A Re-Reading of Ferdinand de Saussure

Copyright © Jim Loter, 1996

In Formalism and Marxism, Tony Bennett reassesses the various works of the so-called Russian Formalists. He argues that although the Formalists "tended to be apolitical in their approach, viewing the aesthetic effect of defamiliarization to which works of literature were said to give rise as an end in itself, divorced from political considerations or consequences," (27) their attempt to define the "literariness" of texts actually rose from a desire to define a discrete object of study for an autonomous science of literature which would then allow for a socio-historical account of literary change (34). This socio-historical, even political, side of the Formalists is apparent in that "new literary forms are called into being ... by the need to challenge and disrupt those forms and conventions ... whose cutting edge has been dulled through overuse" (34). In other words, far from being apolitical, the Formalists, in Bennett's view, treat the work of isolating literature as an autonomous realm is a first step necessary to identify the particular historical, social, and economic influences affecting that realm. Bennett relates the position of the Formalists to that of Louis Althusser who also "conceives of society ... as consisting of a number of relatively autonomous 'instances' or levels of social practice" (Bennett 36). Bennett does not conclude that the Formalists were nascent Marxists, but he does suggest that a "meaningful dialogue between Formalism and Marxism" is now possible with this reconsideration of the political face of the Russian Formalists.1

The mediating paradigm through which this dialogue can take place, Bennett argues, is Saussurian linguistics. "Of the many influences which coalesced in the formation of Russian Formalism," Bennett writes, "that exerted by Saussure ... was arguably of primary importance" (44). For Bennett, Saussure's attempt to isolate la langue (the language system) from la parole (individual uses of language) mirrors the attempt by the Formalists to isolate literary language from everyday language. Furthermore, Bennett defends Formalism against the claim that they "tended to fetishize the literary device" (50). On the contrary, the Formalists believed that literary defamiliarization depended not only on static form but also on dynamic usage. Therefore, according to Bennett, "the Formalists were again following Saussure in the distinction he maintained between the synchronic and diachronic levels of analysis" (51). In short, then, the Russian Formalists shared with Saussure not only a similar need for a methodological division between elements of stasis and of change, but they also agreed that only the static, or synchronic, side of the division could be studied as a science. For Saussure, this object of study was la langue; for the Formalists, it was literature.

Of course, Saussure's famous edict that his object of study -- la langue -- is "immutable" causes major problems for Bennett whose entire argument about the political side of Formalism rests on the assertion that the Formalists were actually concerned with explaining literary change. If their project really did line up neatly with Saussure's, their attempts at isolating literature -- or literariness-- really would be "'underlabourers of the device'" as Trotsky pronounced them (Bennett 29). Bennett's solution to this problem is to criticize Saussure's declaration of the immutability of la langue and offer up Formalism (and the historical poetics of Mikhail Bakhtin in particular) as the corrective to Structuralist literary analysis which is based heavily upon Saussure and, therefore, remains impotent in the face of the historical fact of change.

Bennett's critique of Saussure rests upon a misunderstanding of both Saussure's conception of la langue and the significance of his sweeping elimination of linguistic change from it. The consequences of this error do not so much undermine Bennett's attempt to reclaim the long-vilified Formalists for Marxist literary criticism as they point to a deeper problem in Saussurian-inspired semiotics. Rather than see the Saussurian influence on the Russian Formalist project as a "double-edged sword," as Bennett would have it (74), I would suggest that Saussurian linguistics, by the two very premises Bennett links it with Formalism, is generally not applicable to the study of literature or any other use of a language system. 2

In his Course in General Linguistics, Ferdinand de Saussure defined "semiology" as "a science that studies the life of signs within society;" this new science "would show what consitutes signs, what laws govern them" (16). In order for it to do so, however, Saussure separates la langue -- the syntagmatic axis of language (langage) -- from la parole, or the paradigmatic axis. He writes: "Taken as a whole, speech [parole] cannot be studied, for it is not homogeneous" (Saussure 19). In the interests of science, therefore, Saussure declares the constantly changing, kinetic, daily uses of speech out of the realm of semiology. The true object of semiotic study is la langue which, according to Saussure, "is a well-defined object in the heterogeneous mass of speech facts. It can be localized in the limited segment of the speaking-circuit where an auditory image becomes associated with a concept" (14). Additionally, la langue "is a system of signs in which the only essential thing is the union of meaning and sound-images, and in which both parts of the sign are psychological" (15). It is with these two strategies -- the identification and separation of two distinct realms, and the focus on only one of these realms -- wherein Bennett locates Formalism's debt to Saussure.

