Cultural group selection (CGS) — the spread of group-beneficial cultural traits because groups carrying them outcompete other culturally distinct groups — is a leading explanation for human cooperation among large numbers of non-relatives, but its actual causal importance has never been quantified. The 2014–2016 Behavioral and Brain Sciences exchange (Richerson et al.'s target article, Lamba's critical commentary, and the authors' response) converged on a precise open problem: for particular cooperative traits and institutions, causally identify and quantify the relative contribution of the three CGS processes — differential group proliferation/extinction, selective imitation of successful groups, and selective migration — against individual-level mechanisms such as kin selection, reciprocity, and partner choice. This requires behaviorally grounded measures of culturally heritable between-group variance (cultural analogues of FSTF_{ST}) and archaeohistorical, ethnographic, or genomic time-series suitable for mechanistic model fitting; genomic identification of culture-mediated selection on behavioral traits is a further desideratum of the same exchange rather than a required input to the attribution. A decade of subsequent work has advanced components (field tests of CGS predictions, the Seshat moralizing-gods program, cultural-distance measurement, institutional-evolution frameworks) but no study has delivered the attribution itself, and reviews through 2020 describe the question as unresolved.

Background

Humans cooperate at scales — hundreds to millions of mostly unrelated individuals — that no other primate approaches, and evolutionary anthropology has long debated why. Candidate explanations divide into two families. Individual-level mechanisms operate on the fitness or payoffs of individual actors: kin selection (helping relatives because shared genes make aid genetically profitable), direct and indirect reciprocity (cooperation sustained by repeated interaction or by reputation, where helpful individuals are chosen as partners), and partner choice (competiting to be a valuable collaborator). Cultural group selection (CGS) instead invokes selection acting on variation between culturally distinct groups: if socially learned norms and institutions differ among groups, reduce within-group free-riding, and make groups differentially successful in survival, growth, imitation by outsiders, or attraction of migrants, then group-beneficial cultural traits can spread even when costly to individuals. Richerson and colleagues distinguish three CGS processes: differential group proliferation or extinction (groups with favorable cultures survive, expand, or absorb others), selective imitation (members of less successful groups copy the practices of more successful ones), and selective migration (individuals move preferentially into successful groups). Institutions — culturally transmitted systems of rules, monitoring, and sanctioning, from gossip and taboos to legal codes — are the plausible carriers of such group-level variation because they structure whole-group payoffs.

The debate crystallized in the target article by Richerson et al. (2016), which assembled a qualitative case: culture is a faithful inheritance system; measured between-group cultural variance is often large; institutions are group-level cultural traits; intergroup competition is widespread; and culture-led genetic change is documented (as with lactase persistence under dairying, a canonical case of gene–culture coevolution, i.e. genetic change driven by a culturally transmitted practice). A key quantitative benchmark is cultural FSTF_{ST}, the fraction of total variance in a cultural trait that lies between rather than within groups, directly analogous to the population-genetic fixation index; Bell, Richerson and McElreath (2009) compiled survey-based cultural FSTF_{ST} estimates showing between-group cultural variance often exceeds what genetic group selection could overcome. Lamba's (2016) commentary, drawing on his fieldwork with Mace showing that cooperative and bargaining behavior tracks current demography and ecology rather than stable group norms, argued that the evidence does not support CGS: the FSTF_{ST} values come from self-reported opinions rather than costly behavior; classic intergroup-competition cases (pastoralist raiding, historical firm and religious expansion) show correlation between cultural traits and group success, not that the traits caused the success; and the genomic evidence covers physiologically simple, large-effect loci rather than behavioral dispositions. The authors' response acknowledged these as legitimate priorities: quantifying the heritable cultural-variance component, case-by-case causal demonstration, the relative strengths of the three CGS mechanisms, genomic identification of behavioral loci shaped by culture-led selection, separation of geographic clines from discrete group boundaries, and model comparison of CGS against individual-level alternatives. Later commentary (Turchin and Currie 2016) made the methodological point explicit: CGS hypotheses need tests against real-world data, not plausibility arguments. Smith's (2020) systematic review added a conceptual obstacle — CGS is used in at least three nonequivalent senses (a 'maladaptive' version relying on conformist social learning, a 'normative' version relying on punishment-backed norm psychology, and a 'mechanism-neutral' version relying only on intergroup competition), so any attribution must state which variety is being quantified — and concluded that the extent to which CGS actually shaped human cooperative evolution remains unclear. Fieldwork since has tested single predictions (Handley and Mathew 2020 showed that readiness to cooperate between Kenyan pastoralist groups tracks their cultural similarity), and the Seshat Global History Databank program attempted the kind of archaeohistorical causal time-series test this problem calls for on moralizing gods, with instructive failure: the original Nature paper (Whitehouse et al. 2019) was retracted after data-quality disputes, and the reanalyses (Whitehouse et al. 2022; Turchin et al. 2022) remain contested. The same exchange raises three further questions that are distinct from the attribution sought here: on what units CGS operates in concrete cases (organizational institutions, ethnolinguistic groups, meme-complexes — clusters of culturally transmitted ideas that spread together — or nested combinations; Richerson et al. 2016, response sect. R8); when CGS processes were most important (in the Pleistocene versus the Holocene, the epoch since agriculture began; response sects. R10 'Ancient preadaptations' and R11 'Pleistocene adaptations generated by CGS'); and under what conditions institutional networks freeze versus restructure (response sect. R5, on the 'frozen plasticity' idea that as societies become more institutionally complex, interactions between them make dysfunctional institutions increasingly hard to change until large-scale events produce a punctuated restructuring). Work on the cultural evolution and ecology of institutions (Currie et al. 2021) and on the scale of cooperation in Pleistocene (the pre-agricultural epoch, ending roughly 11,700 years ago) foraging societies (Boyd and Richerson 2022) addresses the questions about institutional rigidity and timing but not the attribution itself.

