The target article muddles current theoretical issues regarding the evolution of human cooperation, and in the process creates an empty set of “strong reciprocity theorists.” To begin, it makes little sense to oppose weak versus strong reciprocity. Weak reciprocity is a particular class of theoretical evolutionary models. By contrast, “strong reciprocity” is a label and summary description for a set of empirical regularities that emerged from work in the United States and Europe (it is not an evolutionary theory). To explain these regularities as well as much ethnographic and cross-cultural evidence (Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr and Gintis2004) – which are not well handled by weak reciprocity theories (Chudek & Henrich Reference Chudek and Henrich2010, Fehr & Henrich Reference Fehr, Henrich and Hammerstein2003) – Boyd, Richerson, Fehr, Gintis, and Bowles (Guala's “strong reciprocity theorists,” hereafter BRFGB) and others have proposed a wide range of cultural and genetic evolutionary models. These models represent hypotheses about what the important mechanisms might be that sustain social norms. In particular, much work has focused on understanding the various ways that cultural evolution can harness and extend aspects of our evolved psychology (e.g., kin psychology) to create stable prosocial norms that could be favoured by cultural group selection (Alvard Reference Alvard2003, Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010a, Richerson & Boyd Reference Richerson, Boyd, Eibl-Eibesfeldt and Salter1998). Only a subset of these models involve the diffuse costly punishment (DCP) referred to by Guala and observed in some public goods games.
Numerous contributions from BRFGB illustrate that they are in no way wedded to DCP. A decade ago, Gintis et al. (Reference Gintis, Smith and Bowles2001) modelled how signalling (not DCP) could favour the provision of public goods in a manner aimed at explaining ethnographic observations among the turtle-hunting Meriam (Smith et al. Reference Smith, Bleige-Bird and Bird2003). In 2004, Panchanathan and Boyd showed how cultural evolution could stabilize norms by linking a dyadic helping game to a public goods game (Panchanathan & Boyd Reference Panchanathan and Boyd2004). There is no punishment there, let alone DCP. In his Nature perspective on this, Fehr (Reference Fehr2004) emphasizes the importance of reputational mechanisms – not based on DCP – to stabilize social norms. In Reference Boyd, Gintis and Bowles2010, Boyd et al. showed how signalling can coordinate punishment to stabilize social norms. Guala mentions this paper approvingly, while not noticing that it is written by the same authors that he says adhere only to DCP.
In these models, as in all models involving DCP, sanctioners have higher payoffs/fitness than non-sanctioners (under the appropriate conditions). There is no magic; these are evolutionary models that explore which strategies are favoured by selective processes and under what conditions. Guala seems to suggest that models based on DCP require that punishment have a net long-term cost. That is false. Sometimes sanctioning costs are “paid-for” via inter-group competition (Boyd et al. Reference Boyd, Gintis, Bowles and Richerson2003, Guzmanet al. Reference Guzman, Rodriguez-Sickert and Rowthorn2007), and sometimes these systems are just mutually self-reinforcing (Boyd & Richerson Reference Boyd and Richerson1992). Such costs have to net out somewhere, or be borne by some plausible constraint (weak reciprocity models all exploit mutational constraints; see Henrich (Reference Henrich2004). Much evolutionary modelling has sought to identify how informal institutions (sets of norms or reputational systems) might reduce or eliminate these costs. Cultural group selection will often favour those mechanisms that more effectively incentivize the sanctioning of prosocial norms while sustaining internal harmony (Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010a).
On the empirical side, the best evidence against the species-wide importance of DCP comes from a large-scale comparative project, initiated under the leadership of BFGB (Richerson not participating here), involving both experiments and ethnography in 24 different small-scale societies. Phase II of this project showed that community size is strongly positively associated with costly punishment. The analysis reveals that communities below a size of about 50 engage in little or no costly punishment (Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010a, Marlowe et al. Reference Marlowe, Berbesque, Barr, Barrett, Bolyanatz, Cardenas, Ensminger, Gurven, Gwako, Henrich, Henrich, Lesorogol, McElreath and Tracer2008). It is also the case that people from larger ethnic groups punish more. Going back a decade, these results confirm findings from Phase I of the project, in which people from three small-scale societies refused to reject low offers in the Ultimatum Game (Henrich Reference Henrich2000, Henrich et al. Reference Henrich, Boyd, Bowles, Gintis, Camerer, Fehr and McElreath2001).
The approach in this work not only explains the absence of DCP in many of the smallest-scale societies, it also accounts for why such punishing motivations emerge in larger-scale societies. Measures of punishment have not only been strongly associated with the size of stable communities and the success of ethnic groups, but they correlate strongly with gross domestic product (GDP) per capita across nations and with norms of civic cooperation and the rule of law (Herrmann et al. Reference Herrmann, Thöni and Gächter2008). In Ethiopia, measures of conditional cooperation, which have been closely linked to individuals' willingness to punish, predict monitoring and effective commons management (Rustagi et al. Reference Rustagi, Engel and Kosfeld2010). Once properly theorized, behavioural game measures – including those related to punishment – readily link to real-world sociality (Henrich et al. Reference Henrich, Heine and Norenzayan2010b).
