position that implies it must have moved through an intervening, but occluded, solid platform (Spelke, 1991). Longer looking by infants is widely accepted as an indication that they have seen an event that violates their expectations.

There is also, not surprisingly, considerable learning about the physical world during infancy. For example, younger infants are sensitive only to large and obvious conflicts between the barrier and obscured block and will not notice smaller discrepancies, such as when the barrier stops 30 degrees too early (Baillargeon, 1995). Thus, they need to learn to calibrate the geometry of physical events with their consequences. Similarly, an appreciation that an unsupported object will fall down takes time to develop (Spelke and Kyeong, 1992). Other researchers argue that competing ways of thinking about objects during infancy are resolved only gradually, over a period of several months (Munakata et al., 1997). Regardless of the details of how quickly infants gain a single clear view of the nature of the physical world, there is substantial agreement that, by the end of the first year of life, they have expectations about objects that fit with many principles governing the behaviors of bounded physical objects.

By 12 months of age, infants are capable of taking into account physical dimensions and magnitudes and their consequences for events. As noted earlier, they consider the angle of movement of a swinging barrier as it relates to the size of an object behind it. They also understand that bigger objects in motion are likely to have bigger consequences. For example, when 11-month-olds observe a cylinder roll down a ramp and move another object through collision, they infer that a larger novel cylinder will move the object more and that a smaller cylinder will move it less (Kotovsky and Baillargeon, 1994). More broadly, while starting early with some very general ideas about solidity and spatiotemporal continuity (Spelke et al., 1992, 1995), infants are constantly refining those ideas into forms that enable more subtle inferences about objects and their behaviors over intervals of time and space (Baillargeon, 2004). By the end of the first year, infants also have a clear sense of causation as opposed to mere correlation or contiguity. Thus, even if one event reliably occurs before another one, infants may not infer causation unless there is also some degree of plausible mechanism, such as one object launching another through collision (Leslie and Keeble, 1987). They can also make inferences about unreasonable versus reasonable hidden causes of the motion of an inanimate object (Saxe, Tenenbaum, and Carey, 2005).

Ongoing research is now asking how a 1-year-old’s mental representations of the world should be best characterized. Should her ability to anticipate the behaviors of physical objects be seen as her having beliefs like those of an adult, or could it reflect mental processes that are less explicit and belief-like in nature? (See Leslie, 1994, for further discussion of these



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement