A Hidden Assumptions in Bell's Theorem?

[PeterDonis]

I don't have any reference.

Then I would strongly suggest reviewing the literature to see if you can find one. We allow somewhat more latitude in this forum, but even here, the kinds of claims you are making should be supported by references. If you claims are really as obvious as you seem to think, you should have no trouble finding them discussed in the literature.

I can explain it in the language of cause and effect and realism. I will never understand why such language would be off limits

Because such language makes assumptions that we know nature violates. The standard "language of cause and effect and realism" leads to models that satisfy the assumptions of Bell's theorem and therefore cannot explain violations of the Bell inequalities. That was one of Bell's main reasons for deriving the theorem.

He clearly did not rule out realistic non-local interpretations in general.

Of course not: the Bohmian interpretation itself is an example of one.

However, Bell to my knowledge never claimed that such interpretations could be described unproblematically using "the language of cause and effect and realism". He recognized that the nonlocality aspect is highly significant. His example of the Bohmian interpretation was meant to illustrate just what one has to commit to if one is going to take a realistic (but nonlocal) view of QM.

 

[kurt101]

Then I would strongly suggest reviewing the literature to see if you can find one. We allow somewhat more latitude in this forum, but even here, the kinds of claims you are making should be supported by references. If you claims are really as obvious as you seem to think, you should have no trouble finding them discussed in the literature.

Ok, I will look harder. And if I can't find one, maybe I will try contacting the authors of the paper https://link.springer.com/article/10.1007/s10701-021-00511-3#Sec21 and see what they think on why what I think is obvious has not been published.

And I think some extra leeway should be allowed in this case if you are going to allow Dr. Chinese to claim that the entanglement swapping disproves cause and effect interpretations. It is hard to refute him with tied hands.

Because such language makes assumptions that we know nature violates. The standard "language of cause and effect and realism" leads to models that satisfy the assumptions of Bell's theorem and therefore cannot explain violations of the Bell inequalities. That was one of Bell's main reasons for deriving the theorem.

but we don't know that "cause and effect" and "realism" violates nature. We know either "realism" or "non-locality" is violated. If we are having conversations in the context of realism, then we should be holding non-localism sacred. That seems like a combination that should be allowed to discuss on this forum as it is very narrow and familiar to everyone and can easily be moderated. And if we deny realism, then you can say anything, but that is clearly absurd to allow for discussion on this forum and so forcing people to use credible references definitely makes sense. Anyways that is my opinion for whatever it is worth.

However, Bell to my knowledge never claimed that such interpretations could be described unproblematically using "the language of cause and effect and realism". He recognized that the nonlocality aspect is highly significant. His example of the Bohmian interpretation was meant to illustrate just what one has to commit to if one is going to take a realistic (but nonlocal) view of QM.

Maybe a good question to ask is how entanglement swapping is handled in the Bohmian interpretation since in this interpretation the photon has a definite position. In particular the case where the BSM is done after the measurement of 1 & 4 might be interested to understand.

 

[DrChinese]

I do not intend to enter into discussions of interpretations, but I would like to comment on the state of the community. I had the pleasure of cohosting Gregor Weihs last week in the colloquium of our department and besides funny/tragic stories about what discussions with Joy Christian are like, he also speaks quite clearly about Zeilinger's take on interpretations and Zeilinger's view is a modern information-theoretic version of Copenhagen which is heavily on the update-of-information side (https://arxiv.org/abs/quant-ph/0005084 and follow-up papers). It goes by the name Brukner-Zeilinger interpretation and a (non-Shannon) information vector is fundamental. The idea of finiteness of information is a golden thread through much of Zeilinger's work.

Quoting Zeilinger in support of a statement claiming that a future action changes the past in entanglement swapping experiments misrepresents Zeilinger's point of view. This becomes clear when reading Zeilinger's own papers on delayed choice entanglement swapping (https://arxiv.org/abs/1203.4834).

He closes stating:
" If one views the quantum state as a real physical object, one could get the seemingly paradoxical situation that future actions appear as having an influence on past and already irrevocably recorded events. However, there is never a paradox if the quantum state is viewed as to be no more than a “catalogue of our knowledge”. Then the state is a probability list for all possible measurement outcomes, the relative temporal order of the three observer’s events is irrelevant and no physical interactions whatsoever between these events, especially into the past, are necessary to explain the delayed-choice entanglement swapping."

Whether or not one considers this convincing is everyone's own decision.

That is a great anecdote about Weihs and Christian (whom I think is off the rails). I can imagine those discussions... so thanks for sharing. And thanks for the Brukner/Zeilinger reference, I had not seen that.

Yes, I get that Zeilinger might say one thing in one place and yet another in the... same... place. "Whether these two particles are entangled or separable has been decided after they have been measured." Which is basically exactly the opposite of the next few sentences.

I found the text following (what you provided) interesting because it is oddly identical to an earlier quote I provided. I guess it might indicate some evolution in his thinking over time.

• Zeilinger et al 2002: ... this paradox does not arise if the correctness of quantum mechanics is firmly believed.
• Zeilinger et al 2012: ...there is never a paradox if the quantum state is viewed as to be no more than a “catalogue of our knowledge”.

I personally don't see how "a catalogue of knowledge" has anything to do with Quantum Mechanics. (Ditto for interpretations where the "agent" is the key.) The experiments are real, and the results objective. We all agree: "the time ordering of the detection events has no influence on the results". Obviously, we agree distance doesn't influence the results either.

But hey, as you say, to each their own.

