A Hidden Assumptions in Bell's Theorem?

[WernerQH]

So the particle does not need to be real here, it is sufficient that it is a suitable concept to describe the physical situation caused by the preparation.

Exactly. It is a way to describe the regularities, the correlations between what is called "state preparation" and "measurement".

A. Neumaier repeatedly tried to explain that it doesn't apply to the continuous evolution in time of a single system (like the universe), but only to ensembles. But with an ontology, I don't see what would stop you to apply it precisely to such a situation. Or maybe I don't grasp the "minimal" in your minimal ontology.

I can't make sense of Neumaier's view at all. For me an ensemble just serves as a "fuzzy" description of its members, (thought) copies of one and the same physical system. Of course, the mathematical description of the ensemble as a whole can have features that are vastly different from that of its members (e.g. continuous and deterministic evolution instead of random jumps, or irreversible instead of reversible dynamics).

P.S.: I doubt that anybody else in this thread had this sort of objection to the position of the "postselection" camp. Your objection seems pretty independent of the kerfuffle to me.

Indeed, I'm not sure to which camp I belong. I just wanted to point out what in my mind is the most "hidden" of the assumptions of Bell's theorem: the existence of particles.

 

[WernerQH]

Of course, that particle doesn't have properties like a classical particle. It has fewer properties, and those fewer properties even can have strange non-local quantum correlations with properties of other particles.

Isn't it absurd to insist on having a "local" theory, and introduce* such bizarre non-local "objects" at the same time?

"Doublethink means the power of holding two contradictory beliefs in one's mind simultaneously, and accepting both of them."

(George Orwell, 1984)

*Actually, they are not introduced, but are leftovers from classical physics, superfluous metaphysical baggage like the ether.

 

[vanhees71]

Exactly. It is a way to describe the regularities, the correlations between what is called "state preparation" and "measurement".

The quantum state operationally describes the preparation procedure, and its meaning is solely probabilistic, i.e., it is a mathematical tool to calculate the probabilities for the outcome of measurements on the so prepared system.

For me, and other followers of the minimal statistical interpretation, this implies that the corresponding probabilistic predictions about the outcome of measurements can only be tested by repeating the measurement on an ensemble of equally prepared systems.

This means that on the one hand the quantum state refers to the single member of this ensemble, because it's the formal description of the properties this system has because of the corresponding preparation procdedure. On the other hand, concerning the meaning for measurements of observables of this system, it refers only to an ensemble.

I know that there are many physicists and almost all philosophers, who try to associate an extended meaning to the quantum state (formally described by a statistical operator) and the probabilities predicted by the quantum formalism. However, for me that's metaphysics, i.e., it's unnecessary for the application of the formalism to the real-world experiments.

I can't make sense of Neumaier's view at all. For me an ensemble just serves as a "fuzzy" description of its members, (thought) copies of one and the same physical system. Of course, the mathematical description of the ensemble as a whole can have features that are vastly different from that of its members (e.g. continuous and deterministic evolution instead of random jumps, or irreversible instead of reversible dynamics).Indeed, I'm not sure to which camp I belong. I just wanted to point out what in my mind is the most "hidden" of the assumptions of Bell's theorem: the existence of particles.

The question what "particles" are in connection with their description with relativistic QFT is another can of worms, which we should not open in this thread but in a separate one if needed.

 

[martinbn]

Isn't it absurd to insist on having a "local" theory, and introduce* such bizarre non-local "objects" at the same time?

I think it is inconvinient to use the same word "non-local" to mean different things. What is confusing is when some do so without specifying what they mean by it. What is absurd is when some specify that they mean one thing by that term, and then use a different meanings in their argumnets!

 

[WernerQH]

I think it is inconvinient to use the same word "non-local" to mean different things. What is confusing is when some do so without specifying what they mean by it. What is absurd is when some specify that they mean one thing by that term, and then use a different meanings in their argumnets!

