One can find in Freud two kinds of theory of the psyche. One concerns the structure and function of its sub-systems. The other is about the nature and management of its contents. Freud's theory of content--the theory of phantasy, of character, of introjects and identifications, of dreams, of affects, and of primary and secondary processes--is entirely original. But his theory of structure and function is in many ways Kantian. There are parallels between Kant's Understanding and Freud's Ego, Kant's moral Reason and Freud's Super-Ego, and, via Schopenhauer and the will, between Kant's inner sense and Freud's Id. Freud's theory of the relation of consciousness and language and therefore his demarcation of the unconscious is also Kantian. So to is his view of the ego and drives as systems and of what in them cannot become conscious. Further, how he saw the unconscious (its timelessness, etc.) resonates with vestiges of Kantian noumena, as he acknowledged. In general, Freud did not acknowledge Kant's model of the mind as an antecedent. That is not altogether surprising. That Kantianism was broadly correct was simply taken for granted among German-speaking intellectuals at the time. Freud may never have had any thought that the facts of psychic life could be conceptualized in any other way.

The theme of the 35th International Congress was where Freud left psychoanalysis at the end of his work and how his last views look now, fifty years later.(1) In such a search for a long view, the schemes of thought which animated Freud's work are of central importance, particularly those that continued to shape psychoanalysis after his death. One such influence was Kant's philosophy of mind, especially his model of the psyche. Even though it not only animated Freud but continued to shape ego psychology in particular long after his death, this Kantian strain in psychoanalytic thought has never received much attention.

1. The main influences on Freud usually identified are German naturphilosophie and Helmholtzian mechanistic materialism.(2) Geisteswissenshaft (the intentional, teleological study of culture) is mentioned and so is Darwin. The role of Machian positivism is sometimes noted. But seldom is the name of Immanuel Kant mentioned.

Yet Kant's models and methods dominated the German-speaking world in the nineteenth century.(3) Moreover, Freud referred to Kant very frequently, more frequently in fact than to any other philosopher.(4) In the form of the Neo-Kantian movement, Kant's views gained their maximum prominence just at the time when Freud was thinking through his own model of the psyche (in (1895) and Chapter VII of (1900)), the model that would guide him for the rest of his life. Herbart, Helmholtz, Meynert and even Lipps were or considered themselves to be Kantians.(5) Further, German intellectual culture was quite isolated from the mainstream of European science and philosophy even into the twentieth century (Beck (1967a), and German-speaking intellectuals tended to be distinctively Germanic. Nothing is more Germanic than Kant, especially his model of the psyche. For all these reasons, Freud must have been influenced by him. Yet this influence is seldom mentioned.(6)

Freud himself is partly responsible for this. He refers to Kant on conscience, the super-ego, etc., half-a-dozen times and took Kant on the categorical imperative as an important antecedent. He also mentions Kant's doctrines of space and time (at least four times) and the noumenal (once, directly). He even refers to Kant on dreams (twice) and on jokes (also twice). (Cf. Appendix A, where I have pulled all these references to Kant together and organized them by subject.) But he never so much as alluded to Kant's model of the psyche. Yet unlike the other topics, here Kant's ideas parallelled his own in close and important ways. These parallels not only illuminate Freud; they also contain ideas that are still rich and plausible today.

That Freud never mentioned Kant's views on the psyche is intriguing. He knew about them. One would assume as much on general intellectual grounds. But we also have direct evidence. In the Schreber case ((1911), 34), Freud referred to the work in which Kant laid out his famous model, The Critique of Pure Reason (1787) by name (the comment Freud cites is to be found on p. 97). Indeed, he owned a copy of it. It has "some marginal markings", though Freud seldom put marks in his books, and he had acquired it in 1882 (Trosman and Simmons (1973), 654 & 651), which was exactly when he first began to think about models of the psyche.

The parallels with his own ideas are so striking that he could hardly have missed them. Three of them are particularly prominent. First, like his own final view, Kant's model is tri-partite. The three elements are Reason, Understanding and Sensibility and they can be readily mapped onto Freud's Super-ego, Ego and Id --to show this in some detail is a major goal of the present paper. Secondly, Kant thought that a large and in many ways the most important part of the operation of the mind is `unconscious'--is not and cannot become open to introspection, cannot become conscious in Freud's sense of the term. Thirdly, for Kant as for Freud something can be or become conscious only if it is described, captured in language. Freud also shared Kant's preoccupation with how the mind can tie experiences together, and even used Kant's term, `synthesis', as his name for the process. To all of this we will return.

There are also, of course, features of Freud's model, in particular his views on the contents of the psyche, which are entirely novel and we will have occasion to mention them, too. Nevertheless, the Kantian parallels are real and major.

Yet Freud did not refer to them or to Kant's model of the psyche as a precursor of his own even once. This is doubtless part of the reason why Kant's influence on him in this area has been overlooked.(7) To us now this omission may look suspicious. But it is not. There was no bad faith on Freud's part. Quite the reverse. In his time the Kantian model of the mind at least in its broad outlines was paramount and over-riding, its correctness simply taken for granted in German intellectual life. So it may never have occurred to Freud even to wonder whether there might be other ways of organizing the facts. Marie Bonaparte once said to him that he was a combination of Kant and Pasteur, apparently intending to pay him an enormous compliment (Jones (1955), 415). This makes it clear that Kant was more than well-known to Freud's circle. His was a name to conjure with. On the general nature of the psyche hardly any German-speaking intellectual was not a Kantian, it was so all-pervasive a paradigm (to use a much-abused term). So there was little pressure to acknowledge Kantian parallels. But the truth of Kant's model is not exactly self-evident to us now and so now we do not notice its role in Kant's thinking. This is to our detriment.(8)

The study which follows is hermeneutic. It ignores most of the history and concentrates on analyzing texts. I focus on drawing out parallels in doctrine and conceptual framework and do not go into the question of why they are there--whether, for example, we have here identifiable historical influences or just parallel independent discoveries. Even at that I can only lay out the broad picture. Though it has already been shown that real historical influences existed, the parallels would be just as clear and almost as interesting if Freud had reinvented Kant's model from scratch.(9)

2. Here is a sketch-map of Kant's position.(10) One can be aware only of one's own conscious states (phenomena). In order to be aware of even these, one must perform a number of intellectual operations on the raw material of them, which Kant called intuitions. First, intuitions must be `run through and held together', which includes locating them in time and (some of them) in space. (The spatial part is what will appear to be external to oneself.(11)) Secondly, earlier ones must be retained in the present so that they can be set beside current ones. Thirdly, they must be organized under concepts (including the famous categories of the understanding). All of this yields `representations' (vorstellungen). Finally, one must tie whole groups of these representations together into systems of representations, using for example causal concepts to do so. Only at this stage does one have experiences, at least of the full-blown conscious variety. The general name for these intellectual operations is synthesis.

