What is to follow is a brief discussion of truth in fiction. At some future date, I should like to expand it into a more substantial paper, but for now consider the following propositions:
- Sherlock Holmes does not have eleven fingers.
- Sherlock Holmes never visited the moons of Saturn.
- 221B Baker Street is not closer to Waterloo Station than to Paddington Station.
What might be the truth value of (1) – (3)? To some, (1) – (3) are, strictly speaking, true. The sentences in question may be uttered as fact in worlds near to our own, in which case, as it may happen, they would be incorrect. Indeed, for them, in order for (1) – (3) to be true in near possible worlds, too many other facts the truth of which we are certain would have to be false. Thus, though (1) – (3) would be true in those worlds, the worlds would be non-normal, in which case (1) – (3) would be false in our normal world.
However, if (1) – (3) are abbreviations of sentences prefixed with the intensional operator, “In the Sherlock Holmes stories, Conan Doyle says… ,” then they would be false, because it is not the case that Conan Doyle wrote, “Sherlock Holmes never visited the moons of Saturn,” or something similar to that effect. (Let us call such sentences f-intensionally prefixed.) Nevertheless, per the above analysis, it would be incorrect to infer from this that it might be the case that Holmes traveled to, say, Titan. More colloquially, it stands to reason that since Conan Doyle set his various scenes in what seem to be non-fictional, modally normal environs (Oxford, London, Salt Lake City, Utah, etc.), the reader ought to assume that, unless explicitly stated, one should understand the possible world of Sherlock Holmes as being similar to our own in the relevant ways, while differing in the ways identified.
This position, however, is much too complicated. Postulating near-possible worlds in order to explicate the non-vacuous truth (if any) of f-intensionally prefixed sentences is to explain the obscure with the obscure. To be more precise, even if we permit some instrumental understanding of possible worlds, viz., a formal means by which we come to an appropriate notion of validity (in logic) or truth (in epistemology), it is not at all clear why we ought to interpret possible worlds as different worlds rather than a part of this world. One may retort that possible worlds are causally disconnected from our own and thus causal processes that span space and time are unable to effect events across worlds. However, given some views, Schwarzschild wormholes may allow for travel to distant parts of our universe or, better still, different universes entirely, which may be causally disconnected from our own universe1. Nevertheless, we would still be wont to consider the possible universes as part of our world, especially considering our laws of physics would be utilized to describe the possible universes and, obviously, the processes that make travel to them possible.
I submit we ought to interpret possible worlds as a heuristic in the instrumental sense and thus give them no metaphysical weight. Having said this does not preclude the possibility of constructing, in a sense, quasi-worlds. Quasi-worlds, in this way constructed via artistic design, need not be causally disconnected, rather, they need only be self-contained according to the propositions true at that world, despite their appearing similarity or connection with our world. In nuce, a quasi-world need only be the set of propositions true at that world. Thus, in Conan Doyle’s Holmes stories, a world with prima facie similarity to our own has been created, wherein certain propositions may be either true, false, neither, or both.
The truth condition of a proposition is determined by the correspondence between the occurrence (or lack thereof) in the quasi-world(s) and the f-intensionally prefixed sentence. E.g., “Sherlock Holmes was an amateur cryptographer” is true because in The Adventure of the Dancing Men Conan Doyle wrote as much. Similarly, the proposition, “Sherlock Holmes never visited the moons of Saturn,” is neither true nor false, because it is not the case that Conan Doyle wrote concerning the matter; we simply do not have enough information to assign truth or falsity to the proposition. Indeed, it may happen that Conan Doyle may write that Holmes both did and did not go to Titan, in which case “Sherlock Holmes never visited the moons of Saturn” would be true and false2. The potential for truth value gaps and gluts in fiction is a palatable incivility because the usual problems for gluts and gaps are noticeably absent from fiction, since the conditions under which truth is assigned are rather more explicit than in non-fiction. This is to say that truth gaps and gluts may lead in our world to expressive limitations, e.g., asserting disagreement if one accepts a true contradiction, but that in quasi-worlds they do not because of the f-intensional operator.
It may be noted that a criticism of possible worlds realism may also be brought against the quasi-worlds interpretation of fiction, i.e., why should we interpret the new worlds as different worlds and not simply parts of this world? The appropriate answer may be: there is no reason. We may, if we like, interpret the new worlds as proper parts of this world. All the propositions that are true in our world are not true in the quasi-world, however, all the propositions that are true in the quasi-world are (vacuously) true in our world. I take this to follow naturally from the intuition that the quasi-world is a proper part of our world.
To conclude, in order to avoid the apparent strong metaphysical commitments certain treatments of possible worlds would have us make, it seems prudent to introduce quasi-worlds, and hence truth gaps and gluts in fiction. What is given up in bivalence is more than gained in metaphysical simplicity: we keep one real world and some vacuously true, false, gappy, and glutty propositions in quasi-worlds, while at the same time dispensing with bloated metaphysical landscapes.
1See M. Morris, K. Thorne, and U. Yurtsever, Wormholes, Time Machines, and the Weak Energy Condition, Physical Review, 61, 13, September 1988, pp. 1446 – 1449 for a discussion on this.
2For an interesting example of inconsistent fiction, see Graham Priest’s Sylvan’s Box in The Law of Non-Contradiction: New Philosophical Essays, edited by Graham Priest, J.C. Beall, and Bradley P. Armour-Garb.Published: Oxford : Clarendon Press ; New York : Oxford University Press, 2004.
P.S. See, also, David Lewis’s ‘Truth in Fiction’, reprinted in Arguing About Metaphysics, edited by Michael Rea. ‘Truth in Fiction’ serves as more or less an impetus for this paper and the position that I criticize.