March 1979
A Proposed Terminology for Discussing Images
J. J. Gibson, Cornell University and The Salk Institute
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 References to these essays must cite them explicitly as unpublished manuscripts.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 term image is loose, both vague and slippery. What are its various distinct meanings? The following suggestions are subject to revision.
1. Pictorial image vs. sculptured image
A picture on the one hand (drawing, engraving, or photograph, either on a movable surface or a fixed wall) or a sculpture on the other hand (free-standing sculpture, model or toy). An image is of something; the latter is said to be “represented”. An artifact.
An image is always of some sort of surface layout: an object (attached to or detached from a surface; composed of a solid, or vaporous; inanimate or animate) or a place (e.g., a terrestrial semi-enclosure) or an event (change of layout or substance, either continuous or discrete). There can be no image of an abstraction, only a symbol for it. Image and symbol must not be confused.
The object, place, or event imaged may be of an existent in the environment or a non-existent, i.e., something that has ceased to exist, or something that is possible, or something that is impossible (a unicorn; a fantastic Utopia). Note that we call the last kind of object, place, or event “imaginary”, with completely circular illogic!
Note that a camera-made image has to be of an existent but that a hand-made image (a chirograph) may be of something that no longer exists, or that is planned, or that is merely fantasized. The object, place, or event imaged may be existent but not terrestrial (e.g., a molecule or a galaxy).
Note that there seem to be intermediates between a pictorial and a sculptural image, in the form of “low” relief and “high” relief. But in general a picture is characterized by having a station-point, whereas a sculpture is not; a picture has only one “face” but a sculpture has many faces. In a painting or photograph only those surfaces that “faced” the point of observation of the picture-maker are “projected” (see below).
Note that any such image must be illuminated in order to be seen. There is, however, a special kind of pictorial image consisting of a semi-transparent sheet with a different opacities in different parts that istransilluminated (see below). We call it a “slide”, since it slides into a “projector”.
Note that there can only be an image of an isolated attached object in the air, or of a sample of a surrounding environment. There can be no image of the whole environment, and hence an “internal representation” of the environment, so-called, is not possible.
Note that an image of a changing object, animal, place, or layout can be devised in the form of a “motion” picture of a “working” model. This extends the familiar usage of the term, which is an image frozen in time (see below).
2. Mental Image
A pictorial or a sculptural image that is supposed to exist in the realm of the mind but not in the realm of matter. Two environments are taken for granted, one internal to the mind (or perhaps the skull?) and the other external to it (the “physical” environment). The mental image does not have to be illuminated or transilluminated in order to be experienced. Whether it is looked at (by the mind’s eye) or not looked at is controversial.
Many perplexities arise from the assumption that a image can be mental. For example, although the mental image is “imaginary”, in one sense, what it is an image can be either “imaginary” or “real” in another sense (see also memory image, after image, and mirror image).
Some students of mental images may find the above definition unwelcome but then they are obliged to find a better definition.
3. Mirror image.
A virtual object, person, place, or event that appears to exist at some definite distance behind a very smooth polished surface. (Virtual vs. real is not the same as imaginary vs. real.) The misperception occurs because visual solid angles coming from the real object have been reflected regularly at the mirror surface in accordance with the rule of equality of the angles of incidence and reflection. The virtual object is not congruent with its real object, but is in point-to-point correspondence with it (projective correspondence; see below).
Note that both the surface-layout and the real surface-layout can be viewed and compared by and observer. He can see objects and his own body (most of it) both “in the air” and “in the mirror.” (Consider the legend of Narcissus in this connection) This puzzling dual awareness may help to explain our historical confusion about the terms “unreal,”, “imaginary”, and “mental”.
Note that the mirror image of an object, place, or event is more nearby comparable to the sculptural image than the pictorial image in so much as perspectives of the virtual object change as the observer (or the mirror) moves. It is also more nearly comparable to the progressive image than the arrested image, since the virtual object moves as the real object moves, but not in just the same way.
Virtual objects are produced by other optical devices than a plane mirror, e.g., a telescope or a microscope. All of them might be called optical images.
4. Retinal image.
A projection of the facing surfaces of an object, place, or event (the “near” surfaces as distinguished from the “far” surfaces) on to the surface of the retina of a vertebrate eye (chambered rather than compound). This projection entails a one-to-one correspondence, traditionally a point-to-point correspondence between radiating atoms of the facing surfaces and focus-points of the retinal image, but possibly a correspondence between the faces and facets of the surfaces and the nested color-patches of the retinal image. (If the former, the correspondence between physical reflectances and phenomenalcolors cannot be explained.)