The higher levels of synthesis are performed by a faculty called apperception. Apperception must be unified in the ways required for it to operate as a single subject of experiences and for it to become aware of itself as a single subject of experience. Apperception, its being unified and the possibility of it becoming aware of itself were extremely important to Kant. But in Freud it played only a ghost of a role, so I will say no more about it. (The ghost of a role is Freud's persisting interest in `double conscience', splits of consciousness, and multiple personality (cf. (1915b), 170-1) and elsewhere).)

Kant called the system which contains intuitions Sensibility. It has two sub-systems, inner sense (roughly, intuitions of one's own psyche and soma) and outer sense (the intuitions which can be spatially organized.)

Understanding is his name for the system which works up these intuitions into representations. It is the language-using system which does the thinking.

The Understanding in turn is governed by Reason. Kant assigned it two functions. One was to generate the regulative principles which the Understanding needed to govern its operations. The other was to house the fundamental principles of morality, which could not be derived from experience. The most fundamental of these was of course the categorical imperative. In Kant's system reason is not a system like the other two and its status is always a bit murky. In this way (and as we will see in others) it is a precursor of Freud's super-ego.

For Kant the principles with which Reason deals are necessary or could not be otherwise and therefore are inescapable, inexorable, for any rational being. Freud was also impressed by the inexorability of moral duty, of course. Though he had a rather different account of the phenomenon and did not, to say the least, think it was a purely rational matter, the fact that he mentioned the categorical imperative a number of times shows that he saw the connection. He seems never to have discussed the other function of Reason, to provide rules of rational thought.

3. If all we can be aware of are states of our own mind, which Kant called phenomena, what about everything else? Indeed, is there anything else? For Kant there was. He called it the `ding an sich' or the noumenal. It had two aspects. It is the source of intuitions, intuitions both of inner sense (of our own psyche and soma) and of outer sense (of the external world); and it is also and crucially whatever it is that works these intuitions up into experiences. Both aspects of the noumenal, both the source of intuitions (in Sensibility) and the apparatus which works them up (the Understanding) must remain forever outside of experience (`unconscious') because both must already have done their job by the time any experience comes into being. They must remain forever merely what produces experiences; they can never be experienced.(12)

Now we come to an important point. The noumenal is the source of the contents not only of outer sense but also of inner sense. Inner sense is our awareness of the Understanding (such awareness of it as we have) and of the rest of our psyche and soma. Freud was certainly interested in the idea that outer sense (our experience of the external world) had a noumenal, `unconscious' source, as we will see. But Kant's other idea, that inner sense has a noumenal base too, should have interested him even more. It is the closest possible anticipation of what he came to call the unconscious. Indeed, his and systems of 1895 together are little more than Kant's noumenal psyche under a different name. Yet he said not a word about it. Here is what he did say on the subject:

Just as Kant warned us not to overlook the fact that our perceptions are subjectively conditioned and must not be regarded as identical with what is perceived though unknowable, so psychoanalysis warns us not to equate perceptions by means of consciousness with the unconscious mental processes which are their object. Like the physical, the psychical is not necessarily in reality what is appears to us to be [(1915b), 171].

In this passage Freud clearly recognizes that his ideas about the external world and its knowability or unknowability are derived from Kant. Indeed, he was by no means immune to the allure even of the radical solipsism into which Kant's position is constantly in danger of slipping, as this passage and others illustrate (cf. (1920), 28; (1938b), 150, 196; (1938c), 300.)

But what about the other side, the internal world? Freud says that psychoanalysis warns us that perceptions by means of consciousness are not to be `equated with the unconscious mental processes which are their objects'. But this thought had already been anticipated by Kant too. For he held exactly the same view about the unknowability of the sources of the internal world as he held about the external world. Here is how he put it:

... inner sense ...represents to consciousness even our own selves only as we appear to ourselves, not as we are in ourselves. ...I ...know myself, like other phenomena, only as I appear to myself, not as I am ... [S]o far as inner intuition is concerned, we know our own subject only as appearance, not as it is in itself [(1787), 166, 167 & 168].(13)

Kant is saying here that much and much the most important part of the Understanding and all other sources (psychic or somatic) of inner sense are opaque to consciousness. Thus Kant anticipates both Freud's ego-unconsciousness in his notion of the noumenal base of the Understanding, and Freud's id-unconsciousness in the idea that the contents of inner sense have a noumenal source. Unlike the unknowability of the external world, powerful arguments can be given for this view.

Fortunately Freud did not accept all the draconian restrictions of Kantian noumenalism. So he could go on to say what Kant could never have said:

We shall be glad to learn, however, ... that internal objects are less unknowable than the external world.

For Kant, of course, the external world is totally unknowable. All we can be aware of are appearances (spatially organized) in our mind.

Furthermore, Freud had some awareness of the Kantian doctrine of the noumenal base of inner sense (though we cannot tell now to what extent he knew that it was developed by Kant). He once quoted Lipps' famous saying that the problem of the unconscious is less a psychological problem than the problem of psychology ((1900), 611; cf. (1938d), 286). Lipps held this view because it was intimately bound up with the views of Kant's about the psyche which he accepted. Indeed, the noumenal psyche is the quintessentially Kantian doctrine.

In short, Freud's notion of unconscious psychic system(s) reflected Kant's doctrine of the noumenal psyche very closely. Indeed, it mirrors both what is plausible in it and what is implausible. And at least in a general way, Freud was probably aware of this even though he never acknowledged it.

4. We could stop there. The above is surely enough to prove the thesis that Freud's view of the psyche contains a lot of Kant. However there is much more. There are still deeper parallels in the two doctrines of the unconscious, as we will see in Sections 7 and 8. But even more interestingly, Freud's view of the structure of the psyche was also anticipated by Kant.