Problem Statement

For particular, named human cooperative traits or institutions, causally identify and quantify the relative contribution of the three cultural group selection (CGS) processes — differential group proliferation or extinction (groups with favorable cultures surviving, expanding, or absorbing others), selective imitation of successful groups, and selective migration (individuals moving preferentially into successful groups) — against individual-level mechanisms (kin selection, i.e. help directed at relatives; direct and indirect reciprocity, i.e. cooperation sustained by repeated interaction or by reputation; and partner choice, i.e. competition to be chosen as a collaborator) in producing those traits, with uncertainty attached to the attribution. The attribution must state which sense of CGS it quantifies — mechanism-neutral (intergroup competition alone), normative (punishment-backed norm psychology), or maladaptive (conformist social learning) — since these are nonequivalent and the quantified result is otherwise not well defined.

The standard set by the exchange is case-by-case rather than field-wide: CGS is 'a plausible force while ... its action needs to be demonstrated in actual cases' (Richerson et al. 2016, response sect. R2), and the authors concede that 'in no particular case can we put all the pieces of a complete CGS explanation together' (response sect. R14). For the trait(s) analyzed, those pieces are evidence that the trait is culturally transmitted (socially learned, and hence a culturally heritable group-level variant) rather than genetically transmitted or independently invented, together with evidence that variation in the trait caused differences in group fitness or demographic success in historical or ethnographic cases rather than merely correlating with group success. The exchange endorses as the means of quantification better estimates of the culturally heritable component of within- and between-group variation — culturally meaningful analogues of FSTF_{ST} (the fraction of total trait variance that lies between rather than within groups), not confined to self-reported opinion surveys and computable for nominal group classifications or for clinal geographic structure (gradual trait change across space rather than discrete group boundaries) at a chosen resolution, together with trait-specific transmission modes — and the fitting or comparison of competing CGS and individual-level models on archaeohistorical, ethnographic, linguistic, or genomic data, aiming at the 'gold standard of accurately estimating the strength of any of the three CGS processes' (target article sect. 7). An attribution need not exclude the individual-level alternatives: the response states that there is 'no need to prove the case to the exclusion of all other processes' and that competing hypotheses or models are often not mutually exclusive (sect. R14).

A rigorous negative result for the analyzed trait(s) is an admissible resolution. The target article frames its case as a syllogism whose legs are questions, adding that 'If the answer to one of these questions is "no," then we must look to other hypotheses' (Abstract). Accordingly, a finding that the behaviour is not culturally transmitted, that observed between-group behavioural variation tracks current demography and ecology rather than stable group norms, or that CGS's estimated contribution to the trait is negligible, each answers the question for that trait, provided the finding is causally identified and its uncertainty stated.

