Since the Voynich Manuscript surfaced in about 1912, many of the best-known codebreaking experts have studied its writing (‘Voynichese’) in depth. Of them, many have concluded that it was written using a cipher system that was (a) stronger than a simple (monoalphabetic) substitution cipher, yet (b) mathematically weaker than a polyalphabetic cipher.
If the University of Arizona’s 2009 radiocarbon dating of the Voynich Manuscript’s vellum (which points to the first half of the fifteenth century) is correct, the most likely reason for (b) becomes blindingly obvious: polyalphabetic ciphers (such as those of Leon Battista Alberti, Abbot Trithemius, and Vigenère) hadn’t yet been invented.
So, does that mean that all pre-polyalphabetic ciphers were easy? Errm… nope. In fact: not even close.
Fourteenth Century Cryptography
Even though Gabriele de Lavinde’s 1379 collection of Vatican ciphers were, at heart, simple (monoalphabetic) ciphers, many also included “nulls” (special cipher shapes that code for nothing at all, and were added into ciphertexts specifically to try to misdirect codebreakers). In the hands of a tricksy encipherer, this can already become not at all straightforward to crack.
Even the very clever CryptoCrack doesn’t have a tool for predicting / identifying nulls in a given ciphertext: and it turns out (I believe) that this is a significantly harder technical challenge than you might think.
Moreover, many of the ciphers in Gabriel de Lavinde’s cipher ledger also contained a nomenclator: this was a list of typically a dozen-or-so shapes enciphering entire words, like a cross between a cipher and a code. (Broadly speaking, a ‘cipher’ enciphers a message a letter at a time, while a ‘code’ encodes a message a word at a time: so nomenclators blur the line between the two).
However, it’s far from clear (to me at least) whether nomenclators were added in the 14th century for security, speed or brevity. I suspect that to insist that it was just a matter of security would be to project principles of Schneieresque computer science onto the codemakers and codebreakers of the 1300s: the true answer would be some vague (and probably unworked-out) combination of all three.
Fifteenth Century Cryptography
At the beginning of the 15th century, however, things started to shift (slightly) in the world of codemaking. 1401 was when a secretary at the Duchy of Mantua produced the following cipher alphabet for corresponding with Simeone de Crema:
Now, in many ways, this is a particularly stupid cipher alphabet, because the top (core) line maps each character in the alphabet to its reversed-alphabet equivalent (i.e. ABCDE –> ZYXUT and vice versa). Yet what is simultaneously clever about it is that it allocates multiple shapes to each of the five vowels.
To be honest, I think it would be a bit of a stretch to infer from this (as David Kahn tries to) that the notion of defending against frequency analysis-based attacks must necessarily have been entering cryptographers’ minds as early as 1401. Rather, it seems many times more likely to me that this trick (now known as “homophonic substitution”) was originally devised for a far more mundane reason: to make it harder for codebreakers to tell which letters are vowels and which are consonants.
Fast forward to the middle of the fifteenth century (probably circa 1450-1455), and we can still see the same palette of tricks in action in the following (undated) cipher alphabet in the Tranchedino cipher ledger from Milan:
Apart from not using the same alphabet backwards as the base cipher alphabet, it would seem that not much has changed since 1401: the vowels are still obfuscated with multiple homophonic alternatives (though with only three different shapes per vowel here, rather than the four shapes per vowel used half a century before).
The more observant among you will also notice that the (formerly Tironian) shorthand abbreviation ‘9’ gets its own cipher shape, as does ℞ (i.e. Rx, if your prehistoric browser can’t render Unicode character ‘U+211E’).
However, the later cipher alphabet also has special cipher shapes for doubled letters, a few other common shorthand abbreviations (p, etc), and a few more nulls than before:
The nomenclator is noticeably beefed up, with this particular cipher boasting more than eighty special entries:
Another Mantuan Cipher (1450)
Given that the 1401 cipher was from the Duchy of Mantua, it’s interesting to have a look at a Mantuan ducal cipher from 1450 in the Tranchedino ledger. This now has two homophonic shapes per consonant (except for x, z, and the ‘9’ shorthand shape), and three homophonic shapes per vowel:
It then has a mini-codebook of common words (Come, Quando, Quanto, Non, etc) and some nulls:
Interestingly, this is followed by an entirely new section, with arbitrary shapes standing in for a whole load of syllable groups (ab, ac, ad, af, ag, etc):
Finally, the page finishes up with roughly the same (small) size of nomenclator as had been in use in Mantua half a century previously:
So, You Call This “Progress”?
There is a long-standing (and widespread) tendency among writers on cryptography to present the development of ciphers in the fifteenth century as a kind of prototype of the modern arms race.
It’s perfectly true that, as the number of parties enciphering messages grew (along with the first flush of modern diplomacy) in the mid-15th century (many historians quite reasonably date this to the 1454 Treaty of Lodi), so too did the number of people who became experienced at cracking them.
However, there seems to me to be no evidence suggesting any kind of awareness of frequency analysis in the West in the fifteenth century. While Leon Battista Alberti’s short book on ciphers (“De Cifris”, 1466/1467) did cover this very well, he appears to have devised the abstract principles himself: and the contents of his book seem never to have been shared with anyone outside the Vatican. Similarly, al-Qalqashandi’s (1412) Arabic encyclopaedia entry on frequency analysis (mentioned in Kahn) appears never to have been transmitted to the West.
Don’t get me wrong, cryptology and cryptography both genuinely advanced in the sixteenth century: but in the fifteenth century, code-breaking had no mechanisms, no abstract methodology to work from: and fifteenth century code-making relied, by and large, on exactly the palette of tricks that were in place by 1450 or so. The only noticeable difference was that of scale: more homophones, more syllables, more nulls, and bigger nomenclators.
What, Then, Of The Voynich Manuscript?
In almost all practical senses, I think it’s fair to note that the Voynich Manuscript stands outside the cipher-making traditions you can see embodied in the cipher alphabets described above. It would seem to have too few cipher shapes to be using homophonic cipher tricks, doubled letters, a nomenclator of commons words, or even nulls.
And yet it dates to this precise period: and – arguably the most telling cryptanalytical feature of all – there is still no modern-day consensus as to which shapes are vowels and which are consonants. Even now, the letters that resemble ‘a’, ‘e’ (sort of), ‘i’, and ‘o’ continue to convince people seeing the Voynich Manuscript with fresh eyes that they ‘must’ not only look like vowels, but ‘must’ also be vowels. However, the closer you look at these, the unlikelier and wobblier this conclusion gets.
So, here’s your paradox for the day: even though the Voynich Manuscript is almost certainly not using the homophonic trick of using multiple letters for each of the vowels that was in use as early as 1401, it very much seems that its author devised or adapted an alternative way of concealing the plaintext’s vowels, i.e. of answering the same basic cryptographic ‘problematique’.
But how did it do that?
Well, Nick: Over the years I’ve laid out the consonants which are combinations of vowels : oi — y aes — s ce – k aes-ceus – excuse + Chris (Christ) —–
just a few
Suppertime ! ciao ! chow !
g -nite !
bd
Nick,
Frequency analysis was apparently quite well known by about .. I don’t know, the eighth or tenth century or something. I take your word for it that it wasn’t known in Europe, but evidently it was already in Baghdad, where a master of the postal service (Ibn Khurradadhbih, perhaps?) explained how to break a secret message by determining letter frequencies.
I wrote about it some time ago, and even reproduced the passage – probably on ‘Findings’. If it interests you I’ll hunt for it.
Diane: Kahn’s “The Codebreakers” covers Arabic cryptology reasonably well (up to al-Qalqashandi). But my point is that from the evidence of the cryptography – that is, the code-making – there was no explicit knowledge in Europe of how to use letter frequencies to crack ciphers during the fifteenth century.
I think Kahn misreads the use of homophony-employed-to-obfuscate-vowels as if it were a technological homophony-employed-to-combat-frequency-analysis step in the putative ‘crypto arms race’: as far as I can tell, there was no explicit frequency analysis in Europe before Alberti, nor indeed after him for a fair few decades (his “De Cifris” was an outlier, and was not diffused into the wider culture).
Nick
Yes, ok.. and I’d guess you read Cryptologia fairly regularly, so you’d know the paper by Abdelmalek Azizi and Mostafa Azizi, ‘Instances of Arabic Cryptography in Morocco’, Cryptologia, Vol.35, No.1 (Jan 2011) pp.47-57.
In case you missed it (which means you didn’t read my post to Findings on Jan.20th., 2012 either :)… here’s the abstract, which mentions the 11thC. Now business connections between mainland Europe and north Africa mightn’t feature largely in the historical record, but I’ve mentioned a contract signed in Marseilles and have touched on Datini too..
In the abstract to that paper published in Cryptologia, is said:
… In this paper, we give an overview of some instances of cryptography which we have found in the Maghreb (from the 11th century to the 17th century): however, there are undoubtedly many other cases to be found in the manuscripts that have not been studied yet.
11thC?
Diane: just because item X in place A predated item Y in place B does not mean that item X necessarily diffused to shape item Y.
Looking at the many hundreds of objects themselves, I see no evidence that explicit knowledge of frequency analysis influenced any fifteenth century codemaking practice. Put another way: while the possibility of diffusion from Arabic cryptology to European cryptography exists, there is no internal evidence to support the notion that it actually did.
Nick,
I see where the crossed-wires is happening. Where you begin from a premise that the written part of the text will have been produced in its present un-readable form in Europe, by an educated Latin Christian, and within the bounds of what we find in Latin Christian works on cipher to the early fifteenth century (how am I going so far with this sentence?), I am asking “where and when, and how, might a written text have been produced which, by the early fifteenth century, looked like the text in Beinecke MS 408?’
Not saying one is good and the other bad; just that your post is written by first taking certain parameters as ‘givens’, and it took me a while to realise that we were asking different questions. I don’t see any reason, for example, for a copy of al-Qalqashandi’s Encyclopaedia to have made it to Europe before anyone on that side of the Mediterranean could use techniques he’d described. All you’d need is to have a business deal going with someone else who’d read it, surely.
Ah – now crossed comments. 🙂
Your saying “just because item X in place A predated item Y in place B does not mean that item X necessarily diffused to shape item Y.”
is *exactly* a point that I’ve had – and am still having – the most enormous difficulty getting across en re the botanical imagery. Well said.
Diane: putting the Voynich Manuscript entirely to one side, the core of my argument here is that there is no internal evidence of direct knowledge of frequency analysis within the cipher alphabets found in Italy in the fourteenth and fifteenth century. As such, I have found no indication that knowledge of Arabic cryptology (such as is found in al-Qalqashandi’s encyclopaedia) diffused into fifteenth century Italian cryptographic practice. I am entirely aware that it might have diffused there: but I have found no evidence whatsoever that it did diffuse there.
And then back to the Voynich Manuscript: I always find it somewhat painful reading when you try to project an argument onto my virtual mouth involving some putative “educated Latin Christian” etc as an ex nihilo pure-point-of-faith premise. For me, the notion of a Northern Italian origin for the Voynich Manuscript is no more than a working hypothesis inferred from a number of characteristic details in the primary evidence that I think are most telling. I didn’t start from Northern Italy and then go out looking for supporting evidence, even if that way round seems to be the Voynichological norm. 😐
Now putting the two parts together: for those people who aren’t you 🙂 and who consider a fifteenth century Northern Italian origin for the Voynich Manuscript to be a reasonable working hypothesis and who are also willing to consider the possibility that it is written in an unknown cipher system, it is surely interesting that (a) notions of explicit frequency analysis did not seem to factor into cipher alphabet design in that time and place; and (b) the core reason driving homophonic cipher adoption seems to have been not disruption of frequency analysis attacks but a desire to obfuscate the vowels.
Not directly related Nick, but is there any source (on your site or otherwise) that lists all of the symbols used in Voynich, with some type of frequency analysis?
Rick: there are lots of sites that do parts of what you want, but no single site that does it all.
http://www.voynich.nu/transcr.html lists the various transcriptions used (I use EVA)
http://voynichportal.com/2016/02/26/how-to-write-voynichese/ is JKP’s take on writing Voynichese
http://www.voynichese.com/ is good for visually analysing Voynichese, but takes a little while to properly grasp.
For creating your own statistics, you might try my JavaScript state machine page: http://www.nickpelling.com/voynich/analyse.htm
Very interesting post, Nick – many thanks.
Kahn is somewhat out of date on Arabic cryptology, and you might find this article interesting if you don’t already know it:
http://www.kacst.edu.sa/en/about/publications/Books/Arabic_Origins_of_Cryptology1.pdf
SirHubert: a very interesting book which I wasn’t previously aware of, thanks for that. Though I’m well aware of the existence and scope of Arabic cryptology, I don’t currently know of any examples in the cryptographic literature that point to the diffusion of its knowledge elsewhere: perhaps this will be covered in the book, or elsewhere within the series of books which this forms the start of.
Nick,
The encoding of rational numbers used by Arabs and medieval can be added to your stirs line. Fibonacci scaled rationsl number n/p by subtracting an LCM 1/m to write 2-term series as often as possible. LE Sigler transliterard the 1202 Liber Abaci in 2002 to include seven distinctions, approaches, to the methid. The most complicated, the seventh distinction selected two LCM 1/m values, for example 4/13 was encoded as a U it fraction series by subtracting 1/4 and 1/18 meant 4/13 = 1/4 1/18 1/468. Wikipedia dies not discuss this later point well. David Eppstrin, UCI suggests that a greedy algorithm was applied.
More importantly, to intellectually understand Arab amd medieval encoding of rational numbers to unit fraction series, Greek methods that also encoded numbers as Ionian Or Ionian alphabetic letters needs to be discussed.
Archimedes, Fibonacci and Gailileo used the same method to encode irrational numbers, even though Archimedes followed a much older Egyotian method that encoded rational numbers n/p by multiplying by LCM m written in the form m/m such a new scaled value mn/mp considered the best divisors of mp that summed to numeratot mn. For example, 4/13 was scaled by 4/4 so that 16/52 inspected the divisors of denominator 52 (52, 26, 13, 4, 3, 2, 1) that summed to numerator 16 (13, 2, 1). In the Egyotian era Ahmes would have hi-listed 13, 2, 1 in red such that Archimedes and Ahmes would have written out the same series in a longhand form that translates to our 21st century by
4/13 (4/4) = 16/52 = (13+ 2 + 1)/52 = 1/4 + 1/26 + 1/52
In different shorthand notations that have confused scholars to this day.
To translate medieval, Arab, Greek and Egyptian unit fraction encoding methods, shorthand writings must add back missing intellectual steps to parse entire encoding methods of each era.
Galileo, Fibonacci and Archimedes used the same vulgar fraction method to encode irrational numbers to eight or more decimal places, when translated into modern base 10 decimal arithmetic. The method stressed the binomial theorem that considered the Quotient (Q), the largest square number less than N, as double the inverse proportion of the Remainder (R), in the form 1/2Q meant the square root of 10 was estimated in three steps, the first step written as
1. (3 + 1/6)(3 + 1/6) = 10 plus an error 1/36
2. To reduced error 1/36 divide by 2(3 + 1/6) in vulgar fraction multiplication
1/36 x 6/38 = 1/228 meant
(3 + 1/6 – 1/228)(3 + 1/6 – 1/228) = 10 + (1/228)(1/228)
Since the binomial theorem works vividly in thus inverse proportion co text,
3. Divide (1/228)(1/228) by 2(3 + 1/6 – 1/228) in vulgar fractions to complete the problem as generally duscussed by
https://www.academia.edu/10608644/Archimedes_Square_Root_of_3_5_6_7_and_29
Hmm, I’m not going to go there — fractions, inverse proportions, binomial theorem, parallelograms, square roots of odd numbers I’ll stick to translating the Spanish and Nahuatl dialogues in B-408 and the Florentine Codex.
Do I sense a huge sigh of relief ?
bd
Kahn’s Classic code breaking book stressed frequency distributions that I learned in the late 1950s working on Russian and Arabic language codes. Kahn’s otherwise excellent book did not stress the use of prime numbers and number theory as medieval, Arab, and Greek used to scale rational numbers, a skill known to the majority of numerate medieval cryptologists.
In passing, our 1585AD base 10 decimal system encoded rational numbers by mixing the binomial theorem with an algorithm as Pascal’s triangle. Read Oystein Ore “Number Theory and its History”, a 1948 book that I studied in 1964, earning a math degree.
Sorry that a few members of this group run from math and rational numbers, when mentioned in passing, a well known skill, and possible key to certain codes, created by numerate medieval and later cryptologists.
My main points remain. Math ideas should be considered , at some point, when hard to decode medieval texts resist frequency distribution and other basic code breaking attacks.
Nick,
Absolutely agree with you that:
“For me, the notion of a Northern Italian origin for the Voynich Manuscript is no more than a working hypothesis inferred from a number of characteristic details in the primary evidence that I think are most telling. I didn’t start from Northern Italy and then go out looking for supporting evidence, even if that way round seems to be the Voynichological norm,”
– except that what you still see as a working hypothesis, I’ve pretty much come to as the end-point after having considered all other aspects of the manuscript (as far as we know them) a- materials, palette, imagery, earlier appraisals and so on – not forgetting the more recent opinions of Alain Touwaide.
I’m rather more fierce than you about those who start from a desired outcome and then go looking for supporting evidence, as I’m sure you know. I call it the “theory-first” approach.
– To describe men of the renaissance elites that won’t raise hackles today, when we pride ourselves on being multi-cultural and religiously tolerant, but I’m defining people of that ‘humanist’ group and not you.
For people (including me) “who consider a fifteenth century Northern Italian origin for the Voynich Manuscript to be a reasonable working hypothesis and who are also willing to consider the possibility that it is written in an unknown cipher system..”
The difference here is that I’m willing to attribute that origin for the manuscript as an artefact – in fact I’d say that’s the opinion I’d give of the artefact, but I cannot attribute first enunciation of what is *in* the manuscript to any humanist ‘author’ of that time. There the internal evidence is opposed to such a view, and the imagery in particular, not only because of what is drawn and the way it is drawn, but the range and common theme through the various sections.
I’m certainly among those willing to consider that the written part of the text is in cipher, especially since the information expressed in the imagery was potentially of such enormous value, but I’m very impressed by the fact that in a century, with the best cryptographers and a fair whack of computer power and high-level decryption programs, the text still hasn’t been shown to be in cipher. And overall, I think an enciphered text would ill-suit the sort of thing contained in the manuscript and its imagery. You must admit, a cipher text accompanied on every page with illustrations is a little unusual. 🙂
As far
Diane: it might not always have been so unusual. Perhaps the other forty-seven enciphered volumes in the same series were all burnt on Savonarola’s bonfire? The caprices of historical happenstance fall far further away than the keenest eye can see. 😐
Diane: given that the floor around the Voynich Manuscript is stacked high with the carcasses of strongly-held certainties that were subsequently proved woefully incorrect, I should add that “working hypothesis” is as far as I’m prepared to go. 😉
This is by the bye, but I’ve stumbled over a lovely history for laypersons in a paper written by Katherine Ellison. Have you seen it?
‘1144000727777607680000 wayes’: Early Modern Cryptography as Fashionable Reading.
It’s in the Journal of the Northern Renaissance. Online, the papers are scholarly though the format looks blog-like.
http://www.northernrenaissance.org/1144000727777607680000-wayes-early-modern-cryptography-as-fashionable-reading/
@ Milo :
Fibonacci — Fun ! Can we use Fibonacci for making what look like snow-flakes?
@ Diane: Sahagun’s ‘diary’ (B-408) was written upon manuscript material from his parents’ supply. He may have gotten additional manuscript material from the nearby university (which he attended after becoming a monk).
After many years of developing a school for boys in “New Spain”, Sahagun was accused, by a monk of another order, of being an heretic. ALL of Sahagun’s written material was confiscated and reviewed by the Inquisition. He was cleared by the Inqusitioners — and most of his written works ended up in Florence. His diary (now known as B-408) was never returned to him.
Several centuries later, Mr. Voynich (a book collector) bought the diary from a small abbey’s store-room which was being closed down by the monks.
The rest of the story rests in Boenicke’s storage/archives.
Somehow, people, ignore references to Roger Bacon and/or Shakespeare. Several years ago Boenicke held an exhibit in the Shakespeare Library which was supposed to be all about the “Voynich” manuscript. Apparently, it was not about the manuscript which apparently was not being displayed; but about Mr. Voynich’s discovery and attempted decoding.
Round n’ round we go…..
bd
Nick
The hypotheses are woeful because they’re usually based on a hope that the answer may lie within the range of whatever they happen to know already.
From the time the ‘hypothesis’ is formed, it seems very difficult for most to keep their mind on the manuscript: defending and expanding the hypothesis and networking to have the hypothesis “deemed fact” then takes most of the effort expended.
Yes, I generalise, but then that’s why we have science.
Diane: I think that what you present there is a very negative and narrow view of hypotheses. In fact, the point of hypotheses is to help focus people’s efforts on determining revealing tests to be performed and good questions to be asked: and if performing / asking those has the effect of killing the original hypothesis, then hooray! Because we now know more than when we started. 🙂
Nick: if fifteenth century ciphers were being strengthened by the addition of nulls and homophones, doesn’t that imply that simple monoalphabetic substitution was no longer thought secure? If so, how were these ciphers being cracked without some kind of frequency analysis, even if that was based on empirical observation?
According to Merriam-Webster
sorry copy and paste
Full Definition of hypothesis
plural
hypotheses
play play \-ˌsēz\
1 a : an assumption or concession made for the sake of argument b : an interpretation of a practical situation or condition taken as the ground for action
2 : a tentative assumption made in order to draw out and test its logical or empirical consequences
3 : the antecedent clause of a conditional statement
It is almost a certainty, that any and all ideas regarding the Manuscript when treated as Hypothesis will inevitably lead to the originators seeking evidence further and further afield. It is easier to seek evidence for a hypothesis in the known than to delve for internal evidence in an unknown.
We currently know quite a bit about the Manuscript but what we don’t know is what, when put together does all of it mean. In other words does A+B=C, and if so why does it equal C!
True, the *theory* holds that
“hypotheses… help focus people’s efforts on determining revealing tests to be performed and good questions to be asked: and if performing / asking those has the effect of killing the original hypothesis, then hooray! ”
– except that in practice this doesn’t happen – not in Voynich studies anyway.
In Voynich studies, any “good questions” have to posed by the same person who first formulated the hypothesis, or by a mate of same.
Otherwise, no matter how good the question might be, nor how good the answer, nor how good the evidence and reasoning between the question and answer, there is one of two things which I’ve seen happen over these past years. (Maybe it was different before 2001-2 or so.
What happens will be either 1) the substance of the argument and/or evidence is lifted and re-formed so that it seems to agree with the hypothesis rather than cast doubt on it [not a technique I’ve seen Nick use, though] or 2) the person who is questioning and/or working to address a “good question” will be- shall we say – very warmly discouraged from continuing.
As example of a good, and fundamental question: Who first researched the Voynich botanical folios, concluding from their comparative studies that the section represented a medical herbal – rather than, say, the pattern book of a tapestry-weaver, wood-carver, or mural painter? Where can one read that person’s research so as to evaluate their evidence and reasoning?
Sir Hubert: there are many ways to solve simple substitution ciphers without frequency analysis. If you do enough of them, you can get to the point where you can almost read them off the page.
Diane: there is no foundational Voynich research, botanical or otherwise. That the herbal pages might be anything apart from herbal pages was something few people questioned, arguably until I proposed in 2006 that the Herbal-B pages might be machine drawings in disguise. 🙂
Well, Nick, I take your ‘fifteenth-century-cryptography’ comment seriously. What bothers me is when some written and illustrated document is assumed to be written in code (cryptography) rather than a pictorial document which was discussed in two (valid) languages. I reiterate: Spanish and Nahuatl appear in every page (folio) of B-408 (which no-one at Boenicke has been able to ‘decode’.
Probably what is causing the most confusion is the defensive actions of the Inquisitioners when they were unable to find ANY heretical remarks/commentary in Fray Sahagun’s diary and/or his students’ translations of twelve books of the History of New Spain. The Inquisition did not return Fray Sahagun’s diary to him. Nor was his twelve-volume “Historia General of New Spain” returned to him.
Many years later, Sahagun’s diary ended up being bought from a small Papal Library , in the twentieth century, by Mr. Voynich. The “Historia General” ended up being called the “Florentine Codex”, after being passed around to various private libraries for a few centuries.
So, no code. Skilled translators (SPANISH/LATIN/MEXICAN) can probably solve the “Voynich Manuscript” within a day or so. I’ve already referred you to Senor Felipe Fernandez Armesto and Senor Bernal Diaz. I can’t find a small book which discussed the ‘colors’ of the New World (and the minerals used for the different colors).
I’m hoping you can find a good translator who can work with the various very confused, but curious-minded “Voynichers”.
bd
The time has,come,” the Walrus said To talk of many things:: Of Forced perspective and of Trompe-l’œil and maybe Epistemological anarchism with Paul Feyerabend.
No were in the Voynich manuscript are there any ideas that lead you out of the dark ages.
Nick,
Thanks for the very civil answer.
I guess my problem is that, in the pragmatic sciences, we hypothesise about implications of known items and do so within a limited range – the limits those of firmly-established data. Without any such groundwork, you get highly elaborate storylines, but they have no more objective reality than a detail in some game-plan for an online role playing game. As long as it fits the group’s preferred narrative, it’s ok. Objective reality seems almost irrelevant. If things were otherwise, we’d be taking the time and trouble to at least investigate and review bdidr’s claimed translation. And Don Hoffmann’s identifying the orthography of the month names as similar to that on astronomical instruments made in France under Persian influence wouldn’t have been dismissed as ‘unoriginal’ when it seems to be to have been very much so. Your ‘machines’ idea was certainly original. 🙂
The French orthography of the month names – in particular the reading of ‘yong’ – was already noted in March 2012 by Thomas Sauvaget on his blog.
Word people think and write differently than number people. Word people think inside out. Number people think outside in. This discussion group seems focused on inside out issues, and not clear abstract codes used over the last 2,000 years.
Concerning my naysayer that suggested a modern Fibonacci can encided a flower, nature’s rhythm sure, JJ Sylvester in 1891 was first to muddle Fibonacci’s 7th distinction as a greedy algorithm. The historical Fibonacci simply ciphered 3-term unit fraction series, such as 4/13 = 1/4 + 1/18 + 1/468.
Medieval abstractions included our base 10 decimal system, irrational numbers Scaled by a binomial theorem, and an algorithm, and Babylonian square root, that I learned in high school, computing one decimal digit at a time, taking 8 steps to approximate an irrational number that Fibonacci achieved in three steps.
Beauty and simplicity of use are muss for any encoding and deciding system. Too bad that the majority of this group think inside out, looking at trees rather than forests, the later a rich world of medieval cryotologists.
I thank Sir Hubert for linking the Historynof Arab Cryotology and Cryotanalysis. Overall the book misses critical math encoding system by relying on the false premise that Arabs created zero. Arabs used the Greek two dots symbol, and ended Greek ciphered numbers recorded as Ionian and/or Dorian letters.
Going back in time, looking for the forests of numbers that Fibonacci was instructed by Arabs to use, precise decoding of the Liber Abaci and other texts, is required. For example Greeks encoded rather numbers by building tables of n/p as the Akhmim Papyrus passed down to us to decode n/3, …, n/31 a text that Kevin Brown and I worked on in 1995. Today it is clear that the o,der Egyptian system of scaling n/p by LCM m, such that mn/mp inspected denominator mp for prime and composite numbers that summed to numerator mn, made simple by working in tables.
Arabs modified the Greek system to a subtraction context as Fibonacci reported by seven dustibctions per LE Sigler’s Incomplete transliteration published in English for the first time in 2002. My summary paper https://www.academia.edu/11810863/Liber_Abaci_2015_update may be worth reading. Zero was not an issue, and was not a Arab innovation.
It is important to actually decode Arab and medieval math texts yourself, and not rely on incomplete work of others.
Others.
As explicit proof of Greek encoding of rational numbers, Kevin Brown’s failed 1995 attempts to decide the Akhmim Papyrus and older Egyptian fraction methods passed down from Egypt, from Ahmes’ 2/n table, to Greeks are summarized by
http://www.mathpages.com/home/kmath340/kmath340.htm
To directly decode pre-Greek unit fraction encoding systems two papers were published. in 2002, the 1900 BCE Egyptian Mathematical Leather Roll -EMLR was analyzed in a manner consistent with the work of Kevin Brown, falsely reported that the 26 line text used an algrbraic identity to split 1/p conversions to awkward and concise fraction series built upon LCM m in a subtraction context.
Only by thoroughly discussing this issue with David Eppstein, UCI, did I learn and read LE Sigler’s 2002 “Liber Abaci” translation from Latin to English. Clearly, all Fibonnaci and Arab seven distinctions were based on rational numbers n/p first subtracted LCM 1/m to obtain two-series as often as possible. When impossible a second LCM 1/m was selected to calculate three-term unit fraction series, as 4/13 = 1/4 1/18 1/468 has been used as an example.
A second paper was published in 2006 that freshly decoded a grain weight and measures unit called a hekat per
https://www.academia.edu/229318/The_Arithmetic_used_to_Solve_an_Ancient_Horus-Eye_Problem
Note that one hekat was multiplied by 64/64 in the LCM m manner used by Ahmes in his 2/n table, and all 87 Rhind Mathematicsp Papyrus problems, the last of which was passed down to the medieval era as “Going to St. Ives”, a mod 7’geometric proportion problem discussed by Fibonacci, and manner others. The hekat was divided into n parts by multiplying
64/64 by 1/m
For example, let n = 3
64/64 x 1/3 = 64/3 + (5/5) remainder (R) 1/64
21/64 + 5/3(1/320) =
(16 + 4 + 1)/64 (hekat) + (1 + 2/3)(1/320) =
(1/4 + 1/16 + 1/64)(hekat) + (1 + 2/3)(ro) since 1/320 of a hekat = ro
Let me stop here. I have gone too deeply into Egyptian fraction forests, issues that muddled Greek shop keepers as Plato discussed in “The Republic”, a link is available for anyone that wishes to read Plato in this context.
NIck: sure, monoalphabetic substitution ciphers are simple, but having learned to crack them (mainly) through letter frequency analysis it’s hard for me to ‘forget’ that technique and imagine trying other approaches.
So I had a look at Simonetta’s Primo, inspiciendum est for a fifteenth-century view. It’s reprinted with French commentary here:
http://www.persee.fr/doc/bec_0373-6237_1890_num_51_1_447615
…and you’re dead right. There’s nothing at all here about letter frequency analysis, and lots about identifying vowels. There are also lists of common, short words to be tried as cribs.
What I find curious, though, is that just about all of Simonetta’s techniques can be neutralized, partly or entirely, simply by removing word spaces. So you can’t rely on Italian words always ending with vowels, or that a one-letter word in Latin is almost certainly the letter ‘a’. I just wonder how likely it is that Simonetta himself wouldn’t have appreciated that.
If letter frequency analysis was allowing Simonetta to read just about anything which came his way, would he really make his knowledge of this technique public? He was clearly familiar with the idea that some digraphs and trigraphs were more common than others – is it really likely that he didn’t appreciate that this applies to letters also, and in a far more useful way?
The Plato link follows:
https://www.academia.edu/25328964/Platos_Mathematics
On Academia.edu a dozen other Egyptian, Greek and medieval unit fraction papers can be found. In the next day a two a formal Arab math from 800 CE will be posted to discuss the entire time period of unit fraction math, from 2050 BCE to its demise in 1600 CE in Europe, a slightly long in the Arabic speaking world when modern Arabic written language erased Ghobar Arabic words that were in use in 800 CE.
Sir Hubert: the idea of removing spaces first appears in the cryptographic timeline at the beginning of the sixteenth century, in Venice. It and frequency analysis are not fifteenth century tricks – that’s kind of the point of the post.
Two one-letter words — ‘A’ and/or ‘I’ Example I am meeting a very interesting……
?
bd
Base 10 decimals were approved by the Paris Academy in 1585 CE tends to hide the deeper encoding of rational numbers considered by Arabs in 800 CE. Placing the Arab innovation in validated historical context consider
http://s3.amazonaws.com/academia.edu.documents/47199356/Arabic_numerals.docx?AWSAccessKeyId=AKIAJ56TQJRTWSMTNPEA&Expires=1468372775&Signature=mK9q88Bdxo4cPtA%2Fy%2BO66kRe0mg%3D&response-content-disposition=attachment%3B%20filename%3DArabic_numerals.docx
Nick: with respect, your post doesn’t mention anything about word-spaces being retained (sometimes? always?) in fifteenth century ciphers, so thank you for clarifying that. But I must also confess that I hadn’t appreciated that Alberti explains frequency analysis quite so precisely – although you do make this clear in your post.
Your assessment of the techniques used to strengthen ciphers during the fifteenth century is interesting and seems very fair, as far as I can judge. I’m sure you’re also right that many of these were primarily aimed at hiding vowels. And you’re quite possibly right that letter frequency analysis wasn’t known, or perhaps its significance not fully appreciated, before Alberti. After all, its original application to Arabic was for the study of the Qur’an and hadith, I think, rather than as a cryptological tool.
But Alberti’s work, written in 1466/1467, is the clearest possible evidence that frequency analysis was understood by at least one person in the West in the third quarter of the fifteenth century. So I’m afraid I find your bold and rather bald claim that you can find no evidence suggesting any kind of awareness of frequency analysis in the West in the fifteenth century very challenging.
For possible interest, I found an edition of Alberti’s work (in Latin) here:
http://www.apprendre-en-ligne.net/crypto/alberti/decifris.pdf
The academia.edu link for the Arab Numerals paper is broken. Ending my time here, unless requested for additional information, the following link should work:
https://www.academia.edu/26935647/Arabic_numerals
Best Regards to all.
Milo
@ Milo :
Because I was a ‘spelling genius’ , and because I couldn’t read the lips of various teachers who would be scribbling numbers or words on the blackboard (with their backs facing the student audience — even simple mathematics –multiplication in particular– was my downfall (though I never had to repeat even a single semester)
Thank you for your kind offerings throughout the years.
And, of course, Nick for his careful reviews of our contributions .
bd 😉
SirHubert: Alberti’s de cifris made no discernible impact before the sixteenth century, but I’ll have to leave the justification for that to a separate 1000-word post. 🙂
Dear all,
I think Nick is the better person to respond to rene zandbergen’s comment of July 11th., 2016, and I hope if I am mistaken on any point that Nick will correct me, but I shouldn’t like readers to mistake the different issues here: ‘first’ versus ‘original’ and so on.
Rene wrote:
The French orthography of the month names – in particular the reading of ‘yong’ – was already noted in March 2012 by Thomas Sauvaget on his blog.
The first person to posit the language as Occitan (Occitania, by default meaning part of modern France) was Jorge Stolfi as I understand it. So he could be said the first to say the month names were of French origin.
Another writer suggested they were in a dialect of Judeo-Catalan that was also spoken in part of northern France, and he referred particularly to the orthography (though not to the writing style or palaeography). That was Artur Sixto, who could be said the first to consider those things.
Nick Pelling, in his book of 2006 took some pains over both orthography and palaeography and some of that is reproduced in posts to this blog, so the reader can decide for him/herself what new (‘first’) observations and original insights should be attributed to Nick.
Then came Thomas Sauvaget, who collected some nice examples, showing that – more or less as everyone expected – the orthography and hand were in a general sense ‘French’.
The work done by Don Hoffmann showed, more exactly, a close correspondence between that orthography (spelling) and the astronomical context – very relevant to the calendar section with its star-holders, and he further contributed the original insight that this occurs on works which we know relate fairly directly to a Persian tradition in astronomical instruments – and ones made considerably earlier than Beineke MS 408.
Don is thus ‘first’ to demonstrate a connection to the astronomical matter; first to show it occurs *in that context* within the region around Picardy (where I had already shown that forms for the standing human archer first occur in Latin works and near where Sixto describes that dialect of Judeo-Catalan having been used and, further, adjacent to regions such as Calais where connection to England and Flanders was especially strong).
Thus, Don’s original research moved the study forward – from Stolfi’s initial “Occitan” or Sixto’s “Judeo-Catalan” to a smaller region and earlier period, while proving that, indeed, the calendar inscriptions may relate to contemporary astronomical studies *influenced from Persia*.
Within that line, Thomas Sauvaget’s study of the hand and any particular items of orthography sit neatly and well deserve credit. If one were not inclined to give that credit, I suppose you could argue that Stolfi had first said the names were written in French style.
This is a bit OT in a very interesting thread.
The web page of Thomas Sauvaget is quite clear and does not talk about a “general sense”. Unfortunately I have recently been unsuccessful in posting here with links included, so I will not try.
Occitan is not French, of course.
Trying to argue who was ‘first’ is not possible if one does not have all relevant material available, and when it is related to theories about the Voynich MS one never has.
At least one could still read the material that *is* available.
I refer to the part of my web page that is concerned with the writing in the MS. There’s a section near the end about these month names, which also has a link to Thomas Sauvaget’s page, and further clarifies that Sergio Toresella already said in 1995 that both the language and the handwriting are French.
I can only recommend to read the page of Thomas Sauvaget, which I had already linked before Don Hoffmann found the reference to the astrolabe in (I believe) David King’s book “The cIphers of the monks”.
I did not realise at the time that the astrolabe was already mentioned by Thomas Sauvaget as well.
Rene: if you replace the first : and the last . with spaces, links get through fine (I reassemble them when moderating the comments).
As far as the French / Occitain / Picard / German zodiac labels go, the words do not match any language perfectly: my own palaeographic argument is that the Pisces month ends with c cedilla, which would be characteristic of Occitain rather than anything else. But I doubt that will cause any other reading not to flourish etc.
Nick: to say that Alberti’s De Cifris had no discernible impact is a different thing – but you may well be right.
I think that anyone who wants to argue that the Voynich Manuscript is enciphered polyalphabetically should be given an Alberti cipher wheel and a copy of the EVA transcription, and then told to encipher the entire thing – without errors. I’d be most interested to see how they got on…
SirHubert: I’m pretty sure we could eliminate polyalphabetic cipher systems from the list of Voynich candidates even had Alberti written de cifris a century earlier. But bear in mind that Alberti’s cipher wheel only steps every few words, so errors are not as critical as with Vigenere etc.
For the general-interest readers.
Apropos of that astrolabe and ciphers, might I mention a post written by Nick in 2010 without seeming sycoph’ic.
http://ciphermysteries.com/2010/07/08/david-a-kings-regiomontanus-acrostic-theory
(for American reader) That date in plaintext is July 8th., 2010.
Nick, with respect to the Voynich Manuscript’s code, consider the following:
1. The VM consists of over 200 pages, written fluidly with no corrections. This indicates that even if the manuscript was copied, the code application and the alphabet must be relatively simple.
2. The left and right hand margins are even. If you study the text carefully you will realize that the author began writing from left to right and when he came to the right hand margin he stopped and continued writing on the next line. There are some single and double letters at the end of one line and the beginning of the next. This indicates that some words were split in two. Folio 2r line 4 has a single ‘8’ at the end of the line and folio 3v line 6 has a mirror image ‘s’ at the end of the line. There are many other examples.
3. Although simple words representing prepositions , conjunctions , definite or indefinite articles are expected, there are no single and few double letter words except at line endings. Were they omitted or included within other words?
4. I am not exactly sure of the average length of the words in the VM’s text, but 5 or 6 letters seems reasonable. What happened to the lengthy words found in all languages? Were they split into two or more words, just as words at the end of many lines seem split.
5. A sequence of words that differ by only one letter occurs fairly frequently in the VM’s text and cause single letter substitution codes to yield babble-like text. For this reason Elizabeth Friedman has stated that simple substitution codes do not work.
6. Why are some of the VM’s symbols or letters, like ‘tl’, found at the beginning of a word, or a ‘g’ at the end of a word? Was the order of the letters in a word not important?
The above observations have caused me to conclude that the VM’s code is probably based on anagrams, with prepositions, conjunctions, definite or indefinite articles included within words and longer words being split. I am well aware that anagrams are degenerate, but the VM only uses anagrams of single words, not like Galileo’s famous anagram of an entire sentence. The number of permutations of an anagram is determined mathematically by [n(n-1)(n-2)…..1], where n represents the number of letters in a word. An anagram of 3 letters would have 3x2x1=6 permutations. For example, the letters ast give six permutations ast, sta, tas, sat, ats, tsa but generates only one sensible word, namely ‘sat’. Typically a six letter anagram will correspond to at most two sensible words.
After reading this take a look at my latest article, The Voynich plant names decoded.
Edith Sherwood
Edith: Philip Neal has similarly pointed to “ordered anagrams” (for want of a better term) as a possible explanation for a number of the features you point to here. Yet (as you already doubtless know I’m going to say) it is an explanation that yields more problems than it solves.
For instance, it does not explain the systematic difference in lengths between A words and B words, nor the different patterns that appear within words in different parts of the text (eg labels).
The late Mark Perakh inferred from his study of the former that there was strong evidence of abbreviation within Voynichese, where (for example) EVA -y might well be a token for ‘truncatio’, scribal shortening of a longer word.
The biggest problem with strongly ordered anagram proposals is that there are thousands of places where such rules cannot hold true. For example, any word such as EVA aror cannot be a strongly ordered anagram unless you treat or and ar as separately parsed tokens.
If you then abandon the strong ordering constraint, you end up with just anagrams: however, it is the strong ordering that you were relying upon in order to explain away the apparent structure of Voynichese words.
So: my counterargument would simply reduce to the observation that you need strongly ordered anagrams (not just any old anagrams) to explain much of the word structure, yet there are many thousands of Voynichese words that are incompatible with strongly ordered anagrams. Hope that helps.
Hello Edith !
Its been a long time since we discussed the palm trees and their use (oil for lanterns, etc, bronze work, ……
Not very long after I lost contact with you, I discovered Nick’s great forum. I still hold firmly to interpreting Fray Sahagun’s native South American scribes and artists works. No cryptology, no African scribes, Fray Sahagun received an education at Salamanca. Not long after finishing his studies, he boarded ship for “New Spain”.
Sahagun tells the rest of his story after arriving in South America. He taught at the School for Boys near Mexico City. Two or three of those boys became interpreters for Fray Sahagun — as well as becoming scribes and illustrators.
The so-called Voynich manuscript was the start of what would become an epic diary and illustrated manuscript of all the things of the “New World/New Spain”.
Compare the contents of the “Florentine Manuscript” with the contents of the “Voynich Manuscript”. I have been having a most fascinating ‘history lesson”, both on Nick’s “Voynich” puzzle pages and another fascinating discussion concerning the mysterious death of Ricky McCormick (twentieth century).
Hang in there ! Fascinating fun!
bd
Nick, You might change your mind about anagrams if you read my article, The Voynich plant names decoded, http://www.edithsherwood.com/index.php
I identify 111 plants by decoding their names using Italian as the language, the AVA alphabet, Florio’s 1611 Italian dictionary and a modified code based on anagrams. Check it out.
Edith: reading your article made me completely certain that there is no practical chance that anagrams had anything to do with the Voynich Manuscript. So it did change my mind, though not in the direction you were hoping.
But at least you’re looking for answers in the right century. 🙂
Edith, I must respectfully disagree with the introduction to your paper – as far as I’ve got so far. You speak as if everyone else described the plants as a ‘mishmash’ – Dana Scott certainly didn’t; neither did John Tiltman, nor do I. But some have recognised that the imagery is constructed to refer to more than a single plant – composite or ‘group’ imagery rather than the sole plant ‘portrait’. I do not see why that sort of economical expression should be misrepresented as if it were impossible, or somehow irrational.
You also say “Ethel Voynich correctly identified folio 9v, as representing a viola and folio 2v, a waterlily, folio 56r is obviously a sundew.”
I’m not sure that anyone can call any of the posited identifications “correct”; we can only say that we and/or others agree, and that the drawing’s interpretation is, or isn’t, adequately explained.
I agree that folio 9v represents a group of the violas, one of which is the tricolor. As I recall, your initial identification was as a pansy – though perhaps memory fails on that point. As for the “waterlily” it is plainly nothing of the sort, for it has a cup-shaped calyx and a protruding stamen, clearly showing the anther and filaments.. more like an hibiscus, really – except I know of no hibiscus with that nasturtium-like leaf – do you? But plainly no water-lily.
As for the drawing on f.56r being “plainly” meant for a sundew… I’d like to see the detailed analysis of the drawing which led to that conclusion. I have yet to see a sundew growing on a stem like that, and frankly to me the flower looks far more like a morning glory. Which is not to say your identification may not one day be proven correct, only that it is not at all “obvious” or self-evident. Sorry if this sounds a little cursory.
PS Edith – It will be no surprise to hear that I fuss about proper attribution and credits. For the past six years, I’ve been trying to correctly cite your qualifications – with a Ph.D it’s usual to add the name of the institution from which the degree was gained. Could you let me have that information – if not here, then perhaps by emailing me – address is on voynichimagery. Thank you.
Edith, sorry to trouble you again, but another credit which I may have to adjust. I had thought that my identification of Sumach (Monday, December 5, 2011) was the first reference to that plant in connection with the botanical folios. At the time, I did survey the current literature, including your website, but perhaps I missed it?
I see that your website is now dated “2015”.
I had identified Sumac[h] in connection with folio 16r, of course, though I recall that about that time your associate “Steve D” was regularly announcing discovery of plants which I had recently identified (others being e.g. the Indian spinach vine, the Kuzu etc.etc.), though always and rather curiously attached to some other folio, without any explanation of reasoning offered. I had thought it was he who announced (after I had) his ‘discovery’ of Sumach.. am I mistaken? Is it your proposal, and if so should your opinion be dated 2015?
Edith’s third and fourth points are directly relevant to this thread. There is a puzzling absence from Voynichese of short words – particles, prepositions, whatever – and indeed of very long words too. Torsten Timm’s analysis is interesting in its findings, although I’m not sure about the conclusions he draws, but he highlights this point very well.
If Nick is right, we should expect a fifteenth century Italian cipher text to preserve word divisions. Including one- or two-letter words ‘within’ longer words, or omitting them entirely as Edith speculates, seems difficult.
Otherwise, while I can see why Philip Neal raised anagramming as a possibility, I can only say that Nick’s explanation of why the structure of Voynichese cannot be explained by strong anagrams (where each letter in a word has to be arranged strictly alphabetically in its anagram) is spot on. Whether some other kind of letter ordering is going on, however, is another matter.
SirHubert: as Cicco Simonetta wrote, we should expect words ending in vowels, e to be the most frequent single letter word, la le li lo to dominate two letter words, and che to dominate three letter words if the plaintext is vulgar Italian.
http://cipherfoundation.org/older-ciphers/voynich-manuscript/cicco-simonettas-regule/cicco-simonettas-treatise-decipherment
Working backwards from there, I predicted that EVA d- is (and) and that EVA qo- is la/le/li/lo. As for che, who knows?
But that’s already too long a throw into the dark for most people, so what can you do?
Nick,
sorry – I’ll make this the last comment for a while – but when you speak of vulgar latin, do you mean one of the Italian dialects or clerical Latin?
From the odd encounters I’ve had with these in the medieval sources, it seems the various dialects sometimes had curious orthography. As example, Goldschmidt’s mention of “conventions of the Venetian dialect… In Venetian spelling x would stand for our consonant “sh,” ch for k…”
Wouldn’t that sort of thing wreak havoc with the way the Voynich statistical data is interpreted?
That article by Goldschmidt, btw:
E. P. Goldschmidt and G. R. Crone, ‘The Lesina Portolan Chart of the Caspian Sea’, The Geographical Journal , Vol. 103, No. 6 (Jun., 1944), pp. 272-278.
Diane: Simonetta used the phrase ‘vulgar tongue’ to refer to Tuscan. Venetian was indeed different, but – in its written form – not irredeemably so. Spoken Venetian, however, is still close to impenetrable for those not born to the islands. 🙂
Nick, Diane, Edith: Again (perhaps tediously) I refer you ALL to Fray Sahagun’s diary (the so-called “Voynich” manuscript) . I am begging you to compare the contents of the “Voynich” aka B-408 with Fray Sahagun’s fabulous manuscript which got traded around several very wealthy men — and ended up being given the name of the archive in Florence — the Florentine Manuscript. Only in the Florentine Manuscript will you find the identification of every artistic element of the so-called “Voynich” manuscript (B-408).
Why I say this, is because Nick has yet to naysay/or validate my various translations — right down to the last folios in the Vms (recipes which illustrations refer to combinations of the various edible specimens — and which DISCUSS whether hot (red) or cold (blue) liquids are to be used. Whether phamaceutical jars are being portrayed or rather measuring cups, will depend on whether hot or cold liquids are to be used. Most of the recipes are Nahuatl/Native American in use of their home botanicals.
bd
bdid1dr,
Within the limits of my ignorance of Nahuatl etc., let me say that I did spend some time looking at the manuscript(s) related to Brother Sahagun’s work in the new world and I agree that there are reasonable comparisons for some stylistic features of the Vms. The thing is that they are only customs of drawing – and I was eventually forced to conclude that when the whole range of imagery was considered those particular elements reflected the influence of “Spain or somewhere southern” and were not derived from native Aztec art. In other words, the comparable items reflect Sahagun’s Spanish education. I may be mistaken, of course, but that’s my considered opinion.
Fray Sahagun’s very first “teaching” experience was a black and white sketch of two persons kneeling on the ground: One person was a monk, the other a native. The written dialogue was Spanish/Latin : “First dig a hole” (ca-ui-tl) ….
I’m hoping someone can find that black and white sketch, with its accompanying dialogue. I can no longer find the URL. I am still 1-dering who was teaching who in that pictorial, and its discussion.
🙂
Diane, My PhD is in chemistry from Imperial College, London, where I held an 1851 Scholarship.
Diane and Edith: Take a look at my discussion of what some people are identifying as a sun dew. Compare with my translation of the entire folio which is discussing the ‘monks-hood”.. Which roots are very invasive in nearby gardens…..
Again, a reiteration and reference to the translated “Na-hua-tl dialogue which has been translated for Fray Sahagun’s later works (now called the Florentine Codex).
b-d-i-d-1-dr (which translates to beady-eyed-wonder) .
Oh, how I wish I could get Professor Leon-Portilla’s attention to the discussions on these pages. Also, I can’t find the published doctoral dissertation of a young woman’s illustrated (full color) speech on “Colors of the New World” — fabulous!
Her illustrated discussion of Cortez’ murdering of the Nahuatl chief/king was pretty scary, but revealing much of Cortez’ brutal acts — including the death of Malinche’s family members.
Again, Nick, I refer you to the web hostess for “Mexico-Lore” — whose home base is London.
bd
Nick,
Good to see you haven’t forgotten the VMs. The problem you put forth, regarding the apparent discrepancy between the sophistication of the Vms text and the reality of historical cypher technology has a simple fix. Take a spanner and loosen the nut that holds the VMs parchment date and the date of VMs composition as identical. Let the date of composition shift forward in time a few decades, perhaps, till the chronology matches your estimation of the author’s precociousness. Then just set it finger tight.
You mention Alberti and the Vatican. I have suggested that VMs White Aries contains a pictorial representation of the historical origins of the red galero as a traditional part of Catholic ecclesiastical heraldry and is substantiated by relevant, heraldic, armorial insignia in the illustration.
Folio 56r written discussion is in re Dianthus caryo-phyllacae and Laconian maids and Cariae district in SW Asia Minor, and Laconiae.
First line of discussion (folio 56r) : Olecax ce Solencentis — crassitudinus creberrimus ——sweet smelling, fragrant, crowded together. NO CODE.
Boy, am I getting tired of countless egotistical contrarian argumentation and self-praise instead of translations of the contents of the “Voynich” / B-408 .
N E U A — I continue to meet the challenges head-on: twenty folios and counting, including the recipes which illustrations are the last portrayals of what appear to be pharmaceutical jars and their contents (red/blue) indicating how much cold or hot water is to be used for any particular botanical item which appears in the Vms. The full discussions for each botanical/pharmaceutical item can be found in those folios which are “bulleted’ with little yellow “stars” asterisks.
The more the discussions get buried under argumentation or claims of being ‘first’, and the less actual identification (SPANISH (latin) and written Nahuatl) , the more befuddled the arguers become. Sorry about that folks.
There is NO code in the VMS. It was a rough draft (Latin and Nahuatl) for Fray Sahagun’s students’ identification and illustration of Fray Sahagun’s diary of his first years in “New Spain”.
Although Fray Sahagun was ‘pardoned” by the Inquisition, none of his written and/or illustrated material was ever returned to him.
Centuries later, the “VMS” was bought (from a small monk’s library) by Mr. Voynich.
What ‘looks like’ an ‘X’ or like a ‘ + ‘ in any fifteenth/sixteenth century document is referring to “Christ”. Thus one will find the use of + in many documents such as letters of approval made for +-opher Columbus to sail the “ocean/sea’ in in search of a route to China — issued by Ferdinand and Isabella.
One can find that document and accompanying discussion on the WWW.
BTW : Columbus never made it to China.
🙂
Howsomever, some monks DID make it to China — and returned home with their walking sticks full of silkworms/and larvae.
😉
Fray Sahagun’s “Florentine Codex” — Natural Things – discusses the silkworm, its larvae, and the leaves of the mulberry tree.
So, if you take a good look at that peculiar botanical item (in the VMS) which people have been guessing for years, with no conclusions being drawn, I am reiterating that it is a fruit of the mulberry tree. The Spanish and Nahuatl discussion which appears both in the VMS and in the Florentine Codex can be found by whichever online ‘reader-application’ (I use Adobe) one chooses.
bd
ps: besides the mulberry fruit’s illustration and discussions there is discussion of the tree’s bark (and the bark of the strangler fig) being made into paper. There is also discussion as to the value of the paper (pure mulberry bark being most valuable).
bd
Nick: Can we be sure as to what Alberti invented and what he learnt from others before him? I note that Piero Alberti, Leon Battista’s, uncle was very close friends with a Francesco, the uncle of an Abbot in the Novara area.
If only we knew what ciphers were being written in Milan prior to 1447.
One does wonder to what extent Simonetta and Alberti were inventors of new techniques or describing existing techniques.
Nick:
I don’t how exhaustive your research into early 15th century Italian ciphers was. What is the chance that there are still examples out there which would illuminate in particular the kind of ciphers being written in Milan at that time?
It would certainly be worth knowing your opinion on this if you have the time.
Nick:
My understanding is that in the early 15th century the profession of “cipher secretary” had not yet come into existence. Prior to this profession who within the government would be responsible for ciphers. Do we know?
Why would someone in the 15th century encode a manuscript of over 200 pages ?
Codes are fine for sending short messages of intelligence , but a whole book ?
xplor: indeed, the idea that someone was using codes or ciphers to “send” a whole book (as if it were a short message) doesn’t seem tenable. But there are plenty of other situations where a code or cipher could have been necessary (i.e. for numerous other reasons).
xplor: As you say I don’t think this was enciphered for the standard reasons of intelligence purposes rather for the personal use; so that the author could read the manuscript, but others could not, thus keeping the contents secret from others than himself.
I think an important question is whether the contents of the Voynich manuscript merited the level of encryption used. If they did then the contents of the manuscript were much more important to have been kept secret than those any other medieval manuscript. This seems unlikely to me especially given the apparent contents.
So I would argue that the author already had a deep knowledge and interest in ciphers and a somewhat paranoid streak. This would explain the motivation to encipher a document that others would not have.
My opinion on the “why” question:
if this question could not be answered, it should give a us a lot to think about.
The situation is that we can find several plausible answers. We just don’t know how close to the truth any of these answers is. And we may never know.
My favourite option (these days) is that it makes the book more interesting.
This is of course completely obviously the case today, but that is unlikely to have been on the composer’s mind.
It could have also made the book more interesting *at that time*.
xplor: I think we have to accept that the author must have been somewhat of a genius in the context of his/her time; at the least a very intelligent man/woman.
