Perception vs. Learning.
Taking in information from the world around us, using vision, audition, touch, or other senses, we may be unaware how universally that world is partitioned into objects. If it weren't for this fact the world might appear as a "blooming, buzzing confusion" of a myriad of more primitive features, such as colors, sounds of different frequencies, shapes, slants, and so forth (James, 1890). Almost all of the these objects are of course learned through experience. Consider the upper two visual displays in Figure l. These appear quite distinct, in meaning, complexity, and on other grounds; for example, if asked to guess, the reader would probably estimate the central pattern to contain more line segments than the upper pattern. Yet these two patterns contain exactly the same line segments, differing only in the spatial arrangement. One factor that leads to this perceptual distinction is probably learning: the fact that we interpret the upper pattern as a 'house'. On the other hand, we cannot too hastily assume that the differences in percept are due to object learning. Another factor is simply the perceptual coherence of the upper figure (related to the Gestalt properties investigated in the early part of this century--e.g see Koffka, 1935), coherence that seems to occur with little or no learning. This is illustrated by the presumably unknown object at the bottom of Figure l, an object whose phenomenology may be closer to the 'house' pattern than the random line pattern.