----------------------------------------------

I will repeat an argument I made earlier: Photons [1 & 2] are initially maximally entangled. Photons [3 & 4] are initially maximally entangled. In these states, neither [1] nor [4] can be entangled with anything else, and certainly not with each other. That's because of monogamy of entanglement (see proof). At a later time, [1 & 4] are maximally entangled (as demonstrated by experiment), clearly an objectively different state than their initial state. I see only 2 options to explain the "DrChinese paradox":

1. Is monogamy of entanglement a flatly incorrect principle? Maybe the proof is wrong, and then my argument fails. If so, a few textbooks may need updating.
2. Or does the entanglement swap executed at [2 & 3] objectively change the state of distant [1 & 4]? (This is of course entirely consistent with every known experiment.)

 

[PeterDonis]

think some extra leeway should be allowed in this case if you are going to allow Dr. Chinese to claim that the entanglement swapping disproves cause and effect interpretations

He has given a number of references to relevant papers.

 

[PeterDonis]

we don't know that "cause and effect" and "realism" violates nature. We know either "realism" or "non-locality" is violated.

These kinds of discussions really can't be done with vague ordinary language. Different people mean different things by "realism" and "locality". (I assume you mean "locality" is violated in the last part of what is quoted above.)

Bell gave a precise definition of "locality" in his papers on Bell's theorem: that the joint probability for measurements on a pair of quantum systems should factorize in a particular way. To my knowledge nobody has produced a model that satisfies that assumption but violates some other assumption of Bell's theorem in order to explain experimental results that violate the Bell inequalities. I think that is why most workers in the field have accepted that non-locality, in the sense of Bell inequalities being violated and that being due to non-factorizability of the joint probability, is unavoidable. If you are going to argue for a position that denies "realism" (and you would need to then point at the specific assumption behind Bell's theorem that you mean by "realism") while preserving "locality" in the sense of the factorizability assumption, you will need to find something in the literature that argues for that position.

Maybe a good question to ask is how entanglement swapping is handled in the Bohmian interpretation since in this interpretation the photon has a definite position.

In the Bohmian interpretation, while the particles have definite positions, those positions are not observable. But those positions determine measurement results, heuristically, by determining
which particular piece of, say, a beam splitter the particle passes through, which in turn determines which output arm the particle ends up in. Since in the Bohmian interpretation effects can propagate instantaneously, the particles can influence each other faster than light and that kind of influence can be used to explain their Bell inequality-violating correlations.

The main limitation I'm aware of with the Bohmian interpretation is that there is no generally accepted relativistic version of it.

 

[kurt101]

He has given a number of references to relevant papers.

Can you post one of the papers that proves the entanglement swapping experiment disproves non-local cause and effect interpretations?

 

[PeterDonis]

Can you post one of the papers that proves the entanglement swapping experiment disproves non-local cause and effect interpretations?

Please read through the thread. (As far as "non-local cause and effect interpretations", I'm not sure there are any other than Bohmian, in which "cause and effect" can be instantaneous.)

 

[kurt101]

The main limitation I'm aware of with the Bohmian interpretation is that there is no generally accepted relativistic version of it.

So how would the entanglement swapping experiment be handled by the Bohmian mechanics interpretation for the case where the BSM test is done after measuring 1 & 4?

The only thing I can think of to explain it is if 2 & 3 are affected by the 1 & 4 measurement in such a way where randomly once in a while 2 & 3 happen to end up in the same state when they enter the BSM. And by argument of symmetry when 2 & 3 end up in the same state 1 & 4 end up in the state required for entanglement if the experiment was run in the opposite order (BSM test first). And this is the argument that I think refutes Dr. Chinese claim that the entanglement swapping experiment disproves cause and effect. @Demystifier is this the correct explanation?

 

[PeterDonis]

how would the entanglement swapping experiment be handled by the Bohmian mechanics interpretation for the case where the BSM test is done after measuring 1 & 4?

That is one of the things I would expect a relativistic version of Bohmian mechanics to address. The non-relativistic version could treat the "cause and effect" between the BSM test and the 1 & 4 measurements as symmetric, i.e., not having a particular direction from one to the other, but I have not seen that kind of thing discussed in what literature I have read on the Bohmian interpretation. (A relativistic version, if it were based on something like quantum field theory, would naturally incorporate such a symmetry, since the operators corresponding to the measurements all commute.)

 

[DrChinese]

Can you post one of the papers that proves the entanglement swapping experiment disproves non-local cause and effect interpretations?

One of us is seriously confused. I say that swapping demonstrates nonlocal action and violates Einsteinian causality. I don’t know anything that matches what you describe.

Bohmian mechanics is nonlocal and realistic, although it is also contextual. Is that what you refer to? If so, I am not arguing for or against.

Hmmm, references to nonlocality? I think I have mentioned there are 5000+ papers with this in the title. Gisin, a top author, wrote a paper simply concluding: “Nature is nonlocal.” He adds that no relativistic theory can fully explain quantum mechanical behavior, and essentially vice versa. (Note this is not in any way a rebuttal to QFT or relativistic QM.)
https://arxiv.org/abs/0912.1475

https://arxiv.org/abs/1401.0419
In this reference, he argues any nonrealistic theory must include some nonlocal components.

I can do these quotes all night long. What I miss is you providing an actual suitable quote supporting your position from someone other than yourself.

 

[martinbn]

I will repeat an argument I made earlier: Photons [1 & 2] are initially maximally entangled. Photons [3 & 4] are initially maximally entangled. In these states, neither [1] nor [4] can be entangled with anything else, and certainly not with each other. That's because of monogamy of entanglement (see proof). At a later time, [1 & 4] are maximally entangled (as demonstrated by experiment), clearly an objectively different state than their initial state. I see only 2 options to explain the "DrChinese paradox":

1. Is monogamy of entanglement a flatly incorrect principle? Maybe the proof is wrong, and then my argument fails. If so, a few textbooks may need updating.
2. Or does the entanglement swap executed at [2 & 3] objectively change the state of distant [1 & 4]? (This is of course entirely consistent with every known experiment.)

