Sean Carroll has a new paper out defending the Multiverse and attacking the naive Popperazi, entitled Beyond Falsifiability: Normal Science in a Multiverse. He also has a
Beyond Falsifiability blog post here.
Much of the problem with the paper and blog post is that Carroll is arguing against a straw man, while ignoring the serious arguments about the problems with multiverse research. The only explanation of the views he is arguing against is the following passage:
a number of highly respected scientists have objected strongly to the idea, in large part due to a conviction that what happens outside the universe we can possibly observe simply shouldn’t matter [4, 5, 6, 7]. The job of science, in this view, is to account for what we observe, not to speculate about what we don’t. There is a real worry that the multiverse represents imagination allowed to roam unfettered from empirical observation, unable to be tested by conventional means. In its strongest from, the objection argues that the very idea of an unobservable multiverse shouldn’t count as science at all, often appealing to Karl Popper’s dictum that a theory should be falsifiable to be considered scientific.
The problem here is that none of those references contain anything like the naive argument that if we can’t observe something, it “simply shouldn’t matter”, or one should not speculate about it, or it “shouldn’t count as science at all.” His reference 7 is to this piece by George Ellis at Inference, which has nothing like such arguments, and no invocation of falsifiability or Popper. Carroll goes on to refer approvingly to a response to Ellis by Daniel Harlow published as a letter to Inference, but ignores Ellis’s response, which includes:
The process of science—exploring cosmology options, including the possible existence or not of a multiverse—is indeed what should happen. The scientific result is that there is no unique observable output predicted in multiverse proposals. This is because, as is often stated by proponents, anything that can happen does happen in most multiverses. Having reached this point, one has to step back and consider the scientific status of claims for their existence. The process of science must include this evaluation as well.
Ellis here is making the central argument that Carroll refuses to acknowledge: the problem with the multiverse is that it’s an empty idea, predicting nothing. It is functioning not as what we would like from science, a testable explanation, but as an untestable excuse for not being able to predict anything. In defense of empty multiverse theorizing, Carroll wants to downplay the role of any conventional testability criterion in our understanding of what is science and what isn’t. He writes:
The best reason for classifying the multiverse as a straightforwardly scientific theory is that we don’t have any choice. This is the case for any hypothesis that satisfies two criteria:
- It might be true.
- Whether or not it is true affects how we understand what we observe.
This seems to me an even more problematic and unworkable way of defining science than naive falsifiability. This whole formulation is extremely unclear, but it sounds to me as if various hypotheses about supreme beings and how they operate would by this criterion qualify as science.
Carroll also ignores the arguments made in my letter in the same issue (discussed here), which were specifically aimed at the claims for multiverse science that he is trying to make. According to him, multiverse theory is perfectly conventional science which just happens to be hard to evaluate:
That, in a nutshell, is the biggest challenge posed by the prospect of the multiverse. It is not that the theory is unscientific, or that it is impossible to evaluate it. It’s that evaluating it is hard.
The main point I was trying to make in the piece Carroll ignores is that the evaluation problem is not just “hard”, but actually impossible, and if one looks into the reason for this, one finds that it’s because his term “the theory” has no fixed reference. What “theory” is he talking about? One sort of “theory” he discusses are eternal inflation models of a multiverse in which you will have bubble collisions. Some such models predict observable effects in the CMB. Those are perfectly scientific and easy to evaluate, just wrong (since we see no such thing). Other such models predict no observable effect, those are untestable. “Hardness” has nothing to do with it, the fact that there is some narrow range of models where tests are in principle possible but hard to do is true but irrelevant.
The other actual theory Carroll refers to is the string theory landscape, and there the problem is not that evaluating the theory is “hard”, but that you have no theory. As for bubble collisions, you have plenty of conjectural models (i.e. “string vacua”) which are perfectly well-defined and scientific, but disagree with experiment so are easily evaluated as wrong. While many other conjectural models are very complex and thus technically “hard” to study, that’s not the real problem, and acquiring infinitely powerful computational technique would not help. The real problem is that you don’t have a theory: “M-theory” is a word but not an actual theory. The problem is not that it’s “hard” to figure out what the measure on the space of string vacua is, but that you don’t even know what the space is on which you’re looking for a measure. This is not a “hard” question, it’s simply a question for which you don’t have a theory which gives an answer.
I do hope someday Carroll and other multiverse fans will some day get around to addressing the real arguments being made, perhaps then this subject could move forward from the sorry state it seems to be stuck in.
Update: Coel Hellier has a blog posting here taking Carroll’s side of the debate.
Update: Yet another new argument for multiverse mania as usual science on the arXiv, this time from Mario Livio and Martin Rees. The same problems with the Carroll article recur here, including the usual refusal to acknowledge that serious counter-arguments exist. They give no references at all to anyone disagreeing with them, instead just knock down the usual straw man, those unknown scientists who think that theorists should only discuss directly observable quantities:
We have already discussed the first main objection — the sentiment that envisaging causally disconnected, unobservable universes is in conflict with the traditional “scientific method.” We have emphasized that modern physics already contains many unobservable domains (e.g., free quarks, interiors of black holes, galaxies beyond the particle horizon). If we had a theory that applied to the ultra-early universe, but gained credibility because it explained, for instance, some features of the microphysical world (the strength of the fundamental forces, the masses of neutrinos, and so forth) we should take seriously its predictions about ‘our’ Big Bang and the possibility of others.
We are far from having such a theory, but the huge advances already made should give us optimism about new insights in the next few decades.
Livio and Rees do here get to the main point: we don’t have a viable scientific theory of a multiverse that would provide an anthropic explanation of the laws of physics. The causes for optimism that they list are the usual ones involving inflationary models that give essentially the same physics in other universes, not the different physics they need for anthropics. There is one exception, a mention of how:
accelerator experiments can (in principle) generate conditions in which a number of metastable vacuum solutions are possible, thereby testing the premises of the landscape scenario.
They give no reference for this claim and I think it can accurately be described as utter nonsense. It’s also (in principle) possible that accelerator experiments will generate conditions in which an angel will pop out of the interaction region bearing the laws of physics written on gold tablets. But utterly implausible speculation with no evidence at all backing it is not science.
The authors note that:
an anthropic explanation can be refuted, if the actual parameter values are far more ‘special” than anthropic constraints require.
The problem with this is that you don’t have a theory that gives you a measure on parameter values, so you don’t know what is ‘special’ and what isn’t. As I keep pointing out, the fundamental problem here is even more basic that not having a probability measure on possible universes: we have no viable theory of what the space of possible universes is, much less any idea of how to calculate a measure on it. And no, we are not seeing any progress towards finding such a theory, quite the opposite over the past decades.
Truly depressing is that even the best of our journalists see this kind of article, written by two multiverse enthusiasts and giving no references or serious arguments for the other side, as “even-handed”.
Update: Two new excellent pieces explaining the problems with the multiverse. Ethan Siegel in particular explains the usually ignored problem that the kind of inflation we have any evidence for doesn’t give you different laws of physics, and ends with
The Multiverse is real, but provides the answer to absolutely nothing.
Sabine Hossenfelder explains four of the arguments generally given for why the Multiverse is science, answering them each in turn, with conclusions:
1. It’s falsifiable!
So don’t get fooled by this argument, it’s just wrong.
2. Ok, so it’s not falsifiable, but it’s sound logic!
So don’t buy it. Just because they can calculate something doesn’t mean they describe nature.
3. Ok, then. So it’s neither falsifiable nor sound logic, but it’s still business as usual.
So to the extent that it’s science as usual you don’t need the multiverse.
4. So what? We’ll do it anyway.
so you are allowed to believe in it. And that’s all fine by me. Believe whatever you want, but don’t confuse it with science.