I'm with the school of thought that we don't use one function in isolation when evaluating a situation, that parsing out which function is the culprit can get tricky.
Probably so, unless you actually speak with the person and find out what internal values are the reasoning behind whatever opinion.
You're right that an S-type person could have better N than an N-type. A brilliant ESTP is probably better at all of his functions than a mentally retarded ENTP.
But again, look at the posts on this thread. It's no coincidence that lots of Ns have a definite communication gap with Ss in general. This doesn't mean it's impossible for an S person to exercise brilliant iNtuition, just that most of them don't exercise on it a high enough level to relate well to strongly N-type people, and vice versa. Ns just feel left out a lot of the time because we are less common than Ss, not because we think our preferred method of data-gathering is inherently better than theirs.
Like I was saying earlier, it's all about averages. The feelings of being ostracized, not fitting in, etc. in the real world might very well be due to S vs. N differences, even if some Ss have better iNtuition than some Ns. I'm sure the inverse is also true because intelligence is a variable here, too.
Even without labelling specific functions as the culprits behind particular situations, we can still get enough information so as to be useful by directly interviewing a person regarding his functional preferences. We may not be able to say, "He's acting this way in situation x because of function y", but we can say with a reasonable amount of certainty that, given appropriate self-descriptions, he tends to prefer either Sensing or iNtuition more often than the other...and that's all MBTI really seeks to do.
It's four independent variables about personal preferences in data gathering and decision making. It doesn't need more proof than that because, as I said before, all it really states on its own is the existence of those four preference scales.