Item

Abstract

We investigated the role of static and dynamic object properties in a categorisation task. The question was whether dynamic properties can be used to the same extent as static properties. Sixteen participants were presented with computer-generated objects varying systematically in four attributes. The static attributes were 3-D shape and colour, and the dynamic attributes were path on a 2-D plane and motion pattern (eg rotating, swinging). In the learning phase, participants were instructed to learn two categories via feedback from a sample of six objects consisting of two prototypes, two objects with one static and two objects with one dynamic cue from the other category. In the test phase, participants had to categorise objects from a new sample of eight objects with four objects having one attribute from the other prototype and four objects with two attributes from each prototype. Overall, participants used both static and dynamic cues for categorisation, showing no bias towards one or the other. There was, however, a significant bias towards the type of action used, showing that subjects paid more attention to motion pattern than to the path. Our results are discussed with emphasis on the relative importance of static and dynamic cues in object categorisation.