My thoughts are a bit scattered this morning as I got little sleep last night, but nonetheless I wanted to underline something with respect to my last post. There I suggested that every system– where “system”, in my thought, is a synonym for “object” or “substance” –is haunted by entropy. Entropy measures the degree of disorder within a system. The short clip below gives a sense of the nature of entropy:
When the particles are first shot into the box they exist in a highly ordered state (i.e., they are strongly localized in the box). As a consequence, the particles begin with a very low degree of entropy, which is to say that there is a high probability that they will be found in one particular region of the box. As the system evolves, however, the degree of entropy or disorder within the system increases. It becomes more or less equally probable that the particles will be found anywhere within the system.
The claim that every social system faces the problem of entropy is the claim that every social system faces the question of how to maintain its organization over time. The relations that constitute a social system or object 1) establish or produce a low degree of entropy such that it is improbable that social actors will be found anywhere within the social system, and 2) are relations between external or independent objects. Take the geographical distribution of wealth in the city of Chicago. On the South Side you tend to find rather poor individuals, while on the North Side you find wealthy individuals. These concentrations represent a low degree of entropy insofar as the probability is low that you will find poor people evenly distributed throughout the city, just as the probability is rather low that you will find wealthy people evenly distributed throughout the city.
The mystery, then, for social and political thought, is why entropy doesn’t increase in such systems. With the passage of time, why doesn’t the distribution of poor and wealthy people become evenly distributed throughout the system. Put differently, what are the mechanisms at work within the social order that maintain a low degree entropy? Gratton maintains that our disagreement over structuralism is merely a disagreement over nomenclature. In short, the social and political theorist should begin with the premise that every social order is improbable.
Gratton likes the word “structure”, I prefer terms like “regimes of attraction” and “feedback loops”. However, I don’t think these are mere differences over nomenclature. Nowhere, among the great structuralist thinkers, will you find reference to anything like the problem entropy. The reason for this is very simple. As Derrida reminds us in Differance when discussing Saussure, structural relations are differential relations without positive terms. The phoneme “b” is literally nothing independent of the phoneme “p”. Or rather, we should say that there is no phoneme “b” or “p”, only the phoneme b/p constituted by internal differential relations. Such a thesis also holds for how Levi-Strauss analyzes kinship relations and myths, or how Althusser analyzes social structures.
However, if you begin from the premise that social relations are differential relations without positive terms, then you’ve already erased the problem of entropy. Why? Because there are no independent or positive terms being related that constantly threaten to fly apart increasingly the probability of distribution throughout the system overall. Because the being of the terms is already constituted entirely by the differential relations there is no issue of how a system maintains an improbable organization over time. It is only when you begin from the premise of a strange mereology where larger scale objects are composed of smaller scale objects that are themselves independent of the larger scale objects that they constitute that the problem of entropy or why the smaller scale objects don’t fly apart destroying the larger scale objects emerges.
The debate, then, is not whether or not there are patterns that reproduce themselves in time, but rather whether or not relations relate positive terms, terms that could be detached, or whether there are only negative or differentially constituted terms. If you take the first route, then the problem of entropy comes into full view and you’re faced with the mystery of how low probabilities of equal-distribution are maintained. If you take the latter route, the problem of entropy doesn’t appear at all as there are no independent terms that could fly apart. As a consequence, the latter route leads you to look for a “supplement”, “event”, “subject”, “act”, etc., because it’s impossible to conceive an immanent entropic dissolution of systems as there are no positive terms for a system to dissolve into. I’m not being unfair to the structuralists here– with whom I’ve worked obsessively for over a decade now in my research –but am taking them at their word with respect to the ontology that they themselves embrace.
In response to my post the other day, Mel worried that my thesis that every system or object is haunted by entropy might strike others as pessimistic. However, I think precisely the opposite is the case. The entropic dimension of every system entails that, as a matter of principle, every system can be dissolved, can be otherwise, can fall apart. Ontologically, then, there is no such thing as a natural social order (nor a “natural” natural object, if by “nature” one means “incapable of being otherwise”). Rather, every social order is a temporary victory over entropy that is perpetually threatened by dissolution. If the essence of ideology lies in treating as natural and therefore inevitable a historically contingent social order, then the entropic dimension of every social system marks the ruin of any inevitable or ineluctable social order. As such, the entropic dimension of every social system marks the ontological ground for revolutionary hope.
