I don't know why this is bothering me but it is so I will waste some time talking about it so that some of the shit I've seen on this forum will stop irritating me.
What if we stopped thinking about functions in terms of Ne vs. Ni or Ni vs. Si or whatever and viewed them all as gradients on a graph. For example if we wanted to see the difference between

and

we could view them as at opposite ends of the Ji spectrum like so:

<----------|---------->
Under conventional socionics an inxj would be -1 and an ixfj would be +1.
Now we apply magnitudes to functional development. Not all people of type xxxx use their functions to their fullest extent right? Some of them have more "developed" capabilities. So let us say that person A has x

and person B has 2x

. We would show that as:
|
| o
| o
| o o
| o o
------------
A B
Ok simple. Now we look at two people of the same type and two of their functions simultaneously. Instead of creating a different graph for each function we can show both on the same plane.
Two infps:
[img][/img]
infp one has 2 Ni and 1 Fe and the other has 4 Ni and 3 Fe.
Note that the same diagram can be used to explain enfjs if you take having more Fe than Ni to equal being enfj and having more Ni than Fe to equal infp.
[img][/img]
Now what about the rest of the functions? The problem is in what you see as the "opposite" of property x (if anything). If we want to describe a person in terms of all 8 of their functions problems arise. For example is the "opposite" of an intj (TiNe) an esfp (SeFi) when an intj supposedly has access to "esfp space" in their consciousness (though in a weak format)? Is the greater distinction one of conscious vs. unconscious (functions 1-4 vs. 5-8) or pleasure vs. pain (1-2 vs. 3-4)? Also does one "oppose"" you more if they actively seek that which is not of interest you or if they place the information upside down? Eg. Does an intj irritate an isfp more or does an esfp irritate an isfp more? One lacks any sort of of common ground (1-2 by 6-5 [intj]) the other uses their functional preferences in the other direction. In visual terms:
[img][/img]
or
[img][/img].
It is obvious that you
could overlay them all and average them in different ways but unfortunately in areas such as this PEOPLE MIMIC THE MODEL (once they are aware of it) rather than trying to "find" their "true" interaction with other people (if such a state is).
To solve such problems I suggest polarizing all aspects of the model against itself. As such:
[img][/img]
{i meant for that to be a cube btw}
where the intersections of the model stand for certain values while the connections between intersections stand for interaction between those values.
For example if 1 = Fe, 2 = Si, 3 = Ni, and 4 = Te
then [1,2,4,3] = esfj, [2,1,3,4] = isfp, [1,3,4,2] = enfj, etc etc. In other words the informational flow for an esfj would be a 1 > 2 > 4 > 3.
Now let us see the sum of the interactions that our language base will allow. For that let us place the functions in any position that we choose until we arrive at a formulation that closely approximates our actual behavior.
In the above example I juxtaposed consciousness with unconsciousness in model a (JePi/PiJe = functions 1-4 while 5-8 were left blank but which will now represent JiPe/PeJi). Let us have 5-8 be Ti Ne Se Fi.
In that scenario:
left/right = consciousness/unconsciousness
top/bottom = beta/delta
front/back = alpha/gamma
3-5-6-4/7-1-2-8 = NT/SF
5-4-1-8/6-3-2-7 = J/P
etc etc
Note that you can order the dimensions in any manner that you wish such as classic Jung: