4. The Reality Effect in the Writing of History, According to Roland Barthes

Roland Barthes's (1915-1980) field of investigation is theory of literature rather than philosophy or philosophy and history. But both in France and in the United States the distinction between literary theory and philosophy (of language) is becoming blurred. This is justifiable inasmuch as literary language is the most complex and interesting kind of language we know and thus preeminently deserves the attention of the linguistic philosopher. A philosophy of language that confines itself to the most elementary forms of language not only succumbs to the dogma that complex forms can be deduced from elementary ones, but also obstructs its view of a number of problems which it wishes to study. In this connection Danto has pointed out that "literature sets up obstacles to the passage of semantical theories [especially with reference to fictive entities] which would go a good deal more easily if literature did not exist."^[33] Barthes is often grouped with the structuralists,^[34] but he does not seem particularly interested in general statements about the structure of language. Both the essays discussed in this section are even described by Barthes as antistructuralist.^[35] Barthes's lack of interest in sweeping statements requires an important qualification, however. For it is his persistent concern in showing that the text is the vehicle of a morality, of an ideology, or of a view of reality unsuspected by writer and reader alike—in short, of what Barthes likes to call, rather dramatically, "mythology." Indeed, the rhetorical aim of the text is to present that mythology as a quasi-natural phenomenon. But the text is the creator of this quasi-natural reality rather than its ideological reflection (here Barthes differs from Marxism).

The central idea in both essays is that the reality of the past must be linked to a so-called reality effect, an effet de réel which is created by irrele-

[32] R. Barthes, "Le discours de I'histoire," in Barthes, Le bruissement de la langue , Paris 1984; R. Barthes, "L'effet de réel," in Barthes, Le bruissement de la langue , Paris, 1984. The second, and also more important, of these two essays is also available in English: R. Barthes, "The Reality Effect," in T. Todorov, ed., French Literary Theory Today, Cambridge, 1982, 11-18. When discussing this essay, I shall refer to the English translation.

[33] A. C. Danto, The Philosophical Disenfranchisement of Art , New York, 1986, 142.

[34] See, for instance, J. Sturrock, "Roland Barthes," in Barthes, Structuralism and Since , Oxford, 1979.

[35] Barthes, "Reality," 11.

vant details mentioned in the historical text.^[36] The reality of the past is an effect caused by a tension in and between historical texts. Barthes shows how in one of his novels Flaubert describes the room of his main character and mentions a pyramid of boxes and cases standing under a barometer. These kinds of details are called notations by Barthes; he contrasts them with the main outline of the story, which he labels predictive, probably because on this level we can make certain predictions about the development of the story. Using Michelet's reference to certain details in the execution of Charlotte Corday, Barthes points out that a similar tension between prediction and notation can be demonstrated in the writing of history.^[37] He then goes on to develop a surprising theory about these notations. First of all, and contrary to what we would expect, they are said to embody the highest degree of perfection that language can attain. This is already indicated by the fact that animals do have something reminiscent of predictive language—for instance, bees have "a predictive system of dances, used in foodgathering"^[38] —but the animal world has no equivalent of the linguistic noise or static to which the word notation refers. Only human beings can chat. More importantly, the history of rhetoric and literature confirms this idea. It was not until the Alexandrian rhetoric of the second century A.D.—about a thousand years after Homer's epics—that the literary tradition of ekphrasis and hypotyposis arose. Ekphrasis and hypotyposis were rhetorical compositions describing ways of life, periods, and places (read: historical themes) as elegantly as possible and purely for the sake of description itself. The description, that is to say, did not form a link in some or other comprehensive, predictive argument. We are dealing here with an early form of notation and we see how it deliberately breaks away from predictive language for the first time.^[39]

Barthes is concerned with what this means for the writing of history. We associate the reality of the past, he says, with notation rather than with prediction. The predictive is for us a meaning conceived or created by the historian; in notation, ekphrasis, or hypotyposis, on the contrary, the past reveals itself as it really was. "Unvarnished 'representation' of 'reality,' a naked account of 'what is' (or was), thus looks like a resistance to meaning, a resistance which confirms the great mythical opposition between

[36] One would not wish to exaggerate the originality of Barthes's insights. A comparable point of view is found as early as the twenties in an article by R. Jakobson, "On Realism in Art," in L. Matejka and K. Pomorska, eds., Readings in Russian Poetics: Formalist and Structuralist Views , Cambridge (Mass.), 1971, 38-46. Essential elements are also found in nineteenth-century novelists like Flaubert, Baudelaire, or Vogfüé and even in eighteenth-century critics and literary theorists. See I. Watt, The Rise of the Novel , Reading (Eng.), 1957; especially chapter 1.

