Humans often learn preferentially from ingroup members who share a social identity affiliation, while ignoring or rejecting information when it comes from someone perceived to be from an outgroup. This sort of bias has well-known negative consequences – exacerbating cultural divides, polarization, and conflict – while reducing the information available to learners. Why does it persist? Using evolutionary simulations, we demonstrate that similarity-biased social learning (also called parochial social learning) is adaptive when (1) individual learning is error-prone and (2) sufficient diversity inhibits the efficacy of social learning that ignores identity signals, as long as (3) those signals are sufficiently reliable indicators of adaptive behaviour. We further show that our results are robust to considerations of other social learning strategies, focusing on conformist and pay-off-biased transmission. We conclude by discussing the consequences of our analyses for understanding diversity in the modern world.

It is commonly argued that humans have generalised predispositions for within-group favouritism and between-group animus (i.e. that humans are parochially altruistic), leading to higher levels of internal conflict in societies with greater diversity. Other research, however, has questioned both the ubiquity of parochial altruism and the role of diversity per se in causing social discord. Here, we use ethnographic, social network and experimental economic game data to explore this topic in two multi-ethnic Colombian communities. We examine the extent to which Afrocolombian and Emberá residents express parochial altruism, finding appreciable variability between communities, and across individuals within communities. When present, parochial altruism appears to be driven by divergent perceptions of group-based economic need, not group identity per se. Our results suggest that diversity may be less likely to cause social discord than past work has suggested, as long as group-based inequalities in wealth, well-being and representation – that can destabilise positive inter-group relationships – are minimised.

Striking a balance between individual and social learning is one of the key capabilities that support adaptation under uncertainty. Although intergenerational transmission of information is ubiquitous, little is known about when and how newcomers switch from learning loyally from preceding models to exploring independently. Using a behavioural experiment, we investigated how social information available from a preceding demonstrator affects the timing of becoming independent and individual performance thereafter. Participants worked on a 30-armed bandit task for 100 trials. For the first 15 trials, participants simply observed the choices of a demonstrator who had accumulated more knowledge about the environment and passively received rewards from the demonstrator's choices. Thereafter, participants could switch to making independent choices at any time. We had three conditions differing in the social information available from the demonstrator: choice only, reward only or both. Results showed that both participants’ strategies about when to stop observational learning and their behavioural patterns after independence depended on the available social information. Participants generally failed to make the best use of previously observed social information in their subsequent independent choices, suggesting the importance of direct communication beyond passive observation for better intergenerational transmission under uncertainty. Implications for cultural evolution are discussed.