I'm not sure if people understand what I'm saying...
If we make a correlation, "Person A makes Person B uncomfortable", then we don't actually know if Person A actually has much to do with the reason why Person B is uncomfortable, or not. It could be correlated, or it could not.
For example, we might say that "A white person hates a black person, because he's a racist". But if all we do is make observational correlations, then we might conclude that white people and black people do not get along, and they're "conflictors". Or even if we gave a superficial "reason", then we might say that "White people have racist functions, and black people have racist PoLR", then that's even worse.
It's pretty obvious that the "because he's a racist" part is a very important information and an explanation. It had very little to do with those two particular individuals per se. It had almost everything to do with this complex idea, "racism", which works outside of those two particular individuals.
And yet that's exactly what Socionics is doing. It's all just about making correlations, without coming up with any explanations.
It's a mistake to say "Socionics affects us like gravity, regardless of what we think about it", because we don't actually know the objective reason for why people are affected in certain ways.