[yor_on]

Abstractions is what will define the experiment, then there will be mathematics supporting it. If the experiment fails the mathematics was wrong, here. It doesn't discuss whether those mathematics will fit 'somewhere else' though. you can look at it at least two ways. One in where whenever a mathematical proof becomes true, it also has a possibility of existing 'somewhere', even if not here. The other is experimental physics, sometimes the mathematics won't exist, so you will have to invent them, to fit the experiment, or logic.
==

Actually this is even more true when it comes to theoretical mathematics. As I gather Perelman had to invent a lot of new ways to describe it, and as I remember, so it was with Einstein.