I just realized you're taking Jungs descriptors in Chapter X too seriously - he actually notes not to do just that in Psychological Types, noting they are no more than Galtonesque portraits of the types (ie, pure forms of the types, and therefore unrealistic).
I don't really need to explain the model I prefer:
YouTube
It's new, basically in a beta-stage, so I'm not exactly running around touting as the gospel truth - but what I can say, with confidence, is that it's producing consistent results across numerous people that are learning the method.
I've got nothing against people like yourself that prefer to be entirely subjective, create your own little definitions, think you're "improving" systems etc - that's cool - I personally prefer logical external consistency - which is to say, I'm only interested in models that work for many people - so that everyone (that wants to) can learn the method, apply it, and achieve consistent results - consistent, as in, the same results as everyone else.
Note here - this is out of a possible 512 types, so the odds of actually pulling it off are statistically very low - but, I and others learning the method are doing great.
Compare this to how typology is right now. People can't even agree on S or N. That's like, 2 things, and there's no consensus, there'll be fights about it, arguments about terminology and how things should be defined - conversations like this, that derail into one person trying to "correct" another persons understandings, etc etc.
Within the model I prefer (even though I stuck to a Jungian typing in my first response here, only elaborating on this other model because you asked) - this doesn't happen, people have a concern, others just point out what it says on the checklist, or refer to class material or YT vids etc - understandably - this means to begin with, people are bound by how the creators of the method define things - some people find this difficult to accept for a number of reasons.
The most glaring one I see, is people are bound by MBTI understandings, which is absurd considering MBTI doesn't even test for functions - the method is more true to Jung than any other method (Socionics included) yet it's not forced into only a Jungian approach because - as you've noted, as I've noted (elsewhere) and as many others have noted - that stuff is like 100 years old, and is definitely out-dated - there've been significant amendments since then and there are multiple models slide right into Jungs work and in some respects, replace it as they make more sense, are still in the same spirit yet are consistently trackable.
One such example is Tony Robbins Human Needs - this method makes use of them - four of them essentially being used to understand and track the four outer-temperaments (IJ, IP, EJ, EP) - and, it works, it actually works - it's consistent, people are learning it, applying it, producing the same or similar results as others are - it's proving (to me) that it's got something real to it - some people might be like "wut, Tony Robbins? That's not Jung or Myers" but if it works, well, progress is being made, imo - similar perspective you likely have about your own understandings and definitions.
I do consider Jung an authority on some things (the functions specifically, he's still a great resource for, but you have to understand that Chapter X of Psychological Types is intentionally presenting warped descriptors) - however when new theories and models are being proven to be more "real" ie, objectively trackable (much of Jungs work is not measurable and so he's barely mentioned in modern psychology degrees which have a far more scientific bent than the days of old).
Anyway that's enough of that, if you've got any questions feel free to PM me.
The general consensus changes because the more vocal members wind up brainwashing others into believing their own subjective understandings, I don't believe for a second there is any complete shift in understandings as time rolls by.