“But what should we do with a fiction that is not blatantly false, but false only because an author has been forgetful”? (truth and fiction, 46) In other words, why can’t we ascribe the same subject-predicate form as parallel descriptions of real-life characters; such that it would be false to say “Holmes wears a silk top hat” in the same way that it would be false to say “Nixon wears a silk top hat.”

As it stands now, several difficulties arise in doing so:

• In some ways, Holmes is a real-life person of flesh and blood as is Nixon.
• How many members are there in Sir Joseph’s chorus, five? fifty?
• What restricts ranges of quantification.
• Finally, why are fictional characters immune to inferred consequences?

. . . the Meinongian must tell us why truths about fictional characters are cut off, sometimes though not always, from the consequences they ought to imply (37).

The answer to this question, argues Lewis is that we must first abstract away from the original to several of the possible revisions that remain closest to the original.

Figure 1: Closest Worlds

Thest Lets say that at $$w_{1}$$, “Baker Street” is closer to “Paddington” than it is to “Waterloo.” Also at $$w_{1}$$ Sherlock Holmes has $$10^{5}$$ number of hairs on his head. While we might think that in the actual world $$w_{0}$$, it is also true that “Baker Street” is closer to “Paddington” than it is to “Waterloo,” but this gives us no reason to think that Sherlock Holmes has $$10^{5}$$ number of hairs on his head. Therefore there are a number of truths $$\psi^{1-n}$$ which hold at $$w_{1}$$.

Figure 2: Relations

The above is wrong because Sherlock Holmes does not appear to have $$10_{5}$$ number of hairs on his head. What this unfortunate sequence looks like is: