Practice Requires Time

Well, we've made it to the end of chapter 2, and I think I may hang up the spurs at this point.  I've thoroughly enjoyed all this writing; I've never been so involved in a book, and it's fortunate that the book rewards involvement to an almost unique degree.  While I may continue to write more in this space, I need to take a step back and figure out what the point of that might be.  If I return, I suspect it will be in a completely different form than the line by line reading I've been doing. 

But before signing off, I'd like to pay my respects to the reader whose joke about live-blogging Difference & Repetition was so funny that it launched 127 posts -- 1 per page of text we've covered -- over the past two years.  Hooocoodanode!?  As I've said before, here at FPiPE we are deeply committed to answering reader's questions so thoroughly that they are bored to death by the response.  So when I was asked, "What's the point of reading a book where you only cover 0.174 pages per day?" we all knew that no simple answer was going to satisfy ... 

One thing is fairly obvious from the start.  You don't read a book like this looking for information.  If the goal were to efficiently transmit knowledge -- either factual knowledge or description of a metaphysical system -- the book would be a complete and total failure.  I mean, 0.174!?  This is the kind of baud rate you expect from a dial-up opera.  

It's natural at this point to ask questions like, "What is Deleuze's thesis?"  or, "What is his metaphysical system like?"  And of course we can surely think of various ways to summarize the thesis and describe the system.  Actually, I could probably reel off quite a list of possibilities here.  Every thing contains the entire universe within it. The core building block of reality is difference, not identity, though this can't really be described as a block.  Difference thought in itself is a verb -- the act of differentiating.  The only things that repeat are questions.  Thought has been civilized and colonized, domesticated, by an obsession with paternity and property.  Etc ... 

These are all fine as far as they go, but I think they are mostly unhelpful in explaining what you get out of a book like this.  Many people, even many philosophers, seem to think of the history of philosophy as a sort of series of propositional flash cards -- Plato said X; Kant believed Y; Heidegger thought Z.  If you're feeling ungenerous, you might look at these cards as just a pile of disorganized beliefs that show no progress towards truth. The whole series has no universally accepted criterion for data and no first principle of reasoning, and is therefore a bunch of wanky and meaningless speculation.  Or you might look at the cards as a progression of ideas structured by arguments and refutations, giving it at least some sense of continuous forward motion even if it's not clear where it's headed.  Either way, the goal in studying philosophy is to know what each philosopher believes.  We extract this information and write it down on a flash card and compare that to what other philosophers believe and what we ourselves believe.  

This isn't the only approach to learning, however.  In fact, when you think about it more, learning mere information is actually the very shallow end of the pool of thought.  Do you remember memorizing state capitals in middle school?  I imagine being able to state propositionally what we think Deleuze believes is about that important.  Even to ask why he believes what he does, what are his arguments for believing one thing versus another, is to miss the point of the book.  In fact, I don't think there's been anything so far that we could describe as an argument or justification.  The problem is that a learning that focuses on either what or why we should believe something is one that implicitly takes for granted the idea that the goal of learning is reproducing a true belief in our heads.  We know by now that Deleuze isn't interested in truth as a correspondence between our ideas and the world, or in the reproduction of a truth in the form of its transmission from one brain to another or from the world to our brain.  He doesn't have some packet of information or belief to give us.  There's nothing we can copy from the model Deleuze provides.

Luckily, we have another familiar model of learning.  How do we learn to swim, or play tennis, or meditate?  Learning a craft or a skill is not the same as learning knowledge.  Learning how is not learning that.  How exactly these are different is a difficult question, but I think we'd all accept that we don't learn a skill by copying a model in any simple sense.  As we saw with swimming, you don't learn by standing next to the pool and imitating the instructors' motions, even if they give you feedback about how well your imitation matches what they think it looks like when they do it.  Instead, something about their words and motions resonates with some process of your own and enables you to feel how to manipulate the medium for yourself. 

Deleuze wants to teach us how to swim.  Or perhaps how to meditate.  He wants to teach us a craft of thinking, how to fabricate thoughts for ourselves, how to sculpt in whatever medium thoughts are made of, or at least to appreciate this sculpture as an art form.  This has nothing to do with truth or belief.  Or perhaps, as Deleuze put it in the case of the Eternal Return, it has to do not with true beliefs, but with the truth of belief.  What do the concepts we fabricate and believe in enable us to feel and do once we bring them into the world?  What new experiences do they open up?  Deleuze changes the idea of what we are meant to learn in philosophy.  We are not supposed to "do as he does" but to "do with him", and to evaluate what happens to us for ourselves.

Once we relate the book to this other model of learning, we immediately appreciate why it takes so long to read.  Learning skills requires practice, and practice takes time.  We already accept this notion with physical skills.  No one expects to play the piano or win the US open or swim the English channel after watching just one youtube video. You have to watch 3 or 4 at least.  The repetition isn't accidental or optional, but necessary, precisely because you are not repeating the same thing every time.  But we don't normally think of learning mental skills in this light, at least beyond early education.  In fact, we don't normally think of intellectual activities as skills at all.  So we find ourselves surprised when they take time to develop.  We wonder whether there isn't some short cut.  For some reason, we assume that knowledge can be transmitted instantaneously.However, if our goal is to learn a skill, or acquire a habit of thinking, that changes our lives, then we need time in which to practice.   

This then is the point of reading 0.174 pages per day, the same point as meditating for 20 years -- to practice.  Of course, this practice isn't self-justifying or self-contained; the ultimate goal is to live better and more fully through the practice.  This is a pretty open-ended goal though, so we are never really done practicing unless we are also done growing and changing.  A good teacher may help speed this practice on its way, but there will never be a substitute for time.  No matter how fast we imagine acquiring a skill, there will still be steps to move through that define a qualitative change from before to after.  Practicing skills literally requires the form of time.  Ultimately, for us humans, maybe this takes the form of the time required for neural plasticity.  What we're after in this form of learning is not the production of a particular brain state, but a whole process of rewiring that changes the structure of the brain.  New connections need to grow in order for us to make something different happen.  It's not so much that this takes time, as that it is time -- the possibility of change, the practice of thinking differently.

Socrates was a Sophist

In the same way that Marx described his own philosophy as "Hegel stood on his head" we might think of Deleuze as an inverted Platonist.  That, at least, is the upshot of this very last section of chapter 2 (pg. 126-128).  We spent an enormous amount of time on Plato generally and especially The Sophist back in chapter 1, and I promise we will not reprise all that.  The thing that most interests me in this last section is actually the image Deleuze set up; I think it might be the first time in his philosophy we get a clear image of the rhizome.

Plato's scheme is based on a set of abstract Ideas.  Particular things in the real world are copies of those Ideas that resemble the model to one degree or another.  The Ideas, being abstract, are self-identical; they are identity-in-itself.  The copies are different from those model identities, but this difference is calculated along a chain of lost or modified identity.  In other words, difference in this scheme is (partial) lack of identity.  We can see how this sets up a tree image.  The Idea is the trunk which then branches off into different numbers of copies at different distances from the ground.  As we get further and further from the ground the branches get more numerous and the resemblance to the initial Idea declines.  Eventually we reach the leaves at the tips of the branches, which are all things that claim to be derived from that single trunk.  The goal of the scheme is to organize the leaves we find in the world in order of their distance from the trunk.  That way we figure out which things are closest to the Idea, who is closest to the true Lover, Statesmen, Philosopher, etc ...