In order to bring this parallel into focus, we must first make a crucial distinction. Freud left us not just two or more theories of the psyche, but two distinct kinds of psychic theory. One was the theory of structure and functioning, for example the theory of the unconscious psychic system which we've just been examining and of the forces which drive it. The other kind is the theory of psychic contents--phantasies, affects, dreams, introjects, character, etc.--and of how such contents are managed--the primary and secondary process. It is Freud's theory of structure that Kant anticipated. Of Freud's theory of contents we find hardly a whisper. (The only exception is what Freud called secondary process. This Kant explored extensively and brilliantly).

How different the two kinds of theory are was never made clearly clearer by Freud than in 1923. The remarkable shift that occurs between Ch.'s I and II and Ch. III of The Ego and the Id is what I have in mind. Ch. I gives a list of ego functions and Ch. II (to simplify) a theory of how the system which performs these functions works and how it relates to perception. This is all familiar theory of structure and function. But at the end of Ch. II something entirely new suddenly appears. It is the fascinating idea that the ego is first a bodily ego, a projection (a model) of the surface of the body. This is not a point about ego functions. It is the first move in an entirely new discussion, a discussion of ego contents, a discussion of the representations, images (vorstellungen, bildes) it contains. Ch. III then articulates what may be Freud's first full theory of contents, beginning with the famous idea that a key part of content, namely character, is "a precipitate of abandoned object-cathexes". That the two kinds of theory differ radically could hardly be clearer.

It is important to realize that this fundamental distinction between the structure and function of the `ego-machine' (to coin an ugly but fitting phrase), on the one hand, and its content of representations on the other, runs throughout Freud's work. Here is a particularly clear instance of it from an earlier period. Freud has just warned against the dangers of picturing the psyche topographically:

We can avoid any possible abuse of this method of representation by recollecting that ideas, thoughts and psychical structures in general must never be regarded as localized in organic elements of the nervous system but rather, as one might say, between them, where resistances and facilitations provide the corresponding correlates. Everything that can be an object of our internal perception is virtual, like the image produced in a telescope by the passage of light-rays. But we are justified in assuming the existence of the systems (which are not ... accessible to our psychical perception(14)) like the lenses of a telescope, which cast the image. And, if we pursue the analogy, we may compare the censorship between the two systems to the refraction which takes place when a ray of light passes into a new medium [(1900), 611].

This seems to mark the clearest possible distinction between representations, whose character had to wait until 1923 for clear elucidation, and the structures and functions which produce and contain and manage them. (The two great management principles are, of course, the primary and secondary processes.)

There is no antecedent in Kant for Freud's theory of the psychic contents (nor of course for what might be his most completely original idea, the notion of primary process). But the distinction between contents, ie. representations, and structure, ie. the forces and functions which generate and manage content, is Kantian.

So--most significantly--is that idea that we are not conscious of structure itself but only of the results of its operations, its contents, ie. representations. In fact, the passage from (1900) which we just examined suggests very strongly that this may have been the source of Freud's life-long view that there is both a system which is unconscious and many representations which are unconscious, drive derivatives for example. The lenses in the 1900 analogy are the system and the virtual images are its contents, the representations. It was a distinction that he never gave up (cf. (1924)).

The distinction between structure and content should not be pushed too far. Character is a case in point. Though it was Freud's entre into content theory, it actually has a foot in both camps. It is clearly a matter of contents, representations--what else would abandoned object-cathexes be? Yet it is more than just content, because character shapes and controls how content (experiences, thoughts, desires) is managed. Perhaps a computer analogy might be helpful. Computers have not two main elements but three. They have data (the analogue of content) and they have hardware (the analogue of structure). But they also have programming. This programming gives the hardware its functions, it capabilities. But it is also content. Character seems in a similar way to be both function and content.

The element of force or energy also muddies the water. It is not function or structure. But it is not content either. That is too big an issue to go into here, but this distinction between structure or function on the one hand and contents on the other itself gives rise to two questions. First, in comparison to Kant what did Freud think that a psychic system was like? Secondly, if we cannot be conscious of systems but only of contents, what about consciousness itself?

5. Although Freud's view of the structure of the psychic system underwent shifts and developments, it remained fundamentally the same from 1895 to 1938. This was because he never abandoned the essentials of the model he developed in 1895. Every subsequent model had a drive system like the endogenous Q of 1895 (libido, the Id) and an ego system like the ego of 1895 (the ego of 19ll and 1917, parts of the systems Ucs and Pcs, or the Ego of 1923). All of them also had a self-judging element, though it only took its final shape as the Super-Ego in 1923. And all of them remained drive-discharge theories. What I find significant in this is that Freud's image of the psyche was fundamentally tri-partite from start to finish, though this also only became fully explicit in 1923.

Now tri-partite models of the psyche have been around at least since Plato. But it was Kant who brought tri-partite modelling to prominence in German thought, and gave it a distinctively Germanic shape. Few images of the psyche have ever been more influential. Freud's tri-partite model, especially the final model of 1923, corresponds to it in detail. It even exhibits many of the same obscurities and ambiguities. The diagram which follows lays out the structure of Kant's model and shows how it relates to Freud's. I have set it up so that the two psyches are facing each other. The arrows link each element of Freud's model to the element in Kant's to which it corresponds most closely. Not all the correspondences are perfect, of course, but I will comment on this after we examine the diagram. Here, then, is Kant's model and how it relates to Freud's:


Noumena Phenomena Consc. Pre-cs. Ucs.

Categ.Imperative <-------------> S U P E R -E G O...>

REASON ____________________

Regulative Princ. <-------------> (?) ____________________ (Self-Consciousness) <-------------> |Cs.|


____________________ <-------------> E G O ............>

Outer Sense <-------------> (Ext. World)


Inner Sense <-------------> ___________ Will <-------------------------------------------> I D (Schopenhauer) _

Figure 1: Kant's model face to face with Freud's.

Brackets indicate elements whose location in the model in question was never made clear, consciousness of self in Kant and perception of the external world in Freud. From top to bottom, here is how the two models relate to one another.

Noumena--Unconsciousness: The section labelled `Noumena' has been left blank because Kant taught that we cannot say anything about it, not even whether mental systems and the external world are discrete entities or an undifferentiated whole there. Literally all we can know is that it exists. Though Freud's notion of the unconscious was certainly influenced by Kant's doctrine, he never thought it to be either as mysterious or as opaque as Kant held the noumenal to be. We will return to this topic in Section 8.

