I picked up on this a few weeks ago, but thought that the combination of Google Translate and my screen-weary eyes was somehow deceiving me – for who on earth would be interested in going along to a week-long Voynich summer camp in a sparsely-furnished flat in Budapest [site in Hungarian]? The site had to be some kind of prop for an Hungarian RPG, surely?

However, it turns out that (despite the slightly jarring pen-name “Hartree-Fock Cares”) this is no joke, and is actually starting very shortly (Monday 24th August 2009). After I left a brief comment on the site just now, Miklos Vincze enthusiastically replied on behalf of the “organizing committee”:-

About the camp: This is an amateaur, self-organizing little summer school, with a few enthusiastic MA, MSc and Grad school students, and even postgrads – linguists, coders, physicists, who decided to spend a week with some research on this manuscript, just for the fun of it. For this purpose we rented a flat right in the middle of Budapest, which we will occupy permanently for a week, starting with the evening of August 24. We are planning to do some projects based mostly on word structures and on various properties of word distributions in the text. Also we would like to identify some of the plants as well, with the help of one of Hungary’s leading herbarists. I myself will try to expand Landini’s work on the autocorrelation patterns in the manuscript, and compare it with different types of medieval texts. And we will work on a plenty of projects like this in our little summer school.

Sadly, all the cheap EasyJet seats between London and Budapest were snapped up long ago, so it would be a bit pricy to fly over for the day (however tempting, particularly as I like Budapest so much): but I’ve offered to do one or two Instant Messaging Voynich Q&A sessions for them during the week, as that seems like a nice compromise. As Barry Norman used to say, “and why not?

If any other Voynich experts reading this would like to donate an hour of their time for some similar IM-based Voynichological Q&A, I’m sure our Hungarian chums would be similarly delighted at the offer. Who knows, we might also even find some way of squeezing in some international Voynich IM’ing at the pub meet on Sunday. 🙂

Though I have a good-sized review of James Amelang’s fascinating book “The Flight of Icarus” in the pipeline, I couldn’t resist posting about one tiny Provençal / Occitan item that popped up there…

One of the many mysteries sustaining the Voynich Manuscript mythology is something which you’d have thought would be easy to sort out: a set of month names apparently added by a later owner to what VMs researchers call the “zodiac pages”. Here’s a set of annotated images I made of them a few years ago (which I ought to move over to this site soon), together with an explanation of why I think these are written in some form of Occitan.

When in 1997 Jorge Stolfi had much the same thought, he asked some Occitan researchers if they had any 15th-16th century texts with month names in: even though there are almost no such extant texts (because official documents were all written in French or Latin, while Occitan itself was in sharp decline), the closest match produced by this was from 15th century Toulon. Hence, for years I’ve wondered… what document was that, then?

Well… it looks like it’s listed in James Amelang’s book (p.281) – the bourgeous farmer Jaume Deydier from Ollioules (Var) near Toulon, who wrote his livre de raison (family chronicle) between 1477 and 1521. Helpfully, Amelang lists some page references (40-42, 230, 250-251) in Charles de Ribbe’s (1879) Les familles et la société en France avant la Révolution. Deydier’s first entry begins: “En nom de Nostre Senhor Dieu Jésus-Christ, et de la siena gloriosa Mayre, et de la sancta Cori ce-restial de Paradis, invocant loqual in tota bona et perfiecha obra si deu invocar, car del processis tout ben, nobilitat et profiech, Estament de mi, Jaume Deydier, natiff de Tholon, aras abitant en aquest present luoc d’Olliol.

Digging deeper, de Ribbe’s (1898) La société provençale à la fin du moyen âge: d’après des documents inédits‎ has more on Deydier. For example, pp.453-454 has an extract from Deydier’s 1521 will, beginning: “Per memoria als successors de mi Jaume Deydier, expressamen à Jacques, mon obeyssant fils.

Rather more recently than de Ribbe, there’s a 54-page book/article by Paul (or Gustave?) Roux titled Le Livre de raison de Jaume Deydier: un document d’une grande importance pour la Provence (1983), which appears to be an offprint from ‎Bulletin de la société des amis du vieux Toulon et de sa région, n°105‎.

