Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

A Misunderstanding Eberhard's paper on FTL communication?

  1. Mar 28, 2016 #1
    As far as I can tell Eberhard and Ross didn't "prove" quantum entanglement can't implement FTL communication in "QUANTUM FIELD THEORY CANNOT PROVIDE FASTER-THAN-LIGHT COMMUNICATION", 1988 (http://link.springer.com/article/10.1007/BF00696109). But it's often cited as such a proof, in threads on PF and elsewhere.

    E & R describe a standard gedanken with sender and receiver (today we would say "Alice and Bob") outside each other's light cones, performing measurements on a pair of entangled particles.

    Section 2.2.2 says:

    "According to Hypothesis 2, the field operators phi-sub-j(xs,ts) at points (xs,ts) in [sender's light cone] commute (or anticommute for Fermion fields) with field operators phi-sub-j(xr,tr) at points (x,t) in [receiver's light cone], when ts < tr. This follows from the fact that two such points are outside each other's light cone. ... It follows that the measurement operators M and Ms in the Heisenberg representation commute."

    Here M and Ms refer to the two measurements of sender and receiver, which may be of observables which would be non-commuting if not space-like separated; such as Z and X axes' spin, or position and momentum. And hypothesis 2 is simply the statement, from standard QFT, that measurements made outside each other's light cones commute (or for fermions anti-commute). That's true - IF you assume FTL communication is impossible (which QFT, being relativistic, does). Thus they appear to assume what they're trying to prove!

    They go on to show that, given that M and Ms commute, probabilities of Bob's eigenvalues can't be affected by Alice's measurement. Then in Section 3 they show the same is true of the joint probability distribution of M and Ms (here called M1 and M2). It's inevitable given their assumptions, but as I say they appear to commit the fallacy of petitio principii.

    But finally they conclude, in 4.3.2:

    "... the possibility of faster-than-light communication is not unthinkable. [But] It is in contradiction with quantum field theory, which is the only known relativistic quantum theory. [Therefore] Justification for any effect providing faster-than-light communication should not be looked for in theories that abide with orthodox quantum field theory but in theories that allow some deviations from it."

    So the point of the paper seems to be: if FTL communication can, in fact, be achieved via entanglement, any relativistic version of QM - namely, QFT - must be wrong. In fact that's what the title of the paper indicates. But this is no proof against FTL communication - as claimed by many writers, including posts below. E & R merely say that if you assume (following SR) that FTL communication is impossible, then FTL communication is impossible.

    Since experts often cite the paper as "proof" against FTL communication via entanglement, there must be something wrong with my reasoning. What is it?
     
  2. jcsd
  3. Mar 28, 2016 #2

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    2016 Award

    afaict: They are only saying that the mathematics that gives rise to the entanglement also forbids FTL... ie FTL is inconsistent with the model used for entanglement and EVERYTHING ELSE.
    Physics is, ultimately, empirical. That is the best we can do.
    But I look forward to what the others say.
     
  4. Mar 28, 2016 #3
    Thanks Simon,

    Since others don't disagree no doubt your view is accepted by them (and, of course, me).

    - Yes, and it's worth emphasizing
     
  5. Mar 29, 2016 #4

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    2016 Award

    I suppose something can be inferred from the number of views without comment, but you cannot take lack of contradiction to be evidence of assent ;)
    Give it a bit of time - the experts are probably just checking their resources.

    [edit: since posting, the reply post #2 got a like ...]
     
    Last edited: Mar 29, 2016
  6. Mar 29, 2016 #5

    Strilanc

    User Avatar
    Science Advisor

    You're confusing possible FTL signals inside the theory with engineering the observable effects of the theory into an FTL communicator. They prove the latter is impossible, but you're complaining about the former.

    Even in a classical simulation of quantum mechanics, where the simulation has a de-facto preferred frame and entanglement is implemented by instantaneous global changes, entities within the simulation can't use those effects to communicate faster than light. And the reason they can't use the possibly-FTL-behind-the-scenes effects to communicate comes down to the fact that space-like separated actions commute. If that's not the case, then it's not QFT being simulated.
     
  7. Mar 29, 2016 #6
    Strilanc, I don't think FTL communication can be done with entanglement. If you analyze delayed-choice experiments, or other specific experiments, it's clear they can't do it. However Eberhard's general "proof" of impossibility seems flawed.
    - You're right: this assumption, built into QFT, simulations, and indeed much of modern physics, is the basis of Eberhard's proof. Now, remembering that "Physics is, ultimately, empirical", we ask, "how do we know "space-like separated actions commute"? The answer: GR and SR say no signal can go FTL. Why? Because all the signals they could imagine back in 1900-1920 were physical entities: mainly EM radiation, but also material objects and gravitational waves. But here we have a "new" mechanism which was not considered when deciding that "space-like separated actions commute". Thus the principle is not established until we decide by experiment that entanglement also doesn't violate it.

