--------------------------
Comment on: Testing the speed of ‘spooky action at a distance’.
(Johannes Kofler, Rupert Ursin, Časlav Brukner, Anton Zeilinger)
-
In a recent experiment, Salart et al. addressed the important issues of the speed of hypothetical communication and of reference frames in Bell-type experiments. The authors report that they "performed a Bell experiment using entangled photons" and conclude from their experimental results that "to maintain an explanation based on spooky action at a distance we would have to assume that the spooky action propagates at speeds even greater than the bounds obtained in our experiment", exceeding the speed of light by orders of magnitude. Here we show that, analyzing the experimental procedure, explanations with subluminal or even no communication at all exist for the experiment.
In order to explain the violation of Bell inequalities within the view where, to use the author‟s wording, "correlated events have some common causes in their shared history", one needs to assume hypothetical communication between the observer stations. This communication must be faster than light if the outcome at one station is space-like separated from all relevant events at the other station.
In the experiment pairs of time-bin entangled photons were sent over 17.5 km optical fibers to two receiving stations, located in Jussy and Satigny, both equipped with a Franson-type interferometer and detectors. The out-comes were observed space-like separated from each other. The phase in the interferometer, i.e. the setting, in Jussy was continuously scanned, while the setting at Satigny was kept stable.
However, if the setting at one side remains unchanged, the results at both observer stations can be described by a "common-cause" without having to invoke any communication at all, let alone superluminal spooky action at a distance. This is signified, e.g., by the fact that no formulation of a bipartite Bell type inequality exists which does not use at least two settings at each side. Therefore, contrary to the claim in the paper, no Bell test was performed.
Furthermore, had the experiment been repeated with a second stable setting at Satigny, a "common-cause" explanation would still be possible. This is because in order to exclude subluminal communication, it is crucial that the outcome event on each side is space-like separated from the setting choice on the other side – which was not done in Ref. [1]. Thus, such experimental data – even if they were taken with two measurement settings at Satigny and even granting the fair-sampling assumption – could be explained by a "common-cause" model. In other words, the experiment tests the superluminal speed of hypothetical influences between outcome events under the assumption of no, not even subluminal, hypothetical influences between setting choices and outcome events.
We also remark that in a Franson-type experiment like the one reported in Ref. [1] the considered Clauser-Horne-Shimony-Holt Bell inequality is not applicable even with perfect detectors because of the inherent postse-lection.2 One would (i) have to use a chained Bell inequality2, (ii) achieve fast switching with a rate depending on the geometry of the interferometer, and (iii) reach a better visibility than the one reported in Ref. [1]. None of these three issues is covered by the experiment.
We would like to stress that this comment should not be seen as a defence of local realism. And neither do we demand that Ref. [1] must present a loophole-free Bell test. However, it is the purpose of our comment to point out "common-cause" explanations of an experiment which aims at putting "stringent experimental bounds on the speed of all such hypothetical influences".
In a recent experiment, Salart et al. addressed the important issues of the speed of hypothetical communication and of reference frames in Bell-type experiments. The authors report that they "performed a Bell experiment using entangled photons" and conclude from their experimental results that "to maintain an explanation based on spooky action at a distance we would have to assume that the spooky action propagates at speeds even greater than the bounds obtained in our experiment", exceeding the speed of light by orders of magnitude. Here we show that, analyzing the experimental procedure, explanations with subluminal or even no communication at all exist for the experiment.
In order to explain the violation of Bell inequalities within the view where, to use the author‟s wording, "correlated events have some common causes in their shared history", one needs to assume hypothetical communication between the observer stations. This communication must be faster than light if the outcome at one station is space-like separated from all relevant events at the other station.
In the experiment pairs of time-bin entangled photons were sent over 17.5 km optical fibers to two receiving stations, located in Jussy and Satigny, both equipped with a Franson-type interferometer and detectors. The out-comes were observed space-like separated from each other. The phase in the interferometer, i.e. the setting, in Jussy was continuously scanned, while the setting at Satigny was kept stable.
However, if the setting at one side remains unchanged, the results at both observer stations can be described by a "common-cause" without having to invoke any communication at all, let alone superluminal spooky action at a distance. This is signified, e.g., by the fact that no formulation of a bipartite Bell type inequality exists which does not use at least two settings at each side. Therefore, contrary to the claim in the paper, no Bell test was performed.
