That's supposed to be easier to understand?
With this clarification, the speaker-hearer's linguistic intuition suffices to account for the traditional practice of grammarians.
OK, I'll level with you.

My replies were made up from a random selection of phrases under headings of Initiating phrases, Subject phrases, Verbal phrases and Terminating phrases. It's called creating fog. The end result is something that looks as if it could mean something, but means nothing. e.g.
Suppose, for instance, that.....a case of semigrammaticalness of a different sort.....is not quite equivalent to...... a corpus of utterance tokens upon which conformity has been defined by the paired utterance test.