Is Quantum Entanglement Just Correlation or a Real Physical Process?

[ThomasT]

... given the experimental setup there is a limit to the correlation of
the findings that can be explained by experimental setup/local factors.

I agree with this, and, afaik, this is the mainstream view. But we're discussing whether the
existence of nonlocality has been conclusively demonstrated.

To paraphrase Bell, the "crucial assumption" in the lhv formalism is locality. One
interpretation of the physical meaning of experimental violations of Bell inequalities is, again:

Locality is represented vis separability of the joint state.
Separability of the joint state excludes statistical dependence as well as nonlocality.
Therefore, inequality violations might be due to either.

Statistical dependence doesn't imply nonlocality, but requires only the interdependence of
sample spaces and detection events.
The interdependence of sample spaces and detection events in entanglement experiments
is produced in the pairing process via local interactions/transmissions.
Therefore, inequality violations don't imply nonlocality.

Further, even though an explicit local formalism (of the joint state) is excluded, the fact that entanglement experiments produce statistical dependence of individual pairs via local interactions/transmissions suggests that the aggregate correlation of the global variables (joint detection rate with angular difference) is the result of local interactions/transmissions.

A higher correlation must be due to nonlocal behavior.

The 'highest' correlation would be the linear one predicted by the archetypal lhv formalism. The 'theorem' from this says that if the separate filters are measuring identical incident disturbances, and if only local interactions and transmissions are at work, then the rate of coincidental detection should vary in direct proportion to changes in the angular difference of the filters.

However, the optics of crossed polarizers suggests that this expectation, as stated, is
wrong. Asking what, exactly, might be wrong with an lhv ansatz that produces such an
unlikely correlation coefficient leads back to the separability requirement.

This is the common view supported by most
scientists.

What? That the observed and qm-predicted correlations in entanglement experiments are due to nonlocal behavior?

On the contrary, iirc, the resident experts on Bell stuff here at PF hold that the existence of
nonlocality hasn't been conclusively established. The mainstream view is that explicit
locality and 'realism' are, wrt quantum entanglement, formally incompatible -- which tells us
nothing about the qualitative nature of an underlying reality which everybody supposes
exists, but about which nobody can say anything definitive.

Since your view contradicts this common view you must supply hard
evidence in the form of citations from peer reviewed articles ...

Hard evidence of what? The argument is there to be criticized.

... otherwise it’s just like crackpot science (and of course forum
rules are made to prevent crackpot science being discussed).

We're discussing the meaning of Bell's theorem and violations of Bell inequalities, and standard qm -- attempting to determine whether or not they imply nonlocality.

While the way I've learned to think about this might be wrong or incomplete, the science
involved is mainstream (not to minimize the ingenuity of the experiments) -- not crackpotty ... ish.

 

[RUTA]

To paraphrase Bell, the "crucial assumption" in the lhv formalism is locality. One interpretation of the physical meaning of experimental violations of Bell inequalities is, again:

Locality is represented vis separability of the joint state.
Separability of the joint state excludes statistical dependence as well as nonlocality.
Therefore, inequality violations might be due to either.

However, the optics of crossed polarizers suggests that this expectation, as stated, is
wrong. Asking what, exactly, might be wrong with an lhv ansatz that produces such an
unlikely correlation coefficient leads back to the separability requirement.

While the way I've learned to think about this might be wrong or incomplete, the science
involved is mainstream (not to minimize the ingenuity of the experiments) -- not crackpotty ... ish.

It is widely accepted that the violation of Bell's inequality implies nonseparability and/or nonlocality. I infer from this post (and others) that you are merely pointing out the possibility of nonseparability without nonlocality, and that is consistent with a mainstream view. If it's a "crackpot idea," then at least you're in good company

 

[StandardsGuy]

Since you're view contradicts this common view you must supply hard evidence in the form of citations from peer reviewed articles, otherwise it’s just like crackpot science (and of course forum rules are made to prevent crackpot science being discussed).

Wow! So on this forum you don't want to hear alternatives to your mainsteam theories? But if he gets two other 'crack-pots' to review his paper, then it's OK? This is not the way to make progress (the Catholics wouldn't accept Galileo's assertion because it wasn't mainstream). Some scientists in the past were considered crack-pots at first, but then accepted. I think this forum's leaders need to re-think this policy.

