Local mechanism for nonlocal anticorrelations inside spin theory

[PeterDonis]

There is a third option: Accept Bell's theorem as such, but that claims it's premises are not valid in the situation, because one seeks not a local mechanism for explaning the results, but just to explain the correlation.

Bell's theorem proves that no explanation of the correlation that builds on "hidden values" will work, but it does not reject a theory that may have a mechanism, but only correlations of unknown values are required. I think this type of explanation need not be (or probably even can't be) "classical".

Do you have any references for these claims? Please bear in mind that, even in this forum, personal speculation is not permitted.

 

[iste]

Yes, proposing such a mechanism would contradict Bell's theorem, since it is precisely the kind of mechanism that Bell's theorem applies to. What you call "determined from source and kind of 'remembered' by both particles" is what is called "hidden variables" in Bell's theorem.

The memory thing isn't a hidden variable like that because it is the correlation that is determined from source and remembered, not the outcomes. The mechanism also already exists in quantum experiment and formalism in the sense that all that is being talked about is a phase shift. You can see in the Jung paper of original post that the phase that determines the correlation from source is something they have gotten from previous influential papers. You can find an interferometric version of this phase in papers by Horne et al. from late 80s/early 90s. Dr. Chinese posted an entanglement paper in a circular polarization thread not long ago that also mentions the phase at source and how changing it can change from vv hh correlations to vh hv correlations. The issue is physically interpretating some mechanism that already exists in quantum mechanics rather than adding some distinctly novel mechanism that may or may not reproduce the correct behavior - so I don't think it would contradict Bell.

Classical models are irrelevant here, since no such model can produce correlations that violate the Bell inequalities.

Well the same type of classical hydrodynamic system as the last paper has been shown to be able to violate Bell inequalities in the case where there is no barrier. The authors are I believe looking at the barrier case so that is an open question I think.

https://scholar.google.co.uk/scholar?cluster=9229154603514546961&hl=en&as_sdt=0,5&as_vis=1

No, it isn't. The fact that "stochastic quantization" is given as one of the assumptions in the Wikipedia article you reference should be a huge red flag to you that your claim here is not valid.

I think this is a case of two different things with the same name. The assumptions in stochastic mechanics come from what you would ordinarily think of as classical stochastic process.

 

[DrChinese]

True, but I think the time-synchronization required for the independent sources is kind of analogous in that it is required to obtain the new entanglement state through coherence.

As said before, this mechanism in original post doesn't account for all aspects of photon polarization correlation but it seems to me it can produce correlations analogous to entanglement swapping.

For two "entanglements" of orbiting particle pairs (e.g. photons 1 & 2 vs. photons 3 & 4), there could be initially no overall correlation between the pairings when you consider that the outcomes in photons 1 & 2 are independently random from those of photons 3 & 4 so that all possible pairs of outcomes for 2 & 3 (or 1 & 4 respectively) are possible.

But when you consider coincidence detection or measurement at 2 & 3 and condition on the outcomes, then it seems natural that the photons 1 & 4 must be correlated because…

What you are suggesting cannot be true, and the papers themselves say the same thing: there are no correlations between 1 & 4 unless and until the decision is made to execute a swap. Let me know if you need a quote/reference on that. Monogamy of Entanglement prevents that, for one thing.

But you are missing a critical point in your use of the word “correlations”. They are *perfect* correlations, which are only possible with entangled systems. On the other hand, in a coherent laser beam, photons are correlated but only have the same polarization at 4 angles. Big difference, right! There cannot be any deeper correlation even between photons from the same source, and certainly not from independent (but phase-locked) sources.

And the most important element is yet again bypassed: The perfect correlations only appear if the decision to execute the swap is made. This is dependent on whether the 2 & 3 photons are indistinguishable (leading to entanglement) or not. That decision and its execution occur remotely. Where is the local mechanism to address this?

As I say: claims of a local mechanism to explain entanglement are often made, but application to modern swapping experiments is sadly lacking. (Hand-waving does not count in light of experimental evidence to the contrary.)

 

[PeterDonis]

The issue is physically interpretating some mechanism that already exists in quantum mechanics

There is no "mechanism" in QM. That's why there is not one single generally accepted interpretation of QM. People can "explain" a given QM result with different interpretations that use completely different, inconsistent, and incompatible "mechanisms".

