Writer (and University of Bristol PhD student) Gerard Cheshire has recently been asking people to look at his paper “Linguistic missing links: instruction in decrypting, translating and transliterating the only document known to use both proto-Romance language and proto-Italic symbols for its writing system“. (Note that this is actually a draft, but dressed up to look as though it is to be published in “Science Survey (2017) 1” when, as far as I can tell, there is no such journal as “Science Survey”.)

His paper breathlessly reveals that Voynichese is nothing more than Vulgar Latin (though without any obvious grammar or structure). He then proposes a scheme mapping Voynich letters to normal letters (though this lacks “b/f, c/k, ch/sh, g/gh, h/j/ym v/w, x/z” [p.17]), which he then uses to “transliterate” some sentences (though shaped out of strings of words assembled from God-only-knows-how-many different European languages) into something approaching modern-day English. These sentences ‘demonstrate’ that the Voynich Manuscript is (running counter to the radiocarbon dating) actually from the 16th century, and is nothing more than a courtly woman’s health and bathing manual, a fact which every other Voynich Manuscript researcher to date has been too short-sighted to see or recognise, bla bla bla bla bla bla bla.

Errrm… really? Really? Really?

Vulgar Latin

First thing I have to point out is that there is no such (single) thing as Vulgar Latin: rather, the phrase denotes a vast family of vulgar / pidgin / hybrid Latin-ish spoken languages sprawled across all of Europe and over most of a millennium.

Every single version of Vulgar Latin was a purely local affair, nobody spoke them all at the same time – Vulgar Latin wasn’t a universal lingua franca, it was a heterogenous set of hacky vulgar dialects that helped people get by locally. And I simply don’t believe for a moment Cheshire’s implicit claim (completely necessary to his argument, but not expressed anywhere I can see) that this kind of Vulgar Latin had no structure, that each specific instance of Vulgar Latin was no more than language expressed as a diarrhoeal deluge of words that listeners teased meaning out of.

As a result, the entire linguistics mindset running through Cheshire’s paper (i.e. the comparison between a single concerted instance of a script and a vast cloud of unwritten potentialities diffusely surrounding a huge family of languages, each of which is presumed to have no structure) seems utterly wrongheaded.

As such, it makes no sense at all to compare a single slab of written Voynichese text (which gives every sign of having been written in a single time and place) with a wide set of different language potentialities (that, further, were almost never written down, and – further still – would in every instance have had a basic rationale and structure [because that’s how language works] that he requires to be absent).

A Monstrous Mash-up

Even though Cheshire puts forward his speculative translations (which he repeatedly calls “transliterations”, as if that somehow brackets out the mile-wide interpretational chasms he repeatedly has to swing across) of several sections of the Voynichese text, I’m going to give as my example here the top three lines of f82v that he discusses on pp.20-21. This is because f82v is a nice, bright, easy-to-read page in the “Balneo” quire (Q13), which means that the various EVA transcriptions speak almost with a single voice:


Cheshire’s own transcription of these lines (according to his conversion-to-letters-scheme) is as follows:

molor orqueina doleina dolinar æor domar om nar nar or æina,
dolina ræina domor nor æina æina na nas omina eimina rolasa,
nais oe eina domina domeina etna domar doma dolar dolina ro.

Let’s take each line apart in turn to see what he’s trying to get at:

molor = mollor = (soften/calm/pacify) [Latin] – because molor (grind/mill/wear) [Latin] “would be inappropriate”
orqueina = ?
doleina = therapeutic [Catalan]
dolina/dolinar = bath/bathe [Romance languages]
æor = ?
domar = to tame/control [Catalan and Portuguese]
om = hom (homine) = man [Latin]
nar nar = foolish/crazy/up-tight [Romansch]
or = ?
æina = wife [Catalan]

Cheshire’s “reasonable transliteration” (i.e. speculative translation) for this first line is: “Calming with therapeutic bathing is always certain to tame the tense man and wife“.

dolina/dolinar = bath/bathe [Romance languages]
ræina (reina) = queen [Romance languages]
domor = [domar] = to tame/control [Catalan and Portuguese]
nor = daughter-in-law [Aromanian]
æina = wife [Catalan]
æina = wife [Catalan]
na = ?
nas = ?
omina = omen [Latin]
eimina = to eliminate [Spanish and Portuguese]
rolasa = ?

His translation of the second line is: “A queen’s bath always relaxes the daughter-in-law and wife to eliminate the omen, for it to happen“.

nais = to begin/commence/create [French]
oe = ?
eina = ?
domina = lady [Latin]
dome[i]na = domain/room [Latin]
etna (ætna) = to heat/burn [Latin/Greek]
domar = to tame/control [Catalan and Portuguese]
doma = ?
dolar = ?
dolina/dolinar = bath/bathe [Romance languages]
ro = abbreviation for rogo = to ask/request [Latin]

His third line of translation runs: “Begin now the method for the lady’s domain, and heat the room to make the bathing smooth, please!

