Here’s what I think the Voynich Wikipedia page ought to look like. Enjoy! 🙂

* * * * * * *

History of a Mystery

Once upon a time (in 1912) in a crumbling Jesuit college near Rome, an antiquarian bookseller called Wilfrid Voynich bought a mysterious enciphered handwritten book. Despite its length (240 pages) it was an ugly, badly-painted little thing, for sure: but its strange text and drawings caught his imagination — and that was that.

Having quickly convinced himself that it could only have been written by one particular smart-arse medieval monk by the name of Roger Bacon, Voynich then spent the rest of his life trying to persuade gullible and/or overspeculative academics to ‘prove’ that his hunch was right. All of which amounted to a waste of twenty years, because it hadn’t even slightly been written by Bacon. D’oh!

Oh, so you’d like to see some pictures of his ‘Voynich Manuscript’, would you? Well… go ahead, knock yourself out. First up, here’s some of its ‘Voynichese’ script, which people only tend to recognize if they had stopped taking their meds a few days previously:

A nice clear example of Voynichese

Secondly, here’s one of the Voynich Manuscript’s many herbal drawings, almost all of which resemble mad scientist random hybrids of bits of other plants:-

Finally, here’s a close-up of one of its bizarre naked ladies (researchers call them ‘nymphs’, obviously trying not to mix business and pleasure), in this case apparently connected up to some odd-looking plumbing / tubing. Yup, the right word is indeed ‘bizarre’:

voynich f77v central nymph Q13 and Voynich balneology sources...?

Did any of that help at all? No, probably not. So perhaps you can explain it now? No, I didn’t think so. Don’t worry, none of us can either. *sigh*

Back to the History Bit

Anyhow, tucked inside the manuscript was a letter dated 1665 from Johannes Marcus Marci in Prague, and addressed to the well-known delusional Jesuit polymath Athanasius Kircher. Marci’s letter said that he was giving the manuscript to Kircher both because of their friendship and because of Kircher’s reputation for being able to break any cipher. The manuscript seems then to have entered the Jesuit archives, which is presumably why the Jesuit college near Rome had it to sell to Wilfrid Voynich several centuries later, just as in all the best mystery novels.

But hold on a minute… might Wilfrid Voynich have forged his manuscript? Actually, a few years back researchers diligently dug up several other 17th century letters to Kircher almost certainly referring to the same thing, all of which makes the Voynich Manuscript at least 200 years older than Wilfrid Voynich. So no, he couldn’t have forged it, not without using Doc Brown’s flux capacitor. (Or possibly the time machine depicted in Quire 13. Unless that’s impossible.)

Incidentally, one of those other letters was from an obscure Prague alchemist called Georg Baresch, who seems to have wasted twenty or so years of his life pondering this curious object before giving it to Marci. So it would seem that twenty lost years is the de facto standard duration for Voynich research. Depressing, eh?

So, Where Did Baresch Get It From?

Well… Marci had heard it said that the Holy Roman Emperor Rudolf II had bought the manuscript for the ultra-tidy sum of 600 gold ducats, probably enough to buy a small castle. Similarly, Wilfrid Voynich discovered an erased signature for Sinapius (i.e. Jacobus Horcicky de Tepenecz, Rudolf II’s Imperial Distiller) on its front page. You can usefully assemble all these boring fragments of half-knowledge into a hugely unconvincing chain of ownership going all the way back to 1600-1610 or so, that would look something not entirely unlike this:-

Which is a bit of a shame, because in 2009 the Voynich Manuscript’s vellum was radiocarbon dated to 1404-1438 with 95% confidence. Hence it still has a gap of roughly 150 years on its reconstructed CV that we can’t account for at all – you know, the kind of hole that leads to those awkward pauses at job interviews, right before they shake your hand and say “We’ll let you know…

Hence, The Real Question Is…

Fast-forward to 2012, and Wilfrid Voynich’s manuscript has ended up in New Haven at Yale University’s Beinecke Rare Book and Manuscript Library. Yet many Voynichologists seemed to have learnt little from all that has gone before, in that – just as with Wilfrid himself – they continue to waste decades of their life trying to prove that it is an [insert-theory-here] written by [insert-historical-figure-here].

If repeatedly pressed, such theorists tend to claim that:
* the ‘quest’ is everything;
* it is better to travel than to arrive; and even
* cracking the Voynich might somehow spoil its perfect inscrutability.
All of which, of course, makes no real sense to anyone but a Zen Master: but if their earnest wish is to remain armchair mountaineers with slippers for crampons, then so be it.

Yet ultimately, if you strip back the inevitable vanity and posturing, the only genuine question most people have at this point is:

How can I crack the Voynich Manuscript and become an eternal intellectual hero?

The answer is: unless you’re demonstrably a polymathic Intellectual History Renaissance Man or Woman with high-tensile steel cable for nerves, a supercomputer cluster the size of Peru for a brain, and who just happens to have read every book ever written on medieval/Renaissance history and examined every scratchy document in every archive, your chances are basically nil. Zero. Nada. Zilch. Honestly, it’s a blatant exaggeration but near enough to the truth true: so please try to get over it, OK?

Look, people have been analyzing the Voynich with computers since World War Two and still can’t reliably interpret a single letter – not a vowel, consonant, digit, punctuation mark, nothing. [A possible hyphen is about as good as it gets, honestly.] Nobody’s even sure if the spaces between words are genuinely spaces, if Voynichese ‘words’ are indeed actual words. *sigh*

Cryptologically, we can’t even properly tell what kind of an enciphering system was used – and if you can’t get that far, it should be no great surprise that applying massive computing power will yield no significant benefit. Basically, you can’t force your way into a castle with a battering ram if you don’t even know where its walls are. For the global community of clever-clogs codebreakers, can you even conceive of how embarrassing a failure this is, hmmm?

So, How Do We Crack It, Then?

If we do end up breaking the Voynich’s cipher, it looks unlikely that it will have been thanks to the superhuman efforts of a single Champollion-like person. Rather, it will most likely have come about from a succession of small things that get uncovered that all somehow cumulatively add up into some much bigger things. You could try to crack it yourself but… really, is there much sense in trying to climb Everest if everyone in the army of mountaineers that went before you has failed to work out even where base camp should go? It’s not hugely clear that even half of them even were looking at the right mountain.

All the same, there are dozens of open questions ranging across a wide set of fields (e.g. codicology, palaeography, statistical analysis, cryptanalysis, etc), each of which might help to move our collective understanding of the Voynich Manuscript forward if we could only answer them. For example…
* Can we find a handwriting match for the marginalia? [More details here & here]
* Can we find a reliable way of reading the wonky marginalia (particularly on f116v, the endmost page)? [More details here]
* Can we find another document using the same unusual quire numbering scheme (‘abbreviated longhand Roman ordinals’)? [More details here].
* Precisely how do state machine models of the Voynich’s two ‘Currier language’s differ? Moreover, why do they differ? [More details here]
* etc

The basic idea here is that if you can’t do big at all, do us all a favour and try to do small well instead. But nobody’s listening: and so it all goes on, year after year. What a waste of time. 🙁

A Warning From History

Finally: I completely understand that you’re a busy person with lots on your mind, so the chances are you’ll forget almost all of the above within a matter of minutes. Possibly even seconds. And that’s OK. But if you can only spare sufficient mental capacity to remember a seven-word soundbite from this whole dismal summary, perhaps they ought to be:

Underestimate the Voynich Manuscript at your peril!

Now ain’t that the truth!?

I have to say that I’m pretty humbled by the hit stats for the Wikipedia Voynich page: when the xkcd webcomic spike happened in June 2009, the Voynich page got a quite shocking 77k hits in a single day. In fact, its daily traffic has gone up from 500-1000 hits at the start of 2009 to 1000-2000 hits as of now (May 2010) (though interspersed with occasional 10k days).

And so I wondered if it was just about time for Jim Davis’ Garfield to take on the Voynich Manuscript: if he did, might it look something like this?

For any passing lawyers, I’m neither passing this off as an original Garfield strip by Jim Davis nor trying to make money from his intellectual property, it’s just one of his strips unconvincingly hacked to show what my idea of a Voynich-themed Garfield strip would look like for the benefit of my Cipher Mysteries readers. And yes, I’ll happily take it down if you ask me to. 🙂

If I had a pound for each time I’ve been disparaging about the Wikipedia Voynich Manuscript article, I’d probably be able to pay off my mortgage: but this is not because I’m negative (when I’m actually the complete opposite), but rather because despite being the most-linked page in the whole Voynich infosphere (Google has it as #1, for example), it’s miserably unhelpful on a truly grand scale.

A few days ago, I decided to go through it to see how much of it I could fix… but after a while, what struck me was that I was just sticking plasters on a gaping wound – that there was something so badly wrong with the way it was structured that no amount of textual hacking would ever fix it. That is, the problem isn’t that it is awkwardly written, but that it is epistemologically broken. I therefore added a new section to the Wikipedia Voynich talk page to say:

[…] But on reflection, it strikes me that what is fundamentally wrong with the whole thing is that it doesn’t really separate objective from subjective; and furthermore that its overall structure helps perpetuate this mingling of fact and speculation. I think that if we were to restructure it more sensibly, we might yet turn this whole article into a genuinely useful resource, rather than the sprawling sequence of speculative stuff it currently is.

