In Gettier's example, Smith is justified in accepting [Jones is the man who will get the job and Jones has ten coins in his pocket] despite the fact that (1) is false. It has been argued, however, that if a person is justified in accepting a proposition h, then h is true. If this contention is correct, Gettier's examples would be effectively short-circuited. But the contention is dubious.
The suggestion is due to Robert Almeder, and here's why his contention is thought to be dubious:
First, people have what would normally be regarded as excellent evidence for propositions that turn out to be false. For instance, at one time sme people had good evidence for believing the Ptolemaic account of the universe and so had excellent evidence for the false proposition that the planets travel in circular orbits. The kind of justification involved in such a case is what we typically call inductive. One feature of such cases is that the evidence that constitutes the justification is compatible with the denial of the proposition justified ... If the content under question where correct, then it appears that many inductive justifications would simply not qualify as justifications at all, and this surely flies in the face of accepted views about justification
Is this really a good objection? It assumes that the properties necessary for S's X-ing to be justified are properties necessary for S to have a justification. Thus, if S's X-ing can be justified only if S's X-ing has F, S can have a justification only if S's justification has F. The context suggests that S's justification can only be S's evidence. That assumption is dubious. If we think of a justification as that which someone would cite in explaining or defending themselves from the charge of wrongdoing, that their actions were permissible will not be part of the justification itself. It is, however, a condition necessary for being justified.
Suppose I know, on the basis of good inductive evidence, that the sun is going to come out on Saturday. Now imagine someone with the same evidence that believes this mistakenly. If we say that the second subject's belief isn't justified, it doesn't seem to follow that we had different justifications for our beliefs. If, however, we said that truth were a condition necessary for having a justified belief, it would follow that in spite of our having the same justifications, only on of our beliefs would be justified.
Here's the second objection:
It is easy to imagine two cases which are identical in all respects, i.e., same evidence and same circumstances, expect that the proposition justified is true in one case and false in the other. For instance, we might suppose another example just like Gettier's in which (1) is true, rather than false. If the contention in question were correct, we should have to say in the one case that the belief is justified, but in the other that it is not. What explanation could there be for this difference? Surely not simply that the proposition justified is true in the one case, but false in the other!
Fair enough, that's not an explanation. But, here's an explanation. The belief in the one case conforms to the norms governing belief, but the other does not. Or, the belief in the one case gives us evidence that we might use to justify further beliefs, but the other does not.
Whatever the merits of such a response, it's sort of surprising that it's not considered. It's sort of surprising to see how transparent the evidentialist assumptions are. It's more than just sort of surprising to see on the next page this passage:
The following principle does seem right ... If h is justified for S on the basis of evidence, e, then no elements of e that are essential for that justification are false.
Right, so if we put all of this together, we get this. If some proposition is false, it is not included in someone's evidence. If, however, some proposition is justifiably believed, it can be false. So, there are justified beliefs in propositions such that those propositions are not part of our evidence.
I don't like that last bit. I thought that if you were sympathetic to internalism facts about our evidence are supposed to be accessible from the armchair. But, unless you're just opting for the view that propositions or facts are not the right sort of beasts to be pieces of evidence, I can't see how you can consistently maintain both:
NFE: No false propositions are part of our evidence.
JFB: The falsity of the belief that e is true is not itself going to prevent the belief that e is true from being justified.