The “Uncanny”

by Sigmund Freud

part three




In the course of this discussion the reader will have felt certain doubts arising in his mind; and he must now have an opportunity of collecting them and bringing them forward.


It may be true that the uncanny [unheimlich] is something which is secretly familiar [heimlich-heimisch], which has undergone repression and then returned from it, and that everything that is uncanny fulfils this condition. But the selection of material on this basis does not enable us to solve the problem of the uncanny. For our proposition is clearly not convertible. Not everything that fulfils this condition — not everything that recalls repressed desires and surmounted modes of thinking belonging to the prehistory of the individual and of the race — is on that account uncanny.


Nor shall we conceal the fact that for almost every example adduced in support of our hypothesis one may be found which rebuts it. The story of the severed hand in Hauff’s fairy tale [p. 244] certainly has an uncanny effect, and we have traced that effect back to the castration complex; but most readers will probably agree with me in judging that no trace of uncanniness is provoked by Herodotus’s story of the treasure of Phampsinitus, in which the master-thief, whom the princess tries to hold fast by the hand, leaves his brother’s severed hand behind with her instead. Again, the prompt fulfillment of the wishes of Polycrates [p. 239] undoubtedly affects us in the same uncanny way as it did the king of Egypt; yet our own fairy stories are crammed with instantaneous wish-fulfillments which produce no uncanny effect whatever. In the story of ‘The Three Wishes,’ the woman is tempted by the savoury smell of a sausage to wish that she might have one too, and in an instant it lies on a plate before her. In his annoyance at her hastiness her husband wishes it may hang on her nose. And there it is, dangling from her nose. All this is very striking but not in the least uncanny. Fairy tales quite frankly adopt the animistic standpoint of the omnipotence of thoughts and wishes, and yet I cannot think of any genuine fairy story which has anything uncanny about it. We have heard that it is in the highest degree uncanny when an inanimate object — a picture or a doll — comes to life; nevertheless in Hans Andersen’s stories the household utensils, furniture and tin soldiers are alive, yet nothing could well be more remote from the uncanny. And we should hardly call it uncanny when Pygmalion’s beautiful statue comes to life.


Apparent death and the re-animation of the dead have been represented as most uncanny themes. But things of this sort too are very common in fairy stories. Who would be so bold as to call it uncanny, for instance, when Snow-White opens her eyes once more? And the resuscitation of the dead in accounts of miracles, as in the New Testament, elicits feelings quite unrelated to the uncanny. Then, too, the theme that achieves such an indubitably uncanny effect, the unintended recurrence of the same thing, serves other and quite different purposes in another class of cases. We have already come across one example [p 237] in which it is employed to call up a feeling of the comic; and we could multiply instances of this kind. Or again, it works as a means of emphasis, and so on. And once more: what is the origin of the uncanny effect of silence, darkness and solitude?


Do not these factors point to the part played by danger in the genesis of what is uncanny, notwithstanding that in children these same factors are the most frequent determinants of the expression of fear [rather than of the uncanny]? And are we after all justified in entirely ignoring intellectual uncertainty as a factor, seeing that we have admitted its importance in relation to death [p. 242]?


It is evident therefore, that we must be prepared to admit that there are other elements besides those which we have so far laid down as determining the production of uncanny feelings. We might say that these preliminary results have satisfied psycho-analytic interest in the problem of the uncanny, and that what remains probably calls for an aesthetic enquiry. But that would be to open the door to doubts about what exactly is the value of our general contention that the uncanny proceeds from something familiar which has been repressed.


We have noticed one point which may help us to resolve these uncertainties: nearly all the instances that contradict our hypothesis are taken from the realm of fiction, of imaginative writing. This suggests that we should differentiate between the uncanny that we actually experience and the uncanny that we merely picture or read about.


What is experienced as uncanny is much more simply conditioned but comprises far fewer instances. We shall find, I think, that it fits in perfectly with our attempt at a solution, and can be traced back without exception to something familiar that has been repressed. But here, too, we must make a certain important and psychologically significant differentiation in our material, which is best illustrated by turning to suitable examples.


