The philosopher David Lewis was best known for his theory of possible worlds. There are, according to him, a plurality of worlds, each one as actual to its inhabitants as this world. Moreover, these worlds really exist. They hum and buzz with implication. Their vile blows and buffets incense the reckless to spite them too. They aren’t the actual world; only this world is the actual world. On the other hand, it is the actual world only to us. For those inhabiting a possible world, their world is actual to them. The word actual merely indexes the inhabitants’ relationship to their own world; it points to nothing absolute or necessary about it. Actuality ought not to be confused with reality. Possibilities are also real; they just aren’t actual.
Lewis used to say that his theory of possible worlds was met with a standard response—the incredulous stare. But he was entirely serious. Any time a hypothetical or conditional (or as Lewis liked to call it, a counterfactual) is spoken, a real possibility—a possible world—is created:
Daidle deedle daidle
Daidle daidle deedle daidle dum
All day long I’d biddy-biddy-bum
If I were a wealthy man.
Stare incredulously if you must. The advantage of the theory is that it provides another method to assess the truth of statements without having to submit them to “proof.” Its usefulness in talking about fiction should be clear, although Lewis denied that fictional worlds were possible worlds in any simple sense.
In his 1978 essay “Truth in Fiction,” Lewis analyzes the value of such fictional statements as “Sherlock Holmes lived in Baker Street” and “Holmes lived nearer to Paddington Station than to Waterloo Station.” His basic solution is to prefix the intensional operator “In such-and-such a fiction” to a fictional statement to form a new sentence, which is no longer self-evidently false. In fiction, the prefix is dropped; to recover its truth, however, the prefix must be reattached.
The operation of the prefix is universally understood by readers of fiction, if the words that I quoted on Friday afternoon from Sir Philip Sidney’s Apologie for Poetrie are true: “What child is there that, coming to a play, and seeing Thebes written in great letters upon an old door, doth believe that it is Thebes?” Inside the old door, however, sentences have a different kind of truth. They are true inside the door, and remain true outside if and only if the prefix “In such-and-such a fiction” is made explicit.
Lewis works through several consequences of this view before arriving at the following analysis:
What follows is that storytelling occurs in at least two worlds—our own actual world, where it is a pretense of relating fact, but also the possible world where the fiction is related as fact rather than fiction. In short, fictions have what Lewis calls a “trans-world identity.” They are the same, word for word, in at least two worlds where they are treated very differently.
The more these two worlds resemble each other—the more that the “collective beliefs” in them overlap—the closer the fictional statements approach truth, at least when compared to another possible world where the fiction is attested to be known fact but nevertheless its assertions are patently untrue.
Truth in fiction is a product of the fictional world’s resemblance to the actual world in which the pretense of fiction was originally carried out—our own world, perhaps, or more likely a different cultural milieu, at a different time, in a different place. Forster’s assertion that “Even when we love people, we desire to keep some corner secret from them, however small: it is a human right: it is personality” is less immediately true in a fictional world where there are no human persons, but only cyborgs, say, for whom “love” is, I don’t know, a maintenance function. (Bill Benzon has speculated about “artificial love” over at the Valve.) Not only must the fiction have a trans-world identity; so must certain beliefs. If and only if these beliefs are held in every possible world of the fiction’s origin can its assertions be true.
The consequence? Fiction cannot secure the truthfulness of its statements intrinsically. The second half of Bowen’s famous aphorism (fiction “must, therefore, contain uncontradictable truth, to warrant the original lie”) is half-right. The uncontradictable truth must also be contained in the community where the fiction was originally written. Fiction’s truth may only be secured extrinsically. More than the fictional world alone must exist for fiction to enter the service of truth.
What else? Realism can now be seen as the genre of fiction that increases its resemblance to the actual world in its concrete details, its buzzing and humming, its biddy-biddy-bumming, for the sake of improving its chances to shame the devil. Fantasy dispenses with surface resemblances to pursue truth, as Ellison’s Invisible Man phrases it, “on the lower frequencies” (or the higher).
And while Lewis’s analysis contradicts my earlier hypothesis that fictional truth-assertions are validated by authority, it doesn’t entirely demolish it. John Churton Collins once said that “The writer of a single good book is soon forgotten by his contemporaries; but the writer of a series of bad books is sure of reputation and emolument.” As sympathetic as I am to such writers of one good book as Henry Roth, Ralph Ellison, or Michael Shaara, it strikes me that a series of books, no matter how bad, is more likely to bolster a writer’s authority. For there may also be a trans-world identity between the writer’s different books; they create different worlds in which the same beliefs are held. Or a writer may derive authority from different books not her own. What Julia Kristeva called “intertextuality” is better understood as a trans-world identity between fictions by different authors. Francine Prose establishes the truth of a girl’s grief and first experience of love by showing that her world differs less from The Mill on the Floss than does any world where girls grieve and love, but not like they do in fiction.
 Originally published in the American Philosophical Quarterly 15 (1978): 37–46. Reprinted with replies to critics in Lewis’s Philosophical Papers (New York: Oxford University Press, 1983), 1: 261–80.
 John Churton Collins, “The Present Functions of Criticism,” in Ephemera Critica: Or Plain Truths about Current Literature (New York: Dutton, 1902), p. 16.