Is the moon there if nobody looks? Bell inequalities and physical reality

[Richard Gill]

Dear all

I would be interested to hear group members’ responses to the latest shots fired in a long-running battle between, on one side, Marian Kupczynski, and on the other side, myself and Justo Pastor Lambare (from Paraguay). 

See https://www.qeios.com/read/DRHFO9 for his recent evaluation of one of my recent writings.

And check out his own latest published paper, see below.

Richard

arXiv:2302.08061 quant-ph physics.hist-ph; published in Frontiers in Physics:

https://doi.org/10.3389/fphy.2023.1117843

Response: Commentary: Is the moon there if nobody looks? Bell inequalities and physical reality


Authors: Marian Kupczynski


Abstract: We reject unjustified criticism of our published article [2209.07992] by Gill and Lambare [arXiv:2211.02481, arXiv:2208.09930]. They completely misinterpret the content and conclusions of this article. They construct a counterfactual probabilistic model in which random variables representing outcomes of four experiments performed using incompatible experimental settings are jointly distributed. Thus, CHSH inequalities trivially hold for all finite samples generated by their model. Their model defines a probabilistic coupling for our model describing only the raw data from Bell tests. The existence of this coupling does not invalidate the derivation of the contextual probabilistic model describing the final data from Bell tests. Only these final data are used to test Bell inequalities. Inequalities cannot be derived because our model violates statistical independence. Our contextual model allows to explain in a local and causal way the violation of inequalities and the apparent violation of no-signaling reported in these experiments.


Submitted 15 February, 2023; originally announced February 2023.


Comments: 3 pages, 3 equations and 33 references


Journal ref: Front. Phys., 13 February 2023, Sec. Statistical and Computational Physics Volume 11 - 2023 

[Jan-Åke Larsson]

Kupczynski's model is nonlocal.

/Jan-Åke

--
You received this message because you are subscribed to the Google Groups "Bell inequalities and quantum foundations" group.
To unsubscribe from this group and stop receiving emails from it, send an email to Bell_quantum_found...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/Bell_quantum_foundations/CA46B1AA-6ED9-4442-987E-5AD6073438E9%40gmail.com.

--
Jan-Åke Larsson
Professor, Head of Department

Department of Electrical Engineering SE-581 83 Linköping Phone: +46 (0)13-28 14 68 Mobile: +46 (0)13-28 14 68 Visiting address: Campus Valla, House B, Entr 27, 3A:512 Please visit us at www.liu.se

[Alexandre de Castro]

This Marian Kupczynski's commentary [https://doi.org/10.3389/fphy.2023.1117843] was edited by Theo Nieuwenhuizen and reviewed by me.

Its fundamental point is Marian's question:

"Can we explain the data from Bell tests without evoking quantum non-locality and quantum magic?"

My answer has been YES. I have no doubt that there is a loophole in the theory because it is possible to show that the probabilities derived by applying the Born rule are compatible with local models of probability distribution, which makes Bell tests inconclusive.

Alexandre

--

[Jan-Åke Larsson]

Marians answer is WRONG. You should not have let this through.

/Jan-Åke

To view this discussion on the web visit https://groups.google.com/d/msgid/Bell_quantum_foundations/CAPcOYa%3D059gAtEbMO%2BbGC6hwi3S1Zma66-wdNJ0KMWU8fUJHCw%40mail.gmail.com.

[Richard Gill]

Hi Alexandre

Yes, I know the paper was reviewed by you and Theo. You both missed the mistakes in the paper. Justo Pastor Lambare and I have written a short refutation (already published in Frontiers), and a longer one (under review for a different journal). So far, Marian has complained but not shown us mistakes in *our* maths. I’m afraid that he does not understand Fine’s theorem and is quite confused about many other issues.

He is building on earlier papers, repeating what he thinks he proved before, but slightly changing the definitions and the assumptions. I think it has evolved into a big muddle. I believe he has suffered from a lot of poor reviewing and too fast publication of too many special issues of not very high quality journals.

