Semantic knowledge guides innovation and drives cultural evolution

Preprint: Semantic knowledge guides innovation and drives cultural evolution Anil Yaman*, Shen Tian*, Björn Lindström. https://arxiv.org/abs/2510.12837. (*equal contribution)

Abstract

Cultural evolution allows ideas and technology to build over generations, a process reaching its most complex and open-ended form in humans. While social learning enables the transmission of such innovations, the cognitive processes that generate innovations remain unclear. We propose that semantic knowledge-the associations linking concepts to their properties and functions-guides human innovation and drives cumulative culture. To test this, we combined an agent-based model, which examines how semantic knowledge shapes cultural evolutionary dynamics, with a large-scale behavioural experiment (N = 1,243) testing its role in human innovation. Semantic knowledge directed exploration toward meaningful solutions and interacted synergistically with social learning to amplify innovation and cultural evolution. Participants lacking access to semantic knowledge performed no better than chance, even when social information was available, and relied on shallow exploration strategies for innovation. Together, these findings indicate that semantic knowledge is a key cognitive process enabling human cumulative culture.

Figure: Semantic knowledge guides human innovation. (a) Experimental task. The semantic and non-semantic conditions had the same innovation rules, but items were depicted with meaningful images in the semantic condition and abstract symbols in the non-semantic condition. (b) Participants generated larger cultural repertoires (total unique innovations per group) when semantic knowledge was available. Without semantic knowledge, human performance was comparable to random bots. (c) The difference between semantic and non-semantic conditions was moderated by the number of social learning attempts in the group, where more social learning resulted in larger differences between conditions. Bots had a fixed social learning probability and were therefore excluded from this analysis. (d) Individual participants also generated more innovations when semantic knowledge was available and social learning was possible. Dots and lines represent predictions from negative binomial regression models, and error bars and bands indicate model-derived 95% confidence intervals. *** p < .001, ** p < 0.01, n.s. p > 0.05.