Here, at last, is an argument about—and, indeed, against—endings: Nineteenth-century novels were written for the professional and bureaucratic classes, for the kinds of people, in other words, who keep reports on the rest of us; good twentieth-century novels were not; and it is endings that most distinguish the one from the other. Readers of Juan Rulfo and Nathalie Sarraute have to make their peace with enduring uncertainty, relinquishing the expectation that a fiction’s final pages will notarize and tidily file away all that’s come before. In the classical text, by contrast—so Barthes—truth arrives “at the end”; “truth is what completes, what closes”; “the profound is what is discovered at the end,” “closing off the infinite repetition of dialogue,” “bringing the interplay of languages to an end.” And here’s where things get tricky for the literary historian trying to make sense of S/Z. Anyone reading Barthes on “the classical text” is bound to feel, at some point or another, that they have just read some of the most penetrating pages ever written on prose fiction. I’m thinking especially of mini-essay #60, in which is explained the dilemma of any literature trying to establish its place in the epistemological society. In any “civilization of enigma, truth and decipherment,” realist novels have no choice but to pursue a certain ruse, producing artificial opacities which they will later offer to illuminate: Who is Pip’s benefactor? Who started the fire at Thornfield Hall? For if they were genuinely committed to truth, there would be no suspense—no tenterhooks, perhaps no story—since they could just authoritatively answer any riddle as soon as it arose: Look, the convict he once helped is sending him money from Australia. Watch out, he keeps his first wife in the attic. Hey, you know that singer? She’s a dude. Most narrators and all novelists are in a position, more than any overly detailed trailer or whispering aisle-mate, to ruin the coming plot-twist. They can always solve the mystery before the detective. The trick of realist fiction in a culture dominated by knowledge is to withhold the truth while seeming to be honest and truth-driven.
So again, “classical texts” do this. “Modern” ones don’t. It’s a brilliant argument whose only drawback is that it can’t possibly be right, at least not as a general proposition about narrative as such. One can hardly doubt that the description serves for some mid-shelf fiction, but do we really think it will hold for every story published before 1922? Where exactly are to we look to find these canons of uniformly closed storytelling? Are we to look to the ancient epic, to “the classical text” in its accustomed meaning? That won’t help. The Iliad ends so abruptly—with the burial of a single Trojan warrior in the middle of an ongoing war—that first-time readers typically feel cheated. Anyone finishing book twenty-four of Homer knows that there is more story left to tell, and woe to the great-books lecturer who has to break it to his students that there will be no wooden horse. For the fall of Troy, a reader will have to wait for the Aeneid, which itself terminates so suddenly, well before the conquest of Italy and founding of Rome, that scholars have generally concluded that Virgil died before finishing the thing.
So do we look instead to French neoclassicism? That might be one of the configurations Barthes had in mind when he grabbed the word “classical.” The problem here is that prose fiction, unlike eighteenth-century history paintings and seventeenth-century neo-Pindarics, is almost never described as “neoclassical,” since prose-writers were much less likely than tragedians or epic poets to invoke whatever few ancient models they had. The most obvious exception to this observation—that would be d’Urfé’s L’Astrée (1607-1627), a pastoral love story, with bonus side adventures, set in the fifth century—clocks in at some 5,400 pages, which means that it can hardly be accused of possessing a precipitate drive to closure. That prose fiction forgoes the claim to classicism can be glimpsed in the form’s accepted names; the European languages furnish two, and they both mark long prose fiction as innovative. This, I’ll grant, might at first be hard to see, since the most common term for “novel” outside of English is a word that also means “a citizen of antiquity’s great Mediterranean empire.” It is a trick of etymology, then, that a roman is precisely not Roman, but rather a story-no-longer-in-Latin, a vernacular tale, recounted in some post-imperial creole or another. “Novels,” meanwhile, are new stories about new things.