If you can see any mistakes in our maths, I would like to know, so I can retract my papers and apologise to Marian.

I have emailed to Theo about this but get no reply. Please forward this email to him. Maybe he will read emails from you.

Richard

Our papers:

https://www.frontiersin.org/articles/10.3389/fphy.2022.1024718/full

https://arxiv.org/abs/2208.09930

Marian of course has immediately published a response to our Frontiers paper in Frontiers, and longer responses on arXiv.

[Richard Gill]

Well, Kupczynski's model changes from time to time.

His earlier models were definitely local (and naturally, satisfied Bell CHSH inequalities). He said that it was not allowed to prove them, but we did prove them. 

His most recent models are non-local.

Why the change? Well, at some point he had local models but needed the detection loophole to “fake” quantum correlations. (The name “detection loophole” was changed to “photon detection loophole” in order to give a scientific motivation for this). So in his absolutely local model, the outcomes X, Y lie in {-1, 0, 1} and he looks at the correlations between X and Y conditional on settings A =a, B = b, and XY .ne. 0. But conditional on XY .ne 0, the joint distribution of the fourtuple (A, B, X, Y) changes. Variables which initially were uncorrelated can become correlated. Settings which were independent can become dependent. The correlations can become signalling, ie Alice’s outcome can depend on Bob’s setting. 

Kupczynski has noticed this fact of probability theory in actual data, that is why he transitions to a blatantly non-local model because he think it is necessary in order to model the actual data of actual experiments. But the actual data he is now modelling is the post-selected data got by removing all particle pairs with non-detections. This is no longer done in good experiments. So he’s a bit behind the times.

At least, that’s how I see it right now. 

[Jan-Åke Larsson]

Oh, absolutely, he is behind. I agree he used to use the detection loophole earlier.
I've explained my 1998-1999 papers that describe the detection loophole to him on several occasions.
Complete with a local realist model that reproduces the maximal violation of CHSH for each value of the detection probability, and an upper bound for the violation for each value of the detection probability.
Perhaps some of that actually reached through.

I was talking about his present model, which is explicitly nonlocal. He claims it is local but it is not.

/Jan-Åke

[Kupczynski, Marian]

Dear Richard and Jan Ake

I knew, that Richard do not understand my papers, but I thought that you Jan-Ake , you should. Correlations in all EPRB scenarios are by definition nonlocal as you like to insist ( and I agree) , but it does not mean that they cannot be explained without spooky influences, retrocausality or  superdeterminism.

 My model is neither local or non-local.    The outcomes are created in a locally deterministic way, but the variables describing distant measuring set-ups may be correlated. My model is contextual and not local realistic  It violates statistical independence. In some Bell - Tests suffering, as you like to say from the detection loophole and an apparent violation of no-signalling the violation of statistical independence is due to setting dependent data post-selection. In Munich experiment one may give other plausible arguments, which do do require direct inflences between distant  measuring set-ups. I am working on it.

The data with disturbance referred above, are deduced from a contextual model M1 describing  the raw data in these experiments. G-L constructed a probabilistic coupling only for this model but this construction is completely irrelevant for the model describing the final data  and for the conclusions of my paper(s). I never told thattheir construction was wrong. I told that I could not derive CHSH inequalities directly using thea ssumptions of my model M1.

Saying that my papers and that my crticism of G-L papers and their defamatory statements  are correct: is misleading and also defamatory.

Enough is enough. Richard and Jan-Ake , please stop  dissemnation of false and defamatory statements . Stop also harassing the reviewers of my paper in  FRONTIERS and a referee of  my reply to  G-L Commentary ( sorry Theo,  Alexander and others). If you Jan-Ake incorrectly believe,  that my particular paper is wrong plese send  me a detailed criticism line by line pointing out what was wrong in my paper. If not,  please stop your one line comments, because they are false , misleading and defamatory.  