Phenomena--Pre-conscious: Though Kant did not specifically split representations into those one is aware of having (consciousness) and those one is not but could become aware of having (pre-consciousness), he knew of and used the distinction ((1787), 134, 142 fn.), and like Freud contrasted both with something vastly larger of which we cannot become aware, namely structure and function. Unlike Freud he seems to have had no interest in whether there were representations which were not just pre-conscious but truly unconscious. However he certainly believed that intuitions, the raw material of representations, had to be.

Categorical Imperative--Super-ego: As we have seen, Freud himself was well aware that his Super-Ego was a descendant of Kant's categorical imperative. It goes far beyond the categorical imperative, of course. Kant would never have dreamt of even the fact of parental introjects let alone that they could have anything to do with the categorical imperative! But it is also less than the categorical imperative. Kant used the notion to attempt to ground a certain morality. Not only did Freud not follow him in this but he paid little attention of any sort to the issue of justifying moral beliefs, though he certainly studied their genesis extensively and was himself a rigidly moral person.

I have shown the Super-Ego as extending into the Unconscious. This seems to have been Freud's dominant view. Nevertheless there is a problem here as Freud himself came to recognize. The contents of the Super-Ego are retained and expressed in concepts, word-presentations. But association with word-presentations is Freud's leading way of demarcating the Pre-Conscious from the Unconscious (cf. Section 7). Thus the contents of the Super-Ego have to be Pre-conscious. When Freud finally got the problem clearly into focus, he is how he resolved it:

...large portions of the ego, and particularly of the super-ego, which cannot be denied the characteristic of preconsciousness, none the less remain unconscious in the phenomenological sense... [(1938b), 162].

Here is one way to read this remark. The Super-Ego consists of representations, contents, not structure or function. But many of these representations cannot be introspected; that is, I cannot become aware that I have them. (Unconsciousness in the `phenomenological' sense must be the old `descriptive' unconsciousness.)

Regulative Principles--?: Kant's regulative principles of Reason have to do with a drive to completeness and generality. There is nothing parallel to this in Freud's writings, as I have indicated with a question-mark. If Freud had dealt with them he would have included them in the Ego.

Self-Consciousness--Cs: Probably the most mystifying aspect of both these models is how consciousness appears in them. For in both of them it appears twice--in Kant's as a heading (in the form of Phenomena) and as an aspect of the Understanding and in Freud's as a heading and as a sub-system of the Ego. This is an interesting parallel which deserves to be explored.

Freud first. As is well-known, he treated consciousness as both a sub-system of the Ego (Cs; the (omega) system of 1895), and also as a quality that certain of the contents of any system can have (though few contents of the Id actually have it). Though he was never very clear about the matter, he seems to have held that a representation gains the quality of being conscious by becoming available to the system Cs, the system (or sub-system) which is sensitive to this quality.

Consciousness as a quality is in a direct line of descent from Kant's consciousness of phenomena, that is, of representations. (Freud must have known this.) It contrasts with representations of which one is not aware (Freud's pre-consciousness) and of course noumena or the non-conscious. Items in any system can have or lack the quality of consciousness. So, like Kant's phenomena, it and pre-consciousness go at the top as a heading.

Cs the system descends from something quite different, Kant's apperceptive function of the Understanding. (Where Freud talked about systems, Kant usually though not always talked about functions.) For Kant apperception begins with the application of concepts to intuitions of objects. Consciousness begins here too. But this by itself would only be what Freud called pre-consciousness. In order to get what Freud called consciousness, apperception must make one aware not just of the thing but also of being aware of it, of oneself. Awareness of the thing would be normal perception. Awareness of being aware of it would be what Freud called consciousness. For Freud's consciousness is really consciousness of self, awareness of oneself and one's psychic states. It is the difference between perceiving something (as animals do) and being aware of perceiving it. In Kant such awareness of one's perceivings is a result of apperceiving. Since the system Cs was that in us which perceives perceptions and other psychic states ((1915b), 171), it is exactly like this aspect of apperceiving.

Kant and Freud both distinguished having representations of the inner (oneself and one's states) or the outer (the external world) on the one hand, from awareness of having them on the other. And they made the distinction in a similar way. Awareness of having them, Kant's self-consciousness, and only this counted as (the quality of) consciousness for Freud. All other representations were merely pre-conscious. For Kant the apperceptive function which yielded self-consciousness was always second-order. It began with an already-existing representation and involved doing something extra to it. Like Freud, Kant was never clear about what more was involved, and like Freud he left us hints that it might be something to do with attention. At any rate, with respect to self-consciousness inner sense is just like outer sense or perception--not a part of it but something to which it is applied.

In Freud, this takes the following form. We have perceptions of the external world. And we have various states which contain representations of the inner world--affects, drive representatives, representations of the soma and of course beliefs, desires and thoughts. The system Cs, which is distinct from all of these, gives the quality of consciousness to all of them. Otherwise they are merely pre-conscious or unconscious.

Actually, the picture is a bit more complicated than this. For Freud changed his mind about perception in the early 1900's and as a result left us two views on how it relates to Cs. In his earlier view (1895 and 1900, Ch. VII) the system or Cs is one system, the system or Pcpt. is another. But later on he connected them and called the resulting notion Pcpt.-Cs. (The first occurrence of this term seems to be in ((1920), 24) but the idea and even a version of the term went back at least to ((1917), 232).)

This however was not so much a change of mind about how perception relates to consciousness as it was a change of mind about what to label with the word `perception'. In the earlier view what Freud called perception (the system) was prior to consciousness (pre-conscious?) and needed extra processing in the system to take on the quality of consciousness. And he made provision for a similar pre-conscious capacity for representing the inner, based on inner `excitations'. These correspond closely to the outer and inner senses of Kant's Sensibility. And the system was of course a separate system. It corresponds to Kant's self-consciousness, the capacity to perceive perceptions and became the (sub-)system Cs. In the later account, however, `perception' became the name not for pre-conscious access to the outer world but for awareness of these representations themselves, for awareness of perceiving. Or rather, it became the term for both--for both access to the world and our awareness of that access. And in the latter usage it covered not only awareness of access to the world but also awareness of all other inner states--affects, desires, etc. Thus Freud came to use `perception' as a name both for our access to the world (which he also called `excitations' ((1920), 24) and, in the form `Pcpt.-Cs.', as the name for self-consciousness, for perception of that access and other inner states. Indeed, for Freud awareness of representations of the inner was not radically different from awareness of representations of the outer in any way whatsoever.

