Talk:Bell test

History note

• The text added in this edit at 17:24, 9 August 2017‎ was developed in and copied from User:InformationvsInjustice/sandbox. Anthony Appleyard (talk) 12:04, 23 December 2018 (UTC)

Adding the BIG Bell test to the article

Last november, the Institute of Photonic Sciences in Barcelona performed a Bell test experiment with random input generated through volunteers through internet [1], [2], with "preliminary results confirming violations of Bell’s inequality, and thus refuting Einstein".[3] May someone knowledgeable add it to the article? Diego (talk) 17:10, 30 January 2017 (UTC)

Proposed new lede and new "Overview" section

Hi guys, thanks for your contributions! I have written a new lede to the article which I feel is more accessible to a lay reader along with a new Overview section.

Please see them here and give me any feedback you see fit. Thanks. Informata ob Iniquitatum (talk) 04:40, 30 July 2017 (UTC)

Let's give it a shot (bearing in mind that we have a whole other article on Bell's theorem with lots of details that wouldn't fit into an overview here). XOR'easter (talk) 17:47, 9 August 2017 (UTC)

Loopholes : Super-determinism not a scientific hypothesis

I think it needs to be pointed out that the fact that super-determinism is not falsifiable (as the Loopholes section rightly claims) makes super-determinism a non-hypothesis (or a non-scientific hypothesis, to be precise). https://wiki.riteme.site/wiki/Hypothesis specifies that in order to be scientific, a hypothesis needs to be testable. — Preceding unsigned comment added by 2620:113:80C0:5:0:0:0:2222 (talk) 10:46, 19 January 2018 (UTC)

“…we will discuss how it may be possible to test this hypothesis …” (linked article title: Rethinking Superdeterminism)

— Fractaloid (talk) 04:17, 11 September 2023 (UTC)

Many-Worlds

When undoing PaddyLeahy's edit my summary got lost somehow, so I'm pasting it here: you're replacing it with a flat assertion that Many-Worlds is at first sight manifestly nonlocal. Furthermore, you're misrepresenting the positions of Griffiths, Vaidman, Wallace, and Timpson. They do not defend that Many-Worlds is nonlocal. I really don't like when someone puts a gigantic pile of references to support some statement without reading them, you're just wasting the time of everyone else. Tercer (talk) 13:57, 15 July 2023 (UTC)

The question of whether many-worlds is local or nonlocal seems off-topic to me. What does it have to do with Bell's theorem? Because Bell assumed experiments have single outcomes, it is outside the scope of what Bell did. Roger (talk) 17:15, 15 July 2023 (UTC)

Technically speaking you are correct: it suffices to say that Bell's theorem does not apply in Many-Worlds because of the single-outcome assumption. However, any human being will be curious to know whether it also local. And it is obviously local, that is not a controversial point. Tercer (talk) 18:58, 15 July 2023 (UTC)

Roger: I am not persuaded by the dismissal of Bell's theorem in this context on the grounds that it assumes that experiments have a single outcome. In any given branch of the multiverse, experiments do have a single outcome, and the violation of the inequalities is a real effect. In my view, the inference to non-locality is sound, so perhaps it's no surprise that I'm in the "non-local" MWI camp. Therefore, I don't think there really is a Many-Worlds loophole. On the other hand, if MWI really was local and correctly describes the world, then this certainly would be a massive loophole: in fact, it would completely destroy the argument.

In this discussion, you have to recognise that 'locality' is not a straightforward concept. There are many ways of defining it, which do not all agree, which accounts for a lot of the debate. D&H use a new and quite arcane definition, namely of the locality of information. In their version of QM, the apparent non-local correlations between spatially-separated qubits are stored locally but inaccessible to the observer. PaddyLeahy (talk) 14:16, 16 July 2023 (UTC)

This article is about Bell's theorem. The theorem is about hidden variable theories. If MWI were a hidden variable theory, then the question of whether it is local would be relevant. But MWI is not, and it is not even clear how locality applies.

