From Work to Text?

An essay by Lisa Smith

Geoffrey Thurley asserts that "Barthes's essay ["From Work to Text"] is no more than an expression of a determination not to acknowledge the old `bourgeois' literature, and we should waste our time chasing shadows if we tried to assign his phrases any concrete meaning" (Thurley 228). Though this comment was meant as a criticism of Barthes's work, the second part of the sentence ironically approximates Barthes' own formulation of his essay. His essay offers "not argumentations but enunciations, `touches', approaches that consent to remain metaphorical" (Barthes 192). Meaning, in Barthes' essay, is established by difference - his entire text consists of a play of signifiers which have meaning, not in themselves, but in relation to other signifiers. The concept of "text" is defined only as it is juxtapositioned with the concept of "work." To attempt to assign a `concrete meaning' to "From Work to Text," to attempt to find the kernel of meaning behind this essay would be unBarthean in both spirit and approach. That does not mean that this text is fundamentally incomprehensible. It does mean that, while one might attempt to grasp the sense of what a "work" might be or what a "text" might be, it would be going against the grain of Barthes's essay to attempt to discover the essence of the text and the essence of the work. Accordingly, in this essay, I will attempt to delineate rather than to define the signifying fields of "work" and "text" as developed in Barthes' "From Work to Text."

Before I begin to describe what a "work" and what a "text" might be, however, it is important to determine whether these terms reflect an approach to literature or whether they reflect some qualitative aspect of literature itself. In other words, can any piece of writing be viewed as either a work or a text, or are there some texts that are characteristically texts and others that are characteristically works?

On one hand, Barthes implies that there is a concrete quality to some writing which identifies it as "text" and not as "work." When discussing the issue of whether texts can be seen as a product of modernity, he comments: "There may be `text' in a very ancient work, while many products of contemporary literature are in no way texts" (Barthes 193). This text, however, has no objective reality, but exists only as praxis: "The Text is experienced only in an activity of production" (Barthes 193). The text, then, is not the physical entity that sits on a bookshelf, but something that comes into existence only as it is done.

According to Barthes, only certain types of works can be produced as texts. Some works tend to represent an attempt to circumscribe meaning; others require the reader to co-author the text. "I can delight in reading . . . Proust, Flaubert, Balzac . . . But this pleasure . . . remains in part . . . a pleasure of consumption; for . . . I know that I cannot rewrite them" (Barthes 197). The Text, insofar as it invites the reader to rewrite it, "is bound to jouissance, that is to a pleasure without separation" (Barthes 197).

On the other hand, he claims that anything could be read as a text. Despite his later claim that Proust's work is not a text, he writes that Proust's life can be read as text. "It is the work of Proust, of Genet which allows their lives to be read as text" (Barthes 195), he writes. He seems to indicate that the applicability of the terms "work" and "text" is determined, not by some inherent quality within a piece of writing, but by ways of seeing literature in general. For instance, he proposes that a work is read under the auspices of the "myth of filiation" (Barthes 195) while a text is read as a part of the vast network of intertextuality. "The work is caught up in a process of filiation . . . As for the Text, it reads without the inscription of the Father" (Barthes 195). Whether one chooses to subscribe to the `myth of filiation' or not seems to be largely a function of one's reading methodology and not of one's discernment of qualities inherent in the specific work/text.

Thurley is inclined to dismiss this aspect of his theory as "profound uncertainty" (Thurley, 228). I do think, however, that Barthes's statements work together to form, a tenable, if problematic position. His view of the text does set up a methodology of reading. At the same time, Barthes is recognizing that some texts are more amenable to this methodology than others. All pieces of writing are "texts" - some, however, are more text-like than others because they intentionally resist closure. The text written as a text consciously submits to the symbolic nature of language. "Thus it is restored to language" (Barthes 194), Barthes writes.

In this essay, I will be primarily concerned with exploring how Barthes delineates a methodology of reading. This is more revolutionary than merely claiming that some works are written with the text in mind while others are merely written as works. For not only the way in which we express meaning, but our conception of the very nature of meaning itself is transformed as we follow the "epistemological slide" (Barthes 192) that marks the shift of focus from the work to the text.

