previous chapter
Chapter 5 Literature and Ideology
next sub-section

Pêcheux: The Concept of Discursive Practice

In his "Letter on Art in Reply to André Daspre" (in Althusser 1971), Althusser speaks of the peculiarity of art in terms of its ability to "make us see, make us perceive, make us feel something which alludes to reality. . . . What art makes us see, and therefore gives to us in the form of 'seeing,' 'perceiving,' and 'feeling' (which is not the form of knowing ) is the ideology from which it is born, in which it bathes, from which it detaches itself as art, and to which it alludes" (Althusser 1971, 222). For Althusser, art is thoroughly ideological; it is a form of ideological practice, grounded in the immediate, lived experience of subjects, yet somehow different from this ideological background. At the same time, the way art "alludes" to its ideological materials is clearly distinct from the way science perceives them. Unfortunately Althusser's terminology is vague, and terms such as seeing, perceiving, knowing , and alludes raise questions that must be answered before his argument can be assessed. Restricting myself to the specific example of literature, I will attempt to develop Althusser's insights in a more rigorous fashion.

Perhaps the proper place to begin is with the problem of discourse itself. Literature "alludes" but doesn't "know," Althusser contends, yet it nonetheless makes us "see, feel, and perceive." The possibility of such a paradox results, presumably, from two sets of differences, the difference between ideological discourse and scientific discourse on the one hand and the difference between literary discourse and other types of ideological discourses on the other. Michel Pêcheux, in Language, Semantics, and Ideology , undertakes what he calls a materialist theory of discourse, which is both a Structural Marxist critique of the dominant forms of contemporary linguistics and an attempt to specify the interrelationship between language and its social environment.[2] Pêcheux focuses on the semantic element of linguistic practice because, in his opinion, semantics is a source of contradiction for both Saussurean linguistics and Chomskyan "generative grammar": Saussurean linguistics is unable to explain historical changes to the linguistic system (langue ), the relation of this system to the social formation as a complex whole, the nature of individual utterance (parole ), or the existence of the subject who speaks; Chomsky's generative grammar attempts to derive all semantic possibilities from syntactic "deep structures," but this move eliminates social factors altogether. We cannot pursue the details of Pêcheux's critique here; however, his proposal for a new theory of discourses, discursive practices, and discursive formations and his discussion of the relation of these to the "subject-form" and the


interpellation of subjects are essential to our investigation of ideological, scientific, and literary discourses.

The major difficulty with existing approaches to these problems, Pêcheux observes, is their uncritical assumption of a pre-given subject who speaks. The self-evidentness of the subject, which Pêcheux identifies as the "spontaneous philosophy of linguists," must be rejected in favor of analyses of the "subject-form (and specifically the subject of discourse) as a determinate effect of a process without a subject" (Pêcheux 1982, 51). To accomplish this task, Pêcheux makes a distinction between discourse (semantic elements) and language (syntax, enunciation). Discourse presupposes language but is not determined by it; discourse is rather a function of historical determinations on semantic processes. As a discursive practice, language (like any ideological practice) does not have its source in the speaking subject; rather, it finds its product in individuals constituted as speaking subjects. Discourse, in other words, interpellates subjects, and it is this insight that informs Pêcheux's investigation of the subject-effect of language and the possibility and nature of meaning. "All my work," Pêcheux explains, "links the constitution of meaning to that of the constitution of the subject which is located in the figure of interpellation" (Pêcheux 1982, 101). For Pêcheux, ideology supplies the self-evidentness of meaning—while at the same time masking the relationship of meaning and interpellation by means of the "transparency" of language.

The "material character of meaning" is a function of what Pêcheux calls discursive formations and interdiscourse . He offers the following two theses by way of explanation:

1. The first consists of the proposition that the meaning of a word, expression, proposition, etc., does not exist "in itself" . . . but is determined by the ideological positions brought into play in the socio-historical process in which words, expressions and propositions are produced. . . . [W ]ords, expressions, propositions, etc. change their meaning according to the positions held by those who use them , which signifies that they find their meaning by reference to those positions, i.e., by reference to the ideological formations . . . in which those positions are inscribed. . . . I shall call a discursive formation that which in a given ideological formation, i.e., from a given position in a given conjuncture determined by the state of the class struggle, determines "what can and should be said. " . . . I shall say that individuals are "interpellated" as speaking-subjects (as subjects of their discourse) by the discursive formations which represent "in language" the ideological formations that correspond to them.

2. Every discursive formation, by the transparency of the meaning constituted in it, conceals its dependence on the "complex whole in dominance " of discursive formations, itself imbricated with the complex of ideological


formations . . . I propose to call this "complex whole in dominance" of discursive formation "interdiscourse," with the qualification that it too is subject to the law of unevenness-contradiction-subordination . . . characterizing the complex of ideological formations. (Pêcheux 1982, 111-13)

The interpellation of subjects is realized within interdiscourse by means of two mechanisms that Pêcheux identifies as preconstruction and articulation . The preconstructed corresponds to the "always already there" quality characteristic of all ideological interpellation. Ideology interpellates the individual as the subject of discourse, yet the interpellation is impenetrable to the subject, who appears to himself or herself as always already a subject. Pêcheux calls this the "Munchausen effect" after the immortal baron, who proposed to extricate himself from a swamp by pulling on his own hair. The preconstructed element of discursive practice is always already there, Pêcheux contends, because it is a function of interdiscourse itself and not of any particular discourse.

