Int. J. Psycho-Anal., 26:137-143 (IJP)
Primitive Emotional Development1
D. W. Winnicott
It will be clear at once from my title that I have chosen a very wide subject. All I can attempt to do is to make a preliminary personal statement, as if writing the introductory chapter to a book.
I shall not first give a historical survey and show the development of my ideas from the theories of others, because my mind does not work that way. What happens is that I gather this and that, here and there, settle down to clinical experience, form my own theories and then last of all interest myself in looking to see where I stole what. Perhaps this is as good a method as any.
About primitive emotional development there is a great deal that is not known or properly understood, at least by me, and it could well be argued that this discussion ought to be postponed 5 or 10 years. Against this there is the fact that misunderstandings constantly recur in the Society's scientific meetings, and perhaps we shall find we do know enough already to prevent some of these misunderstandings by a discussion of these primitive emotional states.
Primarily interested in the child patient, and the infant, I decided that I must study psychosis in analysis. I have had about a dozen psychotic adult patients, and half of these have been rather extensively analysed. This happened in the war, and I might say that I hardly noticed the blitz, being all the time engaged in analysis of patients who are notoriously and maddeningly oblivious of bombs, earthquakes and floods.
As a result of this work I have a great deal to communicate and to bring into alignment with current theories, and perhaps this paper may be taken as a beginning.
By listening to what I have to say, and criticizing, you help me to take my next step, which is the study of the sources of my ideas, both in clinical work and in the published writings of analysts. It has in fact been extremely difficult to keep clinical material out of this paper, which I wished nevertheless to keep short so that there might be plenty of time for discussion.
The following is my highly condensed personal statement.
First I must prepare the way. Let me try to describe different types of psycho-analysis. It is possible to do the analysis of a suitable patient taking into account almost exclusively that person's personal relation to people, along with the conscious and unconscious phantasies that enrich and complicate these relationships between whole persons. This is the original type of psycho-analysis. In the last two decades we have been shown how to develop our interest in phantasy, and how the patient's own phantasy about his inner organization and its origin in instinctual experience is important as such.2We have been shown further that in certain cases it is this, the patient's phantasy about his inner organization,
My especial thanks are due to Dr. W. Clifford M. Scott for his help both in the work on which this paper is based and in the preparation of the paper itself.
1 Read before the British Psycho-Analytical Society, November 28, 1945.
2 Chiefly through the work of Melanie Klein.
- 137 -
that is vitally important, so that the analysis of depression and the
defences against depression cannot be done on the basis only of consideration of
the patient's relations to real people and his phantasies about them. This new
emphasis on the patient's phantasy of himself opened up the wide field of
analysis of hypochondria in which the patient's phantasy about his inner world
includes the phantasy that this is localized inside his own body. It became
possible for us to relate, in analysis, the qualitative changes in the
individual's inner world to his instinctual experiences. The quality of these
instinctual experiences accounted for the good and bad nature, as well as the
existence, of what is inside.
This work was a natural progression of psycho-analysis; it involved new understanding but not new technique. It quickly led to the study and analysis of still more primitive relationships, and it is these that I wish to discuss in this paper. The existence of still more primitive object relationships has never been in doubt.
I have said that no modification in Freud's technique was needed for the extension of analysis to cope with depression and hypochondria. It is also true, according to my experience, that the same technique can take us to still more primitive elements, provided of course that we take into consideration the changes in the transference situation inherent in such work.
I mean by this that a patient needing analysis of ambivalence in external relationships has a different phantasy of his analyst and the analyst's work from the one who is depressed. In the former case the analyst's work is thought of as done out of love for the patient, hate being deflected on to hateful things. The depressed patient requires of his analyst the understanding that the analyst's work is to some extent his effort to cope with his own (the analyst's) depression, or shall I say guilt and grief resultant from the destructive elements in his own (the analyst's) love. To progress further along these lines, the patient who is asking for help in regard to his primitive, pre-depressive relationship to objects needs his analyst to be able to see the analyst's undisplaced and co-incident love and hate of him. In such cases the end of the hour, the end of the analysis, the rules and regulations, these all come in as important expressions of hate, just as the good interpretations are expressions of love, and symbolical of good food and care. This theme could be developed extensively and usefully.
Before embarking directly on a description of primitive emotional development I should also like to make it clear that the analysis of these primitive relationships cannot be undertaken except as an extension of the analysis of depression. It is certain that these primitive types of relationship in so far as they appear in children and adults come as a flight from the difficulties arising out of the next stages, after the classical conception of regression. It is right for a student analyst to learn first to cope with ambivalence in external relationships and with simple repression and then to progress to the analysis of the patient's phantasy about the inside and outside of his personality, and the whole range of his defences against depression, including the origins of the persecutory elements. These latter things the analyst can surely find in any analysis, but it would be useless or harmful for him to cope with principally depressive relationships unless he was fully prepared to analyse straightforward ambivalence. It is at least as true that it is useless and even dangerous to analyse the primitive pre-depressive relationships, and to interpret them as they appear in the transference, unless the analyst is fully prepared to cope with the depressive position, the defences against depression, and the persecutory ideas which appear for interpretation as the patient progresses.
I have more preparatory remarks to make. It has often been noted that, at five to six months, a change occurs in infants which makes it more easy than before for us to refer to their emotional development in the terms that apply to human beings generally. Anna Freud makes rather a special point of this and implies that in her view the tiny infant is concerned more with certain care-aspects than with specific people. Bowlby recently expressed the view that infants before six months are not particular, so that separation from their mother does not affect them in the same way as it does after six months. I myself have previously stated that infants reach something at six months, so that whereas many five months' infants grasp an object and put it to the mouth, it is not till six months that the average infant starts to follow this up by deliberately dropping the object as part of his play with it.
In specifying five to six months we need not try to be too accurate. If in a certain case a baby of three or even two months or even less reaches the stage of development that it is convenient in general description to place at five months, no harm will be done.
In my opinion the stage we are describing, and I think one may accept this description, is a very important one. To some extent it is an affair of physical development, for the infant at five months becomes skilled to the extent that he grasps an object he sees, and can soon get it to his mouth. He could not have done this earlier. (Of course he may have wanted to. There is no exact parallel between skill and wish, and we know that many physical advances, such as the ability to walk, are often held up till emotional development releases physical attainment. Whatever the physical side of the matter, there is also the emotional.) We
- 138 -
can say that at this stage a baby becomes able in his play to show that he
can understand he has an inside, and that things come from outside. He shows he
knows that he is enriched by what he incorporates (physically and psychically).
Further, he shows that he knows he can get rid of something when he has got from
it what he wants from it. All this represents a tremendous advance. It is at
first only reached from time to time, and every detail of this advance can be
lost as a regression because of anxiety.
The corollary of this is that now the infant assumes that his mother also has an inside, one which may be rich or poor, good or bad, ordered or muddled. He is therefore starting to be concerned with the mother and her sanity and her moods. In the case of many infants there is a relationship as between whole persons at six months. Now when a human being feels he is a person related to people, he has already travelled a long way in primitive development.
Our task is to examine what goes on in the infant's feelings and personality before this stage which we recognize at five to six months, but which may be reached later or earlier.
There is also this question: how early do important things happen? For instance, does the unborn child have to be considered? And if so, at what age after conception does psychology come in? I would answer that if there is an important stage at five to six months there is also an important stage round about birth. My reason for saying this is the great differences that can be noticed if the baby is pre-mature or post-mature. I suggest that at the end of nine months' gestation an infant becomes ripe for emotional development, and that if an infant is post-mature he has reached this stage in the womb and one is therefore forced to consider his feelings before and during birth. On the other hand a premature infant is not experiencing much that is vital till he has reached the age at which he should have been born, that is to say some weeks after birth. At any rate this forms a basis for discussion.
Another question is: psychologically speaking, does anything matter before five to six months? I know that the view is quite sincerely held in some quarters that the answer is 'no'. This view must be given its due, but it is not mine.
The main object of this paper is to present the thesis that the early emotional development of the infant, before the infant knows himself (and therefore others) as the whole person he is (and they are), is vitally important: indeed that here are the clues to the psychopathology of psychoses.