I therefore propose that before the (currently first) “Content” section, a new section should be added called “Evidence”, with suggested subsections “Codicology” (facts about the support material [including radiocarbon dating], inks, paints, construction, quire numbers, foliation), “Palaeography” (facts about the ductus, Currier ‘Hands’), “Art History” (techniques used in the construction, such as parallel hatching) and “Statistics” (facts about the letter-patterns and word-patterns). The idea is that by doing this, we can separate out the discussion of what the Voynich Manuscript is from the discussion of what the Voynich Manuscript might be, so that people coming to the topic for the first time can gain an accurate picture untainted by the currently rather high level of embedded speculation. My Cipher Mysteries pages on codicology, quire numbers, parallel hatching may well be fruitful references for some of these topics.

I have elsewhere contrasted the existing speculation-centric “Voynich 1.0” approach with this “Voynich 2.0” history-centric approach: though the main argument against this used to be that we knew too little about the Voynich Manuscript to produce a useful non-speculative introduction to it, my opinion is that this is no longer true. Yet however sensible this change may be from my perspective, it is almost certainly too significant a difference to impose unilaterally on a well-linked article without some kind of sustained debate here first. So… what do you think? Nickpelling (talk) 09:11, 20 May 2010 (UTC)

This is an issue that really affects our whole research community (rightly or wrongly, the Wikipedia page is seen as the front face of the Voynich Manuscript), so please leave your comments either here or by editing the Wikipedia talk page by hand (but log in first). Have I got this right, or do you think the page is just fine as it is?

Something further to reflect upon is the differentiation drawn by Michael Shermer in a recent New Scientist article: he draws a line between skepticism and ‘denialism’, in that ‘denialists’ start from an ideological point of faith and deny / undermine all the evidence that’s inconsistent (i.e. they support their own weak ideas by trying to weaken the evidence supporting other competing ideas), whereas skeptics (as Shermer is the publisher of “Skeptic” magazine, these are the good guys) use “extensive observation, careful experimentation and cautious inference” to separate “the few kernels of wheat from the large pile of chaff”.

Within this conceptual framework, it strikes me that the overall Voynich research programme has had a particularly denialist methodology:

  1. Find a starting point (usually a similarity between something in the VMs and something you happen to know about)
  2. Construct a personal leap of faith (“because these are similar, the VMs must have been written by…”)
  3. Collate all the miscellaneous scraps of evidence that seem consistent with your personal leap of faith
  4. Find rhetorical ways of being dismissive about all the other evidence that is broadly inconsistent
  5. Construct ways of undermining the methodology behind any conflicting evidence (“radiocarbon dating is inaccurate”, etc)

What Shermer omits, though, is that in the real world the trickiest issue of all is how to tell denialists apart from skeptics. After all, they use the same techniques and rhetorical stylistics: who can say where Popperian falsification stops and apologetics starts?

Actually: I can. Firstly, there’s a further division to be made between a cynic and skeptic, insofar as I think a cynic is a denialist whose own personal leap of faith is that “no theory is possible”: you can therefore think of a cynic as an ideological pessimist, who actively denies the possibility of there ever being an achievable answer. By way of comparison, a skeptic is optimistic enough to believe that an answer is possible, but realistic enough to know that the road to such knowledge can be a long and hard one.

To me, a skeptical methodology should be a far more holistic thing (insofar as its goes from the general to the specific), far closer to the kind of thing intellectual historians do:

  1. Collect all the relevant evidence you can
  2. Find ways of assessing the reliability, pliability, and nature of each of them
  3. Assess (and continually reassess) which are the key pieces of falsifying evidence – the ones that let you reject hypotheses
  4. Construct a number of hypotheses that your key falsifying evidence pieces do not manage to kill
  5. Devise research questions to try to limit / constrain / refine / kill your hypotheses
  6. Keep trying (knowledge is hard)

The central differences are therefore (a) the overall direction of research, (b) skeptics understand that most evidence is not absolute, and (c) that skeptics attempt to invalidate all hypotheses, rather than just confirm their own. Further, skeptics know that correlation is not the same as causation, and that causation is 100x more elusive. So… are you a skeptic, a cynic, or a denialist?

Applying all this to the Wikipedia VMs page, at no point does the article try to collate or list all the relevant evidence, or even to give an idea of how reliable or (im)precise each fragment is. Rather, it seems to be a long sequence of theories abutted by denialist oppositions to them. Given that my bias is towards the kind of skeptical methodology I describe above, my overall position is that the page offers little or no assistance to someone who comes in and wants to understand the object prior to forming an opinion: rather, the page offers a load of pre-formed opinions and rebuttals, with fragments of information teasingly wedged between them. It’s rather like trying to understand wildebeest anatomy by looking at the scraps of meat left on a lion’s teeth.

Perhaps I’m barking up completely the wrong tree here: perhaps the “Wikipedia” model for knowledge is innately denialist, in that each article seeks to find a position of dynamic balance between actively-held apologetic positions. I suspect that the epistemological error underlying Wikipedia is that it mistakes a Mexican stand-off for consensus, when skeptical knowledge should ultimately be organized in a quite different way: all the Voynich article does is to highlight this error in a fairly extreme way.

What do you think – can Wikipedia ever be fixed?

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. 🙂