To my mind, there are two basic types of Voynich Manuscript researchers: (a) those who view Voynichese as a language composed of clearly legible individual letters (and who therefore tend to treat it either as a confounding linguistic puzzle or as an exercise in pure cryptology); and (b) those who believe that you would first need to work out how to parse groups of glyphs into tokens before you can even begin to make any sense of the text.

Despite having made the case for (b) back in “The Curse of the Voynich” (2006), I don’t honestly believe that this second group’s camp has ever had more than my tent in it. (An occasional marauding bear, perhaps, but that’s about it as far as it goes, company-wise.)

Why is “Camp B” so empty?

Strongly-paired Glyphs

The argument starts with the difference between strongly-linked glyph pairs and weakly-linked glyph pairs.

In Voynichese, EVA ‘q’ is almost always followed by EVA ‘o’ (5186 times, compared with about 120 for all other occurrences of ‘q’). The strength of this link suggests the presence of an underlying orthographic rule (i.e. “q is always followed by u”), and also that a fair few of the other (non-qo) instances may well prove to be copying slips.

Similarly, if we see the first half of a strike-through ‘ch’ character (i.e. ‘c’) in front of a gallows character, it is almost always matched by the second half of a strike-through ‘ch’ character (i.e. ‘h’). This too suggests that c+gallows+h is following some kind of underlying orthographic rule:

* cth 905:33
* ckh 876:26
* cph 212:6
* cfh 73:6

However, it then turns out that Voynichese is full of families of strongly-linked glyph pairs, and that (though I don’t have precise statistical evidence for asserting it) it is these strong links that drive much of the structure and statistical behaviour of Voynichese.

* ‘qo’
* ‘ol’, ‘al’, ‘or’, ‘ar’
* ‘ee’, ‘eee’, ‘eeee’
* ‘aiv’, ‘aiiv’, ‘aiiiv’
* ‘air’, ‘aiir’, ‘aiiir’
* ‘ok’, ‘ot’, ‘op’, ‘oh’
* ‘dy’ (though I suspect dy works in a different way to the others)

That is, the amount of genuine information inside these groups is very small: which conversely, in my opinion, means that we should not be trying to look for information inside these groups at all. The real information in the text lies in the choice between these strongly groups, not inside each strongly-linked group.

Reading Jelly vs Parsing Foam

As a result, when I look at Voynichese words such as ‘olchedy’ and ‘olcheey’ (which occur a respectable 71 and 17 times respectively), I can only sensibly parse them as “ol-ch-e-dy” and “ol-ch-ee-y” before even beginning to try to make sense of what is going on with them. And even once you have parsed them, they remain just as inscrutable as before.

All of which is to say that I think we cannot yet parse Voynichese reliably, which is the starting point for the single-tent Camp B described at the top of the post. Yet this does not mean that all is lost: it just means that we are still trying to find a reliable and strong way to get started on a difficult road.

But linguistically, this isn’t how languages work. Orthography is driven by issues such as consonance and assonance: but what we appear to be seeing here is more like a jelly of letters (i.e. more structured than soup, but still quite plastic), joined together into words by deeper rules we are still unaware of.

Yet perhaps a more useful (and visual) way of viewing Voynichese is as a ‘foam’ of small glyph-group bubbles, (e.g. ‘ol’, ‘qo’, etc), empty of meaning in the middle but with all the semantic content on their outside at the point where they touch other bubbles. What I’m trying to do is to decompose the foam of words into its constituent bubbles.

89 thoughts on “Voynichese letters vs glyph-groups….

  1. You’re not entirely alone…

  2. D.N. O'Donovan on October 9, 2016 at 7:57 am said:

    You must forgive the marauding bears. A bear who acknowledges the rights of a mere human loses the Ursine moral high ground, you see.

    There are only two kinds of Voynich researchers – those who define research into a fifteenth century manuscript as study of its written text, and … all the rest.

  3. Nick, I think if you review very slightly the Voynich alphabet, you will realize that there is indeed a lot of visible and legible words. For example olchedy may be simply the Greek word ολκοτης – itself ?
    Besides, I found only 38 occurrences per .

  4. Ruby Novacna: in my own review of the Voynich alphabet, it seems that there are about ten English words and ten French words that randomly appear. However, I think it would be somewhat surreal to conclude from those statistics that each set of ten words ‘solves’ 0.001% (or whatever) of the Voynichese dictionary, and that all we would need to do is find the other 99998 languages that each contributed their own set of ten words to Voynichese.

    Given that Stephen Bax’s own attempts at translation yielded slightly fewer words, I think the correct inference to draw is that Voynichese’s word generation ‘engine’ will likely match about ten words of any given language purely randomly.

    I used Takeshi Takahashi’s transcription for that example.

  5. bdid1dr on October 9, 2016 at 3:29 pm said:

    Ho hum, Nick. I see that my translations are not to be considered, even if I have translated the Spanish and Nahuatl dialogues which accompany every pictorial item in the so-called “Voynich” Manuscript. Mr Voynich would probably be mortified by the usage of a code system called the EVA.
    Ambassador Busbecq would probably feel insulted that all of his efforts to return a couple of hundred manuscripts to Europe ended up being viewed as ‘puzzles’ to be played with — including the development of a ‘decoding’ alphabet which obscures, even more, the meaning and the language of a manuscript which was not written in any kind of code.
    You may be able to learn more about the Sahagun family if you are able to correspond with the priests and citizenry of the very small town of Sahagun which is/was within “Leon Province” Spain. The nearby university there (Salamanca) has an enormous archive. I’ve reiterated several times that there is no code in the manuscript (which we now identify as Boenicke manuscript 408).
    The very best translating aid can be found in the “Florentine Codex” — which was Fray Sahagun’s last effort to educate the Mexica of South America — in their own language.

  6. bdid1dr on October 9, 2016 at 3:43 pm said:

    What does not appear in either mss B-408 nor the Florentine Codex — any mention of Ferdinand and Isabella’s murderous campaign to eradicate the Jewish populations of Spain and Portugal. Also, there is a letter from Ferdy and Izzy telling X-topher Columbus to sail the oceans to find a sea route to China — and to claim for Spain any other lands he happens to discover.
    Yes, folks, the letter spells Christopher Columbus’ name as X (Christ)-pher


  7. bdid1dr: if I had a dollar for every time people disagreed with my ideas… I’d be feeelthy rich by now. Rather than complaining, why not think how rich you would be too? 🙂

  8. Josef Zlatoděj Prof. on October 9, 2016 at 5:34 pm said:

    Hi. Hi. 🙂
    I like it. ( Bear, Ursini ). Ursine word is good.
    (Oldřich z Rožmberka). Ulrich von Rosenberg asserted itself. The genus comes from Italy. And it is a genus Ursini. Of course he lied. Because I discovered, so Rosenberg family came from Poland.
    Founder of the family, was Vítek z Prčice and Plantenberga.

    Otherwise, the Rosenberg family was the first after King.
    Nick, you know at least that appointment Hadbank Michael Voynich, as the author of the manuscript ??
    Everything is written in his letter, which is in the Beinecke Library. Michael, manuscript worked for about 16 years. And he found it quite naturally.

  9. D.N. O'Donovan on October 10, 2016 at 12:46 am said:

    Surely Talbot and Tucker are the best people for you to get in touch with? They certainly believe the work is partly Nahuatl.

    It is very frustrating to have no distinction made between an “idea” and what is fairly called a large body of work and the conclusions drawn by the researcher. But that is an entrenched habit, now probably ineradicable, and I doubt that either of us will change it.

    Getting in touch with Talbot and Tucker, or with the Comeguys brothers who first made the suggestion and researched it (though T&T have so far fallen quite a bit below the normal standards by failing to provide fair or properly transparent acknowledgements there, I’m sad to say).

  10. Zodiac on October 10, 2016 at 7:03 am said:

    Nick, is there any metric (or perhaps many metrics) which we could use to test whether the chosen parsing has resulted in any clearer understanding?

    Or put another way, if someone handed you a transcription of Voynichese guarenteed to be parsed correctly, what metrics would converge to a particular value only if the parsing was correct?

    I’m wondering if it’s possible to brute force different ways of parsimg the Takahashi transcription for example, to see if anything makes more sense. It should be possible to try thousands or millions of combinations to see if anything sticks..or use a genetic algorithm to more quickly converge on a result.

