A key to understand the VMS

[Torsten]

Torsten: so what you are saying is that even though Voynichese is autocopied from itself, the "itself" from which it is autocopied (i.e. that seeds the copy) contains near-universal adjacency rules which are then preserved in the autocopying?

And that these near-universal adjacency rules that are in the 'seeding' part of the text (and then preserved in the autocopied part of the text) may or may not be language but you're not interested?

No, adjacency rules are not preserved from the seeding part they are a result of the autocopying process. If a word is copied and some glyphs are replaced with similar ones the order of the glyphs stays unchanged. If you copy for instance 'chol' and replace 'ch' with 'Sh' and 'ol' with 'ar' you get 'Shar'. On first view this word looks different but the adjancency rules are still in place. If then 'Shar' is used as source for generating a new word 'chol' is still a possible outcome. See for instance 'Sheol chol Shar chol' in line f106v.P.25. 

Even if you change something while copying a word some elements of the source word remain. Therefore it is no surprise that the autocopying process results in adjacency rules.

Recursion can result in very interesting effects:
   

And that all the word variations arise not from a conscious attempt to produce randomness, but instead from the autocopyist assembling many small blocks that retain those near-universal adjacency rules?

There is no randomness. I wouldn't expect from someone before 1500 that he would be able to generate something random. In my opinion the VMS was written before 1500. This means that the scribe didn't know what randomness means. Moreover, as far as I know cognitive processes have virtually no way of generating truly random events. It would be like trying not to think of something. Therefore I didn’t find it surprising that something written in medieval times resembles more to human language then to a random text generated by a computer.

If you want to fill a page with text it is much easier to copy text then to generate new text. In this way the text in the VMS is in my eyes a medieval form of the 'Lorem ipsum' text [see You are not allowed to view links. Register or Login to view.].

So... why is it that eol is 25x more frequent than ole, again?

'ol' is more frequent as suffix then as prefix. The most frequent word with 'ol' after 'e' is 'cheol'. It is used 172 times. The most frequent word using 'e' after 'ol' is 'olkeedy'. It is used only 42 times. Since 'cheol' if more frequent then 'olkeedy' the autocopying process leads to more words related to 'cheol' then to 'olkeedy'.