Dear Jean,
Your paper is very well written, and a pleasure to read. I thank you.
However, I think your arguments are more than a decade out of date. I analyzed the nonlocality issue of EPRB correlations in Chs. 23 and 24 of my book CONSISTENT QUANTUM THEORY (http://quantum.phys.cmu.edu/CQT), where I considered EPRB correlations, and demonstrated that they are local in the sense that a choice of measurement on Alice’s side has no effect whatsoever on the particle on Bob’s side. Some more recent discussions are in “Quantum Locality,” Found. Phys. 41 (2011) 705; arXiv:0908.2914 and “EPR, Bell, and Quantum Locality”, Am. J. Phys. 79 (2011) 954. arXiv:1007.4281. You might also take a look at the item I just posted on Shan’s site, “What Does Bell’s Theorem Really Tell Us?” Regarding the Found. Phys. item, I note that it was refereed by David Mermin, who was unable to poke any holes in it. Granted, neither he nor I are as good mathematicians as you are, and you may be able to find a flaw. But I do think you need to take serious arguments seriously, and not simply dismiss them with a single quote from Gell-Mann’s popular writing.
With reference to Einstein’s boxes, that is easily solved. If you adopt a framework in which it makes sense to ask whether the particle is in one or the other box, the answer is that it was, with probabilities given by the Born rule. If you adopt a framework in which the superposition state between the two boxes is a quantum property, then you cannot discuss which box it was in for the same reason that if a spin half particle has the property that S_x = +1/2 it is meaningless to ask for the value of S_z. I am here using the consistent histories approach, which has no measurement problem. If you have some solution to the quantum measurement problem, I would be interested in knowing what it is. If not, then you might want to take a look at mine.
Best wishes! Bob Griffiths