Richard Woodfield, 'The Reality Effect: An alternative account' more

THE REALITY EFFECT: AN ALTERNATIVE ACCOUNT Richard Woodfield For the memory of Professor Wiadyslaw Tatarkiewicz Roland Barthes' essay L'effet du réel offends my sense of logic and my sense of history. What follows is an alternative account prefaced by a short summary of his views for the benefit of the reader unfamiliar with them. THE ARGUMENT1 Citing fragments from Flaubert's Un Coeur Simple and Michelet's Histoire de France, Barthes argued that each contains details whose presence is irrelevant to the larger concerns of the narrative. These may be "salvaged as part of the structure" by being regarded as "fillers, padding (catalyses)" as he had done in his essay Introduction à l'analyse structurale des récits.2 But even so, an exhaustive analysis leaves certain details which are utterly meaningless. Although the piano and pile of boxes in Un Coeur Simple function as signs of bourgeois status and the disorder of the Aubain household, the barometer has no similar status it is simply there and is not properly notable. Michelet's detail of the soft knock at the door is similarly meaningless. Description simply has an additive effect and does not contribute to narrative flow, which is essentially predictive. Turning to history he argued that description has long had an aesthetic function, at its earliest associated with epideictic rhetoric and subsequently, in the second century AD, with ekphrasis which, in the Middle Ages, was practiced without regard for truth or verisimilitude. In Flaubert the aesthetic purpose of description is still very strong and although the description of Rouen in Madame Bovary might have been objectively correct, its real purpose was to create an aesthetic image. The description of Rouen "is constructed so as to associate Rouen with a painting: it is a painted scene taken on by language". The aesthetic function confers a meaning on "the set piece" and stops a "downward spiral into detail". The "irreducible residues of functional analysis", being without a meaning, constitute a representation of naked existence which stands ideologically independent of meaning. Reality resists meaning: "What does it matter that a detail has no function in the account as long as it denotes `what took place'? `Concrete reality' becomes a sufficient justification for what is said". Similarly, to draw Michelet back into the argument, "the `real' and history have gone together (seemingly truthful), but this helped to oppose it to the 'vraisemblable', that is, to the very nature of narrative (imitation or `poetry')". The detail of the soft knock at the door is the guarantee of the truth of Michelet's piece of history, for being opaque to meaning it must have happened. Finally, "the `concrete detail' is constituted by the direct collusion of a referent and a signifier; the signified is expelled from the sign, and along with it, of course, there is eliminated the possibility of developing a form of the signified, that is, the narrative structure itself. ...This is what may be called the referential illusion". All references are to the translation "The reality effect" in. T. Todorov (ed.), French Literary Theory Today (1982), pp 11-17. 2 Published in translation as "Introduction to the Structural Analysis of Narratives" in S. Heath (trans.): Roland Barthes, Image-Music-Text (1977), pp 79-125. 1 THE REFERENTIAL ILLUSION It has been pointed out that the referential illusion requires that the reader of a text "believes he is faced not with words but with things, as if the referent were actually there, in the statement. What in fact he is faced with is, of course, not things, but a rhetorical category (la catégorie du réel). But Barthes' hypothetical reader does not' know this; he is bewitched by the stratagems of writing into confusing category with thing...".3 Prendergast goes on to claim that Barthes could not have possibly believed this thus it could only have been a rhetorical stratagem on his part. But there is more to the matter than that. To understand Barthes' view of the referential illusion one has to come to terms with the real possibility in his system of the expulsion of the signified from the sign. Section 11.2.1 of the Elements of Semiology, which should have provided guidance on the nature of the signified, is marked by extreme vagueness. He did not like Saussure's notion that "the signified of the word ox is not the animal ox, but its mental image", not because the notion is logically indefensible but because "these notions still bear the stamp of psychologism". He preferred the Stoic notion of the lekton (the utterable) which being "neither act of consciousness, nor a real thing, it can be defined only in the signifying process, in a quasi-tautological way: it is this `something' which is meant by the person who uses the sign".4 Being circular, not quasi-tautological, this is neither clear nor helpful. If, as must be presumed, this `something' is separate from the word `ox', questions arise as to the nature of the connection between the word and the `something'. Barthes' account of the relationship between the signified and the signifier is open to precisely the same kinds of objections that Aristotle made against Plato's theory of forms. Barthes tried to find a way out of the dilemma by proposing a functional equivalence between the meaning of words and the meaning of things. This equivalence was of fundamental importance to his account of the semiotic system. It was this association of the meaning of words with the meaning of things which led directly into the possibility of the expulsion of the signified. Thus in attempting to explain the nature of the signified he used a model drawn from the language of clothes: "One can say, for instance, that a certain sweater means long autumn walks in the park". One can go for a walk without wearing the sweater or wear the sweater without going for a walk. The relationship between signifier and signified is, therefore, contingent. But in contrast one cannot talk of a bachelor without talking of an unmarried man and vice versa; the connection is logical and the analogy between words and things is false. Furthermore, the `language' of clothes does not have a grammar or syntax, it has what can loosely be described as a structure; there is nothing in the semiotic system that adds up to a sentence. The notion that the signified may be expelled from the signifier is based on a faulty analogy. On Barthes' account, when we read through a text and stumble on an inessential `ox', the meaning of `ox' is supposed to have dropped out of sight and we are to imagine ourselves in the presence of an ox itself: "Flaubert's barometer, Michelet's little door, say, in the last analysis, only this: we are the real". If we are to make any sense out of the reality effect we need to drop Barthes' false epistemological claims and introduce a notion which would have struck at the heart of his semiotic project: the distinction between the meaning of a word and the meaning of a thing. Thus as we read a text and come across the inessential detail of an ox, the sentence has sense because we 3 4 C. Prendergast, The Order of Mimesis (1986), p 71. R. Barthes, Elements of Semiology (1964), p 43. understand what the word `ox' means, but the presence of the ox in the story has no meaning as it is an inessential entity. This sounds rather less splendid than Barthes' way of putting the matter, but doesn't involve any weird claims. L'EFFET DU RÉEL As a literary theorist Barthes would have claimed that the inessential detail is in the text, not for what it might contribute to the narrative, but for its creation of a sense of literary truth, the vraisemblable. The reason for this, he maintained, is that the classical world left us with the legacy that reality is opaque to meaning; one doesn't have to believe this, as one doesn't have to believe Barthes' other obiter dicta. To prove it wrong would engage one in a full-scale history of Western European thought, and that is not necessary. With less rhetoric and more caution one might prefer to say that a novelist induces a sense of believability into his text by incorporating insignificant details. The nature of the detail is, however, restricted by the beliefs of the reader - which proves, again, that they cannot be meaningless. Daily life tells us that there is an element of truth in the notion of the reality effect. Proving that one has been to a place, or witnessed a scene, depends upon one's ability to describe its unanticipated details. One knows that Nottingham is a city in England, but proof that one has been there depends upon one's ability to describe incidental details of its appearance: "There is a postbox standing diagonally across the road from the old station clock-tower in front of the Victoria Centre". Such a sentence is marvellously useless and is the kind of thing that travellers' tales and newspaper reports are made of. The detail is the mark of "the having-been-there of things".5 It follows from this that the reality effect has no intrinsic aesthetic value as it is present in ordinary mundane forms of speech. A different account of it would have to be put forward to explain its use in the realist text. Detail does not, of itself, make for realism in the novel; otherwise the works of Alain RobbeGrillet could be regarded as the height of realism, which they're not. Carrier's suggestion that the sentence "The long-haired black cat was on the nineteenth century kelim" is more realist than "The cat was on the mat" because it creates an excess of information6 is fundamentally misguided because realism is a quality of literary text, not of sentences. There is another kind of reality effect, which is important for oral literature, and that is created by getting the form of words right, itself