QM: Interesting View - Get the Inside Scoop

[facenian]

Local quantum physics in the sense of Haag's book is local in a very well-defined, meaningful sense. In spite of Bell's theorem (which looks for underlying hidden variables and proves that these hidden variables would have to be nonlocal) and the corresponding experiments (which agree with local quantum physics).

Once more, let me agree with you to a certain extent. The Bell theorem should have one unambiguous interpretation: local non-conspiratorial hidden variables cannot mimic QM or describe the real world.
The controversial part of the Bell theorem is its implications regarding QM's locality. The other issue that should not be controversial is that the last point is controversial.
I hope we can at least agree on that.

 

[PeterDonis]

The controversial part of the Bell theorem is its implications regarding QM's locality.

There are no such implications (although many discussions of Bell's theorem claim, implicitly or explicitly, that there are). All Bell's theorem tells us about QM is that QM cannot be a local hidden variable theory of the kind that Bell's theorem applies to. Bell's theorem does not show that a local hidden variable of that type is the only possible local theory; it simply doesn't address that general question either way.

 

[A. Neumaier]

The controversial part of the Bell theorem is its implications regarding QM's locality. The other issue that should not be controversial is that the last point is controversial.
I hope we can at least agree on that.

No, I am on the opposite side of you in this.

I find it obvious that Bell theorem says nothing regarding QM's locality, since the latter does not work with hidden variables. Once you drop these, Bell's theorem is vacuous.

Whatever controversy there is, it is because the concepts are used in a muddled way.

 

[facenian]

No, I am on the opposite side of you in this.

I find it obvious that Bell theorem says nothing regarding QM's locality, since the latter does not work with hidden variables. Once you drop these, Bell's theorem is vacuous.

Whatever controversy there is, it is because the concepts are used in a muddled way.

I have no doubts that for you is obvious. But saying that it is not controversial is denying the existence of others for whom that opposite is obvious or dismissing them as crackpots. I hate to rely on authority but I think that nobody would consider someone like Gerardus t' Hooft a crackpot. Well, he takes very seriously the nonlocality implied by the Bell theorem to the point that the resolves the nonlocality problem by recoursing to a violation of statistical independence(one of the two hypotheses necessary to derive the Bell inequality). Please do not tell me that he does not understand QFT.

One last logical point I would like to point out. The local character of a given natural phenomenon should be possible to evaluate in a theory-independent way. I don't find it cogent that the local character of a given phenomenon is dependent on the class of theory used to evaluate it. The spookiness of the phenomenon is present or absent independently of which theory describes it.
See for instance this paper https://arxiv.org/abs/2102.07524

 

[A. Neumaier]

Gerardus t' Hooft a crackpot. Well, he takes very seriously the nonlocality implied by the Bell theorem to the point that the resolves the nonlocality problem

Of course once someone looks for an alteration of quantum theory in terms of hidden variables, which is a perfectly meaningful enterprise, then Bell's theorem is relevant. But only then!

On the other hand, I always thought that not the nonlocality problem but the measurement problem motivated t' Hooft to look for hidden variable theories. Do you have clear evidence to your alternative view of his motivation?

The local character of a given natural phenomenon should be possible to evaluate in a theory-independent way.

No, because even to tell what 'local' means needs some theory! It is there where Bell and Haag differ.

this paper https://arxiv.org/abs/2102.07524

It just says that there are two different definitions of locality that are not in contradiction. This reinforces the fact that the notion of locality is not theory-independent.

 

[facenian]

It just says that there are two different definitions of locality that are not in contradiction.

I agree.

This reinforces the fact that the notion of locality is not theory-independent.

I disagree with this part. If you want to test the local character of a theory, the concept must be applicable to it.
In the example that concerns us here, QM does not pass the test for "local-causality" as defined by Bell and shown in the paper. However, QM passes the test for signal-locality.
I strongly opposed the vacuous dictum "QM is local because the Bell theorem is a classical result". That kind of loose talking has led to heated debates that lead nowhere like the exchange between Tim Maudling and Reinhard Werner.
The misunderstanding is worsened by popular and widespread incorrect derivations of the Bell inequalities like the ones based on counterfactual definiteness.

