On Literature

by Umberto Eco

  Until I thought of what was one of the most successful fakes of Western history, namely, the Letter of Prester John. This idea fructified a series of memories and reading experiences. In 1960 I edited the Italian edition (Le terre leggendarie) of Lands Beyond by Ley and Sprague du Camp for Bompiani. There was a chapter on Prester John's kingdom and another on the lost tribes of Israel. On the cover they had put a skiapod, or shadow-footed monster (from a fifteenth-century engraving, I think, with fake coloring and done with a stencil). Years later I bought a colored map from an Ortelius atlas that had been cut up, the very map representing the lands of Prester John, and I hung it in my study. In the 1980s I read various versions of the letter.† In short, Prester John had always intrigued me, and I was attracted by the idea of making the monsters that populated his Kingdom come alive again, as well as those spoken of in the various Alexander Romances, Mandeville's travels, and a whole series of bestiaries. And finally it was a good opportunity to return to my beloved Middle Ages. So my seminal idea was that of Prester John. But I had not started with this idea, I had simply arrived at it.

  Perhaps all this would not have been enough for me had the letter not been attributed (it is one possible hypothesis) to the Imperial Chancery of Frederick Barbarossa. Now Frederick Barbarossa was another magical name for me, because I was born in Alessandria, the city that was founded in order to oppose the emperor. That led to a series of instinctive decisions, like a chain reaction: to discover a Frederick that went beyond the traditional clichés, seen by a son rather than by his enemies and courtiers (and off I went with further reading on Barbarossa), to tell the origins of my city and its legends, including that of Gagliaudo and his cow. Years earlier I had written an essay on the foundation and history of Alessandria (entitled, as it happens, "The Miracle of San Baudolino"),* and from there came the idea to have this history lived through by a character called after the patron saint of the city, Baudolino, to make Baudolino the son of Gagliaudo, and to give the story a popular, picaresque thrust—thus creating a sort of counterpoint to The Name of the Rose, since the latter was a story of intellectuals talking in the high style, whereas this was a tale about the people and military men who were on the whole rather crude, talking in a style that was almost like dialect.

  But here too, what was I to do? Make Baudolino talk in his Po Valley twelfth-century pseudodialect, when we have very few vernacular documents from that period, and none of them from the Piedmont area? Make a narrator talk, and have his modern style spoil Baudolino's spontaneity? However, suddenly at this point another obsession came to my rescue, one that had been going through my head for some time, without my ever thinking that it would be of use to me on that occasion: narrate a story set in Byzantium. Why? Because I knew very little about Byzantine civilization, and I had never been to Constantinople. To many this might seem a rather weak motive for deciding to narrate something that happened in Constantinople, all the more so since Constantinople had only a tangential link with Frederick Barbarossa's story. But sometimes one decides to tell a story only to get to know it better.

  No sooner said than done. Off I went to Constantinople. I read many things about ancient Byzantium, mastered its topography, and came across Nicetas Choniates and his Chronicle. I had found the key, the way to articulate the "voices" of my story: an almost transparent narrator recounts the discussion between Nicetas and Baudolino, alternating Nicetas's learned, high-flown reflections with Baudolino's picaresque tales, without Nicetas, or the reader, ever being able to tell if and when Baudolino is lying, the only fixed point being that he maintains he is a liar (the paradox of the liar and the Cretan Epimenides).

  I had this play of "voices," but not Baudolino's voice. Here I contradicted the second of my two principles. When I was still reading the chronicles of the Crusaders' capture of Constantinople (and I decided that I would have to narrate the story of that event which already appears so novelistic in the texts of Ville-hardouin, Robert de Clary, and Nicetas), just to pass the time I wrote out in pen, in the country, a sort of diary by Baudolino, in a hypothetical twelfth-century Po Valley pidgin, which then became the opening of the novel. It is true that I rewrote those pages several times in subsequent years, after consulting historical and dialect dictionaries, and all the documents I could lay my hands on, but already in that first draft, through its linguistic style, it became clear to me how Baudolino would think and speak. Thus, in the end, Baudolino's language was not born from the construction of a world, but a world was created from the stimulus of that language.

  I do not know how to solve this dilemma in theoretical terms. All I can do is quote Walt Whitman: "Do I contradict myself? Very well then I contradict myself." Except that this use of dialect probably took me back to my childhood and my native area, and therefore to a preconstructed world, at least in memory.

  Constraints, and time

  And yet (world versus language or language versus world) it is not that you spend two or three years constructing a world as though that world existed on its own, independently of the story you want to set in it. This "cosmogonical" phase goes hand in hand with (and in a way that I really could not reduce to a formula or program) a hypothesis about the supporting structure of the novel—and of the world you are creating. This structure consists essentially of constraints and temporal rhythms.

  Constraints are fundamental in every artistic operation. A painter who decides to use oil rather than tempera, or a canvas rather than a wall, is choosing a constraint; likewise the composer who opts for one tonality at the outset (he may then modulate it all he likes, but he has to return to that opening tonality), and the poet who builds what is a cage of rhyming couplets or of hendecasyllables. And do not think that avant-garde painters, composers, and poets—who seem to avoid those constraints—do not construct others. They do, but you may not be obliged to notice them.

