Let us attempt the easiest answers first. We are obliged to attribute the differentiation of ego and id not merely to primitive man, but to organisms that are far simpler still, since it is a necessary manifestation of the influence exerted by the external world. As for the super-ego, we described it as having its very origins in the experiences that led to totemism; the question whether it was the ego or the id that underwent those experiences and acquired moral attributes quickly proves to be pointless. The next consideration that presents itself to our mind is that the id cannot experience or undergo any external pattern of events except via the ego, the sole representative of the external world that it possesses. None the less, we cannot in fact claim that there is a hereditary process operating directly within the ego. What we encounter here is the yawning gulf between actual individuals, and our notion of the species. Furthermore, we must not view the difference between the ego and the id in unduly rigid terms; we must not forget that the ego is part of the id, albeit differentiated from it in a special way.61 The ego's experiences seem to be lost to heredity to begin with; however, if they recur often and strongly enough in numerous successive generations of individuals, they transform themselves so to speak into id experiences, and their impact is then preserved through heredity. The heritable id accordingly harbours within it remnants of countless numbers of previous egos, and when an individual ego evolves its super-ego from the id it is perhaps merely bringing older ego forms back to light, and back to life.
The manner in which the super-ego comes into being makes it readily comprehensible that early conflicts of the ego with the object-cathexes of the id can be continued later on in conflicts with their successor, the super-ego. If the ego has botched the task of overcoming the Oedipus complex, then its62 energy-cathexis, which derives from the id, will reassert itself in the reaction-formation of the ego-ideal. The abundant communication between the ego-ideal and these Ucs drive-impulses serves to explain the puzzling fact that the ideal itself can remain largely unconscious, and inaccessible to the ego. The battle that had previously raged in the nether depths, but had never come to any final resolution through a rapid process of sublimation and identification, is now carried on at a higher level, rather as in Kaulbach's painting of the Battle of Châlons.63
IV
The Two Types of Drives
We have already made the point that our proposed division of the psyche into an id, an ego and a super-ego can only signify a real advance in our knowledge if it also proves to be the means to a deeper understanding and more accurate description of the dynamic relations at work in the life of the psyche. We have also come to appreciate that the ego is particularly subject to the influence of perception, and that in broad terms one can say that perceptions have the same significance for the ego that drives have for the id. At the same time, however, the ego is also susceptible to the influence of drives, just like the id – of which, of course, it is but part, albeit a specially modified one.
On the subject of drives, I have recently (in Beyond the Pleasure Principle) elaborated a view that I shall first recapitulate, and then use as the basis for the next stages of the argument. On this view, we need to distinguish two types of drives, one of which – the sexual drives, or Eros – is far more conspicuous, and far more accessible to our knowledge and understanding. It includes not only the uninhibited sexual drive itself and the goal-inhibited and hence sublimated drive-impulses deriving from it, but also the self-preservation drive that we perforce ascribe to the ego, and that at the very outset of our psychoanalytical work we had good reason to regard as contrasting sharply with the sexual object-drives. Demonstrating this second type of drive caused us considerable difficulty; our solution in the end was to regard sadism as representative of it. On the basis of theoretical considerations underpinned by biology, we posited a death drive charged with the task of causing animate organisms to revert to an inanimate state, whereas Eros pursues the goal of maximizing the complexity of life – and thereby of course preserving it – by an ever more catholic combination of the particles into which living matter had been fragmented. In pursuing their respective goals both drives behave in a strictly conservative manner, in that they seek the restoration of a state that was disrupted by the emergence of life. According to this view, the emergence of life is therefore the cause both of the urge to carry on living and, simultaneously, of the urge for death, while life itself is a battle and constant compromise between these two urges. Considered thus, the question as to the origin of life remains a cosmological one, while the question as to the purpose and intention of life is answered in dualistic terms.64
A particular physiological process is attributable – so the argument goes – to each of the two types of drive (anabolism and catabolism65); both drives are active in each and every piece of living substance, albeit in varying proportions, with the result that any such substance is capable of taking on the role of Eros.
Precisely how drives of the two types connect, combine and blend with each other remains entirely unimaginable – but that such a thing happens, routinely and on a very large scale, is a postulate crucial to our whole framework of ideas. We can hypothesize that as a consequence of the fusion of unicellular elementary organisms into multicellular organisms, the death drive in the individual cell was successfully neutralized, and its destructive impulses diverted to the external world through the mediation of a special organ, to wit the musculature; the death drive accordingly now finds expression – though in all probability only in part – as a destruction drive directed against the external world and other organisms.
