Knowledge ascriptions seem context sensitive. Yet it is widely thought that epistemic contextualism does not have a plausible semantic implementation. We aim to overcome this concern by articulating and defending an explicit contextualist semantics for 'know,' which integrates a fairly orthodox contextualist conception of knowledge as the elimination of the relevant alternatives, with a fairly orthodox ''Amherst'' semantics for A-quantification over a contextually variable domain of situations. Whatever problems epistemic contextualism might face, lack of an orthodox semantic implementation is not among them.Keywords Epistemic contextualism Á Knowledge Á Semantics Á Context Questioning is the cutting edge of knowledge; assertion is the dead weight behind the edge that gives it its driving force…. Information may be the body of knowledge, but questioning is its soul….[A]sking a question means envisaging alternatives, and only one at most of these alternatives can really exist. (Collingwood 1924, pp. 78-79) Knowledge ascriptions seem context sensitive. For instance, if Ann was wondering who stole the diamonds and then found Claire's fingerprints on the safe, then it would seem true for Ann to say:1. I know that Claire stole the diamonds J. Schaffer (&) Rutgers University, New Brunswick, NJ, USA e-mail: jonathan.schaffer@rutgers.edu Z. G. Szabó Yale University, New Haven, CT, USA e-mail: zoltan.szabo@yale.edu
123Philos Stud (2014) 168:491-543 DOI 10.1007 Yet if Ann was instead wondering what Claire stole and then found Claire's fingerprints on the safe, then it would seem false for her to say 1. After all, the fingerprint evidence bears on the question of who stole the diamonds, but not on the question of what Claire stole.Epistemic contextualism is the thesis that things are as they seem: knowledge ascriptions are context sensitive. But contextualism has faced sustained criticism on several fronts. Perhaps the main criticism is that contextualism is semantically implausible. Contextualists seldom articulate an explicit compositional semantics for 'know,' but usually just draw analogies with indexical pronouns, gradable adjectives, and/or quantificational determiners. These analogies are at best partial. And so many philosophers have grown openly skeptical of epistemic contextualism. As Stanley (2005, p. 47) writes: ''[T]he alleged context-sensitivity of knowledge ascriptions has no other parallel among the class of uncontroversial contextsensitive expressions.'' Call this the semantic problem.We aim to solve the semantic problem by articulating and defending a contextualist semantics for 'know.' Essentially, we will argue that 'know' may be treated as an A-quantifier (alongside adverbial quantifiers and modals). Our semantics integrates a fairly orthodox contextualist conception of 'know' as requiring the elimination of the relevant alternatives, with a fairly orthodox ''Amherst'' semantics for A-quantifiers as ranging over a contextually variable domain of situations. We call this a comparativist se...