B Is an experiment planned to discern determinism and randomness in QM

[Demystifier]

The hidden variables are certainly measurable.

I said measured/measurable. When they are measurable, they are called hidden because they are assumed to exist even when they are not measured.

 

[vanhees71]

EPR is based on only 2 premises:

1. locality (space-like events do not cause each other)
2. The predicted perfect correlations exist.

As far as I can tell, nothing you said here, about delayed choices, has anything to do with these two premises. So, the conclusion of the argument (hidden variables exist) necessary follows.

Delayed choice has of course nothing to do with it. This could be possible for local HV theories too. The point however is that Bell's inequality follows from such HV theories, and these are violated within QT, and all experiments with very high significance show that indeed Bell's inequalities are violated in perfect agreement with the predictions of QT.

What's of course not fulfilled within QT but in HV theory a la EPR is that in the latter they conclude from the existence of the correlations that the values of the corresponding observables are predetermined but just unknown. Within QT the values of these observables are objectively not predetermined but "really" random.

 

[PeroK]

I absolutely agree that it's complete FAPP. Bell himself (who coined the FAPP acronym) also often emphasized that. But the whole point of quantum foundations is to say something beyond FAPP. It's impossible to discuss quantum foundations and discard all aspects which are beyond FAPP, at the same time. A very dishonest thing to do is to accept only those beyond FAPP aspects which fit your own philosophical prejudices and discard all the others by claiming that you only care about FAPP.

It seems to me there is a difference between a philosophical prejudice that predated subatomic experimental results: namely, determinism; and, a philosophical prejudice that is tailored to the experimental results of the 20th century.

That doesn't mean either is right or wrong but personally I'm inclined towards the random philosophy because it's not trying to fit new experimental results to an a priori philosophy.

If, for example, subatomic results were the only results we had and there was no preconception of determinism, would you still want a deterministic interpretation of QT?

 

[AndreiB]

If, for example, subatomic results were the only results we had and there was no preconception of determinism, would you still want a deterministic interpretation of QT?

Of course, because we have good reasons to believe in locality and locality implies determinism via the EPR argument.

QT itself is in part deterministic (the unitary evolution), the presumed indeterminism being confined to the measurement process which, given our lack of knowledge regarding the exact state of both the system and instrument is not at all unexpected.

 

[vanhees71]

No! Locality (fulfilled in local relativistic QFT) does not imply determinism. The EPR argument is disproven by observation!

We don't have a lack of knowledge about a state for system, which we can prepare in pure states, and this is possible to a high degree of accuracy for photons, atoms/molecules in traps etc. Within QT the randomness is not alone due to the incomplete knowledge about the state of the measurment device (which is unavoidable, because we cannot have complete knowledge about macroscopic systems) but inherent in the system, even when it is prepared in a pure state (which is the most complete possible determination of a system). This is particularly clear for entangled states. Having, e.g., two entangled photons in a Bell state the polarization of the single photon within the pair is utmost indetermined (unpolarized) while the system as a whole is in a completely determined pure state.

 

[AndreiB]

Delayed choice has of course nothing to do with it.

OK, do you agree with the conclusion that EPR proves that locality implies hidden variables?

The point however is that Bell's inequality follows from such HV theories, and these are violated within QT, and all experiments with very high significance show that indeed Bell's inequalities are violated in perfect agreement with the predictions of QT.

Yes, and this restricts the class of allowed local hidden variable theories to the ones that do not satisfy Bell's statistical independence assumption, such as 't Hooft's cellular automaton.

Within QT the values of these observables are objectively not predetermined but "really" random.

QT does not postulate they are "really" random, just gives their probability. I don't even think a "randomness" postulate is even possible since the concept is notoriously difficult to define.

From an experimental point of view "true randomness" is indistinguishable from pseudorandomness. The digits of PI would pass all statistical tests for randomness, yet they are certainly predetermined by a quite simple algorithm.

 

[vanhees71]

Well, according to our knowledge today QT is valid, and there is "true randomness". The time at which a radioactive nucleus decays is "truely random" (with the known probability for survival being approximately $\exp(-\Gamma t)$).