You say that the state of 14 is objectively different. But you also agreed that measurements on 14 alone cannot tell the difference if 23 were measured or not. These are incompetible. The problem is that you omitt something important. That after the many trials you need to select a subset of the pairs 14, not the full set. Here is my chalange to you. If the state of 14 is objectively different after the measurement of 23, write it down. Show us what it was and what it became. Also explain how a system can be in different states but have the same outcomes on mearements.

 

[Demystifier]

So how would the entanglement swapping experiment be handled by the Bohmian mechanics interpretation for the case where the BSM test is done after measuring 1 & 4?

The only thing I can think of to explain it is if 2 & 3 are affected by the 1 & 4 measurement in such a way where randomly once in a while 2 & 3 happen to end up in the same state when they enter the BSM. And by argument of symmetry when 2 & 3 end up in the same state 1 & 4 end up in the state required for entanglement if the experiment was run in the opposite order (BSM test first). And this is the argument that I think refutes Dr. Chinese claim that the entanglement swapping experiment disproves cause and effect. @Demystifier is this the correct explanation?

I see nothing Bohmian in your explanation.

 

[martinbn]

I see nothing Bohmian in your explanation.

What is the BM explanation? Is it just a combination of the usual delayed choice plus entangelment explanations?

 

[Demystifier]

What is the BM explanation? Is it just a combination of the usual delayed choice plus entangelment explanations?

It's standard explanation, including the entanglement with the measuring apparatus (a'la von Neumann), minus collapse (a'la many worlds), plus Bohmian ontology (particles, fields, or whatever).

 

[martinbn]

It's standard explanation, including the entanglement with the measuring apparatus (a'la von Neumann), minus collapse (a'la many worlds), plus Bohmian ontology (particles, fields, or whatever).

Ah, I see, by the method of rigorous handwaving.

 

[Simple question]

You say that the state of 14 is objectively different.

No. QM say that if one can swap entanglement, the state 14 can be objectively different. So when a theory make predictions, scientific minded people that prefer hard science and not quibbling about philosophy, meanings of words or interpretations, and that are also ready to challenge their long held intuition/beliefs about NATURE .... build some experiment

But you also agreed that measurements on 14 alone cannot tell the difference if 23 were measured or not.

Correct, that is what QM says.

These are incompetible.

It does not follow. This is a beleif not logic.

The problem is that you omitt something important. That after the many trials you need to select a subset of the pairs 14, not the full set.

The problem is that you think it is omitted, while it actually is very important, and in plain sight.

If the state of 14 is objectively different after the measurement of 23

First mistake... the causal ordering is irrelevant.

, write it down.

... partially entangled

Show us what it was

Completely non-entangled

and what it became.

... partially entangled

Also explain how a system can be in different states but have the same outcomes on mearements.

Second mistake... in QM the initial system is unique. So only people using interpretation about causality and locality would think that 1&4 is in "more than one state".

Furthermore, the explaining needs to be done by people making the widest claims, like "everything is local and causal". QM is not, Bell's showed that in theory, and this was confirmed by experiment.
And if you think "un-realism" solved that issue, then fine. "whatever happens happens", may be a fine way to explains things. But none of that is in contradiction with DrChinese reporting (not interpretation) of standard QM results.

You will not be able to get rid of that paradox in your thinking, unless you drop one of those assumptions in Bell's theorem.

I myself equate Non-realism with incompleteness, you do that too, but unknowingly.

Meanwhile other people bet on non-locality. A significant bet, because they'll use teleportation or swapping, to physically protect information handling, and they want it to be real, to be secured... 100%

 

[Fra]

Below isnt a "claim" just painting a picture for balance and motivate the information or agent interpretation. Some you may convert to radical qbism ;)

Yes, I get that Zeilinger might say one thing in one place and yet another in the... same... place. "Whether these two particles are entangled or separable has been decided after they have been measured." Which is basically exactly the opposite of the next few sentences.

I found the text following (what you provided) interesting because it is oddly identical to an earlier quote I provided. I guess it might indicate some evolution in his thinking over time.

• Zeilinger et al 2002: ... this paradox does not arise if the correctness of quantum mechanics is firmly believed.
• Zeilinger et al 2012: ...there is never a paradox if the quantum state is viewed as to be no more than a “catalogue of our knowledge”.

I personally don't see how "a catalogue of knowledge" has anything to do with Quantum Mechanics. (Ditto for interpretations where the "agent" is the key.) The experiments are real, and the results objective.

If the information update is localized to the agent/observer (and as an agent/observer is just made of matter so it must be encoded there), then any information update from any observation of spatially separated events are always local by construction. All the paradoxes typically start when freely mixing up information acquired by different observers, ignoring the physics of communication between agents. Even to define "objectivity" operationally requires nothing less than agent-agent communication. How else can you speak meaningfully objectivity?

Any any interaction between information processing agents are by construction also constrained by communication. An agent cant process information it does not have and thus does not respond to it (for example non-clonable HV). And if the agent is a lump of matter I expect the same to hold for matter-matter interactions.

Of course there are cans of worms all over here, but its a picture at least. The dark side isnt that bad!

/Fredrik

 

[gentzen]

No. QM say that if one can swap entanglement, the state 14 can be objectively different. So when a theory make predictions, scientific minded people that prefer hard science and not quibbling about philosophy, meanings of words or interpretations, and that are also ready to challenge their long held intuition/beliefs about NATURE .... build some experiment

Correct, that is what QM says.

It does not follow. This is a beleif not logic.The problem is that you think it is omitted, while it actually is very important, and in plain sight.First mistake... the causal ordering is irrelevant.... partially entangledCompletely non-entangled... partially entangledSecond mistake... in QM the initial system is unique. So only people using interpretation about causality and locality would think that 1&4 is in "more than one state".

