In the preface to the English edition of Cinema 2, Deleuze claims that cinema is a repetition, in speededup form, of an experience that has already occurred in the history of philosophy.  This notion of repetition recalls the biological notion of the ‘recapitulation’ of phylogeny in ontogeny: individual development recapitulates, or replays in speeded-up form, the development of the species. Haeckel noted that this recapitulation was strongly in evidence at the embryonic stage, so that one can see the human embryo at a certain point appearing to be on the verge of developing a tail that subsequently disappears as the embryo develops. So, on this metaphor, cinema, an apparently new and unprecedented phenomenon in the modern world, nevertheless only develops through recapitulating an arduous development already undergone elsewhere. Now Deleuze’s claim is that cinema recapitulates a movement already undergone in philosophy. Why philosophy, and not visual art, or some other discourse, or perhaps the history of civilization in general? Why does cinema recapitulate a historical passage in the life of the mind?
Deleuze says that the development in philosophy that cinema recapitulates concerns the nature of the notion of time from the Greeks to Kant. Whereas philosophy before Kant thinks of time in relation to movement, Kant subordinates movement to time.  Before Kant, the world was seen as made up of changing, moving bodies, and time referred to our way of measuring rates of change in the physical world. The notion of time was thus subordinated to the demand for measurement of moving bodies. For instance, in the Aristotelian world-view, time is secondary to the general cosmic movement from potentiality to actuality. In the Christian world-view, there is an eternal order opposed to a temporal realm, where time is fundamentally referred to the end of the world, or apocalypse. Deleuze also has in mind cyclical conceptions of time based on the passage of the seasons. In all these cases, time is subordinated to an already given movement of the physical world. Kant, on the other hand, inaugurates modern thinking about time. Kant makes time the transcendental condition of all of our experience, so that it is the structure of time itself, as stretched out, projected and synthesized by a human subject, that in the ﬁrst place conditions our experience of moving bodies, and not vice versa. So time conditions movement. As we will see, however, Deleuze has an unusual reading of Kant’s conception of time, and his ultimate aim is to bring to light ‘a precise moment within Kantianism, a furtive and explosive moment which is not even continued by Kant, much less by post-Kantianism’,  the consequences of which nevertheless reverberate within modern philosophy as well as outside it, in domains such as the cinema. Deleuze’s contention is that we have still not fully realized the consequences for our conceptions of subjectivity and selfhood of the endless, merciless line of time uncovered in its purity by transcendental philosophy.
How might this relate to cinema? What is the simplest deﬁnition we can give to cinema? We can say at least that the fundamental unit of cinema is the moving image. Cinema is composed of images which move, or self-moving images.  Deleuze is suggesting with his ‘recapitulation thesis’ that cinema develops in two main phases. In a ﬁrst phase, time is subordinated to movement. Cinema thus operates with movementimages, and recapitulates traditional ideas about time. Deleuze’s privileged example here is Eisenstein, who develops a form of montage able to express the dialectical totality of the world. In the second phase, cinema arrives in philosophical modernity and comes to terms with time itself, not just with movement. Deleuze’s privileged examples here are Welles, Resnais and Robbe-Grillet (Last Year in Marienbad is the ﬁlm Deleuze constantly returns to when expounding the dimensions of the time-image) and Godard. The development of cinema thus recapitulates in image form the path leading up to a fundamental moment in philosophical modernity – the realization that time is the condition of the world, that it has no beginning and end, and we are at the mercy of it. Cinema for Deleuze is possessed of a singular power in that not only is it a fundamentally temporal art form, but it is always potentially a mass art form as well, and thus is in a perfect position to crystallize a nascent human coming-to-consciousness of the fundamental character of time in the post-Kantian world.
We should comment on the justice of this apparently entirely philosophy-centric view of the cinema. Is Deleuze’s claim, then, that cinema is a kind of spatiotemporal incarnation of ideas that have their pure form in philosophy? What would it mean to answer ‘yes’ to this question? On the plus side, if cinema is the spatio-temporal incarnation of a set of ideas about space and time, doesn’t that mean that cinema, rather than being parasitic upon philosophy, assumes a powerful autonomy as a realization of philosophy? It would complete philosophy’s speculation by realizing it in practice. So what philosophy gives to cinema, it gets back by realizing itself in more concrete form. However, this may seem to many to give philosophy a ridiculously exaggerated role in the internal logic of the development of cinema. So Deleuze qualiﬁes this idea a little. If cinema in its second phase confronts time in all its purity, and overcomes the traditional ideas about time as movement that were holding it back, this moment is triggered by a speciﬁc set of socio-historical conditions. Speciﬁcally, cinema only enters its second phase after the Second World War.
The new cinema records the ruins of the old world, and depicts characters who can no longer rely on traditional, habitual ways of life, who can no longer react in the way they used to. The period after the Second World War is also marked by a new phase of capitalist development: not only are people uprooted or deterritorialized from their traditional forms of life (as in the ﬁrst phase of capitalism), but their desires are now manipulated and deterritorialized by the new consumer society. Not only are old ways of living and working abolished, but people’s interior lives, their very desires, are deterritorialized. Western societies become radically cut off from their past. We enter a new phase of history, governed by the tendency towards absolute deterritorialization. It is these social conditions that allow the Kantian theory of time to become relevant for everybody. And cinema is the privileged place where we can become spectators of the process of this transformation. The darkened space of the cinema auditorium, populated by bodies whose sensory-motor life is suspended along with their social being, provides the ideal space for the unfolding of what Deleuze calls ‘the pure form of time’, a form of time in which the temporal syntheses of memory and anticipation are permitted to detach themselves from their ballast in everyday active social experience.
This is the strong central thesis that undergirds Deleuze’s Cinema. It implies an evaluation, as it implies that ﬁlms which remain caught up in mere movement-images must be seen as outmoded. It also has an ethical component in that it shows that the great modern directors were attempting to come to terms with, and imagine ways of dealing with, life in a world with a profoundly new temporal structure. Deleuze’s Cinema is thus a great progressive work of aesthetics. But we must note it was written in the early 1980s – that is, in what perhaps now looks like the twilight of the great age of European cinema. So perhaps here as well the owl of Minerva only ﬂies at dusk.
In this article I explore two things. First, I try to explain what Deleuze means by this ‘pure form of time’ that he holds to be essential for philosophical modernity and for the cinema. I will do this by referring back to the philosophy of time Deleuze sets out in his earlier works, notably Difference and Repetition (1968), where we ﬁnd Deleuze claiming that it is Hölderlin and Kierkegaard who have most clearly seen the consequences implicit in the Kantian account of time. It is notable that the Cinema books (1983 and 1985) mark a return for Deleuze to some of the most intriguing aspects of his earlier work, work that had apparently been buried following Deleuze’s earlier collaborations with Félix Guattari on Anti-Oedipus (1972) and A Thousand Plateaus (1980). The Cinema books develop and clarify the theory of time at the core of Difference and Repetition, underlining the ongoing interest of the latter work.
