Evolved priors for ethnolinguistic categorization: A case study from the Quechua–Aymara boundary in the Peruvian Altiplano

Ethnic categories uniquely structure human social worlds. People readily form stereotypes about these, and other social categories, but it is unclear whether certain dimensions are privileged for making predictions about strangers when information is limited. If humans have been living in culturally-structured groups for much of their evolutionary history, we might expect them to have adaptations for prioritizing ethno-linguistic cues as a basis for making predictions about others. We provide a strong test of this possibility through a series of studies in a field context along the Quechua–Aymara linguistic boundary in the Peruvian Altiplano where the language boundary is not particularly socially meaningful. We find evidence of such psychological priors among children and adults at this site by showing that their age, and the social categories’ novelty affect participants’ reliance on ethno-linguistic inductive inferences (i.e. one-to-many predictions). Studies 1–3 show that participants make more ethno-linguistic inferences when the social categories are more removed from their real-world context. Additionally, in Study 4 when the category is marked with acoustic cues of language use, young children rely heavily on ethno-linguistic predictions, even though adults do not.