What is meant by the “self” in self-control or self-knowledge? When a man jams his hands into his pockets to keep himself from biting his nails, who is controlling whom? When he discovers that a sudden mood must be due to a glimpse of an unpleasant person, who discovers whose mood to be due to whose visual response? Is the self which works to facilitate the recall of a name the same as the self which recalls it? When a thinker teases out an idea, is it the teaser who also eventually has the idea?
The self is most commonly used as a hypothetical cause of action. So long as external variables go unnoticed or are ignored, their function is assigned to an originating agent within the organism. If we cannot show what is responsible for a man’s behavior, we say that he himself is responsible for it. The precursors of physical science once followed the same practice, but the wind is no longer blown by Aeolus, nor is the rain cast down by Jupiter Pluvius. Perhaps it is because the notion of personification is so close to a conception of a behaving individual that it has been difficult to dispense with similar explanations of behavior. The practice resolves our anxiety with respect to unexplained phenomena and is perpetuated because it does so.
Whatever the self may be, it is apparently not identical with the physical organism. The organism behaves, while the self initiates or directs behavior. Moreover, more than one self is needed to explain the behavior of one organism. A mere inconsistency in conduct from one moment to the next is perhaps no problem, for a single self could dictate different kinds of behavior from time to time. But there appear to be two selves acting simultaneously and in different ways when one self controls another or is aware of the activity of another.
The same facts are commonly expressed in terms of “personalities.” The personality, like the self, is said to be responsible for features of behavior. For example, delinquent behavior is sometimes attributed to a psychopathic personality. Personalities may also be multiple. Two or more personalities may appear in alternation or concurrently. They are often in conflict with each other, and one may or may not be aware of what the other is doing.
Multiple selves or personalities are often said to be systematically related to each other. Freud conceived of the ego, superego, and id as distinguishable agents within the organism. The id was responsible for behavior which was ultimately reinforced with food, water, sexual contact, and other primary biological reinforcers. It was not unlike the selfish, aggressive “Old Adam” of Judeo-Christian theology, preoccupied with the basic deprivations and untouched by similar requirements on the parts of others. The superego—the “conscience” of Judeo-Christian theology—was responsible for the behavior which controlled the id. It used techniques of self-control acquired from the group. When these were verbal, they constituted “the still small voice of conscience.” The superego and the id were inevitably opposed to each other, and Freud conceived of them as often in violent conflict. He appealed to a third agent—the ego—which, besides attempting to reach a compromise between the id and the superego, also dealt with the practical exigencies of the environment.
We may quarrel with any analysis which appeals to a self or personality as an inner determiner of action, but the facts which have been represented with such devices cannot be ignored. The three selves or personalities in the Freudian scheme represent important characteristics of behavior in a social milieu. Multiple personalities which are less systematically related to each other serve a similar function. A concept of self is not essential in an analysis of behavior, but what is the alternative way of treating the data?
The best way to dispose of any explanatory fiction is to examine the facts upon which it is based. These usually prove to be, or suggest, variables which are acceptable from the point of view of scientific method. In the present case it appears that a self is simply a device for representing a functionally unified system of responses. In dealing with the data, we have to explain the functional unity of such systems and the various relationships which exist among them.
The unity of a self. A self may refer to a common mode of action. Such expressions as “The scholar is Man Thinking” or “He was a better talker than plumber” suggest personalities identified with topographical subdivisions of behavior. In a single skin we find the man of action and the dreamer, the solitary and the social spirit.
On the other hand, a personality may be tied to a particular type of occasion—when a system of responses is organized around a given discriminative stimulus. Types of behavior which are effective in achieving reinforcement upon occasion A are held together and distinguished from those effective upon occasion B. Thus one’s personality in the bosom of one’s family may be quite different from that in the presence of intimate friends.
Responses which lead to a common reinforcement, regardless of the situation, may also comprise a functional system. Here the principal variable is deprivation. A motion to adjourn a meeting which has run through the lunch hour may show “the hungry man speaking.” One’s personality may be very different before and after a satisfying meal. The libertine is very different from the ascetic who achieves his reinforcement from the ethical group, but the two may exist side by side in the same organism.
Emotional variables also establish personalities. Under the proper circumstances the timid soul may give way to the aggressive man. The hero may struggle to conceal the coward who inhabits the same skin.
The effects of drugs upon personality are well known. The euphoria of the morphine addict represents a special repertoire of responses the strength of which is attributable to an obvious variable. The alcoholic wakes on the morrow a sadder and wiser man.