  11. Zodiac: that is the question! When people talk about cryptology, parsing rarely comes up: in many ways, that’s because parsing operates in the domains of steganography, writing systems and orthography rather than any code or cipher system. As a result, there is – to the best of my knowledge – no academic literature on what one might call “parsing-to-make-ready-for-cryptology”.

    So… no, I don’t have a metric for parsing success, and that’s something I ought to post about separately. 🙂

  12. Nick,
    Something came to my mind about Voynichese structure and I thought I’d ask you about it first, since in a way it’s about encoding.
    Has it ever been tried the othef way around? See what has to be done to a text in order to obtain somethong Voynich-like? I reckon it would be incredibly hard with a plain text, and might reveal a thing or two. Do you think it is possible at all?

  13. Koen: I’ve tried a fair few experiments, and my conclusions were that a Voynichese word has a lot less ‘content’ than a plaintext word in pretty much any language, however you try to transform it. And then you have the problem of where ‘and’ and ‘the’ might be to contend with too, never mind ‘of’ etc. Oh, and where the numbers are as well. 😐

  14. Thomas F. Spande on October 10, 2016 at 8:32 pm said:

    BD, In response to your post of 10-9, I would like to comment with the following: 1492 was a watershed date for the Western world.

    1. Ferdinand (from Navarre) and Isabella ( who was originally Portuguese) had finally achieved their last of many campaigns against the Moors by their victory at Grenada. They were actually on scene and during this time they were approached by the Jewish treasurer of the Royal mint who proposed bankrolling a voyage by a Genoese mapmaker and mariner, Christopher Columbus, (who incidentally allegedly also had a home on the island of Chios). The proposition put to their Magesties was that the Genoese mariner, leaving from Barcelona, but turning to the West might find a new route to the spices of the East, that would evade the Arab control of this trade.
    2. In gratitude for their victory at Grenada, their Royal Highnesses did something really stupid; they expelled the cream of their artisan class, their Jews who were their diamond cutters, paper makers, etc. Evidently some were allowed to convert.
    3. As a consequence of the expulsion of the Moors and Jews from Spain, the Pope gave the Spanish, a dispensation from the usual requirement of fish on Friday. The Spanish to show their super-Catholicism doubled down on fish eating. The bonus for modern travelers in Spain is that a good fish meal is always at hand, every day of the week.
    4. Columbus found that the Americas unfortunately lay athwart any spice route BUT gold could be had. The royal focus now shifted to gold which they blew threw in a hurry funding mercenary troops fighting Protestantism in Northern Europe, particularly The Netherlands where the Jews from Spain had ended up, taking advantage of the religious freedom the Dutch offered. Soon the Dutch were a major European power, from immigration of skilled artisans, including Protestant cloth dyers and bleachers from France and also active trading in the spice trade in Indonesia’s Banda islands.
    5. The Spanish were not alone in seeking a new and safe route to the spices of the East. In 1498, the Portuguese explorer Vasco da Gama, using a route around the Cape of Good Hope reached Ceylon and the Malabar coast of India and their spice riches. Also they and the English developed clove plantations on Zanzibar and an adjoining island.
    6. The Age of Discovery was fueled mainly by a search for cheaper, faster routes to obtain spices.
    7. A working hypothesis is that part of the VM might be a glossary of spices, their curative properties and horticulture, from Indian or other Far Eastern venues, and is written in a code borrowing some from Tamil or Hindi (e.g. the gallows glyphs) but also using Latin that was their native language. The reason for encryption was to evade Arabic nosiness. I think a Genoese creator fits the bill.
    8. One key question remains and that is: We know the vellum was created in the early 15C but I remain unconvinced that the text was necessarily created later. Since most agree that the VM text is a copy (by two scribes, working together), the text could predate the vellum creation. Because the original plain text or even an original coded text has so far not surfaced, the task remains to date the writing by some embedded textural or delineation clues. A search for some text abbreviations, like “etc” and “&” whose use could be dated or some wrinkles in the Tironian notation (that Nick has commented upon) could be researched. Then drawing techniques like the parallel hatching that Nick has commented upon, nymph hair styles, maybe the use of luffa sponges as might be seen in one bathing scene, etc.
    9. One could date coloration and the pigments used, ink composition, including maybe carbon-dating of any included sugar or glycerine would be useful but is not relevant to the question of WHEN THE TEXT WAS WRITTEN, presumably initially in a non-encrypted plain text.
    10. I apologize for the rambling nature of this post but I think it addresses an important question that so far has just been greeted by assumptions.

    Cheers, Tom

  15. Thomas F. Spande on October 11, 2016 at 12:37 am said:

    Dear all, A last punctillio is that perhaps we can date any putative original plain text by the formatting of the text as in the paragraphing. Some folios are not paragraphed (e.g. 17v) but most are, either by the loose or tight scribe. Paragraphing is ancient (Greek “paragraphos”, 300BC onward) until paragraphing marks came in c. 1100 AD, starting with a small C, then C with one vertical line through it, then two lines and then the C part is blackened to form what we now call a pilcrow, sort of a reversed and darkened “P” with two ascenders. As Nick first indicated, the pilcrow led, with the advent of printing to the paragraph indentation, leaving room for the compositors to insert the pilcrow sign later.

    My daughter suggested to me years ago that the gallows glyphs might serve the role of a pilcrow and might have been arranged in a sequence for some reason, maybe numbering the paragraphs. I tried to find some reason for the gallows but too many problems persisted, such as gallows being too close to one another to have been used for a paragraph or even a very short sentence (like one word!). One scribe might have preferred one gallows form, like the two forms with the single ascender while another would prefer the double ascenders. Both would use often use mixes.

    Formatting of hand written ms or printed text, when using spacing between paragraphs or obviously an ending sentence shorter than the preceding lines, such as found in the VM, will not use indentation. The spacing is sufficient without relying upon indentation. No indentation may also just indicate that the VM ms was not heading for the printer. So no joy there!

    The word pilcrow comes from 1440 AD French, although I found one example, c. 1400 in German. I think paragraphing and the search for pilcrows is a dead end and text formatting studies are not going to help with the dating of any putative original VM plaintext. Cheers, Tom

  16. Hello Nick,

    there is a logic behind the so called “copying slips”. If you count them, there numbers form a geometric series. There are 5425 ‘q’-glyphs within the VMS. Using the transcription of Takahashi in 5291 cases ‘q’ is followed by ‘o’.
    ‘qe’ occurs 66 times,
    ‘qc’ occurs 23 times,
    ‘qk’ occurs 19 times,
    ‘qy’ occurs 8 times,
    ‘qa’ occurs 6 times,
    ‘qs’ occurs 2 times,
    ‘qp’ occurs 2 times,
    ‘qt’ occurs 2 times,
    ‘qf’ occurs 1 time,
    ‘ql’ occurs 1 time,
    one time ‘q’ is standing alone and 3 times the glyph next to ‘q’ was unreadable.

    Similar geometric series also exist for every other glyph. For instance, for ‘d’:
    ‘dy’ (6850), ‘da’ (4076), ‘do’ (548), ‘dc’ (355), ‘ds’ (185), ‘de’ (129), ‘dl’ (82), ‘dk’ (33), ‘dd’ (23), ‘dr’ (14), ‘dg’ (9), ‘dp’ (9), ‘dt’ (7), ‘dm’ (6), ‘di’ (5), ‘df’ (1), ‘dx’ (1) and 635 times ‘d’ is the last glyph.

    Because of this series it is not possible to apply rules like ‘q’ must be followed by ‘o’ or ‘d’ must be followed by ‘y’ or ‘a’.
    If you start to call for instance ‘ql’ a “copying slip”, what is with glyph sequences like ‘qf’, ‘qt’, ‘qp’, … and in the end ‘qe’? Are they all copying errors in your eyes?

  17. Torsten Timm: I might be wrong, but you seem to be assuming that if there is some kind of underlying orthographic or systemic rule that means ‘q’ should always be followed by ‘o’, then the only way another letter can follow ‘q’ is if an ‘o’ has been mutated into that letter.

