About violation of Bell's inequalities

[RandallB]

Well, the ``non-locals'' are as lonely as the ``locals''

Are your kidding (or maybe fooling yourself)??
I don’t consider even one of the subgroups of Non-local; like QM, BM, MWI, Strings, etc. as even small let alone lonely!

Concerning Bell: …… (b) it is still possible to violate the inequalities though by a local realist theory with a clear interpretation (and no predeterminism is needed for that at all).

Again what “local realist theory” does that? I’ve never seen one by you or anywhere else that does so unless they distort their “clear interpretation” of local to include a non-local to bell version of it in the manner of BM, MWI etc.

As you indicate ‘predeterminism’ has nothing to do with Bell. The only issue with Bell is a fixed “determined” local value, set for both photons at the time they split apart, that remains intact for each till measured at some later time far away from each other, i.e. a true LHV.
(Above edited - Thanks DrC )

2nd edit: Sorry, may have given you too much credit for being logical in the above edit. DrC responses following are fine – I’d only be redundant to try and say more; except to say that you have me convinced – logically you cannot be a local realist, what I don’t know what, but not a local realist.

 

[DrChinese]

... working from a misunderstanding of ‘predeterminism’ & Bell. The philosophical ideas of a predestine future, or current result from the past, has nothing to do with Bell. The only issue with Bell is a fixed “determined” local value set for both photons at the time they split apart that remains intact for each till measured at some later time far away from each other, i.e. a LHV.

I agree RandallB, but I think Careful understood that clearly as well. His point, I believe, was that predestination had been brought up by another poster as being an additional requirement (i.e. part of local realism): it's not.

 

[Careful]

Are your kidding (or maybe fooling yourself)??
I don’t consider even one of the subgroups of Non-local; like QM, BM, MWI, Strings, etc. as even small let alone lonely!

No, I think you are kidding yourself ! Actually, most physicists would say QFT is local and that the only thing it is good for is calculating expectation values of local observables ; most of them feel very uncomfortable with the idea of non-locality lurking around the corner when going to a single event interpretation and do not want to consider the problem as such. There is however a big difference between this and acknowledging local realism, that is why it confuses most.

Again what “local realist theory” does that? I’ve never seen one by you or anywhere else that does so unless they distort their “clear interpretation” of local to include a non-local to bell version of it in the manner of BM, MWI etc.

True I never proposed one on this forum (I would first write it on paper ), but if you think about it, any theory containing more degrees of freedom than the ones we observe is capable of violating the Bell inequalities. The difficulty is to construct such natural theory, not to construct a theory in the first place. It is probably not necessary (and impossible) to recuperate the ideal singlet state ``predictions'', these are *clearly* wrong.

As you indicate ‘predeterminism’ has nothing to do with Bell. The only issue with Bell is a fixed “determined” local value, set for both photons at the time they split apart, that remains intact for each till measured at some later time far away from each other, i.e. a true LHV.
(Above edited - Thanks DrC )

I think I have explained this at least 10 times to you (I even referred you to Bell's book and the Clauser Horne Shimony comments ...) but nothing seems to help. Logically speaking ueit is correct, in a deterministic LOCAL world you cannot assume the detector settings to be random, that is uncorrelated from the source. Once such correlations occur, Bell's theorem goes down the drain.

Careful

 

[DrChinese]

Logically speaking ueit is correct, in a deterministic LOCAL world you cannot assume the detector settings to be random, that is uncorrelated from the source. Once such correlations occur, Bell's theorem goes down the drain.

Careful

You can't be serious. The detector settings can be adjusted fully independently (and randomly) and outside the light cone of the entangled particles. The results will still violate the Bell Inequality. To suggest otherwise would mean that there are so many new and previously unknown physical effects that it is ridiculous. For example:

Have 2 radioactive samples used to provide 2 random yes/no signals (by whatever mechanism you could imagine). Each would be placed distant to the respective detectors but also away from the source (so they are outside each other's light cones). The signals control one detector setting for the entangled particles at the source, which are then adjusted mid flight. You will still get a normal Bell test result when you run the experiment. Yet the radioactive samples have no way of knowing that they are supposed to emit a pulse in such a way as to yield suitable statistics for the observed correlations... unless there are new and previously unknown physics we are not aware of. If they exist, where are they? And why do they only show up in Bell tests? How does a radioactive sample - an aggregate of many particles - manage to operate in strict concert with a similar sample far away? And in concert with photons also emitted far away?