a product of repetition. As Ong pointed out, children expect their parents to get their stories right. Every child knows that the nasty wolf in the Three Little Pigs "huffed and he puffed, and he puffed and he huffed, and he huffed and he puffed'.7 The realism of the account is guaranteed by its truth, which is the outcome of its being the same every time. In the world of adults it is the mark of the liar that he should be caught out in inaccuracies of repetition. If a person had been to Nottingham and had remarked on the location of the post box, he wouldn't make mistakes about redescribing its location. How many times have we heard a person say: "Of course it was there! How many times have I told you?" Without wishing to sound logically formal about the matter, one could suggest that the first kind of reality effect depended on truth by correspondence and the second kind of reality effect "The reality effect", p 15. D. Carrier, "Theoretical Perspectives on the Arts, Sciences and Technology", Leonardo, 17 (1984), p. 290. 7 W. J. Ong, Orality and Literacy (1982), p 67. 6 5 depended on truth by coherence. In the first case, believability depends on the detail which corresponds to the truth about the real world and in the second case, believability depends on the detail which is expected out of the process of story telling. In the case of fiction, which is parasitic upon true stories, the strategies for creating the reality effect draw upon everyday usage.8 Neither reality effect is intrinsically aesthetic; both are simply part of the mechanisms of linguistic communication and analogous to the effects of stimulants on visual perception. One can accept the notion of l'effet du réel as a product of certain uses of description without being tied to the notion that description "is not justified by any purpose of action or communication".9 Indeed one has to. While narration might be the dominant element in story-telling it does not have that dimension in ordinary discourse. Communication entails identification of actual persons, things, events and places; without having identifiable characteristics, individuals would be indistinguishable. While in story-telling the incidental detail may be the mark of `having-been-there', in ordinary discourse it is the proof of `having-been-there'. Furthermore, elaborate description is necessary in explaining the nature of things, places, events, etc. Although story-telling may focus on narration, description is essential to bring it into line with ordinary speech, giving it the character of the vraisemblable. ORIGINS OF DESCRIPTION It is simply-a rhetorical trick to associate description with ekphrasis when it pervades a rich spectrum of pre-classical texts; Barthes made it seem as if it was invented in the second century AD. What is interesting, and what Barthes missed entirely, is the change in function of descriptive writing in the pre-classical to classical past, and the variety of descriptive functions left from classical antiquity. Ancient stories like the Epic of Gilgamesh, Dumuzi's Dream and the Iliad gave accounts of the doings of men and the gods; the reality effect of the first two stories depended upon recognition through repetition and the Iliad depended upon recognition through the real. Our first example is drawn from Dumuzi's Dream: "With head-crushing maces bound to their loins, / The two sons from Ur moved against the king. / Having shining clothes on the quai, / The two sons from Nippur moved against the king".10 In the same way that during the Middle Ages saints and personifications were given attributes to mark them off from each other, so the two pairs of individuals were marked off from each other in terms of their attributes. If the two sons from Ur were different from the two sons from Nippur, what were their differentia? The two sons from Nippur were the two sons having "shining clothes upon the quai", as opposed to the two sons with "head-crushing maces bound to their loins"; those were the facts that had to be remembered and that is the story. It is difficult to argue that the descriptions were not an integral part of the narrative text as "head- -crushing maces" show the intent of the sons from Ur to kill Dumuzi and "having shining clothes on the quai" identifies the sons from the cultically clean city of Nippur.11 It might be thought that oral literature is the most likely to carry redundant information as there is a basic requirement to generate words for a metrical fit; but the argument that the words which were The philosophical literature on fiction is considerable. Two useful articles in this context are M. Eaton, "The Truth Value of Literary Statements", British Journal of Aesthetics, 12 (1972) and D. Novitz, "Pictures, Fiction and Resemblance", British Journal of Aesthetics, 22 (1982). 9 "The reality effect", p 12. 10 B. Alster, Dumuzi's Dream (1972), p 67. 11 Alster's observation, op. cit., p 43. 