Six Walks in the Fictional Woods

by Umberto Eco

  How does a person become a second-level model reader? We must reconstruct the sequence of events that the narrator virtually lost, in order to understand not so much how the narrator lost it but how Nerval leads the reader to lose it.

  In order to understand what has to be done, we must refer to a fundamental theme of all modern narrative theories, the distinction that the Russian Formalists made between fabula and sjužet—terms that I shall translate in the commonly accepted way as story and plot.

  The story of Ulysses, both as told by Homer and as reformulated by James Joyce, was probably known to the Greeks before the Odyssey was written. Ulysses leaves burning Troy and, with his companions, gets lost at sea. He meets strange peoples and horrible monsters—Laestrygonians, Polyphemus, the Lotus Eaters, Scylla and Charybdis; he descends to the underworld, escapes from the Sirens, and finally is captured by the nymph Calypso. At this point the gods decide to help him return to his homeland. Calypso is forced to liberate Ulysses, who returns to the sea, is shipwrecked, and tells Alcinoüs his tale. Then he sets sail for Ithaca, where he defeats the suitors of Penelope and is reunited with her. The story proceeds in a linear fashion from an initial moment, T1, toward a final moment, Tx (see Figure 6).

  Figure 6: Story

  The plot of the Odyssey, however, is quite different. The Odyssey begins in medias res, at a moment T0, when the voice that we call Homer begins to speak. We can identify this moment, as we please, either with the day on which Homer allegedly began to narrate or with the moment we begin to read. What matters is that the plot starts at a moment T1 when Ulysses is already Calypso’s prisoner. Between this moment and a moment T2, which corresponds to Book 8, Ulysses escapes from Calypso’s amorous advances, is shipwrecked among the Phaeacians, and tells his tale. But at this point the story goes backward to a time that we shall call T3 and deals with Ulysses’ previous adventures. This flashback lasts for a large part of the epic, and only in Book 13 does the text bring us back to where we were in Book 8. Ulysses concludes his reminiscences and sets sail for Ithaca (see Figure 7).

  Figure 7: Plot

  There are some tales, such as fairy tales, which are called “simple forms” because they have only a story, without any plot. “Little Red Riding Hood” is one of these. It begins with the little girl’s leaving home and entering the wood, and ends with the death of the wolf and the girl’s return home. Another example of simple form may be Edward Lear’s limericks:

  There was an old man of Peru

  who watched his wife making a stew;

  But once by mistake

  In a stove she did bake

  That unfortunate man of Peru.

  Let’s try to tell this story as it might be reported by the New York Times: “Lima, March 17. Yesterday Alvaro Gonzales Barreto (41, two children, accountant at Chemical Bank of Peru) was erroneously cooked in a shepherd’s pie by his wife, Lolita Sanchez de Medinaceli . . .” Why is this story not as good as Lear’s? Because Lear tells a story, but the story is the content of his tale. This content has a form, an organization, which is that of the simple form, and Lear does not complicate it with a plot. Instead, he expresses the form of his narrative content through a form of expression, consisting of the metrical patterns and playful rhymes typical of the limerick. The story is communicated through a narrative discourse (see Figure 8).6

  We could say that story and plot are not functions of language but structures that can nearly always be translated into another semiotic system. In fact I can recount the same story of the Odyssey, organized according to the same plot, by means of linguistic paraphrase, as I have just done, or in a film or comic book, since there are flashback signals in these two semiotic systems as well. On the other hand, the words with which Homer tells the story are part of the Homeric text and cannot be paraphrased or translated into images very easily.

  Figure 8: Text

  A narrative text may conceivably lack a plot, but it cannot possibly do without story or discourse. Even the story of Little Red Riding Hood has come down to us through different discourses—Grimm’s, Perrault’s, our mother’s. Discourse is also part of a model author’s strategy. Lear’s indirect pathetic comment telling us that the old man of Peru was “unfortunate” is an element of discourse, not of story. In a certain sense, it is discourse, not story, that lets the model reader know whether he should be touched by the old man’s fate. The very form of the limerick, which cues us to view the content as absurd, ironic, and tongue-in-cheek, is also part of discourse, so that in choosing this form Lear is telling us that we can laugh at a story which might make us cry if told in the discursive mode of the New York Times.

  When the text of Sylvie says “While the carriage is going up the hills, let us recollect the memories of the days I went there so often,” we know that it is not the narrator but the model author who is speaking to us. It is clear that at this moment the model author reveals Itself in the way It organizes the story: not by means of a plot, but through a narrative discourse.

