There are at least two broad kinds of avowed atheists who take two distinct kinds of stances on the status of their belief. One is the atheist in the widest possible sense-one who claims to passively lack any belief in any gods by simply refraining from believing in them rather than outright making a metaphysical claim that "gods do not exist." The other broad kind of avowed atheist also lacks belief in all gods but is willing to say she either disbelieves in gods or believes (or even knows) based on the preponderance of evidence that there are no gods.
Dan then goes on to develop what you might call a functional test for these kinds of atheism. Essentially, if you act like these groups, so that there is no practical difference between you and them, it may make sense to call you an atheist. Or conversely, if an atheist acts like a theist even while claiming not to believe in a god, that person can perhaps be classified as an atheist. This is the heart of Dan's post, actually, and he leaves it open-ended: should we classify someone as X if their actions make them indistinguishable from other X-ers, even if the person swears up and down that she's a not-X?
Dan gives two examples to help us think about this. First, there's the problem of grief. Dan asks, "To what extent do [the theists'] normal fears of death and deep mourning of lost ones they ostensibly believe in heaven betray functional disbelief in heaven?" Basically, as I understand him, Dan's saying that if you believe in an afterlife of some kind you shouldn't be sorry that a loved one has gone on to it, or be afraid of dying yourself. The fact that the theist does feel these things proves that she isn't really convinced there's an afterlife after all, Dan suggests. (He also offers a similar analysis of covert theists, that is, people who claim to be atheists but when in distress wonder whether they should pray.
Now, I have some problems with this account. First of all, I don't agree with Dan's point that it's inconsistent to believe in an afterlife and still grieve the death of a loved one or fear your own death. Essentially I think that bodily life is a good thing cut short by an early death, and that you don't gain anything by going on to any kind of an afterlife sooner rather than later. So whether or not it's wrong to believe in heaven and hell or their various analogs in any other religion, there's not really anything inconsistent in those two beliefs. I also have my doubts about functional definitions because you can get the same result with two very different beliefs (for example, someone who thinks she should turn in a found wallet to the police and acts on that belief, versus someone who thinks she should keep the wallet but lacks the strength of character to act on that belief).
But those points aside, there's a really interesting issue in the distinction Dan draws between these two types of atheists. As I understand it - and keep in mind, this is the pre-coffee Marta, which shall be purchased and consumed on my way to campus in a bit - Dan is talking about people in both cases who have a specific concept of a god and we don't think anyone meets that definition. One type rejects the idea that any of the usual (or unusual, for that matter) candidate is in fact a god; the other makes the positive claim that "no gods exist." But in both cases, you have an idea of what would constitute a god, and you're taking a position that nothing actually meets that definition.
The trouble is, as soon as you have an idea in mind, I'd say you're already letting God exist in a sense. This is Anselm's basic point: to even talk about something (including God) meaningfully, you have to know what it is you're talking about, but to do that you have to have an object of thought. At a minimum a god "exists" inside your mind even if it doesn't exist anywhere else. The thought is that it's wrong at that point to say "There is no God" unless you're very careful to talk about the concepts as what other people mean by them rather than ideas you've internalized.
Anselm explains the distinction here through the painter analogy:
For, it is one thing for an object to be in the understanding, and another to understand that the object exists. When a painter first conceives of what he will afterwards perform, he has it in his understanding, but he does not yet understand it to be, because he has not yet performed it. But after he has made the painting, he both has it in his understanding, and he understands that it exists, because he has made it.
Anselm has just said that whenever we hear or say a word – if proper communication is occurring – we must know what we're talking about, implying that we have the idea in mind. We can disagree about whether there's anything in the world beyond our mind, sure, just like the painter can wonder whether the idea he has in his mind actually corresponds with an idea he has in mind or whether it's just an object of thought at that point; but it's an object of thought, whatever else it might be.
The way around this is that if you're going to say the concept of God is ill-formed. Basically, if you think it's a concept that involves a contradiction, so it's theoretically impossible that anything meets the definition (as opposed to it being possible but as it works out no one actually meets the definition), then you aren't communicating in the normal way. When you say "God" it doesn't describe a concept you have in your mind. Rather, you're using it in a way my adviser calls parasitically: you're taking on someone else's concept and using it on their terms, but you're not really saying it makes sense to you. So if you think (e.g.) God can't create a rock so big He couldn't lift it so nothing could be omnipotent, you're likely to say that any conception of a god who has the characteristic of omnipotence isn't even possible. Meaning you can't have Him as an object of thought, though you could still talk about other people's (faulty) conceptions of Him.
I personally think this distinction is helpful toward making sense of language. Consider the statements (a) "Legolas is the Prince of Ithilien" and (b) "William is the prince of England." (b) is objectively true since there's an individual named William and, yes, he does hold that title. (a) turns out to be false but for entirely different reasons. There is a fictional character called Legolas, and he is a prince, but he's not the prince of Ithilien. However, none of this depends on the objective reality; it's just a set of fictions developed by an author and known by fans of the book. The statement is false because if you denied any of the claims I listed above you wouldn't be describing the situation depicted in the book. But it's not like we can consult the official records of Ithilien and fail to find Legolas listed among them. There's a different standard of truth; indeed, a different concept of what it means to be true is at work. Similarly, a statement that works with a contradictory concept wouldn't be true or false in the same way either of these are; if in fact you could say true or false apply at all.
None of this is intended as a criticism of Dan's position, btw. As I said this is Marta pre-coffee (soon to be rectified), and it's possible I've misunderstood what Dan was getting at with his post. But even if I am misreading him, I do think it's useful to distinguish between the atheist who simply doesn't believe anything in the world satisfies the definition of God, versus someone who says the concept itself is nonsense. The kind of evidence we'd expect in the two cases would be completely different.
Now off to get my java-fix. Feel free to tell me what you think of all this in the comments.(Originally posted at LJ; please comment there.)