 

[ThomasT]

I infer from this post (and others) that you are merely pointing out the possibility of nonseparability without nonlocality ...

Yes, and also suggesting that the assumption that we are part of a locally causal universe is the more reasonable one.

 

[ajw1]

Statistical dependence doesn't imply nonlocality, but requires only the interdependence of sample spaces and detection events.
The interdependence of sample spaces and detection events in entanglement experiments
is produced in the pairing process via local interactions/transmissions.
Therefore, inequality violations don't imply nonlocality.

Further, even though an explicit local formalism (of the joint state) is excluded, the fact that entanglement experiments produce statistical dependence of individual pairs via local interactions/transmissions suggests that the aggregate correlation of the global variables (joint detection rate with angular difference) is the result of local interactions/transmissions.

I am really not in a position to judge the content of your statements. In fact I would be very interested to know what realistic local effect could cause the correlations found. Maybe you should be more clear on what local interactions/transmission causes the samplespaces to be interdependent in your view.

I can only argue that the literature I find (like the one below) suggests that it is no mainstream view you're representing.

On the contrary, iirc, the resident experts on Bell stuff here at PF hold that the existence of
nonlocality hasn't been conclusively established.

I don't know about the PF experts, but let me cite some lines from an article mentioned before in this thread

Any realistic explanation must therefore include nonlocal interactions, for example s could change in response to a measurement performed on the idler photon. This explanation seems to be preferred by most researchers, and an experimental Bell inequality violation is sometimes described as a “disproof of the principle of locality.”

(public available http://arxiv.org/PS_cache/quant-ph/pdf/0205/0205171v1.pdf" )

 

[DrChinese]

Wow! So on this forum you don't want to hear alternatives to your mainsteam theories? But if he gets two other 'crack-pots' to review his paper, then it's OK? This is not the way to make progress (the Catholics wouldn't accept Galileo's assertion because it wasn't mainstream). Some scientists in the past were considered crack-pots at first, but then accepted. I think this forum's leaders need to re-think this policy.

No one is trying to stifle crackpots. The internet is very egalitarian in that respect, as it offers everyone a chance to set up their own site (I did). But that does not mean that PhysicsForums needs to give it equal time. Clearly, this is a site for mainstream discussions. If you want to post your original ideas, try somewhere else. In the scientific community, a great place for such ideas - in the form of papers - is arxiv.org. You will find more diverse theoretical presentations without peer review. (However, you must be sponsored to be included. You will find "non-standard" papers there.)

 

[jtbell]

Wow! So on this forum you don't want to hear alternatives to your mainsteam theories?

Please read the Physics Forums Global Guidelines, in particular the section Overly Speculative Posts.

The mission of PF is to provide a place where people can help each other learn physics in its current status. It is not a place for hashing out new theories that are not already a part of current discussion among physicists.

 

[StandardsGuy]

Hi Dr Chinese,
Thanks for your answer and tip, however I am not the one with the paper or theory to post. If I was, is seems that even arxiv.org would not let me because of no sponsor. It seems that on something as cloudy as QM that some leway would be granted, but I understand your position.

 

[ThomasT]

I am really not in a position to judge the content of your statements.

I'm probably not either -- at least not fully. Anyway, the argument has been around for a while, and I'm not the originator of it. I forget where I got wind of it. On it's face, it seems reasonable enough (and simple enough that even I can understand it), and this thread seemed like a good place to sort it out. Maybe it's been refuted. I don't know.

I can only argue that the literature I find (like the one below) suggests that it is no mainstream view you're representing.

Some do take experimental violations of Bell inequalities as a "disproof of the principle of locality". But that's not the mainstream interpretation.

The standard way of putting it is that, wrt the formal representation of quantum entangled states, either locality (vis separability) or realism (vis inclusion of variables which determine individual detections) has to be excluded.

However it's put, the mainstream interpretation of Bell stuff has to do with what is or isn't allowed formally, not with what does or doesn't exist in the quantum underworld.

Maybe you should be more clear on what local interactions/transmission causes the samplespaces to be interdependent in your view.

