PERCEIVING ANIMACY AND AROUSAL IN TRANSFORMED DISPLAYS OF HUMAN INTERACTION

When viewing a moving abstract stimulus, people tend to attribute social meaning and purpose to the movement. The classic work of Heider and Simmel [1] investigated how observers would describe movement of simple geometric shapes (circle, triangles, and a square) around a screen. A high proportion of participants reported seeing some form of purposeful interaction between the three abstract objects and defining this interaction as a social encounter. Various papers have subsequently found similar results [2,3] and gone on to show that, as Heider and Simmel suggested, the phenomenon was due more to the relationship in space and time of the objects, rather than any particular object characteristic. The research of Tremoulet and Feldman [4] has shown that the percept of animacy may be elicited with a solitary moving object. They asked observers to rate the movement of a single dot or rectangle for whether it was under the influence of an external force, or whether it was in control of its own motion. At mid-trajectory the shape would change speed or direction, or both. They found that shapes that either changed direction greater than 25 degrees from the original trajectory, or changed speed, were judged to be "more alive" than others. Further discussion and evidence of animacy with one or two small dots can be found in Gelman, Durgin and Kaufman [5] Our aim was to further study this phenomenon by using a different method of stimulus production. Previous methods for producing displays of animate objects have relied either on handcrafted stimuli or on parametric variations of simple motion patterns. It is our aim to work towards a new automatic approach by taking actual human movements, transforming them into basic shapes, and exploring what motion properties need to be preserved to obtain animacy. Though the phenomenon of animacy has been shown for many years, using various different displays, very few specific criteria have been set on the essential characteristics of the displays. Part of this research is to try and establish what movements result in percepts of animacy, and in turn, to give further understanding of essential characteristics of human movement and social interaction. In this paper we discuss two experiments in which we examine how different transformations of an original video