The step taken by Bennett which is not justified, however, is the conflation of the methodological separation of categories with the ontological identity of those categories. In other words, Bennett constructs a fallacious relation, to wit:

la langue is to literature as la parole is to common speech

In so doing, Bennett launches his critique of Saussure's conception of la langue as unchanging. If la langue corresponds to literature and is unchanging, then how can Sausurrian semiotics account for historical changes in literary forms? For Bennett, this question entails a revision of Saussure's notion of la langue. He finds this in Volosinov who declares that the "proper object of linguistics ... [is] the ways in which the rules comprising [language] are used, modified, and adapted in concrete utterances..." (Bennett 76). Saussure is quite clear not only in his distinctions between la langue and la parole, but also in his distinction between linguistics and semiotics. "Linguistics is only a part of the general semiology," he writes in Chapter III of the Course. "To determine the exact place of semiology is the task of the psychologist. The task of the linguist is to find out what makes language a special system within the mass of semiological data..." (16). Linguistics, therefore, is not a science and has as its task the study of speech-acts, la parole, "the executive side" of human language (langage) (13). It is difficult here to see how different Volosinov's conception of linguistics is from Saussure's, and, therefore, quite difficult to understand how Bennett sees this conception as a "violent exception" (75). In the above terms, Volosinov's "theoretical shift" from the rigid framework of the unchanging langue to a viewpoint that "dialogic relationships must be placed at the centre of analysis" amounts to nothing more that his decision to work within linguistics rather than semiotics in his study of literary texts. Since a literary text belongs to la parole -- it certainly being "an individual act ... wilful and intellectual" (Saussure 14) -- Volosinov must operate in the realm of linguistics and not semiotics.3

So semiology as a science remains unscathed by Bennett's critique, but in so demonstrating this, we point to another problem inherent in Structuralist adoptions of Saussure. In light of Volosinov's return to linguistics for properly accounting for individual acts of meaning-making (texts), it seems as if Saussurian semiology is decidedly useless for the study of such acts. If we are to take Saussure at his word (as Bennett and Volosinov both claimed to do) we are left with a science that recognizes that "language is a convention" but also observes that "the nature of the sign that is agreed upon does not matter" (10). In this sense, linguistic (or literary) change is irrelevant because changes "only affect the material substance of words.... Determining the causes of phonetic changes may be of interest ... but none of this is essential" (Saussure 18). The object of the science of semiology is not this heterogenous, radically mutable parole, but the system of language itself: la langue.4

A clearer understanding of la langue can help us further the claim that semiology, as Saussure defines it, is not designed for the formal study of literary texts. In his famous rending of the linguistic sign into the "signifier" (the material marker) and the "signified" (the concept), Saussure also declared that the relationship of the one to the other is "unmotivated, i.e. arbitrary" (69). The signifier and signified are, nonetheless, "intimately united, and each recalls the other" (66). Such are the rules of la langue. These rules are, according to Saussure, immutable; in fact, "the arbitrary nature of the sign is really what protects language from any attempt to modify it" (73). Within the realm of la langue, "there is no reason to prefer soeur to sister," (73); or, in other words, the processes by which signification occurs do not dictate any compulsion to choose a given signifier over another. The choice is not based on the resemblance of a signifier to the thing it signifies; it is not based on any natural, reasonable, or scientifically quantifiable relation at all. 5

Bennett criticizes Saussure on this point explicitly. He writes,

as a body of rules pre-existing the system, la langue is conceived of as a totally unitary system. The strorehouse and embodiment of an undifferentiated societal collective consciousness, it gives rise merely to the concept of the ideal-typical speaker and has no room for the concept of class-based linguistic practices.... (72)

In one sense, of course, Bennett is correct: the isolation and study of the syntagmatic, enduring system of la langue cannot account for changes along its paradigmatic access -- literary or quotidian. But this critique is misguided as Saussure nowhere states any pretension to account for these changes. The rules of la langue, far from "pre-existing the system," actually are the system. Furthermore, the study of la langue not only has "no room for ... class-based [or any other] linguistic practices" -- as Bennett correctly observes -- but neither does it not have room for an "ideal-typical" speaker either. Bennett's critique essentially amounts to a mere expression of frustration at not having the right tool for the job he wishes to do.

Bennett's critique of Saussure ultimately faults the semiologist for even attempting to methodologically sever the syntagmatic from paradigmatic axes of language. He sites Fredric Jameson who writes, "Once you have begun by separating diachronic from synchronic ... you can never really put them back together again" (64). True or not, it is this same logic by which Bennett defends the Russian Formalists in their attempt to isolate "literature" as a discrete object of study. As I mentioned earlier, Bennett argues that the Formalists attempted to describe "literature" as something akin to what Althusser later called "relatively autonomous levels of social practice." Their identification of an autonomous object of study was the necessary first-step in a program to account for literary change. Granted, the Formalists formulated their split of the syntagmatic (literature) from the paradigmatic (convention) more dialectically than Saussure conceived of his own, but upon close inspection, the paradigms described by the Formalists seem to be made of paper tigers against which the concept of ostranenie appears more radical.

