None of these writers, ancient or modern, sat down to write “to express himself.” They sat down to write this kind of story or that, or to mix this form with that form, producing some new effect. Self-expression, whatever its pleasures, comes about incidentally. It also comes about inevitably. The realistic writer may set out to conjure up the personality of his aunt, creating for her, or copying from life, some story through which her character is revealed, and thus he reveals his strong feelings about his aunt; that is, he expresses himself. The fabulist—the writer of nonrealistic yarns, tales, or fables—may seem at first glance to be doing something quite different; but he is not. Dragons, like bankers and candy-store owners, must have firm and predictable characters. A talking tree, a talking refrigerator, a talking clock must speak in a way we learn to recognize, must influence events in ways we can identify as flowing from some definite motivation; and since character can come only from one of two places, books or life, the writer’s aunt is as likely to show up in a fable as in a realistic story. Thus the process by which one writes a fable, on one hand, or a realistic story, on the other, is not much different. Let us look more closely at the similarities and differences.

If we carefully inspect our experience as we read, we discover that the importance of physical detail is that it creates for us a kind of dream, a rich and vivid play in the mind. We read a few words at the beginning of the book or the particular story, and suddenly we find ourselves seeing not words on a page but a train moving through Russia, an old Italian crying, or a farmhouse battered by rain. We read on—dream on—not passively but actively, worrying about the choices the characters have to make, listening in panic for some sound behind the fictional door, exulting in characters’ successes, bemoaning their failures. In great fiction, the dream engages us heart and soul; we not only respond to imaginary things—sights, sounds, smells—as though they were real, we respond to fictional problems as though they were real: We sympathize, think, and judge. We act out, vicariously, the trials of the characters and learn from the failures and successes of particular modes of action, particular attitudes, opinions, assertions, and beliefs exactly as we learn from life. Thus the value of great fiction, we begin to suspect, is not just that it entertains us or distracts us from our troubles, not just that it broadens our knowledge of people and places, but also that it helps us to know what we believe, reinforces those qualities that are noblest in us, leads us to feel uneasy about our faults and limitations.

Some writers—John Barth, for instance—make a point of interrupting the fictional dream from time to time, or even denying the reader the chance to enter the fictional dream that his experience of fiction has led him to expect. We will briefly examine the purpose and value of such fiction later. For now, it is enough to say that such writers are not writing fiction at all, but something else, metafiction. They give the reader an experience that assumes the usual experience of fiction as its point of departure, and whatever effect their work may have depends on their conscious violation of the usual fictional effect. What interests us in their novels is that they are not novels but, instead, artistic comments on art.

The reason the approach seems to me not ideal is that, except in the extraordinary case, it wastes the writer’s time. It instructs him to do something he cannot realistically be expected to do well—and here I mean “well” in the always urgent artist’s sense, not the more casual, more gentlemanly way in which we do things badly or well in other university programs. Let me explain. True artists, whatever smiling faces they may show you, are obsessive, driven people—whether driven by some mania or driven by some high, noble vision need not presently concern us. Anyone who has worked both as artist and as professor can tell you, I think, that he works very differently in his two styles. No one is more careful, more scrupulously honest, more devoted to his personal vision of the ideal, than a good professor trying to write a book about the Gilgamesh. He may write far into the night, he may avoid parties, he may feel pangs of guilt about having spent too little time with his family. Nevertheless, his work is no more like an artist’s work than the work of a first-class accountant is like that of an athlete contending for a championship. He uses faculties of the mind more easily available to us; he has, on all sides of him, stays, checks, safeties, rules of procedure that guide and secure him. He’s a man sure of where he stands in the world. He belongs on sunlit walkways, in ivied halls. With the artist, not so. No critical study, however brilliant, is the fierce psychological battle a novel is. The qualities that make a true artist—nearly the same qualities that make a true athlete—make it important that the student writer never be prevented from working as seriously as he knows how to. In university courses we do exercises. Term papers, quizzes, final examinations are not meant for publication. We move through a course on Dostoevsky or Poe as we move through a mildly good cocktail party, picking up the good bits of food or conversation, bearing with the rest, going home when it comes to seem the reasonable thing to do. Art, at those moments when it feels most like art—when we feel most alive, most alert, most triumphant—is less like a cocktail party than a tank full of sharks. Everything’s for keeps, nothing’s just for exercise. (Robert Frost said, “I never write exercises, but sometimes I write poems which fail and then I call them exercises.”) A course in creative writing should be like writing itself; everything required should be, at least potentially, usable, publishable: for keeps. “A mighty will,” Henry James said, “that’s all there is!” Let no one discourage or undermine that mighty will.

[...] Good description is symbolic not because the writer plants symbols in it but because, by working in the proper way, he forces symbols still largely mysterious to him up into his conscious mind where, little by little as his fiction progresses, he can work with them and finally understand them. To put this another way, the organized and intelligent fictional dream that will eventually fill the reader’s mind begins as a largely mysterious dream in the writer’s mind. Through the process of writing and endless revising, the writer makes available the order the reader sees. Discovering the meaning and communicating the meaning are for the writer one single act. One does not simply describe a barn, then. One describes a barn as seen by someone in some particular mood, because only in that way can the barn—or the writer’s experience of barns combined with whatever lies deepest in his feelings—be tricked into mumbling its secrets.

[...] Does fiction, in fact, have anything whatever to do with truth? Is it possible that this complicated instrument, fiction, studies nothing but itself—its own processes?

A common answer at the present time is that that is the question the serious writer spends his whole life trying to work out by means of the only kind of thinking he trusts; that is, the fictional process. For the moment, we must let that answer stand—with only this reservation: Great fiction can make us laugh or cry, in much the way that life can, and it gives us at least the powerful illusion that when we do so we’re doing pretty much the same things we do when we laugh at Uncle Herman’s jokes, or cry at funerals. Somehow the endlessly recombining elements that make up works of fiction have their roots hooked, it seems, into the universe, or at least into the hearts of human beings. Somehow the fictional dream persuades us that it’s a clear, sharp, edited version of the dream all around us. Whatever our doubts, we pick up books at train stations, or withdraw into our studies and write them; and the world—or so we imagine—comes alive.