References

  1. Lamba, Shakti (2016). 'The empirical evidence that does not support cultural group selection models for the evolution of human cooperation.' Behavioral and Brain Sciences 39, e44. DOI: 10.1017/S0140525X15000175. https://doi.org/10.1017/S0140525X15000175
  2. Richerson, Peter; Baldini, Ryan; Bell, Adrian V.; Demps, Kathryn; Frost, Karl; Hillis, Vicken; Mathew, Sarah; Newton, Emily K.; Naar, Nicole; Newson, Lesley; Ross, Cody; Smaldino, Paul E.; Waring, Timothy M.; Zefferman, Matthew (2016). 'Cultural group selection plays an essential role in explaining human cooperation: A sketch of the evidence.' Behavioral and Brain Sciences 39, e30. DOI: 10.1017/S0140525X1400106X. https://doi.org/10.1017/S0140525X1400106X
  3. Richerson, Peter; Baldini, Ryan; Bell, Adrian V.; Demps, Kathryn; Frost, Karl; Hillis, Vicken; Mathew, Sarah; Newton, Emily K.; Naar, Nicole; Newson, Lesley; Ross, Cody; Smaldino, Paul E.; Waring, Timothy M.; Zefferman, Matthew (2016). 'Cultural group selection follows Darwin's classic syllogism for the operation of selection' (response to open peer commentary). Behavioral and Brain Sciences 39, e58. DOI: 10.1017/S0140525X15000606. https://doi.org/10.1017/S0140525X15000606
  4. Smith, Daniel (2020). 'Cultural group selection and human cooperation: a conceptual and empirical review.' Evolutionary Human Sciences 2, e2. DOI: 10.1017/ehs.2020.2. https://doi.org/10.1017/ehs.2020.2
  5. Handley, Carla; Mathew, Sarah (2020). 'Human large-scale cooperation as a product of competition between cultural groups.' Nature Communications 11, 702. DOI: 10.1038/s41467-020-14416-8. https://doi.org/10.1038/s41467-020-14416-8
  6. Bell, Adrian V.; Richerson, Peter J.; McElreath, Richard (2009). 'Culture rather than genes provides greater scope for the evolution of large-scale human prosociality.' Proceedings of the National Academy of Sciences 106(42), 17671-17674. DOI: 10.1073/pnas.0903232106. https://doi.org/10.1073/pnas.0903232106
  7. Whitehouse, Harvey; François, Pieter; Savage, Patrick E.; Currie, Thomas E.; Feeney, Kevin C.; Cioni, Enrico; Purcell, Rosalind; Ross, Robert M.; Larson, Jennifer; Baines, John; ter Haar, Barend; Covey, Alan; Turchin, Peter (2019). 'Complex societies precede moralizing gods throughout world history' [retracted article]. Nature 568(7751), 226-229. DOI: 10.1038/s41586-019-1043-4. https://doi.org/10.1038/s41586-019-1043-4
  8. Turchin, Peter; Whitehouse, Harvey; Larson, Jennifer; Cioni, Enrico; Reddish, Jenny; Hoyer, Daniel; Savage, Patrick E.; Covey, R. Alan; Baines, John; Altaweel, Mark; Anderson, Eugene; Bol, Peter; Brandl, Eva; Carballo, David M.; Feinman, Gary; Korotayev, Andrey; Kradin, Nikolay; Levine, Jill D.; Nugent, Selin E.; Squitieri, Andrea; Wallace, Vesna; François, Pieter (2022). 'Explaining the rise of moralizing religions: a test of competing hypotheses using the Seshat Databank.' Religion, Brain & Behavior 13(2), 167-194. DOI: 10.1080/2153599X.2022.2065345. https://doi.org/10.1080/2153599X.2022.2065345
  9. Whitehouse, Harvey; François, Pieter; Savage, Patrick E.; Hoyer, Daniel; Feeney, Kevin C.; Cioni, Enrico; Purcell, Rosalind; Larson, Jennifer; Baines, John; ter Haar, Barend; Covey, Alan; Turchin, Peter (2022). 'Testing the Big Gods hypothesis with global historical data: a review and "retake".' Religion, Brain & Behavior 13(2), 124-166. DOI: 10.1080/2153599X.2022.2074085. https://doi.org/10.1080/2153599X.2022.2074085
  10. Currie, Thomas E.; Campenni, Marco; Flitton, Adam; Njagi, Tim; Ontiri, Enoch; Perret, Cedric; Walker, Lindsay (2021). 'The cultural evolution and ecology of institutions.' Philosophical Transactions of the Royal Society B 376(1828), 20200047. DOI: 10.1098/rstb.2020.0047. https://doi.org/10.1098/rstb.2020.0047
  11. Boyd, Robert; Richerson, Peter J. (2022). 'Large-scale cooperation in small-scale foraging societies.' Evolutionary Anthropology 31(4), 175-198. DOI: 10.1002/evan.21944. https://doi.org/10.1002/evan.21944
  12. Richerson, Peter J.; Boyd, Robert; Henrich, Joseph (2010). 'Gene-culture coevolution in the age of genomics.' Proceedings of the National Academy of Sciences 107 (suppl. 2), 8985-8992. DOI: 10.1073/pnas.0914631107. https://doi.org/10.1073/pnas.0914631107
  13. Laland, Kevin N.; Odling-Smee, John; Myles, Sean (2010). 'How culture shaped the human genome: bringing genetics and the human sciences together.' Nature Reviews Genetics 11(2), 137-148. DOI: 10.1038/nrg2734. https://doi.org/10.1038/nrg2734
  14. Lamba, Shakti; Mace, Ruth (2011). 'Demography and ecology drive variation in cooperation across human populations.' Proceedings of the National Academy of Sciences 108(35), 14426-14430. DOI: 10.1073/pnas.1105186108. https://doi.org/10.1073/pnas.1105186108
  15. Beheim, Bret Alexander; Bell, Adrian Viliami (2024). 'Why cultural distance can promote – or impede – group-beneficial outcomes.' Evolutionary Human Sciences 6, e14. DOI: 10.1017/ehs.2024.8. https://doi.org/10.1017/ehs.2024.8
  16. Turchin, Peter; Currie, Thomas E. (2016). 'Cultural group selection is plausible, but the predictions of its hypotheses should be tested with real-world data.' Behavioral and Brain Sciences 39, e55. DOI: 10.1017/S0140525X15000278. https://doi.org/10.1017/S0140525X15000278