Whether this kind of validation can prove that a test has meaning will hinge on what you feel it means for a test to have meaning.

In that case we should probably taboo "meaning" (in this context) and talk directly about whatever it is we want a test to do — make clinically useful predictions, carve reality along its natural joints, etc.