Guide to Reading Deleuze’s Cinema II: The Time-Image, Part I: Towards a Direct Imaging of Time to Crystal-Images
[Here's a continuation of my series on reading Deleuze's Cinema I & II . . . ]
If the greatest impediment to understanding Deleuze’s concepts in The Movement-Image is confusion over what he means by the word image, then a similar roadblock occurs in The Time-Image in relation to his use of the word time. Understand what he means by the word image, and Cinema I is a massively easier read; understand what he means by the word time, and the same thing happens for Cinema II.
Bergson’s Critique of Clock-Time
Deleuze’s conception of time in Cinema II is taken almost directly from Bergson, and with Bergson, it’s easiest to start out by describing precisely what time does not mean for both of them. Bergson’s philosophy finds its genesis in the critique of clock time, and in favor of the lived time of duration. Clock-time, for Bergson, is a way of spatializing time, and as such, it really isn’t time, it’s a form of space. When we think of the time captured by clocks, we think of each moment as a self-contained entity, complete unto itself, separate from the others. Time is then best diagrammed as pearls on a string, with each pearl a separate moment. With each tick of the clock, we move from one pearl to the next.
For Bergson, this is a highly misleading and ultimately false image of time. Lived time, time that endures, is time that flows, time in which the past and future penetrate into the present in the form of memory and desire. Time stretches when it seems to move more slowly (ie: when bored), and compacts during moments of crisis, and we seem to dip deeper into memory at some points (ie: moments of dreaming, fantasy, reverie), and more shallowly during moments of action.
Between Virtual and Actual
For Bergson, the present is a dynamic interpenetration of past and future. The aspect of our lived world that is here for us right now, in the present moment as that which ‘feels most real to us’, is what he calls the actual. When I hold an object in my hand, say, a coffee mug, it feels more real than the memory of a coffee mug, or an image of a coffee mug on a TV screen. That feeling of being more real is what allows us to tell an actual coffee mug from one which is less real, so to speak, or more virtual. An image of a coffee mug in memory, or in a film, is thus a virtual image, while the one we hold in our hand at any given moment is an actual image (and remember, everything is an image for Deleuze, because when he says image it is, for him, and following Bergson, a way of saying a ‘slice of the world’). Worldslices, or images, come in many shades of actuality, and some are more actual and less virtual, or more virtual and less actual, than others.
During periods of stress, in which we are focused on action, we find ourselves immersed in the present moment, it’s needs and exigencies. At this point, we exist mostly in the actual, there is very little virtuality in our world. Since the past and the future, represented in our present as memory and desire/fantasy/anticipation, are relatively weak at this moment, we can say that when concentrating on action, we exist mostly in the actual.
But as we dive deeper into memory and/or fantasy, that is, the realm of the virtual, we leave the present and its needs ever more behind. This is why it is perhaps best to equate the actual with the present, and the virtual with the past/future, or future/past, whichever you prefer. Because the actual will always feel more real, more present, than aspects of the past/future (except for in cases of hallucination). So, at least for humans and in relation to issues of time, the virtual is the past/future, and vice-versa. This isn’t to say that there may not be other examples of the virtual. For example, for Deleuze, an actual coffee mug produces a virtual image when reflected in a mirror. But as we will see, for Deleuze, images in a mirror have a peculiar temporal relation to the actuals they reproduce, more of which will be said in a bit.
Why Time is Freedom
Here we also see why it is that Bergson and Deleuze equate the virtual with freedom. For my dog, a creature of instinct, the actual almost always leads directly to a preprogrammed action. But for me, an actual impression may lead to an instinctual action, or reverie or fantasy or recollection, and with a much greater degree of latitude than my dog. When my dog sees his food, he is unlikely to be thinking of Proust the moment afterwards, while humans have this happen all the time. Which is why we only sometimes do what our instincts tell us, because the past may interrupt the present, and present novel ways of reading the present which may influence our future. Likewise, we may have all sorts of desires which draw particular aspects of our past into contact with our present in ways that disrupt the chain of instinct. The virtual past/future infused into the actual is what produces freedom from being enslaved to the moment. If rocks are fully enslaved to the moment, plants slightly less so, animals a little freer, only humans, as far as we know, can gain significant freedom, and this is because of our complex brains. Brains which store our futurepasts so as to use them to increase our options.
To sum up, the present is more or less the actual, and the past/future is more or less the virtual. Thus, there are two axes to time, not only. Yes, time presses ever further into the future, and we know this because our stores of memory increase over time. This could be thought of as the movement of time horizontally. Such movement, however, isn’t like moving from one pearl to another along a string of pearls, but rather, as a sort of increase in the memory store of the past as the future flows into it via the doorway of the present. However, in addition to a horizontal axis to time, there is also a vertical axis. The closer one is to the present, the closer one is to what in math would be the x-axis, the line of horizontal movement along which past, present, and future are distributed. But the further one dips into the virtual, the past/future, the more one expands upon the y-axis.
Let us say, for example, that in the middle of an action, for example, a morning walk, you encounter some animal you’ve never seen before. What is that, you wonder. As you dip into past memory to search for something that resembles this, you finally find some memories that seem to fit. This is the process of recognition. Recognition that is relatively automatic, and becomes habit, requires less depth of digging around in the past, but when you need to dig more deeply, there is a greater degree of the virtual in the present (more expansion on the y-axis). While it may also take more time to dig around in memory like this (and hence expand on the x-axis), this is not always necessarily the case.
Three Basic Time-Images: Recognition, Recollection, and Dream
Recognition is the lowest level of digging into the depth, so to speak, of the present, and into the future/past. Recollection is the next level of depth, in which one keeps a more tenuous connection to the present, but dives into memory to reconstruct a scene from yesterday, or last year. One is less present in the present moment, so to speak. And finally, when one is dreaming, or fantasizing, one has barely a connection to the present at all. We get lost in reverie, for example, and we may trip as we walk because we are caught up in our memory-fantasies. Or when we get totally involved in a dream, we find that we are sleeping, with barely any connection to the present at all. Bergson even hypothesizes that perhaps death is what happens when the cord is fully disconnected, breaking the link between virtual and actual completely.
It’s important to note that the virtual is not merely the past, and memory, but also the future, and fantasy. For when we fantasize about something, say, we imagine what food we want for dinner, we do so by assembling memories into aggregates. I imagine a wonderful dinner, but the image I have of this that anchors my fantasy is composed of bits and pieces of memories hauled from the past. Likewise, when I recollect something, this is as much a recreation, and hence, full of fantasy and the future and present as much as the past. Memories wouldn’t distort were this not the case. Furthermore, even the present moment of recognition is infused with the future. For when I recognize something in front of me, I use not only memory, but desire, namely, the desire that impels me to action. When I walk down the street, my desire is what impels me, what reaches into my store of memory to retrieve images to meet the present and help me recognize what is in front of me. The past can’t be activated without the future dipping into it. The virtual is this interpenetration of the past and future by means of the present.
It is for this reason that Bergson and Deleuze also describe the virtual as the potential for difference, for creation, for the radically new. The actual is in a sense dead, it can only be what it is. But the virtual is the opening of what is onto the possibility to be different in the future, to have been different in the past, and for desire and memory to impact the present so as to alter it’s relation to itself and the world around it.
Thus, we can come up with a semi-equation to help us. The virtual=past/future=freedom=the new/creation=difference=time, while the actual=the present=necessity=the same=repetition=space.
Now as any who’ve studied some Deleuze know, the distinction between difference and repetition is essential to him. The virtual is associated with difference, pure difference, and the actual with that which repeats, which stays the same, with repetition.
What is a Time-Image?
What then is a time-image? A time-image, for Deleuze, is an image which is infused with time. That is, it is an image which is different from itself, which is virtual to itself, which is infused with past/future. What types of images are these?
Humans use time-images all the time. We call up images in our memories or fantasies to help us navigate the world. We don’t think these images are as real as those provided by our senses at the present moment, but they exist for us nevertheless. When I recognize a coffee mug on my table, I do so by pairing it up in my mind with virtual images of mugs and cups past. Recognition is the pairing of virtual and actual images. Habit occurs when this process becomes semi-automatic, but whenever I encounter something new or different, the process becomes more extended. Any image I draw from the past and/or synthesize with others so as to help me with my process of recognition is called by Deleuze, following Berson, a recognition-image. Were I to drag fragments of memory out of my past to reconstruct an entire scene, for example, of what I did when I last saw my friend two years ago, the image I created of the past, a flashback, essentially, would be a recollection- image. And were I to dream of that meeting, and perhaps then have fantastic things happen, like we meet a cartoon character for dinner later that night, we would have a dream-image.
Such images are never merely images, for Deleuze, That is, they are images which stand for, or in relation, to other images. This is why these images also function as signs, virtual signs of actual images from the present which calls them up in the first place. This is why in Cinema II we see, for the first time, images referred to as signs. Deleuze calls recognition and recollection images forms of mnemo-signs, basically, memory signs. And he calls dream-images types of oneiro-signs. (It is worth noting that relation-images, proto-time-images discussed towards the end of Cinema I, show relation but not via consciousness, and are in a sense ancestors of op- and son-signs . . . )
In film, we often see the process of recognition, recollection, or dream depicted for us. If someone in a film sees something, and then we see a memory of the past flash on screen, followed by an act which shows that now the character recognizes the object in front of them, the image drawn from the past functions in the film in question as a recognition- image. When a flashback occurs in a film, it provides us with recollection-images. And when someone falls asleep and dreams, or hallucinates, we have dream-images.
What distinguishes these three types of images from the more actual images of Cinema I is that they are always not fully what they are. That is, they are virtual, they function as signs. An image of an object in a dream is not ‘fully real’, because it is just a dream. We know it is just a dream because in some other part of the film, we are told this, or this is somehow indicated. When we see the person wake up from the dream later in the film, or go to sleep before, this context is virtually present in the dream-images, and this virtual presence makes these images feel less real to us. Thus, the images in a dream are more virtual, and less actual, than others, because they are suffused with context, with that which is not themselves. That is, they are suffused with difference, otherness, they are only partly there. And here we see why it is that context, difference, time, representation, and relation are all linked to the notion of virtuality for Deleuze.
Any image which functions like this, which helps us recognize, recollect, or dream, is a type of time-image. And there were time-images before WWII, in the period of cinema that was dominated by the movement-image. But after WWII, time-images become ever more prevalent, particularly in avant-garde or non-mainstream, non-Hollywood film. Hollywood film remains stuck in the action-image, while film that really explores new potentials for both filmic and human consciousness began to explore the time-image directly. For the time-image in fact showed itself in two forms before WWII. The first is in prewar recognition, recollection, and dream-images. But there was also the indirect imaging of time via montage. Attempts to capture and image movement used cuts, and cuts indicate a form of pure difference which registered and impacted the images they connected. This is why Deleuze says that montage is an indirect image of time, a version that speaks through the movement-image. Pre-war recollection, recognition, and dream images are direct images of time, yet weak ones. For in fact, they are filtered through human forms of consciousness. They are in a sense cloaked or clothed (to use terms employed by Deleuze in Difference and Repetition) by the sensory-motor schema that dominates the cinematic image before World War II. Thus, we see flashbacks, and dreams, but they are always clearly demarcated. The only odd exception, for Deleuze, are films which depict fantastic worlds that are like dreams, but aren’t dreams as such. The prewar musicals of Busby Berkeley and Vincent Minnelli are prime examples of the world as dream (a form which takes off in the immediate post-war period with the uber-trippy Powell and Pressburger opera-spectacle film from 1951, Tales of Hoffman).
The Post-War Context, and the Direct Image of Time
But after World War II, non-mainstream cinema begins to explore direct-images of time in a manner which is free from the relation of time to action. Films depict dreams and fantasies and memories in abrupt fashion, often not indicating to us we are in dream or fantasy till much later in the film, suspending our ability to tell what is actual or virtual until later. But then there are time-images which simply aren’t tied down to human consciousness. Such direct time-images are uncloaked, so to speak.
