[Article] Orlov YN - Language recognition methods and Voynich Manuscript analysis

[RobGea]

So if i got this right, this guy uses 4 independant statistical methods.
(VMS = Voynich manuscript)

1. Using the Logarithmic symbol distribution model developed by S.M. Gusein-Zade:
  - an argument is given in favor of the fact that VMS is written in some language without the use of vowels.

2. Using the Hurst exponent,
  - shows that, the manuscript contains entries in various languages.

3. Taking the following assumptions:
a1. The manuscript is a bilingual text with a common alphabet.
a2. Vowels were removed from the text before recoding.
a3. Recoding consisted in the unambiguous replacement of a letter with a symbol.
a4. Spaces in text are not considered characters.

and doing some statistical analysis of frequencies in modern texts gives us:
  - the statistical hypothesis that [the VMS was written in two languages but in the same alphabet] can be considered quite acceptable.

4. Use of spectral portraits of bigram matrices.
  - points to text without vowels
  - points to more than one language

Overall these 4 independent statistics indicating that the interpretation of the VMS as a composite manuscript is quite admissible.
 - 'composite' here means (i think) the VMS text has no vowels and uses 2 (or maybe more) different languages.