Curiously enough, Freud's running together access to the external world with awareness of that access might also have been a reflection of something in Kant, namely his idea that perceptions and other mental states not only make us aware of their object but are also the basis for awareness of themselves. How Kant saw the relation of self-consciousness to perception, of apperception to perception, and how these all connect to the Understanding never becomes clear. And this at least Freud certainly reflects. How the system Cs is supposed to fit into the Ego and how it relates to pre- or non-conscious representations of inner and outer is unclear in just the same way, though perhaps the missing metapsychological paper on Consciousness resolved this unclarity. They also both held that whatever this system or function is like, it is but a tiny part of the total psyche.

Freud's innovation was to argue that it can even make us aware of much less in that total psyche than Kant had allowed. And that was little enough. For both of them, there is no reason to think that Cs the system will be conscious even of its own working, despite the fact that it is the system for consciousness! Certainly it will never be conscious of everything about itself.

Understanding--Ego: I have shown the Ego as corresponding both to Understanding and to Sensibility in Kant. The reason is that Freud included inner and outer sense in the Ego while Kant treated them as separate, not part of the Understanding. Since we have already examined inner and outer sense, let me here take up the connection to the Understanding. The cognitive side of the ego, what Freud called the ego-functions, corresponds quite closely to Kant's Understanding though Kant's theory of these functions is fuller than Freud's. In computers they would be the programming which allows the hardware to manage the data. They are what allows the ego as system to manage representations. (Prior to the growth of the ego and such secondary processing, representations are unbound, manifest the primary process.) Freud talked about ego functions a number of times, including in (1923), 17 and (1926), 196.

The latter is one of many places where the ego is described as "an organization characterized by a very remarkable trend towards unification, towards synthesis". Synthesis is a central notion in Kant's account of the Understanding. Freud not only retained Kant's concept but even his term. Freud's interest in it was life-long. As early as 1895, he had noted that acts of judgment bring about a "psychological unity" by synthesizing stimuli into experience of objects (p. 384) and over forty years later this fact still interested him ((1938a), 276).

Outer Sense--External Perception: Here I will add only one comment to what I have already said--Freud did not ever really develop a complete theory of perception, either of the inner or the outer. In 1895 he treated outer perception as a separate system, the system, distinct from both consciousness and the ego. This is what I have reflected in Figure I. Later the only sort of perception he explicitly distinguished was the sort which involved (self-)consciousness, the system Pcpt.-Cs., though he seems to have continued to use the term to cover pre-conscious access to the external world too as we have seen. But to the extent that he did have a theory of outer and inner perception, it was based on Kant's Sensibility.

Inner Sense--Drive Representatives: What I have called `drive representatives' are those processes which are representations of drives in the (pre-)consciousness. Freud also called them psychical representatives ((1915a), 122). Together with affects, representations of soma and beliefs, desires, thoughts, etc. (which I have not specifically identified in Figure I), they correspond in Kant to inner sense. Except in 1895 Freud said even less about what the mechanisms of pre-consciousness representation of oneself and one's states are like than he said about access to the external world.

?--Id: Nothing directly comparable to Freud's Id can be found in Kant. As a source of representations (rather than a vehicle of representations like inner and outer sense), it would have been rolled into his notion of the noumenal. I have included it because Freud developed the Id from libido, which in turn was developed from the endogenous sources of quantity of 1895, and this latter notion does have an exact parallel in Kant. It corresponds to his notion of the noumenal source of the inner sense representations of desire, inclination and affect.

But that is not all. Kant's idea of the noumenal also underwent a development though not by Kant. It was the source of Schopenhauer's notion of the Will. Kant provided the `representation' part of Schopenhauer's world as will and representation in roughly the way the Project provided the ego part of Freud's ego and id. Though Freud denied that Schopenhauer's work had any influence on him ((1925), 59), Schopenhauer's Will is strikingly similar to the Id. The excerpt from Schopenhauer's writings which Strachey reprinted in S.E. 19 makes this very clear. Here is a short passage from it:

...the sexual drive is the kernel of the will to live... therefore... I have called the genital organs the focus of the will. Indeed, one may say that man is concrete sexual drive [(1819), 3, 313-4; reprinted in S.E. 19, 223-4. `Sexual drive' is my translation of `Geschlechtstrieb'].

Yet he both considered himself to be and was a true Kantian. He saw his work as merely developing what was implicit in Kant.

In fact, Schopenhauer anticipated more in Freud than that sexuality is central. He also anticipated Freud's idea that it is a binding force. As he put it, it "holds together [man's] phenomenal existence". In fact, sexuality first began to take on this character for Freud around the time when by his own admission he was reading Schopenhauer, about 1920.(15) Prior to then, he had seen sexuality as merely a disruptive force and it was only stimuli that were bound. Moreover, the kind of binding involved was quite different from the new binding by Eros. The old binding was done by the ego, not by sexuality. Far from being innate, it was learned slowly and painfully under the tutelage of unpleasure. And it bound by controlling, not by joining into larger wholes. The 1920 addition of a kind of innate erotic binding which joined and built (and of its opposite the death drive) was a major and controversial one. I discuss this further in a forthcoming paper o death, eros, sadism, sleep and Freud's last drive theory so here I will restrict myself to just one point. All its essentials had already been laid out by Schopenhauer. He called it the will to live and even saw that it might elucidate our competing tendencies to preserve the individual and the species.

Representations of Id elements--of urges, desires and affects--present an interesting problem for the structure/content distinction. For they are clearly content, representations. Yet they have force or energy--they do work, move the psyche to do things in the way aspects of structure do. This of course raises the vexed problem of psychic energy, of which differing strengths of desire are a particularly clear example. Once raised, the issue must be addressed. But how to do so is far from obvious, at least within a Kantian/Freudian framework. Certainly labelling it a metapsychological myth is not adequate.

6. There is one fundamental difference between the two models, a difference which is not shown in Figure I. Kant thought that Reason and the Understanding must exist full-blown and ready for action prior to the first experience, since the functions they perform (organizing intuitions and capturing them in concepts) are necessary to have experience. They therefore cannot be acquired from experience or learned and so must be innate. Indeed, more is innate in Reason than just its cognitive functions. Kant thought its moral side was innate too. Here he had in mind the categorical imperative, the foundation of morality. The only element of the psyche that was not innate for Kant was experience itself, inner and outer. Freud, by contrast, held that the Ego and everything it could do were acquired from experience, specifically from having to defend against conflicts generated by the drives. Morality, the contents of the Super-ego, were similarly not innate but acquired, by introjections and narcissistic projections. For Freud the only innate element, except for a general disposition to keep content down (the unpleasure principle), was that part of the content of inner sense caused by the drives. And that is precisely part of the only acquired element in Kant, namely the content of inner and outer sense! (At any rate this was Freud's picture until 1920, at which point Eros was added as a new innate factor.) Thus Freud's view of what is innate and what is acquired was almost the exact opposite of Kant's.(16)

A consequence of this difference is that they had rather different views of child development. For Kant Reason, Understanding and the structure at least of Sensibility had to be there from the beginning. All a child acquires from experience are the contents of experience. For Freud by contrast only the drives and the disposition to avoid unpleasure are innate, at least until 1920. Everything else is acquired, including the ego-functions and morality. (Perhaps the propensity to learn them is innate, a result of evolution.)

Related to this, Freud viewed his models as empirical, not a priori: as discovered experientially, not inferred using `transcendental arguments' concerning the necessary conditions of experience. As a result he had little interest in claims as to how things had to be. He was happy enough if could even discover how they in fact were. On these issues Kant was still a Rationalist, whereas Freud was already a child of the new Weltanschauung of biological empiricism which is with us still. Interestingly, they both began as materialists about the mind.

7. There is yet another highly significant correspondence between Kant and Freud. They viewed the relation between perception and description in the same way. The significance of this is that Freud used this view to define the unconscious. Just about the most famous remark Kant ever made was this:

Thoughts without content are empty, intuitions without concepts are blind [(1787), 93].

I want to focus on the second half of this saying, that intuitions (percepts) without concepts are blind. Freud held precisely the same view. He put it this way. A `thing-presentation' (a percept, a representation of an object or a state of affairs) cannot become conscious unless it is linked to `word-presentations'. "A presentation which is not put into words ... remains ... in the Ucs." ((1915b), 202). Thus this view of how percepts relate to concepts also defined the unconscious for him. He gave the same account, more plausibly and with arguments, in 1895 (364ff.). More plausibly because he there made clear something that he did not even mention in 1915, that it is only memories that require words to become conscious. We can be aware of perceptions of the outer, sensations of pain or unpleasure and current bodily movements, it is there quite clear, without putting them into words. (Kant would not have agreed to this exception, but here Freud has the more plausible view.)

In 1915 Freud did not argue for the link he claimed between language and consciousness. Nor is it easy to find an argument for it in Kant, plausible though it is. So the fact that Freud does argue for it in 1895 is noteworthy. His argument is complicated, but worth examining nevertheless.

Freud takes cognizing, coming to understand something we perceive, as the example with which to begin. To cognize, he says, we must pay attention to (hypercathect) the perception in question and then see to which mnemic images its energy (Q) flows. Put less mechanistically, to recognize what qualities we are perceiving, we must compare the perception to our memory of qualities we have encountered before and so can recognize. The process Freud is describing her of coming to recognize by comparing what is seen to memories of what was seen earlier is very much like Kant's Synthesis of Recognition in a Concept ((1787), 133f.).

Reverting again to Freud's mechanistic jargon, to compare a perception to memory requires that the energy from the perception select (cathect) a mnemic image of a quality. But it must select the right mnemic image, one that is actually similar to the qualities of the perceived object. Just any old image won't do. Thus we have to be able to discriminate one mnemic image from another. This in turn requires that the images themselves be recognizable, that is, trigger conscious states (his qualities; qualia or sensa in more modern jargon). But this does not happen automatically; in fact, most memory-traces do not enter consciousness at all. So what is required for a memory-trace to become conscious?

Now comes some moves which are crucial for understanding Freud's theory of consciousness and just about as obscure as anything ever written. For energy from the perception to be directed to the right mnemic image, namely one similar to it, the mnemic image must itself be perceptible, enter consciousness (p. 364). In addition, the information from the image thus perceived must be discharged back to the Ego, to the mechanism of attention which guides the flow of energy from the original perception to the correct memories. So far there is no reason to think that language is required.

But now comes the crucial move. The correlation between what is perceived in the memory image and what is discharged about it to attention (the `indication of quality') must be tight and reliable. Input and output are systematically and tightly related in this way only in language. Only with speech is what we discharge (what we say) correlated in a systematic (Freud says exclusive) way with what we perceive (what we hear). With all other forms of perception, no such correlation exists. What we see is not correlated in a predictable way with any other kind of behavior. So if memories are to help us recognize what we see, we must recall them in the form of descriptions of what is remembered, which "have the closest association with motor speech-images" ((1895), 365).

Note what this argument does and does not establish. It does not establish that language is necessary for all forms of conscious recall of memories. But it does establish that language is necessary for conscious recall in any form which could be used by the secondary process. Perhaps the best way to put the point is this: a memory can only be consciously managed if what is remembered can be described.(17) But Freud thought he had proven more. Certainly he felt that (1915b) established the stronger claim, that memory-traces without associations to words can never enter consciousness at all. To be sure, he applied even this stronger claim only to memory and not to perception, unpleasure and bodily/motor sensation. So even it is much weaker and more plausible than Kant's claim, which was that all `intuitions without concepts' are `blind'. Nonetheless, the line of descent is clear.(18)

In fact, the element of disagreement between them may be more apparent than real. While Freud thought that we can have such psychic states without having words for them, we nevertheless cannot think about any of them without words ((1895), 367). That is, we cannot focus on them, relate them to other psychic states, identify them as instances of already-familiar experiences and so on. But it is not clear where merely being aware of an experience stops and thinking about it begins. Since it is also not clear that Kant would have denied that one can be merely aware of a psychic state without concepts, it is not clear whether he and Freud really disagree.

