Confused by nonlocal models and relativity

[vanhees71]

QFT is microcausal and Poincaré covariant, but it doesn't obey EPR realism and Decorrelating explanations (a part of Reichenbach's common cause principle).

The first is nothing more than what is implied by the Kochen-Specker theorem (i.e. no values pre-exist the measurement/the measurement creates values) and the latter is just the multiple sample spaces point we've spoken of before.

But isn't EPR realism ruled out by the results of the Bell tests? The ironic point is that EPR realism isn't realistic at all, but Q(F)T is!

 

[Heikki Tuuri]

I recall @Demystifier saying that one can extend the Bohm model to relativistic theories.

Since quantum mechanics is a wave phenomenon, and we need to know the entire wave function to do calculations, a local realistic theory which just has N particles at sharply defined positions and velocities is hopeless. Newtonian mechanics is not a wave, and there local realism is successful.

 

[Elias1960]

But isn't EPR realism ruled out by the results of the Bell tests? The ironic point is that EPR realism isn't realistic at all, but Q(F)T is!

QFT does not even claim to define an ontology, thus, is certainly not realistic.

And EPR realism is, of course, not ruled out. Only together with Einstein causality. The idea to reject EPR realism is, of course, quite natural for all those who think questioning Einstein causality is anathema. They prefer giving up realism as well as causality itself (reducing Einstein causality to signal causality) to prevent a preferred frame.

The reference for relativistic Bohmian field theory is

Bohm.D., Hiley, B.J., Kaloyerou, P.N. (1987). An ontological basis for the quantum theory, Phys. Reports 144(6), 321-375

In fact, the Bohmian version of QED was already part of Bohm's original paper.

 

[DarMM]

But isn't EPR realism ruled out by the results of the Bell tests? The ironic point is that EPR realism isn't realistic at all, but Q(F)T is!

Yes it is ruled out. This is part of why I wish "real" wasn't in the name of these things, because as you're essentially saying then we end up saying the most accurate theory we have is "not real" or some nonsense like that.

EPR realism should be "device independent certain value attribution" or something.

 

[bhobba]

Good old EPR again. The theorem is of course true but what you think it is telling us is 'discussed' a lot. I agree with Vanhee's - its rather difficult to accommodate the usual views with QFT - such is even suggested by one of its founding principles - the Cluster Decomposition Property (it is unnatural to apply it to correlated systems ):
https://www.physicsforums.com/threads/cluster-decomposition-in-qft.547574/
I do not strictly think its actually impossible but I have no idea how you would do it. I simply take it at face value - QM is a Generalized Probability theory so it is not surprising it has probabilistic features different than standard probability theory. I think it simply shows in QM correlations have different statistical properties than ordinary probability theory and leave it at that. But I do strongly object to people saying it shows QM is non-local. It 100% for sure does not.

Thanks
Bill

 

[bhobba]

This is part of why I wish "real" wasn't in the name of these things, because as you're essentially saying then we end up saying the most accurate theory we have is "not real" or some nonsense like that.

I think I may have posted Wienberg's paper before in this thread but in case I haven't here it is again:
https://www.physics.utah.edu/~detar/phys4910/readings/fundamentals/weinberg.html
Here is the part I want to highlight:
'I remarked in a recent article in The New York Review of Books that for me as a physicist the laws of nature are real in the same sense (whatever that is) as the rocks on the ground. A few months after the publication of my article I was attacked for this remark by Richard Rorty. He accused me of thinking that as a physicist I can easily clear up questions about reality and truth that have engaged philosophers for millennia. But that is not my position. I know that it is terribly hard to say precisely what we mean when we use words like "real" and "true." That is why, when I said that the laws of nature and the rocks on the ground are real in the same sense, I added in parentheses "whatever that is." I respect the efforts of philosophers to clarify these concepts, but I'm sure that even Kuhn and Rorty have used words like "truth" and "reality" in everyday life, and had no trouble with them. I don't see any reason why we cannot also use them in some of our statements about the history of science. Certainly philosophers can do us a great service in their attempts to clarify what we mean by truth and reality. But for Kuhn to say that as a philosopher he has trouble understanding what is meant by truth or reality proves nothing beyond the fact that he has trouble understanding what is meant by truth or reality.'

