Perhaps the most difficult problem faced by behaviorism has been the treatment of conscious content. Are we not all familiar with colors, sounds, tastes, and smells which have no counterparts in the physical world? What is their place in a behavioristic account? I believe the answer is to be found in the special role assigned to stimuli in an operant analysis. It calls for a certain amount of technical detail, and I shall treat it in some depth.
In the traditional view a person responds to the world around him in the sense of acting upon it. Etymologically, to experience the world is to test it, and to perceive it is to capture it—to take it in and possess it. For the Greeks, to know was to be intimate with. A person could not, of course, capture and possess the real world, but he could make copies of it, and these were the so-called data—the givens—with which, in lieu of reality, he worked. He could store them in his memory and later retrieve and act upon them more or less as he might have done when they were first given.
The opposing view—common, I believe, to all versions of behaviorism—is that the initiating action is taken by the environment rather than by the perceiver. The reflex was a conspicuous example, and a stimulus-response version of behaviorism kept to the same pattern, as did information theory and some computer models. A part of the environment entered the body, was transformed there, perhaps was stored, and eventually emerged as a response. Curiously enough, this differed from the mentalistic picture only with respect to the initiator of action. In both theories the environment penetrated the body: in the mentalistic view, it was taken in by the perceiver; in the stimulus-response view, it battered its way in. The two formulations could be combined—“an image of the outer world striking the retina of the eye activates a most intricate process that results in vision: the transformation of the retinal image into a perception.” Both formulations directed attention to the inner representation of reality in its various transformations. A basic question could be put this way: What becomes of the stimulus?
In an operant analysis, and in the radical behaviorism built upon it, the environment stays where it is and where it has always been—outside the body.
The environment affects an organism after, as well as before, it responds. To stimulus and response we add consequence, and it is not just a third term in a sequence. The occasion upon which behavior occurs, the behavior itself, and its consequences are interrelated in the contingencies of reinforcement we have already examined. As the result of its place in these contingencies, a stimulus present when a response is reinforced acquires some control over the response. It does not then elicit the response as in a reflex; it simply makes it more probable that it will occur again, and it may do so in combination with other conditions affecting probability, such as those discussed in the preceding chapter. A response reinforced upon a given occasion is most likely to occur on a very similar occasion, but because of a process called generalization it may appear on occasions sharing only some of the same properties. If, however, it is reinforced only when a particular property is present, that property acquires exclusive control through a process called discrimination.
The role of the stimulus gives operant behavior a special character. The behavior is not dominated by the current setting, as it appeared to be in stimulus-response psychology; it is not “stimulus-bound.” Nevertheless, the environmental history is still in control; the genetic endowment of the species plus the contingencies to which the individual has been exposed still determine what he will perceive.
Many of the issues discussed in the preceding chapter extend to the stimulus control of operant behavior. For example, perception is in a sense purposive or intentional. A person is not an indifferent spectator soaking up the world like a sponge. An early objection to John Locke’s theory of human understanding was that stimulation seemed to be coldly engraved on the tabula rasa of the mind, and efforts were made to supplement the theory by saying that a person “beheld things as liked or disliked, approved or disapproved, or pleasing or displeasing,” or that a person “judged” the world as he perceived it. But expressions of that sort simply assign to fanciful inner processes what is to be found in genetic endowment and personal history. We are not merely “mindful” of the world about us; we respond to it in idiosyncratic ways because of what has happened when we have been in contact with it. And just as operant conditioning does not mean that a person “infers what will happen when he acts,” so the control exerted by stimuli does not mean that he “infers what exists in the world around him.”
It is often pointed out that a person who has been driven over a route as a passenger cannot find his way as well as one who has himself driven the route an equal number of times. Animals carried about in a given setting do not then move about in it as well as animals who have already moved about. Both have been exposed to the same visual stimuli, but the contingencies have been different. To ask why the passenger and the animal carried about have not “acquired knowledge of the setting” is to miss the point. They have not acquired behavior under the control of the setting.
