May 1971
Do We Ever See Light?
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 References to these essays must cite them explicitly as unpublished manuscripts.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.
This memo stems from the paper by Mark Carter for the seminar on Ecological Optics, March 1971.
The various theories of sensation-based visual perception are not only contradictory to one another but also vague. Failure to understand them is not the fault of the student of vision but of the theories themselves, for they are not clear. They all take it for granted that we see light. This assumption is not as self-evident as it sounds. It is questionable, and there are reasons for rejecting it.
The orthodox theories of visual perception also take it for granted that we have sensations of light, and that these sensations are necessary for perception, although the theories do not agree on how sensations enter into the perceptual process. After examining the reasons for believing this assumption, it will also be rejected.
It is important to note at the outset that we are not raising the question of whether or not inputs of the optic nerve, afferent impulses, are necessary for visual perception. That is a different question. Sensory physiologists tend to identify afferent in[puts with sensations but this is surely a mistake, for they are not cognate. The former may be entailed in perception without the latter.
Do we ever see light? I will suggest that we do not, and that the reasons for thinking so are mistaken. The argument will appeal to phenomenology but to a somewhat more disciplined variety than is often practiced, governed by logic, ecology, and common sense. There seems to be only three classes of environmental things that one actually sees: 1) surfaces the emit light (luminous surfaces or substances that are sources of light); 2) illuminated surfaces that are textured or structured so that they “scatter” the incident light (and then may be solid, liquid, or vaporous); and 3) shadows that are cast on or “attached” to illuminated surfaces. When we see a fire, or flame, or a lamp, or the sun we are perceiving a surface that emits light. The translucent surface of a lightning “fixture” is also of this sort. The moon is a reflecting surface that (for good reasons) appears to be luminous.
When we see a solid surface, or a liquid, or a cloud, or a fog-bank it is only because it has occluding edges and is textured; the perfect plane of glass or water is not seen. And when one sees a shadow it is always (I think) the seeing of an unlighted or darkened surface relative to adjacent lighted surfaces. We see the lightedness of a surface but not the light. In none of these cases does one see light as such, i.e., radiant light.
What about seeing a beam of light? It is only because one sees the reflecting particles in the transmitting medium; this fact applies also to seeing the “rays” or “shafts” of sunlight through clouds.
What about seeing the air? I maintain that we do not. What about the empty sky? There is more justification for this claim, but I would suggest that we only see the cloudless sky relative to the earth, asnothing in contrast with the something of the terrain, separated by the horizon. The emptiness of the sky is analogous to the emptiness of a window, aperture, or hole relative to its occluding edges of true surface.
What about seeing “space”? Again, I maintain that we do not see it, as little as we see “depth” of the “third dimension,” and that what we call the perceiving of space is actually the seeing of slants of occluding edges, or one thing behind another.
What about seeing “the illumination,” or at least registering it, in opposition to seeing “darkness” or “the dark”? My experience suggests, again, that we see the illuminatedness of things but not the illumination as such. Illumination implies something to be illuminated.
Anything that is perfectly transparent is invisible, like clean air and pure water. Only a semitransparent surface is visible as when, for example, the plate of glass or the mirror is dirty. The necessary condition for seeing anything is a structured optic array. Only when something imposes its own structure on the optic array does one see it, for only then is there information for its existence.
Consider next the assertion that we have sensations of light. There are supposed to be sensory impressions that light makes on the retina by analogy with the depth or impressions that objects make on the flexible skin. These are supposed to be the data for sense perception. But I will argue that we do not have sensations of light. A spot or point of light, so-called, in ambient darkness is seen as a small, far-off, and luminous thing, however indefinite. A flash of light, so-called, is seen as a luminous event. Neither one is a sensation of light. We get a genuine sensation, to be sure, when the radiant energy entering the eye is very intense but this is a sensation of dazzle or glare. We get an aftersensation or “afterimage” as a result of such strong stimulation. We also get ocular sensations when the eyeball is squeezed or mechanically jarred, and when the retina is stimulated by an electric current. We sometimes see entoptic phenomena or “spots before our eyes.” All these are subjective sensations, and they are referred to the eyes, but none of the halos or spots or scintillations that are sometimes said to appear to be outside the eye is a sensation of light. The sensation is that of something wrong with the eye. It is a sort of visual itch. The dazzle sensation can pass over into visual pain.
In short there can be sensations that arise from abnormal or excessive retinal stimulation of the receptors, including excessive stimulation by light, but they are cases of sensations, that is, of bodily sensation. Body sensation is not the kind of sensation that has been supposed to be the data perceptionña perceptual sensation. Such experiences cannot possibly be the basis of perception, or the elements of perception. They are the results of exciting the retina abnormally, or artificially, or inappropriately, or occasionally the result of disease.
Sensations can also arise from abnormal ocular adjustments, or from exceeding the capacity of the system to adapt. The impressions that accompany vertigo are of this sort. But they are incidental to the act of perception, not essential to it; symptoms of the physiology of perception, not components of it. The various sensations of the eye are in any case not sensations of light.
To sum up, the seeing of the environment is not based on the seeing of light. And even if there were such a thing as the seeing of light it would not be a matter of having sensations of light if the term sensation is taken to mean a subjective bodily experience.