Jimmy Stephens argument: 

P1. If fallibilism, fallibilism is knowable.
P2. If fallibilism is knowable, there exists a reason against non-fallibilism.
P3. If fallibilism, all putative reasons against non-fallibilism beg the question against non-fallibilism.
P4. No cases of question begging are reasons.
P5. If fallibilism, fallibilism is not knowable.
C: If fallibilism, fallibilism is knowable and not knowable.

Premise two is a claim about the nature of knowledge. Its just to say that because knowledge includes a justificatory element – call it “a reason” – knowing that fallibilism is true requires reasons. Since fallibilism is a theses in competition with infallibilism, contextualism, and skepticism, the reasons involved will need to “overrule” those views, falsifying them so that we’re left with a rational preference for fallibilism over and above its alternatives. So knowing fallibilism means having reasons to doubt its alternatives.

Premise three points out that fallibilism has an entailment about all reasons, as we defined them. It entails that all knowledge involves beliefs that are fallible, correctable, open to revision, whatever, because the reason does not guarantee the belief’s truth. (Repeat with the appropriate categories for non-propositional knowledge.) So reasons for fallibilism will involve or include reasons against non-fallibilism, but the further belief *that* those reasons *are* reasons and *fallible* ones will be open to revision.

However, counting those reasons *fallible* will require a further belief that we are not already in possession of stronger (maximally strong, in fact) reasons that are infallible or that are infallible knowledge. In this way, we cannot adduce reasons against, say, infallibilism unless we assume that it is false in the first place.

To illustrate, just take the Christian claim that all men know God in virtue of His revelation. Let us suppose that this revelation entails infallibly justified beliefs about God, man, the world, and so forth. This falsifies fallibilism if true.

Now the fallibilist will have to provide reasons why this view is false. However, by counting any of his reasons open to correction, they are open to the correction that the Christian claim is true, he cannot prefer any of them unless he already assumes he does not know the Christian claim is true. Thus, he has to start by assuming fallibilism to then offer up any reasons in its favor.