The Evolution of Brains as Predictive Models

Richard Dawkins offered a deceptively simple argument for why nervous systems evolved the capacity for Mental simulation: overt trial takes time and energy; overt error is often fatal. An organism capable of running internal forward models — testing candidate behaviors against a representation of the world before executing them — gains a decisive adaptive edge. The brain, on this account, is not primarily a stimulus-response device but a prospective organ, one whose core function is anticipatory rather than reactive.

This framing converges powerfully with Karl Friston's Free energy principle and the framework of Active inference. Here, cognition is formalized as the continuous minimization of surprise — or more precisely, the minimization of variational free energy, a measure of the discrepancy between an organism's generative model of the world and its incoming sensory signals. The brain actively generates predictions at every level of its hierarchy and propagates Prediction errors upward when those predictions fail. Crucially, behavior itself becomes a form of inference: the organism acts not merely to achieve goals but to confirm predictions or to sample the information needed to refine its model.

The implications extend beyond perception and action into a theory of life itself. Maintaining biological organization — remaining far from thermodynamic equilibrium — requires an accurate enough world-model to navigate an environment full of uncertainty. Prediction, on this view, is not a cognitive luxury but a thermodynamic necessity, the mechanism by which living systems resist dissolution.