It is strange that we have no problem about using concepts like truth and reality in everyday language - even those deeply involved with trying to figure out what they mean do it. But scientists also use it in the way everyday language does. Occasionally they are careful about using it, but for the most part do not worry about it. Science has made enormous progress doing this. Perhaps not worrying too much about it is the better strategy. Some things just seem so fundamental precisely pinning them down may be impossible and perhaps not leading anywhere.

Technically counterfactual definiteness is the right term to use in discussions about EPR, but if you use real does it matter that much? It seems to me to be just a way of fobbing off those concerned about such terms, which IMHO shouldn't really be worried about by scientists anyway. That said when people challenge me on things like that I sometimes resort to counterfactual definiteness type answers.

Thanks
Bill

 

[bhobba]

QFT does not even claim to define an ontology, thus, is certainly not realistic.

I think the author of the bible on QFT, Weinberg, might not agree with you on that.

Personally I think it is similar to the discussions on if the quantum state is real or not. This is related to similar discussions on is probability real or not. Those that use probability, such as Actuaries, simply accept the Kolmogorov Axioms and human beings using abstraction can connect them to actual applications. Its a view you will find in more detail in the beginning pages of Fellers classic on probability. But then again we have rigorous probability theory that simply looks at what the axioms imply. Its like Euclidean Geometry - it speaks of lines, points etc, but how do we know what a line or a point is when we see it. Students proving geometric theorems seem to have no problem. I think its all simply part of a humans ability to abstract away inessentials in applying a theory. Using that view I think such issues are rather meaningless.

Thanks
Bill

 

[Elias1960]

Technically counterfactual definiteness is the right term to use in discussions about EPR,

No. Counterfactual definiteness only follows from EPR realism taken together with Einstein causality. Without Einstein causality, you have no base assuming that there is no causal influence of the choice of measurement at A on the situation in B. But the absense of such influences is a precondition which is explicit in the EPR criterion.

It is strange that we have no problem about using concepts like truth and reality in everyday language - even those deeply involved with trying to figure out what they mean do it. But scientists also use it in the way everyday language does. Occasionally they are careful about using it, but for the most part do not worry about it. Science has made enormous progress doing this.

In particular, it is worth not to worry about what philosophers like Rorty who are deeply confused about theories of truth think. The standard common-sense theory of truth - correspondence with reality - is fine for science, except for those who propose to reject realism.

I simply take it at face value - QM is a Generalized Probability theory so it is not surprising it has probabilistic features different than standard probability theory. I think it simply shows in QM correlations have different statistical properties than ordinary probability theory and leave it at that. But I do strongly object to people saying it shows QM is non-local. It 100% for sure does not.

To repeat myself: In the objective Bayesian interpretation, the rules of probability theory are simply rules of consistent logical thinking in a situation where we have insufficient information. Probability is simply part of logic, the logic of plausible reasoning. Such laws of reasoning are the base for evaluating scientific theories, thus, they have to hold in scientific theories, and if they don't, the theory should be rejected as logically inconsistent.

So, whenever you hear that the rules of classical probability do not hold in quantum theory, you follow a logically inconsistent interpretation of quantum theory. There is no "quantum logic", and there is also no "quantum logic of plausible reasoning". There is only logically inconsistent confusion. This may be a misinterpretation of some different mathematical structures describing something completely different. This may be a confusion of the meaning of words in common language used to formulate logical claims. Whatever it is, it is not a new logic. Logic is what we use to think about nature, it is not an empirical theory which could be falsified by some experiment, but is what we have to use to find out that some experiment really falsifies a theory.

If quantum theory would be really in conflict with the logic of plausible reasoning, it should be simply rejected as logically inconsistent. It isn't. Interpretations that have no conflict with classical logic and probability theory, as well as with classical realism and classical causality, exist so that there is no such internal inconsistency of QT or QFT.

But nothing prevents interpretations of QT as well as of QFT from being logically inconsistent. Peer-review does no longer help, given that there are a lot of publications about not even well-formulated interpretations like many worlds.

 

[DarMM]

Technically counterfactual definiteness is the right term to use in discussions about EPR, but if you use real does it matter that much? It seems to me to be just a way of fobbing off those concerned about such terms, which IMHO shouldn't really be worried about by scientists anyway. That said when people challenge me on things like that I sometimes resort to counterfactual definiteness type answers

Counterfactual definiteness is strictly weaker than EPR realism. This is part of the difference between EPR realism and KS noncontextuality, the former being the stronger assumption.

