Questions about socionics which make me rather carefull and remote to the subject.

1) How can either 'E' or 'I' be defined reliably in socionics when it's estimation is allowed to be derived from having both 'I' parameters as well as 'E'? The E/I always seems a compromise of having the most 'E' or 'I' which in turn ignore all lesser but still existing parameters in a character. It seems domination and binary based. I really would like to hear INTj's asnwers that one and hear them justifying socionics being able to be a good model while being based on very estimative or dominating datasets. (note that the same applies to E/I N/S T/F J/P). Being 49% introverted or something would not translate into an absolute ENTj in my oppionion. Same goes for the N/S. If this leads to crosstypes, then shouldn't socionics be less absolute in type defenitions and doesn't that make it's branches too binary or it's 'datasets' too estimative?

2) If socionics is an estimative observation, who then decides what are the absolutes everything is derived from?

3) How would you describe the limitations of socionics if it claims capable of describing my type when I'm borderline E/I S/N?