  Many theories of literature have insisted that the voice of the model author should be heard solely through the organization of facts (story and plot); such theories reduce the presence of discourse to a minimum—not as if it were not there, but as if the reader should not be aware of its indications. For T. S. Eliot, “the only way of expressing emotion in the form of art is by finding an ‘objective correlative’; in other words a set of objects, a situation, a chain of events which shall be the formula of that particular emotion.”7 Proust, although he praises the style of Flaubert, nevertheless faults him for writing phrases such as “those good old inns which always have something country-like about them.” He quotes the line “Madame Bovary approached the fireplace” and observes with satisfaction, “It was never said that she was cold.” Proust wants “a tightly knit style, of porphyry, without any cracks, without any additions,” in which we see a mere “apparition” of things.8

  A term like “apparition” reminds us of the “epiphany” of Joyce. In The Dubliners, there are some epiphanies in which the mere representation of events tells readers what they must try to understand. On the other hand, in the epiphany of the girl-bird in A Portrait of the Artist as a Young Man, it is the discourse, not the simple story, that orients the reader. This is why I think it is impossible to translate the apparition of the girl in A Portrait into a film, whereas John Huston managed to render the atmosphere of a story such as “The Dead” (in his film of the same name) by simply dramatizing the facts, the situations, and people’s conversations.

  I have been obliged to make this long digression on the various levels of the narrative text because the time has come to answer a very tricky question: If there are texts that have only a story and no plot, isn’t it also possible that some texts, like Sylvie, might have only a plot and no story? Is Sylvie simply a text that says how impossible it is to reconstruct a story? Does the text ask the reader to fall ill, like Labrunie—incapable of distinguishing dreams, memories, and reality? Doesn’t the very use of the imperfect perhaps say that the author wanted us to get lost, not that we should analyze his use of the imperfect?

  It is a matter of choosing between two statements. In one, Labrunie ironically says (in a letter to Alexandre Dumas which appears in Les Filles du feu) that his works are no more complicated than Hegel’s metaphysics, adding that “they would lose their charm if they were explained, if that were possible.” The other is certainly by Nerval and appears in the last chapter of Sylvie: “Such are the delusions [the French text says chimères] which charm and lead us astray in the morning of life. I have tried to set them down without too much order, but many hearts will understand me.” Should we take Nerval to mean that he hasn’t followed any order, or instead recognize that the order he has followed is not readily apparent? Should we assume that Proust—who analyzed in such detail Flaubert’s use of verb tenses and who was so alert to the effects of narrative strategies—asked Nerval to do nothing more than seduce him with his imperfects, an
d thought that Nerval was using this cruel tense, which presents life as something ephemeral and passive, merely to inspire his readers with vague sadness? And is it likely that Labrunie could have taken such pains to order his work, yet did not want us to perceive and admire the devices he had used to cause us to lose our way?

  I have been told that Coca-Cola tastes good because it contains some secret ingredients that the wizards in Atlanta will never reveal—but I do not like such Coke-oriented criticism. I am reluctant to think that Nerval would not have wanted the reader to recognize and appreciate his stylistic strategies. Nerval wanted us both to feel that the periods of time were blurred and to understand how he had managed to blend them.

  One objection might be that my notion of literature does not correspond to Nerval’s, and perhaps not even to Labrunie’s; but let us return to the text of Sylvie. This story—which starts with a vague “Je sortais d’un théâtre,” as if wanting to create an atmosphere similar to that of a fairy tale—ends with a date, the only one in the book. On the last page, when the narrator has lost all his illusions, Sylvie says, “Poor Adrienne! She died in the convent of Saint-S. . . around 1832.”

  Why should there be this imperious date which appears at the very end, the most strategic point of the text, and which seems to interrupt the spell with a precise reference? As Proust said, “One is constantly obliged to turn back to an earlier page to see where one is, if it is the present or the past recalled.” And if we do go back, we realize that the whole narrative discourse is studded with temporal hints.

  They are invisible on the first reading, but on the second they are quite obvious. At the time he tells his story, the narrator says that he has already been in love with the actress for a year. After the first flashback he refers to Adrienne as a “face forgotten for years,” but he thinks about Sylvie and wonders, “Why have I forgotten her for three years?” At first the reader thinks that three years have passed since the first flashback, and gets even more lost, because if that were the case the narrator would be a mere boy instead of a pleasure-seeking young man. But at the beginning of the fourth chapter, at the start of the second flashback, when the carriage is going up the hill, the text opens with “Some years had passed.” Since when? Probably from the childhood days that were described in the first flashback. The reader may think that some years had passed from the time of the first flashback to the time of the second, and that three years have elapsed from the time of the second until the time of the journey . . . During the second flashback, it is clear that the narrator stays in that place for one night and the following day. The seventh chapter (which has the most confused temporal sequence) begins with “It’s four in the morning,” and the following one tells us that the narrator arrives in Loisy toward dawn. From the moment the narrator returns to Paris and begins his love affair with the actress, indications of time become more frequent: we are told that “months passed”; after a particular event, “the following days” are mentioned; then we read of “two months later,” then of “next summer,” of “next day,” “that evening,” and so on. Perhaps this voice telling us of time links may want us to lose our sense of time, but it also encourages us to reconstruct the exact sequence of events.

