By Sigmunf Freud, 1937


Experience has taught us that psycho-analytic therapy – the freeing of someone from his neurotic symptoms, inhibitions and abnormalities of character – is a time-consuming business. Hence, from the very first, attempts have been made to shorten the duration of analyses. Such endeavours required no justification; they could claim to be based on the strongest considerations of reason and expediency. But there was probably still at work in them as well some trace of the impatient contempt with which the medical science of an earlier day regarded the neuroses as being uncalled-for consequences of invisible injuries. If it had now become necessary to attend to them, they should at least be disposed of as quickly as possible.

A particularly energetic attempt in this direction was made by Otto Rank, following upon his book, The Trauma of Birth (1924). He supposed that the true source of neurosis was the act of birth, since this involves the possibility of a child’s ‘primal fixation’ to his mother not being surmounted but persisting as a ‘primal repression’. Rank hoped that if this primal trauma were dealt with by a subsequent analysis the whole neurosis would be got rid of. Thus this one small piece of analytic work would save the necessity for all the rest. And a few months should be enough to accomplish this. It cannot be disputed that Rank’s argument was bold and ingenious; but it did not stand the test of critical examination. Moreover, it was a child of its time, conceived under the stress of the contrast between the post-war misery of Europe and the ‘prosperity’ of America, and designed to adapt the tempo of analytic therapy to the haste of American life. We have not heard much about what the implementation of Rank’s plan has done for cases of sickness. Probably not more than if the fire-brigade, called to deal with a house that had been set on fire by an overturned oil-lamp, contented themselves with removing the lamp from the room in which the blaze had started. No doubt a considerable shortening of the brigade’s activities would be effected by this means. The theory and practice of Rank’s experiment are now things of the past – no less than American ‘prosperity’ itself.

I myself had adopted another way of speeding up an analytic treatment even before the war. At that time I had taken on the case of a young Russian, a man spoilt by wealth, who had come to Vienna in a state of complete helplessness, accompanied by a private doctor and an attendant.¹ In the course of a few years it was possible to give him back a large amount of his independence, to awaken his interest in life and to adjust his relations to the people most important to him. But there progress came to a stop. We advanced no further in clearing up the neurosis of his childhood, on which his later illness was based, and it was obvious that the patient found his present position highly comfortable and had no wish to take any step forward which would bring him nearer to the end of his treatment. It was a case of the treatment inhibiting itself: it was in danger of failing as a result of its – partial – success. In this predicament I resorted to the heroic measure of fixing a time limit for the analysis. At the beginning of a year’s work I informed the patient that the coming year was to be the last one of his treatment, no matter what he achieved in the time still left to him. At first he did not believe me, but once he was convinced that I was in deadly earnest, the desired change set in. His resistances shrank up, and in these last months of his treatment he was able to reproduce all the memories and to discover all the connections which seemed necessary for understanding his early neurosis and mastering his present one. When he left me in the midsummer of 1914, with as little suspicion as the rest of us of what lay so shortly ahead, I believed that his cure was radical and permanent.

¹ See my paper published with the patient’s consent, ‘From the History of an Infantile Neurosis’ (1918b). It contains no detailed account of the young man’s adult illness, which is touched on only when its connection with his infantile neurosis absolutely requires it.

In a footnote added to this patient’s case history in 1923, I have already reported that I was mistaken. When, towards the end of the war, he returned to Vienna, a refugee and destitute, I had to help him to master a part of the transference which had not been resolved. This was accomplished in a few months, and I was able to end my footnote with the statement that ‘since then the patient has felt normal and has behaved unexceptionably, in spite of the war having robbed him of his home, his possessions, and all his family relationships’. Fifteen years have passed since then without disproving the truth of this verdict; but certain reservations have become necessary. The patient has stayed on in Vienna and has kept a place in society, if a humble one. But several times during this period his good state of health has been interrupted by attacks of illness which could only be construed as offshoots of his perennial neurosis. Thanks to the skill of one of my pupils, Dr. Ruth Mack Brunswick, a short course of treatment has on each occasion brought these conditions to an end. I hope that Dr. Mack Brunswick herself will shortly report on the circumstances. Some of these attacks were still concerned with residual portions of the transference; and, where this was so, short-lived though they were, they showed a distinctly paranoid character. In other attacks, however, the pathogenic material consisted of pieces of the patient’s childhood history, which had not come to light while I was analysing him and which now came away – the comparison is unavoidable – like sutures after an operation, or small fragments of necrotic bone. I have found the history of this patient’s recovery scarcely less interesting than that of his illness.

I have subsequently employed this fixing of a time-limit in other cases as well, and I have also taken the experiences of other analysts into account. There can be only one verdict about the value of this blackmailing device: it is effective provided that one hits the right time for it. But it cannot guarantee to accomplish the task completely. On the contrary, we may be sure that, while part of the material will become accessible under the pressure of the threat, another part will be kept back and thus become buried, as it were, and lost to our therapeutic efforts. For once the analyst has fixed the time-limit he cannot extend it; otherwise the patient would lose all faith in him. The most obvious way out would be for the patient to continue his treatment with another analyst, although we know that such a change will involve a fresh loss of time and abandoning fruits of work already done. Nor can any general rule be laid down as to the right time for resorting to this forcible technical device; the decision must be left to the analyst’s tact. A miscalculation cannot be rectified. The saying that a lion only springs once must apply here.

The discussion of the technical problem of how to accelerate the slow progress of an analysis leads us to another, more deeply interesting question: is there such a thing as a natural end to an analysis – is there any possibility at all of bringing an analysis to such an end? To judge by the common talk of analysts it would seem to be so, for we often hear them say, when they are deploring or excusing the recognized imperfections of some fellow-mortal: ‘His analysis was not finished’ or ‘he was never analysed to the end.’

We must first of all decide what is meant by the ambiguous phrase ‘the end of an analysis’. From a practical standpoint it is easy to answer. An analysis is ended when the analyst and the patient cease to meet each other for the analytic session. This happens when two conditions have been approximately fulfilled: first, that the patient shall no longer be suffering from his symptoms and shall have overcome his anxieties and his inhibitions; and secondly, that the analyst shall judge that so much repressed material has been made conscious, so much that was unintelligible has been explained, and so much internal resistance conquered, that there is no need to fear a repetition of the pathological processes concerned. If one is prevented by external difficulties from reaching this goal, it is better to speak of an incomplete analysis rather than of an unfinished one.

The other meaning of the ‘end’ of an analysis is much more ambitious. In this sense of it, what we are asking is whether the analyst has had such a far-reaching influence on the patient that no further change could be expected to take place in him if his analysis were continued. It is as though it were possible by means of analysis to attain to a level of absolute psychical normality – a level, moreover, which we could feel confident would be able to remain stable, as though, perhaps, we had succeeded in resolving every one of the patient’s repressions and in filling in all the gaps in his memory. We may first consult our experience to enquire whether such things do in fact happen, and then turn to our theory to discover whether there is any possibility of their happening.

Every analyst will have treated a few cases which have had this gratifying outcome. He has succeeded in clearing up the patient’s neurotic disturbance, and it has not returned and has not been replaced by any other such disturbance. Nor are we without some insight into the determinants of these successes. The patient’s ego had not been noticeably altered and the aetiology of his disturbance had been essentially traumatic. The aetiology of every neurotic disturbance is, after all, a mixed one. It is a question either of the instincts being excessively strong – that is to say, recalcitrant to taming by the ego – or of the effects of early (i.e. premature) traumas which the immature ego was unable to master. As a rule there is a combination of both factors, the constitutional and the accidental. The stronger the constitutional factor, the more readily will a trauma lead to a fixation and leave behind a developmental disturbance; the stronger the trauma, the more certainly will its injurious effects become manifest even when the instinctual situation is normal. There is no doubt that an aetiology of the traumatic sort offers by far the more favourable field for analysis. Only when a case is a predominantly traumatic one will analysis succeed in doing what it is so superlatively able to do; only then will it, thanks to having strengthened the patient’s ego, succeed in replacing by a correct solution the inadequate decision made in his early life. Only in such cases can one speak of an analysis having been definitively ended. In them, analysis has done all that it should and does not need to be continued. It is true that, if the patient who has been restored in this way never produces another disorder calling for analysis, we do not know how much his immunity may not be due to a kind fate which has spared him ordeals that are too severe.