Cheshire sums up what these three lines mean as follows:

So, the passage appears to be advice for the mother (queen) of a prince to impart to her daughter-in-law as guidance for seducing her son and becoming pregnant.

Like a badly mislabeled lift, this is wrong on so many levels. Nobody reading the above should need to look through Latin, Catalan, Portuguese, Romansch, Aromanian, French, Greek, and “Romance languages” dictionaries to find words to describe this fantastical nonsense. (Though you might find Partridge’s “Dictionary of Slang and Unconventional English” most fruitily germane to the task.)

Disastrous Dog’s Dinners

What Cheshire has been seduced by here is the beguiling notion that the numerous textual difficulties that Voynichese presents might all be magically explained away by a wave of the polyglot fairy’s wand, e.g. that the Voynich’s tightly-knit buzz of similar words might simply be a result of a large number of active component languages somehow feeding into the plaintext. However, it should be no surprise that these polyglot sirens appear rather different when you take a closer look at them:

For all the undoubted cleverness of Leo Levitov PhD, his particular polyglot reading of the Voynich was (as seasoned Voynich Manuscript researchers will happily attest) more or less exactly the same kind of dog’s dinner as Cheshire’s is. And this was for exactly the same reason, which is that the Voynich Manuscript’s curious text presents so many different kinds of non-language-like behaviours all at the same time that trying to read it as if it were a simple language (even a polyglot mash-up “simple language”) is never, ever going to work.

Specifically, the kind of challenging textual behaviours I’m talking about here are:
– 1) Low entropy (highly predictable, babble-like text)
– 2) Highly structured letter placement rules (e.g. highly stylized word beginnings and endings)
– 3) Two or more significant language variants
– 4) A surprisingly high (dictionary size) : (corpus size) ratio.
– 5) A generative dictionary (i.e. covering many more permutations than normal languages do)
– 6) Only sporadic word adjacency pattern matches
– 7) Neal keys (both vertical and horizontal)
– 8) Where are common words like “the” and “and”?
– 9) Where are the number shapes, number clusters, or number patterns?
– (etc)

My point here is that while it is possible to construct a proof-of-concept plaintext language to partially get around one or two of these issues, all the other pesky behaviours will then cause that ‘solution’ to sink like a Chicago Mafia whistleblower. This is all pretty much what Elizebeth Friedman was talking about in 1962 about people seeking such solutions being “doomed to utter frustration”: it’s a horrible shame that in 2017 people continue to fail to even begin to grasp what is such a basic message.

In the case of Cheshire, a polyglot Vulgar Latin reading would aim to get around points 4) and 5), but would then collapse in a miserable heap at the hands of all the other points. Anyone following Stephen Bax’s miserable lead to try to come up with their own ingenious linguistic reading of Voynichese should wise up to the whole list, because – unless you are even trying to satisfy all these oddly non-language-like constraints all at the same time – you’re plainly wasting both your own time and that of everyone else you try to convince.

Laughable linguistics

When I read nonsensical papers like this (and I can assure you that this is not an outlier, because there are plenty more of them out there), I feel a deep sadness for historical linguistics. Even for unbelievably bright people such as George Steiner (who at his peak was clearly a hugely inspirational speaker, and whose books oddly summon to mind Ioan Couliano’s syncretic layerings), far too many linguists lard their writing with speculative etymological riffs anyone else would be embarrassed to put their name to, even if they were walking home from a beer festival drunk and wearing a foolish hat. (For his sins, Cheshire throws a fair few of these soggy prawns onto his linguistic barbecue.)

And whenever I see linguistics people rap about Ur-languages while constructing metronomically-timelined millennia-spanning etymology trees (yet again), I just despair. All the while modern linguists can’t construct solid etymologies for the words we use in the 21st century, what chance do historical linguistics people really stand going back X hundred years? Honestly, some things lie beyond the limits of useful reconstruction, and trying to claim otherwise is a collective (and discipline-wide) failure.

To me, the structural problem with historical linguistics, then, is that if you remove all the brazenly bullshit stuff, what little is left is perilously close to a tree-less tundra: it remains an academic discipline, sure, but one whose grasp of history is all too often paper-thin (as is its actual use to historians), and whose pretensions to science are largely laughable.

And so I really don’t think that Gerard Cheshire should feel bad about having ended up down a garden path here, when it’s actually historical linguistics that has marched down that garden path en masse. The entire conceptual toolkit that he brought to bear on the Voynich Manuscript was as much use as a Swiss Army Knife made of soft-set jelly: sorry to have to say it in such flat terms, but the poor bugger never really stood a chance.

