Opinions on: line as a functional unit

[Jorge_Stolfi]

the table is part of an interactive heatmap that visualizes recurring patterns in the Voynich Manuscript ( based on frequency of occurrence ). See here: You are not allowed to view links. Register or Login to view.

Thanks!  

So your syllable sets are
First: {Ch, Che, Sh, She, ot, ok, yt, yk, qot, qok, ote, oke, ota, oka, qoke, qoka, qokee, qo, ol, Cho, Sho, Cheo, da, dai, ai, ...}
Second: { ty, ky, hy, hey, hedy, hdy, hol, eey, Chy, eol, eo, ey, eedy, ir, or, ol, ar, al, in, iin, ain, aiin, dy, ody, edy, ... }

I don't understand why you have second syllables starting with "h", since "h" is almost always attached to a previous C, S, I, CT, or CK -- and you don't have any first syllables that end with these combinations.  In fact your table lists 22 occurrences of cheo+hy -- what does this mean?

Excluding the "circles" {a,o,y}, and the initial "q", which in my model are treated separately, your syllables would be 
First: { Ch, Che, Sh, She, t, k, te, ke, kee l, d, di, i ...}
Second: { t, k, h, he, hed, hd, hl, ee, Ch, el eed, r, l, in, iin, d, ed ...}

The t,k in the second syllable come ony from ty and ky, which I would leave as part of the first syllable. Likewise the "i" in the first syllable set could be treated as part of the second syllable.  That would leave something like
First: { Ch, Che, Sh, She, t, k, te, ke, kee, l, d ...}
Second: { e, ee, ed, d, l, r, Ch, in, iin, ...}

Using the glyph classes from my table, they would reduce to
First: { X, H, D }
Second: { X, D, N }

Then this would be very distantly similar to my model [ D^d X^x H^h ] [ X^z D^w N^n ] where d,x,h,z,n are 0 or 1, and w is 0, 1,or 2.  Namely your first syllable would span the first bracket pair, my  D^d X^x H^h, and the second syllable would span the second bracket, my X^Z D^w N^n.  

But maybe there is a better scheme than both, which better accounts for words where w = 2 without allowing too many "invalid" words...

All the best, --stolfi