Gilles Deleuze & Félix Guattari and the Rhuthmoi of Language – Part 2

Pascal Michon
Article publié le 18 July 2021
Pour citer cet article : Pascal Michon , « Gilles Deleuze & Félix Guattari and the Rhuthmoi of Language – Part 2  », Rhuthmos, 18 July 2021 [en ligne].

Previous chapter

Hyperpragmatism vs. Pragmatics

In order to support this pragmatic claim, they quoted John Langshaw Austin’s (1911-1960) famous essay How to Do Things with Words (1962). Description was not “all there is” in language. Speech was not only closely linked with action, most of the time it was, by itself, action.

Austin’s famous theses clearly demonstrate that the various extrinsic relations between action and speech [la parole] by which a statement can describe an action in an indicative mode or incite it in an imperative mode, etc., are not all there is. There are also intrinsic relations between speech [la parole] and certain actions that are accomplished by saying them (the performative: I swear by saying “I swear”), and more generally between speech [la parole] and certain actions that are accomplished in speaking [en parlant] (the illocutionary: I ask a question by saying “Is ... ?” I make a promise by saying “I love you ...”; I give a command by using the imperative, etc.). (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 77)

Deleuze and Guattari drew from the theory of the performative and its extension to the illocutionary a series of remarkable conclusions intended to cut definitively with structuralism and to provide new evidence for a new pragmatic perspective: Language was not a “code” nor a means to communicate “information”; it was primarily pragmatic; speech could no longer be defined simply as the extrinsic and individual use of primary significations, or the variable application of a preexisting syntax; speech was not secondary to la langue, it was the essential part of le langage.

The theory of the performative sphere, and the broader sphere of the illocutionary, has had three important and immediate consequences: (1) It has made it impossible to conceive of language [le langage] as a code, since a code is the condition of possibility for all explanation. It has also made it impossible to conceive of speech as the communication of information: to order, question, promise, or affirm is not to inform someone about a command, doubt, engagement, or assertion but to effectuate these specific, immanent, and necessarily implicit acts. (2) It has made it impossible to define semantics, syntactics, or even phonematics as scientific zones of language independent of pragmatics. Pragmatics ceases to be a “trash heap,” pragmatic determinations cease to be subject to the alternative: fall outside language, or answer to explicit conditions that syntacticize and semanticize pragmatic determinations. Instead, pragmatics becomes the presupposition behind all of the other dimensions and insinuates itself into everything. (3) It makes it impossible to maintain the distinction between language and speech [la distinction langue-parole] because speech can no longer be defined simply as the extrinsic and individual use of a primary signification, or the variable application of a preexisting syntax. Quite the opposite, the meaning and syntax of language can no longer be defined independently of the speech acts they presuppose. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, pp. 77-78)

This discussion on Austin entailed two opposite consequences. On the one hand, by contrast with all their predecessors in the rhythmic constellation—even Barthes who had moved directly from hard semiotics to “the pleasure of the text” (1973) while retaining strong structuralist bases—Deleuze and Guattari recognized the crucial role of pragmatic activity and context in language. This was a significant step towards a new linguistic theory that would not bear the weight and shortcomings of structuralism.

But, on the other hand, they interpreted this role in such a radical way that language activity was dissolved into the context itself. The theory of language lost its independence to the benefit of the theory of being, linguistics was finally reduced to philosophy, and the linguist Benveniste to the philosopher Austin.

In order to properly assess the far-reaching consequences of this leap over linguistics, we have to go into details. The previous comments on Austin showed that Deleuze and Guattari explicitly endorsed his extension of the “performative,” that is, the “actions that are accomplished by saying them,” to the “illocutionary,” that is, any action “accomplished in speaking” such as questioning, promising, or commanding [my emphases]. By doing so, Deleuze and Guattari initiated a trend of thought which developed throughout the 1980s and which made Austin into a precursor of contemporary continental Nietzschean philosophy. The ordinary language philosophy was reinterpreted in the light of a radicalized pragmatism. At the end of the decade, Jacques Derrida (1930-2004), for example, considered Austin’s speech act theory as backed by an implicit Nietzschean conception of historicity. His interpretation shed light retrospectively on the deep stakes of Deleuze and Guattari’s discussion. In Austin’s theory, Derrida claimed, language was reducible to speech acts and those speech acts were totally independent from “the authority of the value of truth, from the opposition true/false,” and referred to the sole “value of force, of difference of force (illocutionary or perlocutionary force)” (Derrida, 1990, pp. 37-38, my trans.). In other words, language only transferred movements or forces.