Furthermore, the explaining needs to be done by people making the widest claims, like "everything is local and causal". QM is not, Bell's showed that in theory, and this was confirmed by experiment.
And if you think "un-realism" solved that issue, then fine. "whatever happens happens", may be a fine way to explains things. But none of that is in contradiction with DrChinese reporting (not interpretation) of standard QM results.

You will not be able to get rid of that paradox in your thinking, unless you drop one of those assumptions in Bell's theorem.

I myself equate Non-realism with incompleteness, you do that too, but unknowingly.

Meanwhile other people bet on non-locality. A significant bet, because they'll use teleportation or swapping, to physically protect information handling, and they want it to be real, to be secured... 100%

I have no idea what you mean by this; it seems like word salad to me.

 

[kurt101]

I see nothing Bohmian in your explanation.

My explanation assumes:
The photons have a definite position at all times. That is Bohmian, right?
That cause and effect are preserved (i.e. no retrocausality). That is Bohmian, right?
That the photons have a state. That is Bohmian, right?

So maybe all that is lacking is how the mechanism of entanglement actually works or is described. I don't know how that is done in Bohmian mechanics, but I don't believe it matters to my argument.

I am just trying to understand how Bohmian mechanics can explain the ordering of photons and their states in entanglement swapping, in the case where the BSM test with photons 2 & 3 are done well after the measurement of photons 1 & 4. Bohmian mechanics needs to explain how the entanglement between 1 & 4 is established without resorting to an explanation involving retrocausality. Bohmian mechanics can't look into the future and see what will happen with photons 2 & 3, right?

 

[kurt101]

One of us is seriously confused. I say that swapping demonstrates nonlocal action and violates Einsteinian causality. I don’t know anything that matches what you describe.

That is very possible, but I have given you many opportunities to explain what exactly you mean and often you just ignore my questions about clarity. For example I asked what you meant by violating "Einsteinian causality". And you never gave me an answer. I will ask it again, is this what is your definition of Einsteinian causality? Is this your definition?:
(posted from google search):
In both Einstein's theory of special and general relativity, causality means that an effect cannot occur from a cause that is not in the back (past) light cone of that event.

Furthermore I have stated many times my position on non-locality and this experiment in many ways. It should be very clear to you what my position is by now. Here it is again:

1) In the case where the BSM test is done in the absolute past, I agree with you, this demonstrates non-locality.
2) In the case where the BSM test is done in the absolute future, I disagree with you (at least I think I do) that this demonstrates non-locality.

I can't prove for certain 2 is correct, but I don't think you can prove that it is wrong. That is my disagreement with you. I don't know how I can state it any clearer than that.

Bohmian mechanics is nonlocal and realistic, although it is also contextual. Is that what you refer to? If so, I am not arguing for or against.

And maybe we just have a misunderstanding, but if that is the case, I find your statements very misleading (and I will dig them out of history if you want me to explain more). They are misleading to me because you seem to imply that the entanglement swapping experiment excludes cause and effect interpretations, when that is not true. It only eliminates local interpretations (or realism), something Bell already proved to us a long time ago.

Hmmm, references to nonlocality? I think I have mentioned there are 5000+ papers with this in the title. Gisin, a top author, wrote a paper simply concluding: “Nature is nonlocal.” He adds that no relativistic theory can fully explain quantum mechanical behavior, and essentially vice versa. (Note this is not in any way a rebuttal to QFT or relativistic QM.)
https://arxiv.org/abs/0912.1475

https://arxiv.org/abs/1401.0419
In this reference, he argues any nonrealistic theory must include some nonlocal components.

I can do these quotes all night long. What I miss is you providing an actual suitable quote supporting your position from someone other than yourself.

Again, I agree with non-locality and I have stated that like every other time I post. So I don't know why you are bringing this up.

 

[Demystifier]

The photons have a definite position at all times. That is Bohmian, right?

It is Bohmian within non-relativistic QM. But here some photons are destroyed, after which they don't have definite positions. That's why you need to reformulate Bohmian mechanics in relativistic QFT.

As I said before, the simplest way to reformulate it is to forget about ontological particle positions and postulate the ontological fields having definite values at all times. Even when the photons are destroyed, the fields have some values $\phi({\bf x},t)$. The "velocity" $\dot{\phi}({\bf x},t)$ at any spatial point ${\bf x}$ nonlocally depends on $\phi$ at all points at the same time $t$. This dependence on time is not Lorentz covariant, but $\phi({\bf x},t)$ is a "hidden variable", i.e. not measurable in practice. The measurable quantities are Lorentz invariant. If you wonder how Lorentz non-covariant dynamics of "hidden variables" may lead to Lorentz invariant measurable predictions, see my https://arxiv.org/abs/2205.05986

In this language, you don't longer talk about entanglement between particles. Instead, you talk about entanglement between fields at different positions. The so called "particles" are no longer simple pointlike objects, but complicated configurations of fields.

 

[Demystifier]

That cause and effect are preserved (i.e. no retrocausality). That is Bohmian, right?

I suspect you are confused about the notion of cause and effect. For simplicity, consider classical mechanics without dissipation. For example, billiard balls without friction. How do you distinguish the cause from the effect?

The point is that, at the microscopic level, not only that the the future can be determined from the past, but also the past can be determined from the future. It's only on the emergent macroscopic level that the latter cannot be done. The absence of retrocausality is emergent at the macroscopic level. The hidden variable theory is, however, a microscopic theory in which retrocausality is possible. This should resolve your issues.

