Does EPR definitely rule out locality?

[greypilgrim]

Hi,

Translated from German magazine "Spektrum der Wissenschaft", September 2009, p. 33 (original see below):

Since Bell used the assumption of "local realism", many believe he proved that QM violates either locality or realism. Thus the world could be local, if it violates realism. Though this is a misconception: The original argument of EPR rules out quantum locality without invoking Bell's realism.

Da Bell von der Annahme ausging, die Welt verhalte sich »lokal realistisch«, glauben viele, er habe bewiesen, dass entweder die Lokalität oder der Realismus verletzt wird. Demnach könnte die Welt lokal sein, wenn sie den Realismus verletzt. Doch das ist ein Missverständnis: Da
ursprüngliche Argument von Einstein, Podolsky und Rosen schließt die Möglichkeit der Quantenlokalität aus, ohne den von Bell verwendeten Realismus zu bemühen.

That doesn't sound right. EPR even have "physical reality" in the title, though they might not mean with it the exact same thing as Bell.

Why would we need Bell's argument if EPR already and definitely rules out locality?

 

[vanhees71]

It's wrong. Relativistic QFT describes local interactions (microcausality principle) as well as the quantum correlations due to entanglement of far-distant parts of composed quantum systems.

 

[DrChinese]

As vanhees71 says, your quoted material is not correct. EPR absolutely presupposes locality. And it rejects subjective realism as "unreasonable".

On the other hand: Bell rejects* either locality or objective realism, or both.*As incompatible with the predictions of QM.

 

[Lynch101]

As vanhees71 says, your quoted material is not correct. EPR absolutely presupposes locality. And it rejects subjective realism as "unreasonable".

On the other hand: Bell rejects* either locality or objective realism, or both.*As incompatible with the predictions of QM.

I don't want to hijack this thread but I have some similar questions and I feel they are in keeping with the OP.

Am I wrong in saying that the thrust of the EPR paper was to demonstrate that either QM violates locality or that there must be hidden variables that QM doesn't account for?

Does Bell's Theorem then say that if there are hidden variables then we must give up either realism, locality, local realism, or "free will"?

 

[vanhees71]

As vanhees71 says, your quoted material is not correct. EPR absolutely presupposes locality. And it rejects subjective realism as "unreasonable".

On the other hand: Bell rejects* either locality or objective realism, or both.*As incompatible with the predictions of QM.

To put it in a milder form: Bell provides a measurable property of any local deterministic (I think "realistic" has the meaning "deterministic" in Bell's paper) hidden-variable theory that contradicts the predictions of QT. The established fact that QT provides the correct description in such cases rules out the validity of any local deterministic hidden-variable theory.

In my opinion this has little to do with the EPR paradox though it's also closely related to entanglement, and the EPR paper is only famous, because Einstein is on the author list, and ironically Einstein himself said it missed the point by burrying it "under erudition". His real quibble has indeed more to do with non-separability of QT in the context of entanglement and thus is indeed closely related to Bell's analysis.

In the EPR setup there's no paradox whatsoever, because what's determined in their two-particle preparation (i.e., letting decay an unstable particle originally at rest wrt. an inertial frame) are the compatible observables of relative momentum and center-of-mass position (in a non-relativistic formulation). There's no contradiction whatsoever to begin with: since these momentum and position observables are compatible, they can be (pretty well) determined by some state preparation, and the example with a decaying unstable particle is one possible preparation procedure preparing such a state.

 

[DrChinese]

1. Am I wrong in saying that the thrust of the EPR paper was to demonstrate that either QM violates locality or that there must be hidden variables that QM doesn't account for?

2. Does Bell's Theorem then say that if there are hidden variables then we must give up either realism, locality, local realism, or "free will"?

1. Really, they thought that the second half of your statement was correct: That there must be a more complete specification of a system possible (than QM provided) because there were hidden variables of some type. They took locality as a given.

2. Bell demonstrated that if there are hidden variables, we must give up locality. Usually, "hidden variables" and "realism" go hand in hand - although some distinguish these two phrases.

 

[vanhees71]

No, Bell didn't demonstrate this. He showed that deterministic local hidden-variable theories lead to predictions that contradict the predictions of QT, i.e., he made the question decidable on a scientific ground, whether there can be such a kind of HV theory or whether QT provides the correct descriptions. Today it's clear that QT is right and local deterministic HV theory are wrong.

This a priori doesn't rule out the existence of a non-local deterministic HV theory. An example is Bohmian mechanics within non-relativistic QT.

 

[Lynch101]

1. Really, they thought that the second half of your statement was correct: That there must be a more complete specification of a system possible (than QM provided) because there were hidden variables of some type. They took locality as a given.

2. Bell demonstrated that if there are hidden variables, we must give up locality. Usually, "hidden variables" and "realism" go hand in hand - although some distinguish these two phrases.

Thank you DrChinese. I just finished watching a lecture where it was outlined how EPR assumes locality through the notion of "disturbance of a system". (@greypilgrim this lecture talks about EPR and specifically talks about how it presupposes locality.)

I have trouble keeping track of all the different arguments with regard to QM, but with regard to the completeness argument, I believe there are a number of physicists who remain proponents of this idea. Am I right in saying that they remain proponents because they say they believe that QM doesn't predict the outcomes of individual experiments; is there also something about QM not explaining what is going on in the experimental set-up between the time a particle/wave leaves the detector to the time it shows up on the screen?