By contrast, Benveniste, who defended the linguist’s perspective, was critical of Austin’s extension of the performative to the illocutionary. He contended that the pragmatic result was unessential in performative phrases. Certainly, for Benveniste as for Austin, the performative did not describe something that existed outside and before language activity, and it proved that language was not only a mirror that would reflect states of affairs or states of the soul. But he did not consider either that the performative, as Austin claimed, and Deleuze, Guattari or Derrida after him, “produced or transformed a situation,” in short, that it “operated” and shattered the self-identity of the being in a perpetual surge of energy.

From the outset, Austin, who reflected on this issue as a moral philosopher, considered the action from the viewpoint of its completion and social outcome. He was interested in the “illocutionary” dimension of language, that is to say, in what we do when we say something, or more generally, in the action accomplished by the speaker when he or she says something (see the discussion of this point by Searle, 1969, p. 14-25 in the French trans.). Thus, for Austin, the term “performative” itself indicated “that the issuing of the utterance is the performing of an action” (Austin, 1962, p. 6, my italics). Elsewhere he also declared that “to issue such an utterance is to perform the action” [1].

By contrast, instead of taking the viewpoint of action—of the act once completed—Benveniste, as a linguist, was interested in the act itself. He considered that a performative constituted an act by itself.

The utterance I swear is the very act which pledges me, not the description of the act that I am performing. [...] The utterance is identified with the act itself. (Problems in General Linguistics, 966, trans. Mary Elizabeth Meek, 1971, p. 229)

Therefore, in his eyes, the pragmatic result in no way defined the performative. The case of the imperative was a significant example: “While Come here! is indeed an order, linguistically it is something other than saying, ‘I order you to come here’” (p. 237). The result was certainly the same, but the linguistic act by which it was produced was totally different. Benveniste continued.

There is no performative utterance unless it contains the mention of the act, namely, I order. The imperative, on the other hand, could be replaced by any procedure that would produce the same result, a gesture, for example, and would no longer have a linguistic reality. (Problems in General Linguistics, 1966, trans. Mary Elizabeth Meek, 1971, p. 237)

The conclusion was unambiguous. The “empirical result” was not what mattered. The extraordinary significance of the performative relied on the fact that it constituted “by itself an act.”

We must not be deceived by the fact that the imperative produces a result, that come here! actually makes the person to whom one spoke come. It is not this empirical result that counts. A performative utterance is not performative in that it can modify the situation of an individual, but in that it is by itself an act. (Problems in General Linguistics, 1966, trans. Mary Elizabeth Meek, 1971, p. 237)

What characterized performative phrases, therefore, was not that they extend language outside of itself, and “perform an action,” as Austin claimed, but that they are “self-referential.”

This leads us to recognize in the performative a peculiar quality, that of being self-referential, of referring to a reality that it itself constitutes by the fact that it is actually uttered in conditions that make it an act. (Problems in General Linguistics, 1966, trans. Mary Elizabeth Meek, 1971, p. 236)

We see what distinguished Benveniste’s conception of language implied by this analysis from that drawn from Austin by Deleuze and Guattari, his deconstructionist commentator, and, I would add, all pragmatist theorists—including philosophers like Richard Rorty (1931-2007) or sociologists like Jürgen Habermas (1929-)—who endorsed and extended the concept of illocutionary. For Benveniste, language could certainly not be reduced to its referential function; it was not a simple semiotic mirror of the world. But language could not be defined either by assuming a plain continuity with the world and by including it into a larger pragmatic order. Language could not be integrated into a universal theory of action without losing its specificity. Therefore, it was not possible for him, as some could argue based on Austin’s suggestions, to abandon the universality of language to the benefit of a universality of force. Our language condition was not part of a larger condition of being-in-the-world, into which we would have been thrown, as a vast field of conflicts between wandering energies that would come to express themselves in completely unconditioned ways. On the contrary, for Benveniste, it is necessary to think of being-in-the-world by starting from being-in-and-through-language, because it is language that allows us to instantiate, through acts, points of reference which found our subjectivity as well as our perception of space and time (I-here-now, deictics, verbal tenses), our referential activity (I/you-he or she) and our relation to others (I/you). It was therefore the language that founded both our humanity and our historicity and not the other way around (for more details, see Michon, 2010, Chap. 6).

