On the Difference between Perception and Proprioception
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 References to these essays must cite them explicitly as unpublished manuscripts.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.
Simplifying a distinction made by Sherrington, I propose to use the term perception for an experience of the environment surrounding the body of an animal and the term proprioception for an experience of the body itself (Boring’s somaesthesis, which includes Sherrington’s interoception). Now it is a fact that an environment implies something that is surrounded, and therefore an awareness of the environment implies an awareness of the body existing in the environment. Equally, an awareness of the body entails some feeling of its relation to the surroundings. So there can be no perception without an implied proprioception, nor can there be any proprioception without some awareness, however dim, of the environment in which the body exists. Attention can be directed almost exclusively to either the body or the world but the other is always detectable.
Both perception and proprioception depend on stimulation of the nervous system. But stimulus information (I suggest) is information about the environment and is relevant to perception. It is the “exterospecific component” in the stimulus flux over time. How the “propriospecific component” in the stimulus flux is separated out poses a problem, for we now know that it cannot simply be the input from Sherrington’s “proprioceptors.” Von Holst’s theory of feedback detection does not solve the problem, for not all proprioception is “reafferent.” The answer probably depends on an understanding of “variance-detection” in the nervous system as against “invariance-detection”.
The functional distinction between perception and proprioception, as contrasted with the receptor-based distinction, will clarify confusion. For example, there are many experiments concerned with the experience that we call “the vertical.” This awareness of what Koffka called the “framework” of space is a perception. The accompanying awareness of the axis of the body is a proprioception. Objects, edges, and surfaces are seen as upright or tilted with reference to the ground plane. Similarly, the body is “seen” and “felt” as upright or tilted with reference to this basic feature of the environment. The posture of the body axis (and also of the body-parts) is detected in relation to space (usually gravity and the ground plane). Only under conditions of vertigo and disorientation does this detection break down. The perception of the (stable) environment and the proprioception of the (mobile) body are reciprocal. Each implies the other. When one fails, so does the other.
Now, the information for the perception of gravity (the vertical) and the ground (the horizontal) is given by invariants that are picked up by the vestibular system, and the skin; joints, muscles, and eyes in corroboration of one another. These systems compose the postural systems, or what I call the basic orienting system (Ch. 4); they are normally active and, when so, are little subject to illusions. But when the vestibular system or the eyes (or both) passively receive imposed stimulation illusions may result; or if the visual and the vestibular systems are given discrepant information equivocal perception may ensue. There are a great many experiments of this sort. But perception and proprioception have been confused in interpreting the results.
An error in the detection of the posture of the body relative to real gravity may be brought about in several experimental ways. Equally an error in the perception of what the true vertical of the environment is may be induced (for example, by artificially tilting the whole room and thereby the whole ambient array of light). An error in the perception of “the vertical” (as indicated by the setting of a rod) will necessarily entail an error in the proprioception of one’s bodily tilt, since one can only feel tilted or upright relative to “the vertical”.
The results of these puzzling and interesting experiments should be interpreted in terms of (1) information-seeking perceptual systems which normally get multiple information about the permanent or rigid environment and (2) proprioceptive or posture-regulating systems which have to detect the postures of then body relative to the permanent environment in order to control action. The two functions are concurrent. The organs of perception, the ocular and the vestibular for example, normally serve both functions. The effort to isolate the sensations from the separate organs and theorize about how they are combined is probably doomed to failure. It is a mistake to apply the terms “subjective” or “objective” to either proprioception or perception, since it encourages the confusing of different problems. Whether or not they are correct is one thing; what they are about is another. Proprioception is about the body; perception is about the world. One accompanies the other, but this does not mean that awareness of the body necessarily “influences” or “determines” awareness of the world.
A special kind of proprioception is the detecting of the movement of a body member relative to the whole, and of the whole body relative to the environment. Both are “kinesthesis” in the liberal meaning of the term, although the latter kind of movement is called locomotion. Either may be an active or a passive movement (and note that only in the former case, not in the latter, does the term “feedback” apply — a point not always recognized). I am implying that the detection of locomotion, by whatever system of receptors and organs, ocular, vestibular, cutaneous or other, is a form of proprioception. It should not be confused with the perception of moving or changing objects and surfaces relative to the rigid and permanent overall environment. I assign the term “movement” to the observer and “motion” to objects. The separate registration of both the movements of the observer and the motions of objects, in relation to “space,” is required for the control of his behavior. And now we can note how intimately connected the fact of locomotion perspective is with the fact of space perception. The flow of the ambient optic array specifies locomotion and is propriospecific; at the same time certain features of the flowing array specify the environmental layout and are exterospecific. During locomotion, one is always aware of both his own unique moving point of observation and the surrounding surfaces, each in relation to the other. (And if there is a moving object in the world, one is aware of that, too, probably on the basis of changing occlusion.)
According to this formula, the visual sensations of optical motion that can be experienced during locomotion are incidental to, first, the visual proprioception of the body relative to the world, and, second, the visual perception of the world with the body in it. A piecemeal analysis of the sensations will explain nothing unless these different (if reciprocal) psychological acts are considered. The “optical motions” in the field of view probably ought not to be called motions at all since they are not comparable to either the locomotions of the body of the motions of objects.
Howard and Templeton, Human Spatial Orientation. Gibson, Senses Considered, Ch. 2, 3, and 4, p. 296 ff.
Geldard, The Human Senses, p. 265 ff. on illusions of the vertical.
Merleau-Ponty, Phenomenology of Perception, on the experience of existence in the environment.
Gibson, Visually controlled locomotion. Brit. J. Psychol,. 1958, 49 182-194.
Gibson, What gives rise to the perception of motion? Psychol. Rev., 1968, 75, 335-346.