Liaisons in the Voynich text

[quimqu]

V101 has é but I haven't checked if all of them are recorded. There are also e converted to other gallows (or gallows with a wobbly leg like the t on f23v.1) but they are not recorded in any public transliteration AFAIK.

@nablator thanks, you are right, v101 has it: f1v.10 is “@233;ol” there (é). GC uses é only 3 times, so I put them next to ZL, plus the f101r.3 one:

• f1r.24: GC é where ZL reads “far”, next word “cthol”
  
• f1v.10: GC é where ZL reads “pol”, next word “chodar”
  
• f116r.19: GC é where ZL reads the p of “porchey” (GC splits it “por,chey”)
  
• f101r.3: first word of a paragraph. Its not a normal p or t, its an odd gallows with a c in it (ZL and RF write @183, GC transcribes both, e and t), then “ol”. Next word “{c@133;h}eol”, so it starts with a c too
  

Two things I find interesting:

1. if the e (or c) had stayed, the words would start with e (ear, eol, eor) or be “col”, which doesnt exist either. Words almost never start with e (0.24%) so the fix makes sense, and in f101r.3 its a paragraph start that wants a gallows anyway.
   
2. in all four, what comes right after starts with a bench (cthol, chodar, chey, c?heol). The first stroke of a c/ch is exactly an e curve. So it looks like the scribe skipped the gallows word, started the next word with its c, noticed, and wrote the gallows on top. By chance the next word starts with a bench about 30% of the time, so 4 of 4 is <1%. Only 4 cases and I picked the rule after seeing them, so take it with salt.
   

If its right, its an eye-skip error, the kind you make when copying from something that already exists.

Your other e->gallows ones (like the wobbly t on f23v.1), does the next word start with c/ch/sh too? That would be a clean test, since I havent seen those.