All the same, you should be aware that there is a sizeable (if somewhat isolated?) French-language literature on the whole livre de raison genre, of which Deydier’s chronicle is merely one example. For example, if you go through this 2002 paper by Jean Tricard, you’ll find a decent-length literature list (Note 1). Tricard also comments (Note 5) that although de Ribbe launched the study of this genre, this was “pas toujours avec la plus grande rigueur scientifique d’ailleurs“, which I’m sure you can translate for yourself. 🙂

So – finally – back to the month names. Throughout all the fragmentary quotations from Deydier on the web, the only Provençal / Occitan month name I’ve found mentioned is “Septembre” (which is good, but which we knew anyway): however, without any idea of Deydier’s orthography etc, it’s a bit hard to tie this down any further.

Incidentally, Amelang cautions that many early modern popular / artisan historical autobiographical texts (the subject of his book) suffer both from a lack of a critical edition and from ‘selective’ (if not blatantly misleading) early translations: and so it is hard to be sure what we would find were we to look at Deydier’s chronicle for ourselves. Hence you might reasonably ask: is there a facsimile edition (or scans) out there of Deydier’s livre? Almost certainly not. Do I even have any idea of which library or collection holds Deydier’s original ms yet? No.

Even so, knowing what we don’t know is probably a (tiny) step in broadly the right direction.

Today’s New Scientist brings with it the vastly uninformative claim that…

“A STATISTICAL method that picks out the most significant words in a book could help scholars decode ancient texts like the Voynich manuscript – or even messages from aliens.”

True, it “could” – but you have to say that the overwhelming probability is that it (just like most other methods before it) will fall flat on its hairy statistical arse when applied to truly difficult things. And what are the real chances we’d be able to decode a putative alien language based purely on some statistician’s semantic say-so? Practically nil, I confidently predict: come on, people – decode dolphin language first and we’ll be convinced you’re onto something. (Hint: it has 200 words for “tasty” and 600 words for “fish”).

Honestly, statisticians like these simply don’t ‘get’ the Voynich Manuscript at all – they misread the EVA transcription (for example) as a regular, smooth surface for analysis, while completely failing to see that the VMs is an object that has been constructed to mislead, not to communicate.

As Alfred Korzybski noted long ago (but few now bother to consider), “the map is not the territory“. *sigh*

Here’s a 2009 Voynich theory for me to add to the big list, courtesy of George Hoschel Jr (of California?). George uploaded his solution as a jpeg’ed scan of his transliteration notes (1) followed by four annotated versions of parts of f80v (2, 3, 4, 5): the basic idea is that Voynichese enciphers what he calls “Old Latin”.

Just so you know, George reads Voynichese as a combination of simple letters (a = EVA a , e = EVA e, o/u = EVA o, m = EVA iii, n = EVA ii, z = EVA q, etc) with a lot of plausible-looking specific abbreviations (er = EVA ch, fr = EVA sh, eri = EVA sh [but with the plume offset to the right], eric = EVA she [which George reads as short for ericius, “hedgehog”], and so forth).

Put it all together, and you get a translation of f80v that reads like:

WINE STRONG DRY TO MAGPIE
SAVED CRUMBLED DRIED TO HOOPOE KIDNEY
TO BOILED KIDNEY FRIED / IN
GOAT VEGETABLE BREAD USING
USING HEDGEHOG KIDNEY TO BREAD

In some ways, to me this reads like “Ready Steady Cook” as directed by Tim Burton – “right, teams, you have 30 minutes to produce a restaurant-quality meal using only hoopoe kidney, goat, vegetables, bread, and strong wine.” Feel free to interpret this in quite a different way!

Here’s a Voynich Manuscript short story to sustain you through those long dark Northern hemisphere summer months (you know, that time when you have lots of things to do outside that don’t involve endlessly surfing the ‘net).

It’s called “I Am Darknesse” by Jez Thorpe, and is a fairly enjoyable stab at a drug-tastic horror-trip take on the VMs’ dark secrets. Sure, his characters are a bit cartoony and thin: but it’s nicely written, and (apart from a  “Frank Newbold” gaffe) manages to get the VMs side of things pretty much straight.

Just so you know, Jez is 36, married, lives near Cambridge, has taken at least one Creative Writing course, and recently found himself unemployed.

I mentioned a few days ago that Google Trends showed a huge surge in searches for “voynich” triggered by the xkcd webcomic’s Voynich theory gag. While I guessed that the spike was over 50x baseline, I wondered to myself whether there was a reliable way of accessing accurate hit statistics on Wikipedia pages, because that would be a pretty nice graph to see.

