What the function is? The function is a "type", and it's part of a "typology". There is no "is", if that makes sense. Also, there is no "does". The function is a description of a classification.
Some aspects of that classification are extremely abstract and can only be observed/experienced, and not explained precisely with words. Trying to say what "Ni is" is like trying to describe "what sugar tastes like".
And don't go and say, "That's completely Si!" or whatever. It's an ANALOGY, a metaphorical description. You're supposed to use it to help you figure out what I'm getting at, you aren't supposed to use the analogy as a concrete example to reason from as if it were some sort of statement of fact that can be proved or disproved.
That was an ANALOGY. I'm putting things DELIBERATELY in VERY SIMPLE TERMS so you can then build the abstract ideas from it.
That's because you're taking my description way too literally.
I am not saying, "Ni thinks ONLY in terms of what things do." I'm not saying that Si NEVER thinks in terms of what things do.
I am saying that the framework that Ni tends to build tends to be mostly in terms of what things do, of their roles, etc. Of their "functionality". The Si framework instead tends to be built on what things "are". The distinction is subtle. Si thinks of "serial killer" (a concrete label) and reasons from there. Ni thinks of "tends to kill people often" (an abstract function) and reasons from there.
Yes, I know it all sounds "concrete". That's because it's a very simple example, and we tend to conflate "concrete" with "simple". Si types are quite capable of using complicated and abstract reasoning. Ni types aren't "better" at abstract reasoning overall, so much as they're very good at particular aspects of it. (Ne types are good at other aspects of it.)
Ni types (and Ne types) are good at taking an abstract point and using it to figure out other abstract points, without having to translate the abstractions into concrete ideas. When we're asking to explain our reasoning, it can be deceptive, because then we're forced to translate each abstract idea into something concrete: the reasoning isn't concrete, but the explanations tend to be. That's why it's kind of freeing when you can just handwave and describe the abstraction without being concrete to another intuitive type: they get it without having to laboriously translate the abstract to the concrete.
Si types, conversely, are great at reasoning from concrete idea to concrete idea. They can reason in ways that we'd describe as "abstract", too, but they tend to need to start from something concrete, and then launch into the abstraction. The hops from abstract idea to abstract idea are much more unfamiliar to them and tend to be a blind spot in their reasoning.
~
That's why I veer away from the "unconscious" part of the description of Ni. It makes it look too mysterious, in my opinion. I would say that to most Ni doms, our thinking is very conscious, where the "unconscious" and "irrational" part is that we see patterns in the world that other types don't see. Those patterns we "just see" are predominantly functional patterns, descriptions of what things do, and how they connect in an active cause-and-effect kind of way. Because we "just see" them as obvious, it is difficult to explain how we "know" them. If we were Si types, we would just say that we "remember" them, because that word exists for what Si types do. We Ni types also "remember", but we remember these patterns, such that when we see them again, we know exactly what they are and how they work.
I think what is happening here is simply the perspective clash that these conflicts highlightng. One who's main ego world-view is Ni, who understands what it is, but by the nature of the function can't really explain it.
This is exactly what was pointed out in a one on one discussion years ago with another INTJ (who's very knowledgable and trustworthy with these things), even down to me often taking "analogies" too literally.
Ni was described as inferring "what's being left out" of an existing arrangement of elements. (where Ne infers external elements in terms of a larger arrangement that gives them meaning; and you can see right here, where this can be confused with the Ni "bowl" analogy; being that the bowl could be considered a "larger arrangement that gives them meaning", which is why I find some of these analogies ambiguous).
Ni was portrayed as dealing with "what doesn't yet have a home or a voice in a larger pattern", and thus will reach for an analogy that will give that left-out something a way of being perceived by others, and in order to flesh it out, one has to find things in the outer world that are sort of like it, to make an analogy and get it out of the head and into words.
And significantly enough, here:
"That's why INTPs and INTJs are often in conflict."
Ti wants "to hammer in the stakes of a tent that offers the best shelter for others" [i.e. in my own way of framing it "This is the TRUTH (T), and since I realize it (i), then I think others would want this truth as well" ("subjective" perspective projecting onto others)], but the Ni type is "usually cutting a hole in the side of a tent to peer into the dark and make out something else that's moving on a distant landscape."
So all of this would seem to agree with what you've just said. But the problem is,
people are still not understanding these analogies (rather than clarifying them, which I and others tend to expect them to do), and the biggest part of this is likely due to the strong presence of Inquiring Awareness types (mainly the NP's, and would include SJ's) for whom Ni is a
shadow function, and who thus who interpret in their own introverted perceptive perspective, which is "concrete".
So left this way, that method of defining the function would become like a sort of "in-house" understanding among the NJ's; and you seem to disagree with how I put it, but then not everyone can understand it through those sorts of analogies like you can. (The illustrations end up pointing the wrong way, or even sounding like another function, or at least becoming ambiguous.
THAT's what makes it "mysterious" to us!)
So my strategy was to go after what exactly makes it "introverted" (or "subjective"); what's really internal about it (which is what would distinguish it from Ne; and of course focusing on what's "abstract" about it distinguishes it from Si), and so that is the "unconscious impression", as much and as long as I tried to avoid that term myself.
(And I didn't think you were saying "Ni thinks ONLY in terms of what things do." or "Si NEVER thinks in terms of what things do." It wasn't which does what, and how much they do it; I just thought the description sounded a bit too "rational" in function).
