The relationship between analogy and categorisation in cognition

Abstract

This central topic of this thesis is the relationship between categorisation and analogy in cognition. Questions of what a straightforward representation of a concept or category is, and following from that how extra-categorical associations such as analogy and metaphor are possible are central to our understanding of human reasoning and comprehension. However, despite the intimate linkage between the two, the trend in cognitive science has been to treat analogy and categorisation as separable, distinctive phenomena that can be studied in isolation from one another. This strategy has proved remarkably effective when it comes to the cognitive modelling of extracategorical associations. A number of compelling and detailed models of analogy process exist, and there is widespread agreement amongst researchers studying analogy as to what the key cognitive processes that determine analogies are.However, these models of analogy tend to assume some kind of fully specified category processing module which governs and determines ordinary, straightforward conceptual mappings. Indeed, this assumption is required in order to talk about analogy and metaphor in the first place: few theorists actually define analogy and metaphor per se, but all agree that analogical and metaphoric judgements can be defined in contrast to ordinary categorisation judgements.This thesis reviews these models of analogy, and evidence for them, before conducting a detailed exploration of categorisation in relation to analogy. A theoretical and empirical review is presented in order to show that the straightforward notion of categorisation that underpins the distinctive phenomena approach to the study of analogy and categorisation is more apparent than real. Whilst intuitively, analogy and categorisation might feel like different things which can be contrasted with one another, from a cognitive processing point of view, this thesis argues that such a distinction may not survive a detailed scientific examination.A series of empirical studies are presented in order to further explore the 'no distinction' hypothesis. Following from these, further studies examine the question of whether models of analogical processing have progressed as far as they can in artificial isolation from categorisation, a process in which the processes that are normally deemed 'analogical' appear to play a vital role.The conclusion drawn in this thesis is that the analogy / categorisation division, as currently formulated, cannot survive detailed scientific examination. It is argued that despite the benefits that the previous study of these phenomena in isolation have brought in the past, future progress, especially in the development of cognitive models of analogy, is dependent on a more unified approach.