A constitutional strength of instinct and an unfavourable alteration of the ego acquired in its defensive struggle in the sense of its being dislocated and restricted – these are the factors which are prejudicial to the effectiveness of analysis and which may make its duration interminable. One is tempted to make the first factor – strength of instinct – responsible as well for the emergence of the second – the alteration of the ego; but it seems that the latter too has an aetiology of its own. And, indeed, it must be admitted that our knowledge of these matters is as yet insufficient. They are only now becoming the subject of analytic study. In this field the interest of analysts seems to me to be quite wrongly directed. Instead of an enquiry into how a cure by analysis comes about (a matter which I think has been sufficiently elucidated) the question should be asked of what are the obstacles that stand in the way of such a cure.

This brings me to two problems which arise directly out of analytic practice, as I hope to show by the following examples. A certain man, who had himself practised analysis with great success, came to the conclusion that his relations both to men and women – to the men who were his competitors and to the woman whom he loved – were nevertheless not free from neurotic impediments; and he therefore made himself the subject of an analysis by someone else whom he regarded as superior to himself. This critical illumination of his own self had a completely successful result. He married the woman he loved and turned into a friend and teacher of his supposed rivals. Many years passed in this way, during which his relations with his former analyst also remained unclouded. But then, for no assignable external reason, trouble arose. The man who had been analysed became antagonistic to the analyst and reproached him for having failed to give him a complete analysis. The analyst, he said, ought to have known and to have taken into account the fact that a transference-relation can never be purely positive; he should have given his attention to the possibilities of a negative transference. The analyst defended himself by saying that, at the time of the analysis, there was no sign of a negative transference. But even if he had failed to observe some very faint signs of it – which was not altogether ruled out, considering the limited horizon of analysis in those early days – it was still doubtful, he thought, whether he would have had the power to activate a topic (or, as we say, a ‘complex’) by merely pointing it out, so long as it was not currently active in the patient himself at the time. To activate it would certainly have required some unfriendly piece of behaviour in reality on the analyst’s part. Furthermore, he added, not every good relation between an analyst and his subject during and after analysis was to be regarded as a transference; there were also friendly relations which were based on reality and which proved to be viable.

I now pass on to my second example, which raises the same problem. An unmarried woman, no longer young, had been cut off from life since puberty by an inability to walk, owing to severe pains in the legs. Her condition was obviously of a hysterical nature, and it had defied many kinds of treatment. An analysis lasting three-quarters of a year removed the trouble and restored to the patient, an excellent and worthy person, her right to a share in life. In the years following her recovery she was consistently unfortunate. There were disasters in her family, and financial losses, and, as she grew older, she saw every hope of happiness in love and marriage vanish. But the one-time invalid stood up to all this valiantly and was a support to her family in difficult times. I cannot remember whether it was twelve or fourteen years after the end of her analysis that, owing to profuse haemorrhages, she was obliged to undergo a gynaecological examination. A myoma was found, which made a complete hysterectomy advisable. From the time of this operation, the woman became ill once more. She fell in love with her surgeon, wallowed in masochistic phantasies about the fearful changes in her inside – phantasies with which she concealed her romance – and proved inaccessible to a further attempt at analysis. She remained abnormal to the end of her life. The successful analytic treatment took place so long ago that we cannot expect too much from it; it was in the earliest years of my work as an analyst. No doubt the patient’s second illness may have sprung from the same source as her first one which had been successfully overcome: it may have been a different manifestation of the same repressed impulses, which the analysis had only incompletely resolved. But I am inclined to think that, were it not for the new trauma, there would have been no fresh outbreak of neurosis.

These two examples, which have been purposely selected from a large number of similar ones, will suffice to start a discussion of the topics we are considering. The sceptical, the optimistic and the ambitious will take quite different views of them. The first will say that it is now proved that even a successful analytic treatment does not protect the patient, who at the time has been cured, from falling ill later on of another neurosis – for, indeed, of a neurosis derived from the same instinctual root – that is to say, from a recurrence of his old trouble. The others will consider that this is not proved. They will object that the two examples date from the early days of analysis, twenty and thirty years ago, respectively; and that since then we have acquired deeper insight and wider knowledge, and that our technique has changed in accordance with our new discoveries. To-day, they will say, we may demand and expect that an analytic cure shall prove permanent, or at least that if a patient falls ill again, his new illness shall not turn out to be a revival of his earlier instinctual disturbance, manifesting itself in new forms. Our experience, they will maintain, does not oblige us to restrict so materially the demands that can be made upon our therapeutic method.

My reason for choosing these two examples is, of course, precisely because they lie so far back in the past. It is obvious that the more recent the successful outcome of an analysis is, the less utilizable it is for our discussion, since we have no means of predicting what the later history of the recovery will be. The optimists’ expectations clearly presuppose a number of things which are not precisely self-evident. They assume, firstly, that there really is a possibility of disposing of an instinctual conflict (or, more correctly, a conflict between the ego and an instinct) definitively and for all time; secondly, that while we are treating someone for one instinctual conflict we can, as it were, inoculate him against the possibility of any other such conflicts; and thirdly, that we have the power, for purposes of prophylaxis, to stir up a pathogenic conflict of this sort which is not betraying itself at the time by any indications, and that it is wise to do so. I throw out these questions without proposing to answer them now. Perhaps it may not be possible at present to give any certain answer to them at all.

Some light may probably be thrown on them by theoretical considerations. But another point has already become clear: if we wish to fulfil the more exacting demands upon analytic therapy, our road will not lead us to, or by way of, a shortening of its duration.

An analytic experience which now extends over several decades, and a change which has taken place in the nature and mode of my activity, encourage me to attempt to answer the questions before us. In earlier days I treated quite a large number of patients, who, as was natural, wanted to be dealt with as quickly as possible. Of late years I have been mainly engaged in training analyses; a relatively small number of severe cases of illness remained with me for continuous treatment, interrupted, however, by longer or shorter intervals. With them, the therapeutic aim was no longer the same. There was no question of shortening the treatment; the purpose was radically to exhaust the possibilities of illness in them and to bring about a deep-going alteration of their personality.

Of the three factors which we have recognized as being decisive for the success or otherwise of analytic treatment – the influence of traumas, the constitutional strength of the instincts and alterations of the ego – what concerns us here is only the second, the strength of the instincts. A moment’s reflection raises a doubt whether the restrictive use of the adjective ‘constitutional’ (or ‘congenital’) is essential. However true it may be that the constitutional factor is of decisive importance from the very beginning, it is nevertheless conceivable that a reinforcement of instinct coming later in life might produce the same effects. If so, we should have to modify our formula and say ‘the strength of the instincts at the time’ instead of ‘the constitutional strength of the instincts’. The first of our questions was: ‘Is it possible by means of analytic therapy to dispose of a conflict between an instinct and the ego, or of a pathogenic instinctual demand upon the ego, permanently and definitively?’ To avoid misunderstanding it is not unnecessary, perhaps, to explain more exactly what is meant by ‘permanently disposing of an instinctual demand’. Certainly not ‘causing the demand to disappear so that nothing more is ever heard from it again’. This is in general impossible, nor is it at all to be desired. No, we mean something else, something which may be roughly described as a ‘taming’ of the instinct. That is to say, the instinct is brought completely into the harmony of the ego, becomes accessible to all the influences of the other trends in the ego and no longer seeks to go its independent way to satisfaction. If we are asked by what methods and means this result is achieved, it is not easy to find an answer. We can only say: ‘So muss denn doch die Hexe dran!’ – the Witch Metapsychology. Without metapsychological speculation and theorizing – I had almost said ‘phantasying’ – we shall not get another step forward. Unfortunately, here as elsewhere, what our Witch reveals is neither very clear nor very detailed. We have only a single clue to start from – though it is a clue of the highest value – namely, the antithesis between the primary and the secondary processes; and to that antithesis I shall at this point turn.

If now we take up our first question once more, we find that our new line of approach inevitably leads us to a particular conclusion. The question was whether it is possible to dispose of an instinctual conflict permanently and definitively – i.e. to ‘tame’ an instinctual demand in that fashion. Formulated in these terms, the question makes no mention at all of the strength of the instinct; but it is precisely on this that the outcome depends. Let us start from the assumption that what analysis achieves for neurotics is nothing other than what normal people bring about for themselves without its help. Everyday experience, however, teaches us that in a normal person any solution of an instinctual conflict only holds good for a particular strength of instinct, or, more correctly, only for a particular relation between the strength of the instinct and the strength of the ego.¹ If the strength of the ego diminishes, whether through illness or exhaustion, or from some similar cause, all the instincts which have so far been successfully tamed may renew their demands and strive to obtain substitutive satisfactions in abnormal ways.² Irrefutable proof of this statement is supplied by our nightly dreams; they react to the sleeping attitude assumed by the ego with an awakening of instinctual demands.

¹ Or, to be perfectly accurate, where that relation falls within certain limits.

² Here we have a justification of the claim to aetiological importance of such non-specific factors as overwork, shock, etc. These factors have always been assured of general recognition, but have had to be pushed into the background precisely by psycho-analysis. It is impossible to define health except in metapsychological terms: i.e. by reference to the dynamic relations between the agencies of the mental apparatus which have been recognized – or (if that is preferred) inferred or conjectured – by us.

The material on the other side is equally unambiguous. Twice in the course of individual development certain instincts are considerably reinforced: at puberty, and, in women, at the menopause. We are not in the least surprised if a person who was not neurotic before becomes so at these times. When his instincts were not so strong, he succeeded in taming them; but when they are reinforced he can no longer do so. The repressions behave like dams against the pressure of water. The same effects which are produced by these two physiological reinforcements of instinct, may be brought about in an irregular fashion by accidental causes at any other period of life. Such reinforcements may be set up by fresh traumas, enforced frustrations, or the collateral influence of instincts upon one another. The result is always the same, and it underlines the irresistible power of the quantitative factor in the causation of illness.

I feel as though I ought to be ashamed of so much ponderous exposition, seeing that everything I have said has long been familiar and self-evident. It is a fact that we have always behaved as if we knew all this; but, for the most part, our theoretical concepts have neglected to attach the same importance to the economic line of approach as they have to the dynamic and topographical ones. My excuse is therefore that I am drawing attention to this neglect.

Before we decide on an answer to this question, however, we must consider an objection whose force lies in the fact that we are probably predisposed in its favour. Our arguments, it will be said, are all deduced from the processes which take place spontaneously between the ego and the instincts, and they presuppose that analytic therapy can accomplish nothing which does not, under favourable and normal conditions, occur of itself. But is this really so? Is it not precisely the claim of our theory that analysis produces a state which never does arise spontaneously in the ego and that this newly created state constitutes the essential difference between a person who has been analysed and a person who has not? Let us bear in mind what this claim is based on. All repressions take place in early childhood; they are primitive defensive measures taken by the immature, feeble ego. In later years no fresh repressions are carried out; but the old ones persist, and their services continue to be made use of by the ego for mastering the instincts. New conflicts are disposed of by what we call ‘after-repression’. We may apply to these infantile repressions our general statement that repressions depend absolutely and entirely on the relative strength of the forces involved and that they cannot hold out against an increase in the strength of the instincts. Analysis, however, enables the ego, which has attained greater maturity and strength, to undertake a revision of these old repressions; a few are demolished, while others are recognized but constructed afresh out of more solid material. These new dams are of quite a different degree of firmness from the earlier ones; we may be confident that they will not give way so easily before a rising flood of instinctual strength. Thus the real achievement of analytic therapy would be the subsequent correction of the original process of repression, a correction which puts an end to the dominance of the quantitative factor.

Thus far our theory, which we cannot give up except under irresistible compulsion. And what does our experience have to say to this? Perhaps our experience is not yet wide enough for us to come to a settled conclusion. It confirms our expectations often enough, but not always. One has an impression that one ought not to be surprised if it should turn out in the end that the difference between a person who has not been analysed and the behaviour of a person after he has been analysed is not so thorough-going as we aim at making it and as we expect and maintain it to be. If this is so, it would mean that analysis sometimes succeeds in eliminating the influence of an increase in instinct, but not invariably, or that the effect of analysis is limited to increasing the power of resistance of the inhibitions, so that they are equal to much greater demands than before the analysis or if no analysis had taken place. I really cannot commit myself to a decision on this point, nor do I know whether a decision is possible at the present time.

There is, however, another angle from which we can approach this problem of the variability in the effect of analysis. We know that the first step towards attaining intellectual mastery of our environment is to discover generalizations, rules and laws which bring order into chaos. In doing this we simplify the world of phenomena; but we cannot avoid falsifying it, especially if we are dealing with processes of development and change. What we are concerned with is discerning a qualitative alteration, and as a rule in doing so we neglect, at any rate to begin with, a quantitative factor. In the real world, transitions and intermediate stages are far more common than sharply differentiated opposite states. In studying developments and changes we direct our attention solely to the outcome; we readily overlook the fact that such processes are usually more or less incomplete – that is to say, that they are in fact only partial alterations. A shrewd satirist of old Austria, Johann Nestroy, once said: ‘Every step forward is only half as big as it looks at first.’ It is tempting to attribute a quite general validity to this malicious dictum. There are nearly always residual phenomena, a partial hanging-back. When an open-handed Maecenas surprises us by some isolated trait of miserliness, or when a person who is consistently over-kind suddenly indulges in a hostile action, such ‘residual phenomena’ are invaluable for genetic research. They show us that these praiseworthy and precious qualities are based on compensation and overcompensation which, as was to have been expected, have not been absolutely and fully successful. Our first account of the development of the libido was that an original oral phase gave way to a sadistic-anal phase and that this was in turn succeeded by a phallic-genital one. Later research has not contradicted this view, but it has corrected it by adding that these replacements do not take place all of a sudden but gradually, so that portions of the earlier organization always persist alongside of the more recent one, and even in normal development the transformation is never complete and residues of earlier libidinal fixations may still be retained in the final configuration. The same thing is to be seen in quite other fields. Of all the erroneous and superstitious beliefs of mankind that have supposedly been surmounted there is not one whose residues do not live on among us to-day in the lower strata of civilized peoples or even in the highest strata of cultural society. What has once come to life clings tenaciously to its existence. One feels inclined to doubt sometimes whether the dragons of primaeval days are really extinct.

Applying these remarks to our present problem, I think that the answer to the question of how to explain the variable results of our analytic therapy might well be that we, too, in endeavouring to replace repressions that are insecure by reliable ego-syntonic controls, do not always achieve our aim to its full extent – that is, do not achieve it thoroughly enough. The transformation is achieved, but often only partially: portions of the old mechanisms remain untouched by the work of analysis. It is difficult to prove that this is really so; for we have no other way of judging what happens but by the outcome which we are trying to explain. Nevertheless, the impressions one receives during the work of analysis do not contradict this assumption; indeed, they seem rather to confirm it. But we must not take the clarity of our own insight as a measure of the conviction which we produce in the patient. His conviction may lack ‘depth’, as one might say; it is always a question of the quantitative factor, which is so easily overlooked. If this is the correct answer to our question, we may say that analysis, in claiming to cure neuroses by ensuring control over instinct, is always right in theory but not always right in practice. And this is because it does not always succeed in ensuring to a sufficient degree the foundations on which a control of instinct is based. The cause of such a partial failure is easily discovered. In the past, the quantitative factor of instinctual strength opposed the ego’s defensive efforts; for that reason we called in the work of analysis to help; and now that same factor sets a limit to the efficacy of this new effort. If the strength of the instinct is excessive, the mature ego, supported by analysis, fails in its task, just as the helpless ego failed formerly. Its control over instinct is improved, but it remains imperfect because the transformation in the defensive mechanism is only incomplete. There is nothing surprising in this, since the power of the instruments with which analysis operates is not unlimited but restricted, and the final upshot always depends on the relative strength of the psychical agencies which are struggling with one another.