In Chapter I, we suggested that Plato's thought turned upon a particularly important distinction: that between the original and the image, the model and the copy. The model is supposed to enjoy an originary superior identity (the Idea alone is nothing other than what it is: only Courage is courageous, Piety pious), whereas the copy is judged in terms of a derived internal resemblance. Indeed, it is in this sense that difference comes only in third place, behind identity and resemblance, and can be understood only in terms of these prior notions. Difference is under- stood only in terms of the comparative play of two similitudes: the exemplary similitude of an identical original and the imitative similitude of a more or less accurate copy. This is the measure or test which decides between claimants

[

Note that this tree is not the same as the Aristotelean or Linnaean tree constructed from genera and species.  This is more like a family tree of descent from a common ancestor, a demonstration of how distantly we are all related to Genghis Khan.  These two trees would only get related after the discovery of DNA coding.  Which seems significant in some way I'm unable to put my finger on right now.  Deleuze touched on this point in chapter 1

[The method of Platonic division's] point of departure can therefore be either a genus or a species, but this genus or this large species is understood as an undifferenciated logical matter, an indifferent material, a mixture, an indefinite representing multiplicity which must be eliminated in order to bring to light the Idea which constitutes a pure line of descent. The search for gold provides the model for this process of division.

]

Deleuze asks us to back up one step and see the deeper problem this scheme is meant to address.  We don't see the trunk or know exactly how the branches run.  Those are abstractions.  We just see the leaves.  And our initial goal isn't really to try and organize all possible leaves into a single total tree.  We're talking about Plato here, not Aristotle.  Really we just start off wanting to separate the good leaves from the bad leaves.  Who should we follow as a better Statesman to guide us?  The merchant, the farmer, the shepherd?    Who should we listen to as a Philosopher?  The sage, the orator, the dialectician?  The true goal here is to figure out how to live the good life, how to select the better option.  Philosophy started as a question about how to live well, not about how to correctly represent the truth.  It begins with the friend or the lover of wisdom, not its possessor.  Which is to say that it ultimately poses a moral question.  Today we might call it an ethical or ethological question, but it amounts to the same thing.  We want to mark a difference between living well and not.

Plato's scheme answers to this problem.  It uses the idea of a model and copy to try and organize these questions about the difference between the good and the bad life.  The Ideas aren't fundamentally there to represent reality.  They are a way to help us choose.  The seemingly central distinction between model and copy is actually just a device to help us with this problematic moral distinction between good and bad.  

More profoundly, however, the true Platonic distinction lies elsewhere: it is of another nature, not between the original and the image but between two kinds of images [idoles], of which copies [icones] are only the first kind, the other being simulacra [phantasmes]. The model- copy distinction is there only in order to found and apply the copy- simulacra distinction, since the copies are selected, justified and saved in the name of the identity of the model and owing to their internal resemblance to this ideal model. The function of the notion of the model is not to oppose the world of images in its entirety but to select the good images, the icons which resemble from within, and eliminate the bad images or simulacra.

[

The translator's footnote is helpful here in connecting this back to our more typical understanding of Plato as talking about the distinction between (mere) appearance and reality.

The Greek term phantasma, for which Deleuze often uses the French equivalent phantasme, is frequently rendered in English translations of Plato as 'appearance'. I have preferred to use 'phantasm' in order to retain the connection with 'phantasy' in the preceding section of the text, where Deleuze uses the same French word as equivalent to the Freudian term Phantasie.

] 

What appears then, in its purest state, before the logic of representation could be deployed, is a moral vision of the world. It is in the first instance for these moral reasons that simulacra must be exorcized and difference thereby subordinated to the same and the similar. 

How can just thinking about our life make a difference in our life?  How can it help us distinguish between the things that really matter and the fake shit in our world, our society, and our own minds.  This question is so deep it knows no bottom.  It is the absolutely universal core of philosophy and religion.  Plato actually responds to this question.  He feels its full weight the moment before philosophy gets converted into a mere search for truth and accuracy.  The only problem is that he tried to answer the question, once and for all, rather than allow us to keep asking it.  He decided that the best way to test the different options before us was to see how closely they resembled an ideal model.

For this reason it seems to us that, with Plato, a philosophical decision of the utmost importance was taken: that of subordinating difference to the supposedly initial powers of the Same and the Similar, that of declaring difference unthinkable in itself and sending it, along with the simulacra, back to the bottomless ocean. However, precisely because Plato did not yet have at his disposition the constituted categories of representation (these appeared with Aristotle), he had to base his decision on a theory of Ideas. 

I find this a really intriguing interpretation of Plato.  He does bring up the question of how to live and examine our life in such a way that it is not determined in advance by the received wisdom of state, religion, or tradition but instead depends on our own thinking to make a difference.  This is what makes him the father of philosophy.  But then he has no frame of reference here, he's like a child who wanders into the middle of a movie and wants to know ... my point is he gets lost in the immanent universe of possibilities that thinking of difference creates.   As a result, he betrays it, shackles it to the Same, and restores the transcendence he just deposed.  The positive and creative difference we can make by thinking is dismissed as mere phantasy, and the only difference that survives his moral test becomes the difference between model and copy.  In short, we get reduced to merely remembering the Ideas we forgot in our last life. 

Once Plato works his magic, any difference that doesn't correspond to one or another model, that can't be located somewhere on the tree of copies branching from the trunk of an Idea, becomes a mere appearance.  It's not even a degraded copy.  It doesn't resemble anything real at all.  It's just illusion, confusion, seductive simulacra.

Simulacra or phantasms are not simply copies of copies, degraded icones involving infinitely relaxed relations of resemblance. The catechism, so heavily influenced by the Platonic Fathers, has made us familiar with the idea of an image without likeness: man is in the image and likeness of God, but through sin we have lost the likeness while remaining in the image ... simulacra are precisely demonic images, stripped of resemblance.

The difference we can make ourselves is monstrous.  Dangerous.  Demonic.  It only proves that we've fallen from the Ideas, that our thoughts no longer resemble anything pure, anything that exists truly and in-itself.  Still, we might wonder where all these illusions come from and why there are so many of them.  Also, why are we always so seduced by them?  Why is it so hard to figure out what is a copy and what a simulacra?  Why do the Ideas have to be hidden?  

Actually, the simulacra too have a logic to them.  It's the illegitimate logic of mixture.  The simulacra are the impure and monstrous result of the Ideas inter-breeding.  Instead of being the direct descendant of an Idea, the simulacra is the bastard stepchild of more than one idea.  Plato was obviously not a goer.  Or a golfer.  These orgies are obviously immoral.

Or rather, in contrast to icones, they have externalised resemblance and live on difference instead. If they produce an external effect of resemblance, this takes the form of an illusion, not an internal principle; it is itself constructed on the basis of a disparity, having interiorised the dissimilitude of its constituent series and the divergence of its points of view to the point where it shows several things or tells several stories at once.

Deleuze is clearly referring here to the structure of the disparate that we've spent so much time unravelling.  The simulacra is built from the resonance of two series of difference, the two chains of descent originating in two different Ideas.  They don't resemble any one thing, don't live within one Idea as a degraded copy, but contain within themselves the intersection of two lines of descent.  For Plato, this is the source of their power to confuse.  It's so hard to distinguish the bastard step-child from the dunce prince who fell so far from the family tree.  For Deleuze, the inter-breeding of Ideas is where all the action is at.  The Simulacra aren't bad copies of the Idea-in-itself.  They are always original 'copies' of difference-in-itself.

