My evil paper was accepted for PPR today. I’m rather pleased about that,

because I do like the paper, and there’s not much I could do with it if PPR

didn’t take it, because few journals will take long commentary pieces on

papers published in other journals. It was a little harder to write than may

appear at first. Trying to write a reply to a paper by Adam Elga is non-trivial in a few

respects, because one has to keep up both in terms of the quality of the

philosophy and in terms of the quality of the writing, especially of the

humour. There’s one compensating benefit, which is that because Adam’s always

so clear, there’s no painful exegesis to do before launching into philosophy.

But still trying to keep up is no trivial matter. I think I mostly succeeded

this time, but then I would think that, wouldn’t I?

Behind all the details, and all the jokes, there is a relatively serious

matter to the paper. I think, following Keynes, there’s a very important

difference between risk and uncertainty. It’s a slight exaggeration, but one

way of conceptualising how important the distinction is is that from my

perspective the person who how probable p is has more in common

with the person who knows whether p is true than she has with the person who

has no evidence either way as to whether p is true. Indifference principles

threaten this neat picture, because they suggest that we can get to a real

probability of p, not its objective chance but still a single precise

probability that is objectively correct (relative to a body of evidence) on

the basis of ignorance about the evidence.

That, I think, was the real problem with traditional indifference principles,

not the mere problem that they were inconsistent. The inconsistencies look

like a technical problem that need a technical solution, and after a few

false attempts I think Adam’s solution is the right one. (I still make some

technical objections in the paper, but you could live with them if they

didn’t hint at a deeper problem.) If there was only a technical problem here,

it would be solved. But there isn’t, there’s a philosophical problem here

too, and no amount of care and attention to the details will solve it, only a

theory that blocks any inference from ignorance to probability.