Yes, it's the perfect method for making doublethink less obvious.

 

[vanhees71]

That's why I insist in using local/non-local in the clear mathematical way of relativistic QFT: The theory/interactions are local, because local observables/the Hamilton-density operator obey the microcausality principle, i.e., local-observable operators commute if their space-time arguments are space-like separated. This implies that there's no FTL-signalling and thus no causal connection between space-like separated measurement events (like detector clicks).

Then, of course, the in this sense local QFT should also describe the correlations between far-distant parts of a quantum system in an entangled state, and indeed there is no problem whatsoever, and these correlations have indeed nothing whatsoever to do with any "non-locality" (in the above sense), i.e., they describe not "spooky interactions at a distance" but the "inseparability" of entangled quantum systems.

One should note that the "spooky interactions at a distance" can only be an argument against QT, if one insists on a "collapse of a quantum state" as a physical event. There is no collapse in general, and it's only a description of how an observer, who takes note of an outcome of a filter measurement, i.e., a projection, i.e., it describes the update of the knowledge of this observer rather than a physical process on the measured system.

E.g., in the here discussed case of entanglement swapping: The observer who does the projection measurement of the pair (2&3) to the polarization-singlet state knows that he can describe the four-photon state in the so perpared subensemble (1/4 as big as the original full ensemble) by the state $\hat{\rho}_{14} \otimes \hat{\rho}_{23}$ (where $\hat{\rho}_{jk}=|\Psi^{(-)} \rangle\langle \Psi^{-}|$ are the polarization-singlet two-photon states).

However nothing can have happened to the photons 1 and 4 (at least not instantaneously) due to the measurement on the pair (2&3) due to the microcauslity constraint valid in QED. Indeed, to select the subensemble due to the filter measurement on (2&3), one has to communicate about the outcome of measurements on photon 1, of the filter measurement on (2&3), and on photon 4, which can only be done by comparing the corresponding measurement protocols, and this can be done only with usual communication, which can only be done with "signals" which propagate at most with the speed of light.

 

[Demystifier]

"Doublethink means the power of holding two contradictory beliefs in one's mind simultaneously, and accepting both of them."

(George Orwell, 1984)

Copenhagen interpretation is full of doublethinks, here are some:
- A scientific theory only talks about what can be measured, the uncertainty principle is a fundamental scientific principle that says something about intrinsic properties of quantum systems irrespective of our measurements.
- The state $\psi$ provides a complete physical description of a single system, it determines a probability which is physically meaningful only for an ensemble of systems.
- The Hamiltonian describes a deterministic evolution of the state. Since the Hamiltonian is local, the probabilistic measurement outcomes cannot violate locality.

 

[vanhees71]

Well, yes, Heisenberg is utmost confusing, even more so than Bohr. SCNR.

The first point doesn't make sense. What's meant is that only such observables take determined values, if the system has to be prepared in a state describing this property. E.g., the helicity of a photon takes the determined value $h=1$ if this photon has been prepared to be in a left-circular state. All other observables may be indetermined. All the preparation in a given state tells are the probabilities for the value of an observable, when this observable is measured.

Concerning the second point I think there's some truth in this "doublethink", as I said in #303:

The quantum state operationally describes the preparation procedure, and its meaning is solely probabilistic, i.e., it is a mathematical tool to calculate the probabilities for the outcome of measurements on the so prepared system.

For me, and other followers of the minimal statistical interpretation, this implies that the corresponding probabilistic predictions about the outcome of measurements can only be tested by repeating the measurement on an ensemble of equally prepared systems.

This means that on the one hand the quantum state refers to the single member of this ensemble, because it's the formal description of the properties this system has because of the corresponding preparation procdedure. On the other hand, concerning the meaning for measurements of observables of this system, it refers only to an ensemble.

The third point is completed clear and valid. Due to the microcausality condition there cannot be a causal connection between a measurement in one region of spacetime and one in another region of space time that is spacelike separated to the former.