There are three processes which seem to me to start very early: (1) integration, (2) personalization, and (3), following these, the appreciation of time and space and other properties of reality-in short, realization.
A great deal that we tend to take for granted had a beginning and a condition out of which it developed. For instance, many analyses sail through to completion without time being ever in dispute. But a boy of nine who loved to play with Ann, aged two, was acutely interested in the expected new baby. He said: 'When the new baby's born will he be born before Ann?' For him time-sense is very shaky. Again, a psychotic patient could not adopt any routine because if she did she had no idea on a Tuesday whether it was last week, or this week, or next week.
The localization of self in one's own body is often assumed, yet a psychotic patient in analysis came to recognize that as a baby she thought her twin at the other end of the pram was herself. She even felt surprised when her twin was picked up and yet she remained where she was. Her sense of self and other-than-self was undeveloped.
Another psychotic patient discovered in analysis that most of the time she lived in her head, behind her eyes. She could only see out of her eyes as out of windows and so was not aware of what her feet were doing, and in consequence she tended to fall into pits and to trip over things. She had no 'eyes in her feet'. Her personality was not felt to be localized in her body, which was like a complex engine that she had to drive with conscious care and skill. Another patient, at times, lived in a box 20 yards up, only connected with her body by a slender thread. In all practices examples of these failures in primitive development occur daily, and by them we may be reminded of the importance of such processes as integration, personalization and relization.
It may be assumed that at the theoretical start the personality is unintegrated, and that in regressive disintegration there is a primary state to which regression leads. We postulate a primary unintegration.
Disintegration of personality is a well-known psychiatric condition, and its psychopathology is highly complex. Examination of these phenomena in analysis, however, shows that the primary unintegrated state provides a basis for disintegration, and that delay or failure in respect of primary integration predisposes to disintegration as a regression, or result of failure in other types of denfence.
In any case, integration starts right away at the beginning of life, and in our work we can never take it for granted. We have to account for it and watch its fluctuations.
An example of unintegration phenomena is provided by the very common experience of the patient who proceeds to give every detail of the week-end and feels contented at the end if everything has been said, though the analyst feels that no analytic work has been done. Sometimes we
- 139 -
must interpret this as the patient's need to be known in all his bits and
pieces by one person, the analyst. To be known means to feel integrated at least
in the person who knows one. This is the ordinary stuff of infant life, and an
infant who has had no one person to gather his bits together starts with a
handicap in his own self-integrating task, and perhaps he cannot succeed, or at
any rate cannot maintain integration with confidence.
The tendency to integrate is helped by two sets of experience: the technique of infant-care whereby an infant is kept warm, handled and bathed and rocked and named, and also the acute instinctual experiences which tend to gather the personality together from within. Many infants are well on the way toward integration during certain periods of the first 24 hours of life. In others the process is delayed or set-backs occur because of early inhibition of making greedy attacks. There are long stretches of time in a normal infant's life in which a baby does not mind whether he is many bits or one whole being, or whether he lives in his mother's face or in his own body, provided that from time to time he comes together and feels something. Later I will try to explain why disintegration is frightening, whereas unintegration is not.
In regard to environment, bits of nursing technique, faces seen and sounds heard, and smells smelt are only gradually pieced together into one being to be called mother. In the transference situation in analysis of psychotics we get the clearest proof that the psychotic states of unintegration had a natural place at a sufficiently primitive stage of the emotional development of the individual.
It is sometimes assumed that in health the individual is always integrated, as well as living in his own body, and able to feel that the world is real. There is, however, much sanity that has a symptomatic quality, being charged with fear or denial of madness, fear or denial of the innate capacity of every human being to become unintegrated, depersonalized, and to feel that the world is unreal. Sufficient lack of sleep produces these conditions in anyone.3
Equally important with integration is the development of the feeling that one's person is in one's body. Again it is instinctual experience and repeated quiet experiences of body-care that gradually build up what may be called satisfactory personalization. And as with disintegration so also the depersonalization phenomena of psychosis relate to early personalization delays.
Depersonalization is a common thing in adults and in children, it is often hidden for instance in what is called deep sleep and in prostration attacks with corpse-like pallor:?She's miles away, ' people say, and they're right.