As you say, experiments have a single outcome in a given branch. However it is impossible to relate that outcome to the probabilities and statistics that Bell uses. All branches occur. None are known to be any more likely than any other. You can repeat an experiment, and get a series of outcomes, but other series of outcomes are in other branches. It is impossible to say whether the experiment confirms or violates Bell's inequality. There is no way to apply Bell's analysis. Roger (talk) 17:58, 16 July 2023 (UTC)

Like many, you seem to be underestimating the significance of Bell's theorem. It does not assume any particular theory, only 'locality' in the sense set out by Einstein (not in the EPR paper). See Tim Maudlin's polemical but readable article 'What Bell Did'.

If what you say about the MWI is correct it would be a worthless theory in gross contradiction with the facts, since it would fail to explain the accurate statistical predictions of quantum mechanics, including for Bell tests. Anyone who advocates MWI is signing up to Everett's view that in typical branches quantum predictions are statistically correct. It's true that it is extremely difficult to come up with a non-circular definition of 'typical', but MWI stands or falls on the success of that project. PaddyLeahy (talk) 18:46, 16 July 2023 (UTC)

Bell's theorem does assume that measurements have a single outcome. This is not a controversial point. Maudlin, in particular, does not dispute that.

There's also no difficulty in giving a non-circular definition of typical. Everett did that in his original paper. He defined typical as having large relative 2-norm. There are many objections to his take on probability, but circularity is not one of them. Tercer (talk) 19:49, 16 July 2023 (UTC)

Maudlin is not a supporter of MWI and is not particularly fair to it in this article (which I just re-read after a few years - I'd forgotten that bit). Single outcomes of experiments are only an assumption of the proof in that it is intended to describe the empirical world, where experiments do have outcomes. According to MWI there are other branches yielding different sequence of results, but ('nearly') all statistically agree with QM, violate the inequalities and so require nonlocality. Let's not get into what exactly is the problem with probabilities in MWI. Many solutions have been proposed, all of which have problems of their own IMHO. PaddyLeahy (talk) 20:46, 16 July 2023 (UTC)

"Single outcomes" is assumed everywhere, from the very formulation of the notion of locality to the development of the proof. If you can't understand that too bad for you. In Wikipedia we care about what the sources say, and they say that Bell's theorem doesn't apply to Many-Worlds. Tercer (talk) 21:59, 16 July 2023 (UTC)

Please cite these sources, other than papers pushing what you admit is the controversial position of Deutsch & Hayden. PaddyLeahy (talk) 00:51, 17 July 2023 (UTC)

What is controversial about Deutsch and Hayden is their separable formulation of Many-Worlds. Not their take on Bell's theorem. That is pretty much unanimously agreed upon. See page 19 of Brown and Timpson, where they also cite Wallace, Blaylock, Maudlin, and Wiseman as supporting this position. Tercer (talk) 08:14, 17 July 2023 (UTC)

Yes, MWI is a worthless theory that fails to explain any experimental evidence. Maudlin has some interesting ideas, but they are not accepted by mainstream physicists. He is the one to complain that everyone else is wrong about Bell. Roger (talk) 21:49, 16 July 2023 (UTC)

First of all, you seem to have misunderstood what I wrote. At first sight doesn't mean it is blindingly obvious that; it means without thinking carefully about it, the first impression is. Per my edit summary, the purpose of this edit was to raise a flag that Deutsch & Hayden's approach is controversial. In contrast, the current text states that MWI is local, with no qualification. This is obviously your position but the references you deleted show that it is not the position of all significant thinkers with standing on the issue. Obviously I put the references in because I expected objections from the "MWI is local" brigade, but obviously failed to convince. Let's go through them one by one. I note that you don't mention three references which flat-out contradict Deutsch & Hayden, nor Deutsch's reply to one of them, which would not exist if there was no controversy.