No doubt it is desirable to shorten the duration of analytic treatment, but we can only achieve our therapeutic purpose by increasing the power of analysis to come to the assistance of the ego. Hypnotic influence seemed to be an excellent instrument for our purposes; but the reasons for our having to abandon it are well known. No substitute for hypnosis has yet been found. From this point of view we can understand how such a master of analysis as Ferenczi came to devote the last years of his life to therapeutic experiments, which, unhappily, proved to be vain.

The two further questions – whether, while we are treating one instinctual conflict, we can protect a patient from future conflicts, and whether it is feasible and expedient, for prophylactic purposes, to stir up a conflict which is not at the time manifest – must be treated together; for obviously the first task can only be carried out in so far as the second one is – that is, in so far as a possible future conflict is turned into an actual present one upon which influence is then brought to bear. This new way of stating the problem is at bottom only an extension of the earlier one. Whereas in the first instance we were considering how to guard against a return of the same conflict, we are now considering how to guard against its possible replacement by another conflict. This sounds a very ambitious proposal, but all we are trying to do is to make clear what limits are set to the efficacy of analytic therapy.

However much our therapeutic ambition may be tempted to undertake such tasks, experience flatly rejects the notion. If an instinctual conflict is not a currently active one, is not manifesting itself, we cannot influence it even by analysis. The warning that we should let sleeping dogs lie, which we have so often heard in connection with our efforts to explore the psychical underworld, is peculiarly inapposite when applied to the conditions of mental life. For if the instincts are causing disturbances, it is a proof that the dogs are not sleeping; and if they seem really to be sleeping, it is not in our power to awaken them. This last statement, however, does not seem to be quite accurate and calls for a more detailed discussion. Let us consider what means we have at our disposal for turning an instinctual conflict which is at the moment latent into one which is currently active. Obviously there are only two things that we can do. We can bring about situations in which the conflict does become currently active, or we can content ourselves with discussing it in the analysis and pointing out the possibility of its arising. The first of these two alternatives can be carried out in two ways: in reality, or in the transference – in either case by exposing the patient to a certain amount of real suffering through frustration and the damming up of libido. Now it is true that we already make use of a technique of this kind in our ordinary analytic procedure. What would otherwise be the meaning of the rule that analysis must be carried out ‘in a state of frustration’? But this is a technique which we use in treating a conflict which is already currently active. We seek to bring this conflict to a head, to develop it to its highest pitch, in order to increase the instinctual force available for its solution. Analytic experience has taught us that the better is always the enemy of the good and that in every phase of the patient’s recovery we have to fight against his inertia, which is ready to be content with an incomplete solution.

If, however, what we are aiming at is a prophylactic treatment of instinctual conflicts that are not currently active but merely potential, it will not be enough to regulate sufferings which are already present in the patient and which he cannot avoid. We should have to make up our minds to provoke fresh sufferings in him; and this we have hitherto quite rightly left to fate. We should receive admonitions from all sides against the presumption of vying with fate in subjecting poor human creatures to such cruel experiments. And what sort of experiments would they be? Could we, for purposes of prophylaxis, take the responsibility of destroying a satisfactory marriage, or causing a patient to give up a post upon which his livelihood depends? Fortunately, we never find ourselves in the position of having to consider whether such interventions in the patient’s real life are justified; we do not possess the plenary powers which they would necessitate, and the subject of our therapeutic experiment would certainly refuse to co-operate in it. In practice, then, such a procedure is virtually excluded; but there are, besides, theoretical objections to it. For the work of analysis proceeds best if the patient’s pathogenic experiences belong to the past, so that his ego can stand at a distance from them. In states of acute crisis analysis is to all intents and purposes unusable. The ego’s whole interest is taken up by the painful reality and it withholds itself from analysis, which is attempting to go below the surface and uncover the influences of the past. To create a fresh conflict would thus only be to make the work of analysis longer and more difficult.

It will be objected that these remarks are quite unnecessary. Nobody thinks of purposely conjuring up new situations of suffering in order to make it possible for a latent instinctual conflict to be treated. This would not be much to boast of as a prophylactic achievement. We know, for instance, that a patient who has recovered from scarlet fever is immune to a return of the same illness; yet it never occurs to a doctor to take a healthy person who may possibly fall ill of scarlet fever and infect him with scarlet fever in order to make him immune to it. The protective measure must not produce the same situation of danger as is produced by the illness itself, but only something very much slighter, as is the case with vaccination against small-pox and many other similar procedures. In analytic prophylaxis against instinctual conflicts, therefore, the only methods which come into consideration are the other two which we have mentioned: the artificial production of new conflicts in the transference (conflicts which, after all, lack the character of reality), and the arousing of such conflicts in the patient’s imagination by talking to him about them and making him familiar with their possibility.

I do not know whether we can assert that the first of these two milder procedures is altogether ruled out in analysis. No experiments have been particularly made in this direction. But difficulties at once suggest themselves, which do not throw a very promising light on such an undertaking. In the first place, the choice of such situations for the transference is very limited. The patients cannot themselves bring all their conflicts into the transference; nor is the analyst able to call out all their possible instinctual conflicts from the transference situation. He may make them jealous or cause them to experience disappointments in love; but no technical purpose is required to bring this about. Such things happen of themselves in any case in most analyses. In the second place, we must not overlook the fact that all measures of this sort would oblige the analyst to behave in an unfriendly way to the patient, and this would have a damaging effect upon the affectionate attitude – upon the positive transference – which is the strongest motive for the patient’s taking a share in the joint work of analysis. Thus we should on no account expect very much from this procedure.

This therefore leaves only the one method open to us – the one which was in all probability the only one originally contemplated. We tell the patient about the possibilities of other instinctual conflicts, and we arouse his expectation that such conflicts may occur in him. What we hope is that this information and this warning will have the effect of activating in him one of the conflicts we have indicated, in a modest degree and yet sufficiently for treatment. But this time experience speaks with no uncertain voice. The expected result does not come about. The patient hears our message, but there is no response. He may think to himself: ‘This is very interesting, but I feel no trace of it.’ We have increased his knowledge, but altered nothing else in him. The situation is much the same as when people read psycho-analytic writings. The reader is ‘stimulated’ only by those passages which he feels apply to himself – that is, which concern conflicts that are active in him at the time. Everything else leaves him cold. We can have analogous experiences, I think, when we give children sexual enlightenment. I am far from maintaining that this is a harmful or unnecessary thing to do, but it is clear that the prophylactic effect of this liberal measure has been greatly over-estimated. After such enlightenment, children know something they did not know before, but they make no use of the new knowledge that has been presented to them. We come to see that they are not even in so great a hurry to sacrifice for this new knowledge the sexual theories which might be described as a natural growth and which they have constructed in harmony with, and dependence on, their imperfect libidinal organization – theories about the part played by the stork, about the nature of sexual intercourse and about the way in which babies are made. For a long time after they have been given sexual enlightenment they behave like primitive races who have had Christianity thrust upon them and who continue to worship their old idols in secret.

