> Well, if you take a 2000 word write-up about your own type, there may be some thought-provoking generalizations either about your own type or about patterns that are consistent when people of different types interact... things which derive from what you believed about yourself but are consequences of the MBTI's generalizations about human interactions.

Right, but how many of those generalizations are actually verified? The specifics are studied to be consistent, but there are very few generalizations backed up by data.