Second, I give Deleuze’s argument a new presentation. His claims about the modern form of time in cinema and philosophy rest on a difﬁcult but potentially powerful reading of the relationship between the Kantian subject and time. In Difference and Repetition Deleuze’s strategy is to generate from the bare possibility of repetition in time three different levels of temporal synthesis (habit, memory and thought), arriving at the Kantian account only at the end. Deleuze seems to follow Kierkegaard in allotting an unexpectedly large philosophical role to the notion of repetition. In one text, Kierkegaard states that repetition is the key to formulating the basic ‘collision between ideality and reality’, because any moment of cognition is a re-cognition, a repetition of something that has been before.  Deleuze’s ‘ideal genesis’ of the types of repetition, however, beginning with the ‘passive syntheses’ of habit and memory, lays itself open to criticism, partly due to his reluctance to declare the details of his method. It is hard to escape the thought that the accounts of habit and memory are fundamentally psychologistic.  But even if this is not true, there are other good reasons for trying to invert Deleuze’s actual procedure, and begin instead with his account of the synthesis of thought. Such a reading has an advantage, in so far as it places Deleuze’s thought more squarely in the same normativist space (i.e. the level of thought) inhabited by most thinkers in the Kantian and postKantian traditions today. In this way, the argument can also be reformulated along the lines of Kant’s own Transcendental Deduction, upon which it is generally agreed the success of Kant’s Copernican revolution turns. Kant’s Deduction starts out from a normativist account of the ‘I think’ that must be able to accompany all my representations, and then works from there to a synthetic and a priori account of the relation of thought and intuition. Might it not be possible to trace a similar ‘Deduction’ in Deleuze’s work, starting with his claim that Kant is the philosopher who truly ‘introduces time into thought’?  Thus I sketch out an alternative Deleuzean Transcendental Deduction which also leads from the ‘I think’ and arrives, if not at the pure form of intuition in general, then at the pure form of time. On that basis, I then conduct a regressive argument that ﬁlls in further transcendental conditions for temporal experience by appealing to Deleuze’s interpretation of Bergson’s theses about memory.  Although the path and the result are radically different to Kant’s own Deduction, the reconstruction may cast light on what can truly claim ‘transcendental’ status in Deleuze’s theory of temporal synthesis.
There is another reason for proceeding in this way.
For, as we have noted, in the Cinema books, we ﬁnd Deleuze also ascribing a historical dimension to his notion of the pure form of time. The claim is that cinema has provided a space for the contemplation of a profound restructuring of our temporal structure, a reconﬁguration that was ﬁrst mooted in the philosophy of Kant. If there is any truth in this, that means that we have been undergoing, or living, this reorganization as the form of historical actuality, albeit in the alienated form of the image. The embryonic movements taking place in the darkened space of the cinema auditorium have turned out to be the harbingers of a new form of subjectivity which will no longer be able to be unaware of its own internal, temporal structure.
Deleuze’s transcendental arguments must therefore be reoriented to take account of the consequences of their being brought to the surface. The article concludes by bringing up some possible consequences of this interpretation for cinema today.
the fractured ‘i’: kant and the pure form of time
We know that Kant has a number of theses about time. I will mention three, and then show how Deleuze goes on to exploit a paradox that emerges from their conjunction. First, Kant claims that ‘time is the a priori condition of all appearances in general.’  Not all events must appear through outer sense (there are non-spatial mental events), but all events must appear through inner sense, and time is the form of inner sense. Time is thus a substratum (as persistent form of inner intuition) [in which] both simultaneity and succession can alone be represented. The time, therefore, in which all change of appearance is to be thought, lasts and does not change; since it is that in which succession or simultaneity can be represented only as determinations of it. 
However, ‘time cannot be perceived by itself’, nobody has ever ‘met’ time, it is the form of appearance, not itself an appearance. If we wish to represent permanence, therefore, we are obliged to ﬁnd a substitute among the objects of perception, which can serve as a ‘substance’ upon which to pin properties.  Second, for Kant, this time can only be conceived as single and inﬁnite.  There is no beginning or end to time itself.
Third, Kant claims that our experience is made up of temporal syntheses. We make sense of our sensible experience through the rule-governed use of concepts, which are ordered implicatively in judgemental and inferential networks, by what Kant calls the understanding. But judgements and inferences require a unity of consciousness in order to function coherently: a transcendental subject. Because Kant has excluded the possibility of epistemic realism, he claims that the order (or objectivity) of experience can arise from nothing other than the synthetic activity of combining such representations. Thus there is only coherence and unity in our experience because we actively relate, via normative rules, our past experiences to our present ones, and anticipate on that basis our future experiences. The unity of temporal experience thus depends on a subject, a transcendental ‘I think’, characterized by the spontaneous capacity for apperception or reﬂexivity, which does the work of synthesizing temporal representations.Before we encounter the ‘paradox’ that Kant discovers when he goes on to relate the ﬁrst two theses abouttime to his account of the subject, let us look a little closer at the latter. Kant famously says that ‘The I think must be able to accompany all my representations.’  As Henry Allison comments, this identical ‘I think’ does not take the form of some sort of numerically identical ‘Cartesian’ consciousness; rather, all that we need to be aware of is ‘the “fact” that this identity must be presupposed as a necessary condition of possibility.… What this principle really asserts is the “necessity of a possibility”.’  However, following Robert Pippin, Allison has emphasized that Kant’s claim that the unity of consciousness requires a consciousness of unity also implies a ‘consciousness of synthesis … which is an essential ingredient in the ﬁrst-order activity’ of judgement.  As Pippin puts it,
In any, say, remembering, thinking or imagining, while the object of my intending is some state of affairs or other, I am also aware as I intend that what I am doing is an act of remembering, thinking or imagining, and that I bring to these acts a subject identical with the subject of prior acts of intending. 
Every act of judgement is at least implicitly reﬂexive, in the same way that playing a game requires rules, for in games ‘I can consciously follow a rule without consciously applying a rule.’  Such rule-following is still normative, in that the subject takes itself as able in principle to give an account of what it is doing, and thus in some sense as binding itself to the rules.
Let us concur with this strict ‘spontaneist’ or ‘activist’ interpretation of the Kantian subject, in order to highlight Deleuze’s move all the more starkly, which turns upon the paradox that Kant discovers when attempting to give a full account of the activity of the subject. The paradox is that if inner sense is the form of all appearance, then ‘this presents even ourselves to consciousness only as we appear to ourselves’.  But how can this apply to the transcendental ‘I think’, which after all is the transcendental condition for there being order among temporal appearances? Kant writes that I do not have yet another self-intuition, which would give the determining in me, of the spontaneity of which alone I am conscious, even before the act of determination, in the same way as time gives that which is to be determined, thus I cannot determine my existence as that of a self-active being, rather I merely represent the spontaneity of my thought, i.e. of the determining, and my existence always remains only sensibly determinable. 
Pippin takes the penultimate clause (‘rather I merely represent the spontaneity of my thought’) to indicate that ‘I do not have an “intuition” of my active “determining” capacities … but I am “conscious of the spontaneity of it”.’  Thus Kant is claiming that although we can only experience ourselves and our thinking activity in time, that does not rule out a peculiar ‘consciousness of spontaneity’ which is not itself an experience.  For Deleuze, on the other hand, when Kant says ‘I merely represent the spontaneity of my thought’, this should be taken to mean that ‘the activity of thought applies to a receptive being, to a passive subject which represents that activity to itself rather than enacts it, which experiences its effect rather than initiates it, and which lives it like an Other within itself.’  So what remains paradoxical in Kant’s account for Deleuze is that while the spontaneity of the subject is held to ground the normativity of the processes of conceptual recognition, not only does its spontaneity turn to passivity as soon as I try to apprehend it in time, but I can never truly recognize that spontaneity as my own. Kant’s paradox issues in ‘a ﬁssure or crack in the pure Self of the “I think”, an alienation in principle, insurmountable in principle’. 
