A family of grammars for Voynichese

[Mauro]

I only became convinced that many final ir (note, not just any final r) are scribal errors for iin

If this is true, I would expect the ratio of occurrences of a word written with final -iin vs. the occurrences of the same word but written with -ir to be more or less constant (this assumes a simple model in which the scribe randomly miswrites an -iin as -ir with some probability). Did you check that?

Here are the frequencies of the "coda" elements of my model, as of last week:

    113.500000  0.00091 {n}     868.250000  0.00697 {m} 
   1665.500000  0.01336 {in}     40.000000  0.00032 {im} 
   3779.000000  0.03032 {iin}    15.000000  0.00012 {iim}
    159.000000  0.00128 {iiin}    1.0       0.00001 {iiim}
     
    487.750000  0.00391 {ir}
    130.500000  0.00105 {iir}
      1.0       0.00001 {iiir}

In the n series the most frequent is {iin}, whereas in the r series the most frequent is {ir}. The ratio {ir}:{iir} = ~3.7 matches neither {iin}:{iiin} = 23.6 nor {in}:{iin} = 0.44, but at least is greater than 1, like the former.

All the best, --stolfi

Not exactly the kind of test I had in mind. I made a quick check (very quick, so excuse me if I made mistakes, and with the limit that it refers to the whole text of the VMS), with the occurrences of the 8 most frequent words which end in 'aiin' vs. the corresponding 'air' word.

daiin vs. dair = 834/109 = 7.65
aiin vs. air = 529/78 = 6.78
qokaiin vs. qokair = 259/21 = 12.33
okaiin vs. okair = 152/25 = 6.08
saiin vs. sair = 129/26 = 4.96
qotaiin vs. qotair = 83/5 = 16.6
kaiin vs. kair = 63/16 = 3.93

They look to me a bit too much spreaded out to support the hypothesis that 'iin' is randomly replaced by 'ir' with some probability (but I ran no statistical tests)