31 thoughts on “Gerard Cheshire, Vulgar Latin, and the siren call of the polyglot…

  1. Wasn’t it the Cheshire Cat who said: “If you don’t know where you are going, any road will get you there.”

  2. J.K. Petersen on November 11, 2017 at 8:15 am said:

    “… any road will get you there.”

    And if you ignore the logical fallacies, you’ll get there even quicker.

    As for playing two ends against the middle… Cheshire’s paper simultaneously claims the VMS is “Vulgar Latin” (which he essentially describes as broken-down, simplified, degraded Latin) and proto-Latin/proto-Italic (which existed before there was Latin), so apparently the road that gets you there includes a time machine.

  3. Nick
    Reading the post about ‘native Indian’ languages, it struck me that I’ve not seen anyone talk about whether they suppose the Vms’ written text was intended to be spoken, or read. As your post illustrated so nicely, orthography has to be fairly regular if the text is to be read, but not if to be spoken or sung.

    I’ve seen this reflected even in some early medieval Latin texts, where a line will run on with no discernable space between words until a breath is needed.

    By the same token, a text meant to be spoken aloud could be broken pretty much at random and means used for shortening the text might differ, as might the grammatical structure. As example, have a look at a practiced cook’s notes for a recipe, a technician’s notes-to-self … and I suppose as the extreme, scientific and mathematical notations.

    I’m not offering a theory – it just occurs to me that the default expectation has been that ‘underneath it all’ the Voynich text should be one of the ‘meant to be read’ sort. What do you say?

  4. Vulgate was a written language. The Bible used it for 1000 years. Columbus took a copy with him in 1492.

    Please correct your interesting but fuzzy narrative, and stop attacking people by creating false narratives of one type or another.

    Stick to the text, and the stick to the linguistic and numerical patterns. No where have I seen that the 800 year use the pre-base 10 decimal numeration system built upon the 800 AD Arab innovation of scaling rational numbers to 2-term unit fraction series, whenever possible, considered. Medievals worked from 1454 to 1585 to end the Arab unit fraction system that ended ciphered Greek letters that represented the counting numbers, with 1 = alpha, 2 = beta, …, with unit fractions denoted by placing (‘) after the ciphered Ionian or Doric letter.

    Fibonacci’s 1202 AD “Liber Abaci ” rigorously exposed the system that used subtraction of a LCM 1/m such that the greedy algorithm:

    (n/p – 1/m) = (mn – p)/mp set (mn – p) = 1 whenever possible.

    When impossible, I.e. 4/13 a second LCM 1/m was chosen such that 4/13 = 1/4 + 1/18 + 1/468 as I have previous detailed on the blog.

    I will repost Fibonacci’s seven distinctions, methods, the began the Liber Abaci, as Ahmes began his math text with a 2/n table, that generally scaled rational numbers n/p to 2-term and 3-term unit fraction series, a topic that you will likely continue to ignore, whenever you desire.

    Medievals were skilled encoders of words and numbers. Look at both sides of their intellectual lives and stop throwing out babies with tbe bath water.

    In closing Vulgate was a non-academic language, a spiritual language, that was in use for at least 1,400 years, ended in the English speaking world with the full acceptance of the King James Bible.

    Best Regards,

    Milo Gardner

  5. Milo: “the Vulgate” is the 4th century translation of the Bible into Latin – it was not written in Vulgar Latin, but in Latin. You are creating your own false narrative by confusing two completely different things – even Wikipedia is able to tell them apart:


  6. Diane: plenty of people have suggested that Voynichese might be primarily from a (typically ‘lost’) oral speaking tradition, but without anything so useful as evidence or even helpful observations to back the notion up.

    Conversely, the many curious regularities and patterns that Voynichese presents (and indeed the general regularity of the writing) do collectively seem to suggest that it is primarily a writing system, though at the same time somehow managing to follow a deep logic that doesn’t seem to match anyone’s expectations of how languages are, should be, or even could be written. Oh well, doubtless there’ll be yet another Voynich language theory rolling towards us all down the very same rails before very long. :-/

  7. NIck

    My specialization at university was historical linguistics, so first I feel compelled to add some nuance to the final bit of this post. Of course, I mostly agree: yes, obviously Ur language theories are ridiculous; written language (i.e. all we know for sure) only covers a tiny, tiny section of human history. So we can only peer through the fog a few meters around us, but everything beyond a certain point in history is unknowable forever.

