A new day breaks here in the suburbs, bringing with it birdsong, A-road traffic noise, and yet another Voynich theory to bang my head against.
On paper, Professor Stephen Bax certainly has the combination of big brain, linguistic experience and personal ambition that you’d think would be needed to crack open the Voynich Manuscript’s crab-like shell. But… then again, so did poor old Professor William Romaine Newbold; and his Voynich non-decryption ended up enraging Charles Singer so much (justifiably, it has to be said) that he was still angry thirty years later.
All the same, Bax believes that he has tentatively identified a number of words in the Voynich Manuscript, and has posted a 62-page PDF on his website describing his findings. His initial press release has been picked up by BBC News, the Bedfordshire on Sunday, and the irrepressible Daily Grail amongst many others. He has a lecture arranged for 25th February 2014 in Luton (if you happen to be nearby and interested), and is even planning a small Voynich conference in London in June 2014 to try to get other academics involved in his Voynich research programme.
Yet as Rene Zandbergen likes to point out, the most difficult thing about Voynich research is developing chains of reasoning while avoiding big mistakes. And while I hate to be the one to unplug the sound system just as it’s starting to really get the party started, I’m quite certain that every single one of Stephen Bax’s conclusions to date have been built upon a long sequence of easily demonstrable mistakes.
In fact, even though he is trying to use a sensible sounding methodology to elicit his results, I can’t think of a single piece of Voynich Manuscript evidence or secondary historical evidence he uses that I’d agree is a sound starting point: and I’m not convinced that any of his conclusions could be right either. I’ll go through a whole load of points, you’ll see what I mean soon enough.
1. “Initial Words On Herbal Pages Should Be Names”. Errrm…
As Bax rightly points out, you might reasonably expect the unique-looking first word on each of the Voynich Manuscript’s herbal pages to be the name of the plant depicted on that page, because that is indeed how many medieval herbals were laid out. This is not a new observation or idea: Leonell Strong assumed this as part of his Voynichese decryption in the 1940s (he thought the plaintext was written in English, but enciphered using a curious repeating offset into a local substitution alphabet).
But there’s an immediate problem: almost all the Voynich’s Herbal A pages start with one of the four gallows letters: EVA ‘p’ (53 times), EVA ‘t’ (24 times), EVA ‘k’ (21 times) or EVA ‘f’ (10 times). Which for simple substitution ciphers, broadly as John Tiltman pointed out roughly 50 years ago, would mean that the name of pretty much every plant in the Herbal section must start with one of three or four letters. Which would be nonsensical. (Leonell Strong was fine with this, because he thought the cipher scrambled all that stuff up a little: basically, he didn’t think it was a straightforward language.)
Yet Bax persists, and asserts that all of these gallows glyphs simultaneously map to plaintext C or K (in order to keep his ‘oror’ mapping intact, see  below), and as a result almost all of the plant names he considers start with the letter C – Centaurea, Cotton, Kaur, Crocus, etc. I’m sorry, but this whole notion is directly contradicted by the immediate statistical evidence. This isn’t something to build on, it’s something to abandon and leave far behind while you find some genuinely useful historical evidence to work with.
2. Bax’s proposed Voynichese alphabet has three letter R’s
This too flies in the face of supposed common sense. The Voynich Manuscript has a limited and compact alphabet, with roughly 18-22 characters occurring with particular frequency: and yet Bax concludes from his multi-language linguistic analysis that three of these (EVA r, EVA m, and EVA n) encipher the letter ‘R’. Come on: this is surely close to as unsystematic a system as could be constructed, a giant Red Flag of Non-Believability being waved in front of his train of reasoning.
3. The Voynichese word “oror” = the Hebrew word “arar”, meaning ‘juniper’
f15v not only has “oror” on, it has “oror or” and “or or oro r” immediately above each other on the first two lines. Did Bax not notice this when he picked this out? This is terribly selective and unconvincing. Moreover, “arar” itself is twice as common in the Voynich Manuscript than “oror”: while Bax himself points out good reasons why it shouldn’t be “oror”.
So… why does he persist with “oror” == juniper? “oror” appears throughout the Voynich Manuscript, while “or” appears extraordinarily frequently. This just seems a hopeful (and unsystematic) stab in the dark in exactly the wrong kind of way.
4. Bax thinks that EVA “kydain” = ‘centaur’ – but has he not noticed “dain” everywhere?
Now this is just ridiculous. One of the genuine mysteries of the Voynich Manuscript is the repeated presence of what look extraordinarily like medieval page references (EVA “aiin”): and here’s one apparently embedded in a word right at the top of f2r. So is there any real chance this also happens to encipher “kentaur” in the way he thinks? No, none whatsoever, I think.
5. “doary” = Taurus. Oh, really?
The reason people have in the past suspected the label by the “Pleiades”-like group on f68r might be “Taurus” was because of the late-medieval “-9” style Tironian nota at its end, preceded by a letter that looks like “r”. But both those correspondences remain a bit of a stretch, and so this seems basically unworkable in the way he hopes.
6. Reading EVA “keerodal” as “coriander”.
Ask people who have been working with the Voynich Manuscript’s “Voynichese” language for a few years and they’ll probably tell you (as I’m saying now) that this word is almost certainly a copying error by the Voynich Manuscript’s scribe. It is extremely rare to see “eer” (while “ar” and “or” are both extremely common), so I’m confident that this should instead have been written “karodal”, which closely matches how the Voynich Manuscript’s “labelese” often parses out in pairs, i.e. “k.ar.od.al”. Hence I have practically zero faith that this word could be a natural language version of “coriander” in the way Bax suspects – he has misparsed and miscategorised it.
7. Relying on Edith Sherwood’s hopeful plant identifications.
Oh, come on. Edith tries hard to do her thing, but remember that we’ve had real herbal authorities (such as the fantastic Karen Reeds) look closely at the Voynich’s herbal drawings, and they haven’t seen even 10% of what Edith Sherwood thinks she has seen.
So, in summary: of the nine words Bax claims (in his Appendix 1) to have identified, I disagree with the evidence, reasoning, and linguistic rationale for every single one. I am also sure that his letter assignments are fatally flawed. Contrary to the title of his paper, I honestly don’t believe that through his efforts he has yet identified a single “plausible” word in the Voynich Manuscript.
For me, this isn’t even a matter for Ockham’s blessed Razor: to be even remotely workable, a hypothesis needs to have a single example of evidence that chimes with it in a way that can actually be seen to work. And on the above showing of evidence, what he has presented so far is not yet a workable hypothesis in any obvious way, sorry.