Some have argued that the generality problem Feldman and Conee raise for reliabilism will plague every theory of justification. I don’t recall the precise nature of these arguments, but I wasn’t convinced. I’m worried, however, even though not convinced. Here’s why.
I’ll start by using Chisholmian internalism (conceived as a version of evidentialism) as a test case.
Let’s distinguish between propositional and doxastic justification in terms of the basing relation: you can have the former without the latter if your belief isn’t based properly. So what we want to know is when your belief is doxastically justified (since that is the epistemically non-defective state with respect to justification). For it to be so, it will have to be based properly, and proper basing, for Chisholmians, will require a true epistemic principle that links the basis of the belief with the content of the belief.
Here there’s a problem, though. Take the simple principle if you’re appeared to F-ly and have no grounds for doubt that something is F, then you are justified in believing that something is F. Relative to this principle, what does it take to properly base your belief that something is F? Should it be based on the experience itself, or must it be based on the experience-together-with-a-lack-of-grounds-for-doubt? As I’ve noted here before, the second idea is too strong: background conditions such as lacking grounds for doubt typically play no role in the holding of a belief. And if we go with the former suggestion, why is the belief not properly based when grounds for doubt are present?
It is tempting to me to say the following: isolate what is causally, or explanatorily, responsible for the belief, and then if there is a Chisholmian principle the antecedent of which includes that material, then the belief is properly based. But that is just the “accede to the scientists” response that Feldman and Conee reject for reliabilists. Their fundamental objection, as I see it, is that what is explanatorily relevant need not be justificatorily relevant. If so, however, I doubt we can use explanatory relevance as a clue for proper basing. Without some clues, I don’t see how to be sanguine that evidentialists don’t have a problem about generality when it comes to proper basing. Am I missing something?