The Vertigo of Philosophy: Deleuze and the Problem of Immanence

Christian Kerslake

Published in Radical Philosophy no. 113, May/June 2002.

One of the few terminological constants in Deleuze's philosophical work is the word 'immanence' and it has therefore become a foothold for those wishing to understand exactly what 'Deleuzian philosophy' is. That this ancient and well-travelled notion is held to have been given new life and meaning by a Deleuzian approach is evidenced in much recent secondary literature on Deleuze, and, significantly, in one central theoretical section of Hardt and Negri's Empire, which takes up the theme of 'the plane of immanence'. Yet on closer inspection it becomes clear that what is at stake in Deleuze's contribution to the history of this term is actually quite elusive. I will claim here that 'immanence', despite appearing to connote philosophical transparency, is in fact a problem for Deleuze; indeed perhaps it is the problem inspiring his work. Not for nothing does Deleuze suggest that 'immanence is the very vertigo of philosophy.'
Can a preliminary definition of immanence be given at the outset? I would suggest that two features ñ one formal, the other ontological ñ are pre-eminent. Formally, a philosophy of immanence is a philosophy that does not appeal to anything outside the terms and relations constructed and accounted for by that philosophy. Ontologically, we might say that in a philosophy of immanence, thought is shown to be fully expressive of being; there is no moment of 'transcendence' of being to thought. Such general criteria, however, could be said of a multitude of philosophies from early Greek cosmology onwards. By which criteria, then, could a philosophy be said to be 'more' immanent than another?
Hardt and Negri, by focussing explicitly on what they take to be an exhaustive opposition between immanence and transcendence, claim that there is something specifically modern about the notion of immanence. 'The primary event of modernity', they say, is 'the affirmation of the powers of this world, the discovery of the plane of immanence'. For them, the characteristic of this-worldiness appears to sanction the step of equating immanence with materialism. Modernity achieves its apogee in the powers of affirmation liberated by Spinozism, rather than in the deepening of the powers of reflexivity and self-consciousness liberated by Kantianism. Indeed, they complain that the 'relativity of experience' introduced by Kant 'abolishes every instance of the immediate and absolute in human life and history. Why, however, is this relativity necessary? Why cannot knowledge and will be allowed to claim themselves to be absolute?'
These words will seem strange to those coming from the Kantian tradition. Whilst the complaint is reminiscent of Hegel, the word 'immediate' suggests otherwise. Rather than raising the Kantian stakes as Hegel does, Hardt and Negri seem to retreat from them altogether. But, the post-Kantian might say, isn't it with Kant that the claim to immanence is first truly justified? The purpose of the Kantian critique is surely to ask how immanence is to be achieved, to ask how it is possible, and to secure it by right against the transgressions of theology and metaphysics. The ancient metaphysical idea of immanence must yield to the project of immanent critique. Hardt and Negri seem to suggest that immanence is something that can be immediately affirmed, without any prior investigation into its possibility. Things become odder still for the post-Kantian philosopher when Hardt and Negri suggest that although 'Hegel restores the horizon of immanence ... [this] is really a blind immanence', in which all activity is subordinated to a divine teleological order. Again, it is easy to see how from an Hegelian perspective it is Hardt and Negri's notion of immanence that is blind, in that they are not concerned with the critical questions of the justification of structures of knowledge and action that occupy Hegel in the Phenomenology and serve to secure the Hegelian right to absolute immanence.
In this essay I will claim that Deleuze's views on immanence are far removed from those espoused by Hardt and Negri, and in fact are much closer to the Kantian tradition than is generally suspected. I will also call into question Deleuze's apparent Spinozism regarding the question of immanence. Deleuze does hold that thought can immanently express being, but nevertheless he crucially holds to the Kantian distinction between thought and experience. This is also the key to situating Deleuze between Kant and Hegel: for Deleuze, to claim that the absolute is open to thought does not, as it does for Hegel, imply that it is open to experience.
This said, I will also suggest that if the word 'immanence' appears continuously throughout Deleuze's work, this is not because it is a sign of philosophical continuity, but because it designates the site of an enduring problem. When Deleuze finally comes explicitly to elaborate the notion of immanence in his late works, it has undergone radical change. This essay will take an eccentric path because it attempts to reconstruct and defend Deleuze's early approach to immanence, as opposed to his final views. Despite the absence of explicit discussion of 'immanence' in his magnum opus Difference and Repetition, I claim that it is there that we find Deleuze's most defensible formulation of a new philosophy of immanence.

Deleuze, Hyppolite and Hegel
In 1955 Deleuze wrote a review of his teacher Jean Hyppolite's book Logic and Existence in which he both makes clear how much he accepts of Hyppolite's reading of Hegel and provides the only published plan, to my knowledge, in which he lays out the aims of his future philosophical project. Deleuze begins by saying that Hyppolite's main theme is that 'Philosophy must be ontology, it cannot be anything else; but there is no ontology of essence, there is only an ontology of sense'. He adds 'that philosophy must be ontology means first of all that it is not anthropology'. Let us first unfold Hyppolite's interpretation of this notion of sense.
The use of the word 'sense' (Sinn) does not seem especially central in Hegel's own work, but Hyppolite makes clear that he is identifying it with the more familiar 'notion', or 'concept' (Begriff). Why does he do this? While there is undoubtedly a Husserlian inspiration at work, this move also draws out the sense in which the concept in Hegel is a philosophical reality, it expresses reality. Hyppolite cites Hegel's Lectures on Aesthetics:
Sense is this wonderful word which is used in two opposite meanings. On the one hand it means the organ of immediate apprehension, but on the other hand we mean by it the sense, the significance, the thought, the universal underlying the thing. And so sense is connected on the one hand with the immediate external aspect of existence, and on the other hand with its inner essence.

For Hegel these two opposite meanings signify a common source; they signify that the universal will be generated in the sensible; that the universal concept and the singular intuition are two aspects of the self-differentiation of the absolute. The intelligible articulation of the structure of self-differentiation is what Hyppolite will call sense, while the movement itself can be called expression. For Hegel, the problem with Kant's critique is that the concept remains too external to the thing itself: 'the categories are no fit terms to express the Absolute'. Moreover, the concept as such is never merely possible in Hegel. A Kantian possible concept (eg. of '100 thalers') is for Hegel not really a concept, but merely 'a content-determination of my consciousness'; that is, it is merely a representation. A concept, rather, is ultimately and intrinsically neither representational nor referential, but expressive of a reality. This couple sense/expression will be taken up by Deleuze. Both Hegel and Deleuze are against philosophies of representation because such philosophies claim to express what should be genuinely universal within a framework that remains relative to subjective representational experience (ie. which has only been justified anthropologically), so that the concept of expression doesn't ever gain its full extension, and thought is denied its rightful access to being. The notion of the thing-in-itself is symptomatic of Kant's contradictory position: he forbids himself to say anything determinate about it, yet insists that it has essential content for thought. Kant therefore is only partially aware of the transition to which he is midwife: 'from the being of logic to the logicity of being'. For Hegel, there will ultimately be nothing outside the concept: absolute idealism will transparently and immanently express every aspect of being. It is for this reason that Hyppolite says that 'immanence is complete' in Hegel.
Now, Hyppolite also gives primacy to the notion of sense because he wants to lay priority on the special character of the Logic in Hegel's system. For Hyppolite, the Logic is the expression of being itself; it is the high point of Hegel's system in which 'the concept, such as it appears in dialectical discourse, is [unlike in the Phenomenology] simultaneously truth and certainty, being and sense; it is immanent to this being which says itself'. Hegel's logic is a logic of sense, in which the sense of being itself is said through the genesis of concepts produced by the philosopher. Attempting to avoid the anthropomorphic view of Hegel promoted by Koj've, Hyppolite tries to restore the high metaphysical status of the Hegelian system. Hence, like Deleuze, his anti-humanism is an echo of the claims of classical philosophy. In an important passage for Deleuze, Hyppolite says that
Hegel is still too Spinozistic for us to be able to speak of a pure humanism; a pure humanism culminates only in skeptical irony and platitude. Undoubtedly, the Logos appears in the human knowledge that interprets and says itself, but here man is only the intersection of this knowledge and this sense. Man is consciousness and self-consciousness, while at the same time natural Dasein, but consciousness and self-consciousness are not man. They say being as sense in man. They are the very being that knows itself and says itself.

The implication of Hyppolite's reading here is that the phenomenological and historical parts of Hegel's system are anthropological entries into the system. Hyppolite is influenced by Heidegger's 'Letter on Humanism': man is the 'place', the structural possibility that Being can reveal itself as such, and express its sense through 'man'. After man has been broken down and introduced into the absolute by the Phenomenology, the Logic, absolved of humanism, retraces the ideal genesis of the sense of being. This would be the meaning of Hegel's statement that the content of the Science of Logic 'is the exposition of God as he is in his eternal essence prior to the creation of nature and a finite mind'.
In his review of Hyppolite, Deleuze affirms fully this reading of Hegel. Two passages are of particular importance. The first places Deleuze's development of the notion of difference explicitly within the context of Hegelian self-differentiation:
[T]he external, empirical difference of thought and being [in the Kantian system] has given way [in Hegel] to the difference identical with Being, to the difference internal to the Being which thinks itself.... In the Logic, there is no longer, therefore, as in the empirical, what I say on the one side and on the other side the sense of what I say ñ the pursuit of one by the other which is the dialectic of the Phenomenology. On the contrary, my discourse is logical or properly philosophical when I say the sense of what I say, and when in this manner Being says itself.

Deleuze will never leave behind this image of a 'properly philosophical' discourse. That is, his philosophy will be a philosophy of the absolute; it will accept the move from the perspective of the limitations of knowledge in Kant to the claim that dialectical thought can express the absolute and in turn ground knowledge. Deleuze shares none of the reservations about Hegelian immanence that are exhibited by his fellow post-war French philosophers. He has no bad conscience about the notion of immanence and he does not construct a philosophy of difference in order to subvert immanence (and introduce some notion of 'irreducible otherness' into it), but rather in order to fulfil it - precisely as Hegel does. Our problem will be to explain how and why Deleuze returns to elements in Kant to carry out this aim.
