A few days ago, I noticed a post in the forum that surprisingly matched my own observations—and that's when it got really interesting for me. The post (based on a paper by Andreas Timm, later developed further together with Schinner) looks at the Voynich vocabulary as a network. What it shows is not a series of clearly separate words, but a dense network of forms that often differ only minimally, but connect almost the entire lexicon.
At first, this confused me a little, but then I realized that it actually confirms what I myself have been observing for some time (in a much simpler form).
As I have written here several times, I assume that the text was copied from an older source that was already difficult to read. In some places, it seems quite clear to me that the writer wrote down several possible readings of the same basic form one after the other – apparently because certain glyphs were not easy to distinguish. Even if this is only partially true, it would help explain why Voynichese contains so many almost identical word forms and why variants are often very close to each other. You are not allowed to view links.
Register or
Login to view.
This approach was not unusual in the Middle Ages, although alternatives were usually separated by a special character or by "vel". This is missing here – or the character is simply no longer visible.
As I mentioned, my own tests suggest that the text functions less and less like a coherent recipe or prose text. To me, it looks more like a tabular or parametric description that works with a small number of fixed templates that are slightly varied over and over again. Such a system would naturally require far fewer different words and verbs – which is exactly what we observe in the Voynich manuscript. One possible application could be a series of similar but not identical descriptions of the plant's vascular system.
When I combine these two ideas, the results of the network become at least partially understandable to me: a highly structured source on the one hand, combined with uncertainties in the interpretation of the glyphs on the other. The result would be a text full of closely related forms and local clusters – without having to assume a classic continuous text, an encryption that massively inflates the wording, or an intentional hoax.
But maybe I'm wrong about that... ist difficult...
@Bluetoes Your example could fit this idea. The process might have gone something like this: He writes the word on the left, then thinks, “Hmm, no, a y makes more sense,” changes it, becomes uncertain again, and writes it again as a second word.
But of course, that's just a diedeledely theory.