Now motivations can in part be explained through the MBTI or function analysis but they're insufficient to the task if any clarity is required. This is where you need to know the person not just the type of person.
As for the idea that people need to be understood or classified for proper understanding of human nature, this is proper poppycock as pretty much everyone quoted as understanding the human nature have been people relatively unconnected to psychology and certainly I've never heard of anyone claiming to have a system which explains people.
To me the answer is fairly obvious, function analysis is trying to be what the MBTI is not. It's trying to answer questions where as the MBTI merely provides a platform for discussions and analysis without ever trying to pin people down.
As to the argument of using function analysis because it's what the MBTI is based upon, this is a rather simplistic and naive viewpoint IMO. The MBTI is based upon surveys and averages and bunches of people looking at correlations, the result of which is not then directly related to an individuals results. Ergo you may use Ti well and a lot, that does not mean it's a preferred function. People can be trained away from their personal preferences, "educated" to act differently. There are several examples on this forum alone. Where does function analysis lead you with such people? The MBTI has in built compensation (note this is inbuilt to the training of the testers not the test itself as far as I know) for such things. My most often used case example is my brother in law who was forced to organise from an early age, trained to be ordered and that things must be kept in order. Hence he tests as an ENFJ but he's actually an ENFP. The trained examiner who saw his test noticed a correlation which I think was he tested as "kind of" J which in the experience of the examiner mean't he was actually a P because Js tend to test much more certainly in terms of J/P (something for those who are INTx or whatever to think about).
It's one thing for a bunch of people to analyse a load of test results and find that those who have Ti as their highest used/ capable function and Ne as their second tend to have common traits, it's a whole other ball park though to say that someone with Ti, Ne as their highest functions conforms to those traits. Yes the same thing runs for the MBTI and yes the types do overlap to a certain extent but saying that function analysis is any more precise or capable is kind of like saying that the chassis you got off a Ford Escort is better than a Ford Escort because it allows you a more detailed approach. Sure it does but forgive me if I see it as an element of a Ford Escort and not an entire object in it's own right. Similarly forgive me if I seem to think you're trying to reinvent the Ford Escort with your superior approach.
Anyhow the purpose of this thread was to question the use of function analysis and it's validity as a typing tool. Basically you cannot show a direct and causal link between function use and the kind of person you're looking at without going through the MBTI or engaging in a bunch of terms like hero and witch and ending up sounding like Mystic Meg on acid. Trying to get to someone's MBTI type through function analysis seems similarly fraught with problems and seems most often used by those without enough experience of type in the real world to type people using just the MBTI types. Also even if you manage to employ function analysis to reach someone's type and the journey has given you additional details about how their type works on a more component level, you're still just looking at one facet of what makes a person's nature. You still have a long way to go before you understand them and more probably than not your decision to take such a precise approach probably means you're the kind of person who won't understand anyway.