Similarity and categorisation: getting dissociations in perspective

Dissociations between similarity and categorization have constituted critical counter-evidence to the view that categorization is similarity-based. However, there have been difficulties in replicating such dissociations. This paper reports three experiments. The first provides evidence of a double dissociation between similarity and categorization. The second and third show that by asking participants to make their judgments from particular perspectives, this dissociation disappears or is much reduced. It is argued that these data support a perspectival view of concepts, in which categorization is similarity-based, but where the dimensions used to make similarity and categorization judgments are partially fixed by perspective.