There Is Nothing Magical about Possible Worlds

Abstract

Alexander Rosenberg' has posed the following trilemma to modal realism: either the theory is ungraspable, or it reduces to ersatz modal realism, or it covertly relies on modal notions and so its explanations are circular. Rosenberg is not alone in his attempt to find a chink in Lewis's armour. Recently Naylor2 and Yagisawa3 have tried to foist the embarrassing conclusion on Lewis that he must, if he wishes to be consistent, extend his realism to impossible worlds. Lycan4, paralleling the third horn of Rosenberg's trilemma, alleges that the benefits of Lewis's position, which Lycan affectionately dubs 'Mad Dog Modal Realism', are illusory since the explanations Lewis offers depend on an unanalysed notion of possibility. Devitt and Sterelny5 dismiss genuine modal realism's explanations because they are not causal explanations. Previous attacks on genuine modal realism have failed. Sharlow6 has shown that genuine modal realists do not have to embrace impossible worlds, and I have shown that Lewis requires no modal notions to explicate possible worlds.7 Rosenberg's attack makes it imperative that I defend Lewis again. His argument, if it succeeded, would kill rather than merely wound Lewis's theory. And genuine modal realism is not merely Lewis's position, I accept it as well, and so have motives beyond mere intellectual exercise. The crux of Rosenberg's argument is his claim that disconnected spatio-temporal regions are incomprehensible. He rightly points out that many philosophers past and present have thought it necessarily true that there is only one complete space and time. Also he rightly points out that a multiplicity of disconnected spaces and times is indispensible for the modal realist. I freely grant that this multiplicity is indispensible for possible worlds and at least seems to clash with familiar intuitions about space and time. The nettle must be grasped. The intelligibility of spaces and times unconnected with our own must be defended. I think the conflict between the singularity of space and time and modal realism is more apparent than real, however. Unless I misread the history of this issue, what philosophers have had in mind when they have argued that there is necessarily a single space and time has been that there could be no more than one actual space and time. And this is something genuine modal realists do not need to dispute. Genuine modal realists hold that actuality is 'indexical', so from an