    That is not even remotely how I see “copying slips”. A word-start like ‘qot-‘ might have the ‘o’ omitted, yielding ‘qt-‘: or a different word (say, ‘lchedy’ in Herbal-B) beginning with ‘l-‘ might have been miscopied as beginning with a ‘q-‘ (i.e. ‘qchedy’). Furthermore, Herbal-B has some instances of what I call ‘freestanding q’ characters (and which I suspect may be an improvised cipher for the word ‘water’ in the plaintext), which further complicate the picture.

    Capturing statistical distributions is one thing: but interpreting them as necessarily being geometric series is an interpretative step, one that I find too reductionist at this relatively early point in our understanding of Voynichese.

  18. Hello Nick,

    the type of change doesn’t matter for my argumentation. My point is, that it is not enough to argue that a glyph sequence is occurring less frequently and was therefore probably “miscopied”.

    It is a fact that the number of ‘qe’-pairs (66) is more frequent then the number of ‘qa’-pairs (6) and that the ‘qa’ pairs are more frequent then the ‘qt’-pairs (2). This geometric series shows that your “miscopied” words and “copying slips” follow the logic of the VMS. They are not errors they are part of the system.

    Or to say it with other words: The VMS is full of spelling variations. This is the most intriguing feature of the VMS. There are all types of variations occurring with all types of frequencies. Therefore, the question arises, how can you distinguish between normal “spelling variations” and “copying slips”?

  19. Torsten: my point about ‘qo’ was that it occurs ~5186 times throughout the VMs, while all the other q+[not-o] pairs occur about 120 times altogether, which is two orders of magnitude difference… a very big difference indeed. Though I would tend to agree that the VMs does appear to be full of spelling variations, the fact that the ‘qo’ is such a dominant – and apparently such a rarely misspelled – letter pair would seem to suggest that some parts of the writing system are less susceptible to misspelling than others. 🙂

  20. Thomas F. Spande on October 12, 2016 at 6:00 am said:

    Nick, The stats you relate above for c-x-h where x is one of the four gallows glyphs assigned to a fixed English consonant seems at variance with our earlier shared conclusion that the extreme variability of the single- and double-stemmed gallows in the VM text meant that the glyph represented by “x” cannot always be immutably the same? Has your thinking on this changed?

    One very simple glyph pair that appears to have slipped past everyone in this potentially fruitful attempt at code-cracking is the very simple “89” combo. At the risk of being a chipped recording or CD, I will reiterate an argument I have put forward for years. yep, it involves the Armenian language.

    The argument is simple: Armenian was created ca. 400 AD and unlike most of the Middle Eastern languages is read from left to right, not like Hebrew or Arabic. Furthermore it is a language using 39 glyphs and no accent marks. Some glyphs, like the tipped “2” that occurs in the VM amount to the Czech tiny v above a “c” ,for example, that makes “kovac” into “kovach”.
    The paleography and letter shapes of the language has changed drastically over the years but the underlying language has not. Their glyphs for “o” and “f” were added in the 14th C.

    Anyway I think it can be proved that “89” using indo-arabic numbers, not true arabic numerals is equivalent to the Latin “et”. Many early languages used letters to stand for numbers, when their alphabets were laid out from “a to z” Armenian, Hebrew and Arabic among them. When Armenian is arranged starting with their equivalent of “a” which is assigned “1” we end up with “8” and “9” for the eighth and ninth letters of their alphabet. When Latinized, those Armenian letters are translated into 8=e and 9=t and the resulting glyph pair “89” then was the Latin “et” or the English “and”. So this thin wedge can be used to assign 8 as always e and 9 always as t although I think there is redundancy with Voynich “e” also being supplied by some other letter combinations like a pair of “c”s and possibly the same with “t” But 8 is always e and 9 is always t. This combo can stand for the Latin conjunction “et” but may also be used as the letters “et” in the middle of a “word”. If one accepts this simple assignment of the “89” glyph pair, then some Voynichese can be cracked. Other Armenian glyphs that occur in the VM are the “&” like character without the bottom line crossing and that is the Armenian “f” when Latinized. So 8a”&” becomes “eaf” and may with a preceding glyph be “leaf”; 8a”2″ where the “2” is tipped may be “each”. Here is agree with you, that some English has found its way into the VM.

    I have sponsored the idea that some Armenian glyphs appear in the VM, and while Bax seems also to have accepted this possibility, he seems not to have understood employing it, judging from his translation attempts.

    I don’t think however that Armenians had a hand in the actual composition of the VM, just that certain of their glyphs were used. Incidentally “4” is an Armenian letter not a number. Armenians were unlikely to have adopted the yin/yang symbology, that I think can be discerned in the botanicals, nor their interest in the nude bathing and calendar nymphs and the avoidance of any obvious religious iconography but who knows, maybe the exercise of composing the VM released some pent up weirdnesses?

    Cheers, Tom

  21. Nick, why oppose the followers of the theory of the plaintext and followers of encoded text?
    The possible presence of the plaintext is not detrimental to the idea of coding.
    I think all researchers, without exception, should first belong to Team A, since finding the plaintext is simpler. Everyone should participate in the building of some basic rules.
    And Team B will be composed of the elite of Team A, as it claims to have a more global vision.

  22. Ruby: it is Voynichese itself that opposes most straightforward suggestions (such as “it’s written in a single, simple language”, “it’s encrypted using a simple substitution cipher”, “it’s a purely random hoax”, “it’s encrypted using a polyalphabetic cipher”, etc), not me.

    Moreover: it has been widely known for over fifty years that these suggestions all fail, so I’m hardly being oppositional or difficult by pointing this out.

  23. bdid1dr
    Is the Florence codex in origin of the Voynich manuscript or not , I’ll like to read. Can you give me some useful links?
    Especially that on the VM first page any ligatures can be read as “σεβασμος ζωοποιοσ σταυρος” (see my post “Venerabilis and vivifica crux”). Is this, perhaps, a toponym, such as Vera Cruz or Santa Cruz?

  24. D.N. O'Donovan on October 12, 2016 at 11:52 am said:

    Dear Nick,
    It’s just as well no-one has resolved the problem of the text. Judging by the regularity with which the well-oiled plagiarist machine is working, they would find that their solution had been floated as a vague possibility by a single comment made twenty years ago in some private conversation, and that would be justification enough to assert that the person who did the work, demonstrated that their method worked, and rightly interpreted the text, was being ‘derivative’ and saying no more than what “everyone knows”.

    So effective is this ploy that by the time my two volumes come out, almost every one of my original discoveries will be found ( and credited to who-knows-whom) on sites like

    So my advice is – don’t.

  25. Nick: thanks for your response, it is interesting that you have tried to encode “towards” Voynichese, but that this is near impossible. I would have thought so.

    Such experiment shows the problem with most current attempts to read the text. I believe there are two possible explanations:
    1) The method of “encoding” (in the broadest sense) is completely different than what we currently imagine. For example, it could be one with a certain amount of information loss.
    2) Whatever lies underneath is not the kind of text people imagine.

    Or both… Probably both 🙂

  26. Koen: abbreviating shorthand reduces the amount of information per word, which is why I’m comfortable with the suggestion that it (or something very much like it) is one of the mechanisms in play here. But even that isn’t enough to explain Voynichese. :-/

  27. Diane: I look forward to the day when we can read even 1% of the Voynich reliably. The bonfire of the theories will last for days, possibly weeks.

  28. The scientific results of scientific research are published in scientific revues. But blogs are made of the theories bonfire.

  29. Ruby: I’ve seen plenty of so-called scientific research into the Voynich Manuscript that would be better in a bonfire than in any journal. And I’ve seen some very smart blog posts too. All I conclude from that is that studying the Voynich Manuscript can bring out the best or the worst in you, whatever your disposition. =:-o

  30. bdid1dr on October 12, 2016 at 2:52 pm said:

    8 = aes 9 = ceus / geus
    8 9 = excuse

  31. D.N. O'Donovan on October 12, 2016 at 4:02 pm said:

    Honestly, it’s not the manuscript that does it.

    Give me five people with the right qualifications and experience, and more interest in work than in piggy-backing, and no interest whatever in theories, and we could have it done before Christmas.

    “Breve breviarium breviter abbreviatum sufficit intelligenti” R.B.

