Here is one way to run the New Evil Demon argument against externalist theories of justification. Let S be a normal person, with a justified belief that p, and S* her brain-in-vat twin.

1. If S’s belief that p is justified, then S*’s belief that p is justified
2. If S*’s belief that p is justified, then externalism is false.
C. So externalism is false.

Now how might we motivate premise 1? One way is by something like the following argument.

1. The same reactive attitudes are appropriate towards S and S*’s doxastic states.
2. Whether a belief is justified supervenes on the reactive attitudes that are appropriate towards it.
C. So if S’s belief that p is justified, then S*’s belief that p is justified

‘Reactive attitudes’ is a fairly broad term here. But not so broad as to be trivial. I take it that the anti-deontological conception of epistemology, tracing to important work by William Alston, basically thinks that this argument fails at step 2, although step 1 is false. One of the points I wanted to make in my “epistemic deontology paper”:http://brian.weatherson.org/ddd.pdf is that really we should be challenging it at step 1. In particular, I think that to the extent that this kind of reasoning is popular, it is because of too simple a picture of doxastic attitudes. (And I think the is popular, both among internalists who endorse all the steps, and externalists who think it works if we have a deontological conception of justification.)

To see the problem with the argument, let’s think about a simpler case where we have clear reactive attitude, namely sports. We’ll do this with some YouTube help.

Let’s consider three great plays from different sports. First Shane Warne’s “ball of the century” to dismiss Mike Gatting (it’s the first delivery discussed on this video).

Second Diego Maradona’s “goal of the century” against England.

And finally Willie Mays’ great catch in the 1954 World Series.

Now consider a nearby possible world where there are three similar people, Warne’, Maradona’ and Mays’. None of them quite make the plays that their actual counterparts make, although they are intrinsic duplicates of those counterparts. But in the other world, Gatting changes his shot at the very last second and gets a little bat on the ball; the trailing defender makes a great lunge to stop Maradona’s goal going on; and a crazy fan throws a beer bottle that deflects the ball at the last instant away from Mays’ glove. In such a world, Warne’, Maradona’ and Mays’ are not praised anywhere nearly as much as Warne, Maradona and Gatting are praised.

Of course we don’t just praise Warne, Maradona and Mays for their results. People take wickets, score goals and make catches all the time without getting that kind of praise. Rather, what we are praising is a great performance that led to a great result. Since the counterparts did not have the same result, they don’t get praised as heavily.

What makes the argument I started with look more plausible than it actually is is, I think, the thought that a person and their BIV twin are equally blameworthy for anything they do. I’m not sure that’s right, but it is plausible. But there’s a praise/blame asymmetry here. It isn’t plausible that praiseworthiness supervenes on intrinsic features.

I think the same holds true for doxastic cases. Imagine a murder case where there are two prima facie reasonable responses to the evidence. One detective thinks that the doctor did it, the other that the lawyer did it. Both beliefs can be supported by the evidence. It seems to me that the detective who gets it right is praiseworthy in a way that the other detective is not. They aren’t praiseworthy if they are just guessing. But one is praiseworthy for getting the right result the right way.

Now consider that detective’s BIV twin. She can’t be praiseworthy for getting the right result, since by hypothesis her beliefs are systematically false. So to the extent that praiseworthiness is success-dependent, and I think the sporting cases should already make us think that it is, we shouldn’t praise the BIV. So it isn’t true that in general we should have the same reactive attitudes towards the BIV and the person who gets it right.

Of course this doesn’t immediately show that the evil demon argument fails. All it directly does is undermine one argument for one premise. But I think once we see that some of our reactive attitudes are tied to success, the new evil demon argument looks weaker than it might do.