So what you're calling "what things are vs what things do" is basically what, to go back to Jung, is "what it is" vs "where it's
heading". A bunch of alternatives I came up with is "is (actual) vs could (potential)" [i.e. as in "could do"], "experience vs story" [a "story" is a synopsis of what things do], "practice vs theory" [theory is an understanding of what things do], and Mark Bruzon's "static (items) vs motion (process)" [again, "motion"="doing"].
So again, I just saw "what things do", and from a Ti perspective, this is a more "rational" type of description.
(While what you described as "what's unconscious about it" covers the N ["seeing patterns others don't see]; which could hold for Ne as well, but the difference is that Ne infers them directly from the object, which can be
shown to others); remember, the "i" part is called "unconscious" as well. That's what I meant by Ni being doubly "unconscious".
So while "what things do" might technically describe it, I wouldn't say
the function becomes "conscious" when preferred; what's "conscious" in a type stack [and I believe this point really needs to be taken and understood more] is the
ego state (or a particular "complex") that marks the type and the function's position. Immediate physical sensation is definitely "conscious" for everyone, but when we say "Se" is "unconcious" for an SJ, NP, or an NJ who hasn't developed it yet; it's a particular ego state of awareness that either pays more attention to it or not that is being emphasized
So this is why Ni is often portrayed as "sudden insight" that comes up in an "aha moment" fashion. I don't think it's something itself consciously reasoned out, like referencing an "understanding of how something works". Of course, all these functional perspectives are present in all data, and thus implicit, so there is always
judgment that fills in a perception process, an adds that rational understanding).
A simple example that I've encountered in real life: some students are VERY good at math, and can get 100% on a calculus exam, and be experts in manipulating trigonometric functions - far better than I ever have been. Trigonometry and calculus are weird parts of math that are helped very much by memorizing very specific formulas and methodologies: it isn't obvious that one expression is the same as the other, or that if you substitute things "just so", a complicated integral becomes a simple integral. BUT ...
I can take those students who are used to using "m*g*sin(theta)" to get the force on an inclined plane, and instead of measuring the angle from the horizontal, I mark it as being measured from the vertical. And I can point out that this is not the normal angle theta, but its complement, so I'm being fair and letting them know that something is up, and not being sneaky/tricky about it. It doesn't matter. They'll still use "m*g*sin(theta)" and not "m*g*cos(theta)". Why? Because the "right answer" is "m*g*sin(theta)". That's what they've memorized. They know it's reliable and correct. They can do all this complicated and abstract math, integrating and differentiating trig functions and so on, but a simple problem that involves nothing complex, just an understanding of what sin means and what cos means, is beyond them.
That isn't to say that they can't learn that, too, but only in the sense of it's being a variation of the original problem. I.e., they memorize the new solution as applicable to the new case. They're weak at figuring out how to figure out things from first principles if I changed yet again something else in the problem. (Same physics principles, almost the same math, slightly different set-up.)
I also see this in an ISTJ coworker of mine, who is an excellent software developer. If it's a pattern he's followed before, everything is tic-toc-logical-beautiful. But if there is something new, or it's a new technology, he tries to apply the old patterns in the new paradigm, and wonders why he's having so much trouble. I can explain and teach to him the new patterns, and he eventually figures it out with practice, but he doesn't "just see" those patterns which are obvious to me. He sees the form and works with the form, but he doesn't see the function except as a side effect of the form. I see the function (all of those patterns that I can "just see" without having to think hard), but I don't see the forms he sees except as a side effect of the function.
As a side point, all of this are more good examples of why Ni/Se would be termed "Realizing" and Si/Ne would be "Inquiring" (or as I put it, involve more "comparison"), to use Berens/Montoya's new terms.
So I'll have to think more on this. Perhaps I can find the simpler alternative to "unconscious" I've been looking for from it.
I am constantly comparing concepts/contexts/experiences against the info that's built up over time in my head, then using that comparison to come up with new future possibilities/plans/solutions. I see my Si-dom sister doing the same thing but with different data/criteria. Her plans/visions/possibilities can be just as unrealistic as mine, but in a different, more concrete area of data.
To use the salad and bowl idea, I might see a bowl of fruit and compare it to the bowl of salad that I've stored in the framework in my head, see that they have a bowl in common, and wonder what other commonalities I can find between the contents, and then look for novel ways that those commonalities might be used or related or have meaning. My perception of the bowl of fruit is not so much about the bowl of fruit, but about the concept of a bowl of fruit and how it might relate to concepts already in my head.
So, to me, Ni is about storing concepts of observations that then get triggered and connected by a sensory stimulus of some sort. To my thinking, the process of Ni is similar to Si, but it focuses on a different set of data. In my mind it might be easier to understand Si because it deals with more concrete looking concepts, then what's learned can be applied to understanding Ni. And that in itself, to my thinking, is an example of Ni at work.
Anyone can compare things, but from what you're saying, you do the comparing
once you've already perceived the data, in order to extenalize it (in this case, doing something with it; which fits what I mentioned above). You don't have to compare
in order to do the perceiving to begin with. That's what I was trying to say. (And again, the process of comparing the bowls sounds a bit too conscious and rational, to the point that I can even identify with them in my Ne/Si perspective, but then, maybe that's another one of those analogies that shouldn't be taken too literally. Either attitude of N is about the "concept of" things, but again, Ne will be more external focused [the
things themselves], while Ni, as described, is "the concept of the concept" [ i.e. "meta"]).