My dislike for the use of "real" is nothing to do with a reluctance to use it with QM, it's literally the exact opposite. It's because when we say we reject "realism" it makes it sound as if we are rejecting QM being "real" in the everyday sense of the word rather than the highly technical sense meant by EPR realism.

The reason I wished it was called something else is because I want to use "real" for QM. As I said above:

because as you're essentially saying then we end up saying the most accurate theory we have is "not real" or some nonsense like that

 

[martinbn]

QFT does not even claim to define an ontology, thus, is certainly not realistic.

And EPR realism is, of course, not ruled out. Only together with Einstein causality. The idea to reject EPR realism is, of course, quite natural for all those who think questioning Einstein causality is anathema. They prefer giving up realism as well as causality itself (reducing Einstein causality to signal causality) to prevent a preferred frame.

The reference for relativistic Bohmian field theory is

Bohm.D., Hiley, B.J., Kaloyerou, P.N. (1987). An ontological basis for the quantum theory, Phys. Reports 144(6), 321-375

In fact, the Bohmian version of QED was already part of Bohm's original paper.

Slitely off topic, but can you explain the difference between Einstein causality and signal causality.

 

[DarMM]

To repeat myself: In the objective Bayesian interpretation, the rules of probability theory are simply rules of consistent logical thinking in a situation where we have insufficient information. Probability is simply part of logic, the logic of plausible reasoning. Such laws of reasoning are the base for evaluating scientific theories, thus, they have to hold in scientific theories, and if they don't, the theory should be rejected as logically inconsistent

This is a claim you've made before, that rejecting Kolmogorov's axioms is "logically impossible" or something like that.

We have now constructed generalized probability theories, not just QM such as PR box probabilities, so this just is not true. Your contention that classical probability is the only logically possible such theory is contradicted by the literature.

Other probability theories simply reject some of Cox's assumptions.

 

[vanhees71]

QFT does not even claim to define an ontology, thus, is certainly not realistic.

And EPR realism is, of course, not ruled out. Only together with Einstein causality. The idea to reject EPR realism is, of course, quite natural for all those who think questioning Einstein causality is anathema. They prefer giving up realism as well as causality itself (reducing Einstein causality to signal causality) to prevent a preferred frame.

The reference for relativistic Bohmian field theory is

Bohm.D., Hiley, B.J., Kaloyerou, P.N. (1987). An ontological basis for the quantum theory, Phys. Reports 144(6), 321-375

In fact, the Bohmian version of QED was already part of Bohm's original paper.

Theoretical physics is about describing what we objectively observe in nature, and relativistic QFT is the most comprehensive model we have today about how nature behaves on the most microscopic scale we have hitherto observed, no more no less. Whether or not this provides and "ontology" is irrelevant as far as physics is concerned.

EPR realism is ruled out by the highly accurate observations of the violation of Bell's inequality, while Einstein causality is very well confirmed by all observations.

I like the de Broglie/Bohm theory as a non-local alternative interpretation of non-relativistic QM. It's, however, neither needed to describe the corresponding observations in atomic and solid-state physics, where the non-relativistic QM approximation of relativistic QFT is valid, nor has it been convincingly extended to relativistic QT.

 

[atyy]

But I do strongly object to people saying it shows QM is non-local. It 100% for sure does not.

Standard textbook QM is nonlocal if the quantum state is real: https://arxiv.org/abs/0706.2661

Whether the quantum state is real is a matter of interpretation, so one cannot say that 100% for sure QM is not nonlocal.

 

[DarMM]

Standard textbook QM is nonlocal if the quantum state is real: https://arxiv.org/abs/0706.2661

Whether the quantum state is real is a matter of interpretation, so one cannot say that 100% for sure QM is not nonlocal.

Textbook QM with $\psi$ real is inconsistent though. If you want $\psi$ to be real you need MWI or hidden variables.

 

[Demystifier]

Newtonian mechanics is not a wave, and there local realism is successful.

Newtonian gravity is nonlocal.

 

[atyy]

Textbook QM with $\psi$ real is inconsistent though. If you want $\psi$ to be real you need MWI or hidden variables.

