Bell's Inequality: Must we ditch locality, realism or something else?

[DrChinese]

Probably referring to the ion trap experiment,

http://www.nature.com/nature/journal/v409/n6822/abs/409791a0.html

That's the one.

 

[DrChinese]

And specifically what QM predictions does it not match and why would the De Raedt et al, simulation need to match *all* of the predictions of QM? It is a classical model. One matching prediction should suffice.

Fred
moderator sci.physics.foundations

Ah, you are miles off the mark. As I say, this is a complex subject and I have found from experience that my time is likely wasted trying to convince you that your characterization is incorrect. I think it does a grave disservice to other readers who are not aware that your personal opinion is not consensus science. So let me point out that my objective is not to change your mind, just how you phrase what you do write.

This being a moderated board, and your written statements being non-standard is the issue. Posters here have restrictions here that should be followed. Not sure what the standards are at sci.physics.foundations. Although I am not a moderator, I will point out those elements of your comments that I consider inappropriate.

I will explain some of the relevant issues in a follow-on post.

 

[JeffKoch]

My perspective is as a physicist who knows quite a bit about some areas of physics, and not so much about others, but who knows a great deal about the concept of what I know and understand and what I don't. If I don't understand something, I know how to ask and perhaps answer questions so that I do understand it, and I can even get a pretty good feel for how far away I am from understanding it, and therefore how intrinsically hard to understand it is in comparison to other things that I do understand.

Bell's inequality is a mess, and I wager that few people who claim to understand it really comprehend the magnitude of the problem - I sure don't understand it, and would classify it as one of those really intrinsically hard things to understand. It's fairly simple in the sense of the second amendment to the U.S. constitution, a single sentence that has been analyzed and reanalyzed absolutely to death for over 200 years with no end in sight. And therein lies the complexity, because you immediately start asking questions about what it really means, and inevitably someone goes to an example of what it means, and perhaps another, and you think about what hidden conditions are or are not required, and how close any real experiment can get to a mathematical idealization, and pretty soon you're lost. So, my fallback is to ignore everything but data - computational results can be very interesting and can encourage experimentation, but carefully obtained data, analyzed carefully with an understanding and quantification of assumptions and possible errors, has to lead the way. That's where these sorts of internet discussions become pointless, though perhaps interesting as long as they remains civil and courteous - there appears to be an interesting avenue for future work here, but from my albeit limited reading nothing has been proven or disproven.

And that's how science works - someone looks at something in a different way, publishes a result, and other people follow with other research that might or might not move consensus science in a different direction.

 

[harrylin]

Bell's theorem is generally thought to show that the world cannot be both local and real.

http://en.wikipedia.org/wiki/Bell's_theorem

In simplistic terms, Bell derives an inequality which allegedly must be satisfied if the world is both local and real. In practice, it is found in numerous experiments that Bell's inequality is actually violated - leading to the conclusion that either the locality assumption, or the reality assumption, (or both) must be rejected.

But the following paper allegedly provides a counterexample - a hypothetical situation which involves assumptions that are most definitely both local and real, and yet the scenario described would also violate Bell-type inequalities if analysed in a manner similar to that used for Bell's theorem.

http://rugth30.phys.rug.nl/pdf/aipqo0-KRM.pdf

Conclusion: Violation of Bell-type inequalities does not necessarily always imply that either locality or realism assumptions are incorrect?

Comments?

That conference paper refers to their journal paper "Extended Boole-Bell inequalities applicable to quantum theory". Recently we have been discussing - without coming to a clear conclusion - that paper here:

https://www.physicsforums.com/showthread.php?t=499002

One thing is clear however: many experiments that are claimed to prove "non-locality" in fact fail to do so.

 

[harrylin]

[..]
Second, I have studied the de Raedt et al simulation, and in fact it does as you describe. However, it does not follow QM in that the basic Malus rule is not followed. [..]

That claim of yours is denied by De Raedt et al: according to them, their simulations reproduce QM - incl. the basic Malus rule. Probably you refer to an old simulation (possibly for another purpose) that did not include that rule.