    That said, we have a really good idea of some languages that precede modern ones: Latin, Sanskrit, Gothic… (although Gothic is more like an uncle). By comparing Gothic and the oldest attested forms of other Germanic languages, we also have a clear picture of what their ancestor looked like. Still, we call this Proto-Germanic and mark all used forms (*), to indicate that these forms have been reconstructed and do carry a certain amount of uncertainty.

    Language change is unpredictable, but it does follow surprisingly strict rules. What I mean is, we have no way of knowing whether or not English /p/ will shift to /f/ in a century. What we *do* know is that a shift from /p/ to /f/ is a realistic one, since only one aspect changes (plosive becomes fricative). On the other hand, a direct systematic shift from /f/ to /g/ is impossible because three properties of the sound change.

    Since we know how languages evolve, we can compare the oldest known offspring of a language and reconstruct it from there. Even though the reconstructed proto-language remains hypothetical, the results are scientific and valuable. Proto-Indo-European was still spoken some three thousand years BCE if memory serves, which is peanuts on the scale of humanity’s existence.

    But going beyond Proto-Indo-European is a bit floaty and extremely speculative. It’s a fun exercise, but we just don’t have the data (and never will) to be even a bit certain. This is different for languages like Proto-Italic, since here we have a wealth of information about its direct offspring.

    Now to the point…. I have also been “chosen” to receive Cheshire’s paper and he doesn’t know what he’s talking about. He knows not much about history, even less about linguistics and has clearly not read much about the VM. His methodology is rubbish and he doesn’t even use the correct basic terminology. (He calls folios “spreads” or in the best case “portfolios”. He thinks herbal remedy and homeopathy are synonyms etc.)

    As I told him, it is impossible that the VM contains Proto-Italic. The oldest Latin writings we have are Old Latin, and the language was in an inscriptions stage at that point – compare it to attestations of Runic script, carved on objects or stones. That was before I realized that he doesn’t have a clue what “Proto-Italic” means.

    Anyway, I sent him a detailed explanation of the flaws in his paper, assuming he’d appreciate some pointers, but he replied that it’s a shame I put so much time in it since he knows he’s right – so no need to correct him.

    Apart from this, I also see a surprising amount of ethical difficulties here. As you indicate, he faked magazine publication. By stressing that he’s a PhD student, one also gets the impression he’s a linguist, which is far from the truth. He also lied that “proper linguists” agree with him (impossible) while he gets criticism only from Voynich enthusiasts. Furthermore, he presents his paper as a purely linguistic endeavor, in which the VM is only marginally relevant. His title doesn’t contain any reference to the VM and he doesn’t use the word “Voynich”. He told me this is because he doesn’t want to be associated with, well… with us basically 😉 But the contents of his paper is nothing more and nothing less than yet another Voynich-Latin theory.

    He’s been sending me some more translations, which indicate that he doesn’t even grasp the basics of Latin. Whenever his method produces a pseudo-Latin second person verb, he translates it as a first person.

    Anyway, given his complete deafness, lack of any relevant knowledge, lack of research ethics and tendency to implicitly insult me, I’ve decided to stop replying. Aint nobody got time for dat.

  8. Koen: thanks for your detailed comment, much appreciated. The bit that got my goat the most was his repeated use of “transliteration” as a fake academese replacement for “unbelievably speculative simulated translation”. Ho hum.

  9. Nick,

    Transliterations are commonly used by academics, thereby only exposing what was written on the page. There are topics, math being one, that written texts are only shorthand notations. To read the mental and written texts, as originally written, in their entirety, missing math steps need to be added back.

    A case in point is the 1202 “Liber Abaci”, Europe’s only math book for 250 years, ended in 1454 when the Ottoman Empire overran Constantiople, ended the Byzantine Empire. Oddly the Latin text was not fully translated into English until 2002, with the life work of L.E. Sigler. Sadly, many sections of the texts were only transliterated, thereby missing the intellectual nuggets that remain hidden to academics that make no attempt to complete Sigler’s life work.

    Milo Gardner

  10. For more on Milo Gardner’s analysis of the Liber Abaci: http://liberabaci.blogspot.co.uk/

  11. Milo: the author of the paper I’m reviewing here tries to pass off all his extraordinarily subjective speculations and overoptimistic interpolations as “transliterations”, as if by calling them that he can give his work a halo of “science-iness”.

  12. Milo
    “1202 “Liber Abaci”, Europe’s only math book for 250 years”.. not sure what you include with ‘math’, but copies of Euclid were still being made and used as late as the fifteenth century and inform secondary works to the present day. The same is true for Boethius’ Arithmetic, believe it or not. And the first mention I recall reading of that in Latin Europe is the comment by Gerbert d’Aurillac that he found a copy bound with another work in the north Italian monastery where he was abbot until elected Pope.