Let us take the uncanny associated with the omnipotence of thoughts, with the prompt fulfillment of wishes, with secret injurious powers and with the return of the dead. The condition under which the feeling of uncanniness arises here is unmistakable. We — or our primitive forefathers — once believed that these possibilities were realities, and were convinced that they actually happened. Nowadays we no longer believe in them, we have surmounted these modes of thought; but we do not feel quite sure of our new beliefs, and the old ones still exist within us ready to seize upon any confirmation. As soon as something actually happens in our lives which seems to confirm the old, discarded beliefs we get a feeling of the uncanny; it is as though we were making a judgment something like this: ‘So, after all, it is true that one can kill a person by the mere wish!’ or, ‘So the dead do live on and appear on the scene of their former activities!’ and so on. Conversely, anyone who has completely and finally rid himself of animistic beliefs will be insensible to this type of the uncanny. The most remarkable coincidences of wish and fulfillment, the most mysterious repetition of similar experiences in a particular place or on a particular date, the most deceptive sights and suspicious noises — none of these things will disconcert him or raise the kind of fear which can be described as ‘a fear of something uncanny.’ The whole thing is purely an affair of ‘reality-testing,’ a question of the material reality of the phenomena.**


The state of affairs is different when the uncanny proceeds from repressed infantile complexes, from the castration complex, womb-phantasies, etc.’ but experiences which arouse this kind of uncanny feeling are not of very frequent occurrence in real life. The uncanny which proceeds from actual experience belongs for the most part to the first group [the group dealt with in the previous paragraph]. Nevertheless the distinction between the two is theoretically very important. Where the uncanny comes from infantile complexes the question of material reality does not arise; its place is taken by psychical reality. What is involved is an actual repression of some content of thought and a return of this repressed content, not a cessation of belief in the reality of such a content. We might say that in the one case what had been repressed is a particular ideational content, and in the other the belief in its (material) reality. But this last phrase no doubt extends the term ‘repression’ beyond its legitimate meaning. It would be more correct to take into account a psychological distinction which can be detected here, and to say that the animistic beliefs of civilized people are in a state of having been (to a greater or lesser extent) surmounted [rather than repressed]. Our conclusion could then be stated thus: an uncanny experience occurs either when infantile complexes which have been repressed are once more revived by some impression, or when primitive beliefs which have been surmounted seem once more to be confirmed. Finally, we must not let our predilection for smooth solutions and lucid exposition blind us to the fact that these two classes of uncanny experience are not always sharply distinguishable. When we consider that primitive beliefs are most intimately connected with infantile complexes, and are, in fact, based on them, we shall not be greatly astonished to find that the distinction is often a hazy one.


The uncanny as it is depicted in literature, in stories and imaginative productions, merits in truth a separate discussion. Above all, it is a much more fertile province than the uncanny in real life, for it contains the whole of the latter and something more besides, something that cannot be found in real life. The contrast between what has been repressed and what has been surmounted cannot be transposed on to the uncanny in fiction without profound modification; for the realm of phantasy depends for its effect on the fact that its content is not submitted to reality-testing. The somewhat paradoxical result is that in the first place a great deal that is not uncanny in fiction would be so if it happened in real life; and in the second place that there are many more means of creating uncanny effects in fiction than there are in real life.


The imaginative writer has this license among many others, that he can select his world of representation so that it either coincides with the realities we are familiar with or departs from them in what particulars he pleases. We accept his ruling in every case. In fairy tales, for instance, the world of reality is left behind from the very start, and the animistic system of beliefs is frankly adopted. Wish-fulfillments, secret powers, omnipotence of thoughts, animation of inanimate objects, all the elements so common in fairy stories, can exert no uncanny influence here; for, as we have learnt, that feeling cannot arise unless there is a conflict of judgment as to whether things which have been ‘surmounted’ and are regarded as incredible may not, after all, be possible; and this problem is eliminated from the outset by the postulates of the world of fairy tales. Thus we see that fairy stories, which have furnished us with most of the contradictions to our hypothesis of the uncanny, confirm the first part of our proposition — that in the realm of fiction many things are not uncanny which would be so if they happened in real life. In the case of these stories there are other contributory factors, which we shall briefly touch upon later.


The creative writer can also choose a setting which though less imaginary than the world of fairy tales, does yet differ from the real world by admitting superior spiritual beings such as daemonic spirits or ghosts of the dead. So long as they remain within their setting of poetic reality, such figures lose any uncanniness which they might possess. The souls in Dante’s Inferno, or the supernatural apparitions in Shakespeare’s Hamlet, Macbeth or Julius Caesar, may be gloomy and terrible enough, but they are no more really uncanny than Homer’s jovial world of gods. We adapt our judgment to the imaginary reality imposed on us by the writer, and regard souls, spirits and ghosts as though their existence had the same validity as our own has in material reality. In this case too we avoid all trace of the uncanny.