When a stimulus is weak or vague, it is often clear that other conditions are affecting the probability that a person will see a thing in a given way. The lover “thinks he sees” his beloved in a crowd but only if the visual stimulus is fleeting or obscure. (The effect is studied in the laboratory by exposing a stimulus for a very short time, say, or near the edge of the visual field, or in faint light.) A slight noise at night is heard as a burglar or a mouse by those who respond vigorously to burglars or mice. Level of deprivation makes a difference; one mistakenly “hears the telephone” if a call is important, and the sexually deprived see phalluses or vaginas in objects bearing little geometrical similarity to those organs. In other words, a person sees one thing as something else when the probability of seeing the latter is high and the control exerted by the former is low.
The importance of the history of the perceiver is clear when a chess master looks at a game in progress. What he sees is very different from what is seen by one who does not play chess or who has not played it long. For the master, the setting is an occasion upon which many different moves have been made with good or bad results in games with which he is familiar. To the person who is just learning to play, the setting may be an occasion for a number of moves but moves which have not been much affected by consequences. To the completely naive, the board and its pieces are a visual setting to be described only through possible resemblances to situations in his non-chess-playing history.
We recognize the importance of a history of reinforcement when we undertake to make it more likely that a person will see a particular thing—or, in other words, that he will engage in a particular kind of seeing. We can present a thing suddenly or conspicuously or in a novel and hence surprising way, and we can point to it if our subject has learned to follow a point—that is, if he has learned to behave effectively under contingencies in which a thing indicated plays an important part. But we can also arrange that a particular object will be seen by establishing contingencies which can be met only by responding to it. Traffic signs are designed to be easily seen, but we see them or ignore them largely because of the contingent consequences. Measures of this sort are often said to increase a person’s awareness, or to expand his mind or consciousness, but they simply bring him under more effective control of his environment.
The structuralists have tried to explain perception in terms of the form, or configuration, of what is perceived. Gestalt psychologists may be said to have argued that certain kinds of patterns force the organism to perceive them in certain ways. Some illusions, for example, seem irresistible; we see what we know is not really there. Some examples seem to be reasonably explained in terms of natural selection: it is not surprising that when we see a bird fly behind a tree trunk, we behave as if it continued to exist when out of sight, and even see it move from one side to the other as we see a traffic light jump from red to green. Small gaps in orderly patterns are neglected with profit as we “neglect” the blind spots in our eyes. We do not need to postulate structural principles to explain these characteristics. Contingencies of reinforcement also contribute to irresistible perceptions: a rotating trapezoid which refuses to appear to go around is made more effective by representing it as a window frame.
The great differences in what is seen at different times in a given setting suggest that a stimulus cannot be described in purely physical terms. Behaviorism is said to be at fault in failing to recognize that what is important is “how the situation looks to a person” or “how a person interprets a situation” or “what meaning a situation has for a person.” But to investigate how a situation looks to a person, or how he interprets it, or what meaning it has for him, we must examine his behavior with respect to it, including his descriptions of it, and we can do this only in terms of his genetic and environmental histories. To explain how the real world is converted into an internal iconic representation, one authority has suggested the following: “For perception to go beyond the evidence of the senses the brain must have stored information, allowing it to use available sensory data to choose between possibilities derived from past situations. Behavior is not controlled directly by stimuli … but by the brain’s hypotheses of what probably lies in outside space and in the immediate future.” (This is an example, by the way, of a current practice of avoiding dualism by substituting “brain” for “mind.” The brain is said to use data, make hypotheses, make choices, and so on, as the mind was once said to have done. In a behavioristic account it is the person who does these things.) But we observe simply that a person responds to a current setting (“the evidence of his senses”) because of his exposure to contingencies of which the setting has been a part. We have no reason to say that he has stored information which he now retrieves in order to interpret the evidence of his senses.
Some of the history relevant to perception may have occurred during the evolution of the species. What is seen seems to “depart from the object world,” for example, in the illusions mentioned above, in some of which the mind is said to “infer and predict reality from incomplete data,” but we should say instead that because of his genetic endowment a person responds in a possibly effective way to what seem to be fragmentary stimuli.
The psychophysicists have most rigorously explored the correspondence between experience and reality. Early psychologists, like Wundt and Titchener, tried to discover what a person saw (or heard, felt, and so on) under the pure control of current stimuli, free of the effects of previous exposure. A trained observer was to describe his sensations without making the “stimulus error”—that is, to describe what he was looking at as if he had never seen it before or could never have learned anything about it. He was to see a “patch of color” rather than an object; he was to have a salty taste rather than taste salt; he was to feel warm rather than the warmth of the sun on his skin.