 

[harrylin]

[..] This being a moderated board, and your written statements being non-standard is the issue. Posters here have restrictions here that should be followed. Not sure what the standards are at sci.physics.foundations. Although I am not a moderator, I will point out those elements of your comments that I consider inappropriate.

I will explain some of the relevant issues in a follow-on post.

The pot is calling the kettle black? The standards of this board can be found here:
https://www.physicsforums.com/showthread.php?t=414380

 

[zonde]

It won't help the time coincidence problem. See De Raedt et al's, model and computer simulation.

http://rugth30.phys.rug.nl/dlm/

According to De Raedt model you can't match all photon detections if you do not change coincidence time window.
So we should not be able to get coincident detection rate up by using efficient detectors.

So basically model predicts that you can't satisfy conditions that are necessary to close detection loophole.

 

[DrChinese]

According to De Raedt model you can't match all photon detections if you do not change coincidence time window.
So we should not be able to get coincident detection rate up by using efficient detectors.

So basically model predicts that you can't satisfy conditions that are necessary to close detection loophole.

That's one issue, which should be obvious is a serious problem with the simulation.

 

[DrChinese]

That claim of yours is denied by De Raedt et al: according to them, their simulations reproduce QM - incl. the basic Malus rule. Probably you refer to an old simulation (possibly for another purpose) that did not include that rule.

The de Raedt work is well known to those in the field, and it is NOT generally accepted as overturning Bell in any way. Which, as I have said, is the point. A lot of work is published in various places, and I would not call their work as being published in a suitable peer reviewed Physics journal. As ZapperZ already said, that would be big news (if Bell were overturned). According to Wikipedia, well-written words indeed

No physical theory of local Hidden Variables can ever reproduce all of the predictions of Quantum Mechanics.

Please note the word ALL in the above. So far, none has been presented. And if it were, it could potentially be falsified by experiment.

 

[JeffKoch]

No physical theory of local Hidden Variables can ever reproduce all of the predictions of Quantum Mechanics.

This is exactly my earlier point - you have not stated Bell's inequality, you have tried to use words to describe your interpretation of what Bell's inequality means.

 

[DrChinese]

This is exactly my earlier point - you have not stated Bell's inequality, you have tried to use words to describe your interpretation of what Bell's inequality means.

Yes, true. This is why it is important. And this harkens back to the ideas of the 1935 EPR paper, which asserted that a more complete description of physical reality is possible. It isn't, as Bell shows. More specifically, the Heisenberg picture is the limit.

 

[DevilsAvocado]

I have studied the De Raedt et al, papers extensively. We had a big discussion of them on the sci.physics.foundations newsgroup. It appears that they have successfully invalidated the EPRB type experiments with photons. If the time coincidence window is taken out of the experiments, then the experiments do not produce the QM results.

Is this a joke? How could you expect to get anything useful out of EPRB photon experiments without coincidence counting?

They also have a successful computer simulation that does produce QM results when a time coincidence window is used and produces Bell's results when it is not used

This is confusing? They invalidated real EPRB photon experiments, and real QM results, by removing coincidence counting. Then they construct a computer simulation, which is supposed to reproduce real QM results, by reintroducing coincidence counting...

What are you trying to say??

it is not supposed to be possible for a computer simulation to produce QM results

Sure, just plain logic and common sense. To produce true computer simulation of QM results, you need a quantum computer, right? But this is not available to De Raedt. Nevertheless you claim this is exactly what De Raedt et al have successfully achieved, right? So, what’s your explanation? They invented a groundbreaking shortcut for quantum computing!? They invented a new ingenious programming language (QM++) as a shortcut to quantum computing!? Or, this is not a true simulation of the EPRB/QM results, just a trial & error + fine-tuning to mimic 'desirable' results...?

(The last is probably something that even I could pull off, given some time...)

IMHO, they have shown that Bell's theorem does not and can not match physical reality.