So maybe “the classical text” is actually just Barthes’s idiosyncratic gloss on “the realist novel” after all. Maybe this is what makes realist novels distinctive—and distinctively odious—that they go in for strong closure when so many forms do not, that they are the blind alleys on literature’s otherwise open map. But then how could anyone wishing to make that argument choose as his paradigm Balzac, the steamboat captain of the roman-fleuve, of the riverine novel-that-never-ends, the great novel-flux? The shady southern law student you read about in one novel reappears five years later, as a secondary character in another book, except now he’s under-secretary of state. The property speculator at the center of Zola’s Curée, from 1871-72, is still kicking around eighteen novels later. Any novel that appears in a cycle—any Leatherstocking tale, any Barchester chronicle, any episodio nacional—is liable to have its ending amended by some later entry in the series. The biggest shock is to realize that Roland Barthes is mistaken even about his proof text—not just about Balzac broadly, but about “Sarrasine” in particular. Let me read back into the record my summary of Balzac’s novella:
In southern Italy in the 1770s, a French sculptor falls in love with a pretty soprano without knowing that she is actually a cross-dressing and castrated man; that mistake ends up getting the sculptor killed. In Paris, a half century later, the now elderly castrato unnerves the Bourbon Restoration’s smart set, crashing their balls, sidling wordlessly up to their women.
What you’ll want to bear in mind at this point is that S/Z is often read, with some justice, as an early exercise in queer literary criticism. The Anglo-Foucauldian vocabulary of “the queer” may not have been available to Barthes in ’68 and ’69, but he is plainly interested in how “the classical text” deals with characters who fall outside of its governing schemes of sexual classification, and in pursuing that matter he ends up furnishing a later queer criticism with one of its signature arguments. The painter is killed; the queer love story is terminated; it leaves behind only images of itself—a statue of the singer, and then paintings made of the statue: “something dangerous,” writes Barthes, “has been contained, exorcised, pacified.” Queer desire has been foreclosed; Western discourse has opted to liquidate, as unreadable, the emasculated man, &c. But even my rushed précis should be enough to show you that Barthes has got this wrong, since “Sarrasine” is not one story but two, a narrative and a frame narrative—that structure is common enough—and what is pacified in the one lives on in the other. Sarrasine dies; the singer lives.
Nor is this a subtle point, the literature professor’s usual exercise in competitive exegesis. Barthes bizarrely implies that attenuated images of the castrato are all that remain of him: “the sinister story of La Zambinella grows distant, no longer exists save as a vague, moon-struck enigma, mysterious without being offensive.” (208) But Balzac’s non-realist idiom makes the queer soprano’s uncanny survival hard to overlook. The castrato doesn’t, in fact, just intrude upon the parties of the rich; he haunts them, deaf, desiccated, in outdated ruffles and hose, a queer “phantom,” a zombie of the ancien régime—a “vampire,” perhaps, a “walking corpse,” the story says. So yes, queer illegibility is in a sense killing, at least to a normal; the encounter with the queer term can be lethal. But Barthes has botched the bit about closure: The story does not know how to wish the queer away. The exorcism does not take. The Gothic language—realist? classical?—guarantees repetition.
So what’s going on here? How has the greatest work of post-structuralist literary criticism made such a hash of its central claims? Fortunately, Barthes himself has engineered all the tools one needs to work out an answer. Semiotics, after all, instructs us not to treat any writing as a simple report on the world; every account, every argument, every description involves a more or less arbitrary coding of the data, and our task now is simply to extend that insight to semiotics itself, to stop reverencing S/Z as a special vehicle of glossematic wisdom and to see it instead as cutting capriciously into the literary field. I can be more specific than that: Barthes, unlike many of his predecessors, insists that the codes in any text will always be multiple. That is, he forgivingly calls off the search for the master code to which all others reduce. At the same time, though, he argues that any given story will be organized around a central antithesis, a patterning of its information into some specially charged dichotomy (in whose circumference other codes might nonetheless float free). Indeed, it was in S/Z that Barthes furnished post-structuralism with one of its canonical attacks on binary thought, on any Manichean and stalled antithesis that carves the cosmos into ontologically polarized densities, “two plenitudes set ritually face to face like two fully armed warriors … the given opposition, eternal, eternally recurrent.” The “symbolic” universe posited by nearly all writing can understand the breached antithesis or union of opposites only via the conceptually impoverished category of “paradox,” to be strictly distinguished from the dialectical antithesis-in-motion, though Barthes also suggests that for a story to begin, for it to be experienced as necessary, there has to exist some such absurdity, something excessive, something that upsets the stagnant balance of the binary, an unresolved third term that will eventually and in unprecedented fashion mediate between A and B … or will be eliminated as a menace … or will survive its attempted purge and so open-endedly stalk the binary as free radical and unpaired abortion. The fate of third terms simply is the stuff of narrative; it is the only story anyone has to tell.
