I defend the following three claims:

That knowledge is nonaccidentally-true belief, That the justification condition in J.T.B theories of knowledge essentially equates with the cogniser's use of truth-conducive procedures; and That justification (insofar as it is equated with generally truth-conducive procedures) is necessary for knowledge

The main body of this work is divided into two parts. In part I, I defend the first of these claims.

The second claim is defended in both part I and part IL The third claim is defended in part II.

My analysis of knowledge as nonaccidentally-true belief provides an externalist account of knowledge. It is developed by examining cases of true belief and justified true belief which do not constitute knowledge with the purpose of understanding what such cases lack that successful knowledge claims do not lack. I contend that these unsuccessful claims do not succeed because they are accidentally-true beliefs, whereas successful claims to knowledge are successful because they are nonaccidentally-true beliefs and that is what knowledge actually is. I go on to explicate the notion of nonaccidentality in terms of the notion of truth-conducivity. In particular, I appeal to the cogniser's use of generally truth-conducive procedures to secure the true belief. However, I also add the proviso that there can be no special circumstances obtaining such as to make these same procedures nontruth-conducive under these conditions. Claim 2 constitutes my reply to the claim that there is no commonality to be found in those various theories of justification which philosophers have put forward to meet the justification condition in J.T.B theories of knowledge. I argue that such theories (internalist and externalist alike) aim to provide accounts of generally truth-conducive procedures. Claim 3 constitutes my reply to the claim that justification is not necessary for knowledge. I argue that insofar as justification is equated with the cogniser's use of generally truth-conducive procedures (claim 2 above), justification is necessary for knowledge. That it is necessary, is a claim that follows from my analysis of knowledge as nonaccidentally-true belief.