The statistical interdependence of sample spaces and detections has to do with the pairing process. The authors of papers on Bell experiments usually mention the method(s) they use. If you want more detailed info, then you might have to contact the authors or at least begin Googling stuff like 'coincidence circuitry in Bell experiments' or 'time-to-amplitude analyzer', etc.

Statistically, at the outset of a run, the sample spaces are, within the limitations set by the experimental design, maximal. Coincidence logic circuits are usually initiated by a detection at one end or the other. A detection event at one end immediately reduces the sample space at the other end because it's necessary to pair detection attributes that correspond to quantum optical disturbances that were emitted at the same time.

It's assumed that paired detection attributes correspond to quantum optical disturbances that have identical or closely related properties relevant to the global measurement (ie., filtration) parameter. Where this underlying entanglement is produced in the causal chain from emission to detection is objectively unknown.

However, the assumption that the underlying physical relationship is produced via emission and that it therefore exists before filtration isn't contradicted by the design, execution, or results of the experiments, or standard qm -- or violations of Bell inequalities.

There's just no way, currently, to know, objectively, exactly when or where or how the relationship is produced -- or it's precise qualitative physical nature. The upside to this is that none of it needs to be precisely known to predict rates of coincidental detection. Keep in mind that the sequence of coincidental detection is as unpredictable as the sequence of individual detection.

Presumably, if the sequence of individual detection could be accurately predicted, then a local realistic model of quantum entangled states that agreed with experimental results would be possible.

The bottom line, for now (and if you subscribe to the CI, then forever), is that there's no way to know if Nature is local or nonlocal. Period.

I don't know about the PF experts, but let me cite some lines from an article mentioned before in this thread.

From Dehlinger and Mitchell:
Any realistic explanation must therefore include nonlocal interactions, for example s could change in response to a measurement performed on the idler photon. This explanation seems to be preferred by most researchers, and an experimental Bell inequality violation is sometimes described as a “disproof of the principle of locality.”

They don't say anything in their interpretation of their results that contradicts the argument that nonlocality in Nature isn't implied by experimental violations of Bell inequalities.

 

[RUTA]

The standard way of putting it is that, wrt the formal representation of quantum entangled states, either locality (vis separability) or realism (vis inclusion of variables which determine individual detections) has to be excluded.

Locality (no superluminal causation) is distinct from separability.

 

[ajw1]

Imo. we can exclude 'non-separability of the quantum state' from the possible answers because without further explanation these words have no ontological meaning at all.

Nonlocality only means we find correlated behavior between particles that cannot be explained by local factors (either properties of the particles or other local hidden variables).

So there is an apparent superluminal cause, but no other ontological meaning.

@ThomasT: It's probably me, but I still have to gues the exact meaning of you view. First you seem to indicate that there might be a flaw when selecting the paired particles from the samples. Then you seem to suggest that the particles might even have the related properties when produced.

When I mention the time-reversed causality as one of the possible explanations of non-locality (without judging its validity), can you give a clear example of a possible local explanation you have in mind?

 

[RUTA]

Imo. we can exclude 'non-separability of the quantum state' from the possible answers because without further explanation these words have no ontological meaning at all.

Nonlocality only means we find correlated behavior between particles that cannot be explained by local factors (either properties of the particles or other local hidden variables).

It is difficult to understand a non-separable ontology, I've spent a few years trying to do so with Relational Blockworld and I still make mistakes, but we're finding that it has many conceptual, as well as formal, advantages. To help picture a non-separable ontology, consider a source of "entangled particles" (call it B) and their two detectors (call them A and C). Draw two overlapping circles and put A to the left, B in the central overlap region and C to the right. Let these three regions represent local M4 regions in a curved spacetime per GR. GR says A and B are non-separable (occupy same M4), B and C are non-separable, but A and C are separable (occupy different M4 regions). A non-separable view of QM might say the emission event at B and the detection events at A and C occupy a single M4 frame while the devices A, B and C themselves occupy M4 frames as classically depicted by GR. Therefore, the set of three distinct M4 frames serve as a classical approximation per GR that neglects the true underlying non-separable nature of the whole spacetime.

 

[ThomasT]

Locality (no superluminal causation) is distinct from separability.