Examples of stalled antitheses are easy enough to find. Cowboy v. Indian, the grazia and ingegno of Renaissance culture v. the gross rusticity of medieval art, prudish mortal v. sabertoothed, libertine superman. Here are a few more—or rather one more variously named: the classical text v. the modern one; the readable text v. the writerly; closed texts v. open ones. These all come from Barthes, and their proliferation and near-redundancy suggest how central this opposition is to his system. Modern equals writerly equals open, though there is one term, common enough in Barthes’s writing, that we will want to keep out of this series, and that’s “plural,” since the semiotician himself remarks that all texts are plural by necessity, while only some of them are open. “Plural,” in other words, does not mean “open.” Cellblocks are variously populated but no less locked for that. Polyphonic but nonetheless closed—that idea is, in fact, Barthes’s great contribution to the fight over the novel, his signal innovation over Bakhtin. The realist novel might be a compendium of many voices, but if Barthes is right, they are all muttering slogans you’ve already heard. Choruses, we know, sing from off the same score. The Bakhtinians, then, are the ones who cannot distinguish between polyphonies, disregarding all the devices deployed by texts to weight and rank their voices, like sound engineers at a mixing board, amplifying this channel, fading out that, always ready, should a performer prove uncooperative, to cut the mic. Any time a critic uses the word “heteroglossia,” you can be pretty sure he can’t tell the difference between lead vocal and background sha-na-na.
If, therefore, I am reading S/Z as a semiotician, and not as a Bakhtinian, then I should be able to note that Barthes’s writing is plural—varied in its literary-critical methods, tessellated into numbered pseudo-fragments—and still not feel that this plurality counts for much, because he himself has told us that it doesn’t (or that it needn’t). S/Z, despite having a lot to say about literary plurality, is nonetheless anchored in precisely the kind of antithesis that stabilizes and organizes multiplicity: classical and not modern; readable and not writerly. To ply the distinction between texts open and closed is one good way to stopple your own argument. S/Z is less an open text than a closed one in service of the open. Old-fashioned stories, Barthes tells us, can’t accommodate freaks or “monsters”—his word—characters who fall outside the narrative’s coding of the universe into zero and one, those who are “outside any classification,” “outside the code,” “outside the norms.” Terminator 3, Silence of the Lambs, Thor, others I have yet to name: Any movie that both ends and doesn’t is the monster that S/Z doesn’t want to face, open and closed at once, the prodigal term that a bisected semiotics cannot accommodate.
There’s actually a pattern here. More than once, Barthes identifies one of the pitfalls of the “classical text” and then heedlessly falls straight into it, replicating in his strenuously modernist pamphlet the putatively outmoded iniquities of the old bourgeois literature and so undoing the reassurance he wishes to provide that the rhetorical failings of this latter can be identified and historically quarantined. So antithesis was one such pitfall; here’s another. If there is anything Barthes likes less than binary thinking, it is the related problem of common sense or ideology. No surprise there, though his account of how ideology gets produced is fairly distinctive. What needs tracking, he says, is the routine processing of new scholarly claims into anonymous and uncredentialed wisdom. Literature and journalism are the key devices in this transformation—the transmutation of expertise into street epistemology—seizing hold of what had briefly been specialized arguments, repeating them without attribution, and so lending them the insidious authority of the commonplace. A novel, by opening itself to other registers and discourses, doesn’t make itself more plural; it merely absorbs whatever in the dominant intellectual scene was already least thinking, its triteness and decay. Or worse, literature and journalism are the making-platitudinous of once contested positions. A claim, once reproduced in a “classical text,” is no longer an argument that one might disagree with or feel challenged by; it is merely a topos and will pass back into general circulation in this imbecile form. Such, then, is the business of yet another of the coding mechanisms that Barthes theorizes in S/Z, “the cultural codes,” he calls them; you might think of them as the venders of chestnuts and sedative bromides. Among the fundamental elements in any piece of writing are the things that other people have told its author are true, unchallenged beliefs that are present in its paragraphs only as paraphrase and copied language.