A problem related to that of personalization is that of the imaginary companions of childhood. These are not simple phantasy constructions. Study of the future of these imaginary companions (in analysis) shows that they are sometimes other selves of a highly primitive type. I cannot here formulate a clear statement of what I mean, and it would be out of place for me to explain this detail at length now. I would say, however, that this very primitive and magical creation of imaginary companions is easily used as a defence, as it magically by-passes all the anxieties associated with incorporation, digestion, retention and expulsion.
Out of the problem of unintegration comes another, that of dissociation. Dissociation can usefully be studied in its initial or natural forms. According to my view there grows out of unintegration a series of what are then called dissociations, which arise owing to integration being incomplete or partial. Take the quiet and the excited states. I think an infant cannot be said to be aware at the start that while feeling this and that in his cot or enjoying the skin stimulations of bathing, he is the same as himself screaming for immediate satisfaction, possessed by an urge to get at and destroy something unless satisfied by milk. This means that he does not know at first that the mother he is building up through his quiet experiences is the same as the power behind the breasts that he has in his mind to destroy.
Also I think there is not necessarily an integration between a child asleep and a child awake. This integration comes in the course of time. Once day-dreams are remembered and even conveyed somehow to a third person, the dissociation is broken down a little; but some people never clearly remember their dreams, and children depend very much on adults for getting to know their dreams. It is normal for small children to have anxiety dreams and terrors. At these times children need someone to help them to remember what they dreamed. It is a valuable experience whenever a dream is both dreamed and remembered, precisely because of the breakdown of dissociation that this represents. However complex such a dissociation may be in child or adult, the fact remains that it can start in the natural alternation of the sleeping and awake states, dating from birth.
In fact the waking life of an infant can be perhaps described as a gradually developing dissociation from the sleeping state.
Artistic creation gradually takes the place of dreams or supplements them, and is vitally important
3Through artistic expression we can hope to keep in touch with our primitive selves. It is from here that the most intense feelings and even fearfully acute sensations derive, and we are poor indeed if we are only sane.
- 140 -
for the welfare of the individuals and therefore for mankind.
Dissociation is an extremely widespread defence mechanism, and leads to surprising results. For instance urban life is a dissociation, a serious one for civilization. Also war and peace. The extremes in mental illness are well known. In childhood dissociation appears for instance in such common conditions as somnambulism, incontinence of faces, in some forms of squinting, etc. It is very easy to miss dissociation when assessing a personality.
Let us now assume integration. If we do, we reach another enormous subject, the primary relation to external reality. In ordinary analyses we can and do take for granted this step in emotional development, which is highly complex and which, when it is made, represents a big advance in emotional development, yet is never finally made and settled. Many cases that we consider unsuitable for analysis are unsuitable indeed if we cannot deal with the transference difficulties that belong to an essential lack of true relation to external reality. If we allow analysis of psychotics we find that in some analyses this essential lack of true relation to external reality is almost the whole thing.
I will try to describe in the simplest possible terms this phenomenon as I see it. In terms of baby and mother's breast (I am not claiming that the breast is essential as a vehicle of mother-love) the baby has instinctual urges and predatory ideas. The mother has a breast and the power to produce milk, and the idea that she would like to be attacked by a hungry baby. These two phenomena do not come into relation with each other till the mother and child live an experience together. The mother being mature and physically able has to be the one with tolerance and understanding, so that it is she that produces a situation which may with luck result in the first tie the infant makes with an external object, that is external to self from the infant's point of view.
I think of the process as if two lines came from opposite directions, liable to come near each other. If they overlap there is a moment of illusion-a bit of experience which the infant can take as either his hallucination or a thing belonging to external reality.
In other language, the infant comes to the breast when excited, and ready to hallucinate something fit to be attacked. At that moment the actual nipple appears and he is able to feel it was that nipple that he hallucinated. So his ideas are enriched by actual details of sight, feel, smell, and next time this material is used in the hallucination. In this way he starts to build up a capacity to conjure up what is actually available. The mother has to go on giving the infant this type of experience. The process is immensely simplified if the infant is cared for by one person and one technique. It seems as if an infant is really designed to be cared for from birth by his own mother, or failing that by an adopted mother, and not by several nurses.