 

[Demystifier]

I am just trying to understand how Bohmian mechanics can explain the ordering of photons and their states in entanglement swapping, in the case where the BSM test with photons 2 & 3 are done well after the measurement of photons 1 & 4. Bohmian mechanics needs to explain how the entanglement between 1 & 4 is established without resorting to an explanation involving retrocausality. Bohmian mechanics can't look into the future and see what will happen with photons 2 & 3, right?

At the microscopic level, the state at any time $t$ determines both the future and the past from $t$. So in a sense, yes, Bohmian mechanics can "look" into both the future and the past, in the same sense in which classical mechanics can do that.

The above is just a consequence of determinism. But it should be distinguished from superdeterminism, which is, roughly speaking, determinism plus fine tuning of initial conditions. Classical and Bohmian mechanics are deterministic, but not superdeterministic. With superdeterminism one might, in principle, avoid quantum non-locality. Bohmian mechanics needs non-locality because it's only deterministic, not superdeterministic.

 

[kurt101]

The point is that, at the microscopic level, not only that the the future can be determined from the past, but also the past can be determined from the future.

No, these are just words with underlying assumptions that I don't have access to. Can you give me an example of one experiment that demonstrates what you mean by the past can be determined from the future?

 

[mattt]

No, these are just words with underlying assumptions that I don't have access to. Can you give me an example of one experiment that demonstrates what you mean by the past can be determined from the future?

It just means that in Bohmiam Mechanics (and also in Classical Mechanics) the set of all the points (in Phase Space) that a system will be in (both past and future) are fully determined by any single point.

Said in another way: the possible trajectories of the system in Phase Space are disjoint.

 

[mattt]

That is a great anecdote about Weihs and Christian (whom I think is off the rails). I can imagine those discussions... so thanks for sharing. And thanks for the Brukner/Zeilinger reference, I had not seen that.

Yes, I get that Zeilinger might say one thing in one place and yet another in the... same... place. "Whether these two particles are entangled or separable has been decided after they have been measured." Which is basically exactly the opposite of the next few sentences.

I found the text following (what you provided) interesting because it is oddly identical to an earlier quote I provided. I guess it might indicate some evolution in his thinking over time.

• Zeilinger et al 2002: ... this paradox does not arise if the correctness of quantum mechanics is firmly believed.
• Zeilinger et al 2012: ...there is never a paradox if the quantum state is viewed as to be no more than a “catalogue of our knowledge”.

I personally don't see how "a catalogue of knowledge" has anything to do with Quantum Mechanics. (Ditto for interpretations where the "agent" is the key.) The experiments are real, and the results objective. We all agree: "the time ordering of the detection events has no influence on the results". Obviously, we agree distance doesn't influence the results either.

But hey, as you say, to each their own.

----------------------------------------------

I will repeat an argument I made earlier: Photons [1 & 2] are initially maximally entangled. Photons [3 & 4] are initially maximally entangled. In these states, neither [1] nor [4] can be entangled with anything else, and certainly not with each other. That's because of monogamy of entanglement (see proof). At a later time, [1 & 4] are maximally entangled (as demonstrated by experiment), clearly an objectively different state than their initial state. I see only 2 options to explain the "DrChinese paradox":

1. Is monogamy of entanglement a flatly incorrect principle? Maybe the proof is wrong, and then my argument fails. If so, a few textbooks may need updating.
2. Or does the entanglement swap executed at [2 & 3] objectively change the state of distant [1 & 4]? (This is of course entirely consistent with every known experiment.) Paradox if you understand that is Segen

I see no paradox if you understand a state as a statistic constrain that must be satisfied by any ( sufficiently large) set of potential measurement results (on a sufficiently large ensemble of identically prepared copies of the system).

The 14 (sufficiently large) set of potential measurement results is "the same" ( satisfy the same statistics) no matter if something is done to 23 or not.

But if the BSM 23 is done, and you put your attention only to the 1 and 4 partners of the successful BSM 23s, you are considering a different set of 14s.

The whole 14 set satisfy a different statistic than the (by means of 23 BSM) selected 14 set.

I don't think Zeilinger is contradicting himself, even if his elections of words can make you think so.

 

[mattt]

By the way, I am not saying that it is not perplexing, but (at least for me) it's not more perplexing than the CHSH Theorem, that QM violates the Bell inequality, and the posterior experimental confirmation of quantum mechanics in this respect.

 

[Demystifier]

No, these are just words with underlying assumptions that I don't have access to. Can you give me an example of one experiment that demonstrates what you mean by the past can be determined from the future?

Consider billiard balls without friction. If you know the positions and velocities of the balls in the future, you can determine their positions and velocities in the past.

 

[lodbrok]

Consider billiard balls without friction. If you know the positions and velocities of the balls in the future, you can determine their positions and velocities in the past.

This has absolutely nothing to do with "causation", it's about information processing.

 

[Demystifier]

This has absolutely nothing to do with "causation", it's about information processing.

How do you define "causation"?

 

[kurt101]

Consider billiard balls without friction. If you know the positions and velocities of the balls in the future, you can determine their positions and velocities in the past.

Are you saying that is the essence of your Relativistic Bohmian / QFT Interpretation?

It sounds superdeterminstic to me, like every object has to know what the end state of the universe is and collaborate in just the right way so they don't violate any of the rules as they arrive at that end state (assuming there is even an end state!). Or do you have some characteristic that cuts off future knowledge, like it only has to know the future to a certain point? Or maybe you are just going to say the future knowledge is emergent, but not provide any deterministic explanation on how that emergence actually works.

And I did read a bit of your latest paper and I feel I made a mistake in bringing you into my disagreement with @DrChinese (at least I think we disagree). I thought you held a Bohmian interpretation that had definite paths for photons. My argument requires an interpretation with the characteristics of reality, cause and effect, and non-locality. And now I don't think your interpretation fits those charateristics.

