A Further Note on the Perception of the Motion of
Objects as Related to the Perception of Events
J. J. Gibson, Cornell University
The World Wide Web distribution of James Gibson’s “Purple Perils” is for scholarly use with the understanding that Gibson did not intend them for publication. References to these essays must cite them explicitly as unpublished manuscripts. Copies may be circulated if this statement is included on each copy.
It has been assumed by psychologists that the first thing to investigate in studying the visual perception of motion was the impression of velocity or speed (Spigel, 1965). Brown (1931), Piaget (1946), Mashour (1964), Cohen (1964), and others have contributed experiments, and Fraisse (1963) makes much of the perception of speed in considering the perception of time. Speed is space (displacement) per unit of time. But the only kind of motion in the physical world that is really characterized by speed is the linear translatory motion of a body in empty space. Rotation (spin) and orbital motion and many other familiar kinds are not so much characterized by speed as by periodicity (rate or frequency).
This preoccupation of experimenters and theorists with Newtonian linear speed and its equation v=s/t seems to me a blind alley. It is an abstraction. The “continued motion of a body in a straight line” never occurs in the real world — not even for the legendary apple which fell on Newton’s head, since that was accelerated. Even rigid mechanical motions are full of complications, higher derivatives, and discontinuities. As for elastic or viscous motions, which are much more common in the natural environment, they are even further removed from this abstraction. The textbooks of kinetics and dynamics are not a good guide to a psychophysics of perceived motion.
In any case, this preoccupation seems to involve unexamined assumptions about the relation of optical motion to physical motion, i.e. of the stimulus to the source — assumptions that are either fallacious or questionable. It is assumed that moving elements of stimulating light (the ambient array) are comparable to moving particles of water. It is assumed that two-dimensional displacements of light rays (or points in the retinal image) constitute the only stimulus information for perceiving motion, and that the perception of three-dimensional displacements must be added on this. It is therefore assumed that motions or rotations of a body within the frontal plane of the observer are easier to perceive than motions or rotations of a body in depth. Hence only frontal speed as a function of frontal displacement over time has been studied by experimenters, commonly with a window having a moving endless belt behind it.
But if we suppose that stimulus information can specify its source in the world without necessarily being like it, there is no special virtue in studying frontal motions in the world. Optical transformations of the array are not two dimensional but multidimensional. In Optical Motions and Transformations as Stimuli for Visual Perception (1957) and the film that went with this paper, I tried to make the point that optical change (the stimulus) is governed by different laws than physical change (the source). I would now suggest more boldly that alterations of an optic array caused by movement of the observer himselfare not motions at all. That is, they are not capable of being analyzed in terms of the laws of motion that underlie mechanics—distance, time, velocity, etc. The specificities that hold between optical changes and environmental events need mathematical analysis in their own right. And it will be less confusing if we think of the optical changes in an array as quite unlike the motions of bodies, particles, substances, or matter.
A prime example of this kind of optical change that carries information but that is not a “motion” is the change that accompanies the occlusion or disocclusion of one thing by another in the world. Michotte calls the resulting experience the “screening effect” and it happens at the beginning and end of his so-called “tunnel effect” (Michotte, Thines, & Crabbe, 1964). The perception is that of seeing an object “go behind” or “come from behind” (Reynolds, 1966). Another related example is the type of motion parallax that occurs when one surface that is behind another moves parallel to the occluding edge. It is analogous to a physical motion of “shearing” but any such analogy is dangerous as already noted. (This also may occur as information in the “visual cliff” experiments that have been carried out with animals.) And something similar to these changes occurs whenever a solid object is turned, thereby causing one face of the object to be occluded by another. The quadrilateral face of a turning cube, for example, vanishes into an edge and, on the other side, a new face grows from an edge.