 

[AndreiB]

No! Locality (fulfilled in local relativistic QFT) does not imply determinism.

As explained above, QFT is compatible with both local and non-local views, hence QFT is not a counterexample to EPR.

The EPR argument is disproven by observation!

What observation?

We don't have a lack of knowledge about a state for system

This is true only if you assume completeness. Such an assumption is incompatible with locality (EPR) so we can conclude that it is most likely false.

 

[AndreiB]

The time at which a radioactive nucleus decays is "truely random" (with the known probability for survival being approximately $\exp(-\Gamma t)$).

This "evidence" does not make any sense. Let's assume, for the sake of the argument, that the time of decay is determined by the quark distribution inside the nucleus. In order to predict that time you need to know that distribution and you also need to have the deterministic law that describes the quarks' behavior. Since you, most likely, have none of the above, how would you expect to predict the time of decay? Based on what?

Determinism implies that you can make perfect predictions (in principle) IF you know the initial state and the deterministic law. If you know neither you can't make any prediction. So, your inability to predict decay times is equally expected in a deterministic or a random world.

 

[Demystifier]

Well, according to our knowledge today QT is valid, and there is "true randomness".

Absence of evidence is not evidence of absence.

 

[Demystifier]

OK, do you agree with the conclusion that EPR proves that locality implies hidden variables?

He does not, because he does not agree with the EPR criterion of reality.

 

[Demystifier]

The EPR argument is disproven by observation!

No logical argument has ever been disproved by observation. An observation may disprove an assumption in the argument, but not the argument itself. A logical argument can only be disproved by logic (and its derivatives, such as math).

On the other hand, the assumptions of the EPR argument are QM and locality. I'm sure you don't think that any of those assumptions has been disproved.

The thing in the EPR argument with which you actually disagree is the EPR criterion of reality. Or more precisely, that's what you would disagree with if you understood the EPR argument (without chaning your general opinions about quantum foundations). I know it, because I understand both you and EPR (I studied both for many years).

 

[vanhees71]

Can you then explain what you think is wrong in my understanding of the EPR criterion and with what you think I actually disagree?

Here's what I think:

What I disagree (in agreement with QT) with is the conclusion that due to the 100% correlations described by entangledment the corresponding observables must take determined values before measurement and thus QT were incomplete. One way out is the assumption of hidden variables which we don't know and thus can only describe with probabilities but which always determine the values of all observables.

In the EPR paper, as far as I understand it, they mean by reality indeed that the values must be predetermined, as I take from the statement: "...when the operators corresponding to two physical quantities that do not commute the two quantities cannot have simultaneous reality." I'd not say "cannot have simultaneous reality" but "usually cannot be both determined". Indeed usually two non-commuting self-adjoint operators do not have common eigenvalues, and thus there's no state where the corresponding observables both take determined values.

[Of course, there are exceptions: E.g., two non-collinear components of orbital angular momentum are represented by non-commuting self-adjoint operators but in states with $\ell=0$ all angular-momentum components simultaneously have the determined value 0, but that's not the point here.]

What I conclude from this is that EPR mean by "an observable has reality" that this observable takes a determined value.

 

[PeterDonis]

The hidden variables are certainly measurable. It is what you actually measure.

Bell makes no such assumption. A hidden variable model can have the hidden variables as measurable, but it is not required to.

 

[PhilDSP]

Fast Vacuum Fluctuations and the Emergence of Quantum Mechanics
Gerard 't Hooft
Found.Phys. 51 (2021) 3, 63
DOI: 10.1007/s10701-021-00464-7

"Fast moving classical variables can generate quantum mechanical behavior. We demonstrate how this can happen in a model. The key point is that in classically (ontologically) evolving systems one can still define a conserved quantum energy. ..."

May I ask: what does 't Hooft mean by "fast moving classical variables"? Is he including or requiring or implying superluminal transportation of some form of information for his argument?

