Les espaces conceptuels

Abstract

Conceptual spaces. Within cognitive science, there are currently two dominating approaches to the problem of modeling representations. The symbolic approach starts from the assumption that cognition is essentially involving symbol manipulation. The second approach is connectionism which models associations by artificial neuron networks. In this article, I advocate a third form of representing information that is based on using geometrical structures rather than symbols or connections between neurons. Using these structures similarity relations can be modelled in a natural way. A theory of conceptual spaces is outlined. Such a space is built up from geometrical representations based on a number of quality dimensions which are often derived from perceptual mechanisms. It will be outlined how conceptual spaces can represent various kind of information and how they can be used to describe cognitive processes like concept formation and induction. A "geometric" model of concept formation will be proposed and, among other things, its relation to prototype theory will be discussed.