The main problem is that quantum mechanics is telling us that systems evolve in two, apparently contradictory, ways:
- isolated systems evolve through unitary time evolution; and
- coupled systems evolve through an application of the measurement process.
Decoherence is an attempt to reconcile these two processes, but it's not clear that it works for systems with spatialy seperated entagled states. However, it all works out if you define a theory with emergent locality.
Yes, but the crucial question is: isolated from what? Isolation is realative, not absolute. If you put Schrodinger and his cat [1] (or Wigner and his friend) in a box, they are isolated from you but not from each other. (c.f. the Fruachinger-Renner paradox).
---
[1] This is intended as subtle ironic humor. Schrodinger actually hated cats, which is the reason he chose one to be the subject of his famous thought experiment.
