You say, "Your argument and list make it look like I have no access to S and F (or E and J for that matter either), and that ESFJ's have no access to I, N, T and P, precisely as Berens said, above."
But I say nothing of the kind. And when Berens says, "Presenting [the dichotomies] as either/or led people to think they could not access the other side of the dichotomy," she's engaging in the same kind of silly straw-manning.
Type is about what are often referred to as preferences, Eric. I know that, and you know that.
The MBTI-related aspects of personality that you tend to find in INTJs are the result of the fact that, on four dimensions of personality that the data suggests tend to be somewhat hardwired and relatively stable, INTJs have temperamental tugs that tend to pull them in the direction of (cause them to favor) stuff on the I, N, T and J sides of those dimensions. And not only are there multiple meaningful aspects of personality that each of those four preferences tend to contribute to, but there are also specific aspects of personality that more than one of those preferences can combine to contribute to — with the result that there are meaningful things to be said about typical NJs, and typical NTs, and typical INJs (who the function folks would call "Ni-doms"), and so on.
That dichotomy-centric perspective leads to the expectation that an INTJ and an ESFP won't have any MBTI-related aspects of personality in common, because every relevant preference contribution — from single preferences and preference combinations both — puts them on opposite sides of the applicable divide. (But that dichotomy-centric perspective most certainly does not say that INTJs have "no access" to E, S, F or P — whatever the heck that's supposed to mean — or that ESFPs have "no access" to I, N, T or J.)
By contrast, Berens' function-centric perspective, combined with her (non-Jungian) functions model, leads her to the expectation that INTJs and ESFPs will tend to exhibit similar "Cognitive Styles," because she thinks the underlying contributors to their personalities include four "cognitive functions" in common — not in the limited sense of just having "access to" those functions, but in the sense of favoring them in a way that puts INTPs and ESFJs on the other side of the applicable divides.
And I'm here to tell you that in the real world of, you know, facts and stuff, those contrary expectations can't both be true. INTJs and ESFPs either have MBTI-related personality characteristics in common (that neither shares with INTPs or ESFJs) or they don't. And that's how science works, Eric — and personality psychology is a science (albeit a "soft science"). People come up with theories that lead to contrary expectations and then studies get done that put those expectations to the test.
And, as further discussed in this long INTJforum post, we now have decades of data that provide respectable levels of support for the validity of the four MBTI dichotomies — including lots of meaningful correlations with various dichotomy combinations. And on the other hand, the so-called "cognitive functions" — which James Reynierse (in a 2009 article described in the linked post) refers to as a "category mistake" — have barely been studied. And the reason they've barely been studied is that, unlike the dichotomies, they've never been taken seriously by any significant number of academic psychologists. The third edition of the MBTI Manual was published in 1998 and, according to that Reynierse article, it cites a grand total of eight studies involving "type dynamics" (i.e., the functions model) — which Reynierse summarizes as "six studies that failed, one with a questionable interpretation, and one where contradictory evidence was offered as support." He then notes, "Type theory's claim that type dynamics is superior to the static model and the straightforward contribution of the individual preferences rests on this ephemeral empirical foundation."
If it turns out that neither you nor Berens nor anyone else can ever come up with Enhancing™ descriptions that both INTPs and ESFJs relate to (and that INTJs and ESFPs don't), and Orchestrating™ descriptions that both INTJs and ESFPs relate to (and that INTPs and ESFJs don't), then that will be a strong indication that those labels represent a "category mistake." Contrary to your latest post, the assertion that INTJs and ESFPs have no MBTI-related personality characteristics in common (that neither shares with INTPs or ESFJs) and the assertion that INTJs and ESFPs share a "Cognitive Style" that makes them different from INTPs and ESFJs can't both "be true at the same time"; and they are not just "different angles of looking at the same things."
Saying otherwise isn't being "holistic"; it's being illogical.
Links in INTJforum posts don't work if you're not a member, so here are replacements for two of the links in that long INTJforum post: