Excerpts from Must Read Books & Articles on Mental Health Topics
Articles- Part XXIII

Sigmund Freud (1926)


THE time has come to pause and consider. What we clearly want is to find something that will tell us what anxiety really is, some criterion that will enable us to distinguish true statements about it from false ones. But this is not easy to get. Anxiety is not so simple a matter. Up till now we have arrived at nothing but contradictory views about it, none of which can, to the unprejudiced eye, be given preference over the others. I therefore propose to adopt a different procedure. I propose to assemble, quite impartially, all the facts that we know about anxiety without expecting to arrive at a fresh synthesis.
     Anxiety, then, is in the first place something that is felt. We call it an affective state, although we are also ignorant of what an affect is. As a feeling, anxiety has a very marked character of unpleasure. But that is not the whole of its quality. Not every unpleasure can be called anxiety, for there are other feelings, such as tension, pain or mourning, which have the character of unpleasure. Thus anxiety must have other distinctive features besides this quality of unpleasure. Can we succeed in understanding the differences between these various unpleasurable affects?
     We can at any rate note one or two things about the feeling of anxiety. Its unpleasurable character seems to have a note of its own--something not very obvious, whose presence is difficult to prove yet which is in all likelihood there. But besides having this special feature which is difficult to isolate, we notice that anxiety is accompanied by fairly definite physical sensations which can be referred to particular organs of the body. As we are not concerned here with the physiology of anxiety, we shall content ourselves with mentioning a few representatives of these sensations. The clearest and most frequent ones are those connected with the respiratory organs and with the heart. They provide evidence that motor innervations--that is, processes of discharge--play a part in the general phenomenon of anxiety.
     Analysis of anxiety-states therefore reveals the existence of (1) a specific character of unpleasure, (2) acts of discharge and (3) perceptions of those acts. The two last points indicate at once a difference between states of anxiety and other similar states, like those of mourning and pain. The latter do not have any motor manifestation; or if they have, the manifestation is not an integral part of the whole state but is distinct from it as being a result of it or a reaction to it. Anxiety, then, is a special state of unpleasure with acts of discharge along particular paths. In accordance with our general views we should be inclined to think that anxiety is based upon an increase of excitation which on the one hand produces the character of unpleasure and on the other finds relief through the acts of discharge already mentioned. But a purely physiological account of this sort will scarcely satisfy us. We are tempted to assume the presence of a historical factor which binds the sensations of anxiety and its innervations firmly together. We assume, in other words, that an anxiety-state is the reproduction of some experience which contained the necessary conditions for such an increase of excitation and a discharge along particular paths, and that from this circumstance the unpleasure of anxiety receives its specific character. In man, birth provides a prototypic experience of this kind, and we are therefore inclined to regard anxiety-states as a reproduction of the trauma of birth.
     This does not imply that anxiety occupies an exceptional position among the affective states. In my opinion the other affects are also reproductions of very early, perhaps even pre-individual, experiences of vital importance; and I should be inclined to regard them as universal, typical and innate hysterical attacks, as compared to the recently and individually acquired attacks which occur in hysterical neuroses and whose origin and significance as mnemic symbols have been revealed by analysis. It would be very desirable, of course, to be able to demonstrate the truth of this view in a number of such affects--a thing which is still very far from being the case.
     The view that anxiety goes back to the event of birth raises immediate objections which have to be met. It may be argued that anxiety is a reaction which, in all probability, is common to every organism, certainly every organism of a higher order, whereas birth is only experienced by the mammals; and it is doubtful whether in all of them, even, birth has the significance of a trauma. Therefore there can be anxiety without the prototype of birth. But this objection takes us beyond the barrier that divides psychology from biology. It may be that, precisely because anxiety has an indispensable biological function to fulfill as a reaction to a state of danger, it is differently contrived in different organisms. We do not know, besides, whether anxiety involves the same sensations and innervations in organisms far removed from man as it does in man himself. Thus there is no good argument here against the view that, in man, anxiety is modelled upon the process of birth.
     If the structure and origin of anxiety are as described, the next question is: what is the function of anxiety and on what occasions is it reproduced? The answer seems to be obvious and convincing: anxiety arose originally as a reaction to a state of danger and it is reproduced whenever a state of that kind recurs.
     This answer, however, raises further considerations. The innervations involved in the original state of anxiety probably had a meaning and purpose, in just the same way as the muscular movements which accompany a first hysterical attack. In order to understand a hysterical attack, all one has to do is to look for the situation in which the movements in question formed part of an appropriate and expedient action. Thus at birth it is probable that the innervation, in being directed to the respiratory organs, is preparing the way for the activity of the lungs, and, in accelerating the heartbeat, is helping to keep the blood free from toxic substances. Naturally, when the anxiety-state is reproduced later as an affect it will be lacking in any such expediency, just as are the repetitions of a hysterical attack. When the individual is placed in a new situation of danger it may well be quite inexpedient for him to respond with an anxiety-state (which is a reaction to an earlier danger) instead of initiating a reaction appropriate to the current danger. But his behaviour may become expedient once more if the danger-situation is recognized as it approaches and is signalled by an outbreak of anxiety. In that case he can at once get rid of his anxiety by having recourse to more suitable measures. Thus we see that there are two ways in which anxiety can emerge: in an inexpedient way, when a new situation of danger has occurred, or in an expedient way in order to give a signal and prevent such a situation from occurring.
     But what is a 'danger'? In the act of birth there is a real danger to life. We know what this means objectively; but in a psychological sense it says nothing at all to us. The danger of birth has as yet no psychical content. We cannot possibly suppose that the foetus has any sort of knowledge that there is a possibility of its life being destroyed. It can only be aware of some vast disturbance in the economy of its narcissistic libido. Large sums of excitation crowd in upon it, giving rise to new kinds of feelings of unpleasure, and some organs acquire an increased cathexis, thus foreshadowing the object-cathexis which will soon set in. What elements in all this will be made use of as the sign of a 'danger-situation'?
     Unfortunately far too little is known about the mental make-up of a new-born baby to make a direct answer possible. I cannot even vouch for the validity of the description I have just given. It is easy to say that the baby will repeat its affect of anxiety in every situation which recalls the event of birth. The important thing to know is what recalls the event and what it is that is recalled.
     All we can do is to examine the occasions on which infants in arms or somewhat older children show readiness to produce anxiety. In his book on the trauma of birth, Rank (1924) has made a determined attempt to establish a relationship between the earliest phobias of children and the impressions made on them by the event of birth. But I do not think he has been successful. His theory is open to two objections. In the first place, he assumes that the infant has received certain sensory impressions, in particular of a visual kind, at the time of birth, the renewal of which can recall to its memory the trauma of birth and thus evoke a reaction of anxiety. This assumption is quite unfounded and extremely improbable. It is not credible that a child should retain any but tactile and general sensations relating to the process of birth. If, later on, children show fear of small animals that disappear into holes or emerge from them, this reaction, according to Rank, is due to their perceiving an analogy. But it is an analogy of which they cannot be aware. In the second place, in considering these later anxiety-situations Rank dwells, as suits him best, now on the child's recollection of its happy intra-uterine existence, now on its recollection of the traumatic disturbance which interrupted that existence--which leaves the door wide open for arbitrary interpretation. There are, moreover, certain examples of childhood anxiety which directly traverse his theory. When, for instance, a child is left alone in the dark one would expect it, according to his view, to welcome the re-establishment of the intra-uterine situation; yet it is precisely on such occasions that the child reacts with anxiety. And if this is explained by saying that the child is being reminded of the interruption which the event of birth made in its intra-uterine happiness, it becomes impossible to shut one's eyes any longer to the far-fetched character of such explanations.
     I am driven to the conclusion that the earliest phobias of infancy cannot be directly traced back to impressions of the act of birth and that so far they have not been explained. A certain preparedness for anxiety is undoubtedly present in the infant in arms. But this preparedness for anxiety, instead of being at its maximum immediately after birth and then slowly decreasing, does not emerge till later, as mental development proceeds, and lasts over a certain period of childhood. If these early phobias persist beyond that period one is inclined to suspect the presence of a neurotic disturbance, although it is not at all clear what their relation is to the undoubted neuroses that appear later on in childhood.