  Choosing the seven trumpets of the Apocalypse as a scheme for the succession of events can be a constraint. But also setting the story at a precise date: you can make some things happen then, but not others. It can be a constraint deciding that, to indulge the magic obsessions of your characters, the number of chapters in Foucault's Pendulum has to be 120, not one more or one less, and it has to be divided into ten parts, like the Sephiroth in the Kabbalah.

  Constraints then gradually determine a temporal sequence. In The Name of the Rose, if the story had to follow a sequence based on the Apocalypse, plot time could coincide (except for substantial digressions) with the time of the story: the story begins with the arrival of William and Adso at the abbey, and ends with their departure. Easy (and easy to read).

  With Foucault's Pendulum the very pendular movement of the eponymous device forced me to adopt a different temporal structure. Casaubon arrives at the Conservatoire one evening, hides there, and conjures up past events, then the story returns to its starting point, etc. If for The Name of the Rose I had gradually built up a kind of sequential timetable or calendar, trying to work out what was to happen each day for a week, for Foucault's Pendulum I created a sort of serpentine structure, which registered the shifts back to the past and the anticipations of the future. Like a measuring scale, or orthogonal Cartesian axes. The character is here now, but remembers what happened at time X in the past.

  The beauty is that these schemes are rigid when you think of them at the time, but I have drawers full of schemes I constantly redid as the novel progressed. I mean, the beauty of the story is that you have to create constraints, but you must feel free to change them in the course of writing. Except that at that point you have to change everything and start again from scratch.

  Besides, one of the constraints in the Pendulum was that the characters had to have lived through 1968, but since Belbo then writes his files on computer—which also plays a formal role in the whole story, since it in part inspires its aleatory and combinatory nature—the final events absolutely had to take place between 1983 and 1984 and not before. The reason is very simple: the first pers
onal computers with word-processing programs went on sale in Italy in 1983 (or perhaps 1982). And this is worth bearing in mind by all those who try to explain the success of The Name of the Rose by insisting that it was written by computer. In 1978–79 you could scarcely find in America those cheap little computers called Tandy, and no one would ever have dared write more than one letter on them.

  In order to make all that time elapse from 1968 to 1983,1 was forced to send Casaubon somewhere else. Where? Memories of some magic rituals I had witnessed there led me to Brazil (there I knew what I was talking about and what the shape of that world was). And this was the reason and the auspicious origin of what many found too long a digression, and which for me (and for some benevolent readers) was essential, because it allowed me to make happen in Brazil to Amparo, in shortened form, what would happen to the other characters in the course of the book. If IBM, Apple, or Olivetti had started selling word processors six or seven years earlier, my novel would have been different; there would have been no Brazil, to the relief of many superficial readers, but from my point of view it would have been a great loss.

  The Island of the Day Before was based on a series of historical constraints and rigid novelistic restrictions. The historical constraints evolved from the fact that I needed Roberto to participate as a young man in the siege of Casale, to be there at Richelieu's death, and then to arrive at his island after December 1642, but not later than 1643, the year in which Tasman went there, even though this was some months earlier than the time in which my story was set. But I could set the story only between July and August because that was the period when I saw the Fiji islands, and a ship took several months to get there: this explains the mischievous novelistic insinuations that I made in the final chapter, to persuade myself and the reader that perhaps Tasman had come back later to that archipelago without saying anything to anybody. Here one sees the heuristic usefulness of constraints that force you to invent silences, conspiracies, ambiguities.

  You will ask me: why all these constraints? Was it really necessary for Roberto to be present at Richelieu's death? Not at all. But it was necessary for me to set myself constraints. Otherwise the story could not have gone along under its own steam.

  As for the novelistic constraints, Roberto had to be on the ship, not be able to get off it, and try in vain to learn to swim in order to reach the island. In the meantime, as he reflected on life and death, he would have to invent bit by bit, and then reject, through lack of intelligence, all the philosophical thought of that century. For the benevolent reader this would be more than just a constraint, placed just to receive stimuli: it would be the very essence of Desire. I would be the last to deny it. But since I am speaking about how I wrote and not what the reader could or should find in what I wrote (because, to say the latter, either the novel is enough on its own, or I have been wasting my time writing it and you reading it—which is not impossible), what I mean to say is that, on the one hand, it is the constraint that allows the novel to develop according to a particular Sense, and on the other it is the still-unclear idea of this Sense that suggests the constraints. Since one of these cannot function without the other, we talk of constraints rather than Sense, which is not something an author should pronounce on a posteriori.

  A parenthesis. A rude hack—who wanted to mock Eco the novelist in order to punish Eco the politically committed polemicist—defined the novel as one long act of masturbation. In his crudeness, which was also lexical, the unwitting hack was dead right: the condition of a shipwreck separated forever from the object of desire is certainly, and by definition, onanistic. Except that the scribbler I am talking of, anchored in his own obtuse carnality, saw manna falling from heaven and read it as excreta from foul birds of prey. Nor did he catch the nature of "something mental"—and in the end metaphysical—in that solitary virtue, in that attempt to generate Being by spilling in disordered fashion the seed of a soul exasperated by solitude, until it reaches the point of vision.

