Roderick Chisholm has offered a new attempt to define knowledge in the second edition of Theory of Knowledge. The purpose of this paper is to present an objection to that definiton.Here is the proposed definition :D6.4 h is known by 5 =df h is accepted by S; h is true; and h is nondefectively evident for 5.To understand D6.4 we need to know what it is for a proposition to be nondefectively evident for a person. That has the following definition:D6.3 (...) h is nondefectively evident for S=df Either h is sertain for S, or h is evident for S and is entailed by a conjunction of propositions each having for S a basis which is not a basis for any false proposition.There are four technical expressions in D6.3: “entailed,” “certain” “evident,” and “basis.“. (shrink)