The End of the Armchair for Conceptual Analysis?

While conceptual analysis is traditionally seen as a prototypical armchair, or a priori, project, this view has come under fire from a number of directions. Externalists have claimed that important facts about natural kind terms (among others) are a posteriori (Kripke, 1972). Others have argued that conceptual analysis must be informed by empirical findings on how concepts are structured in our minds (Ramsey, 1998, Stich & Weinberg, 2001). And experimental philosophers have tried to show that conceptual analysis should be based in part on data gathered from the folk, which of course must be obtained outside of the armchair (e.g. Knobe, 2003). Each of these, however, has been disputed by defenders of traditional armchair conceptual analysis (e.g. Fumerton, 1999, Wheatherson, 2003, Kauppinen, 2007), and none has been entirely successful in convincing conceptual analysts to abandon their armchairs. This is because each relies on premises that those staunchly committed to the a priori nature of conceptual analysis will find controversial. In this paper, I provide a new line of argument for the conclusion that conceptual analysis is not an armchair project. My argument does not start from disputed claims about language, concepts, or our minds, but rather from uncontroversial facts about the ordinary practice of conceptual analysis. I argue that these facts imply that conceptual analysis is not purely a priori. Instead, the ordinary, traditional practice of conceptual analysis requires holding empirical views. What’s more, there is a specific set of empirical claims endorsed (possibly tacitly) by mainstream conceptual analysts, which I will argue are unwarranted and dubious.