The proposed solution - strict predestination - is beyond implausible. And it explains exactly nothing; I am not sure it can even qualify as science. As I have said before, you may as well propose that the active hand of God is manipulating the results to make it appear as if the Bell Inequality is violated.

The fact is, no local realist should EVER invoke detector-detector interaction/communication as an explanation of Bell test results. It is logically inconsistent with a belief in Einsteinian locality.

 

[Careful]

You can't be serious. The detector settings can be adjusted fully independently (and randomly) and outside the light cone of the entangled particles. The results will still violate the Bell Inequality.

The problem with you is that you cannot simply say ``logically it is possible'', I did not say plausible. The thing is that in a deterministic theory, randomness does not exist and two past lightcones generally intersect each other. I made it clear I think that I deem such explanation to be unlikely, but it is still LOGICALLY possible.

To suggest otherwise would mean that there are so many new and previously unknown physical effects that it is ridiculous.

Well recent detailed analyses in nonlinear classical dynamical systems do indeed reveal the possibility of large scale synchronisation. Therefore, the suggestion that we do understand non linear classical dynamical systems is ridiculous ! We are just at the beginning of understanding those...

For example:

Have 2 radioactive samples used to provide 2 random yes/no signals (by whatever mechanism you could imagine). Each would be placed distant to the respective detectors but also away from the source (so they are outside each other's light cones). The signals control one detector setting for the entangled particles at the source, which are then adjusted mid flight. You will still get a normal Bell test result when you run the experiment. Yet the radioactive samples have no way of knowing that they are supposed to emit a pulse in such a way as to yield suitable statistics for the observed correlations... unless there are new and previously unknown physics we are not aware of.

Your very reasoning strictly denies predeterminism; in the latter it would be possible that particles on the flight already knew of the adjustments to be made. Again, I do not ``believe'' that but strictly speaking it is possible.

The proposed solution - strict predestination - is beyond implausible. And it explains exactly nothing; I am not sure it can even qualify as science. As I have said before, you may as well propose that the active hand of God is manipulating the results to make it appear as if the Bell Inequality is violated.

That is a difficult question, but non local theories aren't science either for quite similar reasons as you are concerned with here and neither is solipsism...

The fact is, no local realist should EVER invoke detector-detector interaction/communication as an explanation of Bell test results. It is logically inconsistent with a belief in Einsteinian locality.

No, it is not logically inconsistent : it is not because you believe in locality of interaction that no large scale correlations can appear; actually there are many places in classical physics where this happens. In case of the EPR setup, it is hard to swallow, because it involves smart electrons or very delicate (read unlikely) mechanisms. But almost any solution to the EPR paradox (hereby I mean appearant violation of Bell inequalities, not necessarily agreement with singlet state prediction) will go beyond the obvious.

Careful

 

[DrChinese]

The problem with you is that you cannot simply say ``logically it is possible'', I did not say plausible. The thing is that in a deterministic theory, randomness does not exist and two past lightcones generally intersect each other. I made it clear I think that I deem such explanation to be unlikely, but it is still LOGICALLY possible.

OK, I think I follow better now. I didn't think I had misunderstood your normal position to that extent!

Well, my view is this: I often see someone say that theory AA is "possible" as a counter to QM postulate A, and theory BB is "possbile" as a counter to QM postulate B, etc. Yet AA and BB are mutually inconsistent. So to seriously advance a local realist position against Bell, you must really be able to go the entire way.

I understand that it can be postulated that strict predetermininsm means that every effect shares a common source cause. But as I say, this is hardly a theory. For it to be advanced as a theory, it must be made logically consistent AND it must agree with experiment. As you can imagine, there are many difficulties in this regard as to details.

I personally doubt that it is possible to develop a such a theory. And more specifically, it would still fall under Bell's Theorem anyway IF Bell's assumptions of locality & realism end up present in that hypothetical theory. But I guess I can grudgingly acknowledge that a theory might be possible with strict predestination IF either locality or realism are dropped as assumptions (so you are not running afoul of Bell). This might be possible if a different definition of locality were used, for example.

Please note that Bell requires: that the detector setting at one spot is independent of the result at the other spot. Presumably this is negated when there is strict predeterminism.

 

[RandallB]

True I never proposed one on this forum (I would first write it on paper ), but if you think about it, any theory containing more degrees of freedom than the ones we observe is capable of violating the Bell inequalities.

I think I have explained this at least 10 times to you ... but nothing seems to help.
... Once such correlations occur, Bell's theorem goes down the drain.

Well of course none of that helps - every thing you suggest IMO ultimately comes down to “a theory containing more degrees of freedom than the ones we observe”!
At which point you are non-local and no longer the local realist you claim to be.
I find absolutely no foundation to accept “Bell's theorem goes down the drain” based on such non-local ideas.