     Only a few of the manifestations of anxiety in children are comprehensible to us, and we must confine our attention to them. They occur, for instance, when a child is alone, or in the dark, or when it finds itself with an unknown person instead of one to whom it is used--such as its mother. These three instances can be reduced to a single condition--namely, that of missing someone who is loved and longed for. But here, I think, we have the key to an understanding of anxiety and to a reconciliation of the contradictions that seem to beset it.
     The child's mnemic image of the person longed for is no doubt intensely cathected, probably in a hallucinatory way at first. But this has no effect; and now it seems as though the longing turns into anxiety. This anxiety has all the appearance of being an expression of the child's feeling at its wits' end, as though in its still very undeveloped state it did not know how better to cope with its cathexis of longing. Here anxiety appears as a reaction to the felt loss of the object; and we are at once reminded of the fact that castration anxiety, too, is a fear of being separated from a highly valued object, and that the earliest anxiety of all--the 'primal anxiety' of birth--is brought about on the occasion of a separation from the mother.
     But a moment's reflection takes us beyond this question of loss of object. The reason why the infant in arms wants to perceive the presence of its mother is only because it already knows by experience that she satisfies all its needs without delay. The situation, then, which it regards as a 'danger' and against which it wants to be safeguarded is that of non-satisfaction, of a growing tension due to need, against which it is helpless. I think that if we adopt this view all the facts fall into place. The situation of non-satisfaction in which the amounts of stimulation rise to an unpleasurable height without its being possible for them to be mastered psychically or discharged must for the infant be analogous to the experience of being born--must be a repetition of the situation of danger. What both situations have in common is the economic disturbance caused by an accumulation of amounts of stimulation which require to be disposed of. It is this factor, then, which is the real essence of the 'danger'. In both cases the reaction of anxiety sets in. (This reaction is still an expedient one in the infant in arms, for the discharge, being directed into the respiratory and vocal muscular apparatus, now calls its mother to it, just as it activated the lungs of the new-born baby to get rid of the internal stimuli.) It is unnecessary to suppose that the child carries anything more with it from the time of its birth than this way of indicating the presence of danger.
     When the infant has found out by experience that an external, perceptible object can put an end to the dangerous situation which is reminiscent of birth, the content of the danger it fears is displaced from the economic situation on to the condition which determined that situation, viz., the loss of object. It is the absence of the mother that is now the danger; and as soon as that danger arises the infant gives the signal of anxiety, before the dreaded economic situation has set in. This change constitutes a first great step forward in the provision made by the infant for its self-preservation, and at the same time represents a transition from the automatic and involuntary fresh appearance of anxiety to the intentional reproduction of anxiety as a signal of danger.
     In these two aspects, as an automatic phenomenon and as a rescuing signal, anxiety is seen to be a product of the infant's mental helplessness which is a natural counterpart of its biological helplessness. The striking coincidence by which the anxiety of the new-born baby and the anxiety of the infant in arms are both conditioned by separation from the mother does not need to be explained on psychological lines. It can be accounted for simply enough biologically; for, just as the mother originally satisfied all the needs of the foetus through the apparatus of her own body, so now, after its birth, she continues to do so, though partly by other means. There is much more continuity between intra-uterine life and earliest infancy than the impressive caesura of the act of birth would have us believe. What happens is that the child's biological situation as a foetus is replaced for it by a psychical object-relation to its mother. But we must not forget that during its intra-uterine life the mother was not an object for the foetus, and that at that time there were no objects at all. It is obvious that in this scheme of things there is no place for the abreaction of the birth-trauma. We cannot find that anxiety has any function other than that of being a signal for the avoidance of a danger-situation.
     The significance of the loss of object as a determinant of anxiety extends considerably further. For the next transformation of anxiety, viz. the castration anxiety belonging to the phallic phase, is also a fear of separation and is thus attached to the same determinant. In this case the danger is of being separated from one's genitals. Ferenczi [1925] has traced, quite correctly, I think, a clear line of connection between this fear and the fears contained in the earlier situations of danger. The high degree of narcissistic value which the penis possesses can appeal to the fact that that organ is a guarantee to its owner that he can be once more united to his mother--i.e. to a substitute for her--in the act of copulation. Being deprived of it amounts to a renewed separation from her, and this in its turn means being helplessly exposed to an unpleasurable tension due to instinctual need, as was the case at birth. But the need whose increase is feared is now a specific one belonging to the genital libido and is no longer an indeterminate one, as it was in the period of infancy. It may be added that for a man who is impotent (that is, who is inhibited by the threat of castration) the substitute for copulation is a phantasy of returning into his mother's womb. Following Ferenczi's line of thought, we might say that the man in question, having tried to bring about his return into his mother's womb by using his genital organ to represent him, is now [in this phantasy] replacing that organ regressively by his whole person.
     The progress which the child makes in its development--its growing independence, the sharper division of its mental apparatus into several agencies, the advent of new needs--cannot fail to exert an influence upon the content of the danger-situation. We have already traced the change of that content from loss of the mother as an object to castration. The next change is caused by the power of the super-ego. With the depersonalization of the parental agency from which castration was feared, the danger becomes less defined. Castration anxiety develops into moral anxiety--social anxiety--and it is not so easy now to know what the anxiety is about. The formula, 'separation and expulsion from the horde', only applies to that later portion of the super-ego which has been formed on the basis of social prototypes, not to the nucleus of the super-ego, which corresponds to the introjected parental agency. Putting it more generally, what the ego regards as the danger and responds to with an anxiety-signal is that the super-ego should be angry with it or punish it or cease to love it. The final transformation which the fear of the super-ego undergoes is, it seems to me, the fear of death (or fear for life) which is a fear of the super-ego projected on to the powers of destiny.
     At one time I attached some importance to the view that what was used as a discharge of anxiety was the cathexis which had been withdrawn in the process of repression. Today this seems to me of scarcely any interest. The reason for this is that whereas I formerly believed that anxiety invariably arose automatically by an economic process, my present conception of anxiety as a signal given by the ego in order to affect the pleasure-unpleasure agency does away with the necessity of considering the economic factor. Of course there is nothing to be said against the idea that it is precisely the energy that has been liberated by being withdrawn through repression which is used by the ego to arouse the affect; but it is no longer of any importance which portion of energy is employed for this purpose.
     This new view of things calls for an examination of another assertion of mine--namely, that the ego is the actual seat of anxiety. I think this proposition still holds good. There is no reason to assign any manifestation of anxiety to the super-ego; while the expression 'anxiety of the id' would stand in need of correction, though rather as to its form than its substance. Anxiety is an affective state and as such can, of course, only be felt by the ego. The id cannot have anxiety as the ego can; for it is not an organization and cannot make a judgment about situations of danger. On the other hand it very often happens that processes take place or begin to take place in the id which cause the ego to produce anxiety. Indeed, it is probable that the earliest repressions as well as most of the later ones are motivated by an ego-anxiety of this sort in regard to particular processes in the id. Here again we are rightly distinguishing between two cases: the case in which something occurs in the id which activates one of the danger-situations for the ego and induces the latter to give the anxiety-signal for inhibition to take place, and the case in which a situation analogous to the trauma of birth is established in the id and an automatic reaction of anxiety ensues. The two cases may be brought closer together if it is pointed out that the second case corresponds to the earliest and original danger-situation, while the first case corresponds to any one of the later determinants of anxiety that have been derived from it; or, as applied to the disorders which we in fact come across, that the second case is operative in the aetiology of the 'actual' neuroses, while the first remains typical for that of the psychoneuroses.