 

[DrChinese]

I have no idea what are you trying to say. For example, how does Bohmian mechanics imply non-determinism?

It doesn't. Contextuality *implies* non-determinism... but not strictly so. After all, it is the measurement context that provides us the information about correlations, and nothing else. Surely it must seem odd that no one has ever been able to pinpoint experimentally any prior cause that explains the apparently random outcomes. Again, this is not a strict argument.

 

[DrChinese]

EPR is based on only 2 premises:

1. locality (space-like events do not cause each other)
2. The predicted perfect correlations exist.

As far as I can tell, nothing you said here, about delayed choices, has anything to do with these two premises. So, the conclusion of the argument (hidden variables exist) necessary follows.

No, it is based on more than these 2. You skipped the following crucial assumption (quoting from EPR):

"Indeed, one would not arrive at our conclusion if one insisted that two or more physical quantities can be regarded as simultaneous elements of reality only when they can be simultaneously measured or predicted. On this point of' view, since either one or the other, but not both simultaneously, of the quantities P and Q can be predicted, they are not simultaneously real. This makes the reality of P and Q depend upon the process of measurement carried out on the first system, which does, not disturb the second system in any way. No reasonable definition of reality could be expected to permit this."

It was precisely this assumption, and no other, that Bell challenged - and it was found wanting. This assumption is essentially that reality is not subjective... i.e. is NOT contextual.

 

[AndreiB]

No, it is based on more than these 2. You skipped the following crucial assumption (quoting from EPR):

"Indeed, one would not arrive at our conclusion if one insisted that two or more physical quantities can be regarded as simultaneous elements of reality only when they can be simultaneously measured or predicted. On this point of' view, since either one or the other, but not both simultaneously, of the quantities P and Q can be predicted, they are not simultaneously real. This makes the reality of P and Q depend upon the process of measurement carried out on the first system, which does, not disturb the second system in any way. No reasonable definition of reality could be expected to permit this."

You are right that EPR argued for the simultaneous existence of hidden variables corresponding to non-commuting properties. This argument fails for the reason you exposed earlier, QM is a contextual theory. So, I've tried to make clear that I only look at the properties that are actually measured. So, if the X-spins are measured we can conclude that the X-spins existed before measurement. If the Y-spins are measured we can conclude that the Y-spins existed before measurement, and so on for any other orientation.

If the A measurement of the X-spin does not disturb B (locality) it follows that the A measurement should let B in the same state it was before the A measurement. this is what "not disturb" means. We know that after the A measurement B is in a state of well defined spin, so, B had to be in the same state, of well defined spin, (since it was not disturbed) even before.

It was precisely this assumption, and no other, that Bell challenged - and it was found wanting. This assumption is essentially that reality is not subjective... i.e. is NOT contextual.

Contextual means that the property of interest depends on the experimental context. The experimental context is simply the way the experimental parts are arranged, it's as objective as you can get.

A theory is contextual when the objects described by that theory interact. In order to predict Earth's future state it's not enough to know its present position and momentum, you also need "contextual" information, like what other massive objects are, where they are and where they are going. You need to know about the Sun, Moon, Jupiter, the rest of the galaxy and so on. QM is the same. It has nothing to do with subjectivity. the outcome of an experiment only depends on how the experimental parts are arranged, not on the mood of the experimenter. A two slit experiment does not show interference if a particle detector is placed at the slits even if the experimenter does not know about the detectors. It's the objective description of the experiment that matters.

 

[PeroK]

In order to predict Earth's future state it's not enough to know its present position and momentum, you also need "contextual" information, like what other massive objects are, where they are and where they are going. You need to know about the Sun, Moon, Jupiter, the rest of the galaxy and so on. QM is the same.

QM is not the same as classical mechanics! There never has been any viable classical model of the atom. QM didn't replace CM at the subatomic level, because CM was never successful in that case.

Forget about entanglement and the Bell theorem, and focus solely on electron spin. First, there must be a lot of hidden variables. You cannot model electron spin by specifying a single axis of rotation. There must be a hidden variable for every conceivable measurement angle.