Separability models both locality and statistical independence. The statistical dependence (nonseparability) that sufficiently violates Bell inequalities, is, by definition, due to the pairing process. The underlying cause of the entanglement remains unknown, and can't be inferred from what's known.

 

[RUTA]

Separability models both locality and statistical independence. The statistical dependence (nonseparability) that sufficiently violates Bell inequalities, is, by definition, due to the pairing process. The underlying cause of the entanglement remains unknown, and can't be inferred from what's known.

I'm differentiating "QM state separability" from "ontological separability." The QM non-separable (entangled) state is responsible for violations of Bell's inequality. Whether the underlying ontology is that of nonlocality (superluminal signaling/causation) or nonseparability (essentially violating Einsteinian realism) or both is the distinction I'm trying to point out.

 

[ThomasT]

Imo. we can exclude 'non-separability of the quantum state' from the possible answers because without further explanation these words have no ontological meaning at all.

There are two possibilities. The question is, is Nature exclusively local or isn't it? Quantum nonspeparability doesn't imply locality or nonlocality. Violation of Bell inequalities doesn't imply locality or nonlocality. So, it remains an open question -- and a matter of interpretation, and how one weighs the extant evidence from all sources, as to which assumption one feels more comfortable with.

Nonlocality only means we find correlated behavior between particles that cannot be explained by local factors (either properties of the particles or other local hidden variables).

From Wikipedia: "In physics, nonlocality is a direct influence of one object on another distant object, in violation of the principle of locality."

... there is an apparent superluminal cause ...

No. The problem is that there isn't any apparent cause -- and neither nonlocality nor locality can be definitively inferred.

@ThomasT: It's probably me, but I still have to guess the exact meaning of your view.

The exact meaning is that neither nonlocality nor locality can be definitively inferred. You're free to assume one or the other.

First you seem to indicate that there might be a flaw when selecting the paired particles from the samples.

There are technical challenges to deal with (the coincidence loophole), but this is irrelevant for the purpose of our discussion.

Then you seem to suggest that the particles might even have the related properties when produced.

Where/when the related properties are produced, and their precise qualitative nature, is unknown. That's the physical problem. That's why nonlocality can't be ruled out.

When I mention the time-reversed causality as one of the possible explanations of non-locality (without judging its validity), can you give a clear example of a possible local explanation you have in mind?

There isn't a definitive causal explanation for how the correlations are produced. There isn't a nonlocality problem because nonlocality can't be definitively inferred. There is a real, maybe insoluble -- physical -- measurement problem.

 

[ThomasT]

I'm differentiating "QM state separability" from "ontological separability." The QM non-separable (entangled) state is responsible for violations of Bell's inequality. Whether the underlying ontology is that of nonlocality (superluminal signaling/causation) or nonseparability (essentially violating Einsteinian realism) or both is the distinction I'm trying to point out.

QM nonseparability is due to statistical dependence. We're asking what the underlying cause(es) of the correlation between angular difference and coincidental detection is(are) -- which cause(es) can't be definitively inferred from anything that's objectively known, including the statistical dependence.

There's currently no way, either logically or empirically, to say conclusively that Nature operates exclusively according to the principle of locality or that it doesn't (ie., that there exist nonlocal or superluminal causal connections).

Or maybe I didn't understand your point.

 

[ThomasT]

It is difficult to understand a non-separable ontology, I've spent a few years trying to do so with Relational Blockworld and I still make mistakes, but we're finding that it has many conceptual, as well as formal, advantages. To help picture a non-separable ontology, consider a source of "entangled particles" (call it B) and their two detectors (call them A and C). Draw two overlapping circles and put A to the left, B in the central overlap region and C to the right. Let these three regions represent local M4 regions in a curved spacetime per GR. GR says A and B are non-separable (occupy same M4), B and C are non-separable, but A and C are separable (occupy different M4 regions). A non-separable view of QM might say the emission event at B and the detection events at A and C occupy a single M4 frame while the devices A, B and C themselves occupy M4 frames as classically depicted by GR. Therefore, the set of three distinct M4 frames serve as a classical approximation per GR that neglects the true underlying non-separable nature of the whole spacetime.

A nonseparable (ie., seamless, contiguous), underlying ontology at the deepest level wouldn't necessarily preclude the universe operating exclusively according to the principle of locality, would it?

