Check out the concept of a "Everett-Wheeler telephone". If quantum mechanics had a non-linearity and... some other conditions, then you could use it to send, u

Check out the concept of a "Everett-Wheeler telephone". If quantum mechanics had a non-linearity and... some other conditions, then you could use it to send, uh, some information between two sides of a branch point, uh, once, I think. At least, you could send a single bit from one side of the branch to the other. I think the idea is, if there are non-linearities in time evolution, then either [the thing described above] is possible, or sending messages back in time is possible? So, that's, I suppose, a reason one might expect time evolution to be completely linear? (beyond just, it being not completely linear being like, super ugly.) But, also, I don't think the "duplicating the universe requires duplicating all the matter" really follows from QM stuff. These "splittings", they aren't... It is all just linear combinations of things. Ok, so, idk how to handle this in like, a relativistic context, but in a non-relativistic context, if you are just talking about unitary time evolution, with no born rule discrete jump stuff imposed on top (which, if you are using MWI , which I assume is what you mean because you are talking about the splittings, is the thing), then it's not like you have some single time at which you go from one universe to two universes? It's all just continuous time evolution of the wavefunction. The worlds are just projections of this wavefunction into parts that can be considered independently because of like, decoherence and such? Something like that. Maybe you could think of it as less "the world if the measurement outcome is x" and more "the world insofar as the measurement outcome is x" ?