Why is textbook QM with a real quantum state inconsistent? It has a measurement problem, but once the classical/quantum cut is chosen, it is consistent, as far as I know.

 

[DarMM]

Why is textbook QM with a real quantum state inconsistent? It has a measurement problem, but once the classical/quantum cut is chosen, it is consistent, as far as I know.

You can exploit level inconsistency between observers to derive a contradiction in extended Wigner's Friend scenarios.

Now neither Bohr, Heisenberg or several others viewed $\psi$ as a physical degree of freedom, so the extended Wigner's Friend set ups don't affect interpretations people actually hold like Copenhagen.

 

[martinbn]

Why is textbook QM with a real quantum state inconsistent? It has a measurement problem, but once the classical/quantum cut is chosen, it is consistent, as far as I know.

What does real quantum state mean?

 

[atyy]

You can exploit level inconsistency between observers to derive a contradiction in extended Wigner's Friend scenarios.

Now neither Bohr, Heisenberg or several others viewed $\psi$ as a physical degree of freedom, so the extended Wigner's Friend set ups don't affect interpretations people actually hold like Copenhagen.

The cut prevents you from having the inconsistency. In a sense it is just observer dependent reality. The wave function is real for the observer who is using it. To that observer, there cannot be any other observers in the quantum part of the cut.

 

[DarMM]

The cut prevents you from having the inconsistency. In a sense it is just observer dependent reality. The wave function is real for the observer who is using it. To that observer, there cannot be any other observers in the quantum part of the cut.

Are you saying there is a single objective cut defined over all of space? Different observers aren't permitted to have different cuts?

 

[Tendex]

Slitely off topic, but can you explain the difference between Einstein causality and signal causality.

Ha, that's a good one. Let me try and untie a little this cute knot. Signal causality is what's left when you take SR usual causality structure(i.e. Einstein causality or regular EPR local realism ) and in order to make it fit into a quantum theory with no classically predetermined measure values you only leave operational "no faster than light" signaling.
But, you do it by appealing to an essential element of Einstein causality of SR(in what is called "(anti)commutation of spacelike separated operators), i.e behaviour outside the light cone, which of course implies there must be an inside the SR light cone structure somewhere but chose to ignore it for the moment to the effect of rejecting causal realism as QFT is part of quantum theory and its experiments violate Bell's inequalities constructed according to causal realism.
And then of course people has a bit of confusion with all these terms like "real", " causal" and "local" which is not surprising since they usually are not even aware of all this mess.

 

[vanhees71]

Standard textbook QM is nonlocal if the quantum state is real: https://arxiv.org/abs/0706.2661

Whether the quantum state is real is a matter of interpretation, so one cannot say that 100% for sure QM is not nonlocal.

It depends on, what you mean by "nonlocal". Non-relativistic quantum mechanics is of course non-local since there are interactions at a distance as a rule rather than an exception. Relativistic quantum field theory by construction has only local interactions, i.e., the Hamiltonian density commutes with any local observable at space-like distances of their arguments ("microcausality").

All QT is "nonlocal" in the sense that there is entanglement which describes correlations which can be between far-distant parts of a quantum system. Einstein invented for this kind of "nonlocality" the notion of "inseparability". That's much more concise than to call that "nonlocality".

I'd plead to always say what one means by "local" or "nonlocal" (usually one shortly says relativistic QFT is "local" because of the locality of interactions) or better avoid this expression if not clearly specified.

 

[atyy]

Are you saying there is a single objective cut defined over all of space? Different observers aren't permitted to have different cuts?

Yes, different observers cannot have different cuts. So there is a single cut defined over all of space. Whether the cut is objective or subjective is not relevant. At this level, there is a single user or single pair of users (Alice and Bob). To them the quantum formalism will be consistent, and also predict the violation of the Bell inequalities. If they take the wave function to be real, then the formalism is nonlocal to them.

 

[Tendex]

Probability is simply part of logic, the logic of plausible reasoning. Such laws of reasoning are the base for evaluating scientific theories, thus, they have to hold in scientific theories, and if they don't, the theory should be rejected as logically inconsistent.

I agree with this provided one leaves room for the possibility that the current formal mathematical logic can be extended to a more general one for physical theories if at some point experiments lend themselves to a more comprehensive form of scientific theory. This is of course not my idea but was contemplated by people like Poincaré, Weyl and Brouwer.