Rene: Obviously this whole subject is at this stage highly speculative.
The question as to whether the purpose of encipherment was to make the book more interesting to an audience is along the lines of the thinking of the “hoax” theorists it seems to me i.e. to make the book more saleable. For that purpose a “hoax” is arguably better than a cipher.
Personally I am not of the opinion that the author wanted to make the book more interesting to an audience. However I am of the opinion that he wanted to make it more interesting to himself. By this I mean if he was someone with a great interest and skill in ciphers it would be very tempting for him to write a manuscript that he already planned to right, in cipher. I suspect he would have wanted to keep the contents secret and what better way than to use his mastery of ciphers.
Note I am not saying that it is not an important manuscript merely that it seems highly unlikely that it’s importance in proportion to the difficulty of the cipher.
Of course one thing my thinking leads me to is that most probably he had developed skills of encipherment long before writing the Voynich.
Another question that I wonder about is where he would have learnt cipher skills assuming he didn’t invent the cipher completely independently of medieval cryptographic knowledge. I assume this knowledge or access to this knowledge would not have been commonplace amongst the general population at that time and as far as I understand medieval texts on the subject of cryptography were all written at a later date than the manuscript. So I would have thought he must have had some association with the diplomatic world in order to have acquired this knowledge. Do you think that is the case?
Mark: I suspect that speculating about where the Voynich Manuscript creator’s cipher knowledge came from will be fruitless. He (I say “he” for simplicity) seems to have approached ciphering from an angle quite different to the mainstream cryptography of the day.
Mark: as far as Alberti goes, his book gives us a very clear idea of what he invented, when, and why. And this leads no further backwards in time than 1466.
Simonetta, however, I believe was copying a pre-existing document that had grown up within the cipher section of Milan’s Chancellery, and which I would date to the 1450s or 1460s.
Nick: Thanks for the clarification regarding Alberti. I have not researched him in detail.
What you say about a “pre-existing document” is very interesting. Regarding this document what do you know? How do you date it to the 1450s or 1460s? Who do you think the author was?
Again our lack of knowledge regarding what was going on prior to 1447 in Milan, in particular, when it comes to ciphers raises many questions I think.
I think is quite probable that speculating about where the Voynich Manuscript creator’s cipher knowledge came from will be fruitless; however I don’t intend to spend a lot of time on this, but a little speculation might have some value.
We don’t know how far the cipher differs from ciphers of that time. Clearly it must differ significantly, but by how much it seems to me hard to say given that it has not yet been deciphered.
Mark: I covered why I concluded that Simonetta copied (rather than originated) the document in Curse 2006. I dated it by comparing its model for cryptography against the ciphers in the Milanese cipher ledger, which contains a large number of ciphers of its time.
However, we know from its [i.e.the Voynich Manuscript’s] structure and from the numerous statistical experiments and analyses that have been carried out that it is based on a radically different (though not necessarily much stronger) approach to cryptography from that which we have evidence for.
Nick: Sorry I will need to reread that.
One wonders whether any such document would have been based on an even earlier pre-1447 document or knowledge from that time passed on by word of mouth.
The question of what was going on in the early 15th century in Milan keeps rearing its ugly head. And one wonders what exposure and knowledge of ciphers Simonetta acquired in his earlier life in Milan.
Nick:
I am always suspicious of things appearing out of nothing. However we do find examples of people who made major advances in knowledge. So it is possible the author made a completely radical advance or change in cipher design from anything around at that time.
However the idea that the cipher is in no way influenced by contemporary cryptography is hard for me to swallow.
Mark: again, we know a lot of what Simonetta was doing in his earlier life, which was basically marching around the top half of Italy as an administrator in Francesco Sforza’s mercenary army. There is a 1440 cipher ledger from Urbino which seems to have a significant overlap with the kind of ciphers the Sforzas used post-1450 in Milan.
Mark: I’m always suspicious about that too. But in the examples you mention, we have actually quite good knowledge of what was going on – Alberti, for example, knew very little about ciphers pre-1466.
Nick: Did you see the email I sent you with 2 spellings of the same word?
The reason this struck me other than the similarities between the spellings is the following:
Through my “map of a journey” analysis I was lead to believe that the text corresponded to the same word(s) mostly likely the River Ticino. I could see the similarities and it may sound strange, but this really disturbed me. The idea that two symbols could correspond to the same letter was something completely unfamiliar to me and frankly shocking and mischievous it seemed to me on the behalf of the author. You must remember I have no background in Medieval ciphers.
So subsequently at a later stage having read that the use of 2 separate symbols for one letter of the alphabet was something recommend by Simonetta was a bit of a revelation. Now this could all be coindence and I daresay that will be the standard view of this kind of thing. Nevertheless I can’t help, but find this very curious.
The use of null characters seems very likely and is also consistent with Simonetta’s advice.
As far as the use of more than 1 language goes there is no concrete evidence of this I believe, but it seems very plausible.
The Humiliati were in Milan at the right time.
They wrote a lot of things in cipher could the Voynich be one of them.
Nick: I wonder what the most advanced, practical cipher one could come up making use of all the known or possibly known, state of the art, cipher techniques of the time. Clearly there could very likely have been a technique of the author’s own devising added to the mix.
When it comes to the statistics I wonder as I have written before what they say about single word isolated labels and the scope for their explanation, independent of sentence text, by the techniques of the time.
The scope for throwing a lot of noise into sentence text makes statistics pertaining to it less interesting as the scope for distorting the statistics is much greater.
I wonder also about the quest for finding a transformation or more appropriately a reverse transformation which once applied to all single words has the effect of altering the statistics in an interesting way. So for example one could run an algorithm to look through a set of transformations corresponding to a specific technique and then calculate the statistics and somehow identify those with interesting, potentially more like the statistics of normal text, features.
Clearly this idea is very vague and may not be computational feasible, but it has the potential advantage of providing an intermediate step towards solving the cipher without having to go the whole way.
It did occur to me regarding for example the Alberti cipher that the author could have discovered/invented the same idea or a variant of it independently. That is far from unusual. For an example for that broad period one merely needs to think of the priority dispute between Newton and Leibniz over who invented calculus; most modern historians believe they invented it independently.
The Humiliati was known to use the atbash cipher. The problem is knowing what alphabet was used.
. “No one expects the Spanish Inquisition! Our chief weapon is surprise, fear and surprise; two chief
weapons, fear, surprise, and ruthless efficiency! Er, among our chief weapons are: fear, surprise, ruthless
efficiency, and near fanatical devotion to the Pope! Um, I’ll come in again…”
You might find the following of interest:
https://www.theguardian.com/books/2017/jul/05/author-of-mysterious-voynich-manuscript-was-italian-jew-says-scholar
The evidence Skinner uses to support his theory include the lack of Christian symbolism in the manuscript . Also Christians would have avoided the symbolism if the were translating Greek, early Roman or Etruscan. All having more gods than you can shake a stick at.
The Humiliati had the modus operandi and money to have the book written the book. Did they use passagini ? All of the men ended up being heretics.
Nick, regarding your statement to Diane that there is no foundational botanical research on the Voynich Mansucript. I have it on good authority that two well prepared and recognized botanists have a book contract signed with a major American publisher. No names, of course, but their initials are T & J.
As for no documented Arabic influence in European code making, I would remind us all that, according to a French saying, Africa begins at the Pyranees.
All the best, John
John: I would welcome some reliable plant identifications, but I am not holding my breath.
John: I of course look forward to the appearance of any well-researched account of the Voynich Manuscript, botanical or otherwise. However, given that it has to date made fools of nearly everyone who has tried, it is a high wire that has far, far more downside than up. 🙁
Nick: What do you know of the role of the secret chancery? Would ciphers have been managed by the secret chancery?
Mark: quite a bit, that what was I spent a lot of time reading about back in 2005-2006. Do you mean specifically in Milan? The answer to the second question is basically yes, in Milan and elsewhere.
Nick: I am gradulating drilling into to the close family members of the person who I am interested in, many of whom where in Milanese government. I have isolated the following:
1) The most famous, and glamourous(interesting life story), who effectively establish the Milanese Chancery under Gian Galeazzo Visconti.
2) Another who was a member of the Secret Chancery and Ducal Secretary under Filippo Maria Visconti.
3) Another who was a prominent Ducal Secretary under Filippo Maria Visconti and subsequently Milan’s Ambassador to Rome under Filippo Maria Visconti and continuing under Francesco Sforza. I am excited at getting a look at his enciphered letters.
4) Another who was an Apostolic Protonotary, which was a top level Vatican Bureaucrat during the Visconti era. I have no idea who would have been responsible for writing Vatican ciphers.
5) There are others most likely less important, as I can find out less about what they did, who were Ducal secretaries in the Visconti era.
6) There were also descendants who were prominent secretaries in the Sforza government and a family member who married the brother of Cicco Simonetta.
7) There was a later one, which is most likely incidental, who was the envoy to Holy Roman Emperor Charles V.
The individual who I am interested unlike most of his relatives, though not all, adopted the religious life, although he played a very significant political and diplomatic role in the region in which he was based.
One thing I must admit that I didn’t anticipate and that I find very frustrating is that I find that people are frequently referred to by different names such that it can be quite a job to figure out that these are not 2 separate people, but one person.
Lets take you. In this context you could be:
Nick Pelling
Nicholas Pelling
Nicky Pelling
Nic Pelin
John Pelling
Jonathan Pelling
Jon Pelling
Nicolus Pellingus
Nicolas Pellingis
and more…
(I am not including Orlando Pilchard as fortunately I don’t have to worry about pseudonyms.)
This is a real pain for searching electronically.
Working with the mere scraps of information that I can find through meticulous online research is trying.
I am pursuing different avenues simultaneously, but one focusses on slowly getting as deeper into the world of Visconti ciphers in its own right, but also the connections to family members of the person I am interested in.
I suppose one thing which might help me better understand where to look for records like the Filippo Maria Visconti era Papal letters in other archives is the precise mechanics of how the cipher production process, implementation, transfer etc. actually happened. Whether there are likely to be ciphers used within a state between different towns, like say Alessandria, Novara or Pavia, and Milan. I aim to gradually get better at learning where to look for Visconti era cipher correspondence if it exists.
I would note that my proposed author was named without knowing many of his family members where closely associated with the world of ciphers. His life is also much more consistent with writing a herbal manuscript than that of his relatives. However this could be a coincidence; distinguishing between pattern and chance is very difficult. Anyway if even if my author is not the correct author understanding the Visconti era ciphers is a valuable endeavour in its own right.
So if you have advice it would be greatly valued.
Nick: One thing I also find annoying is dealing with mutually contradictory sources.
Nick: An example
In one source it says:
Nick Pelling son of George Pelling born in 1964 in Canterbury.
In another source it says:
Nick Pelling brother of George Pelling and son of John Pelling born in 1954 in Camberly.
So what do I conclude on the basis of this information. Well at the moment I look what the general opinion is amongst sources. In the future I may need to explore the most primary of sources
Nick: Thinking about it I wonder things like how many Ducal Secretaries there were roughly 4 or 15 or 40 or ? How many members of the Secret Chancery there were roughly? This would help me establish the significance of these positions. I suppose I could undertake the very ardous task of trying to compile a list of members of the Visconti Secret Chancery over time.
It is lookings like Ambassorial records from Rome seemingly show the only examples of Milanese ciphers from the pre-Sforza era that I know of, although it is true you mention the “In Milano” cipher reference in Meister. (I don’t know if someone has made electronic copies of these letters and so when I will be able to see them first hand. Like so many things they could be very illuminating or nothing special. Obviously the dream is finding distinctive “Voynich” cipher letters other than the 4o outside of the Voynich somewhere though this may well be a forlorn hope.)
I wonder if there are similar earlier records of Milanese enciphered correspondence in the Vatican archives and if there are how I would go about finding them.
Could there be similar Milanese Ambassorial records in the archives of other states from this period?
I suppose I probably ought to compile lists of Ambassors from Milan (I don’t know when the role of Ambassor was introduced.)
Let me know if you have any thoughts, if are not busy.
(Maybe you already know of other pre-Sforza examples of Milanese ciphers that I have missed.)
Nick: Sorry for going on.
The guy I am focussed on I think played an important political role in addition to his other roles in the Western buffer zone of the State of Milan. I would only guess that political disputes with the Duchy of Savoy would be the mostly likely ones though I haven’t studied the changing political dynamics in Western Novara in the Early 15th Century.
So I wonder if there could possibly be enciphered correspondence by him in the records of Savoy???
I have been trying to research the earlier part of his life, but it looks like I am really working with scraps of information though I may be able to find more on the ground.
Mark: General opinion amongst mutually contradictory Chaucers, if that be what you’re seeking, might just prove to be nothing more than another old Canterbury Tale; perhaps therefore to be taken carefully with a fine grain of Milanese salt.
Mark: there was only one Ducal Secretary in Milan for decades, and that was Cicco Simonetta – everyone else in the Chancery (which was a surprisingly small part of the Castello Sforzesco) worked for him. There is an entire archive file of enciphered documents in Milan (you occasionally see it mentioned), something I’ve been meaning to visit to survey for over a decade now…
Nick: Maybe we are talking at cross purposes and I didn’t make it clear that my focus is really only on the Visconti in Milan at this time in my research.
My understanding is that Cicco Simonetta was not a Ducal secretary in the Visconti era. I don’t believe unless I am mistaken there are any enciphered documents from the Visconti era in Milan due to the destruction of the castle where the Secret Chancery was based.
From what I have read Simonetta, though I must confess that I have not studied his life in detail, worked directly for sometime for Francesco Sforza prior to him becoming Duke as well as once he became Duke and was not a or the Ducal secretary to Filippo Maria Visconti. I believe he did not achieve his political stature immediately on Sforza becoming Duke, but over a period of time.
I wonder if there is a distinction between Ducal Secretary and a secretary working for the Duke, if so that is very interesting and relevant to my research,. I have assumed they were one and the same, but maybe they were not. There certainly seem to have been more than one person working as a secretary to the Duke at one time.
The fact that you say the Chancery had only a few people working in it is a helpful sign, so thanks for mentioning that.
I may be wrong on some of these things so I greatly value your direction and correction.
Thanks so much for your input!
Nick: To clarify the era I am focussed on in this part of my research is largely from about 1390 to 1447 in Milan.
Nick: I have just read on Treccani that Simonetta was Governor of Lodi in 1449, so he can’t of been Ducal Secretary at this time. I think he was really much more closely tied to the career of Francesco Sforza than he was to Milan per se and only began his role in Milanese government when Sforza came to power.
The family I have focussed on have very long ties specifically to Milan.
Nick: From what I have read, again I am not an expert the career of Cicco Simonetta:
There was NOT only one Ducal Secretary in Milan. The was one First Ducal Secretary, but they are multiple people who are described as Ducal Secretaries.
I have read that Simonetta had 2 secretaries and 12 chancellors working for him.
Please correct me if I am wrong.
Nick: I wonder to what extent Cicco Simonetta, despite having written or commissioned the writing of a historically important text on cryptography, was an expert on ciphers or whether in practice the secretaries working for him were responcible for the writing and understanding of ciphers.
Mark: there’s a difference between job titles and actual power – Simonetta had both. 😉
Mark: Simonetta was part of Francesco Sforza’s entourage from a young age, so had nothing directly to do with the Visconti.
Mark: I suggested in Curse that Simonetta didn’t seem to be hugely proficient in ciphers, but who knows for sure?
Nick: I completely agree with you when you say “there’s a difference between job titles and actual power”. My aim was to understand what was going on within the government and there are a number of people who are said to be Ducal Secretaries and in some cases members of the Secret Chancery, so it would be nice to get a idea of the precise nature of their role and how deep their involvement with the writing of ciphers was. Having cipher writing skill does not necessarily mean having huge power, but it is more relevant to the writing of a cipher manuscript. A family with many members who had important roles in Milanese government and for whom many may have had a great familiarity with the writing of ciphers becomes interesting with references to those family members not working directly in Milanese government and who may have learnt the skills from brothers, uncles etc.
For example, if Simonetta wasn’t writing Milanese government ciphers then specifically who was/were? Who had that know how and what were their job titles?
Ultimately we are interested in the permeation of cipher expertise to someone who might also have herbal and astrological expertise or interest whilst also possibly having a diplomatic regional role where cipher writing skills could be of use.
As Simonetta was not involved in the Visconti governments he becomes a figure of much less interest to me. I hoped by repeatedly saying I was focussing on the “pre-Sforza era” in Milan I had made it clear that I was primarily interested in the Visconti period i.e. before 1447 in Milan. I believe that one is mostly likely to see some of the most recognisable features of the Voynich cipher, even if not on the same technical level, in Milanese ciphers from about the period 1390 to 1435; that being possibly other familiar cipher characters or distinct certain simple structural features of the cipher. The problem one faces is the sparcity of Milanese cipher records in Milan for that period though there may be some records of Milanese cipher correspondence in the records of other states.
Mark: the technical difficulties of researching Visconti-era Milan has, paradoxically, made the topic more attractive to historians in recent years, so you should be able to find plenty of books and articles covering the period.
Nick: I have found some stuff and made some relevant contacts, but it looks likely that there is plenty more research to be done.
I hugely appreciate your suggestions.
Nick: As a follow to my comment on Milanese enciphered letters, I thought I would mention this page.
I believe here in your discussion you only mention diplomatic ciphers, yet you allude, in your comment against your other post that there were numerous non-diplomatic ciphers being written at the time. Is it the case that the non-diplomatic ciphers written at the time were less sophisticated than the diplomatic variety?
The more that I learn of Northern Italian politics in the early 15th century the more complicated it gets and it appears that texts on the subject don’t cover the level of complexity such as:
The different alliances and peace agreements.
The various conquests, by hired mercenary armies, and frequently changing boarders.
The complex involvement of the church, not just as the Papal States, but also with regard to religious loyalties of priests to their own state versus their loyalities to the Pope. (And also the affect of the antipope and other heretical movements.)
The influence of the Holy Roman Emperor.
The involvement of foreign powers.
The internal politics of given states
And so on…
Given all this the use of ciphers as part of diplomacy is hardly surprising.
Having just perused this article again, I believe it would be mistaken to think that many of the techniques used after 1450 were not first applied some time earlier. Obviously dating where and at what time each innovation emerged would be hard to say, given the sparce records across the board, and more interestingly who were the main innovators as presumbably innovations spread from one city state to another.
Anyone know what is in->
Costamagna, Giorgio, Tachigrafia notarile e scritture segrete medioevali in Italia, Rome, Edizioni dell’Associazione nazionale archivistica italiana, 1968
Can’t seem to find any information about it.
I guess the following would be of no interest to me:
Le Zifere o della scrittura segreta de G. Battista Della Porta
Mark:
Since Hathitrust provides only a limited online view (search only), you should contact a library https://www.worldcat.org/title/tachigrafia-notarile-e-scritture-segrete-medioevali-in-italia/oclc/624419448&referer=brief_results
Mark Knowles
Giorgio Costamagna, 1916-2000 published literally dozens of books and articles about medieval secret scripts and tachygrapy, the book you quote is about several of these systems
Helmut: Thanks for that!
Yes, I have seen that there are other books by him. I note quite a few focus on Genoa.
I guess I am going to have to have a look at them to see what is in these books. Obviously my focus is on the first half of the 15th century.
(Side note: the Republic of Genoa was conquered by Filippo Maria Visconti, which may or may not be relevant.)
Just got Lydia Cerioni book for 3 weeks; I very much doubt I will need it that long.
I looks like, at first glance, much or all of it is in the Tranchedino files Nick has already shared. However the dates and who the cipher alphabet is associated with appear separately making it much easier to read them.
There are clearly some letters associated with the great nephew of the person that I am interested in. However I doubt they will be of much relevance as they are dated 50 years or more later; still its nice to see them.
Unfortunately there appear to be no early letters though there are a couple dated to 1444 and 1447; I believe.
Has anyone seen if there are any diplomatic cipher alphabets in the Giorgio Costamagna books and if so when those alphabets date from?
Mark: to answer such a tightly focused question, the chances are surely high that you’ll have to walk that road yourself, sorry. 🙁
Nick: Don’t worry. I just asked in case someone knew the answer. I have a feeling that the further I go the more I will have to walk alone as there will be no Voynichero who can help. Anyway thanks for replying, I will trudge on.
Giorgio Costamagna was the Director of the State Archives of Genoa. It appears that he wrote about 15th century ciphers. I hope his writings can give useful insights to not only diplomatic ciphers, in general, of the first half of the 15th century, but hopefully also Milanese ciphers of that period given that Genoa was under Milanese control for some of that time.
A thought about copyright:
I would like to make available to the Voynich as much as possible the cipher records that I discover. However some may be in books which may or not may have passed their copyright date.
Purely as an example, take the Lydia Cerioni book which was published in 1970. This contains cipher records scanned/photocopied/photographed straight out of the cipher ledger and so the records themselves are not subject to copyright; however the author compiled them. So I think that if one ONLY copies the images of the records they are unfortunately still under copyright. But also in my example of Cerioni, 1970 is a long time ago, so I am not quite sure when the copyright expires. If I had the records located and rescanned, obviously they would not be subject to copyright, however this would be a hassle to do.
Nick,
A couple comments (although I’m not a crypto guy by training — like you if I’m not mistaken, I work in computer vision with a robotic vehicle focus):
“homophones were instead added as a steganographic defence”: is steganographic the right word here? My understanding of steganography is that it comprises methods to conceal the very existence of the ciphertext (sometimes under a cover text such as a Baconian biliteral or a Cardan grille) — that’s not really what you’re describing here.
“Actually, the fact that by 1460 homophones were typically added to all letters (though usually with one more homophone for vowels than for consonants, as per the Napoleone Orsini cipher key above) further implies that even by then there was still no conception of frequency analysis driving the choice and number of homophones.”: I’m worried this projects an essentially 20th century post-Index-of-Coincidence/Shannon understanding — that the “right” thing to do is flatten the frequency distribution to maximize the entropy (rather than just disrupt the typical distribution of the underlying language) — onto a much earlier time.
“As Simonetta’s treatise points out, one specific feature made Italian (Tuscan) highly vulnerable to trivial code-breaking: because every word of the language ended in a vowel, it was extraordinarily easy to just read the vowels off the page. And once you had identified the vowels, other common 2- and 3-letter word patterns made it relatively easy to work out the consonants, not unlike solving a crossword. Consequently, the thing Italian cipher-makers circa 1400 needed most was a simple trick to help mask the vowels at the end of words: and this is precisely what we see homophones being used for in Simeone de Crema’s cipher key.” — Doesn’t mask them much, since the characters/glyphs at the end of words will still be the homophones for the vowels. May make telling an Italian ciphertext from a Latin one more difficult, though.
Karl
One additional thought: _computationally_, is the primary effect/purpose of homophones flattening the frequency distribution, or is it increasing the combinatoric complexity of solving the cipher? Has anyone looked at whether solving substitution ciphers with homophones is NP-hard (probably by reduction from some bin packing or knapsack-like problem)?
Karl
Karl: my point was made in contradistinction to Kahn’s reading – I argued in my paper that homophones were added not as some kind of frequency flattening statistical defence, but as a means of concealing the (otherwise blatantly obvious in Italian) vowels at the end of words.
My second point was also made in contradistinction to Kahn’s account, which would appear to be projective in the way you describe.
Your third point is also my point, i.e. that vowel homophones helped a bit but not a lot. All the same, the Quattrocento was just starting. 😉
Nick; I contacted you by email regarding the about 70 cipher keys at the back of Volume 2 of Lydia Cerioni, listed as “Tavoli”. I assume that you do not have a copy of these. I can share them with you on Dropbox, Google Drive, whatever you prefer.
Is that something that you would like me to do?
I think it would be a good idea that you have a copy as you would be best placed to share it with anyone who may be interested in the future, saving them from the annoying task of tracking down the Cerioni Volumes.
Nick: Do you know what the rationale was for not photocopying all the pages of the Tranchedino?
Mark wrote: “Nick: Do you know what the rationale was for not photocopying all the pages of the Tranchedino?”
Maybe the people at Adeva who spent time and money creating the facsimile would like people to buy the book so they can feed their families.
Mark, one of the things that one notices when looking through the Tranchedino codes is that a large number of them are built on the same concept.
The glyph shapes and assignments vary, but the concept is the same: one-to-many character assignments, plus quite a few nulls, plus a list or two of common names and words that are compressed into one glyph. It’s a bit of a “if you’ve seen one you’ve seen them all situation”.
.
It’s almost the opposite of the structure of the VMS in terms of the raw number of different glyphs that are available and how they vary internally when composed as sentences.
JKP: The Lydia Cerioni volumes I describe are Out of Print and I could not find a second hand copy for sale anyehere. So if someone is expecting to feed their family off the revenue from the book it is a very poor strategy.
Mark, when you said, “Nick: Do you know what the rationale was for not photocopying all the pages of the Tranchedino”, I thought you were talking about the 1970 facsimile of the Tranchedino codes (which is still available new).
JKP: I have addressed some of your concerns in a comment elsewhere. I will try and track it down.
Nick, I owe my thanks to Edith Sherwood for her comment above ( July 21, 2016 at 8:21 pm) which I’ve only just seen. Apologies to Edith for earlier failure to respond.
I’m only just reading the July 2016 message from Edith Sherwood for the first time, but since these issues still come up quite frequently, I’d like to address some of the points…
“Nick, with respect to the Voynich Manuscript’s code, consider the following:
1. The VM consists of over 200 pages, written fluidly with no corrections. This indicates that even if the manuscript was copied, the code application and the alphabet must be relatively simple.”
I would agree that it’s written fairly fluently, but there are many corrections and a number of mistakes and “slips”. One does get the feeling that the text may have been at least partly prewritten (perhaps on wax or on discarded parchment) and then copied fresh because the errors are not frequent.
“3. Although simple words representing prepositions , conjunctions , definite or indefinite articles are expected, there are no single and few double letter words except at line endings. Were they omitted or included within other words?”
I don’t know why people say this. There are hundreds of single-glyph tokens and many of them are midline. EVA-y and EVA-r, in particular, can frequently be found midline. In addition, there are about two thousand double-glyph tokens with “or” “ol” and “dy” being some of the most common.
“5. A sequence of words that differ by only one letter occurs fairly frequently in the VM’s text and cause single letter substitution codes to yield babble-like text. For this reason Elizabeth Friedman has stated that simple substitution codes do not work.”
There are indeed many tokens that differ by one glyph. This kind of repetition is more common in old middle eastern and Asian languages than in western languages, but certainly not as common as in the VMS.. A more important reason why substitution codes will not work, however (in conjunction with the homogeneity), is the positional rigidity and glyph distribution.
“6. Why are some of the VM’s symbols or letters, like ‘tl’, found at the beginning of a word, or a ‘g’ at the end of a word? Was the order of the letters in a word not important?
The above observations have caused me to conclude that the VM’s code is probably based on anagrams, with prepositions, conjunctions, definite or indefinite articles included within words and longer words being split…”
I think it’s quite possible that some tokens have been split (and that some may be joined), but anagraming still doesn’t bring Voynichese entirely into line with natural languages. If you take tokens like chy and ohdy that are near neighbors to something like shy and okdy (i.e., tokens that are similarly short and repetitive), combining the tokens PLUS anagraming still does not yield the kind of variety that is characteristic of natural languages (and certainly not of narrative text).
There are quite a number of possible explanations for glyph repetition and homogeneity at the beginnings and ends of tokens. They might be markers, modifiers, joiners, splitters, misdirection (e.g., making the text deliberately resemble Latin), numerals, or abbreviations (e.g., in languages that use Latin scribal abbreviations the “9” shape at the end of a word expands out to a number of endings).
Anagraming is only one of many possible explanations and only seems likely if done according to somewhat predictable patterns, otherwise it results in one-way ciphers.
-JKP- a neat summary of the many things that have been said by various people over the years. I must say it speaks well for your range of reading that you also include Nick’s observation that the text may have first been written on wax tablets. I won’t try to add footnotes for each of the other (accurate) re-statements. 🙂
Diane, none of that post is based on my “range of reading”…
My response is based on an intimate knowledge of the VMS text. I have created three full transcripts (the first in 2008), and have just about completed a unique fourth one. I have studied every single character in the manuscript in detail. I have a disk full of statistical charts that I generated myself.
.
Also… my mention of wax tablets is completely independent of others. I didn’t even know they existed 10 years ago when I first toyed with the idea.
I made a rational guess that if I lived in the Middle Ages and wanted to “prewrite” something (like an unfamiliar alphabet, language, or cipher) before committing to expensive parchment, there must be ways to do it and came up with the following:
1) write it in sand (imagine a little low dark-colored box filled with a thin layer of light-colored sand),
2) press it into butter, lard, or agar (this would only work at certain temperatures),
3) pre-write it on parchment or fabric with pigment that washes out easily, then clean the sheets as they are copied so they can be re-used, or
4) press it into wax that can be warmed and re-used.
It wasn’t until I met the Voynich community that I heard about wax tablets, so I Googled “wax tablet” and had a “d’oh!” moment when I discovered they not only existed in the Middle Ages, but were “standard issue” for upper-class school boys.
.
Every time I write something that comes from my own research, you pat me on the back for doing a “survey” or “summary” of other people’s research. It may look that way to you, but you are mistaken.
.
In fact, my research is frequently at odds with others….
As I mentioned upthread, people constantly claim there are very few single or double glyphs in the VMS and that they are mostly at the beginning and ends of lines. I KNOW these statements are untrue because I have exact numbers that I mined from the VMS that PROVE it’s not true.
I see frequent claims that there are no mistakes or corrections in the VMS. There are HUNDREDS of them. Years ago I started documenting them, but it was too time-consuming—it’s better to study what is there than what was changed.
.
If my ideas agree with others it’s because we think alike or have walked the same path. I live for the thrill of discovery. I lose that if I spend my time on historical research and I’m almost invariably disappointed when I get talked into it (e.g., Strong, Brumbaugh, Newbold, etc.)
.
Nick seems to grasp what the VMS text is and what it isn’t (which is why I thumbs-up his blog) but most people can’t seem to break away from simple-substitution-code mindsets or impractical ideas. Reading about their failed attempts is not a productive use of time.
JKP: It is useful that the Tranchedino is available online, I did not know that. However what I was sharing with Nick was not from the Tranchedino specifically, but rather other cipher keys from a end of the Cerioni book; now I don’t know if that is available for sale online though I would be surprised if it were. The Cerioni book also contains the Tranchedino. However the Peter Kazil scans that Nick shared with me seem to cover only up to page 100. I am perfectly happy to pay for things as generally it saves me time and therefore it saves me money overall as opposed to doing the research. Unfortunately I think the cipher alphabets/keys that I am most focussed on trying to compile are not available for sale, so I have to do my best to find and compile them myself.
JKP & Nick: Also I should say that these days I greatly prefer books in electronic form as I find them much easier to work with and I can study them on the move etc. That is why I am looking forward to the day when NIck makes the Curse available to buy in electronic form; the paper copy is excellent, but an electronic copy would make it even more handy. (I don’t know if Nick would be happy if at some stage I make my own scans purely for my personal use.)
JKP: I was wondering what you thought of the Adeva facsimile. Is it really good quality?
Mark, the quality is good. I’m not disappointed with how they put it together, but the content, the codes themselves, are not as varied as I expected. There are many pages of codes, but…
As I mentioned in a previous post, most of the codes follow the same basic template. There is an alphabet at the top, with substitutions, sometimes two, three, or four substitutions per letter.
Under the alphabet there is usually a line of characters used as nulls.
Beneath that are a couple of columns of place names and dignitaries that have been compressed into one character. Thus, a king’s name might be represented as a plus-sign, for example.
This results in a one-to many cipher for basic text and many-to-one cipher for common names and place-names, using a great many variations on glyph shapes.
For the most part, each code in the book is the same basic format, only the shapes of the glyphs change (with just about every shape under the sun represented—Latin abbreviations, Greek letters, alchemical and astronomical symbols, math symbols, etc.).
So, it depends what one wants to get out of it…
For a historian/cryptanalyst trying to match up actual coded messages with a SPECIFIC configuration of one of the ciphers, it would be an essential time-saving reference.
For a basic understanding of the substitution systems collected by Tranchedino, you only need to see about six pages of the book to grasp their basic method, so the bang-for-the-buck is probably not high enough to justify purchasing it new.
Another possibly useful reference:
Volpicella, I cifrari in gergo. Su « Miscellanea di studi storici» in onore di A. Luzio. Vol. II, pag. 379., 1933 Firenze.
Nick: I have sent you an email, so you can download the rest of the Tranchedino that I have scanned.
Nick: As part of my research having inevitably come across cipher alphabets from outside of Milan and from periods before or after the 15th century it is quite apparent that the many similarities between symbols in the Voynich which I see in alphabets from Milan in the 15th century do not exist or exist to a much lesser extent in cipher alphabets from other times and places. So as examples if one looks at the 16th centiry cipher alphabets in “Die Geheimschrift…” by Meister or the cipher alphabets from the other city states in “Die Anfange…” by Meister or the not diplomatic cipher alphabets in “Ubersicht…” by Bernhard Bischoff then the differences become more apparent. In fact cipher alphabets outside of Milan in the 15th century serve a useful purpose in illustrating the relative similarities and dissimilarities with the Voynich symbols. You have tended to focus on the 4o symbol, which is understandable, but I think it can ignore the broader similarities between symbols or more specifically the other identical symbols that are shared with the Voynich. Take for example the Codex Urbinate, this shares the 4o symbol, but there is little in common with the other symbols there and those in the Voynich. Likewise with the Florentine cipher alphabet from 1424 in Meister, it shares the 4o symbol with the Voynich, but again we don’t seen the commonality amongst other symbols.
I will go through each of the symbols in detail at some stage in my writeup, however I have plenty of more research to do into this area as you well know, having been somewhat waylaid by the Cerioni books.
Nick: Another thought, study of the cipher records and enciphered communication in the Cerioni books should give a clearer insight into with who and how diplomatic communication was taking place i.e. to what extent the ambassor or envoy communication model was the status quo and if communication structures were more complex. This might then shed light on how enciphered communication operated in an era with fewer ambassors or envoys in the “Primo Quattrocentro”.
Rene: You mentioned an Appendix to Meister that doesn’t appear to be there in the free download. Are you sure it exists? If so do you have an idea of what is in it?
Nick: I imagine that Meister picked a selecrion for each city states over a range of time periods for his book to give an overview of Italian ciphers in the 15th century, so I wonder if there were more cipher keys found in a given collection that he chose not to include.
Also I was thinking about the process of converting an enciphered letter to a cipher key/alphabet. Clearly for a short enciphered letter one would only arrive at a partial alphabet. However for a long letter I would have thought that this could be used to produce an almost complete cipher alphabet. Ultimately I am interested in cipher symbols and so cipher alphabets therefore my interest in the contents of a given enciphered letter is in most cases very low.
Nick: I have uploaded quite a bit of Lydia Cerioni – Volume I. There looks to be more of interest to me there than I at first suspected. I don’t yet know if I will end up scanning it all or not.
Nick: I was thinking that the complexity we see in the “gallows” characters is akin to the complexity we sometimes see in the special characters corresponding to a single word or thing such “Pope”, “Duke of Milan”, “War”, “Peace” etc.
Now I am certainly not suggesting that the “gallows” characters operate in that way at all, but rather that symbolically that appears to be the closest parallel. In fact I doubt that there was the need for a character corresponding to a single word in the Voynich as there wasn’t the same requirements as in diplomatic ciphers.
Thanks a lot, Nick, for the Senatore scans! Very kind of you.
I will work my way through Senatore’s text. Regarding the quote:
“The earliest examples of Milanese ciphers dating to 1447-48
were rather backward compared to those of the other Italian states.”
Whilst I am confident that Francesco Senatore is a real expert on this subject, I am somewhat dubious of the accuracy of the statement as from some ciphers of the the pre-1448 period that I have seen there seems to be quite a high level of sophistication when compared to those that I am aware of from other city states.
I think this may be a case where one of your favourite phrases applies: “Absense of evidence isn’t evidence of absense.”
However Senatore may persuade me otherwise.
Nick: Clearly the Voynich cipher is not the same as what we see in the diplomatic ciphers, though we have no idea how much in common they have.
However as an experiment has anyone tried to solve it as if it were a diplomatic cipher and then see where it comes unstuck. Have you ever tried to decipher an enciphered letter yourself without the use of a cipher key i.e. by constructing your own cipher key for the letter? One can refer to Simonetta’s advice for suggestions on how one might do it.
It might be a useful exercise and if one can identify what if anything is in common with diplomatic ciphers we might be able to better understand what is different.
Mark: people have indeed tried to solve the Voynich Manuscript’s text as if it were a fifteenth century cipher of exactly the kind we are discussing here, though without any obvious luck whatsoever.
I have never seen a list of traditions or lone documents we know of that use any of the voynich gylphs. It would be really helpful to have such a list if for nothing more than understanding the context.
I think one of the biggest problems with analysing the Voynich Manuscript is that there are layers of uncertainty about how to break it down. First off we have transcription, both our reading of the characters as well as errors made when it was written. Second is that beyond Currier A & B there seem to be other less drastic changes where the usage/system seems to change in ways that aren’t just explained by the underlying subject matter being different.
The author almost surely had some way to tell other than purely the memory of each section.
Mark, the glyphs in the VMS are positional. Certain characters at the beginning, certain characters in the middle, certain characters at the end. I have never seen a natural language substitution code (which covers most of them) or any other kind of cipher that has this specific underlying structure.
Nick & JKP: I guess that I am wondering if anyone here has practical experience deciphering 15th century diplomatic enciphered letters. As a hypothetical if I were to email someone an enciphered letter can they solve it. I think someone who is in the habit of cracking these ciphers might have a better handle on solving the Voynich.
Obviously we know the Voynich cipher is not the same kind of cipher, but the question that I wonder about is the extent of overlap i.e. how much or how little they have in common.
T Anderson: I think it would help if we were to have some kind of list as to where Voynich characters can be found in other texts.
JKP often says that the Voynich characters can be found amongst the latin abbreviations, though without any sources to back up this statement.
Mark: as far as I know, all the other fifteenth century enciphered documents that have been brought into the light beneath modern historians’ collective gaze have been comprehensively solved.
Nick: I think you miss my point.
I, myself, have never attempted to decipher an enciphered diplomatic letter, even one that has been already deciphered by someone else, but the decipherment of which I have of course not seen. Maybe I should try to do that as an exercise, with one of the more sophisticated enciphered letters it does not sound that easy.
I was wondering if someone highly experienced in deciphering these diplomatic ciphers would be able to spot things in the Voynich others can’t; maybe they wouldn’t spot anything more than we can.
I don’t know how much experience you have in that.
How easy is it to identify the language in the cipher? Or possibly languages as Simonetta describes.
Mark,
I am afraid that it is rather a case where you missed Nick’s point than the other way round. I am writing this without any intention of animosity, but from everything you write it is clear that your understanding of the Voynich MS text properties is insufficient to judge a number of things.
When you write that “we don’t know how much the Voynich cipher and the Italian diplomatic ciphers have in common” (or something similar) it is not entirely clear who is “we”. It is certainly not true that nobody knows, and in fact the three people who have been discussing with you here (Nick, JKP and myself) have a very clear vision of this.
While we (all three of us) may not know how the Voynich text was generated, we have a very clear understanding of how it was NOT generated.
Experience in decrypting Italian diplomatic letters will help absolutely nothing in solving the Voynich MS text. The only problem with these letters is that they present relatively short texts.
The ciphers in Tranchedino, when applied to an Italian (or Latin) plain text will do three things:
1) They increase the alphabet size from about 25 to at least 60
2) They flatten to some extent the single character distribution
3) They flatten (to some extent) the Zipf curve
None of the three are seen in the Voynich MS text. There is no commonality.
To make it worse, the (very) unusual things that are seen in the Voynich MS text are completely incompatible with such ciphers.
People who are intimately familiar with all historical ciphers have been unable to crack the Voynich MS code (f there is one), and in fact they have clearly stated that the Voynich MS text is something different altogether.
Rene: Don’t worry I take no offense.
It is true that I have spent almost no time studying the Voynich text properties.
Can you tell me if the Voynich contains any null characters? Can you tell me if there are any symbols which correspond to the same character in the Voynich?
It short you can’t tell me what if anything the Voynich has in common with diplomatic ciphers as you have no idea how the Voynich cipher works. What you tell me is that in many respects it appears quite different.
So I think I can repeat with confidence that:
We don’t know how much the Voynich cipher and the Italian diplomatic ciphers have in common.
All you know with certainty is that they are not the same, but the fundamental degree of similarity or degree of difference in the core cipher you don’t and can’t know.
You have no idea what the less common characters in the Voynich do or correspond with.
You are comparing the statistical output you see from the Voynich with the statistical output you see from diplomatic ciphers, just because they appear very different that doesn’t mean that there isn’t significant similarity internally.
Rene: As an example you might look at a banana tree and a human being and say that the way they look and function is completely different and that there is really no similarity. However we know that there is an absolutely huge similarity between a banana tree and a human being at the cellular and genetic level. There is a big difference from comparing how things appear externally with how they function internally.
I think all we can say with certainty is that the Voynich and diplomatic ciphers are not the same, but the waters become much more murky when addressing the question of degree of similarity.
Hi Mark,
it is entirely possible to make statements about the Voynich MS text without knowing how to decipher it. The question about nulls and homophones that you give as examples, concern a level of detail much below the actual issues with the Voynich MS text.
Let me give an example. Suppose someone came to you arguing that he could translate Russian to English just by substituting the Russian characters to latin ones, and the result was english.
You would say that this is wrong, and it cannot work.
He could then argue that, since you don’t know Russian, you cannot judge that.
This is quite a similar discussion.
Mark,
from all I can see in your many contributions it becomes clear to me that you obviously have no understanding of the paleography of the Middle Ages.
Every text, whether enciphered or not, using the latin alphabet in the frame of its contemporary writing system shows lots of similarities to others texts of the same time period. Italian diplomatic ciphers written for example in a Gothic Cursive or an early Bastarda differ from non enciphered texts only by the use of single letters and common abbreviations as a substitution. The writing system remains the same for all of them with the exception of a few individual letters/glyphs representing the personal style of the scribe. As far as I remember, JKP has explained these paleographical basics to you several times.
There is absolutely no need to look for internal similarities or making lists of supposed similar glyphs. From a paleographical point of view such an effort will be senseless because as a result it would only show the well known obvious.
Hi Rene
I agree it is possible to make SOME statements about the Voynich text without being able to decipher it; I have never said otherwise. I have said however that there are other statements that it is not currently possible for you to make.
When you say: “The question about nulls and homophones that you give as examples, concern a level of detail much below the actual issues with the Voynich MS text.”
I completely disagree, they are vital examples of questions pertaining to the degree of similarity with diplomatic ciphers. As I have said, there are certainly very significant actual issues outside of that, but that doesn’t negate the question of the degree of commonality.
Yes, I like your Russian example as I know some Russian and in many cases by substituting Cyrillic characters for Latin characters you can render a word which looks meaningless into something very comprehensible. Now obviously that does not work in very many other cases, but what I am arguing in the context of diplomatic ciphers is not a general point, but a question of degree of commonality.
Maybe we are talking at cross purposes and you don’t understand my point.
Hi Charlotte
You say: “There is absolutely no need to look for internal similarities or making lists of supposed similar glyphs. From a paleographical point of view such an effort will be senseless because as a result it would only show the well known obvious.”
I completely disagree. I wish someone who claims all the Voynich characters can be found elsewhere would produce examples in each case, I haven’t seen you, JKP, or anyone else do this. In fact my research indicates that very different character sets can be found being used dependent on the time period and region amongst diplomatic ciphers. So tell me if this is well-known obvious why phenomena like this are the case. You can’t make blanket statements without providing the empirical basis for them.
Charlotte,
I will give you an example:
Can you tell me a document in which you can find the character that Nick says is normally referred to as cPh?
If you can’t then I think my point stands.
Mark,
all would be fine if you wrote: “I don’t know” instead of “we don’t know”….
Let me just leave it at that.
And we don’t know who won the big prize at the Oxford sushi joint either; or at least, I don’t bloody well know. Does anyone in the know know.
Sorry Mark,
why should I provide you with “examples in each case” and why on earth should I waste my time with such a fruitless work?
The empirical basis is self-evident and lies in professional paleography. If you don’t accept statements from people with a sound background in that field, nobody amongst them could help you further.
You seem to be absolutely convinced that you’ll find the key to the Voynich in an Italian diplomatic cipher. If it were that simple, the mystery would have long been solved. There are many professionals who know exactly that this is not the way to the solution, but almost none of them would waste his/her time to proof it on a blog for an amateurish public.
I’m afraid, but I don’t feel the need for providing you with documents, examples or whatsoever.
Rene,
I agree I don’t know, but I contend that you don’t know either and as far as I have heard you haven’t presented a persuasive argument as to how you do know.
I am sorry, I am not trying to be contrarian and I certainly am ignorant of many aspects of the Voynich that are not relevant to my current line of research.
However there are people, who it seems to me parade expertise, they don’t really have. As I have said previously, it is my experience, that there is nobody who we can say with certainty is an expert on the Voynich.
I think Nick, for example, has said many very interesting and I think highly intelligent things indeed, but there is still a chance that most of his theories about the Voynich are wrong, so given that I am reluctant to say even he is an expert.
I think when more is actually understood about the Voynich we can start to talk about experts.
I admit that I do find appeals to authority on the Voynich frustrating. It seems to be used as an alternative to presenting an actual argument. I am happy to listen to argument, but someone just telling me that they know better does not persuade me.
A list of questions which would indicate the degree of similarity to diplomatic ciphers that WE don’t know the answer to:
Does the Voynich contains any null characters?
Are there any symbols which all correspond to the identical character in the Voynich?
Are there any symbols which correspond to whole words, maybe rare characters?
Do any characters, if not all characters, map to a given letter of the alphabet?
Do any characters correspond to a couple of letters say for example “ap”? (Much less likely, but possible)
Are multiple languages used together as Cicco Simonetta suggests?
I am sure I can think of more if necessary.
Now we might be able to speculate on the likelihood of these question being true or not, but WE don’t know and therefore WE don’t know the degree of similarity or difference between the Voynich cipher and diplomatic ciphers.
Charlotte:
I repeat, you say: “There is absolutely no need of making lists of supposed similar glyphs.”
I argue that the greater degree of shared symbols with the Voynich the more likely an association, so if you claim as JKP does that you can find more similarity elsewhere the only way you can do this is by finding examples in each case from given texts.
You may think that fruitless, I don’t. I believe if there is a text using all the same characters as the Voynich that is relevant.
I studied Maths at Cambridge University and studied cryptography as part of that. Does that give me the justification to say to you, Charlotte, that I have much more of a background in cryptography than you, so your opinions are of no value?
Appeals to authority don’t count for much to me. If you make a statement you need to justify it or at least provide a link or a specific reference that can be found online that justifies that statement.
You seem keen to tout your professional expertise, let me know where I can find your CV which demonstrates your expertise on the Voynich.
I am sorry, but I find self-proclaimed experts annoying. I am sure you are a very likable person. But if you regard me as an amateur, but you as an expert then I have my doubts.
Charlotte,
I have googled “Charlotte Auer” and I don’t seem to able to find a reference to anyone except a Psychologist, which either means that you not very well known in your profession or you have a different name.
Look, I don’t want this to get nasty, but it does get my goat when someone says they are a professional and my opinions are of no value and then refuses to provide evidence to justify their arguments.
Mark: I think you are using “expert” in a shallow and fairly unhelpful way. The late Stephen Bax was accustomed to deriding just about everyone who didn’t agree with his particular view of the Voynich, and one of his preferred ways was to attack their supposed expertise, on the (somewhat specious) grounds that if all you self-proclaimed experts cannot read a word, how come you think you are experts? But I thought this was utter bilge (and indeed told him so): there were plenty of people who were experts in the nature and structure of hieroglyphics before Champollion did his thing, even if the things they knew had not yet come together into a unified and organized body of knowledge. And so it is with Voynichese too, I believe: we just haven’t met our Champollion yet.
Mark. Expert on MS 408. It’s me. 🙂
I will give an example. ( Perhaps you can understand that ).
MS 408 folio 1v. ( Plant leaf ). The text is written :
The green leaves are 14.
The golden leaves are six and six.
And that’s my birthday.
__________________________________________
There are several characters on one sheet of the plant.
Those letters are = J.T. ( J = 1. T = 4 ).
Jewish substitution : 1 = A,I,J,Q,Y. 4 = D,M,T. 🙂
Eliška ( Elizabeth of Rosenberg ) writes on this page when she was born.
1466.
Eliška also writes : Czech words.
Nick:
Yes, I recognise the value of a term can indicate that someone has spent time looking at a topic and is familiar with the history of debate and knows pretty certain facts about it.
My purpose is not diminishing or deriding you or anyone else. However I feel at times there are people who use their claim of expertise as an alternative to argument, I am not talking about you in this regard as I cannot think of an example of you doing that as you always seem happy to justify your arguments even if I don’t agree with them. However some people provide arguments about the Voynich that I find pretty unpersuasive and then fall back on the statement that they are a professional or expert. So I am very wary of people using the term for that reason. However I know enough about Voynich research to know that there is a lot that is not known and much less that is known, so when someone claims to know a lot about something I am careful before I accept that.
I don’t fundamentally object to use of the term “expert”, but when it comes to the Voynich currently the idea should be used very gingerly I think.
The internet tells me that the definition of expert is:
“A person who is very knowledgeable about or skilful in a particular area.”
It is hard I think honestly to say that anyone is very knowledgeable about most aspects the Voynich.
I think you are an interesting case in point. Many of your ideas I think are very intelligent like the block-paradigm and if that counts as expertise that is all well and good. Without wanting to flatter you, I have found your input to my research by far and away the most helpful not only because of specific comments, but also actually a lot of your old blog posts. I feel some people just comment to say that I am wrong without any detailed justification.
I am always very happy to acknowledge and thank someone who tells me something about the Voynich that I do not know and is relevant to my line of enquiry.
If I am supposed to accept Rene or Charlotte or JKP or Diane or the man with the ants are right about something, because they say they are right about something without an argument I find satisfying then I am not going to blindly accept, this is not how operate.
I do fear that we are not necessarily quite so far forward as the experts on hieroglyphics before Champollion.
Anyway I don’t want this to get into a confrontation. If someone disagrees with me, but does not have the time or willingness to justify their disagreement then I suggest they do not bother telling me that they disagree with me. I can find that there are too many unconstructive fights and far too little effort to advance different avenues of research. I personally do not get on this blog to get into fights with anyone. I value people’s opinion, but am not obliged to agree with that opinion.
Sorry for being belligerent on this point, but I feel quite strongly that I am not prepared to accept someone commenting to effectively tell me to shut up as I am an amateur and they an expert without strong justification! (I am not talking about you.)
Mark Knowles wrote: “A list of questions which would indicate the degree of similarity to diplomatic ciphers that WE don’t know the answer to:
“Does the Voynich contains any null characters?”
I’ve been looking into this for 10 years. When you decrypt something, null patterns are one of the first things you look for.
The VMS might have null characters, but it’s difficult to prove (and to determine which ones they might be) and they would have to be much less frequent than they are in diplomatic character sets because the VMS glyph-set is too small to incorporate more than a few. I have some ideas about what MIGHT be nulls but even though it’s an obvious question, it’s not a trivial one to solve. I am still studying this.
—–
“Are there any symbols which all correspond to the identical character in the Voynich?”
Yes. As I’ve said numerous times, they can pretty much all be found in medieval Latin (mostly) and Greek character sets, including the VMS rare characters. Very few of the shapes are unusual, if you take into consideration that medieval alphabets included a wide array of abbreviations in addition to the alphabet, it’s the way they are arranged and combined that is unusual.
And… sometimes the way they are arranged is not unusual either…
I’ve mentioned all this before, but I’ll say it again…
• In languages using Latin characters, EVA-m is most often found at the ends of words and the ends of paragraphs. This pattern holds true in the VMS.
• In Latin languages EVA-y is usually at the ends of words, sometimes at the beginning, and occasionally in the middle. It’s the same in the VMS.
• In Latin, EVA-y is sometimes curved (-cis), sometimes straight (-ris). It’s the same in the VMS.
This is why I sometimes wonder if the VMS was deliberately crafted to resemble Latin (at least superficially) while, in fact, being something else.
—–
“Are there any symbols which correspond to whole words, maybe rare characters?”
Yes, Latin includes many one- and two-character abbreviations that correspond to whole words (hundreds of them) and some of the shapes in the VMS, like EVA-r, EVA-y, and EVA-s are similar shapes and which sometimes stand alone.
The “rare” characters in the VMS are not rare glyphs, they are found in Latin and Greek, they are simply rare in the VMS (not used frequently). I’ve already mentioned that the really weird one that looks like a little line with a dot over it and a roof attached to the side of it is a Greek abbreviation. It may look very strange to modern eyes, but there’s nothing strange about it in classical texts.
—–
“Do any characters, if not all characters, map to a given letter of the alphabet?”
What kind of question is this? Are you asking about the shapes or the meaning? As I’ve said, the VMS glyph-shapes correspond to shapes in Latin and Greek (letters of the alphabet + scribal abbreviations), whether you want to believe me or not. That does not mean they map to the same MEANING of those shapes in Latin and Greek alphabets. If they did, we would have figured it out.
—–
“Do any characters correspond to a couple of letters say for example “ap”? (Much less likely, but possible).”
Yes, as I’ve said repeatedly, some of the shapes are the same as those used as scribal abbreviations. In Latin, if you take an “r” and combine it with the abbreviation for “is” (which is a loop and a descender) you get the shape that looks like EVA-m that is usually at the ends of words or lines in Latin. In Latin, it means “ris” (r + is) or “rem” (r + em) so that you can spell words like paris or harem by writing pa+EVA-m and ha + EVA-m.
In some manuscripts the -ris/-cis/-tis abbreviations are so common you’ll find them several times per paragraph. Some scribes abbreviated more than others.
Are multiple languages used together as Cicco Simonetta suggests?
I very much doubt it except for the possibility of some loanwords. Assuming the text is meaningful, one might expect to find some Latin plant names and some Arabic star names. Note how many of the “star names” (if they are star names) start with EVA-ot as do many of the zodiac-section words. Many Arabic star names start with Al-. But… the mapping never works perfectly and if it did, the VMS would have been solved. There is something else going on. It’s not a simple substitution code.
The reason why it’s unlikely that the VMS is polyglot (except for possibly a few loanwords) is because the structure of the VMS tokens is very rigid and repetitive. If it were polyglot, there would be more variation. There isn’t even enough normal variation for one language.
“Now we might be able to speculate on the likelihood of these question being true or not, but WE don’t know and therefore WE don’t know the degree of similarity or difference between the Voynich cipher and diplomatic ciphers.”
You mean you don’t know. The collective “we” assumes we haven’t looked into these questions (we have) and that we don’t have the expertise to evaluate them, which I don’t think is true. No one is a “Voynich expert”, it requires knowledge of too many disciplines, but some of us do have expertise in certain specific areas that enables us to evaluate some of these questions.
Mark:
You say:
“It is true that I have spent almost no time studying the Voynich text properties.”
…and I think that is the problem. Because you simply can’t replicate them by applying any known fifteenth century cipher of the type you describe to a conventional plaintext. You just can’t. Besides, code breakers of the ability of Tiltman and Friedman would have cracked such a code in hours.
I don’t expect you to take my word for it, but some time looking at the Analysis of the Text section of Rene’s website will explain more clearly than I can.
I was asking about a listing of when and where the characters in the VM are found because it would help contextualise the VM for those who aren’t among the handful of people in the entire world who are experts on the paleography of 15th century europe.
It seems the author of the VM made a creative leap when writing it since we don’t have any precedent. We’re more likely to figure the VM out if we understand the ideas the author was exposed to as fully as we can.
