Pritchard on Hinge Propositions

On p. 226 of Epistemic Luck Duncan defines a hinge proposition thusly:

"P is a hinge proposition iff P is as certain as anything in the context."

I take it that this means that it is not the case that there is some proposition Q such that Q is more certain than P (rather than for any proposition Q P is more certain than Q).

Then on p. 242 he says, speaking of such hinge propositions as that I have hands:

"one cannot offer grounds in support of one's belief in a hinge proposition--because these grounds will be more questionable than the belief that they are meant to be grounds for."

Given the definition of a hinge proposition (hereafter HP) there is technically the option that they are equally questionable, but presumably that wouldn't help.

Assuming that the middle term of demonstration here is the essence of HP's it seems Duncan is assuming the following sort of principle (which I think Rich pointed out in discussion).

(*) G can ground P only if G is more certain than P.

Depending on what certainty is, I think (*) is quite questionable and will be rejected by many traditional internalist foundationalists. The reason is that certainty seems to attach only to propositions whereas traditional foundationalists will allow for entirely non-doxastic grounds. ("Entirely" because A. the grounds are not themselves doxastic states, B. nor are they they the conceptualized objects of a conscious doxastic state. I.e. they can be an experience as if P where the host does not hold the belief I have an experience as if P.)

Furthermore, the statement quoted above seems false apart from reliance on (*). Consider:

Rod: I have hands.
Skip: What are your grounds for that belief?
Rod: I see that I have hands.

Rod's suggested grounds don't seem more questionable that his original assertion. But maybe this smacks of McDowell account Duncan argues against in this chapter. So consider this variant.

Rod: I have hands.
Skip: What are your grounds for that belief?
Rod: I have an experience as if I have hands.

Again, Rod's second statement does *not* seem "more questionable than the belief [it is] meant to ground." Its role of justifier relies on the truth of an epistemic principle connecting such experiences with justified beliefs with the same content, but *that* surely isn't the kind of problematicity that Duncan has in mind.

To demand that a subject be able to justify the rules which justify his beliefs would be to walk right into the trap that the Tortoise lays for Achilles in Carroll's dialog (Carroll's Paradox).