We started from the question of how we can shorten the inconveniently long duration of analytic treatment, and, still with this question of time in mind, we went on to consider whether it is possible to achieve a permanent cure or even to prevent future illness by prophylactic treatment. In doing so, we found that the factors which were decisive for the success of our therapeutic efforts were the influence of traumatic aetiology, the relative strength of the instincts which have to be controlled, and something which we have called an alteration of the ego. Only the second of these factors has been discussed by us in any detail, and in connection with it we have had occasion to recognize the paramount importance of the quantitative factor and to stress the claim of the metapsychological line of approach to be taken into account in any attempt at explanation.

Concerning the third factor, the alteration of the ego, we have as yet said nothing. When we turn our attention to it, the first impression we receive is that there is much to ask and much to answer here, and that what we have to say about it will prove to be very inadequate. This first impression is confirmed when we go further into the problem. As is well known, the analytic situation consists in our allying ourselves with the ego of the person under treatment, in order to subdue portions of his id which are uncontrolled – that is to say to include them in the synthesis of his ego. The fact that a co-operation of this kind habitually fails in the case of psychotics affords us a first solid footing for our judgement. The ego, if we are to be able to make such a pact with it, must be a normal one. But a normal ego of this sort is, like normality in general, an ideal fiction. The abnormal ego, which is unserviceable for our purposes, is unfortunately no fiction. Every normal person, in fact, is only normal on the average. His ego approximates to that of the psychotic in some part or other and to a greater or lesser extent; and the degree of its remoteness from one end of the series and of its proximity to the other will furnish us with a provisional measure of what we have so indefinitely termed an ‘alteration of the ego’.

If we ask what is the source of the great variety of kinds and degrees of alteration of the ego, we cannot escape the first obvious alternative, that such alterations are either congenital or acquired. Of these, the second sort will be the easier to treat. If they are acquired, it will certainly have been in the course of development, starting from the first years of life. For the ego has to try from the very outset to fulfil its task of mediating between its id and the external world in the service of the pleasure principle, and to protect the id from the dangers of the external world. If, in the course of these efforts, the ego learns to adopt a defensive attitude towards its own id as well and to treat the latter’s instinctual demands as external dangers, this happens, at any rate in part, because it understands that a satisfaction of instinct would lead to conflicts with the external world. Thereafter, under the influence of education, the ego grows accustomed to removing the scene of the fight from outside to within and to mastering the internal danger before it has become an external one; and probably it is most often right in doing so. During this fight on two fronts – later there will be a third front as well – the ego makes use of various procedures for fulfilling its task, which, to put it in general terms, is to avoid danger, anxiety and unpleasure. We call these procedures ‘mechanisms of defence’. Our knowledge of them is not yet sufficiently complete. Anna Freud’s book (1936) has given us a first insight into their multiplicity and many-sided significance.

It was from one of those mechanisms, repression, that the study of neurotic processes took its whole start. There was never any doubt that repression was not the only procedure which the ego could employ for its purposes. Nevertheless, repression is something quite peculiar and is more sharply differentiated from the other mechanisms than they are from one another. I should like to make this relation to the other mechanisms clear by an analogy, though I know that in these matters analogies never carry us very far. Let us imagine what might have happened to a book, at a time when books were not printed in editions but were written out individually. We will suppose that a book of this kind contained statements which in later times were regarded as undesirable – as, for instance, according to Robert Eisler (1929), the writings of Flavius Josephus must have contained passages about Jesus Christ which were offensive to later Christendom. At the present day, the only defensive mechanism to which the official censorship could resort would be to confiscate and destroy every copy of the whole edition. At that time, however, various methods were used for making the book innocuous. One way would be for the offending passages to be thickly crossed through so that they were illegible. In that case they could not be transcribed, and the next copyist of the book would produce a text which was unexceptionable but which had gaps in certain passages, and so might be unintelligible in them. Another way, however, if the authorities were not satisfied with this, but wanted also to conceal any indication that the text had been mutilated, would be for them to proceed to distort the text. Single words would be left out or replaced by others, and new sentences interpolated. Best of all, the whole passage would be erased and a new one which said exactly the opposite put in its place. The next transcriber could then produce a text that aroused no suspicion but which was falsified. It no longer contained what the author wanted to say; and it is highly probable that the corrections had not been made in the direction of truth.

If the analogy is not pursued too strictly, we may say that repression has the same relation to the other methods of defence as omission has to distortion of the text, and we may discover in the different forms of this falsification parallels to the variety of ways in which the ego is altered. An attempt may be made to raise the objection that the analogy goes wrong in an essential point, for the distortion of a text is the work of a tendentious censorship, no counterpart to which is to be found in the development of the ego. But this is not so; for a tendentious purpose of this kind is to a great extent represented by the compelling force of the pleasure principle. The psychical apparatus is intolerant of unpleasure; it has to fend it off at all costs, and if the perception of reality entails unpleasure, that perception – that is, the truth – must be sacrificed. Where external dangers are concerned, the individual can help himself for some time by flight and by avoiding the situation of danger, until he is strong enough later on to remove the threat by actively altering reality. But one cannot flee from oneself; flight is no help against internal dangers. And for that reason the defensive mechanisms of the ego are condemned to falsify one’s internal perception and to give one only an imperfect and distorted picture of one’s id. In its relations to the id, therefore, the ego is paralysed by its restrictions or blinded by its errors; and the result of this in the sphere of psychical events can only be compared to being out walking in a country one does not know and without having a good pair of legs.

The mechanisms of defence serve the purpose of keeping off dangers. It cannot be disputed that they are successful in this; and it is doubtful whether the ego could do without them altogether during its development. But it is also certain that they may become dangers themselves. It sometimes turns out that the ego has paid too high a price for the services they render it. The dynamic expenditure necessary for maintaining them, and the restrictions of the ego which they almost invariably entail, prove a heavy burden on the psychical economy. Moreover, these mechanisms are not relinquished after they have assisted the ego during the difficult years of its development. No one individual, of course, makes use of all the possible mechanisms of defence. Each person uses no more than a selection of them. But these become fixated in his ego. They become regular modes of reaction of his character, which are repeated throughout his life whenever a situation occurs that is similar to the original one. This turns them into infantilisms, and they share the fate of so many institutions which attempt to keep themselves in existence after the time of their usefulness has passed. ‘Vernunft wird Unsinn, Wohltat Plage’¹ as the poet complains. The adult’s ego, with its increased strength, continues to defend itself against dangers which no longer exist in reality; indeed, it finds itself compelled to seek out those situations in reality which can serve as an approximate substitute for the original danger, so as to be able to justify, in relation to them, its maintaining its habitual modes of reaction. Thus we can easily understand how the defensive mechanisms, by bringing about an ever more extensive alienation from the external world and a permanent weakening of the ego, pave the way for, and encourage, the outbreak of neurosis.

¹ [Reason becomes unreason, kindness torment.’]

At the moment, however, we are not concerned with the pathogenic role of the defensive mechanisms. What we are trying to discover is what influence the alteration of the ego which corresponds to them has upon our therapeutic efforts. The material for an answer to this question is given in the volume by Anna Freud to which I have already referred. The essential point is that the patient repeats these modes of reaction during the work of analysis as well, that he produces them before our eyes, as it were. In fact, it is only in this way that we get to know them. This does not mean that they make analysis impossible. On the contrary, they constitute half of our analytic task. The other half, the one which was first tackled by analysis in its early days, is the uncovering of what is hidden in the id. During the treatment our therapeutic work is constantly swinging backwards and forwards like a pendulum between a piece of id-analysis and a piece of ego-analysis. In the one case we want to make something from the id conscious, in the other we want to correct something in the ego. The crux of the matter is that the defensive mechanisms directed against former danger recur in the treatment as resistances against recovery. It follows from this that the ego treats recovery itself as a new danger.