 

[iste]

There is no "mechanism" in QM. That's why there is not one single generally accepted interpretation of QM. People can "explain" a given QM result with different interpretations that use completely different, inconsistent, and incompatible "mechanisms".

Well, what I mean by 'mechanism' is just something in quantum mechanics regardless of quantum mechanics.

 

[PeterDonis]

what I mean by 'mechanism' is just something in quantum mechanics

Which makes this discussion pretty pointless; basically you're saying "quantum mechanics is quantum mechanics". True, but hardly worth discussion.

 

[martinbn]

What you are suggesting cannot be true, and the papers themselves say the same thing: there are no correlations between 1 & 4 unless and until the decision is made to execute a swap.

And even then there is no correlation. There is correlation of the results of a subset of the trials.

 

[Fra]

There is no "mechanism" in QM. That's why there is not one single generally accepted interpretation of QM.

Agreed. Or rather there ia no known mechanism behind QM that helps deep understanding of QM weirdness.

But that isnt the same as to say its meaningless to look for one. For some of us a least, the reward at the end of this path of inquiry is not just another "pure interpretation" that makes no difference, but the hope of an interpretation that will guide us into the real open questions regarding unification of forces. This is very different that looking for a "Bell class" HV explanation.

/Fredrik

 

[PeroK]

Well, what I mean by 'mechanism' is just something in quantum mechanics regardless of quantum mechanics.

It's worth noting that the idea of a fundamental mechanism isn't sustainable in physics generally. Newton's second law, to some extent, allows the concept of a mechanism in each case. But, Newton's law of gravity - by Newton's own admission - is purely a mathematical relationship that has no possible underlying mechanism. Likewise, GR has no "mechanism" for the stress-energy tensor determining the geometry of spacetime. And, Maxwell's mathematics, again not a specific mechanism, drives the propagation of EM radiation etc. I believe that 19th physicists were slow to accept Maxwell's EM theory because it wasn't mechanical enough. They, like you, wanted a mechanism; not just mathematics.

And, in general, if we are using a Lagrangian or Hamiltonian approach, then we have a mathematical, rather than a mechanical description of nature at all scales, from the microscopic to the cosmological.

You could argue that QM takes this a stage further, but you are never going to find a mechanism in any sense of the word, with microscopic cogs and wheels and levers, at the core of modern physics.

 

[Fra]

It's worth noting that the idea of a fundamental mechanism isn't sustainable in physics generally.

Agreed, this is one of the arguments in favour of ABM (agent based models) instead of the "Newtoian Scheme" System dynamics. Both are mathematical models, and both paradigms can often model the same phenomenan, but they offert perspectives.

In system dynamics, of course there is no "mechanism" we just have evolution laws on system level.
In Agent based interactions, the mechanism is at the focus, and system dyanmics is emergent.

This point is I think also at the root of Smolin path of inquiry where he argues - against the prevailing paradigm in physics - that we need a paradigm change. Sometime he received also plenty of critique for (not too surprisingly!)

/Fredrik

 

[Morbert]

What do you mean by "local interactions"? Depending on how you interpret that phrase, it could apply to all interpretations, or none.

I mean the same that @DrChinese means: If A and B are spatially distant, an external influence on A cannot have an immediate effect on B.

 

[Morbert]

No, there are no such explanations I have ever seen or heard of - and I’ve seen a lot. What there are a lot of are *claims* of such, and usually they focus on demonstrating equivalence to QM mathematically. A full remote swapping context features 2 separated measurements (violating a Bell inequality) and a third remote spot where a decision can be made (with the free will of the experimenter) to create entanglement or not.

Anyone who can pitch that as local action successfully would be of great interest to me. Just show me a paper with a diagram. :)

Do you take issue with accounts for conventional EPR experiments with local interactions? E.g.

As to the EPRB setup, consider an analogous classical situation: a bomb, initially at rest, explodes into two fragments carrying opposite angular momenta. Alice and Bob, far away from each other, measure arbitrarily chosen components J1 and J2 [...]

The distribution of bomb fragments is given by a Liouville function in phase space. When Alice measures J1, the Liouville function for J2 is instantly altered, however far Bob is from Alice. No one finds this surprising, since it is universally agreed that a Liouville function is only a mathematical tool representing our statistical knowledge. Likewise, the wave function ψ, or the corresponding Wigner function (Wigner, 1932) which is the quantum analogue of a Liouville function, are no more than mathematical tools for computing probabilities.

The essential difference [...] is that the classical Liouville function is attached to objective properties that are only imperfectly known. On the other hand, in the quantum case, the probabilities are attached to potential outcomes of mutually incompatible experiments, and these outcomes do not exist “out there” without the actual interventions. Unperformed experiments have no results.

https://arxiv.org/abs/quant-ph/0212023

A similar exercise can be carried out for your systems of interest, where entanglement is established between photons that have never been close together. You have not accepted such exercises in previous threads, so I think your issue with local accounts is larger than swapping experiments.

 

[DrChinese]

1. Do you take issue with accounts for conventional EPR experiments with local interactions? E.g. https://arxiv.org/abs/quant-ph/0212023

2. A similar exercise can be carried out for your systems of interest, where entanglement is established between photons that have never been close together. You have not accepted such exercises in previous threads, so I think your issue with local accounts is larger than swapping experiments.

1. Of course I take issue with Peres’ description, which he of all people should have known better. First, that’s an old reference (2002) and I keep specifying modern experiments. They didn’t even start those until about 2002. Peres certainly knew of the appropriate theory though at that time, and specifically he well knew of delayed choice options. So I would criticize his description in particular.

Second, that is once again ignoring what I am challenging. Which is: remote swapping.

2. No, claiming a “similar exercise” without detailing further is merely hand waving. The devil is in the details, let’s see those. If this were class, I’d say “show your work”.

 

[physika]

You could argue that QM takes this a stage further, but you are never going to find a mechanism in any sense of the word, with microscopic cogs and wheels and levers, at the core of modern physics.

...what limited language.

Mechanism:
Process by which something takes place by a set of rules/law etc.
(Obviously, describable by mathematics).

....

 

[iste]

there are no correlations between 1 & 4 unless and until the decision is made to execute a swap

Yes, I say this in the post, literally in what you quote. There is no correlation unless you condition on the outcomes via measurement and I explain why that is in the quote.

They are *perfect* correlations, which are only possible with entangled systems.

Yes, the correlations from this model are perfect. Again, This mechanism isn't accounting for everything about photon polarization correlations but they are perfect. If you have two particles, each traveling on orbital motions where they rotate at the rate and the phase shift between them is fixed (e.g. 90 degrees), then you can get perfect anti-correlations in the sense that if you measure one polarized vertically the other one is always going to be horizontal and vice versa, or for any other direction of motion. You can then get two particles moving away from each other, doing their little orbiting spiralling motions as they go and so long as their rotations are not interrupted, you should always gets the same correlations even if they are very far apart. If you do a phase shift by 0 or 180 degrees, then it will always be horizontal-horizontal, vertical-vertical, matched diagonals etc etc.

Where is the local mechanism to address this?

Its just statistical conditioning. If there are perfect anticorrelations (just looking at regular entanglement swapping in general here) then if 2 is H and 3 is H, then 1 and 4 must both be V and there is a perfect relationship between 1 and 4 too. 2 is H, 3 is V? Then 1 is V and 4 is H so there will be another perfect anticorrelation. If you dont condition then 1 & 4 can take on all possible combinations of HH, VV, HV, VH and so you wont see any correlation unless you condition on outcomes of 2 & 3 - then they suddenly become perfect just because you have conditioned statistically on those outcomes. Obviously, experimentally you need to ensure that you can get the correct coincidences.

 

[iste]

There is no "mechanism" in QM. That's why there is not one single generally accepted interpretation of QM. People can "explain" a given QM result with different interpretations that use completely different, inconsistent, and incompatible "mechanisms".

Well, what I mean by 'mechanism' is just something in quantum mechanics regardless of quantum mechanics.

Woops, I meant regardless of interpretation.

 

[iste]

Which makes this discussion pretty pointless; basically you're saying "quantum mechanics is quantum mechanics". True, but hardly worth discussion.

