Nonlocality without nonlocality

TL;DR

This paper challenges the traditional interpretation of Bell's theorem by showing that violations of statistical independence imply nonlocal correlations, suggesting quantum phenomena might be modeled by local theories.

Contribution

It demonstrates that denying statistical independence leads to nonlocal correlations and constructs Lorentz-invariant theories that do not rely on this assumption, proposing a local explanation for quantum phenomena.

Findings

Violations of statistical independence imply nonlocal correlations.

Lorentz-invariant theories without statistical independence are possible.

Quantum phenomena may be modeled by local theories.

Abstract

Bell's theorem is purported to demonstrate the impossibility of a local "hidden variable" theory underpinning quantum mechanics. It relies on the well-known assumption of `locality', and also on a little-examined assumption called `statistical independence' (SI). Violations of this assumption have variously been thought to suggest "backward causation", a "conspiracy" on the part of nature, or the denial of "free will". It will be shown here that these are spurious worries, and that denial of SI simply implies nonlocal correlation between spacelike degrees of freedom. Lorentz-invariant theories in which SI does not hold are easily constructed: two are exhibited here. It is conjectured, on this basis, that quantum-mechanical phenomena may be modeled by a local theory after all.