It is the preconstructed element in discourse that creates the subject's identification with himself. The second element, articulation, creates the relationship of the subject to other subjects and to the Subject. Articulation, for Pêcheux, refers to the linear system of discourse: the system of co-references that clarify the operation of discourse with respect to itself (what I am saying now in relation to what I said before) such that a "thread of discourse" is established as the discourse of a subject and, as such, is recognizable by other subjects. Because this is an operation of discourse in relation to itself, Pêcheux refers to it as intradiscourse . Intradiscourse, he explains, "crosses and connects together" the discursive elements constituted by preconstruction which are the raw material, the primary stuff of discourse. However, the mechanisms of this process are such that the primacy of the preconstituted is reversed; rather than appearing as determined by interdiscourse, intradiscourse "forgets" this determination and appears autonomous. "The subject-form (by which the subject of discourse identifies with the discursive formation that forms him) tends to absorb-forget interdiscourse in intradiscourse, i.e., it simulates interdiscourse in intradiscourse" (Pêcheux 1982, 117). As a result of this simulation, what should be at issue, namely, the stable identity of the subject (or other referents), comes to be guaranteed by the thread of the discourse that connects them.

By all of these mechanisms, discourse simultaneously produces an identification of the subject with himself or herself and an identification of the subject with the other subjects. Pêcheux refers to these identifi-


cations as the subject effect and the inter-subjectivity effect wherein, respectively, the subject is self-evident, and individual subjects recognize and ratify one another as a mirror image guarantees the existence of the reflected object. Thus discourse masks its own determinative action under the guise of the autonomy of the subject. Subjects cannot recognize their subordination to the Subject because this subordination-subjection is realized in subjects in the form of autonomy: the freedom of speaking subjects to say whatever they want. It is in this way that in language, as in other ideological practices, the process without a subject creates-constitutes-realizes the subject: the subject is created by "forgetting" the determinations by which he or she is produced.

We may now return to our original question concerning the distinctions between literary, scientific, and ideological discourses and reformulate the question more precisely in terms of the subject-form. If ordinary discourse, the basic everyday form of ideological discourse, operates in the manner we have just discussed, how does it differ from literary and scientific discourses? With regard to the first, it is certain that literary discourse is also deeply indebted to the coincidence between the subject and self and between the subject and other subjects. Indeed, Pêcheux goes so far as to argue that literary fiction represents "the pure idealist form" of the subject effect and that the "novelistic effect of presence finds its origin in the subject-form and, as a result of it, is able to mask the materiality of aesthetic production" (Pêcheux 1982, 119). However, despite its reliance on the subject-form, literary discourse remains distinguishable from ordinary discourse by virtue of the fact that it is a second level of discourse—one that assumes the basic preconstruction-articulation relation of ordinary ideological discourse and produces another new discourse from it. What Pêcheux's remarks fail to make clear is the process whereby ordinary discourse becomes, as it were, the preconstructed for this new discourse, the "always already there" of literature, which is "forgotten" in the very "creative" process which is its effect. As in the case of ordinary discourse, this invisibility leaves its mark, what Derrida calls its "trace," but the shadow of this absence falls somewhere outside literary discourse. This explains why even the most self-conscious and determined of literatures, Brecht, for example, cannot break the illusion of presence at its fundamental level, the level of the subject-form. At most it is able only to break with or undermine presence at the second level, the level of literary discourse itself, while relying even more strongly on the illusion of presence at the first level in order to achieve this effect.

We shall return to literary discourse shortly, but perhaps things will


be clearer if we turn for a moment to scientific discourse. If, as Althusser claims, ideology has no outside for itself, how can it be, as he further contends, nothing but outside for science? What can a scientific discourse be? Pêcheux's method excludes several unsatisfactory approaches: the use of the subject form to break out of subjective discourse either by an individual effort (Husserl) or by a collective one (Habermas); the use of symbolic logic to guarantee absolute knowledge either positively (Carnap, the Vienna Circle) or negatively (through falsifiability, as in Popper); treating scientific discourse pragmatically as the most convenient ideology at a given moment and in a given circumstance, convenience being posited in the form either of a "game" (Wittgenstein) or a "paradigm" sanctioned by the consensus of scientists (Kuhn). Pêcheux contends that the concept of the epistemological break, taken in conjunction with Althusser's thesis that all practice takes place through subjects and therefore in ideology, provides the point of departure for an alternative explanation.

