behavior, hold that “human morality is first and foremost altruistic” (sect. 2.1.1, para. 4). This is not the case. In various publications (see Boehm Reference Boehm2011; Bowles & Gintis Reference Bowles and Gintis2011; Fehr & Gintis Reference Fehr and Gintis2007; Gintis et al. Reference Gintis, Bowles, Boyd and Fehr2005; Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath, Alvard, Barr, Ensminger, Hill, Gil-White, Gurven, Marlowe, Patton, Smith and Tracer2005; and references therein), we offer the following account of human social behavior: The human agent can be modeled as having a preference function that he maximizes subject to material and informational constraints, subject to his beliefs concerning the effect of his actions on social and personal outcomes. We call this the Beliefs, Preferences, and Constraints (BPC) model. The BPC model is a version of the rational actor model (Savage Reference Savage1972), except that beliefs may be constituted by the agent's position in a network of minds with distributed cognition, rather than being simply a personal subjective prior. Agents are genetically predisposed to value certain social and personal outcomes and devalue others, although this predisposition can be amplified and/or attenuated through social experience. Human preferences are conditioned by personal biological, welfare-related, and fitness-related needs (we call these self-regarding interests), but they generally have important elements that relate to the well-being of others (other-regarding preferences), and still others that are purely of a moral nature (such character virtues as honesty, loyalty, courage, considerateness, and worthiness of esteem).
In this framework, individuals are constantly faced with making trade-offs, not only among self-regarding goals (such as consumption and leisure), but also among self-regarding, other-regarding, and character virtue goals. Note that, in this model, individuals get pleasure from satisfying not only their self-regarding preferences, but also their social preferences, by which we mean their other-regarding preferences and their valued character virtues.
We suggest that two human genetic predispositions are particularly crucial. The first is the combination of conditional altruistic cooperation and conditional altruistic punishment, or strong reciprocity, according to which humans are predisposed to cooperate with unrelated others towards achieving collective goals, to punish those who free ride on the sacrifices of others, without an expectation of being repaid in the future for one's efforts. Both cooperation and punishment are conditional, in the sense that a sufficiently high level of defection leads agents to abandon cooperation, and in many situations individuals will participate in altruistic punishment only if there is a sufficient number of punishers (Boyd et al. Reference Boyd, Gintis and Bowles2010). Thus, individuals with strong other-regarding preferences will generally be on guard to detect cheating and self-serving activity of others.
The second crucial human predisposition is the capacity to internalize norms through the socialization process (Boehm Reference Boehm2011; Gintis Reference Gintis2003). The norms that are internalized appear as arguments in the individual's preference function, and include self-regarding elements (such as personal hygiene, ability to defer gratification), other-regarding elements (such as showing empathy for others), and character virtues (such as honesty and courage). In fact, most humans (sociopaths aside) have a conscience which they constantly deploy to evaluate their own behavior, and often curb immediate impulses to conform to the behavioral standards (self-regarding and social) to which they subscribe.
The background condition for the evolution of these human predispositions is our hunter-gatherer past, in which humans carved out a niche involving extremely high levels of cooperation among large numbers of non-kin, under rapidly varying environmental conditions requiring flexible adjustment of social practices to novel environmental challenges (Richerson & Boyd Reference Richerson, Boyd, Heyes and Huber2000). This, of course, is exactly the sort of “mutualistic cooperation” stressed by Baumard et al. Not surprisingly, all of the human behaviors affirmed by the authors fit nicely into the BPC model, and are in no way in conflict with our stress on altruistic cooperation and punishment. Since we are in broad agreement, I will simply suggest some amendments to their arguments.
The authors argue that moral values must be “real” rather than opportunistically feigned because people are not very good simulators and will eventually be unmasked unless their values are genuine. However, if there were a fitness benefit from dissimulation of morality, humans would have doubtless evolved the ability to dissimulate morality. My explanation in Gintis (Reference Gintis2003) is that morality is important for fitness maximization (including personal hygiene, deferred gratification, commitment to skill acquisition), and once humans evolved the capacity to internalize self-regarding virtues, the same psychological mechanisms could be “hijacked” for other purposes, including inculcating social preferences. Moreover, because humans generally suffer from excessively short time horizons (often called “weakness of will”), agents will behave inappropriately when the gains to moral behavior lie in the future, unless there is an immediate benefit from acting morally. Conscience supplies that immediate benefit for moral behavior (Durkheim Reference Durkheim1915, cf. Boehm Reference Boehm2011).
The authors' critique of our stress on altruistic punishment is not well founded. As Bingham (Reference Bingham1999) and Boyd et al. (Reference Boyd, Gintis and Bowles2010) have stressed, the ability of inflicting low-cost punishment on violators of social norms is the very defining feature of our species, and has no counterpart in other species. Of course, lethal punishment is rare, but it is universal in hunter-gather societies (Boehm Reference Boehm1999; Reference Boehm2011; Wiessner Reference Wiessner2005). Moreover, even ostracism and shunning involve strictly positive costs for those who participate. It is for this reason that other species do not include shunning and ostracism in their repertoire of behaviors. There is no good reason, of course, for Baumard et al. to question these well-established facts, as their thesis is not affected one way or another thereby.
