Visual perception of intentional motion.

Abstract

A series of experiments were performed to investigate how motion sequences provide information about the intentional structure of moving figures or actors. Observers had to detect simulations of biologically meaningful motion within a set of moving letters. In the first two experiments a factorial design was used, with type of instruction as a between-subject factor and six movement parameters (number of items, speed and directness of target and distractors, and 'relentlessness' of target movement) as within-subject factor; in the final two experiments, the visibility of the goal towards which the target moved and the use of a tracking movement to distinguish the target were varied. In such displays search time increases with increasing number of stimuli. It was found that (a) the more direct the motion, the more likely it was to be interpreted as intentional; (b) intentional motion was much easier to detect when the target moved faster than the distractors than when it moved more slowly; (c) recognition of intentionality was impaired but not abolished if the goal towards which the target was moving was invisible; and (d) participants did not report intentional movement when the target was distinguished by brightness rather than the manner in which it moved. We argue that the perception of intentionality is strongly related to observers' use of conceptual knowledge, which in turn is activated by particular combinations of features. This supports a process model, in which intentionality is seen as the result of a conceptual integration of objective visual features.