     We see, then, that it is not so much a question of taking back our earlier findings as of bringing them into line with more recent discoveries. It is still an undeniable fact that in sexual abstinence, in improper interference with the course of sexual excitation or if the latter is diverted from being worked over psychically, anxiety arises directly out of libido; in other words, that the ego is reduced to a state of helplessness in the face of an excessive tension due to need, as it was in the situation of birth, and that anxiety is then generated. Here once more, though the matter is of little importance, it is very possible that what finds discharge in the generating of anxiety is precisely the surplus of unutilized libido. As we know, a psychoneurosis is especially liable to develop on the basis of an 'actual' neurosis. This looks as though the ego were attempting to save itself from anxiety, which it has learned to keep in suspension for a while, and to bind it by the formation of symptoms. Analysis of the traumatic war neuroses--a term which, incidentally, covers a great variety of disorders--would probably have shown that a number of them possess some characteristics of the 'actual' neuroses.
     In describing the evolution of the various danger-situations from their prototype, the act of birth, I have had no intention of asserting that every later determinant of anxiety completely invalidates the preceding one. It is true that, as the development of the ego goes on, the earlier danger-situations tend to lose their force and to be set aside, so that we might say that each period of the individual's life has its appropriate determinant of anxiety. Thus the danger of psychical helplessness is appropriate to the period of life when his ego is immature; the danger of loss of object, to early childhood when he is still dependent on others; the danger of castration, to the phallic phase; and the fear of his super-ego, to the latency period. Nevertheless, all these danger-situations and determinants of anxiety can persist side by side and cause the ego to react to them with anxiety at a period later than the appropriate one; or, again, several of them can come into operation at the same time. It is possible, moreover, that there is a fairly close relationship between the danger-situation that is operative and the form taken by the ensuing neurosis.
     Since the differentiation of the ego and the id, our interest in the problems of repression, too, was bound to receive a fresh impetus. Up till then we had been content to confine our interest to those aspects of repression which concerned the ego--the keeping away from consciousness and from motility, and the formation of substitutes (symptoms). With regard to the repressed instinctual impulses themselves, we assumed that they remained unaltered in the unconscious for an indefinite length of time. But now our interest is turned to the vicissitudes of the repressed and we begin to suspect that it is not self-evident, perhaps not even usual, that those impulses should remain unaltered and unalterable in this way. There is no doubt that the original impulses have been inhibited and deflected from their aim through repression. But has the portion of them in the unconscious maintained itself and been proof against the influences of life that tend to alter and depreciate them? In other words, do the old wishes, about whose former existence analysis tells us, still exist? The answer seems ready to hand and certain. It is that the old, repressed wishes must still be present in the unconscious since we still find their derivatives, the symptoms, in operation. But this answer is not sufficient. It does not enable us to decide between two possibilities: either that the old wish is now operating only through its derivatives, having transferred the whole of its cathectic energy to them, or that it is itself still in existence too. If its fate has been to exhaust itself in cathecting its derivatives, there is yet a third possibility. In the course of the neurosis it may have become re-animated by regression, anachronistic though it may now be. These are no idle speculations. There are many things about mental life, both normal and pathological, which seem to call for the raising of such questions. In my paper, 'The Dissolution of the Oedipus Complex' (1924d), I had occasion to notice the difference between the mere repression and the real removal of an old wishful impulse.
     When, in an earlier part of this discussion, we found that the danger of castration was of importance in more than one neurotic illness, we put ourselves on guard against overestimating that factor, since it could not be a decisive one for the female sex, who are undoubtedly more subject to neuroses than men. We now see that there is no danger of our regarding castration anxiety as the sole motive force of the defensive processes which lead to neurosis. I have shown elsewhere how little girls, in the course of their development, are led into making a tender object-cathexis by their castration complex. It is precisely in women that the danger-situation of loss of object seems to have remained the most effective. All we need to do is to make a slight modification in our description of their determinant of anxiety, in the sense that it is no longer a matter of feeling the want of, or actually losing the object itself, but of losing the object's love. Since there is no doubt that hysteria has a strong affinity with femininity, just as obsessional neurosis has with masculinity, it appears probable that, as a determinant of anxiety, loss of love plays much the same part in hysteria as the threat of castration does in phobias and fear of the super-ego in obsessional neurosis.


WHAT is now left for us is to consider the relationship between the formation of symptoms and the generating of anxiety.
     There seem to be two very widely held opinions on this subject. One is that anxiety is itself a symptom of neurosis. The other is that there is a much more intimate relation between the two. According to the second opinion, symptoms are only formed in order to avoid anxiety: they bind the psychical energy which would otherwise be discharged as anxiety. Thus anxiety would be the fundamental phenomenon and main problem of neurosis.
     That this latter opinion is at least in part true is shown by some striking examples. If an agoraphobic patient who has been accompanied into the street is left alone there, he will produce an anxiety attack. Or if an obsessional neurotic is prevented from washing his hands after having touched something, he will become a prey to almost unbearable anxiety. It is plain, then, that the purpose and the result of the imposed condition of being accompanied in the street and the obsessional act of washing the hands were to obviate outbreaks of anxiety of this kind. In this sense every inhibition which the ego imposes on itself can be called a symptom.
     Since we have traced back the generating of anxiety to a situation of danger, we shall prefer to say that symptoms are created in order to remove the ego from a situation of danger. If the symptoms are prevented from being formed, the danger does in fact materialize; that is, a situation analogous to birth is established in which the ego is helpless in the face of a constantly increasing instinctual demand--the earliest and original determinant of anxiety. Thus in our view the relation between anxiety and symptom is less close than was supposed, for we have inserted the factor of the danger-situation between them. We can also add that the generating of anxiety sets symptom formation going and is, indeed, a necessary prerequisite of it. For if the ego did not arouse the pleasure-unpleasure agency by generating anxiety, it would not obtain the power to arrest the process which is preparing in the id and which threatens danger. There is in all this an evident inclination to limit to a minimum the amount of anxiety generated and to employ it only as a signal; for to do otherwise would only result in feeling in another place the unpleasure which the instinctual process was threatening to produce, and that would not be a success from the standpoint of the pleasure principle, although it is one that occurs often enough in the neuroses.
     Symptom-formation, then, does in fact put an end to the danger-situation. It has two aspects: one, hidden from view, brings about the alteration in the id in virtue of which the ego is removed from danger; the other, presented openly, shows what has been created in place of the instinctual process that has been affected--namely, the substitutive formation.
     It would, however, be more correct to ascribe to the defensive process what we have just said about symptom-formation and to use the latter term as synonomous with substitute-formation. It will then be clear that the defensive process is analogous to the flight by means of which the ego removes itself from a danger that threatens it from outside. The defensive process is an attempt at flight from an instinctual danger. An examination of the weak points in this comparison will make things clearer.
     One objection to it is that loss of an object (or loss of love on the part of the object) and the threat of castration are just as much dangers coming from outside as, let us say, a ferocious animal would be; they are not instinctual dangers. Nevertheless, the two cases are not the same. A wolf would probably attack us irrespectively of our behaviour towards it; but the loved person would not cease to love us nor should we be threatened with castration if we did not entertain certain feelings and intentions within us. Thus such instinctual impulses are determinants of external dangers and so become dangerous in themselves; and we can now proceed against the external danger by taking measures against the internal ones. In phobias of animals the danger seems to be still felt entirely as an external one, just as it has undergone an external displacement in the symptom. In obsessional neuroses the danger is much more internalized. That portion of anxiety in regard to the super-ego which constitutes social anxiety still represents an internal substitute for an external danger, while the other portion--moral anxiety--is already completely endopsychic.
     Another objection is that in an attempt at flight from an impending external danger all that the subject is doing is to increase the distance between himself and what is threatening him. He is not preparing to defend himself against it or attempting to alter anything about it, as would be the case if he attacked the wolf with a stick or shot at it with a gun. But the defensive process seems to do something more than would correspond to an attempt at flight. It joins issue with the threatening instinctual process and somehow suppresses it or deflects it from its aims and thus renders it innocuous. This objection seems unimpeachable and must be given due weight. I think it is probable that there are some defensive processes which can truly be likened to an attempt at flight, while in others the ego takes a much more active line of self-protection and initiates vigorous counter-measures. But perhaps the whole analogy between defence and flight is invalidated by the fact that both the ego and the instinct in the id are parts of the same organization, not separate entities like the wolf and the child, so that any kind of behaviour on the part of the ego will result in an alteration in the instinctual process as well.
