Deleuze for the Desperate #13c: Content, expression and the strata

Dave Harris

I have left this plateau to the end of the sequence on language, even though it actually appears before the other two, as Plateau 3. It must have put off so many people from reading further! For one thing, Deleuze and Guattari hide behind a puppet spokesperson developing the narrative in this plateau. They call him Prof Challenger, a Conan Doyle character. The Wikipedia entry describes Challenger as ‘a pretentious and self-righteous scientific jack-of-all-trades'. Deleuze and Guattari obviously feel defensive, perhaps having been criticized for discussing topics where they have no actual expertise. Much of the discussion  includes examples of debates in geology and biology, for example. The references are almost entirely in French. I found it difficult to do any more than read through hoping to get the basics of the argument. I have more detailed notes on my website -- here.
I must say I much prefer the discussion in DeLanda's book on science and deleuzian philosophy (DeLanda 2002), especially the first parts of Chapter 2, which is considerably clearer and less interwoven with high-powered academic asides. That discussion mostly references Deleuze's earlier work Difference and Repetition.
While I am here, it is always worth mentioning Guattari’s book The Machinic Unconscious, which has a useful chapter on language, especially chapter 2. The rest of the discussion is formidably elaborate though.  Guattari’s book emphasizes Hjelmslev’s distinction between content and expression, as before, with a bit more detail. Generally, the argument is that human language is not the only kind of language. Like animal languages it is an actualization of the same abstract machines linking contents and expressions. Human worlds participate in the same machinic phyla and planes of consistency as other forms. It follows that there is nothing unique about human subjectivity.

Plateau 3 of A Thousand Plateaus (ATP)  also illustrates examples of nonhuman forms of communication, tries to restore the role of content and its own forms, and denies human linguistic forms of expression as the only ones. Deleuze and Guattari are also going to cite the work of Hjemslev to argue content has an independent role, that it is not just exhausted by conventional human semiotic forms of expression, that it has substances and forms of its own. For that matter, content and expression are often found connected together in complex ways in communication. As examples, expressions themselves become forms of content for subsequent expressions, but there is also an argument that content actually generates expressions so that expressions are better understood as forms of content.

The key concept here is the stratum. Strata are found throughout nature and they are important in communication. Channeling Spinoza, D&G say that  'To express is always to sing the Glory of god. Every stratum is a judgment of god; not only do plants and animals, orchids and wasps, sing or express themselves, but so do rocks and even Rivers, every stratified thing on earth' (ATP p. 49).  If,like DeLanda,  we think of Deleuze and Guattari as offering a version of complexity theory, rather than theology, that might make more sense. We can get to the underlying mechanisms of complexity theory, virtual forces, attractors, phase spaces and the rest by looking equally at sedimentary rocks or at biochemistry, or anywhere else. Incidentally this raises the issue of Deleuze's 'univocity', the claim that Being always speaks with a single voice, central to Badiou's critique of Deleuze (Badiou 2000).

So stratification is a form of communication, first in a general sense, that all communication takes the raw materials of matter and forms them — human communication does this when it constructs things and words in deleuzian terminology . Secondly human communication itself depends on these prior processes where content has already formed itself into actual substances. There is a lot of discussion of these points and work in geology and then biochemistry to illustrate them.  The patient read-through suggested above eventually gets to three major types of stratum, outlined on pages 65–68. They are:

First type A stratum composed of molecular and molar components, with molecular content and molar expression. The two levels can be linked by a term they use a lot elsewhere, but which I must say I find still a bit elusive — 'resonance' (p. 64). They also talk of ‘reciprocal presupposition’ (p.75).  The molecular level still shows a difference between substance and form, since molecules can interact among themselves to become formed substances. The molar level shows both statistical aggregates and a more definite 'state of equilibrium'. Expression carries the properties of the molecular or micro level to the molar or macro level. There are different possibilities since intermediate strata, 'epistrata'  can have an effect and so can exterior forces, applying especially to expression. There might be cases at the extremes, where, for example the molar level reflects exterior forces as a mould, or it actively modulates these exterior forces, and lots of other possibilities.

