June 1969
Homogeneous Optical Stimulation and its Implications for Visual Perception
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 References to these essays must cite them explicitly as unpublished manuscripts.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.
The problem of what an observer sees when he is confronted with a homogeneous total field of light-stimulation (Ganzfeld) dates from Metzger’s experiment of 1930. The experiment has been repeated in various ways but the implications of the results are not agreed upon.
The observer usually reports an impression like that of a “space-filling fog” in front of his eyes. Koffka took this to mean that a three-dimensional experience of space instead of a two-dimensional experience of color arose spontaneously even under these “simplest possible conditions of stimulation” (Principles, p. 115). But this conclusion only implies that depth perception is innate, or intrinsic to perception; it takes sides against empiricism, in an ancient and fruitless controversy without resolving it (e.g. Natroulas, 1963).
Twenty years later, having formulated what I called a “surface theory” of space perception instead of an “air-theory,” I argued that a “retinal texture” was the necessary stimulus condition for the perception of a surface. I interpreted the impression of a “space-filled fog” to mean only that the observer experienced no surface, or no thing, that is, nothing. This was neither two-dimensional nor three-dimensional but indeterminate. Part of this theory were the postulates that retinal texture, and gradients of density of this texture, were “stimuli” for the perception of surface and for the slant of this surface.
The theory in its early form was weakened by not having the concept of the optic array, and by not making the distinction between the stimulus energy of light and the stimulus information in light; in short, the early theory was not based on ecological optics, but only on the retinal image conceived as a set of focused points of luminous energy. Let us reexamine the experiments on homogeneous optical stimulation, distinguishing between the light entering the eye and the light stimulating the retina.
Metzger has faced his observer with a large smooth plaster wall, reflecting surface similar in principle to a theatrical “cyclorama.” My observations were made with a hemisphere of diffusing glass in front of the face, and later with a translucent eye-cap covering each eye (Gibson and Dibble, 1952; Gibson and Waddell, 1952). The important fact to note about this apparatus compared to Metzger’s is that it makes accommodation impossible. The light coming to the nodal point of the eye has no discontinuities of intensity in different directions (although it need not have uniform intensity in all direction). Hence the light is unfocusable, and the lens can only “hunt.” This condition of the eye has been called “empty field myopia.”
What this demonstrates, I now suggest, is that when the light that is available for the eye is wholly undifferentiated a retinal image cannot exist. It shows that the ambient light at a point of observation is more fundamental for the theory of vision than is the retinal image. If the light entering the eye is unfocusable, the light stimulating the retina still constitutes stimulation for the retina but does not constitute stimulusinformation for it. An eye can extract information from ambient light only when the ocular system can form an image and, for this the ambient light must be structured, that is, it must constitute an optic arrayhaving an arrangement.
An optic array can be defined as a set of adjacent angles at a point of observation, not as an infinitely dense set of light rays intersecting at a point. The definition implies differences of intensity in different directions, not different intensities in different directions, that is, it implies margins or contrasts in the array not a distribution of intensities over the array at what we have called a “Boynton point.” The advantage of this definition is that margins or contrasts are invariant with changes of illumination.
The theory of image-formation in classical optics is based on the assumption of correspondence between each object-point and its image-point. The light entering the eye is conceived as a set of focused pencils of rays, e.g., the light has to be focusable. If it is not focusable, (as the light behind a diffusing translucent eye-cap is not focusable) the classical theory of image-formation fails since no image is formed. It appears that we must begin to think of the retina in two different ways, as a receptor-mosaic, and as an organ. Conceived as a set of photoreceptors the retina, when stimulated by homogeneous light, sends impulses up the optic nerve and produces sensations (or a sensation) of light. Conceived as part of the system for extracting information from an array of light the retina, when stimulated by homogeneous light, is inactivated. The eye cannot accommodate, cannot fixate, and cannot explore, despite the fact that impulses in the optic nerve are occurring, and that there is “input” to the system. In short, the system cannotperceive although it is sensitive to light. Note that, with homogeneous stimulation of the retina, no amount of “central processing” of the input can bring about a perception. No observer wearing diffusing eye-caps has ever perceived anything. The observer cannot see; he is just as “blind” as he could be in utter darkness. For the field of stimulation when it is undifferentiated, unstructured, continuous, or leveled out is also informationless. The main implication for visual perception of homogeneous light-stimulation is that we should study heterogeneous stimulation.