TF-IDF values of categories

[MarcoP]

Here are some more facts to baffle you. In quire 13 20% of the text consists of just nine words. Five of these are 4o- words. Three are -c89 words. In quire 20 20% of the text is made up of sixteen words. Even if you examine separately the pages that form the two language clusters of that quire 20 you get much the same list.

While the many similar words are a peculiar feature of the Voynich ms, I don’t think that the 9 most frequent words making up 20% of the text is unusual.
According to You are not allowed to view links. Register or Login to view., the two most common English words ‘the’ and ‘of’ make up 10.5% of the corpus examined. Unless I miscounted something, in King James bible, the top 4 words cover 20.7% of the text.

the 8.09
and 6.53
of 4.38
to 1.72

I also agree that one would expect the most frequent words to be reasonably consistent across different texts. This puzzling feature makes it impossible toYou are not allowed to view links. Register or Login to view. as the frequent words that occur in all sections, which would be a very convenient attack avenue, under the assumption that Voynichese words correspond to words in a natural language.