Deleuze concludes his review with some pregnant questions for Hyppolite after summarising the mains claims of the book:
Following Hyppolite, we recognise that philosophy, if it has a meaning, can only be an ontology and an ontology of sense. The same being and the same thought are in the empirical and the absolute. But the difference between thought and being is sublated in the absolute by the positing of the Being identical to difference which, as such, thinks itself and reflects itself in man. This absolute identity of being and difference is called sense.... The richness of Hyppolite's book could then let us wonder this: can we not construct an ontology of difference which would not have to go up to contradiction, because contradiction would be less than difference and not more? Is not contradiction itself only the phenomenal and anthropological aspect of difference?

We thus have four criteria laid out in 1955 for Deleuze's future philosophy. Firstly, like Hegel, he believes that Kantian critique must at a certain point be subordinated to a philosophical affirmation of the logicity of being. Secondly, he affirms that as the philosophy of immanence concerns the absolute, therefore all differentiation found in it will be internal, self-generated, differentiation. Thirdly, this philosophy must be able to 'say its own sense', and through this reflexive act, coincide with the sense of Being itself. Finally, we also have the suggestion that the absolute claims of Hegelian philosophy must be purified of dependence on phenomenal and anthropological content, and that this latter category, for some as yet unspecified reason, includes the concepts of contradiction and negation. The decisive problem for Deleuze's project will lie in consistently articulating the third criterion along with the others.
Now, if we look for an actualisation of this project, we appear to find it not in Difference and Repetition, but in Spinoza and the Problem of Expression, also published in 1968. It is in Spinoza that Deleuze finds the fullest flowering of an alternative model of immanent self-differentiation that remains faithful to the Hegelian schema, but which also presents a notion of difference without contradiction. However, the place of Spinoza in Deleuze's philosophy turns out to be extremely complicated, and he remains just as haunting and irresolvable a presence for Deleuze as he was for the work of the post-Kantians.

Spinoza and the 'best plane of immanence'
In the Spinoza book of 1968, Deleuze fashions a history of the philosophy of immanence, from the Neoplatonists through to Duns Scotus, which culminates in Spinoza. He also reaffirms in 1991 that it is Spinoza who sets out 'the 'best' plane of immanence'. I will claim shortly that the meaning of immanence has nevertheless undergone a radical shift between these dates.
Much of Spinoza and the Problem of Expression is concerned with the theological history of the notion of immanence. For Deleuze, Spinoza's contribution is to claim that there is no transcendent God, only a God immanent to nature, whose attributes must be conceived not as 'eminent' to natural attributes, but as 'univocally' sharing the same meaning. But once the theological issue of the identity of God with nature has been achieved in principle, one is still left with a set of purely ontological questions. How is the specific structure of this ontology to be defended? In what form will the nature of being express itself in thought? Why would Spinoza's philosophy be 'more immanent' than Hegel's for instance, when Hyppolite has given strong reasons for affirming that immanence only becomes truly 'complete' in Hegel?
We come to close to an answer if we follow Deleuze's attempt to enact a philosophical construction of absolute immanence in his reconstruction of the first part of Spinoza's Ethics. Deleuze presents an account of absolute difference that is formally coherent and provides a foil to the Hegelian view that difference is primarily negation, and that the self-differentiation of the absolute must be conceived in the form of a totality. I will only mention the gist of the argument here, as my aim is rather to assess its role and status in Deleuze's theory of immanence.
The first few propositions of the Ethics state that 'two substances having different attributes have nothing in common with each other' (E1P2), because an attribute is 'what the intellect perceives of a substance as constituting its essence' (E1D4), and a substance is 'conceived through itself' (E1D3). Substances, moreover, cannot be distinguished from one another by their 'modes', but only by their attributes. No substance can therefore be in a relation of limitation or causality with another. We thus start with a bare plurality of substances with one attribute, each of which has nothing to do with the other. Deleuze points out that it would be incoherent to introduce a unifying, eminent substance 'behind' these substances-with-one-attribute. This would be a merely 'modal' or 'numerical distinction', as it would presuppose a division between substances that share something in common. This would go against the definition of substance, which therefore requires a rigorous logic of 'real distinction'. The universality at work in this picture is distributive rather than collective; it concerns the 'each', rather than the 'all'. Spinoza's next big move is to argue that there can only be an absolute infinity of these really distinct substances-with-one-attribute. But in this case, the notion of 'substance' should really be resituated at the level of absolute infinity itself; therefore the framework is now reconceived so that there is one substance composed of the set of really distinct attributes. The attributes are univocally affirmed of the absolutely infinite substance; there is no transcendent genus or substance 'behind' them, to the extent that it is their univocal affirmation that constitutes their status as substance. Only the real distinction of the attributes, taken to infinity, dispels the need for an eminent unity, or a spurious collective totality of the components of the absolute. Only through this theory of 'real distinction', or pure difference, can Spinoza think absolute immanence, 'the absolute identity of Being and difference'.
At strategic points in the book, Deleuze appears to imply that all the aspects of Hegelian immanence are to be found in Spinoza: expression, the absolute, self-differentiation, genetic method. However, for the presentation of absolute difference to be more than formally coherent, Deleuze would need to commit himself to an account of the relation between the logical (or formal) and the real. Immanence must be realised. In an important phrase, Deleuze claims to have revealed “the only realised ontology'. Now Spinoza's version of the realisation of immanence fundamentally rests on a recapitulation of the traditional ontological argument ('it pertains to the nature of a substance to exist', E1P7). But will Deleuze himself rely on the ontological argument to fulfil the four criteria mentioned above for his own philosophy of immanence? There are three problems with this possibility.
