Is there room for armchair theorizing in epistemology?

Quine argued, of course, for the rejection of the very notion of the a priori,¹ and in “Epistemology Naturalized,” in 1969, he argued for a general approach to epistemological questions which would be straightforwardly empirical.² Precisely what that program might be, however, was very much in dispute, both among those who wished to reject it and those who were eager to sign on. In the 1980s and 1990s, largely under the influence of Alvin Goldman,³ the possibility of a thoroughly empirical approach to epistemology once again came to the fore, and the question of whether epistemological theorizing might profitably be approached a priori, and what, if anything, the empirical sciences might contribute to epistemology, was much discussed.⁴ Those who wished to defend traditional approaches to epistemological questions typically acknowledged the relevance of empirical results to certain questions about knowledge, but then went on to minimize the importance of this fact. Richard Feldman, for example, is typical here. He comments,

Obviously, empirical work is relevant to “the study of human knowledge.” But this shows its relevance to epistemology only if epistemology is as broad as the study of human knowledge. The complete study of human knowledge would, presumably, include historical studies of what people knew when; studies in neuroscience concerning the ways the brain processes information, sociological studies about the ways knowledge is transmitted in societies, and so on. While some philosophers may think that they have something to say from their armchairs about many of these topics, no sensible person could think that all such inquiries can succeed without scientific input. So it is hard to imagine any disagreement with the view that methodological naturalism is true given such a broad interpretation of what counts as epistemology.⁵