Theoretical physics is also very much a thing.
> and should therefore be testable by definition.
That sounds nice but the definition of testable is hard to pin down and changes over time.