Nature Physics on quantum foundations

[Demystifier]

About 2), why is that controversial? I thought that is just QM and interpretation independent?

It is standard that entanglement is caused by local interactions, but it's not standard that resulting correlations are caused by local interactions. Entanglement is a property of the abstract state in the HiIlbert space, while correlation is a property of the concrete measurement outcomes. Creation of entanglement is clearly deterministic, while creation of outcomes is seemingly random. Some interpretations say that correlations are caused by non-local action at a distance, some that they are not caused at all, etc.

 

[vanhees71]

But entanglement implies the corresponding correlations. So if entanglement is caused by local interactions, then also the correlations, described by it, are caused by local interactions.

If one accepts standard micocausal relativistic QFT, there cannot be non-local action at a distance.

 

[martinbn]

It is standard that entanglement is caused by local interactions, but it's not standard that resulting correlations are caused by local interactions. Entanglement is a property of the abstract state in the HiIlbert space, while correlation is a property of the concrete measurement outcomes. Creation of entanglement is clearly deterministic, while creation of outcomes is seemingly random. Some interpretations say that correlations are caused by non-local action at a distance, some that they are not caused at all, etc.

How can there be a non-local action at a distance and no faster than light signaling at the same time?

 

[martinbn]

How can 3') contradict 3), when we are obviously preparing single systems to form the ensemble?

Because the state describes the ensemble not the single system.

 

[Demystifier]

For me the minimal interpretation is the standard interpretation,

Something called "standard" should be universal, not "for you".

How can 3') contradict 3), when we are obviously preparing single systems to form the ensemble?

If you don't accept Ballentine's argument that 3') contradict 3), there is no hope that you would accept mine.

2) follows from local/microcausal QFT.

Except that there is no reference claiming it, making the claim 2) nonstandard.

 

[vanhees71]

Then you imply that one member of the ensemble influences any other, but that can be excluded in modern experiments (e.g., only single diphotons in Bell-test experiments, building an ensemble of really independent single-system realizations).

 

[martinbn]

Then you imply that one member of the ensemble influences any other, ...

How is that implied?

 

[vanhees71]

Because then you say that the quantum state cannot be determined by preparing a single system, i.e., that the ensemble is defined by the probabilistic properties of a single-system preparation procedure. For me this contradicts the findings of many Bell-test experiments, where it is ensured that always single systems are measured.

 

[Demystifier]

How can there be a non-local action at a distance and no faster than light signaling at the same time?

Signal is transmitted by controlled action, while action in general does not need to be controlled. For example, non-local action at a distance may be transmitted by some hidden variables, and variables which are hidden obviously cannot be controlled. Another example is random collapse, which involves non-local action at a distance but cannot be controlled because it's random.

 

[martinbn]

Because then you say that the quantum state cannot be determined by preparing a single system, i.e., that the ensemble is defined by the probabilistic properties of a single-system preparation procedure. For me this contradicts the findings of many Bell-test experiments, where it is ensured that always single systems are measured.

It can but only after you know what state the procedure produces. How do you know that by just one preperation?

 

[Demystifier]

For me this contradicts the findings of many Bell-test experiments

"For me" is the key. It's your view, not standard view.

 

[martinbn]

Signal is transmitted by controlled action, while action in general does not need to be controlled. For example, non-local action at a distance may be transmitted by some hidden variables, and variables which are hidden obviously cannot be controlled. Another example is random collapse, which involves non-local action at a distance but cannot be controlled because it's random.

Ok, suppose we have the standard Alice and Bob scenario. We do it 1000 times, Alice makes a measurement on either the first 500 or on the last 500, the other 500 she does nothing. Bob's task is by measuring his particle to determine when Alice measured, in the first half of the trials or in the second. Can he do it?

 

[vanhees71]

It can but only after you know what state the procedure produces. How do you know that by just one preperation?

Of course, I can't know this, but I have to prepare ensembles and do statistics with them. To completely determine the state you need several different measurements on the ensembles. The ensembles are defined by prepratation procedures of single systems. That's why it is so important to clearly distinguish "states" and "measurements". A state is a preparation procedure (for single systems), allowing to form well-defined ensembles. The outcomes of measurements are random and the statistics of these outcomes is given by the corresponding statistical operator according to Born's rule.

 

[Demystifier]

Ok, suppose we have the standard Alice and Bob scenario. We do it 1000 times, Alice makes a measurement on either the first 500 or on the last 500, the other 500 she does nothing. Bob's task is by measuring his particle to determine when Alice measured, in the first half of the trials or in the second. Can he do it?

He can't do it. This demonstrates that Alice by performing measurement cannot send a signal to Bob. But it does not imply that measurement by Alice doesn't have any influence on Bob's results.

 

[vanhees71]

Ok, suppose we have the standard Alice and Bob scenario. We do it 1000 times, Alice makes a measurement on either the first 500 or on the last 500, the other 500 she does nothing. Bob's task is by measuring his particle to determine when Alice measured, in the first half of the trials or in the second. Can he do it?

No, all Bob will see is a stream of random results with statistics determined by the prepared state. The correlations due to entanglement can only be verified by exchanging information about A's and B's measurement results on each individual entangled system.

 

[vanhees71]

He can't do it. This demonstrates that Alice by performing measurement cannot send a signal to Bob. But it does not imply that measurement by Alice doesn't have any influence on Bob's results.