1. Wouldn't Deleuze have to make more effort to defend this kind of ontological argument from well-known criticisms such as Kant's? For Kant, 'existence' cannot be predicated of the absolute in a formal argument, as to say that something exists requires an extra-logical moment (for instance the presence of an intuition). Now if Deleuze wishes to appeal to the expressivist theory of concepts mentioned earlier in relation to Hegel, then this would be circular, as the validity of that theory depends on a successful demonstration of an internal relation between being and thought. And while Hegel often speaks highly of the ontological argument, the weight of his theory of expression does not rest on a return to that argument, but on other more post-Kantian anti-skeptical arguments about the relation of thought and being, presented in the Phenomenology. Yet there is no Phenomenology at all in Deleuze, no 'introduction to the System'.
2. For Deleuze, the presentation of absolute difference is 'an immediate and adequate expression of an absolute Being that comprises in it all beings.' To cite a phrase Deleuze uses elsewhere, it involves a 'static genesis' of the structure of the absolute. Hegel's Science of Logic, on the other hand, performs a 'dynamic genesis' of 'the logicity of being' in such a way that 'it says its own sense' (accounts for itself through the concepts it has generated) through the very movement of thought presented step-by-step in the book itself. The Logic therefore enacts the complete and immanent interpenetration of the logic of being with the logic of thought. For instance, the movement from being to nothingness and then to becoming at the start of the Logic is simultaneously a movement of thought in which the bare thought of being reveals itself to be nothing determinate. Moreover, it is also through this approach that Hegel completes his response to the Kantian critique of the ontological argument: by arguing that the notion of bare 'existence' or 'being' cannot be conceived without introducing some determinacy into it: to be is to be something.
Now Hegel's articulation of the logicity of being is of course only made possible by the claim that difference must be fundamentally understood as negation. We know that Deleuze disagrees with this, but is the necessary consequence of this disagreement that he also has to give up on a determinate and genetic account of the development of thought? If so, then he will have concomitant problems defending his account of immanence against Hegel's. Hegel manages to generate a lot of determinate possibilities out of the structure of negation: it is hard to see what determinate possibilities can be strictly generated from 'difference in itself'. In the Spinozist account, there is no direct movement from the real distinction of the attributes to the position that thought and extension are two of these attributes.
3. Let us return to the issue of the 'immediate' genesis of absolute immanence. Can Deleuze's formal demonstration of absolute difference by itself present a criterion of absolute immanence that can serve as a standard by which to criticise other philosophies of immanence as failures? It is sometimes suggested that Hegelian immanence introduces an illegitimate transcendence by the mere fact of presenting an order for absolute self-differentiation, or by presenting this order as teleological (see the remarks of Hardt and Negri above). Although here a materialist impulse tends to confuse the argument (the animus being against any claim to hierarchy in the absolute), the idea seems to be that if only one appeals to the notion of immanence itself, as rigidly oppositional to transcendence, that is enough to dispel any spectres of God, teleology, etc. Now, such an approach does not answer the questions above concerning the realisation of immanence, which Hegel has arguably answered better. Nevertheless, might it not be possible to perform an initial theoretical affirmation of the structure of absolute difference that, by illuminating the mere formal possibility of a structure of difference that would avoid negation, opens the possibility of seeing reality in such a way? I believe this thought is definitely being ventured by Deleuze, but it is not clear that this is the path that could lead to 'the only realised ontology'. It is important to remember that Spinoza thinks he is demonstrating the structure of the absolute, and would be critical of any interpretation of 'affirmation' which suggested voluntarism. Spinozism is not a kind of inverted Pascalian wager by which one bets that a transcendent God does not exist. If absolute immanence is to be affirmed, it cannot be as a possibility, but as a necessity. And that requires that it defeat the other ontological possibilities.
We come here to a crossroads. On the one hand, it could be that the Spinozist argument is really a model of absolute difference that is put to work elsewhere by Deleuze in the service of another, more hidden, theory of immanence which will be able to compete with post-Kantian theories of immanence. On the other hand, it is equally clear that Deleuze did indeed go on to affirm the Spinozist theory of immanence as 'the best plane of immanence' in works such as What is Philosophy? Nevertheless, in the following passage it is clear that something has changed:
Spinoza was the philosopher who knew full well that immanence was only immanent to itself ... He is therefore the prince of philosophers. Perhaps he is the only philosopher never to have compromised with transcendence and to have hunted it down everywhere ... He discovered that freedom exists only within immanence. He fulfilled philosophy because he satisfied its prephilosophical presupposition. ... Spinoza is the vertigo of immanence from which so many philosophers try in vain to escape. Will we ever be mature enough for a Spinozist inspiration?

Firstly, the immanence/transcendence opposition is now taking on all the work. Moreover, this notion of transcendence is highly unusual in that it includes not only concepts of entities such as God, but even the notions of subject and object. As Deleuze elaborates in his last ever published article, the short opuscule entitled 'Immanence: A Life', both the subject and the object are not transcendental, but 'transcendent', whereas the field of immanence itself is 'an impersonal pre-reflexive consciousness, a qualitative duration of consciousness without self'. Here Deleuze in fact appeals to the later Fichte, and he seems very close to the philosophy of pre-reflexivity found in Fichte by Dieter Henrich in his seminal article 'Fichte's Original Insight'. However, the suggestion that 'immanence is related only to itself', yet must be considered to be pre-reflexive is a difficult one, as how is the 'self-relation' supposed to be justified if it has no intrinsic connection with reflexive self-consciousness?
