Here’s a little paradox. Suppose we propose to meet for a drink at The Passenger after work, and I tell you that it’s on 7th just above Massachusetts Ave in Northwest. Perhaps being acquainted with my spotty geographic instincts, you ask if I’m certain. And of course I am, I go there after work often, and just to humor you I’ll double check it on my phone—I’m absolutely certain. Or perhaps, out of an excess of caution, I say that at any rate I’m about as sure as I can be about anything—call it 99 percent sure. Alternatively, we can imagine it’s been a while, and I’m really only 70 percent sure—but at any rate, sure enough to assert it with some confidence while allowing that I might, of course, be mistaken.
After a few drinks at the bar—because really, when else would this happen?—you bring up that old Philosophy 101 chestnut: How can I be sure I’m not a brain in a vat, inhabiting an elaborate Matrix-style simulated world? That I haven’t always been? Even that I’m not a simulated intelligence brought into existence mere moments ago, with a suite of false memories pre-installed, including the memory of saying earlier that I was sure about the bar’s location, and all the supporting memories of having been here before?
I’d have to allow that I can’t rule out the possibility. Not only that, I can’t even meaningfully tell you how confident I am—90 percent? 50? 10?—that none of these is the case. The way the question is framed, nothing in my experience could really count as evidence either way. I don’t, in practice worry about these things—I take it pretty much for granted that I have a certain real history that’s taken place a real external world. But if you force me to focus on the question, to bring this background assumption into the foreground by framing it explicitly, I have to admit it can’t itself be justified. My attitude toward it—for the purposes of our discussion, if not in everyday life—must be one of radical uncertainty. It’s not just that I can’t be certain, it’s that I can’t meaningfully assign any particular level of confidence to the belief. It is—I might say if I wanted to be a bit Fancy Town about it—intrinsically unknowable.
Aha! you say. How can I be absolutely certain The Passenger is on 7th Street—or 99 percent sure, or even 60 percent—if I can’t even be certain I or the bar or 7th Street or my memories of them are real at all? Doesn’t that radical uncertainty affect every belief I have about the world? Mustn’t I be radically uncertain, too, about where the bar is?
It’s a cute dorm room puzzle, but the answer is that of course I need not be, because these are questions at very different levels. When I say I am confident about the bar’s location, I’m not talking metaphysics. My assignments of confidence to beliefs are, we might say, local—they’re internal to a system of reasoning and other beliefs that collectively are the grounds for asserting or denying any particular proposition. The brain-in-a-vat question, and variants where my capacity for making or identifying logically valid inferences has been manipulated, are non-local. To be sure, we can imagine things that would count as evidence that the brain-in-a-vat hypothesis is true after all. Maybe some stylish superpeople in mirrored shades and leather trenchcoats release me from my vat. But of course, the question could still be posed—perhaps with more practical urgency!—in the world into which I’d been released. Until then, at any rate, I at least have no internal reason to think I need to take this abstract possibility very seriously. So it would seem odd to declare myself “agnostic” when it comes to garden variety propositions about where bars are located or who won the latest World Cup match.
With that in mind, it should be clear what’s wrong with this Slate essay by Ron Rosenbaum clucking its tongue at the so-called New Atheists and calling for a more humble New Agnosticism:
Faith-based atheism? Yes, alas. Atheists display a credulous and childlike faith, worship a certainty as yet unsupported by evidence—the certainty that they can or will be able to explain how and why the universe came into existence. (And some of them can behave as intolerantly to heretics who deviate from their unproven orthodoxy as the most unbending religious Inquisitor.)
There are a couple claims at issue here, and throughout the piece. One is just the commonplace observation that Richard Dawkins and Sam Harris &c. can come across as arrogant jerks, which is fair enough, but then, who else is going to really proselytize for the absence of a belief? It’s like starting a non-chess-players club; plenty of people fit the membership requirements, but only those with an active hostility to the game are going to feel the need to make a point of joining. In any event, this is at most an observation about a particular group of people; it doesn’t have much to do with the soundness of an atheist position as such.