     This study of the determinants of anxiety has, as it were, shown the defensive behaviour of the ego transfigured in a rational light. Each situation of danger corresponds to a particular period of life or a particular developmental phase of the mental apparatus and appears to be justifiable for it. In early infancy the individual is really not equipped to master psychically the large sums of excitation that reach him whether from without or from within. Again, at a certain period of life his most important interest really is that the people he is dependent on should not withdraw their loving care of him. Later on in his boyhood, when he feels that his father is a powerful rival in regard to his mother and becomes aware of his own aggressive inclinations towards him and of his sexual intentions towards his mother, he really is justified in being afraid of his father; and his fear of being punished by him can find expression through phylogenetic reinforcement in the fear of being castrated. Finally, as he enters into social relationships, it really is necessary for him to be afraid of his super-ego, to have a conscience; and the absence of that factor would give rise to severe conflicts, dangers and so on.
     But this last point raises a fresh problem. Instead of the affect of anxiety let us take, for a moment, another affect--that of pain, for instance. It seems quite normal that at four years of age a girl should weep painfully if her doll is broken; or at six, if her governess reproves her; or at sixteen, if she is slighted by her young man; or at twenty-five, perhaps, if a child of her own dies. Each of these determinants of pain has its own time and each passes away when that time is over. Only the final and definitive determinants remain throughout life. We should think it strange if this same girl, after she had grown to be a wife and mother, were to cry over some worthless trinket that had been damaged. Yet that is how the neurotic behaves. Although all the agencies for mastering stimuli have long ago been developed within wide limits in his mental apparatus, and although he is sufficiently grown-up to satisfy most of his needs for himself and has long ago learnt that castration is no longer practised as a punishment, he nevertheless behaves as though the old danger-situations still existed, and keeps hold of all the earlier determinants of anxiety.
     Why this should be so calls for a rather long reply. First of all, we must sift the facts. In a great number of cases the old determinants of anxiety do really lapse, after having produced neurotic reactions. The phobias of very young children, fears of being alone or in the dark or with strangers--phobias which can almost be called normal--usually pass off later on; the child 'grows out of them', as we say about some other disturbances of childhood. Animal phobias, which are of such frequent occurrence, undergo the same fate and many conversion hysterias of early years find no continuation in later life. Ceremonial actions appear extremely often in the latency period, but only a very small percentage of them develop later into a full obsessional neurosis. In general, so far as we can tell from our observations of town children belonging to the white races and living according to fairly high cultural standards, the neuroses of childhood are in the nature of regular episodes in a child's development, although too little attention is still being paid to them. Signs of childhood neuroses can be detected in all adult neurotics without exception; but by no means all children who show those signs become neurotic in later life. It must be, therefore, that certain determinants of anxiety are relinquished and certain danger-situations lose their significance as the individual becomes more mature. Moreover, some of these danger-situations manage to survive into later times by modifying their determinants of anxiety so as to bring them up to date. Thus, for instance, a man may retain his fear of castration in the guise of a syphilidophobia, after he has come to know that it is no longer customary to castrate people for indulging their sexual lusts, but that, on the other hand, severe diseases may overtake anyone who thus gives way to his instincts. Other determinants of anxiety, such as fear of the super-ego, are destined not to disappear at all but to accompany people throughout their lives. In that case the neurotic will differ from the normal person in that his reactions to the dangers in question will be unduly strong. Finally, being grown-up affords no absolute protection against a return of the original traumatic anxiety-situation. Each individual has in all probability a limit beyond which his mental apparatus fails in its function of mastering the quantities of excitation which require to be disposed of.
     These minor rectifications cannot in any way alter the fact which is here under discussion, that a great many people remain infantile in their behaviour in regard to danger and do not overcome determinants of anxiety which have grown out of date. To deny this would be to deny the existence of neurosis, for it is precisely such people whom we call neurotics. But how is this possible? Why are not all neuroses episodes in the development of the individual which come to a close when the next phase is reached? Whence comes the element of persistence in these reactions to danger? Why does the affect of anxiety alone seem to enjoy the advantage over all other affects of evoking reactions which are distinguished from the rest in being abnormal and which, through their inexpediency, run counter to the movement of life? In other words, we have once more come unawares upon the riddle which has so often confronted us: whence does neurosis come--what is its ultimate, its own peculiar raison d'etre? After tens of years of psychoanalytic labours, we are as much in the dark about this problem as we were at the start.


ANXIETY is the reaction to danger. One cannot, after all, help suspecting that the reason why the affect of anxiety occupies a unique position in the economy of the mind has something to do with the essential nature of danger. Yet dangers are the common lot of humanity; they are the same for everyone. What we need and cannot lay our finger on is some factor which will explain why some people are able to subject the affect of anxiety, in spite of its peculiar quality, to the normal workings of the mind, or which decides who is doomed to come to grief over that task. Two attempts to find a factor of this kind have been made; and it is natural that such efforts should meet with a sympathetic reception, since they promise help to meet a tormenting need. The two attempts in question are mutually complementary; they approach the problem at opposite ends. The first was made by Alfred Adler more than ten years ago. His contention, reduced to its essence, was that the people who came to grief over the task set them by danger were those who were too greatly impeded by some organic inferiority. If it were true that simplex sigillum veri [i.e., simplicity is the seal of truth.] we should welcome such a solution [Losung] as a deliverance [Erlosung]. But on the contrary, our critical studies of the last ten years have effectively demonstrated the total inadequacy of such an explanation--an explanation, moreover, which sets aside the whole wealth of material that has been discovered by psycho-analysis.
     The second attempt was made by Otto Rank in 1923 in his book, The Trauma of Birth. It would be unjust to put his attempt on the same level as Adler's except in this single point which concerns us here, for it remains on psycho-analytic ground and pursues a psycho-analytic line of thought, so that it may be accepted as a legitimate endeavour to solve the problems of analysis. In this matter of the relation of the individual to danger Rank moves away from the question of organic defect in the individual and concentrates on the variable degree of intensity of the danger. The process of birth is the first situation of danger, and the economic upheaval which it produces becomes the prototype of the reaction of anxiety. We have already traced the line of development which connects this first danger-situation and determinant of anxiety with all the later ones, and we have seen that they all retain a common quality in so far as they signify in a certain sense a separation from the mother--at first only in a biological sense, next as a direct loss of object and later as a loss of object incurred indirectly. The discovery of this extensive concatenation is an undoubted merit of Rank's construction. Now the trauma of birth overtakes each individual with a different degree of intensity, and the violence of his anxiety-reaction varies with the strength of the trauma; and it is the initial amount of anxiety generated in him which, according to Rank, decides whether he will ever learn to control it--whether he will become neurotic or normal.
     It is not our business to criticize Rank's hypothesis in detail here. We have only to consider whether it helps to solve our particular problem. His formula--that those people become neurotic in whom the trauma of birth was so strong that they have never been able completely to abreact it--is highly disputable from a theoretical point of view. We do not rightly know what is meant by abreacting the trauma. Taken literally, it implies that the more frequently and the more intensely a neurotic person reproduces the affect of anxiety the more closely will he approach to mental health--an untenable conclusion. It was because it did not tally with the facts that I gave up the theory of abreaction which had played such a large part in the cathartic method. To lay so much stress, too, on the variability in the strength of the birth trauma is to leave no room for the legitimate claims of hereditary constitution as an aetiological factor. For this variability is an organic factor which operates in an accidental fashion in relation to the constitution and is itself dependent on many influences which might be called accidental--as, for instance, on timely assistance in child-birth. Rank's theory completely ignores constitutional factors as well as phylogenetic ones. If, however, we were to try to find a place for the constitutional factor by qualifying his statement with the proviso, let us say, that what is really important is the extent to which the individual reacts to the variable intensity of the trauma of birth, we should be depriving his theory of its significance and should be relegating the new factor introduced by him to a position of minor importance: the factor which decided whether a neurosis should supervene or not would lie in a different, and once more in an unknown, field.
