The adaptationist approach of George Williams (Reference Williams1966) revolutionized evolutionary biology. It replaced a loose willingness to assume that every trait of an organism had an evolved function with, instead, a rigorous focus on characterizing adaptations – traits of organisms shaped by natural selection to solve an adaptive problem. Not all traits are adaptations: some are merely incidental by-products of adaptations and need no special adaptive explanation. Blood contains adaptations for the transport of oxygen and carbon dioxide; its transport properties were crafted by natural selection. Its red color when oxygenated is a by-product of its chemistry and needs no adaptive explanation. In Williams' words, “adaptation is a special and onerous concept that should be used only where it is really necessary” (p. 4). Researchers must carefully test whether a trait is in fact an adaptation and not recognize adaptation in “purely fortuitous effects” or invoke natural selection “to resolve problems that do not exist” (p. 4). The adaptationist program has been enormously productive: By the 1990s, after 30 years of work, the fruit of this approach already filled textbooks (e.g., Alcock Reference Alcock1998; Krebs & Davies Reference Krebs and Davies1993), and it is now the background assumption of virtually all research in animal behavior (Alcock Reference Alcock2001).
Williams had a further insight: Selection is most efficient at the lowest, most faithfully replicating levels in the hierarchy of life. Genes reproduce more faithfully than individuals, who reproduce more faithfully than groups. Thus, adaptions generally arise in the service of transmitting the genes that build them, not in the service of perpetuating the groups where individuals carrying those genes reside. But, as Williams noted, a herd of fleet deer is also a fleet herd. So, how was fleetness built? Parsimony says that, because genic selection is more powerful, unless herd fleetness shows special design that could not be preserved simply by the differential reproduction of speedier deer, group selection is not doing any explanatory work. Group selection should only be invoked with clear evidence of group-level adaptation, not merely group-level phenomena, an error repeatedly made by the target article.
Theories of cultural group selection, gene-culture coevolution, and dual inheritance are now at least 30 years old (e.g., Boyd & Richerson Reference Boyd and Richerson1985). Thus, we think it is fair to question their empirical yield. Have these theories made detailed, unique, and novel predictions confirmed with evidence showing specifically group-functional adaptation? Or has group-functional adaptation been applied when not necessary to solve problems that do not exist?
Our reading of the target article's section 6 suggests the latter. The best evidence offered for cultural group selection is the retrodiction that humans have language, sometimes punish each other, and have religion. There are no novel predictions made by cultural selection theory. (Simply re-describing known facts with the terminology of cultural group selection is not a novel prediction.) More critically, no evidence is offered that these traits arise from cultural group selection; we argue that apparent group-level effects are by-products arising from individual-level benefits when this alternative is taken seriously.
The evoked culture and cognitive niche (a.k.a. improvisational intelligence) hypotheses that Richerson et al. discard are two instances of this alternative. Evoked culture hypotheses propose facultative cognitive mechanisms that modulate behavior based on available information (Tooby & Cosmides Reference Tooby, Cosmides, Barkow, Cosmides and Tooby1992). These mechanisms evolve when, ancestrally, different behaviors are best in different physical and social ecologies. This was surely the case for punishment, as the utility of using coercive power to get one's way would have ancestrally varied between people (e.g., aggression pays more for the strong; Sell et al. Reference Sell, Tooby and Cosmides2009) and between ecologies (e.g., aggression pays more when there are gains it can recruit; Krasnow et al. Reference Krasnow, Cosmides, Pedersen and Tooby2012). Cognitive niche hypotheses propose mechanisms that extract information from the environment and make appropriate inferences (Tooby & DeVore Reference Tooby, DeVore and Kinzey1987).
Importantly, these theories are not the straw-men reported in the target article. Why should a mechanism for exploiting the cognitive niche rely on solitary innovation when a clearly superior design would also target information in the minds of others? Why should a mechanism for evoked culture be unable to operate outside the parameter range under which it evolved? Our mechanisms of diet choice gravitate to big macs and cheesecake despite none existing during our evolutionary history. Although police forces and the rule of law did not exist in Paleolithic society, our evolved psychology of personal aggression can be exploited by these institutions to evoke cultures that never existed before. To the extent that our evolved social psychology is richly articulated, features of this evoked culture can act as further inputs to other facultative mechanisms and evoke further novel cultural features (Tooby & Cosmides Reference Tooby, Cosmides, Barkow, Cosmides and Tooby1992).
A final note: We are surprised that a paper discussing the taxonomy of group selection and cultural coevolutionary models can so repeatedly mistake simple cultural change for cultural group selection. There is no dispute that cultures change over time. This is, in fact, the predicted result when creatures exploiting the cognitive niche leave enduring traces in their societies that evoke new cultures down the line (Pinker Reference Pinker2012). But as Williams argued, adaptation is an onerous concept. It is no less onerous for culturally group selected adaptations. The burden has not been met here.
In our view, the jury is still out on the explanatory and predictive power of cultural group selection theory. Luckily, it is not the only approach on offer. About 30 years ago another theoretical approach emerged as the confluence of adaptationism and cognitive science: evolutionary psychology (e.g., Daly & Wilson Reference Daly and Wilson1988; Tooby & Cosmides Reference Tooby and Cosmides1989). Although Richerson and colleagues criticize evolutionary psychological approaches, it is evolutionary psychology that has inspired scores of novel, specific, and empirically confirmed predictions. The many empirical successes of evolutionary psychology have filled textbooks, handbooks, and hundreds of journal articles (e.g., Buss Reference Buss2005; Reference Buss2014; Buss & Reeve Reference Buss and Reeve2003). Although cultural group selection theory may yet yield insights, at the moment it is evolutionary psychology that provides a generative, testable, and productive framework for understanding the human mind and the cultures it creates.