Mark,
when researching the Voynich MS with a certain theory in mind, the hope is to find evidence that supports the theory, but it is crucial how one handles evidence that contradicts the theory. In serious research, you would look at this closely and reconsider some of your own assumptions.
In bad research, there are several options:
1) Denial
The evidence should be ignored. It can be swept under the carpet.
(I gave you three good, straightforward reasons -numbered 1-3 – why the Voynich MS text is not the result of application of an Italian diplomatic cipher to an Italian or Latin plain text. You first said that this does not count, and now say that no evidence has been presented).
2) Discrediting the source of the information
(People discussing with you are described as self-proclaimed experts and there is no reason to believe them).
There are many people interested in the Voynich MS who have some idee-fixe about it and will say anything in order not to have to budge an inch from their standpoint
Do you want to be such a person?
Having people who wish to discuss with you is quite a luxury, you know.
In my mischievous mind’s eye, I’m forever conjuring up pictures of Lily, with her dull old shears, cutting up small pages from old Delhi velum end runs. Such ritual being preporitory to working her magic with quill and brush to replicate the little blonde nymphs in her own imagined full likeness. Going through such age old rituals and rites designed to enhance the fertility that she had been unfairly denied, though always longed for. Of course she, the logical creator remains the one and only Voynich expert and, self assuredly remained so during the course of her long life; perhaps not completely satisfied that others of equal intellect, could not see through her well meaning jest. Readers take caution and don’t be too disappointed when all is finally revealed; we wouldn’t, wish to see any more self appointed Voynich experts, having to fall on the pointy portion of their collective assegais.
JKP:
You have not understood most of my questions, so I will clarify:
“Are there any symbols which all correspond to the identical character in the Voynich?”
Are there characters in the Voynich which look different and yet function in the same way such that they are interchangable?
“Are there any symbols which correspond to whole words, maybe rare characters?”
Can you say that certain characters in the Voynich correspond to whole words? My question was not whether this is the case in Latin abbreviations.
“Do any characters, if not all characters, map to a given letter of the alphabet?”
Again you do not understand my question. Are there any characters in the Voynich which can be mapped to a given letter of the alphabet? I am not talking about whether they look similar or not.
“Do any characters correspond to a couple of letters say for example “ap”? (Much less likely, but possible).”
Again you don’t understand my question. I am asking if there are any characters in the Voynich that map to a specific pair of letters in the Voynich?
As you say the Voynich characters all correspond to Latin scribal abbreviations, please tell where I can see one document described as “cPh”?
*Again you don’t understand my question. I am asking if there are any characters in the Voynich that map to a specific pair of letters in the alphabet?
As you say the Voynich characters all correspond to Latin scribal abbreviations, please tell where I can see one document containing the character described as “cPh”?
SirHubert:
I am not saying that you can replicate the Voynich text properties by applying any known fifteenth century cipher of the type I describe to a conventional plaintext.
I have looked at the Analysis of the Text section of Rene’s website.
The question I am addressing is that of commonaility, shared features. It is perfectly possible for there to be things in common without them being the same.
The questions I have listed are questions that would indicates the degree of commonality. I have looked at properties of the diplomatic ciphers and asked if any of those can be found in the Voynich.
Mark: when I trawled through the Tranchedino cipher ledger more than a decade ago, the closest visual match I found to the strike-through EVA ch pair was the 1455 Ludovico Petronio Senen strike-through “Subscriptio” (Curse, p.177). There’s much more to be written on this specific subject, but my spare time to write it up is currently so close to zero as to make no difference. 🙁
Nick: Thanks for that.
We should soon have some more cipher keys to add to those listed by Lydia Cerioni. I have already been emailed scans of a cipher key that I requested and made a few contacts at the Milan State Archives. I will try to get as many scans of relevant cipher keys or enciphered letters, though as I have said my focus is more on the earlier ones rather than the later ones. If there are limits to the scans that I can order remotely then I will have to go in person. I will make enquires about the Atti Ducali section that you have referred to as well as the Visconti correspondence. Similarly I will get what I can request scanned remotely from other relevant archives, e.g. there is some Visconti diplomatic correspondence kept in the Pavia archive. I haven’t yet investigated whether other archives handle requests in the same way.
Mark: When I think of that so often maligned ‘F’ word, my thoughts go back in time to Woodstock and good old Country Joe and his Fishs’ unsavory interpretation of the letter, for good personal, historical reasons. So your query relating to a single Romany that might likely correspond to a similar Voynich character, it would have to be the one I would choose. In many respects the lesser form of the uppercase gallows character could only be an ‘F’ realistically. So, what do you know, good old Ethel Voynich was known to use a fair representation when signing off, in her letters back home from her travels abroad, prior to Wilfred’s passing in 1930. NB: Of course after that sad event, she had no reason to continue her own VM research agender on the continent, being quite satisfied to wile the while in the big Apple, knowing that her little secret was safe with Anne.
Realisticly speaking, me thinks my word select format sucks.
Mark wrote: “You have not understood most of my questions, so I will clarify:”
Thank you. Perhaps you could have been clearer to begin with. 🙂 Mark, I didn’t misunderstand them as much as you might think. All those questions have been investigated and are still being investigated and even if people have opinions about them, their points of view have not yet been proven.
—–
“Are there any symbols which all correspond to the identical character in the Voynich?”
Are there characters in the Voynich which look different and yet function in the same way such that they are interchangable?
Quite a few people besides myself have been trying to determine this for years. Look at some of Emma Smith’s blogs and some of the computational attack threads and blogs. Researchers have been investigating precisely this question.
—–
Are there any symbols which correspond to whole words, maybe rare characters?”
Can you say that certain characters in the Voynich correspond to whole words? My question was not whether this is the case in Latin abbreviations.
Latin abbreviations ARE relevant to answering this question!
A number of the glyphs in the VMS that can (and do) stand alone correspond in shape to Latin glyphs that can and do stand alone. If it’s a coincidence, it’s a pretty big one. Which means the VMS glyphs MIGHT correspond to whole words. However, no one has definitively proved it.
—–
“Do any characters, if not all characters, map to a given letter of the alphabet?”
Again you do not understand my question. Are there any characters in the Voynich which can be mapped to a given letter of the alphabet? I am not talking about whether they look similar or not.
Mark, if Voynich researchers and cryptographers knew the answer to this, we would have solved it by now, but the answer, if you mean one-to-one correspondence is, in my opinion, a resounding NO. It’s not a straight substitution code.
The statistical properties of the text do not support the mapping of individual VMS glyphs to letters of any alphabet and expert code-breakers can solve that kind of cipher in minutes. Even if it’s a one-to-many plus many-to-one cipher (like the diplomatic ciphers), code-breakers can usually figure it out in a reasonable length of time.
—–
“Do any characters correspond to a couple of letters say for example “ap”? (Much less likely, but possible).”
Again you don’t understand my question. I am asking if there are any characters in the Voynich that map to a specific pair of letters in the Voynich?”
I did understand your question. I understood the other ones too. If we knew answers to these kinds of questions, we would probably have solved it by now.
The reason I framed my answers in the context of Latin characters, and Latin ligatures and pairs, was to illustrate that combining (as in pairs) was common practice in the Middle Ages and thus it’s quite possible that the VMS functions this way, but no one has yet proved that it actually does.
All your questions are cryptographically basic, but when you have a 5,000-piece jigsaw puzzle with all the pieces turned upside down (imagine they are all the same shape so that you can’t use shape to solve it) and you only have three pieces turned right-side up, it’s pretty difficult to know if you have positioned them correctly. You have to gradually create enough internal context for the process to be self-confirming.
—–
“As you say the Voynich characters all correspond to Latin scribal abbreviations, please tell where I can see one document described as “cPh”?”
I never said “all”. Usually I say “almost all” or “virtually all”. I’ve repeatedly said that a small percentage is Greek.
In Greek, Pi-Rho can be written as EVA-ePh with pi crossing the stem of Rho (which looks like a “p”). I have posted examples. If it’s cursive writing, it looks even more like a benched P.
I’m doubtful that benching plays the same role in Voynichese as it does in natural languages. In Latin, benching is similar to a long macron and stands for missing letters.
The VMS seems to have combined flexible scribal utilities into four very specific shapes. To put it another way, the VMS glued scribal Lego blocks together so you can’t rearrange them.
JKP: The fact that as you say many or most of them haven’t been answered proves my point. My point is the more of those questions and other questions, that I could raise, that can be answered “Yes” the more in common the Voynich cipher has with the diplomatic ciphers. Latin abbreviations are a whole different discussion.
As far as there being common symbols between Latin abbreviations and the Voynich, I have looked at your website and seen few examples, now it may be that I haven’t spotted all your articles that refer to them, but without real tangible examples from documents preferably containing as many as possible different Voynich characters in one document then I remain to be convinced. Unfortunately, you telling me it is so isn’t sufficient, I have to see it with my own eyes. I have looked at Capelli and I haven’t observed the level of commonality that you claim. You previously had said that you hadn’t done this, because you hadn’t had the time and they were your secrets; I can respect that, but without evidence it is hard to persuade me.
This is my point if someone makes a statement, generally, I am not necessarily inclined to believe it unless I can verify it. Obviously there are things I must take on trust like the carbon dating. But if someone says to me “I have been studying the Voynich for 50 years” that is just not enough for me to accept that as true.
Mark:
If you accept that Voynichese cannot be the product of a fifteenth century Italian diplomatic cipher system, that raises a fairly obvious question. I’ll leave matters there and wish you the very best of luck in your research.
Rene,
You say:
“when researching the Voynich MS with a certain theory in mind, the hope is to find evidence that supports the theory, but it is crucial how one handles evidence that contradicts the theory. In serious research, you would look at this closely and reconsider some of your own assumptions.
In bad research, there are several options:
1) Denial
The evidence should be ignored. It can be swept under the carpet.
2) Discrediting the source of the information
(People discussing with you are described as self-proclaimed experts and there is no reason to believe them).”
I completely agree.
However I don’t these apply
1) You to have repeatedly misunderstood what I said and I thought that I made it very clear. I did NOT say the Voynich MS text is the result of application of an Italian diplomatic cipher to an Italian or Latin plain text. I did NOT say that that does not count. I repeatedly said that I was referring to common, shared features. I have provided a list of questions that would indicate the degree of commonality between the Voynich and diplomatic ciphers. It seems to me that NO evidence has been provided that will answer most of those questions. You seemed to imply you thought those commonalities an irrelevancy, I don’t. The 3 textual properties don’t contradict what I am saying.
2) On the contrary, it seems to me that there has been an attempt to discredit my opinion by implying that I am an amateur and not an expert, whilst other people are experts. If someone says they know something and I find their arguments inadequate then them appealing to their authority on the subject does not persuade me. There are facts I must take on trust like the carbon dating. But if possible with other statements I like to be able to verify in the context of the Voynich what someone are telling me.
Obviously, I believe having an open mind is important in Voynich research as in life. It makes no sense to stick to an opinion just through stubbornness, ultimately that only harms oneself. Truth has to be better than comfortable delusions, though in life will all operate under some delusion though may not like to think it.
I try my best not to be stubborn and be prepared to change my mind. Changing one’s mind when it is the right thing to do can be revelatory and can push one forward in life rather than hold one back.
I have changed my mind significantly in the past on the Voynich. When I first looked at the Voynich I was trying to determine where the 9 rosette foldout might be a map of. Nick said that the “city” in the top right rosette represented Milan. I was not persuaded as the drawing of the “city” was surrounding by what I interpreted and still interpret as water. Now as Milan is inland that did not seem to make sense, so I considered many other locations. However I became frustrated as I could not find anywhere that I thought a good match. Now I had come to the conclusion early on from what Nick had said that he was a clever guy. But Milan didn’t make sense, so I thought I would try to debunk that. However the more I looked at it the more I thought Nick was right and that I could tackle my previous objection in a way I felt reasonable. As someone with no acquaintance with medieval cipher alphabets I had no idea of the weird and wonderful symbols that they use, so like most people who first vome to the Voynich I assumed was a language in an unknown script. However looking at Nick research and becoming aware of his book and learning about the statistical properties I became persuaded that it was indeed a cipher.
I have gradually been warming more and more to Nick’s block-paradigm idea, though I was never opposed to it. I think the viability of it depends on how inventive the author was, if the author was as inventive with his diagrams as he was with his cipher then then we will probably be out of luck.
Having said this it would be folly to change my mind when I don’t think it justified as in this case.
Maybe one day I will say Rene was right and I was wrong, but at the moment I don’t believe that is correct.
I would also advise you to reflect on whether you might change your mind about your ideas; it is a two way street.
I agree that you have things to do in life and giving your time to the Voynich community to support people doing research is a noble goal.
I was raising a subject and was not trying to generate a confrontation; that is actually something that I am not fond of and generally I try to avoid confrontations with people in life.
I am inclined witb hindsight that this discussion has taken up your time and my time both of which could have been used more productively on other areas of Voynich research.
I am sorry if this has wasted your time, but I did not try to draw you in to the discussion. Time for all of us is a luxury.
All the Best!
There was a lot of discussion again.
Basically, I have to agree with JKP, especially the signs of the endings which are identical. They never come in the word either. Whether they have the same meaning as in Latin, let’s just leave. In German it could also be (-heit, -keit, -ung, -nis, -tum).
If one compares VM with the Latin (Schriftstill), are signs standing alone, one can assign them to words.
Why is there a self-sound after a pseudo? Well hidden, but certainly is c = e. This also applies to the various manuscripts.
With all the possibilities I have already worked through, I see the problem in the language, and not in the encryption anymore.
Based on my experiments, it must be a rather rare form of Latin, where I certainly find no textbook. Maybe this species is no longer available today.
One of 200 Alpine Latin spoke.
If I look at the history of the region and the possible development of language, multilingualism does not even have to be an intention.
SirHubert:
I don’t accept that Voynichese cannot be the product of a fifteenth century Italian diplomatic cipher system. I think you don’t quite understand what I am saying. I am saying that I accept that Voynichese cannot SOLELY be the product of a fifteenth century Italian diplomatic cipher system.
So the reason why I think Italian diplomatic ciphers are relevant is to do with the questions of commonality. Certainly the question of commonality in the characters used, but potentially in the cipher.
I think people are thinking in very binary terms:
1) It is different
or
2) It is the same
There are other possibilities, such as the might be differences and similarities.
Anyway I greatly appreciate your encouragement with my reseatch. Thanks!
SirHubert:
Just to make it even clearer:
Not being able to replicate the Voynich text properties by applying any known fifteenth century cipher of the type I describe to a conventional plaintext.
Doesn’t imply:
That Voynichese cannot be the product of a fifteenth century Italian diplomatic cipher system.
There is subtle difference I think.
You are a product of both of your parents. That is not the same as saying you are a clone of one of your parents.
Mark Knowles wrote: “As far as there being common symbols between Latin abbreviations and the Voynich, I have looked at your website and seen few examples, now it may be that I haven’t spotted all your articles that refer to them, but without real tangible examples from documents preferably containing as many as possible different Voynich characters in one document then I remain to be convinced.”
I don’t need to convince anyone. I’m conversing with you as a courtesy. People who read medieval texts are familiar with shapes that match EVA-y, EVA-m, EVA-d, EVA-k, EVA-s, EVA-r, EVA-l, EVA-n, EVA-ch, EVA-sh (and of course EVA-a, o, c, and i, and the tails on the ends of chars, which are meaningful in Latin).
Approximately 98% of the VMS is written with sixteen shapes that are commonly found in medieval Latin (more, if you include Greek letters, like delta, and benched ligatures).
Think about that—only sixteen shapes are necessary to produce an almost-complete page of Voynich text.
Dear all,
I think there are two Voynich glyphs that can confidently be mapped to two Latin Roman letters. This is an argument I have made several years ago but I reiterate.
1) Many venerable languages (Coptic, Ethiopic, Hebrew, Syric. Aramaic and OF Course, (Armenian, the subject of this post} used letters for numbers. Armenian used “e” and “t” for “8” and “9” respectively. So working backward, we can confidently map an “e” when we encounter a Voynichese “8” and a “t” when a “9” is seen.
2) So the frequently encountered glyph pair “89” is the Latin abbreviation “et” as all know is an abbreviation or contraction for the English equivalent “and”.
3) As icing on the cake, because Armenian uses a phonetic alphabet (currently 39 glyphs), the International Phonetic Alphabet uses an inverted “e” for “8”. This is pronounced as the “e” in the English word “quiet”. The IPA equivalent for “t” is “t’ ” a phoneme that has brings the “t” behind the teeth during pronunciation.
I am confident that we DO KNOW what the Voynichese “8” and “9” map to; simply “e’ and “t”.
Now the question naturally arises, what is even a small proportion of Armenian doing in the Voynich manuscript? It WAS one of the major languages of commerce in the middle ages and parts of the VMS may have included widely accepted Armenian abbreviations used in commercial transactions or just used, for example, in plant descriptions of herbs and spices. I do not mean to imply that Armenian was any major component of the VMS, but only to indicate that there are two glyphs that can be deciphered with confidence as originating from one of the oldest languages in the world, Armenian. Cheers,
JKP:
You say; “I’m conversing with you as a courtesy.”
I would say that the same applies with me.
I think I ought to say that if I feel that if a line of discussion with a specific individual at a specific time is not going to be productive for either of us I will try in the future to make that clear politely.
I have a limited amount of time and I fear in this instance I have lost a lot of time getting embrolled in a discussion going nowhere. I have expressed my thoughts and others are welcome to have their own opinion.
I think most people researching the Voynich are intelligent people, as I believe I am, so I don’t wish to cause offense. I try to respect others’ opinions and I hope others can do the same of mine.
I does feel like at time there can be more of a focus on pulling down others’ theories than building up one’s own.
I value this blog, the only Voynich blog that I post comments on. I comment on it to get others’ opinions, but often as I believe it valuable to leave a record of one’s thoughts which may or may not be of interest to others now or in the future, in the same way as the record of historic email correspondence is of value.
Time will eventually resolve all disagreements.
@ Thomas F. Spande
That’s all well explained, but how far will it stand if you apply it to the rest of the book? It also has to work if I separate the e from the t and use it individually.
Most Latin written books use a single character for et.
et. mirrored (S). Funnily enough I find the symbol also exactly at the places in the VM where one could suspect it in a text.
I’m pretty sure, 89 is the meaning of tum. Divided into the single character (t) and (um).
(um) is also the abbreviation for (umum = one) is used for singular. If it’s up front, it means (one). Is it all alone so it’s just one.
Only when I try the game on all sides and the thread does not break, I can assume that I’m on the right track.
Here’s an example.
This is a characteristic of a very specific kind. He writes a word, a word later he writes it split, and yet it makes sense.
https://www.facebook.com/photo.php?fbid=2036558949900082&set=gm.1577951085648020&type=3&theater&ifg=1
If I were to split all the individual symbols and break them down into individual parts in order to perform a character count, then c would be the most widely represented. Thus, c = e.
This in turn fits precisely into the word formation.
I would never write anything if I had not played all the possibilities. Based on the results, it is the most deceptive.
Dear All,
I should have sourced my “Armenian connection”. Armenian is a alphabet-based language, not one based on syllables. If the alphabet be laid out starting with the Armenian equivalent of “a”, that resembles a “u” with a tail, One can see in “Armenian numerals (Wiki) that “Arabic” “7” is a long “e”; “8” is as mentioned above, an upside down “e” pronounced as “e” in the English word “quiet” and “9” is a “t'”, as pronounced in the International Pronunciation Alphabet. The IndoEuropean-Arabic numbers for Armenian are in order of the first ten glyphs: Upper Case A,B,G,D,E,Z. lower case long e, the upside down e. t’ and another upper case Zed. I meant to stress in my above post of 5-19 that these are Latinized or Romanized equivalencies, NOT the glyphs or characters that Armenians themselves used in counting.
Just for grins, I plan a word frequency comparison of “and” and “89” (“et”). Complicating this will be the occasional occurrence of the ampersand “&” in Voynichese. Here Voynichese gets tricky in that the Armenian cursive “f” resembles the ampersand.
Just to reiterate, I hope for the last time: the mapping of Voynichese “8” and “9” to “e” and “t” is NOT encryption, It is a straightforward substitution dictated by Armenian numerals and amounts to shorthand.
I think this substitution complements the macron techniques (macronization; does such a word exist in English?) that Nick (chiefly) has expounded upon. Cheers,
Thomas, two questions apropos your theory…
1. In most languages, the letter-pair “et” occurs in any position in a word. In Latin, it is used not only as “et” (and) but as a syllable, a basic building block of the language. In contrast, in the VMS, “et” occurs with unusual frequency at the ends of words, which is not the usual pattern in Latin, Armenian, or any other language that I know. The VMS doesn’t have enough glyphs to assign multiple glyphs per letter, so one can’t easily say that maybe “e” and “t” are represented differently when they are positioned within words.
2. If you reverse the two letters, so it is “te” instead of “et” (thus 98 instead of 89), there should be many instances of 98 throughout the VMS text, since “te” is a common pair in many languages (one which occurs in numerous positions in words), but 98 is infrequent in the VMS and tends to occur only at beginnings or as follows, at the ends:
989 or 8989
which means if you map the characters as you suggested, the VMS would have to be a language that uses the “te” pair almost entirely at the end of words, occasionally at the beginnings, and uncommonly in the middle. This does not fit the profile of Latin, most Romance languages, English, or most other languages.
In fact, French, Dutch, Persian, Greek, et al, “te” is a free-standing word and occurs frequently by itself and yet I can find only one instance of free-standing 98 in the VMS (at the end of a line on f36r).
.
If you propose that 8 = e and 9 = t by substitution, then you have to explain why these letters do not occur in normal ways within words, and why the reverse (“te”) does not occur in the VMS as is common to a wide variety of languages (both within words or by itself).
If I look at the symbol & in the VM and take it apart, I get again a C, a loop and attached). The loop also occurs as a single letter in the VM.
I take again the mirrored S and take it apart so I have again C and attached one). C = E, and ) = T. Combination ET.
Example &.
C = E, and the loop probably S, and again (= T, so we have EST.
This can stand alone and be used as an extension. But is rarely in the middle of the word.
The VM certainly consists of a combination system if the classical encryption were more of a chessboard system. C5 = A, D2 = V etc.
This goes from 5×5 fields = 25 (1 alphabet) to XXL x XXL.
What Peter is describing is how it works in Latin.
In Latin, you take the letter r, i, c, or e and you add a tail (so it resembles EVA-r or EVA-s) and it extends that letter it into a syllable or a word.
I’ve already blogged about i or r with a tail. Since the concept is the same with the letters c or e, I didn’t feel it necessary to go into detail about those, but I’ll give a couple of quick examples here…
In Latin, if you write the letter e or c and add a tail (to create EVA-s), then e-with-tail usually becomes ei, et, or er (depending on context). Similarly, if you write c and add a tail (this also looks like EVA-s), you create ce, ci, cer, or cir (and sometimes others, depending on context and language). Since many languages use Latin scribal conventions, the same shape can stand for whatever is common to a specific language.
.
In Latin, the tail is, in a sense, an abbreviation symbol, like an apostrophe. It is, however, much more flexible than a modern apostrophe in that it can stand for quite a number of different missing letters even when added to the same letter.
@ J.K. Petersen
Thank you, wonderful explained.
To JKP and others,
You are correct in that “et” can occur within words (like, e.g. “set) and I think this does occur in the VMS. I think “8989” can represent “etc.” as it is still used in some modern languages (e.g. Hungarian) as “and, and”. I also agree that “98” can represent “te”.
I think that “words” in the VMS are of arbitrary length and do not represent real words. I think this was originally Nick’s position although when I last paid attention to this point, I think Nick had revised his thinking and now believed that VMS words were as laid out in the text.
On the ampersand. Note that there seem two representations of this glyph. They occur together at the end of the first line of f3r. The first glyph has more of a rocker at the base, the second has a smooth semi-circle as our current depiction of “&”. Unfortunately the Armenian glyph for “f” resembles the ampersand closely, with the diagonal now a straight descender. Cheers,
Thomas F. Spande: the idea of getting rid of (or visually concealing in some way) the spaces between words in order to hinder code-breakers’ efforts first appeared early in the sixteenth century in Venice. Anything from the fifteenth century (such as the Voynich Manuscript) would therefore be highly unlikely to be making use of this trick.
This leads to the somewhat awkward (but unavoidable) conclusion that words in the Voynich Manuscript are almost certainly complete words, even though they appear to have far too little information (unpredictability) to be words. The suggestion that Voynich words are abbreviated or shortened in some way therefore seems to me to be one of the few more solid hypotheses going.
Nick: The “Subscriptio” is great, but I think I have found more similar than that, but not identical, which is what I pine for. In the 1447 cipher key I have been looking at there are, I think, significantly more similar characters, in fact it seems to me that the appearance of the symbols in this are particularly inventive.
One notices that in some ciphers keys there are a series of weird a wacky symbols, whilst in other numbers or sequences of related symbols are used, so you might see:
Duke of Milan b1 Or b*
The Pope b2 Or b+
Holy Roman Emperor b3 Or b#
This sequencial numerical approach, especially when applied to the letters of the alphabet, would seem to be easier to crack than the use of unusual symbols unless these symbols are not grouped together.
I imagine that over time there was a gradual move away from weird symbols to random numbers or letters, though I haven’t traced this process.
Mark,
Voynicheros have previously noted that gallows like characters were used in a document dated 1172 from the monastery of San Savino at Piacenza (see Capelli). I noted that Piacenza was under the control of Filippo Maria Visconti, Duke of Milan and from 1416 to 1470 the town was called Castel Visconti so it may be of interest to you. I note that Piacenza has an archive that is searchable through the internet.
Byron,
Thanks a lot for your reference. I have already seen that image, but I hadn’t spotted the link to Piacenza. I have found a gallows character like that in a few places in the Tranchedino cipher ledger and also other symbols reminiscent of the gallows characters in Milanese cipher symbols. So in the context of my hypothesis I think that is less likely to be the point of origin. However it would be foolish of me to ignore the possibility of more than one influence. Certainly Piacenza is very near Milan and it lies between Tortona and Cremona where my author had close connections. What would be really significant is if someone finds a symbol which looks like “cPh” or what I term “c4sc”.
Anyway thanks again!
As I have suggested, to some considerable controversy, that
within my framework of analysis it seems likely that Milanese diplomatic cipher techniques had some influence on the Voynich cipher. In the context of my theory the Benedictine Dots for Vowels system that Nick describes is also a not implausible influence.
I am very suspicious of the idea that the Voynich cipher had no influences to on and that it was conjured up out of nothing; that just doesn’t make sense. It seems logical to me that the author must have had some knowledge of and experience of ciphers. It also seems logical that it is very likely that some of that specific knowledge and experience found its way into the Voynich. This is not to say that there are not very significant unique aspects to the Voynich cipher that were purely invented by the author.
One question that I haven’t explored is what the sum total of cipher knowledge there was at that time. It is not inconceivable that the author drew very widely for his/her influences. I wonder what the most sophisticated cipher one could produce was given the sum total of contemporary cipher knowledge without any invention on top of that,
Clearly human beings are capable of great invention and there is no reason to believe the author was not, so a lot of the cipher could be pure invention. Without doubt the author had a great interest in ciphers and had been honing his/her skills and developing his/her cipher ideas many years before writing the Voynich. I believe he/she was a lover of ciphers and that is the primary reason the Voynich was written in cipher. I don’t believe his/her thought process was “I am writing a book which needs to be secret, so I must learn about ciphers.”. Rather I believe it far more likely that he/she thought “I want to write this scientific book and it would be nice to be able to keep it secret and this would be an excellent excuse to use my cipher skills.” or even “I would like to write a book in cipher and it would be nice to write it about herbs and astronomy.”.
I was wondering about the size of the rule book (cipher key), how many rules and complication they were and how much information there is in the rule set.
If you look at a more complex diplomatic cipher key, whilst some might understandably believe the rules are intrinsically simple, there sheer number of symbols used and their correspondence is very large.
Purely as a hypothetical if the Voynich rule book was as long as the Voynich itself I would think it uncrackable unless the rules could be more concisely summarised. I think I can say that we don’t how big that rule book could be. I would think we can says that it must be bigger than a certain size. If you have a lot of rules I think it starts to become a nightmare to decipher.
Bearing in mind the seemingly mundane, innocuous nature and theme of the manuscript; subjects being presumably of a legitimate general interest, topical and educative type, what then could the compositors hope to gain from such fine work by using indecipherable code to define it all.
I wonder if the Voynich manuscript has implications for cryptography.
Assuming it contains a cipher then it seems to imply that constructing a cipher which cannot be broken by the likes of William Friedman etc. is relatively easy. I say relatively easy as this was achieved in the 15th century, so unless the author discovered a new highly advanced cryptographic method this cipher must have been easy to design in the context of modern times.
If one didn’t know about the Voynich one would be inclined to presume that any 15th century cipher could be broken by modern cryptographers, but of course that is not the case.
Whilst as we know from this website many ciphers have not been broken, but in those cases the problem of a much smaller quality of material is a significant factor I think.
I wonder how difficult it would.be to design a “simple” cipher that modern cryptographers could not crack even if they were presented with hundreds of pages of text.
@Mark
It’s not just the decryption, it’s the language.
Code x language = possibility
Therefore, it is important to first determine the place of origin and to limit the possibilities of the language.
On page 171 of “Die geheimschrift im dienste der Päpstlichen kurie von ihren anfängen bis zum ende des XVI jahrhunderts” by Aloysius Meister you can find a list of 27 of Gabriele de Lavinde’s 1379 collection of Vatican ciphers.
Nick: I was thinking about what you were saying about cipher keys being changed over time and the date on the cipher key not necessarily being consistent with the contents. However it seems to me that if a cipher key was added to over time then of course it would have to be added to by the corresponding ambassador/envoy etc. In which case why not just send them a new cipher key rather than cipher key updates?
It is important for me whether the date on a cipher key can be seen as reliable i.e. can the whole contents of that key be dated to that date?
I suppose my question to you would be what evidence do you have to support your assertion that the date can be unreliable?
And if you believe that we cannot rely on the dating then what leeway do you think we should allow and why?
Nick: Having been touch again with Francesco Senatore he says regarding his book:
“I did not mean that Milanese ciphers were not sophisticated in 1447-48, I just said there were no evidence for the previous period. ”
My assertion based on the evidence that I have seen is that the sophication of Milanese diplomatic ciphers did not improve significantly subsequent to Francesco Sforza becoming Duke and remained at about the same level of complexity until the end of the 15th century; though it may well be that the usage of diplomatic ciphers in Milan greatly increased after Sforza became Duke.
I think I have figured it out, the often repeated words in the Voynich stand for ‘like”, a vastly overused word particularly familiar to viewers of “Love Island” in the UK, I have been lead to believe, though I must confess that I have not watched that television programme myself.
More research strongly supports the idea that there was no significant improvements in the sophistication of Milanese Diplomatic ciphers from the end of the era when Filippo Maria Visconti was Duke until near the end of the 15th century. I would therefore be inclined towards the view that most of the advances in diplomatic ciphers techniques in Milano during the 15th Century occurred during the reign of Filippo Maria Visconti though it is conceivable that they pre-date that, however I think that unlikely, as it would imply that Milanese ciphers at that early period were vastly more advanced that those of their rivals. Certainly it does appear from what I have seen that the ciphers of the era of Gian Galeazzo Visconti were not particularly advanced, though more evidence could change my viewpoint on that.
I think it worth mentioning again that the “4o” symbol can be found in early 15th century cipher alphabets.
So I think one can argue that the presence of the “4o” symbol in the Voynich and the level of sophistation of the cipher are just as consistent with a pre-1447 date for the Voynich as a post-1447 date.
Nick: Lydia Cerioni lists thoroughly the names of the envoys or ambassadors to other states and the names of those states and when their term started. So there is a clear list of states for which there was enciphered communication subsequent to 1450. However I naturally wonder the extent of the communication with other states prior to 1447. You mention the treaty of Lodi and I am inclined to the view that there was a increase in the amount of enciphered communication by the Milanese state throughout the 15th century, but the extent of it in the early 15th century over time is less clear to me.
Mark: Both the Yakuts and Miwoks, both of whom had settled around present day Lodi in pre Columbian times were known to have had treaties with the Spanish, though not as early as the mid fifteenth century from memory. Perhaps there has been some confusion between the Northern California Yakuts and the Turkic tribe of that name from Central Asia. Whether either are related to the our own Nahuatl Voynichero nominates from the southern regions remains to be seen. I did happen to swing by that area some years ago for wine tasting and I can tell you one thing for sure, I’d hate to get stuck in Lodi again.
John Sanders: I think we are talking about a different Lodi, but then I think you know that. Lodi, Italy not Lodi, California. Lodi, Italy looks quite pretty, though I have never been there and can’t say what it would be like to get stuck there.
Mark: I always have had a great deal of time for self assured folk, who go about their own thing confidently in their own way and are not detered by critisism. It’s nice to see someone who can also share a corny, though well intended joke at their expence, whithout being distracted. I wish you all the best in your efforts which deserve to be rewarded.
John Sanders: Well that it is very kind of you, though my self assuredness could be a sign of false confidence. I hope that I will take valid criticism onboard, though distinguishing between valid and invalid criticism is always hard. I have certainly worked hard researching the Voynich. Unfortunately I am sure that there are many people who have put in effort researching the Voynich whose efforts unfairly will not have been rewarded as many different theories are mutually contradictory. I try to be diligent in the research I do, only time will tell the extent to which I am right and the extent to which I am wrong. Anyway, I thank you for your best wishes.
Even if there were not significant advances in the nature of Milanese ciphers from the end of the reign of Filippo Maria Visconti, through the Ambrosian republic and into the era of Francesco Sforza that does not mean that there were not significant advances in decipherment techniques in the Sforza era. One interesting question is regarding Cicco Simonetta’s suggestion of using more than one language in a cipher, the extent to which that approach was used in the Sforza and Visconti eras I have not determined.
Regarding the status of Visconti ciphers I think it worth noting that Francesco Barbavara, Gian Galeazzo Visconti’s Chancellor and subsequently serving Filippo Maria Visconti, is said to have created the first modern foreign office. (See Garrett Mattingly) One would suspect that the advances in the conduct of diplomacy resulting from this foreign office would be accompanied by advances in the use of cipher techniques and therefore that pre-Sforza ciphers would have been more advanced than those of other city states at that time.
To quote Aloys Meister:
“Even if I did not succeed in finding very old documents containing any Milanese ciphers, this cannot be proof that they did not exist. The work on the Milanese manuscript treasures is still far from complete, so that for our purposes there are still surprising discoveries remaining for the future.”
“It must be left to the future whether, on the basis of ample material, the circumstances under which the beginnings of the Milanese ciphers had evolved can be ascertained.”
Another quote from Meister:
“It would raise the question of whether the lost Visconti ciphers were more advanced than those of the Venetians, or instead if they themselves were influenced by more advanced Venetian cryptography. In any case, we must resign ourselves to the fact that for the time being we cannot judge the time before the Sforza.”
As far as pre-Sforza ciphers, I have an electronic copy of a reproduction of a complex cipher key dated to 1447 and headed Filippp Maria Visconti, so I would doubt the cipher key was added to significantly subsequently as Filippo Maria Visconti died in 1447, so I think there is a very strong case for deducing that these kind of complex ciphers were already in use by 1447. I have recently seen other complex Milanese cipher keys dated to before 1450, when Francesco Sforza became duke. So I will restate that it appears to me that in fact there was little innovation in Milanese ciphers from when Francesco Sforza became duke until the very end of the 15th century. I will continue to amass more evidence which will either strengthen this hypothesis or weaken it, but at the moment the situation looks pretty clear. The more significant question remains, how did Milanese ciphers evolve during the first half of the 15th century? Did the complex ciphers we see in 1447 only emerge in 1440s or were they in use significantly earlier?
I should add that I have downloaded from the BNF website an enciphered letter from the time of Gian Galeazzo Visconti which does not appear to be very complex, though I don’t know from which period of the reign of Gian Galeazzo the letter dates. (Thanks to the Voynichero who recommended the BNF to me.) So my hypothesis is that it is most likely that the big advances in Milanese ciphers date from the reign of Filippo Maria Visconti. Though it is conceivable, but much less likely I think, that significant advances occurred during the brief, turbulent and unsuccessful reign of Giovanni Maria Visconti. Filippo Maria Visconti reigned from 1412 to 1447, so most of the period from which the Voynich has been dated to covers this period which if the links to Milan exist then it would be consistent with a period of creative cipher development.
I would contest that the importance of ciphers to the city states after the Treaty of Lodi, in 1454, and the 40 subsequent years of peace was less than during the prior period. The wars in Lombardy, the first of which began in 1421 and the last, which was the fourth war, ended in 1440. So the incentive to develop and improve the cipher techniques used, I would argue was much greater during the period 1420 to 1438 than after 1454.
Mark: the mainstream historical narrative since the mid-nineteenth century has been that Francesco Sforza’s entourage (including Cicco Simonetta) was the source of much of the cryptography practice of the day. You can hear Meister gamely battle against this story in the quotes you extracted from him.
Now, I don’t know whether this is true (though there is certainly indirect evidence that suggests it might be): most of the underlying ideas seem to have been in place from 1400-1410, so this discussion isn’t so much about theory as praxis.
Nick: I guess then I am disputing the mainstream historical analysis. I think as you are right to say “Absense of evidence is not evidence of absense.” I think it is an assumption easily made that due to the scarcity of evidence of pre-1447 ciphers they must have been very simple and yet we know why most of these have been lost to history and that loss has nothing to do with the simplicity or complexity of the Milanese ciphers. As you say in one sense the essential nature of these ciphers did not change over time, however more complexity was added with symbols corresponding to letter pairs, in the case of one cipher key that I have seen two distinct symbols for the same letter pair. In addition we see a growing number of symbols for common words etc. Of course when you say the underlying ideas seem to have been in place from 1400-1410, in the case of Milan we don’t really know. Anyway I think I am gradually shedding light on pre-1447 Milanese ciphers. I am not yet aware of any evidence that indicates that pre-1447 cipher techniques were simpler than those up to the 1480s and I am aware of evidence to the contrary. If you are aware of a source that has evidence to support your point that would be fascinating.
Ciphers may have been used much more in the era of Cicco Simonetta, this is possible, though Milan was ahead of the game when it came to diplomatic activities in the first half of the 15th century from what I have read.
On another topic did you manage to download my horrendous annotated Voynich 9 rosette map? I am working on my writeup and I will do the best to finish it as soon as I am able. Then should you wish, you will be well-place to highlight the strengths, if you feel there are any, and weaknesses of my analysis.
Nick: Contrary to popular opinion on the subject, from Rene, Charlotte and I daresay others, I am suspicious that there is more in common between the Voynich ciphers and the diplomatic ciphers than is generally believed. I have some ideas which I hope to flesh out at some point soon as to how a different take on those ciphers techniques could explain the properties we see in the Voynich, although even if my line of thinking is broadly speaking correct I think the process of decipherment remains daunting.
Nick:
When you say “the mainstream historical narrative since the mid-nineteenth century has been that Francesco Sforza’s entourage (including Cicco Simonetta) was the source of much of the cryptography practice of the day. ”
Do you have sources that you can refer me to that make that statement?
When you say that “there is certainly indirect evidence that suggests it might be.” What evidence are you refering to?
You say “most of the underlying ideas seem to have been in place from 1400-1410”
When we look at the 1401 Simone de Crema Mantua cipher one sees a marked difference in complexity from the later ciphers. Do you have any examples of ciphers between 1400-1410 that display the level of complexity we see in the ciphers of the era of Francesco Sforza?
Nick: I am sure you appreciate that I am not so trying to tear down aspects of your theory, but rather explore where my hypothesis contradicts yours and what the factual basis for that disagreement is. My assertion is that there was no significant changes in the features of diplomatic ciphers in Milan from the time when Filippo Maria Visconti was Duke until about the 1480s.
It may be that the Sforza administration disseminated and popularised those techniques outside of Milan. It may be that the Sforza administration had many more ambassadors and envoys than before. As to these two points I have not come to a clear determination. However as far as there being a technical leap in the sophication or complexity of Milanese ciphers with the advent of Francesco Sforza and Cicco Simonetta the evidence that I have seen does not support this assertion.
I have read that in fact Cicco Simonetta’s text on cryptography was not appropriate for cracking the modern ciphers of that time and so out of date; I don’t know how true that was as I haven’t yet investigated. And I believe you point out there does not appear to be any evidence of him writing ciphers or that he even wrote the text himself. That is not to say that he didn’t have an highly important and influential role rather just that neither he nor Tranchedino advanced the cipher techniques used. I know I may seem belligerent on this point, but this has some significance to my research and I would think to yours.
Nick: There is one vision of the Milan of Filippo Maria Visconti’s administration as a cipher backwater where very simple ciphers were used occassionally for diplomatic communication. This is a Milan which was behind its rivals in its sophisticated knowledge of ciphers.
There is another vision of the Milan of Filippo Maria Visconti’s administration as being a dynamic area of the development of more and more advanced cipher techniques, starting from the basic ciphers that we see in the Mantua 1401 cipher, that then spread to other city states like Venice, Urbino etc.
These two visions are diametrically opposed. The truth is surely somewhere between the two.
The first vision doesn’t stack up with the evidence.
The second vision fits very well with my overall analysis of authorship and the vision that I think is closest to the truth.
Nevertheless more evidence is required to determine the extent to which the second vision represents the truth.
Mark: these visions of yours sound a lot like historical hypotheses. And only in very rare circumstances do those prove anything in and of themselves – rather, all they do is help direct your attention to the kind of evidence you should be amassing to test them. Remember, it’s the evidence that ultimately does the heavy lifting (i.e. proof), not the hypothesis. :-/
Nick: You are spot on they are historical hypotheses. I fully agree hypotheses do not prove anything in and of themselves, but that does not make them of no value. I agree they help to direct ones attention to the kind of evidence that one needs to test them that is why one forms hypotheses. I absolutely agree that evidence is the key not the hypothesis, but the hypothesis is of value as part of the process of compiling evidence.
(As I have said elsewhere I think “proof” is the wrong word and wrong paradigm. I think the key concepts are likelihood i.e. “probability”. If you mean by proof a statement which has 95% probability of being true that’s great. If you are talking about certainty then I think that is the wrong mindset for these kinds of historical questions.)
I think evidence is vital that is why I replied to your previous comment with:
When you say “the mainstream historical narrative since the mid-nineteenth century has been that Francesco Sforza’s entourage (including Cicco Simonetta) was the source of much of the cryptography practice of the day. ”
Do you have sources that you can refer me to that make that statement?
When you say that “there is certainly indirect evidence that suggests it might be.” What evidence are you refering to?
You say “most of the underlying ideas seem to have been in place from 1400-1410”
When we look at the 1401 Simone de Crema Mantua cipher one sees a marked difference in complexity from the later ciphers. Do you have any examples of ciphers between 1400-1410 that display the level of complexity we see in the ciphers of the era of Francesco Sforza?
——
I fully appreciate you are a busy guy and therefore may not have time to address these points, but these are clearly all statements about evidence that you made and I was just asking where this evidence is. Anyway please don’t feel under any pressure to justify those statements as I imagine you have other priorities.
I am compiling evidence and so far it is pointing firmly towards no significant increase in the level of sophistication of ciphers from the era of Filippo Maria Visconti to those that one sees in the era of Francesco Sforza, more than that I cannot yet say.
I ought to clarify these “visions” are both ends of a spectrum of hypotheses i.e. the truth almost certainly lies somewhere between the two. The question is where on the spectrum the truth lies. I believe it is closer to the second “vision”/hypothesis on the basis of the evidence that I have seen.
Nick: You may be interested in the following:
http://www.ep.liu.se/ecp/149/007/ecp18149007.pdf
This is an enciphered letter to Cicco Simonetta from Tranchedino from before 1450.
Mark: when I wrote Curse, I worked mainly with modern, more balanced accounts of Francesco Sforza’s dukedom (all listed in the Curse bibliography online, e.g. Senatore). These all sought to present pictures of him that discarded the heroic mask cast for him by Victorian-era historians, part of which involved Simonetta’s presumed cipher supremacy.
You can still hear echoes of this in Meister and elsewhere, but it would take a while to root out the ultimate source.
For me, the cipher innovations (e.g. homophones, nomenclators, nulls, twins, verbose cipher, syllabaries, etc) of the fifteenth century seem to have no obvious single source. Hence the suggestion that it was Milanese cryppies who brought these all together remains no more than a nice story, in the absence of actual evidence.
Nick: So this implies that Cicco Simonetta was in the habit of reading and writing ciphers. Of course this in no way undermines what I was saying. However it leads me believe that these techniques were widespread before Francesco Sforza became Duke.
Mark: thanks very much for that, it’s a good paper, one I’ll need to review and discuss more thoroughly than margins such as these allow. 🙂
Mark: indeed, one of the most interesting ciphers in the 1440 Urbino cipher ledger is from Francesco Sforza. But more on that another day. 🙂
Nick: I have downloaded all of the Codex Urbinate (66 cipher keys) from the Vatican website, you can find it on my OneDrive. I emailed you about it and posted the link to it.
Nick: I will get in touch with the author of the paper who seems to have already done much other work on Medieval ciphers. I am curious as to what she knows about other 15th century Italian cipher keys in the Moscow archive.
Nick: I agree that there is no evidence that the cipher innovations (e.g. homophones, nomenclators, nulls, twins, verbose cipher, syllabaries, etc)
of the fifteenth century have an obvious single source or obvious multiple sources.
I agree that given this the suggestion that it was Milanese cipher writers who brought these all together remains no more than a suggestion,
in the absence of actual evidence. Fundamentally I think we have very little evidence which state was the first to innovate by introducing a
new feature of the cipher which then spread to other states. It is perfectly possible that Milanese cipher writers introduce very many techniques,
but likewise it is possible that they were responcible for very few cipher innovations. I think it may be very difficult to determine who innovated
techniques who merely copied them.
Nick: In one sense one could argue that ciphers when Francesco Sforza was Duke of Milan, or in any other city state than Milan, were less sophisticated than under Filippo Maria Visconti as I don’t believe that I have seen elsewhere any examples of more than one symbol being used for a letter pair.
This is consistent with the idea of Filippo Maria Visconti’s administration being a cipher innovator rather than a cipher backwater. I realise this is not a huge difference, but nevertheless I think it is worth stating.
Nick: One aside of note, when originally looking through the Codex Urbinate I did not focus on the headings of each cipher key, but rather on the many symbols contained within for parallels with the Voynich. There are about 60+ cipher keys in the codex dating from as early as the 1440s up to the 1470s. Looking through them all there was one cipher key that I thought was most reminiscent of the Voynich, though far from identical. I said to myself I wonder is this is a cipher key associated with Milan and sure enough this was the only cipher key headed “Mediolanus”. Now that is anecdotal. However given a probability of less than 1/60 of selecting that cipher key (Less than 1.7%) it is quite remarkable if that was a coincidence. The cipher key was dated 1467.
Having looked at the cipher key for Francesco Sforza it is quite simple with no striking parallels to the Voynich. So whilst the other cipher key is dated to the reign of Francesco Sforza in Milan, there is no cipher key for Milan with an earlier date, I would argue that it is very much a Milanese cipher key rather than a Sforza cipher key.
Another point is that if these were cipher keys generated internally by the Urbino Chancery for ambassadors or envoys then would not expect the cipher key for Milan to show Milanese symbols. However if the cipher key was generated in conjunction with or even wholly by the Milanese Chancery for their mutual communication it would make perfect sense for their to be Milanese symbols within the cipher key. This fits in with my broader thesis that envoys and ambassadors were not always used and with some states enciphered communication was direct, either because a state could not afford to have an envoy or ambassador in that state or simply that the two states were on such good terms it was viewed as a problem. I would guess that when a cipher key is headed by the name of the state e.g. “Marquis of Mantua” then enciphered communication was direct whereas when the cipher key is headed by the name of the ambassador/envoy then of course enciphered communication was through that person. I suppose it is possible that the cipher key could be titled with the head of state, but be transmitted through the ambassador or envoy, though why in other cases would the cipher key be headed by the name of the ambassador/envoy if this was the case?
Nick: We have a match!
I decided to see if the “Mediolanus” cipher key in the Codex Urbinate was the same as the “Urbino” cipher key in the Tranchedino cipher ledger and they are a match.
Well, so what…
This means that I can potentially find bona fida Milanese cipher keys from the time of Filippo Maria Visconti in the archive of other city states where those cipher keys are headed something like “Duc Visconteo” or “Mediolanus”, but not when they are headed with name of their ambassador or envoy to Milan, I would think. The question in this instance where we are dealing with shared cipher keys is who was responsible for generating the cipher key? I would guess that on average the most significant state of the two would be the one responsible for the cipher key and therefore authenically Milanese cipher keys would be most likely to be found in the archives of city states significantly more minor than Milan.
I think this is a significant result. Do tell me what you think.
Checking back I think the shared cipher key is dated 1457 not 1467.
Nick: I would think if states often had shared cipher keys for their mutual communication this could be to some extent responsible for the transmission of cipher technology or cipher symbols from one state to another.
Nick: It is worth noting that whilst the contents of the two cipher keys is the same, the handwriting appears to be quite different.
Mark: you’re starting to look at the evidence in the right kind of way. 🙂 The two ledgers would have been handwritten by the two city-states’ respective cipher secretaries, hence the different hands. The difficulty with the Milan ledger is that it is a third copy of the ciphers, and so we cannot see the corrections and extensions added to the ciphers over time. However, the Urbino cipher ledger looks to be original. There’s a lot to be pulled out from these if you are careful! 🙂
For the information of anyone interested:
Page 14 of the Tranchedino cipher ledger
has the same cipher key as
Page 26v of the Codex Urbinate (Page 62)
Nick: Yes, I feel like I am making process. I think I am gradually better understanding the process by which enciphered communication occurred.
Yes, what you say regarding the two cipher secretaries sounds correct to me. I wonder precisely how that happened, presumably one of the cipher secretaries designed the cipher key and the other reproduced it; these minutiae may seem unimportant, but I see them as potentially useful as part of the process of knowing where to look for Milanese cipher keys. In this instance I suspect that this cipher key was designed by the cipher secretary from Milan rather than the one from Urbino. They could have come up with it together, but this seems more complicated than necessary.
By the way, where have you read about it being the third copy as that is something that I was not aware of and the source for that information may have other insights?
You mention that we cannot see the corrections and extensions added to the ciphers over time in the Milan ledger. I wonder about the degree to which there were extensions and corrections as I do not think we see any corrections to the cipher keys in the Codex Urbinate and as you say that looks like it might be the original. Any extensions that were applied do not seem to be obvious. The changes that I have seen are those in some the headings where the name of the recipient has been crossed out and the name of another recipient has been written next to it, presumably due to a change in the envoy used in that case. If we do not see any changes to the “Mediolanus” cipher key in the Codex Urbinate then I would think it the case that there were not changes to the “Urbino” cipher key in the Tranchedino cipher ledger as they should always mirror one another.
I am gradually amassing cipher keys with the focus on the reign of Filippo Maria Visconti.
However for example I was wondering where the archive of Francesco Sforza’s documents is to be found from before he was the Duke.
If I can work out the headings of the Filippo Maria Visconti cipher keys on the basis of the names of the states that they were communicating, but also the names of their envoys or other individual that they were communicating with, in cipher, as they may have Milanese cipher keys in their family archive. For every Milanese cipher key there was an identical twin held by the recipient, so if I can find the twins I can reconstruct the Milanese cipher ledger. Now I appreciate many of the twins were destroyed, though it is possible that the recipient did retain an enciphered letter(s) from which the twin can be deduced.
Nick: It is an obvious point that both cipher keys from a pair shared between two parties communicating with one another must always be synchronised with one another. If they are not in sync this could render a message from one party to another indecipherable, which of course is not desirable. So if there is a change to one cipher key of this pair that change must be made to the other cipher key in order for them to remain in sync, when you are dealing with different states as in the Urbino-Milan case it seems rather an arduous and awkward process to update both cipher keys and one wonders whether in such an instance it would be more sensible to start over with a new shared cipher key.
Nick: There are a couple of characters in the Tranchedini Simonetta letter very reminiscent of one of the unusual rare characters(the V with smoke coming out of it).
I believe one ignores rare characters at one’s peril just because they have a low frequency. As long as one can with some confidence identify them as a geniune character I think they are important.
The “smoke” coming out of the “v” is a common Latin abbreviation symbol.
Most of the time it stands for er/re/ir/ri. In medieval manuscripts, you will frequently see the word “virgin” written (in several languages) as v’gin with the apostrophe being the wiggly vertical shape (which is often written above the “v” rather than beside it). It is also commonly found in words like “verse” and “versus”. In fact, it is used so often to denote “verse” that it is sometimes written v’ without the “se” if the meaning is clear from the context.
Sometimes it barely has a wiggle, sometimes it looks like a lightning bolt, sometimes it has several wiggles. A few scribes will write it vertically to mean er/re and horizontally to mean ir/ri but most don’t make a distinction and will indicate both with the vertical squiggle.
The character is only rare in the VMS (occurs only once). It is not rare in manuscripts that use scribal abbreviations.
JKP: I genuinely wonder if there is any character that man could conceive that you would not describe a consistent with Latin abbreviations, Greek characters, Mathematical and Alchemical symbols, were it to be found somewhere in the Voynich. Now I am sure you will answer no and say that the Voynich characters fit with that tradition and that other characters do not. I remain sceptical. As I believe you say in your blog you decided straight away that the Voynich characters were derived from Latin abbreviations and then sought evidence to support that hypothesis; by doing this you inevitably discounted all the alternative hypotheses. Another possibility is that the set of Latin abbreviations is so vast that it could encompass a myriad of characters. In the context of your theory why is it that I see much more similarity between the Voynich symbols and Milanese diplomatic cipher symbols than I see between the Voynich symbols and those of other city states.
So common in Latin, yet just one example in the whole Voynich text; how strange. Seems like it might well be yet another example of our clever Lily’s smokin’ gun taunt..
Mark,
Cappelli, page 399, gives an example of a “V” with a “smoke trail” that dates from the 15th Century. Cappelli notes “(vtm) vocativum XIV f.” and the “f” might give more useful information but I can’t find the meaning.
Why is that strange? In the diplomatic ciphers the Latin shapes don’t match the frequency of Latin-based text—they are cipher characters used for other purposes.
In the VMS, the shapes are also used for other purposes, whatever those may be. If they weren’t we would be able to read it.
Mark Knowles wrote: “As I believe you say in your blog you decided straight away that the Voynich characters were derived from Latin abbreviations and then sought evidence to support that hypothesis…”
No, that does not accurately reflect my attitude about the glyphs at all.
I had no hypothesis about the origin of the glyphs. I had curiosity about it. Many aspects of the VMS glyphs (a high proportion) are immediately recognizable as Latin as well as a small proportion as Greek. Surely you can see that the letters “a” and “o” resemble Latin characters. To someone familiar with medieval Latin, many of the other shapes are just as familiar as “a” and “o”.
But I did not ASSUME they were derived from Latin. Not at all. What I did was throw the idea of Latin on the back bench and spend several years investigating hundreds of other alphabets to see if there was something out there that I might not know about that might have inspired the glyph shapes.
As I have mentioned repeatedly, I already knew how to read Korean, Japanese, a little bit of Chinese, Russian (Cyrillic), and of course several languages that use Latin characters. I further taught myself enough Amharic, Gujarati, Georgian, Armenian, Greek, Syrian, Hebrew, a tiny bit of Arabic, and many other alphabets, to the point where I could at least read simple words in old manuscripts. I even managed to puzzle out most of a page in Syrian and to figure out quite a few of the Greek scribal conventions that are different from those that were adapted for Latin.
If you can read simple words, it means you are familiar enough with the forms to recognize how they are drawn and how they are used. If you can read simple words, you can usually distinguish between strokes that are meaningful and those that are embellishments.