In doing so, he was to see the irreducible elements of mental life, but even so, sensation seemed different from reality because changes in stimuli did not produce comparable changes in what was seen. A psychophysical function was said to represent the relation between the two worlds; but we could say instead that it represents facts about the discriminative control of stimuli. The position of conscious content grew weaker when methodological behaviorism, together with operationism and logical positivism, questioned the usefulness of sensations as scientific data, and psychophysicists then turned to the process of discrimination, as we have seen. But it was possible to study discrimination while believing in the existence of a world of experience.
Further studies of discrimination, particularly research on the sensory processes of animals, were responsible for further progress. In 1865 Claude Bernard had contended that “experimental studies of sense organs must be made on man because animals cannot directly account to us for the sensations they experience,” but there is now an elaborate “animal psychophysics,” in which stimulus control is analyzed with great precision. It is still likely to be said that the experimenter has “taught the animal to report what it sees,” but the results can be much more consistently formulated in terms of the control set up by specific contingencies of reinforcement. Of all the great mentalistic explanations, the “understanding” or “knowledge” of the British empiricists has suffered the most ignominious fate: it has been reduced to the physiology of the eye and ear.
The distinction between a physical and a mental world, most often found in Western cultures, presumably arose, as in Plato’s supposed discovery of the mind, in the effort to solve the dimensional problem of mental life; there was not enough room in the body for the copies of the world a person seemed to possess. Later, with the rise of science, a different kind of discrepancy appeared. Were the qualities of images and ideas to be found in nature at all? To use a well-worn example, did a falling tree make a noise if no one heard it? Light might be a matter of corpuscles or waves, but it certainly did not seem to be a matter of colors; green was not a wave length of light. This was not a serious problem for early philosophers, who had no reason to question the fact that they lived in a world of colors, sounds, and so on. Nor is it a problem to millions of people today, who also believe that they do so. Nor is it a problem for a behaviorist.
To argue that layman and scientist are simply looking at two aspects of the same thing is to miss the point, because aspect is what causes trouble: people see different things when they have been exposed to different contingencies of reinforcement. Like everyone else, the scientist sees green, but he also responds in other ways to the same setting. It is a mistake, however, to say that the concepts of science are constructed from personal sensory experience. Both layman and scientist respond—in similar or different ways, depending upon the contingencies—to the features of a given setting. (I shall return to the personal knowledge of the scientist in Chapter 9.)
The stimulus control of behavior is subject to severe limitations. Our genetic endowment restricts control to electromagnetic radiation in the visible range, for example, and to sonic sounds, and even within these ranges the eye and ear have their defects. Their faults are not, however, a matter of faulty inferences. The discrepancies are not in a correspondence between experience and reality but in stimulus control.
It is easier to make the point when reality is more complex. When an unfortunate war is attributed to “misperception” or a seminar is devoted to the “discrepancy between the reality and perception of technological change,” translation is mandatory. How are we to perceive the reality of the war or the technological change in order to discover that it has indeed been misperceived? We are always “dealing with reality,” although the term must be taken to include more than a current presentation. The important differences are among behaviors, and these in turn are explained by differences in past contingencies.
Those who believe that we see copies of the world may contend that we never see the world itself, but it is at least equally plausible to say that we never see anything else. The copy theory of perception is most convincing with respect to visual stimuli. They are frequently copied in works of art as well as in optical systems of mirrors and lenses, and hence it is not difficult to imagine some plausible system of storage. It is much less convincing to say that we do not hear the sounds made by an orchestra but rather some inner reproduction. Music has temporal patterns, and only recently have copies been available which might lend themselves to a mental metaphor. The argument is wholly unconvincing in the field of taste and odor, where it is not easy to imagine copies distinguishable from the real thing, and it is seldom if ever made in the case of feeling. When we feel the texture of a sheet of paper, we feel the paper, not some internal representation. Possibly we do not need copies of tastes, odors, and feelings, since we are already physically intimate with them, and for presumably the same reason we are said to feel internal states like hunger or anger rather than copies.
The trouble is that the notion of an inner copy makes no progress whatsoever in explaining either sensory control or the psychology or physiology of perception. The basic difficulty was formulated by Theophrastus more than two thousand years ago:
… with regard to hearing, it is strange of him [Empedocles] to imagine that he has really explained how creatures hear, when he has ascribed the process to internal sounds and assumed that the ear produces a sound within, like a bell. By means of this internal sound we might hear sounds without, but how should we hear this internal sound itself? The old problem would still confront us.
