How exactly does spooky action at a distance violate SRT?

[greypilgrim]

Hi.

It's often stated (for example in Wikipedia) that spooky action at a distance would violate SRT. But how, exactly?
As far as I understand, SRT does neither assume nor predict that the speed of light is an upper limit (it even allows for hypothetical tachyons, but that's not the point).

I know that FTL communication would cause causality problems. But it's demonstrable that entanglement alone cannot be used for communication.

I'm not trying to defend spooky action at a distance, but it seems to me that if you want to argue against it using SRT, you'd need further assumptions or interpretations that minimal SRT doesn't entail.

 

[Vanadium 50]

I don't think you have even defined "spooky at a distance". It is certainly the case that QM does not violate SR.

 

[PeterDonis]

It's often stated (for example in Wikipedia) that spooky action at a distance would violate SRT.

Please give a specific reference. We can't answer a vague question about a vague set of things somebody said somewhere. We need a specific statement in a specific reference to have a valid basis for discussion.

 

[vanhees71]

That's easily answered. The only consistent relativistic QT is local (sic!) relativistic QFT, which by construction excludes any "spooky actions at a distance". Case closed: There's no violation of relativistic causality, because there are no faster-than-light-acting interactions. That's why you have to use a QFT description of relativistic QT to begin with.

The very fundamental consequences are empirically very well estabilished:

(a) for each "particle" there's an "anti-particle" (except for strictly neutral particles, where particle and anti-particle are the same, but also then particle number is not conserved, e.g., for photons).

(b) spin-statistics relation: particles with integer (half-integer) spin must necessarily by bosons (fermions).

(c) CPT invariance: any process that can happen in nature according to local relativistic QFT implies that there must be another process also possible in nature, which exchanges any particle with its anti-particle, is viewed in a mirror and is time reversed. Note that neither of the other combinations of these "discrete symmetries" (P, T, C, CP, CT, PT) is a symmetry of Nature, i.e., it's violated by the weak interaction.

 

[Morbert]

Say you measure the spin of one particle entangled with another in the Bell state uu + dd and you observe u.

If the following claim is true, then we can infer FTL causality: "A measurement on the spin of the other particle is guaranteed to yield a result u. But, all else being equal, if I had not carried out my measurement, then a measurement on the spin of the other particle would not be guaranteed to yield a result u".

This claim is not true under some interpretations.

 

[vanhees71]

If Alice measures here particle to show up, nothing changes for Bob. He'll still doesn't know that he'll necessarily must also get up. This he can know only, if Alice tells him about the outcome of her measurement. That she can do at best sending a signal with the speed of light. So there's no violation of relativistic causality.

 

[Morbert]

If Alice measures here particle to show up, nothing changes for Bob. He'll still doesn't know that he'll necessarily must also get up. This he can know only, if Alice tells him about the outcome of her measurement. That she can do at best sending a signal with the speed of light. So there's no violation of relativistic causality.

This is true, but people will make a distinction between an epistemic statement "Bob knows that his outcome will be up" and an ontic statement "Bob's outcome will be up". They will say that even though there has been no epistemic change for Bob, there has been an ontic change for Bob's particle.

I think we can deny both an epistemic and and ontic change.

 

[greypilgrim]

So there's no violation of relativistic causality.

That's my understanding. Even in a hidden variable interpretation that allows for superluminal signals, there is no information conveyed that could potentially cause causality problems.

So, why do I still see statements like (Wikipedia page on Quantum entanglement):
Einstein and others considered such behavior impossible, as it violated the local realism view of causality (Einstein referring to it as "spooky action at a distance")[5] and argued that the accepted formulation of quantum mechanics must therefore be incomplete.

Like Einstein, Schrödinger was dissatisfied with the concept of entanglement, because it seemed to violate the speed limit on the transmission of information implicit in the theory of relativity.

Or, even more dramatically, translated from the German version:
According to this, two photons, which were entangled with respect to polarization, would have to communicate with at least 10,000 times the speed of light, if they were to send the result of the measurement of polarization on one photon to the other. Such a communication would blatantly contradict the theory of relativity and would mean, among other things, that time loops are possible.