 

[martinbn]

Copenhagen interpretation is full of doublethinks, here are some:
- A scientific theory only talks about what can be measured, the uncertainty principle is a fundamental scientific principle that says something about intrinsic properties of quantum systems irrespective of our measurements.
- The state $\psi$ provides a complete physical description of a single system, it determines a probability which is physically meaningful only for an ensemble of systems.
- The Hamiltonian describes a deterministic evolution of the state. Since the Hamiltonian is local, the probabilistic measurement outcomes cannot violate locality.

Why are these double think? What exactly contradicts what?

 

[Demystifier]

Why are these double think? What exactly contradicts what?

Compare the italic words.

 

[martinbn]

Compare the italic words.

That is not double think, it's just being imprecise. Otherwise it is everywhere. For example: number theory, prime numbers are numbers that cannot be factorized. Primes of the form $p=4k+1$ factor in two distinct factors in the ring of Gaussian integers.

 

[Demystifier]

That is not double think, it's just being imprecise. Otherwise it is everywhere. For example: number theory, prime numbers are numbers that cannot be factorized. Primes of the form $p=4k+1$ factor in two distinct factors in the ring of Gaussian integers.

Fair enough! Can you rewrite my examples of "doublethinks" as precise statements?

 

[LittleSchwinger]

- The Hamiltonian describes a deterministic evolution of the state. Since the Hamiltonian is local, the probabilistic measurement outcomes cannot violate locality.

Well usually what is meant there is that the evolution of the quantum state is a Cauchy problem, i.e. the state at time t=0 fixes the state at later times, but that state itself describes probabilities. It's no different from classical stochastic processes like Black-Scholes. I don't really see it as double think.

Like martinbn I don't think the rest are doublethink either. In fact you're really just rephrasing common statements purposefully using the ambiguity of a single English word to make them appear silly.

 

[WernerQH]

I think you provided a perfect example, speaking of stochastic processes as deterministic and probabilistic at the same time.

 

[LittleSchwinger]

I think you provided a perfect example, speaking of stochastic processes as deterministic and probabilistic at the same time.

In books on Stochastic processes it's usually phrased with something like:
The initial probability state/distribution $P(0)$ determines/fixes the probability distribution $P(t)$ at later times...

To me it's as clear as day that this is referring to the evolution being a Cauchy problem and doesn't contradict in any way that the state encodes probabilities. If other people think this is "doublethink" fair enough.

 

[Simple question]

I try to avoid mincing words, on the contrary do I try to understand the meaning behind. But your editing above still makes we wonder. Am I right to think that you by "FTL correlation" means "spacelike correlation"?

Yes, that is what this means.

Using the FTL word seems to imply a communication, why else use the term.

Certainly not. There is no current explanation on how that phenomenon happens. There is Bell's theorem that allows you to make hypothesis, but not incompatible ones.

FLT means Faster Then Light. Nothing more, nothing less. When experiments are done to precisely label(measure) all those space-time precise events, we measure that the lower bound for "somewhat classically moving communication" would be 10^4 time faster than light. That's not an interpretation issue. I see no reason to quibble with those 3 letters.

The only communication I see going on here is the between the observer Victor to the Observer that compares observations from Alice, and Bob (1&4)

Light will not cut it. And Nature does not wait for Victor to key in, nor some conscience to collapse "state of knowledge". I am more worried of people adopting weird contradictory view, and start questioning basic vocabulary, because observation are spooky=>fail to match some preconceived notions.

and the KEY info from Victor that is required to define the postselected ensemble.

Spooky isn't it ? The EM field of Victor will allow him to measure and create an exact sequence of number, about EM field events that can be in his past, his future, or neither, and that originated from two totally unrelated sources (hint Victor do not observes Socks)

This is supposedly a classical message. Without receiving this key, no observer can infer any entanglement.

