Fascinating points. However, it seems to me that the underlaying definition of "real" (which is a bit unclear) is not satisfying in an ontological sense. Do mean that our models/patterns fitted on observations are somehow real?

Maybe if a thing is a part of a pattern it is less interesting than before we found this particular pattern. OTOH, the pattern maybe more interesting than separate things. In any case, it is unclear to me in what sense this pattern is "real".