The best examples would be the geological stratum that the Plateau  begins with. If we think of this as a stratum of sedimentary rock in a geological formation, we can notice that it has two processes at work. First of all, sand, pebbles and other bits and pieces are laid down on a seabed, and they eventually get compacted together to form a more solid layer. Over time, that layer is subjected to another process of organization. It might be further compressed, folded, perhaps even baked as the sedimentary rocks around Plymouth UK have been folded and baked by the volcanic intrusion that formed Dartmoor. This is a double process, one example of what they are going to call a double articulation, one articulation arises from loose collections of material , and one arises from tighter and more extensive forms of compression and aggregation. In ordinary language, the point is that if you look at the stratum of sedimentary rock, it contains information. You can sometimes see sand and bits of material still visible in the rock and get information from them and then work out what has happened in the different stages. You can eventually read actual visible shapes in the landscape like hills and valleys as products of the underlying geology.

Other examples include the way crystals grow from ‘seeds’ or the way chemicals interact in physical systems to from molecules of increasing complexity and length.

M Harris

The same could be said of the development of social strata. These are patterns of collective action, including those that are particularly powerful in the formation of social classes or elites. Social strata are not specifically discussed in this plateau, but the sociologist Gabriel Tarde receives a 'homage' in Plateau 9 on micropolitics (pp. 240 – 241). There we are told that in his great debate with Durkheim, Tarde insisted that the 'great collective representations' presuppose 'exactly what needs explaining, namely, "the similarity of millions of people"'. There are other references to Tarde's work in some of the books written by Deleuze.

Latour (2002) has a useful essay illustrating Tarde's influence. Latour quotes Tarde in describing social strata as composed of individual components, where: '"the attributes each element owes to its incorporation… Do not form its entire nature… The collective power, of which it is a part… Is only an artificial being, made only of sides and facades of beings"'. There seems to be support for the important role of pragmatics, argued earlier. Referring to the apparent structures in language, Latour argues that for Tarde 'the structure is only one of the simplified routinized repetitive elements of one of the locutors who has managed to include his or her local tradition into the general idiom' (page 8).

Deleuze and Guattari argue that Tarde is not talking about human individuals in the usual sense, but rather about flows which connect individuals, such as 'flows of imitation' (p.241).  Actual individuals are heavily influenced by social strata, as we saw in discussion of the paradox of the subject.

End M Harris

Second type those strata linking organic and inorganic matters. The organic layer still preserves or even amplifies the inorganic matter at the molecular level, as in cellular chemistry, but there is one difference. The boundaries between the layers become thresholds, and elements can align so that 'expression becomes independent in its own right'(p. 66).  Different forms combine to produce real effects. Expression can take a linear form (apparently the nucleic sequence in biology joins together nucleic acids and proteins and this is a crucial stage for the development of organisms). This independent expression means that the organism is capable of greater levels of deterritorialization, including the ability to reproduce.

There are lengthy debates in biology, both embryology and evolution, covered in this plateau. Again I have to depend on DeLanda. I was struck by the argument in his book that modern embryology describes the development of the internal organs of the embryo in terms of initial collections of general cells, possibly stem cells, which are then exposed to a local environment of chemical flows, including hormones, and as a result cross thresholds and go on to develop specific forms like bones or organs. Those in turn affect the environment of the embryo which triggers other flows and other developments, like clustering and folding of cells, once other thresholds of formation have been reached. This is in contrast to the old view, current when I last did elementary biology as a schoolboy, that there was some sort of blueprint which guided the development of the embryo. More generally, discussing the early formation of life we are told that a ‘prebiotic soup’ (p.55) is a stratum composed of inorganic materials. A crucial intermediate layer, an epistratum, the membrane, further articulates these materials, producing a living aggregate, a ‘structuration’ (p.68) , a kind of dynamic structuring process, connecting a living creature and its milieu.

There is also an implication for conventional views of the evolution of species. Again to be very brief about this, and no doubt indicate my own highly limited understanding, it is not that individual members of species or types encounter environments which directly produce modifications. Instead, the species itself is the unit of evolution, and it contains within it all sorts of potentials for development. Darwin was really developing a ‘science of multiplicities’ (p.54).  This is another example of an abstract machine at work (p. 56), possibly partly revealed in a general embryonic form shared by all animal species. As with all abstract machines, some potentials are actualized or realized, while others remain latent. The plane of consistency will regulate the development of potential to some extent, although here we have a tweak saying that the plane of consistency itself can be 'reworked' by particular developments like 'rogue particles' (62).  The discussion then goes on to consider a number of  concrete complexities and developments.