The real argument is a variant on the familiar point that, of course, you can’t prove with mathematical certainty that there isn’t a God, supposedly making atheism (at least) as unsupportable a position as theism. But then, are there many self-identified atheists who would really pretend to such apodictic certainty? Believers sometimes claim proudly to have access to some kind of special revelation that obliterates the possibility of error. But when I say that I think there is no God, I don’t mean anything so grandiose. I mean just that I see no good reason to think that there is, and that all the various stories told about deities appear to me equally likely to be mythical. I don’t believe in basilisks or psychic powers either—probably neither do most religious believers—but few of us, on reflection, would be so bold as to say this is a belief we are absolutely certain about. It’s possible we could be mistaken, even if the possibility seems too remote to bother much about or, indeed, take all that seriously.
Presumably, Rosenbaum gets this much. We do not declare ourselves basilisk agnostics just because we have to admit we could possibly be mistaken. We just say, without qualification, that we don’t believe in basilisks, with the implicit understanding that, of course, one might always be wrong. If that’s all “agnosticism” added, it would indeed be little more than what Rosenbaum calls “polite atheism”; an agnostic here would for practical purposes be an operational atheist who makes a point of saying “but I may be wrong” a little more frequently, or demurely refrains from asserting his secular background worldview too explicitly.
So Rosenbaum’s looking for a metaphysical skyhook that will let him elevate that mere polite atheism to the brain-in-a-vat realm of radical uncertainty. His preferred candidate is another old philosophers’ quandary, and another non-local question: Why is there something instead of nothing?
I’m inclined to say that the question is meaningless—it has the form of a meaningful, even a scientific question, but it can always be framed in a way that places it outside any system of causal explanation. It’s a kind of grammatical misfire, like “This sentence (or proposition) is false.”
To the extent that it is a meaningful question, I have no reason to expect that science either eventually will, or even in principle could answer it. But I am not sure why I am supposed to care, except insofar as it’s interesting to mull over, if you go for that sort of thing. Suppose I allow that it is a genuine mystery—radically uncertain, even. It’s outside the realm about which we can talk meaningfully or offer evidence. So what? If there were some part of the world about which we couldn’t even in principle gather information, would I have to declare myself a basilisk agnostic because, after all, they might be there?
Rosenbaum’s mistake is to suppose that atheists are committed to providing some kind of utterly comprehensive worldview that explains everything in the way religious doctrine sometimes purports to. But why? Can’t we point out that claims made on behalf of one brand of snake oil are outlandish and unsupportable without peddling an even more wondrous tonic?
I don’t know why there’s something instead of nothing, if the question is even intelligible, any more than I can prove I’m not a brain in a vat. These are interesting facts to reflect on in an epistemology seminar. They have very little to do with my ordinary assertions about how to get to The Passenger or whether the details of any particular cosmology seem persuasive, or whether praying to Mecca or confessing to a priest seems like a sensible thing to do. The question of whether there’s a God is only really interesting or a live debate in practice because its embedded in these more particular traditions. Punting to the non-local question of why there’s anything at all is, ultimately, just changing the subject—a fact that may be obscured by gesturing at the realm of mystery and calling the question mark that lives there God.
I see no good reason to think that there are basilisks, or Olympian gods, or even that rather minimalist watchmaker God more often encountered in philosophical treatises than any actual, practiced religious tradition. The existence of dark spots on the physical, metaphysical, or epistemological map is no evidence for any of them. So Rosenbaum’s challenge—explain, atheist, why there is something instead of nothing!—may well be unanswerable, but it doesn’t require an answer. There’s still no reason to treat God talk as anything more than another bit of human storytelling, and no reason to add elaborate hedges to the assertion that The Passenger is on 7th north of Mass.