Deleuze and Guattari tried to discredit this particularly strong position by accusing Benveniste of refusing “a generalized pragmatics,” which was true, and of merely reversing the real order between subjectivity and speech acts, which was quite false and made him into a naive subjectivist.

For it is always possible to thwart that move [the extension to the illocutionary]. The performative can be walled in by explaining it by specific syntactic and semantic characteristics avoiding any recourse to a generalized pragmatics. According to Benveniste, for example, the performative relates not to acts but instead to a property of self-referentiality of terms (the true personal pronouns, I, YOU..., defined as shifters). By this account, a preexistent structure of subjectivity, or intersubjectivity, in language, rather than presupposing speech acts, is adequate to account for them. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 78)

To oppose Benveniste, they quoted the French linguist and enunciation specialist Oswald Ducrot (1930-) who had published in 1972 Dire et ne pas dire. Principes de sémantique linguistique. According to them, Ducrot “reversed” Benveniste’s perspective and advocated the following thread of reasoning: “Certain statements are socially devoted to the accomplishment of certain actions,” therefore, “collective assemblages of enunciation” or “juridical acts or equivalents of juridical acts” explain the “illocutionary” which, in turn, explains “language self-referentiality” and, consequently, the so-called “subjectivity.”

Oswald Ducrot has set forth the reasons that have led him to reverse Benveniste’s schema: The phenomenon of self-referentiality cannot account for the performative. The opposite is the case; it is “the fact that certain statements are socially devoted to the accomplishment of certain actions” that explains self-referentiality [p. 73]. The performative itself is explained by the illocutionary, not the opposite. [...] And the illocutionary is in turn explained by collective assemblages of enunciation, by juridical acts or equivalents of juridical acts, which, far from depending on subjectification proceedings or assignations of subjects in language, in fact determine their distribution. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 78)

However, one wonders if Ducrot was a better reader than Deleuze and Guattari. First, although Benveniste claimed that the performative offered one example—among others—of “self-referentiality” in language, he did not advocate any structuralism-style closure of language upon itself, nor the symmetrical hyperpragmatist primacy of social context upon language promoted by Deleuze and Guattari. By emphasizing the self-referentiality of some speech acts, some pronouns, deictics and verbal tenses, he showed in fact that both views were partial and unable to describe the language in its specific pragmatic being.

First, Benveniste never forgot that, in order to be valid, a performative utterance should be pronounced by somebody socially entitled to and in the socially required conditions. In Benveniste’s own words, self-referentiality meant referring to a reality that an utterance constituted “by the [very] fact that it is actually uttered in conditions that make it an act” (p. 236). In other words, I certainly can say “I sentence you to be hanged until you are dead” to my kitchen refrigerator but this won’t work because I am not a judge and empirical conditions are not entirely suitable. As a matter of fact, Deleuze and Guattari had to recognize a few pages below that Benveniste clearly mentioned this condition, but they did not give him credit for it and argued, on the contrary, quite sophistically, that he had recognized, implicitly against his own premises, the primarily social aspect of the performative.

Benveniste clearly demonstrates that a performative statement is nothing outside of the circumstances that make it performative. Anybody can shout, “I declare a general mobilization,” but in the absence of an effectuated variable giving that person the right to make such a statement it is an act of puerility or insanity, not an act of enunciation. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 78)

Second, by saying that subjectivity was inherent in language, Benveniste did certainly not suggest that the latter was an extension of the self, the ego, the psyche, etc., nor that, as Deleuze and Guattari claimed symmetrically, “subjectifications are not primary but result from a complex assemblage” (p. 79, my emphasis). Just like in the previous case, Benveniste opposed both views and emphasized that subjectivity would emerge only through the activity of language which was immediately social and intersubjective—what Humboldt had called Thätigkeit or energeia.

