Several reasons cast doubt on cultural group selection (CGS) as the key to prosocial religiosity. First, CGS is committed to a high degree of cultural homogeneity. However, not only can different populations share the same religion, but within one and the same population, several religions may coexist; when one group dominates another, several distinct possibilities may occur (homogenization, syncretism, sect explosion, etc.). Second, strong CGS (where behaviors are detrimental for the individual but beneficial for the group) seems to take place in the timescale of centuries (Soltis et al. Reference Soltis, Boyd and Richerson1995).
Other options can be considered: the explanatory role of coercion (Diamond Reference Diamond1997) or the stabilizing role of preexisting maladaptive biases (André & Morin Reference André and Morin2011). Here, we focus on a traditional contender to group-selectionist hypotheses: “greenbeard effects.” “Greenbeards” in evolutionary dynamics are arbitrary traits transmitted in conjunction with other traits that induce some form of preferential treatment to those carrying those same arbitrary traits (Gardner & West Reference Gardner and West2010). In the human cultural arena, this mechanism has been applied to social selection (Baumard Reference Baumard2010; Nesse Reference Nesse, Schaller, Heine, Norenzayan, Yamagishi and Kameda2009) and to partner choice (Noë & Hammerstein Reference Noë and Hammerstein1994). Similarly, it also readily suggests an account of why individuals who share the same religious beliefs – regardless of the content of those beliefs – have a preference to coordinate or cooperate with each other (Viciana Reference Viciana2014).
No doubt, appeals to putative greenbeards (in the form of ethnic markers) have been too common, and the evolutionary consistency of this mechanism has often been oversold in explanations of human prosociality. However, the basic evolutionary game-theoretical setting of this process remains relevant, if the possibility of free-riders – through the decoupling of the arbitrary trait (the “greenbeard”) and the cooperative behavior itself – is properly addressed. This is especially required when cultural evolution is involved, where agents keep some degree of rational choice over many of the behaviors they adopt (Fehr & Fischbacher Reference Fehr and Fischbacher2005).
When is such a form of tag-based cooperation to be expected? One possibility is when cultural traits are acquired in such a way that they reliably correlate with underlying behavioral tendencies (McElreath et al. Reference McElreath, Richerson and Boyd2003). Another possibility arises depending on outside opportunities of subgroups of agents in a population and the appearance of costly traits that reliably signal eagerness to enter into cooperative relationships (Iannaccone Reference Iannaccone1994). In what follows, we develop this second type of solution as a powerful way to explain the strategic adoption of beliefs and practices with increased prosociality.
In our model, agents choose to become part of groups that have three characteristics that are public and known: (1) an entry cost, that one could see as the cost of carrying a certain greenbeard or costly tag (i.e., a religious belief); (2) the degree of redistribution and participation in the production of public goods inside that group; and (3) the recent productivity of that group. Under this model, agents produce resources following a random sequence every “year” or cycle of the model. They share a portion of their production with the group and keep the rest for themselves. Therefore, each agent benefits from the average of resources shared by the group. Fitness is directly related to the resources obtained, but with diminishing returns (see online appendix, available at http://www.normalesup.org/~viciana/bbs_appendix.html). At each cycle, before sharing, the individuals know their productivity for this year, and they can decide either to stay in the same group or leave for a group with different past success and different sharing expectations. In this model, two kinds of groups survive in the long run: no-entry-cost groups with low levels of redistribution, and entry-cost groups with high levels of redistribution. Variability in individual opportunities may turn high costs attractive if they open the door to cooperative enterprises. As a result, the most “committed” individuals may find each other in a situation where they have little to gain from free-riding, as we see in Figure 1.
Figure 1.
Proportion of simulations in which groups with an entry cost survive (red), mean sharing in entry cost groups (solid contour lines), mean sharing in groups with no entry costs (dashed contours), as a function of σ (the larger σ, the larger the year-to-year resource variation) and λ (diminishing returns).
What does this model of the evolution of cooperation have to do with beliefs in the supernatural? Perhaps very little, except that one could say that there may be religious CREDs (Henrich Reference Henrich2009) that increase the credibility of the signaler insofar as there are landscapes of available opportunities (profitability of cooperative activities, lack of outside options) that may enhance the credulity of agents over certain religious forms. As an analogy of how partner choice mechanisms may operate here, think of the so-called Nigerian Internet scam (Herley Reference Herley2012). This scam self-selects the perfect victims with the minimum amount of cost dedicated to partner control (maximizing the degree of credulity of those answering the e-mail). In a similar fashion, certain social circumstances and evolving religious practices can self-select the most committed coreligionists, creating an environment for distributed degrees of credulity that may promote cooperative ventures. This does not need to be a fully intentional process.
The specific answer for why particular religious forms have come to play this role may be historical as much as psychological. From what we currently know, metaphysical beliefs about religious dogma tend to be stronger and more widespread in those populations where existential insecurity is higher (Inglehart & Welzel Reference Inglehart and Welzel2005). This finding holds as much across societies as it does inside societies (Norris & Inglehart Reference Norris and Inglehart2011). Those benefiting most from cooperation can have a strategic interest in adopting certain beliefs, under certain circumstances. From this point of view, the facilitation of cooperation through religious priming, found in experiments, may be partially mediated by the previous economic and sociological context of participants. In other words, do most religious primings induce substantial prosocial effects in the absence of previous coalitionary dynamics as the ones here described?
In summary, a shift in emphasis from group-functionalism to the specifics of coalitionary dynamics is required. Our model suggests that, in studying the prosociality of religious forms, one should pay as much attention to the design features of the credibility-enhancement devices as to the evolutionary landscapes of available opportunities that enhance the credulity of agents.