The point is not just that ‘the conceptual activity through which the mind represents an object, including itself as an object, cannot itself be given to it as an object.’  Deleuze’s argument undermines the normativist claim that being implicitly aware of our rule-following implies also that we bind ourselves to it. The ﬁrst result of Deleuze’s suggestion is that there may indeed be a spontaneity that grounds the activity of the understanding, but we cannot assume that we are identical with it. Whatever else it does, at the very least this move introduces an ineradicable heteronomy into the space of judgement and reason. Moreover, as we will see shortly, it is the temporal aspect of the fracture in the ‘I’ that is at the heart of the matter. For Deleuze, paradoxically, this temporal aspect, rather than vitiating the transcendental project, will actually provide the hidden key to it.
Expressed most radically, in this picture the self becomes something like the Freudian ego, a satellite of the true spontaneity, which must now be opposed to a ‘Cogito [which] incorporates … an unconscious of pure thought.’  Isn’t something like this conﬁrmed in Kant’s section on the Paralogisms, where, after having proven the necessity for an ‘I think’ to synthesize our judgements, Kant turns his gaze upon the transcendental subject itself? There we ﬁnd him taking to its conclusion the idea that the transcendental subject whose self-consciousness is necessary in order for experience to be possible is entirely lacking in substance, and we really have no way of knowing what it ‘really’ is, or how to think the internal character of its spontaneity.
The transcendental subject is a whole empty representation I, of which one cannot even say that it is a concept, but a mere consciousness that accompanies every concept. Through this I, or He or It (the thing) which thinks, nothing further is represented than a transcendental subject of thoughts = x. 
The problem is not just that we do not know noumenally what the ‘I’ is made of; it is that there is nothing about the ‘transcendental subject = x‘, even taken in its bare spontaneity, which ultimately licenses the ascription of self-consciousness to it. Transcendental self-consciousness is revealed to be so absolutely formal that it must in fact be of a different kind to our everyday empirical self-consciousness. The transcendental ‘I think’, by virtue of the severity of its abstraction, must become an Other, which is as it were feeding me thoughts, which appear to me through the medium of time. 
Deleuze therefore goes in the opposite direction to the early Fichte here, who thinks that if selfconsciousness is to be primary for the Kantian turn, one must take it as univocal – for Fichte, at least in the ﬁrst Wissenschaftlehre, it is legitimate and indeed necessary to focus in on the act of self-consciousness, to use the kind of self-consciousness that we actually have access to, in order to secure the goals of the Kantian project. For Deleuze, on the other hand, the lesson of the Paralogisms is that the transcendental act of self-consciousness is a sterile, impassive act; its spontaneity occurs in another scene. This is one of the key differences between Deleuze and the German Idealists: Deleuze is suspicious of using our experience of self-consciousness as a direct model for the transcendental subject.
Now Deleuze in fact goes so far as to say that ‘the correlation between the passive self and the fractured “I” constitutes the discovery of the transcendental, the element of the Copernican Revolution.’  This is an obscure claim; why might this be? On the face of it, this ‘correlation’, such as it is, seems to bode rather ill for Kant’s whole argument in the Transcendental Deduction. The Deleuzean Deduction indeed must follow a very different path from Kant’s. But let us suspend this thought for a moment and now take into account that the deduction of the role of transcendental self-consciousness is only the ﬁrst step of Kant’s argument in the Transcendental Deduction, and the second step is the argument that space and time, as inﬁnite wholes, must themselves be subject to the unity of self-consciousness – that is, an intelligible passage needs to be hollowed out that runs from pure concept to pure intuition, in order ﬁnally to secure the synthetic a priori validity of Kant’s project. Once this latter step of the Deduction is completed, it can be ﬁlled out by the various speciﬁcations of the Transcendental Schematism. What Deleuze wants to suggest, however, is that in a sense there is a more fundamental a priori encounter between the ‘I think’ and time: the paradox of inner sense already reveals to us, at least potentially, time in its ‘pure form’. The ‘fractured I’ itself, the ‘I’ as fractured, can serve as the site for the encounter with a form of time that is more primordial than time taken abstractly as inﬁnite whole. Deleuze is suggesting that there is a pure form of time that is prior to the form (inﬁnite wholeness) that for Kant will guarantee the correspondence of intuition with uniﬁed self-consciousness.
So let us trace out an alternative path for the Transcendental Deduction along these lines. Let us imagine that, after having secured the necessity of the ‘I think’ in the ﬁrst stage of the Deduction, Kant now proceeds to reﬂecting on the attributes of this new subject he has discovered. Here he encounters the Paralogisms, and the paradox of inner sense. His basic task is to ﬁnd a passage from this ‘I think’ to the pure spatio-temporal manifold. But he now faces the obstacle that the subject can only experience its spontaneity passively through a temporal self, as if it were fractured. Now what happens if we take this apparent obstacle as a positive condition? For isn’t another way of stating Kant’s paradox to say that although the transcendental subject is entirely formal and universal, its activity can only be transmitted through irreducibly local temporal perspectives? Isn’t Kant’s paradox telling us that there is no actual, permanent standpoint that I can take upon my own cognitive and practical acts, due to the impersonality of the subject = x? If, as Deleuze suggests, ‘the “I think” affects time, and only determines the existence of a “self” that changes in time and presents a certain degree of consciousness at every moment’,  then there is no transcendental standpoint that I can take on my own past or future, whereby my memories and hopes can be once and for all linked together to form the story of one life, acknowledged by myself as subject as my own. In Karl Ameriks’ formulation, ‘the persistent representation of an “I” need not be the representation of a permanent “I”.’  At any given moment, when I am synthesizing my experiences together, it is always going to be from a local, temporal standpoint. Doesn’t it follow that the way I understand my past now might be different from the way I understand that same past in the future? My past is thus in principle open to being recoded in the future. It is as if, in principle, I must live in the future anterior – I can no longer say ‘I was this’, only ‘I will have been this’. Conversely my hopes are hostage to my current capacity to interpret my past. What I hope for now, what the future is for me now, depends on my understanding of my past. Here we start to glimpse the vertiginous nature ascribed to time in its pure form by Deleuze.  He compares the labyrinthine structure of time that emerges from this interpretation of Kant to Borges’s vision of time in his story ‘The Garden of Forking Paths’. 
Deleuze’s argument applies at three distinct levels.
First, it applies at the level of the analysis carried out by the transcendental philosopher: when he attempts to grasp the nature of the spontaneous subject ‘in person’, he is left describing an indeterminate ‘subject = x‘. This may or may not be a problem, depending on one’s view of Kant’s project.  Second, it erodes the basis in the Deduction for claiming that objective judgements are strictly normative, in so far as it erodes the identity implied in implicitly self-reﬂexive judgements. Thus in so far as one adheres to a ‘spontaneist’ reading of transcendental apperception, then the ground of the normative space of reasons becomes undermined by the possibility of radical heteronomy. But, third, it perhaps applies most effectively to the attempt by the subject itself to grasp its own unity in inner experience. For the transcendental subject provides absolutely no help in bringing about personal coherence inside the medium of inner sense (thoughts, memories, fantasy images and affects). It is this latter level that is of further interest for Deleuze. 