There must be some misunderstanding. Bell's theorem is a purely abstract mathematical theorem stating that:

No physical theory of local hidden variables can reproduce all of the predictions of quantum mechanics.​

If you are claiming there’s something wrong with the physical reality, you have to go after QM. Bell's theorem has nothing, absolutely nothing, to say about the physical reality. All Bell's theorem say, is that the predictions of QM [if correct] are not compatible with LHV.

The proof of a mathematical theorem cannot involve experiments or other empirical evidence in the same way such evidence is used to support scientific theories (like QM). The key attribute for a scientific theory is that it is falsifiable, that is, it makes predictions about the natural world that are testable by experiments.

Hence, if you feel there’s something wrong here, you have to refute the predictions of QM, not Bell's theorem.

Good luck!

Not surprising since Joy Christian has also "disproved" Bell's theorem.

Not much is surprising about Joy Christian, is it? Maybe you could http://www.science20.com/alpha_meme/quantum_crackpot_randi_challenge_help_perimeter_physicist_joy_christian_collect_nobel_prize-79614" ?

Disproved is in quotes because you can't really disprove a mathematical theorem but what he has disproved is that Bell's theorem doesn't match physical reality same as De Raedt et al. Basically, Bell and its variants missed that you have to match pairs up in time.

Eh, you agree about the 'nature' of theorems... then you contradict yourself in mumbling about a "physical reality" and "match pairs up in time"??

Exceptionally confusing...

Again; Bell's theorem has nothing, I repeat, NOTHING to say about the "physical reality", "match pairs up in time", "African doctors", "epidemics", etc, etc, etc.

ALL it says is that [the predictions of] QM is not [mathematically] compatible with LHV.

That’s ALL there is. NO 'mathematical' description on HOW to perform physical experiments, nothing about the validation of QM, nothing about the nature of reality, nothing about African doctors, etc, etc.

Capiche?

For some years now, I’ve seen folks presenting the most hair-raising 'refutations' one can ever think of, including tons of cranky PDF’s, at the most quasi-complex level possible, when the fact is that the whole thing is very simple and beautiful.

The premise that Albert Einstein and Niels Bohr had during their 20 years long debate about the EPR paradox, was (naturally):

1 + 1 = 2

This is the natural/classical assumption we all have, and the only thing LHV/Local Realism can ever 'produce'.

Finally in 1964 John Stewart Bell showed how to once and for all settle the matter by formulating the complete predictions of QM:

1 + 1 = 3

This is ALL that John Bell told us.

(And this 'little' fact will probably change everything in our everyday life, in the future...)

This, is in comparison to all the other over-extensive mumbo-jumbo and about impossible triplets from pairs, epidemics, African doctors, etc, is just hilarious.


Thanks for the attention!

/DA
conservator entangled.states.of.avocados

 

[DevilsAvocado]

It won't help the time coincidence problem. See De Raedt et al's, model and computer simulation.

http://rugth30.phys.rug.nl/dlm/

Looks like the time coincidence problem is a problem only for De Raedt et al:

http://arxiv.org/abs/quant-ph/0703120

Comment on "A local realist model for correlations of the singlet state'' (De Raedt et al., Eur. Phys. J. B 53: 139-142, 2006)
Michael Seevinck, Jan-Ake Larsson

Abstract: De Raedt et al. (Eur. Phys. J. B 53: 139-142, 2006) have claimed to provide a local realist model for correlations of the singlet state in the familiar Einstein-Podolsky-Rosen-Bohm (EPRB) experiment when time-coincidence is used to decide which detection events should count in the analysis, and furthermore that this suggests that it is possible to construct local realistic models that can reproduce the quantum mechanical expectation values. In this letter we show that these conclusions cannot be upheld since their model exploits the so-called coincidence-time loophole. When this is properly taken into account no startling conclusions can be drawn about local realist modelling of quantum mechanics.

