Abstract

Philosophical accounts of the semantics of fiction have tended to be problematic in one of two ways: either they have denied that items used in fictional discourse have their plain meaning, introducing complications into otherwise satisfactory accounts of semantics, or they have posited special kinds of entities, introducing complications into otherwise satisfactory accounts of ontology. Accounts that tried to avoid these problems by positing mere possibilia as fictional entities were thought to be hopeless inasmuch as it was thought impossible to pick out unique mere possibilia, and there could be no causal connection to them of the sort that was thought to be necessary for reference. I proposed to try to deal with the first problem by using counterfactuals, which each pick out a relatively small, constrained set of possible worlds, and presumably could so limit the possibilia contained in them to make picking out unique ones possible. I improve on David Lewis' counterfactual semantics for fiction by altering the analysis to make the experience of a hypothetical listener central, rather than the actions of a hypothetical narrator. As for the second, I show bow the causal theory of names can plausibly be expanded to a counterfactual theory of names, which would allow for naming across possible worlds. I conclude that fictional names rigidly designate mere possibilia, which allows for a more normal semantics for fiction than had seemed possible