Not so Basic Instinct: Learning, Evolution, and the Behavioural Hypercycle Learning and Instinct Coevolution

Traditional ideas about cognitive evolution often treat learning and instinct as zero-sum substitutes, where increased plasticity necessarily displaces hard-wired routines. We challenge this "substitution-only" view using economic concepts and by considering evolutionary implications of “behavioural hypercycles”: i.e., teams of modular component behaviours coordinated toward a functional task. This shows that, while intermediate brain sizes can indeed favour the substitution of instinct for flexible learning, larger nervous systems trigger a sort of "Income Effect" which changes the optimal allocation between how growing neural resources are dedicated between learning and instinct. Rather than displacing one another, sophisticated learning and extensive instinctive repertoires can evolve as adaptive complements under identifiable conditions. We further show that this trade-off is likely level-specific: evolution preferentially canalizes reusable, high-burden primitives (i.e., genetically assimilating behavioural components that are especially costly or difficult to learn) while leaving task-specific links plastic (e.g., subject to goal-directed control). Our analysis suggests that hypercyclic organization is a fundamental principle of complex agency, where instinct provides the reliable scaffold that makes sophisticated learning affordable. Our model is consistent with several lines of evidence (behavioural, genetical, neurological), likely applies broadly to any animals capable of complex behaviour, and points to a range of empirically testable predictions.