While we agree that models relying on DCP (e.g., Henrich & Boyd Reference Henrich and Boyd2001) are not consistent with how norms are actually stabilized in small-scale societies, a variety of other kinds of informal sanctioning mechanisms are, including those that coordinate or incentivize ostracism and punishing. Thus, when Guala pulls quotations from BRFGB's work that refer to “punishment,” he implies that BRFGB refer only to DCP. If instead BRFGB mean “punishment” broadly defined, which results from community condemnations of norm violations channelled through local informal institutions (e.g., kinship or reputational systems), then Guala has misunderstood. Not only have BRFGB provided and promoted theoretical models not involving DCP, they also helped lead the project that have assembled the best evidence against the importance of DCP in small-scale societies. If by “punishment” they meant only DCP, then they would have to have been implicitly dismissing (1) some of their own theoretical models and (2) the fruits of their own anthropological collaboration. This seems unlikely.
We also agree with Guala's “narrow interpretation” of experiments, which was a central methodological element in the aforementioned large-scale comparative project. In the Phase I synthesis, one of the major points was that people bring motivations (values or heuristics) into the experimental games from everyday life (Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath, Alvard, Barr, Ensminger, Henrich, Hill, Gil-White, Gurven, Marlowe, Patton and Tracer2005). The authors drew on interviews, ethnographic observations, and local history to interpret their experimental results.
Much of the research program pursued by BRFGB converges with that favoured in the target article, as a central thrust is to understand the origins of the formal and informal institutions that govern life in both small-scale and modern societies. However, progress might be better pursued by focusing on points of actual difference, such as why relying on fully specified evolutionary models is preferable to invoking the folk theorem (where many equilibria are dynamically unstable), or why it is crucial to consider how formal institutions interface with social norms to endow people with internalized motivations (Chudek & Henrich Reference Chudek and Henrich2010).
The target article muddles current theoretical issues regarding the evolution of human cooperation, and in the process creates an empty set of “strong reciprocity theorists.” To begin, it makes little sense to oppose weak versus strong reciprocity. Weak reciprocity is a particular class of theoretical evolutionary models. By contrast, “strong reciprocity” is a label and summary description for a set of empirical regularities that emerged from work in the United States and Europe (it is not an evolutionary theory). To explain these regularities as well as much ethnographic and cross-cultural evidence (Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr and Gintis2004) – which are not well handled by weak reciprocity theories (Chudek & Henrich Reference Chudek and Henrich2010, Fehr & Henrich Reference Fehr, Henrich and Hammerstein2003) – Boyd, Richerson, Fehr, Gintis, and Bowles (Guala's “strong reciprocity theorists,” hereafter BRFGB) and others have proposed a wide range of cultural and genetic evolutionary models. These models represent hypotheses about what the important mechanisms might be that sustain social norms. In particular, much work has focused on understanding the various ways that cultural evolution can harness and extend aspects of our evolved psychology (e.g., kin psychology) to create stable prosocial norms that could be favoured by cultural group selection (Alvard Reference Alvard2003, Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010a, Richerson & Boyd Reference Richerson, Boyd, Eibl-Eibesfeldt and Salter1998). Only a subset of these models involve the diffuse costly punishment (DCP) referred to by Guala and observed in some public goods games.
Numerous contributions from BRFGB illustrate that they are in no way wedded to DCP. A decade ago, Gintis et al. (Reference Gintis, Smith and Bowles2001) modelled how signalling (not DCP) could favour the provision of public goods in a manner aimed at explaining ethnographic observations among the turtle-hunting Meriam (Smith et al. Reference Smith, Bleige-Bird and Bird2003). In 2004, Panchanathan and Boyd showed how cultural evolution could stabilize norms by linking a dyadic helping game to a public goods game (Panchanathan & Boyd Reference Panchanathan and Boyd2004). There is no punishment there, let alone DCP. In his Nature perspective on this, Fehr (Reference Fehr2004) emphasizes the importance of reputational mechanisms – not based on DCP – to stabilize social norms. In Reference Boyd, Gintis and Bowles2010, Boyd et al. showed how signalling can coordinate punishment to stabilize social norms. Guala mentions this paper approvingly, while not noticing that it is written by the same authors that he says adhere only to DCP.