Similarly, as a modern authority has pointed out, it is as difficult to explain how we see a picture in the occipital cortex of the brain as to explain how we see the outside world, which it is said to represent. The behavior of seeing is neglected in all such formulations. It can take its proper place only if attention is given to other terms in the contingencies responsible for stimulus control.
When a person recalls something he once saw, or engages in fantasy, or dreams a dream, surely he is not under the control of a current stimulus. Is he not then seeing a copy? Again, we must turn to his environmental history for an answer. After hearing a piece of music several times, a person may hear it when it is not being played, though probably not as richly or as clearly. So far as we know, he is simply doing in the absence of the music some of the things he did in its presence. Similarly, when a person sees a person or place in his imagination, he may simply be doing what he does in the presence of the person or place. Both “reminiscing” and “remembering” once meant “being mindful of again” or “bringing again to mind”—in other words, seeing again as one once saw. Explicit techniques of “calling to mind” are techniques of strengthening perceptual behavior, as we shall see in Chapter 7.
Behaviorism has been accused of “relegating one of the paramount concerns of the earlier psychologists—the study of the image—to a position of not just neglect, but disgrace.” I believe, on the contrary, that it offers the only way in which the subject of imaging or imagining can be put in good order.
Seeing in the absence of the thing seen is familiar to almost everyone, but the traditional formulation is a metaphor. We tend to act to produce stimuli which are reinforcing when seen. If we have found the city of Venice reinforcing (we refer to one reinforcing effect when we call it beautiful), we may go to Venice in order to be thus reinforced. If we cannot go, we may buy pictures of Venice—realistic pictures in color of its most beautiful aspects, although a black-and-white sketch may be enough. Or we may see Venice by reading about it if we have acquired the capacity to visualize while reading. (Technology has made it much easier to see reinforcing things in their presence and hence has reduced the chance to see them in their absence. Two or three generations ago a child read, or was read to, from books with few or no illustrations; today he watches television or reads books with colored pictures on every page, and he is therefore much less likely to acquire a repertoire of seeing under the control of verbal stimuli.) With no external support whatsoever, we may simply “see Venice” because we are reinforced when we do so. We say that we daydream about Venice. The mistake is to suppose that because we create physical stimuli which enable us to see Venice more effectively by going to Venice or buying a picture, we must therefore create mental stimuli to be seen in memory. All we need to say is that if we are reinforced for seeing Venice, we are likely to engage in that behavior—that is, the behavior of seeing Venice—even when there is very little in the immediate setting which bears a resemblance to the city. According to one dictionary, fantasy is defined as “the act or function of forming images or representations in direct perception or in memory,” but we could say as well that it is the act or function of seeing in direct perception or in memory.
We may also see a thing in its absence, not because we are immediately reinforced when we do so, but because we are then able to engage in behavior which is subsequently reinforced. Thus, we may see Venice in order to tell a friend how to find his way to a particular part of the city. If we were together in the city itself, we might take him along a given route, but we can “take ourselves along the route visually” when we are not there and describe it to him. We can do so more effectively by pointing to a map or a sketch of the route, but we do not consult a “cognitive map” when we describe what we see in “calling the city to mind.” Knowing a city means possessing the behavior of getting about in it; it does not mean possessing a map to be followed in getting about. One may construct such a map from the actual city or by seeing the city when absent from it, but visualizing a route through a city in order to describe it to a friend is seeing as (not what) one sees in going through the city.
Claude Bernard might also have said that it is impossible to get animals to report the things they are imagining, but there is no reason why the contingencies under which a person sees things which are not there should not be effective with other species. It is possible to get animals to respond to after-images, and by increasing deprivation we can induce a pigeon to respond to a square “as if it were a triangle.” There is no reason why, with such measures, we could not get it to respond to a blank surface when it has previously been reinforced only when the surface had a triangle projected upon it. The design of “verbal” contingencies which would permit it to tell us that it “saw” a triangle would be an interesting exercise.