 

[Lynch101]

No, Bell didn't demonstrate this. He showed that deterministic local hidden-variable theories lead to predictions that contradict the predictions of QT, i.e., he made the question decidable on a scientific ground, whether there can be such a kind of HV theory or whether QT provides the correct descriptions. Today it's clear that QT is right and local deterministic HV theory are wrong.

This a priori doesn't rule out the existence of a non-local deterministic HV theory. An example is Bohmian mechanics within non-relativistic QT.

Thank you vanhees. I am having a little trouble though seeing how this differs from what DrChinese said about Bell's theorem demonstrating that locality must be given up. There might be a nuance that I am missing.

 

[PeterDonis]

Moderator's note: Thread moved to QM foundations and interpretations forum.

 

[bhobba]

I have trouble keeping track of all the different arguments with regard to QM,

With regard to entanglement a lot of 'confusion' can be done away with when you recognize in an entangled state between two particles we no longer have separate particles. We just have a single system - the entangled system. When I watched Lee Smolin in the link you gave he was talking about, for entangled states, one particle being here and the other over there. You can't do that. What's going on when not observed you can't say anything about the separate particles - all you can say is we have an entangled system. When you make an observation to find a property of one of the entangled particles the rules of QM mean the observation also tells us about the other particle because we have set it up so when its observed they are correlated - but saying anything at all before being observed you can't do. Bells locality assumption is you cannot affect particle b by your choice of measurement of particle a. The fallacy here is assuming for an entangled pair there is a particle a or particle b. There is only an entangled system - that's it - that's all. So obviously Bell's assumption is meaningless. What you can do is test - can we consider it a meaningful question - ie do the particles behave as if they are separate particles while entangled. The answer is no - that's what testing bells locality really means - your intuition about there being two separate particles is wrong not QM is inherently non local - it isn't.

I have mentioned before that I personally do not think discussions of locality is meaningful for correlated systems by the cluster decomposition property of QFT so questions like Bell locality do not happen. You can reject my position of course - after all its just a way of viewing it. But if you reject it then, as the above shows, you can get yourself into a whole lot of semantic trouble.

Regarding Smolin's view we need to understand the mechanism of entanglement - I think we do
https://arxiv.org/abs/0911.0695
Its part of the generalized probability theory view of QM:
https://arxiv.org/abs/1402.6562
I must add this is an information theory view of QM. Some knowledgeable posters on this forum caution on viewing QM this way. If you want to pursue it however please start a new thread because I personally have no issues with it. In such a thread I would do more listening than participating because I do not understand the problem.

Thanks
Bill

 

[Elias1960]

Does Bell's Theorem then say that if there are hidden variables then we must give up either realism, locality, local realism, or "free will"?

Bell's theorem (together with some variants) tells us that we have either to give up realism (even in the very weak form of the EPR criterion of reality - this is the original theorem) together with causality (in any version which includes Reichenbach's principle of common cause) as well as recently even the logic of plausible reasoning (the objective Bayesian interpretation of probability) or we have to reject Einstein causality (often misnamed "locality" even if there could be theories with are completely local but with a maximal speed of information transfer greater than c).

 

[EPR]

Giving up realism spells doom for science(our understanding how the classical world works) while giving up locality spells doom for doing physics(as it's currently understood). This is an impasse with no recourse to mend in sight.

 

[vanhees71]

Thank you vanhees. I am having a little trouble though seeing how this differs from what DrChinese said about Bell's theorem demonstrating that locality must be given up. There might be a nuance that I am missing.

The very important point to keep in mind is that relativistic QFT is a local theory, i.e., interactions are local, but it is of course not a local deterministic (hidden-variable) theory. It's important, because this shows that there's really a physical way to decide between Q(F)T and local deterministic HV models.

Given the stronger-than-classical correlations of far-distant parts of entangled quantum systems (e.g., two polarization-entangled photons) relativistic QFT is the only local and only successful theory we have to describe them in accordance with the causality structure of relativistic spacetime.

 

[bhobba]

Giving up realism spells doom for science(our understanding how the classical world works) while giving up locality spells doom for doing physics(as it's currently understood). This is an impasse with no recourse to mend in sight.

There is no doom involved. I think you might benefit from reading Feynman's - The Character Of Physical Law. He points out some people say this is necessary for science, or that is necessary. The trouble is mostly they are wrong. Science is about correspondence with experiment - that's it, that's' all.

Thanks
Bill

 

[Lord Jestocost]

Giving up realism spells doom for science...

Realism and anti-realism in the philosophy of science are merely different approaches regarding the meaning of physical theories. Doing empirical science isn't in need for a realist's point of view.

 

[EPR]

Determinism(aka 'science' - the systematic pursuit of cause and effect relationships in nature) presupposes and requires realism. FAPP emergent 'determinism' is not the same as determinism. There is no such realism/determinism in the micro world as the one observed on the scale of the measuring apparati.
Instrumentalism is just the bare minimum that qualifies as science since it leaves so many questions up in the air.

 

[Lord Jestocost]

Instrumentalism is just the bare minimum that qualifies as science since it leaves so many questions up in the air.

From an instrumentalist’s point of view (in physics), one is not too bothered about the accuracy or reality of one’s ideas, concepts and definitions as long as they work and allow to make accurate predictions about what will happen next. And what are the many questions “instrumentalism leaves up in the air”: Maybe, such questions are meaningless from an instrumentalist’s point of view (in case they cannot be answered through empirical means and physical experiments).