[dvt]

This is very cool! When I was a (philosophy) undergrad at UCLA, I took several graduate seminars (most of them in logic), one of which covered Rhetorical Structure Theory and Kehler's Theory of Grammar. In my term paper[1], §3.1, I posit that the "Parallel Relation" (analogies would be closely related, if not even categorically fall under PR) could possibly be modeled by Euclidean distance, such that dist(A, B) ≈ 0 when A and B retain "strong" linguistic coherence.

Anyway, this is really awesome stuff and I'm happy to see more rigorous work done w.r.t. linguistic motifs.

[1] https://dvt.name/logic/prelation.pdf