Gibson's Ecological View of Psychology

from Gregory, Richard L. (ed.) (1987). The Oxford Companion To The Mind. p. 294. Oxford: Oxford University Press.

GIBSON, JAMES JEROME (1904-79). A distinguished American experimental psychologist, whose work on visual perception was and remains unusually influential. Following an appointment at Smith College, Northampton, Massachusetts, where as a young professor he was considerably influenced by Kurt Koffka, who had recently emigrated from Germany, Gibson for many years ran a major department at Cornell University. His life-work was investigating visual perception of form and motion; his early work including the discovery that distorting lenses produce negative (reversed) adaptation to curvature even with free eye-movements. But his main work was to challenge the approach of Helmholtz and suggest a very different account of perception.

Gibson first of all moved away from the traditional experiments with pictures, and what is seen with a single static eye, towards the observer moving around freely and viewing moving objects in natural conditions. He was led to this by considering pilots landing on fields, where the 'flow lines' of motion are important for seeing the landing-point and estimating height and speed. From such considerations of 'visual flow', and texture gradients, he developed what he called 'Ecological Optics'. This almost ignored retinal images, and active brain processes, in favour of regarding perception as 'picking up information from the ambient array of light'. This is very different from the Helmholtzian notion of perceptions as Unconscious Inferences from sensory data and knowledge of objects. Gibson tried to explain object perception by supposing that some 'higher order' features are invariant with motion and rotation and are 'picked up' with no perceptual computing or processing being required. His search for such invariances, and for just what the visual system uses under various conditions (though this was not Gibson's intention) has proved useful for developing computer vision. His general philosophy, however, that perception is passive pickup of information, is hard to follow. His important books are: The Perception of the Visual World (1950), and The Senses Considered as Perceptual Systems (1966). Much of his work was done with his wife, Eleanor, a distinguished developmental psychologist.

from Hunt, Morton (1993). The Story Of Psychology, pp. 466-468, 474-477. New York: Doubleday. (See original for footnotes and citations.)

pp. 466-468:

There is yet another modern theory about depth perception, one that is neither specifically neural nor specifically cognitive. Not that its proponent tactfully combined the two; on the contrary, he virtually ignored the neural theory and dismissed the cognitive theories as unnecessary and based on wrong assumptions.

Only a thoroughgoing maverick would reject a century's worth of depth perception research and claim to have found a totally different and correct approach. Only a true nonconformist would assert that we perceive depth neither by neural detection nor inference from cues but "directly" and automatically. Only a brash individualist would present a radical epistemology in which the physics of light is said to be such that what comes to our eyes gives us an accurate, literal experience of depth, and that we need not interpret what we see because we see what is as it is.

Such a one was the late James J. Gibson (1904-1980), whose admirers consider him "the most important student of visual perception of the twentieth century" and "the most original theoretician in the world in the psychology of perception," but whose theory is considered by the majority of perception specialists "extremely implausible" (one reviewer even called it too "silly" to merit discussion) and has few advocates.

It did not help Gibson's cause that as he developed his radical concept of perception, between 1950 and 1980, he became ever more scornful of existing perception theories and of mainstream psychology in general. Typical of his views are these comments: "Psychology, or at least American psychology, is a second-rate discipline . . . The main reason is that it does not stand in awe of its subject matter," and "Psychology [is] the effort to find answers to the wrong questions; the study of problems chosen to be convenient to study, instead ot relevant." Yet despite his waspish opinions and a severe hearing loss, he was always so good-natured, engaging, and affable that almost everyone he came in contact with liked him, and one enthusiastic chronicler describes him as "fascinating, animated, and lovable."

Gibson, born in a river town in Ohio and reared in various parts of the Midwest, was the son of a railroad surveyor. This gave him many a chance to ride the trains and to experience what became a core element of his theory of visual perception. As he notes in a brief autobiography, "At the age of eight I knew what the world looked like from a railroad train and how it seemed to flow inward when seen from the rear platform and expand outward when seen from the locomotive."

Gibson went to Princeton University but felt out of place among the club members and preferred to associate with what he called "the eccentrics." For a while he vacillated between philosophy and acting (he was wavy-haired, square-jawed, and good-looking enough for leading roles), but in his senior year he took a course in psychology and at once heard the call. While doing graduate work in psychology at Princeton, he came under the influence of behaviorists. In 1928, however, he received a faculty appointment at Smith, where he met Kurt Koffka, and, though he did not become a Gestaltist, was much influenced by the Gestaltist concepts of organization and structure.

For some years, Gibson was interested in both social psychology and relatively traditional perception research. Then, during World War II he was asked by the Army Air Corps's Aviation Psychology Program to develop depth-perception tests for determining who had the visual aptitudes needed for flying, particularly for making successful take-offs and landings.