     Moreover, the fact that while man shares the process of birth with the other mammals he alone has the privilege over them of possessing a special disposition to neurosis is hardly favourable to Rank's theory. But the main objection to it is that it floats in the air instead of being based upon ascertained observations. No body of evidence has been collected to show that difficult and protracted birth does in fact coincide with the development of a neurosis, or even that children so born exhibit the phenomena of early infantile apprehensiveness more strongly and over a longer period than other children. It might be rejoined that induced labour and births that are easy for the mother may possibly involve a severe trauma for the child. But we can still point out that births which lead to asphyxia would be bound to give clear evidence of the results which are supposed to follow. It should be one of the advantages of Rank's aetiological theory that it postulates a factor whose existence can be verified by observation. And so long as no such attempt at verification has been made it is impossible to assess the theory's value.
     On the other hand I cannot identify myself with the view that Rank's theory contradicts the aetiological importance of the sexual instincts as hitherto recognized by psycho-analysis. For his theory only has reference to the individual's relation to the danger-situation, so that it leaves it perfectly open to us to assume that if a person has not been able to master his first dangers he is bound to come to grief as well in later situations involving sexual danger and thus be driven into a neurosis.
     I do not believe, therefore, that Rank's attempt has solved the problem of the causation of neurosis; nor do I believe that we can say as yet how much it may nevertheless have contributed to such a solution. If an investigation into the effects of difficult birth on the disposition to neurosis should yield negative results, we shall rate the value of his contribution low. It is to be feared that our need to find a single, tangible 'ultimate cause' of neurotic illness will remain unsatisfied. The ideal solution, which medical men no doubt still yearn for, would be to discover some bacillus which could be isolated and bred in a pure culture and which, when injected into anyone, would invariably produce the same illness; or, to put it rather less extravagantly, to demonstrate the existence of certain chemical substances the administration of which would bring about or cure particular neuroses. But the probability of a solution of this kind seems slight.
     Psycho-analysis leads to less simple and satisfactory conclusions. What I have to say in this connection has long been familiar and I have nothing new to add. If the ego succeeds in protecting itself from a dangerous instinctual impulse, through, for instance, the process of repression, it has certainly inhibited and damaged the particular part of the id concerned; but it has at the same time given it some independence and has renounced some of its own sovereignty. This is inevitable from the nature of repression, which is, fundamentally, an attempt at flight. The repressed is now, as it were, an outlaw; it is excluded from the great organization of the ego and is subject only to the laws which govern the realm of the unconscious. If, now, the danger-situation changes so that the ego has no reason for fending off a new instinctual impulse analogous to the repressed one, the consequence of the restriction of the ego which has taken place will become manifest. The new impulse will run its course under an automatic influence--or, as I should prefer to say, under the influence of the compulsion to repeat. It will follow the same path as the earlier, repressed impulse, as though the danger-situation that had been overcome still existed. The fixating factor in repression, then, is the unconscious id's compulsion to repeat--a compulsion which in normal circumstances is only done away with by the freely mobile function of the ego. The ego may occasionally manage to break down the barriers of repression which it has itself put up and to recover its influence over the instinctual impulse and direct the course of the new impulse in accordance with the changed danger-situation. But in point of fact the ego very seldom succeeds in doing this: it cannot undo its repressions. It is possible that the way the struggle will go depends upon quantitative relations. In some cases one has the impression that the outcome is an enforced one: the regressive attraction exerted by the repressed impulse and the strength of the repression are so great that the new impulse has no choice but to obey the compulsion to repeat. In other cases we perceive a contribution from another play of forces: the attraction exerted by the repressed prototype is reinforced by a repulsion coming from the direction of difficulties in real life which stand in the way of any different course that might be taken by the new instinctual impulse.
     That this is a correct account of fixation upon repression and of the retention of danger-situations that are no longer present-day ones is confirmed by the fact of analytic therapy--a fact which is modest enough in itself but which can hardly be overrated from a theoretical point of view. When, in analysis, we have given the ego assistance which is able to put it in a position to lift its repressions, it recovers its power over the repressed id and can allow the instinctual impulses to run their course as though the old situations of danger no longer existed. What we can do in this way tallies with what can be achieved in other fields of medicine; for as a rule our therapy must be content with bringing about more quickly, more reliably and with less expenditure of energy than would otherwise be the case the good result which in favourable circumstances would have occurred of itself.
     We see from what has been said that quantitative relations--relations which are not directly observable but which can only be inferred--are what determine whether or not old situations of danger shall be preserved, repressions on the part of the ego maintained and childhood neuroses find a continuation. Among the factors that play a part in the causation of neuroses and that have created the conditions under which the forces of the mind are pitted against one another, three emerge into prominence: a biological, a phylogenetic and a purely psychological factor.
     The biological factor is the long period of time during which the young of the human species is in a condition of helplessness and dependence. Its intra-uterine existence seems to be short in comparison with that of most animals, and it is sent into the world in a less finished state. As a result, the influence of the real external world upon it is intensified and an early differentiation between the ego and the id is promoted. Moreover, the dangers of the external world have a greater importance for it, so that the value of the object which can alone protect it against them and take the place of its former intra-uterine life is enormously enhanced. The biological factor, then, establishes the earliest situations of danger and creates the need to be loved which will accompany the child through the rest of its life.
     The existence of the second, phylogenetic factor is based only upon inference. We have been led to assume its existence by a remarkable feature in the development of the libido. We have found that the sexual life of man, unlike that of most of the animals nearly related to him, does not make a steady advance from birth to maturity, but that, after an early efflorescence up till the fifth year, it undergoes a very decided interruption; and that it then starts on its course once more at puberty, taking up again the beginnings broken off in early childhood. This has led us to suppose that something momentous must have occurred in the vicissitudes of the human species which has left behind this interruption in the sexual development of the individual as a historical precipitate. This factor owes its pathogenic significance to the fact that the majority of the instinctual demands of this infantile sexuality are treated by the ego as dangers and fended off as such, so that the later sexual impulses of puberty, which in the natural course of things would be ego-syntonic, run the risk of succumbing to the attraction of their infantile prototypes and following them into repression. It is here that we come upon the most direct aetiology of the neuroses. It is a curious thing that early contact with the demands of sexuality should have a similar effect on the ego to that produced by premature contact with the external world.
     The third, psychological, factor resides in a defect of our mental apparatus which has to do precisely with its differentiation into an id and an ego, and which is therefore also attributable ultimately to the influence of the external world. In view of the dangers of [external] reality, the ego is obliged to guard against certain instinctual impulses in the id and to treat them as dangers. But it cannot protect itself from internal instinctual dangers as effectively as it can from some piece of reality that is not part of itself. Intimately bound up with the id as it is, it can only fend off an instinctual danger by restricting its own organization and by acquiescing in the formation of symptoms in exchange for having impaired the instinct. If the rejected instinct renews its attack, the ego is overtaken by all those difficulties which are known to us as neurotic ailments.
     Further than this, I believe, our knowledge of the nature and causes of neurosis has not as yet been able to go.


IN the course of this discussion various themes have had to be put aside before they had been fully dealt with. I have brought them together in this chapter so that they may receive the attention they deserve.


