Category Archives: Historical Ciphers

The news rattling the bars of the Voynich research cage loudest right now is surely the publication of a paper by Marcelo Montemurro and Damián H. Zanette called Keywords and Co-Occurrence Patterns in the Voynich Manuscript: An Information-Theoretic Analysis, deftly summarized in New Scientist as New signs of language surface in mystery Voynich text.

M&Z’s abstract brings out a lot of what they were trying to do – and also points exactly to their mistakes.

“Here we analyse the long-range structure of the manuscript using methods from information theory. We show that the Voynich manuscript presents a complex organization in the distribution of words that is compatible with those found in real language sequences. We are also able to extract some of the most significant semantic word-networks in the text. These results together with some previously known statistical features of the Voynich manuscript, give support to the presence of a genuine message inside the book.“

Central Assumption: the authors implicitly hypothesize that they can get meaningful results for long-range comparisons because Voynichese is homogeneous across all its sections.

…The Problem: this assumption is false (or very nearly so), because there are significant macro-level differences in the way the language in different sections works (Currier A, Currier B, labels) as well as many mid-level differences (Herbal-A, Q13-ese, etc).

Central Conclusion: the authors believe that their language-centric statistical machinery has identified “The thirty most informative words in the Voynich manuscript”.

…The Problem: I’m pretty sure that the authors have in fact very probably identified arguably the thirty least informative words in the Voynich Manuscript. (That may be an independently useful result, but it’s probably not really what they were hoping for.)

I’ll explain.

Voynichese is extremely predictable at a letter-level: it has many rigid letter-level adjacency rules (‘4’ is almost always followed by ‘o’, etc) and position rules (4o- is consistently word-initial, -89 is consistently word-final, etc) and a high level of letter-context predictability.

Yet at the same time, it also has a very large dictionary relative to its text size. I often criticize Gordon Rugg for suggesting historically incorrect Cardan grille-like tables (i.e. they’re a century too late for the Voynich’s construction dating) and for inappropriately back-projecting his modern CompSci mindset onto the early Renaissance (i.e. it’s 500+ years too early for the kind of table-driven hackery he proposes). However, he is absolutely right that a reconstructed Voynichese “dictionary” would, to a modern computer scientist’s eyes, look very much as if it had been generated or permuted by some means.

The paradox is therefore that these two apparently opposite aspects of Voynichese are able to coexist: how on earth can we reconcile its letter rigidity & predictability with its wild word variability?

I think the key to resolving this is to grasp that there is some kind of generative or confounding principle at work within a rigidly predictable framework. That is, that even though there are lots of rules, these rules act as a kind of “container” for semantic or cryptographic variability to exist within.

Hence I believe that Montemurro’s statistical machinery is identifying “words” that fall within the container layer rather than in the confounded content layer. Hence these are arguably the thirty least informative words in the Voynich Manuscript.

It’s a hard point to understand, let alone accept: the confounding trick (some kind of transposition cipher? some kind of paper cipher machinery? some kind of cipher wheel?) driving Voynichese’s inherent variability remains as profoundly unreachable now as it has been for over 500 years.

My apologies to Montemurro and Zanette, but the central challenge we face isn’t to find new language-based statistical tests to apply to the Voynichese corpus, however clever they may be. Rather, it is to find ways of resolving the Voynich Manuscript’s central paradox: how is it that Voynichese is both letter-rigid and word-variable at the same time?

Incidentally, M & Z conclude in their paper that results point to a semantic link between the Recipe and Astro sections, and between the Herbal and Pharma sections. Actually, had they been more aware of the codicology analyses that have been done, they would have seen that their results are consistent with the writing phase order.

In fact, there are many indications that what I call Voynichese’s ‘container’ layer above evolved during the writing, with the most obvious evolution being between Currier A and Currier B. I suspect that what their statistical machinery has imperfectly captured is therefore simply a snapshot into the evolution of the container layer, and not anything ‘semantic’ as such.

In short, the aspect of Voynichese that is most nearly homogeneous across all its sections is its “container” layer: so what Montemurro and Zanette have done is make long-range comparisons between evolutions of the container layer. Currently, my best guess is that these are likely to be almost entirely composed of cipher system meta-tokens (shorthand tokens, transposition cipher placeholders, etc) rather than the semantic contents, which appear instead to have been confounded by some means.