In these models, as in all models involving DCP, sanctioners have higher payoffs/fitness than non-sanctioners (under the appropriate conditions). There is no magic; these are evolutionary models that explore which strategies are favoured by selective processes and under what conditions. Guala seems to suggest that models based on DCP require that punishment have a net long-term cost. That is false. Sometimes sanctioning costs are “paid-for” via inter-group competition (Boyd et al. Reference Boyd, Gintis, Bowles and Richerson2003, Guzmanet al. Reference Guzman, Rodriguez-Sickert and Rowthorn2007), and sometimes these systems are just mutually self-reinforcing (Boyd & Richerson Reference Boyd and Richerson1992). Such costs have to net out somewhere, or be borne by some plausible constraint (weak reciprocity models all exploit mutational constraints; see Henrich (Reference Henrich2004). Much evolutionary modelling has sought to identify how informal institutions (sets of norms or reputational systems) might reduce or eliminate these costs. Cultural group selection will often favour those mechanisms that more effectively incentivize the sanctioning of prosocial norms while sustaining internal harmony (Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010a).
On the empirical side, the best evidence against the species-wide importance of DCP comes from a large-scale comparative project, initiated under the leadership of BFGB (Richerson not participating here), involving both experiments and ethnography in 24 different small-scale societies. Phase II of this project showed that community size is strongly positively associated with costly punishment. The analysis reveals that communities below a size of about 50 engage in little or no costly punishment (Henrich et al. Reference Henrich, Ensminger, McElreath, Barr, Barrett, Bolyanatz, Cardenas, Gurven, Gwako, Henrich, Lesorogol, Marlowe, Tracer and Ziker2010a, Marlowe et al. Reference Marlowe, Berbesque, Barr, Barrett, Bolyanatz, Cardenas, Ensminger, Gurven, Gwako, Henrich, Henrich, Lesorogol, McElreath and Tracer2008). It is also the case that people from larger ethnic groups punish more. Going back a decade, these results confirm findings from Phase I of the project, in which people from three small-scale societies refused to reject low offers in the Ultimatum Game (Henrich Reference Henrich2000, Henrich et al. Reference Henrich, Boyd, Bowles, Gintis, Camerer, Fehr and McElreath2001).
The approach in this work not only explains the absence of DCP in many of the smallest-scale societies, it also accounts for why such punishing motivations emerge in larger-scale societies. Measures of punishment have not only been strongly associated with the size of stable communities and the success of ethnic groups, but they correlate strongly with gross domestic product (GDP) per capita across nations and with norms of civic cooperation and the rule of law (Herrmann et al. Reference Herrmann, Thöni and Gächter2008). In Ethiopia, measures of conditional cooperation, which have been closely linked to individuals' willingness to punish, predict monitoring and effective commons management (Rustagi et al. Reference Rustagi, Engel and Kosfeld2010). Once properly theorized, behavioural game measures – including those related to punishment – readily link to real-world sociality (Henrich et al. Reference Henrich, Heine and Norenzayan2010b).
While we agree that models relying on DCP (e.g., Henrich & Boyd Reference Henrich and Boyd2001) are not consistent with how norms are actually stabilized in small-scale societies, a variety of other kinds of informal sanctioning mechanisms are, including those that coordinate or incentivize ostracism and punishing. Thus, when Guala pulls quotations from BRFGB's work that refer to “punishment,” he implies that BRFGB refer only to DCP. If instead BRFGB mean “punishment” broadly defined, which results from community condemnations of norm violations channelled through local informal institutions (e.g., kinship or reputational systems), then Guala has misunderstood. Not only have BRFGB provided and promoted theoretical models not involving DCP, they also helped lead the project that have assembled the best evidence against the importance of DCP in small-scale societies. If by “punishment” they meant only DCP, then they would have to have been implicitly dismissing (1) some of their own theoretical models and (2) the fruits of their own anthropological collaboration. This seems unlikely.
We also agree with Guala's “narrow interpretation” of experiments, which was a central methodological element in the aforementioned large-scale comparative project. In the Phase I synthesis, one of the major points was that people bring motivations (values or heuristics) into the experimental games from everyday life (Henrich et al. Reference Henrich, Boyd, Bowles, Camerer, Fehr, Gintis, McElreath, Alvard, Barr, Ensminger, Henrich, Hill, Gil-White, Gurven, Marlowe, Patton and Tracer2005). The authors drew on interviews, ethnographic observations, and local history to interpret their experimental results.
Much of the research program pursued by BRFGB converges with that favoured in the target article, as a central thrust is to understand the origins of the formal and informal institutions that govern life in both small-scale and modern societies. However, progress might be better pursued by focusing on points of actual difference, such as why relying on fully specified evolutionary models is preferable to invoking the folk theorem (where many equilibria are dynamically unstable), or why it is crucial to consider how formal institutions interface with social norms to endow people with internalized motivations (Chudek & Henrich Reference Chudek and Henrich2010).