Perhaps because of Gibson's early experiences on trains, he considered the classical cues to depth perception, including shadows and perspective, of little worth. In his opinion they were based on paintings and parlor stereoscopes rather than on three-dimensional reality, and on static images rather than on movement. What seemed to him much more useful and realistic were two other kinds of cues: texture gradient, like the uniforrnly changing roughness of the runway as seen by a pilot during the final leg of an approach; and motion perspective, or the flow of changing relationships among objects as one moves through the environment, including all that a pilot sees during take-offs and landings. These cues soon became, and are today, accepted components of the cue-based theory of depth perception.

Gibson's Air Corps work held the germ of his later view. The crucial mechanism in depth perception (in all perception, according to Gibson) is not the retinal image, with all its cues, but the changing flow of relationships among objects and their surfaces in the environment that the perceiver moves through.

This concept came to dominate Gibson's thinking during the 1950s and 1960s, when he did a considerable amount of research at Cornell that tested his belief in texture gradients. In some experiments he placed diffusing milkglass between an observer and textured surfaces; in others he dilated the observer's eyes to prevent sharp focus on texture; in still others he cut PingPong balls in half and made goggles of them so that what his subjects saw was foglike, without surfaces or volume. From these and other experiments, plus a careful consideration of his research on air-crew testing and training, Gibson came to reject texture gradients and to stress movement by the observer through the environment as the key to depth perception. However large or small the movement, it results in changes in the optic array—the structured pattern of light reaching the eye from the environment—such as is suggested in this drawing:

Figure 31.

How optic array conveys depth

The optic array, rich in information as seen from any point, becomes infinitely richer with movement by the observer. Even minor movements of the head change the array, transforming what is seen of an object and the relationships among objects, and yielding optic flow of one kind or another. Gibson came to believe that there is enough, and more than enough, information in the optic array and flow to convey depth and distance directly, without the need of mental calculation or inference from cues.

This is how Gibson explained depth perception in his sweeping "ecological" theory of "direct perception." The pity is that Gibson, the outsider and eccentric, who, according to a fellow psychologist, was always "terribly tough and uncompromising," felt obliged to throw out the baby with the bathwater. For it is possible to acknowledge both the neural and cognitive views of depth perception as correctly explaining different aspects of the phenomenon and the Gibsonian view as supplementary to them. But it wasn't possible for James J. Gibson.

pp. 475-477:

The direct or ecological perception theory of Gibson and his followers attempts to explain not only depth perception, as sketched above, but visual perception in general. Gibson's doctrine that the perceiver's movement through the environment produces a constantly changing optic array is the heart not only of his explanation of depth perception but of his account of form, size, distance, and motion perception.

Gibson's exposition of his theory in his 1979 book, The Ecological Approach to Visual Perception, is abstruse, argumentative, and difficult to follow, in part because he uses many neologisms. But its limited acceptance by perception psychologists may have been more the result of his dismissal of nearly everything that had been achieved in perception research by others. A typical Gibson comment (from a posthumous article):

The conclusions that can be reached from a century of research on perception are insignificant. The knowledge gained from a century of research on sensation is incoherent. We have no adequate theory of perception, and what we have found in the search for sensations is a mixed batch of illusions, physiological curiosities, and bodily feelings. The implications are discouraging. A fresh start has to be made on the problem of perception.

Gibson's recommendation as to where to make that fresh start sounds sensible enough: with the question (echoing Koffka) "Why do things look as they do?" and continuing:

How do we see where we are in the environment? How do we see whether or not we are moving and, if we are, where we are going? How do we see what things are good for? How do we see how to do things?

To make a truly fresh start, Gibson discarded most of the fundamental assumptions that philosophers and psychologists have always made about visual perception. Among them: sensations are the basis of perceptions; stimuli from the outside impinge on the organism; the organism produces responses; the brain processes, integrates, and interprets the information; the proper study of vision begins with the retina and the brain; optical motion is something for which the visual system must compensate; the environment is represented only in part by the information on the retina; perception depends in large part on inference; and so on and on. A new broom, James J. Gibson.

In place of the assumptions he swept away, Gibson offers his own theory, which is more philosophic than empirical, more conceptual than data-based, and hard to grasp. All the same, he came to his views by means of experiments, first on motion perception and then on perception of form and other real-world qualities. As already noted, the first glimmer of the truth appeared to him in the course of his Air Corps research, of which he said:

We learned more about the perception of objects, I think, than we would ever have done by running standard laboratory experiments on form-perception. For one thing, I got a nagging suspicion that nobody ever really sees a flat form in life, that is, a picture of a thing. One sees a continuous family of perspective transformations, an infinity of forms, that somehow specifies the solid shape of the object.

From this, he went on later to reject not only all research based on pictures and illusions but all theorizing about perception as the result of the mind's interpretation of the imperfect information in the two-dimensional pictures on the retina:

Eventually I came to realize how unlike the pictorial mode of perception is from the natural one. The former is perception at second hand; the latter is perception at first hand. The framed optic array coming from a picture to the eye is quite unlike the natural optic array coming from the world to an eye . . . Eyes evolved so as to see the world, not a picture. Since this became clear to me I have tried to give up any use whatever of the term "retinal image."