  32. Diane: we obviously have very different views on this. If untangling the Voynich Manuscript was just a matter of applying sheer numeric force – whether 5 or 5000 – it would have been solved years ago.

  33. Thomas F. Spande on October 12, 2016 at 4:44 pm said:

    BD. So every page of the VM is an apology for something or other?

    Seems to me that Voynichers do deserve some apology for the possibility of having their collective time disappearing into a rabbit hole!

    One problem with my idea that “89” being simply “et” is that I am certain that the ampersand “&” also appears here and there in the VM text. Why both are there seems like conjunction overkill, I would have to admit.

    I plan a comparison with the occurrence of “&” by the looser and tighter scribes and if anything worth reporting shows up I will inform the membership.

    The name “ampersand” is curious and used to be part of the “ABC’s” learned by children (at least in the US). It was a mispronunciation for “(and per se “and”) as “and” was included with the alphabet. Weird, but “&” survives in the world of commerce and on our computer keyboards to this day.

    Cheers, Tom

  34. Hello Nick,

    it is well known that the second order entropy for letters in the VMS is low (see Benett 1976). With other words the order of letters within words is highly predictable for the VMs in generall.

    ‘qo’ is followed by ‘k’ in 3116 cases.
    ‘qot’ occurs 1130 times,
    ‘qol’ occurs 253 times,
    ‘qop’ occurs 167 times,
    ‘qoc’ occurs 139 times,
    ‘qod’ occurs 126 times,
    ‘qoe’ occurs 124 times,
    ‘qoa’ occurs 53 times,
    ‘qof’ occurs 43 times,
    ‘qor’ occurs 33 times,
    ‘qoo’ occurs 31 times,
    ‘qos’ occurs 15 times,
    ‘qoy’ occurs 14 times,
    ‘qoi’ occurs 12 times,
    33 times ‘qo’ is standing alone and one time the glyph next to ‘qo’ was unreadable.

    ‘qok’ and ‘qot’ occur 4246 times throughout the VMs, while all the other qo+[not-k|t] sequences occur about 1044 times altogether. With other words the ‘qo’ pair is part of dominant sequences like ‘qok’ or ‘qot’.

    You can go even further. In 1387 out of 3116 cases ‘qok’ is followed by ‘e’ and in 1012 cases by ‘a’ whereas all the other qok+[not-e|a] sequences only occur about 717 times altogether.

    In the end you would find that in fact words similar to ‘qokeedy’ (305 times), ‘qokaiin’ (263 times) and ‘qokal’ (176 times) are dominant. Since the spelling variants for common words like ‘qokeedy’ also the spelling variants like ‘qokeey’ (310 times) are common it is only logical that in the end also features of this words like the ‘qo’-pair is common.

    Spelling variants are so common for the VMs that they must exist on purpose. Therefore it is an error to name some of them “copying slips” or “miscopied” words.

  35. D.N. O'Donovan on October 12, 2016 at 6:43 pm said:

    Nick –
    it’s not about numbers . It’s not about cracking ciphers. It’s about provenancing the matter in a manuscript.

    I’d just like to be able to get on with it, and work with people don’t need to be told even once (let along repeatedly over 5 years till they get it) that he cloudband pattern is Asian etc.etc.etc.

  36. bdid1dr on October 12, 2016 at 8:52 pm said:

    ThomS & Nickolas:

    Years ago, I threw away my law dictionaries. Today, I can’t even find an often used expression: et seq … or et sequensentis. It appears that the use of Latin expressions (in or out of court) is becoming more and more obscure with every passing century. My studying ‘legal. ‘law’ terminology was for the benefit of the City Attorney (who was also a teacher of law at a nearby university/ city).

  37. Diane – 5 is really optimistic. I’d have to start with about 10 specialists, though I could certainly fire some after a while 🙂

    Nick – yes, shorthand and abbreviations, perhaps. In a case where a native speaker of one language attempts to write in another, we can also expect a simplification of the phoneme inventory. In a pidgin, there would be all kinds of other simplifications as well, like perhaps a more limited amount of word endings.

    And still, even giving it the best possible odds, it’s still not enough, is it?

    An additional solution I favor is to treat gallows as kind-of-capitals, which solves a few problems, but leaves many…

  38. Diane: as you know full well, I don’t believe that your reading of the historical evidence qua cloudbands is at all correct, and I suspect that you have not yet managed to convince anyone else of this either. But perhaps you’ll find five such disciples eventually, who can say?

  39. Torsten: all the while you present your hypotheses as flat assertions (such as “Spelling variants are so common for the VMs that they must exist on purpose”), you make it very hard for people to want to engage with any of your arguments. The presence of spelling variants doesn’t exclude copying mistakes, nor do copying mistakes exclude spelling variants. A few years back, I estimated the copying error rate: it came out at about 1 to 1.5 per line: this seemed quite independant of spelling variants.

  40. Torsten: …and I should add that calling them “spelling variants” is already a loaded term, in that it presumes that the words are being spelt in a conventional way using words from a conventional kind of dictionary, both presumptions that I genuinely don’t believe hold true.

  41. Hello Nick,

    in your last post you agreed to the statement “the VMs does appear to be full of spelling variations”. With “spelling variations” I mean similar words like ‘qokeey’ and ‘qokeedy’. Therefore it is no problem if we use the more neutral term “similar words”.

    Its rather simple, you say that the VMs contains “copying mistakes”. It’s your assumption. Therefore it’s on you to back it up with arguments. The existence of similar words does indeed not exclude copying mistakes. But that something can’t be excluded doesn’t mean that it is true.

    Maybe my argumentation was not clear enough. On one side the order of letters within words is highly predictable for the VMs in general. On the other side for every common word also similar words exists. For instance for ‘qokeey’ (308 times) the following similar words can be found: ‘qokeedy’ (305), ‘qokedy’ (272), ‘qokey’ (107), ‘qoteey’ (42), ‘qokeeo’ (23), ‘qoeey’ (15), ‘qoekey’ (2), ‘qkeey’ (1) …. That the order of letters is highly predictable can be explained by the existence of so many similar words. If many similar words exists it is no surprise if also features of them the ‘qo’-pair or the ‘eey’ sequence are common.

    If I understand you correct you argue that ‘qkeey’ is more suspicious as copying mistake then ‘qoeey’ or ‘qokey’ because of the missing ‘o’. My point is that it is quite normal for the VMs that similar words exists. For this reason it is impossible in my eyes to distinguish between good and bad similarities.

  42. D.N. O'Donovan on October 13, 2016 at 2:24 am said:

    I have never claimed to be the author of any history of the cloud band pattern, whether in its original home in Asia, or after its introduction to Europe during the Mongol period.

    Nor am I the author of any of the mainstream Dictionaries and Glossaries of Art where the information could have been checked when I first noted the error of an imagined German origin being disseminated. So correct that – it didn’t take five years for the message to get through.. eight years and still running. 🙁

  43. SirHubert on October 13, 2016 at 5:16 pm said:

    Hi Nick,

    Or indeed whether all glyphs or groups were originally intended to be meaningful.

    Just one other general comment, hopefully not too off-topic. It’s actually extremely difficult to fabricate long tracts of plausible gibberish. 240 pages of it would be excruciating. A random word generator is a pretty good way of getting round this and I know something similar has been proposed for the Voynich Manuscript on a number of occasions. But after a while, if I were given that exercise and paid by the page, I’d speed things up by writing generated words from memory sometimes. Who’d know? And who would care about spelling errors?

    I’m not saying I think this is how any or all of the text was created. But I strongly suspect that, if it was, nobody would ever have followed that system slavishly for all 240 pages. And that being so, it’s not such a strong counter-argument to claim that a proposed stochastic process doesn’t explain every observed feature of Voynichese.