 

[RUTA]

A nonseparable (ie., seamless, contiguous), underlying ontology at the deepest level wouldn't necessarily preclude the universe operating exclusively according to the principle of locality, would it?

No, it would not. What you are claiming is true. We can't say the universe is local or nonlocal given the violation of Bell's inequality. It could be nonseparable rather than nonlocal. That's all I've been trying to point out. Sorry for the confusion.

 

[ajw1]

It is difficult to understand a non-separable ontology, I've spent a few years trying to do so with Relational Blockworld and I still make mistakes, but we're finding that it has many conceptual, as well as formal, advantages. To help picture a non-separable ontology, consider a source of "entangled particles" (call it B) and their two detectors (call them A and C). Draw two overlapping circles and put A to the left, B in the central overlap region and C to the right. Let these three regions represent local M4 regions in a curved spacetime per GR. GR says A and B are non-separable (occupy same M4), B and C are non-separable, but A and C are separable (occupy different M4 regions). A non-separable view of QM might say the emission event at B and the detection events at A and C occupy a single M4 frame while the devices A, B and C themselves occupy M4 frames as classically depicted by GR. Therefore, the set of three distinct M4 frames serve as a classical approximation per GR that neglects the true underlying non-separable nature of the whole spacetime.

I thought the non-separability term was only used to represent the orthodox (Copenhagen) view, but apparently it is also used in the Relational Blockworld model. I'm not familiar with this model, but from earlier posts where this model was mentioned I thought I could conclude that it involved time-reversed causality. Is that correct?

 

[ajw1]

From Wikipedia: "In physics, nonlocality is a direct influence of one object on another distant object, in violation of the principle of locality."

I think the problem with this definition is the word 'direct'. Is MWI included? A multidimensional ontological explanation? Time reversed causality?

 

[WaveJumper]

If the concepts of Space and Time, as we perceive them, lose their meaning at the Planck scale, why is it so surprising that we are able to find phenomena in our 3D-looking universe that challenge our perception of physical space and separateness? As Brian Greene points out in The Elegant Universe:

"the familiar notion of space and time do not extend into the sub-Planckian realm, which suggests that space and time as we currently understand them may be mere approximations to more fundamental concepts that still await our discovery.”

I agree with the above statement in as much as we lack a fundamental understanding of the structure of spacetime to be able to account for all the observed phenomena, incl. entanglement, the spread of the wavefunction, length contraction in SR, the instantaneous collapse of the wavefunction, etc.

 

[RUTA]

I thought the non-separability term was only used to represent the orthodox (Copenhagen) view, but apparently it is also used in the Relational Blockworld model. I'm not familiar with this model, but from earlier posts where this model was mentioned I thought I could conclude that it involved time-reversed causality. Is that correct?

Relational Blockworld is acausal. To have so-called "backwards causation" means the future has to be as "real" as the present (so it can have causal influence over stuff happening now). That means the future is as "real" as the present. But, today is yesterday's future so the present and past are also as "real" as one another. That means the past is as "real" as the present. The co-reality of the past, present and future is called blockworld. Nothing is "happening" in a blockworld, everything just "is" in some meta sense (no state of being verb exists for this because our language is designed for use from an internal, not external--whatever that would mean--perspective). So, we don't have a dynamical view at all.

Let me explain how one can interpret quantum field theory in this way, since QFT (not RBW) is accepted physics. QM is just a special case of QFT. This will then allow you to see how QFT (and, therefore, QM) can be viewed holistically (spatiotemporally speaking) and non-separably.