Without receiving this KEY, Alice and Bob can absolutely not, in theory or practice, even acknowledge that there is MORE to be said, about the full ensemble.
You may believe that the arrival of Victor message rewrite history, but I don't. The correlation has(will)happens, the full ensemble changed to "keyed", and all this happened without Einstein causation.

For me Einstein locality just means there are no FTL causations between remote systems?

That is plain wrong as Victor's message arriving at Alice and Bod will demonstrate.
There are events that happens "in sync" at space-like places, and the cause is entanglement, which is obviously FLT.

But where correlations have a previous common cause, they are not a violation of Einstein locality.

One could believe that, especially if not grasping the signification of Bell's theorem. But now that
swapping experiments are a thing, beliefs that spookiness is due to the common cause, which is the preparation procedure, is no more a possible loophole. Nor is the hand waving about un-realistic realization based on state probabilistic-distribution.

While a Bell style HV might have been one way to solve Einsteins original issue. Give his record of doing away with the realism of space and time, not once but twice, had he lived on and digested bells theorem, my bet is that he would have done away with realism of HV as well. Doing away with realism is possibly "spooky".

Possibly ? It does not even mean anything. Do we have to discuss what realism means now ? Again ?

Victor(doing swapping) is not part of the preparation of 1&4. No interpretation insisting on preparation to fully explaining non-local phenomenon (even probabilistic) have been shown incomplete.

BTW Victor does not have "remote chances" to get it right. He only have classic error bars, and 100% certitude.

 

[Simple question]

After projecting (2&3) to the polarization-singlet state also (1&4) are in this state.

After what ?

That's due to the interaction of photons (2&3) with the beam splitter and the detectors.

So the experiment happens because it happens ? Nice.

This is, of course, in principle described by QED since QED of course also applies to interactions of the em. field with matter.

No it does not. You cannot compute anything, because you can't even pick an equation.

There's no need to do any calculations to know that there's no instantaneous interaction between the photons (2&3) and the equipment used to project them to the said polarization-singlet state with photons 1&4 and the equipment used to measure their polarization at their far-distant places.

Well I suppose this is your coherent view on how to project "natural science" in the state "Do not compute anything", and "be loud about it", instead on the only known minimal one.

That's implemented in QED via the microcausality condition. If the registration events of photons (2&3) and 1 and 4 are space-like separated there cannot be causal influences between these measurements.

But, mind you, those events will be correlated, and not trough hand-waving, but trough two space-like fluctuation of the EM field. So the micro-causality condition is quite enough to show that QED is incomplete.

Of course, as soon as for a given photon pair (2&3) the observer at the place knows that also (1&4) are entangled in the polarization-singlet state.

Wrong again. Your philosophy is only concerned about ensemble. You are contradicting yourself.

Of course is an update of knowledge part of the minimal interpretation of QT, which I'm follow as an interpretation. That's not philosophy that's physics!

No, that's philosophy. Philosophy can be good, if you could stick to a coherent one. Collapse of knowledge do not happens in the lab, but in one's head. That's not physics.

Correlation but no interaction/causation! So indeed, finally you agree with what I say for years!

No you don't. You've denied n'th time on this thread alone that entanglement is a-causal. And thus microcausilty cannot explains it.

If there's no FTL (faster-than-light) signalling, then space-like separated detection events cannot cause each other. That's a tautology.

No, that's not what tautology means. You were looking for "circular reasoning". But that's irrelevant as only you is pushing that new incoherent idea.

Others simply use acausal, because... it is.

 

[Fra]

Yes, that is what this means.

Ok, as I thought. I am fine with that you use the FTL word, although it easily cause confusion. But now I know, so we can drop that part of hte discussion, it's not hte important part anyway.

And Nature does not wait for Victor to key in, nor some conscience to collapse "state of knowledge".

But do you really envision that it's "Nature" that infers this correlation? How??

I think one needs to be careful what is beeing done. "Nature" around Alice and Bob, who couldn't care less about Victor or his messages, just sees the output from the full ensemble.

