Realism in the entanglement swap experiment

[vanhees71]

As per usual, your concept of quotes relates to entire books. That’s not a quote! There are many things presented in an entire book or paper. Be specific.

Of course, if you are unwilling to read large parts of a textbook to understand the details of a theory, I can't help you.

I am not self contradictory; you just can’t see any options other than an entrenched position.

Specifically: sure I agree that QFT requires signal locality. And sure I agree that nonlocal action requires a classical signal to perceive. But there is nonlocal action as all of the references say. No paper says anything different.

QFT requires and ensures signal locality, and that's all what locality means. There is no non-local action. At least there's no experiment demonstrating this, and I don't know a single reference that claims this. To the contrary locality in the sense of relativistic QFT is envoked as an argument in Bell tests that there are NO causal connections between the measurements at far distant places, i.e., a large effort is taken to ensure that experiments like teleportation or entanglement swapping etc. are constructed such that the measurement events or even the choice of the measured observables at the far distant places are space-like separated, and then it's argued that there cannot be causal influences among these manipulations at far distant places.

Quantum contexts/events/actions are generally probabilistic as to outcomes. That’s true whether the contexts are local or nonlocal. If you use the word “causal” to describe an experimenter’s choice which must have a determined specific outcome (such as a specific bell state): then yes, QFT would be locally causal - precisely because experiments do not allow the experimenter the opportunity to make that choice.

An experimenter's choice does not need to have deterministic outcome. That's the whole point: The freedom is in the choice of the observable to be measured. Depending on the state these observable may not take a predetermined value before the measurement and that's why the outcome in general is random, and given such a preparation the experimenter cannot choose, what the outcome of the measurement of this observable will take. All this has indeed nothing to do with locality or non-locality. Also non-relativistic QT, which does not fulfill the relativistic locality constraints, is a consistent theory, but it's not describing Nature in all situations while relativistic local QFT does.

But neither I nor most authors require such a strict view when discussing these experiments. Quantum theory predicts an experimenter can choose to execute a swap here and objectively change the state of a pair that is distant. That change is to a state that is randomly occurring, itself outside the control of the experimenter. Further, the experimenter can make his choice before or after 1,4 pair detection. In my book, that’s a violation of strict Einsteinian causality. But of course… not a violation of most tenets of special relativity. That being quantum predictions of relativistic QM.

Yes, but the selection of the subensemble is due to local measurements, which do not causally influence the other two photons which are far distant. The entanglement swap is possible due to the entanglement of the pairs 12 and 34. The entanglement of the subensemble generated by projecting the photons (23) to a Bell state is due to the preparation of the full ensemble in the state, where photons (12) and (34) were both maximally entangled. That is the predition of relativistic local QFT. It's known since 1928 that relativistic QM is not a consistent theory exactly for the reason that it does not obey the causality constraints a relativsitic theory has to obey! That's why Dirac was forced to his hole-theoretical reinterpretation of his equation and thus the introduction of anti-particles. This hole theory is conceptually problematic but nevertheless equivalent to QED, and QED is conceptually much less problematic, because from the outset it assumes the possibility of annihilation and creation processes and realizes the causality constraint via microcausality of local observable-operators.

 

[Morbert]

You have it backwards though! If there is no Bell state, there is no swap. Phase information is perfectly available but not meaningful when a swap does not occur. That because the 1,2 pair does not develop a relationship with the 3,4 pair.

But for those of you out there that don’t believe anything physical happens at the 2,3 measurement that creates the 1,4 entanglement: what difference would that make? As I indicated, the detectors indicate Reflect/Transmit at the projecting splitter and the polarization. What you call phase is simply part and parcel of those same properties.

Seen another way: tell me what happens to 1,4 as a result of a successful swap where the 2,3 photons come close to each other - as opposed to when a delay is introduced and no swap occurs. The coincidence time window has most 2,3 photons detected within 3 ns and usually 1 ns apart. 1 ns is about 1 foot. So they do not precisely overlap. Further, orthogonal photons cannot interact (and that happens in 1/2 the cases). So explain what is different when a delay is inserted to change any readings when our experimenter flips his swap/no swap switch.

I think the choice of vocabulary makes things hard to follow, particularly the parts in bold. For example, when you say "if there is no Bell state, there is no swap" do you mean, "when the instrument that measures the middle photons is set such that it projects onto a mixed state rather than a pure bell state, the corresponding subensembles of experimental runs will not exhibit Bell-inequality-violating inequalities between measurements on the 1,4 pair? If yes then I completely agree. So e.g.

Seen another way: tell me what happens to 1,4 as a result of a successful swap where the 2,3 photons come close to each other - as opposed to when a delay is introduced and no swap occurs.

If our instruments are set such that a measurement on the middle photons is characterised by a projection onto a Bell state, then, after all experimental runs are complete, we can divide our experimental runs into subensembles in accordance with the Bell-state measurement outcomes, and at least two of these subensembles will show Bell-inequality-violating correlations between measurements on 1 and 4.

If instead our instruments are set such that a measurement of the middle photons fails to project onto a Bell-state, we can no longer use the protocol that divides our experimental runs into subensembles that show Bell-inequality-violating correlations between 1 and 4. "No projection onto a Bell state, no protocol for selecting subensembles with bell-inequality-violating correlations."

 

[mattt]

I think the choice of vocabulary makes things hard to follow, particularly the parts in bold. For example, when you say "if there is no Bell state, there is no swap" do you mean, "when the instrument that measures the middle photons is set such that it projects onto a mixed state rather than a pure bell state, the corresponding subensembles of experimental runs will not exhibit Bell-inequality-violating inequalities between measurements on the 1,4 pair? If yes then I completely agree. So e.g. If our instruments are set such that a measurement on the middle photons is characterised by a projection onto a Bell state, then, after all experimental runs are complete, we can divide our experimental runs into subensembles in accordance with the Bell-state measurement outcomes, and at least two of these subensembles will show Bell-inequality-violating correlations between measurements on 1 and 4.

If instead our instruments are set such that a measurement of the middle photons fails to project onto a Bell-state, we can no longer use the protocol that divides our experimental runs into subensembles that show Bell-inequality-violating correlations between 1 and 4. "No projection onto a Bell state, no protocol for selecting subensembles with bell-inequality-violating correlations."

I was writing exactly the same, but I think that you worded it better.

 

[vanhees71]

I think the choice of vocabulary makes things hard to follow, particularly the parts in bold. For example, when you say "if there is no Bell state, there is no swap" do you mean, "when the instrument that measures the middle photons is set such that it projects onto a mixed state rather than a pure bell state, the corresponding subensembles of experimental runs will not exhibit Bell-inequality-violating inequalities between measurements on the 1,4 pair? If yes then I completely agree. So e.g. If our instruments are set such that a measurement on the middle photons is characterised by a projection onto a Bell state, then, after all experimental runs are complete, we can divide our experimental runs into subensembles in accordance with the Bell-state measurement outcomes, and at least two of these subensembles will show Bell-inequality-violating correlations between measurements on 1 and 4.

Exactly, and that's also a very important point. If you do another measurement with 2&3 to select a subensemble, you simply get a different subensemble, which doesn't tell you anything about what might have happened, if you'd do the projection to a Bell state. There is no "counterfactual definiteness", as the quantum-foundationalists call it. To say it simpler with Mermin: QT doesn't tell anything about unperformed experiments. That's already the trap, EPR get caught in.

If instead our instruments are set such that a measurement of the middle photons fails to project onto a Bell-state, we can no longer use the protocol that divides our experimental runs into subensembles that show Bell-inequality-violating correlations between 1 and 4. "No projection onto a Bell state, no protocol for selecting subensembles with bell-inequality-violating correlations."

If the measurement on the pair 2&3 fails for some reason, we can't select it to form the subensembles you want to select in this experiment. Then you simply can't use this specific realization of the procedure at all. That's one of the loop holes, i.e., that you always have a finite probability that there's no definite outcome. E.g., real-world photon detectors have a finite probability to simply fail to detect a photon going through them. This may open the possibility that there's some specific "hidden variable" determining systematically this detector failure and only the photons that get detected behave as predicted by QT. This loophole is pretty well closed by constructing ever better photon detectors.

 

[DrChinese]

If our instruments are set such that a measurement on the middle photons is characterised by a projection onto a Bell state, then, after all experimental runs are complete, we can divide our experimental runs into subensembles in accordance with the Bell-state measurement outcomes, and at least two of these subensembles will show Bell-inequality-violating correlations between measurements on 1 and 4.

If instead our instruments are set such that a measurement of the middle photons fails to project onto a Bell-state, we can no longer use the protocol that divides our experimental runs into subensembles that show Bell-inequality-violating correlations between 1 and 4. "No projection onto a Bell state, no protocol for selecting subensembles with bell-inequality-violating correlations."

So what you’re saying is: the success or failure of the BSM doesn’t change anything for the remote (1,4) pairs (while I say it creates the entanglement); but a swap the experimenter chooses to fail messes up the BSM protocol results. By what logic is the something physical occurring between 2,3 pairs that are orthogonal? Because that is detected regardless of whether the swap succeeds! Are you seriously asserting that adding a delay can change the outcome of the polarization portion of the BSM protocol?

And keep this in mind: for the swap to succeed, the requirement is that 2,3 pairs are allowed to interact and be indistinguishable. The rest of the BSM protocol is executed to identify the resulting Bell state.

Once again I ask: Why would there be a difference in protocol outcomes depending on whether or not the orthogonal 2,3 pairs are near each other in the projecting beam splitter or are farther from each other? What kind of interaction can these two photons have (or not have) that would make any difference at all?

I say that within the full quantum context – which is clearly non-local, and contains portions that are never coexist in common light cones with other portions of the context – the one action that “causes” the entanglement swap is the indistinguishability. The Bell State measurement itself confirms the indistinguishability, as well as identifying such bell states as are possible (since all four bell states cannot be identified with current technology). Since the experimenter in our version has the ability to switch the indistinguishability on or off: that person, then becomes the causal agent of the swap itself. And by causal, I mean: quantum causality. Quantum causality does not respect causal order, and does not respect c. Therefore, it violates strict Einsteinian locality.

 

[PeterDonis]

Thread closed for moderation.

 

[PeterDonis]

At this point all views have been expressed in sufficient detail. Per the guidelines for this subforum, that is the best that can be expected, since interpretation disputes cannot be resolved. Therefore, this thread will remain closed. Thanks to all who participated.