There's been a fair amount of speculation about Gulenko's +/- theory, without much solid discussion of the underpinnings. It seems that any discussion of how far one can go with it, whether one can reverse the signs, and so forth, ought to involve some discussion of what +/- means and why.
First of all, +/-, as originally conceived, appears to be equivalent to the process/result dichotomy, but applied to functions rather than types (i.e., reinterpret the dichotomy as being about the base function of the respective types).
Defined as such, it is merely a mathematical pattern and cannot be removed from its original context (e.g., you can't have Te blocked with Ni with Te being + or Ni being -).
But then there is the semantic interpretation of what +/- *means*, which as with any semantic interpretation, could always be removed from its original context.
As I understand it, the original semantic interpretation presumes a sort of "normal cycle" of processing:
N -> T -> S -> F -> N
...meaning that if one generates (+) ideas or possibilities (N), one has a problem one must solve (-), being that of understanding or organizing these (T). And once one generates (+) that plan, one has the problem to solve (-) of implementing or realizing it (S). And when one has generated (+) that reality (S), one has the problem to solve (-) of making sure that it affects people in the right way (F). And when one generated (+) has the relationship or atmosphere between people (F), one has the problem to solve (-) of figuring out what the meaning of it all is or the potential of the people (N).
It was for this reason, I deduce, that Gulenko assigned + and - signs. Originally, the "-" form of the function was meant not as an "antagonist stance" toward that function, but rather as a problem-solving mode as opposed to a generative mode. I believe that's why Gulenko called + "proactive" and - "reactive" (react to the problems you have to solve).
That also fits in nicely with process/result...process is the generation of something, whereas result is doing something just so that you can get a result (i.e., solve something that's a problem for you).
Where this becomes a problem for some Socionists is that this also seems to be the relation between the base and producing functions. That is, the base function is supposed to be generative, and the producing function solves the problems generated by the needs of the base function.
I believe this is why Smilex said sometimes that the relationship of base/producing functions was reversed for result types.
Another related problem for this theory concerns whether the cycle pictured above is intrinsically the way things are, or if it's just an arbitrary direction.
For example, why couldn't one just as easily have:
N -> F -> S -> T -> N
Here, you have an idea or possibilities and you express their significance, or you have certain values and relationships and so forth that lead you implement various things and so forth. Does that make any less sense than the other cycle, and why?
Those who see the relationship between accepting-producing as "accepting -> producing" for all types will tend to view Socionics in a way which is intrinsically opposed to Gulenko's +/- model. The alternative is to see the relationship between accepting-producing as being modified by whether one is a process or result type.