Well I think i read somewhere that laws explain sets of empirical observations, whereas a theory explains a set of laws.
With regard to the MBTI and trying to fit into one of its categories, part of the problem is that they are attempting to apply Jung's 'laws' (i.e. Thinking opposes Feeling, Sensing opposes Intuition, superiority of the dominant, etc) to a paradigm that is incompatible. You see Jung could say that because he wasn't focused on trying to use the functions to explain all of human psychology. He accounts for the vast differences in people with purposeful vagueness and by concentrating on a holistic view of the psyche, the complexes, the archetypal influences, personas, etc.
MBTI on the other hand, in trying to be empirical, creates a laws out of its own assumptions, but then tries to use Jung's ideas to back up those assumptions. But the two are incompatible. If you are going to use, as MBTI does, the functions as the raison d'etre of the psychological experience, that's fine but real people are not going to play so simply by rigid rules like "if your aux function is Ne your tertiary must be Si" and this kind of stuff. It's just too rigid, it standardizes what should be inherently variable. Jung doesn't care what your aux function is because its not important in his greater picture, its only MBTI and its trying to mix two incompatible ideas where you run into the problem of people seeming to favor both Te and Ti.
In reality, you would actually have to, as Reynierse did with his mutlidimensionality model, break all of the functions down into individual components and just score the strengths of each. You can't apply arbitrary laws like S vs N or J vs P if you are not going to offer any greater or more holistic psychic explanations. You basically have to measure S-N-T-F-J-P (and perhaps I-E) as individual variables and then just score them as a snapshot of percentages like the Five Factor Model. (And J-P absolutely cannot be pointer variables, they can't point to function use just to the definitions of judging-closure-seeking/perceiving-non-closure-seeking). The reason being is if you try to force laws onto something that is by nature variable, you're constantly either going to have to ignore evidence to the contrary (of which there seems to be a lot) or accept that your theory has limitations.
MBTI says things like "people may mistype because of social expectations, and lack of self-awareness" and so on, but that sort of defeats the point. Basically what they are saying is, people cannot be typed correctly unless they fit our model, and it should be the other way around. Jung accounts for this by spending time talking about the ego and the shadow and the personas and the nature of the complexes and so forth. We recognize the difference between the ego and its personas, but with MBTI it appears to be all of these things at once, and none of them. It's like it can't decide whether to just describe behavioral patterns, or personas, or underlying psychology -- it's sort of all the above and none of the above, and again this gives too much power to the functions which were never intended for this. And then we go off with JCF making more rules and theories on top of something that is already philosophical (like dom-tert loop theory) when the evidence that the tert is the same attitude as the dominant seems non-existent. We believe this simply because its what we've been taught. Again the difference between Jungian psychological philosophy and MBTI trying to be a science, but not wanting to play by scientific rules.
To say x person is an Extraverted Intuitive as the Jungians do, leaves the door open for a myriad of possibilities about the person, we're simply saying the person favors using intuition. We're not going as far as to say "...and they will secondarily favor Introverted Judgment and then Extraverted Judgment and these functions will have a heirarchy of strength and the extraverted of the two will have more impact on the outer orientation of the person and on and on and on." That just tries to do too much and then people run in circles trying to figure out if they fit ENTP or ENFP, or maybe they have J tendencies and so forth. It gets ridiculous.
To my knowledge Jung never says we can't use all eight function attitudes, just that conscious orientation will center around the dominant function which will generally be the most differentiated and point in a given direction (and by extension the inferior function will be the conscious expression of the unconscious). This is not describing any of the hard absolute qualities that MBTI and modern type theories interject onto type, like Si will be conservative and past-oriented and Ne will be creative and impulsive. In fact you could take Jung's words to mean anything, he's just sketching out basic processes as a sort of philosophical guide, a way of looking at things, not trying to nail people into solid categories. Maybe we really honestly only use one function, the dominant, or perhaps its the inferior that carries the weight in some people and the other three don't mean much. Jung leaves that possibility open, but MBTI does not.