Does this not mean, however, that if simulacra themselves refer to a model, it is one which is not endowed with the ideal identity of the Same but, on the contrary, is a model of the Other, an other model, the model of difference in itself from which flows that interiorised dissimilitude? Among the most extraordinary pages in Plato, demonstrating the anti-Platonism at the heart of Platonism, are those which suggest that the different, the dissimilar, the unequal -- in short, becoming -- may well be not merely defects which affect copies like a ransom paid for their secondary character or a counterpart to their resemblance, but rather models themselves, terrifying models of the pseudos in which unfolds the power of the false

You can see how this sort of problem would keep Plato up at night.  The whole point of the Ideas was to enable us to distinguish the component parts of mixtures and thereby dissolve them.  A simulacra isn't a thing in itself, but just a confused mix of multiple things, a superposition of the copies of several Ideas that should be understood by breaking it down into its component Ideas.  But if mixture itself has an Idea, if there is a concept of difference in-itself, that fucks up our plan.  We would no longer just be worried about what true ideas are involved in a particular real mixture, but would have to worry about Ideas that are inherently false, confused, and contain a multiplicity of other Ideas within themselves.  Imagine your horror when you climb all the way out to the end of a tree branch and find that the leaf growing there also belongs to another tree.

And now we've reached the point I've promised at the beginning.  The rhizome is constructed as the crossing of multiple branching trees.   Here is a clumsy diagram to illustrate the idea.  Obviously, the focus is the complex overlap in the center.

And here's a more aesthetically interesting one from circa 2005.  Apparently I've had this idea before in a different form.

Compare these to actual rhizomes

And the great rhizome within (courtesy of Ramón y Cajal)

Each of Plato's Ideas creates a tree of descent.  At the limits of those trees, after following the most distant branch to the most degraded copy, the leaves of one tree flutter in the breeze and become confused with its neighbor.  We hoped that if we worked hard enough, that confusing patch that appears at first glance to contain three or four leaves could be disentangled, and we could see how every leaf belongs to one and only one tree.  But in fact, maybe some of these leaves are attached to more than one tree. The trees aren't completely separate and pure unities, but actually form an inherently tangled network with points of intersection.  A rhizome.    Briefly, this makes us wonder whether we shouldn't reverse our perspective and take those points of intersection as a new sort of trunk, so that we can retrace the whole structure in reverse.  If we started with a given leaf, perhaps we could trace it back to the two trunks it shares.   Could each shared leaf serve as a model, a model of the shared, the split, the Other, of which the original trunks would then be the most distant copies?  Or perhaps instead of leaves, we should think of the trunks extending underground into a root system that connects one to the next?  

But no, no matter how we phrase it, this will never work.  The damage has been done once we start to find points of intersection.  With the network, any leaf could be traced to any Idea, perhaps even to every Idea, depending on how the topology of the network is arranged.  We can no longer interpret the whole as a set of intersecting lines of descent from anything.  Nothing can be a copy of anything else.  All we have is the network.  It's all a confused tangle and our only approach is to walk along these paths and try to find out which leaf connects to which and in what order.  

Does this not mean, thirdly, that simulacra provide the means of challenging both the notion of the copy and that of the model? The model collapses into difference, while the copies disperse into the dissimilitude of the series which they interiorise, such that one can never say that the one is a copy and the other a model. Such is the ending of the Sophist, where we glimpse the possibility of the triumph of the simulacra. For Socrates distinguishes himself from the Sophist, but the Sophist does not distinguish himself from Socrates, placing the legitimacy of such a distinction in question.

What becomes, then, of our original problem?  How can our thinking make a difference that distinguishes between the good and the bad life?  Deleuze here echoes his description of difference in-itself from the very first pages of chapter 1.  The simulacra is difference in-itself, it contains difference within itself, but this difference just leads us to other differences, to another simulacra, as we circulate endlessly around the network.  Difference is made in this movement, as we traverse the network and connect things up.  At the limit, we have to map the entire burrow to distinguish the location of a point within it.  From the point of view of this totality, there is nothing special about the location.  There are no endpoints, and everything is on the way to everything else.  But notice also that in some sense the entire network is within our starting point.  It splits and splits within itself to produce all these other differences that might lead it back to itself.  Any random starting point can become the entire rhizome.  In this case the difference between good and bad is made immanently, in terms of blocked passages or the ability to pass.  And the eternal return is the ability to walk the entire rhizome and return to your utterly random starting point.  Traveling in place.

It is no longer the Platonic project of opposing the cosmos to chaos, as though the Circle were the imprint of a transcendent Idea capable of imposing its likeness upon a rebellious matter. It is indeed the very opposite: the immanent identity of chaos and cosmos, being in the eternal return, a thoroughly tortuous circle. Plato attempted to discipline the eternal return by making it an effect of the Ideas - in other words, making it copy a model. However, in the infinite movement of degraded likeness from copy to copy, we reach a point at which everything changes nature, at which copies themselves flip over into simulacra and at which, finally, resemblance or spiritual imitation gives way to repetition.

And One More Thing ...

When we get out to the furthest reaches of Deleuze's ideas, out in the vicinity of the eternal return and univocal being, I always feel like my understanding of them starts slipping away.  There's a danger that no matter how much I write about these ideas I have missed some key ingredient in explaining them, and that the most crucial point has somehow been left unsaid.  In some ways I think that inducing this feeling of continual grasping and slipping and regrasping is part of Deleuze's style.  After all, this is a pretty good description of the relationship between difference and repetition and difference.  Nevertheless, I still feel the need to go back over the system Deleuze has been describing once more and specify exactly where the slipping happens and the circle does not quite close.  I'm imagining this in terms of a question about the structuralist ontology that started with the discussion of coupled oscillators: where did the initial oscillators come from?

My initial answer to this was to imagine that the forced movement of the coupled system itself constituted a new oscillator that could then be coupled to other oscillators that had been built in the same fashion.  Let me first explain how I saw that working.

The system constituted by the eternal return begins with two series of differences.  If we think of these as series of the simplest possible binary difference 01010101 played out over time, then we have a square wave oscillator.  The two series then begin to resonate because they share a singularity.  That is, their differences are distributed in such a way that they become related.  This singularity, or dark precursor, is a strange beast.  In some sense shared by the oscillators; the nature of a singularity is its capacity to make two different things resonate through some sort of interaction.   We might cheekily say that a singularity is never alone, never naked.  But the two oscillators don't share it by being two examples of it.  They don't partially resemble one another or the singularity by virtue of this sharing.  The singularity is not a property of some abstract thing called "oscillators in general" that happens to be shared by these two oscillators in particular.  Instead, the singularity is created by the relationship between the series as much as it creates it.  It's the attractor, as it were, of the dynamic feedback loop that develops between two different oscillators.  Our square wave analogy starts to fail at this point, since the frequencies of the oscillations are in fact properties of all oscillations, and the points of resonance can be calculated in advance.  Remember that an oscillator here is just standing in for a series of differences, which could be chemical concentrations, or phonemes, on the one side, and digital differentiation, or meanings on the other.

The singularity in itself is always hidden (hence dark) and we only know of its existence because the original series begin to resonate, and in fact resonate in a feedback loop that actually welds them into a single system.  At this point there is a forced movement that goes beyond the two original oscillators, which now appear as merely the two distinct parts necessary for this new movement.  The identity of the individual oscillators dissolves into this overall system (hence the death instinct connection).  So it seems that we've produced another, higher level, oscillation that has an alternation between the two original oscillators as its 0 and 1 levels, so to speak.  This oscillator could presumably then resonate with other second degree oscillators constructed in the same manner in order to produce third degree oscillators, and so forth.  The endpoint of the system becomes the conditions from which we begin another system, just like the third synthesis of death liberates a desexualized energy that can later become bound in the first synthesis of habit.

I still think this describes the basic structure of the "systems constituted by the eternal return" as Deleuze puts it.  After thinking about it more though, I believe that the way I've laid it out here actually inverts the direction we are meant to read it in.  Instead of building larger and larger units on top of the external interactions of existing units, we should think about smaller and smaller differences being created within an existing difference.  The point is not that the interaction of two differences creates a new higher level difference.  We should see it running in the opposite direction. The point is that every series of difference is already split into two series inside itself.  Instead of 1+1=2+1=3+... we have 1=1/2+(1/4+(1/8+(...)))

In a way, you might say that this perspective is no different from the first one, and that it was implicit in it.  If any two oscillators can form a third composite oscillator, then it stands to reason that one given oscillator can be broken into two smaller sub-oscillators.  Either way, it's turtles all the way down, right?  Actually though, from the perspective of the role of identity and repetition in the structure, the fractal infinite that plunges within is a profoundly different way of looking at things.  The first way of looking at things takes the identity of the initial unit for granted as an unquestioned point of origin.  The form of repetition that defines the whole is then the reproduction of this unit.  This numerical repetition requires a starting point, and it also requires someone to do the counting, something outside the system to keep track of the number of increments (exactly the problem we saw with Hume at the start of the chapter).  If we move in the other direction and see whatever series we happened to start with as already divided within itself we avoid both of those problems.  There's no unity to take for granted and no assignable origin because every starting unit is already more than one thing.  And the form of repetition is no longer the contingent and external addition of another copy, but the essential internal repetition of a process of continuing fractionation.  There's a huge difference between whether the One is taken for granted at the beginning or whether it is only produced at the very end of the process, as the limit of difference differentiating itself again and again.  

I think it's this fractal recursion that Deleuze has in mind with difference-in-itself.  Difference is immanently within any system.  You might even say that difference is within time, in the sense that one is already two, etc ...  There is no atomic difference.  Looking at things this way definitely produces some strange paradoxes.  If difference is constantly differentiating within itself, then we can (and must) begin in the middle, at any level.  To understand that single arbitrary starting point though we have to move up and down the levels of difference infinitely in both larger and smaller directions.  That single difference becomes an entire universe.  In other words, every difference-in-itself, repeats a whole, and they all repeat a different whole but the same way.  So we say that the One only appears at the end of the chain, after thinking through an infinite series of differentiations, but in some sense it is there implicitly at the beginning in each and every difference.  And the same type of paradox appears with our original question: where did the initial oscillators come from?  In fact, there are no "initial oscillators", if we mean by this some smallest unit that is indivisible.  Every oscillator has two series within it, ad infinitum.  But by the same token, every level we might start with is inside of some other oscillator, in an endless series of nested vibrations.  In which case the identity of that oscillator is actually constructed by something outside itself, of which it is merely a part or an internal differentiation.  The immanence of difference refers immediately to an outside.  And the singularity that differentiates difference refers to all the other singularities in an indefinite multiplicity.

Everything is upside down in the world of the eternal return.  The only anything is everything.

The Abyss

As we would expect, this section ends with another discussion of the eternal return.  This time Deleuze presents it as the law governing the system of difference he has been describing.  Or perhaps it would be better to describe it as a law that continually produces differences as coexisting, and therefore constituting a system, which then appears to follow this law.  It's a strange sort of law that governs the production of chaos.  A law that allows for, even requires, everything, and prohibits nothing, even the things that would seem to break it.  The governor and the governed, the law of eternal return and the system derived from it, are not different things here.  They too coexist in this most abstract chaos.

It is therefore proper to say that the system excludes the assignation of an originary and a derived as though these were a first and second occurrence, because the sole origin is difference, and it causes the differents which it relates to other differents to coexist.  It is under this aspect, without doubt, independently of any resemblance, that the eternal return is revealed as the groundless 'law' of this system. The eternal return does not cause the same and the similar to return, but is itself derived from a world of pure difference. Each series returns, not only in the others which imply it, but for itself, since it is not implied by the others without being in turn fully restored as that which implies them. The eternal return has no other sense but this: the absence of any assignable origin - in other words, the assignation of difference as the origin, which then relates different to different in order to make it (or them) return as such.

The repetition in the eternal return is as close as we can come to the principle of identity.  It's as close as we can come to embracing the chaotic totality of all the differences as the One, or the Same, or the Universe.  As we've seen though, to define a unit of repetition, we actually need to see it three times, first as thing, then as process, then as the difference between and coexistence of process and thing.  Sure enough, the final paragraph of this section lays out three versions of the identity of the eternal return, as if to illustrate that "the" eternal return is already a misnomer.  The only thing that is the Same in the eternal return is its perpetual difference from itself.  

In the first sense, the Same designates a supposed subject of the eternal return. In this case it designates the identity of the One as a principle. Precisely this, however, constitutes the greatest and the longest error. Nietzsche correctly points out that if it were the One which returned, it would have begun by being unable to leave itself; if it were supposed to determine the many to resemble it, it would have begun by not losing its identity in that degradation of the similar. Repetition is no more the permanence of the One than the resemblance of the many.

This first sense puts to rest the idea that eternal return is a law in the normal sense of a principle from which the behavior of a system else can be derived.  Laws always require the permanent transcendence of the form of the law, and the pre-existent identity of the contents to which it applies.  In this sense the law is what always reappears or stays put in every possible incarnation of the world.  It doesn't matter whether we conceive of this law as scientific or spiritual.  I think the description above alludes to the law of the Hegelian dialectic in which Absolute Spirit does its endless backflip.  Neither form of law does anything to account for why there is any world at all for them to apply to.

Alternatively, in the second sense, the same and the similar are only an effect of the operation of systems subject to eternal return. By this means, an identity would be found to be necessarily projected, or rather retrojected, on to the originary difference and a resemblance interiorised within the divergent series. We should say of this identity and this resemblance that they are 'simulated': they are products of systems which relate different to different by means of difference (which is why such systems are themselves simulacra). The same and the similar are fictions engendered by the eternal return. This time, there is no longer error but illusion: inevitable illusion which is the source of error, but may nevertheless be distinguished from it.

The second sense of the eternal return is an improvement.  It takes a world of difference as its starting point, and asks instead why we have any idea of unity and law at all in such a world.  Instead of wondering why God made the world, we wonder why the world ever invented the illusion of God.  While I imagine it's mostly related to Deleuze's definition of the simulacra as a sign that "interiorizes the conditions of its own repetition", the mention of simulation is intriguing.  Here, we cannot define simulation in our usual sense as a fake or virtual or abstract version of a thing.  There is no original thing here to be copied or falsified or abstracted.  The simulation actually has to create for itself the unity of the thing simulated by coupling differences in the subject running the simulation to differences in the object simulated in such a way that the two series of differences resonate together.  Like I say, I find this intriguing in the context of the brain, though I'm not totally clear on what all it might mean.  Presumably in a biological context, resonating together means something like the two being connected so that they allow for the simulation to continue reproducing itself.