(a) Resistance and Anticathexis
An important element in the theory of repression is the view that repression is not an event that occurs once but that it requires a permanent expenditure [of energy]. If this expenditure were to cease, the repressed impulse, which is being fed all the time from its sources, would on the next occasion flow along the channels from which it had been forced away, and the repression would either fail in its purpose or would have to be repeated an indefinite number of times. Thus it is because instincts are continuous in their nature that the ego has to make its defensive action secure by a permanent expenditure [of energy]. This action undertaken to protect repression is observable in analytic treatment as resistance. Resistance presupposes the existence of what I have called anticathexis. An anticathexis of this kind is clearly seen in obsessional neurosis. It appears there in the form of an alteration of the ego, as a reaction-formation in the ego, and is effected by the reinforcement of the attitude which is the opposite of the instinctual trend that has to be repressed--as, for instance, in pity, conscientiousness and cleanliness. These reaction-formations of obsessional neurosis are essentially exaggerations of the normal traits of character which develop during the latency period. The presence of an anticathexis in hysteria is much more difficult to detect, though theoretically it is equally indispensable. In hysteria, too, a certain amount of alteration of the ego through reaction-formation is unmistakable and in some circumstances becomes so marked that it forces itself on our attention as the principal symptom. The conflict due to ambivalence, for instance, is resolved in hysteria by this means. The subject's hatred of a person whom he loves is kept down by an exaggerated amount of tenderness for him and apprehensiveness about him. But the difference between reaction-formations in obsessional neurosis and in hysteria is that in the latter they do not have the universality of a character-trait but are confined to particular relationships. A hysterical woman, for instance, may be specially affectionate with her own children whom at bottom she hates; but she will not on that account be more loving in general than other women or even more affectionate to other children. The reaction-formation of hysteria clings tenaciously to a particular object and never spreads over into a general disposition of the ego, whereas what is characteristic of obsessional neurosis is precisely a spreading-over of this kind--a loosening of relations to the object and a facilitation of displacement in the choice of object.
     There is another kind of anticathexis, however, which seems more suited to the peculiar character of hysteria. A repressed instinctual impulse can be activated (newly cathected) from two directions: from within, through reinforcement from its internal sources of excitation, and from without, through the perception of an object that it desires. The hysterical anticathexis is mainly directed outwards, against dangerous perceptions. It takes the form of a special kind of vigilance which, by means of restrictions of the ego, causes situations to be avoided that would entail such perceptions, or, if they do occur, manages to withdraw the subject's attention from them. Some French analysts, in particular Laforgue [1926], have recently given this action of hysteria the special name of 'scotomization'. This technique of anticathexis is still more noticeable in the phobias, whose interest is concentrated on removing the subject ever further from the possibility of the occurrence of the feared perception. The fact that anticathexis has an opposite direction in hysteria and the phobias from what it has in obsessional neurosis--though the distinction is not an absolute one--seems to be significant. It suggests that there is an intimate connection between repression and external anticathexis on the one hand and between regression and internal anticathexis (i.e. alteration in the ego through reaction-formation) on the other. The task of defence against a dangerous perception is, incidentally, common to all neuroses. Various commands and prohibitions in obsessional neurosis have the same end in view.
     We showed on an earlier occasion that the resistance that has to be overcome in analysis proceeds from the ego, which clings to its anticathexes. It is hard for the ego to direct its attention to perceptions and ideas which it has up till now made a rule of avoiding, or to acknowledge as belonging to itself impulses that are the complete opposite of those which it knows as its own. Our fight against resistance in analysis is based upon this view of the facts. If the resistance is itself unconscious, as so often happens owing to its connection with the repressed material, we make it conscious. If it is conscious, or when it has become conscious, we bring forward logical arguments against it; we promise the ego rewards and advantages if it will give up its resistance. There can be no doubt or mistake about the existence of this resistance on the part of the ego. But we have to ask ourselves whether it covers the whole state of affairs in analysis. For we find that even after the ego has decided to relinquish its resistances it still has difficulty in undoing the repressions; and we have called the period of strenuous effort which follows after its praiseworthy decision, the phase of 'working-through'. The dynamic factor which makes a working-through of this kind necessary and comprehensible is not far to seek. It must be that after the ego-resistance has been removed the power of the compulsion to repeat--the attraction exerted by the unconscious prototypes upon the repressed instinctual process--has still to be overcome. There is nothing to be said against describing this factor as the resistance of the unconscious. There is no need to be discouraged by these emendations. They are to be welcomed if they add something to our knowledge, and they are no disgrace to us so long as they enrich rather than invalidate our earlier views--by limiting some statement, perhaps, that was too general or by enlarging some idea that was too narrowly formulated.
     It must not be supposed that these emendations provide us with a complete survey of all the kinds of resistance that are met with in analysis. Further investigation of the subject shows that the analyst has to combat no less than five kinds of resistance, emanating from three directions--the ego, the id and the super-ego. The ego is the source of three of these, each differing in its dynamic nature. The first of these three ego-resistances is the repression resistance, which we have already discussed above and about which there is least new to be added. Next there is the transference resistance, which is of the same nature but which has different and much clearer effects in analysis, since it succeeds in establishing a relation to the analytic situation or the analyst himself and thus re-animating a repression which should only have been recollected. The third resistance, though also an ego-resistance, is of quite a different nature. It proceeds from the gain from illness and is based upon an assimilation of the symptom into the ego. It represents an unwillingness to renounce any satisfaction or relief that has been obtained. The fourth variety, arising from the id, is the resistance which, as we have just seen, necessitates 'working-through'. The fifth, coming from the super-ego and the last to be discovered, is also the most obscure though not always the least powerful one. It seems to originate from the sense of guilt or the need for punishment; and it opposes every move towards success, including, therefore, the patient's own recovery through analysis.

(b) Anxiety from Transformation of Libido
The view of anxiety which I have put forward in these pages diverges somewhat from the one I have hitherto thought correct. Formerly I regarded anxiety as a general reaction of the ego under conditions of unpleasure. I always sought to justify its appearance on economic grounds and I assumed, on the strength of my investigations into the 'actual' neuroses, that libido (sexual excitation) which was rejected or not utilized by the ego found direct discharge in the form of anxiety. It cannot be denied that these various assertions did not go very well together, or at any rate did not necessarily follow from one another. Moreover, they gave the impression of there being a specially intimate connection between anxiety and libido and this did not accord with the general character of anxiety as a reaction to unpleasure.
     The objection to this view arose from our coming to regard the ego as the sole seat of anxiety. It was one of the results of the attempt at a structural division of the mental apparatus which I made in The Ego and the Id. Whereas the old view made it natural to suppose that anxiety arose from the libido belonging to the repressed instinctual impulses, the new one, on the contrary, made the ego the source of anxiety. Thus it is a question of instinctual (id-) anxiety or ego-anxiety. Since the energy which the ego employs is desexualized, the new view also tended to weaken the close connection between anxiety and libido. I hope I have at least succeeded in making the contradiction plain and in giving a clear idea of the point in doubt.
     Rank's contention--which was originally my own--,that the affect of anxiety is a consequence of the event of birth and a repetition of the situation then experienced, obliged me to review the problem of anxiety once more. But I could make no headway with his idea that birth is a trauma, states of anxiety a reaction of discharge to it and all subsequent affects of anxiety an attempt to 'abreact' it more and more completely. I was obliged to go back from the anxiety reaction to the situation of danger that lay behind it. The introduction of this element opened up new aspects of the question. Birth was seen to be the prototype of all later situations of danger which overtook the individual under the new conditions arising from a changed mode of life and a growing mental development. On the other hand its own significance was reduced to this prototypic relationship to danger. The anxiety felt at birth became the prototype of an affective state which had to undergo the same vicissitudes as the other affects. Either the state of anxiety reproduced itself automatically in situations analogous to the original situation and was thus an inexpedient form of reaction instead of an expedient one as it had been in the first situation of danger; or the ego acquired power over this affect, reproduced it on its own initiative, and employed it as a warning of danger and as a means of setting the pleasure-unpleasure mechanism in motion. We thus gave the biological aspect of the anxiety affect its due importance by recognizing anxiety as the general reaction to situations of danger; while we endorsed the part played by the ego as the seat of anxiety by allocating to it the function of producing the anxiety affect according to its needs. Thus we attributed two modes of origin to anxiety in later life. One was involuntary, automatic and always justified on economic grounds, and arose whenever a danger-situation analogous to birth had established itself. The other was produced by the ego as soon as a situation of this kind merely threatened to occur, in order to call for its avoidance. In the second case the ego subjects itself to anxiety as a sort of inoculation, submitting to a slight attack of the illness in order to escape its full strength. It vividly imagines the danger-situation, as it were, with the unmistakable purpose of restricting that distressing experience to a mere indication, a signal. We have already seen in detail how the various situations of danger arise one after the other, retaining at the same time a genetic connection.
