If you look at the evidence from a empirical/non-type-dynamic perspective, I think it's more easily explained: the "functions" are really describing the additive factors of three preferences (so Ne = E + N + P). So, when one describes the function in its "auxiliary" form, there's a mild mismatch. So, for example, an INTP prefers I + N +P, rather than the E + N + P that comprises "Ne". (Though to be fair, I think the function description are a little less weighted to the E/I preference than to the other two grouped preferences, perhaps so the descriptions can cover the "auxiliary" form).
Therefore, both INPs and ENPs will share the additive qualities of N + P, but there will be differences because the E/I preference is different. This means that there should be a pure "N+P" description what will cover much of what "Ne" is, and apply equally well to INPs and ENPs.
One can see this when one hears things like "INTPs seem almost to direct their extraverted intuition inward, being largely focused on perceiving internal, abstract possibilities" and the like.
Of course, all this implies that is the possibility for descriptions of all the preference dyads and triads. For example, the "Te" of introverts is no doubt different than that of extraverts, because of the effects of the introversion preference itself. So, if we teased out T+J, we could then compare it to I+T+J and E+T+J, and so glean more precise descriptions and better insight. Such a description of T+J would apply particularly well to ambiverts (with middling E/I preference) that have a clear T and J preference.