Larval Subjects . 
October 19, 2010 at 3:14 pm
Levi,
Two things: I never said I like the word “structure.” Also, what is related in a structure but “positive” entities? I’ll post some thoughts on Levi-Strauss later, upon whom I’ve previously given little thought since studying him, but … well, I can’t agree at all on that point about entropy, since precisely differential relations are contingent (and the problem is about meaning there, not referents). It’s not that I don’t appreciate what you write above, but where you see a great difference, I don’t. I don’t offer up a priori structures that synchronically explain all, and this was precisely Derrida and Foucault’s attack on previous structuralisms…
October 19, 2010 at 3:20 pm
Let me also posit that you’ve written something like this yourself, regarding Hägglund’s reading of Derrida:
“My thoughts are still developing in this connection, but Hägglund’s remarks suggest a way of thinking the split-nature of substance in split-objects. One reason I’ve been unimpressed by critiques of substance that claim that it is incapable of becoming, change, process, etc., is that I already think of substances as activities or processes (this comes out with special clarity in chapter five of The Democracy of Objects where I discuss temporalized structure and entropy). Hägglund helps me to think about this. Every object is internally fissured by its own temporal structure such that it contains non-identity (withdrawal) within itself.”
October 19, 2010 at 3:46 pm
Peter,
I assumed that when you referred to nomenclature you were referring to a difference in choice of language. As for structuralism, the most basic and elementary thesis of structuralism is the thesis that structures are composed of differential relations without positive terms. This isn’t me making this up, but what Saussure, Jakobson, Barthes, Levi-Strauss, early Lacan, etc argued. The same is true in Derrida’s “Differance” essay, where he explicitly makes this point. The play of structure is this play of differential relations or traces that are internally related, not a play of parcels, units, or substances. As he explicitly says in “Ousia and Gramme”, substance is equated with presence that ignores this internally related system (pp. 39 – 40). This is why, as I argued last week, the heirs of structuralism were led to look for a void point within structure, a supplement, because they needed a FORMAL element of structure to explain change (as there could be no talk of positive terms acting outside structure). Further, if you read Levi-Strauss’s Elementary Structures of Kinship or Lacan’s “Seminar on the Purloined Letter”, or anything in Mythologiques (recall, for example, his gorgeous discussion of musical composition in the introduction to the first volume of Mythologiques, where he discusses music as a logico-mathematico problem to illustrate the workings of structure), the transformations that take place within structure are formal, mathematical permutations akin to steps in a geometrical proof or a topological transformation, not dynamic processes among distinct elements in interactions of feedback (which is the issue that would motivate Bourdieu’s critique of the structuralists in The Practice of Everyday Life, as well as the critique of later Deleuze with Guattari (the major debate between habitus/practice versus structure)). Thus, for example, when Levi-Strauss analyzes the role of the Trickster in certain myths, trying to account for it’s contradictory qualities, in, as I recall, “Structure and Dialectic” in Structural Anthropology, the Trickster solves a logical problem, a formal problem allowing the structure to function despite a contradiction. It is not a question of interactive feedback between entities. Nor is it a question of dynamic interactions between discrete entities. This is a totally different conceptual space than the one I’m developing. It may be that you’re using the term structure in a much looser sense than I am, but my project partially involves responding to Badiou, Zizek, and Ranciere who I believe formulate their political philosophies in the way they do based on a set of axioms about the social they inherit from the arche-structuralists. This is why, also, Badiou can appeal to mathematics to demonstrate the void at the edge of situations because the issue is one of formal and structural features of structure, not of existing social ecosystems.
I do believe there’s something worth redeeming in Derrida, but this requires reading Derrida perversely and against the grain of his own texts and intentions (as he read Husserl and Saussure), not in simply saying that this is what he was arguing. Substance is a bad word in Derrida, which he (wrongly in my view) treats as a synonym of presence. My excitement with Hagglund’s work was that, for me, it suggests an appropriation of certain aspects of Derrida’s thought, but appropriations are also infidelities.