Woops, I didn't even see this.

No, not at all. For instance, you can derive the Schrödinger equation and reproduce all of quantum mechanics from stochastic processes according to stochastic mechanics. Because what is being derived is identical to quantum mechanics, it will not contradict it in anyway and it won't contradict Bell, hence how Stochastic mechanics has been shown recently to be able to reproduce electron spin correlations and Bell violations. However, despite being clearly equivalent, there is the addition of a different underlying interpretation.

So you can analogously give an underlying interpretation of quantum mechanical formalism and experiment, of the phase which is established at source and is important to the eventual measured correlations.

 

[iste]

It's worth noting that the idea of a fundamental mechanism isn't sustainable in physics generally. Newton's second law, to some extent, allows the concept of a mechanism in each case. But, Newton's law of gravity - by Newton's own admission - is purely a mathematical relationship that has no possible underlying mechanism. Likewise, GR has no "mechanism" for the stress-energy tensor determining the geometry of spacetime. And, Maxwell's mathematics, again not a specific mechanism, drives the propagation of EM radiation etc. I believe that 19th physicists were slow to accept Maxwell's EM theory because it wasn't mechanical enough. They, like you, wanted a mechanism; not just mathematics.

And, in general, if we are using a Lagrangian or Hamiltonian approach, then we have a mathematical, rather than a mechanical description of nature at all scales, from the microscopic to the cosmological.

You could argue that QM takes this a stage further, but you are never going to find a mechanism in any sense of the word, with microscopic cogs and wheels and levers, at the core of modern physics.

Personally, I think this is too quick. For me, the existence of stochastic mechanics, the derivation of quantum mechanics from stochastic processes suggests that a kind of mechanistic construction of quantun mechanics is possible in principle.

 

[DrChinese]

Yes, I say this in the post, literally in what you quote. There is no correlation unless you condition on the outcomes via measurement and I explain why that is in the quote.

Yes, the correlations from this model are perfect. Again, This mechanism isn't accounting for everything about photon polarization correlations but they are perfect. If you have two particles, each traveling on orbital motions where they rotate at the rate and the phase shift between them is fixed (e.g. 90 degrees), then you can get perfect anti-correlations in the sense that if you measure one polarized vertically the other one is always going to be horizontal and vice versa, or for any other direction of motion. You can then get two particles moving away from each other, doing their little orbiting spiralling motions as they go and so long as their rotations are not interrupted, you should always gets the same correlations even if they are very far apart. If you do a phase shift by 0 or 180 degrees, then it will always be horizontal-horizontal, vertical-vertical, matched diagonals etc etc.

Its just statistical conditioning. If there are perfect anticorrelations (just looking at regular entanglement swapping in general here) then if 2 is H and 3 is H, then 1 and 4 must both be V and there is a perfect relationship between 1 and 4 too. 2 is H, 3 is V? Then 1 is V and 4 is H so there will be another perfect anticorrelation. If you dont condition then 1 & 4 can take on all possible combinations of HH, VV, HV, VH and so you wont see any correlation unless you condition on outcomes of 2 & 3 - then they suddenly become perfect just because you have conditioned statistically on those outcomes. Obviously, experimentally you need to ensure that you can get the correct coincidences.

Sorry, virtually everything you are saying involves claiming there are local hidden variables. Bell tells us otherwise, which is generally accepted by the physics community.

To be clear: there is no connection whatsoever between 1 and 4 regardless of the polarization measurements of 2 & 3. The only way you get entanglement between 1 & 4 is for 2 & 3 to interact indistinguishably. No interaction - distant to 1 & 4 and optionally delayed - there are no perfect correlations.

Again I say: address an actual experiment all the way though, not via hand waving generalities.

High-fidelity entanglement swapping with fully independent sources
Kaltenback et al (2008)
https://arxiv.org/pdf/0809.3991

See figure 1, noting that distances in this setup are arbitrary. And please, how about addressing my point: there can be no “information update” unless a remote physical interaction occurs between 2 & 3. So obviously something physical does happen remotely, and it’s not simply an information update. Which is why it cannot be statistically conditioned on the outcome at 2 & 3, which can be performed later. Whether there are perfect correlations or not is strictly dependent on whether that physical interaction occurs.

