I've had a chance to read Keith DeRose's very interesting & rich engagement with some of the experimental epistemology literature, and there's a lot in it that's clearly going to be useful to x-phi practitioners to learn from & absorb. (See some nice discussion of it already ongoing here.) But I also think that there are some ways, in a couple of places, in which Keith is subtly underestimating some of the ways in which one can conduct a different kind of investigation with survey methods than one can from the armchair. And I think it may be a mistake that is not just pretty common outside of x-phi, but also made sometimes by practitioners of x-phi themselves, so let me see if I can take a rough bloggy stab at articulating it.
I think people often take the fundamental difference between armchair and experimental methods in philosophy -- call them a-methods and x-methods (following Dave Chalmers' useful abbreviations here) -- is that both primarily use intuitions about specific cases as their data, but in the former the philosophers deploys their own intuitions, and in the latter, the philosophers deploy what they determine to be the intuitions of a sufficiently strong majority of the folk. If you're testing some epistemological theory E, and E predicts that some hypothetical case H is an instance of knowledge, then on this picture x-phi and a-phi both agree that the thing to do is to determine whether H is intuitively a case of knowledge, and the disagree only about how best to go about doing so. Let's call this the pollist picture of x-phi, in honor of Williamson's gloriously high-handed reference to experimental philosophy as "the method of conducting opinion polls among non-philosophers".
But for many epistemological debates, pollism is a mistaken way to think about a possible x-phi contribution. In particular, debates about whether some factor is or isn't partially determinative of our attributions of knowledge (or justification, etc.) do not need to be addressed in a pollist way. The standard way to argue for such a theory, either with a-phi or x-phi construed pollistly, is something like:
(i) find a case A that is indicated to be (via either a-methods or x-methods) knowledge;
(ii) find a case B that is indicated to be (via either a-methods or x-methods) not knowledge
(iii) contend that the only difference between A and B that could explain their different knowledge valence is a difference along factor F.
Recent Comments