 

[martinbn]

Yes, different observers cannot have different cuts. So there is a single cut defined over all of space. Whether the cut is objective or subjective is not relevant. At this level, there is a single user or single pair of users (Alice and Bob). To them the quantum formalism will be consistent, and also predict the violation of the Bell inequalities. If they take the wave function to be real, then the formalism is nonlocal to them.

So, what does it mean to take the wave function to be real!?

 

[Elias1960]

I think the author of the bible on QFT, Weinberg, might not agree with you on that.

Maybe. But if there is not even agreement about what is real in QM, and the minimal interpretation of QM remains silent about this, how can one expect agreement in QFT, which is, last but not least, the same QM applied to fields? The minimal interpretation of QFT will be even more minimal because it will exclude the Schroedinger picture too (given that it does not look Lorentz-covariant, and everything which is not Lorentz-covariant has to be hidden). Moreover, there is some additional disagreement about the ontology, namely particles vs. fields. There will be hardly an agreement about this.

This is a claim you've made before, that rejecting Kolmogorov's axioms is "logically impossible" or something like that.
We have now constructed generalized probability theories, not just QM such as PR box probabilities, so this just is not true. Your contention that classical probability is the only logically possible such theory is contradicted by the literature.
Other probability theories simply reject some of Cox's assumptions.

How many times I have to repeat the same trivialities? That you can invent mathematical structures that violate some of the rules of logic or probability theory is not questioned at all. What is questioned is that they are applicable as the laws of logic, including the logic of plausible reasoning. You cannot simply change the laws of reasoning.

(You can, of course, doubt some laws of reasoning, and refuse to apply them yourself. This will weaken your ability to argue reasonably, thus, will be quite stupid. But to invent new, different laws of reasoning will simply transform your reasoning into nonsense.)

Thank you that by adding "or something like that" you have at least admitted that I have not written the nonsense you have invented out of thin air.

EPR realism is ruled out by the highly accurate observations of the violation of Bell's inequality, while Einstein causality is very well confirmed by all observations.

Wrong. EPR realism is as well very well confirmed by all observations. What is falsified by the violation of Bell's inequality is the combination of EPR realism with Einstein causality, as well as the combination of a meaningful notion of causality at all (which has to include Reichenbach's common cause principle) with Einstein causality.

So, preserving Einstein causality given the violation of Bell's inequality requires one to give up realism in the extremely weak form of the EPR criterion of reality and also to reduce causality to positivistic signal causality, rejecting Reichenbach's principle of common cause. The tobacco industry will be happy with rejecting this horrible principle which forces them to find causal explanations for the correlations between smoking and lung cancer. Everybody else will simply ignore this, so that after this we have two scientific methods, one applicable for essentially everything, which considers it a necessity to find causal explanations of observable correlations, and a special exception for the violation of the Bell inequality, where one can simply ignore them, given that every causal explanation would violate Einstein causality, thus, would be anathema.

To read

... nor has it been convincingly extended to relativistic QT.

in the answer to a post where the standard reference to this extension has been given makes me wonder if giving such references makes sense at all.

Slitely off topic, but can you explain the difference between Einstein causality and signal causality.

Signal causality means if $A \to B$ then you can also send a signal from A to B. That means, doing something at A changes some probability at B. The violations of the Bell inequalities do not allow such signaling, because they allow for two causal explanations, either $A \to B$ or $B \to A$. If this could be used to send a signal from A to B, the explanation $B \to A$ would be impossible.

 

[martinbn]

Signal causality means if $A \to B$ then you can also send a signal from A to B. That means, doing something at A changes some probability at B. The violations of the Bell inequalities do not allow such signaling, because they allow for two causal explanations, either $A \to B$ or $B \to A$. If this could be used to send a signal from A to B, the explanation $B \to A$ would be impossible.

But my question was about the difference between this and Einstein causality. They still seem the same to me. Einstein causality means that if $A \to B$, the A is in tha past lightcone of B, so you can send a signal from A to B. If you can send such a signal, then A is in the past of B. So, what is the difference?