October 19, 2010 at 6:16 pm
A really good example of what a structural analysis looks like is Lacan’s analysis of psychosis in his early work The Family Complexes. There Lacan suggests that psychosis will appear every third generation in certain families because of how the symbolic is structured in these families (hence talk of a complex). The point is that it is the combinatorial logic of the signifier that produces the outcome of psychosis, not how parents treat children, whether or not a father is present, etc. Like solving an algebraic equation, this combinatorial logic ineluctably works itself out across time in much the same way that there are steps in solving the problem. This is the basic lesson of Lacan’s Purloined Letter seminar as well, where the different subject positions have nothing to do with the actions of the individuals, but rather with how the signifier positions us in it’s ineluctable logic. Hence the agency of the signifier. In Elementary Structures of Kinship Levi-Strauss will trace similar multi-generational logics, showing how exchanges unfold according to an ineluctable logic (his structural reading of Oedipus is based on a similar tracing of a combinatorial logic). As a consequence, analysis of economics, family intrigues and stratagems, etc are irrelevant to a discussion of kinship relations in Levi-Strauss as it is this ineluctable logic that drives everything. Lacan formalizes all of this mathematically with his game of zeros and ones and the use of matrices in seminar 2. Fink has a nice and accessible introduction to how all of this works in The Lacanian Subject. There’s nothing like feedback loops or interactions between discrete units in this functioning of structure, just the inevitable unfolding of immanent combinatorial logics behind the backs of individuals. Hence Ricouer’s famous characterization of structuralism as transcendental philosophy without a transcendental subject. Additionally, for Levi-Strauss it’s simply not true that the relations defining a structure are contingent. There is a universal structure that varies itself while all these variations are variations of one structure. For Levi-Strauss all myths are saying the same thing. He sets out to demonstrate this in the four volumes of Mythologiques. This is what allows him to read myths from the most diverse and disconnected cultures alongside one another. These are just the ABC’s of structuralism, and are part of the reason that Derrida’s Structure, Sign, and Play and Lacan’s later lalangue caused such controversy. They undermined the thesis that structure could attain closure or that it’s governed by ineluctable combinatorial logics.
October 19, 2010 at 7:08 pm
[…] Bryant has a post up (I reply below it) on entropy: The debate, then, is not whether or not there are patterns that […]
October 19, 2010 at 7:20 pm
Hi Levi,
Just as an aside, Levi-Strauss does mention the concept of Entropy in a number of works, and in particular Tristes Tropiques where he states: “Every effort to understand, destroys the object studied in favor of another object of a different nature.” Or: “Anthropology could with advantage be changed into ‘entropology’, as the name of the discipline concerned with the study of the highest manifestations of [a] process of disintegration.”
October 19, 2010 at 7:50 pm
I am reminded of the Saussure’s chess game in which it is not the pieces but the place of the pieces afforded by the structure of the game, the “combinatorial logics.” I think this basic ontology is present for the most part in Lacan, Foucault and Zizek, too, in which the pieces themselves are relatively interchangeable, like on the chess board. What matters is the overarching structure in which they appear or are manipulated. Like Graham is fond of saying, this was once a very liberating idea, but it is no longer as liberating as it once was. Now, the structure basically crushes the individual entities out of existence, such that it is the structure itself which has any ontological agency or power. OOO seeks to restore agency to the pieces themselves, such that we can see how pieces have their place but also how, in another real sense, they can never be crushed out of existence by those places, but offer their own power and introduce differences into existence that are not defined by the logic of the system. I’m not saying Peter agrees with this Saussurian view of structure, but I do agree with Levi that this idea of structure is present in much of Continental philosophy. It seems to me that even when Badiou critiques (rightly) constructivism or constructivist thought, he seems to slip into something similar inasmuch as the logic of the situation is so dominating and the backdrop of any discussion of the subject or the event itself. With OOO, the ontological agency of components and their autonomy from the relations they happen to be in is already there from the beginning, is already operative ontologically, rather than coming in from without as in Badiou and others.
October 19, 2010 at 8:02 pm
Levi, I can’t wait for your discussion of this in your The Democracy of Objects to go public, as I think there are some great resources for thinking how both an object and its pieces are objects, and how an object-system uses its own pieces to reproduce itself through time while, without a diminishment of agency, the pieces themselves are able to do this because they have a positive reality in the first place. A system can only use its pieces because its pieces are themselves real and ontologically busy in their own right. Likewise it puts to rest the idea that an object’s life cannot be punctuated and evolving. It is, as you say, precisely because every object is compound-object, object-rich, that we can begin to explain both the object’s genesis, preservation and dissolution at all. As you put it so well, object as legion, as a crowd, not as a system of pure differences or diacritical marks.