This is where Einstein was being slightly disingenuous IMHO. Of course, electron spin could theoretically be governed by hidden variables. But, it's incredibly messy to say how those variables are organised.

Einstein was already on shaky ground even without entanglement. That's why Bohr, Heisenberg and the rest were convinced of QM decades before the tests of Bell's theorem.

The battleground shifted to entanglement because Einstein believed that would undermine QM. But of course, the results vindicated QM and further undermined hidden variables. Not the reverse.

Leaving aside BM, the battleground after the test of Bell's theorem has shifted to ever more bizarre reworking of determinism, such as superdeterminism. Whereas, those who broadly support orthodox QM have hardly had to modify their understanding in a century.

You have this whole debate back to front. There is no solid deterministic edifice. Instead, there is a solid QM edifice, which the determinists by hook or by crook are hoping to topple. But, again leaving BM aside, there is no solid deterministic edifice waiting to replace QM. Only wishful thinking held together by a belief in unknown laws that result in almost magical correlations.

That's why QM is mainstream science and hidden variables are not.

 

[Demystifier]

This assumption is essentially that reality is not subjective... i.e. is NOT contextual.

I strongly disagree with identification subjective=contextual. If something depends on the apparatus, but not on a conscious perception, then I would call it contextual but not subjective. It is only QBism that strictly identifies contextual with subjective.

 

[AndreiB]

May I ask: what does 't Hooft mean by "fast moving classical variables"? Is he including or requiring or implying superluminal transportation of some form of information for his argument?

No, 't Hooft's model is a cellular automaton, an array of cells, each cell updating at some time interval according to some local rule (the state of a cell changes based on its previous state and the states of the surrounding cells). At each "tick" all cells update. This implements a speed limit, so the theory is local.

Nothing actually moves, it's like a TV screen. The pixels change color, but the pixels themselves don't move. The perceived motion is limited at one cell/update time.

By "fast" variables I think he means variables that change their state very quickly (still not faster than the update time). The on/off state of a pixel will change, say 60 times/s, this would be "fast". But the color of the pixel needs not change that fast, so it would be "slow". Or you can think of a piece of wood floating on water. It's position is ultimately determined by the water molecules it interacts with. The motion of those molecules is much faster than the motion of the wood.

 

[gentzen]

So, if the X-spins are measured we can conclude that the X-spins existed before measurement. If the Y-spins are measured we can conclude that the Y-spins existed before measurement, and so on for any other orientation.

Forget about entanglement and the Bell theorem, and focus solely on electron spin. First, there must be a lot of hidden variables. You cannot model electron spin by specifying a single axis of rotation. There must be a hidden variable for every conceivable measurement angle.

Have you tried to work out how this predetermination of the spin works in BM? Let me try myself ... how do I actually model the fact that the X-spins have been measured by both Alice and Bob?

Leaving aside BM, the battleground after the test of Bell's theorem has shifted to ever more bizarre reworking of determinism, such as superdeterminism.

Well, superdeterminism is both a label for an assumption (its absence) in Bell's theorem, and a label for some absurd mechanism people imagined how it could be used to produce the correlations predicted by QM. But there is a deeper issue, namely that our ignorance of initial conditions does not always come with a probability distribution (or improper Bayesian prior) for that unknow part. There is a probability distribution for the particle positions in BM, but note that we are also ignorant about most parts and details of the wavefunction. And there seems to be no canonical probability distribution for that part.

 

[vanhees71]

No, it is based on more than these 2. You skipped the following crucial assumption (quoting from EPR):

"Indeed, one would not arrive at our conclusion if one insisted that two or more physical quantities can be regarded as simultaneous elements of reality only when they can be simultaneously measured or predicted. On this point of' view, since either one or the other, but not both simultaneously, of the quantities P and Q can be predicted, they are not simultaneously real. This makes the reality of P and Q depend upon the process of measurement carried out on the first system, which does, not disturb the second system in any way. No reasonable definition of reality could be expected to permit this."

