The University of Sussex

Scientific explanations of behaviour: the approach through formal task-description

Horst Hendriks-Jansen

In this paper and its companion paper, I contrast two different approaches to scientific explanations of human behaviour. The approach I shall examine here covers all of Cognitive Science with the exception of some very recent work with autonomous robots. It takes its cue from Turing's [1936] strategy for solving Hilbert's Third Problem. Formal task-description and fucntional analysis are assumed to be the best means of identifying fruitful explanatory kinds for a psychological theory. I discuss a number of variations on this approach, including Newell's and Simon's claims for the scientific legitimacy of information processing models of the mind, Marr's claim for computational solutions based on transformations of canonically defined internal representations, and competence theories like Chomsky's generative grammar. I argue that none of these theories have succeeded in discovering valid natural kinds for an explanation of human behaviour, and that the approach to scientific explanation on which they are based is incapable of discovering such kinds. I conclude by suggesting an alternative approach, which I shall develop further in the companion to this paper.

This paper is not available online