Has the many-worlds interpretation been disproven?

Question

There is a pop-sci article going around which mentions a very recent preprint where the authors did a more sophisticated version of the double slit experiment with photons.

In summary, they are saying that through weak measurements, they avoided the complete collapse of the wavefunction and after repeating the experiment many times, they got a statistical result which, according to them, showed that the photons existed physically in two places (which seems very strange to me, since I've always read that the popular portray of a particle pre-measurement being in more than one place is misleading and doesn't understand the probabilistic nature of the wave function).

They claim that the wave function is a real thing and that the fact that the photons actually splitted in two photons proves that, since one of them is not "disappearing" into one of the branches (if the many worlds interpretation), both of them "remain" in the universe, supposedly disproving the many worlds interpretation

I have many issues with this article

1. apparently the authors consider the wave function as a real thing, but this is a thing that many-worlds interpretation also does

2. this is the first time that I see someone claim that a photons is being actually duplicated prior to measurement instead of existing in an undefined probability state, as it's usually portrayed

3. I don't really understand why they would claim to have disproven the many-worlds interpretation since #1 it's an interpretation, and therefore cannot be proven or disproven and #2 if they didn't actually collapsed the wavefunction by performing weak interactions or measurements, why would they expect to see the "branching" proposed by many worlds? Wouldn't this supposedly occur when the measurement is actually done?

Answer 1

You are correct that interpretations cannot be distinguished by experiment when they are interpretations which predict the same physical phenomena.

However, in the case of many worlds, there is a prediction that could in principle be tested, but no existing apparatus could get even close to testing it, and quite likely no apparatus ever will. This is the prediction, for example, that in a Schrodinger cat experiment both possible outcomes have physical existence in any given run of the experiment. This is different from single-world interpretations which hold that, in such an experiment, the cat is either alive or dead at the end, not both. One does not need something as extreme as the death of a cat to illustrate this. An ordinary particle-detector or photoreceptor will do. (What is required is merely irreversibility; many-worlds denies that there is genuine irreversibility.)

The cited preprint does not describe an experiment which could resolve this. Therefore your doubts about the claims in the preprint are largely valid.

Answer 2

In general, the results of an interference experiment depend on what happens to all of the possible states of the system, see Section 2 of this paper for an example:

Machines, Logic and Quantum Physics

If you don't say that there is something in all of those states you don't have any account of what is happening in reality to produce the outcome and this is part of the motivation for the MWI.

The MWI claims that the equations of motion of quantum theory are correct as a description of reality. The variants of it that are advocated now say that parallel universes are an approximation that only holds under decoherence

Decoherence and Ontology, or: How I Learned To Stop Worrying And Love FAPP

Decoherence is a matter of degree and so there is some arbitrariness in saying when you have a new universe.

You say that you can't distinguish between interpretations, but this is false in general. Different interpretations make different claims about what is happening in reality and advocates of non-MWI interpretations sometimes specifically claim their theories are testable versus quantum theory. See

Collapse Models: a theoretical, experimental and philosophical review

Pilot-wave theory and the search for new physics

Advocates of the MWI also claim it is testable:

The Logic of Experimental Tests, Particularly of Everettian Quantum Theory

The Copenhagen and statistical interpretations are a lot vaguer about what is happening in reality. I am not aware of any experimentally practical test of those theories.

Imperfect measurements don't fit in well with collapse or other alternatives to quantum theory that limit the applicability of quantum equations of motion:

What is orthodox quantum mechanics?

As far as I can tell, the experiment as described in the paper doesn't create any problems for the MWI. I haven't read the New Scientist article, but if they have a refutation of the equations of motion of quantum theory, it really ought to be in the paper since that will make it easier for them to get a Nobel Prize.

Answer 3

While the New Scientist (NS) article is unfortunately paywalled, I think your skepticism is justified. The preprint abstract says nothing about a multiverse, and from a brief skim, a lot of the language the paper uses describing the experimental setup talks about the photon being delocalized. While I haven't worked through the detail to see if their descriptions lines up with the MWI "narrative," I read that as a sign against NS' interpretation, because talking about photons being delocalized in general is a sign of leaning into MWI as opposed to a classical realist view - that is, the picture usually placed in opposition to MWI fundamentally starts from the position that photons (etc) have a singular "real" position, so at most it only makes sense to talk about the probability distribution of that real position being more or less localized.

In my experience, for whatever reason, while popular science presentations quite happily hold up the multiverse as yet another wild and wacky idea from the weirdness factory that is quantum physics, they are extremely reluctant to "take it seriously" in terms of the implications it has for the ontology and interpretation of experimental results. It particularly strains things to talk about an experiment that involves weak measurements in an implicitly Copenhagen context: Copenhagen doesn't have any articulated explanation for how wave function collapse happens; it just does. That worked, at least pragmatically, when the "it" was always total collapse and in a sense the same every time, but weak measurements and the possibility of partial collapse demand elaboration that Copenhagen can't really provide. Contra NS's headline, you need either the multiverse or some other much more sophisticated interpretation of the theory for the experiment they're citing to even make sense as the experimenters describe it.