A person is changed by the contingencies of reinforcement under which he behaves; he does not store the contingencies. In particular, he does not store copies of the stimuli which have played a part in the contingencies. There are no “iconic representations” in his mind; there are no “data structures stored in his memory”; he has no “cognitive map” of the world in which he has lived. He has simply been changed in such a way that stimuli now control particular kinds of perceptual behavior.
Seeing in the absence of the thing seen is most dramatically exemplified in dreaming when asleep. Current stimulation is then minimally in control, and a person’s history and resulting states of deprivation and emotion get their chance. Freud emphasized the significance of wishes and fears plausibly inferred from dreaming, but unfortunately he was responsible for emphasizing the distinction between seeing and what is seen. The dreamer engaged in dream work; he staged the dream as a theatrical producer stages a play and then took his place in the audience and watched it. But dreaming is perceptual behavior, and the difference between behavior when asleep and when awake, either in or out of a relevant setting, is simply a difference in the controlling conditions.
Rapid eye movements during sleep seem to confirm this interpretation. When most actively dreaming, people move their eyes about as if they were observing a visual presentation. (The middle-ear muscles also seem to move during dreams involving auditory perception.) It has been argued that eye movement, as well as ear-muscle movement, show that “physiological input” affects dreaming, but such behavior is quite clearly a physiological output. We can scarcely suppose that the iconic representations observed in dreaming are under the eyelids or in the outer ear.
There are many ways of getting a person to see when there is nothing to be seen, and they can all be analyzed as the arrangement of contingencies which strengthen perceptual behavior. Certain practices in behavior therapy, in which the patient is asked to imagine various conditions or events, have been criticized as not genuinely behavioral because they make use of images. But there are no images in the sense of private copies, there is perceptual behavior; and the measures taken by the psychotherapist are designed to strengthen it. A change takes place in the patient’s behavior if what he sees (hears, feels, and so on) has the same positively or negatively reinforcing effect as if he were seeing the things themselves. It is seldom if ever enough simply to instruct the patient to “have feelings,” to ask him to feel sexually excited or nauseated, but he may be shown pornographic or nauseating material or be asked to “visualize as clearly as possible” a sexual or disgusting episode.
That a person may see things when there is nothing to be seen must have been a strong reason why the world of the mind was invented. It was hard enough to imagine how a copy of the current environment could get into the head where it could be “known,” but there was at least a world outside which might account for it. But pure images seem to indicate a pure mind stuff. It is only when we ask how either the world or a copy of the world is seen that we lose interest in copies. Seeing does not require a thing seen.
We saw in Chapter 4 that the word “mind” is sometimes a mere synonym for the person who acts. It can also stand for the person who perceives. When a person is out of touch with reality, his mind is said to be wandering, or possibly absent. The verb “to mind” often means simply to respond. We warn someone to mind the low ceiling, meaning simply that he should see and respond to it. In this sense we ask someone to mind the children, and he may complain that the children do not mind him.
Mind is also sometimes simply the place in which things are seen. Things “come to mind” or are “called to mind,” and one who is suffering a delusion may be told that “it’s all in your mind,” as distinct from being in the real world. As the place in which things are perceived, mind is closely associated with copy theory and was an important part of the psychology of conscious content. When operationism led to the study of the process of discrimination rather than of sensations, a person was regarded as looking at or listening to the real world. He was no longer reporting his perceptions or sensations; he was reporting stimuli. The world was back where it belonged.
The issue is critical when we turn to the difference between seeing a thing and seeing that one is seeing it. If there are no copies of things inside the body at any time, then all that can be seen introspectively is the act of seeing, and this is what one reports when asked, “Do you see that?” It is still possible, however, to discriminate between things which are there or not there to be seen. I could be said to know that this sheet of paper is really there because I pick up a pen and write on it, and that the bright after-image which bothers me is not there because I do not try to brush it away. I have learned the difference between two kinds of seeing. The thirsty man does not reach for the fantasied glass of water, but the dreamer does not know that what he is seeing is “not really there,” and he responds as fully as a person who is asleep can. (Introspective knowledge of dreaming is weak or lacking because the conditions needed for self-observation are lacking, and when such self-knowledge survives into the waking state, it usually disappears quickly as one forgets one’s dreams.) It is also possible to know that you have seen something before. We re-cognize what we have once cognized. In a déjà vu this feature of self-knowledge is defective.
Other kinds of self-knowledge about stimulus control become available when we analyze the contingencies which control our behavior.