Finally, in the third sense, the same and the similar are indistinguishable from the eternal return itself. They do not exist prior to the eternal return: it is not the same or the similar which returns but the eternal return which is the only same and the only resemblance of that which returns ... Although it is the source of the preceding illusion, it engenders and maintains it only in order to rejoice in it, and to admire itself in it as though in its own optical effect, without ever falling into the adjoining error.

In the final sense, we reach the idea that the only unity in the world is the eternal return itself.  Not as a originary law, or a derived simulation, but as a sort of self-simulation, a simulation that produces itself as it simulates.  The final lap around the wheel of repetition always has this dizzying quality of self-reference to it.  Here, it's as if the reality of the One, the Same, Universe as a Whole, were literally brought into being only by my thinking it or speaking it, even though the only way to do that is to embrace the infinite propagation of differentiation that includes even this speaking.  And etc ... The transformative caesura.  The nonsense within sense.  Diagonalization.  Or the eternal return, like Narcissus, falling in love with a world that just happens to be itself.  It gives a whole new dimension to the idea of the abyss gazing back into you.

Chaos contains all the differences

Finally we come all the way back to the third synthesis.  The synthesis of the form of time as a static entity, conceived as a whole once and for all.  The time crystal.  But of course, Delueze's whole point is that this form, while static, while always the same, is fractured, split between before and after and continually splitting within itself.  It's frozen, but vibrating, as Feldman put it.  The incredible chain of events that led to the tremendous now just keeps on happening; this moment that contains everything is just like every other.

This final presentation of the third synthesis though, seems to focus on the peculiar whole created by this infinite series of splittings. In other words, the entire movement-in-place I just tried to describe.  It amounts to chaos.  A totality which contains everything but is never self-identical.  An infinite potential for difference, for further differentiation, conceived as an identity of sorts.  The "beyond" that all the series point to when they resonate together.  The thing, so to speak, that all the differences have in common.  It's hard to talk about this without lapsing into poetry.  That's probably why the muslims use mathematics. 

The essential point is the simultaneity and contemporaneity of all the divergent series, the fact that all coexist. From the point of view of the presents which pass in representation, the series are certainly successive, one 'before' and the other 'after'. It is from this point of view that the second is said to resemble the first. However, this no longer applies from the point of view of the chaos which contains them, the object = x which runs through them, the precursor which establishes communication between them or the forced movement which points beyond them: the differenciator always makes them coexist.

All the series of differences, what we were calling oscillators, are enmeshed in this chaos.  But I think it's crucial not to see this chaos as the mere collection of these different series.  Because in that case, what would be capable of holding these differences together as a collection, instead of a mere unrelated diversity?  Only the concept of identity.  If we treat the series as things, then the chaos becomes the abstract collection of all possible things.  However, if this chaos is a space of possibility, it's not one that pre-exists the things, but one that is drawn by the way they interact.  If this seems to introduce a circularity into the notion of possibility, where somehow the real things arise at the same time as their own possibility, we are on the right track.  Think again of the egg metaphor.  Things differentiate themselves within the developing embryo.  Yet this differentiation depends on some singularity, some potential phase transition that links one set of differences to another set across different scales.  

I think this is why Deleuze relates the chaos which contains all the series to a forced movement created by the resonant singularity of the dark precursor.  The forced movement is chaotic, destructive of individual identity.  It's the death instinct at work.  In this sense the chaos is created by the communication of the series.  But in another sense the singularity that lets the series communicate through resonance has an existence apart from them.  Many different coupled systems might share this same phase transition.  If we think of the chaos as the forced movements that affect the series when they resonate, then it comes after the series.  If we think of it as the collection of singularities that structure this resonance, then it comes before the series.  Ultimately the chaos is a circular place where the law of linear logical causality no longer applies.  There is no before and after either in temporal or a causal sense.  The chaos is neither derived from the series, nor are the series derived from the chaos; there is no first term here.

I've already kicked the analysis up a meta-level here though, where Deleuze only seems to hint at this causal circularity at the very end of the section. First, he discusses the chaos as the coexistence of before and after in time.  This is what we saw as the pure past of the second synthesis.  It is a space beyond any empirical memory that permits two experiences to be situated within it in such a way that one is conceived as a memory of the other.  Both past and present have to independently coexist in the space before we can experience something as the past of this present (ie. an experience marked as memory).  Deleuze illustrates this pure past of temporal coexistence using the classic Freudian idea of childhood memory.  

When Freud shows that a phantasy is constituted on the basis of at least two series, one infantile and pre-genital, the other genital and post-pubescent, it is clear that the series succeed one another in time from the point of view of the solipsistic unconscious of the subject in question. The question then arises how to explain the phenomenon of 'delay' which is involved in the time it takes for the supposedly original infantile scene to produce its effect at a distance, in an adult scene which resembles it and which we call 'derived'.

We know already though that nothing can be "derived" from anything else in Deleuze's world.  There is no first term from which to derive it, no model from which to make a copy.  Derivation can only mean establishing a relationship between two things.  Or rather, establishing a circular relationship that produces two things, differentiating them at the same time as holding them together.  This is always the problem with memory -- given that any two moments have many things in common, we have to explain how we find some particular thing relevant to both moments in a way that joins them.  In the case of Freudian phantasy, Deleuze will interpret the childhood event as exactly the thing capable of joining those two moments and creating our subjective desire.

It is indeed a problem of resonance between two series, but the problem is not well formulated so long as we do not take into account the instance in relation to which the two series coexist in an intersubjective unconscious. In fact the two series - one infantile, the other adult - are not distributed within the same subject. The childhood event is not one of the two real series but, rather, the dark precursor which establishes communication between the basic series, that of the adults we knew as a child and that of the adult we are among other adults and other children.

Remember that in this passage we are talking about erotic phantasy.  You might think about it in Freudian terms as a question of whether your wife looks like your mother, or, on the contrary, it's your mother who looks like your wife.  Do we find our lover attractive because they remind us of another we had, and another we had, in an unbroken chain stretching back to our first Oedipal fixation?  Or is desire more mysterious than this, and it's our current paramour who casts a light into the past to illuminate what we didn't even realize we loved about her predecessors?  The idea is that time doesn't run in just one direction here, but circularly.  Memory isn't an objective reproduction, but a subjective reconstruction.  It's mutability means that my current subjective experience is shaped by a past that it, in turn, shapes.  The subject I am now and the subject I was then are not the same subject.  Both are drawn together from what you could only describe as an intersubjective field.  The experience of memory with its consciousness of before and after, only occurs when these two subjects reflect back and forth off of each other, or resonate together.  The childhood event triggers this resonance and produces the identity of the remembering subject, which now appears to span time.

There is no question as to how the childhood event acts only with a delay. It is this delay, but this delay itself is the pure form of time in which before and after coexist. 

The phantasy is the manifestation of the child as dark precursor. Moreover, what is originary in the phantasy is not one series in relation to the other, but the difference between series in so far as this relates one series of differences to another series of differences, in abstraction from their empirical succession in time.