The information processing here must be considered to understand this, otherwise there is no correlations.
Creating this "effective ensemble" using Victors key, is not a natural process, so what does this have todo with "nature"? It is a choice of processing by an observer, that by using a KEY can "extract" correlations. And this process WILL WAIT for Victor to key in! You can't possibly perform this experiment, and prove otherwise it Victors does not key in. One can't ignore the details of the inference here. (That said, yes Alice, Bob and Victor are part of "nature" too, but I do not think that is what you meant)

/Fredrik

 

[Fra]

After what ?

After victors key(from 2&3 measurments) is received and used to process/select from the data stream. I find this crystal clear.

This is why it does not matter if the key (2&3) is generate before or after 1&4 as per sombodys clock. The information is the same.

/Fredrik

 

[DrChinese]

The information processing here must be considered to understand this, otherwise there is no correlations.
Creating this "effective ensemble" using Victors key, is not a natural process, so what does this have todo with "nature"? It is a choice of processing by an observer, that by using a KEY can "extract" correlations. And this process WILL WAIT for Victor to key in! You can't possibly perform this experiment, and prove otherwise it Victors does not key in. One can't ignore the details of the inference here. (That said, yes Alice, Bob and Victor are part of "nature" too, but I do not think that is what you meant)

I get so confused by your (and others') perspective about information processing, classical signal keys. Basically, we have a nonlocal action - the swap. In 100% of the cases where Victor observes indistinguishability in the [2] and [3] photons, the swap occurs. If and only if the swap occurs, do the distant and previously unentangled [1] and [4] photons (observed by Alice and Bob) become entangled. That is what most people refer to as causality.

Now I say the swap occurs nonlocally (outside a light cone), but cannot transmit a signal because the particular Bell State that occurs (out of 4 possible) is not known. You don't know if the Alice/Bob results will be correlated or anti-correlated (which is a function of the particular Bell State). But you are wrong that you need the signal from Victor to know which [1] and [4] photons are entangled. All they really need to do is arrive within a time synchronized time window (obviously adjusted for path length from the sources). If they arrive within that window, the [2] and [3] photons will have been indistinguishable and a swap occurs. If Alice and Bob are co-located, they know this immediately (ignoring experimental inefficiencies). They just don't know what kind of swap (i.e. which of the 4 Bell States describe the swap).

So when you say that Victor needs to send a classical signal to say what the Bell State of a swap was, we are just acknowledging something we all agree on: the swap cannot be used to send an FTL signal. In the terminology of @vanhees71, if there is no FTL signaling then there can be no FTL forward in time "causality" demonstrated. Because all classical causality predicted by QFT is local. Again, we all agree on this. So he is correct in a strange but irrelevant way. Because a remote swap is not evidence of a deterministically causal action. If a criminal hits me on the head with a hammer, then it would surely kill me. You can't do something classical like that remotely, right?

But who's questioning that point? The entire purpose of this discussion has been to demonstrate that an entirely different type of "hammer" is being used, a quantum hammer if you will humor me the analogy. When the criminal hits me on the head with this quantum hammer, we can't be sure if I will be injured or the quantum hammer will break (leaving me alive and uninjured).

• In the classical world (hitting me with a classical hammer): I will certainly die - and the criminal must be next to me - i.e. local - to cause my death.
• In the quantum world (hitting me with a quantum hammer): Sometimes I die and sometimes I don't - but the criminal can be anywhere - i.e. non-local - to cause one of the possible predicted outcomes.

QFT predicts both of these cases.

So the argument "The information processing here must be considered to understand this" is in fact circular reasoning. We are demonstrating that the swap "quantum causes" the [1] and [4] photons to become entangled to each other, when they were previously were not (as proved by MoE, previously cited). And we see experimentally that distance is not a factor, which again everyone agrees upon.