It is easy to overestimate the unity of a group of responses, and unfortunately personification encourages us to do so. The concept of a self may have an early advantage in representing a relatively coherent response system, but it may lead us to expect consistencies and functional integrities which do not exist. The alternative to the use of the concept is simply to deal with demonstrated covariations in the strength of responses.
Relations among selves. Organized systems of responses may be related to each other in the same way as are single responses and for the same reasons (Chapters XIV, XV, XVI). For example, two response systems may be incompatible. If the relevant variables are never present at the same time, the incompatibility is unimportant. If the environment of which behavior is a function is not consistent from moment to moment, there is no reason to expect consistency in behavior. The pious churchgoer on Sunday may become an aggressive, unscrupulous businessman on Monday. He possesses two response systems appropriate to different sets of circumstances, and his inconsistency is no greater than that of the environment which takes him to church on Sunday and to work on Monday. But the controlling variables may come together; during a sermon, the churchgoer may be asked to examine his business practices, or the businessman may engage in commercial transactions with his clergyman or his church. Trouble may then arise. Similarly, if an individual has developed different repertoires with family and friends, the two personalities come into conflict when he is with both at the same time. Many of the dramatic struggles which flood the literature on multiple personalities can be accounted for in the same way.
More systematic relations among personalities arise from the controlling relations discussed in Chapters XV and XVI. In self-control, for example, the responses to be controlled are organized around certain immediate primary reinforcements. To the extent that competition for reinforcement makes this behavior aversive to others—and to this extent only—we may refer to an anti-social personality, the id or Old Adam. On the other hand, the controlling behavior engendered by the community consists of a selected group of practices evolved in the history of a particular culture because of their effect upon antisocial behavior. To the extent that this behavior works to the advantage of the community—and again to this extent only—we may speak of a unitary conscience, social conscience, or superego. These two sets of variables account, not only for the membership of each group of responses, but for the relation between them which we describe when we say that one personality is engaged in controlling the other. Other kinds of relations between personalities are evident in the processes of making a decision, solving a problem, or creating a work of art.
An important relation between selves is the self-knowledge of Chapter XVII. The behavior which we call knowing is due to a particular kind of differential reinforcement. In even the most rudimentary community such questions as “What did you do?” or “What are you doing?” compel the individual to respond to his own overt behavior. Probably no one is completely unselfconscious in this sense. At the other extreme an advanced and relatively nonpractical society produces the highly introspective or introverted individual, whose repertoire of self-knowledge extends to his covert behavior—a repertoire which in some cultures may be almost nonexistent. An extensive development of self-knowledge is common in certain Eastern cultures and is emphasized from time to time in those of the West—for example, in the culte du moi of French literature. An efficient repertoire of this sort is sometimes set up in the individual for purposes of therapy. The patient under psychoanalysis may become highly skilled in observing his own covert behavior.
When an occasion arises upon which a report of the organism’s own behavior, particularly at the covert level, is likely to be reinforced, the personality which makes the report is a specialist trained by a special set of contingencies. The self which is concerned with self-knowing functions concurrently with the behavioral system which it describes. But it is sometimes important to ask whether the selves generated by other contingencies “know about each other.” The literature on multiple personalities raises the question as one of “continuity of memory.” It is also an important consideration in the Freudian scheme: to what extent, for example, is the superego aware of the behavior of the id? The contingencies which set up the superego as a controlling system involve stimulation from the behavior of the id, but they do not necessarily establish responses of knowing about the behavior of the id. It is perhaps even less likely that the id will know about the superego. The ego can scarcely deal with conflicts between the other selves without responding to the behavior attributed to them, but this does not mean that the ego possesses a repertoire of knowing about such behavior in any other sense.
One of the most striking facts about self-knowledge is that it may be lacking. Several cases deserve comment.
A man may not know that he has done something. He may have behaved in a given way, perhaps energetically, and nevertheless be unable to describe what he has done. Examples range all the way from the unnoticed verbal slip to extended amnesias in which large areas of earlier behavior cannot be described by the individual himself. The possibility that the behavior which cannot be described may be covert raises an interesting theoretical problem, since the existence of such behavior must be inferred, not only by the scientist, but by the individual himself. We have seen that a mathematician frequently cannot describe the process through which he solves a problem. Although he may report the preliminary stages of his investigation, his arrangement of materials, and many tentative solutions, he may not be able to describe the self-manipulation which presumably preceded the required response which he suddenly emits. It is not always necessary to infer that other behavior has actually occurred, but under certain circumstances this inference may be justified. Since authenticated overt behavior sometimes cannot be reported by the individual, we have no reason to question the possibility of a covert parallel.
