Joy Christian wrote:Very well, although I have explained my point in detail in the Appendix D of my paper which triggered this discussion ( i.e., this one:
https://arxiv.org/abs/1501.03393 ), let me explain my point again using the language traditionally used to prove Bell’s theorem ( see Bell’s famous paper of 1964 ), using the “large N limit” ( i.e., integration ) as well as the usual assumption of “local hidden variables”, as insisted upon by the proponents of Bell’s theorem.
Let me reproduce the key part of the derivation of the bounds -2 and +2 on the CHSH correlator from Bell’s paper of 1964. It involves the mathematical identity we discussed previously, namely E(X) + E(Y) = E(X + Y), and an average of the random variable X(a, a’, b, b’, k) with k as a hidden variable (usually denoted by lambda), which is defined as
X(a, a’, b, b’, k) := A(a, k)*B(b, k) + A(a, k)*B(b’, k) + A(a’, k)*B(b, k) – A(a’, k)*B(b’, k),
where A and B are +1 or -1 and a, a’, b and b’ are possible measurement directions that may be freely chosen by Alice and Bob. The mathematical identity of interest is then
Int_K A(a, k)*B(b, k) rho(k) dk + Int_K A(a, k)*B(b’, k) rho(k) dk + Int_K A(a’, k)*B(b, k) rho(k) dk – Int_K A(a’, k)*B(b’, k) rho(k) dk = Int_K X(a, a’, b, b’ ) rho(k) dk ,
where Int_K denotes integration ( i.e., “large N limit” ) over the space K of hidden variables k.
Let me stress that the above is the standard mathematical identity, frequently used in the derivation of Bell’s theorem, and can be found even in Wikipedia. The RHS of the above identity is simply an “average” of the random variable X(a, a’, b, b’, k) defined above.
Now it is claimed by the followers of Bell that X(a, a’, b, b’, k) is just a random variable on the space of all counterfactually possible outcomes that may be observed by Alice and Bob ( as in my dice example ). And — they claim — the bounds of -2 and +2 on the CHSH correlator follows immediately from its “average” on the RHS. Well, they indeed seem to be, because, as noted, the equality between the LHS and the RHS in the above equation is a strict mathematical identity (as frequently stressed by the followers of Bell), and it is easy to see that the RHS of the equation is bounded by -2 and +2. But that is just the problem.
Since the above equality is a strict mathematical identity, we can just take its RHS as our starting point and ask: What is exactly being averaged in it?
Well, what is being averaged is a quantity that cannot possibly exist in any possible physical world, as demonstrated in my paper linked above. Spacetime events such as B(b) and B(b’ ) are only counterfactually possible measurement results, along two mutually exclusive detector directions, b and b’. Since b and b’ are classical, macroscopic directions, no experimenter has the ability to align his or her detector along both b and b’ simultaneously. Therefore, regardless of any physical theory — classical or quantum — it is impossible for Bob to measure both B(b) and B(b’ ) simultaneously, just as it is impossible for a die to land on 3 and 6 simultaneously. Therefore X(a, a’, b, b’, k), which involves sums like B(b)+B(b’ ), is a totally fictitious quantity that cannot possibly exist except in some fantasy world, and therefore the bounds of -2 and +2 derived from X are equally fictitious. They have nothing whatsoever to do with any possible physical world, classical or quantum. They are merely mathematical curiosities. They by no means rule out any local hidden variable theories. They by no means forbid a strictly local, realistic, and deterministic derivation of the correlation E(a, b) = -a.b. Bell’s so-called “theorem” has no relevance for physics whatsoever.