I still think it would be interesting to hear some high level description of how your Bohmian interpretation explains the entanglement swapping experiment or even the simpler EPR experiment like Alain Aspect first executed where the measuring device is modified at the last moment, but maybe that is a large undertaking or low value for whatever reason.

 

[Demystifier]

Are you saying that is the essence of your Relativistic Bohmian / QFT Interpretation?

Yes.

It sounds superdeterminstic to me,

But it's not, as I explained in one of the posts above.

like every object has to know what the end state of the universe is and collaborate in just the right way so they don't violate any of the rules as they arrive at that end state (assuming there is even an end state!).

This does not involve fine tuning of initial conditions, so it's not superdeterminism. After you hit the billiard ball, the classical system of the billiard table and balls "knows" that the ball will fall into the hole before it actually falls into it.

I still think it would be interesting to hear some high level description of how your Bohmian interpretation explains the entanglement swapping experiment or even the simpler EPR experiment like Alain Aspect first executed where the measuring device is modified at the last moment, but maybe that is a large undertaking or low value for whatever reason.

The latter is simpler to describe. Let $x=\{x_1,x_2\}$ be positions of the two entangled EPR particles and let $y=\{y_1,y_2,...\}$ be positions of particles of the measuring apparatuses. Their interaction is described by some potential $V(x,y,t)$. The total wave function $\psi(x,y,t)$ satisfies the Schrodinger equation with this potential. Changing measuring device at the last moment means that $V(x,y,t)$ depends on $t$ such that $V$ changes at the last moment. Whenever $V(x,y,t)$ changes, the $\psi(x,y,t)$ changes accordingly as well, that's what the Schrodinger equation tells us. The Bohmian trajectories are deterministically guided by this $\psi(x,y,t)$, so whenever $V(x,y,t)$ changes, the particle trajectories change as well. That's how Bohmian mechanics explains EPR experiments with the modification of measuring device at the last moment. Does it help?

 

[kurt101]

But it's not, as I explained in one of the posts above.

I reread your previous posts and I think I am understanding better, but I am definitely not there yet. I will keep reading and I may have further questions for you in the future. And thanks for the EPR explanation. It does help a lot.

 

[lodbrok]

How do you define "causation"?

"Causation" derives from "cause" which means ... to be wholly or partially responsible for the occurrence of ... "Event B" is said to be caused by "Event A" if "Event A" is wholly or partially responsible for the occurrence of "Event B".

This is the standard understanding of "causation" in classical physics, special relativity, QFT, etc with the additional implication that "Event B" would not have occurred without "Event A" happening in the past of the light cone of "Event B".

Causation is a physical process (ontic), whereas using information about the current state of a system to "determine" the state in the past or using information about the state of the past to determine the state of the future is anthropomorphic information processing (epistemic). Otherwise, you obtain an absurd situation in which different agents can choose to "cause" or not "cause" the past/future by choosing whether to perform or not perform a mathematical calculation.

Do you define it differently?

 

[Demystifier]

Do you define it differently?

Yes. In your definition it's not clear what "responsible" means. Intuitively I understand what you mean, but this concept is not precisely defined in physical terms. Therefore I don't think that "responsibility" and causality are ontic. I think they are emergent concepts, making sense only at the effective macroscopic level. It only looks like something is responsible for something else, while in reality, on the fundamental microscopic ontic level, there is no responsibility and no causality in this sense.

 

[Demystifier]

I reread your previous posts and I think I am understanding better, but I am definitely not there yet. I will keep reading and I may have further questions for you in the future. And thanks for the EPR explanation. It does help a lot.

Let me also explain what superdeterminism would be in a billiard context. Suppose that Alice and Bob are spatially separated and each plays billiard on her/his table. And suppose that their playing is correlated: whenever Alice hits the ball in the right hole, Bob hits the ball in the left hole, and vice versa. How to explain this correlation? A nonlocal "Bohmian" explanation would be that Alice and Bob have a secret communication faster than light, that's deterministic but not superdeterministic. A superdeterministic explanation would be that there is no secret communication, but initial conditions of the Universe are fine tuned so that the processes in the Alice's and Bob's brain evolve such that they always make the opposite decisions on hitting the balls.

 

[Fra]

Causation is a physical process (ontic), whereas using information about the current state of a system to "determine" the state in the past or using information about the state of the past to determine the state of the future is anthropomorphic information processing (epistemic).

Otherwise, you obtain an absurd situation in which different agents can choose to "cause" or not "cause" the past/future by choosing whether to perform or not perform a mathematical calculation.

What seems absurd is I think interpretation dependent. For me it is not absurd at all:

The logic is simple:

For me I take as one guiding principles of is that the abstractions information processing and representation of states of information should loosely speaking be corresponding to physical internal processes and internal states of an agent/observer.

Then the "causation" is that the agents inference of the probable[information processing=computation] future from its subjective obsevations (which btw are possibly different from what obe calls "observable" in QM), causally influences the agents actions. So in the qbist vieq i would say the causaliy is internal as in my processing of input causally influences my decided actions.

From this follows a "game of expectations"
This is also why one says that the probability in qbism is normative or guiding, not descriptive. Beacause you can only describe the past, and you at best "expect" a certain future.

This is the dark side and apparently not for everyone but I think once you get in, its not absurd at all.

/Fredrik

 

[kurt101]

Let me also explain what superdeterminism would be in a billiard context. Suppose that Alice and Bob are spatially separated and each plays billiard on her/his table. And suppose that their playing is correlated: whenever Alice hits the ball in the right hole, Bob hits the ball in the left hole, and vice versa. How to explain this correlation? A nonlocal "Bohmian" explanation would be that Alice and Bob have a secret communication faster than light, that's deterministic but not superdeterministic. A superdeterministic explanation would be that there is no secret communication, but initial conditions of the Universe are fine tuned so that the processes in the Alice's and Bob's brain evolve such that they always make the opposite decisions on hitting the balls.

