NEWBORN AND OLDER INFANTS PERCEPTION OF PARTLY OCCLUDED OBJECTS

In experiments described by Kellman and Spelke (1983) and Kellman, Spelke, and Short (1986), 4-month-old infants were habituated to a stimulus (usually a rod) which moved behind a central occluder, so that only the top and bottom of the rod was visible. Subsequently, the infants increased responding to a stimulus consisting of two object pieces with a gap where the occluding block had been but did not respond to a continuous rod, suggesting that during the habituation trials, they had been perceiving a connected object, that is, they were "filling in" the unseen portion. The present article describes five experiments which were designed to see if perception of object unity is present at birth. In Experiments 1, 2, and 5, moving, occluded displays were shown to newborn infants. In Experiment 1, the familiarized stimulus was an outline square which underwent translatory motion behind a stationary occluder, and in Experiments 2 and 5, the familiarized stimulus was a rod which moved back and forth behind a stationary occluder. In all three experiments, the newborns subsequently gave a strong preference for a continuous, rather than a broken, stimulus. Experiments 3 and 4 showed, respectively, that newborns perceive both moving and stationary parts of a moving, occluded display, and that the preferences found in Experiments 1, 2, and 5 are best interpreted as novelty, rather than familiarity, preferences. In striking contrast to the above, in Experiment 5, 4-month-old infants, tested under the same conditions as the newborn infants, gave a strong novelty preference for two object pieces rather than a continuous stimulus, a finding which replicates the results of Kellman et al. (1986). These findings argue against the view that infants begin life with a knowledge of the unity and coherence of objects and suggest that infants' understanding of objects changes in the early months of life. Unlike 4-month-olds, newborns appear to perceive only that which is immediately visible, and they seem to be unable to make perceptual inferences from visual input. © 1990.