This leads us to the second change: immanence has become a 'pre-philosophical presupposition'. Now, this move towards a late-Fichtean position has two major consequences for Deleuze's project. Firstly, the apparent embrace of a featureless form of intellectual intuition raises problems with the continuing philosophical affirmation of 'difference' and 'multiplicity'. As we will see, Deleuzian 'dialectical difference' was elaborately and determinately worked out in Difference and Repetition in a way that is antagonistic to any reliance on some source of primal 'indifference'. Secondly, Deleuze can no longer claim to have found 'the only realised ontology', because such a philosophy of immanence could never be realised; its pre-reflexivity precludes this. Thus we come to the conclusion that Deleuze's late affirmation of the Spinozist notion of immanence occurs at a huge cost: immanence is now a 'presupposition' that must be 'pre-philosophically' affirmed. And this surely amounts to a return to Fichte's criterion, that it depends on the kind of person one is whether one accepts this version of things.
I have said that in Spinoza and the Problem of Expression, immanence genuinely appears to be a matter of philosophical construction. I ventured that Deleuze's static genesis of absolute difference could provide a model for the construction of immanence itself. What was needed was an account of its critical validity in relation to other philosophies of immanence. The materials for this are present in Difference and Repetition.

Immanence and Ideas in Kant and Deleuze
It is Deleuze's return to Kant in Difference and Repetition that provides the most powerful approach to a new philosophy of immanence. Kant's own 'plane of immanence' could be said to have two aspects. Firstly, the implication of the whole project of a 'Critique of Pure Reason' is that reason can perform a critical operation upon itself - an immanent critique. However, exactly how this reflexive act is to be accomplished is not clear. Kant at first seems to envisage that there is a pure element of reason that has 'its own eternal and unchangeable laws' and is a 'perfect unity' and that therefore provides the necessary vantage point for an auto-critique of human experience. However, since the thrust of the first Critique is precisely to show the dependence of reason on the other features of cognitive functioning (such as sensibility and the understanding), Kant makes it clear at the protracted end of the work that the 'unity of reason' must be considered rather as a 'single supreme and inner end, which first makes possible the whole'. That is, the fulfilment of an immanent critique systematically requires the teleological projection of an actualised unity of the diverse aspects of cognition. It turns out that the work of the Critique of Pure Reason is to be part of a metaphysics, which 'is also the culmination of all culture of human reason'. Metaphysics in turn is a part of 'philosophy', which is 'the science of the relation of all cognition to the essential ends of human reason (teleologia rationis humanae)'.
The second aspect of Kantian immanence is much more well-known. Kant's method of transcendental argumentation secures an enduring restriction upon all the faculties and features of cognition so that they can only be legitimately used if they conform to the structure of experiential cognition. That is, their immanent use is justifiable, but their transcendent use is shown to be illegitimate. Kant's main use of the term 'immanence' is in fact with regard to the immanent use of the faculties of cognition.
Two related questions are relevant here. First, the procedure of the self-critique of reason and the restriction produced and consolidated by that procedure are related in a mysterious way. The latter is by right the result of the former, but the former is the most obscure. If the wider method of the self-critique cannot be justified, then how can Kant say that he has strictly drawn the line between legitimate and illegitimate cognition? Second, it appears that Kant is guilty of using the notion of 'reason' equivocally. Reason acts both as the subject and object of critique, without it being made clear how reason (as subject) could save a bit of itself from its involvement with the other faculties of cognition (in its role as object of critique). These metacritical issues are encountered in one way or another by the post-Kantians, but the Deleuzian take on them is quite specific, and perhaps closer to Kant than the post-Kantians were prepared to go.
Kant's notion of immanent critique seems to involve an unstable oscillation between noumenal and teleological claims. In the first edition Critique Kant appears to affirm some kind of cognitive access to noumena, for instance in the section on noumenal freedom where the human being is said to be 'one part phenomenon, but in another part ... a merely intelligible object'. This echoes the distinction in the 'pre-critical' Inaugural Dissertation between 'things thought sensitively ... as they appear, while things which are intellectual are representations of things as they are.' Nevertheless, as Kant elaborates his system (particularly under pressure of his development of the theory of inner sense, and of problems in the 'deduction' of freedom), he begins to shift all the metacritical weight of reason's power to criticise itself onto systematic teleology. The claims about the 'culture of human reason' are expanded in the Critique of Judgment, where the functions of experience and knowledge themselves are more explicitly tied up with purposive activity (for instance through the development of the notion of 'reflective judgment').
Now in his philosophical works of the 1950's and 60's, Deleuze too appears to appeal both to some kind of noumenal access, and to a teleology of the cognitive faculties. On the one hand, Deleuze often comes across a high rationalist. He argues in 1956 that it is only by 'determining the differences in nature between things Ö that we will be able to 'return to things themselves' ... If philosophy is to have a positive and direct relation with things, it is only to the extent that it claims to grasp the thing itself in what it is, in its difference from all that it is not, which is to say in its internal difference'. With its quasi-Hegelian appeal to 'internal difference', this desire to 'return to things themselves' is by no means an echo of the trusted phenomenological maxim: on the contrary, Deleuze appears closer to resurrecting the rationalist project of returning to noumena. Elsewhere, Deleuze writes of attaining a 'truly sufficient reason' which will enable us to determine things in themselves in their internal difference.
