Using the various digraphs as single units is certainly interesting and more 'correct'. Of course for the Voynich MS we don't know how to do this, although it is a reasonable guess that combinations like cTh are candidates.
In the HMM model of the UDHR in Abkhaz (that should identify vowels and consonants), the two groups are largely coinciding with two different halves of the PCA plot.
While the texts size of the UDHR is short, it is not *too* short, but it does suffer from what I'd call unusual vocabulary.
I have compared Latin texts of Pliny and Mattioli (16 centuries apart) and their properties almost coincide.
The Latin version of the UDHR is quite different.
In the HMM model of the UDHR in Abkhaz (that should identify vowels and consonants), the two groups are largely coinciding with two different halves of the PCA plot.
While the texts size of the UDHR is short, it is not *too* short, but it does suffer from what I'd call unusual vocabulary.
I have compared Latin texts of Pliny and Mattioli (16 centuries apart) and their properties almost coincide.
The Latin version of the UDHR is quite different.