The therapeutic effect depends on making conscious what is repressed, in the widest sense of the word, in the id. We prepare the way for this making conscious by interpretations and constructions, but we have interpreted only for ourselves not for the patient so long as the ego holds on to its earlier defences and does not give up its resistances. Now these resistances, although they belong to the ego, are nevertheless unconscious and in some sense separated off within the ego. The analyst recognizes them more easily than he does the hidden material in the id. One might suppose that it would be sufficient to treat them like portions of the id and, by making them conscious, bring them into connection with the rest of the ego. In this way, we should suppose, one half of the task of analysis would be accomplished; we should not reckon on meeting with a resistance against the uncovering of resistances. But what happens is this. During the work on the resistances the ego withdraws – with a greater or less degree of seriousness – from the agreement on which the analytic situation is founded. The ego ceases to support our efforts at uncovering the id; it opposes them, disobeys the fundamental rule of analysis, and allows no further derivatives of the repressed to emerge. We cannot expect the patient to have a strong conviction of the curative power of analysis. He may have brought along with him a certain amount of confidence in his analyst, which will be strengthened to an effective point by the factors of the positive transference which will be aroused in him. Under the influence of the unpleasurable impulses which he feels as a result of the fresh activation of his defensive conflicts, negative transferences may now gain the upper hand and completely annul the analytic situation. The patient now regards the analyst as no more than a stranger who is making disagreeable demands on him, and he behaves towards him exactly like a child who does not like the stranger and does not believe anything he says. If the analyst tries to explain to the patient one of the distortions made by him for the purposes of defence, and to correct it, he finds him uncomprehending and inaccessible to sound arguments. Thus we see that there is a resistance against the uncovering of resistances, and the defensive mechanisms really do deserve the name which we gave them originally, before they had been more closely examined. They are resistances not only to the making conscious of contents of the id, but also to the analysis as a whole, and thus to recovery.

The effect brought about in the ego by the defences can rightly be described as an ‘alteration of the ego’ if by that we understand a deviation from the fiction of a normal ego which would guarantee unshakable loyalty to the work of analysis. It is easy, then, to accept the fact, shown by daily experience, that the outcome of an analytic treatment depends essentially on the strength and on the depth of root of these resistances that bring about an alteration of the ego. Once again we are confronted with the importance of the quantitative factor, and once again we are reminded that analysis can only draw upon definite and limited amounts of energy which have to be measured against the hostile forces. And it seems as if victory is in fact as a rule on the side of the big battalions.

The next question we come to is whether every alteration of the ego – in our sense of the term – is acquired during the defensive struggles of the earliest years. There can be no doubt about the answer. We have no reason to dispute the existence and importance of original, innate distinguishing characteristics of the ego. This is made certain by the single fact that each person makes a selection from the possible mechanisms of defence, that he always uses a few only of them and always the same ones. This would seem to indicate that each ego is endowed from the first with individual dispositions and trends, though it is true that we cannot specify their nature or what determines them. Moreover, we know that we must not exaggerate the difference between inherited and acquired characters into an antithesis; what was acquired by our forefathers certainly forms an important part of what we inherit. When we speak of an ‘archaic heritage’ we are usually thinking only of the id and we seem to assume that at the beginning of the individual’s life no ego is as yet in existence. But we shall not overlook the fact that id and ego are originally one; nor does it imply any mystical overvaluation of heredity if we think it credible that, even before the ego has come into existence, the lines of development, trends and reactions which it will later exhibit are already laid down for it. The psychological peculiarities of families, races and nations, even in their attitude to analysis, allow of no other explanation. Indeed, more than this: analytic experience has forced on us a conviction that even particular psychical contents, such as symbolism, have no other sources than hereditary transmission, and researches in various fields of social anthropology, make it plausible to suppose that other, equally specialized precipitates left by early human development are also present in the archaic heritage.

With the recognition that the properties of the ego which we meet with in the form of resistances can equally well be determined by heredity as acquired in defensive struggles, the topographical distinction between what is ego and what is id loses much of its value for our investigation. If we advance a step further in our analytic experience, we come upon resistances of another kind, which we can no longer localize and which seem to depend on fundamental conditions in the mental apparatus. I can only give a few examples of this type of resistance; the whole field of enquiry is still bewilderingly strange and insufficiently explored. We come across people, for instance, to whom we should be inclined to attribute a special ‘adhesiveness of the libido’. The processes which the treatment sets in motion in them are so much slower than in other people because, apparently, they cannot make up their minds to detach libidinal cathexes from one object and displace them on to another, although we can discover no special reason for this cathectic loyalty. One meets with the opposite type of person, too, in whom the libido seems particularly mobile; it enters readily upon the new cathexes suggested by analysis, abandoning its former ones in exchange for them. The difference between the two types is comparable to the one felt by a sculptor, according to whether he works in hard stone or soft clay. Unfortunately, in this second type the results of analysis often turn out to be very impermanent: the new cathexes are soon given up once more, and we have an impression, not of having worked in clay, but of having written on water. In the words of the proverb: ‘Soon got, soon gone.’¹

¹ [‘Wie gewonnen, so zerronnen.’]

In another group of cases we are surprised by an attitude in our patients which can only be put down to a depletion of the plasticity, the capacity for change and further development, which we should ordinarily expect. We are, it is true, prepared to find in analysis a certain amount of psychical inertia. When the work of analysis has opened up new paths for an instinctual impulse, we almost invariably observe that the impulse does not enter upon them without marked hesitation. We have called this behaviour, perhaps not quite correctly, ‘resistance from the id’. But with the patients I here have in mind, all the mental processes, relationships and distributions of force are unchangeable, fixed and rigid. One finds the same thing in very old people, in which case it is explained as being due to what is described as force of habit or an exhaustion of receptivity – a kind of psychical entropy. But we are dealing here with people who are still young. Our theoretical knowledge does not seem adequate to give a correct explanation of such types. Probably some temporal characteristics are concerned – some alterations of a rhythm of development in psychical life which we have not yet appreciated.

In yet another group of cases the distinguishing characteristics of the ego, which are to be held responsible as sources of resistance against analytic treatment and as impediments to therapeutic success, may spring from different and deeper roots. Here we are dealing with the ultimate things which psychological research can learn about: the behaviour of the two primal instincts, their distribution, mingling and defusion – things which we cannot think of as being confined to a single province of the mental apparatus, the id, the ego or the super-ego. No stronger impression arises from the resistances during the work of analysis than of there being a force which is defending itself by every possible means against recovery and which is absolutely resolved to hold on to illness and suffering. One portion of this force has been recognized by us, undoubtedly with justice, as the sense of guilt and need for punishment, and has been localized by us in the ego’s relation to the super-ego. But this is only the portion of it which is, as it were, psychically bound by the super-ego and thus becomes recognizable; other quotas of the same force, whether bound or free, may be at work in other, unspecified places. If we take into consideration the total picture made up of the phenomena of masochism immanent in so many people, the negative therapeutic reaction and the sense of guilt found in so many neurotics, we shall no longer be able to adhere to the belief that mental events are exclusively governed by the desire for pleasure. These phenomena are unmistakable indications of the presence of a power in mental life which we call the instinct of aggression or of destruction according to its aims, and which we trace back to the original death instinct of living matter. It is not a question of an antithesis between an optimistic and a pessimistic theory of life. Only by the concurrent or mutually opposing action of the two primal instincts – Eros and the death instinct -, never by one or the other alone, can we explain the rich multiplicity of the phenomena of life.

How parts of these two classes of instincts combine to fulfil the various vital functions, under what conditions such combinations grow looser or break up, to what disturbances these changes correspond and with what feelings the perceptual scale of the pleasure principle replies to them – these are problems whose elucidation would be the most rewarding achievement of psychological research. For the moment we must bow to the superiority of the forces against which we see our efforts come to nothing. Even to exert a psychical influence on simple masochism is a severe tax upon our powers.

In studying the phenomena which testify to the activity of the destructive instinct, we are not confined to observations on pathological material. Numerous facts of normal mental life call for an explanation of this kind, and the sharper our eye grows, the more copiously they strike us. The subject is too new and too important for me to treat it as a side-issue in this discussion. I shall therefore content myself with selecting a few sample cases.

