Berens doesn't frame them this way, but she does compare both models to the ancient ones. In the old threads we have on the humors, they continue to enjoy relative success in the correlation I have revealed, as people occasionally still posting test tesults. Like we just had an ENFP who is Sanguine-SupinePhlegmatic (Sanguine=ENP [social]; Supine and/or Phlegmatic=NF [leadership]).
This would explain why Keirsey's model would seem to leave some out, or that certain types are more like the assigned temperament than others. SP is supposed to be Sanguine, which is very bright and expressive, and ISFP does not seem to fit that. That's because that type's social style is ISF, which is Phlegmatic. The two temperaments blend and temper each other. So if you only look at it as one temperament, then it will not appear to fit.