We see these first, Deleuze argues, in two early non-Hollywood contexts, namely, the films of Italian neo-realism, and the films of Ozu. In neo-realist film, we often see the camera linger on the scenes of destruction which serve as the setting for many of these films. It is as if the camera is attempting to process these images, but it cannot, so it lingers, and registers the trauma of the image in its pure form, the difference of the image from our ability to integrate it into our world. In Ozu, we often see his famed ‘pillow shots’, in which we pull away from one of his domestic dramas, and Ozu provides us with an image which seems to metaphorically comment upon the scene at hand, but which floats, as it were, outside the direct consciousness of any particular character. Such images which seem to lift up out of the consciousness of the characters in the film, but which also are beyond the traditional establishing shot, are examples of what in literature would be called ‘free-indirect discourse’, namely, something between the voice of the narrator and of the characters. Deleuze taps this term, and uses it to describe these moments of what can ultiamtely be called ‘free-indirect’ camera or vision. These moments Deleuze describes as those of opsigns and sonsigns, or optical and sound images which cannot be integrated into either pure objective or subjective frames.
Deleuze is a bit unclear as to whether or not traditional recognition, recollection, and dream-images should be considered direct images of time (though he does refer to them as time-images, hence my use of the term ‘cloaked’ to make a distinction here). But he is very clear to say that if montage indirectly imaged time via the movement-image, then the op- and son-signs give us direct images of time.
Mirror-Images: From Hyalo-signs to Image Crystals
Beyond this, however, there are also time-images which are further removed from the sensory-motor schema whereby the perception, affection, and action of human bodies in movement dominated film. There are films in which we see mirrorings of various sorts. Mirrors provide us with virtual images of actual entities. A person seeing their actual face in the mirror sees a virtual image of their face. What is the temporal status of their face? It seems as if the time of the mirror exists in a perpetual past-future, or future-past of the present of the actual image if reflects. From here, we see Deleuze develops his notion of the hyalo-sign, or glass- or mirror-image.
Deleuze doesn’t limit hyalo-signs to mirror images in film. Rather, two characters which are similar to each other, that resemble each other, are often referred to as mirror-doubles, particularly in psychoanalytic film criticism. And here we see Deleuze take on psychoanalysis, and show that he is able to outflank it. Mirrorings in film disrupt the otherwise linear flow of time in a film, they create temporal short-circuits. For if time is generally marked by entities like clocks, which use the difference physical difference presented by the world around us to mark relative degrees of change in space, what happens when space starts to resemble itself? Mirrors disrupt time. And when there is a hall of mirrors, which Deleuze calls a crystal (for a crystal is little other than an object made of many little shards of reflecting surface), then we have an entity in which notions of before and after start to literally break down. Films in which parts of the film mirror each other, which are full of short-circuits of this sort, he calls image crystals, or crystal-images.
Time travel films of all sorts are image crystals, as are films which literalize fantasy, hallucinations, and dreams so as to create repetitions of various sorts. Films full of mimicry and doubling might not have overt time-travel in them, but they produce odd temporal short-circuits nevertheless. These are all, for Deleuze, crystal-image films.
I’ve discussed image crystals in several other posts, so I’ll stop here. But needless to say, crystal-images are the way in which Deleuze is able to emphasize the aspect of the virtual which is truly that of difference, beyond any human notion of time. The virtual is difference as such, becoming, and time, particularly human time, is merely one of its forms.
This is why Deleuze says that time-images bring montage into the image. They are pure difference inside an image. When we see crystal image films, we very often will not know, on first viewing, what exactly an image means. One thing you learn watching image crystals is to suspend your judgment of images, because you will never know which aspect of an image will be selected for radical reworking later in the film. In this manner, each image becomes suffused with past/future, time, context, relation, and difference. It becomes virtual, less directly present, pure difference lurks between the very pores of the aspects of the image. What is present is an imaging of time, a depiction of time, of pure difference, in the image itself.
A direct imaging of time.