    Then, if you want to move from geometry and arithmetic to specific applications:
    merchants handbooks where you learned to do mental arithmetic of some complexity and conversions of weights, measures and money

    Plus the various astol-onomical tables which required several maths. skills to work with.

    Then the sort of thing written down in a couple of mercantile handbooks but which had earlier been taught by rote: traders traded before Fibonacci.

    I really can’t agree that the Liber Abaci was the ‘only’ book that you could learn maths from. Before or after the fall of Constantinople.

  13. Nick,

    Let me begin by agreeing with one aspect of your vulgate post. Of course Vulgate was an oral and written Latin language. Oral forms were regional in scope, as you fairly pointed out. Yet oral Vulgate was replaced by Italian, connects to Dante and others hard work , points that your Wikipedia linked did not include. Written vulgate in the 1600s was changed to an academic form compared to the Vulgate known to Jerome, specifics can be detailed by Koen, and trained linguists.

    Koen, you may wish to comment on pertinent written forms of Vulgate that stayed the same, and those words and metaphors that changed, during the Jerome to Columbus periods.

    Nick, your Classical Greek education, as learned in the UK, was oddly and sadly elitist compared to the history of economic thought, during the same time periods, that I studied in the US.

    The British Museum is a case in point. The BM, to this day, oddly places Greek math above its parent, Egyptian fractions and Egyptian math, specifically by only transliterating Henry Rhind’s 1858 purchases in Cairo, the EMLR and RMP. The BM limited its Egyptian arithmetic to additive issues, rather than as Plato and others reported most Greeks learned math by traveling to Egypt. Archimedes could not place a paper in Alexandria’s library without the approval of Eratosthenes, a Greek policy that continued the very old Egyptian practice.

    Greeks practiced number theory learned from Egyptian math texts. Egyptian, Greek and Number theory from any era, including our own, defines multiplication and division operations as inverse to one another. The 1900 BCE Akhmim Wooden tablet multiplied (64/64) by 1/3, 1/7, 1/10, 1/11 and 1/13 into exact n/64 quotients and 1/320 remainders. Two-part answers were returned to 64/64 as proofs by multiplying by 3, 7, 10, 11 and 13, vivid facts missed by Peet (in 1923 by only citing n/320 data) , but acknowledged by Hana Vymazalova in 2002. Peet and the BM oddly conclude, to this day, that Egyptian division was limited to a guessing method named “single false position”, actually a medieval geometric solution to second degree equations…that was not limited to a division definition.

    Let me stop here, having bypassed your UK classical education in ways that probably make your head ache. Be a BM believer, and maintain skeptical views of Egyptian number theory that reached the medieval era, if you wish.

    But you have been informed that your medieval number eye has been blind, and can not be used to properly decid the VM, likely filled with Arab unit fractions.

    Best wishes,


  14. Milo: it is hardly elitist of me to point out that the Vulgate Bible was not written in Vulgar Latin. And you are surely not doing whatever camp you think you’re in any favours by continuing to confuse the two (immensely different, if similarly named) things together, despite my best efforts.

  15. J.K. Petersen on November 12, 2017 at 1:38 am said:

    Milo, did you read Cheshire’s paper?

  16. Milo
    Better than wikipedia, but not so good as the hard-copy version.

    I’m assuming that the oldest western (Latin-) church would know their stuff when it comes to the Vulgate, since their history goes back to the time it – the Vulgate [version of the Latin BIble] – was written.

  17. Koen is right about Cheshire, and also about historical linguistics being an accurate science. The key to it is historical phonology. Proto-Germanic is a hypothesis consisting of the phonological system of the proto-language in the form of a table: labial, labiodental, dental etc times voiced stop, unvoiced stop, fricative etc (similarly for the vowels), and a set of context-dependent rules deriving the phonological systems of attested languages from it. Lists of words (e.g. zwei = twee = two) are the evidence for the hypothesis. The evidence for proto-Germanic is strong, consisting of hundreds of clearly related words and an absence of systematically anomalous ones. Phonological constraints are also what allows us to cite cognate words of similar but not identical meaning (e.g. Zaun = tuin = town) without claiming a licence to do so ad hoc.

    Claims that Indo-European, Uralic and other languages are related as members of the Nostratic superfamily are not constrained in this way. They may possibly be true, but the hypothesis is inherently weaker than the Indo-European hypothesis because it is not a phonological hypothesis and the pool of possibly cognate words is too small to establish one. Is there any prospect of progress? I see one in the analysis of apparent Indo-European loan words in other language families, particularly Sumerian (Whittaker’s Euphratic) and Old Chinese (as reconstructed by Edwin Pulleyblank and Christopher Beckwith): watch this space!

