What would you call an approach that says this:Well, personally, I'm not for nonlocal realism because I don't like the LET assumption or non-forward causality. Basically my view is nonlocal nonrealist.
A system is a physical setup that is established by a preparation and has the function of associating any hypothetical measurements we can describe with a set of probability distributions, including two-point correlations (and higher). Quantum mechanics provides us with the instructions for connecting the preparations to the probabilities.
To me, that description is all we need to use the system concept in science. I see that it does not necessarily invoke realism in the sense of independent parts owning their own unique probabilities that can only be changed by "influences", we only invoke that picture when it is actually working for us (which is, as I said, when the influences are subluminal and can themselves be observed by intercepting them), so that gets called "nonrealist." But it's the "realist" approach that seems unrealistic to me. In any event, what I just said never encounters any issue with nonlocality, either in terms of Bell's theorem in EPR systems, or in terms of fermionic indistinguishability in white dwarfs.