If the measurement events are space-like separated within relativistic microcausal QFT there cannot be any influence of one of the measurements on the other.

 

[Demystifier]

If the measurement events are space-like separated within relativistic microcausal QFT there cannot be any influence of one of the measurements on the other.

Yes, but this does not contradict my claim. Standard relativistic microcausal QFT is one possible model of reality consistent with existing experiments, but it's not the only possible model of reality consistent with existing experiments.

 

[martinbn]

He can't do it. This demonstrates that Alice by performing measurement cannot send a signal to Bob. But it does not imply that measurement by Alice doesn't have any influence on Bob's results.

What kind of influence can there be if you cannot find any!

(By the same logic you could claim that there is a little sprite that looks at Alice's results and changes Bob's accordingly.)

 

[martinbn]

Of course, I can't know this, but I have to prepare ensembles and do statistics with them. To completely determine the state you need several different measurements on the ensembles. The ensembles are defined by prepratation procedures of single systems. That's why it is so important to clearly distinguish "states" and "measurements". A state is a preparation procedure (for single systems), allowing to form well-defined ensembles. The outcomes of measurements are random and the statistics of these outcomes is given by the corresponding statistical operator according to Born's rule.

Yes, and that is why there is a difference between an ensemble and a single system representing the ensemble. The state describes the ensemble.

 

[Demystifier]

But entanglement implies the corresponding correlations.

No, entanglement + Born rule implies the corresponding correlations. The origin of Born rule is controversial, hence the origin of correlations is controversial.

 

[vanhees71]

Yes, but this does not contradict my claim. Standard relativistic microcausal QFT is one possible model of reality consistent with existing experiments, but it's not the only possible model of reality consistent with existing experiments.

I only argue within standard physics, not about speculative future theories. Of course, one can never exclude any possibility that such a new theory is needed in the future.

 

[vanhees71]

No, entanglement + Born rule implies the corresponding correlations. The origin of Born rule is controversial, hence the origin of correlations is controversial.

For me the Born rule is simply a postulate included in standard QT in the minimal interpretation. In this sense there's no "origin" of it at all (as there are no "origins" of any other of the formal postulates) other than the fact that it is empirically very successful.

 

[Demystifier]

What kind of influence can there be if you cannot find any!

What number can violate the Goldbach conjecture if you cannot find any!

 

[Demystifier]

For me the Born rule is simply a postulate included in standard QT in the minimal interpretation.

Which, of course, is controversial.

 

[gentzen]

How can there be a non-local action at a distance and no faster than light signaling at the same time?

I guess that the confusion arises more from the unclear meaning of "action at a distance" than from the word "non-local". If you imagine the randomness itself as being nonlocal, then no contradictions to "no faster than light signaling" arises, and you even get an argument why there must be randomness:

Nicolas Gisin’s short book Quantum Chance nicely explains how the paradox arises that quantum mechanics is local and nonlocal at the same time: The randomness itself is nonlocal, and it must be really random, because otherwise this non-locality could be used for instantaneous signal transmission.

Even so there may be "interpretational dances" that can avoid even "nonlocal randomness", the notion itself is intuitive and unproblematic.

 

[Demystifier]

I only argue within standard physics

Except when you don't, as my 3 examples illustrate.

 

[vanhees71]

Which, of course, is controversial.

Within some philosophers' community may be, but not within the physics community, right?

 

[martinbn]

What number can violate the Goldbach conjecture if you cannot find any!

If there is such a number it can be found at least in principle. The action of Alice measurement cannot be found even in principle.

 

[vanhees71]

I guess that the confusion arises more from the unclear meaning of "action at a distance" than from the word "non-local". If you imagine the randomness itself as being nonlocal, then no contradictions to "no faster than light signaling" arises, and you even get an argument why there must be randomness:Even so there may be "interpretational dances" that can avoid even "nonlocal randomness", the notion itself is intuitive and unproblematic.

This is simply resolve by saying that relativistic microcausal QFT is local but it allows correlations between far-distant parts of a quantum system (described by entanglement).

 

[Demystifier]

Within some philosophers' community may be, but not within the physics community, right?

Wrong. Many physicists, including you, do philosophy.

 

[vanhees71]

Except when you don't, as my 3 examples illustrate.

So what's "standard physics" with regard to QT? Copenhagen with collapse? de Broglie-Bohm? Many worlds?

 

[Demystifier]

The action of Alice measurement cannot be found even in principle.

No, it's not a matter of principle. It's only a FAPP phenomenon, very much like the 2nd and 3rd laws of thermodynamics.

 

[Demystifier]

So what's "standard physics" with regard to QT? Copenhagen with collapse?

Yes. You and me both dislike it, but it's standard.

 

[gentzen]

This is simply resolve by saying that relativistic microcausal QFT is local but it allows correlations between far-distant parts of a quantum system (described by entanglement).

I guess I understand what is meant by "relativistic microcausal QFT is local". I don't get what is meant by "it allows correlations between far-distant parts of a quantum system (described by entanglement)", and how it would give me a notion as intuitive and unproblematic as "nonlocal randomness". So I get a feeling like "maybe those are nice words, but what do you want to tell me with those words". Somehow it feels "too abstract" to me.

 

[martinbn]

No, it's not a matter of principle. It's only a FAPP phenomenon, very much like the 2nd and 3rd laws of thermodynamics.

Are you saying that in principle Bob could figure out whether Alice has done or not something with her particle?