I am sorry, that Richard and Jan-Ake are filing your mailboxes with messages you did not ask for.  If you want to have an opinion about my papers, you dont need Richard or Jan -Ake to tell you what they are about ,just read them and if you have comments please send them to me privately.  

I am not going to answer any future  messages from Richard and Jan Ake . No matter, what they are going to post about my papers. I just advice not react to it. Just read my papers and plese contact me directly, if you have some comments. If Jan- Ake and Richard want to discuss with me, they will have enough time in Vaxjo. It would be more descent  than what they are doing now.  

Best greetings from Canada

Regards,

Marian

---

From: bell_quantum...@googlegroups.com <bell_quantum...@googlegroups.com> on behalf of Richard Gill <gill...@gmail.com>
Sent: Tuesday, May 9, 2023 12:25 PM
To: Jan-Åke Larsson <jan-ake...@liu.se>
Cc: Bell Inequalities and quantum foundations <Bell_quantum...@googlegroups.com>; Alexandre de Castro <alx...@gmail.com>
Subject: Re: [Bell_quantum_foundations] Is the moon there if nobody looks? Bell inequalities and physical reality

 

Attention: L’émetteur de ce courriel est externe à l’Université du Québec en Outaouais.

To view this discussion on the web visit https://groups.google.com/d/msgid/Bell_quantum_foundations/DBC44AE1-E4DC-4431-8813-E7D25CD54535%40gmail.com.

[Jan-Åke Larsson]

Marian, Stop this nonsense.

Your model is explicitly nonlocal.

I do not care that you *believe* otherwise.

I have been very decent to you, only discussing scientific points.

But you do not listen.

/Jan-Åke

[Jan-Åke Larsson]

In clear text to you others that might read this:

Marian's most recent model has different probability spaces for different pairs (A,B), (A,B'), (A',B), (A',B') of settings at Alice and Bob.

In this model, the outcome for Alice's measurement for setting A must either be from the probability space for (A,B), or for (A,B').

- If Alice's outcome for setting A is allowed to depend on whether Bob used B or B', then the model is nonlocal (and can violate CHSH). Similar for A' at Alice and B and B' at Bob.
- If Alice's outcome for setting A is the same independent of if Bob used B or B', and similar for A' at Alice and B and B' at Bob then the model is local (and cannot violate CHSH).

Marian uses the first alternative and at one point called this "local contextual", now apparently only "contextual". It is contextual in the Kochen-Specker sense so this is not wrong. Unfortunately, the spatial separation also makes the model nonlocal because of the dependence of an outcome on the remote setting.

Marian's claim that using four disjoint probability spaces would invalidate this reasoning is groundless. The model is manifestly nonlocal.

/Jan-Åke

To view this discussion on the web visit https://groups.google.com/d/msgid/Bell_quantum_foundations/caf2bdf1-f2ba-7c85-2954-5dccf834ba75%40liu.se.

[Kupczynski, Marian]

Dear Jan-Ake,

It is not a nonsense, what I am saying. Did you read my explanation. Evidently we define differently what nonlocal is. Perhaps, if for you it is  Bell-nonlocal, then  I agree. 

 I have to admit,  that in my paper : .Frontiers | Is the Moon There If Nobody Looks: Bell Inequalities and Physical Reality (frontiersin.org)

Is the Moon There If Nobody Looks: Bell Inequalities and Physical Reality Bell-CHSH inequalities are trivial algebraic properties satisfied by each line of an Nx4 spreadsheet containing ±1 entries, thus it is surprising that their violation in some experiments allows us to speculate about the existence of non-local influences in nature and casts doubt on the existence of the objective external physical reality. Such speculations are rooted in incorrect interpretations of quantum mechanics and in a failure of local realistic hidden variable models to reproduce quantum predictions for spin polarization correlation experiments (SPCE). In these models, one uses a counterfactual joint probability distribution of only pairwise measurable random variables (A, A′, B, B′) to prove Bell-CHSH inequalities. In SPCE, Alice and Bob, using 4 incompatible pairs of experimental settings, estimate imperfect correlations between clicks registered by their detectors. Clicks announce the detection of photons and are coded by ±1. Expectations of corresponding random variables—E (AB), E (AB′), www.frontiersin.org

In equations (26-27) the normalisation constants are missing. Right hand sides should be divided by C_xy= P( Ax no 0 and  and By no 0) . It does not change the conclusions of the paper because there is no quantitative use of these Equations in this article.   I will post Erratum in Frontiers and also in other journals, if in my articles published there  similar constants were missing. Lack of these constants did not have any impact on my conclusions.   Thank you Richard for noticing it.  