Once we have accepted the notion of a merging of the two types of drives, we are then also confronted by the possibility of a - more or less complete – de-mergence of them. I would suggest that in the sadistic component of the sexual drive we see a classic example of a purposive merging of drives, while in sadism qua autonomous perversion we see an exemplary instance of de-mergence, albeit not one where the process has been taken to extremes.66 This in its turn affords us fresh insight into a large mass of facts that have not previously been considered in this light. We can readily see that, in order to effect release, the destruction drive is routinely put at the service of Eros; we suspect that epileptic fits are produced by, and indicative of, a de-mergence of drives; and we are beginning to realize that amongst the effects achieved by certain serious neuroses, e.g. the obsessional neuroses, the de-mergence of drives and the appearance of the death drive deserve special consideration. By way of a rapid generalization: we are inclined to think that libido regressions, for instance regression from the genital to the sadistic-anal phase, are rooted essentially in a de-mergence of drives, and that, inversely, progression from the early to the definitive genital phase is dependent on an accession of erotic components. The question also arises whether ordinary ambivalence – which we so often find to be particularly marked in those who are constitutionally disposed to neurosis – should not be regarded as the result of a de-mergence; this latter process is so primal, however, that it must rate instead as a merging of drives that remained incomplete.
Our interest quite naturally turns to two particular questions: first, whether we shall not perhaps discover revealing connections between the structures we have postulated – the ego, super-ego and id – on the one hand, and the two types of drives on the other; second, whether we shall be able to show that the pleasure principle, the mechanism that controls psychic processes, stands in a firm and clear relationship to the two types of drives, and to the forms into which the psyche has differentiated. Before we enter upon this discussion, however, we need to deal with a doubt that challenges the very formulation of the question itself. While there can be no doubt about the pleasure principle, and whilst our division of the ego is soundly based on clinical evidence, yet our grounds for distinguishing between the two types of drives seem not altogether strong enough, and it seems quite possible that facts evinced by clinical analysis might rob them of all credibility.
There does appear to be just such a fact. We might reasonably substitute the polarity of love and hate for the antithesis constituted by the two types of drives. Whereas of course we have no problem showing how Eros is represented, it comes as quite a relief that we are now able to identify the destruction drive – which takes its lead from hate – as representing the highly elusive death drive. Clinical observation, however, clearly shows us not only that hate is an unexpectedly regular accompaniment of love (ambivalence), and is very often its precursor in human relationships, but also that in certain circumstances hate changes into love, and love into hate. If this transformation involves anything more than just temporal succession, that is, one thing simply taking the place of the other, then clearly we are left with no basis for making such a fundamental distinction as that between erotic drives and death drives – a distinction premised on the notion of physiological processes that run directly counter to one another.
Now cases where we first love someone and then hate them (or vice versa) because they themselves have occasioned the change, clearly have no bearing on our problem; nor do those cases where love that has not yet become manifest reveals itself first through hostility and a tendency to aggression, for here the destructive component may simply have run on ahead during the process of object-cathexis, before being joined in due course by the erotic component. But a number of cases are known to us from the psychology of neuroses in which there are much stronger grounds for supposing that a transformation does indeed take place. In paranoia persecutoria the patient resists an excessive homosexual attachment to a particular individual in such a way that this most deeply loved individual turns into a persecutor against whom the often dangerous aggression of the patient is directed. We can legitimately interject that a previous phase had served to convert the love into hate. Regarding the genesis of homosexuality, indeed also of desexualized social feelings, psychoanalytical study has only recently revealed to us the existence of intense feelings of rivalry leading to aggressive tendencies, feelings that have to be overcome before the hated object can become the loved object, or become the object of an identification.67 The question arises whether we can assume that in these cases the hate is converted directly into love. After all, it is a matter here of purely internal changes, precipitated in no way by any change in behaviour on the part of the object.
However, another possible mechanism is familiar to us from our psychoanalytical study of the change that occurs in paranoia. Here, an ambivalent attitude is present from the outset, and the transformation is brought about through a reactive displacement of cathexis, whereby energy is withdrawn from the erotic impulse, and added to the hostile one.
Something very similar, albeit not quite the same, happens in the process that leads to homosexuality, namely the overcoming of hostile feelings of rivalry.68 Hostility is an attitude with no prospect of gratification, and in consequence – for economic reasons, in other words – it is replaced by the attitude of love, which offers better prospects of gratification, that is, the possibility of release. In neither of these cases, therefore, do we need to assume a direct transformation of hate into love, which would be incompatible with the notion of a qualitative difference between the two types of drives.
It has not escaped our notice, however, that in drawing on this other mechanism whereby love changes into hate, we have tacitly made a further assumption – one that deserves to be made fully explicit. We have based our argument on the supposition that there exists within the psyche – whether in the ego or the id is still uncertain – a displaceable energy which, though indifferent in itself, can join forces with a qualitatively differentiated erotic or destructive impulse and increase its overall cathexis. We simply cannot get anywhere without positing a displaceable energy of this kind. But we are still left wondering where it comes from, who69 it belongs to, and what it signifies.
The problem of the quality of drive-impulses, and how that quality is maintained throughout the sundry vicissitudes that drives are prone to, remains decidedly obscure, and to date barely any attempt has been made to tackle it. In the case of the sexual partial drives, which lend themselves particularly well to observation, one can see a number of processes that follow a similar pattern. Thus, for instance, the partial drives to some extent communicate with each other; a drive originating from one erogenous source is capable of surrendering its intensity in order to reinforce a partial drive originating from another; the gratification of one drive can serve another in place of the latter's own. Further similarities could be cited – all of which inevitably encourages us to venture certain kinds of hypotheses.