    Even then of course it's only right provisionally, like all physical laws, theories and principles. There may always be some new phenomenon discovered tomorrow that invalidates them.

    It's been said that entanglement comes from QFT, so how can it violate it? No, entanglement - the fact, not the representation within a theory - comes from Alain Aspect's (and other's) experiments.

    I'm aware that some dispute its significance entirely: for instance "Coherent Histories", others; but I'm ignoring them.

    Clearly if it did work, FTL communication (via any mechanism at all) must violate the principle. If Alice can send a message "Hi Bob", she must do it first, he must read it second. If he read it first, he would read garbage. In this case space-like separated actions DO NOT commute.

    In conclusion you seem to be agreeing with my main point. When we assume "space-like separated actions commute" we can easily prove FTL communication impossible - using any mechanism, not just entanglement. But the principle holds only for mechanisms known 100 years ago. It doesn't apply to new phenomena, until it's proven by new experiments and/or other theories which don't assume what they're trying to prove.

    Of course I could be wrong.
     
    Last edited: Mar 29, 2016
  8. Mar 29, 2016 #7

    Strilanc

    User Avatar
    Science Advisor

    Actually, raw commutativity is not quite strong enough to prove lack-of-FTL.

    Suppose Alice and Bob have magic message-swapping boxes. When you put a message in box A, it pops out a message that was/will-be put into box B. When you put a message in box B, it pops out a message that was/will-be put into box A. It doesn't matter if Alice hasn't yet put her message into box B, Bob can pull it out when putting his message in. When multiple messages are involved, they queue.

    (If you try to create a paradox, you get a "TRANSMISSION FAILURE. DO NOT MESS WITH TIME!" error message. The boxes are extremely good at thwarting counterfactual attempts to derive paradoxical information from whether a failure message is triggered or not.)

    If Alice and Bob are space-like separated and have no mechanism to coordinate besides these two boxes, their actions commute. Yet you can still use the boxes for FTL communication.

    Now you're confusing the question of whether reality allows for FTL signalling with whether quantum field theory allows for FTL signalling. Quantum field theory maps quantum mechanics onto fields in space in a particular way. Given the details of that mapping, you can prove that actions that are space-like separated will commute.

    If we find out that QFT is wrong, that it doesn't describe reality, the proof that QFT doesn't allow FTL will still stand. We'll simply have realized that the proof doesn't apply to reality, because QFT didn't apply to reality. Maybe the next theory will allow for FTL communication, maybe it won't. That has no bearing on the validity or soundness of the proof we're talking about.
     
  9. Mar 29, 2016 #8
    Actually, raw correlations are not quite strong enough to prove anything. When you read loophole free , call Columbo, Vera or Barnaby and let them investigate. You would be surprised ...
     
  10. Mar 29, 2016 #9

    Simon Bridge

    User Avatar
    Science Advisor
    Homework Helper
    Gold Member
    2016 Award

    "Ohh, thers no ef teh ell pet. Sorre abut thaat."
    -- DCI Vera Stanhope

    We cannot prove there is no ftl in Nature... that is not how empiricism works.
    We can prove a particular model excludes it... in this case that no modification is needed to qft to preserve causality in the face of apparent ftl communication via entanglement.
    I say "apparent" bc communication only appears if you treat the particles classically... the communication is an emergent effect of the statistics. I think that's the hardest part to grasp. It's just not a classical system, but our intuition preferrs classical.
     
  11. Mar 29, 2016 #10
    - I could happily concede this point, but in fact I did cover it:
    - "others" include Transactional Interpretation, Two-State Vector Formalism, Double Inferential State-Vector Formalism, time-symmetric Wheeler–Feynman absorber theory, and any and all others (and appurtenances thereof) which feature retrocausality. Such interpretations don't change my basic point, although it would need to be stated differently.
    - No doubt I'm a bit confused, and wrong about some point or other; but ...

