Is Lewis' Modal Realism modal realism?

Ross Cameron, University of St Andrews (9 Oct 2004)

Second SPPA Seminar Day, University of Glasgow

In this paper I discuss and develop the modal irrelevance objection to Lewis' modal realism. The objection, put forward by van Inwagen and Michael Jubien, says that Lewis' modal realism is unjustified so long as it leaves unanswered the question 'what do Lewisian worlds have to do with modality?'

I argue that the modal irrelevance objection is one that Lewis does indeed have to answer and which, as a matter of necessity, he cannot answer; and as such, Lewisian realism cannot be justified. There is an obvious parallel between the modal irrelevance objection and Moore's open question objection to naturalism, and many would view that as a reason to disregard the modal irrelevance objection. I offer a way of understanding these objections which avoids the usual rejoinders to them; I understand them as methodological challenges - as attempting to challenge the justification for a theory rather than attempting to expose some internal adequacy of the theory.

So understood, Moore is challenging the naturalists basis for believing their theory, and van Inwagen and Jubien are challenging Lewis' basis for believing his. As I develop it, the objection is this: Lewis asks us to believe in modal realism because of the benefits the theory has if it is true. One of those benefits is that it offers a reduction of modal notions to non-modal notions. This is only a benefit if Lewis' theory is a realist theory of modality (in a sense to be explained): for otherwise the theory does not give a reduction of the modal to the non-modal but instead eliminates the modal in favour of the non-modal. But one will only believe that Lewisian realism is a realist theory of modality if one believes that there is the right sort of connection (again: to be explained) between modal facts and facts about Lewisian worlds. And, I argue, one will only believe this if one is a Lewisian realist.

The crux of the argument then is that there is a vicious methodological circularity in Lewis' arguments for his modal realism. Lewisian realism is only warranted if it affords the benefits Lewis claims of it. But one only has warrant to believe that it does afford such benefits if one has warrant to believe in modal realism. Hence, one must have a warranted belief in modal realism in order to acquire a warranted belief in modal realism; and hence warrant for modal belief is unobtainable.

Crucial to the argument is the claim that Lewisian realism is only a realist position on modality if one is a modal realist. I attempt to establish this by adopting a variant of the definition of realism given by Michael Devitt. Devitt defines realism about the Fs as the doctrine that Fs exist and are mind-independent. But this makes anti-realism about mental entities trivially true. I give an alternative characterisation: realism about the Fs is the doctrine that the Fs exist and that they do not constitutively depend on the wrong kind on thing. Lewisian realism is modal realism, then, iff facts about worlds are the right type of facts for modal facts to depend upon. This is what the modal irrelevance objection questions, and what I argue only the Lewisian realist should accept.

To return to the page from which you came, use your browser's "Back" function

[PROCEEDINGS OF THE SPPA]