Well… it turns out (thanks to a post today from the Feral Graphing blog) that there is precisely such a thing: Wikirank. This shows that the Wikipedia VMs page got 68631 hits on 5th June 2009, compared to a normal daily average of (say) 1000 or so, so a roughly 68x spike.

OK, it’s not hugely important, but it’s a nice resource to know about. 🙂

One of Google’s more interesting experimental engines is Google Trends: this aggregates data on keyword searches, to let you compare the relative popularity of different keywords over time: for example, “Paris Hilton” and “Star Wars” are (Google-wise) just about as popular as each other. From the graph, you can see that interest in Star Wars spiked up in May 2005, which Google guesses (correctly) was from the Star Wars film opening: while the Paris Hilton volume spiked somewhat when she left jail in June 2007. This reveals other non-obvious search aspects, such as the apparent cargo cult worship of Ms Hilton in Indonesia and Mexico. 🙂

But I digress.

For the graph for “Voynich”, Google Trends’ algorithms gamely suggest to the SciAm 2008 online re-release of Gordon Rugg’s 2004 article (marked “[A]”) as a possible cause of Voynich interest (but this is plainly wrong). The twin peaks actually correspond (a) to a surge in Voynich interest in France caused by Jean-Claude Gawsewitch’s (2005) “Le Code Voynich”, and (b) to the (2009) Voynich gag in webcomic XKCD.

google-trends-voynich

If we look solely at the June 2009 numbers, the scale of the XKCD peak is even clearer: Google search traffic for “Voynich” apparently spiked by more than 50x over baseline traffic levels. Whoosh.

google-trends-voynich-June-2009

This massive XKCD spike is what lies behind the battle raging in the Wikipedia Voynich Manuscript Talk page. On the one hand, you have Wikipedia editors who think the Voynich page is basically OK (yes, there used to be a section on VMs in popular culture, but it got culled over a year ago) – and on the other, you have an army of vociferous XKCD fans who think that there should be at least some mention of XKCD squeezed in there, surely?

I think it’s important to point out that neither side is entirely blameless in this dispute. Wikipedia editors deliberately use its neutral voice and juxtaposing header templates as a way of fusing (achingly) high and (shudderingly) low cultural references together under a banner of supposed universality: though this syncretism helps to differentiate it (as a project) from traditional encyclopaedias, nobody involved is quite sure whether ultimately to privilege high or low culture – both are important at different times and for different reasons.

Similarly, the XKCD fanbase (which seems centred on Seattle, according to Google Trends?) has done itself no favours by the large number of inane troll-like edits attempted on the Voynich page. And none of them has so far really explained (in the Discussion page) why it is they collectively feel the urge to stitch XKCD’s web of cultural referentiality into Wikipedia’s fabric. From the outside, it certainly looks like a kind of drive-past L337 grafitti being daubed on Wikipedia’s walls: if there is a genuine point to the whole activity, I’ve yet to find it.

Ultimately, I suspect that the basic problem is that there is no consistently useful dividing line between high and low culture: when you have Wikipedia pages on Immanuel Kant and Brian Cant (and every silly cant inbetween), who’s to say where to draw it?

Finally, a brief Wikipedia aside. A fair while ago, someone (I’m pretty sure it was “Syzygy”, Elmar Vogt’s Wikipedia editor nom-de-plume) very kindly added a mention of my ‘Averlino’ theory to the Wikipedia Voynich page. I’m pleased that it is mentioned there, because – unlike a lot of theories – I did try hard to produce a working hypothesis consistent with the facts, rather than blatantly defying them (which seems to be the norm some days, sadly). Even if you happen to disagree with it, it does at least have the merit of pointing towards a sensible template: I’m quite sure that, if not Averlino, the real author will turn out to be remarkably similar.

Yet Elmar’s description of my theory wasn’t hugely accurate: and so I thought I ought to take the opportunity to correct and update the final two sentences. Just in case anyone is tempted to revert the changes, here are links to the research I’ve published since “The Curse of the Voynich” to back it up:-

If Pelling is right, then the manuscript is enciphered with an extremely convoluted cascade of methods, mixed together to make the resulting cipher text appear to have the properties of an unknown medieval language (such as consonant-vowel pairing, folio references, etc).

I discussed this archaic language covertext in more detail in this recent blog post. And also:-

He claims most of the non-zodiac marginalia were originally added by the document’s author(s), but have ended up unreadable because of incorrectly-guessed alterations superimposed by multiple later owners.

