None of this has to do with measurements and I’m not sure why they’re being brought into the conversation . . . .
Instead it has everything to do with whether and how much we can / should trust our brains to accurately and reliably perform nonsimultaneous A/B comparisons when it is notoriously inaccurate and suggestible in that respect. <<
The connection between those two thoughts is that measurement systems are repeatable and precise and often accurate and are not suggestible. Therefore, they can help us overcome the issues with relying on the brain/ear system entirely. (I think you already know and appreciate that position.)
I would hope that even those who talk about "training" the ears would agree that an objective foundation is needed for such training. Otherwise, the whole situation becomes quite murky, as one can be "trained" to prefer highly colored sound. In a sense, we all get trained to prefer our current systems, as has been pointed out many times.