Theories of knowledge as credit for true belief, or as cognitive achievement, have to face the following objection: in the famous Barn façades case, it seems that the truth of Barney's belief that he is in front of a barn is to be explained by the correct functioning of his cognitive capacities, although we are reluctant to say that he knows he is in front of a barn. Duncan Pritchard concludes from this that a safety clause, irreducible to the conditions a true belief must satisfy in order to be credited to the believer, constitutes a necessary condition of knowledge. Because those who define knowledge as a cognitive achievement have not convincingly answered this objection, I intend, in the first section of this paper, to refute it by showing how practical and doxastic achievements differ. Then I defend the idea that we should also distinguish two kinds of environmental epistemic luck, and, correlatively, two kinds of safety, that are not equally incompatible with knowledge. Finally, I argue that this distinction is inevitably vague and I indicate some consequences of this vagueness for our knowledge‐intuitions.

We are naturally inclined to judge that it is better to know that p than to merely truly believe that p. How to account for this intuition? In this paper, I examine Williamson, Goldman and Olsson, and Pritchard’s answers, and agree with Pritchard that it cannot be consistently claimed that (a) knowledge is epistemically superior to mere true belief, and that (b) truth is the only finally valuable epistemic good. Contrary to Pritchard, I argue that the latter claim is deeply mistaken. I do so by showing that mere true beliefs have no epistemic value at all. I sketch the consequences of this point concerning the epistemic value of knowledge, and answer some objections to the thesis that mere true beliefs do not possess epistemic value.

J’expose et critique l’interprétation reçue de la controverse classique entre William Clifford et William James sur ce qu’est l’éthique de la croyance. Je défends la position de Clifford en soutenant que sa fameuse maxime selon laquelle « on a tort, partout, toujours et qui que l’on soit de croire que ce soit sur la base d’éléments de preuves insuffisants » doit être comprise comme énonçant que toute croyance qui est le produit de la corruption de notre jugement par nos désirs est fautive. J’indique ce qu’il s’ensuit à propos des croyances religieuses en particulier.

No abstract