Maybe now it makes more sense to return to the meta-level I mentioned earlier.  The structure of the pure past (with its memories coexisting in time), is the same as the ontological structure of chaos (in which all the series coexists logically), and both are the same as the eternal return (which makes each of Deleuze's syntheses of repetition coexist as part of 'the same' idea).  Or this, I hypothesize, is why the second chapter has such a strange structure that repeats at biological, psychological, and ontological levels.  

Regardless of the merit of this stylistic hypothesis, it's clear that we've been building towards a definition of the repetition of "the Same" phrased only in terms of difference.  Whatever we call it  -- the unconscious, chaos, the third synthesis, or univocal being -- the point is that all things only belong to this largest totality on account of their difference.  All they share, their whole common ground that brings them together as a coexistent group, is difference.  

For it is in the same movement that the series are understood as coexisting, outside any condition of succession in time, and as different, outside any condition under which one would enjoy the identity of a model and the other the resemblance of a copy. When two divergent stories unfold simultaneously, it is impossible to privilege one over the other: it is a case in which everything is equal, but 'everything is equal' is said of the difference, and is said only of the difference between the two.

The coexistence of things is not based on one common root from which they are derived, either through resemblance or analogy.  Even their Being is not "the Same".  Instead, they are held together in sharing a process of differentiation, and they express this, their coming into being, their becoming, in a single and same resonant voice.  I can only presume it says "ommmmm".

The Refrain

My division of this third repetition of the three figures of repetition (pg. 116-126) get a little more nebulous after the note on Proust on pg. 122.  Deleuze seems to continue discussing the dark precursor in a literary context, but also smoothly shifts into discussing the system that the dark precursor connects or animates.  In a sense I think this blurring may be inevitable and my divisions arbitrary.  It's clear that somehow the creator, the creation, and the space of possible creation (to use yet another way of slicing the three repetitions) are all what Deleuze would call 'the same', that is, are connected series of difference.

In any event, the next section is interesting because it anticipates Deleuze's later ideas about nonsense (developed in The Logic of Sense).  As you can probably guess by now, Deleuze doesn't look at non-sense as the absence of sense or its negation.  In fact, we should probably write it as ?-sense instead of non-sense.  The idea is roughly that the perpetual movement of the dark precursor, the fact that it only exists as a thing by being a process of vortical flow, is a privileged instance which illustrates how the whole system is constructed.  

The structure seems vaguely reminiscent of the various diagonalization arguments in mathematics (Cantor, Gödel) that use self-reference to define a perfectly clear example of something which falls outside of a system you assumed was complete.  In Gödel's case specifically, it doesn't help to add this new item back to the system either, because another escaping instance can always be constructed by the same means.  A special part constructed within the system demonstrates that something will always fall outside it; the system is inherently incomplete.  Likewise, the linguistic dark precursor is an instance of self reference that inherently falls outside the system of sense -- it is nonsense.

The question of whether psychic experience is structured like a language, or even whether the physical world may be regarded as a book, depends upon the nature of the dark precursors. A linguistic precursor or an esoteric word does not have an identity by itself, not even a nominal one, any more than its significations have a resemblance, even an infinitely relaxed one: it is not just a complex word or a simple gathering of words, but a word about words which is indistinguishable from the 'differenciator' of first-degree words and from the 'dissembler' of their significations. Its value, therefore, lies not in the extent to which it claims to say something but in the extent to which it claims to state the sense of what it says. The law of language which operates within representation excludes that possibility: the sense of a word can be stated only by another word which takes the first as its object. Whence the following paradoxical situation: the linguistic precursor belongs to a kind of metalanguage and can be incarnated only within a word devoid of sense from the point of view of the series of first-degree verbal representations.

By now, this structure is familiar to us.  However, I do like the way he sets up the problem here with those first few lines.  We want to know whether our psychological experience, or the world itself, make any sense, have any meaning, in the way a language does.  This may seem like a straightforward question at first, but as soon as you throw your starting point into question things get complicated; do we really understand how a language makes sense to begin with?  How does language get its representational ability?  When the passage continues, it makes clear that this was the problem underlying Deleuze's initial question.  We can agree with Lacan or Galileo that the psyche or the physical world is structured like a language only if we can uncover a dark precursor in those systems that works via the same mechanism of self reference.

This double status of esoteric words, which state their own sense but do so only by representing it and themselves as nonsense, clearly expresses the perpetual displacement of sense and its disguise among the series. In consequence, esoteric words are properly linguistic cases of the object = x, while the object = x structures psychic experience like a language on condition that the perpetual, invisible and silent displacement of linguistic sense is taken into account. In a sense, everything speaks and has sense, on condition that speech is also that which does not speak - or rather, speech is the sense which does not speak in speech.

The final line here is clearly another of Deleuze's koans we need to ruminate over.  He seems to be saying that everything in our inner experience and the outer world at larger means something.  It all makes sense.  But we must understand that sense is created when we find a structure where signs refer only to other signs perpetually.  This is the structuralist conception of language.  Language doesn't mean something because of some one-to-one correspondence between words and stuff, but because the language contains a structure of differential elements (phonemes and words) which can be coupled to the structure of the differential elements in an idea or the world.  Sense is only created by the structure of this coupled system of differences as a perpetually expanding indefinite whole.  It does not reside within any individual element of it.  Any single element that tries to refer to the whole process of sense gets sucked up into a problem of self-reference since it must be merely part of this whole, and can only itself acquire sense on that condition.  The dark precursor or self-referential element is like a part that tries to stand in for the whole it is part of, creating an infinite regress.  This produces nothing but paradox and nonsense at the level of the element, but shows us how sense is produced at the level of the whole system.

In this manner, the conditions under which a book is a cosmos or the cosmos is a book appear, and through a variety of very different techniques the ultimate Joycean identity emerges, the one wefind in Borges and in Gombrowicz: chaos = cosmos.

I often think that the final equation could be practically the summary of Deleuze's whole philosophy.  God = chaos.  The infinite potential.  We seek to get as close to it as we can, which we can only do by producing more of it ... in such a way that we can then continue producing even more of it.  Or as they put it in A Thousand Plateaus:

Were you cautious enough?  Not wisdom, caution.  In doses.  As a rule immanent to experimentation: injections of caution.  Many have been defeated in this battle.

Maybe now we can better appreciate the blurring I mentioned at the beginning.  The dark precursor as an element which joins and differentiates the series blurs into the whole chaotic totality that it depends on to make sense. I find the diagonalization image helpful here, or its psychoanalytic correlate: "the narcissistic libido, the reflux of the libido into the ego", the way the self takes itself as an object in the synthesis of death.  Ultimately though, Deleuze's point is less about this one special moment of self reference than the way it forms a necessary part of a three part system.  In fact, this whole section seems to be built like a fractal where we see repeatedly that the third synthesis itself has three parts, with the third part referring to the three part system created by the relation of the first two, etc ... ad infinitum.  As a result, we seem to be encouraged to shuffle around the various terms for these three parts.  So it's not only the dark precursor's resonant singularity that blurs into the forced movement of the chaotic totality it must represent as nonsense, but that chaos which also blurs into the divergent series of differences we started with.

The basic series are divergent: not relatively, in the sense that one could retrace one's path and find a point of convergence, but absolutely divergent in the sense that the point or horizon of convergence lies in a chaos or is constantly displaced within that chaos. This chaos is itself the most positive, just as the divergence is the object of affirmation. It is indistinguishable from the great work which contains all the complicated series, which affirms and complicates all the series at once.

