Abstract

Juan Comesaña argues that Halloween Party shows that Sosa’s (2002) disjunctive safety condition on knowledge is too strong. Mark McBride agrees, and proposes a modification to that condition in order to evade Halloween Party. I show that that Halloween Party is not a counterexample to Sosa’s disjunctive safety condition. However the condition, as well as McBride’s modification to it, is insufficient for true belief (or acceptance) to be knowledge. Sosa’s condition needs supplementing in some way that would yield a full analysis of knowledge.