By taking away the possibility that the spontaneous subject is in any determinate way related to the self, Kant seems to surrender the self to merely ‘external’ determination. The contents of inner sense appear to be left to be related back to the external events which gave rise to them (according to the logic of the Refutation of Idealism), with their relative importance for the self merely a matter of objective determination according to natural laws. But it would be fallacious to conclude from this that the syntheses constraining the objects of inner sense are therefore exactly the same as those constraining outer sense (space). For even if the content of inner sense originates in outer sense, the rules for the reproduction of these mental objects (considered as objects of inner sense) are surely different from the rules governing objective, external appearances. As well as behaving entirely differently, my mental events are open to a fundamental reinterpretation of their sense in a way that objective events are not. My memories may undergo subtle reinterpretations over time, thus fundamentally changing their nature. Under certain conditions of interpretation, what appeared as a memory can be retrospectively transformed into a fantasy image, and vice versa. Isn’t it this very defeasibility of inner experience that is precisely what is most threatening about the paradox of inner sense?
So, by right, each moment of my past opens itself up to possible reinscription. Forking paths lead out of each moment into different possible future worlds. Deleuze claims that time now assumes a special kind of inexorability: to be in time is to be perpetually open to the breakdown of continuity in my experience, to be perpetually open to a loss of coherence in the narrativity of my life. My memories and hopes are in a profound sense defenceless against this permanent a priori possibility at the heart of time, the a priori temporal labyrinth that must accompany my experience of myself in time. Each memory bears within it the explosive potential to be reinterpreted from an indiscernible point in the future; each hope bears within it the catastrophic possibility that it is a delusion based on an incompletely integrated past. 
Kant’s insistence that the unity of the subject for itself must be grasped in the form of an Idea (the Idea of the self) is itself an acknowledgement of the problematic nature of the gap that opens up between subject and self.  In a sense he is protecting himself from the wound he has uncovered in the subject, in the fractured I. However, on the other hand, if Deleuze wants to avoid postulating an Idea of the self which would serve to wall up the passages opened up by time’s labyrinth, he would seem to be faced with a profoundly negative result. Time would seem to undermine perpetually the ‘I think’ and the situation would seem to be equivalent to something like the radical separation of the I of enunciation from the I of the statement in Lacanian psychoanalysis. The ‘I’ of enunciation would perpetually escape any position within the statements it utters; the most we could do is come to accept that fact. 
However, Deleuze’s aim is different. It is rather to discover a pure ‘analogy’ to the fractured ‘I’ in a special synthesis of time. Appealing initially to the ﬁrst Kantian thesis about time, Deleuze asserts that such a synthesis ‘is necessarily static, since time is no longer subordinated to movement; time is the most radical form of change but the form of change does not change’.  The transcendental conception of time must be stripped down to a pure order of irreversibility, in which ‘the future and the past are … formal and ﬁxed characteristics which follow a priori from the order of time’. Deleuze now claims that this ‘before’ and ‘after’ in all their purity can provide the perfect receptacle for the schematization of the fractured ‘I’, under certain extreme, dramatic conditions. If this is possible, then wouldn’t it fulﬁl the demands of the second step of a Transcendental Deduction, while retrospectively clarifying the proper transcendental level of the ﬁrst? Deleuze’s suggestion is that the transcendental subject might show itself as empty precisely under the condition that the self shows itself incapable of assuring its own continuity, ‘having abjured its empirical content’.  A radical failure of empirical uniﬁcation could actually provide the space for the appearance of the Cogito, as empty, as a pure form, and as Other.
Here obviously Deleuze departs even further from Kant, and he claims in fact that the only post-Kantian thinker really to glimpse the consequences of this ‘furtive and explosive moment’ in Kant’s theory is Hölderlin, in his short text Remarks on Oedipus. In Difference and Repetition it is the case of Oedipus, as read by Hölderlin, which provides Deleuze with an illustration of how the subject can in a sense give body to the pure form of time, although in Cinema 2 he also goes on to appeal to the potential of cinema to present all kinds of dramatic ‘temporal forking paths’, beyond the limit-case of Oedipus. But the story of Oedipus has a privilege for Deleuze due to the purity of its structure as outlined by Hölderlin.  ‘Is it possible’, Deleuze asks, ‘that Kantian philosophy should thus be the heir of Oedipus?’ 
For the ﬁrst half of his life, Oedipus grows up believing that he is the son of a Corinthian nobleman. He hears a prophecy that he will kill his father and sleep with his mother, which terriﬁes him. So he leaves Corinth. On the way he gets into a ﬁght with an entourage of noblemen from another place, and kills them. He arrives at Thebes, where he answers the riddle of the Sphinx and the people elect him as king, thus obliging him to marry the queen, whose husband has recently been killed. Oedipus remembers the prophecy but he knows he is safe as he has not killed his father nor slept with his mother; his mother and father are alive and happy in Corinth. When it starts to dawn upon him what has been happening, he undergoes a massive crisis. He realizes that he is not the person he believed he was. His whole conception of who he has been is completely overturned. He now must realize that he was unwittingly all this time treading the path towards the fate that was spelled out for him. Every single past memory is undermined, and recoded by his realization that he is not who he thought he was.
For Hölderlin, the story of Oedipus at this point manifests a ‘caesura’.  Deleuze suggests that it is nothing other than ‘the caesura, along with the before and after which it ordains once and for all, [which] constitutes the fracture in the I (the caesura is exactly the point at which the fracture appears)’.  Hölderlin writes that in this moment of transition, in the utmost form of suffering, … there exists nothing but the conditions of time and space. Inside it, man forgets himself because he exists entirely for the moment. [Time] is reversed … no longer ﬁtting beginning and end. 
The power of time to overturn one’s most intimate memories thus actually seems to gain body in this caesura that holds apart a pure ‘before’ and ‘after’. The entire ‘before’ and ‘after’ of Oedipus’ history is structured around the caesura of his transformation. Oedipus’ memories from this moment on are not his own memories; it is as if they belong to another man. His ability to unify his experience in time is shattered; in a profound sense, he has become subject to time. The fractured ‘I’ has become realized, but at the expense of a man’s coherence.Now although Oedipus is of course a limit-case in the range of human experience, this does not detract from his possible transcendental value. Deleuze’s statement that ‘the correlation between the passive self and the fractured ‘I’ constitutes the discovery of the transcendental’ might in fact mean that, in so far as this correlation involves a direct schematization of the formal emptiness of the transcendental subject onto the empty order of time, even if such a movement can only be sustained by a dramatization (involving the story of a self such as Oedipus), this nevertheless fulﬁls the condition of necessity that it is possible that such a ﬁgure, if not ‘accompanies’, then certainly haunts our representations. The model of Oedipus presents the fractured ‘I’ in person, but in dramatic form. It is in Oedipus that Kant’s account of the subject reveals its deepest meaning as a ‘Cogito for a dissolved Self’. 
In Cinema 2 Deleuze wants to say that such moments are implicit whenever there is a temporal forking. The ﬁlms of Welles, Resnais and Robbe-Grillet all provide further dramatizations of the effects of time’s multiple forking paths. The abandonment of Oedipus as a child is the fork in the Oedipus story. The story fulﬁls the role of limit-case because, after he realizes he has married his mother and killed his father, his entire previous history is rewritten for him; there is nothing left on which to rest his identity. Yet ultimately for Deleuze this condition of reinscription is potentially present throughout experience.
But is there a way to ‘descend’ from this pure form of time back to the empirical experience of time? This can be achieved by excavating the further conditions of this limit-case of temporal experience, the fractured ‘I’, or, more speciﬁcally, by deducing the de jure structures of temporality that have to be in place in order for the limit-case to function as such. It is at this point that we need to turn to Bergson.
Déjà vu: bergson and the transcendental synthesis of memory
We have said that each present in principle opens itself to potential recoding by the future, and that each future in turn depends on the way the past is taken. But clearly this potential for recoding cannot be unlimited. So we are obliged somehow to formulate how each present can be a reservoir of future interpretations, without it falling into indeterminacy. Deleuze appeals to Bergson here because the latter shows us how this problem must lead into another more fundamental one: the problem of how the present passes into the past at all. How does a present become past and preserved as that past if it is open to future interpretations? What gives the past its peculiar weight? It is not enough to say that the past is constituted as such after a new present has taken its place, as then the scope of that past would be restricted to what it signiﬁed for the following present. As Bergson says, according to the point of view in which I am placed, or the centre of interest which I choose, I divide yesterday differently, discovering several very different series of situations or states in it.… Scores of systems of carving are possible, no system corresponds with joints of reality. What right have we, then, to suppose that memory chooses one particular system, or that it divides psychical life into deﬁnite periods and awaits the end of each period in order to rule up its accounts with perception? 
Because the content of each present cannot simply be delimited as soon as the moment has passed, and because it therefore remains open for future reinterpretation, we must assume that the past is somehow formed ‘alongside’ the present: for otherwise we are left without a measure for determining how the past remains that past.  Bergson’s paradoxical resolution, according to Deleuze, is that ‘no present would ever pass were it not past “at the same time” as it is present … The past is contemporaneous with the present that it was.’48 In other words, each actual present is somehow doubled by a virtual ‘shadow’ of itself, which enables it to be re-actualized as the past it will have been. Perhaps the best analogy to use while thinking about this is how something can happen to us which we know now will have signiﬁcance for us at some point in the future, but at the moment we are at a loss to determine precisely how. 
Deleuze takes Bergson as doing transcendental philosophy here; he takes him to have ‘profoundly explored the domain of [the] transcendental synthesis of a pure past’.  What is compelling about Deleuze’s return to Bergson here is the suggestion that the constitution of the past as past needs to be accounted for, and that so far transcendental theory has failed to do this, limiting itself to the conception of memory as reproduction. But Bergson also suggests that a certain experience can bear out his extrapolation here: the experience of déjà vu.  Déjà vu, he suggests, can only be accounted for if we assume that the past is constituted as past at the same time as the present. Under normal circumstances, this ‘double inscription’ of past and present is not experienced as such, because our attention is directed towards the future. But if this latter condition is suspended (due to failures in attending to the present), then déjà vu becomes possible: that is, we experience a paradoxical ‘memory of the present’. The Bergsonian notion of déjà vu provides Deleuze with a paradigmatic example of ‘transcendental empiricism’. Transcendentally speaking, our actual existence … whilst it is unrolled in time, duplicates itself all along with a virtual existence, a mirror-image. Every moment of our life presents two aspects, it is actual and virtual, perception on the one side and memory on the other. Each moment of life is split up as and when it is posited.
Or rather, it consists in this very splitting. 
But if the future-oriented direction of cognition is suspended, ‘we can become conscious of this duplicating’, and experience what Deleuze calls a ‘direct or transcendental presentation of time’. 
Having made this paradoxical move to the assumption of a double inscription of past and present, having been forced into it by the potential of each present to be recoded in the future, Bergson takes a deep breath and makes one more step into this increasingly dark territory. We are obliged, he argues, to assume that the past is preserved in integral layers which accumulate, one on top of the other, as if memory were structured like a cone, with the apex representing the present advancing into the future, while the base grows ever larger. Each present that has passed is akin to a layer of this ever enlarging cone; these layers of the past have a merely virtual existence, in contradistinction to the actuality of the passing present. Now although Bergson does write as if this model implies the integral preservation of each level or layer of the past in an ongoing, ever more bloated synthesis, he does not have to go that far. The important point, as Deleuze comes to realize in the Cinema books, is to help us conceive how attention to an actual object at any given moment may be ﬁlled out by an appeal to different layers of the past where that object is embedded in different remembered contexts, and conversely to conceive how each past may contain more in it than has been actualized by any subsequent attempt to recall it. One doesn’t have to go the whole way and claim that the entire past is preserved in self-subsisting levels; that hypothesis can perhaps simply be sent back to its historical context, the prewar Victorian world of spiritualism and F.W. Myers, with which Bergson was entangled, as shown by his essay on ‘The Soul and the Body’. 
What I want to do now is put together these two facets of Deleuze’s theory of time and explain how they lead towards his theory of repetition, which ultimately completes this theory of time. I have suggested that Deleuze’s Bergsonian take on the past can help us descend from the Kantian notion of the pure form of time and provide us with a model of how the past is able to be reinterpreted by the future. With the cone model of memory, we move towards an account of how the accumulation of experience individuates us, and serves as a backdrop for temporal continuity. With this in mind, we can work our way back up the transcendental ladder, back to the situation of Oedipus. I described how the story of Oedipus in a sense dramatizes the situation of the fractured I. One may justiﬁably respond: what has this to do with us? By turning to the concept of repetition, inherited from Kierkegaard, we will see more clearly what it has to do with us.
The crime of time: kierkegaard and the pathology of repetition
Perhaps the quickest way to approach Kierkegaard’s notion of repetition is through his critique of Kant’s notion of moral agency. Kierkegaard famously claims that ethical action needs to be supplemented by another dimension he calls ‘the religious’, but which he also calls ‘repetition’. He goes so far as to claim that ‘just as [the ancients] taught that all knowing is a recollecting, modern philosophy will teach that all life is repetition.’  Kierkegaard claims that, because moral agency involves the realization of an ideal norm in time, the person who chooses the good freely is always subject to a vertigo which ethics itself, considered purely as demand, is unable fully to deal with. With the thought of our freedom, the ﬁeld of possibilities opens up.  We realize that what happens in the future is up to us. But Kierkegaard claims that anxiety permanently accompanies freedom because we inevitably doubt our capability to actualize moral actions in the future because of our failures in the past. ‘Anxiety is the dizziness of freedom, which emerges … when freedom looks down into its own possibility.’  We doubt whether we are up to being good, to subjecting ourselves to universal law, because past failures weigh us down. What we have been, and what we feel like we are because of what has been, make us hesitate to become what we might be.
But Kierkegaard’s point only becomes strong if we accept that these past failures are not contingent but somehow necessary and uncircumventable. In other words, guilt must be shown to be a precondition of the exercise of freedom. We cannot say that we committed past misdeeds out of ignorance, because we did not know better. Past misdeeds must somehow be shown to involve the active denial of autonomous free action. And this is why guilt must be thought of as indicating sin – an active defection from the moral law. With this reformulation of the notion of original sin, we arrive at an interesting encounter between Kierkegaard’s thought and psychoanalysis. Just as Freud postulates a ‘ﬁrst error’ or proton pseudos in childhood, an original experience of pleasure at the hands of a parent which inevitably becomes recoded as complicity in an act of seduction,  so also does Kierkegaard posit that an experience of guilt necessarily arises at the beginning of conscious life because as soon as one understands oneself as responsible one is committed to interpreting one’s temporal experience morally, and one’s moral experience temporally. However, on Kierkegaard’s picture, although we have an account of the co-originality of morality and temporality, we still do not quite have a speciﬁc reason why the human being must necessarily experience its past as a source of guilt. On its own terms, Kierkegaard’s claim that original sin is necessary is left crucially hanging.
Yet this is not true for Deleuze, who can refer at this point to the Bergsonian account of the past just outlined. If the past is contemporaneous with the present it has been, any present is by right accompanied by its potential to be reinterpreted from the future. But, conversely, future reinterpretations of the past really do reconﬁgure the meaning of that past as a past. Inevitable infantile acts of hesitancy, complicity or confusion will thus become recoded in the light of moral order; I will not be able to help seeing myself as responsible for my actions even when I wasn’t. The induction into moral experience is thus necessarily accompanied by the sentiment that we are always already guilty. It is this effect of the ‘weight’ of the past that gives power to the Platonic pathos of recollection, and that Kierkegaard will want to combat with repetition. In the former, the present is always grounded in the past, but in the latter it is the realization of the very contingency of the past as well as the present that motivates one’s decision to will it again, as if for the ﬁrst time, thus in turn releasing one’s capability to act successfully now.  ‘There is no other crime than time itself.’ 
Kierkegaard provides highly nuanced accounts of how assuming such a necessary guiltiness can conspire to scupper moral action. For instance, it can lead very easily to our taking an ironic stance on the possibility of achieving the good. The ironic person is someone who has a consciousness of the ideal, of what man should be striving for, but is convinced of our inability to attain it. He imagines anyone who seriously strives for the good as somewhat laughable. Another possibility is what Kierkegaard calls the demonic. In the story of Agnes and the Merman in Fear and Trembling, the Merman is so transﬁxed by his sinfulness that he denies himself the possibility of achieving the good as a punishment. He has a thorn in his side and he gets a savage pleasure from driving it further in.  These possibilities (or, we could say, these pathologies of time) show that ethics has to be supplemented by another dimension, the forgiveness of sins. Ethics cannot forgive sins, as it wants people to be ideal; it would be paradoxical for ethics to guarantee forgiveness of sins, as it needs to be strict, to keep one striving. Only another discourse, religion for Kierkegaard, can forgive sins. But Kierkegaard is quite speciﬁc about what religion must do. It must overcome memory, overcome recollection, overcome any idea that one’s capability for the good pre-exists one’s action. This is what repetition must ultimately mean: ‘getting back’ the inﬁnite sense of possibility lost in the course of temporal existence, as a result of the necessary weight of the past.  As we have seen, the modern, Kantian notion of time at least allows for the possibility of a radical break with one’s understanding of one’s own past. Oedipus is the negative example of this. But Deleuze suggests that the same structure shows the possibility of a repetition that permits the overcoming of guilt in order to act, by refusing determination by the past.
In Either/Or, Kierkegaard imagines an unhappy person who is never satisﬁed by any realization of their will because they have already experienced in thought what they are hoping for, while they are unable to accept the past as past because they refuse to accept that what has happened has happened. Thus they constantly remember something they should be hoping for, and hope for something that they should remember.  On the Deleuzean theory of time I have presented so far, shouldn’t we conclude that this singular unhappy person is everybody, at least potentially, and that to reclaim the possibility of agency in modernity means confronting the labyrinth of time and passing through it by virtue of a repetition?
The modern fact is that we no longer believe in this world. We do not even believe in the events which happen to us, love, death, as if they only half concerned us. It is not we who make cinema; it is the world which looks to us like a bad ﬁlm. 
For Deleuze cinema is the art form that has the most potential to dramatize the multiple ways of inhabiting the modern form of time. Cinema permits a montage of temporal relays quite different from, but based on the same temporal syntheses as, the experience of human beings.  However, in an interview Deleuze comments that the real reason he felt drawn to writing about the cinema is because of the possibility it has of representing ‘spiritual life’, the way it has of dramatizing spiritual acts, or acts of choice.  And if, as Deleuze says, ‘the philosophy of repetition is pathology’,  then that dramatization will also include temporal maladies of agency, such as those depicted by Kierkegaard. Deleuze describes how Kierkegaard’s tales, such as Agnes and the Merman, are akin to sketches for ﬁlm scripts, as if they each lay out different possibilities of ethically inhabiting time.  Cinema has the means to present all the vicissitudes of ethico-temporal experience, whether it be the experience of fatality in Bresson, or the choice to choose in Rohmer. Bressonian fatality, for instance, may be a dead end from the purely ethical point of view, but when presented as cinema it not only gains the power to move us because it allows us to inhabit the ethico-temporal situation of its characters, but it also cannot fail to call forth the power of repetition almost in recoil to what is being depicted.
At the outset, we encountered Deleuze’s suggestion that cinema recapitulates in speeded-up form a movement undergone in philosophy. The claim was that the embryo of cinema repeats the revolutionary movement already undergone in Kantian philosophy. In Difference and Repetition, Deleuze writes that ’embryology already displays the truth that there are systematic vital movements, torsions and drifts, that only the embryo can sustain: an adult would be torn apart by them.’  If philosophy was not in fact able to live this moment, if it could not sustain the movement demanded of it, then could its most beautiful daughter, the cinema?  Deleuze suggests that for mass audiences in darkened spaces across the rubble-strewn landscape of the midto late-twentieth century, cinema was indeed sustaining movements that could not otherwise be lived by its spectators.
But the suspicion was also ventured at the outset that Deleuze’s Cinema project itself appeared at what now looks like the moment of the twilight of European cinema. Could it be that cinema has now undergone the fatal experience of being born, so that its powers as an embryo are now becoming something less than a memory? What can this mean for the relation of Deleuze’s analyses to future cinema?
If Deleuze is right about the destination he retrospectively discovers in cinema, then surely there is no other option but to seek out and to produce ever more sophisticated realizations of the possible pathologies of repetition. Fredric Jameson was wrong to see the collapse of narrative coherence in individual, society and art as a ‘fragmentation’ with analogies to schizophrenia. What is going on in aesthetic works of fragmentation is rather an exploration of the new temporal reality. On the other hand, it could be objected that Deleuze’s argument depends on the retention of the intrinsic power of the past to form a ‘weight’ against the present. Jameson argues that this is precisely what is lost: ‘the present … comes before the subject with heightened intensity … the breakdown of temporality suddenly releases this present of time from all activities and intentionalities.’  Moreover, if there is an ongoing historical actualization of the fractured ‘I’, as Deleuze suggests in Cinema 2, then that will participate in an absolute deterritorialization which tends towards the accomplishment of the de-substantialization of the past. As is indicated in the Bergsonian transcendental argument above, the pre-existence of the past is in any case ultimately grounded only on its openness to the future. It is the fact that we never know what is implicated for the future in any single present that is the ﬁnal reason for the constitution of the past as such. But with the historical collapse of the boundary concepts of God and Self, initiated by Kant, all recollection now tends to reveal its internal dependence on the future. This is why ‘we no longer believe in this world’.
If ﬁlms such as Lynch’s Lost Highway or Eternal Sunshine of the Spotless Mind testify to an ongoing interest in cinematic temporality, one could nevertheless argue that they also tend to follow one of two tendencies: on the one hand, towards increased focus on the trauma at the source of temporal divergence (Lost Highway as model); on the other hand, to the extent that they articulate a movement towards repetition, they can appear as oddly glib and sentimental (Eternal Sunshine). Perhaps the ﬁlms and spectators of today are no longer ‘spiritually’ able to sustain the movement. Or perhaps, just as religion was the ﬁgurative, alienated, yet penultimate form taken by Hegel’s absolute subject, cinema as alienated image of the Deleuzean subject will increasingly tend to show its inability to ﬁgure a pure act of repetition, so that it must react against its very status as image, or be complicit in its entrapment in it. In this case, Deleuze’s writing of the Cinema books would have been an act of memory, preserving the historical drama of cinema itself from oblivion.
