Notes on the Voynich Manuscript - Part 19 [1992 December 4] ----------------------------------------- Yes, I'm still here, and with the end of term approaching might even get my head above water soon! Is the Voynich manuscript written in Arabic? I once looked at this conjecture, and alas rather think not. Let me share those thoughts. My starting point was the star names, so many of which began with 'opq-'. It seemed an obvious guess that this stood for "al-". Now, that doesn't tell us anything about the language, since we too use the Arabic star names, but it seemed worth a little speculation. Such a provenance is not impossible. If it's Arabic, the place of origin is surely Muslim Spain - al-Andalus - and very likely Catalonia or the Balearic Islands. Now, that civilisation produced some major work in astrology and herbal medicine. It also produced a linguist called al-Zubaydi, who proposed a simplified form of Arabic, with most of the grammar regularised or eliminated, that in many ways anticipated Peano's "latino sine flexione". Well now... consider a small, esoteric school, who carried through al-Zabaydi's plan, designed a synthetic script and grammar, and translated a few popular texts into it, as a proof of concept. Or even, took a copy of a real Arabic book, after the illustrations had been drawn but before the text, and wrote the text instead in their new language. Is that why some folios have the pictures in the right margin? Anyway, one or more of these documents fell into the hands of James I, King of Aragon, when he conquered the territory in 1229. And one of his clerics took such a document to the first Council of Lyons, in 1245, where he gave it to that noted English scholar and Arabist, Robert Grosseteste (1175-1253), who, of course, was the tutor of Roger Bacon. You see, you can connect anything to anything, if you try hard enough! So, what about the text? Well, if 'oqp-' is "al-", then we have an obvious guess at the infamous 4: it does indeed mean 'and', and so has the value "wa-" or "w'", and '4oqp-' is read "w'al-". However, the definite article isn't always written (or pronounced) "al-". Before some consonants, it mutates; the L turns into the following consonant. So we have "al-Kitab" ('the book'), but "ad-Din" ('the faith', ie Islam), "ar-Rahman" ('the Merciful'), "an-Nur" ('the light'), and so on. In all, there are 13 such consonants in the Arabic alphabet. [Note: including, of course, 'z', hence "az-Zubaydi", but it seems most western texts don't bother with this rule.] Well, let's suppose the other gallows letters represent such duplicated consonants, so that 'olp-' stands for "ar-R", which by frequency is plausible. Unfortunately, there aren't enough gallows letters, and their frequencies are a little too low. This, of course, is a problem with any Voynich speculation - there seem to be just too few consonants to make sense, both in terms of numbers of letters and their frequencies. The point at which the entire speculation collapses, however, is the text groups. These cannot be Arabic words - they show no sign of triliteral roots, internal vowel mutations, or indeed any feature typical of Arabic. They look much more like Arabic with the prefixes and suffixes, but with the roots taken out of the middle - an eccentric and not very useful encryption system. Or, of course, an Indo-European language with initial stems and end inflexions, which our analyses keep throwing up in our faces. And, finally, why would they write from left to right? So, unless I've missed something big, this too was a dead end. Robert