Freud's argument probably would not stand up to scrutiny. For one thing, it seems to require that even thinking of a word give off, discharge, a small quotient (a very small quotient ((1895), 367) of energy from one psychic system to another. But it is an argument. And that is noteworthy. Few views of Kant's are more plausible or have had more influence than the view that thinking and self-aware experiencing (consciousness in Freud's sense) necessarily involve the use of language. Yet very few arguments for the view exist. That is what makes the fact that Freud actually argued for it so noteworthy.

The idea that consciousness requires language is connected to one further Kantian strain in Freud. Like Kant but unlike most psychologists he put memory front and centre in his theory. He even used Kant's terminology, speaking for example of `reproductive' thought ((1895), 319) (Kant: `reproductive imagination' ((1787), 132). Freud's theory of memory richly deserves a proper study.

8. To close, let me return to Freud's attraction to noumenalism. It went beyond even the extensive range of interests we have already explored. For Kant, the noumenal cannot be spatial nor temporal. Nor can it cause anything else. These are all imposed on experience by the mind and therefore could not be part of noumenal reality. This view is generally held nowadays to be nonsense based on confusions. Nonetheless, it was probably the source of Freud's notion that the unconscious is timeless. He once almost said as much himself:

As a result of certain psychoanalytic discoveries, we are to-day in a position to embark on a discussion of the Kantian theorem that time and space are [merely] `necessary forms of thought'. We have learnt that unconscious mental processes are in themselves `timeless'. This means in the first place that they are not ordered temporally, that time does not change them in any way and that the idea of time cannot be applied to them. ... On the other hand, our abstract idea of time seems to be wholly derived from the method of working of the system Pcpt.-Cs. and to correspond to a perception on its own part of that method of working. ... I know these remarks must sound very obscure ... [(1920), 28].

What is interesting about what Freud says here is that he himself did not in fact take the timelessness of the unconscious literally; but Kant would have.

Freud did not ever deny, for example, that unconscious processes start (in time), can be modified by psychoanalysis (in time), and end with death (in time). By their timelessness he seems mainly to have had in mind that these processes tend not to change over time, in particular are not modified by maturation, and they can appear out of `normal' temporal sequence in bizarre ways. Kant by contrast would have held that any idea of unconscious processes (as contrasted with subsequent conscious or pre-conscious representations of them) beginning or ending would be sheer nonsense, that the concept of time quite literally cannot be applied to them. This is just what Freud said; but, thank goodness, did not mean, not literally. I do not have space--or time--to go into why Kant did.

What Freud says here about time he later said about space too. He held that it too is a projection of our own psyche. In one of the last small pieces he ever wrote we find this. "Space may be a projection of the extension of our psychical apparatus. No other derivation is possible." ((1938c), 300). This is slightly different from his view of time; we get the concept of space from how the psyche is built, not as with time from how it works. Further, he directly contrasts his idea that space is an acquired concept with Kant's view of it as a priori--innate and not learned. But these differences are not significant. What he accepts from Kant is what is important, namely Kant's idea that space is imposed on reality by the mind.(19)

If Freud had been completely serious about this, he would also have had to agree with Kant that nothing outside of the mind is temporal or spatial or even available to consciousness. As we have seen, he did not. While Kant's vision of the world was his template, he never took its full metaphysical rigor literally. If he had, solipsism would have awaited him, as it did Kant. Nevertheless, in a great many ways his views of the mind, knowledge and world were Kantian.


Freud's References to Kant and Kantian Doctrines

Subject Reference

Critique of Pure Reason (by name).......... S.E. 12, p.34

Categorical Imperative .................... S.E. 4, p.68

S.E. 13, pp.xiv & 22.

S.E. 19, p.167

S.E. 22, pp.61 & 163

Dreams (and madness as waking ....) ..... S.E. 4, pp.70-1 & 90

Jokes ..................................... S.E. 8, pp.12 & 199.

Space and/or Time ......................... S.E. 5, pp.503-4

S.E. 18, p.28

S.E. 19, p.231

S.E. 22, p.74 & n.1

S.E. 23, p.300

Synthesis (incl. of dreams) ............... S.E. 3, p.46, 51n.

S.E. 4, pp.310ff

S.E. 5, p.597

S.E. 7, p.88

S.E. 20, pp.97-100,

112, 196

S.E. 22, p.76

S.E. 23, p.276

The Noumenal (not by name) ................ S.E. 14, p.171


Amacher, P. (1965). Freud's Neurological Education and Its Influence on Psychoanalytic Theory. New York: International Universities Press.

Bakan, D. (1975). The authenticity of the Freud Memorial Collection. J. Hist. of the Beh. Sci. 11:365-67.

Beck, L.W. (1967a). German Philosophy. In The Encyclopedia of Philos ophy, ed. Paul Edwards. New York: Macmillan Publishing Co., Inc. and the Free Press, vol. 3, pp. 291-309.

--------- (1967b). Neo-Kantianism. In Ibid., vol. 5, pp.468-473.

Brook, J. A. (Forthcoming). Death, eros, sadism and sleep: the Project and Freud's last drive theory.

----------- (Forthcoming). Splitting and repression in Freud's Work.

Jones, E. (1953). The Life and Work of Sigmund Freud, Volume I. New York: Basic Books

-------- (1955). ----------------------------------, Volume II.

Freud, S. (1895). Project for a scientific psychology. S.E.1.

--------- (1900), The interpretation of dreams. S.E.4 & S.E.5.

--------- (1911). Psychoanalytic notes on an autobiographical account of a case of paranoia (dementia paranoides). S.E.12.

--------- (1915a). Instincts and their vicissitudes. S.E.14.

--------- (1915b). The unconscious. S.E.14.

--------- (1917). Metapsychology supplement to the theory of dreams. S.E.14

--------- (1920). Beyond the Pleasure Principle. S.E.18.

--------- (1923). The ego and the id. S.E.19.

--------- (1924). A note upon the `mystic writing pad'. S.E.19.

--------- (1925). An autobiographical study. S.E.20.

--------- (1926). The question of lay analysis. S.E.20.

--------- (1938a). Splitting of the ego in the process of defence. S.E.23.

--------- (1938b). An outline of psychoanalysis. S.E.23.

--------- (1938c). Findings, ideas, problems. S.E.23.

--------- (1938d). Some elementary lessons in psychoanalysis. S.E.23.