In the path integral formalism for QFT, you exponentiate the action S and integrate over all field configurations to obtain the transition amplitude Z (which squared gives the probability). The action is an integral over all spacetime whose integrand is a function L (Lagrangian density) of the field(s) F. So, we have Z = integral[dF exp[iS]] ... leaving off the normalization factor ... and S = integral[d^4x L[F]]. If you go to a discrete spacetime lattice (normal in QFT), the operators on F in the S integral turn into a difference matrix, K, so S looks like F.K.F + F.J where . means "dot product" and J is a "source." Now Z = integral[dF exp[F.K.F + F.J]]. Notice that F is an integration variable, so it doesn't appear in Z. In fact the answer to this is Z ~ exp[-J.K^-1.J], where the inverse of K is called the propagator in QFT. [There are missing factors of i and 1/2 here, I'm just giving you the form for the purposes of this discussion.] Of course, at some stage the spacetime lattice spacing is allowed to go to zero so as to recover a result in continuous spacetime (called regularization), but that's not the issue. Seeing how this is done you should see that Z can be interpreted as a functional on K+J, rather than "a sum over all paths." Since K+J (+ as in a Clifford algebra place holder) is distributed throughout all spacetime, you have a blockworld view of the experiment (which, therefore, includes outcomes). To get non-separabilty you simply recognize there is a relationship between K and J, KF = J is classical field equation corresponding to the QFT. So, Z can be viewed as a measure of the symmetry of K+J, which describes the experiment in a spatiotemporally holistic, non-separable fashion.

QM is just spatially discrete QFT (obtained by letting temporal spacing on lattice go to zero but keeping space discrete), so you can see how an emission event at the source and the detection event(s) at the detector(s) can be viewed as occurring without mediating fields, waves, particles, etc., i.e., they're non-separable.

The stuff about spatiotemporal holism in QFT is just to help you digest the acausal view that replaces the dynamic view of source-detector relationships, i.e., to help you transcend the idea of causation as fundamental.

 

[WaveJumper]

Nonlocality only means we find correlated behavior between particles that cannot be explained by local factors (either properties of the particles or other local hidden variables).

So there is an apparent superluminal cause, but no other ontological meaning.

With respect to ontology, it couldn't be any other way now, could it? If i was to quote Einstein:

"What really interests me is whether God had any choice in the creation of the world."


so if QM was a local theory and the electrons did not have non-local wave properties, they would be in constant motion around the nucleus in orbits and would be subjected to gravitation and the laws of motion. According to Maxwell theory any electron revolving in an orbit must radiate energy continuously. Hence the electrons revolving around a positive nucleus would spiral into the nucleus and atoms would collapse. This is simply no way to build a universe that would last billions of years. I'd say without non-local effects(e.g. the spread and collapse of the wave function), our universe is unthinkable and inconceiveable. It'd be doomed to failure and I guess this answers Einstein's question with a "No". There doesn't appear to be any other way to build a universe and while not explaining the mechanism behind "spooky action at a distance" it sets a condition for any universe that is supposed to last immense periods of time - non-locality or immediate end of universe.

 

[ajw1]

With respect to ontology, it couldn't be any other way now, could it? If i was to quote Einstein:

"What really interests me is whether God had any choice in the creation of the world."


so if QM was a local theory and the electrons did not have non-local wave properties, they would be in constant motion around the nucleus in orbits and would be subjected to gravitation and the laws of motion. According to Maxwell theory any electron revolving in an orbit must radiate energy continuously. Hence the electrons revolving around a positive nucleus would spiral into the nucleus and atoms would collapse. This is simply no way to build a universe that would last billions of years. I'd say without non-local effects(e.g. the spread and collapse of the wave function), our universe is unthinkable and inconceiveable. It'd be doomed to failure and I guess this answers Einstein's question with a "No". There doesn't appear to be any other way to build a universe and while not explaining the mechanism behind "spooky action at a distance" it sets a condition for any universe that is supposed to last immense periods of time - non-locality or immediate end of universe.

This is an interesting thought. How exactly does non-locality contribute to the stable 'orbits' of electrons in an atom, in your opinion?

 

[WaveJumper]

This is an interesting thought. How exactly does non-locality contribute to the stable 'orbits' of electrons in an atom, in your opinion?

What orbits? I think you have misunderstood something as your question doesn't make sense.

 

[ajw1]

What orbits? I think you have misunderstood something as your question doesn't make sense.

That is probably right: I was thinking in terms of entangled particles within the atom. But you didn't mean that, ofcourse.

 

[StandardsGuy]

"What really interests me is whether God had any choice in the creation of the world."

I guess this answers Einstein's question with a "No". .

Give us a break. Surely you don't think that all the constants in the universe have to be exactly their present values. As far as electrons "must lose energy continuously", maybe Maxwell's theory needs an adjustment. What if they 'remember' how much they 'use' and give it all up at once? How would anybody know? Maybe that could be part of how atoms decay.