[split] Why the VMS text is meaningful or meaningless?

[DonaldFisk]

OK, I guess I need to read your work in detail (which I haven't time to do yet).

But in a century, no one has found any convincing evidence of meaningful information encoded in the text.

That's not true. I won't mention the work by Montemurro & Zanette, because, as I noted, you suggest that you were able to reproduce the like results with the random text (however, if one works onwards from the pre-analyzed statistical properties of the text it might be trivial that s/he finds co-occurrences to sustain across the text). But, apart from that, there is quite a number of tiny indications. My favourite one is that two most frequent "Voynich stars" (labeled objects of f68r1 and f68r2) - otol and odaiin are both mentioned in You are not allowed to view links. Register or Login to view. and both in the same paragraph. I wonder what would be the probability for that in a random meaningless text. Another one is that You are not allowed to view links. Register or Login to view. (the supposed Dracaena - a plant that was largely known in Europe by that time only by hearsay, since Europeans did not yet frequent the Canaries or the Socotra back then, in other words  - definitely a rare plant from the perspective of the VMS author, if a European he be) is the only (!) botanical folio not containing any occurrence of a label vord.

There are also clear distribution shifts of some vords to certain sections - such as balneo or recipe. (I can only wonder why this fruitful field is largely unexplored still). Not sure if this would be a characteristic of a meaningless text generated by picking cards.

In a 240 page manuscript of meaningless text, you're bound to find those sort of co-occurrences somewhere.   Also, the plants are poorly drawn and few have been positively identified.

Yes, it is a characteristic of a meaningless text generated by picking cards.   I proposed that each section has its own state transition table.   The probabilities contained in them can't be determined without reverse engineering.   This also explains Montemurro & Zanette's finding.   Their word associations  (figure 2 in their paper) are because the same table was used.   Their section associations (figure 4) are because similar tables were used.