Manvir Singh’s target paper contains much new material and some that is less new. Unusually, this comment focuses on what is less new, not to diminish Singh’s own contribution but to provide some historical context that partly explains why the path he proposes for cultural evolution has not been taken so far.

A hundred years ago, Max Weber laid out an ambitious program for the social sciences (e.g., Weber & Tribe, Reference Weber and Tribe1921/2019). To simplify, Weber’s program rested on one assumption and one goal. The goal was a comparative study of political, legal, and religious institutions across cultures. The assumption was that social change is chiefly a consequence of people’s subjective motivations. A widely successful example of the Weberian approach is, of course, his analysis of capitalism. In the beginning, capitalistic accumulation had to be chosen by individuals who, by devoting their time and energy to the storing of wealth, went against their society’s values. They selected a capitalist lifeway for reasons that made subjective sense to them. Understanding those reasons was, Weber argued, the core of the social scientist’s job. This general theoretical orientation, methodological individualism, had a considerable posterity far beyond Weber (Heath, Reference Heath, Zalta and Nodelman2024). All major theoretical schools in the social sciences had to take a stance towards it.

Cultural evolution’s relationship with methodological individualism is complicated. The two approaches have much in common, starting with an atomistic view of social phenomena (built into the discipline by the routine use of agent-based models) and a wariness of holistic approaches. Yet, the core tenet of methodological individualism, the centrality of individuals’ subjective reasons for action, is not exactly a cornerstone of the discipline. Why? The answer differs from one school to the next.

Cultural group selection is a heterogeneous family of models and theories (Morin, Reference Morin2016; Richerson et al., Reference Richerson, Baldini, Bell, Demps, Frost, Hillis, Mathew, Newton, Naar, Newson, Ross, Smaldino, Waring and Zefferman2016). Some of its proponents are staunch supporters of methodological individualism (e.g., Gintis, Reference Gintis2007). Yet, as Singh rightly notes, the most famous and impactful group-selectionist models tend to be those where a cultural practice becomes widespread even though it runs counter to the immediate interests of those adopting it, like « strong reciprocity » (e.g., Gintis et al., Reference Gintis, Bowles, Boyd and Fehr2003). Relatedly, proponents of cultural group selection have emphasized the possibility that cultural practices (like food-processing technologies) may spread even though the individuals adopting them do not know why the practices are useful or even that they are useful at all (Derex et al., Reference Derex, Bonnefon, Boyd and Mesoudi2019; Henrich, Reference Henrich2016; see Lewens, Reference Lewens2024, and Osiurak & Reynaud, Reference Osiurak and Reynaud2019). If these models are accurate, cultural change can happen even if it goes against the grain of clear individual preferences or if its effects are not even registered by individual minds. In other words, cultural evolution can be blind to individual motivations. I agree with Singh that there are too few clear examples where cultural evolution acts like a « blind watchmaker », short-circuiting cognition. If Singh’s framework can serve as an antidote to cognition-blind cultural evolution, so much the better.

Cultural attraction’s lack of fit with methodological individualism has other roots. That theory stresses the importance of individual cognition for cultural change. The psychological factors that shape culture include goal-directed action and what accompanies it: subjectively available beliefs, desires, and reasons for acting. But the list does not stop at this level. One important motivation for cultural attraction as a research program is to draw the consequences, for social science, of psychological discoveries that were not available when social science’s main theoretical frameworks were formed. The founders of methodological individualism, like early theorists of Homo Economicus, built their frameworks on a view of human behavior in which action is driven by beliefs and desires that are conscious, coherent, and have causal power. They did not know about the cognitive unconscious, the modularity of mind, or bounded rationality. Since then, much has been learnt about psychological mechanisms that may not be conscious, or coherent, or causally efficient. Much of our cognitive life is but poorly integrated within the set of conscious beliefs and desires that form our subjective sense of self; for that reason, Sperber (Reference Sperber, Boudon, Chazel and Bouvier1997) made a plea to replace methodological individualism with an « infra-individualism ».

Cognitive mechanisms are not necessarily consciously accessible – hence, they are not always subjectively assessed. Most languages have a harmonic word order, where numeral adjectives occupy the same position as other adjectives (as in “two chairs”/“blue chairs”) (Culbertson et al., Reference Culbertson, Franck, Braquet, Barrera Navarro and Arnon2020). This is an easier syntactic structure to learn, even for speakers who grew up learning a non-harmonic language. Yet, speakers have little insight, if any, into the mechanisms that makes an harmonic order more learnable. Here, cognition impacts culture without conscious selection.

Contrary to another central tenet of methodological individualism, the individual judgments that shape culture may not form a coherent set of beliefs; alternatively, they may be coherent but not causally efficacious – that is, post hoc rationalizations that do not drive behavior. This has been well studied in the case of “theological correctness” – the mismatch between official religious dogmas and the beliefs, often implicitly held, that drive behavior even in rigorously trained religious specialists (Barrett, Reference Barrett1999). For instance, believers in Christianity or Hinduism explicitly endorse the view that God(s) are not located in space, yet routinely act and speak as if the opposite were true. The contradictory nature of these beliefs is no obstacle to their adoption: on the contrary, the possibility of switching between two contradictory interpretations of the same concept, one more intuitive than the other, can make religious concepts easier to manipulate and memorize (Boyer, Reference Boyer2007). Here, a cultural trait spreads, but for reasons that agents subjectively reject.

The point here is not to dispute the fact that people have irreplaceable insight into what drives them. Not all beliefs or desires are conscious, coherent, and causally efficient, but some are all three, and they matter for cultural change. The point is to adopt a finer-grained and more accurate view of cognition, and in so doing to illuminate some blind spots of standard social science.