Actually the correspondence between parts of the environmental layout and parts of the retinal image might better be called introjective than projective.
5. Physiological image.
The momentary pattern of excitation of the retinal receptors or receptive units caused by the retinal image of a fixated eye. It is not an optical image but an anatomical image. This is the image (if any) that is transmitted to the brain, not the retinal image. The pattern of the physiological image changes whenever the eye turns but the pattern of the retinal image does not change, since it is anchored to the environment. The physiological image fades away when the retina stays fixed relative to the retinal image whereas the retinal image does not (see afterimage below).
This distinction between retinal image and physiological image does not seem to be recognized in physiological optics, although it is crucial for an understanding of vision.
6. Screen image.
A projection in the literal sense of throwing (casting) the shadow of an object on a homogenous diffusing surface. The shadow of an object has been taken to be its “form,” in accordance with Plato’s parable of the cave. The shadow is cast by radiant light, from a source, either a very distant source (which yields a parallel projection) or a nearby source (which yields a polar projection). The object may be an opaque solid or a semi-transparent sheet with differential transmittances a “projector” being used to cast differential shadows on the screen. The relation between the transilluminated picture and the screen picture in one of projective correspondence.
This relation can be abstracted and generalized (neglecting color and texture) to yield the ghosts of flat surfaces, the pure planes that studied in projective geometry. This discipline is convened with the transformations from one lane to another and the invariance under transformation, but not with the non-change underlying change in the actual environment.
7. Arrested image.
A so-called “still” picture
8. Progressive image.
A so-called “motion” picture
9. Afterimage.
The aftereffect of intense or prolonged stimulation of the fixed retinal receptor mosaic (with a trade-off between them). An aftersensation following the end of a strong local sensation. It may be a positive or negative afterimage with respect to color/brightness. It is a physiological image that persists instead of being temporary. It is phenomenally comparable to a transparent shadow of the object that was the source of stimulation superimposed on the surface of the environment being looked at. Since the afterimage “moves with the eye”, like other entopic phenomena, we are tempted to say that it is “projected” upon the surface by the eye. Like a screen-image it gets larger as the surface becomes more distant (in accordance with Emmert’s law). But this notion that an eye projects a physiological afterimage as a lantern projects a lantern slide is profoundly misleading. It can be extended to other doctrine that an image can be projected outward from the mind upon the environment, a mental image or a memory image.
10. Memory image.
The aftereffect of a terminated sense-perception of a object, person, place, or even, as distinguished from the aftereffect of a localized sensation. The primary memory image is thought to be short-term continuation of the percept whereas the secondary memory image is thought to be “stored” so that it can later be “retrieved.” The words trace or engram for a memory image imply that it is pictorial rather than sculptural, and arrested rather than progressive, but these conclusions are surely false; we have memories of solid objects and events. But how could the images of sculptures and motions be stored in the brain?
Note that the pictorial image, the sculptural image, the mirror image, and the screen image can be scrutinized, looked at, and scanned (the parts can be fixated successively), whereas I suggested that the memory image, the mental image, the physiological image, the afterimage, and the retinal image cannot. There is no illumination by which any of the last six kinds of image can be seen.
11. Eidetic image.
A kind of memory image exhibited by some children under some circumstances that can supposedly be projected on a screen and scrutinized as if it were the object or picture remembered. The claim has been made that the eidetic child can pay attention to features of details in the memory image that were not noticed in the original perception. But the evidence for this claim is questionable. The assertion that the child makes exploratory eye-movements between fixations of the parts of an eidetic image is also questionable. The evidence purporting to show that the memory image of an object may be indistinguishable from a perception of that object under certain circumstances (Perky’s experiment from Titchener’s laboratory, 1910) is very dubious. These experiments need to be repeated.
12. Image making.
The making of an artifact that enables another observer to become aware, in part, of what the image-maker has observed, remembered, planned, imagined, or thought, without the use of words or numbers. One can move, chip, or otherwise shape a substance; one can make traces on a surface; and one can employ a dark chamber with a lens (or pinhole) and surface of photopigments or photosensitive elements in a great variety of technological complicated ways. The artifact provides an optic array, providing information for the perception of the artifact as such. Thus there is a dual awareness, of the image as such and of what the artificer was aware of.