At the level of the history of science(s), Pêcheux asks, how is it that science emerges from the historical process? He begins with the recognition that there is no pre-scientific "state of nature" where knowledge effects are completely absent from the field of discourse (Pêcheux 1982, 129-37). However, "what is peculiar to the knowledges (empirical, descriptive, etc.) prior to the break in a given epistemological field is the fact that they remain inscribed in the subject-form, i.e., they exist in the form of a meaning evident to the subjects who are its historical supports, through the historical transformations that affect that meaning. The result of this for discursivity . . . is that, this being so, the knowledge effect coincides with a meaning effect inscribed in the operations of a discursive formation . . . that constitutes it" (Pêcheux 1982, 136). The historical process that opens up the conjuncture of an epistemological break is characterized by Pêcheux as the gradual formation of a "block" such that, within the complex of a particular discursive formation, the identity of knowledge and meaning no longer "works." Contradictions within this identity accumulate and begin to repeat themselves in a circular manner until "the very structure of the subject-form (with its circular relation of subject/object) becomes the visible 'limit' of the process" (Pêcheux 1982, 136). In short, the historical moment of the break inaugurating a particular science is necessarily accompanied by a challenge to the subject-form and the "evidentness" of meaning that is part of it. "What is specific to every break is . . . that it inaugurates, in a particular epistemological field, a relationship be-


tween 'thought' and the real in which what is thought is not, as such, supported by a subject" (Pêcheux 1982, 137).

It is the absence of the subject-form that distinguishes scientific concepts from ideological representations (or "meaning" in the strict sense of an articulation between preconstituted subjects). Concepts of science do not have a meaning for Pêcheux; rather, they have a function in a process without a subject of meaning or knowledge. "In the conceptual process of knowledge, the determination of the real and its necessity, a necessity independent of thought, is materialized in the form of an articulated body of concepts which at once exhibits and suspends the 'blind' action of this same determination as subject effect (centering-origin-meaning)" (Pêcheux 1982, 137). This is what Althusser implies when he characterizes scientific practice as a process without a subject. This result is startling at first, for Pêcheux seems to be denying the existence of scientific discourse—at least in the sense of a discourse between the subjects of science. Of course, he does not deny that scientists as well as others discourse about science, that is, speak "scientifically," and communicate knowledge effects. They do so, however, only on the condition of returning to the discourse of the subject-form, the essential condition of all discourse. However, the appropriation of concepts by discourse introduces a contradiction into the latter, what Pêcheux calls disidentification . Disidentification is not the abolition of the subject-form, which is impossible, but a "transformation-displacement" of this form; in effect, disidentification is an interpellation of a new type. The ideological mechanism of interpellation does not disappear, of course—there is no "end of ideology" as a result of science—but it does begin to operate in reverse, on and against itself, through the "overthrow-rearrangement" of the complex of ideological formations and the discursive formations that are imbricated with them. In short, Pêcheux maintains that the appropriation of scientific concepts by the subject-form tends to undermine ideological identification in a way that other ideological discourses, for example, literature, cannot since they are trapped within a field of representation-meaning constituted by and for the subject-form.

We should not, Pêcheux cautions, conclude that the process of disidentification occurs automatically or necessarily: "The disidentification effect inherent in the subjective appropriation of knowledges is achieved in different ways (and may in the limit case not be achieved at all) according to the nature of the discursive formation which provides this effect with its raw material" (Pêcheux 1982, 163). Since no subject


can be established by science, where representation-meaning is suspended, the conflict between identification and its opposite always takes place in ideology and, in theoretical form, in philosophy. Precisely because of the natures of science and ideology, science can never break with ideology "in general" but only and always with a specific ideology. Furthermore, this break inaugurates a continuous struggle that is never over or won. At the level of discursive practices, the subjective appropriation of scientific concepts is continually threatened by the subject-form of discourse and by the ideological apparatuses that interpellate all social subjects.

The conceptual-experimental (scientific) operation, [which] in forms specific to each branch of the production of knowledges materializes necessity-real as necessity-thought (and to that extent locally forecloses meaning and the subject-form), never exists "in a pure state," in a form disjunct from its notional-ideological counterpart. Consequently the appropriation of the real by thought cannot consist of a de-subjectification of the subject, but presupposes a work of subjective appropriation in and on the subject-form, i.e., amongst other determinations in and on the subject-form of discourse. (Pêcheux 1982, 191)

The ideological effect of conceptual knowledge consists, then, of an interpellation, but of a distinct and different type since it works to undermine existing interpellations, at least locally, at the level of discursive practice itself. We must keep in mind that contradictions at the discursive level are neither the most important contradictions on which the possibility of social change hinges, nor are they the only contradictions at work within the social formation or even in the ideological instance. If they are arguably the most fundamental contradictions, in the sense that all social practice rests on the subject-form, they are nonetheless overdetermined by other relations, in the last instance by struggles between classes. The complex, overdetermined nature of interpellation explains why its "success" in subjugating the individual to the status quo can never be complete even at the level of discourse. The uneven development of ideological interpellation produces new contradictions at the same time that it reproduces the contradictions of the existing relations of domination. This is why ideology must be defined in terms of the transformation as well as the reproduction of social relations and why the conflict between identification and disidentification in discourse, like all ideological conflicts, is determined by relations of power outside its local terrain.


previous chapter
Chapter 5 Literature and Ideology
next sub-section