Shklovsky's description of the technique of "defamiliarization," for example, rests on the assumption that "normal" perception is familiar, economical, and can be described by "the process of 'algebrazation'.... Either objects are assigned only one proper feature ... or else they function as though by formula and do not appear in cognition" (12). In short, ordinary perception is just ordinary, familiar. "The technique of art," on the other hand, "is to make objects 'unfamiliar,' to make forms difficult" (12). For Shklovsky, ordinary speech is the communication of ordinary perception and, likewise, is "economical, easy, proper ... the 'direct' expression of a child," whereas "poetic speech" is concomitant with art's goal to defamiliarize (23). It is easy to see from these brief examples that Skhlovsky defines the object of Formalist attention -- poetic language -- against an illegitimately impoverished definition of prose. Thus, though the two sides of the Formalist split do operate in a more symbiotic relationship than do Saussure's langue and parole, and, consequently, they do not completely exclude the matter of change as does semiology, one must be willing to accept a certain amount of reductionism from the Formalists for the sake of this increased interrealation.

In sum, Tony Bennett's attempt to open a critical exchange between Russian Formalism and Marxism appeals to the dialogic "revisions" of Saussure's concept of la langue offered by Valentin Volosinov and the dialogic criticism perfomed by Mikhail Bahktin. In this embryo of a paper, I have tried to show that although Bennett is quite correct that Saussure's formulation of the science of semiology does not take into account linguistic change, this criticism is misplaced. Saussure, in my understanding, never claimed that semiology was to be used as a method for accounting for the changes brought about by common usage and speech (la parole). Moreover, in addition to this critique's being misplaced, it is also contradictory in that Bennett's point of dispute with Saussure (the isolation of la langue in order to study the static system of language) is logically akin to Bennett's own defense of the Formalists (the isolation of the unchanging element in literature in order to study the process of literature). Though the isolation of the literary in Formalism is not quite as rigorous as Saussure's isolation of la langue, it is based on an unsubstantiated purging of all that is poetic from non-"literary" language. The consequences of all this seem to point to the fact that Saussurian semiology in its exclusive emphasis on the general system of language and not on any particular use of language is uniquely not suited as a method for studying texts.


1 Critics of Althusser have noted the immense difficulty in trying both to maintain an autonomous, distinct realm of inquiry and to account for its mutability by appealing to outside causes. Paul Hirst has written that such a division of "relative autonomy" leads Althusser back to economic determinism and, again, to a simplistic base-superstructure model, or it must eliminate any relation between the separate realms. (This position is summarized in Lapsley and Westlake's Film Theory: An Introduction; and Terry Eagleton's Literary Theory: An Introduction). Based on the strength on critiques of Althusser, it seems to be an unfortunate move for Bennett to draw such close parallels between Formalism and Marxism at the Althusserian juncture, but it is not the focus of this paper to investigate that strategy any further.

Return to the essay

2 This is not to say that Saussurian linguistics is of no relevance to the realm of literary or film analysis. In fact, though certain of Saussure's insights (which I hope to flush out) demonstrate that the existence of his semiology leaves virtually unchanged the project of critical analysis, it is by this very insight that such a project is more accurately defined. In other words, Saussure's general linguistics defines textual criticism more by explaining what it is not than what it is.

Return to the essay

3 The idea that a literary text is produced by an "individual act" may be criticized in this context as Bennett's point is that such ideas fail to account for the social pressures exerted upon the production of a text. Indeed, Saussure's declaration that for an act in la parole "the individual is always its master" seems to offer a na´ve intentionalty as a guide to the meaning of an utterrance. It must not be forgotten, however, that Saussure continually insists that the parole/langue distinction is strictly methodological and, therefore it can be deduced, that Saussure believes that in practice, individuals do not exist outside their socio-historical context.

Return to the essay

4 Another objection may be raised that literary changes and phonetic changes are not the same thing. It is true that because of the value placed upon literary texts, changes in form and language within them may be seen as more important than simple shifts in sounds, dialects, etc. At the conceptual level, however, the smallest meaningful segment is a phoneme and, therefore, any change whatever in such an entity changes the meaning of a text (cf. Benveniste). In other words, the difference between "Her sister was called Tatyana..." (to cite Shklovsky's example of "roughened" form) and "`Twas brillig, and the slithy toves..." is, in a linguistic sense, a difference of phonemes. Expanded, perhaps, a linguistic analysis can perhaps hit upon some social and political influences on such a shift, but it must be remembered that to identify a difference is fundamentally to identify a linguistic difference. Semiology, in Saussure, is not concerned with such changes of parole.

Return to the essay

5 Of course, Saussure does spend a bit of time on the mutability of the sign. He defines "change" as not simply "phonetic changes undergone by the signifer, or changes in meaning which affect the signified concept" but as "a shift in the relationship between the signified and the signifier" (75). Though this tends to point to the fact that even as speech changes, it always changes within the binary structure of the signifier/signified dyad (and, hence, backs up my definitions of la langue and la parole), Saussure does write about individual word changes which seem to belong firmly on the side of parole and outside of his science. His recourse simply to "time" as the agent of change is also quite a problem, but I don't have the time to go into it here. For this paper, I will bracket aside the part on the "Mutability of the Sign."

Return to the essay