Here is one instance. It is well known that at all periods there have been, as there still are, people who can take as their sexual objects members of their own sex as well as of the opposite one, without the one trend interfering with the other. We call such people bisexuals, and we accept their existence without feeling much surprise about it. We have come to learn, however, that every human being is bisexual in this sense and that his libido is distributed, either in a manifest or a latent fashion, over objects of both sexes. But we are struck by the following point. Whereas in the first class of people the two trends have got on together without clashing, in the second and more numerous class they are in a state of irreconcilable conflict. A man’s heterosexuality will not put up with any homosexuality, and vice versa. If the former is the stronger it succeeds in keeping the latter latent and forcing it away from satisfaction in reality. On the other hand, there is no greater danger for a man’s heterosexual function than its being disturbed by his latent homosexuality. We might attempt to explain this by saying that each individual only has a certain quota of libido at his disposal, for which the two rival trends have to struggle. But it is not clear why the rivals do not always divide up the available quota of libido between them according to their relative strength, since they are able to do so in a number of cases. We are forced to the conclusion that the tendency to a conflict is something special, something which is newly added to the situation, irrespective of the quantity of libido. An independently-emerging tendency to conflict of this sort can scarcely be attributed to anything but the intervention of an element of free aggressiveness.

If we recognize the case we are discussing as an expression of the destructive or aggressive instinct, the question at once arises whether this view should not be extended to other instances of conflict, and, indeed, whether all that we know about psychical conflict should not be revised from this new angle. After all, we assume that in the course of man’s development from a primitive state to a civilized one his aggressiveness undergoes a very considerable degree of internalization or turning inwards; if so, his internal conflicts would certainly be the proper equivalent for the external struggles which have then ceased. I am well aware that the dualistic theory according to which an instinct of death or of destruction or aggression claims equal rights as a partner with Eros as manifested in the libido, has found little sympathy and has not really been accepted even among psycho-analysts. This made me all the more pleased when not long ago I came upon this theory of mine in the writings of one of the great thinkers of ancient Greece. I am very ready to give up the prestige of originality for the sake of such a confirmation, especially as I can never be certain, in view of the wide extent of my reading in early years, whether what I took for a new creation might not be an effect of cryptomnesia.

Empedocles of Acragas (Girgenti),¹ born about 495 B.C., is one of the grandest and most remarkable figures in the history of Greek civilization. The activities of his many-sided personality pursued the most varied directions. He was an investigator and a thinker, a prophet and a magician, a politician, a philanthropist and a physician with a knowledge of natural science. He was said to have freed the town of Selinunte from malaria, and his contemporaries revered him as a god. His mind seems to have united the sharpest contrasts. He was exact and sober in his physical and physiological researches, yet he did not shrink from the obscurities of mysticism, and built up cosmic speculations of astonishingly imaginative boldness. Capelle compares him with Dr. Faust ‘to whom many a secret was revealed’. Born as he was at a time when the realm of science was not yet divided into so many provinces, some of his theories must inevitably strike us as primitive. He explained the variety of things by the mixture of the four elements, earth, air, fire and water. He held that all nature was animate, and he believed in the transmigration of souls. But he also included in his theoretical body of knowledge such modern ideas as the gradual evolution of living creatures, the survival of the fittest and a recognition of the part played by chance (????) in that evolution.

But the theory of Empedocles which especially deserves our interest is one which approximates so closely to the psycho-analytic theory of the instincts that we should be tempted to maintain that the two are identical, if it were not for the difference that the Greek philosopher’s theory is a cosmic phantasy while ours is content to claim biological validity. At the same time, the fact that Empedocles ascribes to the universe the same animate nature as to individual organisms robs this difference of much of its importance.

¹ I have based what follows on a work by Wilhelm Capelle (1935).

The philosopher taught that two principles governed events in the life of the universe and in the life of the mind, and that those principles were everlastingly at war with each other. He called them ????? (love) and ?????? (strife). Of these two powers – which he conceived of as being at bottom ‘natural forces operating like instincts, and by no means intelligences with a conscious purpose’¹ – the one strives to agglomerate the primal particles of the four elements into a single unity, while the other, on the contrary, seeks to undo all those fusions and to separate the primal particles of the elements from one another. Empedocles thought of the process of the universe as a continuous, never-ceasing alternation of periods, in which the one or the other of the two fundamental forces gain the upper hand, so that at one time love and at another strife puts its purpose completely into effect and dominates the universe, after which the other, vanquished, side asserts itself and in its turn defeats its partner.

The two fundamental principles of Empedocles – ????? and ?????? – are, both in name and function, the same as our two primal instincts, Eros and destructiveness, the first of which endeavours to combine what exists into ever greater unities, while the second endeavours to dissolve those combinations and to destroy the structures to which they have given rise. We shall not be surprised, however, to find that, on its re-emergence after two and a half millennia, this theory has been altered in some of its features. Apart from the restriction to the biophysical field which is imposed on us, we no longer have as our basic substances the four elements of Empedocles; what is living has been sharply differentiated from what is inanimate, and we no longer think of the mingling and separation of particles of substance, but of the soldering together and defusion of instinctual components. Moreover, we have provided some sort of biological basis for the principle of ‘strife’ by tracing back our instinct of destruction to the death instinct, to the urge of what is living to return to an inanimate state. This is not to deny that an analogous instinct already existed earlier, nor, of course, to assert that an instinct of this sort only came into existence with the emergence of life. And no one can foresee in what guise the nucleus of truth contained in the theory of Empedocles will present itself to later understanding.

¹ Capelle (1935), 186.

In 1927, Ferenczi read an instructive paper on the problem of the termination of analyses. It ends with a comforting assurance that ‘analysis is not an endless process, but one which can be brought to a natural end with sufficient skill and patience on the analyst’s part’. The paper as a whole, however, seems to me to be in the nature of a warning not to aim at shortening analysis but at deepening it. Ferenczi makes the further important point that success depends very largely on the analyst’s having learnt sufficiently from his own ‘errors and mistakes’ and having got the better of ‘the weak points in his own personality’. This provides an important supplement to our theme. Among the factors which influence the prospects of analytic treatment and add to its difficulties in the same manner as the resistances, must be reckoned not only the nature of the patient’s ego but the individuality of the analyst.

It cannot be disputed that analysts in their own personalities have not invariably come up to the standard of psychical normality to which they wish to educate their patients. Opponents of analysis often point to this fact with scorn and use it as an argument to show the uselessness of analytic exertions. We might reject this criticism as making unjustifiable demands. Analysts are people who have learned to practise a particular art; alongside of this, they may be allowed to be human beings like anyone else. After all, nobody maintains that a physician is incapable of treating internal diseases if his own internal organs are not sound; on the contrary, it may be argued that there are certain advantages in a man who is himself threatened with tuberculosis specializing in the treatment of persons suffering from that disease. But the cases are not on all fours. So long as he is capable of practising at all, a doctor suffering from disease of the lungs or heart is not handicapped either in diagnosing or treating internal complaints; whereas the special conditions of analytic work do actually cause the analyst’s own defects to interfere with his making a correct assessment of the state of things in his patient and reacting to them in a useful way. It is therefore reasonable to expect of an analyst, as a part of his qualifications, a considerable degree of mental normality and correctness. In addition, he must possess some kind of superiority, so that in certain analytic situations he can act as a model for his patient and in others as a teacher. And finally we must not forget that the analytic relationship is based on a love of truth – that is, on a recognition of reality – and that it precludes any kind of sham or deceit.

