A family of grammars for Voynichese

[nablator]

I don't know too: all the innumerable discussions and comparisons on different grammars (including those in this thread) serve this purpose. Using AI is a very nice idea, I actually had fancied to ask ChatGTP4 to generate for me a random sample of Voynichese and see what happens, never did it because I never subscribed to ChatGPT (I have a strong allergy for subscriptions) and because it works in a way that it could have easily generated text by simply assembling together actual Voynich snippets, which is rather trivial.

It's interesting that all LLMs tokenize the text. These tokens are not syllables. Training an LLM with a VM transliteration generates tokens similar to the chunks that a slot model produces, for example: ['f', 'a', 'chy', 's', 'ykal', 'ar', 'a', 'taiin', 'shol', 'shor', 'y', 'cth', 'r', 'es', 'yk', 'or', 'shol', 'dy', 's', 'or'].

From the partial ordering principle at work in slot sequence models it seems that Voynichese is made of ordered chunks. In my opinion these chunks were assembled by the scribes using various tricks, mostly: some spaces were removed, gallows added between words or inserted to make the words more believable: there would be many "oe", "oa" without them. So they are probably decorative, not part of the chunks generated by the slot sequence model. Ignoring them solves the problem of weird words like qko. Other red herrings probably exist. Is q simply a variant of ch? Sometimes unique words drop hints: