A rationalist should strive to have a given belief if and only if that belief is true. I want to be a theist if and only if theism is correct.
You’re assuming that “no God” is the null hypothesis.
Not really. Bayesian reasoning doesn’t have any notion of a null hypothesis. I could just as well have said “I want to be an atheist if and only if atheism is correct”.
Without any empirical evidence either way, is it more likely that there is some sort of Deity or that there isn’t?
One can talk about the prior probability of a given hypothesis, and that’s a distinct issue which quickly gets very messy. In particular, it is extremely difficult to both a) establish what priors should look like and b) not get confused about whether one is taken for granted very basic evidence about the world around us (e.g. its existence). One argument, popular at least here, is that from an Occam’s razor standpoint, most deity hypotheses are complicated and only appear simple due to psychological and linguistic issues. I’m not sure how much I buy that sort of argument. But again, it is worth emphasizing that one doesn’t need control of the priors except at a very rough level.
It may help if you read more on the difference between Bayesian and frequentist approaches. The general approach of LW is primarily Bayesian, whereas notions like a “null hypothesis” are essentially frequentist.
You’re right that prior probability gets very, very messy. It’s a bit too abstract to actually be helpful to us.
So, then, all we can do is look at the evidence we do have. You’re saying that the argument is one-sided; there is no evidence in favor of theism, at least no good evidence. I agree that there is a lot of bad evidence, and I’m still looking for good evidence. You’ve said you don’t know of any. Thank you. That’s what I wanted to know. In general I don’t think it’s healthy to believe the opposing viewpoint literally has no case.
In general I don’t think it’s healthy to believe the opposing viewpoint literally has no case.
Do you think that young earth creationists have no substantial case? What about 9/11 truthers? Belief in astrology? Belief that cancer is a fungus(no I’m not making that one up)? What about anything you’ll find here?
The problem is that some hypotheses are wrong, and will be wrong. There are always going to be a lot more wrong hypothesis than right ones. And in many of these cases, there are known cognitive biases which lead to the hypothesis type in question. It may help to again think about the difference between policy issues (shouldn’t be one-sided), and factual questions (which once one understands most details, should be).
You’re right that prior probability gets very, very messy. It’s a bit too abstract to actually be helpful to us.
You cannot escape the necessity of dealing with priors, however messy they are.
So, then, all we can do is look at the evidence we do have.
The available evidence supports an infinite number of hypotheses. How do you decide which ones to consider? That is your prior, and however messy it may be, you have to live with it.
Not really. Bayesian reasoning doesn’t have any notion of a null hypothesis. I could just as well have said “I want to be an atheist if and only if atheism is correct”.
One can talk about the prior probability of a given hypothesis, and that’s a distinct issue which quickly gets very messy. In particular, it is extremely difficult to both a) establish what priors should look like and b) not get confused about whether one is taken for granted very basic evidence about the world around us (e.g. its existence). One argument, popular at least here, is that from an Occam’s razor standpoint, most deity hypotheses are complicated and only appear simple due to psychological and linguistic issues. I’m not sure how much I buy that sort of argument. But again, it is worth emphasizing that one doesn’t need control of the priors except at a very rough level.
It may help if you read more on the difference between Bayesian and frequentist approaches. The general approach of LW is primarily Bayesian, whereas notions like a “null hypothesis” are essentially frequentist.
You’re right that prior probability gets very, very messy. It’s a bit too abstract to actually be helpful to us.
So, then, all we can do is look at the evidence we do have. You’re saying that the argument is one-sided; there is no evidence in favor of theism, at least no good evidence. I agree that there is a lot of bad evidence, and I’m still looking for good evidence. You’ve said you don’t know of any. Thank you. That’s what I wanted to know. In general I don’t think it’s healthy to believe the opposing viewpoint literally has no case.
Do you think that young earth creationists have no substantial case? What about 9/11 truthers? Belief in astrology? Belief that cancer is a fungus(no I’m not making that one up)? What about anything you’ll find here?
The problem is that some hypotheses are wrong, and will be wrong. There are always going to be a lot more wrong hypothesis than right ones. And in many of these cases, there are known cognitive biases which lead to the hypothesis type in question. It may help to again think about the difference between policy issues (shouldn’t be one-sided), and factual questions (which once one understands most details, should be).
You cannot escape the necessity of dealing with priors, however messy they are.
The available evidence supports an infinite number of hypotheses. How do you decide which ones to consider? That is your prior, and however messy it may be, you have to live with it.