I repeat that G-L criticism was of a different nature and it was irrelevant. 

Best 

Marian

---

From: Jan-Åke Larsson <jan-ake...@liu.se>
Sent: Tuesday, May 9, 2023 3:10 PM
To: Kupczynski, Marian <Marian.K...@uqo.ca>; Richard Gill <gill...@gmail.com>

[Jan-Åke Larsson]

Dear Marian,
Perhaps we do agree then. Bell-nonlocal it what I am after.

However, I would hesitate to call anything "local" where the probability space itself depends on the remote setting. For me, that means the pair source must have access to both settings at pair emission. This is a nonlocal influence that makes Alice's outcome depend on Bob's setting.

Best regards,
Jan-Åke

[Mark Hadley]

Epr is explicitly designed so that contextuality has to be non local.

To view this discussion on the web visit https://groups.google.com/d/msgid/Bell_quantum_foundations/YT3PR01MB56520986B06915AF2397903F9A769%40YT3PR01MB5652.CANPRD01.PROD.OUTLOOK.COM.

[Jan-Åke Larsson]

Not EPR, but Bell's setup.

The EPR setup on its own is not enough, there exists a local hidden variable model for that, see Bell 1984. Measurement at equal settings is not enough, you need the Bell inequality that uses a chain of measurement settings at skewed intervals. Perhaps I should write a paper that points this out in its entirety.

 /JÅ

[Kupczynski, Marian]

Dear Jan -Ake

I am splitting a probability space into: variables describing locally distant system, which are  statistically dependent and variables describing local  measuring set-ups ( contextuality) which are statistically dependent: the cos(theta)  dependence of observed correlations may only be explained using sppecific assumptions on probabilty distributions of instrument variables. 

Since : a=A ( lambda_1, lambda_x)  b=B ( lambda_2, lambda_y) I am calling this  a local determinism. A statistical dependence between  lambda_x and lambda_y introduces                      " nonlocality" . Bell "causal locality" which is violated  it is a different assumption, see nice discussion in Wiseman : Two Bell Theorems.  

Marian

---

From: Jan-Åke Larsson <jan-ake...@liu.se>
Sent: Tuesday, May 9, 2023 4:03 PM

[Jan-Åke Larsson]

Dear Marian,
It is not locally deterministic because the probability space itself needs both settings to be well-defined. That there is no further dependence on the setting within the space(s) is irrelevant.

Best
Jan-Åke

[Richard Gill]

Dear Marian and all

Consider a Bell type experiment when outcomes X, Y take values in {-1, 0, 1} and settings A,B in {1, 2}. Bell locality with measurement independence (corresponding to classical statistical causality à la Pearl book “Causality”) says that there exists U such that: (A, B) is statistically independent of (X, Y, U), and such that conditional on (A,B) = (a,b), X and Y are statistically independent given U, and that the law of X given A, B, U does not depend on B, the law of Y given A, B, U does not depend on A.

All CHSH inequalities hold, and no-signalling holds.

When you normalise a joint distribution of (A, B, X, Y) by conditioning on the event XY .ne. 0 everything can change. The joint distribution of the settings (A, B) can change, the joint conditional distribution of outcomes (X, Y) given (A,B) obviously can change dramatically. CHSH inequalities, and indeed Tsirelson inequality too, can become violated. The violation is possibly so extreme that one can have CHSH = 4.

Independence of settings can become dependence. No-signalling can become signalling.