  18. Mark Knowles on November 12, 2017 at 3:11 pm said:

    Nick: I have devised a test for individuals claiming to be able to translate the Voynich. The test does unfortunately have to rely on the honesty of the test taker and a non-photographic/savant memory of the manuscript. However if the test taker lies to us they are also lying to themselves about their ability to translate it. I think whether you or I believe they can translate the Voynich most people claiming to be able to translate the Voynich really believe they can and so would be happy to pursue the test in all honesty with the conviction that they will do well.

    This is my version, but it could be applied more generally:

    I copied a selection of clearly written labels from the 6 rosette foldout page (this could also be applied to other pages with clear distinct labels) onto a blank image page. Then the distinct items of Voynich text are numbered 1,2,3,4,…..

    The test taker is emailed the list and required to translate the items. Then the test taker is then emailed a copy of the 6 foldout page with labels removed and asked to mark where each label by number belongs.

    Throughout the test the test taker is told not to in any way consult the original 6 foldout page image. (This is where the test taker’s honesty is required. If they cheat obviously they will be able to fix the results. There is no way of preventing them from cheating) There is no time limit for the test though it would be troubling if it were to take them a very long time.

    At the end of the process a comparison of the correlation of the label numbers marked by the test taker on the page and their actual location can be made. One would expect the correlation to be about as a good an assignment as would be achieved by a random process if their translation method does not work. If there is a very close correlation between where they think the labels belong and where they actually belong then I think their analysis deserves further investigation.

    I am sure this test could be refined. I recall the kinds of tests people like James Randi or Richard Dawkins have applied to people claiming to have special abilities or alternative medicine using techniques akin to double blind trials; this is not an exact parallel, but you get the idea. I am particularly reminded of a test Richard Dawkins applied to “Dowsers”, these are people who claim to be able to divine where there is water.

    I think this example can be found at:


  19. Nick,

    We agree with respect to the main topic of this thread. Anyone that only employs transliterations is likely throwing baby out with the bath water.

    To accurately decode any text, written at any time, the cryptanalyst must travel back in time, dropping off modern metaphors that define modern culture, and pick up clues to think in the time period that is under study.

    Vulgate, orally and written, offers many knotty issues that I am unprepared to address, beyond the fact that oral vulgate was so debased by the 1200’s that many books were written on the importance of replacing it for person to person communications. Dante has been given credit, rightly, or wrongly for placing the final nail in the public death of oral vulgate in Itaty. It is said that when an issue of public syntax arises in written Italian that Dante’s advise should be followed.

    Written vulgate was upgraded into an academic context that one of our US Supreme Court justices, Scalia, studied as a youth, rising to the apex of the law, never learning the (vulgate) Latin that was used by his Sicilian father. Near his death, on CSPAN, Scalia said he had one regret in his life, for nit appreciating the imagery and beauties of the oral language of his father, (vulgate) Sicilian spoken in his father’s home town.

    Think outside the modern box, and travel back to the medieval 1300-1400s to attempt to break the VM code. Drop off modern thinking, even the transliterated views of the British Museum, and any like modern culture defending institution, where ever it spews out nationalist biased info.

    I think of Mexico’s national anthropological museum that dehumanizes native peoples by displaying them with blank faces, while filling in Conquestadore faces with blue eyes and full facial features. European and US museum offer subtle biases along the same lines, sad situations that must be erased to fairly study any cultural era not fairly represented in our school books, and cultural institutions.

    Best Regards,


  20. Josef Zlatoděj Prof. on November 12, 2017 at 10:03 pm said:

    Your research is not good. You are always at the beginning. At this rate, you will not achieve a positive result in a hundred years.

    Nick, you want the key to the translation ? ( MS – 408 ).

    Milo. Think about it. Greek alphabet.
    1. Alfa = A.
    2. Beta = B.
    3.Gama = G.
    4. Delta = D.
    5. Epsilon = E.
    Jewish substitution.
    1. = A,I,J,Q,Y.
    2. = B,R,K.
    3. = G,C,S,L.
    4.= D,M,T.
    5.= E,H,N.

  21. Nick,

    Your personalized myopia is amazing per […]

  22. Milo: if you’re going to complain about people’s myopia, you might consider looking to see who actually posted the comment you’re complaining about. [Hint: it wasn’t actually me.]

  23. Nick,

    Again, you are superficially correct. Someone named D posted the idiotic
    claim that the “Liber Abaci” was not Latin Europe’s only math book from
    1202 to 1454. Without D being identified, you may have used the pseudonymn
    …unlikely off course, but your terse erasure of my pertinent points means,
    to me, that you may side with D…both of you were under educated in 20th century classrooms, neither of you had have access to the English transliteration of LE Sigler’s 2002 Engllish version of 1202 Latin text. The 1202 Latin math text
    Determines its own context by statements and proofs, number based facts that are vivid, and can be double checked by anyone with a pencil, paper and a
    Math brain.