I do like your point that a truly good proposed local idea should not be introduced publicly before putting on paper in a journal or at least under copyright. But when you are ready (is that why your grinning) to write that paper; what experimental test is there you can use to support a local theory??

 

[Careful]

I personally doubt that it is possible to develop a such a theory. And more specifically, it would still fall under Bell's Theorem anyway IF Bell's assumptions of locality & realism end up present in that hypothetical theory. But I guess I can grudgingly acknowledge that a theory might be possible with strict predestination IF either locality or realism are dropped as assumptions (so you are not running afoul of Bell). This might be possible if a different definition of locality were used, for example.

Please note that Bell requires: that the detector setting at one spot is independent of the result at the other spot. Presumably this is negated when there is strict predeterminism.

Right, why talk about it or say it CANNOT be ? Just each time someone comes by and says ``he I have a theory which reproduces the singlet state exactly'', then just take a look at it and see where the mistake is made. In the Petrov case, it was just too easy.

Careful

 

[Careful]

Well of course none of that helps - every thing you suggest IMO ultimately comes down to “a theory containing more degrees of freedom than the ones we observe”!
At which point you are non-local and no longer the local realist you claim to be.

That is absolute B.S., such theory is local of course. In principle I can add as many new fields as I want to, as long as the coupled system satisfies hyperbolic equations of motion, I am fine.

I do like your point that a truly good proposed local idea should not be introduced publicly before putting on paper in a journal or at least under copyright. But when you are ready (is that why your grinning) to write that paper; what experimental test is there you can use to support a local theory??

Of course experimentalists will have to look into the BARE data with a different bias, that is rather obvious. Well you know, this problem has beaten many prominent physicists, so I guess some time can be granted, no ? I do not feel like writing B.S. about it, neither do I approve of such papers from others, nor do I feel that some ad hoc LHV model with information loss which leads to Bell inequality violation for our observations is going to help the local realist case and neither do I tolerate dogmatic denial of some possibilities just because it conflicts some world visions. This is just a very hard problem which requires a few new nontrivial insights into nature to really solve it (of course there exists no solution yet), actually given the poor acces to databases of experiments it is even not so easy to say what the problem really is.

Careful

 

[mgelfan]

Bell inequalities don't necessarily have anything to do with questions of locality or realism ... but you are free to interpret them that way.

The inequalities do, necessarily, have to do with arithmetic properties of the lists of numbers that they deal with.

Maybe the discussions of (non)locality and the viability of hidden variables vis violations of Bell-type inequalities will eventually be discovered to have
(really) been about nothing more than some fairly simple arithmetic.

 

[Careful]

Bell inequalities don't necessarily have anything to do with questions of locality or realism ... but you are free to interpret them that way.

The inequalities do, necessarily, have to do with arithmetic properties of the lists of numbers that they deal with.

Maybe the discussions of (non)locality and the viability of hidden variables vis violations of Bell-type inequalities will eventually be discovered to have
(really) been about nothing more than some fairly simple arithmetic.

Hi,

I guess you are referring to the papers of Louis Sica here ? I have to dissapoint you, the latter contain some subtle ``gedanken'' errors in the way he is defining some correlations.

Careful

 

[mgelfan]

Hi,

I guess you are referring to the papers of Louis Sica here ? I have to dissapoint you, the latter contain some subtle ``gedanken'' errors in the way he is defining some correlations.

Careful

Thanks Careful.

I've read those papers. Can you point me to a discussion of the errors?

Anyway, my statement wasn't necessarily based on Sica's stuff ... it's just a notion that I have about the meaning of Bell inequalities when applied to experimental results ... and my notion could well be wrong.

Would you agree that the statement that "a specific orientation of a measuring device (eg. a polarizing filter) corresponds to some definite property of the disturbance incident on the filter" is an essentially meaningless statement?

 

[Careful]

Thanks Careful.

I've read those papers. Can you point me to a discussion of the errors?

I am not aware of any paper discussing in detail the error made in Sica. But you can see that -in the case of the three polarizer settings a,b,b' and three data lists of a,b,b' and identifications between a,b and a,b' (which is allowed, since we measure these things independently) - the author simply defines the identification (and hence correlations) between b and b'. The latter problem can be circumvented when working with four data lists, but then the inequality cannot be produced.

Anyway, my statement wasn't necessarily based on Sica's stuff ... it's just a notion that I have about the meaning of Bell inequalities when applied to experimental results ... and my notion could well be wrong.