[37] Barthes, "Reality," 11.

[38] Barthes, "Reality," 12.

[39] C. Ginzburg, "Ekphrasis and quotation," Tijdschrift voor filosofie 50 (1988): 3-20.

the true-to-life (the living) and the intelligible."^[40] Unlike what is expressed by notation, meaning is constructed and therefore cannot achieve the effect of reality. But we need to consider here that notation is only capable of doing so by its contrast with prediction and meaning.^[41] The origin of notation lies there, after all, and not in an extratextual relationship between a description in the text and a state of affairs in the past. This being so, one wonders whether one should speak of a reality illusion rather than a reality effect.

That brings us to the heart of the matter. Is the reality suggested by the opposition of notation and prediction reality or mere illusion? In the Fregean theory of signs with its strict distinction between language and reality it must be called an illusion. Now a peculiarity of the Saussurian theory of signs generally adhered to by French philosophers (and especially as interpreted by Barthes) is that it does not differentiate between language and reality as far as the reference of the sign is concerned. This puts a different complexion on the matter. Thus Barthes can write:

In the first instance the referent is detached from discourse, it is exterior to it; as its foundation the referent is supposed to determine discourse: this is the time of res gestae, and discourse presents itself as being simply a historia return gestarum [that is Frege]. But secondly it is the signifier itself that is suppressed, confounded with the referent; the referent now comes into a direct contact with the signifier [and that is Saussure].^[42] (my translation)

The obvious thing to do in this kind of situation would be to compare the Fregean and Saussurian theories of signs and judge Barthes's suggestions on that basis. But that is precisely the route I decided to avoid in this argument: instead of measuring the writing of history from a predetermined philosophical standpoint, my aim is to arrive at a philosophical standpoint

[40] Barthes, "Reality," 14. See also Barthes, "Discours," 165: "En d'autres termes, dans l'histoire 'objective,' le 'réel' n'est jamais qu'un signifié informulé, abrité derrière la toutepuissance apparente du référent. Gette situation définit ce que l'on pourrait appeler l'effet de réel. " (In other words, in so-called objective history, the "real" is nothing but an unformulated signified, hidden behind the powerful but apparent presence of the referent. This situation defines what one might call the reality effect .) Also important here is Barthes's later distinction between studium and punctum. See R. Barthes, La chambre claire: Note sur la photographie , Paris, 1981, 49.

[41] In a similar context Ehrenzweig speaks of the "uncompromising democracy" which governs the relationship of the important and the unimportant. See A. Ehrenzweig, The Hidden Order of Art , London, 1973, 43.

[42] "Dans un premier temps. . . le référent est détaché du discours, il lui devient extérieur, fondateur il est censé le [i.e., the discoursel règler: c'est le temps des res gestae, et le discours se donne simplement pour historia return gestarum , mais dans un second temps, c'est le signifié luimême qui est repoussé, confondu dans le référent, le référent entre en rapport direct avec le signifiant," (Barthes, "Discours," 164.

(here concerning the conflict between Frege and Saussure) based on evidence from the writing of history. Existing views on sign, reality, and reference are the object of this enquiry, not its point of departure.^[43]

As often with novel theories, Barthes's argument perhaps raises more questions than it answers. I do not have the pretension of being able to draw up an exhaustive list of these problems—let alone solve them. I shall therefore confine myself to the three questions which seem to me most important. First, is the connection between the writing of history and the (nineteenth-century) realistic novel suggested by Barthes valid and useful? For in his view, both achieve the reality effect. This leads to a second question: Can the historical text be credited with the ability to bring about a reality effect in the way indicated by Barthes? Third—and this concerns Barthes's most spectacular claim—does Barthes establish a credible link between the reality effect and the opposition of notation and meaning?