You have
Access
Several reasons cast doubt on cultural group selection (CGS) as the key to prosocial religiosity. First, CGS is committed to a high degree of cultural homogeneity. However, not only can different populations share the same religion, but within one and the same population, several religions may coexist; when one group dominates another, several distinct possibilities may occur (homogenization, syncretism, sect explosion, etc.). Second, strong CGS (where behaviors are detrimental for the individual but beneficial for the group) seems to take place in the timescale of centuries (Soltis et al. Reference Soltis, Boyd and Richerson1995).
Other options can be considered: the explanatory role of coercion (Diamond Reference Diamond1997) or the stabilizing role of preexisting maladaptive biases (André & Morin Reference André and Morin2011). Here, we focus on a traditional contender to group-selectionist hypotheses: “greenbeard effects.” “Greenbeards” in evolutionary dynamics are arbitrary traits transmitted in conjunction with other traits that induce some form of preferential treatment to those carrying those same arbitrary traits (Gardner & West Reference Gardner and West2010). In the human cultural arena, this mechanism has been applied to social selection (Baumard Reference Baumard2010; Nesse Reference Nesse, Schaller, Heine, Norenzayan, Yamagishi and Kameda2009) and to partner choice (Noë & Hammerstein Reference Noë and Hammerstein1994). Similarly, it also readily suggests an account of why individuals who share the same religious beliefs – regardless of the content of those beliefs – have a preference to coordinate or cooperate with each other (Viciana Reference Viciana2014).
No doubt, appeals to putative greenbeards (in the form of ethnic markers) have been too common, and the evolutionary consistency of this mechanism has often been oversold in explanations of human prosociality. However, the basic evolutionary game-theoretical setting of this process remains relevant, if the possibility of free-riders – through the decoupling of the arbitrary trait (the “greenbeard”) and the cooperative behavior itself – is properly addressed. This is especially required when cultural evolution is involved, where agents keep some degree of rational choice over many of the behaviors they adopt (Fehr & Fischbacher Reference Fehr and Fischbacher2005).
When is such a form of tag-based cooperation to be expected? One possibility is when cultural traits are acquired in such a way that they reliably correlate with underlying behavioral tendencies (McElreath et al. Reference McElreath, Richerson and Boyd2003). Another possibility arises depending on outside opportunities of subgroups of agents in a population and the appearance of costly traits that reliably signal eagerness to enter into cooperative relationships (Iannaccone Reference Iannaccone1994). In what follows, we develop this second type of solution as a powerful way to explain the strategic adoption of beliefs and practices with increased prosociality.
In our model, agents choose to become part of groups that have three characteristics that are public and known: (1) an entry cost, that one could see as the cost of carrying a certain greenbeard or costly tag (i.e., a religious belief); (2) the degree of redistribution and participation in the production of public goods inside that group; and (3) the recent productivity of that group. Under this model, agents produce resources following a random sequence every “year” or cycle of the model. They share a portion of their production with the group and keep the rest for themselves. Therefore, each agent benefits from the average of resources shared by the group. Fitness is directly related to the resources obtained, but with diminishing returns (see online appendix, available at http://www.normalesup.org/~viciana/bbs_appendix.html). At each cycle, before sharing, the individuals know their productivity for this year, and they can decide either to stay in the same group or leave for a group with different past success and different sharing expectations. In this model, two kinds of groups survive in the long run: no-entry-cost groups with low levels of redistribution, and entry-cost groups with high levels of redistribution. Variability in individual opportunities may turn high costs attractive if they open the door to cooperative enterprises. As a result, the most “committed” individuals may find each other in a situation where they have little to gain from free-riding, as we see in Figure 1.
Proportion of simulations in which groups with an entry cost survive (red), mean sharing in entry cost groups (solid contour lines), mean sharing in groups with no entry costs (dashed contours), as a function of σ (the larger σ, the larger the year-to-year resource variation) and λ (diminishing returns).
What does this model of the evolution of cooperation have to do with beliefs in the supernatural? Perhaps very little, except that one could say that there may be religious CREDs (Henrich Reference Henrich2009) that increase the credibility of the signaler insofar as there are landscapes of available opportunities (profitability of cooperative activities, lack of outside options) that may enhance the credulity of agents over certain religious forms. As an analogy of how partner choice mechanisms may operate here, think of the so-called Nigerian Internet scam (Herley Reference Herley2012). This scam self-selects the perfect victims with the minimum amount of cost dedicated to partner control (maximizing the degree of credulity of those answering the e-mail). In a similar fashion, certain social circumstances and evolving religious practices can self-select the most committed coreligionists, creating an environment for distributed degrees of credulity that may promote cooperative ventures. This does not need to be a fully intentional process.
The specific answer for why particular religious forms have come to play this role may be historical as much as psychological. From what we currently know, metaphysical beliefs about religious dogma tend to be stronger and more widespread in those populations where existential insecurity is higher (Inglehart & Welzel Reference Inglehart and Welzel2005). This finding holds as much across societies as it does inside societies (Norris & Inglehart Reference Norris and Inglehart2011). Those benefiting most from cooperation can have a strategic interest in adopting certain beliefs, under certain circumstances. From this point of view, the facilitation of cooperation through religious priming, found in experiments, may be partially mediated by the previous economic and sociological context of participants. In other words, do most religious primings induce substantial prosocial effects in the absence of previous coalitionary dynamics as the ones here described?
In summary, a shift in emphasis from group-functionalism to the specifics of coalitionary dynamics is required. Our model suggests that, in studying the prosociality of religious forms, one should pay as much attention to the design features of the credibility-enhancement devices as to the evolutionary landscapes of available opportunities that enhance the credulity of agents.