October 19, 2010 at 8:12 pm
Levi, I would like to contest your characterization of Lévi-Strauss here and elsewhere. Lévi-Strauss was far more sophisticated in his theorizing of social situations than you often suggest; he was no mere structuralist. In fact he coined terms that refer quite explicitly to the very properties of social systems you mention above. For example, as early as 1959 Lévi-Strauss was using the terms “cold societies” and “hot societies” to refer to the different dynamics inherent in “primitive” (in a non-pejorative sense) and “progressive” social systems.
For Lévi-Strauss “cold” societies were societies “outside of history”, characterized by cyclical time, with strong boundary and maintenance enforcing social institutions and customs, and retaining a mythic mode of thinking. By contrast, the “hot” societies were driven by linear time with their emphasis on “progress”, associated with constant technological change – which thus suppressed the timeless mythic mode of thinking. Incidentally, Lévi-Strauss sees myths in-themselves as “machines for the suppression of time”. However, in the primordial logic of myth also lies evidence of “le pensée sauvage” (untamed though), which, though often suppressed and unconscious, Lévi-Strauss believed is an element of all human thinking.
In addition, in Tristes tropiques Lévi-Strauss discussed entropy quite explicitly – coining the terms entropology and entropologist. In the 1961 book, A World on Wane (p.397) Lévi-Strauss defined entropology as:
His knowledge of both cybernetics and rudimentary thermodynamics was actually also quite extensive (having extensively read and corresponded with Norbert Wiener); as evident in the following statement:
Also, here is Lévi-Strauss riffing on social “units” and local societies as “ensembles” (assemblages?):
All this is to say that there is much more to Lévi-Strauss than what his critics have made of him. As I read it, Lévi-Strauss’ real world “structures” are multiplex patterns and not at all monoliths.
[ also please note: Stanley Diamond’s work on entropy and ‘culture’ in his 1974 book In Search of the Primitive: A Critique of Civilization, among others in the anthropological literature dealing with what you seem to be after… ]
October 19, 2010 at 8:15 pm
I also think Lacan overplayed the more ‘structural’ elements of Lévi-Strauss’ work – to the exclusion of his more empirical dynamics statements…
October 19, 2010 at 10:14 pm
Michael,
That’s all correct. Levi-Strauss had a massive corpus and, moreover, there’s a vast difference in the Levi-Strauss of the ethnographers and the Levi-Strauss apprehended by the French structuralist philosophers. It’s that Levi-Strauss that I’m targeting.
October 20, 2010 at 10:30 am
“The mystery, then, for social and political thought, is why entropy doesn’t increase in such systems. With the passage of time, why doesn’t the distribution of poor and wealthy people become evenly distributed throughout the system. Put differently, what are the mechanisms at work within the social order that maintain a low degree entropy”
There doesn’t seem to be a special mystery about why social systems don’t ineluctably succumb to the second law of thermodynamics. SLT applies for closed systems where the sum of energy is a constant. Social systems are open, importing energy from their milieu which is used to form chemical bonds, lay and repoint masonry, and run databases. Come the Zombie Apocalypse, stuff won’t get fixed, bodies will decay and various other social proprieties and distinctions will go unobserved.
“if you begin from the premise that social relations are differential relations without positive terms, then you’ve already erased the problem of entropy. Why? Because there are no independent or positive terms being related that constantly threaten to fly apart increasingly the probability of distribution throughout the system overall. ”
There are periodic reference to cybernetics in Derrida’s work (See, for example, OG, p.9) but the only extensive engagement to my recollection is with Freud’s neurology in ‘Freud and the Scene of Writing’. I think most of these pretty much take the existence of physical or social processes for granted and are content to explore the implications of micro-changes within a materially embodied system for our metaphysical understanding of the meaning or content expressed by its parts. So Derrida, for example, never denies that a word can be used in ways that are not prescribed by its standard patterns of use. He’s less interested in why this is possible than with its implications for formalist notions of meaning or structure which follow.
October 20, 2010 at 4:11 pm
[…] Limits of Entropy Posted on 20. October 2010 by inregard In this post, Levi Bryant argues that object-oriented ontology takes into account the problem of entropy in a […]
October 20, 2010 at 6:36 pm
fair enough Levi, as i do read many franco-philophiles as having a tendency to obscure Lévi-Strauss’ more anthropological intent: gleaning patterns from the real-empirico (ethnological) world, as opposed to reducing multi-scaled phenomena to abstract structures.