I once thought David Hume's reasons for being skeptical about scientific laws were inconsistent with his arguments against miracles. He argues that we can't know about scientific laws or causes, because all we perceive are one thing happening followed by another thing happening. We don't perceive any causing, just the things we take to be cause and effect. Our taking it to be cause and effect is thoroughly irrational, Hume says, and thus we know nothing about whether there are any causes or scientific laws. For all we know, a ball you throw into the air could come back down, as you expect it, or it could turn into a bird and fly away. We expect it to do the former, but there's no reason we have to think it can't do the latter.
Hume goes on to say that we should never believe in miracles, because you should always proportion your belief to the evidence, and there is zero evidence for miracles. He rules out the very possibility of miracles, it seems, and he does this in the very same work where he has spent so much time setting up worries about whether our entire scientific understanding of the world might be wrong, leaving us with the result that, for all we know, basketballs might turn into seagulls and fly away. How can he consistently say both of these things?
But then I read Hume more closely in subsequent readings, and I came to the conclusion that Hume's approach is consistent after all. What he says in his skepticism about science is that we don't know there are scientific laws of the sort that we believe in if we think one thing makes another happen. He also says that, for all we know, unexpected things that would seem to violate the laws of physics that we believe in could be possible. But he does go on to give a pragmatist account of why we might as well believe in scientific laws anyway, since it's served us well so far, and it's not as if we can help it anyway. It's also not as if we have a choice.
But then in the miracles chapter, he gives a careful argument. He first defines probability as how often something happens in our own personal experience. Then he says that, if you haven't experienced miracles, it follows that miracles have zero probability. But why, then, could he say that plants could sprout legs and start walking around, as far as we know? Isn't that like a miracle? But he's careful here. If we believe that a plant did such a thing, we'd be believing in a miracle. We shouldn't do that, because it has zero probability. It's never happened, in my experience, so I should think it has zero probability. At the same time, I can't rule it out. So it's not impossible, as far as I know. If I did witness it, I'd have to proportion my beliefs with the evidence I then had. But as it is, I shouldn't believe in such things. I should just believe in their possibility, but I shouldn't allow for anything more than zero probability.
The key here is in defining probability in terms of how often it's happened in your experience, while defining possibility in terms of whether it's consistent with your experience. Something could then have zero probability but be well within the realm of possibility. So, because of that, I came to think that Hume's view was indeed consistent, even if it's a strange set of views.
But now I've become convinced again that there's a deep inconsistency in Hume's approach to these two issues. It has to do with his willingness to extend pragmatist arguments toward functioning the way we ordinarily do with respect to the scientific skepticism he begins with, while not extending pragmatism toward functioning the way we ordinarily do with the issue of miracles. He accepts our ordinary views on scientific laws, even though he insists that such beliefs are irrational and not grounded in anything more likely to produce true beliefs than crystal-ball gazing, at least as far as we can be sure. He relies on the testimony of other people in order to believe in regularities in nature that he can rely on to live his life. He refuses to accept the testimony of other people when it comes to miracles, however.
He does have a reason why he treats these two areas differently. He says that he has witnessed regularities in nature himself, and he relies on other people's testimony that coheres with his own observations but extends them. He has not witnessed miracles. He has witnessed people being dishonest or gullible, and so he has higher than zero probability of even honest people lying or even intellectually careful people being deceived, yet zero probability of miracles occurring. The higher probability, even if it's very low, is still higher, and so he should believe any possible explanation that's above zero even if low over the zero probability of miracles.
Will this work, though? I'm not convinced. Why does he give miracles a zero probability? Purely because he hasn't witnessed any himself. That's his criterion for belief. But he hasn't witnessed any of the scientific research that he relies on to accept scientific laws. He hasn't witnessed any of the events history tells of that he's willing to believe in. He relies on the testimony of other people about all manner of things, except miracles. He says the difference is between a kind of events he hasn't personally witnessed and a kind of events he has. But many of the events he hasn't witnessed are of a sort he hasn't witnessed and rely on expertise and specialist knowledge that he would have no access to. He's willing to be a pragmatist in accepting those beliefs even though he knows none of it. He rules out even the possibility of pragmatically accepting the testimony of other people whose specialized experience would support miracles. I'm not sure his pragmatism about many other sorts of beliefs would hold up if he refused to extend his pragmatism to those areas the way he does with miracles. I think his whole pragmatist belief system would fall apart. He's willing to relax his tighter, skeptical approach and go pragmatist with a number of other areas, and doing so with miracles would leave him not insisting on believing in them only if he's ever experienced a miracle himself. The same standards he applies to history, science, and many other areas of belief would leave him without his insistence on zero probability if he hasn't experienced something, and that would end his argument for ruling out miracles from the outset. He would then have to consider testimony about miracles as having at least some positive value in his pragmatic belief system.
None of this is to say that he would end up believing in miracles. He might not. But he wouldn't be ruling them out without consideration. He would be giving miracle reports some level of credence, even if he might ultimately not decide they are credible enough for his pragmatist acceptance to kick in. As his argument actually goes, however, he does seem to me to be treating miracle reports and other kind of specialized experiences differently, and that leads me to conclude that perhaps I was initially right that he is inconsistent, even if my original diagnosis of that inconsistency didn't locate it properly.