It is especially at the start that mothers are vitally important, and indeed it is a mother's job to protect her infant from complications that cannot yet be understood by the infant, and to go on steadily providing the simplified bit of the world which the infant, through her, comes to know. Only on such a foundation can objectivity be built, or a scientific attitude. All failure in objectivity at whatever date relates to failure in this stage of primitive emotional development. Only on a basis of monotony can a mother profitably add richness.
One thing that follows the acceptance of external reality is the advantage to be gained from it. We often hear of the very real frustrations imposed by external reality, but less often hear of the relief and satisfaction it affords. Real milk is satisfying as compared with imaginary milk, but this is not the point. The point is that in phantasy things work by magic: there are no brakes on magic, and love and hate cause alarming effects. External reality has brakes on it, and can be studied and known, and, in fact, phantasy is only tolerable at full blast when objective reality is appreciated well. The subjective has tremendous value but is so alarming and magical that it cannot be enjoyed except as a parallel to the objective.
It will be seen that phantasy is not something the individual creates to deal with external reality's frustrations. This is only true of phantasying. Phantasy has more title to the word primary than has realization, and its enrichment with the world's riches depends on the moments of illusion that I have described.
It is interesting to examine the individual's relation to the objects in the self-created world of phantasy. In fact there are all grades of development and sophistication in this self-created world according to the amount of illusion that has been experienced, and so according to how much the self-created world has been unable or able to use perceived external world objects as material. This obviously needs a much more lengthy statement in another setting.
In the most primitive state, which may be retained in illness, and to which regression may occur, the object behaves according to magical laws, i.e. it exists when desired, it approaches when approached, it hurts when hurt. Lastly it vanishes when not wanted.
This last is most terrifying, and is the only true annihilation. To not want, perhaps as a result of satisfaction, is to annihilate the object. This is one reason why infants are not always happy and
4 I will just mention another reason why an infant is not satisfied with satisfaction. He feels fobbed off. He intended, one might say, to make a cannibalistic attack and he has been put off by an opiate, the feed. At best he can postpone the attack.
- 141 -
contented after a satisfactory feed. One patient of mine carried this fear
right on to adult life and only grew up from it in analysis, a man who had had
an extremely good early experience with his mother and in his home.4
I realize that this is only the bare outline of the vast problem of the initial steps in the development of a relation to external reality, and the relation of phantasy to reality. Soon we must add ideas of incorporation, but at the start a simple contact with external or shared reality has to be made, by the infant's hallucinating and the world's presenting, with moments of illusion for the infant in which the two are taken by him to be identical, which they never in fact are.
For this illusion to be produced in the baby's mind a human being has to be taking the trouble all the time to bring the world to the baby in understandable form, and in a limited way, suitable to the baby's needs. For this reason a baby cannot exist alone, psychologically or physically, and really needs one person to care for him at first.
The subject of illusion is a very wide one that needs study; it will be found to provide the clue to a child's interest in bubbles and clouds and rainbows and all mysterious phenomena, and also to his interest in fluff, which is most difficult to explain in terms of instinct direct. Somewhere here, too, is the interest in breath, which never decides whether it comes primarily from within or without, and which provides a basis for the conception of spirit, soul, anima.
VIIPRIMITIVE RUTHLESSNESS (STAGE OF PRE-CONCERN)
We are now in a position to look at the earliest kind of relationship between a baby and his mother.
If one assumes that the individual is becoming integrated and personalized and has made a good start in his realization, there is still a long way for him to go before he is related as a whole person to a whole mother, and concerned about the effect of his own thoughts and actions on her.