At first a couple of the gallows characters might seem like oddballs, but if you are familiar with scribal abbreviation shapes and the various ways in which they are combined, it’s possible that even a couple of the oddballs may be inspired by Latin and Greek scribal combinations.
.
In the end, after years of trying to find a better match, it became clear that Latin characters and abbreviations not only matched best in terms of glyph shapes, they were also far-and-away the best match for abbreviations, not only in terms of the shapes, but also their position in the tokens. No other alphabet passes all three tests and none of them except Latin (and its Greek predecessors) pass the position test.
In your blog post “The Origin of the Voynich Glyphs”:
You say “When I first encountered the VMS, I recognized most of the shapes from medieval scribal traditions” (I am yet to experience that recognition)
You then say you checked out some other alphabets.
And later “The Latin alphabet and scribal abbreviation conventions can explain almost all the VMS characters. I already knew this, but sometimes you have to look around to appreciate what you already have.”
Your blog post called “What Can Hands Tell Us?” tries to compare the gallows characters with others that have less than a loose similarity.
In your post “Latin’s ‘Om-age’ to Indic Numerals”:
In example #1 of the numerous items that you have circled in red only one has a resemblance to something we see in the Voynich. In example #2 amongst the many circled items I don’t see any similarities. In example #3 again there are no similarities. Likewise with example #4. I then noticed below some circled characters that looked very much like what looked like those that one sees in the Voynich, but of course I realised a split second later that the text was from the Voynich.
Looking at your post “Voynich Script – The Leaning Letter and Why I Never Use the Eva Font”:
Again all your characters comparison show little or no similarity.
As far as I can tell you seem on occasions to be saying that if one combined some penstrokes that are sometimes seen amongst the latin abbreviations then one could in theory construct some of the Voynich characters; that seems pretty weak. Probably if you combined enough penstrokes used in the scribal abbreviations you could come up with a painting of the Mona Lisa, but I doubt Da Vinci did it that way.
I don’t know how you can say that the cipher alphabets aren’t consistent with the Voynich.
So I guess I will focus on cipher alphabets and the extent to which they can be linked to the Voynich cipher and you will explore your own avenue.
I have a feeling that going onwards I will comment on characters and you will say they can be found amongst latin scribal abbreviations and I will then express my doubts and so on.
In the blog post “Letter Patterns, EVA-m (the ‘j’ Shape)” you find a shape which does look like a Voynich character, however this shape is very common in diplomatic ciphers.
Byron: I can’t find the example you refer to in Cappelli, page 399
Mark, you seem to think all my posts are about the similarity between VMS characters and Latin. They are not. In fact, only about 10% of them are on this specific topic.
The one on Indic-Arabic numerals that you referenced above is about the commonalities between Indic and Latin scribal traditions that existed in the Middle Ages, some of which have since disappeared (modern Latin does not use medieval abbreviations). It is not written as a direct comparison between the VMS and Latin. It is a conceptual article about how these abbreviation traditions came about. One needs to understand the underpinnings before one can learn the details.
Yes, the EVA-m shape is very common in diplomatic ciphers and even more common in Latin-alphabet manuscripts. The diplomatic ciphers, as I have already pointed out with a visual example from Tranchedino, uses many Latin abbreviation shapes.
Also, as I have pointed out, each of the diplomatic ciphers uses dozens of characters, some use as many as 200 glyphs for a single cipher and each cipher has a different set. This means the diplomatic ciphers have to use every shape they can think of to meet their need for thousands of different characters. It’s only common sense that in addition to Latin characters and abbreviations they would have to draw from dozens of other sources.
Mark Knowles wrote: “As far as I can tell you seem on occasions to be saying that if one combined some penstrokes that are sometimes seen amongst the latin abbreviations then one could in theory construct some of the Voynich characters; that seems pretty weak. Probably if you combined enough penstrokes used in the scribal abbreviations you could come up with a painting of the Mona Lisa, but I doubt Da Vinci did it that way.”
No, that’s not how it works. If you had the smallest working knowledge of medieval Latin script you would understand why this argument is not valid.
There were conventions about how things were combined. WITHIN these conventions you could be creative.
For example, there are three different abbreviation shapes for certain letter combinations. Which one you use is up to you. A simple word like “compotus” can be written six different ways. MOST scribes chose the more common ways, but if a scribe used a unique (but valid) combination, other scribes might snicker, but they would still be able to read it.
.
If 75% of the VMS glyphs follow common conventions and 15% follow less common (but recognizable), conventions and 10% are different or unique (or more similar to Greek than Latin), are you going to reject the assertion that the majority of glyphs are based on Latin? It seems to me that is what you are doing.
JKP: Well the combination of conventions that you describe seems to leave so much flexibility that of course you can derive some Voynich shapes. You could derive all shapes from Sumerian script given enough flexibility; in fact in one sense the Voynich shapes can be derived from Sumerian.
The idea that the author of the Voynich was widely laughed at for the characters he/she used seems peculiar.
On that basis, then within the framework of the diplomatic cipher alphabets you could derive all Voynich characters.
The argument you decribe appears very weak. A much stronger argument would be if you could demonstrate examples of specific Voynich shapes being used not how they could in theory be constructed.
Mark, you don’t seem to get it. There is only a small percentage of oddballs. You keep talking as though most of the VMS shapes are not standard characters when there are only a few that are head scratchers, and even those are possibly conventional shapes in slightly unconventional combinations.
Saying that the Voynich shapes can be derived from Sumerian is ridiculous. As I’ve said repeatedly, it’s not just the shapes that are consistent with Latin, it is also their behavior and their positions within tokens. Sumerian has nothing in common with VMS characters in shape, behavior, or positional arrangement.
.
How many examples do you need? You don’t seem to have gotten anything out of Cappelli, one of the leading references for paleographers and those who want to read medieval Latin. If Cappelli is not comprehensible to you then anything I post is not likely to be meaningful either.
JKP: You need to be able to account for all of the shapes not most, though as far as I can tell you don’t seem to be able to account for most of the shapes. There are certainly some basic shapes in the Voynich that can be found in the latin alphabet or amongst the arabic numerals, but that hardly justifies your scribal abbreviations argument.
Cappelli is perfectly comprehensible as is your blog, the problem is that they do not persuade me that Latin abbreviations are the source of the Voynich alphabet. Your blog is particularly unpersusasive.
Anyway let’s go back to agreeing to disagree.
JKP: I should say your blog looks particularly nice visually and many of your blog posts are well written, well presented and an entertaining read. However, whilst I have not read all your posts, there are quite a few that I disagree with as in the case of Latin abbreviations.
Mark Knowles wrote: “You need to be able to account for all of the shapes not most…”
I don’t think anyone is going to agree with you on this, Mark.
If I wrote this text:
Onçe upøn å †ime †here waƨ å Little girL whø Lived a† the edge of †he føreƨ†.
Are you going to tell me it does not resemble the English alphabet just because there is not 100% match?
The VMS characters that divurge from conventional Latin (and Greek) are in the minority and even those that seem strange are only somewhat strange, just as substituting ø for o in the above example is only somewhat strange.
Anyway, I’m done. When you learn to read medieval Latin we can talk again.
And by the way, yes I know those are Scandinavian characters in my example.
I chose them because they are easy to type, not because there is some discernible “foreign” component in the small number of VMS characters that diverge a little. The ones that divurge still resemble Latin (and Greek) abbreviation shapes, they are simply combined in a less common way.
I want to correct you, the symbol “smoke” occurs twice. This is the page f1r and one more on f80r (10 lines from the bottom, above the letter “о”). And in the second case, the text cleanup occurred clearly. And, probably, the new word did not fit into the allocated interval, and therefore a “smoke” was used to indicate the omission of several letters.
I have read the posts again and must say J.K.P is right. It does not matter what the gyphers look like, but how they are used.
Abbreviations of names and locations in encryptions have nothing to do with shorthand typing.
All assume that it is real Latin, but do not remember that it is a person who is not really right Latin. eg. Hildegard von Bingen.
For me personally this is the most likely option.
Maybe he could not finish his studies.
To the page F1, interesting that even about a single character is spoken, since it is unique anyway. What about the other 2 characters where are unique.
The question that comes to mind is why the top sign is not left where it really belongs. Everything else is not relevant as it does not help.
PS: If somebody does not know what shorthand is, he would guess in Arabic, but even there endings are just endings.
JKP: It is certainly true that I am not knowledgeable in Medieval Latin. However I doubt that even if I was an expert on the subject I would agree with you as the disagreement that I have with you seem to have little to do with that specifically, but rather to do with the similarities or lack of similarities of the visual appearance of the symbols in the Voynich with the characters in the Latin abbreviations. Again I do not find the appeal to supposed expertise on this subject to justify a theory that visually does not seem justifiable.
The fact that you say that the author of the Voynich if writing using Latin abbreviations would be snickered at by other scribes says it all. I can see why someone might find it laughable that the Voynich characters derive from Latin abbreviations.
JKP: I don’t agree with most of your analysis in your blog post “The Cardinal’s Nymphs”, though some of it does fit with my thinking. This is not, because you have some kind of expertise in Medieval gardens and if I only understood Medieval botanical gardens as well as you do then I would see that you are correct. The same applies with Latin abbreviations; it is just that I don’t find your arguments persuasive not that if I had the level of knowledge of Medieval Latin you claim to have then the arguments that don’t seem logical would suddenly become logical. Rather than appealing to the expertise, that you say you have, your arguments need to be self-contained and justify your position in and of themselves.
I consistently have not tried to draw you into this subject and you have been keen to insert your opinion, so if you are frustrated by this discussion then just recognise that we disagree on this subject. I respect that you are an intelligent person and have spent a considerable amount of time studying the Voynich, but you need to appreciate that I am not obliged to agree with your thinking. Maybe some day I will decide that you are right or you will decide that my perspective is the correct one, we will see. I think you would better use your time developing your arguments on your blog than interjecting your theory as it also wastes my time. Anyway good luck with your research!
Peter M: I think it does
notmatter very much what the symbols look like. How they are used is also certainly important. I think the diplomatic cipher alphabets better fit with better overall with the criteria.Two beautiful minds thrashing it out; one leadiing with unabashed Norse Intellectual superiority and the other countering with most uncommonly modest Welsh reason. Quite enthralling to the point of distraction, though unfortunately far departed from what appears to this fuzzy minded critic, as the simple God’s honest truth. I have a long held, though somewhat guarded respect for both parties to the proceedings; trusting that the eagerly contested outcome may prove to be of benefit to other Voynich adherents, perhaps at some not too distant point in time.
@Wladimir Dulov
The second smoke on page 80v looks more like an ink stain. That was certainly no intention and it is blue. You can find more splashes on the page.
John Sanders: It is very kind of you to say so. I would like to think that we both match your comments regarding our intellect, though I have my doubts. If we were both quite as intelligent as you suggest I am sure one of us would have unravelled the mystery of the Voynich some time ago. As it is we must both plod on exploring the theories that make sense to each of us. I think there is a lot to be said for the fact that we both have different theories, that way there is more chance that one of us is on the right path.
I don’t think I am modest, so much as someone doing their best not to fall into the trap of delusions of grandeur which I think is so easy to do. This can lead to one stubbornly sticking to one’s theory rather than changing one’s mind as one becomes aware of new evidence. There is a very fine line between sticking to your guns when you believe your position is correct and not being prepared to budge for fear of admitting to yourself and others that you were wrong. I try my best to stay on the right side of that line, but I don’t always find it easy to judge which side of the line I am on.
Peter M wrote: “All assume that it is real Latin, but do not remember that it is a person who is not really right Latin. eg. Hildegard von Bingen.”
I don’t assume it is real Latin, if you mean the language.
A number of people have attempted Latin decipherments but I haven’t seen any that are convincing.
Latin abbreviations were used for Latin, French, German, Italian, Spanish, Scandinavian, English, Czech, and even Indic languages have similar shortcuts (even if the shapes are different).
.
I am talking about the origin of the shapes. This may not seem important to some people (and doesn’t necessarily reveal the underlying language, if there is one), but it does matter in terms of discerning the cultural influences on the designer and his or her mindset. Medieval Latin is conceptually more sophisticated than shorthand. This way of thinking could potentially have influenced the way VMS text is “encoded” (by that, I mean designed/arranged).
Voynichese might be a synthetic language, it might be partly or wholly abbreviated, it might be cipher, it might be a combination of letters and symbols, it might be numbers, it might be natural language, or the Latin characters may have been used as a smokescreen (Bacon developed at least one cipher that was designed to look like Latin, so it did happen).
Whatever it is, Latin-abbreviation concepts may have influenced the underlying design. In Latin, a single shape has several interpretations. Most statistical studies compare Voynichese to unabbreviated Latin. This affects entropy values.
.
I’m not making any assumptions about what the text is (although I am hoping it relates to the diagrams), but clearly the strongest influence on the shapes is Latin letters and abbreviations.
I think there is a matter of confusion, though it has been addressed previously. I am not saying that there was no indirect influence on some of the Voynich characters from Latin or Latin abbreviations. What I am saying is that I believe that the direct influence was from the diplomatic cipher alphabets, which to some extent were influenced by latin and probably latin abbreviations. However I believe the idea that Latin abbreviations and greek etc. alone were sufficient to explain the characters I disagree with, which seems to be JKP’s thesis.
Mark wrote: “However I believe the idea that Latin abbreviations and greek etc. alone were sufficient to explain the characters I disagree with, which seems to be JKP’s thesis.”
I have never said that. I don’t know where you get that.
I have never said that Latin and Greek alone were sufficient to account for all the characters. You’re the one who keeps rewording my views as though there is an all-or-nothing requirement. I have said “most” and “almost” until I’m tired of typing them.
Upthread, I said that a small percentage (very small) may have origins in astrology, alchemy, or kabbalah. But even this does not materially change the picture, for two reasons… 1) some of them appear only once in the VMS and do not appear in the labels or the main text. 2) these symbols were known to Latin scribes and included in Latin documents. They are, in a sense, like “loan words”.
.
I do not like blanket statements because there are things we cannot know until we figure out the system for the VMS text. For example, we cannot know
if EVA-ell is inspired by a number or an abbreviation, since it was used for both in Latin documents.
if EVA-d is inspired by a letter or a number, since it was used for both in Latin documents.
We cannot know if EVA-k is from kabbalah or from Latin. It was used in kabbalistic invocations but also commonly used in Latin documents and it’s drawn exactly the same way.
What we can know is that a majority of the VMS glyphs have morphological and positional analogs with Latin characters and abbreviations. It’s easy to prove. Just count them. Unfortunately, to do that, you have to have a good grasp of Cappelli so you can distinguish between a line or loop that is meaningful and one that is an embellishment.
J.K.P.: in my opinion, it’s a difficult balancing act, linking Voynich glyphs with Latin abbreviation shapes.
On the one hand, Capelli lists so many shapes (and from so many eras and phases of Latin usage) while Voynichese uses so few shapes in a systematic way that the asymmetry between the two systems is acutely uncomfortable. From this alone, I find it hard to accept that Voynichese maps between the two in any useful sense.
On the other hand, the way that EVA y has such sharply differing word-initial and word-final behaviour is uncannily reminiscent of the way the Tironian nota 9 was used. This implies to me that the writer was intensely aware of Latin abbreviations, even if Voynichese probably isn’t directly mapping them.
I’m not sure that comprises a middle position, but I thought I’d mention it anyway. 🙂
Basic:
I’m looking for the system of encryption.
I give the symbols a number.
I give the number a value ( letter )
I am looking for the language
It does not matter what the symbols look like. Symbols are chosen by people but have nothing to do with the system.
It’s him, you look for the person. The more agreement, the more probable the person or close contact.
Caesar encryption, Templar code, disk encryption, etc. They’re all different techniques, but they all have something in common.
The same applies to the VM. After so many people have tried to crack it in different ways and failed, the question arises ” what is the problem ”
How was it around 1400? Language? Scripture ? Gramatics?
Alone with over 100 Latin spellings, and someone who knows no grammar and writes everything the way he hears it, and uses expressions where today no one knows more, I also need a completely different approach in decoding.
100 years of VM and not a step further….think about it.
Peter M.: you’re using the argument the late Stephen Bax used to wheel out – that because we have made no progress in breaking its cipher in a century, it therefore cannot be a cipher.
Even a moment’s thought should reveal that the claim implicitly relies on a large number of fairly toxic presumptions (e.g. that the people who have looked at were blinkered cipher specialists rather than linguists, that we have made no progress, etc), all of which are completely false.
Replacing letters with numbers is called gematria, but this only started to be used in a Christian / non-Hebrew context many decades after the Voynich’s radiocarbon dating range.
JKP: So to clarify are you saying that you believe or not that the diplomatic cipher alphabets were the direct source of the Voynich symbols or not?
I wrote “Latin abbreviations and greek etc.”; note the “etc.” I know you normally include Mathematical symbols, Astrological symbols and arabic numerals to this list.
As an aside don’t you think there is a case for saying that symbols derived for the purpose of writing a cipher are a more likely direct source than symbols that were used generally for all sorts of other purposes?
Also you say that scribes would snicker at someone creating the unusual Voynich symbols. By contrast cipher secretaries would find the creation of Voynich style symbols completely normal and not laughable.
To reiterate I think the visual similarities between the Voynich symbols and Latin abbreviations are important and from your blog it is clear that you do not show that visual similarity.
@ Nick
I didn’t know Stephen Bax said anything like that. That was news to me.
I’m not saying it’s not a skiff, on the contrary, I’m convinced it is.
I think the language is the problem, not the code.
For decryption I replaced symbols with numbers, it’s easier for the PC and it removes the confusion.
Number encryption is not what I mean.
According to my experience with different systems and in comparison to the VM, the VM is rather a simple kind of system. A combination.
I have made my experiences with gramatics.
Example: I know someone who emigrated to Turkey after the 4 school year. Today she still writes in German, but to understand it it is more a listening than a reading.
That although she once learned German, but not long enough.
That’s exactly how I imagine the VM problem today
Nick wrote: “J.K.P.: in my opinion, it’s a difficult balancing act, linking Voynich glyphs with Latin abbreviation shapes.
On the one hand, Capelli lists so many shapes (and from so many eras and phases of Latin usage) while Voynichese uses so few shapes in a systematic way that the asymmetry between the two systems is acutely uncomfortable. From this alone, I find it hard to accept that Voynichese maps between the two in any useful sense.”
I don’t think it’s a difficult balancing act. It’s really not an obscure connection.
One thing that Cappelli doesn’t really make obvious but which becomes immensely clear when you read a lot of medieval manuscripts is there is a ratinale to the shapes. It may look like pages and pages of shapes, when, in fact, it can be distilled down to 20 or so general principles AND most scribes choose a dozen or so of the most common abbreviations and stick with those throughout the manuscript AND the ones that are most common have a high correspondence to EVA-y, EVA-k, the three versions of EVA-m (each one means something different in Latin and each one occurs in the VMS), EVA-ch, EVA-r, EVA-s, the macrons on EVA-q, reverse-c and the tails on daiin. I even have samples of EVA-sh (but it is not as common).
“On the other hand, the way that EVA y has such sharply differing word-initial and word-final behaviour is uncannily reminiscent of the way the Tironian nota 9 was used. This implies to me that the writer was intensely aware of Latin abbreviations, even if Voynichese probably isn’t directly mapping them.”
EVA-y is so common in Latin-alphabet manuscripts (primarily in the final position, sometimes first position and very occasionally within words) that I have found pen tests, alphabets, where the scribes include EVA-y in the alphabet position after “z”. In fact, if you look closely at the column text at the bottom-right of 1r, you will find a shape like a fancy EVA-y (it was occasionally drawn with a snaky tail) after the barely visible “z”. Whoever wrote the column text also thought of EVA-y as belonging with the alphabet.
EVA-y is positioned the same way and drawn the same way in the VMS as it is in Latin-script texts. Not only is it positioned the same way and drawn the same way BUT, in Latin manuscripts it is sometimes drawn in-line, sometimes superscripted AND in the VMS this also happens (something researchers don’t see if they rely on transcripts rather than looking at the original text). Also, the correspondence between Latin-abbreviation shapes and shapes in the VMS that stand alone (as single characters) is also high. That goes beyond coincidence.
Mark Knowles wrote: “JKP: So to clarify are you saying that you believe or not that the diplomatic cipher alphabets were the direct source of the Voynich symbols or not?”
If you look at my original messages about this, what I said is that you cannot LOGICALLY say one is derived from the other because there is the possibility that BOTH the diplomatic ciphers and the VMS may have gotten ideas from Latin.
The diplomatic codes are really rather obvious and mundane in their construction. One-to-many, many-to-one substitution. That’s all. The VMS is neither mundane nor obvious. Whatever it is (nonsense?, systematic text?, natural language?), it’s clever.
“As an aside don’t you think there is a case for saying that symbols derived for the purpose of writing a cipher are a more likely direct source than symbols that were used generally for all sorts of other purposes?”
We don’t know if the VMS is a cipher. It’s certainly not a typical substitution cipher and the structure of it is about as different from the structure of the diplomatic ciphers as is possible.
Was the VMS writer aware of the super secret diplomatic ciphers? I don’t know. Usually only trusted people in inner circles saw those, so the VMS designer would have to be someone with high social stature and special knowledge. What I do know is that the VMS characters could have been derived from Latin script without knowledge of diplomatic ciphers, so if there is a connection, it’s not a necessary one.
“Also you say that scribes would snicker at someone creating the unusual Voynich symbols. By contrast cipher secretaries would find the creation of Voynich style symbols completely normal and not laughable.”
I don’t think cipher secretaries would snicker at the shapes. I think it would pique their curiosity.
.
I think a study of the diplomatic ciphers is valuable and highly interesting in its own right but I think it is dangerous to assume too soon that there is necessarily a connection between them and the VMS. If there is, it should come out of the data, not out of a preceding theory. Even if there is a connection (in terms of shapes), I don’t think it the diplomatic ciphers had any influence on the way the VMS glyphs are arranged.
Nick: Quite obviously and unapologetically, the current discussions re relative merits of possible archaic language input into make up of the VM language is beyond my feeble attempts at comprehension. However, I couldn’t help but pick up on your seemingly unrelated reiteration of apparent faith in the admittedly flawed ‘initial’ assertions re carbon dating results; these being outlined in the ‘verbal’ press release of late 2009. In spite of the almost instananeous retraction by the good folk from University of Arizona’s research team’s claim of a date match between the (uncommonly small) parchment sheets and the text ink components, we’re yet to hear how such a blunder might have occurred. I may have to stand corrected on maybe missing the awaited excuses. That said and all, I’d still like some sort of explanation; now that the ink has had time to dry.
JKP: Obviously the diplomatic cipher alphabets get some influence from Latin as they contain some latin letters. The extent to which latin abbreviations influenced cipher alphabets is up for debate. I imagine they probably had some influence, but it may well have been very slight.
For me the question of the direct source of Voynich characters is the important one. As I have said, I believe we can trace the Latin alphabet back to the Greek alphabet and that further back to Minoan alphabets and I would imagine they were originally influenced by Cuniform. The Latin abbreviations have their origins and their growth and evolution, everything has an origin. However saying for example that the Voynich characters were derived from Cuniform is less than helpful, though technically in a sense probably true.
So to repeat I am concerned with the direct origin of the Voynich characters, this is the important question.
Now we return to my discussion with Rene and Charlotte about whether Voynichese was influenced by diplomatic ciphers.
We don’t know how clever the Voynich cipher really is. I think we all would like the end solution to be elegant and interesting, but it may not be that.
The questions that I posed to you previously regarding the commonalities between the Voynich and diplomatic ciphers illustrate my point as most of them you said that we don’t know the answer to. There are many different takes on the principles of diplomatic ciphers, some of which I think could be largely or wholly consistent with the Voynich. It is possible that the Voynich took no influence from these advanced substitution ciphers, but I have my doubts. I have my own ideas about how you could implement a diplomatic cipher in a superficially very different way, so that the end result appears to be largely very different from the standard diplomatic ciphers and yet the method is essentially almost the same. Now it is possible that it is more different than that, but I doubt that those who say that diplomatic ciphers and the Voynich cipher are completely different really have the evidential basis to say so, having read up on the different evidence.
You say “If there is, it should come out of the data, not out of a preceding theory.” Not necessarily there are potentially many ways at arriving at answers, the data is only one.
Bee Mark. VM- 408 is not Latin. They’re not even shortcuts.
Mark, so the letters are numbers. All characters are numbers.
The author ( Eliška z Rožmberka ) writes in the manuscript :
I use the Jewish code. And I write in Czech.
This is how Elizabeth of Rosenberg writes in this manuscript. 🙂
( Mark . I’we already written you about side f1v. You could understand what I wrote to you ? )
Mark Knowles wrote: “Obviously the diplomatic cipher alphabets get some influence from Latin as they contain some latin letters. The extent to which latin abbreviations influenced cipher alphabets is up for debate. I imagine they probably had some influence, but it may well have been very slight.”
Mark, they don’t just contain Latin letters, they contain Latin abbreviations. Lots of them.
These shapes: EVA-y, EVA-s, EVA-r, EVA-y, EVA-k are not Latin letters, they are common Latin abbreviations. In Latin, a tail is a connected macron. A macron is a line (an abbreviation symbol) placed over the letter.
The macron was mostly written above the letter, but for quick writing, if the missing letters were near the end of the word (or attached to a single letter, as with EVA-s and EVA-r), then the macron is connected to become a tail and swoops back. Not every scribe connected it to make a tail, but many did. It was a matter of choice.
There were three distinct forms of tails in Latin and there are three distinct forms of tails in the VMS, just as there are three forms of EVA-m in Latin and three forms of EVA-m in the VMS. You can’t see this by looking at a transcript.
Similarly, the “-is” abbreviation is a loop with a descender. One sees it on EVA-m and EVA-k. It has several meanings but when it is drawn like EVA-m, it is primarily at the ends of words and the ends of lines/paragraphs in Latin and is primarily at the ends of words and the ends of lines/paragraphs in the VMS.
.
Latin abbreviation shapes are found in the VMS and they are found in the diplomatic ciphers, so I would say that Latin abbreviations (not just the letters) had just as much influence on diplomatic ciphers as did Latin letters.
Mark Knowles wrote: “We don’t know how clever the Voynich cipher really is. I think we all would like the end solution to be elegant and interesting, but it may not be that.”
It is clever in structure (if there is any content, it’s likely to be mundane rather than clever—I’m not hoping for elegant content, just for SOME content).
There are many common tokens and repeating patterns and yet there is just enough sophistication to the way the whole thing is put together to say, “There might be something here.” It goes beyond what one can readily discern as nonsense text.
Mark Knowles wrote: “The questions that I posed to you previously regarding the commonalities between the Voynich and diplomatic ciphers illustrate my point as most of them you said that we don’t know the answer to.”
Mark, you are doing it again. When I say “most”, you turn it into “all”. When I say “some”, you turn it into “most”.
Yes, there are unanswered questions. Of course there are. But both the diplomatic ciphers and the VMS include numerous Latin scribal letters and abbreviations, so we cannot assume without direct evidence (which hasn’t been found so far), that there is any connection between the ciphers and the VMS.
If Latin is called Source L, they may both have gotten their ideas from Source L rather than from each other.
I note that the German translation of Cappelli (J. J. Webers illustrierte Handbucher. Leipzig, 1901 and 1928) is said to include two thousand more abbreviations than are listed in Cappelli 1929. See: https://kuscholarworks.ku.edu/bitstream/handle/1808/1821/47cappelli.pdf
There might be something of interest in those additional abbreviations,
I also note that there was a French precursor to Cappelli.
“Dictionnaire des abreviations latines et franchises usitees dans les inscriptions lapidaires et metalliques, les manuscrits et les chartes du Moyen age, 5th ed., Paris, 1884. reprinted by Olms. Hildesheim, 1965.”
‘Lapidaires et métalliques?” It sound interesting to me. I assume that Lapidaires means lapidaries (the people who work at lapidary)? Can the word métalliques be translate as metal worker, or metallurgist?
J.K. Peterson, I note that you recently mentioned that the use of abbreviations would tend to change the entropy and that ‘Most statistical studies compare Voynichese to unabbreviated Latin.” My question is, have any entropy measurements been conducted on latin text that is largely abbreviated? I can’t find any such studies but I am a total amateur in this area and I may not have been looking in the right place.
Byron Deveson: texts using scribal shorthand (i.e. using contraction and abbreviation rather than Tironian notae) are difficult to read and are relatively little studied – once someone has struggled to transliterate and reconstitute what the text was supposed to say, few historians have much interest in the original shorthand version. I’ve been searching for more than a decade to find good texts on the subject of fifteenth century Italian scribal shorthands (let alone any entropy studies on them) but have to date found only fragmentary pieces and secondary mentions. 🙁
Byron Deveson: “inscriptions lapidaires et metalliques” would seem to me to mean nothing more complex than “inscriptions on stone or on metal”, e.g. epigraphy.
The manuscripts that are heavily abbreviated are indeed hard to read, but they are in the minority. Most scribes used the most common abbreviations and they are not difficult to read. You just have to know three or four languages and the abbreviation basics.
It’s a busy time of year for me work-wise, but if I can find time next weekend, I’ll post some examples with translations.
JKP: I asked you the following questions->
“Does the Voynich contains any null characters?”
You said we don’t know.
“Are there any symbols which all correspond to the identical character in the Voynich?”
You did not understand my question as it concerned homophones.
“Are there any symbols which correspond to whole words, maybe rare characters?”
You did not understand my question as it concerned repertoire.
“Do any characters, if not all characters, map to a given letter of the alphabet?”
You said we don’t know.
“Do any characters correspond to a couple of letters say for example “ap”? (Much less likely, but possible).”
You did not understand my question as it concerned wherher there are symbols that map to letter pairs.
Are multiple languages used together as Cicco Simonetta suggests?
You doubt it, though I doubt you had much of a basis for saying so.
There are quite a few questions I could pose refering to similarities and differences to diplomatic ciphers.
JKP: My understanding is that where you believe symbols in the Voynich correspond to those used in the Latin abbreviations you do not believe those symbols mean those actual abbreviations. My understanding is that you believe that the author used symbols from Latin abbreviations, but that they have a completely different meaning in the Voynich. Tell me if that is not your position.
My understanding is that you believe those symbols are used in a way, because of their meaning, which is quite different from how they are used by all other Latin scribes. Obviously that sounds rather strange, but the Voynich is rather strange so maybe that is OK. It is worth noting that with diplomatic cipher alphabets it is normal to assign a meaning to a symbol, one does not need to subvert any more standard usage where such a usage exists.
So largely it seems that we are most likely looking for the direct source of the symbols, so what Latin abbreviations correspond to in other contexts is not important as you believe that they do not have that meaning.
I believe the most important question is visual similarity not supposed usage as it seems you reject the standard usage of Latin abbreviations anyway.
The extent to which diplomatic cipher alphabets were influenced by Latin abbreviations or not, is not particularly important, but as I have said the direct source of Voynich symbols is not what I am interested in. And I reiterate that in my opinion your blog does a poor job of demonstrating that visual similarity.
Josef Zlatodej Prof: If the 408 characters are in reality mere numbers, would it be then conceivably possible to look into old George Boole’s algebraic set principles, as a simple means, for we, the mostly unimaginative ants, to translate the VM text.
John Sanders:
It could be that Josef Zlatodej Prof is on to something.
Maybe the Voynich was written by an ant colony. Ants leave pheromone trails that influence how other ants behave when they interact with them. So the complex behave of an colony may have produced the Voynich.
We have all been assuming that the Voynich was written by a human or humans if it was in fact a product of the animal kingdom that would be fascinating.
Mark Knowles wrote: “My understanding is that where you believe symbols in the Voynich correspond to those used in the Latin abbreviations you do not believe those symbols mean those actual abbreviations.”
That is correct. There is correspondence between VMS glyphs and some of the common Latin abbreviation symbols. However, they do not appear to have the same meaning as in Latin. I do believe the VMS designer and probably the scribes had a good awareness of Latin scribal conventions.
“My understanding is that you believe that the author used symbols from Latin abbreviations, but that they have a completely different meaning in the Voynich. Tell me if that is not your position.”
I would not go so far as to say “completely different”. I believe they do not have the same specific expansions as Latin languages, HOWEVER, it is possible that they have similar functions.
For example, in Latin, r with a tail is an abbreviation with certain specific meanings. In Voynichese, it may not have the same specific meaning, but the tail nevertheless MIGHT signify that it is an abbreviation. Voynich tokens are shorter than many languages. This also suggests the possibility that it might be abbreviated text.
Let’s say you wanted to write an Indian language with Latin characters (or something similar to Latin characters to accommodate different sounds). Latin has certain common endings, some of which are indicated with a character shaped like EVA-y. In the Indian language, there would be different common endings, but the EVA-y character could still be adapted to express them. Missionaries frequently invented alphabets using Latin or Latinesque characters so local populations without a written language could be taught to read and write. When they did so, they brought western concepts of shapes and abbreviations with them.
Hildegard von Bingen, as another example, invented her own “common ending” to signify plants and shrubs. This ending was invented, but is, in fact, based on a very common Latin ending, one that she co-opted and gave a different meaning.
So, we have at least two medieval circumstances in which Latin ideas underlay the invention of new forms of communication.
“I believe the most important question is visual similarity not supposed usage as it seems you reject the standard usage of Latin abbreviations anyway.”
That’s not quite how I think about it. I don’t think visual similarity is the most important question. I believe that visual similarity MIGHT indicate that other aspects of Latin conventions were used as well.
I question only the specific letters that abbreviations expand into (at least for now). I do not reject the possibility that similar shapes might have a similar function—shapes that are abbreviations in Latin MIGHT also represent abbreviations (whether it be letters, numbers, or symbols) in the VMS.
John + Mark.
I will write to you again. 🙂
Manuscript page f1v.
One letter ( leaves) of the plant is written in several letters.
———————————————————–
Eliška writes in the manuscript :
The green leaves are 14. And the golden leaves are six and six.
And that’s my birthday.
———————————————————–
Eliška tells you when she was born. 🙂 This is the year 1466.
Every ant looks at the letters that are on the leaves.
Jew substitution : number 1 = a,i,j,q,y. Number 4 = d,m,t.
Leaves are written on letters : J + T. ( J = 1. T = 4 ) 🙂
Can you understand that at least ??
JKP: I do think visual similarity is the most important question in this situation. So there we differ.
You seem to be confused as to whether Voynichese is a cipher or abbreviated language. It seems the implication of what you say is that it is an abbreviated language. However you seem to think the abbreviations are abbreviations for different things. So for example as parallels in English LOL, IOU, USA, ETC …. would have different meanings, but still be abbreviations. So it would seem that you are essentially implying the language is not Latin as it would be completely bizarre to have altered the meaning of all the abbreviations. If it is just an abbreviated known language it should be easy to crack. The idea that it is an unknown abbreviated language is a step too far for me. If it is not an abbreviated language then it does not seem very consistent with your logic as all your arguments about the function of the characters would be lost. If it is not Latin I suppose it would be like you saying that LOL, IOU, USA, ETC are abbreviations for different words in a language. I suppose there are so many different Latin abbreviations that within that framework anything is possible.
Obviously I am in the cipher school of thought, which seems to be fairly inconsistent with your theory as it would be odd for someone to grab abbreviations and put them in a cipher.
Generally I find your theory quite contrived and I do find your blog examples unpersusasive. Your theory just seems at odds with itself. It feels like you are saying it behaves like abbreviated language, but it is not an abbreviated language.
JKP: It is interesting as in a sense I think we are at different ends of the spectrum.
Obviously, an abbreviation is a shorten form of a word. So the implication of your theory seems to be that the text is more compact.
My current attitude is quite the opposite I think the text is verbose.
So from your theory the real text is much longer than it is in my thinking.
Mark Knowles wrote: “You seem to be confused as to whether Voynichese is a cipher or abbreviated language. It seems the implication of what you say is that it is an abbreviated language. However you seem to think the abbreviations are abbreviations for different things.”
Mark, stop trying to shoehorn my thinking into theories and shrinkwrapped packages.
I do not problem-solve that way.
.
I am NOT “confused” about whether it is a cipher or abbreviated language. I do not KNOW if it is a cipher or abbreviated language (and neither do you), those are only two of several possibilities. I MAKE NO ASSUMPTIONS ABOUT what it is (let me repeat that… I MAKE NO ASSUMPTIONS ABOUT it). Eventually the DATA will indicate what it is.
The abbreviation shapes in the VMS are consistent with common abbreviation shapes in Latin languages. In other words, the VMS MIGHT include abbreviations. Notice I use the word MIGHT because it is only ONE of several possibilities (one that needs to be investigated).
There is a big difference between
1) developing theories and then trying to line them up with the evidence (which, in my opinion is bass ackwards) and
2) brainstorming, writing out an open-ended list of possibilities, and then carefully researching each one until the picture emerges FROM THE DATA. There is not enough evidence to form a theory about WHAT the text is (cipher, synthetic language, symbolic text, natural language, etc.). It is not shaping up well as natural language but… that is only when glyphs are studied as individual units.
.
I am working my way through the possibilities. This takes time. People don’t like the way I do things. They want quick answers, quick solutions, absolutes. Some of them want recognition and race to send out news releases before they have even checked their own work to see if it holds up to scrutiny! They don’t have the patience for open-ended questions and slow methodical research. Tough. Science rarely serves up quick answers.
JKP:
I am not trying to shoehorn your thinking, but rather to explore the likely implications of your thinking. In a sense you seem to shoehorn your thinking into the notion that the origin of the Voynich characters are to be found largely amongst the Latin abbreviations. In a sense you should, given your perspective, say that Latin abbreviations MIGHT be the source of SOME of the Voynich characters, so as to avoid confining your theory to shrink wrapped packages. It seems that problem solving by exploring the likely implications of an idea to see if they seem somewhat contradictory to that original idea, is a valid approach.
Both you and I do not KNOW many things about the Voynich. You do not KNOW the Latin abbreviations are the primary source of Voynich characters. This is an assumption that you make though you say that you MAKE NO ASSUMPTIONS ABOUT it.
MARK, YOU DID IT AGAIN. You restated my views and changed what I said.
I make no assumptions about the CONTENT of the text. The CONTENT is different from the shapes. Why do I have to keep repeating that?
Shapes are one line of investigation that may or may not help sort out the content. I DO know that there is a very high proportion of VMS glyphs that match the common Latin letters and abbreviations, far beyond what can be explained by coincidence. It took me 10 years to investigate this.
You have not studied it at all and yet you seem to think your opinion is of equal weight. Well that’s fine, you are entitled to your opinion, but I wish you would learn to separate issues like content and shape and to give up trying to generalize everything into absolutes, or we will always be talking at cross-purposes.
JKP: I am sorry that you feel that I am restating your views and changing what you said. I merely trying to refer to what I understand what your views are. If there is a lack of clarity on your part that is not something that I can deal with. I have carefully read what you have written on this subject and I find your argument unpersuasive.
I have long said that the visual similarity of the shapes is the key issue. And on that front I feel your theory is pretty weak, having read your blog.
I have not addressed the CONTENT of the manuscript here at all, so I can only guess that you are using the word CONTENT in a completely different way to mean whether it contains abbreviations or not. In some instances you may not have stated something, but I can still say that that statement appears to be a consequence of something that you have stated; that is not misrepresenting you.
I am sure that there are people who have studied the Voynich for more than 10 years who have come to different conclusions, so the number of years that you have studied it is unimportant.
You say: “You have not studied it at all and yet you seem to think your opinion is of equal weight.” Now we are back to appeal to authority not evidence and argument. I have read your blog posts on this subject and don’t find your evidence and arguments persuasive; you telling me that you believe that you are an expert on the subject is going to make no difference.
I have made an effort to try and understand your position, but you seem to make absolute statements then when challenged retreat to vague statements and accuse me of misrepresenting you by attributing absolute statements to you.
You make the absolute statement: “I DO know that there is a very high proportion of VMS glyphs that match the common Latin letters and abbreviations, far beyond what can be explained by coincidence.” I haven’t seen you provide the evidence for that absolute statement.
I don’t think writing in block capitals reinforces your argument.
Mark Knowles wrote: “If there is a lack of clarity on your part that is not something that I can deal with. “
Nice try. If you take my word “most” and change it to “all” I don’t think that is because of a lack of clarity on my part.
If you cannot comprehend Cappelli, that is not a lack of clarity on my part either.
“I have not addressed the CONTENT of the manuscript here at all, so I can only guess that you are using the word CONTENT in a completely different way to mean whether it contains abbreviations or not.”
That is not what I mean by content. Here I may not have been clear enough. When I say the CONTENT of the text, I am talking about the underlying meaning (if there is any). For example, if Voynichese were natural language, then the content would be words, phrases, perhaps sentences. If it is numbers, then the content may be coordinates, or a two-step code, or kabbalah, or something else.
The content and glyph shapes may or may not be related. If the glyphs were Chinese or Arabic instead of Latin, then it might be prudent to look for Asian or Middle Eastern influences first. But they are not, and even though I can read a couple of Asian languages, I have not found any evidence of Asian influence on the text, so far (yes, I have looked and continue to look).
“I have long said that the visual similarity of the shapes is the key issue. And on that front I feel your theory is pretty weak, having read your blog.”
Mark, is it a theory that the glyphs “a” and “o” and “r” and “i” and “c” resemble Latin letters?
If your answer is no, then the glyphs that resemble Latin abbreviation shapes match just as well. If I showed EVA-m to someone who can read medieval manuscripts and said ris/cis/tis, that person would nod like it was obvious. If I pointed at EVA-y and said con/com/us/um, that person would nod and shrug as if it were obvious. If I showed them EVA-r and the three different tails, they would nod in recognition. No other conversation would be necessary. That’s not a theory, it’s knowledge of medieval conventions.
Nick,
In my research re: Piacenza I came upon a manuscript that appears to document choreography and/or accompanying music. Some aspects of the manuscript vaguely reminded me of the VM so I thought that the experts should be made aware of it.
La Republica.it Birmingham, scoperto il Corano piu antico al mondo: risale a 1400 anni fa.
http://www.repubblica.it/esteri/2015/07/22/foto/corano_antico_mondo-119597615/1/#1
https://ais.ku.edu.tr/course/19594/Neoplatonic%20and%20Pythagorean%20Notions%20of%20World%20Harmony%20and%20Unity%20and%20Their%20Influence%20on%20Renaissance%20Dance%20Theory.pdf
“NEOPLATONIC AND PYTHAGOREAN NOTIONS OF WORLD HARMONY
The Concept of Harmony and Unity in Renaissance Dance Manuals In the earlier part of the 15th century, a school of Italian Renaissance dance came into existence which was centred at the
Sforza court in Milan, and has therefore been called il ballare lombardo. The foundations of this Lombard School of Dancing were laid by Domenico da Piacenza (or da Ferrara) and developed
further by Antonio Cornazano and Guglielmo Ebreo (who also called himself Giovanni Ambrosio). Together they produced a system of dancing which remained the Italian form of court dance until the end of the 16th century. They wrote three dance manuals,which not only contained the main repertoire of Renaissance court dances, but also a body of theory, which has interesting ramifications with the aesthetics and philosophy of the period.
Domenico was born in Piacenza at the end of the 14th century and worked at the court of Leonello d’Este until 1450. Sometime after this date he transferred to Milan where he arranged, in 1455,
together with his assistants Guglielmo and Cornazano, the festive balls for the engagement of Ippolita Sforza with Alfonso of Calabria. Domenico returned to Ferrara in 1456 and continued to
work there until his death, of which we have no exact date. But Cornazano stayed at the Sforza court until 1465, establishing his master’s system of dancing in Lombardy and taking it, like
Guglielmo, to other Italian courts.”
I note that “….Domenico’s teaching was laid down in a manuscript which was probably produced for the Ducal Library in 1455. The dances included in it were created over some twenty years ….”
So, the VM could be an earlier (pre 1435) workbook or an earlier work that Domenico collected?
JKP: If I have ever confused “all” and “most” in your writing then I apologise, though I very much doubt that I have made that mistake as often as you claim.
Understanding Capelli is the easy bit linking it to the Voynich is the hard part.
It is not a theory that the characters “a” and “o” and “r” and “i” and “c” resemble Latin letters, but it is a theory that most of the other Voynich symbols resemble Latin abbreviations.
Knowledge of medieval conventions does not mean that they match with the Voynich. In the same way that knowledge of Medieval Turkish does not mean one can say that the Voynich characters are from Medieval Turkish. You need to be able to provide a self-contained argument rather than appeal to your claimed expertise or anyone else’s claimed opinion if you wish to persuade. Otherwise you are essentially saying trust me I am an expert therefore my theory is true, I am sorry but that won’t wash with me. I have read your blog posts relating to this question and I have read your comments and that is what I have based my opinion of your theory on. If there are stronger arguments for your theory than those contained in your blog then I would advise you when you have time to document them in your blog; I cannot comment on arguments not presented.
JKP: Anyway this discussion has been going round and round in circles and I really don’t want to waste any more time on it now, so I hope we can shelve it for the time being.
Mark, it goes in circles because you keep dismissing me, even when I provide you with helpful information.
We got into this again because I (helpfully) mentioned that the “smoke” symbol is a common Latin abbreviation. You blew it off instead of taking me seriously.
Here are visual examples (it also includes an example of EVA-sh):
https://voynichportal.com/wp-content/uploads/2018/07/SquiggleApostrophe.png
It is a Koran. I don’t know why I thought it was a manuscript dealing with dance!
JKP:
It is true that I consistently have argued that your theory is not correct. I suppose one could say that by doing so I am providing you with helpful information which you are dismissing.
You have not demonstrated that the “smoke” is a common Latin abbreviation. In fact if one looks at your examples it is clear that they are not the same. The apostrophes that you show are like little “s” ‘s above a symbol. This is quite different from what we see with the smoke. Also the V shape with the “s” above it is not like the V shape in the Voynich.
However it is clear you are trying to build on your argument which is a good thing. My helpful advice to you would be devote more of your time developing your argument on your blog rather than getting embroiled in pointless discussions with me. When you have written another post on this subject you can refer me to your link, nothing more is needed. You can expand and develop your argument rather than wasting either of our times by trying to provide me with what you believe is “helpful information” and I believe is a pointless distraction.
I really don’t want to waste more time on this, no matter how “helpful” you imagine you are being by propagating your theory to me.
I have to come back to the disputed conclusion that the Voynich cipher is most likely what I would call an atypical diplomatic cipher. By the use of the word “atypical” I think I mean an unusual implementation of a diplomatic cipher, not a fundamentally different kind of cipher.
So for example one could have a scenario where a series of Voynich characters correspond to an individual letter or multiple letters of the latin alphabet, such as:
“%&€” = “z”
“%€@” = “z”
“#$&&” = “ap”
“@£@” = “the”
“#%$&@” = NULL
“!*$££” = NULL
Now we don’t see any examples of diplomatic ciphers where multiple separate cipher characters are used this way, so this would be one kind of atypical diplomatic cipher. But there could be other atypical implementations of diplomatic ciphers.
I ask myself what the foundation or kernel of the Voynich cipher is and it seems it must be some kind of substitution cipher. This opinion is less based of the specifics rather on what I am inclined to think is the unlikelihood that the author invented a completely original method of encipherment. That is not to say that it isn’t possible that the author did incorporate some small original feature into his/her atypical diplomatic cipher, such as rules pertaining to character ordering.
I am still inclined to the view that what has been termed single-word labelese is a good point of attack due to it being the Voynich cipher at its simplest. I don’t know if the complete single-word labelese text is easily obtainable for the purpose of performing statistical analysis or there is likely to be quite a bit of work manually identifying each case of specific text as being a label.
The diplomatic ciphers are substitution systems with varying numbers of glyphs corresponding to varying numbers of Latin letters. Each one is a different set.
Your example is also a substitution system with varying numbers of glyphs corresponding to varying numbers of Latin letters… so what is it about your scenario that you consider “atypical” or as an “unusual implementation” in the context of diplomatic ciphers?
Mark wrote: “The apostrophes that you show are like little “s” ‘s above a symbol. This is quite different from what we see with the smoke. Also the V shape with the “s” above it is not like the V shape in the Voynich.”
In Latin, there are about 12 variations on the “smoke” apostrophe (all of which are recognizable to scribes). The fancier the letter they are associated with, often the fancier the apostrophe (in other words, ones that are a little more embellished will sometimes be added to capital letters —I mostly sample lowercase examples because they are the most common). Some scribes even write them horizontally rather than vertically, but they mean the same thing.
Also, the smoke apostrophe can be associated with any letter, not just “v”. It just happens to be common with v (as in “ver” and “vir”). I’ve never said the VMS seagull shape was a “v”, only that the VMS glyph is positioned above and within it in the same manner as is common to Latin scribal conventions. Also, there is a more typical lowercase-style “smoke” apostrophe later in the VMS.
These observations are not based on a single example here or there, they are based on studying all the VMS text and all the apostrophe shapes, which are moderately frequent in the VMS (most transcripts treat them as though they don’t exist) and which are all morphologically analogous to common and conventional Latin apostrophes.
JKP: What makes it “atypical”, “non-standard” or an “unusual” implementation based on my experience of 15th Century diplomatic ciphers is that there are no examples that I am aware of, like this. Normally one character, though sometimes a complex characters, is substituted not a series of characters. So for example I have never seen a whole word in a diplomatic cipher alphabet that corresponds to a null; only individual characters. To me it is very plausible that some of the frequent and repeated Voynich words are in fact nulls. Similarly it seems plausible that series of characters or even words map to a specific character, word or sequences of characters. So fundamentally this would be the same as a normal diplomatic cipher, but implemented in an unusual way, that’s why I have termed it an atypical diplomatic cipher. But there may well be other ways of implementing diplomatic ciphers that could also be called “atypical”.
JKP: I can only go on the evidence that you have provided in your posts on your website. If you are telling me that you have not included the best evidence there then I would advise you to do so. When making an argument you should really present the evidence which most strongly supports that argument rather than tell your reader to trust you that there is even stronger evidence that you have not included.
Generally I have the feeling that in practice I could see you arguing that any character I can conceive of could be found amongst latin abbreviations if you really believed that character were in the Voynich. So I have the sense that you cast your net so wide that you can inevitably find whatever you are looking for there and even amongst your find catch that you present to the reader the similarities can appear weak.
As you know so far so I don’t see a lot of mileage in the whole “latin abbreviations” theory as the basis for Voynich characters.
Nick: Don’t be modest, but is it your opinion that you could devise a non-modern cipher that would be as difficult as the Voynich to crack? I just wonder how difficult it would really have been to produce such a cipher. Maybe it is actually not that hard.
Mark Knowles wrote: “What makes it “atypical”, “non-standard” or an “unusual” implementation based on my experience of 15th Century diplomatic ciphers is that there are no examples that I am aware of, like this. Normally one character, though sometimes a complex characters, is substituted not a series of characters. So for example I have never seen a whole word in a diplomatic cipher alphabet that corresponds to a null; only individual characters.“
Then maybe you need to expand your “experience of 15th Century diplomatic ciphers” because Tranchedino recorded page after page of diplomatic codes that have nulls of 2 or 3 characters and some Romance-language words as long as six characters.
On f31v, we see the Latin words “sic” and “non” used as nulls.
On the following page, the words “dire”, “fare”, “nienta”, and “questro” are nulls.
There are also 2- and 3-character Greek words used as nulls, and numerous common 2- and 3-character Latin abbreviations such as “per”, “cis” and “qui” written as two characters (the same way we might write “lb” for “pound”).
Roman numeral sequences of two or three characters were also used as nulls.
Of the 14 nulls on folio 5v, 8 are 2 or 3 characters, more than 50%.
On folio 71r, all of the 10 nulls are 2 or 3 characters, and five of them are common syllables.
On folio 157v, 6 out of 7 nulls are 2 or 3 characters, including ion, lon, and gros.
On folio 154v, the very common Latin word “vel” is a null, as is “vol”.
On 151v, “hinc”, “istine”, and “heu” are nulls.
On 150v, “sed”, “quid”, “cur”, “quonam”, “ante”, and “frans” are nulls.
I didn’t even look hard to find these. I randomly opened the book at a few places and there they were.
In the diplomatic ciphers, two- and three-character nulls are more common than single-character nulls and some nulls are words of five or six characters.
Mark, I write my blogs to help those who are unfamiliar with Latin paleography see some of the patterns in the VMS, not to argue some theory.
But no one can learn paleography from reading a few blogs. You have to read manuscripts, lots of them. People take years-long courses of study to be able to read abbreviated Latin and Greek script. Blogs are only an introduction.
.
I am not arguing that any Latin abbreviation character can be found among Latin abbreviations (and thus can be interpreted as existing in the VMS).
In fact, the VMS glyph set only includes the most common Latin abbreviation shapes. In that sense, it’s rather dull (and suspiciously regimented).
JKP: You have not understood what I am saying. clearly I have studied the Tranchedino in great detail, but you don’t see Nulls that are a series of cipher characters, though you can see latin words. So we don’t see what I am referring to amongst the diplomatic cipher alphabets. This should be obvious.
JKP: You do present a theory, as your “theory” is not an objective fact as you pretend. “helping people to see patterns” means trying to persuade people of patterns you believe are there.
I have read your blog posts on the subject and most of the time there is a relatively poor similarity between the Voynich characters and those that you have plucked from amongst latin abbreviations. Simply if they exist you should have plucked better examples not just fall back on some appeal to claimed authority.
It looks likely that we again may be veering towards the kind of pointless discussion that from my perspective you seem far too keen on.
Mark Knowles wrote: “JKP: You have not understood what I am saying. clearly I have studied the Tranchedino in great detail, but you don’t see Nulls that are a series of cipher characters, though you can see latin words. So we don’t see what I am referring to amongst the diplomatic cipher alphabets. This should be obvious.”
Yes, there are nulls in the diplomatic codes that are series of cipher characters. Thousands of them. They are not all words, many are invented characters or combinations of characters that are nonlinguistic.
• There are Latin words that are nulls. There are Greek words. There are Italian words that are nulls.
• There are Latin abbreviations that are nulls, there are nonsense tokens that are nulls, and there are cipher sequences (made-up characters) that are nulls.
I have an entire book of them on my lap. I’m looking right at them.
You must have spoken before you looked or not researched them as carefully as you imply because you said, “I have never seen a whole word in a diplomatic cipher alphabet that corresponds to a null; only individual characters,” and your impression CLEARLY does not reflect reality.
There are many whole words in the diplomatic ciphers that correspond to nulls, both linguistic (in several languages) and nonlinguistic (combinations of cipher characters).
JKP: Without being able to attach images to comments in this blog it is difficult ti show you what I mean. I also have the Tranchedino in front of me as well as other cipher keys.
I would suggest you have a think about what I have written and see if you can honestly understand what I am referring to.
Mark, this statement that you made: “I have never seen a whole word in a diplomatic cipher alphabet that corresponds to a null; only individual characters,” does not reflect the general format of the diplomatic ciphers at all.
They make generous use of multiple characters to correspond to nulls—sometimes whole words, sometimes syllables, sometimes nonsense sequences, sometimes repeating sequences, and sometimes cipher-glyphs (invented shapes). Nulls are in no way restricted to individual characters.
So, it’s pretty hard for me (or anyone else) to reflect on what you mean when reality does not concur with what you said.
JKP: As far as I am aware nobody else has stated that they do not understand what I am referring to.
I will try to make it as simple as possible for you.
Imagine we have a series of invented cipher symbols that correspond to a Null or that correspond an individual letter of the latin alphabet then that is distinct from what we see in the 15th century diplomatic cipher alphabets. We can find individual invented cipher symbols corresponding to Nulls or letters or pairs of letters, but not sequences of invented cipher symbols.
Imagine some of the common repeated words we find in the Voynich and imagine that they do correspond to Nulls; we do not see the likes of this in the diplomatic cipher alphabets, though if they indeed are Nulls, that could certainly fit within fundamentally the same framework of diplomatic cipher alphabets.
