I appreciate Singh’s focus on the proximate psychological goals that people employ when evaluating and adopting different cultural traits. However, I would like to make two points regarding Singh’s target article, one related to the broad theory of subjective (cultural) selection and one related to the claims regarding cross-cultural universals (the “cultural manifold”).

First, while commendable, I do not see Singh’s broad framework of subjective selection as so different to many existing strands of cultural evolution research. Singh claims that “[u]nlike cultural group selection or cultural evolution through cue-biased transmission (e.g., imitating the healthy, successful, or prestigious)… subjective selection emphasizes subjective appeal over objective benefits.” Yet prominent theories regarding cue-biased transmission mechanisms already assume that the cues people use to adopt cultural traits are subjective. Take prestige bias, for example. In Henrich and Gil-White’s (Reference Henrich and Gil-White2001) formulation, prestige bias occurs when people copy others whom they perceive as being prestigious. These prestigious people may often possess adaptive (objectively or ultimately beneficial) knowledge or skills, but not necessarily. Rather, prestige is a subjective shortcut to acquiring adaptive cultural traits. Earlier models of prestige-like dynamics (Boyd & Richerson, Reference Boyd and Richerson1985) showed that while prestige (aka “indirect”) bias is broadly adaptive, it can misfire precisely because subjective prestige does not always correlate with objective success. This leads people to copy irrelevant or maladaptive traits exhibited by prestigious individuals. Experiments confirm that people are sensitive to this correlation, moderating their use of subjective prestige cues based on whether they correlate with objective success (Brand et al., Reference Brand, Heap, Morgan and Mesoudi2020) and whether that success was attained in a relevant domain to the trait being copied (Brand, Mesoudi, & Morgan, Reference Brand, Mesoudi and Morgan2021).

Elsewhere, transmission chain experiments show that information that is subjectively appealing, e.g., that evokes disgust (Eriksson & Coultas, Reference Eriksson and Coultas2014), has negative valence (Bebbington et al., Reference Bebbington, MacLeod, Ellison and Fay2017) or concerns social interactions (Mesoudi, Whiten, & Dunbar, Reference Mesoudi, Whiten and Dunbar2006), is preferentially recalled and transmitted. This is not necessarily the “objectively” correct or true information. In my study, made-up social information about fictional characters was transmitted better than factual but non-social information about climate change (Mesoudi et al., Reference Mesoudi, Whiten and Dunbar2006). Models (Tanaka, Kendal, & Laland, Reference Tanaka, Kendal and Laland2009) and real-world data (de Barra, Eriksson, & Strimling, Reference de Barra, Eriksson and Strimling2014) show how objectively ineffective, or even harmful, medical treatments can spread due to the biased subjective perception of their efficacy. And as Singh acknowledges (in Section 6.2. “Addressing assumptions of adaptiveness”), even in experiments which show that cultural evolution can occur without causal understanding of why a trait works (e.g. Derex et al., Reference Derex, Bonnefon, Boyd and Mesoudi2019), there is still the subjective evaluation of alternatives and selection of the best-performing cultural variant.

It is unclear to me how cultural selection could be anything other than “subjective,” except perhaps when driven by blind genetic algorithms or machine learning. The real question is when and why proximate, subjective psychological mechanisms converge or diverge from what would enhance individual inclusive fitness (or perhaps group fitness, under specific circumstances). This brings to mind Durham’s (Reference Durham1991) the distinction between “primary value selection,” where traits are selected based on goals derived from genetic evolution (e.g., the pursuit of food, shelter, or sex), and “secondary value selection,” where the goals are themselves derived from social learning (e.g., the pursuit of money, stamps, or video game points). I suspect Singh’s focus on subjective selection will be most useful in focusing attention on explaining where these latter, secondary, socially learned goals come from, how they themselves culturally evolve over time, and in some cases diverge from primary values.

Second, Singh argues that complex cross-cultural regularities – the “cultural manifold” – typically arise via independent convergence due to universal aspects of human psychology, which act to generate similar subjective selection pressures. While this may well be true for many of the examples he gives, especially shamanism and musical commonalities, other potential explanations for the cultural manifold should also be considered. While Singh dismisses deep/common ancestry as an alternative to independent convergence, he does not fully consider another more likely alternative: payoff-biased diffusion or migration across societies. (As an aside, this is perhaps ironic given the many quotations from Franz Boas in the target article, who was a prominent advocate for diffusionism within anthropology, e.g. Boas, Reference Boas1937) Formal models have shown that both convergence and diffusion can generate identical patterns of cross-cultural regularity (Mesoudi, Reference Mesoudi2021), while historical data show that many traits spread across societal borders via payoff-biased diffusion or migration, from agriculture (Fort, Reference Fort2015) to glass-making (Macfarlane & Martin, Reference Macfarlane and Martin2002) to musical instruments (Nia et al., Reference Nia, Jain, Liu, Alam, Barnas and Makris2015). Payoff-biased diffusion or migration is also a key factor in the spread of complex social institutions, as seen for example in post-feudal Japan’s adoption of Western institutions related to the police force, post office, and print media (Westney, Reference Westney1987) or the spread through history of institutional religions often due to their consequences for cooperation (Schulz et al., Reference Schulz, Bahrami-Rad, Beauchamp and Henrich2019; Stark, Reference Stark1996). Indeed, the more complex and therefore unintuitive or difficult to learn the cultural trait, the less likely it is to be independently reconstructed (Mesoudi, Reference Mesoudi2021). So while studying the psychological mechanisms driving independent convergence is important, we should not be blind to the equally important study of cultural diffusion and migration as drivers and shapers of the complex cultural manifold.