The naturalistic imperative in cognitive science

grantor: University of TorontoThis thesis reviews current progress in cognitive science towards satisfying the naturalistic imperative. Careful attention is paid to distinguishing metaphysical naturalism from methodological naturalism. Following this, methodological naturalism is further clarified as the set of imperatives: analyse, formalise, and mechanise. It is argued that simulation on its own is insufficient for satisfying these imperatives. Instead it must be supplemented and constrained by existing psychological theory which will indicate the relevant features for comparison in any simulation that seeks to be an explanation of human cognition. Therefore, we turn to an examination of current work in cognitive psychology. Specifically, we examine areas where attempts are being made to analyse and formalise cognition in terms of more basic psychological processes. So, the thesis proceeds to examine current work on categorisation and the nature of concepts, memory, problem solving, and inference. It shows how each can be considered a basic process in terms of which the rest of cognition might be analysed and formalised. However, we also discover that in each of these areas the naturalistic imperative is being seriously frustrated by a combination of conceptual and methodological problems. It is further shown that these problems are not independent but stem from a deep problem concerning the relationship between the two central notions of representation and relevance. The thesis argues for a very tight interdependence between the two notions so that whenever one is being used theoretically the other is being presupposed. In turn, it argues that this interdependence is frustrating current attempts to naturalise cognition. What then follows is an examination of attempts to break out of this bind by independently naturalising representation or relevance. These two projects are shown to fail precisely because they cannot break the interdependence between representation and relevance. It is argued that what is behind this representation-relevance problem is a very serious problem in any attempt to naturalise the notion of relevance. Finally, there is a discussion of what follows for philosophy and cognitive science if relevance cannot be completely naturalised.Ph.D