 

[A. Neumaier]

In the example that concerns us here, QM does not pass the test for "local-causality" as defined by Bell and shown in the paper. However, QM passes the test for signal-locality.

This reinforces the fact that the notion of locality is not theory-independent. One theory (hidden variables) says locality means Bell locality (what you call "local-causality" as defined by Bell), and one theory (of quantum field theory) says that locality means spacelike commutativity of fields (which implies what you call signal-locality). Both concepts are theory laden, just in different ways.

 

[AndreiB]

Of course once someone looks for an alteration of quantum theory in terms of hidden variables, which is a perfectly meaningful enterprise, then Bell's theorem is relevant. But only then!

This is true, but you should not forget that without hidden variables physics must be non-local as EPR proved.

Just look at the abstract of bell's paper:

On the Einstein Podolsky Rosen paradox
https://cds.cern.ch/record/111654/files/vol1p195-200_001.pdf

"THE paradox of Einstein, Podolsky and Rosen was advanced as an argument that quantum mechanics
could not be a complete theory but should be supplemented by additional variables. These additional vari
ables were to restore to the theory causality and locality "

OK, so you need hidden variables to restore locality. Bell investigated the only remaining local option, hidden variables. Any other theory is already shown to be non-local (if fundamental/complete).

 

[AndreiB]

one theory (of quantum field theory) says that locality means spacelike commutativity of fields (which implies what you call signal-locality).

If you assume QFT to be complete/fundamental it has to be non-local in the sense that one measurement has a causal influence on the other, space-like measurement. If QFT is seen as a statistical approximation it can be local, no problem.

 

[A. Neumaier]

If you assume QFT to be complete/fundamental it has to be non-local in the sense that one measurement has a causal influence on the other, space-like measurement.

No. Only that correlations are nonlocal. They are expressible in terms of 2-point functions which are manifestly nonlocal.

This is true, but you should not forget that without hidden variables physics must be non-local as EPR proved.

... based on the assumption that the detection events are particle properties. But mathematical models where the detector is treated by QM but light as external classical field predict a photoeffect with Poisson distributed events, just like the standard QM treatment for coherent light. Since the model has no particles, this proves that the detection events cannot be particle properties. They are artifacts of the binary yes-no nature of the detector.

Bell investigated the only remaining local option, hidden variables. Any other theory is already shown to be non-local (if fundamental/complete).

Bell assumed that the things measured are local variables and found a contradiction. But he didn't realize that the variables instead of the dynamics could be nonlocal. Classical mechanics is already full of nonlocal expressions which when measured do not behave like Bell's theorem claims - they don't satisfy his hypothesis.

I completed quantum mechanics without introducing additional variables. Reinterpreting what is already in the formalism was enough.

 

[AndreiB]

No. Only that correlations are nonlocal.

A correlation is a correlation. Two synchronized clocks are correlated. This has nothing to do with non-locality. What can be non-local is the mechanism by which those correlations are enforced. In the case of synchronized clocks the mechanism is local, past communication with a master clock.

... based on the assumption that the detection events are particle properties.

No such assumption is made by EPR. The argument assumes only locality and that the results are correctly predicted by QM. The physical embodyment of the hidden variables does not make any difference. It could be particles, fields, whatever.

The argument is really simple. If you can perfectly predict the outcome of a measurement, that measurement is predetermined. Whatever it is that determines it is called a hidden variable. It need not be a particle.

If the result were genuinely random you could only predict it with 50% accuracy on average. So, the assumption that EPR measurements are random has been falsified by our 100% prediction accuracy.

The only way to avoid hidden variables is to postulate that whatever measurement is first instantly determines the second, so a non-local mechanism.

Classical mechanics is already full of nonlocal expressions which when measured do not behave like Bell's theorem claims - they don't satisfy his hypothesis.

What do you have in mind?

I completed quantum mechanics without introducing additional variables. Reinterpreting what is already in the formalism was enough.

Can you please provide your local explanation of the EPR perfect anticorrelations in terms of your thermal interpretation?

 

[facenian]

This reinforces the fact that the notion of locality is not theory-independent. One theory (hidden variables) says locality means Bell locality (what you call "local-causality" as defined by Bell), and one theory (of quantum field theory) says that locality means spacelike commutativity of fields (which implies what you call signal-locality). Both concepts are theory laden, just in different ways.

Actually, no. Local causality(LC) and signal causality are concepts equally applicable to QM and hidden variables(HV).
If local causality made sense only with HV theories, it wouldn't make sense to say that QM violates it. This is a logical problem. You can't ask what's the taste of the color yellow (excluding synaesthesia). See discussion in section IV of the paper https://arxiv.org/abs/2102.07524, it is quite short and easy to read.
Pardon my insistence but the statement claiming that QM is local because the Bell inequality is a classical result is a vacuous meaningless tautology. It implies that poor John Bell was a moron.

 

[A. Neumaier]

A correlation is a correlation. Two synchronized clocks are correlated. This has nothing to do with non-locality.

Correlations at different points in space are nonlocal, by definition, because they depend on what happens at both points. No matter how you try to explain them.

Can you please provide your local explanation of the EPR perfect anticorrelations in terms of your thermal interpretation?

I discuss locality in Section 4.4-4.5 of Part II of my preprint series, and in Chapter 13.5-13.6 of my book.

 

[facenian]

No such assumption is made by EPR. The argument assumes only locality and that the results are correctly predicted by QM. The physical embodyment of the hidden variables does not make any difference. It could be particles, fields, whatever.

In fact, there is a problem here with regards to the EPR paper. EPR introduced a metaphysical concept, i.e., "elements of physical reality". This has produced much unnecessary confusion to this day. Einstein immediately reacted against this and explained (in a letter to Schrodinger) how to correctly argue against completeness. Bell referred to the EPR paper but never mentioned "elements of physical reality". Bell certainly was a clear thinker.

 

[A. Neumaier]

If local causality made sense only with HV theories, it wouldn't make sense to say that QM violates it. [...] See discussion in section IV of the paper https://arxiv.org/abs/2102.07524, it is quite short and easy to read.

I didn't find there a theory-independent definition of local causality. Without reference to a theory one can neither define cause and effect nor the meaning of local.

EPR introduced a metaphysical concept, i.e., "elements of physical reality".

This is not only a metaphysical concept but the basis of all our physics. If experimental facts are not elements of reality then all of physics is unreal.

 

[facenian]

I didn't find there a theory-independent definition of local causality. Without reference to a theory one can neither define cause and effect nor the meaning of local.

I guess the discussion could go forever. But it is in the nature of discussing ideas that are hard to express accurately in words. All that I mean is this: to ask if a certain theory like QM (or HV) possesses the property called local causality (LC), it should be possible to test the theory against it. i.e., the property should be applicable to the theory.
In other words, it does not make sense to ask whether QM has a property called "tasty" but it does make sense to ask whether it has a property called LC.

Since we are discussing two theories: QM and hidden variables(HV) and we want to know if either theory posses a property called LC, I pointed out that the concept should be applicable to both. In that sense, and only in that sense, the concept should be theory-independent.
The alluded paper(https://arxiv.org/abs/2102.07524) shows that it makes sense to ask whether QM is locally causal and that QM does not pass the test for local causality irrespective of the Bell inequality, so QM's failure to pass the test cannot be blamed on the violation of the Bell inequality as usually is.
The Bell inequality does not prove that QM is not locally causal, it only proves that we cannot fix that with non-conspiratorial common causes.
Of course, we can argue whether LC is the appropriate concept for "locality" but that is a different discussion.
I beg your pardon if I repeated things that you already know, but perhaps someone else will read this post.

 

[A. Neumaier]

All that I mean is this: to ask if a certain theory like QM (or HV) possesses the property called local causality (LC), it should be possible to test the theory against it. i.e., the property should be applicable to the theory.

I agree, but the 'should' is a requirement, not something accomplished. It requires that the property called local causality (LC) must be defined unambiguously and theory-independent. Where is this in the paper you cited?

The Bell inequality does not prove that QM is not locally causal, it only proves that we cannot fix that with non-conspiratorial common causes.

Yes, so it is not a statement about quantum mechanics but a statement about fixing the latter by non-conspiratorial common causes.