The point we won’t want to miss is that the book in which Barthes describes the cultural codes is itself shot through with such codes—that Barthes’s writing, I mean, is an unweeded bed of borrowed wisdom and pilfered idioms. A few of these are especially salient. Indeed, if you follow these you’ll be able to spot for yourself the rhetorical mechanisms by which Barthes codes some endings as open and assigns to their openness an unchanging set of connotations. The ideology of the open text announces itself here:
(1) The goal of literary work (of literature as work) is to make the reader no longer a consumer, but a producer of the text. ★ This sentence says little more than that we should prefer modernist novels to nineteenth-century ones because William Burroughs demands engaged and active readers and Victor Hugo didn’t. This idea is, I think, best understood as encouraging readers to refashion a text or to adapt it to their circumstances, to not be bound by what the author had in mind or indeed by a concern with what a given collection of words “really means.” For our purposes, the important point is that this notion could be phrased in two very different ways: as permission to “play with the text” or as permission to “work with the text.” I’m guessing you would say “play,” which is why you should notice that Barthes does not. Literature is labor; the reader fabricates or manufactures. Barthes has opted for the toilsome version of the active reader, and this in some maximally Marxist way—production!—even though once stripped of these rhetorical choices, his underlying argument is merely professorial and stick-to-it-ive: not that listening to the Mekons will make you want to go out there and start a band, but that reading Robbe-Grillet kind of feels like a job. The Barthes who talks about work is, at any rate, the Barthes we think we know, the Barthes of western Marxism, the Barthes who politicized semiotics, the Barthes who transformed structuralist linguistics into a stick with which to beat the bourgeoisie, the Barthes who thinks a difficult novel enrolls its cultivated readership into the ranks of the proletariat. There is more in this vein: Modernist novels are “production without product,” a utopian factory that need assemble no roadworthy Buicks. Realist novels, by contrast, are literature’s consumer goods, ready-mades arranged behind interchangeably orange spines. If this is the case, then the task of a materialist criticism is to steal from such “products” their finished face, to defetishize old novels the way one would any deceptively settled object, to insist that fiction is something made and that it might yet be re-made by any reader-producer “working back along the threads of meanings,” “observing the reversibility of the structures from which the text is woven.” The trick now for us—as readers not of Balzac but of Barthes—is to hear these sentences as language and solely as language, to register only their memes, and so not to start parsing them as stance or admiring them as argument, because if we are reading S/Z semiotically, then we are prohibited from saying that Marxism is authentically available in its pages as a method. Just tell yourself: Any point that Barthes couches in Marxist terms he could just as well have couched otherwise, without the workerism that makes of the novelist a weaver and of the novel a bolt of frayed organdy. In any particular paragraph, historical materialism announces itself not as an intellectual-political project, but only as a sociolect, the ornaments and epithets of the independent Left, a communism amputated back to its figures of speech, a trophy Marxism (REF. the Marxist code).
