Putting two and two together – a model of learning and memory in cultural evolution

Abstract

Abstract To what extent do the cognitive capacities of learners shape the dynamics of cultural evolution and gene-culture coevolution? To begin to address some of these questions, an individual-based model was built, incorporating the cognitive dynamics of working memory, long-term memory and learning strategies—imitation and learning by induction. By representing cultural values as hierarchically arranged sequences of unit token values, this work showed how the emergence of complex cultural representations depends on chunking, becomes constrained by working memory and is spatially limited by learning distance. Cultural token variants diverged into spatially clustered lineages with agent similarity scaling relative to distance, demonstrating the emergence of cultural systems built upon shared cultural variants. Moreover, by simulating a process of inductive learning, agents converged upon certain cultural primitives and complex tokens, thereby marrying aspects of cultural attractor theory with cumulative cultural evolution. The model offers some computational insight into the roles of imitation and inductive learning across the lifespan, showing that imitation occurred most in the first iterations of an agent’s lifespan and inductive inference throughout the adult lifespan, thereby shedding light on the proximal effects of cognitive capacities in the debate between replication and reconstruction of cultural values.