 

[facenian]

This is not only a metaphysical concept but the basis of all our physics. If experimental facts are not elements of reality then all of physics is unreal.

Allow me to disagree with an enphatic no! QM, our best theory, does not assume that. It is metaphysical because it assumes that physical properties actually exist before measurements. Experimental facts are about measurements, not what existed before the observation. Also, determinism does not imply pre-existence as is usually claimed.
Here again is a source of usual misinterpretation with the Bell inequality: that it assumes "elements of physical reality". The Bell inequality only uses determinism(derived or assumed). It does not assume pre-existence.

 

[PeterDonis]

Allow me to disagree with an enphatic no!

You're agreeing with him, not disagreeing with him. He's saying that experimental facts are elements of reality. So are you.

 

[A. Neumaier]

Allow me to disagree with an enfátic no! QM, our best theory, does not assume that. It is metaphysical because it assumes that physical properties actually exist before measurements. Experimental facts are about measurements, not what existed before the observation. Also, determinism does not imply pre-existence as is usually claimed.

Detectors are elements of our macroscopic reality in Einsteins sense. Measurements are elements of our macroscopic reality once measured. Without elements of reality, there is no objective physics.

The question is whether electrons have such elements of reality. Independent of Bell's experiments I don't think they have, except in semiclassical situations. But quantum fields have such elements of reality. This is explained in my book.

Here again is a source of usual misinterpretation with the Bell inequality: that it assumes "elements of physical reality". The Bell inequality only uses determinism(derived or assumed). It does not assume pre-existence.

I know that Bell does not assume elements of reality. But in his conceptual basis he assumes beables (in the form of hidden variables). These are his strong substitute for elements of reality.

 

[facenian]

You're agreeing with him, not disagreeing with him. He's saying that experimental facts are elements of reality. So are you.

Not in the EPR sense. This is difficult as I pointed out before. You perhaps are right he said "elements of reality" not "EPR elements of reality".

 

[vanhees71]

No. Only that correlations are nonlocal. They are expressible in terms of 2-point functions which are manifestly nonlocal.

What do you mean by "nonlocal" here? That the correlations are observable with space-like separated local measuremurements? That's of course true (e.g., for the polarization measurments on two photons in a Bell state). But also this doesn't contradict anything concerning the locality (microcausality) of QFT.

 

[A. Neumaier]

What do you mean by "nonlocal" here? That the correlations are observable with space-like separated local measuremurements? That's of course true (e.g., for the polarization measurments on two photons in a Bell state). But also this doesn't contradict anything concerning the locality (microcausality) of QFT.

Local = depending on one position.
Nonlocal = depending on two positions instead of one.
That correlations are nonlocal is a triviality and doesn't contradict anything concerning the locality (microcausality) of QFT.

 

[facenian]

Local = depending on one position.
Nonlocal = depending on two positions instead of one.
That correlations are nonlocal is a triviality and doesn't contradict anything concerning the locality (microcausality) of QFT.

Different positions in space are not sufficient for nonlocality. Position in time is also necessary. Events are nonlocal when space-like separated. Correlations can have local common causes or non-local common causes and finally do they have to be causally connected?
I think that by construction QFT is built to comply with necessary conditions of locality, not with sufficient conditions. Besides is built upon the principles of ordinary QM that is manifestly non-local in the sense of Bell's local causality (again this does not imply hidden variables or the Bell inequality)

 

[A. Neumaier]

Different positions in space are not sufficient for nonlocality. Position in time is also necessary. Events are nonlocal when space-like separated.

I was talking for simplicity about a fixed time.

Correlations can have local common causes or non-local common causes and finally do they have to be causally connected?

The definition of correlations is independent of this.

I think that by construction QFT is built to comply with necessary conditions of locality, not with sufficient conditions.

It is thought to be necessary and sufficient, though the mathematical analysis is currently too hard to prove it. But the causal framework of Haag is a sufficient condition.

