Toward a Theory of Literary Nonfiction

What is a historical fact? A spent shell? A bombed-out building? A pile of shoes? A victory parade? A long march? Once it has been suffered it maintains itself in the mind of witness or victim, and if it is to reach anyone else it is transmitted in words or on film and it becomes an image, which, with other images, constitutes a judgment. I am well aware that some facts, for instance the Nazi extermination of the Jews, are so indisputably monstrous as to seem to stand alone. But history shares with fiction a mode of mediating the world for the purpose of introducing meaning, and it is the cultural authority from which they both derive that illuminates those facts so that they can be perceived.

E. L. Doctorow, "False Documents"

What do we mean when we contend that a nonfiction narrative is literary? This question has become increasingly important in light of excellent writing by Norman Mailer, Tom Wolfe, Michael Herr, and others, and in response to literary historians, such as Robert Scholes and David Lodge, who argue that our cultural aesthetic is demanding texts that define "reality" and "realism" in new ways. Critical attention to the New Journalism has succeeded in increasing our understanding and appreciation of particular works, but there remains a great deal of confusion about theoretical issues, such as the distinction between fact and fiction, the qualities of literary status in nonfiction, and the responsibilities of the author in turning history into art. Much of the confusion comes from terms such as "nonfiction novel," grandiose assertions such as "there is no difference between fiction and nonfiction," and simple-minded definitions of artistic nonfiction based on the use of techniques common in fiction. In this essay I will argue that literary nonfiction and fiction are fundamentally different, despite their resemblances in structure or technique, and that this difference must be recognized by any theory that hopes to do justice to powerful nonfiction narratives.

In "The Logical Status of Fictional Discourse" John Searle points out that the distinction we commonly make between factual and fictional statements is based, not on any characteristic of the statements themselves, but on our perception of the kind of statement being intended. Suppose a friend tells an amazing anecdote. If we believe it to be a joke or an invention, we look for a punchline or narrative flourishes; if we think it is a true story, we may formulate questions in our minds, asking for supplementary information. The proper response is indicated by the type of story we think we are being told, and that decision in turn is influenced by factors such as our relationship with the storyteller, the social context, and the antecedent conversation, as well as by properties of the story itself. We may not be sure what kind of story we are hearing, in which case we prepare for a joke so as to avoid being duped into treating invention as fact. In any event, we can never know purely on internal evidence whether the story is meant to be taken as true. Perhaps the teller is insane, in which case he may intend his story to be taken seriously, though our inclination is to doubt it. In that doubt lies a clue to the difference between fiction and nonfiction. It would not make sense to "doubt" a work of fiction. When we claim that something is a "true story," we mean either that it is to be taken in a certain way or that it can serve as an adequate representation of real events. The madman's tale is "true" in the first sense, but not in the second. The first distinction is between fact and fiction, the second between good and bad fact. This difference is important because, as the example of the anecdote illustrates, different sorts of responses are appropriate for fiction and nonfiction.

If Searle's distinction makes sense, it follows that the author is sole determinant of whether a text is fact or fiction, whereas the reader must decide for herself whether a work is good or bad fact...