 

[DarMM]

Yes, different observers cannot have different cuts. So there is a single cut defined over all of space. Whether the cut is objective or subjective is not relevant. At this level, there is a single user or single pair of users (Alice and Bob). To them the quantum formalism will be consistent, and also predict the violation of the Bell inequalities. If they take the wave function to be real, then the formalism is nonlocal to them.

I think it is relevant. The usual understanding is that if the cut is subjective then you get contradictions from extended Wigner's Friend scenarios. If the cut is objective then you have Many-Worlds, just we also have a non-relativistic covariant degree of freedom defined over all of space which specifies where the worlds "stop". So it's either inconsistent or MWI + nonlocal degree of freedom.

There are other problems with textbook QM + $\psi$ directly describing a physical degree of freedom (Dirac-Von Neumann) unlike textbook QM + $\psi$ as a book keeping device for probabilities (Copenhagen). It's not really an interpretation anybody in Foundations holds to anymore.

 

[DarMM]

So, what does it mean to take the wave function to be real!?

Essentially that $\psi$ directly describes a physical degree of freedom in the system. It's a shorthand common in physics.

 

[DarMM]

How many times I have to repeat the same trivialities? That you can invent mathematical structures that violate some of the rules of logic or probability theory is not questioned at all. What is questioned is that they are applicable as the laws of logic, including the logic of plausible reasoning. You cannot simply change the laws of reasoning.

You can change the laws of probability, as evidenced by Quantum Probability. A probability theory that doesn't obey the Kolmogorov axioms or obey Cox's assumptions.

 

[Elias1960]

But my question was about the difference between this and Einstein causality. They still seem the same to me. Einstein causality means that if $A \to B$, the A is in tha past lightcone of B, so you can send a signal from A to B. If you can send such a signal, then A is in the past of B. So, what is the difference?

The difference is that if you have a 100% correlation between A and B spacelike separated, then in normal, classical causality this requires an explanation, either by some direct causal influence $A \to B$ or $B\to A$ or by some common cause $C\to A, C\to B$.

This requirement for a causal explanation of observable correlations has to be given up if one follows those who reject all explanations which use a preferred frame. But once no other explanations exist, one has to give up causality.

The difference is that if you require a causal explanation, then you have only two choices for spacelike separated events A and B: Either $A \to B$ or $B\to A$. Both explanations would violate Einstein causality, thus, Einstein causality would be falsified if this principle would not be given up. A common cause explanation $C\to A, C\to B$ is what the violation of the Bell inequality excludes.

 

[martinbn]

Essentially that $\psi$ directly describes a physical degree of freedom in the system. It's a shorthand common in physics.

That is still unclear to me. How could it be a degree of freedom! Anyway, since I don't know it I will stay out of it.

 

[martinbn]

The difference is that if you have a 100% correlation between A and B spacelike separated, then in normal, classical causality this requires an explanation, either by some direct causal influence $A \to B$ or $B\to A$ or by some common cause $C\to A, C\to B$.

This requirement for a causal explanation of observable correlations has to be given up if one follows those who reject all explanations which use a preferred frame. But once no other explanations exist, one has to give up causality.

The difference is that if you require a causal explanation, then you have only two choices for spacelike separated events A and B: Either $A \to B$ or $B\to A$. Both explanations would violate Einstein causality, thus, Einstein causality would be falsified if this principle would not be given up. A common cause explanation $C\to A, C\to B$ is what the violation of the Bell inequality excludes.

I think I am not asking my question clearly. I just want to know the definitions of the two different causalities that you use. Einstein causality and signal causality. What has to be concluded by Bell's theorem is a separate question.

 

[DarMM]

That is still unclear to me. How could it be a degree of freedom! Anyway, since I don't know it I will stay out of it.

Well how can anything describe a physical degree of freedom? I assume you understand the sense in which the electric field $E$ describes a physical degree of freedom?

Or do you mean it's more natural to take it as describing a probability distribution rather than a physical degree of freedom.

 

[martinbn]

Well how can anything describe a physical degree of freedom? I assume you understand the sense in which the electric field $E$ describes a physical degree of freedom?

Or do you mean it's more natural to take it as describing a probability distribution rather than a physical degree of freedom.

I am guessing I don't know what degree of freedom is. It is intuitive in mechanics. A point in the plane has two degrees of freedom. In electrodynamics the electric field has infinitely many degrees of freedom. But wouldn't have thought that the field itself is called a degree of freedom.