Griffiths, as is well known, does not accept many worlds explicitly, so you might say it's no surprise that he doesn't accept D&H's approach to quantum information either, but their disagreement on this point is not directly related. Vaidman's editorial review asserts that 'Accepting the existence of parallel worlds [2] eliminates randomness and avoids action at a distance, but it still does not remove nonlocality.' The point of the issue of Entropy that this editorial introduced was precisely that this is controversial, and hence the reference supports the statement you deleted. I'm afraid it's you who needs to read Wallace and Timpson more carefully (it's only five pages long!). They are very clear in rejecting D&H's assertion that unitary QM (i.e. the MWI), as usually understood, is a local theory: their point is that the only way D&H can be considered to be local is if it is regarded as (yet another) distinct theory, albeit observationally indistinguishable from standard QM. As such it comes with a lot of metaphysical baggage which W&T describe as an 'unacceptably high price'. I submit that if David Wallace, who has thought harder than anyone about the MWI (see his book), does not accept that it is a local theory, then a reliable encyclopaedia article should not make an unqualified statement that it is, as in the text that you restored. PaddyLeahy (talk) 13:45, 16 July 2023 (UTC)

Now I understand where your confusion is coming from: you've failed to notice the distinction between having local dynamics (i.e. no action at a distance) and local states (also known as separability). It is uncontroversial that Many-Worlds has local dynamics, and it is also uncontroversial that it has nonlocal states.

Deutsch and Hayden did proposal a new formulation of Many-Worlds that has local states, and that is in fact controversial, but that is besides the point. What matters here is their observation that Bell's theorem does not apply in Many-Worlds. Tercer (talk) 20:12, 16 July 2023 (UTC)

Sorry, the confusion is yours. 'Locality' in the context of EPR and Bell's theorem does not mean action at a distance, in the sense of making faster-than-light signalling possible. It means precisely the kind of nonlocality that occurs in QM as entangled states, but the definition is theory-independent (spelled on in great detail by Maudlin, but summarised in Bell's quote from Einstein: "the real, factual situation of the system S2 is independent of what is done with the system S1 which is spatially separated from the former."). Maybe you would even agree that this statement is manifestly violated by MWI? To be honest this point about locality could be made a lot clearer in the article and the corresponding one on Bell's theorem, which exemplify the kind of playing down of the result that Maudlin rails against. PaddyLeahy (talk) 21:04, 16 July 2023 (UTC)

There is no single notion of "locality" in the context of EPR and Bell. There are many different notions, and failure to distinguish between them is the cause of many fruitless arguments. There's the old EPR locality, there's the locality in Bell 1964, that only applies to observables with deterministic outcomes, there is the factorizability from CH 1974, there is the local causality of Bell 1975, and so on and so on and so on.

You should be careful about distinguishing action at a distance from signalling as well. Action at a distance is an ontological notion, and does not imply in signalling, which is an agent-centric notion. Bohmian mechanics, for example, has action at a distance without the possibility of signalling. Tercer (talk) 21:39, 16 July 2023 (UTC)

Locality in the context of Bell's theorem means a local hidden variable theory. Roger (talk) 21:49, 16 July 2023 (UTC)

The article said, before the last edit, that local realism is untenable. i agree that is the consensus. The edit comment says Deutsch and Carroll would dispute it. I doubt it. Local realism has a funny definition. I think it is untenable, the way it is usually defined. Roger (talk) 21:56, 16 July 2023 (UTC)

We shouldn't use the expression "local realism" at all, it is way too vague. The way it's usually defined is as a local hidden-variable theory, and it is indeed untenable, as Deutsch and Carroll would immediately agree. But if one defines local as having local dynamics and realist as having an observer-independent reality then any Many-Worlder will disagree, of course it's tenable, and the violation of Bell inequalities does not threaten it. Tercer (talk) 22:05, 16 July 2023 (UTC)

For once I agree with Tercer. But in this context 'local' should mean local in the sense of the locality that is used to prove the Bell inequality, which is quite different from the sense of Tercer's last sentence. To illustrate, consider an indeterminist theory that was completely local: particles have well-defined trajectories and polarizations and no long-distance interactions, but their polarization sometimes jumps by a random angle (uncorrelated with jumps in other particles). An MWI version explains that at every such jump the universe actually splits into a multitude of branches, one per possible outcome. Such a theory would predict that the Bell inequalities were satisfied in both the quantum jump and MWI version. So Bell inequality violation is picking up a non-locality that is independent of the local nature of the Hamiltonian. PaddyLeahy (talk) 22:48, 16 July 2023 (UTC)

Roger "tends to doubt it"? Deutsch (and the other Everettians) is explicit that MWI is a locally real theory. For example the following: https://www.mdpi.com/1099-4300/21/1/87/htm. Or see this tweet by Deutsch. Clear enough for you? "Non-locality in physics is a myth. Violating Bell's inequalities does not make a theory non-local:" https://twitter.com/DavidDeutschOxf/status/1086244739421274118?t=XqpfQI_mU8-vWoeNBpQG_Q&s=19 Joncolvin (talk) 23:01, 16 July 2023 (UTC)

For a good overview of the consensus of Everettians that MWI is locally real, see here https://www.quora.com/How-can-the-Many-Worlds-Interpretation-be-tenable-if-Bells-theorem-tells-us-that-quantum-mechanics-is-not-compatible-with-both-locality-and-realism/answer/Paul-Mainwood?ch=15&oid=86215114&share=266c6110&srid=gYZf&target_type=answer Joncolvin (talk) 23:18, 16 July 2023 (UTC)

the point is, Everett is a fairly popular interp in the QM physics community, there is consensus that Everett interp is locally realist, and thus it is not true that there is consensus that the universe is not locally real, and the article should not be suggesting there is Joncolvin (talk) 23:22, 16 July 2023 (UTC)

As Tercer explains, the problem is the loaded term "local realism". Yes, the MWI say their theory is local, and that their parallel worlds are real. But of you follow the WP links for local realism, you get something that is untenable. Some people use realism to mean properties expected from classical mechanics. Roger (talk) 23:28, 16 July 2023 (UTC)

Contra Mainwood's Quora comment, and Tercer, there is no consensus that MWI is locally real in the EPR/Bell sense, even among Everettians; I give examples above. The 'Parallel Lives' paper that you linked explicitly contrasts their picture with the "highly nonlocal — and very popular ([Bub & Bub], pp. 119–121) — naïve version of its many-worlds counterpart according to which the entire universe would split whenever Alice pushes a button". Perhaps we should re-label this objection the "Parallel Lives loophole", since it is fundamentally the same as the D&H theory - and almost as bizarre as superdeterminism. PaddyLeahy (talk) 00:44, 17 July 2023 (UTC)

"naive" is the key word there, it is referring to the popular (folk) understanding of MWI, not the one elaborated by Deutsch, Wallace etc. So that's hardly relevant Joncolvin (talk) 00:56, 17 July 2023 (UTC)

Deutsch and Wallace take opposite positions on this point, as noted above. PaddyLeahy (talk) 01:22, 17 July 2023 (UTC)

Don't think that's true. Timpson and Brown are clear they there is NO non locality in MWI https://arxiv.org/abs/quant-ph/0212140 , and Wallace and Timpson refer to this result in https://arxiv.org/abs/0907.5294 Joncolvin (talk) 01:53, 17 July 2023 (UTC)

regardless of whether *you* think it's untenable, there is demonstrably no consensus that Bell context local realism does not obtain in MWI. Per NPOV the article should not be saying that there is such consensus. Joncolvin (talk) 00:50, 17 July 2023 (UTC)

Ok, latest edit refering to "local hidden variable theories" is fine. Joncolvin (talk) 01:56, 17 July 2023 (UTC)

Who cares? There is a consensus that Bell's theorem does not apply to MWI, and that is all that needs to be said for this article. Roger (talk) 03:49, 17 July 2023 (UTC)

But then we would have nothing to argue about Joncolvin (talk) 06:13, 17 July 2023 (UTC)