incidental post – An outstanding problem.

As readers may know, there’s been a lot of talk about ChatGPT and its approach to problem-solving. Some examples in relation to the Voynich manuscript have been raised at Nick Pelling’s ciphermysteries blog fairly recently.

I don’t know whether ChatGPT has been fed the Voynich glyphs, or only one of the several and fairly arbitrary assignments of the glyphs to other systems but here’s a problem some readers might like to consider, with or without AI.


If, as I’ve read this drawing, the two flanking figures represent the primal pair (and possibly as anabibazontes ” – promoted/elevated/prominent ones”), and if the main part of the diagram represents a 5-elements system, then why should the diagram provide 8 labels and not 7?

There’s also the really obvious problem

Problem 2:

Is it possible to link the Voynichese labels (minus the supporting pair) to a list of elements in any recorded script and language known before 1440? How many pre-modern scripts and languages does ChatGPT know?


I also wonder whether those working on Voynichese have an over-optimistic expectation that the written text will observe a standardised orthography and a literary rather than vernacular approach to language. Other factors, too, might affect what is written, such as conventional elipses For example..

Suppose someone were to posit the inscription on the left (below) might have for its final three glyphs the ancient Greek word for a hearth-fire – “pur”. (A researcher named Ruby Novacna has in fact been following up a Greek-language possibility ever since 2012, though she hasn’t posted recently. Her word-list was published in 2022).

So, you might argue for the label on the left (above) that last three glyphs suggested the Greek word ‘Pyr/Pur’ .. but how to explain the first glyph in that vord?

You might posit an addition of the definite article to emphasise this is the element fire, not any old individiual fire, but the definite article for pur in modern Greek is a feminine, and in ancient and classical Greek was the neuter τό. An example is found in Plutarch’s biography of Numa where he refers to the deathless/eternal [hearth-] fire of Vesta/Hestia.

τὸ π. τὸ ἀθάνατον

Another possibility is badly spelled, ungrammatical Greek which is certainly not unknown (parts of the the Periplus of the Erythraean Sea are so un-classical as to be incomprehensible).

But I can’t help wondering how an AI or an amateur could have enough under their belt to recognise such departures from a standardised use of language, to adjust for such influences as local and dialect forms’ being literally rendered, or our text copying that of a semi-literate/archaic original?

It is not hard to envision still more unhappy factors – such as a foreigner’s attempting to write a language which they’d never learned to speak, or which they had learned to speak as best they could, but never learned to write using formal orthography and grammar. I do wonder whether the cryptologists have expected too much from the written text.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.