Reply to Comment on “Experimental Test of an Event-Based Corpuscular Model Modification as an Alternative to Quantum Mechanics”

Abstract

We have presented an experimental result, where we show that the predictions of an Event-Based Corpuscular Model (EBCM) from Ref. 2, a model that belongs to a class of non-ergodic models first described by Buonomano, are inconsistent with our measurements. We concluded that traditional quantum mechanics adequately describes our experiment, where pure single photons were used, and that our findings agree with earlier verification experiments by Summhammer, where a poissonian source of heavy particles was used. In their comment, Michielsen et al. disagree with the above conclusions. The central point of their argument is that EBCM predictions would be compatible with the observation, if the Mach–Zehnder interferometer (MZI) under test is modeled as a single system, and not as two independent, non-communicating beam splitters (BS). Unfortunately, the physics of interaction and a quantitative model of communication among the parts of MZI that would make it possible to reduce the model to effectively one memory register are not offered. Furthermore, no modification for a verification experiment that would guarantee the absence of an unspecified interaction is offered. An argument resorting to communication between the MZI parts is not a surprise. It is identical to invoking a locality loophole in Bell inequalities: when one considers communication among the different parts of the experimental apparatus, it is well known that hidden variable models can reproduce the results of Quantum Mechanics (QM). But by invoking an undescribed locality loophole, the EBCM and its verification experiment largely loses its appeal, because it becomes a sub-class of hidden variable theories falsifiable via a locality loophole free Bell test. In this sense, the class of local variable theories with noncommunicating memories per each optical element was by far more interesting and unique, because they provide an independent way for testing. We point out that the argument of Michielsen et al. begs the question by suggesting that more experiments are needed while providing no details on how one may avoid any and all interactions that may lead to replacing the two independent registers in the two beam splitters (as originally proposed in their model) with one collective register. Ultimately, we believe that it is not possible to suggest an experiment, even in principle, that would allow a test similar to that from Ref. 2 with the two beam splitters outside of the light cone, i.e., where one could guarantee that no sub-luminal communication occurs, because the ‘‘message carriers’’ are photons that travel with a luminal speed. Given that this loophole can never be closed, it is really hard to define what ‘‘independent’’ means. In view of this, it is unfortunate that Michielsen et al. do not offer any discussion on how to ensure that optical parts are ‘‘independent’’. Particularly, no details are offered on how to reduce independent BSs to a single system that guarantees that the functionality of the device before and after reduction is consistent with QM. How would that reduction occur in nature? Is this reduction adiabatic or discontinuous? How can one control the transition from a set of non-interacting EBCM elements to a single EBCM element? In other words, where are new registers located, and why are they to be moved from the BS as initially asserted, and how does such a device reproduce laws of QM before, during and after such a drastic transition? We believe that, by invoking a non-formalized locality loophole, Michielsen et al. commend the experimental efforts of Ref. 1, acknowledging that all other aspects of the experiment are flawless. That is, the authors of Ref. 1 conducted a fair experiment, in strict accordance with their initial theoretical proposal, and free of an unintentional bias. In conclusion, the comment does not explain how and why the EBCM model is modified to overcome the Ref. 1 test. Further, it neither clarifies the relevance of the model in its new version nor proposes new tests of it. By invoking a non-formalized locality loophole, the authors implicitly moved their model to a much larger class of local realism models.