Thanks to Peter Hallward for his comments on an earlier draft, and to Peter Dews for inviting me to the ‘Philosophy and Social Sciences’ conference at the Czech Academy of Sciences, where an earlier version of this article was ﬁrst given. It was written during a Leverhulme Research Fellowship in the Centre for Research in Modern European Philosophy at Middlesex University.
1. ^ Gilles Deleuze, Cinema 2: The Time-Image, trans. H.
Tomlinson and R. Galeta, Athlone, London, 1989, p. xi.
2. ^ Ibid. Cf. also pp. 39 and 271 for further comments indicating the centrality of Kant’s theory of time for Deleuze’s cinema project. For other purely philosophical formulations of this epochal shift, see Gilles Deleuze, Difference and Repetition, trans. P. Patton, Athlone, London, 1994, p. 88; and ‘On Four Poetic Formulas that Might Summarise the Kantian Philosophy’, in Essays Critical and Clinical, trans. D.N. Smith, Verso, London, 1997, pp. 27–9.
3. ^ Deleuze, Difference and Repetition, p. 58.
4. ^ This primary emphasis on the moving (‘kinematic’) image is behind Deleuze’s decision to privilege the term ‘cinema’ over ‘ﬁlm’.
5. ^ ‘In reality as such, there is no repetition. This is not because everything is different, not at all. If everything in the world were completely identical, in reality there would be no repetition, because reality is only in the moment.… When ideality and reality touch each other, then repetition occurs. When, for example, I see something in the moment, ideality enters in and will explain that it is a repetition’, Søren Kierkegaard, Johannes Climacus, or De Omnibus Dubitandum Est, trans. H.V. Hong and E.H. Hong, Princeton University Press, Princeton NJ, 1985, p. 171. This is the most ‘epistemological’ or ‘metaphysical’ of the interpretations Kierkegaard gives to repetition, as usually repetition is an aesthetic concept (in the book Repetition itself) or a religious concept (e.g. in Fear and Trembling, where Abraham acts in order to ‘get back’ Isaac). However, Deleuze does not explicitly mention this text, so my use of it here is speculative.
Deleuze unfortunately never justiﬁes why he thinks the concept of repetition is of such fundamental import, or why he thinks one can generate a whole theory of temporal synthesis from it. I return to the religious notion of repetition below.
6. ^ Deleuze denies that these syntheses are fundamentally psychological. He variously declares them to be organic, transcendental or ontological. But it remains difﬁcult to assess at which theoretical level they are pitched.
7. ^ Deleuze, Difference and Repetition, p. 87.
8. ^ This regressive account could be said to stand to Deleuze’s Difference and Repetition as Kant’s Prolegomena stands to the ﬁrst Critique. See Kant’s distinction between ‘synthetic’ or progressive and ‘analytic’ or regressive approaches in Prolegomena to Any Future Metaphysics, trans. J. Ellington, Hackett, Indianapolis, 1985, p. 8, Akademie edition, p. 263.
9. ^ Immanuel Kant, Critique of Pure Reason, trans. P. Guyer and A. Wood, Cambridge University Press, Cambridge, 1997, A34/B50. This text will cite as usual with ﬁrst and second edition pagination.
10. ^ Ibid., A181/B224. As Deleuze says, ‘Everything that moves and changes is in time, but time itself does not change or move … [it is] the immutable form of change and movement.’ ‘On Four Poetic Formulas that Might Summarise the Kantian Philosophy’, p. 29.
11. ^ Ibid. Thus begins the First Analogy.
12. ^ Kant, Critique of Pure Reason, A32/B48.
13. ^ Ibid., B131.
14. ^ Henry Allison, Kant’s Transcendental Idealism, Yale University Press, New Haven, 1983, p. 140.
15. ^ Henry Allison, ‘On Naturalising Kant’s Transcendental Psychology’, in his Idealism and Freedom, Cambridge University Press, Cambridge, 1996, p. 62.
16. ^ Robert Pippin, ‘Kant on the Spontaneity of Mind’, Canadian Journal of Philosophy, vol. 17, no. 2, June 1987, p. 459. Pippin repeats this account in Hegel’s Idealism, Cambridge University Press, Cambridge, 1987, pp. 21–4.
17. ^ Ibid., p. 460.
18. ^ Kant, Critique of Pure Reason B152.
19. ^ Ibid., B158. Deleuze refers to this passage (as well as to B428–430, where similar views are expressed) in Difference and Repetition, pp. 85f.
20. ^ Pippin, ‘Kant on the Spontaneity of Mind’, p. 454.
21. ^ See also Allison, Kant’s Transcendental Idealism, pp. 275–8, where he comments at length on Kant’s fragment ‘Is it an Experience that We Think?’
22. ^ Deleuze, Difference and Repetition, p. 86, stress added.
23. ^ Ibid., p. 58.
24. ^ Allison, Kant’s Transcendental Idealism, p. 278.
25. ^ Deleuze, Difference and Repetition, p. 174.
26. ^ Kant, Critique of Pure Reason, A346/B404.
27. ^ Deleuze arrives at something like Freud’s view of selfconsciousness by pushing the Kantian view to the limit.
28. ^ Deleuze, Difference and Repetition, p. 86, stress added.
29. ^ Deleuze, ‘On Four Poetic Formulas that Might Summarise the Kantian Philosophy’, p. 29.
30. ^ Karl Ameriks, Kant’s Theory of Mind, Clarendon Press,
Oxford, 2000, p. 135.
31. ^ The vertigo is ampliﬁed if we take into account Deleuze’s claim that Kant’s critique of rational psychology in the Paralogisms coincides with what he calls ‘the speculative death of God’ effected by Kant’s destruction of the Ontological Argument for the existence of God.
32. ^ Deleuze, Difference and Repetition, p. 111. This reference occurs during Deleuze’s later interpretation of the pure form of time from a psychoanalytic point of view. Cf. Gilles Deleuze, Logic of Sense, trans. M. Lester, C.
Stivale and C. Boundas, Columbia University Press,
New York, 1990, pp. 62, 114; Cinema 2, p. 49. In the latter, Deleuze compares the Borges story with the ﬁlms of Joseph Mankiewicz, and the implications I am drawing here are most clearly spelled out in these passages.
33. ^ It may not be a problem if one takes a merely ‘functional’ view of what Kant is attempting to achieve in the Deduction. The non-experiential nature of pure apperception is in that case not so threatening, as one would just be providing a bare model for possible cognition, without specifying how it is actualized, or without claiming that ‘the perspective of the individual who is engaging in various mental activities [is the same as] that of the theorist who is describing those activities’ (Patricia Kitcher, Kant’s Transcendental Psychology, Oxford University Press, Oxford, 1990, p. 83).
34. ^ I leave it open here whether Deleuze’s accentuation of the paradox of inner sense brings down the legitimacy of Kant’s account of the understanding (the second level just referred to). For a strict normativist reading of Deleuze’s Kantianism which accentuates the role of reason and Ideas over that of the understanding, see my ‘Deleuze, Kant and the Question of Metacritique’, in the Southern Journal of Philosophy, vol. 42, no. 4, December 2004. From this point I shall rest the weight of Deleuze’s argument on the consequences of the paradox for inner experience.