Gibson concluded, as we have seen, that the optic array contains all the information about the real world that we need. He admitted that our knowledge of the world is limited by the characteristics of the nervous system, but those characteristics, he maintained, were determined by evolutionary adaptation to the world; we are aware of all those properties of the environment that we need to be aware of. Evolution produced a perceptual system in each creature that enables it to directly recognize the qualities in the environment that will be useful to it—in Gibsonian language, its "affordances," or what uses the properties of any entity will afford the creature. So things look edible, drinkable, walk-onable, swim-inable, and so on, in relation to the observer's size and physical equipment.

Perception, in short, is not a process of interpreting a degraded retinal image but a direct and truthful experience of reality by means of the optic array and optic flow. This, for Gibson, is the central mechanism of perception—not the neural events documented by Hubel and Wiesel (whose work he regarded as irrelevant), and not cognitive processes, which he considered to be based on erroneous and artificial assumptions.

Direct perception is also Gibson's answer to Berkeley. We know that the world is out there and exists independently of us because as we move through the environment, we see things in an ever-changing way and yet we experience their continuity, reality, and independence of us as perceivers. So do all other animals; only philosophers ever doubted that the world really is what it seems to be. Gibson's theory of visual perception thus ventured far beyond vision into epistemology.

And further. By the end of his life, Gibson had come to believe that perception was the cornerstone of all psychology and that his theory of it could bring about radical change in many areas of the science. Such concepts as mind, consciousness, learning, and drive would be replaced by an ecological psychology based on creatures' awareness of the useful and dangerous properties of places, events, and objects, and on their organization and control of their actions to achieve desirable results in the real world.

What progress Gibson might have made toward this empyrean goal we can never know; he died only two years after publishing The Ecological Approach, and although some of his ideas about visual perception, notably optic flow, have become part of accepted doctrine, neither these nor his more ambitious notions have brought about a revolution in perception, much less psychology in general.

It is a pity that Gibson became so intolerant and evangelistic; his concept of optic flow, though it does not make unnecessary much that he thought it did, is certainly of value, but his extremism has kept his contribution from being as widely accepted as it deserves to be.

[By way of contrast, here is a representative cognitive view of perception, Irving Rock's theory of perception as problem-solving and inference-making starting from an impoverished retinal image]

pp. 474-475

Rock spent many years devising and conducting other experiments to test the hypothesis that, more often than not, perception requires higher-level processes than those taking place in the visual cortex. These studies led him, finally, to his present thesis that "perception is intelligent in that it is based on operations similar to those that characterize thought."

And indeed, says Rock, perception may have led to thought; it may be the evolutionary link between low-level sensory processes in primitive organisms and high-level cognitive processes in more complex forms of life. If what the eye sees, he argues, is an ambiguous and distortion-prone representation of reality, some mechanism had to evolve to yield reliable and faithful knowledge of that reality. In his words, "Intelligent operations may have evolved in the service of perception."

This is not to say that all perception is thoughtlike; Rock specifically cites the waterfall illusion as explicable in low-level neural terms. But most facts about motion perception and other kinds of perception seem to him to require high-level processes. Unconscious inference, as in our use of texture gradient cues to sense distance, is only one of them. Description that results in interpretation is another. In the ambiguous old hag-young woman figure by Boring, what one sees is not the result of simply recognizing an image but of describing to oneself what a particular curve is like: like a nose or like a cheek. Many perceived forms or objects are not instantly recognizable; recognizing what they are comes about through such a process.

Perception also often calls for problem solving of one sort or another. One hardly thinks of perception as the solving of problems, but Rock has marshaled a considerable amount of evidence—much from earlier studies by others, some from his own original experiments—to show that in many cases we seek a hypothesis to account for what we see, weigh that hypothesis against other possibilities, and choose the one that seems to solve the problem of making sense of what we see. All of this usually takes place in a fraction of a second.

One example: In a laboratory phenomenon known since the time of Helmholtz, if a wavelike curvcd line is passed horizontally behind a slit, as in this figure,

Figure 34.

Anorthoscopic perception: The dot moves up and down but the mind figures out what is happening.

most observers first see it as a small element moving up and down, but after a while some of them will suddenly see the sinuous line moving at right angles to and behind the aperture. What produces their altered and correct perception? Rock found that one clue they use is the changing slope of the line as it passes the slit; another is the end of the curved line, if it comes into view. Such clues suggest to the mind an alternative hypothesis—that a curve is moving past the slit horizontally, rather than that a small element is moving up and down. This hypothesis is so much better that the mind accepts it and sees the line as it really is.

Rock sums up his theory as follows:

On a theoretical level, at least according to the theory presented here, both perception and thought entail reasoning. In some cases, generalizations or rules are arrived at in perception by induction. These rules are then used deductively as premises from which inferences are drawn. Perception in some cases can be characterized as the result of creative problem solving, in the sense of searching for the grounds (or internal solution) from which a specific interpretation follows. Perception entails decisions, just as does thought. Operations that culminate in perceptual experience are of the same kind that characterize thinking.