We now understand better why Deleuze and Guattari—just as Serres—were so unfair with Benveniste and why, on the contrary, they praised so highly Austin who had anticipated—so they claimed—their own anti-anthropological hyperpragmatism [2]. Actually, Benveniste shared with them—and actually anticipated—much more than they were ready to accept: the critique of the reduction of language to reference, representation, or information; the rejection of the structuralist reduction of language to its formal and semiotic part; the introduction of a new perspective oriented towards activity and empirical context. But, at the same time, he stoutly opposed their view with a powerful pragmatics which involved strong anthropological concerns that did not stem from any essence of humanity but implied, on the contrary, its radical historicity.

By contrast, the limitations of Deleuze and Guattari’s hyperpragmatism appeared quite clearly when they defined language as “the set [l’ensemble] of all order-words, implicit presuppositions, or speech acts.” This implied considering language not as a discourse activity but as a collection of utterances, the meaning of which depended exclusively on social circumstances—just as structuralist thinkers had previously sustained that it depended exclusively on the internal relations of the linguistic code.

The price to pay for skipping the dynamics specific to language activity was high.

First, it involved the denied preservation of the most traditional perspective of la langue, now associated with an ontology of force, which recurred constantly in their discourse.

Order-words do not concern commands only, but every act that is linked to statements by a “social obligation.” Every statement displays this link, directly or indirectly. Questions, promises, are order-words. The only possible definition of language [langage] is the set [l’ensemble] of all order-words, implicit presuppositions, or speech acts current in a language [une langue] at a given moment. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 79)

Second, since the enunciation was not part of the discourse, that is, since it was not an activity per se, it could only be a succession of discrete statements referring to each other in an endless chain of indirect reports. “Language in its entirety [was] indirect discourse.” Therefore, due to the inevitable “presence of a reported statement within the reporting statement,” the meaning was always shattered by an interior split—even the meaning of the discourse which tried to debunk it, one was forced to recognize.

That is why every statement of a collective assemblage of enunciation belongs to indirect discourse. Indirect discourse is the presence of a reported statement within the reporting statement, the presence of an order-word within the word. Language in its entirety is indirect discourse. Indirect discourse in no way supposes direct discourse; rather, the latter is extracted from the former. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 84)

But what Deleuze and Guattari suggested here only amounted to a renovated version, under a Foucauldian guise, of the endless report from sign to sign within the structure of la langue emphasized by Derrida under the name of différance in Margins of Philosophy (1972). Certainly, like Foucault in The Archaeology of Knowledge, Deleuze and Guattari advocated sticking to the objective dispersion of the statements, and unlike Derrida they no longer referred to sign and structure, but the principle of “difference” they introduced into language was similar. It only transposed into the Foucauldian perspective, the Derridean notion that words and signs could never fully summon forth what they meant, but could only be defined through appeal to additional words, from which they differed. Thus, Deleuze and Guattari faced the same logical difficulties as their deconstructionist contemporary. Their theory was self-defeating or marred by a performative contradiction: what sense or, at least, what accuracy could have an assertion according to which any assertion is basically senseless or, at least, inaccurate?

Third, the meaning was not only ambiguous and fleeting, it was also entirely social. The statements combined into superior “assemblages of enunciation” (p. 83) which in turn combined into “a regime of signs” or “a semiotic machine.”

To the extent these variables enter at a given moment into determinable relations, the assemblages [of enunciation] combine in a regime of signs or a semiotic machine. It is obvious that a society is plied by several semiotics, that its regimes are in fact mixed. Moreover, at a later time there will arise new order-words that will modify the variables and will not yet be part of a known regime. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, pp. 83-84)

But this perspective was not that different from the Marxist belief, that they intended to criticize, in the determinism of the “superstructure” and especially “ideology” upon the actual discourse of the individuals. There again, Deleuze and Guattari’s suggestion consisted in translating into the Foucauldian vocabulary a vision supposed to be outdated but which persisted in their discourse which retained the dominant Marxist idea that individual discourses were entirely determined by social forces—except of course their own.

