I was referring to physics not being a 100% complete science yet...and audio would fall under that.
Basically, you're kinda saying that what we know we can measure today, we can prove...and what we can't measure today, should be assumed not to be important....right?
How does one do a null test with a 96kHz file against a 44.1kHz file without converting them first to the same kind of file, in which case, you've altered them to make them the same before doing a null test.....?
Anyway, the point I was really making wasn't that there is no difference between fidelity and preference, but that preference is equally important, and always measuring and relying on math to tell you what you should prefer is not IMO what makes for best scenarios in every case.