Baumard et al. claim that I and my coauthors, in our work on human strategic choice behavior, hold that “human morality is first and foremost altruistic” (sect. 2.1.1, para. 4). This is not the case. In various publications (see Boehm Reference Boehm2011; Bowles & Gintis Reference Bowles and Gintis2011; Fehr & Gintis Reference Fehr and Gintis2007; Gintis et al. Reference Gintis, Bowles, Boyd and Fehr2005; Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath, Alvard, Barr, Ensminger, Hill, Gil-White, Gurven, Marlowe, Patton, Smith and Tracer2005; and references therein), we offer the following account of human social behavior: The human agent can be modeled as having a preference function that he maximizes subject to material and informational constraints, subject to his beliefs concerning the effect of his actions on social and personal outcomes. We call this the Beliefs, Preferences, and Constraints (BPC) model. The BPC model is a version of the rational actor model (Savage Reference Savage1972), except that beliefs may be constituted by the agent's position in a network of minds with distributed cognition, rather than being simply a personal subjective prior. Agents are genetically predisposed to value certain social and personal outcomes and devalue others, although this predisposition can be amplified and/or attenuated through social experience. Human preferences are conditioned by personal biological, welfare-related, and fitness-related needs (we call these self-regarding interests), but they generally have important elements that relate to the well-being of others (other-regarding preferences), and still others that are purely of a moral nature (such character virtues as honesty, loyalty, courage, considerateness, and worthiness of esteem).
In this framework, individuals are constantly faced with making trade-offs, not only among self-regarding goals (such as consumption and leisure), but also among self-regarding, other-regarding, and character virtue goals. Note that, in this model, individuals get pleasure from satisfying not only their self-regarding preferences, but also their social preferences, by which we mean their other-regarding preferences and their valued character virtues.
We suggest that two human genetic predispositions are particularly crucial. The first is the combination of conditional altruistic cooperation and conditional altruistic punishment, or strong reciprocity, according to which humans are predisposed to cooperate with unrelated others towards achieving collective goals, to punish those who free ride on the sacrifices of others, without an expectation of being repaid in the future for one's efforts. Both cooperation and punishment are conditional, in the sense that a sufficiently high level of defection leads agents to abandon cooperation, and in many situations individuals will participate in altruistic punishment only if there is a sufficient number of punishers (Boyd et al. Reference Boyd, Gintis and Bowles2010). Thus, individuals with strong other-regarding preferences will generally be on guard to detect cheating and self-serving activity of others.
The second crucial human predisposition is the capacity to internalize norms through the socialization process (Boehm Reference Boehm2011; Gintis Reference Gintis2003). The norms that are internalized appear as arguments in the individual's preference function, and include self-regarding elements (such as personal hygiene, ability to defer gratification), other-regarding elements (such as showing empathy for others), and character virtues (such as honesty and courage). In fact, most humans (sociopaths aside) have a conscience which they constantly deploy to evaluate their own behavior, and often curb immediate impulses to conform to the behavioral standards (self-regarding and social) to which they subscribe.
The background condition for the evolution of these human predispositions is our hunter-gatherer past, in which humans carved out a niche involving extremely high levels of cooperation among large numbers of non-kin, under rapidly varying environmental conditions requiring flexible adjustment of social practices to novel environmental challenges (Richerson & Boyd Reference Richerson, Boyd, Heyes and Huber2000). This, of course, is exactly the sort of “mutualistic cooperation” stressed by Baumard et al. Not surprisingly, all of the human behaviors affirmed by the authors fit nicely into the BPC model, and are in no way in conflict with our stress on altruistic cooperation and punishment. Since we are in broad agreement, I will simply suggest some amendments to their arguments.
The authors argue that moral values must be “real” rather than opportunistically feigned because people are not very good simulators and will eventually be unmasked unless their values are genuine. However, if there were a fitness benefit from dissimulation of morality, humans would have doubtless evolved the ability to dissimulate morality. My explanation in Gintis (Reference Gintis2003) is that morality is important for fitness maximization (including personal hygiene, deferred gratification, commitment to skill acquisition), and once humans evolved the capacity to internalize self-regarding virtues, the same psychological mechanisms could be “hijacked” for other purposes, including inculcating social preferences. Moreover, because humans generally suffer from excessively short time horizons (often called “weakness of will”), agents will behave inappropriately when the gains to moral behavior lie in the future, unless there is an immediate benefit from acting morally. Conscience supplies that immediate benefit for moral behavior (Durkheim Reference Durkheim1915, cf. Boehm Reference Boehm2011).
The authors' critique of our stress on altruistic punishment is not well founded. As Bingham (Reference Bingham1999) and Boyd et al. (Reference Boyd, Gintis and Bowles2010) have stressed, the ability of inflicting low-cost punishment on violators of social norms is the very defining feature of our species, and has no counterpart in other species. Of course, lethal punishment is rare, but it is universal in hunter-gather societies (Boehm Reference Boehm1999; Reference Boehm2011; Wiessner Reference Wiessner2005). Moreover, even ostracism and shunning involve strictly positive costs for those who participate. It is for this reason that other species do not include shunning and ostracism in their repertoire of behaviors. There is no good reason, of course, for Baumard et al. to question these well-established facts, as their thesis is not affected one way or another thereby.