Infants Attribute Value± to the Goal-Directed Actions of Self-propelled Objects.

Motion is a fundamental source of information for basic human interpretations; it is basic to the fundamental concept of causality and, the present model argues, equally basic to the fundamental concept of intentionality. The model is based on two main assumptions: When an infant perceives an object (1) moving spontaneously and (2) displaying goaldirected action, it will interpret the object as intentional and assign to it the unique properties of the psychological domain. The key property tested was: Do infants attribute value to interactions between intentional objects using criteria specified by the model? We showed infants (average age 52 weeks) computer-generated animations of spontaneously moving "balls," using looking time in a standard habituation/dishabituation paradigm. In two positive interactions, one ball either "caressed" another, or "helped" it achieve its goal; whereas in two negative interactions, one ball either "hit" another, or "prevented" it from achieving its goal. In keeping with predictions of the model, when transferred to a negative condition, infants who had been habituated on a positive condition showed greater dishabituation than those habituated on a negative condition. The results could not be easily explained by the similarity relations among the animations depicting the interactions. The results suggest that well before the age when the child can ascribe mental states or has a "theory of mind," it recognizes the goals of self-propelled objects and attributes value to the interactions between them.