I think that this is actually possible, i.e. it is possible to state “Yahweh doesn’t exist” with certainty: if Yahweh is internally mathematically inconsistent, i.e. Yahweh’s existence would be a proof that 1 = 0.
However, there are probabilities involved in any assertion that Yahweh is inconsistent; unless you have a complete definition of Yahweh (e.g. highlighted phrases in a specific Bible) that involves something as clear as “Yahweh is green” and “Yahweh is not green”, but there are the standard problems due to language being imperfect telepathy (etc) that make this unlikely to be possible.
Well, Eliezer Yudkowsky doesn’t assign probability exactly 1 to 2 + 2 = 4 either, and whereas at first I thought that was nuts, this made me realize he does have a point (as I think any difference between “2 + 2 = 4” and “51 is prime” is only quantitative).
I used to think this, then this paper that lukeprog linked, on logical uncertainty, gave me a coherent model that says otherwise. I’m still not sure which model I like better.
I think that this is actually possible, i.e. it is possible to state “Yahweh doesn’t exist” with certainty: if Yahweh is internally mathematically inconsistent, i.e. Yahweh’s existence would be a proof that 1 = 0.
However, there are probabilities involved in any assertion that Yahweh is inconsistent; unless you have a complete definition of Yahweh (e.g. highlighted phrases in a specific Bible) that involves something as clear as “Yahweh is green” and “Yahweh is not green”, but there are the standard problems due to language being imperfect telepathy (etc) that make this unlikely to be possible.
Well, Eliezer Yudkowsky doesn’t assign probability exactly 1 to 2 + 2 = 4 either, and whereas at first I thought that was nuts, this made me realize he does have a point (as I think any difference between “2 + 2 = 4” and “51 is prime” is only quantitative).
I used to think this, then this paper that lukeprog linked, on logical uncertainty, gave me a coherent model that says otherwise. I’m still not sure which model I like better.
Looks interesting. I’ll have a read of it.