    QFT does NOT allow for FTL signal; reality might. Another key difference: reality is, well, real; whatever it tells us (via the language of experiment) is fact. Whereas QFT is just a theory, to be used within its domain of applicability and discarded / superseded whenever contradicted by the afore-mentioned reality. Where's the confusion?
    - Long ago we called it "second quantization", an awkward term. The wave function is treated as an operator, or better is to say each point (uncountably infinite in number) is an operator, the whole bunch of them a field. A field is derived from an appropriate equation: Dirac for fermions (1/2 spin), Maxwell for photons, Proca for massive spin-1 bosons (weak - W's and Z), for Higgs - an isospinor scalar field (based, I suppose, on Klein-Gordon), for gravitons - a spin 2 massless field (not sure the name of some of these), and more can be thrown in for composite particles, phonons, and so on.

    Notice I didn't include Schroedinger's because when going to field formulation we start using only Lorentz-invariant, relativistic equations. So this is where speed-of-light limit comes in; all these equations are based (one way or another) on D'alembertian and the Minkowski metric. With the loss of absolute time, Hamiltonian is used rarely; we rely on Lagrangian with action principle (Euler-Lagrange) and Noether symmetries shortcuts.

    We also reject global gauge theories for local, and old-fashioned action-at-a-distance ideas (like Ampere's EM theory and, as far as I can tell, Fermi's original weak interaction theory) for local interactions mediated by gauge bosons (virtual and on-shell). Lie algebras are used to implement what we used to call "contact transformations".

    These, I believe, would be the details from which the commutativity of space-like separated operators is proven, within QFT.

    Thanks very much, Strilanc, for the opportunity to clarify some of these points.
     
    Last edited: Mar 29, 2016
  12. Mar 29, 2016 #11
    - That's right. Physical theories, induced from strong correlations seen in reality, are actually "proven" only provisionally.
     
  13. Mar 29, 2016 #12
    just participate to an experiment or analyze carefully datas
     
  14. Mar 29, 2016 #13
    If you mean: just experiment, don't theorize at all - that's taking it too far. Experimental data is the gold standard, but without theories to, essentially, summarize results in clever ways, we'd be swamped by details and unable to calculate our way out of a paper bag. Also theories point the way to the next vital experiments. You need both - in spite of the huge risk of theories being mistaken for reality. In fact a really addicted theorist simply stops caring about reality entirely, becomes a mathematician disguised as a physicist.
     
  15. Mar 29, 2016 #14
    We cannot prove too that there were not green bears on the first galaxies. The kind of FTL we are talking about comes from the non-locality theory that comes from extrem interpretation of the Bohr theory. As a new concept, the non locality theory must exhibit a proof of its physicality. No ? Just give me a reference and I'll say why it's not a proof, I know them all
     
    Last edited: Mar 29, 2016
  16. Mar 29, 2016 #15
    I say just that non-locality is not a little bit proven theory. Datas are full of the fair sampling loophole with an incredible margin when it's not worse ; don't read abstracts only , go deeply in the publications and ask for the datas.
     
  17. Mar 30, 2016 #16
    I've assumed that from Alain Aspect to Hensen et al., Giustina et al., Shalm et al. (2015), experiments more or less proved non-locality - i.e. violation of Bell's inequality. The later experiments claim to cover loopholes.

    But you say the "unfair-sampling" (or, "detection") loophole has not been addressed sufficiently. Therefore I'm mistaken to assume non-locality (a.k.a. retrocausality, or whatever ontological flavor one prefers) is real.

    Unfair-sampling refers (skipping some details) to the fact that when a pair of entangled particles aren't both detected, that datum is discarded. Contrary to expectations the data that are left may not be a random sample.

    The point of this thread - that Eberhard and Ross are guilty of circular reasoning - is, right or wrong, unaffected by the truth of non-locality. But the point becomes much less important, if you're right.

    Now, the physics community thinks non-locality has been demonstrated; that any loopholes left are more or less negligible. That's been the attitude for decades and the latest very careful experiments only reinforce it. Sources like Wikipedia agree.

    By the way I notice you're not a native English speaker, so if you have difficulty understanding me, don't hesitate to ask for clarification. Don't take offense: my French is even worse than your English, so on average you're superior.

    The question is, why are you the only one who knows the unfair-sampling loophole invalidates non-locality conclusion, with "an incredible margin"? You sound rather sure it's no accident - that in fact non-locality is false, and this loophole will never be plugged. Why?

    Unfortunately if you can't back up your claim with citation from respectable journals, by PF rules you're "speculative" and may be censored at any time. So try to support your claim somehow.

    I'm not going to go dig through that data in any big hurry. Life is short and science is long. But I'll keep it in mind and someday, perhaps, look into it.

    However it works out, thanks for the tip.
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted



Similar Discussions: Misunderstanding Eberhard's paper on FTL communication?
  1. FTL communication (Replies: 11)

Loading...