It was precisely this assumption, and no other, that Bell challenged - and it was found wanting. This assumption is essentially that reality is not subjective... i.e. is NOT contextual.

EPR overlook that the entangled particles are one system. The very definition of entanglement is that, well, subsystems are entangled or, as Einstein said in a way better than the EPR paper in 1948, inseparable.

A. Einstein, Quantenmechanik und Wirklichkeit, Dialectica 2,
320 (1948),
https://doi.org/10.1111/j.1746-8361.1948.tb00704.x

I don't know, how you come to the conclusion that QT is not objective though. Are you a qubist now?

 

[AndreiB]

QM is not the same as classical mechanics!

I didn't say that.

There never has been any viable classical model of the atom. QM didn't replace CM at the subatomic level, because CM was never successful in that case.

I guess you wanted to say "QM DID replace CM". Agreed.

Forget about entanglement and the Bell theorem, and focus solely on electron spin. First, there must be a lot of hidden variables. You cannot model electron spin by specifying a single axis of rotation. There must be a hidden variable for every conceivable measurement angle.

Indeed.

This is where Einstein was being slightly disingenuous IMHO. Of course, electron spin could theoretically be governed by hidden variables. But, it's incredibly messy to say how those variables are organised.

I didn't say its easy. The problem is that I have a sound logical argument that links the existence of those hidden variables to locality, and locality is very likely to be true. I think it is perfectly reasonable to choose a "messy" and local theory over a simple and non-local one. You need to take into account that modern physics is not only QM, it's also GR. Making GR indeterministic and non-local does not seem to work, right?

The battleground shifted to entanglement because Einstein believed that would undermine QM.

He proved QM incomplete or non-local. That result still stands. The mistake of EPR was that they insisted on the existence of HV for non-commuting properties, allowing Bohr to escape. But the argument works fine for only the measured values. Had Bohr been presented with this, simplified argument, he would have been defeated.

But of course, the results vindicated QM and further undermined hidden variables. Not the reverse.

What results? EPR assumed QM gives the right predictions, so QM's predictions being confirmed could not possibly undermine the argument. QM predicts perfect correlations. Local randomness would predict correlations 50% of the time. Local randomness is falsified, end of story.

Leaving aside BM, the battleground after the test of Bell's theorem has shifted to ever more bizarre reworking of determinism, such as superdeterminism.

This is not bizarre. Since EPR proved local indeterminism impossible, and Bell proved a type of local determinism impossible, the only local option remains superdeterminism. What I find even more bizarre is the belief in logically impossible fundamental theories, like local and non-deterministic ones.

Whereas, those who broadly support orthodox QM have hardly had to modify their understanding in a century.

This is why the unification program along string theory is such a success. As long as the local/non-local character of QM is irrelevant (as in the case of optimising a superconductor), the theory works fine and one is free to ignore EPR. When locality matters, such is the case of unification program, nature hits the indeterminist hard, again and again.

You have this whole debate back to front. There is no solid deterministic edifice.

Yes, there is. Locality is very likely right, there is no evidence against it, all our theories accept it, including GR. We have a sound argument proving that locality implies determinism. I cannot think of any better evidence for determinism than that. 100% correlations between space-like measurements utterly falsify indeterminism.

Instead, there is a solid QM edifice, which the determinists by hook or by crook are hoping to topple.

You simply assume QM is fundamentally indeterministic, instead of a statistical theory. This assumption is unjustified. What's wrong with a solid statististical edifice?

But, again leaving BM aside, there is no solid deterministic edifice waiting to replace QM.

Indeed.

Only wishful thinking held together by a belief in unknown laws that result in almost magical correlations.

You seem not to understand that determinism is imposed by locality as a conclusion of a sound argument. Determinism is not an assumption. It's not about what one desires or finds philosophically pleasing. The correlations are "magical" only for the non-determinist. For a determinist they are just another case of past common cause, an explanation that always worked till now.

 

[PeroK]

@AndreiB I'm not sure we've ever had anyone on here before who has argued that the 20th Century provided an unbroken and unassailable vindication of deterministic mechanics!