So, rather than finding a “genuine message” (as New Scientist put it), perhaps they have instead found a “genuine container” for the message? This may prove to be a very useful result in its own right, but it’s probably not the smoking gun linguistic proof they were hoping to use to discredit Rugg’s tables.

When I was young, I often used to play Scrabble with my grandmother Win on my way home from school. (By which I mean her maisonette was on my route home, not that we played Scrabble on the bus.) Which probably helps account for the deep-rooted enjoyment I still get from weird and wonderful words, many decades later.

From way back then, my favourite English word has always been “svelte” (though “tergiversate” was nipping at its heels for a couple of weeks last year). The reason I particularly like svelte is that it’s (I’m struggling to describe) ‘productively onomatopoeic’, in that the slow ‘l’-sound in the middle makes it feels elegant (indeed svelte) on the tongue. Really, it’s a word with an unusual (but nicely matching) mouth feel, one that manages to stand out from a dictionary sized pack. With getting too synaesthetic on you, to me it’s a kind of David Gower four of a word, a left-handed ping that’s over the boundary before the fielders even notice it’s gone. Something can’t be half-svelte, it’s either got it or it hasn’t.

Svelte also brings right to the fore the mad ragtag heterogeneity of English, the arbitrary coupling together of chance encounters over the millennia. To some it sounds Svedish Swedish (or perhaps a piece of stray Elvish?) but it’s actually a French word (svelte), from an Italian (svelto, “stretched out”), from Vulgar Latin (ex + vellere, i.e. to stretch + out).

(You might therefore suspect that it shares some kind of origin with “vellum” which is also stretched out, but the latter has its roots in “veal”, i.e. young calves: hence vellum is properly fine calfskin.)

Languages are like that: for all their modern apologists, academies, and syntactic niceties, they’re at heart accidental rather than designed. Esperanto and all the other modern conlangs are all very well, but a good part of the charm of real-world languages is the way stray and mongrel words hop in to fill the semantic gaps that inevitably open up as culture mutates and evolves. English obviously needed a word that expressed presence of svelteness in an object, why else would svelte have succeeded and persisted otherwise?

But (and isn’t there always a but in Cipher Mysteries)… where’s all that in the Voynich Manuscript’s language? Even if William Friedman was completely and utterly wrong about the Voynich’s being an artificial constructed language (which he was), I really can see exactly why he thought & believed that. For Voynichese words show such a strong family resemblance – a strongly interlinked productive grammar, if you will – that it almost precludes anything else. Whatever Voynichese is, there is definitely an artificiality to it, or at least an abundance of artifice. I suspect that anyone trying to map Voynichese onto a direct language base will almost inevitably find (to their eventual embarrassment) that it’s just too artificial to be workable: and that’s pretty much what Elizebeth Friedman concluded too.

So here’s your Voynich paradox for the day. I’m sure that there can be no “svelte” in the Voynichese ‘language’ as we see it, because the overwhelming majority of its words arise from a compact productive grammar quite unlike that of a real, heterogeneous, messy, accidental, historic language: and yet the look of Voynichese so resembles a language that it’s hard not to feel as though you’re perpetually a mini-dictionary away from just reading it.

Of course, for me the resolution of this paradox comes down to a well-chosen bunch of steganographic tricks (such as verbose cipher, shorthand, etc) that serve to conceal the plaintext in a misleading form… but you will no doubt have your own theories about how to slice through such a Gordian knot. 🙂

A correspondent has asked me to summarize the evidence I’ve found in the Voynich Manuscript suggesting bifolio reordering. As long-term Cipher Mysteries readers will know, I laid much of this out in my 2006 book The Curse of the Voynich: but a lot has also emerged in the years since.

A practical starting point here is my long-standing page on the Voynich Manuscript’s codicology. This points to evidence for a whole sequence of fairly direct codicological conclusions:-
(1) The Folio Numbers Are Not Necessarily Correct
(2) The Bifolios Are Not Necessarily The Right Way Up
(3) The Quire Numbers Are Not Necessarily Correct
(4) The Bindings Are Not Necessarily Correct
(5) The Quires Are Not Necessarily In The Correct Order
(6) The Quire Contents Are Not Necessarily Correct
(7) The Paints And Colours Used Are Not Necessarily Original

To this, I’d add some other evidence:-

(8) The quire numbers and the folio numbers are not (quite) consistent.

As John Grove pointed out back in 2002, the Voynich Manuscript’s Q9 (“Quire 9”) was rebound along a different fold after the quire numbers were added but before the folio numbers were added, leaving the Q9 quire number in the wrong place (i.e. not on the back page of its quire). And, as Glen Claston later pointed out, broadly the same thing happened for Q14 (the nine-rosette page): once Q14’s first binding fold collapsed or was damaged, the large multi-folio page was then restitched along a different, less obviously damaged fold, again leaving the Q14 quire number in the wrong place.

Of course, given that most of the (15th century) quire numbers look roughly a century older than the (16th century) folio numbers, a bit of rebinding between the two phases is perhaps to be expected. But all the same, this inconsistency should alert us to the fact that the bifolios were actively being worked on between the quiration and foliation.

(9) Some of the quire numbers’ downstrokes continue within the wrong quires.

I found two clear examples of this (Curse p.18): (a) the downstroke of the ‘9’ in “29” (i.e. ‘secund-us’) continues at the bottom of a page in Q6; and (b) the downstroke of the ‘5’ in “5t9” (i.e. ‘quin-t-us’) continues at the bottom of a page in Q3. In both cases (and particularly in the first of the two), it seems likely that at the time the quire numbers were added, the herbal bifolios were in quite a substantially order from the order we are presented with several centuries later.

(10) Vellum tears with parallel orientation may indicate that those bifolios came from the same tanned skin.

The examples I found (Curse pp.54-56) were on the f16-f9 bifolio and the f10-f15 bifolio, as well as on the f38-f35 bifolio and f36-f37 bifolio. The fact that the bifolios were still immediately adjacent in both cases loosely implies that the basic idea of codicological continuity during construction (i.e. that adjacent bifolios individual sections were probably folded and cut down from larger sheets of vellum) may well be sound. It also suggests that the f28-f29 bifolio (which has a stitched vellum tear) may be out of sequence.

(11) Currier “Herbal A” and Currier “Herbal B” bifolios seem jumbled up.

Back in 1976 or so, US Army cryptanalyst Prescott Currier noted two apparently distinct “dialects” of the ‘Voynichese’ language: he called these “A” and “B”, and pointed out a whole set of curious properties that helps you distinguish them from each other. Moreover, any given bifolio has only “A” or “B” writing on it: this broadly supports the idea that these correspond to two broadly separate writing phases, rather than two separate writers writing in parallel.

(12) The three sunflower-like drawings look to have been separated.

f33v, f40v and f50r all contain pictures of similar sunflower-like plants, and are all “Herbal B” pages: this reinforces the idea that the Herbal B pages should be looked at as a quite distinct content collection from the Herbal A pages. I’d add that the Herbal B “plants” seem far more artificial to me than Herbal A “plants”, some (but not all) of which actually resemble real plants (e.g. water lily, pansies, etc).

(13) Q13 and Q20 may have originally each been two smaller quires that were later merged.

There is a whole heap of content analysis that supports the idea that what we now call “Q13” was originally a ‘Q13a’ and a ‘Q13b’ (as proposed by Glen Claston in 2009) and that what we call “Q20” was originally a ‘Q20a’ and a ‘Q20b’ (as proposed by me in 2010).

This is not so very far from the observation [(3) above] that, given that the jars in the pharma section seem to progress from the end of Q19 to the start of Q17, Q19 originally preceded Q17. In short, we can’t be at all sure that the quire arrangement we see now matches the original quire arrangement or order – quires may well have been formed of smaller original quires that were merged (for whatever reason) before the quire numbers were added.

(14) The two pages with “chicken scratch” marginalia may well have originally been adjacent.

I suggested this in my 2012 Voynich Centenary Conference presentation “Between Vellum and Prague”, which tried to reconstruct how the Voynich Manuscript’s quires had moved around between its original ‘alpha’ state and the final (foliated) state. I think it would therefore be interesting to find out whether f66v (in Q8) and f86v3 (in Q14) were from the same vellum skin.

(15) Some bifolios that were (probably) central to a quire are now not.

f84v-f78r should clearly have been the centre of a Q13 quire (the pictures join across the bifolio’s central fold): but I should add that Rene Zandbergen also pointed out in 2010 that f18v-f23r may well have been the centrefold page-pair of a quire; and that I also suggested much the same of f33v and f40r in 2006 (Curse, p.70), though for a different reason.