If you give me your email I can emailed you an image explaining visually what I mean.
Mark Knowles wrote: “…We can find individual invented cipher symbols corresponding to Nulls or letters or pairs of letters, but not sequences of invented cipher symbols…”
Many of the invented symbols are combined with Latin or Greek letters and abbreviation symbols. If the null is comprised of invented-invented-LatinAbbreviation (for example), do you consider it to be a sequence of invented cipher symbols?
From your statement it sounds like you consider two or more to be a “sequence”. But do you consider it a sequence if they are combined with a second or third character like a recognizable number, letter, or abbreviation?
JKP: Rather than focusing too much on definitions I think it better to get a sense of my perspective.
I am inclined towards the view that there is something in common between the Voynich cipher and diplomatic ciphers. My own theory of authorship points in this direction, but I think there are also other arguments for this.
However it is clear that the Voynich cipher is not identical to the diplomatic ciphers of the time. So I ask myself if there is a way that a diplomatic cipher could have been implemented differently such that it would fit more closely what we see with the Voynich. The example that I have described is one such idea.
Mark Knowles wrote: “However it is clear that the Voynich cipher is not identical to the diplomatic ciphers of the time.”
The Voynich text is dramatically different in structure to the diplomatic ciphers. One has to stretch one’s arms quite far apart to describe them.
The diplomatic ciphers are many-to-one/one-to-many substitution codes with an extremely large glyph-set for each cipher (in comparison, the VMS glyph-set is quite constrained) and, unlike the VMS, they are not highly repetitious or positionally dependent, they are linguistically dependent.
JKP: I think if one considers the kind of implementation of a diplomatic cipher that I suggest it could account for many of the appararent differences.
The presence of null words would explain certain things. Rare characters corresponding to a repertoire as we see in diplomatic ciphers would also fit. Series of characters mapping to individual characters or other series of characters could also explain other features. So some kind of “atypical” diplomatic cipher seems to me a geniune possibility.
As I believe I have said I find the idea that the Voynich cipher is part of a known cipher tradition is more plausible than the idea that a whole new kind of cipher was invented by the author of the Voynich. Whilst the idea that the author invented a unique and sophistated cipher of his/her own is possible and certainly attractive, it would seem less likely.
@Mark
What do you mean by null words. I am a bit confused by the Google Translation.
Peter: By “null” words I means words in the Voynich which have no meaning and can be ignored. The use of “null’ characters is a normal feature of diplomatic ciphers. The repeated words that we find in the Voynich I suggest could be null words.
I return to the question as to why the author of the Voynich wrote the book in a sophisticated cipher whereas nobody else felt the need to. It seems to me that there are a variety of possible reasons:
1) Because he/she could. Clearly the author must have been very adept at ciphers prior to writing the Voynich. I think it unlikely that the author studied ciphers solely for the purpose if writing the Voynich rather I think the author found in the Voynich a convenient excuse for applying his/her skills. I would suspect the author had developed his/her knowledge of ciphers for some time before beginning work on the Voynich. This is I think the most likely explanation.
2) Because the author really believed the contents were so much more important than those of any other contemporary book and therefore he/she must have been to some extent deluded as the contents almost certainly weren’t.
Given that we are not very aware of an unparalleled cipher expert of that time who could have written the Voynich the author can’t have been someone seeking fame for there abilities.
My own candidate for author is relatively obscure which I think is not inconsistent with the authorship reasons hypothesis that I have presented above.
Mark: if there were, Sherlock Holmes -style, only two plausible explanations, we might gain some kind of traction from this kind of approach. In practice, however, it should be possible to construct ten or twenty more explanations that fit the little evidence we have. 🙁
Nick: You are right it is very much speculation, but I think there is a place for speculation. I agree there are probably other hypotheses which could fit. Even if I am correct I am not sure how this theory would help to identify the author. Nevertheless to reiterate I don’t think this should discourage one from speculating as long as one is honest that that is what it is. I am inclined towards the view that it is possible for speculation to be a stepping stone to something more concrete.
@Mark
Maybe you should first focus on where you look.
Maybe it is easier for someone who knows German.
If I look at the words or the 3 sentences in German, and only just with much effort understand what he writes there, even though I speak German. If I compare his spelling with others and with examples from this period, I have to say he writes really shit. No grammatical, no sentence and the sense leaves much to be desired.
Let’s say that German is his mother tongue, and he tries to translate that into Latin, where he may have learned somewhere. Then it will be little shit, a lot of shit.
Think about it. The encryption itself is not the problem, it is quite simple.
If you are talking about the 116v marginalia, Peter, it doesn’t seem likely that the person’s mother tongue is German unless there’s something hindering his or her languages skills. If so, there are numerous possibilities…
• Dyslexia.
• Hearing impairment. Many hearing impairments are partial, not total, and would affect language skills to different degrees. If a person’s mother had measles while pregnant, the child might be born deaf or blind (hearing loss was caused in by other medieval illnesses, as well.
• Lack of formal education. Perhaps it was someone who picked up imperfect, unschooled writing and language skills along the way. Only a small part of the population went to school. Medieval illustrators often could not read or write but rubbed shoulders with scribes who were literate and may have picked up imperfect writing skills.
So, a number of factors could impair language skills in someone whose mother tongue was German.
.
But what if their mother tongue was not German?
• If German was not the person’s mother tongue, fractured German was probably not uncommon. Borders were less stringent in the less populated areas, and colonization was common. There were always a certain number of immigrants imperfectly picking up the local language.
• Scholars migrated frequently between Paris, London, Heidelberg, Padua, Pisa, Salerno, Naples, et al. Instruction was often in Latin, but they would pick up some of the local language at each new location.
• Many people migrated to Vienna and Prague looking for patronage and other job opportunities. Some of them knew Latin before migrating, but several Holy Roman Emperors mandated German, and scholars from outside Germany and the HRE capital didn’t automatically study it and would have to pick it up once they got there, something that doesn’t happen overnight.
.
The 116v marginalia is a mixture of language patterns: German?, Romance?, Latin?, Voynichese. It’s hard to know if the mixture is deliberate or circumstantial. I tend to think it’s deliberate BUT… I have met two people who spoke polyglot… They lived in several European countries, but never in any place long enough to learn the language well, and one of them admitted he had learned bits of so many languages he had lost his mother tongue.
It’s also possible the marginalia writer had no connection to those who wrote the VMS main text. The two Voynichese tokens and the drawing on 116v might have been there before the rest of the text was added. Or, the person might have added Voynichese without knowing what it meant. I’ve seen many examples of notes and “pen tests” on old manuscripts where one scribe has written around or even incorporated the marks of previous scribes.
I’m HOPING the use of different languages is deliberate, but there’s no guarantee. I’m also hoping the marginalia is somehow connected to the main text, but there’s no definitive proof yet.
@ J.K. Petersen
I have considered all 3 pages where German occurs.
Question: Suppose you have learned Latin and have an English mother tongue. Would you write something in german if you are not sure what it looks like?
That is a fundamental question where you have to face.
What would I do if ……..?
I’ve thought about that many times, Peter, and, of course, I have no answers, I have only ideas, only a few of which will fit in a blog comment…
Why would someone weak in German try to write German and somewhat succeed in a couple of places?
• Assuming there was some connection between the marginalia writer and the designer of Voynichese, what if the people who created the text were German but the person who wrote the marginalia were Spanish, French, or Greek, but wanted to communicate at least some of it in German?
• Or what if they were in the court of the emperor and the only languages they had in common were German and Latin?
• What if the marginalia represents a medical recipe or charm? Many of these charms were written in a mixture of German and Latin, in which case someone not fluent in German might try to incorporate some German as a consequence of seeing other charms or remedies.
• What if it’s not German. What if it’s Yiddish?
We don’t have much information on how Yiddish was written in the early 15th century. It was German spoken/written by Jewish immigrants, but some had been there for a number of generations. They spoke Hebrew and Yiddish at home (some spoke Persian and Yiddish). Many of these neighborhoods were 80% Jewish (including the one in which W. Voynich was born), and Yiddish was a blend of German, but since it wasn’t the mother tongue, it was blended differently in different regions—there’s no one template for how it was spoken. I’ve always considered Yiddish as a possibility for the almost-German text, but there isn’t enough to be sure.
Many of the Jewish scholars worked together with scriptoria to do translations. They were well-qualified to do this, with one foot in the European door and the other in the Levant and Middle East (and parts of the corridor stretching up into Silesia). The marginalia writer doesn’t appear to be a professional scribe or a professional translator, but must have had some connection to literacy to have written on the Voynich Manuscript.
@ J.K. Petersen
There is something else where you may not be able to see. He writes in a German dialect. No real High German, but German. Dialects are tied to regions.
This book is cursed. Anyone who reads it just needs too much if and when.
Let’s take a look at this book. There is nothing inside where others would not know. Not really something mysterious.
Did he just encrypt it so nobody could understand his bad Latin so as not to lose his face in front of others.
Or did he want to play with Scripture he had studied in another country?
I do not see a single reference to another culture or religion, and stick to simple ones.
Peter wrote: “There is something else where you may not be able to see. He writes in a German dialect. No real High German, but German. Dialects are tied to regions.”
I think that’s obvious. If it were real High German, we could read it. I’m familiar with quite a few of the medieval German dialects and I still can’t read it. There’s not enough text. The only clear dialect is “nim” for “nimm”, but that particular substitution was used in a number of regions. The other one is “gas” or “gaf” but since we don’t know which one it is, that’s not easy to pin down either.
I consider the 116v text a sideline. I’m interested in it for its own sake, but I doubt that it will tell us much about the structure of Voynichese.
I do not know who invented the nonsense “gas”, but it means “ez as”.
See the third word on 17r. There is the whole readable text “Mather all süez her” = “mother of all sweet her” “Mutter aller süss ihr”
The whole sentence on 116 is …. “nach Verwendung falsch oben so nim ez as mich o”
I do not know what he is referring to, but I think that has something to do with the goat and the mermaid. There is a logical connection.
Peter, good grief!
How did you manage to change a normal”g” (on the last line of 116v) into “ez”? It’s a completely normal conventional “g” and looks nothing like “ez”. No one wrote “z” like that. Many scribes wrote “g” like that. It says “gaf” or “gas”.
And “mather” is not right either. Those are two Gothic ell shapes next to each other on the 17r marginalia, written exactly the same as the ell shapes on 116v, with the sharp loop. It’s “mallier” not “mather”.
This scribe writes “h” with a descending tail as can be seen by the “h” in “anchiton” and the “h” in “mich” on 116v. It’s a conventional way to write “h”, very common, and the same handwriting. Even if it’s a tail-less “h” (which is possible, many scribes mixed both styles), it’s still not an “h”. This scribe uses a round loop for the curve of the “h”, as can be seen on 116v and “mall___” isn’t written that way.
I read medieval manuscripts written in Gothic book and cursive hands every day (not just looking at pictures, I mean actually READING them) and I do not agree with your interpretation of these shapes.
@ J.K. Petersen
You will only find it in German texts.
He writes in a similar dialect as mine, and has made a small spelling mistake, I’ll explain that to you.
Actually the sign belongs to the word “nim” today one would write “nimmts” “nimmt es” / it “take” “take it”
it is spoken like “nimmz” or “nimmez” the emphasis “nimm-ez”
Now it says: “nim ez as mich o” And now the whole sentence makes sense.
But it’s nice if someone wants to explain my own dialect.
I can not only see a mistake, I can hear it too.
Otherwise, stay with your “G”
Peter,
1) It does not say “ez”. That is a “g”, plain and simple. It’s a standard, normal, conventional medieval Gothic bookhand “g”. I’m amazed anyone could call it anything else—it’s one of the plainest letters on the folio.
2) It is not only your dialect. The spelling “nim” was used quite frequently in three different regions of the Holy Roman Empire (three regions that were somewhat distant from each other) and I’m quite sure you weren’t born in all three which means there’s no way for you to know if it’s your dialect or one of the others. And it doesn’t matter anyway. Anyone who reads medieval German manuscripts sees “nim” all the time and knows perfectly well what it means.
3) I think it is very unlikely that there is an “o” at the end of the last phrase…
• That roundish shape is not in line with the preceding text, it is lower.
• It is not the same size as the other “o” letters.
• It has five dots above it, which is not a typical way to write the letter “o”.
• Most importantly, it is not shaped like the other “o” letters (it is lumpy and has a point lower left, it looks more like a stone than the letter “o”).
I’m pretty sure the rounded shape is a small drawing, not a letter. Just as there are other drawings on the left, it’s quite possible for there to be a drawing on the right.
It’s unbelievable, as you can see in the book more than a thousand g + 9, but you do not see the difference.
“o” means nothing else than “auch” because you can ask any Bayer or Austrian, and it is certainly not a drawing
And I know my dialect, and I know the dialect in which it was written.
Not directly mine, but I live only 80km away.
Stay with your G, and this with Gottez Gruss
It doesn’t matter what “o” means. It’s not an “o”. It’s not written like an “o” in any way. It doesn’t resemble the other “o” shapes on the page at all in
1) shape,
2) size,
3) position or
4) context (there are 5 dots above it).
You can’t claim it’s an “o” just because you WANT to turn the whole thing into a valid phrase.
You don’t know the dialect in which it is written because you are reading a “g” as “ez”. You can’t misread the letters and then claim it’s your dialect.
I will post my evidence that it is the letter “g” and you can post your evidence that it is “ez”.
But I’m curious how you want to explain that it is not an “o”.
Maybe the sign “ez” is not written clean, but it is the same as on 17r. And this word means “sweet” suez written.
I admit, I also thought it was a “g”. But if I take the mark of 17r, the whole sentence also makes sense. And that is certainly not a “g”.
In VM we have 3 sentences in German, all three are written in Tyrolean style.
The important thing about sentences is that they want to tell someone something. And for that they should also have meaning.
Look under Tyrolean spelling, and the rest explains itself
Peter, I am familiar with medieval Tyrolean spelling. I’ve been looking at it for quite a few years. I am very familiar with Tyrolean and Bavarian script styles. I have been sampling them for more than 10 years.
There is no “ez” on 17r, just as there is no “ez” on 116v.
The “g” in the last phrase on 116v and the “cz” or “ez” in “luc’z” or “buc’z” on 17r do not look the same at all!
Interpreting the Text
On 17r it says luc’z or buc’z kuc’z. The line above the “cz” is a macron, an abbreviation symbol, very clearly written. But the “z” is probably not the letter “z”. When it is next to a macron like this, it is usually an abbreviation symbol that is written like “z” but which is historically the letter “m” rotated 90 degrees clockwise (look at it sideways and you will see what I mean).
Many scribes wrote this abbreviation symbol similar to “z” but it is not a letter, it is a symbol, a very common one. You have to read medieval texts to recognize it, but once you learn it, you see it frequently (sometimes in every paragraph), usually at the ends of words.
Expanding the Text
The macron means there are one or more letters missing between the “c” and the “z-abbreviation-symbol” (sometimes the macron is offset from where the letters are missing, but this is not messy writing, it is tidy, so I think we can trust that the letters are missing between “c” and “rotated-m-symbol”).
Often the missing letter is “e”, with the “z” shape standing for “m”, but not always, sometimes 2 or 3 letters are missing, sometimes as many as 5 letters. Some scribes even use the rotated “m” symbol to mean “rum” instead of using the “rum” symbol that looks like a “4”.
There are quite a few options for the missing letters. It might be expanded to lucrum, lucorum, lucrem, lucorem or quite a few other things. To expand it properly you have to know the language, because many languages used Latin scribal abbreviations but they didn’t always expand them in the same way). The first letter is not very clear, so it’s hard to know if it’s “l” or “b” or possibly even “k”. There are quite a few other possibilities.
Even if it is a “z” (less likely, but possible), it still doesn’t say “cz” or “ez”, it would be “c-something-z”, and several letters might be missing.
You have to read medieval script from a medieval perspective. Modern scripts have dropped most of these conventions.
What you do not understand, 1 letter and the whole text makes sense. No matter how many reasons you still bring.
u für mich isch äs das xi
Peter, how do you know the text is supposed to make sense? Maybe it’s twin-speak, or a personal language. Maybe it’s polyglot. Maybe it’s partly coded.
You can’t change the letters to something that is not written on the page to FORCE it to make sense.
That’s not research. Research is about finding out what it IS, not what you want it to be.
How should I explain this to you?
Let’s take the word “valdch” on page 116. Actually, this word does not exist that way.
We would write it today “faltsch” or “faldsch” depending on the region. In High German one writes it “falsch” = “wrong”
You can just hear it. Even if I write it with “V” “valdsch” for the ear, it remains the same.
Everyone here somehow writes differently that you can not learn so easily. That’s where you’re born.
You write “luc’z or buc’z kuc’z” but these words do not exist in German or dialect.
The next possible would be “lueg” or “lüeg” to German ” schauen ” = look.
or just “suez” , “süez” = sweet. z=ss
Peter wrote: Let’s take the word “valdch” on page 116. Actually, this word does not exist that way.
We would write it today “faltsch” or “faldsch” depending on the region. In High German one writes it “falsch” = “wrong”
I have no trouble understanding falsch or valsch, it’s the same in Norse and German, and I’m quite familiar with it, but you have no proof that the last two letters are “ch”.
The scribe does NOT write “h” like that. Look at the two other “h” letters on the page and the “h” on fl7r—they don’t resemble the last letter of “vald–” in any way. You don’t know if the second-last letter is c or e either (it looks more like “e”).
You can’t change the letters into whatever you want just to satisfy your desire to turn it into valid words. They might not be real words. They might be anagrammed, they might be code words, they might be abbreviated.
You seem to think it’s okay to ignore what is on the page and to change it to what you want it to be, into something that makes linguistic sense, but we do not know for certain that it is wholly linguistic.
I’m sorry, now they want to explain my language to me again.
As you can see it is not an “n” in “obren” he writes it differently than “nim”, but what a miracle With them they are sure 2x mm with a point in the middle.
I’m sorry to say that, but they really do not understand any of the manuscripts.
No matter what you write, it means …..
Mutter aller, süess ihr vor……
und den muss des
Nach Verwendung falsch oben, so nimm z es mich o
Frage an einen Sprachspeziallisten, am besten ein deutsch Professor.
War das Wort “gas” so wie wir es kennen vor 1500 überhaupt schon im gebrauch ?
Ich kenne dies nur im vergleich mit Luft. Giftige Luft, schlechte Luft usw.
Danke nicht mehr nötig, habe mich gerade selber schlau gemacht.
Thank you no longer necessary, I have just made myself smart.
English:
https://en.wikipedia.org/wiki/Gas
German:
https://de.wikipedia.org/wiki/Gas
my dialect:
https://als.wikipedia.org/wiki/Gas
Yes, Peter, the word “gas” was in use. I found it in a manuscript. I looked for it. “Gaf” was also in use. It’s medieval lowland dialect for “gave”. I found several 15th-century examples of “gaf”.
“Gas”, when it is pronounced “gås” means goose. The word has been around for a long time, mostly in the north, but apparently also in parts of Bavaria.
But there’s a further possibility. I don’t remember seeing this discussed yet…
They frequently dropped endings in the Middle Ages. Letters and grammatical endings were often omitted from words like “nimm”, and a word like “gasse” (lane/alley/path) would sometimes be spelled gass (in the Middle Ages almost all words ending in double-s were sometimes spelled with single s) and thus might have been spelled “gas”. Dropping “e” at the ends of words, and dropping silent letters at the ends of words (which includes an extra “s”), was common. People wrote as they spoke.
Also, regarding “nim”…
If “nim” is actually “min” (my), then it might say “so min gas” (so, my street or so, my path). The “mich” at the end of the phrase isn’t very grammatical, although it does work with “gaf”, but in this context “path” could mean either a physical path/street or a spiritual path.
I think it is probably “nim” (take) rather than “min” (my) because of the position of the dot for the “i”, but dots were sometimes misplaced, so I suppose “min” should be considered. The word “min” (my, belonging to me) was spelled mein, main, meyn, myn, and min (and perhaps other ways, but those are the ones I’ve seen). In the north and parts of Saxony “min” is still in use.
I forgot to mention…
I don’t know who said it first, but “gas milch” has also been suggested, and I think it’s a reasonable idea even though geese obviously do not give milk.
Here is a goose milk reference I came across in Gessner, 1669:
“Die Gänsemilch wird auß dieser Feiste gemacht, wann man dieses Fett mit Milchkochet, und Zucker und Rosinlein darein thut.”
Gänsemilch can also be spelled Gase/Gas milch or Gåse/Gås melk.
.
And then there is the more familiar “Gäsemilch” (sometimes spelled with “G” but more often spelled Käsemilch, cheese milk), which might very easily have been spelled “gas milch” in the Middle Ages since they frequently did not capitalize nouns and frequently dropped endings. I don’t know if “cheese milk” has been suggested, but if “goose milk” is considered, then “cheese milk” should be, as well.
It is spelled “Käs milch” in 18th-century books. Substituting G for K is valid for many words, thus “Gäs milch” was probably a variant. For those unfamiliar with German… “cheese milk” refers to whey, the clear milk that separates out in the cheese-making process).
Once and for all, so that everyone understands a meaningful translation.
If he does something wrong above (recipe) it will catch him self.
He is not caught by a goat (Gäs, Gääs, Gais) or a goose (Gaas). it is also not milk (milk, mälchä, molke) the speech.
Or simply put, it’s shitty if the surgeon himself gets chloroform.
Well, the interpretation of “mich” as “milch” is not mine, but I have been told that “mich” is a valid way to write the way it is spoken in some places and it’s quite possible that is true.
My point was that there are many possible interpretations without changing the letters to something that they are not.
Peter, if the first line does not quite make sense, and the second line does not quite make sense, and the third line does not quite make sense, then how can you be sure the fourth line was intended to make sense? Maybe letters have been deliberately changed or left out of every line and the last line only superficially appears more complete than the others.
Nick: Having reread this post, probably for the nth time, there are a few thoughts I have had.
I think we can conclude that the advances in *mainstream* 15th century cipher techniques occurred in the first half of the 15th century (“Primo Quattrocentro” if you prefer). In the second half of the 15th century the extent to which ciphers were used and the familarity of a wider number of people in those cipher techniques and the size of cipher features like repertoire used may well have increased, but the techniques did not really change.
*When I say mainstream I exclude the work of Leon Battista Alberti as these techniques were not used, as far as I understand, by the diplomatic or wider community at that time.
Again the fact that the first half of the 15th century was a time of frequent war, and as we know from history war time is a period when cipher techniques tend to develop most, further supports the idea of this being a time of significant cipher development.
So what? Well this period of rapid cipher developments coincided with the date at which the Voynich manuscript was supposedly written. Now at times of rapid technological change different avenue of technological development are explored, some falling out of fashion and others becoming core features of the technology going forward.
Now I would have thought that given the presumed complexity of the Voynich cipher to those hypothetical people familiar with it, it would have been deemed too complicated to be used effectively in practice for diplomatic communication by any, but a tiny minority of people. So the Voynich cipher may represent or be connected to some kind of developmental cipher deadend, rejected for its impracticality for use amongst ordinary diplomats/cipher clerks. It could have been that rather than going down the road of more and more symbols for different things an alternative encipherment strategy was employed.
Nice theory, where is the evidence?
Well from my point of view, given that I connect the Voynich with Milanese diplomatic circles, I would say pretty much absent as are almost all cipher records from Milan in the first half of the 15th century. However my current research may prove to support such a theory, we will see.
Anyway it is a theory that I wanted to throw out there, if obviously highly speculative in nature.
I am increasingly converging on the view that the same words are often spelt in different ways.
Prior to being aware that it was common for diplomatic ciphers to allow for multiple spellings of the same word, as characters could be substituted for more than one different symbol, I noticed two words on the rosette page, which I may or may not have discussed, that I believed meant the same thing and were spelt in very similar, though not identical ways. I was shocked by this at the time as felt it a bit underhand to have two spellings of the same word as it seemed to me that there should be a one to one mapping of words. So having later discovered that this was normal amongst diplomatic ciphers. And again we see the same thing when comparing the T/O’s, words with very similar, but non identical spellings.
Now, I am confident that many will argue that this is merely coincidence, as it is common for Voynich spellings to look similar one to another; of course I am not so convinced in that.
Of course if the same word can be spelled in different ways this would reinforce the idea that the Voynich cipher was influenced by diplomatic ciphers.
I occasionally have the nagging feeling that at the end of the process of translating the Voynich we will find we do not have the impressive cipher that we hope for, but rather an annoyingly awkward cipher or something of the kind.
My worries fall I think into these categories:
1) Awkward features of the cipher that make it hard for anyone to decipher even the author. Such as a kind of many to many mapping of characters i.e. Letter “A” maps to “#”, “$” or “%” However Letter “B” maps to “@”, “£” or “$” which means that “$” could represent “A” or “B”; this could represent a nightmarish scenario for decipherment.
2) The author was careless in following his or her own cipher, thus making mistakes such that it is much harder to decipher.
I am hoping for a nice elegant solution, which I still think is a very plausible scenario.
I wonder if there are other examples where there are two words which one is inclined to believe say the same thing even though they have slightly different spellings. I guess this is hard to find as in my examples I had an idea of the word in the first place and clearly looking for two words with a similar spelling is in no way and indication that they are the same word.
@Mark
Most of those working on decryption already make the first mistake while thinking.
They work with the grammar and the spelling of 1600.
The C 14 dating tells me but I should stick to 1400.
This text was written by a monk in the 1400s. He has already been cleaned up.
entia c^rnis , ctiamsi vocatur pcccatum , non titiqnp, qnia
peccatnm est , sed quia peccato facta cst , sic rocalur, sic-
ttt scriptura manns cujusquc dicitur, quod manns eam fe-
cerit; peccata autem sunt, quae secnndum carnis concupis-
centiam vel ignorantiam iilicite fiunt, dicnntUB, cogitan-
tur. Quac transacta etiam reos tenent, si non remittantur.
I have to expect this kind of text when working on decryption in the VM.
I have a few words, but I did not really understand the meaning.
But I also have to know that if another monk writes the same text 50 km away, he will look different again. That’s the curse on grammar and dialect. There is also the vocabulary if he does not write in his mother tongue.
Example: what do you think when reading (hat from home) maybe just house roof.
Therefore, it is so important for me to find the place of origin.
Study the 2 step comparison material of equal time and place. Only now does it make sense if I am doing the decryption.
In medieval times, it was quite normal to spell the same word in different ways, sometimes even in the same sentence.
However, as far as the idea of the same letter being expressed by different glyphs in the VMS… Voynichese does not have the statistical properties one would expect from a many-to-one cipher.
The positional properties of the glyphs would not be so rigid or consistent if it were many-to-one. It seems very unlikely.
JKP: Yes, as I said in the Voynich Ninja comment(it has picked up a different username, which hopefully will get changed), the interesting question to me is how a diplomatic cipher could be modified or applied differently to give the kind of statistical properties that we see with the Voynich. I have made suggestions before about how an “atypical” diplomatic cipher might operate, such as sequences of characters mapping to individual letters or mapping to sequences of letters and null words, i.e. sequences of characters that map to nulls. I am inclined to the view for a few reasons to believe we have some null words.
JKP: Different spellings of the same word and null words are somewhat reminscent of diplomatic ciphers. Obviously for other reasons I believe we are looking at a diplomatic cipher as my 9 rosette theory points to an author who has strong links to the world of diplomatic ciphers. In addition I think the argument we are working with some kind of cipher is strong as there are good arguments against the alternatives i.e. natural language etc. Diplomatic ciphers, to the best of my knowledge, were the most advanced ciphers of the time and used a lot in Northern Italy, a region where we find other associations with the Voynich. It appears from my investigations that the carbon dating of the Voynich fits with a period of rapid diplomatic cipher research and developments in Northern Italy.
So how do we account for the differences from diplomatic ciphers that we see in the Voynich? This it seems to me is the key question.
Well this to some extent depends on how innovative the author was. The author might have been very well versed in diplomatic ciphers, but developed his/her own radically different and unique method of encipherment, such that the similarities are relatively few. Alternatively the author may have developed a cipher with a lot of broad similarities with the kind of diplomatic ciphers that we are all familiar with, but with a distinctive and different implementation of those techniques that can account for the difference in statistical properties. My instincts are that nothing comes completely out of thin air, meaning that I doubt that the cipher is completely unique and original though that still should not be discounted. However we must also take into account that we know from the contents of the Voynich that the author was a highly creative thinker, whicb would be perfectly consistent with a high level of creativity in the design of the cipher.
As far as this area of research goes, which is actually not currently my main focus, I think the most fruitful angle is trying to produce an “atypical” diplomatic cipher design consistent with what else we know about the structure of the Voynich cipher even if not the very specific details of the cipher.
As I have mentioned before, at some stage I will look into labelese research as my suspicion is that this presents the best line of attack on the cipher, but at the moment this is still low on my Voynich TO DO list. This might help bridge the gap between diplomatic ciphers and the Voynich cipher.
@ J.K. Petersen
I agree with you. for a many-to-one cipher, is the VM text to monotone.
Peter: I wonder if the monotony that you describe can be accounted for by a scenario where a string of more than 1 character can map to a letter of a null.
e.g. “$#” -> “g”
e.g. “%&” -> “”
Or even something like:
“%@?” -> “fa
or
“#$%@&” -> “”
This does not happen with traditional diplomatic ciphers, but is essentially the same kind of cipher; hence the use of the word “atypical”.
It is certainly possible that the Voynich is a completely innovative kind of cipher produced by someone with a familarity with diplomatic ciphers, but personally my first line of attack would be to investigate possible similarities with diplomatic ciphers.
@Mark
Basically you should look at a larger section of the text.
A null character would be possible, but would shorten the already short words even more, as long as they are single words. (That’s what I’m assuming, unless it’s spelling mistakes or variants). Otherwise, the whole text would be shortened. That would make no sense in the descriptions of the drawings, where there are only a word or two before.
Therefore, I personally consider null characters in the VM rather than unlikely.
Peter: I don’t see the shortening of words as a problem as spaces may not be “real” separators of words. e.g. “voy nich” looks like two words, but is actually just one “voynich”. I find it hard to imagine how the underlying Voynich text could not be significantly shorter. I would think given the nature of what we see it must be noticeably shorter. In fact it may possibly in reality be very much shorter.
When you say “That would make no sense in the descriptions of the drawings, where there are only a word or two before.” what exactly do you mean? Are you talking about labels?
Peter is right, many-to-one could result in “monotone” cipher, depending on how it is applied (and assuming it’s not also one-to-many as in the diplomatic ciphers). That’s actually quite a good descriptive term for it. The VMS does have this “monotone” property and most people seem to intuitively sense it because MANY of the proposed solutions try to expand individual glyphs into multiple letters BUT…
there is still the all-important component of positionality. WHERE do specific glyphs fall in a token (and why)?
This is very rarely (if ever) accounted for in any VMS “solution” and this is also where VMS tokens differ dramatically from those in diplomatic ciphers. Chars in diplomatic ciphers tend to fall in pretty much any position—it is, after all, a one-to-many, many-to-one substitution cipher. VMS glyphs do not do this.
I know that word-sharing is possible because today’s words are written together, which used to be separate, and vice versa. Mostly it is the writer who sometimes separates and sometimes not.
The question is here, intentionally or not.
I can not tell you what it’s like in English. But in German it is so with the syllables.
I have the most important prefixes; ein- aus- ver- vor- zu-
I’ll take the word “gehen =go” or “sehen=see” once
ein = on, gehen =go / ongo ?? eingehen =enter
ausgehen, zugehen, vergehen, vorgehen = to go out, to go, to pass away, to proceed.
If I take a sign for a prefix, I have to remember 5 characters, but you’re in the ass.
Do I still do the same thing with the endings; -heit, -keit, -ung, -nis, -tum
Now you have a serious problem.
This works well in Latin. Simple but effective. Think about it.
I have read a lot of old German text from the 1400. That to get a feel for the spelling and and the general style.
I recommend you also with English texts. It gives an insight how the writer would translate something (for example) into Latin.
A few thoughts on positionality. I should preface this by saying that I have not studied this question in details, but here goes:
1) I can see the attraction of the anagram of theory as it remove this problem, though I can’t say I am convinced by this.
2) Prefixes on words could tell you something about a word i.e. how many characters long it is, ignoring spaces. However this does seem to quite fit the bill. It does make me think of something like [bold][underlined][italic][times new roman][16 size] i.e. a series of descriptors of the word. Whivh could be something like: [3 syllables][5 consonants][2 vowels], but then one wonders from that kind of information how one gets at the word in question.
3) Words could simply be broken up so that certain characters appear at the beginning of each “voynich word” i.e. where spaces are ignored. Such as:
“Onc eup onatim eth erew asam ann am edmark” -> “Once upon a time there was a man named mark” where I have cut up the words so that the first letter of each word is a vowel.
JKP: Yes, thinking about it positionality could be largely achieved by how you cut up the real text into voynich words. You could cut up the text, so that “Voynich words” always started with a given set of letters and always end with a given set of letters giving the appearance of positionality where there really is none.
However the key thing is labelese.
Someone needs to study the statistical and other properties just of labelese as it is voynichese in its simplest form; I have said this for a while, but it is something I myself have not yet found the time to do.
A simple thought to ponder perhaps, is that most of the VM glyph words with their vowel like letters might well prove to be nulls, apart from an occasional one or two, formed by the scribe inadvertently, that appear to be meaningful but in reality mean stuff all as well.
At first glance, your “onc eup onatim eth erew asam ann edmark” would work.
But not at second glance. The prefixes and endings would shift to the middle of the word. But that does not happen in the VM. The “8g” combination as well as, “qo, iii), etc. are always in the same place.
John Sanders: Yes, I agree. We have to be prepared for the possibility that there may be a lot of junk which just needs filtering out to reveal the true text. There may be quite a bit of verbosity, so that the real text is actually significantly shorter.
@Mark
Here opinions differ. In fact, I assume that the text gets longer. Here, too, I have my reasons.
Peter: You say ->
“The prefixes and endings would shift to the middle of the word. But that does not happen in the VM. The “8g” combination as well as, “qo, iii), etc. are always in the same place.”
I am not clear here what precisely you are saying or if you have understood my point. Can you explain in more detail what you mean?
@J.K. Petersen
What I also have to consider is the writing divergence.
As you surely know, in old texts there is always something with the “c” or “e” and “u” or “v”. Sometimes they look the same and I have to read it to be sure.
But what about “q and 4” write tolerance or intent.
By intention of deception I do not break the monotone for the eye, but the text changes.
Peter: This is something one should be able to determine i.e. the information content of the whole text relative to an equivalent lengthed text in latin, for example.
Peter: If we take as simple examples the following texts:
“$$$$$$$$$$$$$”
That must have a low informational content in and of its self.
Whereas:
“$$$&$&&&$$&$”
That would normally have a higher informational content.
Whereas:
“$@#&&%£*^÷÷}{”
Looks to have normally an even higher informational content.
The somewhat monotonous nature of the Voynich text relative to an equivalent say latin text would seem to lead one to believe the Voynich text has a lower informational content than the latin text. This would seem to mean that there must be some kind of verbosity going on with the Voynich text.
I wonder, would it be possible to make some kind of calculation of the information content concentration of the Voynich relative to a latin text and thereby come up with some kind of number for verbosity?
e.g. 200% verbosity, meaning the true text is roughly half the length of the Voynich text if in latin
This is a text from the VM. The characters are roughly similar.
89 89 9qo4oxo89 & 89 o8o8
This is the right order:
89 89 9 qo4oxo89 & 89 o 8o8
This is the breakdown:
tum tum um exquisitum est tum a tat
That’s the translation:
One or both excellent, and there is a lot
This is what the technology looks like.
Or more simply expressed. What does “Rio de Janeiro” or “Montevideo” mean?
These are geographical features of the Portuguese seafaring.
They mean:
(The river in January), and, (I see a mountain).
The author of the VM, may be a doctor, or pharmacist.
But certainly no secret agent from the pope or emperor.
Because it’s easy, but not understood, you look in the distance, because you do not see what lies in front of your feet.
Nick: A question for you.
Supposing I wished to solve a 15th century diplomatic cipher, where I know the underlying language, I accept that the Voynich certainly isn’t a standard 15th century diplomatic cipher, but that is not what I am focusing on directly to start with. Now it is certainly possible to solve them manually, but supposing one wanted to be able to automate the process of generating the cipher key from a reasonable length ciphertext, i.e. such as can at least be solved by a human, with no crib or a very small crib, do you think it could be done? It seems to me it could, but it seems like it would be a non-trivial task.
If I could produce an algorithm or program to solve standard 15th century ciphers I can see this as being a useful step for me. Now I guess a human being can decipher text by knowing the context of words or familiarity with the subject of the text such as when trying to determine rare characters. But with a “sufficiently long” ciphertext could it be done on a computer, excluding glossary words, which would have to be worked out manually based on the context? What counts as “sufficiently long” is hard for me to guess at?
I guess one would need to access a complete dictionary file/database in the relevant human language. I don’t know quite how well frequency analysis works with a homophonic cipher with letter pairs etc. I suppose you may have to hard code in the standard common words as expecting the algorithm to deduce them is a tall order though possible. I guess one could write a human assisted algorithm though it should not to be very laborious for the human component as I want to automate or almost automate the human out of the process to make it speedy to generate the cipher key from the ciphertext. What do you think?
Mark Knowles: people have been writing automated cipher crackers sicne the dawn of the computer age. But having said that, I suspect you are actually wondering whether these cipher crackers would work with verbose ciphers. My best understanding is that the answer is no. Moreover, I suspect that there is no mainstream cipher cracking software that is able to predict nulls in ciphertexts. So this isn’t as automated as you might suspect. :-/
Nick: On the subject you raised I would have guessed it possible to produce an algorithm that could solve verbose ciphers, but it would be that much harder. I suppose the question is whether it would be a polynomial time or exponential time algorithm, Ultimately you would have to prune the solution tree effectively as obviously looping through all possibilities is not realistic .
I will look into this subject soon, I hope. Of course one would have to have a clear idea of the structure of the cipher.
One of the difficulties with assessing verbose ciphers is deciding where the boundaries might be. If the blocks were of the same length, it would be easier, but that is not necessarily the case.
It’s possible for a glyph-shape to function by itself and to be part of a multiglyph (verbose) pattern (based on proximity, for example, or some other criterion).
There’s quite a bit of analysis and decision-making even before constructing the algorithm.
One definitely needs to know what the problem that one is trying to solve is before constructing an algorithm to solve.
I have a suspicion that maybe in this area there is a tendency to read too much into some features of the text as to whether there is some cunning scheme to engineer this feature or it is symptomatic of some bizarre obscure language. I think some things could be pretty arbitrary or maybe unintended consequences of the author’s behaviour.
Nick: What would be nice if it exists if there was a quick concise checklist to tick off all the criteria that a cipher trying to account for what we see in the Voynich text must satisfy. I have read Rene’s list, though I could do with looking it over a few times, but I don’t know if it covers all the important features. I think it is important to distinguish between things the cipher must do and secondary issues that need explaining at some point, maybe Neal keys would fit into this category. I have a suspicion that if the core issues are addressed it is likely that the more minor questions can be resolved satisfactorily.
I have a suspicion that the Voynich text is truncated in certain cases, as I know has been suggested before, so for example “Wednesday” might become “Wed”. This would be for the case reason that people normally shorten text, i.e. when there isn’t the available space to write the whole word or just for brevity or maybe laziness. I should say this is not shorthand or some kind of abbreviation system that is being used, just simply truncation. Now there could in addition be shorthand or some system of abbreviation in use, but that is not what my thinking leads me to at present.
Mark Knowles: in the fifteenth century, shorthand was in a kind of lull between the lingering late Tironian notae of the Middle Ages and the modern shorthands that started in the late 16th century. What people actually used was abbreviation and truncation etc.
Nick: Yes, from what I have seen so far I think there are very likely instances of truncation going on. I suppose “Wednesday” could in theory be abbreviated to “Wdy” or something like that. However as is known I don’t believe Voynichese is a complicated form of shorthand using latin abbreviations, but rather cipher which has its origin amongst the world of diplomatic ciphers. As is, I think, now clear I believe it is a modification of and/or enhancement on a diplomatic cipher. I have suggested it looks most likely to me, at this time, that it is some kind of verbose diplomatic cipher.
Things like homophones might explain the lack of repeated Voynichese “phrases” or Voynichese “series of words”. Given that a given word can be spelt multiple different ways and therefore a phrase can be spelt a lot of different ways it is not so surprising not to see these kind of repetitions.(I am not refering to sequences of the same word here.)
Mark Knowles: it doesn’t have to be EITHER cipher OR shorthand, it could very easily be both at the same time. Someone wrote about this before, I can’t think of his name… 😉
Nick: You are right it could be. It could be cipher, shorthand, obscure language with an ancient lost script. It could be all those things and more, however that is clearly not where my thinking leads me. The idea that some of the symbols represent shorthand abbreviations in the context of the Voynich as has been suggested by … does not fit with my perspective at all, maybe I will change my mind, but I am far from there at the moment and getting further away it looks like.
Nick: Regarding my perspective I may seem inordinately stubborn and unwilling to adopt other ideas, but I just have my analysis and I am more than happy to change my mind when I think it is justified. I think for the time being I will stick by my guns on the question of latin abbreviations.
I should say that when it comes to the term “shorthand” it depends what you mean, diplomatic ciphers have shorthand implicit within them in a sense i.e. they can have individual characters that represent whole words or multiple letters. However I don’t think that is what is typically meant by “shorthand”. rather it is a system for abbreviating words; I emphasise the term “system”. I think the idea that the Voynich glyphs have any practical correspondence with actual latin abbreviations is a mistake, though they may sometimes have a visual correspondence, as this is basically inconsistent with the way diplomatic ciphers operate and I see no evidence or good reason to believe there is a deviation from them in this way.
One additional reason, amongst the many, to expect that the Voynich is a diplomatic cipher is the following:
Diplomatic ciphers were the most advanced ciphers of the time and it seems reasonable to imagine that of all the traditions of encryption it is most likely to be part of this due to how advanced the Voynich cipher clearly must be. Developments are normally built on other developments and rarely come out of thin air. Given the Voynich links to Northern Italy it makes even more sense that the cipher is linked to the world of diplomatic ciphers, which were so significant in that place and time.
It seems very likely that the author had some connections to the diplomatic world to be so familiar with these techniques. It is worth observing that Giovanni Fontana, I think, had no significant connections to the diplomatic world and produced a much simpler cipher. Even more so one would think it slightly more likely that the author had links to a city state(s) where major cipher development was taking place. Milan and Venice were the most important of these, though it would certainly be very speculative on this basis to conclude links to either of these.
It is worth being aware, I think, that the symbols used in diplomatic ciphers were arbitrary and assigned on an arbitrary basis to individual letters, words etc. There were often symbols that were common from one cipher key to the next, though the associations to letters, words etc. that they were given were generally different. However it is certainly the case that sometimes there were symbols used which were only found in a given cipher key and just an invention of the author. I think we should expect the same with the Voynich cipher.
JKP: You say:
“One of the difficulties with assessing verbose ciphers is deciding where the boundaries might be. If the blocks were of the same length, it would be easier, but that is not necessarily the case.
It’s possible for a glyph-shape to function by itself and to be part of a multiglyph (verbose) pattern (based on proximity, for example, or some other criterion).”
Of course this is a good reason for directionality. If multiglyphs are directional, with low character pair entropy, then it makes it much easier to distinguish the multiglyphs apart. This makes it easier for the author to read back his own work. I should add that in diplomatic ciphers many individual glyphs are directional.
Even with directional biglyphs it is easy to still generate a very large set of distinct biglyphs.
So for example:
Suppose we have 26 glyphs and 13 are always in the first position and the other 13 in the second position. Then we can have 13 x 13 distinct directional biglyphs. This is 169 directional biglyphs from a set of 26 glyphs this would provide more than enough biglyphs for a diplomatic cipher. As long as directionality is maintained one can have any mixture of multiglyphs without a problem. If course I am not saying that the author was necessarily always that rigourous.
You say:
“There’s quite a bit of analysis and decision-making even before constructing the algorithm.”
That is absolutely true.
I should probably say that if one has a triglyph
$&#
Then one shouldn’t also have the biglyph
&#
And similarly not
#
So in addition to directionality they should also clearly be non-overlapping. However as stated I am not sure whether the author has been quite so rigourous. It is important to note that all multiglyphs been directional is not a prerequisite for this kind of cipher, though it has many benefits.
An example of directionality in diplomatic cipher symbols:
It is common to see individual symbols(particularly in letter pair mappings) such as:
b-
b+
b=
b*
However one doesn’t see:
-b
+b
=b
*b
So really this kind of directionality in glyphs isn’t so different from directionality in multiglyphs.
In additional if there is strict directionality in all multiglyphs, which from what I know I doubt, then it should be simple to identify all the distinct multiglyphs used. Then actually we are dealing with something much closer to a standard diplomatic cipher. However I fear it is not that simple. Nevertheless this kind of approach may have some broad use.
I do feel that we should be wary of expecting the cipher to be elegant as it might not, however much we would like it to. I get the impression that the general inclination is to assume it is elegant.
Mark: I guess my own take on your ideas, put in my layman terminology, would be, that we have a not so simple, standardish diplomatic cipher, that meshes not in the slightest with rather inelegant, somewhat ordinary artwork of mixed media themes accompanied by nondescript, meaningless though presumed vulgar context. All of which is haphazardly laid out on very average quality non standard sized, (age certified) velum; within a goatskin covering of a strangely indeterminate vintage. It has all the makings of a quagmire in my inexpert, most ‘umble opinion, if that be the an appropriate description. Should you find yourself in general agreement, perhaps we can move forward with an air of confidence that might well prove to be infectious.
John Sanders: My ideas might be wrong, but I think there has been a tendency to all too readily dismiss the idea that Voynichese could have any connection with diplomatic ciphers of the time; that is deeply mistaken I think.
Diplomatic ciphers come with a whole toolkit of substitution techniques, which makes them a good basis from which to design one’s own cipher, rather than starting from scratch with a totally original design.
If we look at the characters which like;
a)
ai)
aii)
aiii)
There tends to be the assumption that their similarity is indicative of some general rules or principle by which they are related. Howevsr that may possibly not be the case as if it were a standard diplomatic cipher, which to reiterate I accept it is not, then it is typical to find symbols which look striking similar and related, yet have no relationship other than being adjacent on the cipher key.
As an example:
a) represents the letter W
ai) represents the letter X
aii) represents the letter Y
aiii) represents the letter Z
where letters before W in the alphabet have completely different symbols corresponding to them
Now this is obviously purely speculation, but it may provide food for thought.
Re: a/ai/aii/aiii
Whether it’s purely speculation or not, relating these sequences to letters is one of the more obvious things that I (and probably numerous other researchers) have considered (and tried).
This logic is basically week three of “Cryptanalysis 101”. I usually start with Occam’s Razor and work out from there. So, years ago, I wrote out dozens of pages of possible correspondences using exactly this logic. However, since this was years ago and it was difficult to confirm any of them, I also have hundreds of additional pages with other ideas that I think equally possible.
And yet, these signs only appear in the background or alone.
This is the simplest Latin around 1400. Take a close look at it. This is waiting for you if you have the key.
http://www.e-codices.unifr.ch/de/kba/Wett0004/1v/0/Sequence-1727
That’s why most people search for complicated solutions, and that’s because they do not understand the simple ones.
I convert this text into a 1 to 1 encryption = each character gets another symbol, also the ending abbreviation.
Who can not solve that need not worry about the VM text.
Mark, I don’t want to interfere with your thought process. We all start in different places and work through ideas in our own ways.
But my thinking on this (with a/ai/aii/aiii as an example) is that if it takes me a less than three minutes to consider that these sequences might map to individual letters (or numbers, or as multiglyphs), then the Friedmans (and their colleagues), who were GIFTED and TRAINED cryptanalysts, probably went through the same thought process in three seconds.
So I ask myself, if they weren’t able to solve it, what was the greatest impediment to doing so? Certainly not skill or imagination (they had both), and certainly not desire or persistence (they had that too). Knowledge was no impediment, they had education and access to many resources, and they had time (the Work Group looked into it for about four years)… but
…it was difficult to access information about medieval culture and languages. In wartime, it’s not easy to travel or ship books, and many manuscripts were not accessible to the public (like those in private collections and closed or bombed-out libraries), and nothing was digitized.
There are literally hundreds of different languages in Africa and Asia. It greatly helps to know which language group might be most fruitful. Assuming the VMS is natural language, if it’s not one of the major languages THAT would be a significant impediment to skilled cryptanalysts. Even substitution codes are a nuisance to solve if you don’t know the underlying language.
.
I’m not saying this is what has prevented us from figuring out the VMS (there are several other realistic possibilities), but suggestions about very basic cryptanalysis probably aren’t going to further the research.
I don’t mean to say Cryptanalysis 101 in a disparaging way—it’s a place to start—but the Friedmans and their colleagues were probably at the Cryptanalysis 801 level, so it’s probably more fruitful to look beyond basic code-breaking concepts if we want to decipher the VMS.
And while I’m at it, take a look at this page too.
http://www.e-codices.unifr.ch/de/kba/Wett0004/112r
JKP: I don’t think you have understood what I am saying. My point was that there need not be any relationship between them. So as a simple example in a standard diplomatic style cipher we could conceivably have:
a) represents the letter P
ai) represents the letter U
aii) represents the letter X
aiii) represents the letter Z
Note I am not saying there is a one to one mapping of symbols to letters of the alphabet, but providing a simplified example. I provide simple examples to illustrate specific points not necessarily illustrate some general analysis of Voynichese; I think you may not understand the purpose of my examples.
Could we make any headway by getting shut of those archaic capital letter forms, namely those nondescript ‘scaffoldish’ types along with (possibly) their generally spaced prefixes; just on the off-chance that the remaining letters might equate to anything remeniscent of something familiar. Or has that simple simon solution been attempted a million times already and found wanting.
Mark wrote: “Note I am not saying there is a one to one mapping of symbols to letters of the alphabet, but providing a simplified example.”
Yes, I understood that. You are re-stating the obvious.
It seems to me that the majority of currently-active Voynich researchers fail to grasp that EVA was intended only as a way to transcribe Voynichese, specifically so as to allow the exploration of how to parse Voynichese (i.e. how to map EVA symbols to individual tokens in whatever thing underlies it) to be carried out. That is, EVA was supposed to be a stepping stone towards an answer, not an answer in and of itself. Oh well. 🙁
Nick: Good point. If I understand you correctly that is consistent with my position.
JKP: You say:
“Yes, I understood that. You are re-stating the obvious.”
Yet you restate as before:
“But my thinking on this (with a/ai/aii/aiii as an example) is that if it takes me a less than three minutes to consider that these sequences might map to individual letters (or numbers, or as multiglyphs), then the Friedmans (and their colleagues), who were GIFTED and TRAINED cryptanalysts, probably went through the same thought process in three seconds.”
JKP: The point I was illustrating was that the assumption of pattern in many cases may be a mistake, whereas what we may have in some cases is relatively arbitrary associations. From what I have read this is an assumption all too often made. I was drawing on an example from standard diplomatic ciphers to illustrate this point, not making a very specific point about Voynichese.
In the latin alphabet take the letters:
M
N
Now these 2 letters look visually similar and what is more they are next to each other in the alphabet. So somebody not familiar with the alphabet might imagine there is some special relationship or common rule associated with them, whereas in reality they don’t have a lot more in common than the fact that they are both consonants, and of course most letters of the alphabet are consonants.
Yes, Mark. I know. To me, you are stating the obvious.
I have never assumed that similar patterns or similar shapes in the VMS represent related letters (or numbers, or symbols).
Maybe they do, maybe they don’t. Why would one assume one way or the other until there is enough data to support the findings?
JKP: Well, when I read what is written one gets the impression that people imagine or assume it is very likely there is a connection. People seem to often operate on the basis that there must be a reason for every aspect of Voynichese that we see rather than considering the possibility that many features may be just arbitrary. With diplomatic ciphers arbitrary assignments and arbitrary symbol designs are often part of the course.
As I stated before there also appears to be the expectation that the end solution will be elegant with every piece fitting together beautifully as part of an elegant intricate and ingenious system; that may not be what we find. I fear a lot of people will be disappointed when the solution arrives, but who knows maybe they won’t it is still to early to say.
@Nick
In order to work properly with EVA, I would have to completely re-map the glyphs of VM-Text. Since the conversion of the individual glyphs is already wrong. There where I can find on the internet, is shabby shit.
“es, et, est” are single characters, they look almost identical and are recorded as 1 character in the EVA. But they are not.
The differences of the deception are not recorded.
Did I understand that, I do not need EVA anymore.
Since a Korean PC keyboard would be better, because I can capture up to 4 characters in a glyph.
@Nick
What do you think ? Since we are with the VM rather at the beginning of the 15th century. Is not it more useful to search in the 14th century? That for text or encryption.
addendum:
I think, that a long time has to pass before a change becomes noticeable. (In this age)
Peter wrote: “… Since we are with the VM rather at the beginning of the 15th century. Is not it more useful to search in the 14th century? That for text or encryption.”
Why stop at the 14th century?
My assumption is that IF the VMS was created in the early 15th century (I think it probably was), the persons who created it might have been exposed to manuscripts dating back to the 8th century or earlier. Manuscripts in monasteries, noble collections, and various chained libraries encompassed many centuries.
The evidence does not seem to point to early medieval sources (the fashions, zodiacs, cloudband styles, face-style, palaeography, and other details are distinctly late 14th century and 15th century), but that doesn’t mean the person who drew the images was unfamiliar with earlier works.
When I search, I frequently start around 800 CE (and sometimes earlier) and work forward to about early 17th century. I’ve never assumed that someone living in the 15th century would have seen only 15th-century materials. Many of the 15th-century herbal and compotus manuscripts are based on much earlier exemplars.
Even though I do this, the data usually brings me back to the 15th century, but at least I feel I reached that point in an honest way, by searching broadly on both sides.
German, written 1399-1400, Vienna, Latin.
http://www.e-codices.unifr.ch/de/ubb/A-IX-0092/13r
When I see something like this, the VM seems rather easy to me.
It’s the different spellings that though the same time period and same language where the whole thing is so crazy.
How far can I include linguistics at all? Is it even meaningful?
That is also the reason why I search the encryption technique rather on the simple side.
And who actually messed up that beautiful Roman Latin?
addendum (english)
addendum (Latin)
ad dendum to abut
ad dentum the teeth;
ad den dum While the den
ad den tum Then the den
Shit Latin, that’s a miracle that you get the right drug in case of illness 🙂
Rene Zandbergen’s “Analysis of Text” section of his website is pretty thorough in going through various observed aspects of Voynichese, so he is to be commended for that. I have studied this section quite thoroughly and although I disagree with some of Rene’s conclusions this is clearly a very useful reference. However I wonder if there are other similar consolidations of different observed features of Voynichese. Ideally, as mentioned before, a checklist of properties without the statistical analysis by which those features were arrived at, for reasons of brevity, would help. Different features documented in different places is obviously less convenient to work with, though Rene provides links that I should look at.
Peter: You could have answered your own question; multi lingual, vulgar latin oriented, prescriptive doctors and pharmacists from the Indian sub-continent; That’s who!….
Gandhi ?
Nick: I was thinking about what you said previously and I thought it right putting my thoughts here. As per normal, you make a good point about the fact that we should be cautious regarding the consideration of the usage of cipher techniques in the Voynich that we only know of in much later ciphers. However I thought it probably also worth mentioning that in general we should also be wary of the idea that everything we see in the Voynich cipher must have a parallel in ciphers of the period; ultimately if it did we would have solved it long ago. So I think it is plausible that there is/are some original feature(s) for the time in the Voynich cipher; now what that feature(s) may be is certainly unclear. As I have stated many times I have been very much of the opinion that the author of the Voynich was quite an original thinker and so it is not unreasonable I think that many different aspects of the Voynich do not have a contemporary parallel. Purely as an example, whilst I don’t believe there is a parallel with Alberti polyalphabetic ciphers, it is conceivable that the author discovered/invented this technique that we associate with a later period.