Here let us pause for a moment to assure the analyst that he has our sincere sympathy in the very exacting demands he has to fulfil in carrying out his activities. It almost looks as if analysis were the third of those ‘impossible’ professions in which one can be sure beforehand of achieving unsatisfying results. The other two, which have been known much longer, are education and government. Obviously we cannot demand that the prospective analyst should be a perfect being before he takes up analysis, in other words that only persons of such high and rare perfection should enter the profession. But where and how is the poor wretch to acquire the ideal qualifications which he will need in his profession? The answer is, in an analysis of himself, with which his preparation for his future activity begins. For practical reasons this analysis can only be short and incomplete. Its main object is to enable his teacher to make a judgement as to whether the candidate can be accepted for further training. It has accomplished its purpose if it gives the learner a firm conviction of the existence of the unconscious, if it enables him, when repressed material emerges, to perceive in himself things which would otherwise be incredible to him, and if it shows him a first sample of the technique which has proved to be the only effective one in analytic work. This alone would not suffice for his instruction; but we reckon on the stimuli that he has received in his own analysis not ceasing when it ends and on the processes of remodelling the ego continuing spontaneously in the analysed subject and making use of all subsequent experiences in this newly-acquired sense. This does in fact happen, and in so far as it happens it makes the analysed subject qualified to be an analyst himself.

   Unfortunately something else happens as well. In trying to describe this, one can only rely on impressions. Hostility on the one side and partisanship on the other create an atmosphere which is not favourable to objective investigation. It seems that a number of analysts learn to make use of defensive mechanisms which allow them to divert the implications and demands of analysis from themselves (probably by directing them on to other people), so that they themselves remain as they are and are able to withdraw from the critical and corrective influence of analysis. Such an event may justify the words of the writer who warns us that when a man is endowed with power it is hard for him not to misuse it.¹ Sometimes, when we try to understand this, we are driven into drawing a disagreeable analogy with the effect of X-rays on people who handle them without taking special precautions. It would not be surprising if the effect of a constant preoccupation with all the repressed material which struggles for freedom in the human mind were to stir up in the analyst as well all the instinctual demands which he is otherwise able to keep under suppression. These, too, are ‘dangers of analysis’, though they threaten, not the passive but the active partner in the analytic situation; and we ought not to neglect to meet them. There can be no doubt how this is to be done. Every analyst should periodically – at intervals of five years or so – submit himself to analysis once more, without feeling ashamed of taking this step. This would mean, then, that not only the therapeutic analysis of patients but his own analysis would change from a terminable into an interminable task.

At this point, however, we must guard against a misconception. I am not intending to assert that analysis is altogether an endless business. Whatever one’s theoretical attitude to the question may be, the termination of an analysis is, I think a practical matter. Every experienced analyst will be able to recall a number of cases in which he has bidden his patient a permanent farewell rebus bene gestis². In cases of what is known as character-analysis there is a far smaller discrepancy between theory and practice. Here it is not easy to foresee a natural end, even if one avoids any exaggerated expectations and sets the analysis no excessive tasks. Our aim will not be to rub off every peculiarity of human character for the sake of a schematic ‘normality’, nor yet to demand that the person who has been ‘thoroughly analysed’ shall feel no passions and develop no internal conflicts. The business of the analysis is to secure the best possible psychological conditions for the functions of the ego; with that it has discharged its task.

¹ Anatole France, La révolte des anges.

² [‘Things having gone well.’]

Both in therapeutic and in character-analyses we notice that two themes come into especial prominence and give the analyst an unusual amount of trouble. It soon becomes evident that a general principle is at work here. The two themes are tied to the distinction between the sexes; one is as characteristic of males as the other is of females. In spite of the dissimilarity of their content, there is an obvious correspondence between them. Something which both sexes have in common has been forced, by the difference between the sexes, into different forms of expression.

The two corresponding themes are in the female, an envy for the penis – a positive striving to possess a male genital – and, in the male, a struggle against his passive or feminine attitude to another male. What is common to the two themes was singled out at an early date by psycho-analytic nomenclature as an attitude towards the castration complex. Subsequently Alfred Adler brought the term ‘masculine protest’ into current use. It fits the case of males perfectly; but I think that, from the start, ‘repudiation of femininity’ would have been the correct description of this remarkable feature in the psychical life of human beings.

In trying to introduce this factor into the structure of our theory, we must not overlook the fact that it cannot, by its very nature, occupy the same position in both sexes. In males the striving to be masculine is completely ego-syntonic from the first; the passive attitude, since it presupposes an acceptance of castration, is energetically repressed, and often its presence is only indicated by excessive overcompensations. In females, too, the striving to be masculine is ego-syntonic at a certain period – namely in the phallic phase, before the development to femininity has set in. But it then succumbs to the momentous process of repression whose outcome, as has so often been shown, determines the fortunes of a woman’s femininity. A great deal depends on whether a sufficient amount of her masculinity complex escapes repression and exercises a permanent influence on her character. Normally, large portions of the complex are transformed and contribute to the construction of her femininity: the appeased wish for a penis is destined to be converted into a wish for a baby and for a husband, who possesses a penis. It is strange, however, how often we find that the wish for masculinity has been retained in the unconscious and, from out of its state of repression, exercises a disturbing influence.

As will be seen from what I have said, in both cases it is the attitude proper to the opposite sex which has succumbed to repression. I have already stated elsewhere¹ that it was Wilhelm Fliess who called my attention to this point. Fliess was inclined to regard the antithesis between the sexes as the true cause and primal motive force of repression. I am only repeating what I said then in disagreeing with his view, when I decline to sexualize repression in this way – that is, to explain it on biological grounds instead of on purely psychological ones.

¹ ‘”A Child is being Beaten”‘ (1919e).

The paramount importance of these two themes – in females the wish for a penis and in males the struggle against passivity – did not escape Ferenczi’s notice. In the paper read by him in 1927 he made it a requirement that in every successful analysis those two complexes must have been mastered.¹ I should like to add that, speaking from my own experience, I think that in this Ferenczi was asking a very great deal. At no other point in one’s analytic work does one suffer more from an oppressive feeling that all one’s repeated efforts have been in vain, and from a suspicion that one has been ‘preaching to the winds’, than when one is trying to persuade a woman to abandon her wish for a penis on the ground of its being unrealizable or when one is seeking to convince a man that a passive attitude to men does not always signify castration and that it is indispensable in many relationships in life. The rebellious overcompensation of the male produces one of the strongest transference-resistances. He refuses to subject himself to a father-substitute, or to feel indebted to him for anything, and consequently he refuses to accept his recovery from the doctor. No analogous transference can arise from the female’s wish for a penis, but it is the source of outbreaks of severe depression in her, owing to an internal conviction that the analysis will be of no use and that nothing can be done to help her. And we can only agree that she is right, when we learn that her strongest motive in coming for treatment was the hope that, after all, she might still obtain a male organ, the lack of which was so painful to her.

¹ ‘. . . Every male patient must attain a feeling of equality in relation to the physician as a sign that he has overcome his fear of castration; every female patient, if her neurosis is to be regarded as fully disposed of, must have got rid of her masculinity complex and must emotionally accept without a trace of resentment the implications of her female role.’ (Ferenczi, 1928, 8.)

But we also learn from this that it is not important in what form the resistance appears, whether as a transference or not. The decisive thing remains that the resistance prevents any change from taking place – that everything stays as it was. We often have the impression that with the wish for a penis and the masculine protest we have penetrated through all the psychological strata and have reached bedrock, and that thus our activities are at an end. This is probably true, since, for the psychical field, the biological field does in fact play the part of the underlying bedrock. The repudiation of femininity can be nothing else than a biological fact, a part of the great riddle of sex.¹ It would be hard to say whether and when we have succeeded in mastering this factor in an analytic treatment. We can only console ourselves with the certainty that we have given the person analysed every possible encouragement to re-examine and alter his attitude to it.

¹ We must not be misled by the term ‘masculine protest’ into supposing that what the man is repudiating is his passive attitude – what might be called the social aspect of femininity. Such a view is contradicted by an observation that is easily verifiable – namely that such men often display a masochistic attitude – a state that amounts to bondage – towards women. What they reject is not passivity in general, but passivity towards a male. In other words, the ‘masculine protest’ is in fact nothing else than castration anxiety.