James H Reynierse has been a consistent supporter of the usefulness and validity of the MBTI preference pairs, but he has been a consistent critic about assumptions in the 16 MBTI type theory.
The primacy of the MBTI preferences was understated by the MBTI organization in the 1980s in favor of a theory of 16 MBTI preference types which made assumption about ideal relationships between preferences. Independent researchers such as Costa and McCrae found that such assumptions were not valid when compared to the actual empirical data for a person. In other words theory is handy but if empirical data falsifies the theory then revisions to the theory of 16 MBTI types are needed. The MBTI empirically measured preference pairs support the Five-Factor trait model, and provide an independent check on it.
Reynierse puts the whole issue in perspective in Type Theory Revisited. Implications of Complementary Opposites for the Five-Factor Model of Personality and Organizational interventions in the Handbook of Personality at Work, Chapter 16. Routledge, 2013.