The case for the complex and context-dependent nature of gene–culture interactions is well made by Uchiyama et al. However, the dual inheritance model they describe could be extended to provide greater explanatory power by incorporating epigenetic effects. Epigenetic gene regulation can amplify phenotypic variation (mimicking greater gene variation) and accelerate both cultural and genetic evolution.

“Epigenetics” refers to changes in gene expression brought about by chemical modifications that do not change the DNA sequence itself. The processes themselves appear to operate mainly during development rather than adulthood – preparing the individual mammal for the specific situation into which it has been born. Epigenetically regulated genes are disproportionately found in the brain, suggesting selection for influencing behaviour and hence, culture (Keverne, Reference Keverne2014). Epigenetics explains how long-term changes in brain chemistry are programmed by short-term experiences, especially during developmental windows in early life.

These epigenetic tags, such as DNA methylation, are reversible but can potentially persist across two generations if they occur in a woman pregnant with a daughter who carries all the eggs for her lifetime while still in the womb herself. This timescale is much more comparable with cultural change than classical selection for true allelic variation, and so an invaluable bridge between the two. To date, a major obstacle to even modelling gene–culture coevolution has been the different timescales, because culture can change very rapidly compared to gene frequencies.

The authors refer to prosocial norms as an example of a cultural psychological trait that has been modelled as culturally evolving and the associated trait of “social trust” can be provided here as an example of how epigenetics can inform models of gene–culture coevolution. Norms of social trust are also transmitted from birth both vertically and horizontally. A cultural spectrum influenced by norms of social trust runs, for example, from more collective cultures to more individualistic cultures.

Trusting behaviour is regulated by a hub of neurotransmitters and hormones including oxytocin and serotonin (Riedl & Javor, Reference Riedl and Javor2012). There are a number of genes that influence the levels of these chemicals such as the oxytocin receptor gene, OXTR, and the serotonin transporter gene, SERT. Both of these genes are polymorphic, with alleles associated with different levels of influence on trusting behaviour (Feldman, Monakhov, Pratt, & Ebstein, Reference Feldman, Monakhov, Pratt and Ebstein2016; Iurescia, Seriipa, & Rinaldi, Reference Iurescia, Seriipa and Rinaldi2016). At least some of the alleles are also epigenetically regulated (Iurescia et al., Reference Iurescia, Seriipa and Rinaldi2016; Kumsta et al., Reference Kumsta, Hummel, Chen and Heinrichs2013).

In mammalian history, a stressful world would typically have been one with a harsh environment resulting in strong competition for resources in terms of food, mates, and territory. In such an environment, positive trust and prosocial behaviour may have deadly consequences and come at too high a cost. However, the exact costs and benefits of trust may vary locally and over shorter timescales, and classic selection would be too insensitive. The advantage of epigenetic processes are that they enable gene expression to respond developmentally to these changing trade-offs caused by rapid environmental (sensu lato) signals.

The serotonin transporter gene illustrates how much variation in adaptation is possible given a gene that is both polymorphic and epigenetically regulated. The short allele of the SERT gene, “S,” is less expressed than the long allele, “L.” Caspi et al.'s (Reference Caspi, Sugden, Moffitt, Taylor, Craig, Harrington and Poulton2003) study found that this common polymorphism of the serotonin receptor gene affects the probability of depressive episodes as the number of stressful life events accumulates (i.e., there is a gene–environment interaction mediating the risk of depression). S carriers were vulnerable to this effect because the S allele is epigenetically regulated – it is methylated in response to stress in early life. Methylation of the serotonin transporter gene is associated with altered emotional processing mediated by altered brain activity in regions including the amygdala and anterior insula (Frodl et al., Reference Frodl, Szyf, Carballedo, Ly, Dymov, Varisheva and Booij2015). The result is an increased response to fear which promotes mistrust. S carriers are less trusting, and more discriminatory to outgroups – especially under stress.

Rather than viewing the S allele as a risk factor, however, it is better modelled as facilitating a sensitivity that also has positive consequences. For example, S carriers are also more sensitive to social signalling generally – and derive more benefit from social support (Way & Lieberman, Reference Way and Lieberman2010). Social support is an effective buffer against the increased sensitivity to stress experienced by S carriers. Without variable levels of stress in their lives, the different effects of S and L alleles in the carriers would be undetectable. Clearly, an awareness of the different epigenetic regulation of S and L carriers is necessary to fully evaluate heritability estimates of associated traits such as social trust.

Global population differences in S and L allele frequencies are well-documented (Minkov, Blagoev, & Bond, Reference Minkov, Blagoev and Bond2015). S frequencies are consistently higher in East Asian populations than that in North European (70–80% S carriers vs. 40–45%). This prompts the questions: Have population differences in S and L allele frequencies come about by neutral processes or selection? Collectivism in Asia is associated with high S allele frequency, and there is some evidence that the increased social support in collectivist cultures might buffer against the increased risk of stress-vulnerability and depression (Way & Lieberman, Reference Way and Lieberman2010).

Socially sensitive OXTR allele frequencies also vary East to West in a similar way (Luo & Han, Reference Luo and Han2014), further supporting a gene–culture coevolution model involving social trust. An intervention in the United States providing social support to at-risk families has been found to be more effective with S carriers who derive more benefit from social support – and in this case there is also less OXTR methylation (Beach et al., Reference Beach, Lei, Brody and Philibert2018).

So, there appears to be a network of genes that are epigenetically sensitive to stress and possibly social stress in particular. Indeed, it is likely that the detection of the influence of single alleles is made possible because the phenotypic measure actually reflects selection at multiple related loci.

The suggestion that epigenetic regulation of socially sensitive alleles of genes such as SERT and OXTR can influence group level behaviour raises the possibility of epigenetically mediated gene–culture coevolution whereby ecological stress and/or social stress impacts cultural traits such as social trust. This is just one example illustrating how epigenetic effects may impact gene–culture coevolution and suggest the need for a triple inheritance model rather than a dual one, with further implications for sampling and estimating heritability.