The role of continuity cues in the perception of causality by adults and infant

In order to evaluate the claim that causality is perceived directly, the role of continuity cues in causal perception was explored. Two views posit that spatial contiguity, temporal contiguity, and similarity of cause and effect are necessary for the direct perception of causality (Michotte, 1963; White, 1988). Four studies investigated the necessity of similarity and observable spatial and temporal contiguity for causal perception.;In Study 1, adults judged the causality of launching events, events in which one object approaches and makes contact with a second object, which then moves. Adults judged events characterized by spatial and temporal contiguity as causal and those characterized by violations in spatial or temporal contiguity as non-causal, whether or not the objects moved along similar paths. Adults also judged the causality of occluded launching events, in which the spatial and temporal contiguity could not be observed and must be inferred. Adults were sensitive to temporal differences among these events, and judged some as causal and others as non-causal.;Studies 2 and 3 explored infant perception of causality and similarity of movement. In Study 2, 10- and 12-month-old infants responded to changes in the similarity of movement and to changes in contiguity; when familiarized with a single event, they treated as novel events characterized by either a change in similarity or in contiguity. In Study 3, 10-month-old infants responded to events characterized by dissimilar movement as causal and non-causal; infants familiarized with causal events treated non-causal events as novel, and infants familiarized with non-causal events treated a causal event as novel.;Study 4 explored perception of occluded events by 10-month-old infants. Infants, unlike adults, were not sensitive to the temporal differences among the events. After familiarization with any occluded launching event, either causal or non-causal, infants responded to non-causal launching events as more novel than a causal launching event. Thus, it appears that while perceivers as young as 10 months infer that an occluded event is causal, the ability to use subtle temporal differences as a cue to causality emerges with development.