     We shall perhaps be able to proceed a little further in our understanding of anxiety when we turn to the problem of the relation between neurotic anxiety and realistic anxiety.
     Our former hypothesis of a direct transformation of libido into anxiety possesses less interest for us now than it did. But if we do nevertheless consider it, we shall have to distinguish different cases. As regards anxiety evoked by the ego as a signal, it does not come into consideration; nor does it, therefore, in any of those danger-situations which move the ego to bring on repression. The libidinal cathexis of the repressed instinctual impulse is employed otherwise than in being transformed into anxiety and discharged as such--as is most clearly seen in conversion hysteria. On the other hand, further enquiry into the question of the danger-situation will bring to our notice an instance of the production of anxiety which will, I think, have to be accounted for in a different way.

(c) Repression and Defence
In the course of discussing the problem of anxiety I have revived a concept or, to put it more modestly, a term, of which I made exclusive use thirty years ago when I first began to study the subject but which I later abandoned. I refer to the term 'defensive process'. I afterwards replaced it by the word 'repression', but the relation between the two remained uncertain. It will be an undoubted advantage, I think, to revert to the old concept of 'defence', provided we employ it explicitly as a general designation for all the techniques which the ego makes use of in conflicts which may lead to a neurosis, while we retain the word 'repression' for the special method of defence which the line of approach taken by our investigations made us better acquainted with in the first instance.
     Even a purely terminological innovation ought to justify its adoption; it ought to reflect some new point of view or some extension of knowledge. The revival of the concept of defence and the restriction of that of repression takes into account a fact which has long since been known but which has received added importance owing to some new discoveries. Our first observations of repression and of the formation of symptoms were made in connection with hysteria. We found that the perceptual content of exciting experiences and the ideational content of pathogenic structures of thought were forgotten and debarred from being reproduced in memory, and we therefore concluded that the keeping away from consciousness was a main characteristic of hysterical repression. Later on, when we came to study the obsessional neuroses, we found that in that illness pathogenic occurrences are not forgotten. They remain conscious but they are 'isolated' in some way that we cannot as yet grasp, so that much the same result is obtained as in hysterical amnesia. Nevertheless the difference is great enough to justify the belief that the process by which instinctual demands are set aside in obsessional neurosis cannot be the same as in hysteria. Further investigations have shown that in obsessional neurosis a regression of the instinctual impulses to an earlier libidinal stage is brought about through the opposition of the ego, and that this regression, although it does not make repression unnecessary, clearly works in the same sense as repression. We have seen, too, that in obsessional neurosis anticathexis, which is also presumably present in hysteria, plays a specially large part in protecting the ego by effecting a reactive alteration in it. Our attention has, moreover, been drawn to a process of 'isolation' (whose technique cannot as yet be elucidated) which finds direct symptomatic manifestation, and to a procedure, that may be called magical, of 'undoing' what has been done-a procedure about whose defensive purpose there can be no doubt, but which has no longer any resemblance to the process of 'repression'. These observations provide good enough grounds for reintroducing the old concept of defence, which can cover all these processes that have the same purpose-namely, the protection of the ego against instinctual demands--and for subsuming repression under it as a special case. The importance of this nomenclature is heightened if we consider the possibility that further investigations may show that there is an intimate connection between special forms of defence and particular illnesses, as, for instance, between repression and hysteria. In addition we may look forward to the possible discovery of yet another important correlation. It may well be that before its sharp cleavage into an ego and an id, and before the formation of a super-ego, the mental apparatus makes use of different methods of defence from those which it employs after it has reached these stages of organization.


The affect of anxiety exhibits one or two features the study of which promises to throw further light on the subject. Anxiety [Angst] has an unmistakable relation to expectation: it is anxiety about something. It has a quality of indefiniteness and lack of object. In precise speech we use the word 'fear' [Furchf] rather than 'anxiety' if it has found an object. Moreover, in addition to its relation to danger, anxiety has a relation to neurosis which we have long been trying to elucidate. The question arises: why are not all reactions of anxiety neurotic--why do we accept so many of them as normal? And finally the problem of the difference between realistic anxiety and neurotic anxiety awaits a thorough examination.
     To begin with the last problem. The advance we have made is that we have gone behind reactions of anxiety to situations of danger. If we do the same thing with realistic anxiety we shall have no difficulty in solving the question. Real danger is a danger that is known, and realistic anxiety is anxiety about a known danger of this sort. Neurotic anxiety is anxiety about an unknown danger. Neurotic danger is thus a danger that has still to be discovered. Analysis has shown that it is an instinctual danger. By bringing this danger which is not known to the ego into consciousness, the analyst makes neurotic anxiety no different from realistic anxiety, so that it can be dealt with in the same way.
     There are two reactions to real danger. One is an affective reaction, an outbreak of anxiety. The other is a protective action. The same will presumably be true of instinctual danger. We know how the two reactions can co-operate in an expedient way, the one giving the signal for the other to appear. But we also know that they can behave in an inexpedient way: paralysis from anxiety may set in, and the one reaction spread at the cost of the other.
     In some cases the characteristics of realistic anxiety and neurotic anxiety are mingled. The danger is known and real but the anxiety in regard to it is over-great, greater than seems proper to us. It is this surplus of anxiety which betrays the presence of a neurotic element. Such cases, however, introduce no new principle; for analysis shows that to the known real danger an unknown instinctual one is attached.
     We can find out still more about this if, not content with tracing anxiety back to danger, we go on to enquire what the essence and meaning of a danger-situation is. Clearly, it consists in the subject's estimation of his own strength compared to the magnitude of the danger and in his admission of helplessness in the face of it-physical helplessness if the danger is real and psychical helplessness if it is instinctual. In doing this he will be guided by the actual experiences he has had. (Whether he is wrong in his estimation or not is immaterial for the outcome.) Let us call a situation of helplessness of this kind that has been actually experienced a traumatic situation. We shall then have good grounds for distinguishing a traumatic situation from a danger-situation.
     The individual will have made an important advance in his capacity for self-preservation if he can foresee and expect a traumatic situation of this kind which entails helplessness, instead of simply waiting for it to happen. Let us call a situation which contains the determinant for such an expectation a danger-situation. It is in this situation that the signal of anxiety is given. The signal announces: 'I am expecting a situation of helplessness to set in', or: 'The present situation reminds me of one of the traumatic experiences I have had before. Therefore I will anticipate the trauma and behave as though it had already come, while there is yet time to turn it aside.' Anxiety is therefore on the one hand an expectation of a trauma, and on the other a repetition of it in a mitigated form. Thus the two features of anxiety which we have noted have a different origin. Its connection with expectation belongs to the danger-situation, whereas its indefiniteness and lack of object belong to the traumatic situation of helplessness--the situation which is anticipated in the danger-situation.
     Taking this sequence, anxiety--danger--helplessness (trauma), we can now summarize what has been said. A danger-situation is a recognized, remembered, expected situation of helplessness. Anxiety is the original reaction to helplessness in the trauma and is reproduced later on in the danger-situation as a signal for help. The ego, which experienced the trauma passively, now repeats it actively in a weakened version, in the hope of being able itself to direct its course. It is certain that children behave in this fashion towards every distressing impression they receive, by reproducing it in their play. In thus changing from passivity to activity they attempt to master their experiences psychically. If this is what is meant by 'abreacting a trauma' we can no longer have anything to urge against the phrase. But what is of decisive importance is the first displacement of the anxiety-reaction from its origin in the situation of helplessness to an expectation of that situation--that is, to the danger-situation. After that come the later displacements, from the danger to the determinant of the danger-loss of the object and the modifications of that loss with which we are already acquainted.
     The undesirable result of 'spoiling' a small child is to magnify the importance of the danger of losing the object (the object being a protection against every situation of helplessness) in comparison with every other danger. It therefore encourages the individual to remain in the state of childhood, the period of life which is characterized by motor and psychical helplessness.