But you are essentially saying: There can be no "nonlocal quantum causality" because such causality cannot be used to send a signal. No test of "nonlocal quantum causality" could ever be demonstrated with your requirement! That's circular, you have included a requirement that we all agree cannot be met!! But your requirement does not in any manner rule out the situation I am describing. Alice/Bob aren't expecting to know which outcome will occur (i.e. which observed polarizations will be correlated or anti-correlated). To meet your requirement: FTL signaling is required and if that were possible, swapping could be used for... FTL signaling. Circular.

Remote entanglement swapping is a tool different than a signaling tool. Real quantum networks can entangle previously unentangled particles at speeds exceeding the speed of light, regardless of reference frame.

 

[Nullstein]

If and only if the swap occurs, do the distant and previously unentangled [1] and [4] photons (observed by Alice and Bob) become entangled. That is what most people refer to as causality.

You're wrong about this. There exist entangled subensembles even if no BSM is performed. So you can never know whether the BSM was responsible for producing the entanglement in those subensembles that are post-selected after the BSM or whether these subensembles would have been there anyway and due to the lack of a BSM result we just have to no interesting way to post-select them. This leaves open the possibility for a knowledge-based interpretation without action at a distance, such as Zeilinger is endorsing. If you are in conflict with a Nobel laureate, who moreover is an expert in this field, you should sooner or later begin to question your own premises.

 

[PeterDonis]

There exist entangled subensembles even if no BSM is performed.

How would you pick out such subensembles?

If your answer is that you would look at the results of the 1 & 4 measurements and pick out runs on that basis, my response was given in a previous post of mine: that's cherry-picking your statistics. In other words, it's not a valid method of picking a subensemble. The whole point in the entanglement swapping experiments is that the subensemble is picked out without any knowledge whatever of the 1 & 4 measurement results; only the 2 & 3 BSM result is looked at.

 

[PeterDonis]

you can never know whether the BSM was responsible for producing the entanglement in those subensembles that are post-selected after the BSM or whether these subensembles would have been there anyway

Yes, you can test whether they would have been there anyway: just do the initial preparation (in which 1 & 2 are maximally entangled and 3 & 4 are maximally entangled), don't do any BSM measurement on 2 & 3, and then measure the spins of 1 & 4 and look at the correlations.

Your own math, which you have posted multiple times, predicts what the result will be: no correlation between 1 & 4. I.e., no entanglement between 1 & 4.

 

[PeterDonis]

This leaves open the possibility for a knowledge-based interpretation without action at a distance, such as Zeilinger is endorsing.

Yes, I agree Zeilinger is using a knowledge interpretation. Which of course is different from the interpretation @DrChinese is using. So it's to be expected that their interpretations of these experiments will be different.

 

[Nullstein]

How would you pick out such subensembles?

If your answer is that you would look at the results of the 1 & 4 measurements and pick out runs on that basis, my response was given in a previous post of mine: that's cherry-picking your statistics. In other words, it's not a valid method of picking a subensemble. The whole point in the entanglement swapping experiments is that the subensemble is picked out without any knowledge whatever of the 1 & 4 measurement results; only the 2 & 3 BSM result is looked at.

There is no way to pick them out without cherry picking, but nevertheless, they are there. This is all that's needed for a knowledge-based interpretation, such as Zeilinger is endorsing, to be viable. Nature need not give use a way to access all information.

Yes, you can test whether they would have been there anyway: just do the initial preparation (in which 1 & 2 are maximally entangled and 3 & 4 are maximally entangled), don't do any BSM measurement on 2 & 3, and then measure the spins of 1 & 4 and look at the correlations.

Your own math, which you have posted multiple times, predicts what the result will be: no correlation between 1 & 4. I.e., no entanglement between 1 & 4.

There is also no entanglement between 1&4 if you perform a BSM and forget about the results. This is the analog situation. What you are doing is trying to compare the full ensemble to a subensemble, which is not appropriate.