8 generated were in excess of the needs of the narrative has to be proved. It would be rather more to the point to maintain that words were retained for a metrical fit despite their incongruity and that subsequently, with the emergence of a literature which placed greater emphasis on the connection between text and the world, these incongruities were ironed out. The Epic of Gilgamesh has rather more rationality going for it than Dumuzi's Dream, but it was still based on recognition through repetition: "Ninsun went into her room, she put on a dress becoming to her body, she put on jewels to make her breast beautiful, she placed a tiara on her head and her skirts swept the ground".12 These vivid images are strongly underpinned by metre and they are formulaic, but they are still informative and knit into the ethos of the narrative. Metrical requirements remained in Homer's Iliad, of course. Odysseus was variously described as dios (brilliant), polymetis (resourceful) or polytlas dios (long-suffering brilliant) depending on the requirements of the rhythm of the text.13 But there is another dimension present in the Iliad making it very different from the previous two works, and that is the invitation to test the text out against the world, to look for the `how' and `why' of things rather than the `what'. Formulae become rather flat when compared to the following passage: "As when rivers in winter spate running down from the mountains throw together at the meeting of streams the weight of their water out of the great springs behind in the hollow stream-bed, and far away in the mountains the shepherd hears their thunder; such, from the coming together of men, was the shock and the shouting".14 The truth of this description lies in an appeal to the appropriateness of the image to sensory response. The following lines from Dumuzi's Dream would stretch the powers of the imagination to unacceptable limits if tested by the same criteria: "Set up a lament, set up a lament, O Plain, set up a lament, O Plain, set up a lament, O Swamp, set up a cry! O Crabs, set up a lament in the river, O Frogs, set up a cry in the river!"15 But that is not the point of the text. As far as we know, the art of writing fictional prose, to be enjoyed as such, emerged in the first century BC with Chariton's Erotic Tales about Chareas and Callirhoe.16 It has been argued that their source was in Middle Eastern folk tales.17 Be that as it may, it is clear that the ancient novel was produced to satisfy the needs of a leisured and literate public, not unlike the public of the eighteenth-century novel. Possibly the best known of the genre, Longus's Daphnis and Chloe, is saturated with descriptive passages. The writer tells us at the beginning that the story was inspired by a painting he had seen: "It was a painted picture, reporting a history of love. The grove indeed was very pleasant, thick set with trees and starred with flowers everywhere, and watered all from one fountain with divers meanders and rills. But that picture, as having in it not only an excellent and wonderful piece of art but also a tale of ancient love, was far more amiable".18 N. K. Sandars (trans.): The Epic of Gilgamesh (1986), p 74. R. Lattimore (trans.): The Iliad of Homer (1961), p 39. 14 Ibid, p 125. 15 Op. cit., p 53. 16 See T. Hägg, The Novel in Antiquity (1983). I am indebted to Professor Gombrich for his suggestion that I should read the classical novel. 17 G. Anderson, Ancient Fiction (1984). 18 G. Thornley (trans.): Daphnis and Chloe by Longus (1924), p 7. 13 12 Thus the function of the picture within the text was to provide a key to its opening. It is arguable that the descriptive passages were aimed at an audience that enjoyed savouring images and perhaps that explains the existence of Philostratus's Imagines, produced at about the same time as Daphnis and Chloe, which was a set of descriptions of real paintings.19 Barthes' notion of ekphrasis conflated three separate descriptive techniques which didn't coalesce until the fourth century AD: the description of paintings, which "always serves as a simile (usually a proleptic one) of the theme of story or discourse", the oratorical notion of ekphrasis, which is "a narrative passage characterised by visual vividness and useful for speech composition and historiography", and the literary notion which "defines ekphrasis as a digressional description of secluded places in nature".20 His blanket view concealed a number of different strategies. On the basis of a superficial reading of Curtius' European Literature and the Latin Middle Ages, we are asked to believe that throughout the Middle Ages: "description was not constrained by any desire, for realism; truth or even verisimilitude, was of little moment - nobody was bothered when lions or olive trees were placed in a northern landscape. The only constraints that mattered were descriptive ones; plausibility was not referential, but