The adaptationist approach of George Williams (Reference Williams1966) revolutionized evolutionary biology. It replaced a loose willingness to assume that every trait of an organism had an evolved function with, instead, a rigorous focus on characterizing adaptations – traits of organisms shaped by natural selection to solve an adaptive problem. Not all traits are adaptations: some are merely incidental by-products of adaptations and need no special adaptive explanation. Blood contains adaptations for the transport of oxygen and carbon dioxide; its transport properties were crafted by natural selection. Its red color when oxygenated is a by-product of its chemistry and needs no adaptive explanation. In Williams' words, “adaptation is a special and onerous concept that should be used only where it is really necessary” (p. 4). Researchers must carefully test whether a trait is in fact an adaptation and not recognize adaptation in “purely fortuitous effects” or invoke natural selection “to resolve problems that do not exist” (p. 4). The adaptationist program has been enormously productive: By the 1990s, after 30 years of work, the fruit of this approach already filled textbooks (e.g., Alcock Reference Alcock1998; Krebs & Davies Reference Krebs and Davies1993), and it is now the background assumption of virtually all research in animal behavior (Alcock Reference Alcock2001).
Williams had a further insight: Selection is most efficient at the lowest, most faithfully replicating levels in the hierarchy of life. Genes reproduce more faithfully than individuals, who reproduce more faithfully than groups. Thus, adaptions generally arise in the service of transmitting the genes that build them, not in the service of perpetuating the groups where individuals carrying those genes reside. But, as Williams noted, a herd of fleet deer is also a fleet herd. So, how was fleetness built? Parsimony says that, because genic selection is more powerful, unless herd fleetness shows special design that could not be preserved simply by the differential reproduction of speedier deer, group selection is not doing any explanatory work. Group selection should only be invoked with clear evidence of group-level adaptation, not merely group-level phenomena, an error repeatedly made by the target article.
Theories of cultural group selection, gene-culture coevolution, and dual inheritance are now at least 30 years old (e.g., Boyd & Richerson Reference Boyd and Richerson1985). Thus, we think it is fair to question their empirical yield. Have these theories made detailed, unique, and novel predictions confirmed with evidence showing specifically group-functional adaptation? Or has group-functional adaptation been applied when not necessary to solve problems that do not exist?
Our reading of the target article's section 6 suggests the latter. The best evidence offered for cultural group selection is the retrodiction that humans have language, sometimes punish each other, and have religion. There are no novel predictions made by cultural selection theory. (Simply re-describing known facts with the terminology of cultural group selection is not a novel prediction.) More critically, no evidence is offered that these traits arise from cultural group selection; we argue that apparent group-level effects are by-products arising from individual-level benefits when this alternative is taken seriously.
The evoked culture and cognitive niche (a.k.a. improvisational intelligence) hypotheses that Richerson et al. discard are two instances of this alternative. Evoked culture hypotheses propose facultative cognitive mechanisms that modulate behavior based on available information (Tooby & Cosmides Reference Tooby, Cosmides, Barkow, Cosmides and Tooby1992). These mechanisms evolve when, ancestrally, different behaviors are best in different physical and social ecologies. This was surely the case for punishment, as the utility of using coercive power to get one's way would have ancestrally varied between people (e.g., aggression pays more for the strong; Sell et al. Reference Sell, Tooby and Cosmides2009) and between ecologies (e.g., aggression pays more when there are gains it can recruit; Krasnow et al. Reference Krasnow, Cosmides, Pedersen and Tooby2012). Cognitive niche hypotheses propose mechanisms that extract information from the environment and make appropriate inferences (Tooby & DeVore Reference Tooby, DeVore and Kinzey1987).
Importantly, these theories are not the straw-men reported in the target article. Why should a mechanism for exploiting the cognitive niche rely on solitary innovation when a clearly superior design would also target information in the minds of others? Why should a mechanism for evoked culture be unable to operate outside the parameter range under which it evolved? Our mechanisms of diet choice gravitate to big macs and cheesecake despite none existing during our evolutionary history. Although police forces and the rule of law did not exist in Paleolithic society, our evolved psychology of personal aggression can be exploited by these institutions to evoke cultures that never existed before. To the extent that our evolved social psychology is richly articulated, features of this evoked culture can act as further inputs to other facultative mechanisms and evoke further novel cultural features (Tooby & Cosmides Reference Tooby, Cosmides, Barkow, Cosmides and Tooby1992).
A final note: We are surprised that a paper discussing the taxonomy of group selection and cultural coevolutionary models can so repeatedly mistake simple cultural change for cultural group selection. There is no dispute that cultures change over time. This is, in fact, the predicted result when creatures exploiting the cognitive niche leave enduring traces in their societies that evoke new cultures down the line (Pinker Reference Pinker2012). But as Williams argued, adaptation is an onerous concept. It is no less onerous for culturally group selected adaptations. The burden has not been met here.
In our view, the jury is still out on the explanatory and predictive power of cultural group selection theory. Luckily, it is not the only approach on offer. About 30 years ago another theoretical approach emerged as the confluence of adaptationism and cognitive science: evolutionary psychology (e.g., Daly & Wilson Reference Daly and Wilson1988; Tooby & Cosmides Reference Tooby and Cosmides1989). Although Richerson and colleagues criticize evolutionary psychological approaches, it is evolutionary psychology that has inspired scores of novel, specific, and empirically confirmed predictions. The many empirical successes of evolutionary psychology have filled textbooks, handbooks, and hundreds of journal articles (e.g., Buss Reference Buss2005; Reference Buss2014; Buss & Reeve Reference Buss and Reeve2003). Although cultural group selection theory may yet yield insights, at the moment it is evolutionary psychology that provides a generative, testable, and productive framework for understanding the human mind and the cultures it creates.