What Are Nervous Systems For?

Abstract An underlying bias of contemporary cognitive science is that the brain and nervous system are in the business of carrying out computations and building representations. Gibson’s ecological approach, in contrast, is decidedly noncomputational and nonrepresentational. How, then, are we to construe the role of brain and nervous system? We consider this question against the backdrop of evidence for rich achievements in perception and action by agents without brains or nervous systems. If fundamental coordination of perception and action does not require a neural substrate, then what value is added in having one? And if the neural substrate is not in the representational–computational business, then what business is it in? We pursue answers grounded in the constraints of macroscopic, multicellular life and thermodynamics.