Clifford starts by giving the analogy of a shipowner. Basically a man owns a ship some immigrants are going to use to travel across the ocean, and rather than simply investigating whether it's seaworthy he sends them off full of confidence that the ship will survive. The ship is old and he does not have good reason to assume this; he's just optimistic or (more likely) doesn't really care about their safety. Darned one-percenters and all. :-) A few weeks out the ship sinks, and the man is shocked by the "tragedy." After all, he didn't mean for those people to die; he didn't know the ship was so bad as all that.
Clifford points out that this defense is bunk. He may not have known but he should have known and would have known if he'd looked. But then Clifford asks an interesting question: should we hold the unlucky shipowner any more blameworthy than if he'd been lucky and the ship hadn't sank. Clifford thinks not. If you believe something on good evidence and it turns out you're wrong, that may be tragic and not your fault. But if you didn't actually see what was wrong with the situation because you weren't looking, or because you ignored what you saw, well, that is your fault. As Clifford said most memorably, It is wrong always, everywhere, and for anyone, to believe anything upon insufficient evidence.
As far as I can see, Clifford has two real reasons here. The first is that beliefs inspire action. He thinks it's all but impossible to truly believe something and not have that impact the way you act toward different people in some way or another. But he also points to another issue. If you get in the habit of believing things on insufficient evidence, then you lose your ability to be rational and critical. Even if you could believe something and out of luck it never caused you to act wrongly, it would still be wrong because that would build up bad habits in you.
I think Clifford's on to something here in most circumstances. Certainly if Henry Ford believed he needed to put fairy filters on his cars to keep the creatures from gumming up his cars' engines, that would be an irrational action. If he acted on that, I'd even say he was being wasteful and probably hurting someone (himself, his customers, possibly even indirectly his employees since he can't give them the benefits or wages they perhaps deserve). This is because if fairies existed or messed up car-engines, you'd expect to find evidence of this near at hand. Similarly if I ran across a field without looking where I was going, and consequently stepped in a rabbit-hole and twisted my ankle, you'd be quite right to say my belief that I didn't need to be so careful was a bad one. It would still be a bad one even if I got lucky. This is because the assumptions here - that fairies gum up engines, that a certain field is free of holes and other dangers - are easily verifiable and refutable. And in both cases my beliefs were based on insufficient evidence. The fact that I hadn't bothered to go out and gather the evidence doesn't change the fact that they were bad beliefs.
This brings us to the question of God. Clifford isn't talking about philosophy of religion in particular, but the essay is in a philosophy of religion anthology being studied for a philosophy of religion seminar. And it's obviously relevant. If religious beliefs are operating in the same way as everyday beliefs are, then I should need evidence for my claim that God exists, just like I should need it for my claim "This field is safe to run across" - if anything, I'd need more evidence. Extraordinary claims requiring extraordinary proof and all. But as we were discussing this in class I found myself thinking of Karl Barth and D.Z. Phillips and even Kierkegaard (what little I understand of him). These people not only have some kind of a "belief" in God --or if not belief, there's some kind of epistemic state concerning God they want to affirm, whether it's faith or hope or whatever-- but they're quite clear about the fact that they're not offering a hypothesis like a scientist would about the observable world. This isn't the kind of belief (if you want to call it that) that requires any kind of evidence, because the "believer" (what a misnomer that seems at times!) is doing something very different than what Clifford thinks the shipowner ought to do.
Let me put it another way. Clifford says we should not have a belief based on insufficient evidence. This is kind of like the point I raised with the Dawkins scale of a/theism the other day: many people, fideists and the like, don't have a belief where any kind of evidence would be sufficient. This implies our concept of God is the kind of thing that can be proved even in principle, and as I said at the time, to many religious "believers" this is a pale concept of God. So in these cases I'm not sure that the belief has insufficient evidence; it's simply not the kind of thing where you should expect evidence.
I'll take this a step further. If Clifford is concerned (as he seems to be) about the cumulative effect of believing something on insufficient evidence, he should also be concerned about the effect of disbelieving something on the basis of evidence. If our cocnept of God is the kind of thing where we shouldn't expect proof, then what kind of lack of evidence justifies skepticism?
My gut instinct is that the fideists are on to something here, too. By all means, demand sufficient evidence for a scientific hypothesis. But when it comes to the kind of things that don't really fit that kind of analysis --things that are beyond our ability to comprehend-- I think Clifford's approach can lead people to overstate their ability to reason.
The problem of course is that a lot of religious people do talk about their beliefs about God like they're true and false the same way that the statement "my wallet is black" or "my soda is room temperature" are true or false. And people like Clifford (and anyone else) are probably right to criticize religious people for making those kinds of statements. When someone says a statement like a particular wallet is black is true, what I usually hear is that they know it to be true - or at least that they believe it to be true in a way that would lead to knowledge if they had the necessary evidence. But the more I think about it, statements like "God exists" simply don't seem to work that way. It may be misleading to call such statements "beliefs," even. At a minimum, I'd say we need to be very, very careful not to mix them up with scientific hypotheses.
This idea is very much a "baby" concept - just something that occurred to me as I was sitting in class this afternoon. But it seems to me that Clifford's argument only really works against people treating a belief as grounded (supported by good evidence) when they don't have that kind of evidence. I'd go a step further and say it's just as dangerous (and dangerous in a similar way) to say evidence should be possible if we only looked for it, when that kind of evidence isn't really possible - it leads people to reject ideas they're not really in a position to reject, which in itself leads to bad epistemic habits. I'd say both atheists and theists approach religious belief wrong when we treat it as beliefs like scientific beliefs. At least with myself, there's something radically different going on when I call myself a "believer." Maybe we need a new name for what we call religious beliefs, or something. But blaming religious beliefs for not meeting that burden doesn't seem like the answer.
This line of thought really requires more time than I'm able to give it just now. And it may be wrong. But I at least wanted to take the first step of laying these thoughts out.