The Variability of ‘Knows’

Abstract

AbstractMuch recent debate in epistemology has been provoked by the point that whether the ascription, or denial, of knowledge of some matter to an agent is appropriate seems to depend, in a number of ways, not just on their evidence but on variable extraneous features of the context and circumstances—thus Hume’s idea that he knew full well of the existence of an external world when in the company of his drinking companions but not in the philosophical quiet of his study. This is the variability phenomenon. The chapter considers in some detail and compares each of four theoretical proposals to account for it systematically—contextualism, interest-relative invariantism, MacFarlane-style assessment-sensitivity, and regular invariantism—all of which agree that the phenomenon of variability is real but offer very different accounts of the nature of the dependencies and parameters involved. The chapter reviews their offerings in detail and makes a case that all are fundamentally unsatisfactory either failing to predict all aspects of the variability of ‘knows’ manifest in our practice or else predicting uses of ‘know’ and its cognates which seem manifestly outré and incorrect. The recommended conclusion is that our discourse involving ‘knows’ and its cognates is actually subject to no genuinely systematic pragmatics-sensitive variability of assertion conditions. The suggested moral is a form of deflationary account of the concept of knowledge—or better, of the function of ‘knows’. ‘Knowledge’ does not cut epistemic reality “at the joints” as people have come to like to say, and the idea that knowledge should come first in analytical epistemology is accordingly a mistake. Rather, knowledge—the presumed substantive referent of ‘knows’—comes nowhere.