Modal fictionalism and the imagination

Abstract

So far Gideon Rosen's intriguing 'fictionalist' analysis of modality appears to have surmounted the problems that have been directed at it, by Rosen among others (see Rosen 1993 and Noonan 1994). I want to argue, however, that further reflection on a problem that Rosen himself raises in his original paper points to a difficulty which undermines the merits of Rosen's analysis. In order to appreciate the problem which Rosen raises it is important to be clear about the relationship between Rosen's position and that of Lewis. Rosen's basic idea (1990: 332) was that one should add the primitive operator 'According to the modal realist's hypothesis of a plurality of worlds' (PW) as a 'fictional' prefix to Lewis's analysis of modality (in 1986: 5). Thus where Lewis's analysis is exemplified by