  44. bdid1dr on October 14, 2016 at 4:07 pm said:

    The word ‘Espanol’ (with no diacritical marks) would appear as one of those very elaborate so-called pilcrows : Several months/years ago, I gave as an ‘x-m-p-ll’ the word R x ll : as used by a modern-day name for the drugstore chain “Rexall”.
    So: take another look at all of those elaborate “P” words — because that was exactly how Fray Sahagun’s Native American students/artists/writers were taught to create combinative words. Try it out for yourselves: write the word ‘ pilcrow’ — using only that one elaborate “P”

  45. bdid1dr on October 15, 2016 at 12:02 am said:

    PS: ( post script ) You might also like to take a look at Fray Sahagun’s “PSalmodia” (written in both in ESpanol & (and) Na-hua-tl……. mostly about the Church’s celebrations of Holy Feast Days.
    I am hoping that Professor Leon Portilla may be able to tell us what education at what university did Fray Sahagun receive before he was assigned to the “New World – Mexico —- and the exclusive School for Boys.
    PS : My husband has just gifted me with another book written by Professor Leon-Portilla (which he wrote while he was delegated to UNESCO (Paris, France.) Fascinating, and adds considerably to understanding the phonetics of Nahuatl speech and writing.

  46. Davidsch on October 16, 2016 at 10:45 am said:

    Pfoei, why are the replies always longer than the original blog text?

    If you want to use “tokens” are those always fixed at 2 long,
    are they variable, or are they 2,3 or perhaps even 4 long?

  47. Davidsch: probably because I don’t put up Voynich-related posts often enough. 😐

    As far as the length of tokens goes, I think that there are some that look (in EVA) like pairs (qo, or, ol, ar, al), some that look like triples (eee, aiv, air), some that look like quadruples (aiiv, aiir), and some that look as though something odd is going on but I don’t know what (k / ok / yk / ckh / ockh, t / ot / yt / cth / octh, d / y / dy, etc). And so the question is how we can we work out what makes up the correct token-parsing scheme… which I don’t have a good answer to.

  48. bdid1dr on October 17, 2016 at 4:59 pm said:

    @ Nick, a question: When are you going to close down the use of the EVA (as a decoding invention) since there is no code in the “Voynich” manuscript ? What you won’t find are any words ending in ‘ty’ — Nahuatl translators had only ‘tl’ for words ending in ‘ty’ — such as beauty, plenty, country….. Another interesting feature of Nahuatl : the ‘h’ is silent. The letter ‘ J ‘ is the sound of ‘ h ‘ — as in ‘ Jose’ .


  49. bdid1dr: EVA has only ever been a transcription tool, what people choose to do with it beyond that is their own problem. 🙂

  50. bdid1dr on October 17, 2016 at 5:11 pm said:

    Jesus : Nahuatl pronunciation of the holy name: would be sounded as hay-sus — and still is pronounced thus in Mexico/South America/North America (at least in the West Coast, Texas, and New Mexico.

  51. bdid1dr on October 20, 2016 at 5:33 pm said:

    Lecturer Diana Kerpel ” Colors of the New World”
    Professor Leon-Portilla “Fifteen Poets of the Aztec World”
    Walden Browne “Sahagun and the Transition to Modernity”
    Felipe Fernandez-Armesto “Pathfinders — A Global History of Exploration”
    Frances Karttunen “An Analytical Dictionary of NAHUATL
    Quite a few books on the history of the “New World” — one of which is “Race and Class in the Southwest — A Theory of Racial Inequality —Mario Barrera
    Just to name a few of the books my husband gifts to me.

    Because of visual, hearing, and speech deficits, I was considered somewhat ‘dull’ by my elementary school teachers. Hearing aids? Bah! Loud babble, because the teachers (invariably) scribbled on the blackboard while lecturing with their back facing the ‘audience/students’ .
    Thank goodness for “Toastmasters USA” and thanks to the publishers at University of Oklahoma Press (Norman Oklahoma and London) , and Notre Dame.
    Once again I urge y’all to read at least anything written by Professor Leon-Portilla — and “Pathfinders’ by Felipe Fernandez-Armesto.


  52. Davidsch on October 21, 2016 at 10:01 am said:

    @BD, here another one for your Christmas wish list:

    Getting Organized in the Google Era: How to Stay Efficient, Productive (and Sane) in an Information-Saturated World Paperback – May 3, 2011
    by Douglas Merrill (Author), James A. Martin (Author)

  53. Thomas F. Spande on October 22, 2016 at 12:41 am said:

    Dear all, What explanation, if there be any, can be put forward for the odd occurrence of 14 occurrences of c-c o “inverted gamma” in the last five lines of second paragraph on f56v? There are within these, two instances of this three character “glyphlet” occurring consecutively three times.

    Another example can be found in f47r, where ten instances of this three character “glyphlet” occur, one occurring three times in a row.

    An idea occurred to me, in view of the unusually high vowel frequency for “o” in the VM is that some of “o”s are redundant and can be struck out when following c-c, making this “c-c “inverted gamma”” and if c-c be assigned a vowel like “o” or “i”, these then become simply “on” or “in”. The repeats “on on on” or “in in in” are baffling. Maybe just glossolalia?

    It does appear from those useful graphs that “o” usually follows a “c-c” when that pair shows up. Maybe to the scribes this amounts to two vowels in a row with it being understood that the second is redundant? Just added to confuse the uninitiated? Cheers, Tom

    ps. I have not made any concerted search for this triglyph and just noted a few in passing during my search for the frequency of occurrence of the “&”.
    More on that anon. Both the loose and tight scribe use & and it is used in text where the “89” also occurs.

  54. bdid1dr on October 22, 2016 at 4:28 pm said:

    @Thomas (and you too, Nick) (or vice versa) :

    The c – c combination can be read by focusing on which “c” is smaller :
    Face lace or palace
    pneu mo cocci
    eat (a meal or catch a cat or cease what you are doing
    in all cases the smaller ‘c’ is actually “e”


  55. Thomas F. Spande on October 22, 2016 at 11:16 pm said:

    BD, Watch out for the weeds where I think you are headed! The VM scribes were good but not perfect. When examining ampersands or eights or what I think is the Armenian “f”, (has a rocker at the base and the bottom of the glyph is not closed up). a lot of variation is seen.

    My plan is to examine c-c o and c’-c o for their frequency in every folio and compare with the “o” present and absent and to see how this stacks up with the normal frequency of “o” in all known languages. Recall that “o” is way more prevalent in the VM than in Latin, Armenian, etc. That makes me think that many are nulls and if the c-c with or without a curlicue (to use a term you first used) might themselves be vowels. One idea at the moment is that “on on on” might indicate the plant being illustrated is a perennial. Cheers, Tom

  56. bdid1dr on October 23, 2016 at 5:41 pm said:

    @Nick: The “Voynich” manuscript’s alphabet does not -ave the letter h. Nor does it -ave the letters ‘u’ ‘ nor ‘v’ : oe eo — take your pick for words such as ‘vision’ and ‘word’ or ‘world’.

    The difference between the worlds’ words “word” “world” “would” become obvious and quite helpful oen one is trying to translate the so-called ” oue-oi-n-ch m n eo-script. No code: just a phonetical system which can be understood by persons with hearing or vision impairments — I have both.

  57. bdid1dr on October 23, 2016 at 5:58 pm said:

    Back to spinning some several hundred yards of yarn with my Navajo spindle — to make a very small rug (ruglet?). My loom is large enough to make an 8 by 10 foot ‘carpet’ — if I had the heart, heft, and health to do so. Jes’ say’n…….

  58. bdid1dr on October 24, 2016 at 3:59 pm said:

    Somewhere in Fray Sahagun’s rough draft (the Voynich) one can find another peculiar ‘glyph’ : ‘ itius/teus ‘ . It is not often seen until the opportunity seems
    ‘pro-pi- tius’ .

  59. Thomas F. Spande on October 27, 2016 at 6:05 pm said:

    Dear all, At the risk of getting Nick’s nose out of joint, I would first commend his recognition of the unusually high frequency of the pair “qo” in the VM text and his commentary of this glyph pair, BUT if one makes a jump to (you guessed it) Armenian, one finds (see “Armenian alphabet” on Wiki and numerical assignments to the Armenian letters), one finds that 8=e, 9=t (that I have discussed to a fare thee well) but more pertinent to the assignment of “qo” is that the “q” is actually a “4” and this corresponds to the Latinized or Romanized Armenian glyph for “d” so “qo” becomes “do” that I think is a lot easier to deal with. So dump the “q” and grab the “d”! Cheers, Tom

  60. Thomas F. Spande on October 28, 2016 at 5:39 pm said:

    Dear all, I note, as I am sure Nick and others have also, that “4o” is usually followed by a gallows glyph. If those be ignored (heresy!) then “do” is a good Latin verb for “put, give, cause, make, to permit, etc.”