Tentatively, these examples might all be called optical edge-transformations to distinguish them from the perspective transformations or slant transformations studied by Gibson and Gibson (1957) and by Flock (e.g. 1964), and the size transformation studied by Schiff (1965). The edge-effects are transformations only in a general mathematical sense of the term; they are not literally conversions of one form into another any more than they are motions. The fact that light travels in straight lines and that ambient light consists of a sheaf of mutually intersecting rays at a point means that occlusion and change of occlusion are an inescapable feature of vision in a natural environment. All seeing animals, not just intellectual man, are faced with the fact of covering, superposition, screening, hiding, losing sight of, disappearance from view, going behind, and their opposites. (All these optical changes are mathematically reversible.) The animal sees his prey go around a corner; the child sees his companion try to hide. Animals and children hide themselves from the view of another. They also hide valuable things in holes and boxes, and they cover things up and then uncover them. As creatures move about they are accustomed to seeing one vista “open up” from another by disocclusion, as when one tops a rise in the land and begins to see the valley beyond. So the philosophical problem of the phenomenal persistenceof objects despite the disappearance of their sensations, so bothersome to Hume and others, misses the real problem entirely. The basic problem is how the visual system of an individual responds to edge-transformations in his field of view and how perception and behavior conform to the fact of occlusion or adapt to it.
The perception of occluded things self-evidently does not depend on corresponding patches of color in the visual field, that is, on visual sensations. But this perception presumably does depend on the pickup of optical information for occluded things, that is, information over time, or information contained in certain kinds of change. Compare this with the perception of distant things — the ancient puzzle of depth perception. In the latter case the perception might depend on corresponding color-patches or sensations, for at least they exist in the visual field. Theories of depth perception (size constancy) aim to explain how “little” sensations get converted into “big” perceptions. But I argue that the optical information for the perception of distance has nothing to do with the presence of visual sensations of distant object, just as I argue that the optical information for the perception of occluded objects has nothing to do with the absence of visual sensations for occluded objects.
An object that is occluded, like an object that becomes sufficiently far away, is loosely said to “disappear” or “vanish”. But note that the two kinds of optical change are quite different and are easily distinguishable. We do not confuse “going behind” and “going away”. Note also that the two kinds of change, considered mathematically as transformations, are reversible, just as the physical motions of the objects are reversible. Objects reappear from behind, or from “the distance.” These modes of optical disappearance are to be contrasted with the disappearance that goes with the onset of darkness (also reversible). And all these are not to be confused with still another group of disappearances, the physical counterparts of which are note reversible. A solid, light-reflecting object may disappear (vanish, become invisible) by oxidation, solution, disintegration, vaporization, sublimation, digestion, and other physio-chemical processes. Burning, dissolving, crumbling, and exploding are examples. In accordance with the law of entropy, the original objects are never reconstituted. One can imagine the materialization of a vaporized object (or display it by reversing a motion picture film — of sequences by Gibson and Reynolds) but one never sees it in the environment.
Optical wiping and shrinking of certain sorts specify the existence of objects. Optical disintegration of certain sorts specifies the non-existence of objects. It would be interesting to know how early children begin to distinguish between disappearance from sight and disappearance by dissolution (and what species of animals can do so) as evidenced by appropriate behavior. The optical information is available. It is very doubtful that children begin by perceiving impermanent objects (supposedly on the basis of their impermanent sensations) and then construct a concept of permanent objects intellectually. The sensations have nothing to do with the development of perception; the information available has everything to do with it.
Brown, J. F. The thresholds for visual movement. Psychol. Forsch., 1931, 14, 249-268.
Cohen, R. L. Problems in motion perception. Uppsala, Sweden. 1964.
Flock, H. R. A possible optical basis for monocular slant perception. Psychol. Rev., 1964, 71, 340-396.
Fraisse, P. The psychology of time (Tr. by Leith), Harper and Row, 1963.Gibson, J. J. Optical motions and transformations as stimuli for visual perception, Psychol. Rev., 1957, 64, 288-295.
Gibson, J. J. and Gibson, E. J. Continuous perceptive transformations and the perception of rigid motion. J. Exp. Psychol., 1957, 54, 129-138.
Mashour, M. Psychophysical relations in the perception of velocity. Stockholm, 1964.
Michotte, A., Thines, G., and Crabbe, G. Amodal completion and perceptual organization (Tr.), Studia Psychologica, Louvein, 1964.
Piaget, J. Les notions de mouvement et la vitesse chez l’enfant. Presses Univ. de France, 1964.
Reynolds, H. Doctoral thesis, Cornell, 1966.
Schiff, W. Perception of impending collision, Psychol. Monog., 1965, 79, #604.
Spigel, I. M. Readings in the study of visually perceived movement. Harper and Row, 1965.