Perception is the process and experience of gaining sensory information about the physical world. The characteristic questions that perceptual psychology poses and the methods it employs derive from its main theoretical aims and assumptions described below. CLASSICAL PERCEPTUAL THEORY In the classical approach of Hermann \THelmholtz\t, the first step was to divide sensory experience into modalities such as vision, touch, and smell, and to subdivide the modalities into elementary \Lsensation\ls from which all more complex perceptual experiences--such as those of objects and events--were presumed to be constructed. Sensations were to be explained in terms of their physiological bases (the receptor neurons) and the physical energies to which the receptors are specially adapted to respond. The Psychophysical Methods Each noticeably different sensory experience was presumed to rest on a corresponding receptor process; the psychophysical methods (see \Tpsychophysics\t) were quantitative procedures designed to measure and identify such noticeable differences. For example, the eye focuses the physical light from an object into an image on a mosaic of photosensitive receptors (the retina); these photoreceptors provide the basic sensations of light and color as responses. According to classical perceptual theory, perception of the important attributes of the world, such as the relative brightness of an object, are not sensations; rather, they are complex learned perceptions. Depth Perception Depth perception has been similarly explained. Depth perception is the experience of the third dimension of visual space. It includes perception of the distance of an object from the observer (absolute distance) and perception of the distance of objects from one another (relative distance). Since three dimensions cannot be reproduced on the two dimensions of the retina's surface, the question arises of how to explain accurate human and animal depth perception. Helpful supplementary information in the form of depth cues is one answer. Most depth cues available to the stationary eye were listed by Leonardo da Vinci, such as linear perspective, occlusion of a far object by a near one, and aerial perspective or increasing haze. Classical perceptual theory assumed that depth perception was learned from such cues. Perceived distance would result from the visual color and shade sensations associated with memory images of previous muscle-stretch and touch sensations. However, Edward L. Thorndike showed in 1899 that some animals can respond appropriately to visual depth cues even though they have had no prior visual experience, suggesting that some depth perception is innate rather than learned. Subsequent research has corroborated and extended Thorndike's findings. CONSTANCIES, ILLUSIONS, AND ORGANIZATIONAL PHENOMENA Three sets of phenomena cause difficulties for classical perceptual theory and have been responsible for most research in perception: constancies, \Tillusions\t, and organizational phenomena. Perceptions accord more often with objects' properties than with the sensory stimulation; for example, a man's perceived height remains constant even though his retinal image size changes as he approaches an observer. There are many such perceptual constancies, which usually cause one to perceive the world more correctly than would be expected from sensory stimulation (in the previous example, from the changes in retinal image). Illusions are cases in which perception accords neither with how the receptors are stimulated, nor with the characteristics of the objects themselves. For example, in brightness contrast, an object's reflectance--in fact constant--appears to change when its surroundings change. In certain geometrical illusions, the appearance of size or length of horizontal or vertical lines is drastically altered by the addition of a few lines. Whereas experience with the world might teach people to perceive things correctly, as shown by the constancies, it is less evident why experience should result in illusions. Illusions are in fact pervasive phenomena. The organizational phenomena rest on the perceptual distinction between figure and ground: when a contour gives shape to one area, the region bounded by the other side of the contour (ground) usually has no recognizable shape. In such cases, a figure may be perceived as one object or another, but not as both simultaneously. Which area becomes the figure is therefore critical to what object will be perceived. \TGestalt psychology\t opposed classical perceptual theory by considering the form (the Gestalt) of the stimulating energies to be the essential attribute. Gestaltists sought laws of organization such as the "law of good continuation," which states that people perceive the figure-ground organization that interrupts the fewest smoothly continuing lines. Such factors have not been quantitatively or objectively studied. Nevertheless many of them provide impressive demonstrations relevant to the casually observable facts of perception and (it was thought) unrelated to familiarity. Highly familiar objects can in fact be concealed in favor of quite unfamiliar shapes in apparent contradiction to classical theory. Like a melody that remains the same when transposed in key, a particular form might have the same effect on the nervous system regardless of its particular place on the sensory surface, its specific size, and so on. This confirms the Gestaltist explanation for the perceptual constancies, which differs from the classical one, but was never adequately worked out. The accounts of physiological processes to which Gestalt theory attributed its demonstrations have been thoroughly discredited, but there have been continued attempts at objective formulation of the laws of organization by later psychologists using the tools of information theory. The principle here is that one perceives the simplest organization that could be fitted to a particular pattern of stimulation. Despite their central importance to the Gestalt approach, theories of form and shape perception have not progressed far along these lines; in fact, the classical approach has come to assimilate the Gestalt demonstrations as explained in the following section of this article. One explanation of the constancies and illusions that has continued to gain support since Helmholtz is that both reflect the same processes. That is, one perceives those objects or events that would normally be responsible for the sensory stimulation received. In this way, a person's visual system acquires associations that reflect the normal structure of the physical world. For example, the perceptual system learns to take distances into account when estimating the sizes of objects. Such sophisticated inferences are surely not conscious, if in fact they are made at all, so this theory is often phrased as "unconscious inferences based on unnoticed sensations." The theory is difficult to test, because the sensations cannot be directly observed. CLASSICAL PERCEPTUAL THEORY REVISED Because the elementary experiences (sensations) in classical theory must be considered unobservable and unpredictable (as the constancies, illusions, and organizational phenomena demonstrate), Egon \TBrunswik\t restated (1956) Helmholtz's position as follows: Because of the regularities in the physical world, the light at the eye normally contains packets of cues to any property of the physical world. The correlations are usually less than perfect--that is, the cues are only probabilistic. The organism presumably