It seems as though there is much disconcertion on this board about whether socionics is an actual certified honest-a-goodness above-board bona-fide science. But I ask you: what good is it to you personally if socionics is "proved" or "disproved"? That's the lazy man's cop-out. Whether or not socionics is valid, the interest in another human being's essential nature -- intrinsic to the act of typing -- is a wonderful avocation to exercise. To some extent astrology can be useful in the same regard. The diffence of course is that astrology is arbitrary and socionics is based on careful observation. But to the extent they allow us a framework to delve into another's character they both serve the same laudable function.
I don't mean though to give socionics short shrift here by comparing it to such blatant quackery. Whether or not massively parallel processed supercomputers fed reams of juicy data would eventually inductively spit out the boolean value of true or false, it's probably the most fundamental model in use, but from a deductive standpoint.
Here's an entertaining (for me) attempt to (mostly deductively) bisect out to the 16 types:
One of the first discoveries an infant makes is that, unlike in the womb, they can alter their world. If they close their eyes, everything goes black. If they cry, their mother comes to feed them. A baby begins to sense its power and freedom.
This power manifests itself in two forms. One way is the power to alter or adapt one's perceptions, to navigate perceptually the thick and wonderous world around them. The other way is the power and freedom of choice, to realize (not conciously at first, of course) that for every act there is a consequence.
A preference must soon develop between these two modes because the exercise of one mode limits the exercise of the other. How? A choice requires a steady perceptual framework to be meaningful (should I sit or stand on that strangely shaped wooden thing? is influenced heavily by the perceptual framework that interpets it as a chair). Now any perception that might get in the way of making that choice becomes a hindrance. At the same time perceptual freedom limits the ability to make definitive choices because one has the ability to shift perspectives (thank you Labyrinth for this insight) and thus the ability to undermine the absolutism so useful in this endevour.
Thus the first dichotomy solidifies, J vs. P.
The next discovery is the discovery of the self, or the discovery of objectivity. These two things are closely related, in fact different ways of talking about the same yet opposing things. The sense of self requires self-awareness, which requires looking at oneself as if one were not oneself. This not-oneself is the beginnings of objectivity. However this not-oneself-that-is-oneself is in a tenuous position. It is being pulled in two directions, by its two parents, as it were. One parent is the objective world that first awakened it to the realization that we are not alone; the other is the deeply subjective, non-introspective self, the dark sea that knows without understanding that it is completely alone and complete in itself.
The identification of self at this stage of development determines the next dichotomy, I vs. E.
I have to go to sleep now but if you are intrigued I will continue with the next two bisections...