On the other hand, Deleuze is concerned in all of his works up until Difference and Repetition with the notion of teleology. Kant's Critical Philosophy is an explicitly teleological reading of the structure of Kant's system. In an article on Kant's aesthetics from 1963, Deleuze writes that 'in the Critique of Aesthetic Judgment, Kant poses the problem of the genesis of the faculties in their primary free accord. He discovers an ultimate foundation, which is lacking in the other Critiques. Critique in general ceases to be a simple conditioning, to become a transcendental Formation, a transcendental Culture, a transcendental Genesis'. It is at this point, however, that we can locate a crucial development of the Kantian position. In Difference and Repetition, Deleuze attempts to push further the theory of the 'ends of reason' by reconstructing Kant's theory of Ideas of reason, so that the concepts of the understanding are seen to depend fundamentally on the orientation of cognition towards Ideas. My claim in what follows is that Deleuze fuses the noumenal and the teleological in his new notion of 'Idea', in such a way that he can legitimately claim that thought has access to noumenal being (while experience, understood in terms of recognition according to the generality of concepts, does not). This achievement of the immanence of thought to being, however, is achieved critically in Deleuze, rather than metaphysically, as in Kant.
To proceed it is necessary to bring out the general teleological structure of cognition present in Kant's work right from the first edition Critique. The basic aim of the Transcendental Deduction of Categories is to discover an apriori structure that grounds the connection between concepts (as 'functions of unity') and the sensible manifold. It is now recognised that the argument of this Deduction continues well into the 'System of the Principles of Pure Understanding'. However, I would claim that the argument extends even further, right into the further reaches of the Transcendental Dialectic. In fact, it is precisely here that the general task of the Transcendental Deduction meets up with the metacritical status of the Critique, in the teleology of pure reason. Kant in fact is clear about the general importance of Ideas for the basic activity of cognition in the first edition Critique when he suggests at length that a third Deduction ñ a Transcendental Deduction of Ideas - is also necessary. While the apriori forms of the understanding are often taken to be sufficient conditions for the 'coherence' of experience, Kant himself argues directly against such a view. Just as the Deduction of Categories was a response to the possibility that spatiotemporal 'appearances could after all be so constituted that the understanding not find them in accord with the conditions of unity', presenting a mere rhapsody or 'confusion' of sensations (the crucial passage at A90/B123), so does Kant admit that it is conceivable that 'among the appearances offering themselves to us there were such a great variety ... of content ... that even the most acute human understanding, through comparison of one with another, could not detect the least similarity'. Kant now appeals to reason to finally ground the applicability of concepts to experience, and to ground the coherence of concepts in judgments in general. 'For the law of reason to seek unity is necessary, since without it we would have no reason, and without that, no coherent use of the understanding, and lacking that, no sufficient mark of empirical truth'. Kant says that the understanding presents only a 'distributive unity' among appearances, without granting a 'collective unity'. It is only by projecting a 'horizon' or guiding totality that the analytic unity of concepts can be used logically, in such a way that higher and lower 'functions of unity' converge with each other. This would fulfil the fundamental requirement that is at the root of the Transcendental Deduction of the Categories. This horizon, says Kant, must 'direct the understanding to a certain goal respecting which the lines of direction of all its rules converge at one point'.
However, obviously the collective unity (or totality) of appearances, as a 'focus imaginarius' is precisely what can never be experienced as such, so the principle can only be regulative, not constitutive; that is, it is an Idea. Nevertheless, Kant insists that Ideas legitimately project a logical world, a mundus intelligibilis, of complete representation. In fact, the Idea has an anomalous transcendental status: on the one hand it is a peculiar kind of 'problematic concept', which itself does not conform to the usual criteria for concepts (it is not related to an intuition, nor does it serve as a tool for recognition). On the other hand, it is a transcendental condition: it is thus a condition of the possibility of unity in a concept; it gives unity to a concept, by acting as the horizon in which unification can occur. Ideas themselves cannot be known (one cannot know God, or the self, etc.), but they are necessary conditions for the coherence of concepts (and therefore of knowledge and experience).
Two problems arise for Kant. First, how can Ideas be both particular concepts and conditions of concepts in general? Second, while the first stages of Kant's critique demonstrate the constitutive role of pure forms such as the categories and space and time, to go on to affirm the transcendental necessity of the Ideas involves affirming the necessity of something unconditioned. But what grounds this claim? How can this teleology be justified in such a way that it does not merely depend once more on a noumenal postulation about the “essential ends of reason', or the structure of conceivability in general? On the other hand, if the ends of reason are merely “regulative' for finite minds, then how can this teleology be related to the teleology of reason necessary for the self-critique of reason itself to be possible? The weight Kant places on the 'outer limits' of the critique, on teleology, reason and the Ideas, is in danger of producing an implosion in the critical structure.
