Modal Mixture Functional Chord Grammar

I once wrote a post called "A Chord Grammar" that included some rules for generating chord sequences with modal mixture, i.e. chords from both C major and C minor scales. In that post, I didn't describe how the borrowed chords from the minor scale functioned in the major scale. I didn't know. I've tried figuring it out today, just by vibes, with some pretty weird results. So I go back and play a bunch of sequences and try to figure out what has predominant, dominant, or tonic vibes, and keep adjusting things. Or I go back to the chord progressions from the old post and see if they make sense under my current functional assignments. Here are my best current guesses.

First, the functions from C.maj diatonic triads and tetrads:

* C.maj, C.maj7 are tonic.
* D.m, D.m7 are predominant.
* E.m, E.m7 are tonic, maybe sometimes dominant.
* F.maj, F.maj7 are predominant.
* G.maj, G.7 are dominant.
* A.m, A.m7 are tonic.
* B.dim, B.m7b5 are dominant.

Here are diatonic chords from C minor as I think they function in C major chord progressions. 

* D.dim, D.m7b5 are dominant.
* Eb.maj, Eb.maj7 are predominant.

* F.m, F.m7 are predominant.
* G.m is tonic, G.m7 are tonic after borrowed chords, but have almost no function after the chords of the home key.
* Ab.maj, Ab.maj7 are predominant.
* Bb.maj, Bb.7 are dominant.

There are lots of weird things in here, but that's how I hear it. Two more weird things that modify the previous assignments.

1. Sometimes it feels to me like F.maj is tonic in progressions with modal borrowings. 

2. The chord A.m does not feel tonic after Bb.7. It's a weaker as a tonic after several of the modal borrowings, but that one is especially bad.

Also I thought F.m and F.m7 were dominant for a long time and still feel uncertain about those. Like, 

    [F.maj -> F.m -> C.maj]

feels really strong, but maybe that's because F.m and F.maj are both predominant, and we just have a plagal cadence with a predominant prolongation

    [PreD -> PreD -> T]

Or maybe F.m is dominant, and

    [F.maj -> F.m -> G.7 -> C.7],

which also sounds good, has a dominant prolongation

    [PreD -> D -> D -> T]

I just don't know. Going a little crazy. Maybe F.m6 sounds like an F.dim7 passing chord. Which is enharmonic in 12-TET with D.dim7. Which is like a D.7. Which is the secondary dominant of G. Which is ... *wanders off mumbling*.

I hesistate to codify this any further, because my functional assignments keep changing, but let's at least lump things together differently by function rather than chord root for a different perspective.

* C.maj, C.maj7 are tonic.
* E.m, E.m7 are tonic, maybe sometimes dominant.
* G.m, G.m7 are tonic after borrowed chords.
* A.m, A.m7 are tonic, but can't be be preceded by modal mixture chords. If we can assume that a chord with a 7ths degree can keep its function when you remove the 7th degree, then we can write this more succinctly as:

Predominant: [Ab.maj7, D.m7, Eb.maj7, F.m7, F.maj7]

Dominant: [D.m7b5, G.7, B.m7b5, Bb.7]

Tonic: [C.maj7, E.m7, G.m7, A.m7]

with caveats that E.m7 and F.maj7 might have dual functions, and A.m7 and G.m7 have slightly restricted function.

In a lot of my old modal mixture chord progressions, I had things things that went both [PreD -> D -> T] and [D -> PreD -> T]. And that second progression template sounds fine, you can totally play [G -> F -> C] without crying, but somehow I'd gotten under the misapprehension that chord progressions were only supposed to move in the first of those two manners.

So you can't go "Hm, yes, chord Q sounds good after chord M, so Q is  probably further along in the [PreD -> D -> T] line". Things can go anywhere and you just have to get a vibe for which progressions sound more like 4-5-1s or 5-4-1s or 6-2-5-1s or other things. Or maybe you can also take some hints by interlacing borrowed chords with old chords of known function. But it's still mostly vibes.