Note that in “Bertlmann’s socks” Bell explicitly rules out the case of outcomes being able to be zero. In fact he moves to a three party experiment with a third party Charlie.

Charlie has no setting choices. One could say there is only one possible value c0 of Charlie’s setting. Charlie has *outcome* Z and one of its values, say z0, is distinguished.

Bell locality is there exists U such that the distribution of (X, Y, Z) conditional on (A, B, C, U) factors in the same way as in the two-party case.

Under Bell locality, CHSH inequalities for the distribution of X,Y given A, B, C and Z hold, for the post-selected ensemble where Z = z0. 

Richard

Sent from my iPad

On 9 May 2023, at 21:57, Kupczynski, Marian <Marian.K...@uqo.ca> wrote:



[Richard Gill]

PS about the issue of disjoint outcome spaces allowed to depend on setting and owner, I wrote a very short note on the consequences of Marian’s relaxation of setting dependence, something he has moved to in his most recent paper(s). The note is on arXiv, https://arxiv.org/abs/2305.03026. Marian’s favoured notation is hyper compact but dangerously ambiguous. I’m sorry to be pedantic about notation. Sometimes it is necessary. Maybe there are better ways to do the disambiguation than the way I came up with.

Abstract: Kupczynski (2023) claims that Gill and Lambare (2022a, 2022b) misrepresent several of his published papers. This paper shows that the latest version of his "contextuality by default" model of a Bell experiment places no constraints whatsoever on the statistics of observed results in Bell type experiments. It thereby effectively allows arbitrary non-locality, ie direct causal effects of local measurement settings on distant measurement outcomes.

Sent from my iPad

On 10 May 2023, at 01:39, Richard Gill <gill...@gmail.com> wrote:

Dear Marian and all

[Inge Svein Helland]

Dear Richard, dear Marian, dear all.

I have been asked by you, Richard to have a look upon the discussion between yourself and Marian Kupczynsky. I have done so, and the result is the attached note: A form of the Bell model proposed by Marian Kupczynsky must be abandoned if we insist upon a causal model.

It seems like the CHSH inequality follows. I also mention my own recent arguments on trying to explain why this inequality can be violated in practice.

Inge

---

From: bell_quantum...@googlegroups.com <bell_quantum...@googlegroups.com> on behalf of Richard Gill <gill...@gmail.com>

Sent: 10 May 2023 01:55:54
To: Kupczynski, Marian; Alexandre de Castro; Jan-Åke Larsson

Cc: Bell Inequalities and quantum foundations

Subject: Re: [Bell_quantum_foundations] Is the moon there if nobody looks? Bell inequalities and physical reality

--

You received this message because you are subscribed to the Google Groups "Bell inequalities and quantum foundations" group.
To unsubscribe from this group and stop receiving emails from it, send an email to Bell_quantum_found...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/Bell_quantum_foundations/5E309048-BE7C-4AF8-9F26-6904C058A1B2%40gmail.com.

[Alexandre de Castro]

Jan-Åke,
I reviewed Richard's comment in another journal [https://royalsocietypublishing.org/doi/full/10.1098/rsos.201909], and recommended its publication.  Marian is a renowned researcher, and his work deserves to be published too.

Alexandre

[Jan-Åke Larsson]

It is not the merits of the researcher, but the merits of the manuscript that are under consideration in the review process.

/Jan-Åke

[Richard Gill]

Ha ha, Alexandre, then you missed the serious errors in that paper too! Maybe time to read one of my papers carefully? Please, please, tell me if you see an error. I will retract or correct the paper and apologise profusely and publicly and sincerely to whoever was harmed by my error.

Even renowned researchers can make mistakes. Everyone makes mistakes. The best of referees can miss them. The final responsibility is with the author. 

[Richard Gill]

Sorry, delete last email from me (the one quoted below). I’m mixing up different people’s papers in different journals of the Royal Society. It’s getting too late in a too long day,.

[Alexandre de Castro]

no problem, Richard