Trying to make Philip Goff miserable

Cosmic despair for all!

My friend Philip Goff just posted an article about how I’d made him miserable by convincing him, at least temporarily, that there’s no need to explain the fine-tuning of the constants by invoking God, because a multiverse would do just fine. Thankfully, peace and contentment has reasserted itself as Philip is now convinced I’m wrong.

Philip has invited me to reply, so the hope of this post is that I can pull him back to the abyss of existential despair where I’ve made myself at home.

This is a long back-and-forth in a conversation that’s been going on between us for a couple of years now, so I don’t want to go through the whole thing again, but it seems some sort of a summary is unavoidable.

We both agree that the constants in our laws of physics seem to exhibit a surprising degree of apparent fine-tuning for life and complexity, such that if they were just a bit different, then the universe would be devoid of anything interesting.

We both think this is striking enough to demand some sort of explanation.

Where we disagree is on whether a vast multiverse across which the constants vary would explain it, in much the same sort of way that the perfection of earth for life is explained by understanding that there are lots of planets, most of which are not suitable for life, and we must find ourselves on one of the rare jewels that is.

More accurately, I think that apparent fine-tuning is by itself evidence that supports the hypothesis that there is such a multiverse, and Philip does not. Philip’s objection (following others such as Roger White and Paul Draper in particular) is that while this would explain why *some* universe is fine-tuned, it wouldn’t explain why *this* universe is fine-tuned. If we already knew there were such a multiverse, then it would be analogous to the situation with earth and the planets, but since we don’t, it is not.

There are a lot of subtleties going on here, but the crux of the matter turns out to be whether it is correct for a Bayesian to take it as background knowledge that the world exists when updating with the evidence that the world is life-permitting, or that the laws of physics have the sort of structure they do when updating with the evidence that the constants have the specific values they do.

Philip thinks this is fine, and I do not. If we already know the world exists, then I claim we already know that it is life-permitting, because the way that we know the world exists is that we are living creatures embedded within it. The background directly implies the evidence, so it makes no sense to separate them as Philip wants to.

Philip seems to think the problem with my approach is that

you’re not going to be able to put anything in the background information if you’re considering the evidence that our universe has the right numbers for life, as all your knowledge is dependent on our universe having the conditions that allow you to exist

But here I think Philip misunderstands my position. OK, I’m starting from the way he wants to look at the evidence and pointing out a problem with it. But the way I want to look at the evidence is entirely different, not Philip’s way with this one small change. In particular, I don’t think we should be taking the new evidence to be “that the universe has the right numbers for life”. This is also background to me, for just the reason Philip says. All my knowledge is logically dependent on this.

Rather, the way I want to organise things is that the background is that the universe exists, is life-permitting and so on, and the new evidence is that physicists report that the existence of life and complexity are surprisingly sensitive to the value of the constants. It didn’t have to be this way. It’s logically possible, at least as far as I can tell, for the physicists to have found the reverse, that life was pretty much inevitable no matter what sort of universe we dreamt up. The surprising evidence is just how finely the constants seem to need to be tuned to permit life — not the mere fact that the universe permits life! I could not possibly be¹ in a situation where (assuming I’m cognitively normal) I didn’t realise that the universe I’m living in permits me to live in it. But I could easily be in a situation where I didn’t realise how precarious my existence was.

As such, I’m sorry to break it to him, but I don’t think the solution Philip has found to his angst works. Back to the abyss with ye!

1. I don’t want to imply that Bayesian reasoning requires that the way we separate out background and evidence has to be realistic, or even epistemically possible. But I do think that if you go about it in a realistic way, you are more likely to avoid mistakes. Go crazy as long as you get the right answer at the end. However, if two approaches give different answers, the more realistic approach is less likely to have gone astray.