The situation is altered as soon as the writer pretends to move in the world of common reality. In this case he accepts as well all the conditions operating to produce uncanny feelings in real life; and everything that would have an uncanny effect in reality has it in his story. But in this case he can even increase his effect and multiply it far beyond what could happen in reality, by bringing about events which never or very rarely happen in fact. In doing this he is in a sense betraying us to the superstitiousness which we have ostensibly surmounted; he deceives us by promising to give us the sober truth, and then after all overstepping it. We react to his inventions as we would have reacted to real experiences; by the time we have seen through his trick it is already too late and the author has achieved his object. But it must be added that his success is not unalloyed. We retain a feeling of dissatisfaction, a kind of grudge against the attempted deceit. I have noticed this particularly after reading Schnitzler’s Die Weissagung [The Prophecy] and similar stories which flirt with the supernatural. However, the writer has one more means which he can use in order to avoid our recalcitrance and at the same time to improve his chances of success. He can keep us in the dark for a long time about the precise nature of the presuppositions on which the world he writes about is based, or he can cunningly and ingeniously avoid any definite information on the point to the last. Speaking generally, however, we find a confirmation of the second part of our proposition — that fiction presents more opportunities for creating uncanny feelings than are possible in real life.


Strictly speaking, all these complications relate only to that class of the uncanny which proceeds from forms of thought that have been surmounted. The class which proceeds from repressed complexes is more resistant and remains as powerful in fiction as in real experience, subject to one exception [see p. 252]. The uncanny belonging to the first class — that proceeding from forms of thought that have been surmounted — retains its character not only in experience but in fiction as well, so long as the setting is one of material reality; but where it is given an arbitrary and artificial setting in fiction, it is apt to lose that character.


We have clearly not exhausted the possibilities of poetic license and the privileges enjoyed by story-writers in evoking or in excluding an uncanny feeling. In the main we adopt an unvarying passive attitude towards real experience and are subject to the influence of our physical environment. But the story-teller has a peculiarly directive power over us; by means of the moods he can put us into, he is able to guide the current of our emotions, to dam it up in one direction and make it flow in another, and he often obtains a great variety of effects from the same material. All this is nothing new, and has doubtless long since been fully taken into account by students of aesthetics. We have drifted into this field of research half involuntarily, through the temptation to explain certain instances which contradicted our theory of the causes of the uncanny. Accordingly we will now return to the examination of a few of those instances.


We have already asked [p. 246] why it is that the severed hand in the story of the treasure of Rhampsinitus has no uncanny effect in the way that the severed hand has in Hauff’s story. The question seems to have gained in importance now that we have recognized that the class of the uncanny which proceeds from repressed complexes is the more resistant of the two. The answer is easy. In the Herodotus story our thoughts are concentrated much more on the superior cunning of the master-thief than on the feelings of the princess. The princess may very well have had an uncanny feeling, indeed she very probably fell into a swoon; but we have no such sensations, for we put ourselves in the thief’s place, not in hers. In Nestroy’s farce, Der Zerrissene [The Torn Man], another means is used to avoid any impression of the uncanny in the scene in which the fleeing man, convinced that he is a murderer, lifts up one trap-door after another and each time sees what he takes to be the ghost of his victim rising up out of it. He calls out in despair, ‘But I’ve only killed one man. Why this ghastly multiplication?’ We know what went before this scene and do not share his error, so what must be uncanny to him has an irresistibly comic effect on us. Even a ‘real’ ghost, as in Oscar Wilde’s Canterville Ghost, loses all power of at least arousing gruesome feelings in us as soon as the author begins to amuse himself by being ironical about it and allows liberties to be taken with it. Thus we see how independent emotional effects can be of the actual subject-matter in the world of fiction. In fairy stories feelings of fear — including therefore uncanny feelings — are ruled out altogether. We understand this, and that is why we ignore any opportunities we find in them for developing such feelings.


Concerning the factors of silence, solitude and darkness [pp. 246-7], we can only say that they are actually elements in the production of the infantile anxiety from which the majority of human beings have never become quite free. This problem has been discussed from a psycho-analytic point of view elsewhere.



** Since the uncanny effect of the ‘double also belongs to this same group it is interesting to observe what the effect is of meeting one’s own image unbidden and unexpected. Ernest Mach has related two such observations in his Analyse der Empfindungen (1900, 3). On the first occasion he was not a little startled when he realized that the face before him was his own. The second time he formed a very unfavourable opinion about the supposed stranger who entered the omnibus, and thought ‘What a shabby-looking school-master that man is who is getting in!’ — I can report a similar adventure. I was sitting alone in my wagon-lit compartment when a more than usually violent jolt of the train swung back the door of the adjoining washing-cabinet, and an elderly gentleman in a dressing-gown and a travelling cap came in. I assumed that in leaving the washing-cabinet, which lay between the two compartments, he had taken the wrong direction and come into my compartment by mistake. Jumping up with the intention of putting him right, I at once realized to my dismay that the intruder was nothing but my own reflection in the looking-glass on the open door. I can still recollect that I thoroughly disliked his appearance. Instead, therefore, of being frightened by our ‘doubles’, both Mach and I simply failed to recognize them as such. Is it not possible, though, that our dislike of them was a vestigial trace of the archaic reaction which feels the ‘double’ to something uncanny?




Return to Part I


Return to Part II