(2) How can one code be superior to another without abusively closing off the plurality of codes? Only writing, by assuming the largest possible plural in its own task, can oppose without appeal to force the imperialism of each language. ★ These two sentences are a rumpus of political commonplace. If you’re still thinking that Barthes is a Marxist, then the term that just jumped out at you was probably the last, imperialism, around which the literature professor conscripts modernist fiction into the anti-colonial struggle (REF. the code of the Bandung Conference). That gloss is mostly spurious, though, because the passage’s several other ideologemes form around this scrap of liberationist rhetoric a rather different political constellation. Plural … not closed … without force—at a half century’s remove it is easier than ever to see that this is the patois of Cold War liberalism and thus the verbal housing of a US-led and Wilsonian anti-colonialism rather than a Third-Worldist one (REF. the code of the UN Charter). This would be the moment to go back and renew our attention to the word closed, whose essentially Popperite register now looms into view. It is a tag Barthes uses often—realist novels are “closed,” bad readings are “closed,” antitheses are “closed,” actions assigned overfamiliar verbs are “closed,” dictionaries are “closed”—though the opposite of “closed,” in S/Z, is the Schlegelian “infinite” and not the expected “open,” such that one might mnemonically rename the book The Infinite Text and Its Enemies. The word “plural” occurs more frequently still, many dozens of times in the space of some 200 pages, though it, too, acquires an unanticipated antonym: neither “monolithic” nor “homogeneous” nor “uniform,” but “incompletely plural” (REF. the code of the separation of powers and the mixed constitution, of religious tolerationism, of institutionalist social theory, and of Americanism in Europe). Semiotically, then, the distinctive quality of Barthes’s writing lies neither in his extensive recitations from the anti-totalitarian liturgy, nor in his much scarcer deployments of red catchwords, but in his free-form recombination of the two, by dint of which the vocabulary of historical materialism is absorbed into and made to do the work of the open society. Marxist intellectual procedures—the critique of fetishism, the redescription of art and culture as “production”—get pressed into the service of liberal political arguments, the pursuit of the “integrally plural,” “the largest possible plural,” the diversissimo. That’s S/Z.
(3) The difference between feudal society and bourgeois society, index and sign, is this: the index has an origin, the sign does not: to shift from index to sign is to abolish the last (or first) limit, the origin, the basis, the prop, to enter into the limitless process of equivalences, representations that nothing will ever stop, orient, fix, sanction. ★ Another argument about endings: Barthes asks us to dislike old-fashioned narrative—novels, but in this case, one suspects, plays, too—for preferring some denouements to others, for needing to end some particular way, and so for tamping down on our sense of the future as a field unenclosed, changing possibility into plot and routing the morrow through story’s variously convention-bound bottle-necks and border crossings. A novel can reach its foreordained conclusion without even having to deliberate over alternatives, just by relying on a given genre’s sense that this is how such characters are meant to wind up. Sense and Sensibility, after all, might be a more interesting book if one could at least conceive of Elinor Dashwood marrying Willoughby (or Lucy Steele) and not always awkward and dependable Edward Ferrars. Or: It doesn’t matter how many times you re-read Waverley, the Highlanders aren’t going to win this time. Barthes is never more anti-Aristotelian than when he objects that plot structures create, through repetition, an artificial sense of fate or destiny or historical necessity. The word he attaches to this argument is “reversible”: The “classical text” generates a sense of “irreversible order”; it installs an “incompletely reversible system.” But once that word is in play, the argument becomes easy to discard. That we should reject all claims to irreversibility is prima facie unconvincing. Are we really sure that everything is reversible? Every broken jug glueable? Every monarchy restorable? Are you confident that vanished glaciers can be re-frigerated? Have you ever known a Yemeni wedding party to be reanimated or retroactively un-mis-assassinated? Do you have a proposal for getting white people to leave North America? No action is irreversible is the sentiment of a man who has never oversalted his soup. Nor is it really clear that “the classical text” is as rigidly end-directed as all that. I can write that it is tough to imagine a Hamlet in which a briefly resolute prince successfully avenges his father, but as soon as that sentence is typed, one goes ahead and imagines it anyway: Claudius cut down at prayer, Ophelia no longer nunneried, the prince raised to the throne and dithering still.
More to the point: Does Roland Barthes actually believe that all sequences are reversible? Or rather, does he write as though he did? Go back and look again at the sentence quoted here:
The difference between feudal society and bourgeois society, index and sign, is this: the index has an origin, the sign does not: to shift from index to sign is to abolish the last (or first) limit, the origin, the basis, the prop, to enter into the limitless process of equivalences, representations that nothing will ever stop, orient, fix, sanction.