We have to postulate an early ruthless object-relationship. This may again be a theoretical phase only, and certainly no one can be ruthless after the concern stage except in a dissociated state. But ruthless dissociation states are common in early childhood, and emerge in certain types of delinquency, and madness, and must be available in health. The normal child enjoys a ruthless relation to his mother, mostly showing in play, and he needs his mother because only she can be expected to tolerate his ruthless relation to her even in play, because this really hurts her and wears her out. Without this play with her he can only hide a ruthless self and give it life in a state of dissociation.5
I can bring in here the great fear of disintegration as opposed to the simple acceptance of primary unintegration. Once the individual has reached the stage of concern he cannot be oblivious to the result of his impulses, or to the action of bits of self such as biting mouth, stabbing eyes, piercing yells, sucking throat, etc., etc. Disintegration means abandonment of the whole person-object to his impulses, uncontrolled because acting on their own; and further this conjures up the idea of similarly uncontrolled (because dissociated) impulses directed towards himself.6
VIII PRIMITIVE RETALIATION
To go back half a stage: it is usual, I think, to postulate a still more primitive object-relationship in which the object acts in a retaliatory way. This is prior to a true relation to external reality. In this case the object, or the environment, is as much part of the self as the instinct is which conjures it up.7In introversion of early origin and therefore of primitive quality the individual lives in this environment which is himself, and a very poor life it is. There is no growth because there is no enrichment from external reality.
To illustrate a different method of approach to this subject I add a note on thumb-sucking (including fist and finger sucking). This can be observed from birth onwards, and therefore can be presumed to have a meaning which develops from the primitive to sophistication, and it is important both as a normal activity and as a symptom of emotional disturbance.
We are familiar with the aspect of thumb-sucking which the term auto-erotic covers. The mouth is an erotogenic zone, specially organized in infancy, and the thumb-sucking child enjoys pleasure. He also has pleasurable ideas.
Hate is also expressed when the child damages
5 There is in mythology a ruthless figure—Lilith—whose origin could be usefully studied.
6 Crocodiles not only shed tears when they do not feel sad—pre-concern tears; they also readily stand for the ruthless primitive self. (There is much clinical evidence of this.)
7 This is important because of our relationship to Jung's analytic psychology. We try to reduce everything to instinct, and the analytic psychologists reduce everything to this part of the primitive self which looks like environment but which arises instantly out of instinct (archetypes). We ought to modify our view to embrace both ideas, and to see, if it is true, that in the earliest theoretical primitive state the self has its own environment, self-created, which is as much the self as the instincts that produce it. This is a theme which requires development.
- 142 -
his fingers by too vigorous or continuous sucking, and in any case he soon
adds nail-biting to cope with this part of his feelings. He is also liable to
damage his mouth. But it is not certain that all the damage that may be done to
a finger or mouth in this way is part of hate. It seems that there is in it the
element that something must suffer if the infant is to have pleasure: the object
of primitive love suffers by being loved, apart from being hated.
We can see in finger-sucking, and in nail-biting especially, a turning-in of love and hate, for reasons such as the need to preserve the external object of interest. Also we see a turning-in to self, in face of frustration in love of an external object.
The subject is not exhausted by this kind of statement and deserves further study.
I suppose anyone would agree that thumb-sucking is done for consolation, not just pleasure; the fist or finger is there instead of the breast or mother, or someone. For instance, a baby of about four months reacted to the loss of his mother by a tendency to put his fist right down his throat, so that he would have died had he not been physically prevented from acting this way.
Whereas thumb-sucking is normal and universal, spreading out into the use of the dummy, and indeed to various activities of normal adults, it is also true that thumb-sucking persists, in schizoid personalities, and in such cases is extremely compulsive. In one patient of mine it changed at 10 or 11 into a compulsion to be always reading.
These phenomena cannot be explained except on the basis that the act is an attempt to localize the object (breast, etc.), to hold it half-way between in and out; a defence against loss of object in the external world or in the inside of the body. I should say, against loss of control over the object, which occurs in either case.
I have no doubt that normal thumb-sucking has this function too.
The auto-erotic element is not always clearly of paramount importance and certainly the use of dummy and fist soon becomes a clear defence against insecurity feelings and other anxieties of a primitive kind.
Finally, every fist-sucking provides a useful dramatization of the primitive object-relationship in which the object is as much the individual as is the desire for an object, because it is created out of the desire, or is hallucinated, and at the beginning is independent of co-operation from external reality.
Some babies put a finger in the mouth while sucking the breast, thus (in a way) holding on to self-created reality while using external reality.
An attempt has been made to formulate the primitive psychological tendencies which are normal in early infancy, and which appear regressively in the psychoses.