Fourth, this simultaneously Parafoucauldian, Paraderridean and Paramarxist conception of meaning naturally implied a very limited conception of subjectivity. Higher “regime of signs,” intermediate “assemblages of enunciation,” and lower “successions of statements” helped to frame both the individual “utterances”—the speech—and what Deleuze and Guattari termed cryptically the collective “incorporeal transformations attributed to the bodies” (p. 80)—what we may probably call, more simply, the socially accepted significations.

Whereas Benveniste revolutionized the concept of subjectivation by establishing it at equal distance between individualist and holist perspectives, as well as psychological and sociological conceptions, they interpreted his suggestion in the psychological sense of subjectivisation – “subjectification,” and reduce the concept, for their part, to its most common form of assujettissement – “subjection” by the social systems.

There is no signifiance independent of dominant significations, nor is there subjectification independent of an established order of subjection. Both depend on the nature and transmission of order-words in a given social field. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 79)

According to them, since the enunciation was entirely socially determined, the subjectivity involved in it was strictly limited by the “impersonal collective assemblages” that framed its performances.

There is no individual enunciation. There is not even a subject of enunciation. [...] The social character of enunciation is intrinsically founded only if one succeeds in demonstrating how enunciation in itself implies collective assemblages. It then becomes clear that the statement is individuated, and enunciation subjectified, only to the extent that an impersonal collective assemblage requires it and determines it to be so. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, pp. 79-80)

However, Deleuze and Guattari were unable to fully maintain this self-defeating position—what about their own “discourse” and their own “subjectivity”?—and were eventually forced to pay homage to Benveniste. As any other “order-word,” they claimed at first, “I” depended “on a molecular assemblage of enunciation,” in other words, it was just an effect of the statements assemblages—just as it was an effect of the linguistic structure or of the ideology in former theories. However, they could not support this line of reasoning until the end. “To write,” they had to admit, was to “bring to the light,” to “select,” or to “gather” whispering voices, secret idioms “from which I extract something I call my Self.” With this remark, they implicitly joined the long list of authors who had recognized that “Je est un autre – I is another” (Arthur Rimbaud, Letter to Georges Izambard – May 13, 1871) or that “Ah ! Quand je vous parle de moi, je vous parle de vous – When I speak to you about myself, I am speaking to you about yourself” (Victor Hugo, Les contemplations, 1856). But Deleuze and Guattari also recognized in extremis the language activity through which “I,” which is only an empty form available to anybody, is filled with a specific content and the subject built. In this sense, as Deleuze and Guattari claimed, I is an “order-word” coming from others but, contrarily to their opinion, it does not possess any constraining power. The “I” in “I extract something I call my Self” cannot be just an effect of social assemblages of statements; it is obviously an emerging effect of the activity of discourse itself intimately intertwined with the experience and social activity of the speaker.

I always depend on a molecular assemblage of enunciation that is not given in my conscious mind, any more than it depends solely on my apparent social determinations, which combine many heterogeneous regimes of signs. Speaking in tongues. To write is perhaps to bring this assemblage of the unconscious to the light of day, to select the whispering voices, to gather the tribes and secret idioms from which I extract something I call my Self [quelque chose que j’appelle Moi]. I is an order-word. (A Thousand Plateaus, 1980, trans. B. Massumi, 1987, p. 84)

Next chapter


[1Quote drawn by Benveniste from the proceedings of the Royaumont colloquium devoted to analytical philosophy: “Analytical philosophy and language” (1963), in Problèmes de linguistique générale, 1966, p. 269. Trans. Mary Elizabeth Meek, 1971, p. 233.

[2To avoid any misunderstanding, I must say here that, by the terms “ontological pragmatism,” “generalized pragmatism” or “hyperpragmatism,” I am not referring to the Anglo-Saxon tradition which developed in the wake of the work of Charles Sanders Peirce, but to a philosophical position—to which at least part of the Peircean tradition seems, however, to adhere—which holds that language is secondary to energy, force and action (pragma).

Follow site activity RSS 2.0 | Site Map | Private area | SPIP