This is in response to the Carrier post that precedes it. Back in graduate school, the reigning idea was that the problem of skepticism was overrated and that the enterprise of refuting the skeptic was pretty much misguided. At least that was the prevailing attitude of the "naturalistic epistemology" that was popular at that time. My epistemology teacher, Fred Schmitt, used to like to refer to foundationalism and skepticism as "The Bobbsey Twins," that foundationalism attempts to provide a response to skepticism and so takes skepticism seriously in a way that perhaps it should not. I think there are three strategies in response to the threat of skepticism. The problem of skepticism arises classically by 1) the discovery that knowledge requires justified belief 2) attempting to justify beliefs by arguments, 3)noticing that arguments themselves have premises that have to be justified and 4) realizing that you need justification for the justification, and then justification for the justification for the justification, and then justification for the justification for the justification for the justification... and I'm going to disappoint Fred for wimping out here, but I think you get my drift.
It was Descartes, with what I have been calling the Satan Test, who wanted to put all beliefs up to the severest possible tests to see if any of them were justified. But even if, as I have argued he succeed with the existence of a self that thinks, in order to avoid a pretty severe skepticism he had to dumb down the Satan Test. At least that's how I see it.
Descartes' Evil Demon has been replaced by the modern Brain in the Vat, but the problem remains the same. How do we refute the skeptic about, say, rational inference? So long as I feel required, for the sake of my epistemic life, to refute the claim that we are brains in vats, it looks awfully difficult to answer the person who asks me to prove that I am not a brain in a vat. And no, theism won't do the trick here. That's why I do not recommend attacking naturalism on the grounds that we theists can refute the skeptic but the atheist cannot. That's what I call a skeptical threat argument. Whatever I appeal to as my justification for why I am not a brain in a vat, the skeptic is going to attribute to the masterful neuroscientific understanding of the vatkeepers.
It seems to me that two strategies are effective in responding to the skeptic. One is outright dismissal, and the straightforward denial that we need to refute the skeptic. The burden of proof should not be on me for denying that we are brains in vats, the burden should be on the skeptic to prove that we are brains in vats.
But secondly, it can be pointed out that the skeptic's position is self-refuting. If someone says that there are no rational inferences, it doesn't bother me much as a mere statement, but if the skeptic is trying to argue for it, then the skeptic is contradicting himself. Moreover, as Hilary Rodham Putnam pointed out, on a plausible causal theory of reference, the words in the statement "We are brains in vats" are meaningful in virtue of the causal relations between the words and their objects, in other words, if the words "brains" and "vats" stand in an ordinary causal relationship to brains and vats. If they don't, which would have to be true if we WERE brains in vats, then the words don't mean anything. In other words, the BIV hypothesis is meaningful only if it is false.
Now it seems to me that both of these strategies is available to the advocate of AFR. The attempt to argue empirically against the skeptic, which Carrier thinks is somehow undermined by accepting AFR, seems to have serious problems as a strategy against the skeptic anyway, since whatever empricial evidence is provided can be explained in terms of the omnipotent power of either the evil demon or the vatkeepers, depending of what mythos you are using to get the skeptical argument going.