It is wrong or meaningless to say the least, we all would wish it were that easy.

Would you agree that the statement that "a specific orientation of a measuring device (eg. a polarizing filter) corresponds to some definite property of the disturbance incident on the filter" is an essentially meaningless statement?

Of course I agree this is meaningless, but then you have to point out in detail where things go wrong.

Careful

 

[mgelfan]

I am not aware of any paper discussing in detail the error made in Sica. But you can see that -in the case of the three polarizer settings a,b,b' and three data lists of a,b,b' and identifications between a,b and a,b' (which is allowed, since we measure these things independently) - the author simply defines the identification (and hence correlations) between b and b'. The latter problem can be circumvented when working with four data lists, but then the inequality cannot be produced.

I must reference the paper(s). But, for now, I don't understand what you mean by " ... correlations) between b and b' ". I thought that the correlations refer to the relationship between the angular difference (theta) of, say, b and b' and the joint detection rate for that theta. As far as I'm aware the individual detection rates aren't correlated to anything ... they're just random.

Would you agree that the statement that "a specific orientation of a measuring device (eg. a polarizing filter) corresponds to some definite property of the disturbance incident on the filter" is an essentially meaningless statement?

Of course I agree this is meaningless, but then you have to point out in detail where things go wrong.

I thought I was pointing out one place where things go wrong. After all, what does a hidden variable description entail? In other words, I'm saying that any hidden variable description of quantum events is, necessarily, essentially meaningless -- simply because we don't have any way to definitively check it out.

So, in a nutshell, if we talk in terms of hidden variables, then we can calculate individual probabilities based on assumed values of the hidden variables and therefore also represent the joint probability as a product of the individual probabilities -- and then of course the resulting expectation values, while satisfying an appropriate Bell inequality, differ from those calculated with qm (and also, apparently, differ from experimental results).

 

[Careful]

I must reference the paper(s). But, for now, I don't understand what you mean by " ... correlations) between b and b' ".

This is just a shorthand for saying: ``if I would do an (EPR) experiment to measure the latter correlation, then I would get such and such data lists´´. Read the paper and you will see what I mean.

I thought that the correlations refer to the relationship between the angular difference (theta) of, say, b and b' and the joint detection rate for that theta.

Well what relationship, (1 - 2*x/PI) or cos(x), makes a hell of a difference you know.

I thought I was pointing out one place where things go wrong. After all, what does a hidden variable description entail? In other words, I'm saying that any hidden variable description of quantum events is, necessarily, essentially meaningless -- simply because we don't have any way to definitively check it out.

Your phrases become more and more illuminating. Perhaps we should zip you back to the times where atoms, electrons and so on were hidden variables and matter was thought of as vortices in the eather.

So, in a nutshell, if we talk in terms of hidden variables, then we can calculate individual probabilities based on assumed values of the hidden variables and therefore also represent the joint probability as a product of the individual probabilities -- and then of course the resulting expectation values, while satisfying an appropriate Bell inequality, differ from those calculated with qm (and also, apparently, differ from experimental results).

Individual probabilities huh ?! The only thing Bell did was to calculate the right results for the wrong problem. And are you coming up now with nonlocality and/or consciousnes mumbo jumbo ? Moreover you managed to contradict yourself in one message: the aim of Sica was to prove that Bell's inequalities were trivial and always satisfied in experiments (which is what you said too when claiming that they might be purely based on arithmetic), actually he somehow suggests that violation of the Bell inequalities is due to data overdetermination (in either that people would use four columns instead of three, which is trivially true but he somehow seems to think it is not).

Careful

 

[NateTG]

Besides: any theory which agrees with the predictions of QM is not local realistic, per Bell. So if you want to impress with your hypothesis, put forward an ACTUAL theory which contains LHV and matches the predictions of QM. Saying that it is "possible" is like saying string theory can "possibly" unify physics... maybe, maybe not.

Not that it's my favorite, but in the 'deterministic universe' interpretation, the entire universe is Bell-local, and, thus, under this sort of interpretation there are no experiments to which Bell's theorem applies.

I expect that conclusively demonstrating Bell non-locality by experiment would turn modern physics on its head in a big way.

 

[DrChinese]

Not that it's my favorite, but in the 'deterministic universe' interpretation, the entire universe is Bell-local, and, thus, under this sort of interpretation there are no experiments to which Bell's theorem applies.

My point is that there is no such theory which can give results which match QM once you attempt a Bell test. It is easy to say that there is such an interpretation, much more difficult to actually concoct one. Please keep in mind that every particle must carry around a lot of information if it is to respond correctly to all tests that it may be subjected to... since there is no entangled state any longer!

