SIMILARITY AND CATEGORIZATION

The present paper is a review of the literature dealing with current m odels of categorization and similarity estimation. The first part revi ews models of concept representations based on, notions such as featur es and dimensions. In the case of dimensional representations, each st imulus is defined as a point in a multidimensional space and the simil arity between two stimuli is a function of the distance between the co rresponding points in the space. In the case of feature representation s, a stimulus is defined in terms of features and the similarity betwe en two stimuli is a function of their common and distinctive features. In the second part, we analyze the construction of the feature space that subjects refer to when they have to categorize a stimulus or esti mate the similarity between two stimuli. We examine the influence of c ontext and of naive theories on the estimation of similarity. The proc ess of alignment is analyzed. We present a unified view of similarity and categorization based on the notion, of a feature space while expla ining the dissociations obtained between similarity and categorization .