  But let us get back to the point. The point is that Roberto could not leave the ship (except at the end, but for uncertain objectives and outcomes). Consequently everything that had to be narrated but that did not happen on the ship had to come about via memory, unless the plot was to be flattened into just being the story, and end up relating in every detail how a young man, who went to Paris after his adventures at Casale, found himself on a ship, etc. Try if you like, but I can assure you that however vain my labors have been, yours will be even more so.

  This imposed on me not a serpentine temporal sequence, as in the Pendulum, but more one step forward and three back, one step forward and two back, one step forward and one back. Roberto remembers something, and meanwhile something happens on the ship. Something happens on the ship, and Roberto remembers something. Gradually, as Roberto's memories move from 1630 to 1643, events on the ship happen hour by hour. All this up to the arrival of Father Caspar. At that point the story stops, as it were, in the present, for some time. Then Father Caspar disappears in the sea, and Roberto is on his own again.

  What was I to make him do? The novelistic constraints meant I had to make him try various ways of getting ashore. But these had to be slow, recorded day after day, repetitive and monotonous. In the end I still had to write a novel, whose aim—let it be said in the face of every aesthete, and in full respect for the laws of the genre as they have come down to us from the Hellenistic romance to our own time, not to mention Aristotle's Poetics— had to be that of providing narrative pleasure.

  Fortunately I was the victim of another constraint. To conform to the spirit of the seventeenth-century novel, I had to introduce a Double, and I really did not know what to do with him. Suddenly this Double came good: while he tries to reach the island, learning each day to swim better (but never well enough), Roberto imagines the novel about his Double, and thus the one-step-forward-three-back structure could be reproduced, since Roberto cannot reach the island but makes his Double reach it, making him start from the point where he himself started. How nice to see a novel writing itself! I did not know where I would get to, because the novelistic constraint dictated that Roberto should not get anywhere. The novel ends because it heads directly toward its conclusion on its own. This is what I would like my Model Reader to notice. That the novel writes itself, since that is how it happened, and how it always happens, really.

  Speaking of constraints, the finale of Baudolino had to take place in 1204, because I wanted to narrate the conquest of Constantinople. But Baudolino had to be born around the middle of the century (I fixed on 1142 as a point of reference, so as to have my character at the age of reason and consent at the time of many events I wanted to relate). The first mention of Prester Johns letter is around 1165, and I already make it circulate a few years later, but why then does Baudolino, after persuading Frederick to give him permission, not set off immediately for Prester John's Kingdom? Because I had to have him coming back from the Kingdom only in 1204, so he could tell the tale to Nicetas during the burning of Constantinople. And what did I have to make Baudolino do in that interval of almost four decades? It was a bit like the business of the computer in Foucault's Pendulum.

  I make him do many things, and I constantly make him delay his departure. At the time it seemed like a waste to me, it was like inserting a series of temporal stopgaps into the story in order to arrive finally at that damned date of 1204. And yet, when you look at it closely (and I hope, or rather I know, that many readers did realize this), I created the Spasm of Desire (or, rather, the novel created it without my realizing it there and then). Baudolino wants the Kingdom but constantly has to postpone his search. Thus Prester John's Kingdom grows in Baudolino's desire, and in the reader's eyes (I hope). Once more the advantages of constraints.

  How I Write

  At this point one can understand how useless are questions like "Do you begin with notes, immediately write the first or last chapter, write with pen, pencil, typewriter, or computer?" If one has to construct a wor
ld, day by day, and try out endless temporal structures, if the actions the characters perform and have to perform according to the logic of common sense or of narrative convention (or against narrative convention) have to fit with the logic of the constraints (involving constant rethinks, cancellations, and rewrites), there is no uniform way of writing a novel.

  At least for me. I know of writers who wake up at 8:00 A.M., work at their keyboard from 8:30 to 12:00 ("nulla dies sine linea; no day without at least a line"), and then stop and go out and enjoy themselves until evening. Not me. First of all, when I write a novel, the act of writing comes later. First I read, make notes, draw portraits of the characters, maps of the places, and plans for the time sequences. And these are done with a felt pen, or computer, depending on when and where one does this, or on the kind of narrative idea or detail one wants to record: on the back of a train ticket, if the idea comes to you on a train, in a notebook, on an index card, using ballpoint, tape recorder, or blackberry juice if really necessary.

  Then what happens is that I chuck out, tear up, tear into pieces, forget things in different places, but I have boxes full of notebooks, with blocks of pages in different colors, bits of card, even sheets of foolscap. And this chaotic variety of props helps my memory, because I remember that I jotted down that particular note on the letterhead paper of a London hotel, and the first page of that chapter was scribbled down in my study, on an index card with pale blue lines, and using a Mont Blanc pen, whereas the following chapter was initially written down in the country, on the back of a recycled piece of draft paper.