 

[NateTG]

My point is that there is no such theory which can give results which match QM once you attempt a Bell test. It is easy to say that there is such an interpretation, much more difficult to actually concoct one.

It's trivial to produce an non-predictive interpretation that matches QM exactly that is deterministic by simply postulating that, for any measurement, the result was determined at the time of the big bang. (Technically, it's a bit ambiguous whether this interpretation is actually realistic, but that's a semantic discussion.)

I suppose MWI is a more popular example of this sort of theory. Instead of using a classical determinism, it takes the 'branch both ways' approach, and you already know that MWI is consistent with QM.

 

[DrChinese]

It's trivial to produce an non-predictive interpretation that matches QM exactly that is deterministic by simply postulating that, for any measurement, the result was determined at the time of the big bang.

I used to think as you, until I thought about it a while. IF you try to take that a step further, you realize that a Bell test actually requires that there be all kinds of new forces and particle attributes so that the result works out right. If you stipulate that it is to be a local theory, that is... because the measurements are taken at space like separated points and the measuring apparatus setting can be set to trigger off of independent triggers (like radioactive isotopes, or similar) which are random. Trying to explain these as being predetermined requires you to postulate that they are working in concert to achieve the ultimate results, a very complex theory indeed.

 

[NateTG]

I used to think as you, until I thought about it a while. IF you try to take that a step further, you realize that a Bell test actually requires that there be all kinds of new forces and particle attributes so that the result works out right. If you stipulate that it is to be a local theory, that is... because the measurements are taken at space like separated points and the measuring apparatus setting can be set to trigger off of independent triggers (like radioactive isotopes, or similar) which are random. Trying to explain these as being predetermined requires you to postulate that they are working in concert to achieve the ultimate results, a very complex theory indeed.

These 'random' independant triggers are measurements. Ergo, they are predetermined in this trivially deterministic interpertation. In a strongly deterministic interpretation, there is no randomness.

You're welcome to keep bashing your head into the wall about this, but in a strongly deterministic universe with a big bang, Bell's theorem does not apply, because all space-times are Bell-local with each other.

 

[Careful]

These 'random' independant triggers are measurements. Ergo, they are predetermined in this trivially deterministic interpertation. In a strongly deterministic interpretation, there is no randomness.

You're welcome to keep bashing your head into the wall about this, but in a strongly deterministic universe with a big bang, Bell's theorem does not apply, because all space-times are Bell-local with each other.

Look, I am local realist but I do support Dr Chinese in his remarks to come up with a plausible theory. Don't get me wrong, I think it is a huge mistake to support non local or consciousness theories or rubbish alike. Perhaps you should look for something else...

Careful

 

[DrChinese]

These 'random' independant triggers are measurements. Ergo, they are predetermined in this trivially deterministic interpertation. In a strongly deterministic interpretation, there is no randomness.

I understand the concept. But it hardly explains why there are correlations between distant measurements. It only explains the idea that all apparent randomness has a common prior cause. With this hypothesis, you are hard pressed to explain why all measurements of spin (for example) don't yield the same value (as would be expected if they share a common origin). In other words, your hypothetical theory basically says that the entire universe is entangled.

 

[Nacho]

There is another way to look at this, that is actually quite amusing, to me at least.

We don't know how, but we mostly accept it as "the way it is" and don't look any deeper into it when QM (nature) throws us randomness. Yet we almost become indignant and unaccepting (or, at least questioning) when it does almost the opposite to us .. shows us correlations of random events!

Are correlations of random events more of a mystery than **true** randomness?

If and when we get to the bottom of the randomness, might not we get to the bottom of the correlations? Or conversely, is it futile to try to explain why we get correlations of randomness when we can't even explain why we get randomness?

 

[NateTG]

I understand the concept. But it hardly explains why there are correlations between distant measurements.

It's not a theory, it's an interpretation, and it's not a particularly attractive one. From a scientific perspective, it's no different than explaining QM as the action of billions of invisible devils using invisible strings to pull particles along.

Ultimately, the point is that Bell's theorem itself requires multiple unfalsifiable assumptions such as Bell Locality and Bell Realism. Which are probably strictly stronger than the usual notions of locality and realism respectively. Moreover, and this is perhaps surprising, to date, locality and realism are *interpretation* issues. That is to say, they are unfalisfiable.

Providing even a single interepretation of QM that is, in fact local and realistic, no matter how ugly, is sufficient to refute the assertion that QM is intrinsically not local realistic.