However the problem with this line of thought is that it could be used to justify the assertion of the use of almost any technique by the author of the Voynich, no matter how outlandish, so I think we have the hard job of distinguishing between the plausible and implausible innovations the author probably made.
Nick: I think when trying to determine what the chance that the author used a certain enciphering technique not present, to our best knowledge, amongst ciphers of that period then the question has to relate to the extent of the intellectual leap involved. Clearly the author did not implement the RSA algorithm as that would involve a vast mental leap. I don’t even think the period in which we first see a technique implemented is crucial, but rather just whether the author could realistically conceive of such an idea. We can ask the question, could I conceive of that technical advancement and if so is that based on knowledge I have that the author didn’t have at that time? It seems very questionable to imagine that the author was much less creative that we are.
Wish everyone a nice Easter !
https://www.facebook.com/photo.php?fbid=2258066547749320&set=gm.2041923225917468&type=3&theater&ifg=1
Mark: the general rule I work to is that for each time a claimed cipher solution requires history books to be rewritten just to break the cipher, its likelihood of being correct drops by 90% or more.
Nick: That’s a good point one should be very suspicious of that and yet won’t the solution to the Voynich manuscript inevitably caise history books to be rewritten. If we take the following:
1) It is written in an as yet unknown language -> history books need to be rewritten.
2) It is written in a known language in an unknown script -> history books need to be rewritten.
3) It is written in a complex form of shorthand -> history books need to be rewritten.
Then of course we have to consider the situation where it is written in cipher. If it is written in a cipher known to us of the time then surely we must be able to decipher it. If we can’t decipher it then it can’t be in a cipher known to us of the time. If it is written in a cipher unknown to us of the period then it will surely lead to the rewriting of history books.
It seems to me that the discovery of how read of the Voynich manuscript will inevitably and unavoidably cause history books to be rewritten; it seems to me the question is not if, but how they need to be rewritten. (It appears to me that by its very nature the Voynich rewrites history books) I fail to see how we can come to any other conclusion.
In addition the text may not be the only way in which the Voynich will rewtite history books.
Mark: there were a few 15th century ciphers that enciphered some of those few Tironian notae that were still in use, I gave an example of one in Curse. So hypthesizing that Voynichese might be in some way an enciphered shorthand would seem to be one of the few candidate theories that don’t mess with the history books.
I’m not concerned in this context as to what the Voynich Manuscript may or may not contain, I’m only concerned with whether the proposed cipher system requires history books to be rewritten.
Nick: I will reread your Tironian notae writings.
I just mentioned the contents of the Voynich as an aside, which is just part of my general argument regarding the degree of inventiveness of the manuscript, but I take your point that that is somewhat off topic.
Nick: Having looked at what you have written it seems not unreasonable to argue that there may be a few Tironian Notes used and that would certainly be consistent with my theory. However it does depend on the extent to which you think they may be present and the extent to which they replace or supersede the standard letters of the alphabet i.e. if you are thinking of fewer than 10 or many more then that makes a difference I think.
Nick: To me there is a big distinction between the possibility of a cipher containing a handful of Tironian Note shorthands as we find with some Diplomatic ciphers, which is something that I am openminded about as a possibility, and a cipher composed of largely or completely of shorthand.
I don’t know what Nick’s opinion is on this, but I haven’t seen any sign of Tironian Notes in the VMS text. Almost all VMS characters are Latin characters, ligatures, numbers, and abbreviations, except for a few that are slightly more similar to Greek, as I’ve been saying all along.
There are thousands of Tironian note-combinations, so vague similarity is almost inevitable, but there’s no systematic similarity whatsoever between Tironian notes and the VMS. There is a system behind Tironian notes that can be observed and learned, just as with most shorthand systems, and it’s clear that the VMS does not follow these logical patterns and conventions. Any glyph similarity is coincidental (or inherited through Latin conventions).
Nick: As I have made clear in the past, the kind of shorthand model suggested by JKP looks to me to be quite infeasible.
JKP: I think in Diplomatic ciphers any symbol could be attached to a Tironian note. Diplomatic ciphers are non-symbol specific, one could invent any symbol one fancied and assign it to a letter, letter pair, word etc. In that respect the Voynich symbols are hardly unusual.
One difference between Diplomatic ciphers and the kind of shorthand JKP advocates is that he has to root around to try and find shorthand symbols that look vaguely like what we see in the Voynich. There is not the same degree of importance in matching the Voynich symbols to symbols that one sees amongst diplomatic ciphers. Whilst it is certainly nice and a worthwhile endeavour to search for symbols in diplomatic ciphers that resemble what we see in the Voynich, symbols are often invented and some symbols may only appear in one cipher key, so the expectation of finding matching symbols is not the same as for those trawling through latin abbreviations shorthand symbols to find them.
It does seem that given the limited number of specific symbols in the Voynich it is unclear how any kind of latin abbreviations system would be feasible.
Mark: was that meant as some kind of criticism of my enciphered shorthand hypothesis?
Mark Knowles wrote: “One difference between Diplomatic ciphers and the kind of shorthand JKP advocates is that he has to root around to try and find shorthand symbols that look vaguely like what we see in the Voynich.”
I have no idea what you are talking about.
I haven’t advocated any kind of shorthand in regard to Voynichese.
Shorthand is a system. Tironian Notes are a specific system. Pittman shorthand is a specific system. Each has its own rules and “grammar” and logic to which the shapes are assigned. They have to or they would be impossible to learn and remember.
– Latin characters that are similar to Voynichese are not shorthand.
– Latin ligatures that are similar to Voynichese are not shorthand either, they are simply two or more characters combined to make it easier to write them quickly. It’s a stylistic choice for quickness of hand.
– Latin abbreviations aren’t really shorthand either. A high proportion are apostrophes. Tails on the ends of words are apostrophes. Would you call the word “isn’t” English shorthand? No, contractions are just part of the normal repertoire, familiar to everyone. They are not distinct enough to call “shorthand”.
We do NOT have to “root around” to find symbols that match Voynichese. They are readily apparent to anyone familiar with medieval Latin. They are NOT apparent to people who learn modern Latin because these conventions disappeared in the 17th and 18th centuries.
Mark Knowles wrote: “I think in Diplomatic ciphers any symbol could be attached to a Tironian note. Diplomatic ciphers are non-symbol specific, one could invent any symbol one fancied and assign it to a letter, letter pair, word etc. In that respect the Voynich symbols are hardly unusual.”
I really surprises me that you wrote this because it means you haven’t really looked at the Diplomatic symbols (or understood them) even though you’ve been talking about them (and theorizing about them) for a long time.
The vast majority of symbols in the Tranchedino collection are quite specific and most of them are based on Latin characters, numbers, and abbreviations. Even the ones that are different are usually created in a specific way by writing a common Latin character and simply adding a leg or stem. For example, the Latin abbreviation for “qui” (a q with a line through the stem) occurs in the first cipher on Tranchedino 1r and you will see that they simple added an extra line and one more line to create three symbols. But anyone who knows Latin recogizes the base-abbreviation for “qui”. It’s like taking the Latin character “t” and adding an extra cross-stem. Anyone who knows the Latin character set would recognize the base symbol. The system they used to create extra characters is fairly consistent and quite recognizable.
There are also some Greek and Math symbols, but they are readily recognized also. Most scholars learned at least some Greek. The symbols that look like ell shapes or boxes with dots in them are Greek abbreviations (this was a very common way to abbreviate in Greek that did not make it into Latin scribal conventions).
The specific groups of shapes that might look strange to you are the many small symbols with small circles connected with little bars in various combinations, but those are Latin too. These specific symbols are used quite extensively as paragraph markers in manuscripts that are heavily glossed in the margin.
Nick: Not really, although I am not 100% sure what your enciphered shorthand hypothesis is.
As far as I understand it JKP Latin Abbreviations shorthand hypothesis is different from your Tironian notes hypothesis, so I think these need to be addressed separately and independently. There may be some overlap between the two, but I am not yet clear as to this.
I think I understand what JKP is suggesting and that seems problematic, so I was addressing that.
I think the question for me is to what extent you see Tironian notes playing a part in your hypotheses?
Nick: Roughly how many different Tironian notes do you think are represented? e.g. 5, 20, 100
Do you think the symbols only represent Tironian notes or is the whole alphabet represented as well?
Mark: it would seem that you haven’t even begun to understand the structure of what I’m talking about here. Shorthand in the 15th century fell between the long slow death of Tironian Notae (hardly any left by 1450, and basically none by 1500) and the much later rise of modern shorthands (circa 1600 and onwards). So what scribes actually used were abbreviatory tricks: contraction (missing letters out) and truncation (lopping off the end parts of words).
What I proposed more than a decade ago is that the Voynichese writing system might well include some kind of aggressive abbreviation linked with some kind of verbose cipher. So the words appear to get longer (because of the verbose cipher) but then get trimmed down again (because of the abbreviation), yielding a curiously balanced writing system where Voynichese words end up just about the same size as the plaintext.
So, in answer to your question, I don’t believe that Voynichese enciphers even a single Tironian nota.
I do believe the whole plaintext alphabet is represented, but via a tricksy verbose cipher.
JKP: To start with I have looked at the Tranchedino many times. I have also looked at many other sources of cipher keys of the period, some of which are not available in any book or online.
You say:
“I really surprises me that you wrote this because it means you haven’t really looked at the Diplomatic symbols (or understood them) even though you’ve been talking about them (and theorizing about them) for a long time.”
It seems within your framework any possible symbol anyone could think of you would say was a development of latin characters, greek symbols etc.
I propose a test I can email you 10 cipher symbols. Some of which I have made up, not found elsewhere, and some from cipher keys you have not and cannot find, to avoid you looking them up in Tranchedino etc. Then can you tell me which are genuine and why?I can email you the cipher keys after to verify the identifications.
Nick: By aggressive abbreviations, what exactly do you mean?
As an example:
“Elephant”
->
“Elpht” or
“Eleph” or
“$phant” or
“Eleph%” or
Or do you mean something quite different?
Mark: please stop trolling JKP. This is getting boring, and I’d like you to do better than this.
As far as ELEPHANT goes, LPHNT would be one style of contraction (dropping the vowels), while ELEPH- would be one style of trunkation (pun intended), but there are plenty of variations of both. The fifteenth century had many local scribal abbreviation styles, particularly in Italy: and the structure of different languages suggested different abbreviation strategies, e.g. Latin and Tuskan (pun also intended).
JKP: I am not very focused on the semantics of whether you like to call the latin abbreviations you refer to, shorthand or not.
According to the internet, which never lies:
“Shorthand is a method of rapid writing by means of abbreviations and symbols.”
Which looking at Cappelli seems to fit, but if you don’t like that word we can find another.
Even you say: “Latin abbreviations aren’t really shorthand either”, so you aren’t sure.
Anyway semantic discussions are a waste of time.
You say: “We do NOT have to “root around” to find symbols that match Voynichese. They are readily apparent to anyone familiar with medieval Latin. ”
And yet you do not present examples(I have seen the examples you have presented and you can’t come up with examples that fit.)
Anyway I don’t see any point revisiting a discussion we have had at least once before. It sometimes feels like we must be operating with completely different eyes.
Nick: You are right, I’m bored too. As usual I was not planning on getting embroiled in another debate with JKP. I will just summarise:
I do believe that there are many symbols in diplomatic ciphers that cannot be found in Latin abbreviations. There is a few symbols in the Voynich, including the gallows, that I have not seen or from what I know anyone else has seen amongst Latin abbreviations.
The End
As far as the abbreviations you refer to it seems you are talking about the author leaving out some letters in a word as in the examples you gave. Then I think it does depend on which system is used.
Mark Knowles wrote: “It seems within your framework any possible symbol anyone could think of you would say was a development of latin characters, greek symbols etc.”
No, if that’s what you think, you’re way off-base.
There is a difference between OBSERVING patterns and inventing them. I am not inventing anything. I recognize these patterns, just as you would recognize an apple or an orange. People who learn medieval scribal conventions can distinguish a standard shape from a nonstandard shape, just as mathematicians can distinguish a real mathematical formula from a fake formula drawn with invented symbols.
.
And as far as what Nick is talking about and what I’m talking about, those shouldn’t be confused either. Nick is talking about what underlies the VMS text (in terms of possible structure and meaning), at least in the last few posts. I am interested in that too, but when I cite Latin (and some Greek), I am referring to the origin of the glyph shapes, a subject that should NOT be confused with the underlying structure of Voynichese when discussing concepts like shorthand.
What Nick is describing is entirely possible.
.
As for your proposed “test”, 10 cipher symbols is not enough to make a judgment. There are several thousand cipher symbols in the Tranchedino collection from which to distill out the patterns. There are more than 200 sides in the VMS to observe the patterns. This is sufficient data to make a few generalizations (note that I said a few).
10 cipher symbols is not. You should know that. No scientific method would accept a sample set of 10 for something like this.
Besides, if you are entirely ignorant of medieval scribal conventions, you might unintentionally create one that matches medieval conventions in some language without even knowing it.
Mark: I must concur with the boss for once; Persistance with this unproductive line, is becoming somewhat tedious and pointless. Likewise trollish behaviour towards respected team playing contributors can only lead to general distraction and must discontinue hence forth. It would be in our mutual interests that, rather than to become off thread non team playing bully boys, we might otherwise align our VM thoughts to ideally reflect those themes more popular with the much more agreeable status quo. If we can pull together to fullfil such fine ideals, the mystery of the Voynich Manuscript will remain so eternally…and we will indeed have done better Mark wouldn’t you say?..
Mark Knowles wrote: “There is a few symbols in the Voynich, including the gallows, that I have not seen or from what I know anyone else has seen amongst Latin abbreviations.”
Which gallows have you not seen? The EVA-k shape is an extremely common Latin abbreviation. It is especially common at the beginnings of lines and paragraphs. It is also a common ligature in Old French (it stands for “Il”).
The EVA-t shape is sometimes used as a substitute for EVA-k. It is less common than EVA-k, but it exists.
Benched gallows can be found in Greek. In fact, they come in two kinds, rounded bench and straight bench (both of which exist in the VMS). They don’t usually have a double-loop on the “P” shape (it’s usually single loop), but other aspects are the same.
Double loops (as in EVA-p) are not common in Latin, but they do occur in Greek. The small loop added to the upper right of a letter is a common Greek abbreviation symbol that usually means “o” or “e”. This abbreviation is frequently found on the very common word “Peri”, which is often at the beginnings of lines and paragraphs (similar in position and context to the EVA-k abbreviation for “Item” in Latin).
As far as position goes…
• EVA-k is a very common Latin abbreviation found mostly at the beginnings of paragraphs and lines (by lines, I mean either lines or sentences, depending on whether it is narrative text or lists).
• Peri is a very common Greek word that is frequently abbreviated (which could be likened to a Latinized EVA-p) and, when abbreviated is also found mostly at the beginnings of paragraphs and lines/sentences (often at the beginnings of lines full of lists).
In the VMS,
• gallows are frequently found at the beginnings of paragraphs (and for all we know maybe also at the beginnings of lines except we don’t yet know where Voynichese “sentences” begin and end).
• The EVA-y shape (one of the most common Latin abbreviations) is also positionally consistent with Latin.
• EVA-m and EVA-g shapes (both very common Latin abbreviations) are also positionally consistent with Latin.
Similarity in shape is one thing—it can easily be coincidence. Similarity in shape and position has to be investigated more thoroughly because it is less likely to be coincidence.
Nick & John Sanders: I started a conversation where I was trying to understand Nick’s theory and only mentioned JKP theory by way of trying to determine the extent of the commonality or lack thereof between the 2 theories. I did not want to get involved in a discussion that I have had before with JKP and am also somewhat bored by. JKP appears to be significantly more interested in rehashing that discussion. He also suggested that I had not looked at the Tranchedino and have no understanding of diplomatic ciphers, which is absurd. [Remainder deleted]
Nick: As to the author having written some words omitting certain letters I cannot argue that that has happened. In fact as I have mentioned before I believe that some words on the 9 rosette page have been truncated. The question for me is whether that is a consistent pattern of a standard removal of letters from words. This being whether all words vowels have been removed or there may be some instances of that. Similarly if it is argued that all words are truncated or there may be some instances of that. So for me the question is whether there was a system of the same kind of abbreviations in all cases or whether there are instances of abbreviations either of a common form such as truncation or a few different kinds of abbreviations. I am sceptical regarding a system of universal abbreviation, but am very open-minded regarding there being specific instances of almost any kind of abbreviation(excluding the JKP kind of abbreviations). In short I find your argument that abbreviations of the kind you describe may be commonplace in the Voynich very reasonable. Certainly this could resolve “word” length issues as a verbose abbreviated cipher would broadly speaking fit what we observe. However I doubt that all words are abbreviated and it appears to me that there are some instances of unabbreviated words.
Nick: As you know I suggested a system of possible removal and insertion of blank spaces . Now whilst I rejected your historical present argument for this as I don’t believe it would be much of a mental leap for the author as it was not much effort for me to conceive of, the greater simplicity of your abbreviations argument is enticing. I find the argument of frequent, though not universal, truncation of words particularly plausible. I am not saying it is necessarily so, but it is definitely a possibility it seems. How do you reconcile positionality? I obviously have my suggestion, but I am open to others.
Nick: Regarding your abbreviation suggestion versus my misleading spaces I have thought single word labelese a good medium to assess this. If there are more long words in labelese than we see in sentencese then it would support my assertion. If there is much the same length of words it would seem to me to support your assertion. In general, the length of word profile of labelese versus sentencese would be a useful comparison.
One other thing that interests me more and more is how we determine the extent of the verbosity. The 2 character entropy statistics lead me to believe they act as a block i.e. some character pairs do not operate independently and must in someway map as a pair to one individual thing such as a letter of the alphabet. Now however questions about what the 3 character entropy says about whether there are likely to be some 3 character blocks would be interesting. Similarly extending this to calculating the entropy of 4 character plus blocks may possibly be interesting. This could also give us insight into the possibility that spaces are not really spaces. If we can then do our best to identify as many of the blocks of any length as possible that could be interesting. (E.g. individual repeated words would seem to constitute larger blocks.)
Nick: More thoughts. I would think universal abbreviation would be a problem. However frequent abbreviation would not. There are instances where abbreviation can lead to ambiguity as to what the word is and I would think instances where this is even ambiguous given the context of the rest of the sentence. However I think there are many instances where abbreviation would not cause comprehension problems. If and the extent of use of abbreviations in the text may be hard to determine. To argue that there are no abbreviations whatsoever, of the kind that you discuss, in the text stretches credulity.
Mark: to be precise, my hypothesis is that there are probably no Tironian notae in the plaintext, but that the plaintext is probably instead abbreviated by use of contraction and truncation, in line with the kind of scribal abbreviation that was widely used in mid-15th century Italy.
I’m not sure what you think I said, I have tried to say what I said it as clearly as I can for over a decade now.
Nick: I haven’t known you for more than a few years. Truncation by omitting letters seems very plausible to me. Depending on the kind scribal abbreviations you are refering to determines my perspective. If you mean things like i.e. , e.g. , etc. and so on that does not present a problem to me. If you are refering to the kind of thing we see in Cappelli then I find that problematic.
Nick: It is true that I have not read everything that you have written on the subject. The use of standard abbreviations for certain words by removing certain letters in each case seems perfectly possible, but that obviously would not imply a universal rule such as removing all vowels or preclude someone writing a word in full if they wished to.
Mark: in 15th century Italy, each group of scribes had its own local way of abbreviating text, its own local conventions. Contractions, truncations, overbars, hyphens etc were all par for the course, but without some secondary information from the manuscript itself it is very difficult to work out a rigorous way of determining which abbreviation style was / could have been used etc, e.g. if Milan was the locale, you should look at Milanese scribal abbreviations.
However, fifteenth century Italian scribal abbreviation has proved a frustrating topic to research. Everyone with an interest seems to have died a century ago without writing anything down. 🙁
Nick: On that basis you could reasonably argue that the author invented his own form of abbreviations. My concern tends to arise if, and I am not saying this your perspective, one has to introduce many more symbol correspondences to make the abbreviations work. Meaning if we have a special symbol mapping for “ing” or a special symbol for “ion”and many more, even within the context of a verbose cipher I can see that causing problems. However from what I can understand that is not what you are referring to. If you are talking about mapping a handful of symbols like hyphens, dots and so on I can’t see a problem with that.
Fundamentally the question becomes how many mappings you require. By this I mean you must have a mapping to each letter of the alphabet, mappings for nulls, maybe common words, maybe letter pairs, maybe a few Tironian notes. It also seems to me that there can’t be a one to one substitution, so some kind of homophonicity which again increases the number of mappings.
If you are dealing with many 3 character blocks I am not sure the words will be long enough even if abbreviated. If you have 2 character blocks you won’t have enough mappings. I am not suggesting that is true in your case, but it starts to become true if one requires many more symbols.
Nick: I feel we probably greatly underestimate the intelligence and creativity of the author. Take yourself for example you created the game “Frak” out of your own imagination, I came very close to playing Frak once when I was a kid; though the best I could do was watch someone else playing it.
I am sure you or I or many other Voynich researchers could come up with our own unique and hard to crack ciphers, so I don’t think we should imagine our author couldn’t. I feel at time there can be an overemphasis on grounding him/her within the confines of a specific historical tradition. This is not a specific criticism of any particular people, but rather I wonder if we need to be more openminded about what we are dealing with.
Nick: There are some “words” that I have a suspicion about which I will email you at some point. At the moment, when I have time, I am trying to study other people’s analysis of the text.
Nick: I ought to trawl through all the analysis of text that different people have done, but that could take quite a while. If I am not confident of a conclusion on that basis then I will need to work out which tests I think I need to conduct. Then implement those tests and see where those results lead me i.e. other tests, a conclusion.
However in the back of my mind I wonder if fundamentally I will conclude that we need a larger crib (I think I have 7 words, but that is not a lot) and that this is the only way forward, which will push me back on my other lines of research.
So what to prioritise in the time I have for Voynich research I am not sure. What I have demonstrated, I think is that it is perfectly possible to construct a non-standard diplomatic cipher that essentially fits the kind of properties we see in the Voynich and given your point it does not necessarily require features to handle spaces.
Mark Knowles wrote: “What I have demonstrated, I think is that it is perfectly possible to construct a non-standard diplomatic cipher that essentially fits the kind of properties we see in the Voynich and given your point it does not necessarily require features to handle spaces.”
Where did you demonstrate this? Do you have a sample? I remember you said on the forum you were going to post a sample.
JKP: I haven’t made this public, but I have shared it. I may well go public with it, but it will depend on how good an idea I think it is once I have drilled down to very very specific details. It provides an explanation of positionality, directionality and relatively low numbers of symbols, rare symbols and repeated words amongst other things. I think I have probably addressed most of those already in public such as directionality, few symbols, but I know I haven’t discussed positionality.
JKP: The more significant I think my idea is the less likely I will share it soon.
@Mark
I’m very excited about what you have to offer.
No matter what your idea looks like, it also has to be easy to read for a second person. It is not enough if you just leave a few letters and the other does not understand.
@JKP
Do you think that’s the same word?
I still think that the German text is the key to the origin of the VM. I do not see a second person.
https://www.facebook.com/photo.php?fbid=2261016390787669&set=gm.2048969255212865&type=3&theater&ifg=1
Peter, I don’t know. It’s tempting to think it’s valden, but what if the 8-shape is a “d” rather than “s”? On the “portas” line, it seems more likely to be “s”.
Valde/valden works in both German and Latin. It’s not an uncommon reading for this word. It could also be a legitimate alternate spelling for velden, which can mean a number of things in different languages, including “to choose”.
The difficult part is reconciling it with the word that comes after it, which is even more difficult to interpret. It would be nice to find some way to relate them (assuming they are part of the same phrase).
@JKP
I know that it’s shit again. As always, I have the comparison material, but the origin is missing. I’ve been looking for the right place for days now. I know it would be helpful if I had the whole sentence. As always !
I should slowly get used to writing a note.
But I know that I have to search in the South Alemanic.
That (PORTAS, PORTAD) is also such a thing.
I think it’s the short form of (importante, ital., or latin.) I do not really know, it’s just an assumption.
I was just looking at the text of the 9 rosette foldout, the only part that I have not studied in detail. I choose this page as I am very familiar with it and I have my own theories about it already and also working with all the text in the manuscript as whole seems a lot to grapple with. But very importantly it has lots of labels. Isolated single word labels are great, because we don’t need to worry about if spaces are really spaces or what is happening with the beginnings of paragraphs and neal keys and sentences and the like or whether they are simple words like “the”, “and”, “not” etc. Now it is possible, and I think sometimes likely, that a label might be a null word designed to confuse. However if it is not then, certainly in the context of the 9 rosette foldout in my opinion it is a specific word, in fact quite possibly unique in the manuscript. Now there are words that encircle each of the Rosettes and it is more debatable whether these are really isolated single words or constitutes parts of a sentence. There are a few other instances where it is possible that words follow from one another. However there are quite a number of examples which unequivocally are isolated single word labels. Now looking through these there are some which look like generic commonplace text, which could well be nulls, I think, and there quite a few which look like distinctive words with their own very specific spellings.
This raises a question or two. I have long thought, even before I knew there were such things as nulls in ciphers of the period, that it is almost certainly the case that words that repeat 3 times are nulls; there are alternative explanations I can think of, but I think they are implausible, such as repetition for emphasis or repetition as a very verbose way of count e.g. “boil boil boil” i.e boil three times. This makes me ask, if there are null words and/or null sequences of symbols, then how prevalent are they; conceivably they could be very widespread. If they are widespread they could be messing around a lot with the statistics for such things as symbol frequency and so giving us a completely distorted view of what Voynichese looks like. The question then becomes how one filters out nulls as much as possible from Voynich text, so that then we can recalculate the true statistics? I think null words would have, most likely, to be common words or at least have a common frequent feature or marker, as otherwise the author would have to manage a vast library of null words. So one could remove all the most common words from the Voynich text, even though some could be words like “the”, “at”, “and” etc., and then recalculate the frequency statistics to see if they are significantly different. Obviously deciding what are “the most common words” and what don’t quite fall into that category is far from straightforward, though when in doubt I would remove them.
It seems to me we could even have a situation where for example every word that starts with an “o” and is followed by a gallows character is a null.
Am I right in thinking that the 9 rosette page is the only page not on voynichese,com?
Mark: there are at least twenty perfectly (and equally) good explanations as to why Voynichese has repetitions in the text, so I’m doubtful that simply deciding that they are nulls is a great starting point. Similarly for o + gallows: and so forth.
Clearly there are two very large (and very general) problems intersecting here: (a) how to parse the Voynichese text into tokens, and (b) how to decide whether or not a given token is a null.
Nick: I won’t ask you to enumerate twenty perfectly (and equally) good explanations as I can appreciate that could be a job for you, though I would certainly be interested in what you think they are. I know you believe in forming hypotheses and this seems to me a very good one. Of course in addition there are other reasons that I believe the Voynich contains nulls. The o + gallows is much more speculative, I was not saying I necessarily think that is the case rather raising it as a possibility.
@Mark
Null words do not really affect the decryption. The only thing that happens is that the meaning of the text becomes visible only after 13 words as after 10 words.
It looks a bit different with zero ticks.
Especially on the VM text, the question arises.
If I already have to write a monotone text, with little different symbols, and add zeros now. Would I reduce the (possible 15 characters) by 3, then would not the actual text look even stupider?
Nick: Some other possible explanations of repeated words which I find less plausible:
1) The first word is meaningful and this valid word is repeated to confuse.
2) The author got confused or had problems with his/her eyesight.
3) They serve some transformational role on the text.
4) Repeating a word somhow transforms its spellling or meaning.
Nick: If repeated words are nulls then one would expect them to be in general common words or have some standard format. I can cross check this.
A little reading and it seems the repeated words are indeed common words, which fits well with my hypothesis. If they were rare words it would very much go against my hypothesis.
One hypothesis I read was that repeated words could be numbers. Like:
“One One One”
Which would be 111 or one hundred and eleven.
This however seems unlikely as one would not expect such large numbers. Also by observing the patterns of such words it should be possible to further assess the likelihood that they are numbers.
So I am still left with the conclusion that they are most likely nulls.
Mark Knowles wrote: “But very importantly it has lots of labels. Isolated single word labels are great, because we don’t need to worry about if spaces are really spaces or what is happening with the beginnings of paragraphs and neal keys and sentences and the like or whether they are simple words like “the”, “and”, “not” etc. Now it is possible, and I think sometimes likely, that a label might be a null word designed to confuse. However if it is not then, certainly in the context of the 9 rosette foldout in my opinion it is a specific word, in fact quite possibly unique in the manuscript. Now there are words that encircle each of the Rosettes and it is more debatable whether these are really isolated single words or constitutes parts of a sentence. There are a few other instances where it is possible that words follow from one another. However there are quite a number of examples which unequivocally are isolated single word labels. Now looking through these there are some which look like generic commonplace text, which could well be nulls, I think, and there quite a few which look like distinctive words with their own very specific spellings.”
Yes, they might be nulls, words, or parts of sentences. Do you really need 15 lines to list the three most likely possibilities? Don’t forget, most of the people reading this have already spent time studying the rosettes folio and have already given thought to these possibilities. You don’t want to insult their intelligence.
Extra words don’t make people look smarter. In fact, a lot of the time, it makes them look like they haven’t quite got it straight in their own minds and thus are waffling around the subject.
I’m not trying to be snarky. It’s just that your ideas are more likely to be read by others if you use three effective words instead of 20 repetitive or extraneous (or obvious) ones.
I think it worth being aware that most of the isolated single word labels on the Voynich 9 rosette page are much longer than the average Voynich word length. Which is what I would expect as given my map analysis it makes these words often refer to very specific geographical locations that would sometimes have long names.
When it comes to word correspondence on the 9 rosette page, in only a few cases, 7 to be precise, I have a clear idea of what I think the correspondence is. Now one might think given my claimed complete analysis of the page that I would find it easy to identify every word. However in most instances I can’t. The chief difficulty I find is that there aren’t arrows from the words pointing to what they refer to. Now one would assume the word relates to something in geographic proximity on the so called “map” to the word itself, however this might be a river or a mountain or a stretch of land or conceivably something tangentially associated. In the case of the 7 examples I think I can link the text much better. In the other cases I can only come up with a list, maybe a short list, of possible meanings. Without concrete meanings for most of the words now they are not much use, especially as there is the possibility that they could be a null.
So as an example;
If we have a word that we cannot read placed on a world map over Western Ukraine, well what might that say?
It could say “Ukraine” or “Carpathians”(Mountains) or “Dnieper”(River) or Lvov (City) or it could say something more general like “Steppes”(Flat land)
Another thought I had, which I daresay could be said to be obvious, is how difficult it is to often identify which symbol is which or whether there is a space or not. Not having looked at the text as a whole in great detail I would assume that the same problem is prevalent throughout the manuscript. Nevertheless it is somewhat frustrating as when trying to match real words to Voynichese labels it is often unclear what one is matching with what.
On the subject of nulls there is word “okar” according to voynichese.com which appears in quite a few places on the 9 rosette page and frequently throughout the manuscript. Now I think this word must be a null. I can’t conceive of better explanation. A big thing that makes me think that is in one instance it appears as an isolated single word label which means in can’t be a common word like “and”, “the” etc. It also appears very similar or identical to another word that I thought for other reasons was a null.
Now I am sure some would doubt what I have said. Nevertheless I am rapidly becoming more and more suspicious that the Voynich might be littered with nulls.
A fairly long time ago I suggested the following that there might be a crossover or middle ground between the theory that the Voynich is a hoax, i.e. nonsense, and that it is meaningful, this being that there is quite a lot of nonsense text mixed in with the meaningful text. This of course accounts for the fact that on the one hand the Voynich looks like made up text and the other hand it also appears like meaningful text.
The question would then become, what percentage of Voynich text is null?
10% or 20% or 50% or 80%
And if we do have a large percentage of null text how do we identify it and effectively filter it out to reveal the true text?
In the eventuality that null text constitutes a large portion of the Voynich, if it is filtered out it lays the way for a complete reanalysis of the Voynich text. Entropy and most of the other phenomena may be completely different paving the way for a genuine analysis of Voynichese as though for the first time, Distinguishing systematically between common words and null words could be hard, though maybe not be so hard. I begin to wonder if many nulls words have a common format, i.e. they often don’t look very different one from another.
It is conceivable to me that if there is a large amount of null text then once it is removed directionality, positionality and more could vanish. That is highly speculative, but not implausible to me.
Mark: the highly distinctive internal structuring (i.e. adjacent predictability of letters within words) and position dependency are both strong indicators that there are few (or perhaps no) nulls in Voynichese.
If you try to use nulls as a way to explain away all the problems with Voynichese, you’ll soon end up with an easily explainable blank sheet of paper.
Nick: I very much agree it is a lot to hope that many of the difficulties of Voynichese will be resolved once all null text is removed. Having said that the more I look at the text the more I become confident that of all the explanations that I can think of certain words must be nulls and if the 9 rosette page text is representive of the manuscript as a whole then nulls appear to be commonplace; so clearly in this respect our opinions differ markedly.
With respect to Voynichese repeated words are mentioned a lot, but it also seems we have clustered words more generally, so we might see something like:
banana orange banana orange banana orange
where an individual word is not repeated, but words alternate
Essentially we have the same kind of problem.
Et responsum est quaestio de quanto ad Dóminum = 0%% nulla verba.
Just quia Verbum non intelligunt, aut non richtich Decoded, non sequitur propter hoc quod significat non aliquid.
Et nunc legitur vetus German tot scriptis, ut numerare in orthographiam recensere, ut hoc intelligere nisi quinque omne verbum. Et Latine spectat circa idem.
Cum nescio quid verbi quod you mean “akar” exactam possition.
Sed putant, sed omnino verbum facere conari verbum ‘akar “.
Quod si putatis suus ‘scriptum ut’ ubi ‘seu OKAR Okor “.
Cum centum scio temptatis prius per ‘o’ sit Verbum antennas. Cum est normalis LX% + mihi claves ad quod iam Trefferquto. Non materia quantum temporis est verbo.
Scio enim latine K nimis rara, sic suus ‘magis of a “C”.
Ego autem verba, “per cor” aut “cir” Nunc est ad LXXX% rate.
EGO coniecto ius nunc gravibus est. Top Ribbon ius in urbem ad extremum verbum in spiram. Is vultus amo (oPaxAg) (in optical, loop = x, = P A).
Ego semper putavit verbum significat “est posyum” quod esset nefas!
Hoc appellatur “enim possum” User in Germanica et orthographiam “Z ‘in duplici” ss “. Pi = I.
In quibus verbis non habet syllabas vocant, qui complexio, ledo XC% rate.
Cryptology est rem ridiculam, in Sensu spirituali intellectis. Prima quidem hic non.
I must admit to having suspicion whenever I see a word starting “ok” or “ot” as these are so prevalent and I already think some amongst them are nulls, of course that does mean they all are. Some Voynichese words clearly appear to have their own distinctive structure, whilst others appears monotonous and very similar to each other. (This is just based on looking at the 9 rosette text, which I am trying to digest at the moment.)
Mark: trying to generalise 200+ pages of text from solely the small amount of text you find on the nine rosette page seems like an extreme case of the tail wagging the dog. And inferring, from your inability to see how labels on the nine rosette page map to the specific place names your theory predicts, that those labels must include nulls seems like about four or five simultaneous leaps of faith more than the evidence justifies.
I’ll freely admit that Voynichese presents us with a very specific challenge when it comes to making sense of the gallows characters. But it also does that with or/ol/ar/al groups, an/ain/aiin/aiiin/aiiiin groups, e/ee/eee/ch/sh groups, qo- pairs, and -y/-dy groups. If we explain away all of these difficulties as having been caused by nulls, we’ll lose 90+% of the text.
Nick: I thought it instructive to focus at this stage on a manageable amount of text that I can absorb. Obviously this may not be reflective of the totality of Voynich text. I also chose this text as it is a page that I know very well, which I felt could help me appreciate the text better and likewise help me appreciate the page better.
You say:
“inferring, from your inability to see how labels on the nine rosette page map to the specific place names your theory predicts, that those labels must include nulls”
That is not what I have done. The two labels that I argue are nulls happen to have exactly the same word, so it has nothing to do with the specific place names my theory predicts and that word is repeated frequently throughout the manuscript and that page specifically, often in sequence. So my argument that they are nulls is built on firmer foundations than your statement could lead one to believe.
Nick: I am not arguing we should explain all the difficulties we have by resorting to nulls, that’s a bit of a straw man. In my mind, whilst it is certainly not what I currently think, I deem it just within the realms of possibility that 90+% is null.
I was thinking, I am considering the notion that they are two groups of Voynichese words the “monotonous” and the “distinctive”. The monotonous words have relatively similar spelling to a number of other members of the group, so their spellings are grouped into different clusters, and are very common and often are repeated. The distinctive have a wide variety of different and fairly unique spellings and are not repeated and less common.
Now it would be premature to argue that the monotonous correspond to nulls and the distinctive to real words, though that would seem to be the implication of this kind of speculation.
Mark: that would seem to be more your inference than an implication.
Nick: Fair enough Nick, I should have phrased that sentence more carefully.
Thinking about repeated words. If there was only one “word” in the Voynich manuscript and that words was repeated 10000 times, what information could realistically be contained in the text, not much it seems to me? It seems to me that with the very common, often repeated, one letter different words we have a not so different scenario. I will look around to see if I can find another explanation that others have provided which is nearly as plausible as the idea that we are faced with nulls with these words.
If these words are all nulls then how could the author recognise them when writing and reading the text? I guess having a common format will help, so as their spellings are similar they are easier to remember without having to spend a lot of time looking them up from a cipher key. So if deciphered what would they say, would they be meaningless jumbles of letters? Or some variant on something like “blank” or “nothing” like “blonk” or “nathing”? Or could they spell a variant of a real word like “cabbage” or “cobbage” or “babbage” which the author can tell from the context to ignore? Is there something within a word like a couple of symbols flagging it as a null?
It would be easy, as has been suggested, to when in doubt mark a word as null, but of course that would be ill advised. Nevertheless I think one should not assume that there are not a lot of nulls, even if that is a less attractive conclusion. I guess my approach would be if a word satisfies a variety of criteria it should be put on the possible null list, as before, these being:
1) Is the word repeated?
2) Is the word common?
3) Does the word have a very similar spelling, such as by one symbol, to other words that satisfy 1 & 2?
Other questions that could be asked:
4) Can the same word often be found in the same sentence or sequence of words even if not strictly speaking repeated?
5) Can the same word be found adjacent to a word with one symbol difference?
These are closer to being an objective set of criteria, though I daresay they can be improved on. Still this kind of process could help to separate Voynichese into 2 classifications of words if that is advisable
If Voynichese can be split into two sets of words it should be appararent rather than one set verging into the other with an unclear no-man’s land in the middle. One would expect them to be largely distinct and different groups of words. Obviously there may well be common words which are not repeated or with similar spellings to other such words, so whilst having one shared feature they lack the others and therefore do not belong in that grouping.
If the set of words that satisify those criteria turns out to be very large then so be it.
If I develop a various precise well-defined set of criteria then I could implement an algorithm to partition the words into 2 sets. It would probably help to score words to see how well they fit the criteria this would help to identify if they are 2 distinct sets or a continuum. I will have a look more closely at the text so that I can define the criteria with greater precision.
Obviously my perspective is that words that satisfy the criteria would seem to likely be nulls.
Mark: By some reckoning, too many ifs might be a few too many to contend with in problem solving. Though that didn’t seem to worry Kippling who settled for lucky thirteen. Didn’t get him a knighthood but becoming poet laureate was next best thing…. and what’s more my son, he got to walk with Kings, yet keep the common touch..
John Sanders: You make a good point. I was trying not to be too emphatic, so as to avoid the accusation of unproven assumptions.
I am not suggesting that 90+% of the Voynich is null text, I would need quite a lot of evidence to start to believe that. However I think in general we should not be afraid of reaching radical conclusions where and when the evidence supports a radical conclusion. I think there is a tendency to reject the consideration of these kinds of radical conclusions out of hand rather than keeping them on the table of possibilities.
Mark: Radical conclusions are not consistent with time honoured, logic based concepts developed and adhered to by the designated experts amongst us, who would proclaim total rejection of anything we lesser lights might offer that does not support the all knowing intelligencia’s rigid 1421 VM birth concept. I’m thinking that things may be about to change to enable the introduction of more acceptable less radical possibilities to emerge.
Looking at the 9 rosette page it is crazy the number of Voynichese words that start “ok” or “ot”, similarly there are common endings. As a hypothetical if every Voynichese word started “ok” then it would be superfluous, so the degree of monotony in some Voynichese words really means that there can’t be that much information being conveyed to the reader in those cases.
Mark: you’re not looking at this part of the text in the right way. If the Voynichese on rosette / label pages has more ot- / ok- tokens, the right questions to ask are about what it has correspondingly less of. If we can see what ot/ok are replacing, we might gain an idea of how one maps to the other.
Nick: Thanks for your comment.
What I notice is there are quite a few words starting with an “o” followed by a gallows character, especially “t” and “k”, then there are more interesting words, which may start with a gallows character or may have a gallows character in the middle or near the end or two gallows characters or not at all, rather than as is common the 2nd character. It also seems on first inspection that these words are typically more homogeneous and more likely to fit my “monotonous” criteria. In the more interesting words characters like the table symbol (also seen in diplomatic ciphers) occur more frequently as does the bench character with the curl ontop. It is true that “89” or “9” are common endings, though often in some languages common endings are much more typical than common beginnings.
Obviously as I am looking at only one page this is a small sample; still just with this page it feels like there is a lot to get to grips with.
Anyway Nick, I would appreciate it if you can enlighten me as to what it has less of.
One could argue that beginnings of Voynichese words like ot and ok could be something like genders. “Le” or La”, “Der, “Die” or “Das” etc. however this still purposes a problem with regard to their monotonous nature.
Nick: It has correspondingly lower entropy, but I don’t think that is where you are going.
Nick: Viewing a few other pages at random there appear to be quite a lot of ot and ok words there.
@Mark
Could not it be easy that you do not consider the combinations? Deception is a factor in cryptology.
They almost all look alike, but are they really?
a the, from
it, a it it, from it
at, a at but, at the
ad, a ad for, by the
ut, a ut so, from that
et, a et and, by the
id, a id that, from this
I have spent some time studying the text on the 9 rosette page and I feel that there is much much more studying of the text on this one, all be it very large, page for me to do. (I should say in addition to the specific study of this page I have and will continue to make use of general results about the manuscript text to assist my 9 rosette text analysis.) So far I would be very surprised if some of the text is not null i.e. meaningless. Similarly I would be very surprised if some of the text is not meaningful. So the percentage of null text is greater than 0% and less than 100%, I think. To hazard a wild guess as to what percentage of text is null on this page I would say about 60%, however I could easily radically revise this figure in the future.
I think once I have fully analysed this page I will then, likely, be able to recall all the text on this page accurately, that is how intimately I need to know the text, which is slightly daunting. However I know the rest of the page that well, so its not that extreme, the text has always been something I have been unfamiliar with, so I am just filling that gap in my knowledge. It seems too much to hope that a study of just the text on this page will nearly be sufficient to decipher Voynichese, but I think I have found it and will find it very enlightening.
Whilst I am not overly fond of the idea, I suppose it is possible for a word to be meaningful and a null. So a word has a geniune meanng in the context of a sentence in one instance and in the context of another sentence serves as a filler with no real meaning.
For example:
I like my new hat. I am going hat hat to the shops hat now hat hat hat.
In the first sentence the word “hat” has a real meaning. However in the second sentence the word “hat” acts as a filler.
I don’t like this interpretation as it can easily lead to confusion, but more importantly doesn’t fit my initial observations about the text. Nevertheless I think it is a possibility.
If there are Voynichese “words” that represent nulls then can those “words” be deciphered even if the decipherment results in a meaningless series of letters.
As I have mentioned if there are a large number of Voynichese words that act as nulls how in practice can the author look them up, so as to know they are nulls when reading or can be used as nulls when writing? The explanation that seems most likely to me at the moment is that the null words are frequently interrelated i.e. by differing in spelling by one symbol from other null words or less likely have some flag within the word that indicates that it is a null.
That was really a stupid example. Have you ever thought about the sentence definition?
For me where German writes, the English have a wrong sentence and twist everything.
It seems to me, having looked at the 9 rosette text, that claims of positionality appear to have been overemphasised. Whilst it is true that generally there are normally not many different word endings, though in some cases we do have unusual word endings, it seems clear that other symbols can appear anywhere within a word. (I am unsure as to whether any symbol can potentially appear at the end of a word.) It seems to me that positionality looks to be most marked amongst the more monotonous words, but not necessarily a universal feature.
Mark: that’s what happens when you try to scale up observations from less than 0.5% of the total corpus. :-/
@Nick
Maybe I drank too much yesterday. But you could write something like the 10 commandments of the VM.
1. You should not claim to have solved it, if you only have 1 word.
2. You should not talk about Aliens, if you do not know which star they come from.
3. Etc. 🙂
Nick: I am guessing that by positionality we mean a symbol or symbol combination only ever appears in a given position within a word. Or do you we mean that a symbol or symbol combination typically appears in a given position within a word?
If the former then one geniune counterexample would be sufficient to demonstrate that a symbol can in some instances be found in a different position. If the latter then this is not true positionality, I think.
Peter: that’s not fair, now I will have to drink too much as well so that I can finish off the list. :-#
Mark: there you go with your straw man arguments again, building up something false so that you can easily demolish it. 🙁
In the context of the Voynich Manuscript, “positionality” is taken to mean that certain letters have an unusually marked tendency to appear in certain places in words, on lines, and on pages. While less than absolute claims, these very strong tendencies are almost entirely atypical of how language works, and so are difficult to reconcile with ‘pure’ linguistic explanations for Voynichese.
It is possible to construct fragmentary explanations for part of these, e.g. line-terminal ‘-m’ words could possibly be hyphens, linking the last word of the line to the start word of the next line. But as for explaining them all away at the same time, I don’t think that’s presently possible.
Nick: It is not intended or in actuality a straw man argument, but rather an exploration of the notion. I am interest in learning about the subject and exploring the various arguments. I have no interest in demonishing arguments purely for the sake of it, that seems pointless, so why assume I do? I find challenging arguments, especially widespread ideas, a valuable activity.
I was perfectly happy to accept the notion of positionality, but am now beginning to question it more having looked at a certain amount text, all be it a small part of the total.
When you say “unusually marked tendency” from my preliminary findings within the framework that I have been exploring that could be indicative of the patterns found in formation of “null” text rather than patterns in real text; I should say that I am not asserting that that is the case, but rather that presenting it as one possible hypothesis which may explain the phenomenon, at least in part, as it seems more evident in what I have begun to term the “monotonous” text.
As I think you know, I don’t subscribe to ” ‘pure’ linguistic explanations for Voynichese”.
All this of course remains speculation and my study of just text in the 9 Rosette page remains at its early stage and so my thinking is very far from crystallised. I accept that there is obviously great value in working with the text as a whole, but I would argue there is real value in addition in studying a smaller amount of text in minute detail.
Mark: working with a small sample inevitably exposes you to the perils of overfitting (making stuff fit) and daydream delusion (wishing too hard that stuff would fit). Adding arbitrary nulls into that SSS (small simple size) mix is only likely to encourage you to go further along perilous paths, sorry. 🙁
Nick: I don’t think I am guilty of having a “daydream delusion (wishing too hard that stuff would fit)”, I like to think I am more objective than that and trying to arrive at truth. I have I said I am and will be making use of general statistical data and patterns pertaining to the manuscript as a whole, in addition; however I would be pretty confident that nobody has studied the whole text in minute detail, so I believe there is merit in studying part of it in that way to trying to glean aspects which an overview could easily miss. I should say that once I have finished my minute study of this one page I may extend it to a couple of other pages of the manuscript, but many more would be unmanageable, I think. Also one use of the 9 Rosette page in this respect is that it has a large number of labels, which I am inclined to think probably offer more insights than sentence text. If I look at other pages I would likely choose others with a preponderance of labels as well.
I don’t think identifying nulls represents a perilous path. I have a variety of different and I think good reasons to suggest there is some null text on the page, for which I cannot think of or have read of an alternative explanation, some reasons relate to general facts about the text as a whole, The nulls I have added to my very likely null list are added for far from arbitrary reasons. Given that I think some words are very likely nulls the natural question for me to ask is how deep the rot extends, something I do not yet have a clear idea of. You are welcome to argue that there are no nulls in the text, but you shouldn’t assume my arguments are based on a whim, I am merely trying to find what I think is the most plausible explanation for what I have observed, I have no vested interest I arguing for a greater or less preponderance of nulls. When I have finished you can, if you wish, look at my complete arguments and tell me why they are wrong and offer alternative explanations.
Mark: the one place where nulls show themselves to cryptanalysts is in contact tables, where a null would typically (though not necessarily) be expected to have an unstructured adjacency profile (i.e. its usages are not correlated with any particular letter shape before or after itself). However, this is almost exactly the opposite of what we see in Voynichese, where the (apparent) adjacency ‘rules’ are very strongly structured. So the default cryptanalytical position with Voynichese vis-a-vis nulls is that, unless you have some extraordinarily strong argument, there aren’t any.
From my perspective, the only character usage I currently suspect might be a null is a line-initial ‘s-‘ character within a paragraph: but of the 4311 word-initial ‘s-‘ characters (according to voynichese.com), I can’t imagine that many more than 150 of these are null candidates. And that, out of the complete Voynichese corpus, is a very small number indeed. 🙁
Nick: Clearly we have a very different perspective, but ultimately it is up to me to provide strong arguments as to why I believe certain words are nulls, which requires me to delve deeper into the text that I have been looking at. I doubt I will be able to provide an “extraordinarily strong argument” for this as there is very very little about the Voynich for which such an argument can be provided,
Do clarify, if you wish, what you believe are the “(apparent) adjacency rules” that preclude the possibility of nulls. I should clarify, though I have explained it before, that my focus is not on null symbols, but rather on null Voynichese “words”. I am not saying that there are no null symbols or symbol combinations, just that so far I have no reason to believe that is so, though there could possibly be symbol combination flags within a word that mark it as a null.
Mark: adjacency ‘rules’ shape the contextual probability curve, e.g. the probability that EVA ‘q’ is followed by ‘o’. But if anyone were to suggest to me that, even though they can’t read a word or even a letter of a text, they are able to tell that a specific word can only be a null word, I’d say they were dreaming.
Nick: The adjacency ‘rules’ you describe are clearly not problematic in terms of null words.
Well, clearly from your perspective I must be dreaming, obviously I view things differently, but it is certainly a subject I am and I will continue to explore. It seems to me that identifying null words is a lot easier than reading text. Anyway I know I will have more to say to elucidate my ideas as they develop.
Nick: The only alternative to the null word hypothesis that I have heard of, that sounds vaguely plausible to me, is that repeated words represent numbers. However that hypothesis only explains some things and completely fails to explain other things it seems to me. I have no preference for any specific hypothesis that will explain the behaviour of these certain words, but the “null” hypothesis seems to fit far better than any others that I know of or can imagine.
Mark: you’ll have to stop saying “can imagine”, it’s not a flattering look. :-/
Nick: Fair enough, I will try to use that less. I am just trying to be honest flattering or not. My purpose is definitely not to flatter myself. If you can think of a better explanation I would love to hear it.
One thing I can think of is the scenario where you have the same word being different things in different context. Such as:
hse -> house
But also
hse -> horse
And
hse -> hose
But to me this is a crazy scenario and it doesn’t really explain my observations either.
Continued..
I don’t deny the possibility of abbreviations, but if there are a lots of interpretations of a words it would make reading Voynichese difficult for whomever is reading it.
If all the repeated words are numbers then it appears we have many more distinct numbers than would be needed. Also having more than one label that just has the same number attached to it would seem, from my observations, less than useless. Then we are faced with how to interpret the various numbers in the different instances. So for these reasons and others the number hypothesis doesn’t seem to stack up.
Mark: enough with the straw man arguments! You keep on summarizing things badly and then (unsurprisingly) dismissing your bad summary, all of which makes you look not entirely unlike an idiot. You can definitely do better than that, so can you please try, hein?
Nick: A better summary of my suggestion of multiple word interpretations or the number systems explanation, which explain repeated words etc., from you would be welcomed, if you choose to provide it, though I appreciate that you are a busy person.
Mark: both the multiple word interpretations from abbreviation hypothesis (where the plaintext is abbreviated for the purposes of reminding the encipherer of the plaintext) and the (specifically where dai[i][i]n = Arabic digit) number system hypothesis I proposed in Curse back in
20162006, so I’m not really sure I can do much better than to refer to that.There must surely be at least ten other viable explanations for each, if you were to cast your imaginatory net a little more widely.
Nick: One thing which would be nice is if we all can refrain from the ad hominem. Whilst I appreciate that it is your blog and so you can say and do whatever you like, it doesn’t seem quite right to use terms like “prick” or “idiot” when I think you would censor them and probably more if they were directed at you, you have censored me for far less. Anyway throwing insults are a waste of both of our time. I am sorry if you feel that I have misrepresented something, if I come across an advocate for a given position I will certainly asked them to expound on their perspective.(I think you should note that my point regarding the multiple word interpretations was not an attempt to criticise your abbreviation hypothesis, but rather an attempt to briefly explore an argument that occurred to me which could in theory explain some of what I have observed.)
Mark: when you stop mis-summarizing really very basic aspects of Voynich research, stop building up (then tearing down) your own straw man arguments, and stop attacking other people over basically nothing (all three of which behaviours I keep asking you to stop), the powerful urge I get to call you an idiot will surely go away. And it will be better for everyone.
Another possibility that I can think of is that of some whole words mapping to individual letters. However this would still mean that certain labels may consist of a single letter, which would be next to useless for identification purposes. And also that would not help much to explain words repeated 3 or more times.
Nick: Someone is always welcome to point me to what they regard as better summaries or provide better summaries themselves. Being told my summary is wrong without providing any specifics doesn’t take me much further forward.
As far as attacking other people that is quite wrong. I attack arguments not people, though very vigourly at times. Attacking arguments not people is perfectly legitimate in my opinion. I never use words like “idiot” etc., which is attacking a person. I have no interest in getting into arguments over nothing nor do I think I do so. The above argument has come out of you disagreeing with me, which is all fine and good, not my trying to start an argument at all.
I think you are really exercising a double standard in this regard, which is your prerogative, but it puts me in the position of being unhappy to be on the receiving end of unnecessary insults. You can disagree with me or be critical of my approach as much as you want without the need to add personal insults. I don’t think it would be right to allow myself to be consistently insulted. If I tolerate that I am effectively giving someone the green light to insult me directly as much as they like and however they like, nobody should accept that. This all puts me in a difficult position, if you think that behaviour is completely legitimate for you then it means by me commenting here I am accepting of that. I guess you will make it crystal clear where you stand.
Nick: Do you have site rules or specific rules that you apply when vetting comments?
Mark: you have attacked JKP here on a number of occasions in ways which I would describe as ad hominems, which are basically red card offences, moderation-wise. And you certainly don’t help yourself by repeatedly misrepresenting people’s points of view. I flag these things to you repeatedly, because I hope you will come to leave better comments.
But all the while you repeat the same wrong-headed behaviours, I have the choice between deleting your posts or calling you out on them (and sometimes calling you an idiot for it). If you’d rather I do more of the former than the latter, that’s fine by me.
Yes.
But that doesn’t mean I have to tell you what they are, or that I have to rigidly follow them.
Nick: The truth is that telling me why you think I have misrepresented or misunderstood something or why my argument is wrong is the most useful thing to do. If you don’t time for that then that is fine, but calling me an idiot is not going to help much.
