You do realize the Meyer-Briggs typology has been abandoned in favor of more meaningful tests? At this point it's basically a scientific-ish version of astrology.
You're talking about whether the MBTI framework is correct, or verifiable, or scientific. That's interesting because the context for this whole discussion is around a rationality checklist which may not be correct, or verifiable, or scientific.
It seems like the more appropriate question when discussing these frameworks might be whether they are useful. Many people find the MBTI framework a useful way to think about themselves. Just like many people find GTD useful for being organized, or Paleo useful for choosing what to eat, or Agile useful for coordinating software development, or [system/lens/framework here] useful for [thing that people do], even though they are not demonstrably "correct" (or even superior to competing systems).
Correct and useful aren't necessarily the same thing, especially when we're discussing systems that are more of a descriptive worldview than an actual set of predictions.
It's mostly fine to talk about MBTI traits because they are for the most part just a different vocabulary for the same concepts as the current psychological standard.
The Big 5 is the current standard, and the four MBTI dimensions correlate to 4 of the Big 5. E/I is Extraversion in Big 5, N/S is Openness to Experience, and the other two are less orthogonal, but generally T/F is Agreeableness and J/P is Conscientiousness. Neuroticism in Big 5 is essentially not measured in MBTI and doesn't correlate well with any of the MBTI traits.
That comes off as a hand wave, especially considering that it's a model / typology (if you understand the role of a model vs. a theory) and is still used in research. See also Dr. Dario Nardi's neuroscience work at UCLA.
A quick search of Dr. Nardi reveals that he has a financial interest in the MBTI, being the author and/or publisher of many "educational" works on type (see, e.g., here: http://www.radiancehouse.com/psych.htm ).
I did not see any peer reviewed journal articles by him on personality in a Google Scholar author search for his name, only books and a single conference paper.
Finally, looking at the most recent CV I saw -- http://www.darionardi.com/webcv.html -- only one of the papers in the "personality" section is actually from a journal, and it is, lo and behold, a Type-centric journal. And half of the conference papers are for the Association for Psychological Type International.
It seems like the more appropriate question when discussing these frameworks might be whether they are useful. Many people find the MBTI framework a useful way to think about themselves. Just like many people find GTD useful for being organized, or Paleo useful for choosing what to eat, or Agile useful for coordinating software development, or [system/lens/framework here] useful for [thing that people do], even though they are not demonstrably "correct" (or even superior to competing systems).
Correct and useful aren't necessarily the same thing, especially when we're discussing systems that are more of a descriptive worldview than an actual set of predictions.