Holt, R.R. (1963). Two influences on Freud's scientific thought: a fragment of intellectual biography. In Essays on Personality in Honour of Henry A. Murray. New York: Atherton Press, pp.364-387.

Kant, Immanuel (1787). Critique of Pure Reason. Trans. Norman Kemp- Smith. London: Macmillan Publishing Ltd., 1963.

Kanzer, M. (1981). Freud, Theodor Lipps, and "scientific psychology". Psychoanal. Q., 50:393-410.

Lewis, N. and Landis, C. (1957). Freud's library. Psychoanal. Rev., 44:327-354.

Schopenhauer, A. (1819). The World as Will and Representation. Excerp ted in S.E.19, pp. 223-4.

Trosman, H., and Simmons, R. D. (1973). The Freud library. J. Amer. Psychoanal. Assn., 21: 646-687.

Various (1973). Panel on the ideological wellsprings of psychoanaly sis. Psychoanal. Q., 43: 167-242.

2364 Orlando Avenue

Ottawa, Canada, K1H 7K1

1. . A shorter version of this paper was read at the 35th International Psychoanalytic Congress, Montreal, 1987.

2. . Holt (1963) is a typical if early example of such studies.

3. . Another major influence on German-speaking intellectual life in the latter part of the nineteenth century was the Romantic idealism of Hegel and his followers. Interesting, this school of thought seems to have had almost no influence on Freud. Hegel and the Hegelian philosophy are only mentioned twice in the entire Standard Edition (S.E. 4:55 & S.E. 22:177), and the only obviously Hegelian idea in Freud's work is the idea of projection. To a first approximation, German intellectuals seem to have been divided into the Kantians and the Hegelians, the science-lovers and the speculators. In such a split, Freud would clearly have been a Kantian, despite a strong speculative streak in both his youth and his old age.

Freud's library, the part of it he took with him to London, is interesting with regard to the influences on him. (What information we have about the part he left behind is not reliable (Trosman and Simmons (1973); Bakan (1975)).) He took Darwin's Gesammelte Werke (a translation), and most of Mach's works but, curiously, only one late work by Helmholtz and only two by Meynert. No Hegel, no Fichte, no Schelling, only one early work of Schopenhauer's, but Nietzsche's Gesammelte Werke in 23 volumes were included. He also took two books of Kant's work, The Critique of Pure Reason, and Short Writings on the Philosophy of Nature (both in German, of course). The significance of the first of these will emerge shortly.

4. . Here are the figures: Plato (17), Aristotle (19), Schopenhauer (25), Lipps (26), Nietzsche (17), and Kant (28+; 16 by name plus at least a dozen more to his doctrines). Lipps is included, though he was a psychologist, because he was a prominent Kantian. Appendix A provides a complete list by subject of Freud's references to Kant or his doctrines. Many of them are not mentioned in Strachey's indexes, which are not entirely reliable (Kanzer (1981) has noted this too).

5. . Cf. Beck (1967b). Jones ((1953), 375) indicates that Meynert in particular was well-known as a Kantian. Meynert was one of Freud's principal teachers (Amacher (1965). Beck (1967a & especially 1967b) gives a good description of Neo-Kantianism in the German-speaking countries in the late nineteenth century.

6. . Just two examples. Holt (1963) does not mention Kant or Kantianism once. Nor did the Panel on The Ideological Wellsprings of Psychoanalysis (1973).

7. . Another reason might be that the kind of influence Kant's model of the psyche had is at an entirely different level from the influences commonly mentioned such as mechanism and naturphilosophie. A model of the mind is rather different from a grand metaphysic of all reality or a sweeping Romantic epistemology.

8. . Freud's ambivalent attitude to philosophers, and to Kantians such as Lipps, is also relevant here (cf. Kanzer (1981)). This ambivalence would make an interesting study. He can, for example, deny with all sincerity about philosophers in one place things which he clearly acknowledges in other places to be true.

9. . Though that would be enough to make me believe in a Zeitgeist--or a collective unconscious!

10. . It is a condensation of the sections called `Transcendental Aesthetic', `The Pure Concepts of the Understanding, or Categories', and `The Transcendental Deduction of the Pure Concepts of the Understanding', especially the latter, in the Critique of Pure Reason (Kant (1787)). His views on Reason are not well-represented in these sections, but they are not completely presented in any other single place either. The Introduction and `Book I. The Concepts of Pure Reason' of the `Transcendental Dialectic' discuss it further.

11. . There is an interesting problem about what the word `external' in `external world' can mean in Kant. It cannot mean `located outside me' because for Kant the whole of space is a state of me; intuitions are not spatial when they arrive in Sensibility but have spatial location imposed on them. (Kant says the same, even more bewilderingly, of time.) So `external' can only mean `not states of myself'. The notion cannot have any spatial connotation.

12. . Apparently Herbart held the same view (Beck (1967a), 305).

13. . These short passages are from the `Transcendental Deduction', the most central and most famous text in Kant's work, not from some minor, obscure paper. As I noted, Freud owned and marked up a copy of the work in which it occurs.

14. . I have deleted a passage here, "... which are not in any way psychical entities themselves and..." in which Freud seems, bewilderingly, to slip into equating the psychical with representations, and indeed with (pre-)conscious representations, thereby obliterating the distinction between system and content which he has just drawn.

15. . His own pronouncements notwithstanding, there is reason to think that Freud knew of Schopenhauer's ideas long before 1920, in fact early as the 1870's. For Schopenhauer was one of the figures discussed by Brentano in his lectures during the years when Freud attended them (P. Swales, private communication).

16. . Freud rarely discussed the innate vs. acquired issue, except insofar as it was part of the problem of the `constitutional factor'. (But cf. (1900), 549.)

17. . This view of Freud's about the relation of memory and language also had implications for his theory of repression and the unconscious. I explore these in a forthcoming paper on splitting and repression in Freud's work.

18. . Note that the other half of Kant's saying, that thoughts without content are empty, also anticipated a doctrine of Freud's, this time about schizophrenia. In schizophrenia, Freud held, word-presentations have become detached from thing-presentations, that is, from their referents, their content. Moreover, Freud made this claim in the same passage in which he laid out the doctrine of the unconscious which parallels the other half of Kant's famous remark ((1915b), 196-204).

19. . Helmholtz, the great founding father of the school of medical thought in which Freud was trained, also held this view of space (Beck, (1967b), 469).