The conception of the text reflects an alteration of what is considered to constitute meaning. Barthes explains that his notion of the Text is inextricable from a certain way of viewing language. Language, he writes, "is structured but off-centred, without closure" (Barthes 194). There is, then, no meaning, per se, for there is no centre, no locus of meaning to be found. Meaning does not represent the presence of the sign, but rather is a representation of the deferment of meaning: "The work itself functions as a general sign and it is normal that it should represent an institutional category of the civilisation of the Sign. The Text, on the contrary, practices the infinite deferment of the signified" (Barthes 193).

Since there is no locus of meaning to be found in the text, the traditional boundaries placed on meaning in a work are no longer guides to meaning. If meaning is presence, it can be presented in a controlled way by the author and can be seen to occupy a defined space. If meaning is a chimera; if meaning is in fact the absence, the deferment of meaning, then it is fundamentally uncontrollable and refuses to be confined and tied down to a specific signification. The language in a text is not defined by the language in the work which lies between the covers of the book. In Barthes's words, "the work can be held in the hand, the text is held in language" (Barthes, 193).

Accordingly, critics who work within the conceptual space of the work circumscribe this space. The meanings found within the work are determined by the author, by the work's historical place within the world, and by its relation to other works (Barthes, 195). The relation between works, however, is subordinated to the authority of the work itself. For while the sources "influence" the work, the determinations of authorial intention dictate the way in which the sources will be shaped (Barthes 195).

According to the concept of the text, by contrast, the sources are not the background of the work which can be separated from the work itself, but are an inextricable part of that text. The work itself is "the text-between of another text" (Barthes 195), Barthes writes. These sources run through the text and yet are "anonymous, untraceable" (Barthes, 195). Since every text only exists as a part of the vast web vast of intertextuality that forms the Text, the author cannot control the meanings that he/she attempts to present in the text: "It is not that the may not `come back' in the Text, in his text, but he then does so as a `guest' (Barthes 195).

When all of the controls which are found in the domain of the work are removed, interpretation is impossible, for there is not one voice which emerges, but rather, a multitude of voices. The voices that are woven through the text as a result of its existence as a part of the intertextual explode in a cacophony of noise. "The Text is plural," Barthes explains, "Which is not simply to say that it has several meanings, but that it accomplishes the very plurality of meaning: an irreducible (and not merely an acceptable) plural" (Barthes, 194).

Perhaps it would be more in accordance with Barthes's methodology to attempt to demonstrate the difference between viewing a piece of literature as a text and as a work. For Barthes emphasizes that his "few propositions" outlined in his essay "do not constitute the articulations of a Theory of the Text" (Barthes 197) for a theory, in effect, is a metalanguage which claims to exist apart from the Text itself which is an impossibility according to Barthes' own articulation of the Text. "The discourse on the Text should itself be nothing other than text, research, textual activity, since the Text is that social space which leaves no language safe" (Barthes 197), Barthes writes.

In this context, it would be interesting to look at Barthes's own essay as a work and as a text. In many ways, Barthes signals his own separation from the "Newtonian" texts that claim to be works (Barthes 192). In other ways, however, the philosophy of the work inevitably permeates his own claims to avoid its dictates. As a work, as an entity which is separable from the discourse surrounding it, Barthes can claim that his essay is in conformation with the view of the text. Paradoxically, when viewed as a text, this separation from those texts which claim to be works is not achieved.

If I were analyzing Barthes' essay from the point of view of the theory of the work, I would accept his authorial claims to place "text" at the critical centre of attention and to throw the concept of "work" onto the periphery. At the beginning and at the end of his essay, Barthes clearly indicates his intention to separate his own critical formulations from critical formulations of the work. Since meta-language employs logic in order to set its own string of signifiers apart from the text, he will avoid doing so. He sets up the "arguments" and the "logic" which others employ in opposition to his "propositions" which are to be "understood more in a grammatical than in a logical sense" (Barthes 192). During the course of the work, he attempts to maintain the distance between the binary opposition of "work" and "text" by defining each term in contrast to its Other. At the end of his essay, he again insists that his "few propositions . . . inevitably" fail to form a meta-language which would dictate how a text should be read (Barthes 197).

In other ways, however, the essay belongs to the textual world consisting of texts conceived as works. For instance, though he claims to avoid formulating a Theory of the Text, he cannot in fact escape the need to understand language through theorizing. Though he signals his aversion to logical constructs by attempting to assert rather than attempting to explain what constitutes a work and what constitutes a text, he cannot do so without operating according to the dictates of a meta-language.