Deleuze finds a way through these problems precisely by exploiting the Kantian discovery that Ideas must be different in kind to concepts. Kant was onto something when he implied that Ideas are not themselves unified or objects of recognition. Deleuze's ingenious move is to take a peculiarly literal reading of Kant's statement that Ideas are 'problematic'. If Ideas are complete determinations, but concepts are general, then Ideas are problematic because they do not withstand coherent generalisation: this is their quality, that they cannot be recognised or experienced. Nevertheless, they are in principle open to thought, as the necessary horizon of complete determination. Not only this, they are also essential to motivate knowledge at all.
The fact is that [reason] alone is capable of drawing together the procedures of the understanding with regard to a set of objects. The understanding by itself would remain entangled in its separate and divided procedures, a prisoner of partial empirical enquiries or researches in regard to this or that object, never raising itself to the level of a 'problem' capable of providing a systematic unity for all its operations ... [it] would never constitute a 'solution'. For every solution presupposes a problem .

This is really an echo of Kant's theme in the Preface to the second edition Critique that
reason ... compel[s] nature to answer its questions ... Reason, in order to be taught by nature, must approach nature with its principles in one hand, according to which alone the agreement among appearances can count as laws, and in the other hand, the experiments thought out in accordance with these principles.

Knowledge itself is preceded by the posing of questions, that is, by thought. Knowledge should not be understood as simply involving descriptions of states of affairs according to rules; rather knowledge concerns solutions to problems. Therefore, established knowledge, or what permits recognition, is really nothing but the realm of established solutions.
Kant does not spell out explicitly this difference in kind between Ideas and concepts. For him, one of the main criteria for the problematic 'horizon' is that it be unified. But is this a relevant criterion for the structure of problems? The criterion of unity is strictly speaking a function of the understanding. Concepts are 'functions of unity' and empirical cognition or knowledge is the locus of 'unification' through concepts. Kant is therefore presupposing the projected unity of Ideas only as a telos from the standpoint of knowledge, that is, from empirical representation. The power of Ideas is understood in terms of logical representation, in terms of a logical calculus that can only be a pale reflection and amplification of the realm of already established empirical concepts. However, if Ideas are to be thought primarily as problems (according to Deleuze's literal reading), this implies that they must already have their own consistency and form as problems that stand structurally outside achieved empirical knowledge, 'feeding' and conditioning knowledge. Any empirical knowledge is only 'determined by the conditions of the problem, engendered in and by the problem along with the real solutions. Without this reversal, the famous Copernican revolution amounts to nothing'.
Deleuze proceeds to argue that Ideas can be conceived as already possessing the power to synthesise difference in themselves. Again, this thought is familiar from Hegel: the Kantian dialectic is taken by Hegel to be the clue to the real extra-representational structure of the determinable world, a structure which lies beyond the 'concept' in the Kantian sense. Deleuze too is content to use the word 'dialectic' to describe the specific mode of differentiation for Ideas; Deleuze's account of problems is said to explore 'the dialectical half of difference'. Also like Hegel, Deleuze believes that Kantian 'complete determination' is conceivable at the level of thought (if the correct means are used), even if it is not 'experienceable' as such by a finite being. Complete determination is reconceived by Deleuze as the ideal determination proper to a problematic field. However, contra Hegel, he excludes a dialectics of negation as the correct means to undertake an exhaustive determination of the Idea. As mentioned above in the first section, Deleuze believes that the form of contradiction is a 'merely phenomenal' aspect of difference itself. What can this mean?
Again one returns on the rebound from Hegel to Kant. For Kant, although concepts are 'functions of unity' in judgments, synthetic judgments are perpetually amplifying concepts, revising them according to the problem or Idea according to which they are 'focussed'. As a result, concepts are ultimately indefinable. The principle of contradiction in fact refers only to concepts that have already been established and given preliminary definitions, and serves as a rule of unity within experience. But due to the de jure immersion of the concept in the problematic field, in which established concepts and definitions can be broken down and reformed once a problem becomes transformed, the principle of contradiction has only relative significance. Hegel can thus with some justice be said to have failed to plunge deep enough into the nature of difference in the absolute. Instead, for Deleuze the Idea is determined according to a logic of structure, in which contradiction between terms that actualise the structure should not be confused with the relations and transformations set out in the structure itself. If the structure is taken purely in its 'pre-actual' state, as a set of ideal transformations, in which the elements are subject to reciprocal determination, then the contradictions that might arise between the actualised elements and relations remain undecided or unselected. In this pure state, of course, the problem can only be thought, not experienced, precisely because experience functions by means of conceptual recognition.
Such problematic structures may apply to particular fields of knowledge and experience, or may ground the question of what counts as knowledge itself. As an example of a particular structure, Deleuze sometimes refers to the Lacanian school's theories of psychic structure. Take the Oedipus complex: there are a number of possible positions in the structure (mother, father, female child, male child) which can be occupied ('identified' with) in various ways, and thus can become caught in various vectors of desire. The Oedipal structure 'itself' cannot be experienced, although it can be completely determined. If the identifications break down, pathology may ensue, as in Dora's case. Dora may begin to experience her identity as a 'problem', oscillating between subject-positions. While fantasy and dream may be able to give form to and sustain the transformations of thought, the introduction of the problematic field into experience itself, bound by the rules of conceptual recognition and a particular spatiotemporal structure, can only be deeply destabilising, in Kantian terms a 'transcendent' exercise of one's faculties.
Such problematic structures must also extend to the most abstract philosophical levels. The criteria for knowledge itself are set up in response to the 'problem' of knowledge. Again, these criteria themselves cannot be 'experienced' or 'known', and the philosophical exploration of a problematic field cannot itself be judged by the standards of knowledge, as it sets those standards.
It is the sense of the destination of cognitive activity in a horizon that is to remain by right problematic that marks the singularity of Deleuze's extension of the teleology implicit in the Kantian Copernican turn. For Deleuze, indeed, the result of transcendental philosophy will not primarily be the dictum that all philosophy must conform to the conditions for the possibility of experience, that is, enact the immanent use of the structures of experience - in fact, Deleuze encourages their transcendent use or exercise [exercice], as it is precisely this that will critically reveal the limits of experience. For Deleuze, all activities, both voluntary and involuntary, in which thought becomes caught up in a problematic field which undermines the structure of experience, go under the name of 'transcendental empiricism', a phrase which is analogous to the Hegelian notion of 'speculative experience'. Hegel's view that the critical apprehension of limits requires that they be transgressed is thus taken up in a new way by Deleuze. As is the case for Hegel, Deleuze's notion of immanence actually requires the transcendent use of the faculties, and the activity of thought beyond experience. But unlike for Hegel, experience never becomes fully reconciled with thought. This allows Deleuze the space to develop a new, non-Hegelian 'logic of sense' (Hyppolite's phrase), which attempts to express the paradoxical act of thinking problems. In The Logic of Sense Deleuze elaborates on the ability of problematic thought to perform an 'ideal genesis' of its own conditions, and thus to 'say its own sense'.
It is clear that Deleuze's potentiation of Kantian Ideas therefore involves an inversion of Kantianism. It is no longer that the empirical use of Ideas is a transcendental illusion; rather it is our attempts to apply the rules of conceptual representation to problems and Ideas that is the real transcendental illusion. For here, representation transgresses its own limits and treats problems as concepts. Kant had misinterpreted what he discovered: the real illusion is to interpret Ideas as concepts which lack an intuition, and not rather according to the specific logic of problematic, complete determination. Kant's claim that the realm of Ideas was ordered in the form of a purely logical world of representation is in fact an uncritical presupposition, which Deleuze critically rectifies.
Given the destination of cognitive thought in the Idea, the only choice for the critical philosopher is to univocally affirm problematicity as such. But what form can this take? It is at precisely this level that the Spinozist argument for absolute difference finds its true place. Absolute difference is shown to be formally coherent in the Spinoza book, but its existence could not be assumed without recourse to an ontological argument. As we saw, the procedure of 'starting' with absolute immanence risks falling back into 'pre-philosophical presupposition'. But in fact, absolute immanence lies at the 'end' of the system, rather than its beginning: it is the telos towards which cognition and critique move, and which must be philosophically affirmed. Now, the demonstration of the formal coherence of the thought of absolute difference gives us the right to replace the Kantian collective horizon, in which all Ideas converge in a presupposed unity modelled on the concept, with a truly, intrinsically differential horizon, whose only foundation is absolute difference without unity. Reason itself can be remodelled ('a truly sufficient reason'): it is no longer be immediately considered to 'seek unity'. From the ideal notion of collective unity we move to a permanently distributive structure of reason. And while the Kantian 'common horizon' is shattered, chaos or indeterminacy does not ensue; rather the splinters can assume a new formation.
This philosophical affirmation of 'the absolute identity of Being and difference' provides Deleuze with a novel ontological position between Kant and Hegel. For Kant, Ideas are merely problematic, 'merely ideal', while for Hegel, the dialectical Idea is fully actual. However, for Deleuze, Ideas are essentially problematic in themselves. Like Hegel, Deleuze will affirm that there is no noumenal reality that cannot potentially be captured by dialectical thought. Thought can indeed fully express being ñ but (contra Hegel) only through a (non-conceptual, non-negative) form of differentiation that remains intrinsically problematic for experience. Between Kant and Hegel, Deleuze's claim is that Ideas, as problems, are constitutive. That is, they are univocally affirmed of being itself, against the equivocity of Kantian reason.
So why does Deleuze insist that 'immanence is the very vertigo of philosophy'? There are perhaps both manifest and latent answers in Deleuze's work. The manifest answer is that immanence is the telos of reason, which in its full differential and dispersive form, can only signify the undermining of experience on the part of reason. The latent answer invokes structural limits within the very notion of immanence. Since Deleuze's account of absolute difference does not allow for an immanent unfolding of determinate categories (in the way that Hegel's theory does), he must instead take a more crooked path to immanence, involving a complex mixture of transcendental (Kantian) and formal and ontological (Spinozist) argumentation. In other words, it is because Deleuze attempts to construct an immanent theory of difference which escapes the forms of negation and the concept that he must sacrifice the self-generating and self-validating features of Hegel's system of immanence, features that make it not only a philosophy about immanence, but a philosophy that demonstrates at every step its own immanence in its very writing and being read. How, then, is one to adjudicate between Deleuze's and Hegel's systems? Perhaps this question is closer to the 'vertigo of philosophy' Deleuze really had in mind, which may explain his attempts to move beyond his early system. The vertigo would be latent in the problematic notion of immanence itself.