overtly discursive; it was the rules of the discourse genre which laid down the law".21 But the real situation was not as clearcut as that. The reality effect by repetition did have a role to play, but equally there were ekphrases which are demonstrably of real buildings and images.22 Perhaps more interestingly, there were attempts in the twelfth century Byzantine novel to strive for contemporary appeal.23 A DIGRESSION The double reality effect has important implications for the literary sources of art history. It explains how some kind of plausibility may be attached to St. Nilus of Sinai's Letter to Heliodorus Silentarius: "Both of them had their prayers heard, the father in his cave on the mountain, the son in captivity, and, behold, our Plato suddenly appeared on horseback before the young man who was then awake, bringing along another horse without a rider. The young man recognised the Saint because he had often seen his portrait on images. Straightaway (Plato) ordered him to arise from among all the other (captives), and to mount the horse; his fetters fell apart like a spider's web, and he alone was delivered by virtue of his prayer... "24 It has often been suggested that Byzantine descriptions of paintings as being life-like, and the kind of statement made in this fifth-century text, are merely literary gestures the kind of things said as a matter of style and that the writer could not possibly have meant what he was saying to be taken as true. Grigg's view25 that the Byzantines thought their images to be realistic because they were the best they had got, not having much classical art, seems to me to miss the point that the function of images in Byzantine society was different from the function of images in Athenian society. The See K. Lehmann-Hartleben, "The Imagines of the Elder Philostratus", Art Bulletin, 23 (1941). E. C. Harlan, The Description of Paintings as a Literary Device and Its Application in Achilles Tatius (Ph.D. diss. Columbia University, 1965), p 2. 21 "The reality effect", p 13. 22 See H. Maguire, "Truth and Convention in Byzantine Descriptions of Works of Art", Dumbarton Oaks Papers, 28 (1974). 23 Hägg, op. cit., pp 74-75. 24 C. Mango, The Art of the Byzantine Empire 312-1453 (1972), p 40. My emphasis. 25 R. Grigg, "Relativism and Pictorial Realism", Journal of Aesthetics and Art Criticism 42 (1984). 20 19 repetitive and standardised paintings of Saint Plato did the job of differentiating Plato from other possible persons and insofar as it did that, then it was a good painting of Saint Plato. THE EMERGENCE OF THE MODERN NOVEL It may be argued, though this is not the place to do it, that vernacular translations of ancient novels had a formative effect on the creation of the modern novel. Historians of national literatures have tended to look for their roots in earlier vernacular literature.26 The more important point, for this paper, is that the construction of the nineteenth-century novel emerged from considerations of logic and sense in its eighteenth-century predecessors. The eighteenth-century novelist tried to deceive his reader into believing that the stories told were actually true: "Ceci n'est pas un conte".27 It is in the eighteenth-century novel that we are most likely to find a literary use of the Barthesian reality effect. Marivaux's La Vie de Marianne is seminal in this respect. As Stewart28 has pointed out: "Rather than make the mystery of her birth absolute by simply having everyone die and the robbers disappear, Marivaux went to the trouble of giving us even details that add no real information... Inconclusive information of this kind has a ring of authenticity about it: since it is useless, it is being told only because it is true". But the Memoire-Novel had a number of technical limitations. If a novel were true, how could the author have access to other minds or be party to events when he was not there? How could he recall long dialogues in great detail? If a novel were composed of letters or was a journal it would be bound by a temporal order and how could it describe dramatic events at the moment they were occurring? Plausibility created many problems and the realist novel of the nineteenth century emerged out of its solution, not out of the use of the reality effect. MADAME AUBAIN'S BAROMETER Emma Bovary knew her furniture from reading about it in the Mystères de Paris: "She pored over the description of furniture in the works of Eugène Sue".29 Flaubert knew about furniture by observing its use.30 The reference to the barometer in the description of Madame Aubain's house was not gratuitous, it did not create l'effet du réel by the withdrawal of the signified but by contributing to the narrative. Take the description "Un vieux piano supportait, sous un baromètre, un tas pyrimidal de boîtes et de cartons". Its richness in meaning may be appreciated if, instead of registering "the detail of the piano as a sign of the bourgeois status of its owner, and that of the boxes as a sign of disorder and something like a reverse or fall in status", one adopts a more ruminative approach. It may seem scientific to attach notions to words in a monolithic way, but it hardly squares with the process of reading.31 It seems right that Flaubert should have expected his readers to have adopted the same approach to his novels as Emma Bovary did to hers and to have entered into a world of day-dreams set in train by his descriptions. The mind engages "un vieux piano" and wanders, it engages "un For example J. M. Ferrier, Forerunners of the French Novel (1954). The title of a story by Diderot. 