His attempt to distance himself from the `logical' language of literary criticism is only moderately successful. The aphoristic style which he employs to this end has been used before. Wittgenstein's Tractatus Logico-Philosiphicus, for example, is composed entirely of aphorisms. And yet, taken together, they form the logical positivism of the early Wittgenstein. Similarly, Barthes's assertions work together to form a kind of logic which is more or less internally coherent. As I hope my essay has demonstrated, Barthes does have a theory of the text which is permeated by a certain view of what constitutes meaning and language. His enunciations do constitute the articulations of a Theory of the Text.

In addition, according to my reading of "From Work to Text," Barthes compensates for the way in which he attempts to distance his own writing from the conception of the work by connecting his writing with the world of the work. He cannot claim outright that the logic of the text has supplanted the logic of the work, for that would be the equivalent of setting up a tyranny of the meta-text. Instead, he compares the shift from work to text to the shift from the Newtonian to the Einsteinian conception of the universe:

Just as Einsteinian science demands that the relativity of the frames of reference be included in the object studied, so the combined action of Marxism, Freudianism and structuralism demands, in literature, the relativization of the relations of writer, reader and observer (critic) (Barthes 192).
Science establishes a meta-language which is generally accorded a greater degree of objectivity than a non-empirical discourse like literary theory. In using a scientific metaphor, he implies not only that the work is old fashioned, but that the logic behind the work has been superseded.

This comparison between scientific and non-scientific discourse has also been made before. In the preface to the second edition of the Critique of Pure Reason, for instance, Kant makes such a claim. All previous philosophers were blinded by their adherence to the Ptolemaic system of the universe, he claims. By contrast, he has effected the philosophical equivalent of a Copernican revolution:

Failing of satisfactory progress in explaining the movements of the heavenly bodies on the supposition that they all revolved round the spectator, [Copernicus] tried whether he might not have better success if he made the spectator to revolve and the stars to remain at rest. A similar experiment can be tried in metaphysics, as regards the intuition of objects (Kant 22).
Kant is certainly setting up a meta-language here. In fact, his confidence about his ability to set up this meta-language borders on arrogance. "The danger is not that of being refuted, but of not being understood" (Kant 36-37), he writes. When Barthes' essay is viewed in light of Kant's use of this trope, the hint of the authority of science which, as I have argued, adheres to the metaphor itself, is intensified. This authoritative overtone strains against Barthes' own aversion to setting up a meta-language.

According to Barthes' theories, whether Kant's text is seen to flow into Barthes is not a function of whether or not Barthes' writing was directly influenced by Kant. As Barthes writes, the influences of a text are anonymous, untraceable:

The Text . . . [is] woven entirely with citations, references, echoes, cultural languages . . . which cut across it through and through in a vast stereophony. . . The citations which go to make up a text are anonymous, untraceable, and yet already read (Barthes 194-5)
If I can see how Kant's writings can be seen to flow through Barthes' work, that part of Kant's writings which seems to be echoed in Barthes' work becomes a part of what Barthes is saying. There are no boundaries between texts and source texts, but only a vast web of intertextuality.

Meaning in Barthes' essay, then, is irreducible. It is an articulation of the theory of the Text; at the same time, it resists a complete articulation of the theory of the Text. Barthes proclaims his allegiance to the text rather than to the work and yet he can only do so by attempting to cut himself off from the texts which surround his text. This attempt to remove himself from the logic of the work is inevitably unsuccessful.

Any position on how to read texts seems to be inevitably formulated in a meta-textual way. To claim that we can work just within the realm of the text without aspiring to reach beyond it is an illusion. At the same time, I agree with Barthes that meta-narratives are themselves a part of the world of intertextuality. We must work according to the meta-narratives we have created. And yet, these meta-narratives can never really escape from the Text itself.  

Lisa Smith is a student in ENGL 4F70, 1997-98 at Brock University; she is a year 4 Honours English and Liberal Studies combined major.


Barthes, Roland. "From Work to Text" in Modern Literary Theory. ed. Philip Rice and Patricia Waugh. New York: Arnold, 1996.

Kant, Immanuel. Critique of Pure Reason. trans. Norman Kemp Smith. New York, St. Martin's Press, 1963.

Thurley, Geoffrey. Counter-Modernism in Current Critical Theory. New York: St. Martin's Press, 1983.