Abstract
Human observers, including infants, perceive unitary objects by analyzing both motion and static information (e.g., configuration). When aligned edges undergo common motion, infants tend to perceive them as unified, but when aligned edges are stationary, or misaligned edges move together, infants' unity percepts are abolished. The best account of how motion and alignment interact in early object perception is in dispute. One possibility is that early perceptual organization follows a limited set of principles (e.g., edges moving in tandem are always interpreted as connected). On this “top-down” account, misalignment has an indirect impact on unity by reducing motion sensitivity. A second possibility is that object percepts arise from lower-order perceptual skills and experience. On this “bottom-up” account, motion and static information have direct, independent inputs to unity percepts. To address this controversy, 2-month-old infants viewed displays in which edges protruding from above and below an occluder underwent in-phase or out-of-phase motion. Edges were either aligned or misaligned when moving in-phase. Following habituation, the infants saw an identical display, alternating with the opposite-motion display, and looked longer at the novel motion regardless of edge orientation. In a companion experiment, 2-month-olds were probed for unity percepts in displays with aligned or misaligned edges; only infants in the former group perceived unity. These experiments lead to three conclusions. First, configuration has no bearing on motion sensitivity: Infants discriminated in-phase and out-of-phase motion regardless of edge alignment. Second, performance in the first experiment cannot be explained on the basis of responses to unity, because infants in the second experiment perceived unity only when edges were aligned. Third, infants' capacity to discriminate motion and configuration exceeds their capacity to perceive unitary objects, in accord with a bottom-up view.
Supported by NSF grant BCS-0094814