Recent posts on the mystery of the VMs’ unreadable marginalia (for why should they be unreadable, given that marginalia are normally added to explain or remind?) are here and here: but most of the discussion is still relatively unchanged since The Curse.

The story as it appeared in the Guardian (thanks for the tip, Charles!) is refreshingly simple, sans any Dan Brown-esque cliché. Even though the oldest known book-format Bible ended up scattered across the globe, with sections in London, Leningrad, Leipzig and Sinai (the fragments most recently found turned up in Sinai only in 1975), a determined team of scholars decided to bring them all together to form a single, online publication. There was even a conference on it at the British Library this week.

International collaboration aside, the way the data is presented on the website has many nice aspects, such as the fundamentally codicological mindset (the RGB images and raking angle images are just a click away from each other, though it’s a shame that you can’t switch from one to the other while under magnification), the parallel transcription, etc.

For the Voynich Manuscript, broadly similar things have been suggested for years: in fact, I’d say almost every Voynich expert has secretly wished for an Encyclopaedia Voynichiana, a book comprising not only good quality scans of each page, but also a transcription, detailed commentary on that page’s drawings and colours, unusual Voynichese word-patterns, contact transfers, and marginalia. The problem, of course, is that even if such a book could actually be written, it would be almost certainly be 3,000-pages long, which is longer even than Jonathan Strange and Mr Norrell (and with almost as many footnotes).

The only reason that this kind of project is even remotely conceivable is because of the effort and resources the Beinecke Rare Book & Manuscript Library put into scanning the VMs: but to a very great degree over the last couple of years, I’ve come to realize the limits of this knowledge – that while it has enabled us to document the two-dimensional surface of the Voynich Manuscript (almost to the point of fetishizing it!), it has simultaneously blocked us from its inner dimensions, the codicological ‘superstrings’ coiled up within each mark.

Some days, I look at things like the Codex Sinaiticus and think: how wonderfully transparent it would be to work on an object of knowledge that was so gloriously shallow. But then I inevitably return to the sheer mountain face of the VMs, advancing one fragile epistemological fingerhold at a time: perhaps the climber is defined by the hill. 🙂

An intriguing email just arrived on Cipher Mysteries’ virtual doormat: recent blog subscriber Anna Castriota (thanks for writing, Anna!) asks whether I think there is any sign of filigree in the Voynich Manuscript.

Of course, as per normal with the VMs, the answer is a “tentative maybe”. There are good grounds for believing that its author had been exposed to an eclectic range of artistic influences that we can glimpse expressed in different ways, so it would not be a huge surprise if filigree was also on this list. But… was it?

Technically, filigree is formed from twisted gold or silver wires: and if you go looking for its characteristic shape in the VMs, I think you might just about note a resemblance in the exterior lines of the sun face and moon face “calendar” pages.

f67r1-filigree

Voynich Manuscript, f67r1 (“moon face calendar”), detail

f68v1-filigree

Voynich Manuscript, f68v1 (“sun face calendar”), detail

Of couse, if you were to point out that this is a pretty tenuous comparison, I’d have to agree. 😉 

However… I suspect it is instead rather more useful to hunt through the VMs looking for the spirit of Quattrocento filigree decoration, wherein goldsmiths repetitively filled out basic designs with dense filigree twists. If I’ve got the history right, this was a mid-Quattrocento precursor to what Warburg called the horror vacui, the abhorrence of emptiness – a fashion that crept into decorative artworks after the 1480 rediscovery of Nero’s Golden House (Domus Aurea) with its numerous busy images – according to which artists strove to fill every inch of decorated surface with a mass of buzzing details. An accessible source on these ‘grotesques’ is Chapter 7 of Joscelyn Godwin’s “The Pagan Dream of the Renaissance”: but (unless you happen to know better) I happen to think filigree goldsmiths basically got there first.

As far as the Voynich Manuscript goes, I think a relevant piece of forensic inference is that there are good codicological grounds for concluding that many of its later (more sophisticated) drawings were executed in two basic passes: (1) a primary layout pass (probably expressing the basic idea) , and (2) a secondary decorative pass (probably concealing that basic idea, for whatever reason). So, our research question then becomes whether we can detect this filigree design spirit at play in the Voynich Manuscript’s more decorated pages…

And you know, on balance, I think that we probably can: the key example I would give is the pair of “magic circle” pages. These comprise two circular diagrams on adjacent pages, one with a sun shape in the centre (f85r2) the other with a moon (f86v4), but both surrounded by four human figures at 90 degrees to each other.

two-magic-circles-small

What, you can only see the four characters on the left-hand magic circle? Well… I’ll come to that shortly.