35. ^ In my forthcoming Deleuze and Psychoanalysis (Continuum, London, 2006), I make it explicit that one of Deleuze’s aims in Difference and Repetition is to contribute to a new psychoanalytic theory of Nachträglich-keit (deferred action or traumatic causality). For suggestions about possible socio-political schematizations of a similar temporal structure, see Peter Osborne, The Politics of Time, Verso, London, 1995, ch. 4, ‘Modernity,
36. ^ See Kant, Critique of Pure Reason, A335/B392, on the Idea of ‘the absolute unity of the thinking subject’.
37. ^ Cf. Jacques Lacan, ‘The Subversion of the Subject in the Dialectic of Desire’, in Écrits, trans. B. Fink, Norton,
New York, 2002, pp. 288/801, 298/813. 38. Deleuze, Difference and Repetition, p. 89.
39. ^ Ibid., p. 89.
40. ^ There is also something uncannily appropriate in the way this ancient story waits for post-Kantian modernity to reveal its most explosive possibilities. The Erasers, the ﬁrst successful novel by Robbe-Grillet (one of the central ﬁgures in Deleuze’s Cinema) was also an exploration of the temporality of the Oedipus story.
41. ^ Deleuze, Difference and Repetition, p. 87.
42. ^ Hölderlin, ‘Remarks on Oedipus’, Essays and Letters on Theory, trans. T. Pfau, SUNY Press, Albany NY, 1988, p. 102.
43. ^ Deleuze, Difference and Repetition, p. 89.
44. ^ Hölderlin, ‘Remarks on Oedipus’, p. 108.
45. ^ Deleuze, Difference and Repetition, p. 58.
46. ^ Bergson, ‘Memory of the Present and False Recognition’, in Mind-Energy, trans. H. Wildon Carr, Macmillan,
London, 1920, p. 129.
47. ^ Bergson is also right to insist that the past cannot be accounted for in terms of weakened perceptions, which would mean that if one varied the intensity of a perception, at some point it would change tense.
48. ^ Deleuze, Difference and Repetition, p. 81.
49. ^ As mentioned above in note 32, the work of Joseph Mankiewicz is the cinematic convergence point for Deleuze’s Kantian, Borgesian, and now Bergsonian arguments about time. Mankiewicz uses ﬂashback in ﬁlms such as The Barefoot Contessa and All About Eve in such a way that they manifest ‘each point where time forks’ in the narrative. But ‘time’s forks thus provide ﬂashback with a necessity, and recollection-images with an authenticity, a weight of past without which they would remain conventional’ (Cinema 2, p. 50).
50. ^ Deleuze, Difference and Repetition, p. 81.
51. ^ Bergson, ‘Memory of the Present and False Recognition’, in Mind-Energy.
52. ^ Ibid., p. 135.
53. ^ Deleuze, Cinema 2, p. 275. The illusion, Bergson says, can become so complete that the subject not only feels that they have experienced this moment before, but that they can predict exactly what will happen from now on. The subject risks becoming a ‘double’ of himself, ‘present as a simple spectator of what he is saying and doing’ (Bergson, Mind-Energy, p. 109).
54. ^ See Mind-Energy, chapter
2. ^ Bergson there contends that his theories of memory help us to conceive how memories can be preserved after death. Deleuze’s suggestion that ‘the past is pure ontology’ (Bergsonism, p. 56) certainly lends itself to a substantialist interpretation, which should nevertheless be resisted. As Deleuze himself claims, his account is transcendental, and is thus pitched at the level of ‘sense’. It is non-psychological in this particular sense.
55. ^ Søren Kierkegaard, Repetition, trans. H.V. Hong and E.H. Hong, Princeton University Press, Princeton NJ, 1983, p. 131.
56. ^ ‘The possible corresponds exactly to the future.’ Kierkegaard, The Concept of Anxiety, trans. R. Thomte, Princeton University Press, Princeton, 1980, p. 91.
57. ^ Kierkegaard, The Concept of Anxiety, p. 61.
58. ^ Sigmund Freud, ‘Project for a Scientiﬁc Psychology’, in Standard Edition of the Complete Psychological Works of Sigmund Freud, vol. 1, trans. J. Strachey, Hogarth Press, London, 1966, pp. 352f.
59. ^ Kierkegaard develops a point in Leibniz in order to arrive at this idea. Against the fatalistic ‘Lazy Sophism’ that ‘if the future is necessary, that which must happen will happen, whatever I may do’, Leibniz says that ‘it is untrue that the event happens whatever one may do: it will happen because one does what leads thereto’ (Leibniz, Theodicy, trans. E.M. Huggard, Open Court, La Salle IL, 1985, pp. 54, 57). Historical events are contingent truths different in kind from necessary truths. Kierkegaard adds to this that past events must also be thought of as retaining their contingency. Citing Leibniz’s notion of possible worlds, Kierkegaard writes that not only is the future not necessary, but ‘the basis of the certainty of the past is the uncertainty regarding it in the same sense as there is uncertainty regarding the future’ (Philosophical Fragments, trans. H.V. Hong and E.H. Hong, Princeton University Press, Princeton NJ, 1985, p. 80.) The past could have happened otherwise than it did.
60. ^ Deleuze, Cinema 2, p. 37.
61. ^ Søren Kierkegaard, Fear and Trembling, trans. H.V.
Hong and E.H. Hong, Princeton University Press, Princeton NJ, 1983, pp. 94–9.
62. ^ On repetition as ‘getting back’, see Edward F. Mooney, ‘Repetition: Getting the World Back’, in A. Hannay and G. Marino, eds, The Cambridge Companion to Kierkegaard, Cambridge University Press, Cambridge, 1998.
63. ^ ‘The Unhappiest One’, in Either/Or, trans. A. Hannay,
Penguin, Harmondsworth, 1992, p. 216.
64. ^ Deleuze, Cinema 2, p. 171.
65. ^ As a random example, take a scene of Citizen Kane.
After his election defeat Kane has a conversation with his friend Leland, who we know he will fall out with.
Because we know this before we arrive at this scene in the ﬁlm, we can’t help but experience the scene with a sense of déjà vu, as if we have already experienced it.
We realize we are observing the moment which will have been the ﬁrst sign of Kane’s falling out with Leland.
This scene shows how cinema allows us to appreciate the ‘double inscription’ of each moment: as well as being present, it always has a projective, future anterior dimension, which allows it to be experienced in the past or future, depending on how the ﬁlm is organized. See Cinema 2, p. 106.
66. ^ In Gregory Flaxman, ed., The Brain is the Screen, University of Minnesota Press, Minneapolis, 2000, p. 366.
67. ^ Deleuze, Difference and Repetition, p. 290.
68. ^ Deleuze, ‘The Brain is the Screen’, p. 366; and Cinema 1, p. 233.
69. ^ Deleuze, Difference and Repetition, p. 118.
70. ^ In ‘On Four Poetic Formulas that Might Summarise the Kantian Philosophy’, Deleuze writes of the ‘Shakespearean aspect of Kant, who begins as Hamlet and winds up as Lear, whose daughters would be the post-Kantians’ (p. 35).
71. ^ Fredric Jameson, Postmodernism, or, The Cultural Logic of Late Capitalism, Verso, London, 1991, p. 27.