Perception and understanding of effects of gravity and inertia on object motion

Experiments using a preferential looking method, a perceptual judgment method, and a predictive judgment method investigated the development, from 7 months to 6 years of age, of sensitivity to the effects of gravity and inertia on inanimate object motion. The experiments focused on a situation in which a ball rolled off a flat surface and either continued in linear motion (contrary to gravity), turned abruptly and moved downward (contrary to inertia), or underwent natural, parabolic motion. When children viewed the three fully visible motions, both the preferential looking method and the perceptual judgment method provided evidence that sensitivity to inertia developed between 7 months and 2 years, and that sensitivity to gravity began to develop after 3 years. When children predicted the future location of the object without viewing the motions, the predictive judgment method provided evidence that sensitivity to gravity had developed by 2 years, whereas sensitivity to inertia began to develop only at 5±6 years. These findings suggest that knowledge of object motion develops slowly over childhood, in a piecemeal fashion. Moreover, the same system of knowledge appears to be tapped both in preferential looking tasks and in judgment tasks when children view fully visible events, but a different system may underlie children's inferences about unseen object motions. Human adults are sensitive to a variety of effects of gravity and inertia on the motions of objects. In particular, a hand-held object that is released in midair looks natural only if it begins to move downward, an object that falls freely looks natural only if it undergoes appropriate acceleration (Shanon, 1976), and an object that rolls off a cliff looks natural only if it moves downward on a parabolic path (Kaiser, Proffitt & McCloskey, 1985). What are the origins of this sensitivity? Gravity and inertia have constrained the motions of objects throughout the history of the earth, and humans and other animals have evolved a variety of sensory and motor mechanisms that take account of their effects (Howard, 1982; Schone, 1984). It is therefore possible that humans have also evolved perceptual and cognitive mechanisms that are sensitive to effects of gravity and inertia. Alternatively, human adults have a lifetime of experience observing objects, and they may have learned about natural object motions. One goal of the present research is to investigate these contrasting possibilities through studies of infants and children. Despite their sensitivity to the naturalness of perceived physical events, adults are prone to error if they must infer the path or acceleration of a moving object that is hidden (Shanon, 1976; McCloskey, 1983). Moreover, adults often give mistaken explanations for the motions of objects and make erroneous predictions about future object motions (e.g. Piaget, 1976; Clement, 1982). These observations suggest that the tacit conceptions underlying adults' perception of object motion are distinct from the explicit conceptions that underlie their predictions, judgments and explanations. What are the origins of the gap between implicit and explicit knowledge of object motion? Studies of schoolaged children provide evidence that the explicit knowledge guiding judgments about object motion sometimes differs from the implicit knowledge guiding actions on objects in children as young as 5 years. Indeed, the same child may act so as to propel an object correctly, and # Blackwell Publishers Ltd. 1999, 108 Cowley Road, Oxford OX4 1JF, UK and 350 Main Street, Malden, MA 02148, USA. Address for correspondence: Elizabeth S. Spelke, E10-246, Department of Brain and Cognitive Sciences, MIT, Cambridge, MA 02139, USA. Copyright © 2000 All Rights Reserved then report erroneously on the conditions that guided her actions (Piaget, 1976; Krist, Fieberg & Wilkening, 1993). Some previous research with infants also suggests a discrepancy between the knowledge that guides actions on visible objects and the knowledge that guides inferences about hidden objects. When infants reach for a continuously visible moving object, their reaches are `predictive' i.e. aimed ahead of the object's currently visible position, and guided by inertia (von Hofsten, Vishton, Spelke, Rosander & Feng, 1998). In contrast, when infants view an object that moves behind an occluder, their looking preferences between events in which the object reappears at different positions suggests no sensitivity to inertia (Spelke, Katz, Purcell, Ehrlich & Breinlinger, 1994). These findings suggest that the gap between action and judgment extends back to infancy, but they are not decisive for two reasons. First, it is unclear whether preferential looking tasks tap the same kind of knowledge as the verbal judgment tasks given to adults (see Bertenthal, 1996, and Spelke, Breinlinger, Macomber & Jacobson, 1992, for discussion). Second, the studies assessing infants' predictive reaching for fully visible objects used somewhat different events and presentation conditions than those assessing infants' preferential looking at partly occluded objects, and so they cannot be compared directly. A second goal of the present research is to explore the possible divergence between preferential looking methods and verbal judgment methods, as well as the possible divergence between perceptions of and judgments about constraints on object motion, through systematic comparisons of the early development of sensitivity to object motion across different kinds of tasks. A third goal of this research is to investigate the nature of human knowledge of gravity and inertia. Diverse conceptions of object motion have been expressed in the history of science (Duhem, 1954; Kuhn, 1970, 1977) and by contemporary science students (Champagne, Klopher & Anderson, 1980; Clement, 1982). In the history of physics, theories of force, acceleration and velocity have been subject to continuous innovation. Aristotelians, impetus theorists, and classical and relativistic physicists have offered different explanations for object motion and different descriptions of how objects move under particular conditions. Studies of contemporary college students have been interpreted by some investigators as suggesting that students reason intuitively as impetus theorists did (McCloskey, 1983). Other investigators, however, have proposed that students' reasoning is based on piecemeal knowledge rather than on any general conceptions of object motion (diSessa, 1983). The present research aims to shed light on this issue by investigating the development of sensitivity to effects of gravity and inertia on the path of motion of an object that moves off a supporting surface, and by comparing children's performance in this situation to their performance in a situation studied previously, in which an object moved on an inclined, supporting surface with either appropriate or inappropriate acceleration (Kim & Spelke, 1992). If a single conception of gravity underlies infants' reactions to moving objects in both situations, then infants should become sensitive to the natural path of motion for an unsupported object at the same age at which they were found to become sensitive to the natural acceleration of a partially supported object: between 5 and 7 months of age (Kim & Spelke, 1992). If piecemeal knowledge underlies humans' commonsense understanding of effects of gravity and inertia, in contrast, then understanding may emerge at different times in these different situations. This research focused on the development of sensitivity to gravity and inertia in one situation. If a ball rolls down a ramp and then off its edge, it continues to move forward while also moving downward at a steadily accelerating speed. The forward and downward motions combine to form a parabolic path. Experiments 1±5 investigated 7-month-old infants' sensitivity to this effect of gravity and inertia in fully visible events, using Kim and Spelke's (1992) preferential looking method. The findings of all these experiments were negative, providing evidence that the sensitivity to gravity and inertia shown in Kim and Spelke's studies reflected limited, piecemeal knowledge of objects. Experiments 6±8 next investigated sensitivity to the same events at 2 years of age, providing evidence for emerging sensitivity to inertia but not gravity. Experiments 9±12 investigated how this sensitivity develops in 3to 6-year-old children, both with the preferential looking method and with a verbal judgment method assessing children's perception of the naturalness of observed object motions. The two methods provided evidence for the same developmental changes in sensitivity to inertia and gravity, suggesting slow piecemeal development of a single system of perceptual knowledge. Finally, Experiment 13 investigated the development of sensitivity to inertia and gravity, in 2to 6-year-old children, by means of a different verbal judgment task assessing children's predictions about the future position of an object whose motion they have not seen. This last experiment provided evidence for a different developmental sequence, suggesting that the divergence between perception and judgment found in adults begins early in development. 340 In-Kyeong Kim and Elizabeth S. Spelke # Blackwell Publishers Ltd. 1999 Copyright © 2000 All Rights Reserved Sensitivity to gravity and inertia in 7-month-old infants