By Gordon Rugg
There’s a new article about the Voynich Manuscript. It’s by Torsten Timm, and it’s on arXiv:
It’s 70 pages long; the abstract claims that: “As main result, the text generation method used will be disclosed”.
That’s a big claim.
In brief, the mechanism proposed in this article looks fairly sensible at first sight – it basically consists of a way of generating new words from a particular word, using a set of rules about what can be substituted for what. It’s low tech and simple, and it can produce something that looks like Voynichese.
However, as usual, the devil is in the detail, and I’m not convinced that this method provides a good explanation for the odd statistical details of the Voynich Manuscript. For instance, it doesn’t provide a compelling argument for why words in the first half of a line tend to be different in length from words in the second half of a line. I was also unconvinced by the explanation on page 18 for the “dialects” and the handwriting differences in the Voynich Manuscript: ‘The difference between both “languages” may only be that the scribe changed his preferences while writing the manuscript.’
This article also sits awkwardly on the fence with regard to whether the Voynich Manuscript contains only meaningless gibberish, or whether it contains meaningful material. That’s a significant issue with regard to how the text of the manuscript was actually generated, and this doesn’t fit comfortably with the claim that the article shows how the text was generated.
There’s no mention of significant previous work by previous researchers relating to the idea that the Manuscript might contain coded material concealed among gibberish padding text. In addition to my own discussion of this idea (and the problems with it) in Cryptologia, this has also been discussed and investigated in some depth by other Voynich Manuscript researchers who aren’t mentioned in Timms’ article.
So, in summary, it’s an interesting idea, and there are some sensible, interesting suggestions, but there are some major gaps in its references to previous work, and I don’t think it provides a compelling explanation for the statistical oddities that are a key feature of the Voynich Manuscript.