Truth in Fiction,Impossible Worlds,and Belief Revision

We present a theory of truth in fiction that improves on Lewis's [1978] ‘Analysis 2’ in two ways. First, we expand Lewis's possible worlds apparatus by adding non-normal or impossible worlds. Second, we model truth in fiction as (make-believed) belief revision via ideas from dynamic epistemic logic. We explain the major objections raised against Lewis's original view and show that our theory overcomes them.     

The Genesis of Possible Worlds Semantics

This article traces the development of possible worlds semantics through the work of: Wittgenstein, 1913–1921; Feys, 1924; McKinsey, 1945; Carnap, 1945–1947; McKinsey, Tarski and Jónsson, 1947–1952; von Wright, 1951; Becker, 1952; Prior, 1953–1954; Montague, 1955; Meredith and Prior, 1956; Geach, 1960; Smiley, 1955–1957; Kanger, 1957; Hintikka, 1957; Guillaume, 1958; Binkley, 1958; Bayart, 1958–1959; Drake, 1959–1961; Kripke, 1958–1965.     

The Real Foundation of Fictional Worlds

I argue that judgments of what is ‘true in a fiction’ presuppose the Reality Assumption: the assumption that everything that is (really) true is fictionally the case, unless excluded by the work. By contrast with the more familiar Reality Principle, the Reality Assumption is not a rule for inferring implied content from what is explicit. Instead, it provides an array of real-world truths that can be used in such inferences. I claim that the Reality Assumption is essential to our ability to understand stories, drawing on a range of empirical evidence that demonstrates our reliance on it in narrative comprehension. However, the Reality Assumption has several unintuitive consequences, not least that what is fictionally the case includes countless facts that neither authors nor readers could (or should) ever consider. I argue that such consequences provide no reason to reject the Reality Assumption. I conclude that we should take fictions, like non-fictions, to be about the real world.