For the most part, you can't really type other people in socionics, and Sol had a point when he said you can only type people 20% of the time. To be able to really type someone else, you have to completely know their motivations to do what they're doing and assess their competence. Unless you actually have the same or at least similar IEs, you can't do that at all, and saying "you must be X type because I'm Y type and I see X" is a circular argument. Observed behaviors can't directly proceed from "information processing" the same way qualia can't directly proceed from stimuli (hallucinations, everyone actually sees color differently, etc.).
This is also why socionics can't/shouldn't pretend to be a science even if you want to use it. Socionics is ideas without any vague possibility of empiricism. You can't see yourself except reflected in other things like a mirror, and since all you see is your perceptions, you paradoxically have to be constantly changing them around to compensate for your blindspot to have a constant idea. Socionics can be a body of knowledge and useful as long as you don't think all knowledge is ultimately empirical or get too dogmatic with it, but it can only have meaning to the person doing the typing because you can't get outside of your own perceptions. The only way to vaguely type other people is based on your relations with them, but doesn't that sort of bring two people together into one unit with shared thoughts anyways? You can't just sit back and type someone based on their behaviors and say you're being rational about it. It is inherently irrational since empathy is the only way to know how someone would relate to their IEs since everyone displays all of them. You have to figure out how the other person relates to their IEs, which means empathy, and even that depends somewhat on them. So socionics is marvelously the least objective thing there can be, and even a pseudoscience as long as people pretend it's scientific.
This is also why socionics can't/shouldn't pretend to be a science even if you want to use it. Socionics is ideas without any vague possibility of empiricism. You can't see yourself except reflected in other things like a mirror, and since all you see is your perceptions, you paradoxically have to be constantly changing them around to compensate for your blindspot to have a constant idea. Socionics can be a body of knowledge and useful as long as you don't think all knowledge is ultimately empirical or get too dogmatic with it, but it can only have meaning to the person doing the typing because you can't get outside of your own perceptions. The only way to vaguely type other people is based on your relations with them, but doesn't that sort of bring two people together into one unit with shared thoughts anyways? You can't just sit back and type someone based on their behaviors and say you're being rational about it. It is inherently irrational since empathy is the only way to know how someone would relate to their IEs since everyone displays all of them. You have to figure out how the other person relates to their IEs, which means empathy, and even that depends somewhat on them. So socionics is marvelously the least objective thing there can be, and even a pseudoscience as long as people pretend it's scientific.