January 1974
Overt and Covert Attention
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 References to these essays must cite them explicitly as unpublished manuscripts.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 act of attention is primarily an act of orienting, exploring, adjusting, and optimizing, not a process of filtering. It is a circular process of input and output combined, where the input is modified by the output as well as the output being dependent on the input, not a linear process where each stage of the input follows a previous stage. It is, to use an old-fashioned and rejected terminology, a sensorimotor process, not a sensory process. It entails motor responses, although it is not itself a motor response; in fact it is not consistent with the stimulus-response formula in psychology but only with a new formula which substitutes equilibrium-seeking for responses to stimuli. It involves obtained stimulation as contrasted with imposed stimulation, that is to say not only the modifying of the stimulus input but even the seeking of new stimulus input. It depends on overt activity of the perceptual organs not on the passive stimulation of sensory receptors. Overt attention is the characteristic activity of a perceptual system, not a processing of data coming in over a channel of sense. Typically it does not operate on sensations conceived as signals in a sensory nerve-bundle but on what I have called the flowing sea of stimulus energy in which we live. The concept of the information for perception should not be taken from the theory of communication. I have argued all this in the Perceptual Systems.
The above distinction between overt attention and covert attention, with the assertion that the former is primary, does not in the least deny that the latter is possible. There can be “inner” or “mental” attention. Perhaps the clearest example of this is the ability to look at one object in the environment but pay attention to another or others. This ability, far from being simple, is complex and it needs a careful examination.
Consider William James’ description of this ability (Principles, Vol. 1, p. 437). “Usually, as is well known, no object lying in the marginal portions of the field of vision can catch our attention without at the same time ‘catching our eye’…provoking such movements of rotation and accommodation as will focus its image on the fovea, or point of greatest sensitivity. Practice, however, enables us, with effort, to attend to a marginal object whilst keeping the eyes immovable.” Note the distinction between catching the attention and catching the eye. Normally they go together but, with training, the former can occur independently of the latter. One has to learn to pay attention (in the covert sense) to an object in the periphery of the field of view while paying attention (in the overt sense) to an object in the center of the field of view with its image on the fovea.
In this case, which is the object of interest? Presumably not the object at which one is looking. Nevertheless a special kind of interest must be developed in the item being fixated or else fixation will not persist. This is why the act of looking steadily at a fixation point is so difficult for an observer faced with a laboratory display. He has to maintain a kind of interest in a black spot or the center of an X along with another kind of interest in a peripheral item, or in some feature of the peripheral field. Children and animals cannot do it. The required fixation point has no meaning, value, or affordance. The observer must set himself to fixate, accepting an unnatural task, or else his gaze will deviate. The overt attention is not caught or captured but has to be mobilized. Attention to the peripheral parts of the field cannot possibly yield wholly clear perception.
Displays of optical information in the laboratory, whether of large visual angle approaching the whole field of view or of small visual angle, are often presented with this fixation-point method. An alternative is the tachistoscopic method which also prevents exploratory scanning of the display, or the field of view, by not allowing enough time for eye-movement. Thus the optic array is printed on the retina, as it were, in the form of an image or picture. But these methods of freezing the overt attention, of preventing the eye from being “caught”, are unnatural. In life as contrasted with the laboratory what catches the attention is what catches the eye, and they are not separated. It is an item of interest in the optic array, a specific of what the environment affords the individual.
In life the eyes jerk about and make several fixations per second. What gets fixated or foveated? What is meant by an “item of interest” in the optic array? To answer “stimuli” and to talk of the intensity, vividness, color, or motion of such stimuli will not do. I believe that what catches the eye in an optic array is nothing less than an invariant of its structure which needs to be seen clearly. It needs to be seen clearly because it specifies something important in the world. And the application of the fovea (along with the focusing of the lens) improves the clarity with which a structure is seen. The movement of an eye which results in clarification cannot possibly be a response to a stimulus. That would require that an item of interest in the periphery should capture mental attention before it can guide fixation attention, which seems to put the cart before the horse. The saccadic movement of an eye, like the focusing of its lens, must be a cybernetic process under natural conditions of vision.
From the new point of view what happens when the normal exploratory adjustments of the perceptual organs are prevented in a laboratory setting? Presumably the circular activity of combined input and output, with modification of input, is cut short. The process tends to become “sensory” instead of “sensorimotor.” The perceptual system is not allowed to reach an equilibrium, an optimal state for information pickup, and it acts more like a channel for sensory signals. The array of stimulation is imposed on a mosaic of receptors and the organ (eye or hand) can no longer shift about under the array. Efferent impulses to the muscles of the system are either inhibited or have no effect on subsequent afferent impulses. New or altered stimulation cannot be obtained. The traditional notion of a passive sense with activity being confined to the brain alone seems to be confirmed.
The theory of neural filtering of inputs comes into its own, since there can be selection from a whole mass of potential sensations without selection from the outer sea of potential stimulation surrounding the observer. A great deal of evidence fits in here, and goes to show that the inputs of the eyes and ears and skin can in fact be filtered. Attention can be shifted from one part of the retina to another, both being stimulated, or from one feature of a pattern to another. It can be shifted from one area of skin to another. Attention can be shifted from one ear to the other, both being stimulated, or from one message to another in the same ear. It can of course be shifted from one sense to another, as from seeing to hearing. This kind of selection is supposedly “mental” and it has quite generally been taken to be the essence of attention. It is a faculty or function, with a basis in the physiology of the brain, and it consists of the brain’s choosing what is to be admitted into consciousness. But how very different this kind of attention is from the cybernetic and equilibrium-seeking kind of overt attention. Writers on attention like Titchener and Pillsbury describe the facts of covert or neural attention. But James wrote a chapter on attention in the Principles that foreshadows the present emphasis.
Quite different theories of perception are implied by the facts of covert attention and those of overt attention. The first leads to the theory of operations on sensory input and to experiments on what has been called information processing. The second leads to a new theory of the extracting of information from the flowing array of stimulation.