What is particularly curious about these is that the second “moon” magic circle drawing has been obscured by layer upon layer of dense decoration, such that the four figures around the centre (along with a series of other details) are barely visible. That is, while the left one is completely undecorated, the right one is abundantly overdecorated to the point of total coverage.

 second-magic-circle-small

In the immortal words of Rolf Harris CBE, “Do you know what it is yet?” If you’re still having trouble seeing this, here is that same magic circle one last time with all the concealing decorative stuff removed (though fairly inexpertly, and at low resolution):-

 second-magic-circle-stripped-bare

And no, I too have no idea what this denotes or means. But I am pretty sure (from the different inks that were used) that this diagram was drawn in multiple stages, with something not dissimilar to the stripped-out version being the result of the first (expressive) drawing phase, and the preceding image being the result of the second (concealing) phase.

Whatever was originally depicted here, it is hard not to conclude that the author wanted to obscure the structure of this particular page, and so filled in every available space with a dense thicket of misleading detail. Ultimately, this reminds me very strongly of the repetitive decorative nature of filigree decoration – which is why I pointed towards a “tentative maybe” right at the top of the page. 🙂

You can also see this kind of busy, space-filling activity at play on the nine-rosette fold-out page (which is actually on the reverse side of the same folio as the magic circles pages). Here again you can see what appears to be the initial expressive phase (in lighter ink) and a later space-filling phase (in darker ink, and painted in). Quite why this should be so remains a mystery, but I do suspect that here again it was added to distract attention away from the original content of the page.

nine-rosette-filigree

All of which therefore suggests that it might be a (quite literally) revealing exercise to look at the nine rosettes through multispectral filters, to see if we can determine if they similarly began life unadorned, but were then über-decorated at a later stage. Perhaps the right question to be asking of all the drawings hasn’t to date really been asked: what is signal, and what is noise?

Of course, it has been suggested that these filled-out circular designs are merely scribal doodles, time-filling graffiti, a bored student in his medical class, etc: but then again, nobody now looks at the “Academia Leonardi Vi[n]ci” circular knot roundel prints from circa 1500 and calls them mere doodles, so what are the rules in this game, hmmm? Or is this just symptomatic of one of those semantically irregular verbs, that shade away when applied to people you don’t much care for: “I design, you sketch, he/she doodles“?

academialeonardivinci-small

“Academia Leonardi Vici”, print at the British Museum “(After) Leonardo da Vinci”

I sometimes get accused of reading too much into all these fragments: but when you put them together, they very often do afford us glimpses into their secret shared life. This, in the end, is what Art History is all about: for all its forensic aspirations, it is largely a romantic discipline, built on the faint hope that the notion of technique relied upon by each generation of artist will be sufficiently expressed in their works to reconstruct a narrative of continuity across the centuries.

But sometimes, it’s hard to actually put this into practice without feeling somewhat akin to “The Mentalist”: empathetically channelling a torrent of details to bring them back to life as a stream of narrative. But is it only in TV Land that this approach can solve whodunits, or is “The Mentalist” actually a closet art historian? [*]

[*] And yes, I did happen to see “The Mentalist” Episode 1-13 (“Paint It Red”) where Patrick Jane switched a pair of canvases to get a stolen $50m 15th century painting back from a Russian mobster. Be assured that proper art historians don’t do that (well, not often, anyway). 😉

His eyes stinging from all the Google Translate hits popping up on his server logs (what did I tell you?), Elias Schwerdtfeger posted up an English translation of his Voynich Bullshit Index post from a couple of days ago, no doubt cursing me through gritted teeth as he typed. 🙂

By my reckoning, I reckon my Averlino theory in “The Curse of the Voynich” gets:

  • 10 points initial float (i.e. for having any kind of theory)
  • 2 points x 8 years wasted
  • 9 points x 2 identified nine-rosette buildings – St Mark’s Basilica in Venice, Castello Sforzesco in Milan
  • 9 points x 2 buildings identified after visiting them

…that is, a grand total of 62 points. Not bad: but in the words of most school reports, Could Do Better. 🙂