Hypervector Analysis of the Voynich Manuscript

[ReneZ]

Thanks Darrin for this interesting analysis.

I read the linked page, and it leaves me with a number of questions.

What exactly do you mean with:

Languages in non-Latin characters were machine transliterated before normalization.

If one uses Unicode, there is no longer any issue with non-Latin characters. One can still apply the language rules to translate upper case to lower case (for languages that use two cases).

The question of the Voynich MS transliteration is an important point. Now I strongly disagree with JKP's wording, because Eva *is* an alphabet. Just because it is easy to remember doesn't make it a 'mnemonic system'. It is not more a mnemonic system than Currier's alphabet is. It is one of several existing alphabets, and each of them have their own pro's and con's.

The point is that Eva (like the others) is just an approximation and we don't know how good it is, and how much this affects the results. The best we can do is try several alphabets, and see where the 'points' end up in the graph.

Is the size of the alphabet an important property for this analysis? You mention 26, which just happens to be the size of the English alphabet, but most languages have quite different alphabet sizes. Even the different transliteration alphabets for the Voynich MS have different sizes, with Eva at 25, Currier at 36 and extended Eva and v101 well over 100.

Have you looked at the 3rd and 4th component of the PCA as well (which means principal component analysis, rather than peripheral)?

Thanks in advance for any answers.