Mark: once again you misrepresent things. I tell you what you have done wrong AND I call you an idiot. Neither seems to be working.
Nick: One thing I ought to mention is that I think you are under the impression that I have been summarising your ideas from the Curse, that is not what I have been doing, other people have expressed similar though different ideas about number representation. I have just been trying, primarily, to address how I make sense of my observations by exploring different hypotheses. Your own ideas, from my recollection, don’t explain these observation and probably do not claim to explain them. (I only just saw your dai[i][i]n comment.) Null words still seems to me the best explanation and a perfectly reasonable, if unsatisfying, notion.
Mark
How many times do English words come before where to start with th?
If I see the same words in succession, the first guess is that a punctuation mark is missing.
There are no visible in the VM. xxxxx that, that that xxxxx.
That without including a possible configuration. There are languages where this takes over the repetition of a word. Big big big, big, bigger biggest.
Why should it be numbers? It would always be schnapps numbers. 123 123 123.
You write something about feeling, what counts is experience.
You should urgently read old books to get a feel for it, and not fixate on an idea.
By reading things I have noticed, to which not even one thought. Or how is it possible for a person to miss the same word 4 times? But I only see that when I read the whole text.
Once you have an idea, edit it for the first time. If you have survived, and do not end up in a dead end, you can still write about it.
Mark: given that you clearly didn’t understand what I wrote in 2006, you’re not giving me much incentive to explain it again in 2019.
On the 9 Rosette foldout page each circle has a ring of text around it. Well this text could be a series of independent labels. However given the extent to which labels are repeated on a circle this makes little sense to me. If this text corresponds to a sentence or sentences this presents another problem in that it fits perfectly around the circle with no or almost no blank spaces. There is no evidence to suggest that words on some circles are spaced out more to fit all the way around. So if we have a normal sentence or sentences then why does it always fit so neatly around the circle? On some circles we see repeated words or alternating repeated words. If I wanted to write a sentence or two about the contents of a rosette, would it seem likely that It would always fit a given number of characters so closely. When I write a text message do I always use the almost 160 characters available? This again points me towards the idea of filler or null text.
It is my observation that there is a smaller proportion of “distinctive” text overall on the circular rings than amongst the specifically added blocks of text such as we see in the bottom right rosette or in the spiral in the top right rosette or amongst the vertical list to the left of the top left rosette. It makes sense that there would be less null or filler text amongst the text that has been specially added rather than added part of a general pattern or structure. Obviously this observation could only be precisely demonstrated with formal statistics, but anyway that is my observation.
A question that strikes me is when are labels really labels? Now this seems pretty obviously on the face of it. Well when text is in complete isolation from other text that would seem obvious, though it could be argued that even the most isolated text is linked to the nearest text to it, however this seems pretty far fetched to me. There is some text however on the 9 rosette page which could be argued is part of a sentence or an isolated label, in some cases the argument one way or the other is stronger. Text being part of a sentence helps argue away its oddness as one can say that the text can only be interpreted as a block. When a word is an isolated label, but that appears in different contexts and is common and repeated with not much similarity to other words it is hard to argue I think that it has a real meaning and is not just a filler or null text. The alternative is that that text is meaningful, but its meaning is so irrelevant to the context or general that it is hardly worth writing; I don’t find that plausible.
Just been looking at other labels on other pages and what is perplexing is to see the same label in a completely different context, which inclines one to think that either one has null text or at least two different words can be written the same way, which would normally make it infuriating to interpret what the given word is; it would merely act as a clue to what the word might be. Alternatively the word is so general as to seem to me likely fail to communicate much useful or interesting specific information given the varied contexts
So far whenever I look at circles of text around zodiac, rosettes etc. these circles appear to have text that is consistently spaced and of a consistent size from one to the other and encompass the whole circle. If these are a series of separate independent labels then each label could be adding extra information. If however these consist of sentences going around the circle one would expect wider spacing or a gap at the end of the sentence. It feels as if the author, in this situation, deliberately fills the available space with as much text as possible, almost as if it was more of a design feature than meant to impart information. This fits in my mind with the idea of some kind of filler text.
Mark: You may not realise it, but you do make a very com Pelling argument, for a proposed artificial language based on algebraic propositional truth logics. Whilst not necessarily inconsistant with the relevant depicted images, I doubt that it gives a sensible account of what the individual diagrams represent in real life terms. Good thing is, it can likely be computer translated given the proper input criteria and might also be readable in a common language allbeit in a confusing text.
John Sanders: I am primarily at the moment just trying to make observations and reconcile them.
If the text around the circles is just a series of independent labels with no sequential relationship then repeated labels seem to make little sense. In the case of a one to one mapping between the “words”/labels and items specifically within the circles i.e. at the same position/angle this doesn’t fit as in the case of some 9 rosette circles there can be no one to one mapping.
One thing that may or may not go against my argument is that on page 114 (not sure which folio that is) there is a “^” on the top left of the outer circle and a word has been added, which means I think that there can’t have been sufficient space for a necessary word. I think it is highly unlikely that that is a null word unless the author was incredibly sneeky, which does seem far fetched. Also this word is just after what looks like a separator, which seems very odd as if a sentence then it would be the first word, whereas an extra word would be the last word in a sentence, unless of course the author realised that he/she had omitted a necessary word at the start of a sentence and later wanted to add it.
Again and again the text fits perfectly around the circles in the Voynich at the same size and width of spaces, how does one engineer such a perfect fit in so many cases without fillers? In a few cases the text on the outer circles is segmented and clearly separated and maps to segments of the inner circle, which makes it is seem less likely that in the other cases we have a series of independent “words”/labels. Could it be that they are separate “words”/labels and yet at the same time interrelated without being parts of a sentence? It seems hard to “imagine” at this moment quite how that would work and why the precise number of labels that fit perfectly would be required.
I think it is worth noting that on pages with large bodies of text i.e. “paragraphs” the text rarely runs to the end of the line which would lead one to believe that enough fillers in order for the text to fit perfectly were not added at the end, though I don’t think one can conclude on that basis that no fillers were used in that text as a whole. This might lead one to conclude that the last words in the “paragraph” were less likely to be filler words.
In the odd case in different parts of the manuscript the last line of text is deliberately not left aligned where it could be, which could lead one to believe that the last line constitutes some kind of concluding statement or signature.
If one is trying to map Currier A to Currier B and one has null “words” that I would I think rather throw a spanner in the works as trying to map one null word to another would not appear to me to make much sense.
In fact the presence of null words would throw a spanner in a lot of Voynich analysis and thinking.
Slightly oddly there are 3 lines at the top of 58r all of which have not been left aligned where at least two of them could have been; this could easily have little significance, but I think it is worth observing.
What’s with the adding of text is red ink on 67r? Is it is some way more important and so needs to be emphasised?
Having looked at quite a lot of gallows characters I would say that in many cases, if not all cases, the left side must with practically 100% certainty be a “4”, so we have for example “4P” rather than “qp”. It is possible, though I think unlikely, that in a few cases we have a “q” not a “4” and therefore a separate and different character. The left side is almost always much more angular like a “4”, whereas the right side is much more curved like a “P”. Systematically going through a large number of gallows with a left side should be sufficient to demonstrate this conclusively. I think this can have some significance as it gives a bit more insight into the origin of Voynich symbols. I would say that this is an obvious conclusion, though I would expect this to be disputed.
To determine whether the top left of the gallows character is a “4” or not is really a debate as to whether we have a triangle or a circle and the same applies to the right hand side. So the question then becomes do we have a series of triangles or circles in each case. So how do we answer this question rigourously? Well, we could implement an algorithm to do some formal collective image analysis and could come to some kind of aggregate result, though I think this is really overkill as it should be pretty clear on looking at enough characters.
I can see a number of people being disappointed by the idea of null/filler text, if that is indeed the case, as it kind of negates the idea of a pattern in that text, though not the text as a whole, and just makes it arbitrary and meaningless. For that reason I can see why some people could be keen to reject the notion, though if this idea of null/filler text is incorrect they would of course be quite right to reject it.
I think it worth observing that it is not uncommon in diplomatic ciphers to see symbols like:
c
cc
ccc
cccc
where these have no connection to numbers whatsoever they merely map to different letters of the alphabet in a fairly arbitrary way, though it is not unknown for them to map to adjacent letters, which is really a reflection of the laziness of the author.
So whilst numbers systems like this seem enticing and rather clever there is a danger of reading far too much into this. I am of course not saying that there are no numbers in the Voynich, but rather cautioning about building a system around such ideas without strong evidence to support them.
If one looks at the outer circle text on the centre right rosette in the top right corner one can see a 5 symbol long word repeated 3 times and surrounded by a couple of words that are very common and that I have already argued are nulls for other reasons. As with so many cases of outer circle text this fits neatly all the way round the circle. This is illustrative of the kind of problem that I see that makes me conclude that we have some null/filler text.
An alternative that would fit is if the author wrote what is largely gibberish, this would function in the same way as null text in the sense of it being meaningless text. However I don’t believe, for a variety of reasons, that the author was the kind of person that wrote nonsense.
When I have to fill out forms I, like I daresay many people with large and maybe even medium sized handwriting, can sometimes find it hard to fit everything I want to say in the available box. This tends to result in either two situations, either my writing gets progressively smaller so as to fit everything or I end up continuing outside of the box. Now bear in mind that of course these are rectangular straight boxes not curved boxes going around a circle. Now trying to fit a specific meaningful text that I want to write with a constant size of writing and constant spacing and no segmentation, so that it fits perfectly again and again circle after circle with no filler seems to be quite a feat. Then again I am sure it could be argued that my failure to do so is just a function of my idiocy, but it seems without null or filler text of some kind then you are left with all the text round the circle being fairly meaningless, especially when you have repeated words.
I think if it is accepted that there are null or filler words somewhere in the manuscript it then opens the door to the question of the overall extent of null text.
The circle text could represented some strange incantation such as:
sun air fire fire fire water air ….
Which is in a practical sense meaningless, but the words have a meaning, though even an incantation ought to have a predetermined length and not just continue until the space runs out.
Anyway I think the author was more sensible that to surround the circles with nonsense text. Ultimately this could lead us to conclude all the text is nonsense, which I don’t think for a variety of reasons. So I am inevitably drawn back to the idea of null/filler text and it can be found in some places then it can very likely be found in others.
Mark wrote: “Slightly oddly there are 3 lines at the top of 58r all of which have not been left aligned where at least two of them could have been; this could easily have little significance, but I think it is worth observing.”
Not odd, but very common. It’s space left for the rubricator to add an initial (e.g., a “red weirdo”). It’s done in a completely normal way, and the fact that it’s at the top of the page in a new section (in the sense of the content being different from the previous page), reinforces the likelihood.
It’s possible the VMS was never finished and so the initial (or small iconic illumination or whatever was going in there) never got added.
And yes, it has been noted… by many people.
Mark wrote: “What’s with the adding of text is red ink on 67r? Is it is some way more important and so needs to be emphasised?”
That’s usually what red ink means in medieval manuscripts (red pigment was more expensive so it was often reserved for emphasis, plus, it stands out more, which is also suitable for emphasis).
JKP: Thank you very much for your comments on those to questions, though it should be noted that those were the most minor points that I touched on.
@Mark
You mean the 3x “oMo8g”?
Interestingly, you write about the Gallows Divergence, but you do not use them.
You’ve written a lot about theories, but you yourself have never really worked on one of your theories.
It takes an awful lot of time and patience, and many are trying to bring some order to the Caos.
Look carefully, the signs are not the same!
oMo8g oDo8g oMo8g
amatum adatum amatum
amatus a datus a matum
beloved by a mature audience
I’m not sure yet, but that’s about how it works.
There is still much that I do not understand, but I do my work.
Everything is the logical deduction from the word “Taurus” and a lot of time and work.
Peter: I have producing a theory about the 9 rosette page. I have explored ideas about how the Voynich cipher could have roots in diplomatic ciphers of the period.
At the moment I have not been constructing an overarching theory, but trying to make observations and reconcile those observations in what seems to me the most plausible way. Saying that I think it is most likely that there is filler/null text doesn’t explain what is going on overall and that is not what I am trying to now.
The problem that I have at the moment is that how could the author distinguish between null text and real text. If there is null text then it appears to be repeated on multiple pages of the manuscript, so the author can easily identify it. How does he/she do it? I could be a list of null words, but this can’t be unmanageably long. It could be a list of word variants i.e. general format, which would account for single letter variation between words. It could a standard flag e.g. every null word start with a “o” followed by a gallows character. Or it could be some combination of these. Without more specificity vague talk about null/filler text seems like just that, vague talk, though even if one cannot identify how it works one could identify thay it exists. Without having a good way to identify it one cannot filter it out from the text to reveal the true text, but anyway this is still early days.
Mark: While your reconciling with the left gallows 4 loop, you might like to swing by Wiki and check Ethel Boole’s half gallows F in her “Faithfully Yours” sign off, then tell us what you think.
John Sanders: Do you happen to have a link?
It is a very strange thing, that I have only just learnt, that the father-in-law of the man who rediscovered the “Voynich” manuscript was George Boole the really famous Mathematican and Logician, in so far as Mathematicans and Logicians are famous. (For anybody wishing to tell me that I haven’t made a new discovery and that this is already well known to the Voynich community I am not presenting this as such.) That is just another demonstration of the fact that coincidences abound and that we should not necessarily read to much into what could be a coincidence, though of course we should also be careful not to dismiss things as coincidences where there is a real phenomenom; its hard to walk that line.
Mark, Ethel Voynich nee Boole wrote the novel “The Gadfly” and the Soviet apparatus published many copies of this turgid tome, and paid her handsome royalties. The more cynical of us would suspect that the royalties that flowed into Ethel’s bank account found their way to KGB/GRU clandestine activities in the USA. Wikipedia says that the book is “set during the Italian Risorgimento, (and) is primarily concerned with the culture of revolution and revolutionaries.” At least that computes.
I don’t know of the USSR ever paying royalties to anyone else.
Byron: I must admit I have paid very little attention to Wilfred Voynich, so all this is news to me. I don’t believe he wrote the manuscript that bears his name, but from all I have learnt about him he and his wife clearly had a interesting life. There are some subjects directly or indirectly related to the Voynich that I have paid very little attention to and am pretty ignorant of, such as the connection with Emperor Rudolf and herbs; I have always had quite a narrow focus.
Mark: Don’t be mistaken, don’t be misled and don’t allow yourself to be stifled in your efforts to gain a better understanding of the amazing Boole family, including first and foremost, Mary Boole, who taught Wilfrid Voynich everything he knew about the antiqarian book world. Knowledge which she gained whilst employed as a low paid librarian at Queens College from 1864, when she became George’s widow to around 1880, when she was sacked because of her many unpopular progressive views, including the teaching of Boole Algebra in her own inimitable style. She mentored her five home taught polymath daughters and grand children, including the likes of physicist Sir Geoffrey Ingram Taylor, maintaining her rage against intransigent and ineffective learned establishment principles, until her passing in 1916. You have been granted opportunity, allbeit through co-incidence, as you say, to look into this remarkable family at greater depth; take it and I guarantee you will be in a better position to understand what Vm may be all about. Though I’m not suggesting that any changes to your thinking will be looked upon as enlightening or game changing to the die hard status quo.
I don’t claim to have had access to Lily’s bank account, but as she didn’t receive much in the way of Soviet royalties for her book until she was in her late eighties, I doubt whether it would have sufficed to change her known anti Bolshie views, or to affect her rather sparten lifestyle in mid town Manhatten. In her younger days, as a committed revolutionary, she was unhappy, to put it mildly, with the capitalist Czarist regime, championing the cause for its overthrow by any peaceful means. I think the moderate citizen’s favorite Kerensky would have been her choice for change, until his inherent flaws were exposed, dealt with and led to his living the life of ‘Riley’ (pun intended) in asylum lovers Paris. It seems to be that Lily knew nothing about the international populatity of ‘Gadfly’ which she had written in the 1890’s and her modest windful was said to have been as much a surprise to her as it was to the U S Department of Internal Revenue, who kindly allowed it to go untaxed, from memory.
John Sanders: I read a biography of George Boole along time ago. As my line of Voynich thinking still, I’m afraid, revolves around the 15th century, more modern developments don’t figure so highly. Although, naturally, if I have the time, I have interests outside of the Voynich; I am not completely one dimensional.
Whilst I am far from being the historic cipher expert, it seems to me that having a cipher containing filler text is very common.
Take an example I remembered from reading (the much derided)Sherlock Holmes as a boy. Holmes sees a message which reads:
“The supply of game for London is going steadily up. Head-keeper Hudson, we believe, has been now told to receive all orders for fly-paper and for preservation of your hen pheasant’s life.”
If one reads every third word beginning with the first, there is a straightforwardl message:
“The game is up. Hudson has told all. Fly for your life.”
Now most of the text of the original message is filler text. Clearly as this is not written in secret symbols the original message has also to be intelligible, but with the Voynich I think this necessity is not there.
Continued:
So I wonder why we should assume that of the huge number of words in the Voynich not a single word is a filler/null, especially when the evidence makes it seem very plausible?
I accept that the idea that there is a significant amount of null text in the Voynich is not the most exciting explanation, but I guess I find it the most plausible at the moment and truth is more important to me than the thrill of the chase.
If with text going around the circle we had something like the following with respect to weather over time:
dry wet showers wet wet wet humid cold snow dry dry wet
i.e. some kind of chronology
However still this doesn’t really fit as there is no clear or consistent segmentation and it doesn’t explain the other related problems.
To me filler text is the best explanation for what I see.
Mark: Unlike some, I’m all for “the thrill of the chase”, at odds with our moral from the ‘hare and the tortoise’ fable, stipulating that “slow and steady wins the race”. I see you as one prefering to emulate the hard shelled old tortoise who, by making slow but steady progress, crosses the finish line first. Whereas others will take off like rabbits, wind in their faces, eventually straying off the path altogether, much like our thrill seeking march hare, and ending up in last place, having made complete bunnies of themselves in the process.
Mark wrote: “So I wonder why we should assume that of the huge number of words in the Voynich not a single word is a filler/null, especially when the evidence makes it seem very plausible?”
I don’t think people are assuming there can’t be nullwords, Mark.
Your host on this blog has frequently mentioned steganography (I have too, but Nick has been studying the VMS longer than I have, so I’m sure he mentioned the possibility long before me). Steganography can often have large amounts of filler text. Sometimes the filler is meaningful, sometimes not.
In anacrostic-style illuminations (which were popular in ecclesiastical manuscripts), the intervening text could be meaningful or nonmeaningful, but the “hidden” text was sometimes no more than a word. In “unsigned” manuscripts, sometimes there were one or two words (e.g., the name of the scribe who penned the manuscript) hidden somewhere in the text or illustrations.
If there’s meaningful text in the VMS, I doubt that it is as sparse as some of the extreme cases, but I’m sure most cryptanalysts would assume nulls might be present until demonstrated otherwise.
JKP: My point is that if one accepts the possibility that there might be null words then the question then becomes to what extent or what percentage of the Voynich text is null/filler?
Is it?
5%
20%
50%
80%
Is it the case that there are some words that are always null or simply arbitrary filler text is added?
Nick raised the question of whether 90×% is null text. On the basis of my current observations I strongly doubt as large a percentage is null text, though it is within the realms of possibility that my opinion could change. However if 90+% were null this would radically change our analysis of the text such as frequency counts, so really almost all Voynich text analysis could be thrown out the window.
The best guess that I can make given my observations is that very roughly 60% of the text could be null, however this figure as I have said could be radically readjusted on the basis of my having studied the text in more detail. Nevertheless a figure of 60% null text would make a big difference to many of the Voynich textual phenomena that have been recorded. For example, positionality could be largely a feature of null words not real words as could possibly be directionality and frequency statistics could be completely skewed by null words, so it fact much of what has been studied could be the structure and formation of null text. If this is true then the question that arises is how we distinguish between null text and other text i.e. what rules if any pertain to null text. This is obviously speculation, but I think valuable speculation.
John Sanders: I think the problem is that people often project what they want the Voynich to be onto the story and nature of the manuscript, without really focusing on trying to find out what the nature and the story of the manuscript really is. The widespread presence of null text is not most romantic or elegant position to take and some understandably might find the idea that many of the intricacies found in the Voynich are explained away by null text very dull. However I am focused on what I think at any one time is most likely to be true and my opinions certainly might change over time. That is the conclusion that I have come to on the basis of observation and the application of the unfairly maligned Occam’s Razor i.e. the simplest explanation that fits the observations is probably the best even if it appears to be a dull explanation. But who knows in a month’s time my ideas may have shifted significantly.
Mark wrote: ” But who knows in a month’s time my ideas may have shifted significantly.”
And we’ll get to read paragraphs and paragraphs of every nuance along the way.
It took me a while to figure out why you were repeatedly stating the obvious, on a site where your host has thought about and blogged about pretty much everything you’ve been discussing.
I kept wondering, why does he do that? It’s kind of insulting to those who have been researching the VMS for a long time…
and then the lightbulb went on…
You like to think out loud. It took a while to for me to grok that. I get it now.
JKP: I just googled:
“null words” Voynich
And
nullwords Voynich
The references were very sparce indeed. There were three references on this site, two of which were comments by me either recently or some time ago and one which was a comment by Job with respect to Nick’s Blitz Cipher and Nulls post. The only two others references on the internet that I spotted was a comment on Voynich Ninja and some analysis by Davidsch.
So on the face of it, for a subject that has exhaustively been analysed in the Voynich community there is very little evidence. DavidSch argues on the basis of statistics that there may be a large percentage of null words and Job talks a little bit about some of the same ideas that I have discussed. But it appears that most of what I have raised has not been addressed by other people.
JKP: I think the issue is that many people understandably really don’t like the idea of widespread null/filler words in the manuscript and so are very resistant to the notion; it makes Voynichese seem not quite as exciting, clever or interesting to them and they are invested in the idea that it is exciting, clever and interesting. However I don’t think widespread null words precludes Voynichese being interesting.
Mark, it isn’t necessary to write that there are nullwords in general references to something that might be ciphertext because the term “nulls” encompasses null characters, null words, and null phrases without making assumptions about how long the null sequences might be. So those who believe the VMS may be a cipher may only have written “nulls” or “null characters”.
Plus, it goes without saying that there might be nulls, if it is a cipher (a high proportion of ciphers have them), so it’s hardly necessary to mention nulls unless one is specifically reporting on a certain pattern of interest. In other words, the low incidence of your search term might simply be the way you are searching.
There are 500 times as many hits, when you use simply “voynich null”.
Some people will use the term “verbose cipher” to describe text that might include nulls. This is less precise because “verbose” doesn’t necessarily include nulls, but since ciphertext often has nulls, it’s to be assumed that a verbose cipher might.
And then, as I already mentioned, there is steganography. If there are people who suspect steganography, they won’t necessarily use the term “nullwords” to describe the filler text.
I know you dislike it when I bring up word definitions, but they are obviously relevant while trying to search, or while trying to communicate about specific aspects of an unknown script.
The most relevant point that I have seen to what I discussing is the following by “Job”:
“As far as the Voynich goes, there isn’t much room for null characters, but null words are plausible.
On the other hand, a null-word scheme could be as complex as we want it to be. It could lead to shoehorning, e.g.:
1. A word is null if it starts or ends with a specific character.
2. A word is null depending on position.
3. A word is null depending on a combination of characters.
4. A word is null depending on structure.
5. A word is null if i want it to be.”
I have already discussed these points, though with a somewhat different and more detailed perspective on some of them.
I think someone reading Job’s point could conclude that if there are widespread null words then identifying them could be almost impossible. I, as I have mentioned in previous comments, am a bit more optimistic as to the possibility of identifying such words in so far as they exist, though I am not saying it would be easy.
The only terms that I am aware of to describe this topic are “null words” and “filler”/”filter text”. I have been lead to believe that there has been a lot written on this topic in the context of the Voynich, but maybe other terminology has been used to describe it and that is why I haven’t seen if. Otherwise I can only conclude that Voynich researchers have barely scratched the surface of this highly relevant topic and that I am actually saying something new and original.
Keep in mind, Mark, that not everyone divulges what they’ve been doing until they think they have a solution. Many are unknown to the Voynich community until they send out a press release or publish a paper.
And there are probably even more who never divulge any research because they never reach a solution.
JKP: I have not just referred to null words, but been much more specific about them and questions pertaining to them. If this specific topic has been discussed then feel free to let me know where. When I say specific, I mean I have been quite detailed and specific, so I am not referring to a vague allusion to the subject. I have googled:
Voynich null
I have not gone through all 500 yet, though I may, but what I have seen so far are all references to null symbols/characters not null words. Not only have I discussed null words, but I have pursued my own particularly investigation of the subject from my own perspective, so it seems very far from a rehash of something covered in detail by someone else.
JKP: I cannot operate on the basis that someone else might secretly be covering the same ground. If I did then I and you and everyone else in fact should not ever present any research or ideas as we could never be sure that someone else it private had not already done much more research into that area; one has to work on the basis of what has been posted in public. Without any evidence to the contrary one has to operate on the basis that what one is doing is original. If you have evidence that you wish to present that the same research has been published before online then I would be very interested in knowing about it.
Mark, you can also conclude that (almost) every symbol is a null, based on statistical comparisons. It can be proven, however statistics is still not considered and accepted as hard proof, but more as a guideline.
However, it has been proven that an acceptable outcome is far from such a conclusion. It seems people quickly are able to accept one of these sooner:
1) a conspiracy conclusion 2) (another) not verifiable solution.
DavidSch: I do agree that to some people the idea of a considerable amount of null text in the Voynich is unpalatable. To be honest I don’t really like the idea myself that a significant percentage of the Voynich manuscript is null/filler text, however at this stage it seems to me a very plausible scenario.
However I think it very unlikely that it is all null text, so I wouldn’t go as far as you.
Mark: D’Imperio discusses null words on p.37 (4.4.2) of “An Elegant Enigma”, so it’s hardly a new hypothesis. And I regularly hear the suggestion that Voynichese might be some kind of Trithemian ciphertext (e.g. where every other word is nonsense or filler text), which is broadly the same idea. So the question isn’t ‘has the idea of null words been considered?” (because it has, by numerous researchers, and at length), but “what specific evidence can you point to that you think supports the idea of null words being present in Voynichese?”.
For those interested this is what D’Imperio has to say about Null words, that Nick was referring to I believe, when discussing various possible scenarios:
“T.7 Plain text concealed in a much Ionger “dummy” or “cover” text most of which is meaningless.
T.8 A Trithemian or Baconian system. involving the use of some binary or trinary characteristic (closed or open Ietters; tails up or tails down: ligaturing or lack of it; etc.) as the true message-carrying feature in a manner similar to the “dots” and “dashes” of Morse code, applied to a “cover” text or “carrier” text which is meaningless in itself.
As will be shown in Chapter 9, all of the above possibilities were known and used by early practitioners of secret writing. well within the fifteenth and sixteenth century. Roger Bacon mentions a number of them in an often-cited passage in his work entitled “De Mirabili Potestate Artis et Naturae” (Bacon 1859). The methods he lists include made-up alphabets, geometric figures combined with dots, shorthand (“ars notoria” or Tyronian Hand), and dropping vowels from the plaintext. in alchemical treaties attributed to him, Bacon is also thought by some to have employed anagraming, simple substitution (one plain text character to one cipher character). and concealment of a short message within a much longer meaningless “cover” text.”
Not as a criticism of her, I should say this is a brief allusion to this subject and not a detailed analysis along the lines that I have explored. I, at no time claimed, that I had invented the concept as should be quite clear from my previous comments, However, nevertheless I think this passage is a valuable reference on the subject. It is noteworthy that D’Imperio uses the terms “dummy text”, “cover text” and “carrier text” not “null words” or “filler text” as I have used; this may be the standard terminology, though it does have subtly different associations from the term “null words”.
So anyway as I said in an earlier comment:
“I have not just referred to null words, but been much more specific about them and questions pertaining to them. If this specific topic has been discussed then feel free to let me know where. When I say specific, I mean I have been quite detailed and specific, so I am not referring to a vague allusion to the subject. ”
To be continued on your other points Nick…
Mark: D’Imperio nowhere claims to have collected everything on any subject anywhere in her book. Rather, what she is trying to collate is a summary of the enormous amount of work and analysis that had been carried out over the previous decades. In that, I think she succeeded admirably.
Nick: My point was not a criticism of D’Imperio, whatsoever, from what I have read of her book it is very professional and insightful. My point was that I question the assertion that what I have said is a rehash of ideas that have been covered in detail ad nauseum and in the form that I have. I was highlighting that what D’Imperio does is very far from what I have discussed.
I will address the other points you made in your previous post in another comment soon.
In addition I am happy to provide you in the next couple of weeks with my critique of your writing on cipher techniques in the Curse i.e. those in Chapter 10, 11 and 12, after having previously been told that I am too dullwitted to comprehend the sophisticated analysis contained therein. Naturallly I think some of what you have written in interesting and insightful and some I think is quite weak or questionable.
Mark: naturally I will be eager to see the evidence you bring forward to support your points of view, given that there currently seems to be no end to the tsunami of unsupported opinions washing over the beach of my blog. 🙁
As a point of reference Berj Ensanian uses the terms “text” and “anti-text” where I use the terms “monotonuous” and “distinctive” though we approach this topic from different angles, hence the difference in terminology.
My top 4 candidates for null words:
otol
otor
okor
okol
Though there are other words which are very high up on the list.
In general some null words are of the form:
FGSE
Where F = “o”
G = “t” or “k”
S = “o” or “a”
E = “l” or “r”
However I am sure that some or all of these sets F,G, S and E can be expanded to include more symbols.
For example “89” could probably be added to E.
It may be possible to add “p” to G.
I am certainly not suggesting that all null words are of this form.
I wouldn’t be surprised if words that are prefixed by null words such as “otor” can be reduce in length by ignoring the prefix. So then “otor” would not only be a null word, but also null text more generally(though maybe only in the context of prefixes).
Mark: Distinctivly monotonous seems to strike a nerve, but it’s all in the angle of the dangle right?..Hey that’s a joke man, just a little overdone perhaps!..Don’t mean nuthin, drive on OK?..
John Sanders: In case I did not make it clear my previous comment, in earlier comments I had distinguished between two kinds of words in the Voynich the “monotonous” words and alternatively the “distinctive” words. Anyway, you were joking, so I shouldn’t take you comments too seriously.
I have been thinking about 1 symbol difference networks. This is where one has a connection between each word and each other word with one letter different only for words that actually appear in the Voynich manuscript. So “otol” and “otor” would be connected to each other also “otor” and “okor” would be connected to each other. Now the importance of a node word could be calculated on the basis of the frequency of the word in the Voynich and also the number of instances that it is repeated and length of the repetitions in each instance. It should then be possible to identify high score word clusters. I would then argue that those word clusters would most likely be null word clusters.
Well what about words such as in English the following:
Cat
Hat
Rat
Rut
Hut
Cut
But
Well, these are all valid words with one letter different. However are these all high frequency words or repeated words? So most words in this cluster will have low scores. (It may also be wise to consider word length when scoring a word node, where shorter words score less. Also I think it is reasonable to view words as connected if one has the same spelling as the other except with an addition symbol added somewhere in the word.) I think this approach would complement the manual identification of null words.
I imagine this cluster identification approach could be refined with more insight on the basis of observing the other properties of word clusters.
I wonder if words that exhibit more of the standard general observed positionality and directionality should score higher than words that don’t as I think possibly they are more likely to be nulls.
One really would hope broadly speaking that 2 distinct sets would immerge from this process rather than one set verge into another. For various reasoms there might be a few words for which there is some uncertainty as to which set they belong to, but hopefully they will be a tiny proportion.
Another approach to remove all null words, though this approach will most likely also remove many non-null words, that has occurred to me is that one could simply remove say the top 80% most frequent words words leaving behind the 20% of rarest words. I think after that some manual inspection this set to identify possible nulls that might have slipped through the net. As I suspect word endings often including tironian notes then of this remaing set of words I would suggest only selecting say the first 4 symbols, so truncating each word. Then I suggest applying a similar procedure to a comparable long text in another language. After that I would suggest doing tests likely frequency analysis.
One thing that does occur to me about comparison with a text in another language is that if that text contains a completely different vocubulary to the Voynich frequency analysis could be problematic; herbal manuscripts and theological manuscripts may have quite different vocubulary and so quite different letter frequencies.
One thing that occurs to me is that proving that there are no null words is very difficult as one would have to demonstrate in each every case that a word is not a null. Whereas demonstrating that there are null words merely requires one to do this in only in the case of one specific word.
@Mark
You write:
One thing that occurs to me is that proving that there are no null words is very difficult as one would have to demonstrate in each every case that a word is not a null.
I just want to know if the translator plays a trick on me here.
Prove that there are no null words? Is not it to prove that there are some?
I have already read some books from that time. Not just medical, others too.
On closer inspection, the frequency of the same words is the rule rather than the exception. But in medical books it is especially bad.
Each plant is almost the same. Recipes are even worse, because you even get a headache because it is so monotonous.
Regarding labels I think it worth asking that if one saw the following pictures each with exactly the same label against them, what would one conclude?
Marilyn Monroe
Mount Everest
A Bathtub
A Rotten Apricot
Alpha Centauri
A Safety Pin
A Mathematical Formula
A Ray of Light
What would one think about what the label might say?
On the face of it all these items appear to have very little in common. So what might the label say? There are possibilities, but my inclination, as one might guess given my recent statements, is that the most likely possibility is that the label says nothing i.e. is meaningless. Then if one finds another label with only one letter difference from the first label and that second label is attached to a variety of quite distinct images also with seemingly little in common. And then one finds third and fourth labels with one letter difference attached to a different range of images. And finally one notices in a attached text that these words are often repeated multiple times and sometimes followed one by another it would rather increase one’s suspicions that one is dealing with null words. This of course is analogous to what one finds in the Voynich. I can think of other explanations, but in my opinion they are all way down on the plausibility scale.
This experience already tells me when someone indicates a possible translation, whether it is possible or not. And it also tells me if anyone has ever read a similar book before. I say expressly read, not just viewed.
Too much spice breaks the soup.
Gordon Rugg’s idea for generating null words in his hoax was by using a Cardan Grille, I believe. Now obviously my ideas and Rugg’s ideas differ markedly in that he believes that all words are null words and that there are no meaningful words in the Voynich whilst I believe that a significant proportion of words are null, but also a reasonable proportion of words are meaningful(I still cannot say with confidence what those precise proportions are, though I guesstimated a figure of roughly 60% null words, but this could possibly be as low as 15% or as high as 95%.)
However maybe Rugg and myself have the same kind of ideas about how null words might be generated. Personally, at this stage, it doesn’t seem likely. My inclination is that the author added null words on a more ad hoc basis without the use of specific tools like a Cardan Grille and probably without relying on specific procedures to generate null words. However I believe null words have a specific common format making it easy for the author to think of the next null word without relying on a huge null word list.
It is striking that amongst circular text and bounded linear text that the text fits so neatly with the same degree of spacing and size of text in the available space for the text. In the odd example there is a smallish gap at the end of the circular text, however this is rare. This circular text normally has a separator, which supports the idea that this is sentence text not a series of labels as sentence text has a start and an end. Also the words in the circular text normally in no way map to nearby images. There are a small number of instances where there are many separators in circular text and so they do map to nearby images or general structure.
I would find it very difficult to again and again produce relevant sentences that perfectly fit the bounds of a specific circle. However if I could use fillers it would make producing bounded text of this kind much easier. This again supports the idea of the existence of null words, especially as sometimes in this circular text one finds repeated words and adjacent one letter difference words.
The more null words that one can use the easier that it becomes to fit a sentence or two exactly around a circle, so the fact that the fit of text around a circle is so good in most cases would lead me to believe that there quite a few filler/null words used.
In theory it would be possible to with great effort produce text that fits perfectly around a circle without fillers and keeping the text pertinent and useful in its context. However in practice to do this again and again would be nightmarishly difficilt, so this is another reason why I don’t think that’s what the author did.
One could have a situation where all words that have the second symbol as a gallows character are nulls, I am suspicious of this assertion, but it is conceivable. Or there could be some other symbol interaction that flags a word as a null, purely as an example all words begin with “o” and ending with “9” could be nulls. Or the appearance of null words could be more loosely and imprecisely defined.
Obviously identifying the words that are nulls and the words that are not is vital in arriving at a filler word free text or true text. Once the true text is reached the analysis begins anew as though with a blank sheet almost all previous textual analysis being probably close to worthless as it is based on a sizeable quantity of nulls in addition to the real text, so statistics etc. would have been completely skewed.
But how to be confident that one has systematically identified all and only the null words seems at this stage to me to be not an easy task, but we will see.
When it comes to null words the author must have been able to remember without too much difficulty what are nulls and how they are defined, as I doubt he/she had a long null word list. Now it is possible there is more than one independent set of null words and therefore more than one formula to define different groups of null words, but if they were defined in a way that was simple enough for the author to identify then hopefully it will not be too difficult to identify them all.
Hello Mark,
I’ve wondered whether there is a ‘typical’ length of text accompanying plant illustrations in mediaeval herbals. So if you have a picture of a foxglove, for example, do you generally have three or four lines of text describing plant and properties? Or could it vary widely?
It has struck me that this might give a control for studies of Voynichese. If there are roughly the same number of graphemes and words in Voynichese plant descriptions as we find I n contemporary plaintext herbals, this perhaps argues against the idea that Voynichese has a high proportion of null words?
SirHubert: I must confess that I have no idea as to whether there is a ‘typical’ length of text accompanying plant illustrations in medieval herbals, but it is certainly something that it is worth someone looking into.
My study of Voynichese has been largely, though not wholly, confined to studying the labels(labelese) in the manuscript and also the text circles and the bounded lines text. My conclusions are based on my observations of them, but given what I have seen of the text overall it appears to me that the kind of phenomena that I have described are not found only amongst labels, but are applicable to the text as a whole. I should say that I was already suspicious that there are null words in the manuscript, but it is only as result of studying the labels that I have come to the conclusion that there are such a sizeable proportion. However it is also apparent that there are many words with very distinctive spellings, so I am pretty sure there are real words in the text.
Of course it is possible, though I doubt it, that like the books of the Benedictine Abbot Trithemius that the underlying text concerns a completely different subject from the impression we get from the drawings. However I can’t see someone going to the length of drawing all those picture just to mislead.
SirHubert: There are 2 questions I think:
Are there any null words?
If there are some null words then what proportion of words are null?
I have made a clear outline of why I think there are some null words. The question of the proportion of words that are null is a more complicated question.
Now if one believes that there are no null words then I welcome a persuasive alternative explanation that fits all the phenomena that I have described.
In addition, I should say that I have applied the unfairly derided Occam’s razor to this problem i.e. I think the simplest explanation that fits the facts is the best.
Mark: what you are supplying here is effectively a very convincing argument as to why Occam’s Razor is completely useless, sorry. 🙁
Sir Hubert, plant descriptions in contemporary and medieval manuscripts are generally not of the same length. Some plants were considered more important than others, or were thought to have more properties of interest than others.
For example, Plant A might only have one part that was used (seed) and it might only be useful for one thing. Plant B might have several parts of value (roots, leaves, flowers) and they might be for several purposes. The purposes are not always medicinal, sometimes they are culinary, aromatic, dye-related, or talismanic.
So, the length of descriptions often varied.
However, some texts do follow a formula. For example, they state the names of the plant, the properties, and then the medicinal uses. Others are not so organized.
Many medieval herbals have only the name of the plant and sometimes that has been added later rather than at the time the same time as the drawings (many manuscripts went unfinished… plague and famine may have contributed to this).
.
Several years back, I created a very long concordance of VMS text to try to find relationships among plants, and between the plants and the rest of the manuscript. It took more than two years and is more than 1100 pages long. It maps every token to its shape-mates. I even tried token-pairs, with similar results.
It did not yield the kinds of patterns one would expect if the content were natural language related to the drawings. But, the concordance was based on honoring VMS spaces (it was the simplest place to start) and VMS spaces are not necessarily word boundaries.
Mark, you asked:
“Regarding labels I think it worth asking that if one saw the following pictures each with exactly the same label against them, what would one conclude?
Marilyn Monroe
(rest of list deleted for brevity)
What would one think about what the label might say?”
I would suggest:
“Fig.2”
Mark: I’d suggest 8 and 10 for Marilyn Monroe would be the perfect figure….By the way, you won’t agree but I think that you and I are now headed on a steadilly converging set of paths to an eventual link-up. It will bare a slightly bitter harvest for one of us. ie. Your final translation of the Vm text, yet to be determined but close, is now on intercept course with my own known solution, including a well known author. Only one difference being the antecedant origins lie four hundred some odd years apart and I doubt that you’ll be in any strong position for to nominate an author…No matter though, you can borrow mine, she won’t mind at all.
Rene: That occurred to me. In fact I discussed that possibility with my brother over the phone as he is curious as to my interest in the Voynich. The problems I see with that in the context of the Voynich are as follows:
If there is a label such as “figure 2” one would expect that “figure 2” would be referred to in the text eithet explicitly or implicitly such as “Figures 1 to 3” or “Even numbered figures”. If numbered labels are used, but there are no reference to them then they are superfluous.
In addition why would we see repeated words such as “Figure 2, Figure 2, Figure 2”? (Now it true that the text could just say “2” rather than “figure 2” and “2 2 2” could mean say “222”)
If all these repeated labels are numbers then why are they one symbol difference and in a difference symbol position from each other?
In addition, where do we go to find the text that refers to “figure 2” when that word does not occur in the other text on that page?
Why do some labels for exactly the same kind items on a page have figure numbers whilst others have standard words? Why do they have an inconsistent variation in numbers, one would expect them to always be sequential?
How do we resolve the problem of fixed length circular text?
The more one explores the figure number hypothesis the more problems that it throws up when one looks at the text. Yes I suppose one could construct some very complex figure number theory that could explain away all the problems, but the simple explanation that they are nulls seems to be far and away the most plausible.
Hi JKP,
Thank you for this, and I’m sorry for not expressing my suggestion as clearly as I should.
I take your point that a Plant A, to which more properties were ascribed and about which more was then known, would tend to have longer entries than a less interesting/useful Plant B.
What I was suggesting was comparing different entries for just one plant (Plant A, if you like), across a wide range of European herbals produced between, say, 1200-1500. How widely would these entries vary in length? I appreciate, of course, that there would be some variation according to the language used, so that if you took four lines of Latin you might translate those into five lines of English, for example.
My completely off-the-cuff impression is that there seems to be a lot more Voynichese accompanying a plant in comparison with herbals written in Latin or English, where there’s less text. But I’ve no idea if that’s right?
As per medieval scribal conventions, Figure 2 could be written 10 different ways. It could be Fig. 2, Fig’r 2, F’g’r 2, F’gur’ 2, and so on. There was a great deal of variation in how they wrote common (and not-so-common) words.
Even if there weren’t so many spellings and abbreviations, it’s also possible that a label might be Fig. 2, but that the main text might be Figure 2 (or some other non-identical variation).
I have to agree with René’s point, that there is more than one way in which labels might be similar to one another and different from the text.
I try to keep a variety of possibilities on the table until the data indicates which ones might be more relevant.
I did not say that I think that in the Voynich MS the label words mean “Fig.2”, because in the Voynich MS there are no pictures of Marylin Monroe or any of the other things you were listing. For a medieval text there are other possibilities.
You were talking about Occam’s razor…
Between the three options:
– All words are meaningful
– All words are meaningless
– There is some unknown part of the words that is meaningless, and there should be some way of easily detecting them but I haven’t figured that rule out yet
Guess which one should be the first to be discarded by applying Occam’s razor.
Rene: Might I suggest that you search out George Boole’s algebraic propositional truth theory as a starter, should you not be familiar with such an outlandish old hat formula for a higher level of of logical thought transfiguration.
Rene: I mentioned Marilyn Monroe and the others as an analogy, as I said. In the Voynich one can find examples like where the same label is attached to a star, a plant, a nude woman with some kind of water pipes and a segment of a T-O Map amongst other things.
If all words are meaningful how do we account for examples like this, the fact that these words differ by one symbol from other words with similar properties, the fact that these words are found repeated 2 or more times. In addition how do we explain circular bounded sentence and linear bounded sentence as I have discussed.
Considering all words are meaningless, i.e. something like the Rugg hypothesis. Well, amongst these monotonous words we find many labels with distinct spellings that are sometimes not repeated anywhere in the text. Now whilst it is possible that all these words are also null I find it likely that the author invented these specific words. I admit that it is theoretically possible that these “distinctive” words are null. it seems unlikely.
Whilst I by no means discount the possibility that there are parts of some words that are null, in fact on the basis of my research I find it very plausible, what I am discussing at this stage is focused on whole words being null.
I have made suggestions about the kind of rule I think we are looking at and I am developing more ideas along these lines.
I would say that by Occam’s razor the first option should be disguarded as it goes against the evidence. Occam’s razor says to prefer the simplest solution that fits the evidence, the first option does not fit the evidence, so it should be rejected I would say.
Mark was asking a very contrived question that had nothing to do with the Voynich MS, and I believe that the point would have been that, if it could not be answered, this would have some implication for the Voynich MS.
However, the question could easily be answered, so in this case there was no point to be made.
By the way, my favourite version of Occam’s razor is:
“the simplest explanation that fits the facts is the least complicated one”
but I am not so sure that it holds in all cases.
Rene: I think it is important to understand that Occam’s razor does not say the simplest solution is the best regardless of the facts, but given that solution is consistent with the facts. Otherwise we could conclude that the manuscript just appeared by magic, but the problem with that is it doesn’t fit with the known evidence of of matter in the universe.
When one tries to justify the “All words are meaningful” statement in a way that is consistent with the facts the explanation has to become much more complicated than the one that I have proposed, hence the use of Occam’s razor.
Rene: I made it clear that my example was an analogy with the Voynich. I was not asking it as a point of general interest independent of the Voynich, maybe I should have referred directly to the examples in the Voynich. Nevertheless this seems to be a somewhat pedantic point you are making.
Certainly one can find an objection to Occam’s razor, i.e. how does one define what is a simpler or more complicated theory than another, However in the vast majority of instances I think this is pretty straightforward and in the other cases it can be done. Certainly it the case of the Voynich that we are discussing it is very relevant and I would welcome practical objections to it in this context. If complexity of a theory that fits the facts is no bar I can come up with a myriad complicated theories that would fit the evidence. However I have a strong preference for the simpler solutions that fit the facts not the most complex, the null word solution appears to me to be such a solution.
Häufigkeit von Wörtern, Beispiel Blut:
In einer vielzahl von Passagen kommt das Wort “Blut” übermässig oft vor.
1. ” ist das blut heiss, so gib ihm xxx”
2. ” ist das blut schnell, so kann xxx helfen”
3. ” ist das blut stark, so gib xxx”
Die Hauptausage besteht darin das für den Schreiber alle symtome über das Blut kommt.
1. Fieber
2. Puls
3. Blutdruck
Diese Wörter würde man heute benutzen.
Die Optionen, fremde Sprache, nicht gelesen, nicht verstanden machen den Rest.
Mark: Occam’s Razor appeals to simplicity as the single most important ‘X factor’ in the theory beauty parade, but unfortunately there is no working definition of ‘simplicity’ that will steer you consistently away from the heaving, sweating mass of idiot Voynich theorists. The #1 way to elevate arguments is by finding actual evidence, and thereby giving people something to work with, rather than offering up yet more hopeful speculation to throw on the Voynichian pyre.
I’ll give you an example: when EVA ‘s’ appears as the first letter of a line, I can propose a testable hypothesis that I think it forms line-initial words that are rarely (if ever) found elsewhere in the text, and would argue that this implies that line-initial EVA ‘s’ is therefore either a null or some non-textual character. Can that hypothesis be tested in an absolute way? Yes. Would debating this actually move discussion of nulls forward? I would hope so.
Spot the difference.
Nettes Beispiel:
Aber die Frage lautet jetzt, wie würde man das A verstecken?
Marilyn Monroe
Mount Everest
A Bathtub
A Rotten Apricot
Alpha Centauri
A Safety Pin
A Mathematical Formula
A Ray of Light
I have little free time at the moment, due to real-world commitments. However, as a break from that, and to de-fry my poor brain, I dabble from time to time in the mystic art of making sense of voynichese.
I am confident that use of Occam’s razor will supply a solution that demonstrably cuts the mustard.
On the hypothesis that the author wrote a heavily abbreviated set of symbols to be read aloud – most likely as Latin – it is possible that one and the same abbreviation may stand for any one of a set of words, depending on context.
An example from English – I may, in a speech note, write “ct” to stand for, depending on my topic: cat cats, catapult, cut, cite … etc.
Another example: in a speech on automobiles I may refer to my note of “carb” and read it aloud as “carburretor”.
Imagine the delight and joy of a cryptologist 1,000 years hence on discovering my notes – sans illustrations. Imagine his/her frustration on discovering this list of 117 words for which “carb” may be an abbreviation:-
https//www.morewordscom/starts-with/carb/
Again, I may write “t” to remind me to say “it, the, to …” and even “tomorrow”, depending on context.
Call this encryption, abbreviation or data compression as you will: it is exceedingly lossy over time. Take this slightly obfuscated example from my recent blog:
“tdismg n assy ofavvv ng s mr cmpx t fo ranrml 4c blk”
Even I, who wrote it, made a mistake when reading it aloud. Had I written “rassy” I might have remembered to say, not “assembly”, but “reassembly”. The loss of data here leads to only trivial loss of information.
Perhaps the author of the VM was not concerned about the possibility of such trivial errors. Perhaps he was also unconcerned about encoding much in the way of grammar, which can readily be restored in speech.
In speech notes, infinitives (without “to”) are as good as most variants, and e.g. “rng” is as good as “running” in: “w wr rng alng t bch wn w sptd a d wl.”
Contemporaneously, this can be read by the writer as :”We were running along the beach when we spotted a dead whale.” Found later by a person who does not know the context of what is written, or the underlying language, it is so much gobbledygook. Unless there is a sketch, however unartistic, of a whale-like object next to some squiggly lines.
Is anyone up for some Latin speech notes on the health benefits of herbal baths?
Patrick: as I’m certain you know well, 15th century shorthand writers were caught in the death throes of Tironian notae (a few notae remained in use [though adapted], but to be honest not many), but were still a century-plus before the birth of modern shorthand with Bright’s Characterie. Their response was to develop local scribal abbreviation systems (particularly in Italy, but also elsewhere in Europe): for example, Milanese scribal abbreviations were quite different from Roman scribal abbreviations. Perhaps you are not yet aware that there is a specialist technical literature on these?
Nick: I am not sure if your critique of Occam’s razor in your previous comment is related to my theory or just more general.
If your example is intend to contrast with my arguments as being non-specific then I have been very specific. I have identified specific words that I believe are null and presented a precise formula to define such words. Now I am sure I will expand on this formula to cover more specific words that I believe to be null, in fact I could expand it further now. However I am keen to identify the null word structures further by inspection and also computationally to come up with a more general formula/formulae for null words. I have been specific about the criteria that I think define many null words and the specific examples can be found by searching the text for the examples that I have already provided, though I may provide a long list of page and position on page references where you can track these down.
Your first letter of a line theory that you describe sounds very plausible, though I haven’t studied the subject and therefore haven’t formed an opinion.
Both of our hypotheses can in principle be tested in an absolute way. However I think in both cases it would be hard. Unfortunately so many hypotheses about the Voynich can be only be tested in an absolute way once the manuscript is deciphered. Would debating my hypothesis and your hypothesis actually move discussion of nulls forward? I think, so, which is what I have been trying to do.
Mark: even though you have pointed to specific Voynichese words that you think might be nulls, you haven’t put forward any kind of argument as to *why* you think they are nulls to any depth greater than ‘you feel a bit uneasy about them so they must be nulls’.
I’m not saying you’re wrong (because maybe you’re right, who can tell?), but rather that you haven’t put forward any actual evidence as to why you think they are null words.
By way of contrast, the hypothesis I put forward in the comment explains the specific pattern I’m looking at (EVA ‘s’-initial words in line-initial positions), why I suspect these are anomalous (these words appear almost nowhere else in the Voynich manuscript), and what my tentative conclusion is (that these line-initial EVA ‘s’ shapes are nulls).
Your turn to try this.
Patrick Lockerby wrote: “On the hypothesis that the author wrote a heavily abbreviated set of symbols to be read aloud – most likely as Latin – it is possible that one and the same abbreviation may stand for any one of a set of words, depending on context.”
Your original algorithm posted on your blog treated each abbreviation as though it could be unraveled in the same way (in other words, did not factor in context) and I wrote a critique challenging your assumption because medieval abbreviations simply don’t work this way—context is everything when it comes to interpreting abbreviations.
Since I was the only one to point out this problem, it appears that you have read my posts and have altered your thinking about this. So I’m curious, are you planning to release another version of your transliteration?
More specifically…
1) Are you still taking an algorithmic approach? and
2) Have you changed (or are you planning to change) the algorithm and produce another version of the transliteration?
Sir Hubert wrote: “What I was suggesting was comparing different entries for just one plant (Plant A, if you like), across a wide range of European herbals produced between, say, 1200-1500. How widely would these entries vary in length? ”
Since I read the plant descriptions (I don’t just look at the pictures), it’s very easy for me to answer this, and not just for European herbals, I have also looked at Asian and Middle Eastern herbals.
Length of Plant Descriptions
I have never seen a particularly long plant description. More than a folio is uncommon. Some plant illustrations have no labels. Some have only a label (and sometimes that label was added in a later hand). Some have only a label and a few words. It is rare to find more than a paragraph or two.
Everything was written by hand, and writing by hand is tedious. Most plant descriptions were only a few lines. Those of Hildegard von Bingen are a bit longer than many but were not especially long.
By this standard, the amount of text on a VMS plant folio is a little longer than average but it’s not highly unusual. There are some herbals where the plant is on one page and the text is on the other and there is more room for text. On these, the text occasionally fills a whole folio for important plants (more than what one sees in the VMS).
Some herbals dedicated a full folio to each plant drawing/description, but it wasn’t a lot of text. It was large and widely spaced. Those following the traditional Greek model frequently included a long row of vernacular names and only a few lines or a paragraph describing its use. In other words, the descriptive content was rarely more than a short paragraph or two, even when the text filled the page.
Uses
Many of the herbal manuscripts were used as textbooks in universities or as references in hospital monasteries (plants were medicine), and thus were copied by students of medicine and pharmacy when they could get access to them. Some were in royal houses and probably weren’t seen much by outsiders.
Some of the important exemplars originated in the Salerno/Naples schools and were brought to Lombardy (at a time when Lombardy included Florence) and were further copied in Lombardy or the Veneto. One of the very early illustrated herbals made it all the way to England, probably via early Jerusalem pilgrims or home-coming crusaders.
Unillustrated Herbals
Many medieval herbals do not have illustrations (or had very simple ones that could resemble almost any plant). I think unillustrated herbals are very much overlooked in Voynich studies. I have seen far more textual herbals than ones that are illustrated. The information content appears to be much the same but it can sometimes be difficult to identify the plant from the description alone if the vernacular name was used. Sometimes I can figure them out if the plants more-or-less follow the same order as a manuscript that is illustrated, but if they include local plants, it’s sometimes impossible to identify them.
Early herbals were not necessarily alphabetical, often the plants were ordered according to use, form, or other arrangements. To people in the medieval period, form was very important. The shape of the plant was thought to be a signal from God as to which part of the body the plant could help.
Even if the plants were alphabetized (which they often weren’t), vernacular names can vary widely and many of them have been lost as more standardized systems took hold, which makes it quite difficult to identify a plant from a short description with no drawings.
So, to sum up…
• Medieval herbals with images were less common than herbals with only text. This puts the VMS in the minority.
• Herbals with descriptive text covering a full page were less common than herbals with a few lines of text. This puts the VMS somewhat in the minority, but the amount of text is not highly unusual (more than most, but less than some).