This is what I still don't understand about your interpretation. You say that measuring Alice instantly changes the potential which changes Bob's wave function and the state of the apparatus Bob will be measured by, even though Bob's future measurement apparatus is outside of the light cone of Bob at that point and its future configuration is not yet known (actually this later part I am not sure if it is what you are saying). I just don't see the causal connection between Alice's measurement and Bob's future measurement apparatus. I can change lots of other things near Bob's future measurement apparatus; do those things affect the measurement of Bob? Not really, so what is so special about Alice's measurement that can change Bob's future measurement apparatus.

Basically if you said that Alice's measurement changed the wave function of Bob and left it at that, I would be perfectly satisfied with the explanation. I am just confused if you are also implying that Alice's measurement changes the potential and the configuration around Bob's future measurement apparatus which is outside of the light cone of Alice and Bob.

 

[lodbrok]

Yes. In your definition it's not clear what "responsible" means. Intuitively I understand what you mean, but this concept is not precisely defined in physical terms. Therefore I don't think that "responsibility" and causality are ontic. I think they are emergent concepts, making sense only at the effective macroscopic level. It only looks like something is responsible for something else, while in reality, on the fundamental microscopic ontic level, there is no responsibility and no causality in this sense.

It is very clear. An event is responsible for another event if the other event would not have happened without the first event happening in its past light cone.

So what is your definition then?

 

[mattt]

It is very clear. An event is responsible for another event if the other event would not have happened without the first event happening in its past light cone.

So what is your definition then?

Given that in Classical Mechanics and in Bohmiam Mechanics, for an isolated system (take the whole universe) any two world-curves are disjoint in Phase-Space, your definition of causality gives rise to:

A is cause of B iff A and B belong to the same world-curve of the system and the time parameter for A is lower than the time parameter for B.

That's what your definition of causality looks like if you try to apply it to the fundamental, microscopic level (in Classical Mechanics or Bohmiam Mechanics).

Your definition have some utility if you apply it at the macroscale, where a huge number of microstates correspond to the same macrostate ( and a huge number of different microscopic world-curves can pass through different microstates that all corresponde to the same macrostate).

In this context, a macrostate A can be thought of as the cause of macrostate B in the way you described.

 

[DrChinese]

"Causation" derives from "cause" which means ... to be wholly or partially responsible for the occurrence of ... "Event B" is said to be caused by "Event A" if "Event A" is wholly or partially responsible for the occurrence of "Event B".

...
Otherwise, you obtain an absurd situation in which different agents can choose to "cause" or not "cause" the past/future by choosing whether to perform or not perform a mathematical calculation.

The decision to swap - or not - can be made in the future. It is partially responsible for the entanglement effect. Then by your own definition, the future causes the past, which is your "absurd" ("paradoxical" to others). Of course, it's actually the full context - with past and future elements - that "causes" the ultimate outcome. It can't be Einsteinian causality, because time order and distance are not factors, and the outcome is not uniquely determined (a la Born Rule). So there really isn't a good word to encompass what role the BSM takes on in these swaps. "Quantum" cause?

Let's see if I understand your alternative idea correctly.

You say that a) the maximally entangled [1 & 2] pairs - which cannot be entangled with anything else; and b) the maximally entangled [3 & 4] pairs - which also cannot be entangled with anything else; contain c) a subset in which the [1 & 4] pairs are also maximally entangled; and d) there's a subset in which the [2 & 3] pairs are maximally entangled (in the same Bell State as the matching [1 & 4] pairs per c). So you imagine that the revelation of specific [2 & 3] entangled pairs as being in an entangled state (such as Psi-) via the BSM will identify a matching [1 & 4] subset that will violate a Bell inequality. At no time, in your view, does the successful BSM on the [2 & 3] pairs CHANGE anything in the identified [1 & 4] subset, much less change anything in the past (which you consider "absurd").

Do I have this about right?

Because... I have already explained multiple times that your idea is falsified by Monogamy of Entanglement (reference and proof previously provided). [1] can never be maximally entangled with [2] and also maximally entangled with [4]. Or maybe you deny Monogamy of Entanglement?

 

[lodbrok]

For me I take as one guiding principles of is that the abstractions information processing and representation of states of information should loosely speaking be corresponding to physical internal processes and internal states of an agent/observer.

I agree all agents are governed by the same physical laws and processes as other systems.

Then the "causation" is that the agents inference of the probable[information processing=computation] future from its subjective obsevations (which btw are possibly different from what obe calls "observable" in QM), causally influences the agents actions.

Yes, causation within the agent, not the systems being observed or about which information is being processed. Demystifier appears to suggest in post #172 that being able to calculate the past state of a non-dissipative system from the present state is retrocausality. It may be causality in the internal state of the observer but has nothing to do with the causality of the system under observation.
Alice and Bob's analysis of the 1&4 data happens in the future light cone of Victor's measurement. Victor's results are transmitted to Alice and Bob through classical channels. The use of Victor's data to filter [1&4] causes the selection of a sub-ensemble which shows entanglement. Here we have causation within the agents while processing information obtained at sub-light speeds in the proper time ordering consistent with the definition of causation. None of this has anything to do with the actual physical systems of the particle streams [1, 2, 3, 4].

I would be curious to hear a definition of "causality" that suggests otherwise.

This is the dark side and apparently not for everyone but I think once you get in, its not absurd at all.

I'll conjecture that the dark side definition of causality can't distinguish "A retro-causes B" from "B causes A".

 

[Demystifier]

It is very clear. An event is responsible for another event if the other event would not have happened without the first event happening in its past light cone.

