From what I understand of what you said, you are saying that every aspect of socionics is a model that attempts to approximate certain things about real people in the real world. It is entirely disconnected from reality like how a window has nothing to do with the view from it, even through it can be used to structure it, show certain patterns. Or reality does not obey the rules of socionics. Did I misinterpret you?
Do you think one could construct an entirely empirical kind of socionics?