Quote Originally Posted by Phaedrus
Some unsorted personal reflections on this extremely interesting material ...
One of the best things to surface in a while, sure.

From my perspective I have thought that I focus on generalizations. I have associated that with simplification, clarification, and one of my working hypotheses have been that it might have something to do with inductive reasoning in contrast to deductive reasoning. That's one reason why I have thought that Sergei Ganin's INTp Uncovered profile is misleading when he describes the INTps as enemies of generalizations. Are we talking about the same thing or not? I'm not sure, and now Gulenko complicates things a bit too.
It's not so simple. INTp is Ni to begin with. Ni is less specific than Si for example. It's also an internal function to begin with. As a starting point it is already a rule, of something generally happening. It's the end-point of inductive assumption for an IP type. (Whereas for the EJ it's the end-point of trying to deduct something to non-existence.) So an INTp is someone who tries to condense his general knowledge into simple packages of Te. From large vague knowledge to powerful simple assertions.

ILIs are, in essence, synthesizers. I try to explain a synthesis when I say that, despite their apparent differences between how the types are described in Socionics and MBTT, they are nevertheless talking about the same referents, the same groups of people, the same empirical types. I have thought that that means that ILIs don't build deductive theories from axioms or basic assumptions -- that they try instead to explain the external information in more simple terms, but maybe that can be interpreted as almost identical to what Gulenko says about the involutionary (result) types:
Gulenko has a problem here though. He uses analyst as a substitute for Static and synthesizer as code-word for dynamic. I'm dynamic and so I'm a synthesizer, but I also use axioms and deduce stuff out of that. What is critical to understand here that Gulenko refers particularly to analysis of perceiving functions (extrovert-perceiving functions, introvert judging functions). I analyze judging functions, but my nature to the environment, to the actual events in the world is still as a synthesizer, while on a social level, I'm an analyst. (Judging-perceiving difference). Or to put it another way, not every taciturn type is static.

Quote Originally Posted by Gulenko
Observing and comprehending complex phenomena, inductive brain reduces them to the simple, purified of details diagrams and the models. Thus, involutional types in order to be dismantled at the situation, will simplify it. For them is typical the judgment in the reverse order - from the complex to the simple.
So, according to Gulenko, that is what LIIs do -- simplify and reduce -- whereas the ILIs, belonging to the group of evolutionary (process) thinkers, would be the ones using deductive logic and building more complex theories out of simple axioms or assumptions, that is what I have associated with a the expression "a bottom-up" approach. And most people on this forum have automatically associated deductive logic with LIIs, not ILIs. How shall we understand this? Are we contradicting each other or not? What are we talking about? What is Gulenko talking about?
We must understand the difference between process-result and negative-positive. Both INTp and INTj are negativist. They close out ideas and in such a manner proceed from larger to smaller. Neither one is induction-related. But INTj is 'result'. They see things as simple, believe that their first idea of a subject is the totality of the subject, no need to go deep into it. Or if they do, another quick glimpse from another point of view will do the trick. They will not pause to examine it deeply, they will not go into depth, they will not proceed step by step and wait for the result to emerge only at the end-point of a long pondering. This is what it means when it is said that INTps over-complicate and INTjs over-simplify. It is about whether they handle the search of truth as a quick drive-by or a long trek. So INTjs will quickly produce good results in simple matters but will rarely reach the correct end in complex matters while INTps will take a long time to produce answers even to simple questions whereas they remain more capable in producing correct answers to complicated ones.

To complicate things further (if that is that what I'm doing ...) Gulenko also mentions in his article that for the involutionary types "is characteristic the accomplishment of backward motions - from the end at the beginning, from below - upward." I have thought that that is exactly what I have bee noticing in my own behaviour. That I almost always prefer to start with the conclusion, the main thesis, before I get into the details and the question how the conclusion was derived. I tend to ask: What's the point of all this? What (thesis) are you trying to argue for?
Then, you have a problem

Is Gulenko wrong about this?
No. But as I've asserted ad nauseam, no one behaves according to a single type every time. And you say you've only noticed it. So it's a recent phenomenon. Anyway, from what I've seen dialectic thinking works in unison with vortex thinking, so if you do a little bit of that, it's fine.

The LIIs try to simplify, to reduce a complex reality into a few axioms and assumptions, and the theory of the functions in Socionics is an example of that kind of thinking.

The ILIs try to see reality in all its complexity before they even begin to think about reducing things to basic assumptions. So, at least in the initial stage of the investigating process, the ILIs seem to complicate things further and further. But their ultimate goal is to arrive at an all-encompassing understanding of the world that can be captured in a theory.
Let me put it like this... LIIs reduce by discarding unneeded parts. ILIs reduce by connecting threads. Connecting threads is complex because it has to be done in a correct way whereas anything can be discarded and discarding is simple.

Gulenko also seems to focus on the first stage of an LII's approach, whereas I, as said, have focussed on the later stages. Is that a possible dissolution of our apparent differences in views? Do we actually agree on this?
Possible.

An important aspect of this kind of thinking is the tendency to classify, or as Gulenko puts it:

Quote Originally Posted by Gulenko
Algorithmic thinking also solves well problems to the classification, since has the gift of the recognition of complex means.
I'm not entirely sure what Gulenko means with this but I think it might actually be a mistake. Another example of missing the importance of narrative-taciturn dichotomy. ENTjs are pretty good at classifying things and they're not algorithmic thinkers. But I'm not really sure what he means with this. Perhaps he means the way algorithmic thinkers reduce things to the smallest possible form that contains all the relevant information? But the end-result isn't always a class. It can be a process. Hmm... Anyway, I think he's made an error in this.