Embodied cognitive evolution and the limits of convergence.

Comparative psychology seems to be perpetually bogged down in intractable debates about which species have what cognitive capacities, which criteria to use and whether or not the capacities are domain general. The problem arises from lack of conceptual clarity about how to define, measure and compare cognitive capacities. In turn, conceptual vagueness arises from the use of anthropocentric folk-psychological concepts given apparent scientific legitimacy by framing them in cognitivist, computational terms. This 'cognitivist gambit' assumes that cognitive processes necessarily involve representations that are independent of the sensory-motor specializations associated with different body plans and ecological niches. We argue instead that sensory-motor adaptations are not inconvenient confounding variables that should be controlled to isolate cognition, but intrinsic aspects of cognitive evolution. This perspective implies that, because bodies and their sensory-motor control are highly divergent across the tree of life, comparative psychology should pay more attention to phylogenetic constraint and divergent cognitive evolution. It also implies that boiling down neuro-cognitive evolution to brain size or numbers of neurons will fail to capture the richness and complexity of the interrelationships between nervous systems, cognition, behaviour and ecology. If correct, this perspective suggests a need to reconsider the ontological basis of comparative psychology.This article is part of the Theo Murphy meeting issue 'Selection shapes diverse animal minds'.