Again: the 2 & 3 outcomes are the same - HH, HV, etc. - regardless of whether that physical interaction occurs. But if it does not occur, there are no perfect correlations. So that is the difference, and why your explanation cannot work.

I ask that you not fall back to old descriptions that do not work for modern experiments. There cannot be the kind of locality you envision given these experiments from the past 15 years or so. Trace through it carefully and you will see what I’m saying is correct. I still have no idea what the mechanism is that underlies this process. But it is just basic QM, and really should not be anything surprising to anybody.

 

[Morbert]

1. Of course I take issue with Peres’ description, which he of all people should have known better. First, that’s an old reference (2002) and I keep specifying modern experiments. They didn’t even start those until about 2002. Peres certainly knew of the appropriate theory though at that time, and specifically he well knew of delayed choice options. So I would criticize his description in particular.

Before we tackle modern experiments, I want to know if you take issue with such accounts of simpler EPR experiments. E.g. Alice and Bob are spatially distant and perform joint measurements on entangled 2-particle systems. The outcomes exhibit Bell-inequality violating correlations. Without worrying about more sophisticated experiments yet, do you believe these simpler experiments cannot be accounted for with local interactions?

 

[DrChinese]

Before we tackle modern experiments, I want to know if you take issue with such accounts of simpler EPR experiments. E.g. Alice and Bob are spatially distant and perform joint measurements on entangled 2-particle systems. The outcomes exhibit Bell-inequality violating correlations. Without worrying about more sophisticated experiments yet, do you believe these simpler experiments cannot be accounted for with local interactions?

I follow the standard explanation that was the norm after Bell’s paper came out: local realism is ruled out. That means locality is a logical possibility when realism is ruled out.

I don’t know if I would use the term “local interactions“ in such a context. But I don’t argue with the general concept of locality as being feasible.

Of course, all of this before the discovery of remote entanglement via swapping. Just as Bell demonstrated something that had existed all along in quantum mechanics, remote swapping. demonstrates something that has also existed all along. Now that we know about entanglement swapping, it is no longer possible to entertain locality as being feasible.

 

[PeterDonis]

you can derive the Schrödinger equation and reproduce all of quantum mechanics

"All of quantum mechanics" is more than the Schrödinger equation.

That said, as far as I can tell, the stochastic interpretation is a recognized interpretation of QM. So you can adopt it as your preferred interpretation.

What you can't do, though, is assert that whatever meaning you want to give to terms like "locality" and "mechanism" based on the stochastic interpretation must also be adopted by everyone else in the discussion. If you read the guidelines for this subforum, you will see that discussions of QM interpretations are not resolvable because different interpretations make inconsistent claims, and you can't just assert that your preferred interpretation is right and all the others are wrong. There is no "right" and "wrong" in the sense of being able to test interpretations vs. each other by experiment; they all make the same predictions for all experimental results. So interpretation discussions are a matter of opinion. Even if your opinion is that the stochastic interpretation provides a "local mechanistic" account of quantum phenomena, others might not share that opinion, and there is no way for you to show that they must share it.

 

[Morbert]

I follow the standard explanation that was the norm after Bell’s paper came out: local realism is ruled out. That means locality is a logical possibility when realism is ruled out.

I don’t know if I would use the term “local interactions“ in such a context. But I don’t argue with the general concept of locality as being feasible.

The first stepping stone at least.

In a traditional EPR experiment, when Alice carries out a measurement on the 2-particle system, she immediately updates relevant probability distributions for outcomes of possible measurements Bob can (or did) perform on the same system. According to Peres, she can do this without invoking nonlocal interactions as she does not interpret these distributions as concerning objective properties of the 2-particle system.

Similarly, in an entanglement swapping experiment, when Alice carries out a BSM involving both 2-particle systems (The 1,2 system and the 3,4 system), she can immediately update distributions for outcomes of possible joint measurements on the 1,4 system without invoking nonlocal interactions as she does not interpret these distributions as concerning objective properties of the 1,4 system.

 

[PeterDonis]

when Alice carries out a BSM

Alice isn't the one who carries out the BSM. Alice and Bob each have photons they will measure, photons 1 and 4, but someone else, Charlie or Victor or whoever, carries out the BSM on photons 2 and 3. And whether or not that BSM gets carried out determines whether Alice and Bob see correlations between their measurements. And this is true even if Alice, Bob, and the location of the BSM are far apart and all of the relevant events are spacelike separated so no light signals can travel between them.

 

[Morbert]

Alice isn't the one who carries out the BSM. Alice and Bob each have photons they will measure, photons 1 and 4, but someone else, Charlie or Victor or whoever, carries out the BSM on photons 2 and 3. And whether or not that BSM gets carried out determines whether Alice and Bob see correlations between their measurements. And this is true even if Alice, Bob, and the location of the BSM are far apart and all of the relevant events are spacelike separated so no light signals can travel between them.

You can adopt any naming convention, so long as the important comparison is made: In the traditional EPR, an observer carries out a measurement and updates distributions re/ possible measurements of the other observer. In the entanglement swapping experiment, an observer carries out a BSM and updates distributions re/ possible joint measurements on the 1,4 system by the other observers.

 

[PeterDonis]

You can adopt any naming convention

The naming convention isn't the issue. The issue is that your description implied that the person who carries out the BSM is the same as one of the people who makes the measurements whose correlations will be assessed. That is not the case.

 

[Morbert]

The naming convention isn't the issue. The issue is that your description implied that the person who carries out the BSM is the same as one of the people who makes the measurements whose correlations will be assessed. That is not the case.

My description did not imply this.

 

[PeterDonis]

My description did not imply this.

It did for anyone who is familiar with the usual naming convention used for these experiments. From your post #55 I now see that you did not intend that implication, but while you say the naming convention isn't the issue, if you pick a naming convention that's different from what's usually used in the literature, you are inviting confusion.

 

[DrChinese]

1. In a traditional EPR experiment, when Alice carries out a measurement on the 2-particle system, she immediately updates relevant probability distributions for outcomes of possible measurements Bob can (or did) perform on the same system. According to Peres, she can do this without invoking nonlocal interactions as she does not interpret these distributions as concerning objective properties of the 2-particle system.

2. Similarly, in an entanglement swapping experiment, when Alice carries out a BSM involving both 2-particle systems (The 1,2 system and the 3,4 system), she can immediately update distributions for outcomes of possible joint measurements on the 1,4 system without invoking nonlocal interactions as she does not interpret these distributions as concerning objective properties of the 1,4 system.

1. What you are saying makes no sense, and to the extent you say it represents Peres’ views I reject it completely. It is the joint expectation value that is important, and that is strictly a function of a joint context. A nonlocal context…

2. Similarly, this literally makes no scientific sense. What Alice knows after her kind of measurement (on 1) is precisely nothing about 4. And has absolutely no connection of any kind to 4 unless an experimenter elsewhere makes a decision to entangle it with 4. You know, an experimenter who is far away and is… not local.

 

[iste]

"All of quantum mechanics" is more than the Schrödinger equation.

That said, as far as I can tell, the stochastic interpretation is a recognized interpretation of QM. So you can adopt it as your preferred interpretation.

What you can't do, though, is assert that whatever meaning you want to give to terms like "locality" and "mechanism" based on the stochastic interpretation must also be adopted by everyone else in the discussion. If you read the guidelines for this subforum, you will see that discussions of QM interpretations are not resolvable because different interpretations make inconsistent claims, and you can't just assert that your preferred interpretation is right and all the others are wrong. There is no "right" and "wrong" in the sense of being able to test interpretations vs. each other by experiment; they all make the same predictions for all experimental results. So interpretation discussions are a matter of opinion. Even if your opinion is that the stochastic interpretation provides a "local mechanistic" account of quantum phenomena, others might not share that opinion, and there is no way for you to show that they must share it.

I didn't say stochastic mechanics is right, I just stated my disagreement that in principle mechanistic explanations are not sustainable. I just gave an example which is no different from the examples in the post I was replying to - if you can derive quantum mechanics from a "mechanistic" perspective then surely that is a counterexample.

Also, none of the conversations I have had in this thread involving 'mechanisms' and 'locality' have been from a stochastic perspective at all. Nothing in the original post of the thread is a stochastic perspective.