The formation of inorganic substances has been going on long before human beings occupied the earth. Atoms have been forming into molecules, molecules into chains of molecules, chains of molecules into organisms and so on. Minerals have developed in the same way from atoms and molecules, and actual development depends on the atomic structure of elements, such as the flexible molecular combinations of carbon, or the catalytic properties of metals. Again I'm depending on DeLanda here, including a short article he wrote about the importance of metals in forming up the substances that we see around us. This is a literal example of what Deleuze and Guattari mean when they say that ‘The first articulation concerns content, the second expression’ (p.49).

Third type: strata that can modify the external world. This happens once expression becomes a matter of symbols or language. It is possible that this happened originally or only with human beings, especially as human hands developed for tool use, and human faces and larynxes for speech and gesture.  Beneath those is  'the cerebral – nervous milieu' (71). It has developed from the organic stratum as a kind of 'pre-human soup immersing us' (p.71),  and it has led both to hands that can use tools and faces that can vocalize. It is this double development that underpins the subsequent important distinction between 'apparently irreducible categories: things and words' (p. 71).  Both hands and faces together provide for a new level of deterritorialization, especially in particular milieux. Human forms of expression rapidly turned into diverse languages, but there are still limits imposed by the organic elements of larynx, mouth, lips and face, as a kind of substance limiting expression.

There are  interactions between organic, ecological and technological factors and we could map them, in principle. For example, different milieux produce different sorts of speech, and there is a marvelous and typical example here on page 69 where operating in a forested milieu apparently permits softer speech than the shouts or grunts required on the savannah: the conclusive evidence for this is that  'Everyone knows that lumberjacks rarely talk’! This is referenced to a French study written in 1933. With human language we also have new contents and expressions — 'technological content, semiotic or symbolic expression'.

The ability to vocalize introduces superlinearity, endless linearity of expression developing relays between speakers and listeners over time. It also permits translations, not only of other languages but of nonhuman forms of expression: this leads to 'the scientific conception of the world' (p. 69), where all the characteristics of the other strata get translated into human signs. It is not surprising that human language develops 'imperialist pretensions' (p. 70) and people that think that human language is essential to any form of understanding or communication — and these might include Lacan.  I must say I think he has a good case.

But we still have to remember that the biological raw materials that underpin human language are themselves only aspects of an underlying 'technical social machine' with 'states of force or formations of power', while human expression is also the product of 'the semiotic collective machine' (p. 70) that produces signs on all strata. Both machines preexist human variants.  Concrete machines on human strata  can be understood as intermediate states of the abstract machine. There is a tendency for human beings to want to render this abstract machine as something exclusively human, an 'illusion' arising from the tremendous deterritorializing power of human language. The particular social and technical machines also do develop a degree of autonomy, but still as strata. However, o
nly with human signs can the abstract machine fully develop and actively begin to write. I must say I think that is a crucial difference.

In further clarifications (!), signs can be understood as having three components — indexes of territory, symbols which deterritorialize, icons which reterritorialize. Since all strata, including non-human ones, also show deterritorialization and reterritorialization, perhaps we can we say that they all exhibit signs. However, there is no common system of signs [nor a presymbolic semiotic chora, which alludes to Kristreva? p.72] . We have to make sure that we are not seeing human signs as the only type against which others are inferior. Human signs develop only when forms of expression become quite distinct from forms of content, with their own categories. Human signs can be organized into regimes as we saw.  There is still a prior form of 'so-called natural codings' (p.72), which may include 'animal signs 'indicating territoriality. This argument underpins the lengthy Plateau on the refrain, as we saw.


In this Plateau, there is further criticism of the old enemy, structural linguistics. That approach has thought of the relation between signifier and signified in various ways, but none of them give any independent existence to the signified. This gives the signifier a high degree of redundancy, which means here an ability to be endlessly applied to an overlapping range of signifieds 'hence its incredible despotism, and its success' (73). For Deleuze and Guattari by contrast there are separate forms of content and expression, 'real independence and a real distinction'. A form of content is not just a signified. Nor does expression entirely dominate content — there is 'reciprocal presupposition'. It follows that expression is not just a matter of the play between signifiers. There are indeed even forms of expression that do not have signs — the genetic code here.

What usually happens with linguistic imperialism is that both signifier and signified are extracted from words. It is the construction of these words  that provides the link. But we can see that words like names are never simply linked to things. We know this from what has been argued elsewhere about empirical things having nonempirical virtual dimensions, and, when we looked at the haecceity, we saw that apparently isolated things have just been individuated from flows and are connected to other apparently isolated things. Here we have an argument from Foucault saying that an apparent thing like a prison is better understood as a form of content related to other forms of content like schools or factories. Expressions like delinquency emerge reciprocally from this expanded content. It is not single words that we should study but rather statements arising in a specific social field, a stratum, in this case a regime of signs. There will be a specific history and politics to the statements

Both expression and content should be seen as located in multiplicities. They can be seen as sharing an abstract machine, a diagram. In Foucault's case this abstract machine produces  prisons, schools, hospitals and factories [this is argued more fully in Deleuze's book on Foucault]. Specific signifiers and signifieds are 'formalizations' (75). These formalizations can conceal important concrete variation, and things like the actual semi-covert life of prisoners can be cited in support of prison reform (see Deleuze 2006, ch.38). Specific assemblages act to connect content and expression, not some universal characteristic of language. These assemblages need to offer double articulation as described above, organizing content and expression. This operates with regimes of signs and therefore with formations of power — disciplinary power for Foucault.

There is criticism of another unspecified model of language too, those involving bases and superstructures. It is not at all clear why this topic appears at this point.  However, neither content nor expression should be seen as 'the determining factor'. To argue this would be to risk reimposing some hierarchy, lending support to a state apparatus. Language does not fit into the concept of ideology. Again they do not specify who might hold this rather dogmatic view – I doubt if many contemporary marxists did. Both economic base and cultural superstructure have the same combinations of content and expression. This is not improved by sticking in a bit of Lacanian psychoanalysis to get some kind of freudian marxism, possibly a dig at Althusser (p.77), although Althusser is admired in another Plateau as we saw in the discussion of sign regimes.

Finally, there is a denial of any kind of evolution at work with the strata, although the earlier remarks on content being articulated first look awfully like it. The strata themselves do not occupy a fixed order and can act as substratum for each other. For example technical phenomena can provide 'a good soup for the development of insects, bacteria, germs, or even particles' (p. 77). On the plane of consistency, at the virtual level so to speak, everything mixes: elements are only firmly connected in strata. Again there is not even a dualism between strata and plane of consistency, because we find abstract machines at work in both, located on planes of consistency or on strata.

The Plateau ends by reasserting what might be seen as the virtual elements, like abstract machines and planes of consistency. The strata by contrast do what might be seen as the concrete work to link expression and content, regulate flows of particles and signs, and manage deterritorialization and reterritorialization. All these are examples of double articulation. The strata differ from each other in the way in which content and expression are linked, and these are real distinctions between them (page 81). Strata constantly subdivide. The machinic assemblages operate between them as well as within, and they also link to the plane of consistency which enable them to effectuate the abstract machine.

As a final aside, t
here is a fashionable development at the moment called post-humanism or nonhumanist ontology, which wants to fully embrace the idea that things communicate among themselves and with us. This has somehow been encouraged by work on quantum physics and the strange communications that seem to go on between subatomic particles. Barad's work has been influential here, and she finds some implications for feminism.  Latour’s article mentioned before has his own take supporting Actor-Network Theory. Others seem to think that plants or crossroads literally speak to them, that they have not only have their own language but can translate them into human languages too. I am not at all sure the work in Plateau 3 leads to this sort of development. I think the point is to suggest that the same patterns of forces, the same abstract machine, produces both human and other kinds of language. I don't think the argument is that all languages can be seen as like human language, with pragmatics and subjectification processes. That would be the kind of linguistic imperialism they warn about – but see what you think. 


Althusser L (1977) 'Lenin and Philosophy' and Other Essays, London, New Left Books.
Badiou, A.  (2000) Deleuze.  The Clamor of Being.  Minnesota: University of Minnesota Press (my notes here)
Delanda, M.  (2002) Intensive Science and Virtual Philosophy, London: Continuum (my notes here)
DeLanda, M. (1999) Deleuze and the Open-Ended Becoming of the World, Manuel DeLanda Annotated Bibliography, [online]  (my notes here)
Deleuze, G. (2006) Two Regimes of Madness: texts and interviews 1975--95. Ed D. Lapoujade. Trans. Ames Hodges & Mike Taormina. New York: Semiotext(e) Foreign Agents Series (my notes here)
Deleuze, G (2004) [1968] Difference and Repetition, Trans Paul Patton, London: Continuum Publishing Group (notes here)
Deleuze, G.  (1999) Foucault, S. Hand (trans).  London: Continuum (notes here)
Latour, B. (2002) Gabriel Tarde and the end of the social. In P. Joyce (Ed). The Social in Question. New Bearings in History and the Social Sciences, pp. 117–32. London: Routledge.  Online: