Schneier on Security
A blog covering security and security technology.
« GENESIS: NSA Exploit of the Day |
| NEBULA: NSA Exploit of the Day »
February 28, 2014
Decoding the Voynich Manuscript
The Voynich Manuscript has been partially decoded. This seems not to be a hoax. And the manuscript seems not to be a hoax, either.
Here's the paper.
Posted on February 28, 2014 at 6:25 AM
• 20 Comments
To receive these entries once a month by e-mail, sign up for the Crypto-Gram Newsletter.
Nick "Cipher Mysteries" Pelling is less than convinced:
"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."
See here for more.
with all the influences cited, it could likely be a Roma language variant.
The moral here is that if you want your ciphertext to be preserved for centuries even when no one has the key, you need to include lots of plaintext pictures of naked ladies.
So, the theory is that it isn't a 'cypher' but simply a dead language.
I guess that makes sense, but I thought the leading theory was that it was a cypher text written by an alchemist (they tended to keep discoveries to themselves).
It's convincing, sorta, or at least intriguing, when listening to his presentation. If one were to begin taking the V.Ms seriously though,
namely that this is real text in some real language written using some real alphabet, then ISTM one thing one must account for is this question : why do we have exactly ONE manuscript in this or a related language/script, and in an improbably superb state, so superb in fact that it was proposed to and bought by a wealthy Prince -- and NONE other albeit tiny fragments, inscriptions, using the same or comparable script ? It's not like nobody has been looking for such, methinks...
I hope it isn't true; I want people to try to decode it, but fail in the end. Life is more enjoyable with mysteries.
Best, simplest, argument I've read that the thing is simply the product of someone's imagination.
Someone on midevalist.net commented that alchemists liked to keep their stuff secret. Small wonder: "If you can't dazzle them with brilliance; baffle them with bull#%1t" (Americanism - ante 1978).
(Doesn't mean Bruce doesn't get the joke - as above)
If you think that you can decrypt the Voynich Manuscript by ignoring pretty much every statistical and cryptological study ever made, and instead latching onto nine wobbly words scattered through 200+ pages, then knock yourself out, Stephen Bax's presentation has everything you're looking for.
Bruce surely knows better than to inhale the vapors off this one.
Personally, I'd love to see someone analyze the Voynich Manuscript, the Codex Seraphinianus, and the Bible using the same decoding techniques. The basic point is that we've got an unknown, a known pseudo-language, and a document composed by multiple authors. It doesn't have to be the Bible, either. An academic text or even a Medieval herbal or alchemical text of age equivalent to the Voynich manuscript could be used as the known text.
The basic point is that the Voynich too often seems to be analyzed in isolation, and it would be nice to study it in comparison to other documents. For example, how much of the text appears to cluster around the images?
My bet is that Dr. Bax, as an Applied Linguistics expert, has got it right, for the first time in the history of the Voynich. The fact that it gives Nick Pelling hissy-fits of jealousy is nice confirming evidence!
"I hit on the idea of identifying proper names in the text, following historic approaches which successfully deciphered Egyptian hieroglyphs..."
Actually, until the Rosetta Stone, the use of proper names to create translations was woefully unsuccessful. It was too easy for the cryptotranslators to fudge the interpretation.
Observer: "hissy-fits of jealousy" LOL. :-)
Look, a long sequence of Voynich linguists have followed this same 'methodology' of plucking out hopeful word lists (though hardly ever with more than 10 words in) since the 1940s. The fact that this is also exactly as far as Bax has got suggests to me that he hasn't exactly progressed the science of Voynichology past the sticking-pins-in-a-map stage.
Bax 1, Pelling -1 . I give Pelling -1 instead of 0 for not showing any grace.
I still think that the Voynich manuscript is an ancient fake. Seeding the ciphertext with words which sound rather like known plant names in a little-known language would be a very good ploy to convince very good decryption experts that they were indeed on to something, and that the manuscript contained hidden knowledge that with work could and indeed would be decrypted.
The likely scenario is thus that the faker manufactures part of the document (complete with easily-broken interesting names) then shows this to a royal cryptographer, with the tag-line of "I've just bought this, but I cannot decode it. Can you? If so, would you like to buy the rest (though you'll need to front me some money so I can get it first)?
Royal flunky looks at the first fragment, finds "likely" words, and fronts up a bribe. Faker then manufactures the rest of the document complete with more decode-able words in a sea of gibberish, and flogs the rest of the fake to the Royals. After a fair amount of decoding attempts, the manuscript is put to one side and forgotten about as being not amenable to decryption. Actually, if we assume that one-time pads were known of (or even just deliberate fakes manufactured to mislead) this may have happened quite a bit, so there may be quite a few promising but unreadable fakes knocking about.
To make a con like this work, the manuscript has to look old. Ancient wisdom, got to be good stuff, yes? To do this, a faker starts out with actual old parchment, removes the previous writing from it, and uses it for his own new scribblings. Doing this used to be common practice, as parchment was valuable, so some faint presence of previous writing on parchment would not be suspicious.
The Voynich parchment has been C-14 tested as being medieval in date. I wonder if anyone has tried to C-14 test the pigments in the writing instead, or has looked for previous writing incompletely erased by the faker?
Nick Pelling: "plucking out hopeful word lists". That sounds like what early philologists used to do before comparative linguistics recognised that systematic correspondences are evidence for relatedness, whereas chance resemblances are not. Is that why Bax has that appendix listing words for "black" that start with the same sounds in 33 different languages? Given how closely related some of those languages are (he even includes both Hindi and Urdu!), that should surprise nobody, least of all a linguist.
If the opinions of "linguistics experts" arewhat matter, I'd be more interested in Jacques Guy's, but it's probably unprintable.
"Heteromeles • February 28, 2014 6:58 PM
Personally, I'd love to see someone analyze the Voynich Manuscript, the Codex Seraphinianus, and the Bible using the same decoding techniques. "
Actually, someone did.
There is an atheist variety show in London every year (forgot name), and one time a performer did copycat an Israeli "scientist" who "decoded" Bible using crosswords. Performer "decoded" Moby Dick using same technique.
Lots of fun!
Ben Firt: actually, I'm acting remarkably gracefully, considering I've spent the last few weeks being repeatedly attacked online by people who think I'm being mean-spirited to poor old Stephen Bax by pointing out that nine wonky (and very questionable) words out of 200+ pages is arguably a worse hit rate than a null hypothesis.
Richard H: I suspect that a sentence by Jacques Guy would be more genuinely informative than 62 pages of Stephen Bax's Voynich theory. And if Guy swears profusely in that sentence, so much the better. ;-)
Schneier.com is a personal website. Opinions expressed are not necessarily those of Co3 Systems, Inc.