Is this just bad Wikipedia?

 

[vanhees71]

This is very common. It can only come from people, who never thought about the foundations of relativistic QFTs, where locality is implemented from scratch by the microcausality constraints on local observables. Another culprit is the idea of a collapse of the quantum state, which is part of some of the Copenhagen Interpretation flavors, assuming that through the local interaction of a particle with the detector (performing in this way a position measurement) the "wave function" must "instantaneously collapse" to one very much concentrated around the measured position. This is based on Schrödinger's original idea that an electron "is" the wave function (in the same way as light "is" the electromagnetic wave). The wave function rather is, according to Born's ingenious insight, just a description of the probability for the electron being to be found at a position, $P(\vec{x})=|\psi(\vec{x})|^2$.

As long as you strictly stick with the physical probabilistic core of the "interpretation", i.e., that the quantum state provides probabilities for outcomes of measurements, there's no quibble. In the example, if Alice measures her particle having spin up then she knows, given the state of the particles prepared before (!!!) any measurement is made and thus implying the correlations between the particles' spins due to the fact that this state is an entangled state, that Bob must also find spin up (when measuring the spin component in the same direction, of course). Bob himself doesn't know this and thus just randomly gets also spin up, and Alice doesn't know that this is really true. For Bob to know that he must get spin up due to Alices finding, he'd need the information provided by Alice and for Alice to confirm what Bob found, she needs the corresponding information about his outcome. Both can only be provided by some information exchange which, within relativistic physics (and particularly also within relativistic QFT) is only possible with signals propagator with a speed at most that of light (in vacuo).

 

[Demystifier]

I'm not trying to defend spooky action at a distance, but it seems to me that if you want to argue against it using SRT, you'd need further assumptions or interpretations that minimal SRT doesn't entail.

That is true. To get spooky action at a distance, you have to make a reality assumption, namely, to assume that it makes sense to talk about values of some variables before they are measured. Minimal quantum theory, relativistic or not, does not make such a reality assumption.

 

[Morbert]

That is true. To get spooky action at a distance, you have to make a reality assumption, namely, to assume that it makes sense to talk about values of some variables before they are measured. Minimal quantum theory, realistic or not, does not make such a reality assumption.

<insert obligatory mention of consistent histories>

 

[PeterDonis]

<insert obligatory mention of consistent histories>

Consistent histories is an interpretation of QM, and discussion of interpretations belongs in the interpretations subforum, not here. The subject of this thread is minimal QM without any interpretation.

 

[Demystifier]

<insert obligatory mention of consistent histories>

I should have written: ... before they are measured and before a framework is chosen.

 

[Morbert]

Consistent histories is an interpretation of QM, and discussion of interpretations belongs in the interpretations subforum, not here. The subject of this thread is minimal QM without any interpretation.

OP's confusion cannot be addressed without recourse to interpretation.

 

[PeterDonis]

OP's confusion cannot be addressed without recourse to interpretation.

Yes, it can; several posts in this thread have already done so, by correctly pointing out that in minimal QM with no interpretation, there is no issue.

It is true that if you go beyond minimal QM with no interpretation, different interpretations can take different positions on the issue. Discussion of that belongs in the interpretations subforum, as I said before.

 

[PeterDonis]

why do I still see statements like

These statements depend on adopting a particular interpretation of QM, and, as noted in post #15 just now, different interpretations can take different positions on the issue, and discussion of that belongs in the interpretations subforum.

 

[Morbert]

Yes, it can

No, it can't. The OP, in the first post, explicitly asked about the relation between spooky action at a distance and how it violates relativity. Interpretation-less QM does not speak to spooky action at a distance.

 

[PeterDonis]

Interpretation-less QM does not speak to spooky action at a distance.

Sure it does: it defines it as "entanglement", or more precisely as "correlations between spacelike separated measurement results on entangled systems". The OP did not adopt any specific definition, and that one works fine for minimal QM.

 

[PeterDonis]

Thread closed for moderation.