From feudalism, indices, and a world of ordered sequence; to capitalism, signs, and a world without such sequence. We’ll want to note right away that this sentence can’t possibly be saying what it seems to be saying—that Capetian France and the Holy Roman Empire lacked language or that in such societies words really did function like footprints or photographs, as the physical evidence left behind by their referents. Barthes has to be saying that pre-modern people typically thought of words as securely attached to their objects and that moderns to their credit know better, that they know language to be free-floating. So not: from index to sign, but: from sign-mistaken-for-index to sign-recognized-for-what-it-is. But now we have to bracket that clarification, setting aside the substance of Barthes’s historical claim in order to register the simpler, semiotic point that this sentence has borrowed the language of historical periodization (REF. the code of the philosophy of history). In the process of expounding the doctrine of reversibility, Barthes proves perfectly capable of proposing an irreversible series of his own. This isn’t just a gotcha point—the problem is more interesting than that, an incoherence and not just an inconsistency. In S/Z, the society-without-sequence appears as an item within a sequence, and the problem is that if Barthes can identify the shift from one to another, then the shift itself doesn’t exist. Or the same point in reverse: If it’s true that we live in a society-in-which-one-no-longer-thinks-in-terms-of-sequence, then Barthes shouldn’t be able to name the historical stages by which such a society came to be. All you have to do to make this paradox go away is concede that Barthes has written a literary manifesto and not the statement of fact that S/Z sometimes pretends to be. Modernity, Barthes writes at one point, is the time of “confusion,” of “the unbridled (pandemic) circulation of signs, of sexes, of fortunes.” Capitalism, queerness, and modernist literature—we don’t really need to know why he thinks these three belong together; it is enough to know that he codes them all as “modern” (or “bourgeois”) and not as “classical” (or “feudal”). Modernity is the period of generalized circulation, of wealth-, desire-, and language-without-ground. Barthes’s point was not that twentieth-century novelists couldn’t (or uniformly didn’t) attempt sequence or strong endings, but simply that such novelists were out of keeping with a consumer society that in 1970 was the true empire of signs. Realism is to be rejected not for being too bourgeois, but for being insufficiently so. S/Z is a petition to bring European fiction more fully into the market, to subject it irrevocably to the “limitless process of equivalences.” “Limitless process”: The first word in that phrase is going to require special comment, of course. Commercial and semiotic societies are “limitless,” governed, he says on the same page, by an “endless process.” Just follow the bleeping code: Capitalism and language are “limitless” and “endless.” The modern text replicates this condition in literature, “denying [the] final word (denying the end as a word).” Capitalism, language, and modernist literature thus make up a set. All three belong to a semiological modernity that, Barthes tell us, “nothing will ever stop,” an endlessness, then, that far from restoring to us the open future ensures only that it will never come.
So Barthes wants fiction writers to forego the temptations of the tidy ending, to commit to the infinite text as project and program. The only point that still needs making is that he needn’t have bothered, that this recommendation is, in fact, entirely redundant. His point about capitalism is accurate enough, after a fashion. I say again that Hollywood blockbusters go in for unresolved and ambiguous endings, and you reply, perhaps, that that you have a ready explanation for this. Open endings set up sequels, you say; they are just so many dropped cans of Barbasol. This is no doubt true, but if we pause now to mull this observation, the puzzle will merely change shape before us. So let’s say that the open ending is how franchises are built, hence the mark of commerce impinging on narrative form, and that in some unusually forthright way—the necessary re-organization that fiction undergoes if it is to be industrialized: mass-produced and sold in installments. As soon as we say that, we have to give up once and for all on the common belief that popular narrative requires strong closure: wedding bells sounding, bombs defused, families reunited, victory laps haughtily jogged. For popular narrative has for several generations now been produced in two antithetical modes: 1) the old, completed forms of romance (the good guys win!) and comedy (everybody wins!) coexisting with 2) oddly ongoing and undecided versions of same: novels sold chapterwise, novels serialized in magazines, novels serialized in newspapers, comic strips, movie serials, television. English fairy tales end with happily ever after, but German tales go out on and if they have not died, then they are still alive. And if pulp comes with endings both closed and unclosed, then we can no longer treat the open ending as one of modernism’s more sophisticated achievements—the art novel’s defining challenge to its downmarket competitors, the carrier of its abstrusely anti-teleological and anti-totalizing doctrines. Commercial imperatives and experimental writing attack traditional narrative in the same way and to similar effect, and the “modern text” mutates into a universal term without meaningful rivals.