Contra Huemer on Whether You Need Souls to Explain Fine Tuning

To explain fine-tuning, you just need a multiverse, not the assumption that we're souls.

Toddler 1: Invisible tanks are real.

Toddler 2: Oh yeah, I’ve seen them.

—Overheard by my brother when he was working at a summer camp.

Michael Huemer and I both think we are souls. We disagree, however, about whether one needs to think souls exist to explain fine-tuning, based on some version of the inverse gambler’s fallacy charge. I’ve already addressed this basic charge, though Huemer has a more nuanced version of it that he relays in his excellent book Knowledge, Reality, and Value as well as the following blog article:

Arguments for "God" #3: Fine Tuning

Now for the leading objection to the multiverse theory. To explain the objection, we need to introduce a hypothetical example:

Coin Flip: You flip a coin, and it comes up heads ten times in a row. In general, if a coin is flipped ten times, the odds of its coming up heads every time is 1/1024. This is sufficiently improbable that you start looking for an explanation. One hypothesis is that there are thousands of other coins somewhere, each getting flipped ten times.

Question: Does the hypothesis of many other coins explain the fact that you observed ten heads in a row? And do you therefore have evidence for lots of other coins being flipped?

I hope you can see that the answer is “no” to both. (If the answer were “yes”, then every time you observed anything unlikely, that would be evidence for other universes.[58]) Now, why is the answer “no”? Well, it is true that the existence of many other coin flippers would make it more likely that at least one coin would come up heads ten times in a row However, that’s completely irrelevant, because your evidence in the story is not [At least one coin came up heads ten times]. Your evidence in the story – what you observed to be true – was [This coin (the one you just observed) came up heads ten times]. And the probability that this specific coin would come up heads ten times in a row is completely unaffected by the existence or non-existence of other coins flipped by other people. Thus, the observed outcome with this coin also provides no evidence whatsoever for the existence of other coin-flippers.

Now draw the parallel to the multiverse theory. We observe some seemingly improbable parameter values in the universe. The hypothesis of many parallel universes would indeed make it much more likely that at least one universe would have life-friendly parameters. However, the existence of all these other universes would have no effect whatsoever on the probability that this universe would have life-friendly parameters. And our evidence is that this universe has life-friendly parameters. So our evidence provides no support at all to the multiverse theory since the multiverse theory does nothing to explain the evidence. That is what I call the “this universe” objection to the Multiverse theory.

I used to think that objection was decisive and therefore that the Multiverse theory was a complete loser. I was wrong, though; there is a way to defend the Multiverse explanation, but it is one that most atheists are not going to like. Assume that you could have been born (i.e., you had some chance of being born) into any universe that was hospitable to your kind of life. In that case, the probability that you would find yourself alive now would be greater if there were many universes than if there were only one universe. So the multiverse theory can explain the fact that you are alive now, and so your current existence is evidence for the multiverse

On the other hand, suppose you could not have been born in any other universe besides this one, even if there were other universes. In that case, your probability of finding yourself alive now would be completely unaffected by the existence of other universes, and so you have no evidence for other universes.

So the way to defend the multiverse theory as an explanation for our evidence is to claim that persons are not world-bound: A given individual could have existed in any universe that was sufficiently hospitable. Any time a sufficiently hospitable universe exists, you have a chance of being born into it. And maybe that’s true, so maybe the multiverse theory is okay.

Huemer’s basic analogy fails because existence has selection effects. It’s true that you shouldn’t infer that there’s a multiverse from the fact that someone gets a bunch of heads on a coin flip. But the reason for this is straightforward—given that you exist in the circumstances that you’re in, the odds that someone would get 10 consecutive heads is unaffected by the total number of universes. This inference understates the evidence—the more specific version of the evidence takes into account your background information that you exist and are in the scenario that you were in prior to the flips.

Huemer’s next claim is that, if I can only exist in this universe, then positing lots of other universes doesn’t help explain anything, just like, if a cockroach can only survive in one room of a house, you shouldn’t think it has lots of other rooms on the grounds that that explains why there are so many cockroaches. But here he commits a subtle error. It’s true that given that I exist, on the physicalist view, I can’t exist elsewhere. But I don’t know which of the beings in modal space I am.

Suppose that there are two possible states of affairs: first, that there will be a single universe with one person who is qualitatively like me (tall, handsome, brilliant, yet also brooding and mysterious); second, that there will be one hundred people qualitatively like me, each in their own universe. Upon finding out that I exist, should I think there’s a multiverse?

Yes! It’s true that whichever universe I’m in will be the only one that I could have been in if, per impossibile, there aren’t souls. But I don’t know which of the universes I’m in because I don’t know which of the beings qualitatively like me I am. As a result, I should split my credence across the 101 possible people like me that I might be, and conclude that almost certainly the second state of affairs obtained.

It’s true that if the first state of affairs was real then I couldn’t have existed in the second state of affairs. But I don’t know which one is real—which of the possible qualitative duplicates of me I am—so as a result, I should update in favor of the version of the universe that has more people.

We can illustrate this point mathematically. Suppose there are two hypotheses: God and the multiverse. Assume that the God hypothesis predicts just one finely tuned universe (it doesn’t, btw, as I argued in a recent YouTube video—it instead predicts everyone would be created). The multiverse predicts 100 universes, one of which would be life-permitting.

I wake up to discover that, remarkably, I exist. Let’s say that I start out thinking each hypothesis has a 50% chance of being right and that I couldn’t exist in any universe other than the one that I exist in. But I don’t know which universe I’m in. Suppose that the universes are arranged from right to left, with the rightmost universe labeled R1, the second rightmost labeled R2, etc.

I don’t know which universe I’m in. If theism is true, I’d be guaranteed to exist in R1, while if the multiverse is true, I should split my credence across all the possible versions of me that might exist. So I conclude that the odds I’m in w1 is 50.5%, w2 is .5%, w3 is .5%, etc. Thus, my existence does not confirm theism over the multiverse.

Now, maybe one thinks that I could only exist in a universe qualitatively like ours. Well, if this is true, then theism won’t be confirmed over a multiverse because the multiverse would also predict one life-permitting universe. If there are more possible worlds, the odds one would be qualitatively like ours increase.

Now, if we knew we were in the first universe, for example, would theism better explain that fact. Yes! Consequently, if we knew we were in the 50th universe, that would confirm a multiverse. But because we don’t know which of those is true, neither hypothesis gets a boost.

Here’s an analogy. Suppose that some people are looking for a house for me. Assume there’s only one kind of house I could exist in; if they can’t find the right house, I would die. There are two hypotheses—first that they spent a long time looking for a house in Colorado, second that they spent a short time looking in Maryland. Assume that only up to one of these states could have a habitable house. Upon discovering that there’s a house I can exist in, I should get evidence for the first hypothesis. It’s true that whichever particular house they find will be the only one that could house me, but I don’t know which house could house me, so I should think that there are more opportunities to find such a house.

Here’s another one: suppose that the current version of me could only exist now. It still seems like I should third in sleeping beauty (which Huemer agrees with). If there are more people, because I don’t know whether I’m the person of day 1 or 2, then the odds I’d exist would be higher.

Thus, explaining fine-tuning isn’t a good reason to believe in souls. Other considerations are the soul reason to believe in them.