learns to rely on any cue to a degree proportional to the cue's correlation with an object's attributes. In this version of classical perceptual theory, not only are the constancies and illusions examples of the same reliance on cues, but also the Gestalt phenomena are explained. The figure-ground phenomenon is considered to be an inference made by the perceptual system about which side of a line is really an object's edge, and the laws of organization to be merely cues on which those inferences are based; for instance, the law of "good continuation" reflects the extreme unlikelihood that two objects' edges, at different distances, will line up precisely in the retinal image. To be usefully specific and subject to experimental verification, this approach must be based on quantitative knowledge about the correlation between cues and the object-attributes they reflect. Such information could presumably be obtained from "ecological surveys," which in effect explain what perceptual learning has taught the perceiver. Until recently, few such ecological surveys have been undertaken. RECENT PHYSIOLOGICAL FINDINGS Recent physiological and psychophysical study of the nervous system suggests that the latter contains receptive units much more complex than a mere mosaic of photoreceptors (in vision), and that allow for a more direct perceptual theory than the classical version. Ernst \TMach\t and Ewald Hering, contemporaries of Helmholtz, made early proposals to account explicitly for at least some perception of an object's lightness, form, and distance in terms of innate sensory mechanisms. These proposals have recently gained immensely in popularity with the discovery of lateral connections between the receptors, as well as in the higher levels of the nervous system, that provide for more direct response to object properties. For example, neural networks exist that respond directly to the ratio of the light coming from some object relative to the light from its immediate surroundings. Such responses would normally remain constant with changes in illumination, because any change in lighting of both target and background would proportionally change the light that each of them sends to the eye, leaving the ratio itself intact. Furthermore, cells have been found in the retina and higher nervous systems of amphibians and mammals that respond primarily to patterns and relationships in the retinal image, not merely to physical energy. For instance, such cells respond to dark disks surrounded by bright rings, and vice versa; to edges of a particular orientation or direction of movement; and to simultaneous stimulation of corresponding points in the retinas of both eyes. It is not yet clear to what extent, and in what manner, such pattern-sensitive networks, or feature detectors, actually contribute to perception, but their existence lends plausibility to more direct theories of perception. James Jerome Gibson has proposed the most thoroughgoing of such theories, in which the properties of the scenes and events in the perceived world are direct responses to information in the light at the eye. A particular aspect of Gibson's theory--that the visual system registers differences in textual gradients between near and far surfaces and uses these impressions to comprehend depth--has been used successfully in developing computer vision systems. It is not yet known, however, whether these computer vision systems are analogous to the human vision system. Besides gradients, other environmental clues, such as motion, contour, and shape, have been shown as being important to visual perception. NEW DIRECTIONS IN RESEARCH Other lines of research that might test this general theory are in progress. Primary among these is the study of perceptual development. A fair amount of evidence exists that some animals can respond innately to depth cues, thus showing that the classical analysis of sensory processes was at least incomplete in that regard. But evidence that infants' perceptions of size remain constant in spite of changes in object distances (and despite changes in the resulting retinal image size)--which is the issue most central to this question--remains controversial. Research in perceptual development has pushed back earlier and earlier the stage at which the infant is considered perceptually competent, but the classical theory has not been finally dismissed (see also \Tinfancy\t). In the earliest classical theory, a perception of a shape was thought to consist of the memories of the eye movements that would have to be made in order to bring each point on its contour to the center of vision. Such a definition leaves out a great deal (for instance, the organizational phenomena discussed above). But it does raise an extremely important point: because one sees detail only at the fovea (a small region in the center of the retina), the eye makes successive rapid, aimed movements, called saccades, at different parts of any object or scene. With each eye movement, of course, the image of the scene shifts on the retina. The shifting retinal images are normally not noticed, a form of constancy that is often explained as a compensation for the eye movements. Much research was, and is, being done on altering the extent of the compensation through relearning (for instance, by the prolonged wearing of prism spectacles, which change the correlation between where the eye muscles point the eye and the image it receives). Just as the question of how individual successive glances are perceived has been studied, so research on brief "tachistoscopic" glimpses has studied the effects of attention, expectation, familiarity, and motivation. Words with which the viewer is more familiar, has reason to expect, or that accord with his or her interest and concerns, will be detected at briefer exposures. Explanations for these effects remain under debate. In any case, however, such research does not address the most serious problem posed by eye movements--how one uses the sequence of partial glimpses to perceive completed objects and scenes. The process of how one fills out and stores momentary glimpses is close to (and may be identical with) that of mental imagery--that is, experiences of objects not actually stimulating the sense organs. In classical theory, as noted, mental images provided the vehicle for transforming raw sensations into perceptions of the world, but research on imagery proved so unreliable that the problem was set aside for many years. Objective work on imagery has recently increased on many fronts since the 1970s. Neurologists are studying how \Lneural network\ls in the brain process perceptual information, and computer scientists continue work (with varying degrees of success) on developing computer analogs of these networks. Studies of people who have suffered brain trauma have also yielded information on perceptual processing centers in the \Tbrain\t. Finally, the field of \Tcognitive psychology\t has produced important insights into perception. Julian Hochberg Bibliography: Caws, Mary, Perspectives on Perception (1989); Gibson, James Jerome, The Senses Considered as Perceptual Systems (1966) and The Perception of the Physical World (1950); Goldstein, E. Bruce, Sensation and Perception, 3d ed. (1989); Gombrich, Ernst H., The Image and the Eye (1982); Hochberg, Julian, Perception, 2d ed. (1978); Hubel, D.H., Eye, Brain, and Vision (1988); Rock, Irvin, The Logic of Perception (1983); Wilding, J.M., Perception (1983). See also: \Tsenses\t AND \Tsensation\t; \Tsensory deprivation\t.