Prime example is treating the functions as behaviors and dividing between e/i function attitude too much, and then using these behaviors as solid indicators of type. (Lenore suggested that this was the influence of temperament theory, and temperament focuses more on "affective' behavior). Again, this leads to the problem of types not fitting because one thinks they "use" the wrong functions too much.
In Jung's conception, there were four functions, and the orientation was really distinct in itself. When I realized this, then it became easy to understand things such as the Tertiary being either/or orientation, and how the shadows work.