The text of the Voynich Manuscript must be one of:
If it's unencrypted, it should be possible to detect grammar within sentences even if the language is an unknown isolate. Based on the number of glyphs, it is probably written in an alphabetic script or abjad, so the text can be assumed to be at least partially phonetic.
If it's encrypted, I don't expect the encryption to be particularly deep, firstly because only simple encryption mechanisms were known at the time of its creation (see FIFTEENTH CENTURY CRYPTOGRAPHY), and secondly because deep encryption would render it unreadable without first going through a laborious decryption process, which would involve separating the plaintext from the illustrations. Any modern attempt at decryption would have to work on the whole text, unambiguously. This means, inter alia it should not depend on anagrams.
Not a language
I don't know anyone working along these lines, and regard it as very unlikely. A musical score might be one such remote possibility.
It's easy to say but much harder to prove that it's meaningless, as this involves proving a negative. It would still require the identification of a mechanism, available in the first half of the fifteenth century, which could generate the text. Any such mechanism would have to reproduce all the statistical properties of the text.
© Copyright Donald Fisk 2017