    I am thinking at the moment that the gallows glyphs are an albatross around the necks of Voynichers and propose dumping the whole bunch of gallows as the “Mother of all distracting nulls”. They are too erratic to be real consonants. Let’s try some decrypts by striking out entirely ALL the gallows glyphs!

    More on the “&” anon. I think at the moment that they serve the role of punctuation. Cheers, Tom

  61. Thomas F. Spande on October 30, 2016 at 4:31 pm said:

    Dear all, Well I see my proposal of 10-26 that the gallows glyphs be considered nulls, has been greeted by a collective yawn! I cannot believe I am the first to propose this!

    Moving along to my analysis of “&” in the botanicals. I find the following: the looser scribe in 53 folios, uses “&” a total of 21 times with 20 being at the right end of the text, (maybe the end of a sentence?).

    The tighter scribe in 62 folios uses “&” a total of 46 times with 33 occurring at the right end of the text. So the scribes differ in the frequency of their use of “&”.

    When “&” occurs within a line of text but not at the end, it might still indicate the end of a statement, i.e. a “sentence”? I plan to recheck these conclusions as sometimes “tight” and “loose” sometimes are hard to assign and the “&” glyph can show some variation. Thus, the above data should be considered a tentative initial analysis.

    Back to an analysis of “c-c o “inverted gamma””, “c’-c o “inverted gamma”” and for good measure “c-c o “tipped 2″”. Recall that the “8”-like glyph that has a rocker at the base and is not closed up is the cursive Armenian “f”; the tipped “2” is Armenian for “ch” used in pronunciation as written Armenian is a phonetic language with 39 characters. The “o” and “f” that are counted among the 39, were added in the 14th C as they were considered “borrowed” glyphs from other languages ( maybe English) ?

    All for the moment., Cheers, Tom

  62. Thomas: explaining away everything we find difficult about Voynichese as nulls wouldn’t leave us much to work with. 🙂

    And remember that the Voynichese alphabet is very compact, so we don’t really have room for four characters to be nulls without really crippling the ability of the language to express anything, whether language or cipher. 😮

  63. Thomas F. Spande on October 30, 2016 at 9:10 pm said:

    Nick, I thought we had an agreement that the gallows glyphs of the four types DO NOT always represent the same set of four consonants. They are just too wildly variant. I tried a system with the glyphs representing groups of five consonants, with the simplest, (single stem, single-loop) being b,c,d,f,g; the most complex (double stemmed, double-looped) representing one of the last group of five, (t,v,w,x,z). In English, if one subtracts the vowels (including y) one is left with twenty consonants.

    In a recent analysis of mine of f24r, we have NONE of the single stemmed, single looped glyph; only three of the single stemmed, double-looped glyph; 21 of the double stemmed, single looped glyph and 22 of the double stemmed double-looped glyph. A text with mainly the last two glyphs seems more problematic than to dump the whole bunch as nulls. The huge variation and the differences between scribes seems to me that most or all the gallows glyphs cannot be consonants.

    If random, generated by some process (like throwing a dart) we are left with a mission impossible! Many have accepted the idea that their shapes reflect their identification with one consonant or another. I don’t think that works at all. Just the view of “Crazy Down in the Basement! ” Tom

  64. Thomas F. Spande on October 30, 2016 at 10:05 pm said:

    Nick, Sticking with f24r, I count 370 glyphs (not including the gallows) and 46 gallows glyphs indicating the gallows represent a total of 46/370+46 or 11%. In my opinion this leaves a lot of VM code to decipher! I think there may even be more nulls; then I would agree that we have cut too much of the tree down.

    Cheers, Tom

  65. bdid1dr on October 30, 2016 at 10:27 pm said:

    Nick & ThomS : Those ‘c c’ s can actually be read as either ‘c e’ as in the c e nter of the world or any word which is either ‘e c c e n tr i c’ or ‘c e n tl’ (central)’ boring. Get out your Latin (Espanol n this case) dictionaries, folks, and compare with the discussions which appear in both the so-called ‘Voynich Mss” and the huge publication (all 13 ‘books’ in the Florentine Codex ; of which I have “Book ll
    Earthly Things.
    You will then be able to bypass Boenicke Library and Paula Zyatz” unsuccessful efforts at translating the so-called “Voynich Mss”. BTW, the so-called “Voynich” manuscript was returned to Europe (along with a couple of hundred other elderly manuscripts by Austrian diplomat Busbecq. Busbecq was then directed to deliver the manuscripts, Arabian horses, and zoo animals to Rudolph II.

    Shortly after Rudolph was incarcerated (as being crazy) by his relative Max (?) the battle of ‘White Mountain” pretty much threw Europe (and New Spain) into the chaos of Protestant vs Roman Catholic ‘religious’ war.

    BTW: I chose the very large book “Eleven – Earthly Things” (in the Florentine Codex) because Anderson and Dibble (and University of Utah) did a superior job of translating and portraying each and every photo reproduction of the contents of the Florentine Manuscript.
    Just recently my husband gifted me with one of Prof. Leon- Portilla’s earlier works: “Fifteen Poets of the Aztec World”. In this small book are translations of songs/poetry “Nican ………… ”
    Not too long ago, I was able to listen to (and translate a song: “Nican Mopohua” .
    I’ve always wanted to know what the song was about.

  66. Thomas F. Spande on October 31, 2016 at 6:40 pm said:

    Nick, Yet more on f24r that was picked totally at random, not originally to make any particular argument.

    That folio has 20 lines of text with the last line consisting of only 8 glyphs and no gallows. The arrangement of the three gallows present, starting with the simplest, a single stemmed double-looped glyph that folks have assumed stands for “p”. There are two on line 1 and one on line 19 for a total of three occurrences:

    For the more prevalent gallows, the double stemmed, single-looped gallows that has been assigned to the consonant “k”: there is one occurrence in lines 4 through 8, 10, lines 15-17 and line 19.

    The “k” glyph occurs twice in lines 3, 9, 11 and 12 and three times in line 2.

    The “t” glyph ,i.e. the double stemmed, double-looped glyph, occurs once in lines 2, 4, 10, 13 and 19. It occurs twice in lines 5, 7, 12 and 17. Finally “t” occurs three times in lines 3 and 6.

    I think this analysis indicates the gallows glyphs for “k” and “t” are sprinkled liberally throughout f24r with roughly the same frequency.

    Furthermore, if one consults the “letter frequency tables of Wiki, where 15 languages are broken down into the frequency of occurrence of ALL the letters, there are only a few languages where the frequency of “k” is significant. It certainly is not English, French, German, Spanish, Portuguese, or Italian. But Turkish (k>t) and Polish (“k” is approximately the same as “t”) defy the Romance languages. In the remaining languages, Swedish, Danish, Dutch, Finnish, Icelandic and Czech, “k” is frequent but always significantly less than “t”.

    As an English exercise, I took 8 lines (459 glyphs) of “Curse…” from p. 51 and found in them only 2 “k’s” to 40 “t’s”, confirming that “k’ is rare in English.

    Perhaps the current assignment of t, k, p, f to the four gallows glyphs needs to be reexamined?

    p.s. Incidentally, It appears that Bax has bought into Nick’s assignments, judging from his attempts at decryption.

    Cheers, Tom

  67. bdid1dr on October 31, 2016 at 9:46 pm said:


    Frequency of the ciphers are not going to give you even a small clue to what is being discussed. What will help in solving the mysterious scripts and illustrations in the VMS, is to obtain a copy of Fray Sahagun’s Florentine Codex, “General History of the Things of New Spain — Book Eleven –Earthly Things. If you are not inclined to spend any more money or time on the so-called “Voynich” manuscript, this one large soft-bound book will help you solve most, if not All, of the contents of the “Voynich” manuscript — even if you are not able to provide provenance .
    Thirty translations are in my file cabinet. I certainly do not want to be a ‘spoil-sportress”. All of my life, beginning when I was four years old, I’ve had a huge ‘bump of curiousity” when it comes to historical, world-wide events. I began with the “Heidi” books . (All about a young girl and her adventures in Alpine country (Swiss Alps).


  68. Tom: the Voynich EVA letter assignments must be 20 years old by now, and ain’t got not nothing to do with me. As far as Bax goes, I refer the honourable gentleman to the words of James Joseph Brown (1933-2006): “I got mine / He got his”.

  69. bdid1dr on October 31, 2016 at 10:13 pm said:

    PS : If you happen to have online “Adobe Acrobat Reader” you will find the translation process of the entire “Florentine Codex”. You will then be able to read any portion of the entire “Florentine Codex” — all THIRTEEN Parts . Fascinating reading. You will then be able to read and translate any part of the so-called “Voynich” manuscript.
    I don’t recall Adobe Acrobat Reader portraying the ‘nekkid ladies’ in their water-fall and bath-house (as they are portrayed in the “Voynich/B-408.

  70. Thomas F. Spande on November 1, 2016 at 3:23 am said:

    Nick, Maybe I dreamed this, but I thought slightly before Oct. 10, that you put down statistics for four glyph triplets, c-X-h , where X was one of the four gallows glyphs. I now seem unable to find that post but it seemed to me that you had in mind, a specific assignment for each one of the four gallows glyphs?

    If this be privileged information that you wish not to share, that is OK of course, I do not need to know your assignments, just that you might indicate that you do have in mind, a mapping of the gallows to four specific consonants or maybe even vowels? Or is this a variable feast?

    I assumed that since you did once refer to “t, k, p, f” in your initial discussions of the orthography of the VM, that this was still your thinking? Of course time has passed and I have huge gaps in my following of VM posts and I apologize for lumbering you with outdated, likely tentative assignments that are now in the rear view mirror. Cheers, Tom

  71. Tom: I remain unconvinced by any explanation for the gallows (including my own), and certainly don’t have any kind of rigid mapping in mind. Having said that, I do suspect that they typically function as consonants (with e, ee, eee, ch, and sh as vowels), though not in a trivially linguistic sense.

  72. Tom: t k p f is the representation in EVA, it’s used merely as a convenience when discussing the Voynich and does not constitute a claim or proposal as to what the shapes actually denote or signify.

  73. Thomas F. Spande on November 1, 2016 at 2:36 pm said:

    Nick, Thanks for the prompt response and for sharing your thinking on these glyphs. Cheers, Tom

  74. Thomas F. Spande on November 2, 2016 at 12:53 am said:

    Nick et al., Some time ago when I was pursuing the idea that the gallows might serve as start-stop signals, I looked at a few folios in detail. Taking that data and recasting it with another idea in mind, I append my analysis of six folios here in order of their complexity: single stem, single-loop; single stem double-loop; double stem single-loop and double stemmed double-looped.

    TIGHTER scribe: f31v: 3, 4, 38, 16, 11 lines; gallows glyphs per line = 5.6; The single-stemmed glyphs dominate line 6 (6 occurrences), the start of the second paragraph. and also the first glyph of line 1.

    f33v: 1, 0, 31, 13; 11 lines; gallows glyphs per line = 4.1; The single-stemmed gallow (one only) occurs as the second gallows in line 7.

    f34r: 3, 2, 30, 32; 16 lines; gallows glyphs per line = 4.2; three single stemmed glyphs occur in first line.

    – – – – – – – – – – – – – – – – – –

    LOOSER Scribe: F2V: 0, 1, 8, 10; 8 lines; gallows glyphs per line = 2.4; the single single-stemmed gallows is the second gallows on line 1 and enlarged.

    f6r: 2, 3, 16, 18; 11 lines; gallows per line = 3.6; Four single stemmed gallows in line 1.

    f 29v: 0, 0, 23, 30; 12 lines; gallows per line = 3.6.

    I think the single-stemmed gallows tend to be decorative and used at the start of the text or a paragraph. It does appear that the tighter scribe uses more of the gallows glyphs but because the text there is more condensed, I will have to recalculate using an actual glyph count, not just gallows per line.

    All for the moment. I will continue trying to prove that the occurrence of the gallows glyphs are too erratic to be mapped as consonants. Cheers, Tom

  75. bdid1dr on November 2, 2016 at 4:26 pm said:

    So, ThomS, what are they saying? A very interesting and informative discussion of a particular specimen would be the illustration which “looks like” a couple of dark ‘sun dews’. They are, in fact, ‘monks-hood’ plants. The discussion (Espanol/Nahuatl) is all about the invasive nature of the plants (roots) and what uses they have.


  76. Tom: remember that pairs of matched (particularly single-legged) gallows on the top line of a page (usually 2/3rds of the way along) are called “Neal Keys”. This is one of the few places you’ll see a gallows character appear at the end of a word. And in pages dominated by labels, we tend to see ‘o + gallows’ more than ‘gallows’. So there is a lot more to how gallows appear on the page than you might think.

  77. bdid1dr on November 3, 2016 at 12:00 am said:

    I’ll try to be brief: Those large “Pilcrow”s are Spanish/Latin/Nahuatl syllables/combinitive words such as “Bless-us-Padre ” — or Prefix – Presentation – or simply “Padre”.

  78. Thomas F. Spande on November 3, 2016 at 5:42 am said:

    Nick, I have noted also that the single-stemmed gallows do tend to appear in the first line of paragraphs. Sometimes just one type, sometimes both. This holds true of both scribes. I have gone over another 15 folios looking for a pattern in both the single- and double-stemmed gallows and will relay this analysis tomorrow, although I suspect my conclusions are redundant and that I have merely rediscovered the wheel.

    Briefly, the double-stemmed gallows seem all over the map and this might form the basis of an argument that it will be a stretch to assign consonants to them. There is such a generous inconsistency noted that I cannot believe they stem directly from any known language.

    Thanks for the background on those Neal pairs, an observation of which I was unaware.

    More anon. Cheers, Tom, (a disappointed Cleveland Indians fan) !

  79. Thomas F. Spande on November 3, 2016 at 6:01 pm said:

    Dear Folks, At the risk of destroying once and for all, any residual interest people might have in the purpose of those pesky “gallows glyphs” i am listing my analyses of 15 randomly selected folios from the VM botanical section. I use the same ordering as my post of Nov. 2, i.e. single stemmed, single-looped, single stemmed double-looped; double stemmed single-looped and finally the double stemmed double-looped. They follow in two groups, those done by the looser writing scribe (seven), then the more tight enscriber. Today’s post will cover the looser scribe.

    LOOSE: f2v: 2 paragraphs; 0, 1, 8, 10; 8 lines. ratio of the double-stemmed gallows = 0.8; the sole single stemmed gallows is the lead glyph.

    f3r: 4 paragraphs: 0, 6, 15, 19; 20 lines; ratio = 0.79; 4 single-stemmed gallows in lead line of fourth para, one is the lead glyph; one in line 1 of first para; one as lead glyph of second para.

    f3v: i paragraph: 0, 1, 27, 12; 14 lines; ratio = 2.3; sole single-stemmed gallows is lead glyph.

    f6v: 1 paragraph: 0. 1, 25, 19; 21 lines; ratio = 1.3, sole single stemmed gallows is midways in line 1.

    f9r: 2 paragraphs: 0, 2, 13, 23 : 10 lines; ratio = 0.57, 1 single-stemmed gallows in line 3; one as the lead glyph of second paragraph.

    f13r: 2 paragraphs: 0, 7, 21, 10 : 10 lines; ratio = 2.1, 4 single gallows in line 1; 3 in first line of second paragraph. None are lead glyphs.

    f 20v: 3 paragraphs: 4, 7, 8, 19 : 11 lines, ratio = 0.42; The single stemmed, single-looped gallows occurs twice in line 1, one being the lead glyph and twice in the first line of paragraph 2; The single stemmed, double-looped gallows occurs 5 times in line 1; once each in the first line of paragraph 2 and the last line of paragraph 3, but in no cases is this gallows a lead glyph.

    Only a few of the single stemmed gallows can be considered pairs. The wildly varying ratios of the double stemmed gallows makes, for me, a huge impediment in assigning these to consonants. BTW I have measured the number of glyphs of the loose and tight scribe per 10 cm and the difference is not huge being 33 for the loose and 38 for the tight scribe. I will pick a bunch more and average them, but the post of Nov. 2 where I have calculated gallows per line is a pretty good approximation.

    For bd: I have no clue as to why often a single stemmed gallows is drawn as an enlarged glyph. I have assumed it is just one of the many annoying distractions of the VM text as often they are hugely embellished on top of it.

    Will convey some more of my gallows analysis tomorrow, dealing with the tighter writing scribe. I hoped to find some clear cut difference between the scribes, but no “Eurekas” as yet, or anyway nothing significant. Cheers, Tom

  80. Thomas F. Spande on November 4, 2016 at 10:47 pm said:

    Dear all. I have paid more attention to the idea that, I think, lies behind the “Neal pairs”. What follows is one last folio by the loose scribe and 5 of the tighter encscriber. The same gallows glyph order is used as above:

    f47r: 2 paragraphs: 2, 1 16, 9: 11 lines ; ratio = 1.8, one single stemmed, single looped gallows as the lead to para 2; one in line 3; one single stemmed, double looped gallows as the lead glyph in the 2nd paragraph.

    f4r: 2 paragraph: 0, 4, 5, 19: 13 lines: ratio = 0.26, 1 single stemmed gallows in line 1, two in line 2, one as the lead glyph of 2nd paragraph (not consecutive) 1 in line 10.

    f11r: 2 paragraph: 2, 2, 8, 15: 7 lines: ratio = 0.53; 2 single stemmed, single looped gallows appear consecutively in line 1; one single stemmed double-looped appear in line 1, one oversized gallows in the second paragraph, but not the lead.

    f18r: 2 paragraphs,: 1, 1, 25, 15: 14 lines: ratio = 1.7; one gallows in line 1, one gallows with two legs, 2 loops as the lead glyph.

    f13v: 2 paragraphs: 1, 2, 19, 20: 10 lines: ratio = 0.95: one single stemmed, single looped gallows as lead of 2nd para; one single stemmed, double looped gallows in line 2, first para; one in line 1, 2nd para.

    f26r: 2 paragraphs: 3, 7, 19, 18: 10 lines; ratio = 1.5; two in line 1, first line in para 2; 4 in line 1, one as the lead glyph; 3 in line 1, second para, including the lead. One pair of the single stemmed, double looped gallows are consecutive (i.e. Neal pairs as I understand them). Another six to follow later.

    Cheers, Tom

  81. Thomas F. Spande on November 5, 2016 at 4:29 am said:

    Dear All, First a correction of f26r above: The gallows glyph sequence should have been: 3, 7, 19, 13. The rest of the data is unchanged.


    f31r: 3 paragraphs: 4, 1, 31, 11: 16 lines; ratio = 2.8; three consecutive single stemmed, single-looped gallows in line 1; 1 at the end of line 2; second paragraph, line 1 for single occurrence of single stemmed, double-looped gallows glyph.

    f31v: 2 paragraphs: 3, 4, 16, 38: 11 lines; ratio = 0.42; Single-stemmed gallows: 3 consecutive glyphs in line 1 of paragraph 2; ;Lead glyph of first paragraph; 3 in line 1 in paragraph 2, two of which are consecutive. One is the lead glyph of paragraph 2.

    f33r: 1 paragraph: 3, 3, 25, 13: 7 lines; ratio = 1.9; 2 consecutive single stemmed, single looped gallows in line 1, one isolated; 2 consecutive single stemmed, double-stemmed gallows in line 1; lead glyph of line 5.

    f33v: 2 paragraphs, 1, 0, 31, 13; 11 lines, ratio = 2.4; line 1 paragraph 2.

    f34r: 3 paragraphs, 3, 2, 18, 24; 16 lines;ratio = 0.75, single stemmed, single-looped gallows, one in line1; one in line 1 and 7, one in 3nd para; double-looped gallows: lead glyph and another following closely ( a presumed Neal pair).

    f43r: 3 paragraphs, 3, 10, 23, 43, 14 lines, ratio = 0.53; single stemmed single looped gallows, two in lines 1, one in first line of 3rd paragraph. single stemmed, double-looped gallows: 3 in line 1; one in line 6; four non-consecutive gallows glyphs at start of paragraph 2; 2 in first line of paragraph 3, including lead glyph. Only two of the single stemmed gallows are consecutive.

    As can be observed, the ratio of the double stemmed gallows with either the loose or tight scribe vary all over the map and likewise, the single stemmed gallows. Maybe there is a clue to be found in the presence of the single stemmed gallows as they occur most often in the lead line of a paragraph, often as the lead glyph. Frequently they are enlarged or display scribal flourishes. At the moment, to me, they appear mainly to be decorative.

    Cheers, Tom

  82. Thomas F. Spande on November 6, 2016 at 9:29 pm said:

    Dear all, I think it is really unlikely that the single stemmed gallows can be mapped to any character, consonant or vowel. Just too difficult to explain how this can be when nearly all are found in the first lines of paragraphs and often as the lead glyph. They evidently play some key role, but I doubt that being mapped as a language glyph is one of them.

    I have gone over the bulk of the VM botanicals and find that often the single stemmed, single-looped gallows is absent. A partial list of these follows:

    Tighter scribe: f19r ( only one 1 single stemmed, double-looped); f 24r; f41r; f48r; f51r; f52v; f53r; f54v; f55r; f57r; f57v; f87r.

    Looser scribe: f7v; f35r; f35v; 38v; bottom part of f42v (7 lines); f47v; f56r; f87v; f90v.

    Many of the folios with a single-stemmed single looped gallows have only one: f2r; f34r f43v; f55r; f50v; f53r and some seen in the above two posts.

    Three folios have NO single stemmed gallows at all: f15r (15 lines by the tight scribe) (top part of f42v (8 lines; tight scribe; f36r, (6 lines by the looser scribe).

    What to make of this? I am totally baffled by the apparently random appearance of the single stemmed gallows. Nick might have a point that merely because we cannot understand the function of a glyph, gives us no justification for declaring it null and void! Still that does seem a logical possibility.

    If I were formulating a substitution code and wanted to pick a null character, I would go with something exactly like the gallows glyph. Nearly all stick out as often the lead glyph and are often embellished. Most are on the first line of individual paragraphs. Where paragraphing is not evident, a gallows might still indicate a new paragraph has been created. Line endings are often hard to spot as they are integrated with the drawings.

    Cheers, Tom

    ps. I will go over the single stemmed gallows once more to spot what I think might be “Neal keys” Then I plan to get back to spice spotting.

  83. Thomas F. Spande on November 7, 2016 at 12:05 am said:

    Nick, Can you give us an idea of what Neal has in mind for those single stemmed gallows? I have only guessed that what counts might be between identical gallows glyphs and might be something like the name of the plant?. Before I do any analysis of this hypothesis, I’d like to run it by you and get your input as well. Cheers, Tom

  84. Tom: Philip Neal is far too level-headed and sensible to try to force a meaning out of (what we now call) “Neal Keys”. Yet at the same time, I would contend that the presence of both vertical and horizontal Neal Keys is very much like the Gillogly sequence in the Beale Papers (if you know that), in that it is something that almost all Voynich theories would fail to explain or predict.

    That is to say, Neal Keys are useful for making people stop to consider whether their supposedly-complete pet theory might not actually explain everything about the Voynich manuscript after all. 🙂

  85. Tom: once again, I’ll say that I while I don’t believe that the gallows characters map to specific consonants, I do believe that they map (Neal Keys aside) to “confounded consonants” – that is, they are never vowels. More than that, I don’t know. :-/

  86. Thomas F. Spande on November 7, 2016 at 6:16 pm said:

    Nick, Thanks for the prompt reply. I still don’t understand how even something like a “confounded consonant” (whatever that might be), can evidently vary so much, folio to folio.

    Is a “confounded consonant” something that elicits the response “Confound it!!” when encountered?

    We face an election tomorrow that will make the Brexit vote seem like a civilized difference of opinion. I see now how fascism could arise in the US of A.
    Polls favor Hillary but Donald is not going to go quietly into the night if (when) he loses the vote. There will be squeals from him of “vote rigging” and “the system is rigged”. It won’t be pretty!

    Cheers, Tom

  87. Tom: a “confounded consonant” is a consonant… but which particular consonant is something nobody can (currently) say. Maybe one day… 🙂

  88. Thomas F. Spande on November 8, 2016 at 4:43 am said:

    Nick, OK! Darts it is!

    Cheers, Tom

Leave a Reply

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

Post navigation