Similarity of Voynichese glyphs according to their immediate statistical environment

[Jorge_Stolfi]

Based on the observation that it is possible to generate other words, which exist in the VMS, by replacing similar shaped glyphs, it is possible to list the following rules:

"in", "iin" and "iiin" can replace each other (in - iin - iiin) 
"e", "ee" and "eee" can replace each other (e - ee - eee) 
"ee" and "ch" can replace each other (ee - ch) 
"ch" and "sh" can replace each other (ch - Sh) 
"k", "t", "p" and "f" can replace each other (k - t - p - f) 
"chk", "ckh" and "eke" can replace each other (chk - cKh - eke) 
"o" and "a" can replace each other (o - a) 
"o" and "y" can replace each other (o - y) 
"n", "r" and "m" can replace each other (n - R - m) 
"l", "r" and "m" can replace each other (l - R - m) 
"r" and "s" can replace each other (R - s) 
"s" and "d" can replace each other (s - d)

I agree with most of that, if one takes "can replace" to mean "can occupy the same word model slot as". That is, without implying phonetic or semantic equivalence, or similar character/digraph statistics.  ]

My view, from lots of statistics munching, is  that "ee" is a letter of its own that can occupy the same slots  as "ch" and "sh" (but is not necessarily equivalent to either).  And that "eee" is an "ee" modified by "e", like "che" and "she".

However, it seems quite possible that this "ee" is (always, or some of the time) just a "ch" with the ligature omitted by the scribe, or that "ch" is "ee" with a ligature added for clarity.  

This occasional omission of ligature could explain why "eke" is similar to "ckh" etc.

I don't think that "n", "r", and "m" can replace each other.  Rather, I am convinced that "m" is an abbreviation for "iiin" and perhaps other endings, that the Scribe could use when space was tight (or he was feeling lazy).  And the ending "ir" is always a scribal error for "iin", caused by them having similar shapes in sloppy handwriting. Likewise, some  (but probably not all) "ar" and "or" in word-final position may be errors for "ain" and "oin".  

Consequently, I don't think that "l" can replace "r" or "m".  But some "l" may be Scribal errors too.

Not just "s" and "d", but all four of "l", "s", "d" and "r" can replace each other in my word model.  The apparent statistical differences are the result of the r/in errors in word-final position, as per above, and also errors where a sloppy short "k" (but not a "t") was mistaken for a "d" or vice-versa.

I also view forms like "ih" , "ikh" etc. as meaningless variants of "ch" and "ckh" etc, deformed by the Scribe or by the Restorer.

And I consider double-h forms like "cthh" etc. as instances of "cthe" where the ligature was improperly extended by the Scribe.

All the best, --stolfi