It is impossible to conceive the assemblage of a scientific experiment apart from a field that generates plans and topological, mathematical, axiomatic and computational descriptions. But sign-machines can function equally well directly within material and social machines without the mediation of significant processes of subjectivation, something which has become more obvious each passing day. The fact that the common essence of semiotic machines and material or social machines proceeds from the same type of abstract machine is the decisive step we must take in order to found a political pragmatics on something other than good intentions.
Felix Guattari, L’Inconscient machinique: essais de schizo-analyse. Paris: Editions Recherche, 1979. p. 67.
That we underestimate machines is an understatement. Human language itself is a code which produces codes, hence an always already over-coded decoding — and the decoding processes, for their part, go as far as you like. Let us be cautious, then, and attempt to linger for a moment on the side of the symbolic. Every discourse, every instance of language, every explicit “saying” — is also implicitly a kind of abstract program. A program gives us in turn the language in which that program is expressed — and also in which completely new programs can be expressed. Finally, every text also contains an irreducible element of pure ontology, thereby encoding — between the lines — the very principles for organizing discourse itself. Whichever metaphor obscurely prefigures the communicative passage, tracing these interdependent “resemblances,” or “differential” networks of “abstract” models, (or even “ethico-spiritual” traces of traces) necessarily takes us on an adventure outside of the text — but mysteriously or ironically, always into other kinds of texts! This infinite indeterminacy — or antiproductive rupture — is the basis of a “parasitic” logic, the logic of interruption, inequality, a constitutive non-determination.
Hence, in addition to these four distinct but interwoven layers or aspects co-existing in even the shortest text — indeed in a single word — it seems we must also suppose some pre-logical flux of intensity, a matrix of differences, in which these varying aspects would themselves become locally codified and relatively grounded. A diagram needs a space in which to be built and materials from which to be constructed; ideas needs relational fields in which they realize themselves sensibly and and dramatize their “break” into reality to one another — how, why and where they fall to their death onto the depths of bodies — but even this as though organically or by divine judgment. Bodies break the recursive cycle of language through the intervention of a partial object (programmer-parasite.) The parasite, the cold body sucking the warmth, writes new programs, and in doing so inevitably scrambles the meanings of the old instructions. The parasite is ontological rupture or antiproduction, phenomenological transduction — its work, grounding relation, is itself grounded only by an act of invention, translation, dramatization. Grounded in metaphor, in a productive diagram, in an abstract machine. Or, in other words: the parasite, whose provisional ground or counter-network is the minimal subject of the abstract machine, guarantees the consistency of the abstract programs’ specific productive diagram simultaneously as (1) a single variation, which is also (2) a model for variations; yet this is model is at once a (3) variable language of models, as well as the (4) machinic meta-ontology pragmatically governing the organizational principles of languages themselves.
Fractal Ontology 
I wonder if we should use A Thousand Plateaus here in order to replicate their distinction between codes and languages. They use the example of the bee (probably from Benveniste’s example) and show that bees don’t have a language because they lack indirect (what Guattari calls illucotionary) discourse. Other than that, this sounds like an interesting beginning post: it really makes me hope SPEP takes our panel, because I think this sort of project can be combined with what I wanted to do in my paper.
Great stuff buddy!
Yes, I would say it’s likely Deleuze and Guattari have gone further than just about anyone else — other than perhaps Michel Serres — on this question of codes and languages. In the chapter you mention they also write: “An assemblage of enunciation does not speak ‘of’ things; it speaks on the same level as states of things and states of content… In short the functional independence of the two forms is only the form of their reciprocal presupposition, and of the continual passage from one to the other.” (ATP 87)
We are confronted here with the mutually constitutive difference between bodies and language (eating or speaking?) or between signs and acts. D+G have an interesting point to make here; they argue from this that we never simply see the interlinkage of the order-words, nor the pure causual causual relationship between object elements, not at least each separate from one another. But, on the other hand — D+G point out — we don’t quite find one representing the other, either. “On the contrary, the independence of the two lines is distributive, such that a segment of one always forms a relay with a segment of the other, slips into, introduces itself into the other.” (ibid) Is this not the figure of the parasite or the troubadour — perhaps even more broadly, the figure of the stranger or the absolutely Other, the body without image? How, if not through this parasitic diagram, this semiconduction, could we slip endlessly between the two layers — the world of things and the world of words — from order-words to silent objects? One way of approaching this impasse might consider that both bodies and language have codes of their own — implying that beneath or beyond this division, neither layer can be separated from a plane of consistency where words and bodies intermingle.
Yeah, either way, I hope we hear from them soon. 🙂