     So far we have had no occasion to regard realistic anxiety in any different light from neurotic anxiety. We know what the distinction is. A real danger is a danger which threatens a person from an external object, and a neurotic danger is one which threatens him from an instinctual demand. In so far as the instinctual demand is something real, his neurotic anxiety, too, can be admitted to have a realistic basis. We have seen that the reason why there seems to be a specially close connection between anxiety and neurosis is that the ego defends itself against an instinctual danger with the help of the anxiety reaction just as it does against an external real danger, but that this line of defensive activity eventuates in a neurosis owing to an imperfection of the mental apparatus. We have also come to the conclusion that an instinctual demand often only becomes an (internal) danger because its satisfaction would bring on an external danger--that is, because the internal danger represents an external one.
     On the other hand, the external (real) danger must also have managed to become internalized if it is to be significant for the ego. It must have been recognized as related to some situation of helplessness that has been experienced. It may quite often happen that although a danger-situation is correctly estimated in itself, a certain amount of instinctual anxiety is added to the realistic anxiety. In that case the instinctual demand before whose satisfaction the ego recoils is a masochistic one: the instinct of destruction directed against the subject himself. Perhaps an addition of this kind explains cases in which reactions of anxiety are exaggerated, inexpedient or paralysing. Phobias of heights (windows, towers, precipices and so on) may have some such origin. Their hidden feminine significance is closely connected with masochism. Man seems not to have been endowed, or to have been endowed to only a very small degree, with an instinctive recognition of the dangers that threaten him from without. Small children are constantly doing things which endanger their lives, and that is precisely why they cannot afford to be without a protecting object. In relation to the traumatic situation, in which the subject is helpless, external and internal dangers, real dangers and instinctual demands converge. Whether the ego is suffering from a pain which will not stop or experiencing an accumulation of instinctual needs which cannot obtain satisfaction, the economic situation is the same, and the motor helplessness of the ego finds expression in psychical helplessness.
     In this connection the puzzling phobias of early childhood deserve to be mentioned once again. We have been able to explain some of them, such as the fear of being alone or in the dark or with strangers, as reactions to the danger of losing the object. Others, like the fear of small animals, thunderstorms, etc., might perhaps be accounted for as vestigial traces of the congenital preparedness to meet real dangers which is so strongly developed in other animals. In man, only that part of this archaic heritage is appropriate which has reference to the loss of the object. If childhood phobias become fixated and grow stronger and persist into later years, analysis shows that their content has become associated with instinctual demands and has come to stand for internal dangers as well.


So little is known about the psychology of emotional processes that the tentative remarks I am about to make on the subject may claim a very lenient judgment. The problem before us arises out of the conclusion we have reached that anxiety comes to be a reaction to the danger of a loss of an object. Now we already know one reaction to the loss of an object, and that is mourning. The question therefore is, when does that loss lead to anxiety and when to mourning? In discussing the subject of mourning on a previous occasion I found that there was one feature about it which remained quite unexplained. This was its peculiar painfulness. And yet it seems self-evident that separation from an object should be painful. Thus the problem becomes more complicated: when does separation from an object produce anxiety, when does it produce mourning and when does it produce, it may be, only pain?
     Let me say at once that there is no prospect in sight of answering these questions. We must content ourselves with drawing certain distinctions and adumbrating certain possibilities.
     Our starting-point will again be the one situation which we believe we understand--the situation of the infant when it is presented with a stranger instead of its mother. It will exhibit the anxiety which we have attributed to the danger of loss of object. But its anxiety is undoubtedly more complicated than this and merits a more thorough discussion. That it does have anxiety there can be no doubt; but the expression of its face and its reaction of crying indicate that it is feeling pain as well. Certain things seem to be joined together in it which will later on be separated out. It cannot as yet distinguish between temporary absence and permanent loss. As soon as it loses sight of its mother it behaves as if it were never going to see her again; and repeated consoling experiences to the contrary are necessary before it learns that her disappearance is usually followed by her re-appearance. Its mother encourages this piece of knowledge which is so vital to it by playing the familiar game of hiding her face from it with her hands and then, to its joy, uncovering it again. In these circumstances it can, as it were, feel longing unaccompanied by despair.
     In consequence of the infant's misunderstanding of the facts, the situation of missing its mother is not a danger-situation but a traumatic one. Or, to put it more correctly, it is a traumatic situation if the infant happens at the time to be feeling a need which its mother should be the one to satisfy. It turns into a danger-situation if this need is not present at the moment. Thus, the first determinant of anxiety, which the ego itself introduces, is loss of perception of the object (which is equated with loss of the object itself). There is as yet no question of loss of love. Later on, experience teaches the child that the object can be present but angry with it; and then loss of love from the object becomes a new and much more enduring danger and determinant of anxiety.
     The traumatic situation of missing the mother differs in one important respect from the traumatic situation of birth. At birth no object existed and so no object could be missed. Anxiety was the only reaction that occurred. Since then repeated situations of satisfaction have created an object out of the mother; and this object, whenever the infant feels a need, receives an intense cathexis which might be described as a 'longing' one. It is to this new aspect of things that the reaction of pain is referable. Pain is thus the actual reaction to loss of object, while anxiety is the reaction to the danger which that loss entails and, by a further displacement, a reaction to the danger of the loss of object itself.
     We know very little about pain either. The only fact we are certain of is that pain occurs in the first instance and as a regular thing whenever a stimulus which impinges on the periphery breaks through the devices of the protective shield against stimuli and proceeds to act like a continuous instinctual stimulus, against which muscular action, which is as a rule effective because it withdraws the place that is being stimulated from the stimulus, is powerless. If the pain proceeds not from a part of the skin but from an internal organ, the situation is still the same. All that has happened is that a portion of the inner periphery has taken the place of the outer periphery. The child obviously has occasion to undergo experiences of pain of this sort, which are independent of its experiences of need. This determinant of the generating of pain seems, however, to have very little similarity with the loss of an object. And besides, the element which is essential to pain, peripheral stimulation, is entirely absent in the child's situation of longing. Yet it cannot be for nothing that the common usage of speech should have created the notion of internal, mental pain and have treated the feeling of loss of object as equivalent to physical pain.
     When there is physical pain, a high degree of what may be termed narcissistic cathexis of the painful place occurs. This cathexis continues to increase and tends, as it were, to empty the ego. It is well known that when internal organs are giving us pain we receive spatial and other presentations of parts of the body which are ordinarily not represented at all in conscious ideation. Again, the remarkable fact that, when there is a psychical diversion brought about by some other interest, even the most intense physical pains fail to arise (I must not say 'remain unconscious' in this case) can be accounted for by there being a concentration of cathexis on the psychical representative of the part of the body which is giving pain. I think it is here that we shall find the point of analogy which has made it possible to carry sensations of pain over to the mental sphere. For the intense cathexis of longing which is concentrated on the missed or lost object (a cathexis which steadily mounts up because it cannot be appeased) creates the same economic conditions as are created by the cathexis of pain which is concentrated on the injured part of the body. Thus the fact of the peripheral causation of physical pain can be left out of account. The transition from physical pain to mental pain corresponds to a change from narcissistic cathexis to object-cathexis. An object-presentation which is highly cathected by instinctual need plays the same role as a part of the body which is cathected by an increase of stimulus. The continuous nature of the cathectic process and the impossibility of inhibiting it produce the same state of mental helplessness. If the feeling of unpleasure which then arises has the specific character of pain (a character which cannot be more exactly described) instead of manifesting itself in the reactive form of anxiety, we may plausibly attribute this to a factor which we have not sufficiently made use of in our explanations--the high level of cathexis and 'binding' that prevails while these processes which lead to a feeling of unpleasure take place.
     We know of yet another emotional reaction to the loss of an object, and that is mourning. But we have no longer any difficulty in accounting for it. Mourning occurs under the influence of reality-testing; for the latter function demands categorically from the bereaved person that he should separate himself from the object, since it no longer exists. Mourning is entrusted with the task of carrying out this retreat from the object in all those situations in which it was the recipient of a high degree of cathexis. That this separation should be painful fits in with what we have just said, in view of the high and unsatisfiable cathexis of longing which is concentrated on the object by the bereaved person during the reproduction of the situations in which he must undo the ties that bind him to it.