Modal fictionalism -- a simple dilemma

Abstract

Gideon Rosen [5] proposed an account of modal discourse modal fictionalism which promises to secure all the advantages that go with translating it into possible world discourse without having to foot the ontological bill. Where A is a sentence whose principal operator is the usual necessity or possibility operator ('O' or '0'), the possible worlds theorist discerns quantification over worlds and paraphrases it as '(Vw)A*(w)' or '(3w)A*(w)' where A* is just A expanded to include an argument place for a world term. The modal fictionalist proposes to neutralize the ontological commitment to worlds carried by such sentences by prefixing them with a fiction operator, written 'According to PW' where 'PW' denotes some suitable version of possible worlds theory, such as that of David Lewis [2]. The intended effect is that the non-factivity of this prefix ensures that we can assert and believe the resulting statement without incurring a commitment to worlds, just as prefixing 'Holmes habitually took opium' by 'In Sir Arthur Conan Doyle's stories' yields a sentence which we can assert and believe without committing ourselves to the existence of a real individual named 'Holmes'. Rosen [6], moved by a rather subtle objection (also advanced by Stuart Brock [1]), has since expressed doubts about the viability of this account. Peter Menzies and Philip Pettit [4] have argued that the Rosen-Brock objection tells only against the letter, not the spirit, of Rosen's original proposal. Harold Noonan [3] has argued, correctly as far as I can see, that the objection does not even tell against the letter. The purpose of this note is not to prolong this intriguing debate, nor to take up the clearly important question whether modal fictionalism really can secure all, or at least the main explanatory advantages to which modal realists lay claim,1 but to air a much simpler difficulty for modal fictionalism.