What is true is the following: Whether or not we do a BSM: In both cases, there is no entanglement between 1&4 in the full ensemble and there is entanglement between 1&4 in a subensemble. So experimentally, there is no difference and hence no way to distinguish it.

For you, the same advice holds: If you are in contradiction with a Nobel laureate, who is an expert in this field, then you should consider the possibility that the misunderstanding might be on your side and you should consider whether you might need to spend some time trying to understand the argument more carefully.

 

[Nullstein]

Yes, I agree Zeilinger is using a knowledge interpretation. Which of course is different from the interpretation @DrChinese is using. So it's to be expected that their interpretations of these experiments will be different.

How is this an interpretation issue? According to your last post, my understanding is that you believe that it can be experimentally distinguished.

 

[PeterDonis]

There is no way to pick them out without cherry picking, but nevertheless, they are there.

In other words, you think cherry picking is perfectly ok. Sorry, but I disagree. I doubt we're going to resolve that issue.

There is also no entanglement between 1&4 if you perform a BSM and forget about the results.

No, there is no way for you to demonstrate entanglement if you forget about the results. That doesn't mean entanglement isn't there; indeed, by forgetting the results you have given up the ability to test whether it is there or not, which means you cannot justify a claim either way--you can't say there is, and you can't say there isn't.

Which of course is a stupid way to do an experiment.

What is true is the following

You simply cannot assert this as fact. You can say it's what your preferred interpretation says, but that's all. Please review the ground rules for this subforum (which are in a sticky thread at the top of the subforum). As I have already explicitly pointed out in this thread, interpretation discussions are not resolvable.

If you are in contradiction with a Nobel laureate, who is an expert in this field, then you should consider the possibility that the misunderstanding might be on your side

Sorry, but arguments from authority carry no weight here. I have read plenty of Zeilinger, I have a decent understanding of where he is coming from, and what he says makes sense given the interpretation he is using. But that doesn't change the fact that there are multiple interpretations of QM and they say mutually inconsistent things, and such disagreements are not resolvable at our current state of theoretical knowledge.

 

[PeterDonis]

How is this an interpretation issue?

Um, because Zeilinger, by his own admission, is using a knowledge interpretation, and @DrChinese is not?

According to your last post, my understanding is that you believe that it can be experimentally distinguished.

I believe that we can do what I described in post #323 in order to experimentally investigate the question of whether there is any entanglement between 1 & 4 if no BSM is done, yes. To me that seems like the obvious way to do so.

But of course I disagree with you about things like whether cherry picking is a valid way to pick out a subensemble. So I am not at all surprised if you don't share my belief. That just means we have another interpretational disagreement that is not resolvable: what seems to me like an obvious experimental way to investigate something, seems to you to have no bearing whatever on that thing.

 

[Nullstein]

In other words, you think cherry picking is perfectly ok. Sorry, but I disagree. I doubt we're going to resolve that issue.

It depends on what I want to do by cherry picking. If all I want to do is to demonstrate the existence of an entangled subensemble, then yes, by cherry picking, I can do that. If I'm interested in applications, e.g. I want to build a quantum computer and make use of the existing entangled subensemble, then no, cherry picking doesn't help me at all.

No, there is no way for you to demonstrate entanglement if you forget about the results. That doesn't mean entanglement isn't there; indeed, by forgetting the results you have given up the ability to test whether it is there or not, which means you cannot justify a claim either way--you can't say there is, and you can't say there isn't.

Which of course is a stupid way to do an experiment.

The point here is to compute interpretation-independent predictions of quantum mechanics and to compare them to the data from the experiment.

In the BSM case, QM predicts that there is no entanglement between 1&4 in the full ensemble and that there is an entangled subensemble. Moreover, this subensemble can be post-selected based on the result of the BSM.
In the no-BSM case, QM predicts that there is no entanglement between 1&4 in the full ensemble and that there is an entangled subensemble. But this entangled subensemble cannot be post-selected based on the result of a BSM, because there wasn't a BSM to begin with. It is there nevertheless in the sense that it can be found in the data.