Furthermore, had the experiment been repeated with a second stable setting at Satigny, a "common-cause" explanation would still be possible. This is because in order to exclude subluminal communication, it is crucial that the outcome event on each side is space-like separated from the setting choice on the other side – which was not done in Ref. [1]. Thus, such experimental data – even if they were taken with two measurement settings at Satigny and even granting the fair-sampling assumption – could be explained by a "common-cause" model. In other words, the experiment tests the superluminal speed of hypothetical influences between outcome events under the assumption of no, not even subluminal, hypothetical influences between setting choices and outcome events.
We also remark that in a Franson-type experiment like the one reported in Ref. [1] the considered Clauser-Horne-Shimony-Holt Bell inequality is not applicable even with perfect detectors because of the inherent postse-lection.2 One would (i) have to use a chained Bell inequality2, (ii) achieve fast switching with a rate depending on the geometry of the interferometer, and (iii) reach a better visibility than the one reported in Ref. [1]. None of these three issues is covered by the experiment.
We would like to stress that this comment should not be seen as a defence of local realism. And neither do we demand that Ref. [1] must present a loophole-free Bell test. However, it is the purpose of our comment to point out "common-cause" explanations of an experiment which aims at putting "stringent experimental bounds on the speed of all such hypothetical influences".
-
Reply to the: "Comment on: Testing the speed of `spooky action at a distance' "
(D. Salart, A. Baas, C. Branciard, N. Gisin, and H. Zbinden)
-
Quantum correlations cannot be described by local common causes. This prediction of quantum theory, surprising as it might appear, has been widely con rmed by numerous experiments. In our Nature Letter [1] we considered this point as established and addressed another issue: the alternative assumption that quantum correlations are due to supra-luminal influences of a first event onto a second event. For this purpose we believe that it suffices to observe 2-photon interferences with a visibility high enough to potentially violate Bell's inequality, as we reported (over 2 x 17.5 km). Simultaneously closing other loopholes, like the locality loophole as desired by Koer and colleagues, would certainly be an interesting addition, as would be any Bell tests that simultaneously address several of the loopholes.
Indeed, to rigorously exclude any common cause explanation of the observed quantum correlation one should, ideally, simultaneously close the locality and the detection loophole (and assume the existence of independent randomness and that quantum measurements are nished when detectors re or at least when a mesoscopic mass has sufficiently moved as insured in our experiment, see our recent article [2]). This is a formidable task and any progress towards achieving it is most welcome. So far, however, all experiments have addressed at most one of these loopholes; ours is no exception.
Concerning the comment on the use of a Franson interferometer for testing quantum nonlocality, we stress that this is not a fundamental issue. In principle it suffices to replace the entrance beam splitters of each interferometer by a fast switch. In this way the non-interfering lateral peaks observed in the 2-photon interferogram would disappear. However, in practice such switches suffer due to losses of around 3 dB. Hence, with today's technology it is much more
convenient to replace the ideal switch by a passive coupler, as we did in our experiment in a way very similar to [3].
Quantum correlations cannot be described by local common causes. This prediction of quantum theory, surprising as it might appear, has been widely con rmed by numerous experiments. In our Nature Letter [1] we considered this point as established and addressed another issue: the alternative assumption that quantum correlations are due to supra-luminal influences of a first event onto a second event. For this purpose we believe that it suffices to observe 2-photon interferences with a visibility high enough to potentially violate Bell's inequality, as we reported (over 2 x 17.5 km). Simultaneously closing other loopholes, like the locality loophole as desired by Koer and colleagues, would certainly be an interesting addition, as would be any Bell tests that simultaneously address several of the loopholes.
Indeed, to rigorously exclude any common cause explanation of the observed quantum correlation one should, ideally, simultaneously close the locality and the detection loophole (and assume the existence of independent randomness and that quantum measurements are nished when detectors re or at least when a mesoscopic mass has sufficiently moved as insured in our experiment, see our recent article [2]). This is a formidable task and any progress towards achieving it is most welcome. So far, however, all experiments have addressed at most one of these loopholes; ours is no exception.
Concerning the comment on the use of a Franson interferometer for testing quantum nonlocality, we stress that this is not a fundamental issue. In principle it suffices to replace the entrance beam splitters of each interferometer by a fast switch. In this way the non-interfering lateral peaks observed in the 2-photon interferogram would disappear. However, in practice such switches suffer due to losses of around 3 dB. Hence, with today's technology it is much more
convenient to replace the ideal switch by a passive coupler, as we did in our experiment in a way very similar to [3].
Post
Nessun commento:
Posta un commento