Well, I think these assumptions reflect some observations, for instance, why ISFj sensory subtypes resemble ISTjs ie the relative emphasis on in relation to would also relatively increase role function use -- of course maybe it's the wrong way to explain it but it does fit some things.
So that would mean that sometimes super-ego relations might seem closer than they otherwise would. For example, a producing subtype LII would be a little like an EII, while an accepting ESI would also be a little like an EII. So these two types may not seem as conflicting as typical super-ego relations would suggest.