  That’s why I would like you to look at the diagram shown in Figure 9. Please don’t consider it a cruel, unnecessary exercise. It will help us grasp the mystery of Sylvie a little better. On the vertical axis, let’s put the implicit chronological sequence of events (the story), which I have reconstructed, even where Nerval merely gives faint hints of it. On the horizontal axis, we have the sequence of chapters—that is, the plot. The sequence of the explicit story, which Nerval informs us about in the text, looks like a sawed-off, jagged horizontal line on the plot axis; from this line branch off vertical arrows pointing to the past. The continuous vertical arrows represent the narrator’s flashbacks; the dotted ones stand for the flashbacks (hints, allusions, brief recollections) the narrator attributes to Sylvie or to other characters (including himself, when describing his memories to Aurélie). They should start from the virtual present perfect, in which the narrator is speaking, and point to a virtual past perfect. Both tenses, however, are continually disguised by the use of the imperfect.

  At what moment does the narrator speak? That is to say, when is the T0 at which he speaks? Since the text makes reference to the nineteenth century and since Sylvie was written in 1853, let’s take 1853 as the Time Zero of the narration. This is a mere convention, a postulate on which to base my discussion. I could just as well have decided that the voice is speaking today, in 1993, while we are reading. What matters is that once a T0 has been established, we can have an exact countdown, using only the data given by the narrative discourse.

  Figure 9

  If we calculate that Adrienne died in 1832, after the narrator had met her as a youth, and if we consider that, after the night he took the carriage and after he returned to Paris two days later, the narrator is fairly clear in telling us that months and not years have passed since he started his relationship with the actress, it is possible to establish, approximately, that the evening spoken about in the first three chapters and the whole episode of his return to Loisy occur in 1838. If we imagine the narrator at that time as a dandy in his early twenties, and if we take into account that the first flashback describes him as a boy of perhaps twelve years of age, we can establish that the first recollection goes back to 1830. And since we are told that as of 1838 three years had passed since the time referred to in the second flashback, we can assume that the events had taken place in about 1835. The final date, 1832, when the death of Adrienne occurs, is of help to us because in our reconstructed year 1835 Sylvie makes vague allusions leading us to think that at that time Adrienne was already dead (“Votre religieuse . . . Cela a mal tourne”—that is, “It had an unhappy ending,” “It turned out badly”). Thus, having fixed two conventionally precise chronological references—1853 as Time Zero of the narrating voice, and an evening in 1838 as Time 1, in which the game of memories begins—we can establish a regressive series of times leading us back to 1830, as well as a progressive one, which leads to the final separation from the actress, more or less in 1839.

  What does one gain from this reconstruction? Nothing at all, if one is only a first-level reader. One might, perhaps, be able to clear some of the mist, but loses the spell of being lost. Second-level readers, on the other hand, realize that these recollections do have an order and that these sudden shifts in time and quick returns to the historical present follow a rhythm. Nerval has managed to create his misty effects by composing a sort of musical score.

  It is like a melody, which the reader can enjoy first for the effects it elicits, and later by discovering how an unexpected series of intervals can produce these effects. This score tells us how a tempo is imposed on the reader by “shifting gears,” so to speak. Most of the flashbacks occur in the first twelve chapters, which cover twenty-four hours (from eleven o’clock at night, when the narrator leaves the theater, to the following evening, when he leaves his friends to return to Paris the next day). One could say that in these twenty-four hours, eight previous years are embedded. But this depends on an “optical” illusion resulting from my reconstruction. On the vertical axis of Figure 9, I have scored all the steps of the story that Sylvie as a text presupposes but does not explicitly tell—because the narrator is unable to control these temporal strategies. Of these eight years, only a few points, a few scattered fragments, are retrieved. Thus, we have an enormously dilated space in the plot in which to narrate a few disconnected hours of the story, since these eight years are not narratively recaptured and we have to figure them out, lost as they are in the mist of a past which, by definition, cannot be recovered. It is the quantity of pages devoted to the effort of recollecting these hours, without actually reconstructing their actual sequence, it is the disproportion between the time of recollection and the time actually recollected, that produces such a feeling of tender agony
and sweet defeat.

  Precisely because of this defeat, the final events occur rapidly, in just two chapters. We skip months, and suddenly come to the end. The narrator justifies his quickness by commenting, “What can I say now which is not the story of so many others?” There are only two brief flashbacks. One is by the narrator, telling the actress about his youthful vision of Adrienne (and he is now no longer dreaming, but telling Aurélie a story the reader is already aware of); the other comes from Sylvie, like a thunderbolt, establishing the date of Adrienne’s death, as the only real, undeniable fact of the whole story. In the last two chapters the narrator quickens the plot, because there is nothing left of the story to discover. He has given up. This change in pace makes us shift from a time of enchantment to a time of delusion, from a motionless time of dreams to an accelerated time of facts.

  Proust was right when he said that this rainbow-hued picture is evoked by music and that it lies not in the words but among the words. As a matter of fact, it is created by the relationship between plot and story, which commands even the lexical choices of the discourse. If you place the transparent story/plot grid over the discursive surface of the text, you discover that it is exactly in these nodes, when the plot jumps backward in time or returns to the main stream of the narration, that all the changes in tense occur. All these shifts from imperfect to present or past perfect, or from the past perfect continuous to the present and vice versa, are certainly unexpected and frequently imperceptible, but never unmotivated.