A common variation on this is to assert that they can essentially elevate a mild probability hypothesis and treat it as a high probability hypothesis. Certainly when I was a theist I sometimes thought that way.
Currently trendy is the Bayesian argument, which frequently starts with asserting that the only acceptable proper and right-thinking prior probability of God is 0.5 AAAAAARGH
(The problem here being the assumption that we start knowing nothing at all rather than that we know really quite a lot—that being the bit in the argument where a negligible probability is turned into a non-negligible one.)
(Sorry, I just had lunch with some relatively sensible theists and I’m still going AAAAAARGH)
I’ve never heard anyone seriously say that, but I don’t doubt that they exist.
I’m curious what they have to say about the different beliefs of different theist traditions… that is, is the prior for each of them supposed to be .5?
I need to find a written example. I believe a debate Christopher Hitchens lost against a theist involved him being waylaid by an unexpected Bayesian attack. In any case, I have an anecdotal datum that the word is out there.
Something like that—I need to find a written example of the Bayesian probability argument for God. The idea is that when you have no idea, and no basis for an idea, you have no reason to have the probability of A greater or lesser than the probability of not-A.
Note that a large part of the thrust of Dawkins’ The God Delusion is to show the improbability of the God hypothesis (as stated by Dawkins) given what we know already.
The assumption that we know nothing about a question, when we actually know quite a lot, is a common way to turn a negligible probability into a non-negligible one, e.g. as I have noted before about parapsychology.
(slaps self upside of head) And, of course, there’s an entire section of The God Delusion dealing with the Bayesian argument, as first popularised by Stephen Unwin in his 2003 book The Probability Of God, who—ta-dah! - started with 0.5.
Ah, here we are! And another! Not silly enough to actually use 0.5 as the prior, but then these are the sophisticated versions—or at least lengthy.
Another anecdote of the currency of “Bayes, P(God)=0.5”: Armondikov at RationalWiki also ranted recently on his FB (else I’d link it) about theists who’ve discovered the word “Bayes” and start at 0.5.
The important phenomenon to note here is the word “Bayes” achieving currency amongst the not joined-up of thinking, as an excuse for stupidity. Good thing or bad thing?
Most likely a bad thing, given similar past examples. E.g. I’ve talked to atheists who won’t touch Bostrom/Anthropic Bias because they associate “the anthropic principle” with theological fine-tuning arguments. And the general problem of audiences’ first impression of Bayes being that it’s just another clever way to argue for whatever you want to believe or want others to believe.
On the other hand, it would be interesting to have more debates between theists and atheists who are both familiar with Bayes and use it explicitly, if the atheist is good enough at noticing and explaining flawed uses of it, so audiences can become familiar with fallacious uses of it and see that it can be used wisely.
A common variation on this is to assert that they can essentially elevate a mild probability hypothesis and treat it as a high probability hypothesis. Certainly when I was a theist I sometimes thought that way.
Currently trendy is the Bayesian argument, which frequently starts with asserting that the only acceptable proper and right-thinking prior probability of God is 0.5 AAAAAARGH
(The problem here being the assumption that we start knowing nothing at all rather than that we know really quite a lot—that being the bit in the argument where a negligible probability is turned into a non-negligible one.)
(Sorry, I just had lunch with some relatively sensible theists and I’m still going AAAAAARGH)
I’ve never heard anyone seriously say that, but I don’t doubt that they exist.
I’m curious what they have to say about the different beliefs of different theist traditions… that is, is the prior for each of them supposed to be .5?
I need to find a written example. I believe a debate Christopher Hitchens lost against a theist involved him being waylaid by an unexpected Bayesian attack. In any case, I have an anecdotal datum that the word is out there.
That’d imply an interesting cosmology.
Is that because if you treat probabilities of (God or not God) as maximum entropy without prior information you’d get 50/50?
Something like that—I need to find a written example of the Bayesian probability argument for God. The idea is that when you have no idea, and no basis for an idea, you have no reason to have the probability of A greater or lesser than the probability of not-A.
Note that a large part of the thrust of Dawkins’ The God Delusion is to show the improbability of the God hypothesis (as stated by Dawkins) given what we know already.
The assumption that we know nothing about a question, when we actually know quite a lot, is a common way to turn a negligible probability into a non-negligible one, e.g. as I have noted before about parapsychology.
(slaps self upside of head) And, of course, there’s an entire section of The God Delusion dealing with the Bayesian argument, as first popularised by Stephen Unwin in his 2003 book The Probability Of God, who—ta-dah! - started with 0.5.
Lukeprog has posted one in the discussion section a while back.
*two
A Bayesian Argument for the Resurrection of Jesus
A Bayesian Argument for Theistic Fine-Tuning
Ah, here we are! And another! Not silly enough to actually use 0.5 as the prior, but then these are the sophisticated versions—or at least lengthy.
Another anecdote of the currency of “Bayes, P(God)=0.5”: Armondikov at RationalWiki also ranted recently on his FB (else I’d link it) about theists who’ve discovered the word “Bayes” and start at 0.5.
Rejoinder to them is at http://lesswrong.com/lw/jp/occams_razor/ and http://lesswrong.com/lw/19m/privileging_the_hypothesis/.
The important phenomenon to note here is the word “Bayes” achieving currency amongst the not joined-up of thinking, as an excuse for stupidity. Good thing or bad thing?
Most likely a bad thing, given similar past examples. E.g. I’ve talked to atheists who won’t touch Bostrom/Anthropic Bias because they associate “the anthropic principle” with theological fine-tuning arguments. And the general problem of audiences’ first impression of Bayes being that it’s just another clever way to argue for whatever you want to believe or want others to believe.
On the other hand, it would be interesting to have more debates between theists and atheists who are both familiar with Bayes and use it explicitly, if the atheist is good enough at noticing and explaining flawed uses of it, so audiences can become familiar with fallacious uses of it and see that it can be used wisely.
Here’s some non-LW links: Richard Swinburne, Stephen Unwin.