Journal reference: Eur. Phys. J. B 58, 51-53 (2007)
DOI: 10.1140/epjb/e2007-00194-3

 

[FrediFizzx]

Looks like the time coincidence problem is a problem only for De Raedt et al:

http://www.arxiv.com/abs/0706.2957

Reply to Comment on "A local realist model for correlations of the singlet state"

"The general conclusion of Seevinck and Larsson is that our model exploits the so-called coincidence-time loophole and produces sinusoidal (quantum-like) correlations but does not model the singlet state because it does not violate the relevant Bell inequality derived by Larsson and Gill, since in order to obtain the sinusoidal correlations the probability of coincidences in our model goes to zero. In this reply, we refute their arguments that lead to this conclusion and demonstrate that our model can reproduce results of photon and ion-trap experiments with frequencies of coincidences that are not in conflict with the observations."

 

[FrediFizzx]

De Raedt et al, have a new paper on arXiv,

http://www.arxiv.com/abs/1108.3583
"Hidden assumptions in the derivation of the Theorem of Bell"

Abstract:
"John Bell's inequalities have already been considered by Boole in 1862. Boole established a one-to-one correspondence between experimental outcomes and mathematical abstractions of his probability theory. His abstractions are two-valued functions that permit the logical operations AND, OR and NOT and are the elements of an algebra. Violation of the inequalities indicated to Boole an inconsistency of definition of the abstractions and/or the necessity to revise the algebra. It is demonstrated in this paper, that a violation of Bell's inequality by Einstein-Podolsky-Rosen type of experiments can be explained by Boole's ideas. Violations of Bell's inequality also call for a revision of the mathematical abstractions and corresponding algebra. It will be shown that this particular view of Bell's inequalities points toward an incompleteness of quantum mechanics, rather than to any superluminal propagation or influences at a distance."

Fred

 

[Delta Kilo]

De Raedt et al, have a new paper on arXiv,

http://www.arxiv.com/abs/1108.3583
"Hidden assumptions in the derivation of the Theorem of Bell"

Oh noes...

Bell also introduced a variable \lambda to be discussed in detail below and constructed from these facts an inequality equivalent to:
A_a(\lambda)B_b(\lambda) + A_a(\lambda)B_c(\lambda) + A_b(\lambda)B_c(\lambda) \le +1 (4)

[The rest of the paper is concerned with finding flaws and assumptions of eq (4)]

I wonder, is there any point to say AGAIN that this is NOT Bell's inequality? That the final form of Bell's inequality does NOT have \lambda in it and therefore does NOT imply that three different measurements must be somehow obtained simultaneously for the same \lambda?

 

[ThomasT]

Bell's theorem has nothing, absolutely nothing, to say about the physical reality.

This is my current understanding as well. I also want to say that it's now the mainstream view, but I don't know that it is (or that it isn't).

All Bell's theorem says, is that the predictions of QM [if correct] are not compatible with LHV.

We can omit the "if correct" caveat, and I'd put it something like this: Bell's theorem says that any LRHV model/account/description/theory of quantum entanglement which is based on restrictions placed on the formulation of LRHV models of entanglement by Bell, is incompatible with the standard QM treatment of quantum entanglement.

This has been proven mathematically, and isn't disputed as far as I know.

Bell inequalities, which have been tested experimentally, are based on the restrictions that Bell, in his formulation, placed on any LRHV model of quantum entanglement. If Bell's formulation is indeed general, then the experimental results indicate that any and all LRHV accounts of quantum entanglement are ruled out.

Some researchers, such as Joy Christian, have proposed 'LRHV' formulations that contradict the generality of Bell's formulation. I use the scare quotes here because it isn't readily evident that Christian's formulation can be regarded as an LRHV formulation. That is, it's rather mathematically exotic, and I don't think it fits with what most people, including Einstein, would consider an LRHV model of quantum entanglement. So Christian's proposal gets set aside (ie., considered as essentially nonrealistic) until somebody can clearly explain why it should be considered an LRHV model of quantum entanglement -- which Christian himself doesn't seem to be able to do.

While we seem to agree on the meaning and significance of Bell's theorem, I'm not sure what you mean by the following:

Finally in 1964 John Stewart Bell showed how to once and for all settle the matter by formulating the complete predictions of QM:

1 + 1 = 3

??

