WHAT DOES KNOWLEDGE-YIELDING DEDUCTION REQUIRE OF ITS PREMISES?

According to the principle of Knowledge Counter-Closure (KCC), knowledge-yielding single-premise deduction requires a known premise: if S believes q solely on the basis of deduction from p, and S knows q, then S must know p. Although prima facie plausible, widely accepted, and supported by seemingly compelling motivations, KCC has recently been challenged by cases where S arguably knows q solely on the basis of deduction from p, yet p is false (Warfield 2005; Fitelson 2010) or p is true but not known (Coffman 2008; Luzzi 2010). I explore a view that resolves this tension by abandoning KCC in the light of these challenges, and which acknowledges their force but also their limits. Adopting this view helps identify the epistemic constraints that operate on the premises of knowledge-yielding deduction, clarifies the epistemic role of deduction, and allows us to distil the kernel of truth in the motivations that are standardly taken to support KCC.