Besides is built upon the principles of ordinary QM that is manifestly non-local

Ordinary QM of multiple particles is not relativistic, hence has no reason to be local in any sense. Newton's mechanics is also not local! Neither is the heat equation!

in the sense of Bell's local causality (again this does not imply hidden variables or the Bell inequality)

You still lack proof that your LC can be applied to QM. Where is the precise theory-independent definition of LC that contradicts QM?

 

[vanhees71]

In special relativity two events can be causally connected if and only if one event is in or on the light cone of the other event, because such and only such events have a frame-independent temporal order.

Space-like separated events cannot be causally connected, because they don't have a frame-independent temporal order.

Locality of a theory like classical or quantum field theory means that within this theory there cannot be any causal connection between space-like separated events. In other words there cannot be faster-than-light signal propagation.

I don't know what you mean by "ordinary QM". If you mean non-relativistic QM, then locality is not an issue anyway, because here by construction the temporal order of events is absolute anyway and there is thus no constraint on signal velocity at all. Indeed, in Newtonian physics instantaneous actions at a distance are the standard description, and Newton worried about this indeed in connection with his theory of the gravitational interaction. Today we know of course that Newton's worries were quite justified.

 

[facenian]

You still lack proof that your LC can be applied to QM. Where is the precise theory-independent definition of LC that contradicts QM?

Equations (7) and (8) in this paper https://arxiv.org/abs/2102.07524
As I said before, we can make QM local by shifting to another definition. The problem is that people usually don't do that and utter trivial tautologies devoid of any meaning like "QM is local because the Bell theorem is a classical result".

 

[A. Neumaier]

Equations (7) and (8) in this paper https://arxiv.org/abs/2102.07524

These equations assume the hidden variables ##\lambda##, hence they do not apply to QM.

 

[facenian]

These equations assume the hidden variables ##\lambda##, hence they do not apply to QM.

Please read from the initial paragraph above eq (7) "The fact of matter..."
There are no "hidden variables" when you take lambda to be the singlet state. It is just ordinary QM.
(Try to read it objectively without prejudices)

 

[A. Neumaier]

There are no "hidden variables" when you take lambda to be the singlet state. It is just ordinary QM.

It is not just ordinary QM, since you take the state to be the complete hidden variable that each particle carries. (8) only means that the state is not the complete common cause of the detection events. Nothing about locality is implied (or assumed) in such a simple argument!

Indeed, according to the statistical interpretation of QM, the state is not assigned to a single particle but only to an ensemble of identically prepared particles.

 

[facenian]

No - you take the state to be a hidden variable that each particle carries.

But according to the statistical interpretation of QM, the state is not assigned to a single particle but only to an ensemble of identically prepared particles.

Of course, I take the state to be the only hidden variable. That means you do not introduce anything foreign into the theory and you can apply the LC concept to the theory without changing or perturbing it.
There are no particles carrying anything or state assigned to anything. There are only two predictions made according to the laws of quantum physics that notoriously fail to pass the test. It is a laconic mathematical expression that ordinary QM does not pass.

 

[A. Neumaier]

You still lack proof that your LC can be applied to QM. Where is the precise theory-independent definition of LC that contradicts QM?

Equations (7) and (8) in this paper https://arxiv.org/abs/2102.07524

Why are these two equations a precise theory-independent definition of LC?

I see neither an encoding of local nor one of causality. (7) is just a denial of the independence of ##a## and ##b## given ##\lambda##, which is obvious when you set ##a=b## - even classically.

 

[PeterDonis]

Thread closed for moderation.

 

[berkeman]

After a Mentor discussion and some edits, the thread is reopened.

 

[AndreiB]

In fact, there is a problem here with regards to the EPR paper. EPR introduced a metaphysical concept, i.e., "elements of physical reality". This has produced much unnecessary confusion to this day.

I don't see what is confusing about "elements of physical reality". It's a placeholder for some unknown physical property. It can be a field magnitude or a particle property or who knows what. The argument is clear and simple, unlike Bohr's response to it.

The mistake of EPR was to insist on the simultaneous existence of non-commuting properties. This is not necessary. You measure both particles on Z and you conclude that those measurements must be predetermined (hidden variables) or that one measurement determines the other (non-locality). Why bother with the X or Y spin at all? EPR's choice to introduce counterfactuals allowed Bohr to attack the argument.