Anyway, what remains, imo, is for somebody to give a really clear explanation of exactly why viable LRHV models of entanglement (per Bell's restrictions) are impossible, and why Bell's theorem (ie., experimental violation of Bell inequalities) tells us nothing about the reality underlying instrumental behavior that we can't infer without it.

Maybe the latest paper by Hess, De Raedt, and Michielsen does this. I don't know, I haven't read it yet -- and it might take a while for me to understand it. In any case, I have my own understanding of the Bell theorem interpretation conundrum which tells me, quite simply, why LRHVs, per Bell, are ruled out and why BIs don't inform wrt underlying reality.

Bottom line, imo, there's a very reasonable local realistic 'understanding' of quantum entanglement which can't be viably formulated as an explicitly LRHV model of quantum entanglement. That is, the LR program is dead even though we can retain the assumption that quantum entanglement is entirely due to common causes and local (c, and sub c) interactions and transmissions. A sort of Catch-22 for diehard local realists.

And Bell's mathematical demonstration of the incompatibility of LRHV with QM is entirely correct, even though the discussion surrounding its meaning has become a bit 'cloudy' (as in JeffKoch's, it's "a mess") wrt certain commentators and interpreters.

 

[harrylin]

The de Raedt work is well known to those in the field, and it is NOT generally accepted as overturning Bell in any way. Which, as I have said, is the point. A lot of work is published in various places, and I would not call their work as being published in a suitable peer reviewed Physics journal. As ZapperZ already said, that would be big news (if Bell were overturned). According to Wikipedia, well-written words indeed

No physical theory of local Hidden Variables can ever reproduce all of the predictions of Quantum Mechanics.

Please note the word ALL in the above. So far, none has been presented. And if it were, it could potentially be falsified by experiment.

I notice that you backed off from your unverified claim about the Malus law - good!

Now, De Raedt et al's physical model of local Hidden Variables reproduced experimental results that were thought to be impossible for that kind of theory; so, it is perhaps not the "right" model, but it's surely a demonstration of the fact that in science we should never be dogmatic.

 

[zonde]

That's one issue, which should be obvious is a serious problem with the simulation.

That's not a problem that's a feature.
Obviously any LHV model that would aim at explaining experimental results of Bell tests should deviate from idealized conditions implied by Bell theorem i.e. it should exploit some loophole.

That's one reason why Joy Cristian model can be ignored until it becomes clear what loophole it is exploiting.

But if we talk about de Raedt model there are experiments that show very good correspondence between coincidence count rates of downconverted photons and quantum efficiency of photon detectors. There is nice overview about this topic:
"physics.nist.gov/Divisions/Div844/publications/migdall/apopts41.pdf"[/URL]

And that would mean that coincidence rates should be lowered when we insert polarizers in the photon path.

But I would say that there are other problems with the model so that experimental falsification of this model is not really necessary IMHO.

 

[harrylin]

[..]
This is confusing? They invalidated real EPRB photon experiments, and real QM results, by removing coincidence counting. Then they construct a computer simulation, which is supposed to reproduce real QM results, by reintroducing coincidence counting...

No, they successfully simulate the results of a class of real experiments as analysed with variable time windows; and there is nothing confusing about it if you run the demo simulation:
http://rugth30.phys.rug.nl/eprbdemo/

If you think that the simulated real experiments invalidate "real QM results", please clarify.

To produce true computer simulation of QM results, you need a quantum computer, right? [..]

Of course not - a simulation is a calculated imitation of a physical process, it isn't the real thing. For simulations you only need to make many calculations.

There must be some misunderstanding. Bell's theorem is a purely abstract mathematical theorem stating that:

No physical theory of local hidden variables can reproduce all of the predictions of quantum mechanics.​

To claim that a statement about physical theory is "purely abstract mathematical" is a contradiction in terms...

[..] Finally in 1964 John Stewart Bell showed how to once and for all settle the matter by formulating the complete predictions of QM:

1 + 1 = 3

This is ALL that John Bell told us.

??