    Returning to the main thread, Cheshire ,did not offer a bilingual Vulgate
    Written language. What was offered as evidence were fragments of
    Oral words and phrases connected across Southern Europe. Silly is the
    highest value of a well written paper that reached unverified conclusions.


  24. Milo: if you want to know who “D.” is, it’s Diane O’Donovan (who is definitely not me).

    Whatever points that she and I agree on are few and far between, and don’t seem to include anything to do with the Liber Abaci, whatever your interpretation. My erasure of most of your comment was to try to prevent you making a fool of yourself, but it seems that I was somewhat wasting my moderating effort in that regard.

  25. Mark Knowles on November 16, 2017 at 10:28 am said:

    In regard to Gerard’s theory, that of others and our own theories.

    I must say first I have found Gerard from my interactions with him to be a very polite and friendly person albeit with a deeply mistaken confidence in the validity of his own theory. He has clearly invested a significant amount of time and effort in constructing it. Most people who say that they can translate the Voynich believe their theory is correct and can get angry and annoyed when others reject their theory. (I should say of course that I am not expert on Human psychology.)

    I guess admitting that you are mistaken after trumpting your theory can result in embarassment; we humans can often be motivated by ego. People naturally experience disappointment when their theory in which they have invested time and emotional commitment is proven false. I think wanting your theory to be true can often overtake a sober reflection on a theory’s pluses and minuses. This all makes it very hard for people to admit that they are wrong.

    Having said that I believe people who don’t construct theories add little value. Some of us have big theories about the manuscript as a whole some of us have little theories about a small part of it. Ultimately all attempts at really significant contributions are, at this time, theories not certainies. It is true that an individual can make empirical observations about the manuscript, without any interpretation made, but these only take you so far. Progress is made through process of speculation and theory building. So some people can hide behind not constructing theories and thus it is impossible for anyone to say their theories are wrong. Progress can be made sometimes through incremental collective efforts, but other times one individual makes a big leap.

    We all need to turn the mirror inwards to see if our theories stack up.

    Ultimately we are left with the question:

    Am I also someone mistaking these same mistakes or is my theory more than this?

    Often, but not always there is no rigourous objective way to be certain of this. So we inevitably to some extent have to fall back on subjective analysis which often meaning relying on instinct. We need a very health dose of doubt in our theories, but also, which might seem contradictory, a determination and confidence to keep pursuing your theory and its consequences to their limit; being crippled by doubt does not help. One needs to listen to others, but also have the self-confidence to ignore others as we know there have been thinkers who were right, but widely ignored or dismissed. Most of all one always needs an openness to change one’s mind and potentially reject one’s theory and construct a new theory.

    I think what is vital in analysing a theory is the predictions that it makes. So if for example the translator were to translation a line like “I, Prince Gunter of Hamburg, wrote this manuscript in 1412.” (fictional).Then this is a prediction that can be explored. Or if they translate a line a saying “This is thyme”,or in a specific place on an astronomical chart “Monday” or somewhere else “Dusseldorf Castle” and so on. In addition making the rules of their translation method precise and repeatable is important i.e. their translation algorithm. I think very broadly speaking it is much more difficult to assess non translation theories of the Voynich objectively i.e. a theory of the manuscript which does not claim to produce a translation. It is difficult to distinguish whether perceived patterns are more than coincidences which result from normal random chance. Some may feel that it is straightforward to analyse a theory to tell if it is broadly speaking correct or not, in some cases this may be true in others I think it becomes more difficult.

    As some may know I favour the development of a rapid and simple testing procedure for new translators. This should make the assessing of a new theory much quicker and simpler especially given the number of theories out there. I have already designed and implemented one part of this process.
    Anyway, so I ask myself:

    Is my theory fundamentally flawed or is there real mileage in it?

    I think there is mileage in it, but then I would say that or would I?

  26. Nick,

    Fooling yourself that I am the bad guy may come to an end.

    Good-bye. Your personal attacks have not been subtle. It
    Seems number people, like myself, can not communicate
    with word people that defend their word space by throwing
    out personal attacks. I am moving on.

    Living in the present, as you do, and overlaying personalized
    modern views over the medieval VM is an odd practice the you
    will continue, missing many needles in the haystack.

    Transliterations, by anyone, offer very fuzzy windows to the VM era.

    Needed deeper translations of the “Liber Abaci” and other Arab
    and European unit fraction number theory and language exts offers
    reliable ways to visit the past, leaving modern classroom oversights

    A range of valid windows, math and language, to the politicized past
    are needed. Vulgate offers one class of Opportunity that few know
    how to exploit in a professional manner.

    Best wishes,


  27. Milo: you clearly still have no idea what the difference between Vulgar Latin and The Vulgate is. And that is not a “personal attack”, I simply fail to see how I can spell it out any more clearly. 🙁

  28. J.K. Petersen on November 17, 2017 at 10:08 am said:

    Mark Knowles wrote: “Most people who say that they can translate the Voynich believe their theory is correct and can get angry and annoyed when others reject their theory.”

    Since I am amongst the loudest detractors (having written blow-by-blow commentaries on Bax’s, Gibbs’s, Lockerby’s, and Cheshire’s “translations”) I want to make it clear that I am *not* rejecting their theories.

    The theory is not usually the problem (although forming one too soon can hinder rather than help)… it’s the arguments *for* a particular theory and the actual results that either

    1) don’t make any linguistic sense or
    2) yield only a handful of words using a method that doesn’t generalize to
    the rest of the manuscript or
    3) which subjectively manipulate the results by cherry-picking, anagraming,
    or positing one-way ciphers.

    We are constantly bombarded with translations based on unsubstantiated statements and assumptions. That’s not a fault of the theory, that’s a problem with the translator’s knowledge base and his or her method. There is also a notable unwillingness to offer a range of explanations for data that clearly has more than one possible interpretation.

    People who know Latin (or think they know it) argue that the VMS is Latin. Those who know Arabic, argue that it’s Arabic. Those who know Czech, argue that it’s Czech, those who know English, argue that it’s English, and so on…

    How can this happen? Because the VMS is a string of glyphs that resemble vowels carefully interspersed with glyphs that somewhat resemble consonants but which are ambiguous enough to interpret either way. What most translators fail to notice is that a disproportionately large number of glyphs only appear in particular positions.

    Hundreds of pages of text will *always* yield a few words in any language (I have found hundreds) and a few might even be adjacent to drawings that appear to match.

    I’m quite open to theories about it being abbreviated text, or Latin or Czech or Arabic or Georgian or Armenian or auto-copied, or synthetic, or what-have-you. I don’t reject any of these possibilities. What dismays me is that those who claim to have solved it, regardless of their theories, have not made solid, cohesive, defensible arguments.

    If you genuinely study Voynichese meta-structure, you will see it has a unique positional dynamic—one that goes unacknowledged (and perhaps unnoticed) by the majority of researchers who claim solutions. If you don’t study the meta-structure, then whatever you think you see may be an illusion.

  29. JKP: I heartily agree with almost all of your comment. However, I should perhaps caution that the idea that Voynichese has a meta-structure quite distinct from its structure carries an innate presumption that we can necessarily discern structure from meta-structure, and that is quite a problematic claim. But this is arguably just as mucha commentary on those (particularly linguistic-minded) theorists who propose that they have identified Voynichese’s core structure and that all the annoying remainder is some kind of meta-structure which they can airily hand-wave away.

    It may well be that some of what gets typically treated as meta-structure might genuinely be meta-structure: for example, I can quite conceive that Neal keys might well prove to be no more than a textual way of highlighting a short section of text, that was perhaps even written in red ink in the original. However, until we can even start to say definite things about how Voynichese works, we should probably try to avoid prematurely separating structure and meta-structure. 🙂

  30. D.N.O'Donovan on February 8, 2018 at 10:36 am said:

    Dear Milo,

    I am not sure which languages you prefer reading, so I can’t offer you a serious sort of reading list. I have taken this from one of the web-pages which turned up when I googled ‘Medieval mathematics text’.

    From the 4th to 12th Centuries, European knowledge and study of arithmetic, geometry, astronomy and music was limited mainly to BOETHIUS’ translations of some of the works of ancient Greek masters such as NICHOMACHUS and EUCLID. All trade and calculation was made using the clumsy and inefficient Roman numeral system, and with an abacus based on Greek and Roman models.

    By the 12th Century, though, Europe, and particularly Italy, was beginning to trade with the East, and Eastern knowledge gradually began to spread to the West. Robert of Chester translated AL-KHWARIZMI’s important book on algebra into Latin in the 12th Century, and the complete text of Euclid’s “Elements” was translated in various versions by Adelard of Bath, Herman of Carinthia and Gerard of Cremona. The great expansion of trade and commerce in general created a growing practical need for mathematics, and arithmetic entered much more into the lives of common people and was no longer limited to the academic realm.

    The same little article does go on to talk about Leonardo of Pisa (Fibonacci).


    You can still some pre-Fibonacci maths texts if you like. Just look online for the names I’ve written in capitals and add ‘manuscript’ after them. Or go to just about any great library’s scanned books and look for them. Hope this helps.

Leave a Reply

Your email address will not be published. Required fields are marked *

Post navigation