What you are posting is, IMHO, pure nonsense.

The QM model of the atom is the only viable model there has ever been. There's no other explanation for chemistry, nor has there ever been.

 

[AndreiB]

He does not, because he does not agree with the EPR criterion of reality.

i don't think the criterion can be denied. The criterion says that if you an element of reality exists if you can predict that quantity without disturbing the system. Now, if you don't disturb B by your A measurement, and QM says that your A measurement leaves B in well-defined spin state it logically follows that B was in a well-defined spin state before your A measurement. As Tim Maudlin said, the criterion is "analytic", it necessary follows from the meaning of the words. The only way out is to deny that A does not disturb B (and accept non-locality) or deny the QM prediction (that the state after the A measurement is a well-defined spin, opposite to the one found at A.

 

[vanhees71]

Logic only tells you formal rules to deduce from the assumptions on the truth of propositions the truth of other propositions.

An important part we have learned about Nature is the atomistic structure of matter, which implies that it is impossible to measure the "atoms" without disturbing them, because we need at least one other "atom" to measure. There's no way to measure anything without interacting with it in such a way as to disturb the system.

That's also true in a way when measuring far distant entangled parts of a quantum system. Though there is no non-local interaction by measurement at A's position on the part at B's position, nevertheless the system as a whole is disturbed in such a way that the entanglement is destroyed. This is unavoidable. You can also gain information on the outcome of certain measurements at B having measured A, but this does not in any way justify EPR's conclusion that this measured value at B was predetermined before A's measurement. To the contrary, for a maximally entangled system the observables of the parts are both usually maximally indetermined.

 

[AndreiB]

@AndreiB I'm not sure we've ever had anyone on here before who has argued that the 20th Century provided an unbroken and unassailable vindication of deterministic mechanics!

I didn't make such a claim. Local indeterminism has been refuted, not indeterminism per se. If you disagree with the argument, please present your refutation here! here is the argument again:

If the A measurement of the X-spin does not disturb B (locality) it follows that the A measurement should let B in the same state it was before the A measurement. this is what "not disturb" means. We know that after the A measurement B is in a state of well defined spin, so, B had to be in the same state, of well defined spin, (since it was not disturbed) even before.

OK? Explain me what is wrong with the above argument!

The QM model of the atom is the only viable model there has ever been. There's no other explanation for chemistry, nor has there ever been.

This does not prove that QM must be fundamentally indeterministic. A correct, yet statistical approximate theory could do the job.

 

[gentzen]

I guess you wanted to say "QM DID replace CM". Agreed.

Be very careful with such comments. They might indicate a missing effort from your side to try to understand what your communication partner was saying.

The problem is that I have a sound logical argument that links the existence of those hidden variables to locality, and locality is very likely to be true.

Even so the word "local" has a clear meaning, it is a property of something in some context. Asserting "locality is true" is a pretty meaningless statement, as long as no context is given, and even with context you should still try to clarify what it is that is asserted to be local.

For example, in a classical computer, communication of information is normally local. Storage of information can be local too, but sometimes error correction codes are used to explicitly avoid locality of storage, for example on CD/DVDs. The state in QM behaves like such error correction codes in certain ways, and seems to exhibit a similar kind of nonlocality (or inseparability as many people here prefer to say).

 

[Demystifier]

i don't think the criterion can be denied. The criterion says that if you an element of reality exists if you can predict that quantity without disturbing the system. Now, if you don't disturb B by your A measurement, and QM says that your A measurement leaves B in well-defined spin state it logically follows that B was in a well-defined spin state before your A measurement. As Tim Maudlin said, the criterion is "analytic", it necessary follows from the meaning of the words. The only way out is to deny that A does not disturb B (and accept non-locality) or deny the QM prediction (that the state after the A measurement is a well-defined spin, opposite to the one found at A.

I think there is a logical third way out. I can predict what will be the result of measuring B at the time when I measure B, but I can deny that B had any value at all before I measured it. That's indeed how Copenhagenish (non-realist) type of interpretations work.