28 P. Stewart, Imitation and Illusion in the French Memoir-Novel 1700-1750 (1969), p 126. 29 G. Hopkins (trans.): Gustave Flaubert, Madame Bovary (1985), p 54. 30 A useful book in this context is P. Danger, Sensations et Objets dans le Roman de Flaubert (1973). 31 Despite the many claims made for it Barthes' S/Z (1974) is really quite narrow in its approach to Balzac's Sarrasine and the coding seems to be applied quite arbitrarily. Perhaps Scruton is right to see it as another variant of the explication de texte: R. Scruton, "The Impossibility of Semiotics" in The Politics of Culture (1981). 27 26 baromètre" and wanders, it engages "un tas pyrimidal" and wanders; the reader ponders on the kind of household that Madame Aubain occupied. A piano is used for entertainment, it is an instrument of enjoyment, it attracts company, it's the focus of singing, and those things passed with the decease of M. Aubain; perhaps the children would have been happier if M. Aubain were still alive and the piano played. What were the boxes doing on top of the piano, had they been left there the day that M. Aubain died? A pile of boxes indicates a lack of order where there was order before, or was there order before? And there is something deeply dissatisfying with the notion that the word 'baromètre' has no resonance once over the purely logical hurdle of meaning. It is a word chosen by Flaubert, the great student of human bêtise. Being difficult to imagine a noun with no connotations it is doubly difficult to imagine Flaubert robbing a word of its connotational appeal: very little could be more stupid than a barometer, except perhaps an English weather- forecast. More to the point, a barometer is used by people who want to know what the weather will be like when they go out. After her husband's death Madame Aubain did not go out but "sat all day long in a wicker easy-chair by the window"32 where she could happily see the weather pass her by. The barometer indicated the reversal of her fortunes as much as the eight mahogany (note: mahogany and not wicker) chairs, presumably used again for entertaining. According to Flaubert's hostile critics, young ladies spent far too much time day-dreaming over novels, and the sorts of thing that could happen to them were amply demonstrated by the story of Madame Bovary. This is not to suggest that one should open the flood-gates to misguided critics who want to know the number of Lady MacBeth's children. It is, however, to suggest that historians should look into the material conditions in which texts were read. How did people actually read novels and what sorts of experiences did they invoke in their reading? One can be fairly certain that people did not read novels in the nineteenth century as they are read today. The point surely is that well-selected details engage the readers' interests and far from invoking a willing suspension of disbelief trade incessantly on their beliefs, hopes and wishes. CONCLUSION It is unfortunate that the late Professor Tatarkiewicz quoted only the first part of the following passage from Quintillian's Institutes of Oratory as the full text is revealing: "There are certain experiences which the Greeks call phantasiae, and the Romans `visions', whereby things absent are presented to our imagination with such extreme vividness that they actually seem before our eyes. ... When the mind is unoccupied or is absorbed by fantastic hopes or day-dreams, we are haunted by these visions of which I am speaking to such an extent that we imagine that we are travelling abroad, crossing the sea, fighting, addressing the people or enjoying the use of wealth that we do not actually possess, and seem to ourselves not to be dreaming but acting. Surely, then, it may be possible to turn this form of hallucination to some profit".33 Indeed Quintillian's phantasiae were turned to some profit; they formed the stock themes of the ancient novels in which the authors worked hard on credibility. Barthes' l'effet du réel does not work because the signified drops out of sight but because description enhances the narrative effect. 32 33 R. Baldick (trans.): Gustave Flaubert Three Tales (1986), p 17. W. Tatarkiewicz, History of Aesthetics I (1970),` p 252 and Quintillian, Institutes of Oratory VI, 2, 29-30.