There's already been some progress made on the Voynich manuscript — there are two separate 'languages', that much is known. (Though they may be different hands / conventions of the same language — my guess is two different ligature conventions, since the 'gallows' are the only ones with significantly different distributions between the two hands.)

I vaguely recall hearing something about positional frequency: some letters can only appear initially, and some only finally. Tone letters? But no, Nahuatl doesn't have tone. And the word lengths aren't right — unless vowels weren't written...?

The thing has been 'decrypted' as everything from Latin to Manchu (no, really, Manchu) so this is probably as bullshit as the others — but are there South American languages with frequent reduplication and words of about the right length?

Actually, no, it's even more implausible; the vellum has been carbon-dated to the early 15th century — a century before the Codex Osuna, and before the discovery of the New World — and the paints are apparently those used in Europe at the time. Though they're at least aware of the problem:

How was the parchment, which may date to animals killed in the first half of the 15th century, used over a full century later for this manuscript?37 How did putative medieval German script on folio 166v (the so-called “Michiton Olababas page”) get integrated into this manuscript?  Was this a case of European parchment being repurposed?

There's also French month names elsewhere in it.

Relevant:

What they didn’t consider: the demonstrably 15th century vellum in play (radiocarbon dating), 15th century digit shapes (in the quiration), 15th century number forms (in the quiration), 15th century contractions (on the zodiac roundel hand) and 15th century parallel hatching (in several drawings). So, that’s evidence from the domains of codicology, palaeography, and Art History immediately consigned to their great big wastepaper basket of Not Examined Here Stuff.

Any ideas on the Michiton Oladabas page?