These predictions are interpretation independent and can be tested.

Now we have to interpret this. One interpretation is that the entanglement in the subensemble came about by a spooky action due to the BSM. Another viable interpretation is that the subensemble was there from the beginning and we just didn't collect the information necessary to extract it without cherry picking. The second interpretation is the knowledge-based one and it is perfectly viable.

You simply cannot assert this as fact. You can say it's what your preferred interpretation says, but that's all. Please review the ground rules for this subforum (which are in a sticky thread at the top of the subforum). As I have already explicitly pointed out in this thread, interpretation discussions are not resolvable.

This fact follows from the axioms of QM, it is a theorem, so it is not interpretation dependent. I have provided the proof and references. If this were interpretation dependent, then I would agree that it cannot be resolved. However, the question, whether it is interpretation dependent in the first place, can be resolved. If you believe that anything in my math is interpretation dependent, then you should be able to point at one particular equality sign and say: "Look, here, this one is hinges on interpretation XYZ and doesn't follow from the axioms of QM."

Sorry, but arguments from authority carry no weight here.

I agree that it should carry no weight, but on the other hand, you regularly require citations for quite basic results and now you don't accept Zeilinger as a reference.

I have read plenty of Zeilinger, I have a decent understanding of where he is coming from, and what he says makes sense given the interpretation he is using. But that doesn't change the fact that there are multiple interpretations of QM and they say mutually inconsistent things, and such disagreements are not resolvable at our current state of theoretical knowledge.

I agree that there are multiple interpretations that say mutually inconsistent things. But every interpretation needs to agree on the experimental predictions. I'm purposefully only making reference to experimentally testable predictions and their derivation here, in order not to get into the trouble of unresolvable questions.

Um, because Zeilinger, by his own admission, is using a knowledge interpretation, and @DrChinese is not?

Yeah, but at least I am very careful only to talk about the experimentally testable predictions that follow from the axioms of QM, so there should not be any interpretational issues arising. Everyone may use their favorite interpretation, but everyone also needs to agree on the testable predictions.

I believe that we can do what I described in post #323 in order to experimentally investigate the question of whether there is any entanglement between 1 & 4 if no BSM is done, yes. To me that seems like the obvious way to do so.

Well and I believe that what you describe in #323 does not work, because you are comparing the results of a subensemble (in the BSM case) with the results from the full ensemble (in the no-BSM case). What is necessary to resolve this, however, is to compare the results of the full ensemble in the BSM case with the results of the full ensemble in the non-BSM case and the results of the subensemble in the BSM case with the results of the subensemble in the non-BSM case.

But of course I disagree with you about things like whether cherry picking is a valid way to pick out a subensemble. So I am not at all surprised if you don't share my belief. That just means we have another interpretational disagreement that is not resolvable: what seems to me like an obvious experimental way to investigate something, seems to you to have no bearing whatever on that thing.

The reason I disagree with you about your experimental setup is not that we disagree about cherry picking (which we do, I don't deny this), but because I disagree that comparing the results of the full ensemble in one case to the results of a subensemble in the other case, is an appropriate way to test this.

 

[PeterDonis]

you regularly require citations for quite basic results and now you don't accept Zeilinger as a reference

I didn't say the Zeilinger reference was an invalid reference. I said he's using a particular interpretation, so you can't state his position as fact; it's dependent on the interpretation he adopts.

 

[PeterDonis]

In the no-BSM case, QM predicts that there is no entanglement between 1&4 in the full ensemble

Yes.

and that there is an entangled subensemble.

No, QM does not predict this. Cherry picking "predicts" this, but I don't agree that's a valid method.

 

[PeterDonis]

This fact follows from the axioms of QM

We've already been over this multiple times. The math is one thing. Claims about what the math means and what math is relevant is another.

 

[PeterDonis]

Thread closed for moderation.

 

[Dale]

The thread will remain closed