This all sounds like quite a mouthful, but I think the intent is actually understandable however we phrase it.  Every time we're trying to draw the same tortuous, seemingly self-referential, circle that never actually closes on itself, because there was never any self there to close on.  Instead the center of the circle moves around at random, and it's speed of rotation and radius are likewise drawn by the lottery officials at Babylon.  It really gives a new dimension to the old chestnut: "God is an infinite circle whose center is everywhere and whose circumference is nowhere."

The trinity complication-explication-implication accounts for the totality of the system - in other words, the chaos which contains all, the divergent series which lead out and back in, and the differenciator which relates them one to another. Each series explicates or develops itself, but in its difference from the other series which it implicates and which implicate it, which it envelops and which envelop it; in this chaos which complicates everything. The totality of the system, the unity of the divergent series as such, corresponds to the objectivity of a 'problem'.

The Disparate

The words "disparate" and "disparity" have appeared a few times already.  So far I have been reading them as synonyms for simple difference.  But with the introduction of the dark precursor, Deleuze finally gives us a definition of the term, which makes clear that the disparate isn't the same as just an ordinary, first degree, difference.   The disparate is what happens when two different series of differences are related.  It raises difference to a power to become difference in-itself, the differentiation of difference, or second degree difference.

We call this dark precursor, this difference in itself or difference in the second degree which relates heterogeneous systems and even completely disparate things, the disparate.

After introducing the term, he immediately goes on to talk about the question of the size of a difference.  After seeing this, suddenly the earlier mentions of the disparate that we saw in the introduction and chapter 1 began to take on a new meaning for me.  Since the term has been used pretty sparingly so far, I think I'll just collect all the passages in which it has been mentioned so that we can examine it more closely. 

For it is not the elements of symmetry present which matter for artistic or natural causality, but those which are missing and are not in the cause; what matters is the possibility of the cause having less symmetry than the effect.  Moreover, causality would remain eternally conjectural, a simple logical category, if that possibility were not at some moment or other effectively fulfilled. For this reason, the logical relation of causality is inseparable from a physical process of signalling, without which it would not be translated into action. By 'signal' we mean a system with orders of disparate size, endowed with elements of dissymmetry; by 'sign' we mean what happens within such a system, what flashes across the intervals when a communication takes place between disparates. The sjgn is indeed an effect, but an effect with two aspects: in one of these it expresses, qua sign, the productive dissymmetry; in the other it tends to cancel it. The sign is not entirely of the order of the symbol; nevertheless, it makes way for it by implying an internal difference (while leaving the conditions of its reproduction still external). 20

Signs involve heterogeneity in at least three ways: first, in the object which bears or emits them, and is necessarily on a different level, as though there were two orders of size or disparate realities between which the sign flashes; secondly, in themselves, since a sign envelops another 'object' within the limits of the object which bears it, and incarnates a natural or spiritual power (an Idea); finally, in the response they elicit, since the movement of the response does not 'resemble' that of the sign. 22

Everywhere, couples and polarities presuppose bundles and networks, organised oppositions presuppose radiations in all directions. Stereoscopic images form no more than an even and flat opposition, but they depend upon something quite different: an arrangement of coexistent, tiered, mobile planes, a 'disparateness' within an original depth. Everywhere, the depth of difference is primary. 51

Repetition is the formless being of all differences, the formless power of the ground which carries every object to that extreme 'form' in which its representation comes undone. The ultimate element of repetition is the disparate [dispars], which stands opposed to the identity of representation. Thus, the circle of eternal return, difference and repetition (which undoes that of the identical and the contradictory) is a tortuous circle in which Sameness is said only of that which differs. 57

That which is or returns has no prior constituted identity: things are reduced to the difference which fragments them, and to all the differences which are implicated in it and through which they pass. In this sense, the simulacrum and the symbol are one; in other words, the simulacrum is the sign in so far as the sign interiorises the conditions of its own repetition. The simulacrum seizes upon a constituent disparity in the thing from which it strips the rank of model. 67

If it is true that representation has identity as its element and similarity as its unit of measure, then pure presence such as it appears in the simulacrum has the 'disparate' as its unit of measure - in other words, always a difference of difference as its immediate element. 69 

These quotes make clear that the disparate has been a technical term all along.  If we try to put them all together though, we encounter a certain slipperiness in the definition.  The disparate as dark precursor is some sort of force or agent that causes the two series of differences to resonate.  But because they resonate, they then form a single system of which the two series of differences form two disparate sides.  So, 'the' disparate is actually already two things (the system as a whole and its two sides) as befits a something which is nothing but difference.  

Understanding how a difference can already be two differences, which can then give rise to even more difference, helps us connect the coupled oscillators we discussed before with the signs and signaling mechanisms we find in those quotes about the disparate.  Signs are what appear between two oscillators when they become coupled into a single signaling and communication system.  From the perspective of the two series involved, these signs are external, passing from one side to another as exchanges of energy and information.  These signs express the "productive dissymmetry" between the oscillators, which I take to simply mean the fact that there are two different sides required, and not a single one.  However, they tend to cover over or cancel this dissymmetry because the coupling results in a resonant feedback loop that creates a single system and drives the individual parts into behaviors that they would never have been capable of alone.  That is, this resonant forced movement feeds back to produce even more difference.   It produces more quantitative difference in the amplitude of oscillation of each of the original series, and brings into existence a new qualitative difference -- the existence of the disparate system as a whole.   From the perspective of the whole system it creates, the difference is now internal, and the sign is interiorized within the system, becoming the sign of the differentiation of two sides of that single system.  

Now I think the idea that the simulacrum "interiorizes the conditions of its own repetition" makes more sense.  From the perspective of the coupled system, the signs that pass back and forth between the sides, transmitting and retransmitting a resonance, so to speak, are exactly what keeps the whole system resonating as a unit.  It's a bit like a bell whose vibrations make it capable of striking itself; the simulacrum is a sign capable of repeating itself in a feedback loop.  When a sign system becomes self reinforcing like this you can almost think of it as the cause of itself, or at least, to allude our earlier dynamic systems interpretation of the sign, as the trigger of itself.  With a feedback loop you come close to getting something from nothing, or at least getting what looks like a relatively stable thing from a condition of continuous circulation.  Which is precisely to say that you get a repetition (in time) without having had an initial model to copy it from.  The whole simulacrum system looks like it sprung from the head of Zeus fully formed because it hides an internal difference that creates an external repetition.  Hence the description of the disparate as the "unit of measure" of a pure presence, a pure arising of some unit from what we shouldn't call nothingness, but the chaos of difference in-itself.

Finally, if we relate this reflection on the relationship between difference and disparity back to the concept of embryogenesis, we can add something to our understanding of the theme of size, of the question of the large and the small.  The disparate, as the difference of difference, the in-itself of difference, is the differenciator of difference.  It differentiates differences at the same time as it gathers them into a single system.  The metaphor of the coupled oscillators illustrates the gathering that produces a single resonant system as its endpoint.  The metaphor of the embryo illustrates how a single differentiator -- the egg -- already contains within it a multiplicity of potential differences that appear as a stepwise series of coupled differences.  Morphogenesis generally is a process in which quantitative differences in the flow of amniotic fluid or the concentration of Sonic hedgehog become qualitative differences in final morphology, like, say two hands with five fingers each.  This is exactly the differentiation of an initial difference into more difference, the coupling of a series of chemical differences to a series of morphogenetic differences.  

Thinking further along these lines, we also might want to reconsider the idea I proposed that the large and the small might by analogy to harmonic overtones.  I was thinking of the question of size in terms of the difference between the two series involved, as if it were a question of comparing the relative size of their fundamental frequencies.  But now that we're being more careful with the terms, we can see that this would actually be a question about the size of the disparity -- how different are these series of oscillating differences, or how big is the second order difference -- and not the size of the first order difference.  It now seems to me that Deleuze instead intends for large and small to apply directly and absolutely to the first order difference -- difference proper in his terminology.  Large differences are intensive and qualitative.  They can't be divided without transforming into something else.  They are the morphological differences in our example.  By contrast, small differences are extensive quantities that can be continuously divided like the chemical concentrations in the embryo.  

In this reading, every disparity contains within itself both a small and a large difference, corresponding to the two sides that it simultaneously links and differentiates.   The disparate actually defines these two orders of magnitude and in each case causes them to resonate.  So while they may have absolutely different characteristics, they are defined as relative sides of a single movement of their differentiator, the disparate.  In this sense difference (well, technically disparity I guess) is never between the large and the small, since it actually creates them.  I feel like I have just reinvented Spinoza's system of attributes and modes.

Literary Systems

Since I haven't read Roussel or much Joyce, and only finished the first volume of Proust's opus, illustrating the dark precursor with examples drawn from these authors wasn't especially helpful for me.  As a result, this is going to be a short post.  I know, promises, promises.  I think I at least vaguely grasp how the Roussel and Proust materials illustrate the way the dark precursor links series by their difference, rather than their identity. 

Everything I know about Roussel I learned from wikipedia.  Luckily, this turns out to be enough to make some sense of Deleuze's example, which refers to Roussel self-described compositional method.

Roussel kept this compositional method a secret until the publication of his posthumous text, How I Wrote Certain of My Books, where he describes it as follows: "I chose two similar words. For example billard (billiard) and pillard (looter). Then I added to it words similar but taken in two different directions, and I obtained two almost identical sentences thus. The two sentences found, it was a question of writing a tale which can start with the first and finish by the second. Amplifying the process then, I sought new words reporting itself to the word billiards, always to take them in a different direction than that which was presented first of all, and that provided me each time a creation moreover. The process evolved/moved and I was led to take an unspecified sentence, of which I drew from the images by dislocating it, a little as if it had been a question of extracting some from the drawings of rebus. For example, Les lettres du blanc sur les bandes du vieux billard = The white letters on the cushions of the old billiard table… must somehow reach the phrase, …les lettres du blanc sur les bandes du vieux pillard = letters [written by] a white man about the hordes of the old plunderer."

I imagine that this method is anything but obvious from reading the original text, which certainly makes the role of the precursor, difference between b and p, dark enough.  Deleuze's point with this examples is that the story, which consists of two series that culminate in those phrases, is held together not by some similarity in what the two phrases refer to (resemblance of the signified), nor even to a likeness between the sounds of key word that appears in them (nominal identity of the signifier), but are joined by the difference between words (differential character).

The precursor, however, by no means acts by virtue of its identity, whether this be a nominal or a homonymic identity: we see this clearly in the case of the quasi-homonym which functions only by becoming indistinguishable from the differential character which separates two words (b and p). Similarly, the homonym appears here not as the nominal identity of a signifier but as the differenciator of distinct signifieds which then produces secondarily an effect of resemblance between the signifieds along with an effect of identity in the signifier.

Perhaps the overarching point here is that homonyms are a good illustration of the structuralist idea that meaning in language is dependent on differences in context.  The same sound (spoken signifier) can point to several different ideas or things in the world. (signifieds).  Which witch is which is only differentiated by context or (sometimes) by the orthographic differentiation of writing.  

This would contradict the idea that the power of language to point to or represent things stems from it's ability to equate different things through generalization.  I use the word "bug" to refer to every actual bug, even though each one is obviously distinct phenomena.  Does our language identify all those bugs with one another, lump them all into a single concept, simply because of the empirical fact that we don't have sounds for each individual bug?  In essence, the sound of the word "bug" functions as a homonym for all the bugs, and even for the same bug on different days.  We actually saw this idea treated back in the introduction as a type of natural blockage that prevents every individual thing from having its own concept of word.  

We have here a reason why the comprehension of the concept cannot extend to infinity: we define a word by only a finite number of words. Nevertheless, speech and writing, from which words are inseparable, give them an existence hic et nunc; a genus thereby passes into existence as such

But if that's how language works, then why doesn't a homonym also create a new concept or reference, instead of just potential confusion?  

Instead, the power of language to refer to particular bugs, or even the general concept of bug, comes from the overall structure of the language and the various possible contexts in which "bug" is said instead of "pug".  The power of "bug" to function as a homonym, so to speak, comes not from identifying individual bugs with the pre-existing category, and corresponding single sound, "bug", but from the way that sound makes all the different bugs in all the different possible contexts resonate.  That bug over there, and this bug over here, the hairy one with the six eyes, and the one from yesterday that fell into your soup.  If the secret of language is its structure, and the structure of the way it relates to our lived experienced, then the resonance between these bugs doesn't come from the fact that they all sound the same, but from the fact that in each context they are different from pugs.  This at least, is my rough translation of what Deleuze concludes from Roussel's example.

In fact, it is not by the poverty of its vocabulary that language invents the form in which it plays the role of dark precursor, but by its excess, by its most positive syntactic and semantic power. In playing this role it differenciates the differences between the different things spoken of, relating these immediately to one another in series which it causes to resonate.

Proust is meant to provide another example of how two series get connected by something which doesn't exist, or at least has no identity in itself.  We saw before how reminiscence in Proust puts us in touch with an in-itself that never really was a present, as if all our experiences were merely dreams that we were awake.  Deleuze characterizes all of Proust's approach to memory as asking the implicit question of how certain qualities come to get associated with certain memories, the exact qualities that enable us to draw together a present and a past as related by a moment of memory.

No doubt, to remain at a first dimension of the experience, there is a resemblance between the two series (the madeleine, breakfast), and even an identity (the taste as a quality which is not only similar but self-identical across the two moments). Nevertheless, the secret does not lie there. The taste possesses a power only be- cause it envelops something =x, something which can no longer be defined by an identity: it envelops Combray as it is in itself, as a fragment of the pure past, in its double irreducibility to the present that it has been (perception) and to the present present in which it might reappear or be reconstituted (voluntary memory). This Combray in itself is defined by its own essential difference, that 'qualitative difference' which, according to Proust, does not exist 'on the surface of the earth', but only at a particular depth. 

The cookie I'm eating now tastes the same as grandma cookies.  I used the same recipe.  Must be some similarity in the chemical structure of the cookie, right?  Sure, but that empirical similarity is not what really explains why they get related by memory.  Many things went on at grandma's house.  My experience eating this cookie right now is likewise enormously complicated.  How does the particular element of the taste of the cookie come to be a salient variable that attaches to both of these experiences and characterizes them as related?  Why does the taste remind me of grandma's house, and not the red color of my hoodie or the way my dog's smell is similar to the wolf's?  Any two moments of experience can have an infinite number of similarities that might link them.  But it's actually the differences -- the points within each experience that stood out as different or important -- that link the experiences as the same, but different.  "Only differences are alike".  But their likeness comes to us only through their differences, through some process that produces them as different, as now but not now, as memories.