The Case for Atheism

[frasierc]

Thanks for engaging with my responses, your responses are really interesting and helpful. I'll try and respond briefly to your responses on interpretation of evidence. As I said before - discussing the evidence when we don't agree on what evidence is admissible nor how to interpret it is unlikely to be productive. I think this is a generalizable concept on all questions - regardless of the topic.

Well, of course. But beliefs that change in accordance with new evidence aren't assumptions, they're just... beliefs. It's impossible to come to any claim as a totally blank slate, but the idea is to accept new claims if they come with a certain degree of evidence and support, and they don't hinge on logical fallacies to function. It's unfortunate for the theistic side, but every one of the arguments they make, including yours, are either fallacious or not supported to the extent that they should be believed.

Again, this isn't a presumption of naturalism, however. If you could show me a repeatable miracle with a reliable result and I could rule out cheating through natural means, then I'd give equal chances to supernatural origins as anything else. But I would be comfortable in just saying I don't know what caused it until we actually found a cause, too.

This is slightly jumping the gun I think we get to the discussion of how presuppositions relate to evidence further down.

You argued in your previous post you base your conclusions purely on the data. From this response you're now saying you have some assumptions but you're willing to revise these in the light of the evidence.

This is reasonable and I would say also reflects my position. The question which I think we will come to below is how informed your prior is (by which I mean how much more likely do you think your current position about the explanation of the world is than theism). That will determine how much your interpretations are impacted by your prior beliefs and how willing you would be to revise this belief.

What is a flat prior (all weight in the analysis of evidence given to the data and no or minimal prior assumptions) in the context of our discussion? If you want to make no prior assumptions then mathematically you express this by parametising metaphysical naturalism and theism as equally probably explanations of the world. Whether you hold that assumption to be true or not, if you want your interpretation of the evidence to be determined by the data mathematically you have to make that assumption.

I agree. It's just too bad for your side that the more evidence we gather that requires no gods, often to the contrary of what various religious books claim, the probability of a natural universe increases.

Again, I think this jumps the gun - we're still at prior beliefs. Your assertion about the evidence requiring no gods - I think you need to define that more clearly.

Whether this is convincing or not depends very much on your prior belief.
Do you mean by this evidence about the world that can be interpreted as being consistent with a naturalist worldview regardless of whether it can also be interpreted as consistent with a theist worldview? If not please let me know - if this isn't what you mean then by all means ignore the argument below.

If this is your interpretation of evidence then we're back to the burden of proof argument and strongly informed priors that naturalism is true.

If you're assuming a flat prior you have to show that the evidence is better explained by naturalism that theism. Not simply that if its consistent with both that this makes naturalism more likely.

However, if I’m fairly stating your position that the burden of proof is on theism then by definition you don’t have a flat prior (I think you are saying they’re not equally probably explanations of the world).

The burden of proof exists for all ontologically possible claims. Both "there definitely is a supernatural world featuring a god," and "there definitely is only a natural world without any gods," would entail the same burden of proof. However, my position is "I don't know what this universe is yet," and thus doesn't have a burden at all. That's why I- and most of the others in this thread, if I recall- are agnostic atheists and not gnostic ones; the purely gnostic naturalist claim overstresses its case just as much as a theistic one does.

The anostic atheist position, however, doesn't prevent us from finding fault with any given argument that comes in.

Thanks that's helpful clarification. So in that sense your gnostic atheist position is a flat prior? That is, there is great uncertainty whether naturalist or theist explanations of the universe are valid explanations of the world?

But see responses below which seem to me inconsistent with what you've said here.

If I’m being true to your position, this is in Bayesian terms technically called an informed prior – which means you have a view about what explanation is more likely. It would be for you to tell me how much more likely you think metaphysical naturalism is compared with theism. But obviously the more likely you think it is- the greater weight is given to this prior belief and less weight to the data when evaluating the evidence.

Well, I have no experience with the supernatural at all, and plenty with the natural. That's why, say, when there's a storm I'm more willing to accept a claim that it's happening in line with natural meteorological patterns, than that it was created wholecloth by magic. I don't know that there is magic, after all. I do know that storms naturally occur.

Now, if the person with the magic claim can demonstrate the spell, then... fine. Methodological naturalism, not metaphysical naturalism.

That's a helpful illustration. I think this is very useful as this seems clear to me your not using a flat prior to interpret the evidence.

Of course I would also come to the conclusion the storm is due to meterological patterns and that explanation would be totally consistent with my theist position.

The question is why do you think these meterological patterns have a naturalistic cause and don't just reflect the world God creates and sustains?

I think from your illustration (but as I say this may be a misinterpretation so you can correct it) you are saying you hold a strong informed prior that naturalist explanations are more likely.

If you start with a flat prior the illustration doesn't show that either the naturalist or theist position is a better explanation. If you start with a strongly informed prior that naturalism is more likely then you'll conclude naturalism is a better explanation. Which means by definition your conclusion is determined by your prior beliefs.

Your informed prior that naturalism is more likely is probably less stronger than a gnostic atheist but that still doesn't imply that you don't have an informed prior.

What I’m saying is there are at least two opposing informed priors about which explanation is a better understanding of the world. This is quite a common situation – so how do Bayesian analyses take this into account when examining the data? Generally, they analyse the data comparing conclusions using different priors. At a minimum they look at the impact of using each of the opposing priors as well as flat priors. Which is to say the burden of proof argument for me is simplistic and in many senses misleading way of interpreting the evidence.

Only if you're under the impression that the burden of proof only lies on your claim. But it lies on every positive claim made: it's just having evidence for a thing, really. This is simpler than you're making it.

I've tried to show in the above example prior beliefs matter whether your a gnostic atheist, agnostic atheist, or theist. So I guess I'm saying I don't think there is a neutral position on this question - by all means show me that I'm wrong.

I'm a little confused by your responses. On some you are saying you
just want to take the evidence into account and therefore consider the flat prior to reflect your agnostic atheism.

But at the same time the illustration you've given suggests a strongly informed prior that naturalism is more likely. So I suppose I'm not sure what position you're taking regarding prior beliefs before interpreting the evidence.

Its almost impossible to understand how you interpret the evidence if I don't understand your position on prior beliefs.