A man may not know that he is doing something. Absent-minded conduct, unconscious mannerisms, and mechanically habitual behavior are common examples. More dramatic is automatic writing, in which behavior taking place at the moment cannot be described by the “rest of the organism.”
A man may not know that he tenas to, or is going to, do something. He may be unaware of aggressive tendencies, of unusual predilections, or of the high probability that he will follow a given course of action.
A man may not recognize the variables of which his behavior is a function. In the Verbal Summator, for example, the subject often supposes himself to be repeating a verbal stimulus when it is easy to identify variables lying elsewhere in his environment or history which account for the behavior (Chapter XIV). Projective tests are used for diagnostic purposes just because they reveal variables which the individual himself cannot identify.
These phenomena are often viewed with surprise. How can the individual fail to observe events which are so conspicuous and so important? But perhaps we should be surprised that such events are observed as often as they are. We have no reason to expect discriminative behavior of this sort unless it has been generated by suitable reinforcement. Self-knowledge is a special repertoire. The crucial thing is not whether the behavior which a man fails to report is actually observable by him, but whether he has ever been given any reason to observe it.
Self-knowledge may, nevertheless, be lacking where appropriate reinforcing circumstances have prevailed. Some instances may be dismissed without extended comment. For example, the stimuli supplied by behavior may be weak. One may be “unaware” of a facial expression because of the inadequacy of the accompanying self-stimulation. The subject in an experiment on muscle-reading may not be aware of the slight responses which the reader detects and uses in getting the subject to “tell” him the location of a hidden object. The functional relation between behavior and a relevant variable is especially likely to be of subtle physical dimensions. A face in the crowd may be clear enough as a stimulus to generate a mood, but the fact that it has done so may still not be noted. This does not mean that the stimuli are below threshold, for they may be brought into control in other ways. When we point out some part of the behavior of an individual, an occasion is established under which special reinforcement is accorded a discriminative reaction. The fact that the individual then responds to his behavior is what we mean by saying that he was “able to do so” in the first place.
Another case of “not knowing what one is doing” is explained by the principle of prepotency. In the heat of battle there may be no time to observe one’s behavior, since strong responses conflict with the discriminative response. Self-knowledge may also be lacking in certain states of satiation and in sleep. One may talk in one’s sleep or behave in other ways “without knowing it.” Behavior under the influence of drugs—for example, alcohol—may also occur with a minimum of self-observation. The effect of alcohol in reducing the behavior of self-knowledge may be similar to that in reducing the response to the conditioned aversive stimuli characteristic of guilt or anxiety.
It has been argued that one cannot describe behavior after the fact which one could not have described at the time. This appears to explain our inability to recall the events of infancy, since the behavior of the infant occurs before a repertoire of self-description has been set up and therefore too soon to control such a repertoire. The same explanation should apply to behavior unnoticed in the heat of battle. However, it is possible that the rearousal of response on the pattern of the conditioned reflex may supply the basis for a description. In any case it is sometimes impossible to describe earlier behavior which could have been described, and perhaps was described, at the time it was emitted. An important reason why a description may be lacking has still to be considered.
Repression. We have seen that punishment makes the stimuli generated by punished behavior aversive. Any behavior which reduces that stimulation is subsequently reinforced automatically. Now, among the kinds of behavior most likely to generate conditioned aversive stimuli as the result of punishment is the behavior of observing the punished act or of observing the occasion for it or any tendency to execute it. As the result of punishment, not only do we engage in other behavior to the exclusion of punished forms, we engage in other behavior to the exclusion of knowing about punished behavior, in the sense of Chapter XVII. This may begin simply as “not liking to think about” behavior which has led to aversive consequences. It may then pass into the stage of not thinking about it and eventually reach the point at which the individual denies having behaved in a given way, in the face of proof to the contrary.
The result is commonly called repression. As we saw in Chapters XII and XIV, the individual may repress behavior simply in the sense of engaging in competing forms, but we must now extend the meaning of the term to include the repression of knowing about punished behavior. This is a much more dramatic result, to which the term “repression” is sometimes confined. The same formulation applies, however. We do not appeal to any special act of repression but rather to competing behavior which becomes extremely powerful because it avoids aversive stimulation.