In this present discussion, too, I can offer not proof but only a hypothesis. It seems plausible to suppose that this displaceable and indifferent energy, active very probably in both the ego and the id, derives from the store of narcissistic libido, and is thus desexualized Eros; indeed, the erotic drives in general seem to us to be more plastic, more displaceable, more displaceable than the destruction drives. That being so, we can quite logically go on to suggest that this displaceable libido operates on behalf of the pleasure principle, by preventing any undue build-up and facilitating release.70 In so doing it clearly displays considerable indifference as to which particular pathway is adopted by the release process, provided that the actual process itself takes place. We know this trait to be typical of the cathexis processes in the id. It is evident in erotic cathexes, where a marked indifference is displayed with regard to the object; and it is very marked indeed in the transferences that occur in analysis – transferences that have to be effected, regardless of who happens to be their object. Rank has recently produced some splendid examples demonstrating that neurotic acts of revenge tend to be directed against the wrong people. This type of behaviour on the part of the unconscious inevitably reminds us of that comic little story of the three village tailors, one of whom is due to be strung up because the village's sole blacksmith has done a dastardly deed that calls for a hanging.71 Someone has to be punished, even if it's not the guilty party. This same disregard first came to our attention in the displacements characteristic of the primary process in dream-work.72 Whereas in that instance it was the objects that were apparently deemed to be of only secondary importance, in this present context it is the pathways utilized by the release process. If the ego were involved, we would expect to find an insistence on greater precision in the choice of both object and pathway.
If this displaceable energy is desexualized libido, then it may also be termed sublimated, for it would still be firmly adhering to Eros's central objective of being a unifying and binding force, by serving to bring about that unity which – or at least the striving for which – is the ego's most distinctive feature. If we include thought processes in the broader sense amongst these displacements, then of course thinking, too, may be seen to be covered by the sublimation of erotic energy.
This brings us back to a possibility that we touched on earlier, namely that sublimation routinely takes place via the medium of the ego.73 Another circumstance that we might recall here is that this same ego deals with the initial object-cathexes of the id – and no doubt later ones as well – by taking their libido into itself and annexing it to the ego-alteration brought about by identification. Needless to say, this conversion [of object-libido] into ego-libido entails a desexualization, an abandonment of sexual goals. At all events this affords us clear insight into an important function of the ego in its relationship to Eros. By thus commandeering the libido of the various object-cathexes, setting itself up as sole love-object, and desexualizing or sublimating the libido of the id, it operates directly counter to the designs of Eros; it puts itself at the service of the opposing drive-impulses. In respect of certain other object-cathexes pertaining to the id, it simply has to put up with them – to tag along, so to speak. We shall return later to a further possible consequence of this activity on the part of the ego.74
At this point we probably need to make an importan
t addition to the narcissism theory. At the very beginning the entire libido is massed in the id, during the period when the ego is in the process of formation, or formed but still weak. The id sends forth part of this libido for the purpose of erotic object-cathexes, whereupon the ego, having meanwhile gained in strength, seeks to commandeer this object-libido and force itself on the id as a love-object. The ego's form of narcissism is thus a secondary one – one that has been withdrawn from objects.75
Again and again, we find that the drive-impulses that we are capable of monitoring turn out to derive from Eros. If it were not for the arguments set forth in Beyond the Pleasure Principle, and ultimately also the sadistic admixtures encountered in Eros, we would have difficulty in holding firm to our fundamental dualist position. But since we have no alternative, we are driven to the supposition that the death drives very largely remain silent, and that the clamour of life comes mostly from Eros.76
And also from the battle against Eros! There can be no denying the notion that the pleasure principle serves the id as a compass in its battle against the libido, which habitually disrupts the smooth process of life. If the constancy principle in Fechner's sense77 does indeed govern life, which on that view is supposed to be a steady slide into death, then it is the demands made by Eros, that is by the sexual drives, which – manifesting themselves as the needs that drives give rise to – interrupt the downward slide and create new tensions. Guided by the pleasure principle or, to be precise, by the awareness of unpleasure, the id defends itself against them by a variety of means. It does so in the first instance by meeting the demands of the non-desexualized libido as rapidly as possible, that is to say by striving to give gratification to the directly sexual urges. But it does so on a far larger scale by using one particular form of such gratification in which all the constituent demands coincide, in order to rid itself of those sexual substances that are the supersaturated vehicle, so to speak, of the erotic tensions. The shedding of the sexual substances in the sex act corresponds in a sense to the separating-out of soma and germ-plasm.78 This explains why the state that ensues upon full sexual gratification is similar to dying, while in certain lower animals death coincides with the act of procreation. Reproduction is the cause of these creatures' death in the sense that the death drive can effect its aims without let or hindrance once Eros has been removed from the picture through the act of gratification. And finally, as we have seen, the ego makes it easier for the id to assert control by sublimating parts of the libido for itself and its own purposes.