Not applicable in Bohmian mechanics because it's nonlocal so light cones don't play such a role.

 

[Demystifier]

which changes Bob's wave function

There is no such thing as Bob's wave function. In the case of entanglement, there is only one wave function describing, Bob, Alice, their apparatuses, etc.

 

[Demystifier]

I am just confused if you are also implying that Alice's measurement changes the potential and the configuration around Bob's future measurement apparatus which is outside of the light cone of Alice and Bob.

The change of Alice's potential does not affect the Bob's potential. But Alice and Bob share the same wave function, so when the wave function is changed due to the Alice's potential, it has consequences for Bob as well.

 

[kurt101]

There is no such thing as Bob's wave function. In the case of entanglement, there is only one wave function describing, Bob, Alice, their apparatuses, etc.

Does the entire universe share the same wave function in your interpretation?

The change of Alice's potential does not affect the Bob's potential. But Alice and Bob share the same wave function, so when the wave function is changed due to the Alice's potential, it has consequences for Bob as well.

Why does the change in the wave function from Alice's measurement affect Bob? And why does it not effect some other pair of locally prepared entangled photons. I assume the answer is preparation like @vanhees71 might tell me, but what is the answer in your interpretation? And again, why is the wave function of Bob's future measuring apparatus affected or is is only affected once the light cone from Bob reaches it?

 

[lodbrok]

The decision to swap - or not - can be made in the future. It is partially responsible for the entanglement effect. Then by your own definition, the future causes the past, which is your "absurd" ("paradoxical" to others).

The past can't be in the future light cone of the future so that's not my definition at all. The entangled subset is generated by Alice and Bob, in the future relative to Victor's measurements. Victor's data are transmitted to Alice and Bob through classical sub-light speed channels. Therefore the selection of the subset is from information obtained from the past relative to the moment the selection happens.

The rest of your post does not accurately represent what I stated.

 

[kurt101]

Because... I have already explained multiple times that your idea is falsified by Monogamy of Entanglement (reference and proof previously provided). [1] can never be maximally entangled with [2] and also maximally entangled with [4]. Or maybe you deny Monogamy of Entanglement?

I just want to point out that your monogamy argument only works in the case where the BSM test is done before the measurement of 1 & 4. It does not apply to when the BSM test is done after the measurement of 1 & 4 because 1 & 4 no longer exist when the BSM test is done between 2 & 3. In other words, in the case where the BSM test is done after measurement of 1 & 4 you are just reporting what used to be true of 1 & 4, not what is currently true of 1 & 4 and so the monogamy rule is not violated.

 

[Fra]

I agree all agents are governed by the same physical laws and processes as other systems.

Good start :)

Yes, causation within the agent, not the systems being observed or about which information is being processed. Demystifier appears to suggest in post #172 that being able to calculate the past state of a non-dissipative system from the present state is retrocausality. It may be causality in the internal state of the observer but has nothing to do with the causality of the system under observation.
Alice and Bob's analysis of the 1&4 data happens in the future light cone of Victor's measurement. Victor's results are transmitted to Alice and Bob through classical channels. The use of Victor's data to filter [1&4] causes the selection of a sub-ensemble which shows entanglement. Here we have causation within the agents while processing information obtained at sub-light speeds in the proper time ordering consistent with the definition of causation. None of this has anything to do with the actual physical systems of the particle streams [1, 2, 3, 4].

I would be curious to hear a definition of "causality" that suggests otherwise.

I'll conjecture that the dark side definition of causality can't distinguish "A retro-causes B" from "B causes A".

My main point was to just argue against "absurd situation in which different agents can choose to "cause" or not "cause" the past/future by choosing whether to perform or not perform a mathematical calculation".

I see your reseponse but I realize that to try to explain more I need to get into deep water as it goes much beyond the basic concepts of information processing agents, and that we can't talke about. So perhaps I should leave it here or mentors will bash me. But the idea would be in short that a real agent, has limited freedom of choosing which operations to perform or not. Just like there is only so many "actions" a molecule can take, and the ones actually take are not entirely condicidental. But I will stop there.

A paper that is not really all the way to where I aim, but in the right direction is this one, but not sure if you see the connection...
Precedence and freedom in quantum physics
-- https://arxiv.org/abs/1205.3707

/Fredrik

 

[mattt]

This is what I still don't understand about your interpretation. You say that measuring Alice instantly changes the potential which changes Bob's wave function and the state of the apparatus Bob will be measured by, even though Bob's future measurement apparatus is outside of the light cone of Bob at that point and its future configuration is not yet known (actually this later part I am not sure if it is what you are saying). I just don't see the causal connection between Alice's measurement and Bob's future measurement apparatus. I can change lots of other things near Bob's future measurement apparatus; do those things affect the measurement of Bob? Not really, so what is so special about Alice's measurement that can change Bob's future measurement apparatus.

Basically if you said that Alice's measurement changed the wave function of Bob and left it at that, I would be perfectly satisfied with the explanation. I am just confused if you are also implying that Alice's measurement changes the potential and the configuration around Bob's future measurement apparatus which is outside of the light cone of Alice and Bob.

In Bohmian mechanics, the change in disposition of one piece of apparatus can affect the results obtained elsewhere with a distant apparatus (even if the statistical distribution of those distant measurements, taken alone, does not change, so a person there cannot know if a person here has changed orientation of the apparatus or not).

Exactly the same as standard quantum mechanics.

Take into account that the global wave function satisfy the same Schrodinger equation as always, so it's exactly the same as in standard quantum mechanics.

The only addition is the guiding equation, that gives you the velocities of the bohmian particles determined by the wave function, and the equilibrium condition.

A little bit more involved is the treatment of subsystems and their version of what a measurement is.