It is not always knowledge of the form of a response which is repressed, because punishment is not always contingent upon form. Aggressive behavior, for example, is not punished in warfare. Imitative behavior is not often punished so long as it is actually under the control of similar behavior on the part of others. For example, when we emit obscene or blasphemous behavior in testifying to an instance of it on the part of someone else, our testimony may not be entirely free of conditioned aversive consequences, and we may avoid testifying if possible; but the aversive stimulation will be much less than that aroused by the same behavior when it is not imitative. In experiments with the Verbal Summator a subject will often emit aggressive, ungrammatical, obscene, or blasphemous responses so long as he remains convinced that he is correctly repeating speech patterns on a phonograph record. He has been told to repeat what he hears and punishment is not contingent upon the form of his behavior under these circumstances, especially if a few objectionable samples are first presented clearly. As soon as he is told that there are no comparable speech patterns on the record, however, this type of response usually becomes much less frequent. The individual must now, so to speak, take the responsibility for the aggression, obscenity, and so on. In other words, his behavior is now of a form and under a controlling relation upon which punishment is contingent. In such a case the subject will often refuse to acknowledge that earlier stimuli were not of the form he reported.
A variation on the repression of a controlling relationship is sometimes called “rationalization.” The aversive report of a functional relation may be repressed by reporting a fictitious relationship. Instead of “refusing to recognize” the causes of our behavior, we invent acceptable causes. If an aggressive attack upon a child is due to emotional impulses of revenge, it is usually punished by society; but if it is emitted because of supposed consequences in shaping the behavior of the child in line with the interests of society, it goes unpunished. We may conceal the emotional causes of our aggressive behavior, either from ourselves or from others, by arguing that the child ought to learn what sort of effect he is having on people. We spank the child “for his own good.” In the same way we may delight in carrying bad news to someone we dislike “because the sooner he knows it the better.” It is not the aggressive response which is repressed, but the response of knowing about the aggressive tendency. The rationalization is the repressing response which is successful in avoiding the conditioned aversive stimulation generated by punishment.
In Chapter XIV, we saw that a group of responses strengthened by a common variable might not all have the same aversive consequences and that as a result of the principle of summation the response with the least aversive consequence would emerge. In more general terms, we may note that the property of a response which achieves reinforcement need not coincide with the property upon which punishment is based. A response may appear, therefore, which achieves reinforcement while avoiding punishment. For example, the visual stimulation of a nude figure may be reinforcing because of previous connection with powerful sexual reinforcement. But in many societies the behavior of looking at such figures is severely punished. Under special circumstances—as, for example, in an art museum—it is possible to engage in this behavior and escape punishment. The behavior of the artist may show a similar compromise. His art must not be pornographic or too sensual, but while staying within certain limits which avoid punishment, it may nevertheless be successfully reinforcing for biological reasons. In fantasy the individual makes a similar compromise between seeing certain objects or patterns and avoiding aversive stimulation: he daydreams in a given area but in such a way that he does not generate too much guilt.
A symbol, as the term was used by Freud in the analysis of dreams and art, is any temporal or spatial pattern which is reinforcing because of similarity to another pattern but escapes punishment because of differences. Thus an abstract sculpture is symbolic of the human form if it is reinforcing because of resemblances and if the artist would, in the absence of punishment, have emphasized the resemblances. A musical composition symbolizes sexual behavior if it is reinforcing because of a similarity in temporal pattern and if it is emitted in place of such behavior because it is different enough to escape punishment.
The principal realm of the symbol is the dream which occurs when we are asleep. This is a species of private event which is extremely difficult to study and is, therefore, the subject of much conflicting discussion. In a dream the individual engages in private discriminative behavior, in the sense of Chapter XVII. He sees, hears, feels, and so on, in the absence of the usual stimuli. Controlling variables may sometimes be discovered in the immediate environment or in the recent history of the individual. In the perseverative dream, for example, one may dream of driving a car if one has been driving for many hours. More often, however, the relevant variables are harder to identify. The attempt to do so is commonly called the interpretation of dreams. Freud could demonstrate certain plausible relations between dreams and variables in the life of the individual. The present analysis is in essential agreement with his interpretation. The individual is strongly disposed to engage in behavior which achieves such reinforcements as sexual contact or the infliction of damage upon others. These kinds of behavior, however, are precisely the sort most likely to be punished. As a result the individual not only does not overtly engage in such behavior, he cannot engage in it covertly or see himself engaging in it covertly without automatic aversive self-stimulation. In the symbolic dream and in artistic or literary behavior, however, he may engage in discriminative behavior which is strengthened through stimulus- or response-induction by the same variables but which is not liable to punishment. It is often said or implied that some skillful agent engages in a sort of “dream-work” to produce this result, but the result follows automatically from the discrepancy between the properties of behavior upon which reinforcement and punishment are contingent.