As for the descriptive content of these herbals, I’ve wanted to blog about it for years and years because there are four or five basic “templates” for how the information was presented, but I simply can’t find enough time to blog all my research. In a nutshell, many of the plant descriptions included some or all of the following:
name, names in other languages, habitat, Galenistic properties, descriptions of uses (often from Pliny and some of the other historic physicians), remedies (recipes).
The two that seem to get left out most often when the scribe was writing brief notes are the habitat (it was often omitted, or reduced to one word like “mountain”), and the recipes.
Even if the uses of the plant were listed, the exact dosages and mixtures were frequently not included (they may have been carefully guarded as trade secrets by apothecaries and physicians). It wasn’t until the mass-market book industry was on the horizon that professionals began to divulge this information more readily (probably because they saw an opportunity to make money by getting it published).
Nick: I think you may have mistaken my intention. I am not considering shorthand as a formal system: merely as a personally developed memory jogger for use by the author/s and to be read aloud. What I have in mind is a step beyond writing e.g. the initial letters of a poem or speech to the same purpose. Good ideas get to be reinvented over and over, as e.g. your example of Bright’s Characterie being reinvented many times. A gentleman named Pitman comes readily to mind.
If I were to write –
t b r ntb tist q
converted into personally invented glyphs it would take great effort to “decode” it. But in plain letters, a cryptic hint about how unfair life can be in a nunnery …
Any takers?
This will be my farewell to y’all. This is my last chance of trying to get you all “open-minded” for the languages in the so-called “Voynich” manuscript. If you can read and write in Spanish, you will be able to understand the writings of Fray Sahagun’s Nahuatl STUDENTS . You will also be able to identify EVERY item AND their use !
Please don’t fall into the traps presented by so-called ‘military officials”. D’Imperio WAS NOT working for the military, ever, while writing her “biography”/biology”.
DO get two volumes of the “Florentine Codex” : “EARTHLY THINGS” and “THE GODS”
Fray Sahagun’s manuscripts were never returned to him after the inquisitioners (Spaniards) became involved with the war between Suleiman and Spain.
Yes, Patrick, it’s easy (even without the hint). 🙂 Shakespeare. But… do you think an entire manuscript would be written like that?
The problem is that the structure of the Voynich text is very positional, and mnemonic memory joggers are not—they are quite diverse in terms of character-position and variety.
Hi JKP,
Thank you very much for taking the time to write such a detailed and helpful response. You’ll have deduced that I was hoping that comparing the length of the Voynich plant descriptions with those found in contemporary herbals might have given us a clue as to whether we’re looking at some kind of encipherment technique involving nulls/steganography. But seemingly not – which is itself helpful to know, of course.
Patrick: my point is that in the fifteenth century, there were a number of (informal, local) scribal abbreviation systems used for taking shorthand notes. Why conceptualize about a modern reinvention of this basic idea, when you can – with a bit of research – find a number of examples of this in the historical record? Surely these would be more useful a test of what we might expect to see in the Voynich Manuscript than anything else?
Sir Hubert,
I once did the word counts for the 96 (IIRC) recipes (plant descriptions) of the alchemical herbals. The distribution was not dissimilar from the Voynich MS. I can’t show it here, of course.
There’s the caveat that the B-language descriptions are on average clearly longer than the A-language ones. There are 97 plants on the herbal A pages by my count.
Another similarity between the two is that there is a handful of pages with two plants, both in the Voynich MS and the alchemical herbals.
The biggest discrepancy is that the recipes of the alchemical herbals are full of short phrases that keep coming back in many of the recipes. This we do not see in the Voynich MS.
Rene: just because we can’t obviously see them doesn’t mean they’re not there. Putting Herbal B pages to one side, I think there must surely be some underlying stock phrases being reused on the Herbal A pages, but that some specific feature of the writing / enciphering system is disrupting our view of it.
For example, even if gallows characters are some kind of in-page transposition cipher (i.e. where the actual character they represent sits elsewhere on the page), we might be able to devise a matching strategy that will see through it. And even if contractions and truncations are taking place, we might be able to see through those too.
Nick, I do agree with the general point that it may be there even though we don’t see it. However, I haven’t yet been able to come up with any good ideas how this could work, without at the same time flattening out the word frequency distribution.
(This flattening out is something that, as you know, Vigenere ciphers tend to do, and it would be similar for other one-to-many types of substitution).
Rene: ‘flat’ is one thing Voynichese really isn’t. 🙂
I have some good ideas about how to look for patterns that have been somewhat obscured, so I’ll try to give this a go…
Nick:
You say “you haven’t put forward any kind of argument as to *why* you think they are nulls to any depth greater than ‘you feel a bit uneasy about them so they must be nulls’.”
If you have read my comments and maybe you haven’t had time to, other than checking for things you deem offensive, which is fair enough, you are a busy person, but I have supplied very explicit arguments as to why I think certain words are nulls. The idea that I have been perusing the manuscript and every so often I look at a word and get a funny feeling in my tummy and therefore conclude that word is a null is very far from the truth. I have put foreard actual evidence, but I will do so again.
The words I have listed I think are nulls for the following reasons:
1) They appear as isolated single word labels against many different and disparate pictures such as stars in the cosmologucal section, plants in the recipes section, nude women, pipes and features on the 9 rosette page amongst others. As these pictures are so different from each other it is hard to believe that the same word could be applicable as a label.
2) These words each appear repeated in the manuscript, which not a characteristic of normal words.
3) These words differ by one letter from one another whilst each also exhibiting properties 1 and 2.
4) It is not unusual for these 1 letter different words to be found followed by one another and as neighbouring labels.
5) These words exhibit positionality and directionality, something that the more distinctive words don’t and these are features unusual for ciphers and natural languages.
6) These words are common throughout the manuscript, which seems possibly more consistent with nulls.
All of these 6 pieces of evidence combined, not one in isolation, I think present a strong case why the words and their formula, that I have listed and other words that I have not yet listed, I conclude are most likely null words. I have explored a variety of other hypotheses that could explain this evidence, but I have found them all very unsatisfactory.
If you don’t understand what I have written here, as I may not have been sufficiently clear, please ask and I will be happy to clarify.
Nick: Another more general piece of evidence that I think supports the hypothesis that there are null words is the statements I have made about the circular bounded sentence text and linear bounded sentence text that one finds in the cosmological, astrological and 9 rosette pages.(There are good reasons to believe that this is sentence text not labels such as the presence of a start/end separator and no clear mapping of words to adjacent drawings). I think that fact that this text fits so neatly in the available space with text of standard size and spacing, is a strong indictation of the use of filler words i.e. null words.
Again this point like the other points may be explained away in isolation, but all together I think they strongly support the notion of null words.
Nick: I have guesstimated a figure of 60% of the words as they appear in the manuscript being null on the basis of words I strongly believe are null and words that I suspect are null. This figure could easily increase or decrease significantly based on further study, but it is indicative of my sense that there are a lot of null words.
Mark: heaven knows I read your comments. :-/
In every single “null word” case you flag here, you have tangled your observations, speculations, and partial / overlapping hypotheses together in a way that simply cannot be tackled as a coherent argument by someone looking in. You’re not yet in a situation where you can call your mass of ideas about null words a complete theory that can be tested in any useful sense.
You’ve kind of missed the point about science, which is that if you can provide a good argument about a single piece of the stuff you’re thinking about, then you can debate that in a constructive way and see what conclusions can be usefully drawn from it. But as it is, you’re trying to prove too many points all at the same time, when you haven’t yet built up a workable argument for any one of them. Less is more!
It’s a very creative use of the word ‘evidence’…
Nick: I appreciate it is a real nuisance working through all my comments, so thanks a lot for your efforts.
Rather than getting too deeply embroiled in this discussion I think it is worthwhile for me, with the Voynich time I have, continuing to research my theory. I think I have made my ideas on this clear and available to other people for now and for the future.
Rene: I am guessing that is a reference to my statements. In the light of your interpretation of previous statements, vis a vis my “map” theory and Gerard Cheshire, that I have made, I am not sure that this worries me.
Mark Knowles wrote:“They appear as isolated single word labels against many different and disparate pictures such as stars in the cosmologucal section, plants in the recipes section, nude women, pipes and features on the 9 rosette page amongst others. As these pictures are so different from each other it is hard to believe that the same word could be applicable as a label.”
Why do you think they are words?
This is a very big underlying assumption and your entire argument falls apart if they are not.
JKP: What do you think they are?
In one sense it doesn’t matter what they are their occurrence as I have described means they are associated with many completely different things which ties in with the notion of being null, so they could be individual characters and the same would apply. All other interpretations as to what they could be don’t stack up in my opinion, whether it be that they are numbers or adjectives. I have explored alternative hypotheses, which I have not described in detail here, and they might fit in one instance, but without them becoming very complex or convoluted and/or implausible/far fetched, they don’t cover all instances. I think often the most prosaic conclusions are the right ones; this could be another formulation of the foul Occam’s razor.(My “map” theory is probably the most prosaic of the map theories)
I use the term “word” in the Voynichese sense and so I think they are by definition Voynichese words. Clearly as I think they are null I therefore don’t think they are words in the natural language sense as no word in natural languages are null.
But if you ask me directly do I think the non-null labels map to natural language words in their own right? I would say that I think that most likely those words which do not have null-prefixes, map to natural language words. However most of my arguments are not conditioned on the notion of mapping Voynichese words to natural language words.
So when you say:
“This is a very big underlying assumption and your entire argument falls apart if they are not.”
I don’t make the assumption that they are natural language words, though it would be my conclusion that it is highly unlikely that they are not natural language words.
Previously I questioned, understandably, whether spaces between words in the Voynich correspond to real spaces, having studied the labels in detail I conclude that it is pretty likely that Voynichese spaces are natural language word separators, i.e. real spaces.
Mark, you are quite right. The six points of “evidence” is what I meant.
It has long been clear to me that you are not too worried about comments, and I see nothing to be gained from me making any more.
@Mark
Words such as arad,arae, arat, arai, aram, aras, arar, are all normal Latin words. Even oror is one, and comes from burn.
I think slowly You have a bad e-translator.
Mark Knowles: “Previously I questioned, understandably, whether spaces between words in the Voynich correspond to real spaces, having studied the labels in detail I conclude that it is pretty likely that Voynichese spaces are natural language word separators, i.e. real spaces.”
Then you will have to account for the odd statistical properties of the VMS tokens (length, entropy, frequency), and the positional peculiarities of specific glyphs.
I don’t think that declaring that a high proportion of the text as nulls sufficiently accounts for this, especially when the argument itself (for specific tokens you picked out to be nulls) is not convincing—it appears to overlook certain properties of the text that might have other explanations.
Mark: whenever you say (and you’ve taken to saying it a lot) things (as here) along the lines of “I have explored alternative hypotheses […] and they might fit in one instance, but without them becoming very complex or convoluted and/or implausible/far fetched, they don’t cover all instances”, just about everyone else facepalms.
This sounds as though you’re judging your ideas necessarily right because you’ve run out of practical or even combinatorial imagination? I really don’t like being in the position of having to tell you how ridiculously shallow that sounds, but here we are, I guess. 🙁
What are the other explanations for the observations that I have made?
If you could come up with an alternative that explains all the observations as well and is generally consistent and is not vastly more complicated then I would be happy to hear it. (I can contrive much more complicated explanations designed to explain away the observations, but they lose by a long way when faced with Occam’s razor.)
Nick: I am judging my ideas right of course on the basis of my thought processes, everybody does.
If someone can come up with an explanation that I have not considered for the observations that I have made and that strikes me as more plausible then that’s great and that is also consistent with Occam’s razor which you may not value as a principle, but I do. I can’t “imagine” one and I have heard nobody else state one here or anywhere that I have read; of course I haven’t read everything. This is a conclusion that I came to on the basis of observations and was to some extent unexpected to me.
Mark: There are occasions where I have been supportive of your more sound ideas, especially those concerning letter/word nulls and how discriptive dialogue somehow fits too neatly within a set space bar created around an illustration. This supporting the real possibility of natural or God forbid artificial language coming into play. I have also been critical of your seeming uppity habit in ignoring and being dismissiive of any who might include a bit of chummy wry humour when agreeing with some of your ambitious or zaney theoretical propositions. You seem to have developed a recent tendency to go into attack mode when confronted by your several irrellevant bullying, critics, who understandably may be in awe of your sound though mainstream dismissive developments or else have fears that your general lines of inquiry may well prove counter productive of all their own fine efforts when medals are presented. Much to the chagrin of the Vm clique you have by fault gained my enthusiastic support for the present, if only because of my desire for you to succeed in some of the endeavors that you most fervently believe in….Alas my voice is being mostly stifled through admittedly infrequent rejection of my posts, against the generally accepted ‘post or perish norms’ for bloggers who publish without fear or favour. I do have certain obligative bad feelings that I could be responsible for the manic dismissive traits coming to the surface of late i our more or less respected blog master; namely towards nonsensical attributions of Vm authorship pointing to the unthinkable logic based proposition of more recent creation. The dambed hide of mere minnow upstarts pouring water our tried, tested, allbeit to date miserably failed interpretational theories, all based on perfectly sound evidence and science. Those based on a sound understanding of medieval knowledge, passed down for accreditation by those who may have created much of it so as to enhance their own proud legacies.
Mark: I’ve had a think about this overnight, and it has become clear to me that your entire line of reasoning here is based on what is known as the Holmesian fallacy. This is an “appeal to omniscience”, that we can enumerate all possible explanations and then merely score through those that we can disprove, leaving (hopefully) a single answer.
Unfortunately, History makes asses of those who think they can out-reason her in this way: the truth (if it ever emerges) is almost always much subtler and nuanced than we allow for or imagine.
This is, of course, extraordinarily frustrating, because it means that we instead have to content ourselves with explaining smaller pieces at a time, and then – even more patiently – try to fit those pieces together to slowly move towards seeing the larger picture.
Here, you’re trying to use null words as a Royal Road to explain the whole of Voynichese all at once: the problem is that this is not categorically different to Gerard Cheshire’s use of an imaginary (and also fallacious) polyglot proto-Romance language, Torsten Timm’s autocopyist theory, Gordon Rugg’s Cardan grilles theory, or even Richard SantaColoma’s it’s-a-modern-hoax leitmotif. These are all routes which offer the same kind of “complete feeling” (i.e. “I can explain the whole of the Voynich Manuscript in one go”) at the expense of going in the wrong historical direction completely.
For you, null words are particularly appealing because they help you explain away things that don’t feel right to you: and the more things you see that don’t feel right, the more null words you see. But at the heart of this exercise, that’s just explaining away, rather than actually explaining: and there’s an ocean-sized gulf between the two. :-/
john sanders: I’ve approved more than a thousand of your comments, so you’ve had ample bloggish oxygen here. However, I have had to moderate out a fair few of your comments in the last few months not because they were trolling (though this has also happened), but simply because they were just too boring for me to read – and if they’re boring for someone with my level of patience, they can only be extraordinarily boring for other people unfortunate enough to have to suffer through them.
So I’ve allowed this comment through not because I like it (because I don’t, it’s a nasty sniping little piece of work that adds nothing to anything), but as a sorry lesson to others of how not to do it.
Hi Rene,
Thank you for this also. You and JKP have kindly saved me duplicating a lot of work which has previously been done (and which I’d likely have done less thoroughly).
The biggest discrepancy is that the recipes of the alchemical herbals are full of short phrases that keep coming back in many of the recipes. This we do not see in the Voynich MS.
Exactly – plus the lack of ‘unique’ words for plant names and the like, which you really would expect to find in there too (Montemurro/Zanette).
I’d liked the idea of extensive use of null characters as a low-tech way to conceal these features, especially if there were rules about where, when and how specific graphemes could be used as nulls. But if this would leave too short a plaintext once these were removed…looks like I’m back to asking if the text is even meaningful or not 🙂
The entire text being meaningless is a realistic possibility for me.
However, we completely lack any insight into how this meaningless text would have been generated.
That shouldn’t be seen as a disadvantage of this option, because it is also true for the alternative: a partially or fully meaningful text.
(Given that we have at the moment not a single theory that explains everything, Occam’s razor does not yet come into play. Which is better anyway, since I consider it a blunt instrument).
Rene: I doubt that it is necessary to have a theory that explains everything. If we could genuinely explain even one thing, we would have made an extraordinarily large step in the right direction.
Nick: Firstly, my point is what to infer from my observations about a small set of words, particular though not exclusively based on a close study of labelese. So at its core it is not a general theory, but rather a specific localised theory. However on the, I think reasonable deduction, that these small number of words are very likely to be null words it poses the question as to how many null words there are. This is where my thinking becomes more speculative. Having observed that there are a significant number of other words amongst the labels, in particular, that exhibit very similar patterns and structures, while there are other words that deviate markedly from these, I think it is a reasonable working hypothesis that null words are widespread, not something I anticipated. So I am exploring this more general idea. This general hypothesis does not claim to explain everything. I have not yet identified all the words I think that are likely to be null and to do this is a nontrivial, though doable, task, which will take a combination of more detailed observations and computation force and even then the odd example may escape the net, hence I have made it clear that I have made a very imprecise guesstimate of the percentage of null words at 60%. This general hypothesis does not claim to explain any of the patterns found within non-null words, at all, and without having studied those patterns I freely admit I have no clear idea what to expect.
So I am definitely not “trying to use null words as a Royal Road to explain the whole of Voynichese all at once:”. So in fact this is quite different from Gerard Cheshire’s, Gordon Rugg’s, or Richard SantaColoma’s( theoryI don’t know Torsten Timm’s theory, so I won’t comment on that.) They claim to explain everything with how and what Voynichese is. My working hypothesis does not claim to explain all Voynichese, though does claim to explain a sizable proportion of Voynichese.
It is not about what feels right to me, my feelings don’t come info it, it is about trying to construct a theory/hypothesis that I think best explains my observations. I can’t “imagine” what I think is a better explanation and so far I have got the impression that nobody else can imagine a better explanation; now we all might be lacking in imagination, but with that attitude one would reject most ideas, one has to do the best with the imagination that one has to work with whether just mine or collectively.
Contrary to your stated position I think it is perfectly reasonable to consider all possible explanations that one can “imagine” and see which explains the observations without becoming much more complex than is necessary. We all rely on the reasoning processes mentally available to us and if that counts as a claim to “omniscience” then so be it. I am aware that this is a very bold hypothesis that challenges many pre-existing ideas and orthodoxies, but that is not why I have formulated it, though sometimes it is very valuable to challenge orthodoxies.
Mark: the presence of null words is not a “very bold hypothesis”, it’s just one of many thousands of other similarly unverifiable speculative hypotheses that have been put forward over the decades.
By way of contrast, if you had put forward some technical means by which your putative class of null words could be identified, that would be something genuinely new and interesting. But you haven’t.
Nick: The means of identifying null words I would suggest is if they satisify all the criteria that I have listed.
Nick: I guess I’ll have to be man enough to admit that my at times, rather foolish contributions could be considered a tad boring to some; like your average Averlino 1465 bee class supporter for instance. Just so long as I’m able to extend occasional meaningful pleasentaries to the Rich Santacolomas, Rene Zandbergens, Mark Knowles’s, Diane O’Donovans and other poor misguided Vm fools of that order. It does take one to know one Nick, the trick being if one wants to set traps for fools, they better not be set within one’s own blog site. PS: I note that you no longer call me john, but no mind Nick old mate.
John Sanders: Thank you for your support, it is really appreciated. I am sorry if at times I am rather humourless. I am sure at times that I am dismissive where I shouldn’t be. I aim to present ambitious propositions and I think there is sometimes a place for zany propositions. I think I don’t always respond in the best way when faced with unhelpful comments. I think at times I can be subject to anger and somewhat stubborn in changing my ideas. I try to be self-aware when it comes to these things. Similarly in all honesty some of these failing are far from only mine, but also shared, and sometimes to a greater extent, by some of my critics, who may at times be less self-aware. I find your writing style interesting and distinctive, amongst other reasons, so I value comments.
If there are null words in the Voynich this points to a greater likilihood that Voynichese is a cipher as if it were just a natural language with essentially a sum kind of character substitution it wouldn’t contain null words. Likewise if it were just an abbreviation system then it wouldn’t contain null words. Now this, in and of itself, does not preclude the possibility that it is a mixture of cipher and other form like a simple natural language substitution ot a system of latin abbreviations.
Certainly if it is a cipher then this fits neatly with the presence of the “4o” character as well as other characters shared with cipher alphabets.
Another and the initial motivation for exploring the possibility of the presence of null words at this time was that I observed that there was one word on the 9 rosette, which I was quite confident about its meaning, and was not consistent with my text identification and I then found the same word on the 9 rosette page on a completely different label that I was pretty sure does not refer to the same thing. I then noticed that the word was also on other labels in completely different sections of the Voynich manuscript and it exhibited the other properties that I have discussed. Now I obviously don’t expect anyone to accept my text identification at this time, but I thought it worth providing this as one more reason that personally lead me to reach my conclusion.
I would guess that words like “daiin” are also null. It is possible that words beginning with a “4o” are also null, though I haven’t yet studied these words and would need to in order to form a clear opinion.
JKP: You say:
“Then you will have to account for the odd statistical properties of the VMS tokens (length, entropy, frequency), and the positional peculiarities of specific glyphs.
I don’t think that declaring that a high proportion of the text as nulls sufficiently accounts for this, especially when the argument itself (for specific tokens you picked out to be nulls) is not convincing—it appears to overlook certain properties of the text that might have other explanations.”
First of all I don’t need to account for other phenomena outside of a null words theory, as it makes no claim to explain any properties of non-null words. It is very possible that a null word theory may account fully or partially for many, if not all, of the features you describe such as directionality and positionality amongst others as these are things that from my observation appear to be most evident in words which have the null properties that I have described compared with words that don’t have those properties.
You haven’t explained why the argument is not convincing or what the other explanations you suggest are.
Mark: you’re being more than a bit glib here.
Nick: Well, if that ad hominem is the extent of the critique of my hypothesis, then I am doing well.
Mark: sorry about that, what I should have done was made it clear how what you were claiming was glib.
Until you can properly say why you think some words are nulls and others aren’t, your null word theory isn’t explaining anything. The only words your theory has specifically identified as nulls, as I recall, are okar / okal / otar / otal. But given that these make up a tiny fraction of the total words in the Voynich, your theory still has a very long way to go.
So far, then, your null words theory doesn’t even account for itself, let alone for anything else in the manuscript. So I must admit to have fallen into the trap of thinking your comments sound a bit glib.
But perhaps they are.
Nick: I agree my theory has got some way to go. I think I have been pretty transparent what I can say with varying degrees of confidence in each case. I can say and have properly said why I think some words are nulls. I can certainly add words to that very short list. though at this stage it will still remain a relatively small list. However I have presented a hypothesis which is a development of those identifications and seems to fit with those word identifications as there are other words which I think have a lot in common with those words that I have observed even if I don’t have quite the same level of confidence as I do with the other words.
In so far as some of the words that I regard as non-null, I would point more to words which have their own unique spellings and so distinctive in spelling compared to other words. Obviously I would expect there also to be non-null words that are not quite so distinctive, but the more different a spelling is from the spelling of other words in the manuscript the more likely I think to be non-null as this is the opposite of what I see with the words I think are null and more consistent with a natural language. I think it would be possible to introduce a formal method of comparing the similarity in spelling of two different words, i.e. something based on things like relative number of common letters and relative position of common letters. Then I think one could to introduce a reasonable measure as to how similar a word is to each and every other word in the Voynich and then sort words by that aggregate measure. The higher up the list the more likely to be non-null words. I haven’t yet thought of a sure fire way of partitioning the list of words completely into null and non-nulls, but this requires more observation and studying of the text and more imagination and research into algorithmic solutions. Probably partitioning the set would require a combination of computation and manual efforts. Though I have ideas as to how to make progress without partitioning the set completely. Also I have a suspicion, based on observation, that some long words may be non-null words with null text prefixes, if so on wishing to partition to set of words one would need to take account of them. Anyway, a lot more observation and thought is required. However I think I on the basis of a very small number of words there is a strong case that they are null. In fact obviously, it only requires demonstrating that one word is a null to demonstrate that there are null words in the manuscript and therefore we are operating with a cipher. So in conclusion I have varying degrees of confidence/speculation that is why I use the term working hypothesis.
I don’t know precisely what you mean when you say the theory doesn’t account for itself. My point was to challenge those who say these ideas are wrong and to find out why they think they are wrong.
Nick: Even if one identifies exactly which words are null and so removes them from the text one is still left with starting afresh trying to make sense of all the non-null left.
Mark, if you look back at the thread, my comment was in response to you saying this: “…having studied the labels in detail I conclude that it is pretty likely that Voynichese spaces are natural language word separators, i.e. real spaces.”
You seem to have answered something else.
You refer to null words in one breath and then “natural language word separators” in another breath and then say that you don’t need to account for “other phenomena outside of a null words”.
It was you who implied that a study of one gave you insight into the other. In fact, “conclude” is a pretty strong word for something you subsequently say you don’t have to account for.
Nick: I think one could define a measure for each word of the degree to which it is repeat, this would be a function of the length of each repetition sequence and the total number if distinct instances of repetition. So words could be ranked by the degree of repetitiousness using that aggregate measure. The less the degree of repetitiousness the less likely it is to be a null.
Also a figure for the number of times a word appears in the Voynich could also be used as guide to whether a word is null or not; the more often a word appears in the manuscript the more likely it is to be a null everything else being equal.
These measures could by combined with the measure of distinctiveness of spelling to produce one function/formula for estimating the likilihood a word is a null or non-null. Clearly this function will be to some extent arbitrary, but that does not mean that if well constructed it could not be very useful. Obviously if one is lucky enough to find a clear discontinuity in the function then one may be able partition the set of words. I think also looking at the way words with similar spellings cluster together would also be useful in distinguishing the two kinds of words.
If 5% of distinctive words are misidentified and included in the wrong category it would probably be sufficient still to carry out productive analysis of the non-null Voynichese text. So achieving a perfect partition is probably not essential in progressing. In fact progress may possibly be achieved using a very imperfect partition where many non-null words are rejected.
Nick: Another thing one could is with a given word is count the number of times it appears as a label in different sections of the manuscript, so if a word appears as a label in the recipes section and the cosmology section for example then it will receive a score. The higher the score the more likely the word is a null everything else being equal. Again this measure could be incorporate in the general function that I have describdd.
Mark: sorry, but you’re just pulling this whole sequence out of the air. You have no idea how likely any of the things you’re talking about are, and that’s the good part. 🙁
JKP: The question of null words is independent of the question of whether spaces are spaces. The reason I said what I said about spaces was that we find many of the same words in single word labels in sentence text in the Voynich. So in the context of these words it is pretty likely that they are the same word in the sentence as in the label and therefore that the spaces around those words in sentences are real spaces. There are many words for which that is the case. This would mean that many spaces are real spaces. So by extension it seems to me that if many spaces are real spaces then it is pretty likely that all spaces are real spaces.
Words in sentence text in general don’t seem noticeably different in their structure and features to words in the labels, so again this is another reason I think they are the same kind of “word” and so spaces are real spaces.
Nick: Sure I have no idea how likely any of the things I have said are that is why I use the term hypothesis and anyway so many things in this kind of research are not certain. However these ideas of a sequence are consistent with the hypothesis which was and is formed on the basis of observation. Uncertainty is not a reason for not exploring ideas. I think I have been clear about what I am more or less confident of.
I think the thing I find difficult is I feel that you demand proof and certainty from the theories of others that you don’t come close to with your theories and don’t expect of your theories, which does seem to be a double standard.
Mark: it is a hard thing to point out, but drawing inferences from rosette labels you cannot read based on a theory about the rosettes you can glimpse but not prove, and then expanding out those inferences to the way text must surely works on the other 200+ pages of the Voynich Manuscript seems like an almost guaranteed way to fail.
Even if you were to say that there are a hundred different indicators you could possibly use to help point to null words, this wouldn’t help if those indicators also pointed to genuine words. As just about anyone who has actually tried this kind of experiment on Voynichese will tell you: Voynichese words strongly resist being categorised, and it would be a genuine astonishment if anything you can try to tell the difference between null words and genuine words will work for more than a page or two before collapsing in an ignominious heap.
The #1 problem is that you’re not starting from statistical evidence, you’re starting from what seems like annoyance that a rosette label that you had high hopes for as a crib turned out to be a dud (specifically, a label that appears in lots of other places). But that’s not going to get you anywhere, fast or slow. 🙁
Mark: as an aside, I don’t ever “demand proof and certainty”. What I do do is point out when people replicate the many, many, many mistakes I and other Voynich researchers have made over the decades. I don’t know the whole Voynich story (and that is a peculiarly Baxian type of criticism to make, which has zero value), but I do know when people are making methodological mistakes: but what the hey, it’s only your life you’re wasting so why should I care? etc etc
Rather, what I hope for is for people to follow productive lines of inquiry which will help everyone learn more. Perhaps it will turn out that there are indeed null words in the Voynich Manuscript, as you propose: all I’m saying here is that if you carry on without ‘running the numbers’, it won’t be you who finds that out.
Mark Knowles wrote: “First of all I don’t need to account for other phenomena outside of a null words theory, as it makes no claim to explain any properties of non-null words.”
but then Mark Knowles wrote: “…having studied the labels in detail I conclude that it is pretty likely that Voynichese spaces are natural language word separators, i.e. real spaces.”
and Mark Knowles subsequently wrote: “The question of null words is independent of the question of whether spaces are spaces.”
If you believe statements 1 and 3, there is an apparent contradiction between them and statement 2.
They cannot be both null and natural language. And you are accounting for phenomena outside a null words theory if you “conclude that it is pretty likely that Voynichese spaces are natural language word separators…”
Also, as to statement 1, it’s pretty difficult to form a theory about null words without at least some understanding of the properties of non-null words. Otherwise how can you be sure you are correctly distinguishing one from the other? Without understanding how Voynichese works, you might be mistaking meaningful tokens for nulls.
Hi Nick
Did you notice? You used “many” three times in a row.
Which is the zero word? 🙂
Nick: Our theology adopts the little known Macco’s laser principle, which by extension incorporates Johnson’s elementary KISS doctrine; so as to include the simple logic based sub clausal “when you’re on a good thing, stick to it”… I would not in all decency stoop so low as accuse my presumed intellectual betters of being “extraordinarily boring” by definition. A rather unflattering label that I must be prepared to own if I’m to have any realistic chance for Vm input in future. Best hopes for resolution of the current nulls impass which may enable hopefully more agreeable fifteenth century absolute truths to prevail. Meanwhile allow us lowly “confederacy of dunces” to get on with our own frivolous and trivial pursuits svp.
Nick: That is a complete mischaracterisation. My theory is not just based on one rosette labels as I have made crystal clear. I have looked at all the labels in the Voynich.
You have really constructed a quintessential straw man argument.
In terms of what other people have tried, from what I have read or anyone can seem to point to nobody has explored this line of research, so there is no evidence that anyone has tried and failed or determined there is nothing in it. I am sure I can provide statistical analysis, but one is still left with what one infers from that.
You are right when you say “it’s only your life you’re wasting” if indeed I am wasting it, which my assessment is that I am not.
JKP: They are clearly not contradictions, it you cannot see that then I can’t help you. This statement was an aside and I never claimed it was part of the null words theory/hypothesis. In addition it is a complete waste of time discussing this as this is not the subject at hand, but similar to previous discussions we have had.
First of all stating certain words are null is not a statement about non-null words. I have also stated what I view as the distinction between the two notions. The point that I was making was that I don’t need to have any idea about how to decipher non-null words as part of a null words theory.
You say: “Without understanding how Voynichese works, you might be mistaking meaningful tokens for nulls.”, but also a full understand of Voynichese is not necessarily a prerequisite for identifying null words.
I didn’t say “a full understanding”, Mark. You added in the word “full”.
Mark wrote: “The point that I was making was that I don’t need to have any idea about how to decipher non-null words as part of a null words theory.”
Yes, I know that’s the point you were making. And I disagree. I didn’t say “decipher” either. I said “understanding”. It’s possible to study structure and gain a certain understanding of the text’s properties without deciphering it.
JKP: “understanding” is a very vague word in this context, so divining what you meant by that word is difficult. “a certain understanding of the text’s properties” is a bit clearer, but not a lot clearer.
Fine, we disagree.
Nick: You were talking about statistics. What I have been talking about is precisely compiling statistics such as the Levenshtein distance.
Mark: all I’m suggesting is that you should collect statistics before rather than after throwing theories out into the ether. It’s not rocket science.
J.K.Petersen: I’d go much further with that same idea, and say that I think the only way to decipher the Voynich Manuscript will be by understanding lots of individual pieces of structure, and then knitting those pieces into a progressively fuller structure. For example, even though Currier A/B is a good step, it’s only the first of the forty-odd steps that will need to be taken to cross the chasm. :-/
Nick: My theory is based on observation. It is normal to have a working hypothesis and collect statistics to see if they fit with the hypothesis. I have good reasons to believe a certain very small set of words are null words. Now it is possible that all words are null. My inclination is to think not as there are some words on some labels with quite distinctive spellings. On that basis and other observations of a number of words very similar one from the other with what seem like similar properties to the “known nulls” and a number of distinctively spelt words I formed a hypothesis that there are a large number of null words, but not all.
Mark: a hypothesis that says “based on cursory visual examination, I think some Voynichese words might be nulls” would be a bit lightweight and speculative, but fairly uncontentious nonetheless. But a hypothesis that says “based on cursory visual examination (and without even the faintest flicker of statistical analysis), I think 60% or more of Voynichese words are nulls” is just laughably rubbish.
Go to voynichese.com and try some experiments out for yourself. For example, there are only 86 instances of otol, 46 instances of otor, 83 instances of okol, and 34 instances of okor. If these are your big hopes for null words, you’re miles off even 1%, let alone 60%. Statistics can (and indeed should) be the theorist’s friend: but making a quantitative claim based on zero statistics at all makes you look foolish, sorry. 🙁
Nick: By using the word “cursory” you assume it was cursory. How do you know that? How much visual examination of labelese would you say is not cursory? I stated clearly that the 60% figure is a very vague guess. Again we have a straw man.
I have been on voynichese.com and I have tried plenty of experiments. I didn’t say those constituted all null words. I said these were a very small list.
I formed a hypothesis based on specific careful observation, which I think is a perfectly valid approach. What is the point in not addressing what I have said rather than a complete straw man built on what I have not said?
I think statistics are very valuable and I am in the process of compiling statistics. However I think, and I am sure you do, that there is a place for observation.
Ab initio. Aut simul stabunt; aut simul cadent.
Mark: in the end, you have put forward a whole lot of claims about the significance of the putative null words that you think you can observe, and a lot of I-think-if-you-do-this-kind-of-statistical-test-it-will-show-that-I’m-right commentary attached to those claims.
For my (many and varied) sins, I have read all your comments here, but am still absolutely none the wiser about what it is about these words that makes them look like null words to you. If I thought you had any kind of statistical backing to your ideas, I’d test them: but I don’t believe you have.
Just about the only specific observation you have discussed is that you started looking at them when you found that a particular label on the nine rosette page that you thought ‘must’ be connected with your other Voynich theory turned out to be something that appeared on other labels. But even that is not an observation, it’s a whole heap of disappointment that the crib your other theory predicted would be ‘place name X’ turned out to be a bust. So I don’t yet see a theory, a hypothesis, any evidence, or any observation here.
What you can’t seem to see is that your conclusion that this is a bust was solely the result of a whole load of other unwarranted assumptions you previously made about how Voynichese works – that it is a pure language (it isn’t), that it can be understand in straightforward terms (it can’t), etc. So you’re actually building a whole load of false premises for yourself, and then feeling cheated because they don’t all mesh together. And you criticize me for pointing out this nest of basic rookie mistakes? Riiiiiight.
Nick: The key to me more than anything else, though not alone, is the fact that we, not infrequently, have the same labels attached to many different and disparate things: a star, a water pipe, a nude woman, a plant amongst other things. In addition these labels that exhibit those properties normally have one letter difference from each. (At some stage I may make a systematic list of all the examples, though would be very laborious.) What can one conclude from that? when I have asked nobody has answered. So I have considered various explanations, for that, which I think is a perfectly valid approach methodologically speaking. By far and away amongst my alternative ideas and those of other I believe the null word explanation stands out. i.e. these specific label words have no meaning and are merely a filler designed to confuse the decipherer. Now I am inclined to the view, though at this stage the evidential basis is a weaker, that not all labels are null, but those with the more distinctive and unique spellings are meaningful. I could list a range of alternative explanations and why I think each is inadequate either, because it doesn’t work or is ridiculously complicated.
I have stated this point before and I have listed other points that I think support this argument, yet you seem to have latched onto the label on the 9 rosette page as the basis for my whole theory which I mentioned as an aside and I made it clear that this is not core to my theory. I have discussed what I see as the value of studying labelese many times in the past and the 9 rosette label is by far and away not the most significant of my observations. Yes, the rosette label got me thinking about the subject, but I was going to look at labels anyway. What made much much more difference in reaching my conclusion was looking at labels in general not the 9 rosette label, yet you seem to make it the fulcrum on which all my theory is based. I mentioned the 9 rosette in addition to my other points in that it was another reason that I think pointed towards my conclusion, but from far the central plank, I did make this very clear when I mentioned it. You must have seen my previous list and previous discussion of my other arguments.
I must say that prior to studying the labels in general I did not expect to reach the conclusions that I have discussed. Although I should reiterate that this is still a hypothesis and I have not claimed it as a proven result, though I have more confidence in some aspects of the theory than others.
I never said or thought Voynichese is a pure language, that should be obvious as I have long been talking about it being a cipher. One should neither assume Voynichese can be understood in straightforward terms or that it cannot be understood in straightforward terms. I don’t know where you get the idea that I do or have felt that I have been cheated; talk about a load of false assumptions.
What kind of statistical tests do you think would demonstrate to your satisfaction the existences or non-existence of null words?
I guess for me what would be nice is if someone has specfic suggestions how I could verify or discount the hypothesis. Straw man presentations of my ideas don’t help me move forward. You have freely admitted that you don’t know if there are null words, so therefore I assume you acknowledge that the hypothesis that there are null words in the Voynich is worthy if exploration.
Do we know how to contact Berj Ensanian?
Please don’t feel obliged to reply to me, I appreciate that you are a busy guy.
Mark, I believe that Nick’s argument was that what you are doing is far too superficial. Whether that is a correct interpretation on my side or not, it is certainly my opinion. You present a conclusion without having done the ground work.
You write:
“At some stage I may make a systematic list of all the examples, though would be very laborious.” So, you haven’t done that.
Why not present at least some specific examples of labels that occur near four very different objects?
You also write that you have considered various explanations for this behaviour (if indeed it exists to some statistically relevant level). However, we don’t see that.
What if the label words are not nouns (a standard assumption) but properties?
This is entirely possible and would explain why the same word can appear near different objects.
What if they are magical words to be uttered? Same argument.
What if they are abbreviated?
“asp.”, “ast.”, “ser.” could mean any number of things. Just a few examples.
I am trying to be constructive here.
Rene: I have observed a number of examples. I just haven’t gone through them for benefit of others as this is quite a job: i.e. things like
label otol – 1) folio 87v top right corner(star)
label otol – 2) folio 22r centre(water pipe)
(These are hypothetical examples)
In addition I could provide a series of screenshots with specific labels, but this would also require quite a lot of work.
I am not saying that I will not do this, primarily for clarity and the benefit of others, just not immediately. I will try to present one example with references shortly, though I have already listed some just not where to find them.
I am planning to list the explanations that I have considered.
I have considered properties as opposed to nouns i.e. adjactives. The point as I see it is that the same properties would then be shared with completely different things. So one asks what those adjectives could be. They would have to be pretty common adjectives such as “good”, “old”, “beautiful”, so much so that they could provide little or no information of practical value and I am inclined to think the author was more intelligent than that. Then how is it that these repeated “adjectives” have very often one letter difference from one another.
Surely if they are magical words it would make more sense if they were part of a sequence/sentence than isolated and if they are stars that are drawn identically how does one distinguish one from another when saying a magical word. It is possible and I freely admit not an option I have considered, but I think less likely. Given that these words also appear in different places in sentence text do they operate as magic words there. On the rosettes page how and when should one say a magic word, it seems had to conceive as the are so detached, i.e. often form no clear part of any sequence and can be quite distinct. Also on the rosettes page one would have to regard so many of the labels as magic words whether “monotonous” or “distinct”. As an aside I tend to share Nick’s view as the author being hyper rational, so the idea of the author writing gibberish or magic words seems less likely. It then seems to me one would move from a class of nulls and non-nulls to “magic words” and “non-magic words”. Anyway an interesting suggestion.
If they are abbreviated I would think in these different instances it would be nightmarish for the author to work out what they are abbreviated from, so much so as to be next to useless. In addition the one letter difference nature of these labels seem to be a bit suspicious. I like your examples of “asp.”, “ast.”, “ser.” . I have no idea in practice what they could be abbreviated from, so I would find them close the useless.
In addition there are other arguments that I have listed for the presence of null words such as the circular bounded text, which the alternative hypotheses would have to account for.
Then going back to Occam’s razor I think the null words explanation is the simplest that fit my observations. Note the phrase “simplest that fit my observations”, Occam’s razor does not mean the simplest irrespective of the facts, but rather given the facts. I can easily think of very complicated explanations that fit the facts, I could go into some of them, but not now.
You are being very constructive and that is much appreciated. The magic words option was not one that I considered.
Labels in different sections that are similar could most certainly be properties (hot, cold, wet, dry, or perhaps seasons in which the subject is relevant, or ruling planets (which ruled pretty much everything in those days), or the day of the month in which it was best to use or do those things (some manuscripts have pages and pages of days for doing specific things), or any number of different classifications that were popular in the Middle Ages.
They could also be instructions (heat, cut, dry, or perhaps turn, move, rotate).
They could be cross-references that match something in the VMS text (but which might include nulls in the main text to obscure the reference) or in another text. For example, I have seen an herbal manuscript that had cross-references to plants in a completely different manuscript (which I only recognized because I was already familiar with the other manuscript).
Mark: if you have not collected specific observations together in a sufficiently detailed and thoughtful manner, you have not even started. Occam’s Razor does not exempt anyone from doing the necessary work. 🙁
JKP: I have considered cross-referencing and this presents its own kind of problems. Such as in the cases of page with no obvious accompanying blocks of linked text how one matches up the label to relevant text. Yes, there are ways that in theory this could happen like another separate “Voynich” manuscript(Volume II) with all the references or pages of footnotes such as could be the pages of just text or some other complex means of referencing. However these other kinds of cross-referencing theories quickly become very convoluted or complex. There is more that could be said on this hypothesis. Again this theory does not fair well when it comes to Occam’s razor as it inevitably is much more complex.
As far as the properties, I have considered the kinds of things you suggest and I cannot conceive how any adjective, even the most general such as “good” or “bad”, could really work with the specific examples. Then again we have the one letter differences, which seem in addition hard to reconcile with the adjectives.
Nick: I have been careful in making specific observations, what I haven’t done is listed them systematically for the benefit of others. However I think you are right it is clear that in order to better present my theory to you and others I will need to do so as it may be unclear what I am talking about precisely. It should be noted that for my own benefit I have done quite a bit of leg work, so I have done far from a cursory study, though if I want to do so exhaustively there is quite a lot more to be done. I should add on the basis of the significant amount of work that I have done the presence of null words seems by far the most plausible conclusion, but I agree that that is my opinion, which others should certainly not take on trust, though if others can be bothered they can look up the specific words I have mentioned amongst the labels such as “okol” and “otol”.
Mark: when looking at the Voynich manuscript as a whole, otol and okol appear (as I’ve mentioned before) only a tiny handful of times. As a result, forming an argument about how Voynichese works from just these two rare words would be a case of the flea wagging the dog. :-/
Nick: When one starts to factor in variants, then there are a lot more, essentially words of the form:
F = “o”
G = “k” or “t” (and most likely “p”)
T = any character
L = “r” or “dy” or “ry” or “l” and probably others
where the word is FGTL (First Gallows Third Last)
That creates a large number of variants, all of which to varying extents are labels attached to more than one distinct and different drawing.
In fact as a really wild guess I think it possible that all words of 5 or less symbols and where the first symbol is “o”, “y” or “qo” and the second symbols is “k” or “t” (and most likely “p”) are all nulls, but as I say that is a shot in the dark. Obviously outside of that group there are others I would guess are null words.
Nick: In the manuscript as a whole these labels may not appear a large multiple of times, but amongst the labels they do.
Mark: yeah, we need more wild guesses.
Nick: They are not all wild guesses, I flag up them when they are. The properties I describe of “otol” and “okol” amongst the labels are very tangible. As with most things I have a spectrum of confidence, but I try to indicate where on that spectrum ideas lie. For example I would regard any word that has a second letter as a gallows character with suspicion, but I feel very far from confident in saying that all such words are null. Nevertheless I think even wild guesses could present themselves as possibly of interest to some reading these comments now or at some time in the future, I am trying to anticipate what a more general theory of nulls might look like.
As an extreme example, there could have been a 2nd Voynich manuscript, identical to the first we know in the sense that all the images are the same, but written in a natural language and script like say Latin or German. Then to find out what any text or word says in our original Voynich manuscript one merely needs to look it up in the 2nd Voynich manuscript; this be some kind of mega-uber-crib. This could explain the text, however it requires the giant leap of there being a 2nd Voynich manuscript. There could instead be other different and slightly simpler mega-cribs which would explain what we observe. There are, I daresay, a large variety of other complex and convoluted theories which could explain our observations. This is where Occam’s razor becomes important, i.e. the simplest theory that explains the observations. I don’t favour the simplest theory consistent with the evidence, because of lack of imagination on my part, I would argue, I can come up with plenty of complex wacky theories, but rather because I think it the most likely. The idea of there being lots of null words in the text is a rather dull one, which appears to wave away a lot of the interesting aspects of Voynichese, but the question to me is whether that theory is the right one, not whether it is boring.
Mark: Occam’s Razor is a useless, retrogressive, medieval knowledge hack that has been rubbish for at least five centuries. Its recent rise in popularity has been because it has been taken up with enthusiasm by idiot conspiracy theorists, for the simple reason that they almost always have no evidence, argument, or indeed idea for what they would like to claim – all they can point to is the ‘elegant simplicity’ of their attempted explanations (backed up by William of Occam), which isn’t even remotely a good enough reason to waste other people’s time. If you want credibility here and elsewhere, I suggest you quickly ditch Occam’s rusty and blunt razor, and instead start your arguments with evidence and direct observation (and preferably lots of both).
In the case of the Voynich Manuscript, you have complete sets of digital scans, 200+ pages of text and deveral complete transcriptions to work with, so it’s not as if you could claim that it’s hard to access it.
Nick: Yours is certainly very much a minority position amongst scientists from what I can tell. Nevertheless being a minority position doesn’t make it wrong. I imagine the problem with the conspiracy theorists that you refer to is that they incorrectly apply Occam’s razor. Occam’s razor in no way precludes complex or very complex explanations if there are no simpler ones that fit the evidence, it just has a preference for simplicity amongst hypotheses everything else being equal as far as them fitting the observations.
Yes, as I have mentioned before, what constitutes “simplicity” is a slightly difficult subject, though I think in practice it very rarely presents problems.
A few standard quotes that can be trawled up from the web:
“the supreme goal of all theory is to make the irreducible basic elements as simple and as few as possible without having to surrender the adequate representation of a single datum of experience.” Einstein
“It seems better to employ the principle known as Occam’s razor and cut out all the features of the theory that cannot be observed.” Stephen Hawking
“we are to admit no more causes of natural things than such as are both true and sufficient to explain their appearances.” Newton
Now as is well known I oppose appeals to authority, so one could take those quotes with a pinch of salt. Nevertheless I presented them as it hardly follows that one has to be a complete nitwit to think Occam’s razor is useful tool.
In fact one thing that came strongly to mind when thinking of Occam’s razor was the Black-Scholes equation for which that approach was vital.
Now clearly contemporary theories of biology and modern physics can be very complex, but that does not cut against Occam’s razor so to speak.
There are mathematical arguments in favour of Occam’s razor.
I have thought for a while it understandable why you dislike Occam’s razor as your approach often seems to be to try to find the most complex and intricate theory that fits the observed facts. So you read extra meaning and interpretation to darker versus faded penstrokes, you conceive complex numbering systems based on where the tail of a symbol(à la Newbold) crudely seems to terminate rather than viewing it as natural variation and so on. All these ideas are charming and enticing, but so often feel like flights of fancy. I can see why Occam’s razor is anathema to you, as you let you imagination run free, conceiving beautiful byzantine explanations where simpler one’s would suffice. I have been reminded of the interview you did with the Voynich Ninja where you said that you were writing a novel until you were drawn away by the Voynich. The “Curse” is in some ways a gripping novel. Your work inventing computer games is a testament to imagination and creativity. But I would suggest that understanding the Voynich unlike writing a novel or inventing a computer game is a scientific project of trying to discern the truth not invent what one would like to be the truth.
One thing that I did want to mention that does concern me with the null words theory is a kind of Ruggian problem, how did the author know which null words to use and when? There are two possibilities that come to mind:
1) He already knew the general structure and rules that the null words conform to and his imagination did the rest when it came to selecting the precise words to use in each instance.
2) He had a procedure for generating null words, this could be akin to a Ruggian Cardan Grille or there could be many other random number approaches to generating the words e.g. dice rolls.
I think of the two 1 is far more likely as amongst labels there seems to be a tendency, often for two words with one letter difference for neighbouring labels, so unless there was some kind of state machine this would seem to be a whim of the author. The idea of a state machine for null word generation seems an unnecessarily complicated explanation where I think hypothesis 1 would do the trick.
Mark: sorry, but you obviously don’t talk to many scientists or philosophers. (And don’t get me started on Hawking.) And Occam’s Razor not anything that applies to mathematics in any obvious way.
In many ways, my Averlino theory can be simply summarized as “The Voynich Manuscript is, in some way, Antonio Averlino’s little books of secrets“. The problem – as with every other theory, including yours – is that trying to apply that to the Voynich Manuscript quickly becomes tangled. At least I am honest enough to admit to those tangles.
Nick: Occam’s razor does not apply to Mathematics, that’s not what I wrote. Occam’s razor can be justified by probability theory.
I certainly agree that we are all faced with tangles, myself included.
Nick: When it comes Occam’s razor I think what some present as the simplest explanation is actually not.
For example:
The Voynich was written by aliens
Or:
The Voynich was written by God.
On the face of it these seem like nice simple explanations and examples of Occam’s razor in action. However I think that is misleading as in fact these are very complex explanations as “aliens” or “god” are very complex notions. It also raises the question of the empirical evidence outside of the Voynich for aliens visiting earth etc. So I have a suspicion that in the case of the conspiracy theories that you refer to what are claimed as the simplest explanations that fit the facts are nothing of the kind. Like any tool Occam’s razor can be misapplied and conspiracy theorists are known for their gift for misapplying logical approaches. Just, because someone claims to be applying Occam’s razor doesn’t mean they really are
2Mark
Occam’s razor is mostly about description of the process, not about the process itself. So for example in maths it’s applicable to axioms and definitions rather than to theorems. In programming it’s applicable to type and class definitions rather than to the code.
And while the presentation of any theory goes in order axioms/definitions, then theorems, the invention process usually goes in reverse order (strictly speaking it’s iterative) and thus Occam’s razor can be applied only when substantial part is completed.
As for now there is nothing to raze.
Mark: Occam’s Razor was essentially replaced by ideas of probability. Occam’s Razor is nothing to do with probability, it is to do with simplicity (in most formulations), which is why it is useless. It was a crappy medieval knowledge hack that has as much place in the modern world as the supposed spontaneous generation of worms inside cheese.
Nick: If you are looking to Mathematical formalisation and proofs of Occam’s razor Solomonov’s theory is not a bad place to start.
FarmerJohn:
I think you are right in the sense that Occam’s razor has its place in the decision making process, so as to decide where to focus one’s energy amongst competing hypotheses. Yes, Occam’s razor serves as a way of selecting a minimal set of axioms that model the system that you are looking to describe.
As you say in designing a programming language one would think of as simple a language as is fit for purpose, so unnecessary and pointless variable types would be eliminated.
I think it is correct to say the axioms of the model are selected after part of the process of applying Occam’s razor and certainly in building a model there can be an iterative process i.e. if I start with these axioms does the resultant model explain the empirical evidence if not how should I replace or modify these axioms.
When you say: “Occam’s razor can be applied only when substantial part is completed.” then I disagree. One does not need to build a global explanatory model, but can build locally explanatory models as a stepping stone to a globally explanatory model. So for example I might have a model that best explains a specific page of the manuscript given Occam’s razor when compared with other theories of the page.
As another example, whether my null words theory is right or wrong it doesn’t seek to explain all of the words of the manuscript only at its core a small number of words, so as long as it doesn’t contradict other evidence from the manuscript it can serve as a local theory to be compared in simplicity with other local theories of the same evidence.
I think Occam’s razor in the context of the Voynich can restrain the more fanciful and outré theories, keeping ideas more down to earth.
Mark: Solomonov’s theory is just Bayesian, so anyone trying – as the miserably wrong-headed Wikipedia article tries – to connect it with (the entirely non-statistical, and entirely non-Bayesian) Occam’s Razor is barking up completely the wrong tree.
Honestly, I really wish you would stop wasting your precious time trying to justify Occam’s Razor as something that somehow backs up your observations. It doesn’t. It won’t. It can’t. Occam’s Razor is a pile of shit. Conspiracy theorists like to wheel it out because they believe it gives a sprinkle of justificatory ‘science-flavoured’ pixie-dust to even their stupidest theories. But it’s still a pile of shit. Don’t use it here. Or anywhere else, if you’ve got any sense.
I hope I’m making myself clear here.
Mark: if you look at the Voynich theories that invoke Occam’s Razor as their support, you’ll see an extraordinarily long litany of foolishness. So your suggestion that Occam’s Razor “can restrain the more fanciful and outré theories, keeping ideas more down to earth” would strongly seem to run counter the way the world actually works.
…perhaps most famously “I intend to adopt in this article the heuristic of Ockham’s razor…“. And what a pile of shit that was.
Sorry, but it was.
Nick: I had to google that. Whilst I have not read Bax’s article I think it fair to assume that there is relatively little that I would agree with. I think when someone says “I intend to adopt Occam’s razor” to me it is like someone saying “I intend to be logical”, saying it doesn’t make it so. Occam’s razor is a well known principle, so it unsurprising that someone wishing to present a theory as being rigourous would cite it just as someone might cite various principles of logic or probability or statistics etc., citing is the easy part applying them correctly is the hard part. I suppose I ought to read Bax’s article to identify where I think he went wrong in his reasoning, if I can motivate myself sufficiently.
Mark: as Rene says (though I paraphrase), perhaps the only real use of all these terrible Voynich theories is to teach us all a collective lesson on how not to do it. Anyone looking for genuine meat on Bax’s bombastic bones will – in my opinion – go hungry.
Nick: Bax implies that the Voynich being “written in a natural language encoded in an unknown script” is the simplest explanation, I would question that given the evidence. It might seem to be the simplest explanation to a linguist, but might not appear the simplest explanation to someone familiar with Northern Italian 15th Century ciphers and there are evidential reasons in the manuscript to associate it with Northern Italy, i.e. the swallow tail battlements, and Central Europe. In a more universal context I would be inclined to Bax assertion as being the simplest one that fits the evidence, but given the other evidence that we have in the Voynich I would think that becomes questionable. In fact I can see his theory given the context potentially requiring a more complex explanation. Ultimately Occam’s razor is not a bias towards the simplest theory, but rather a bias towards the simplest theory that fits the evidence.
Bax was a linguist and I think, as I have mentioned, if Voynichese is not a natural language in an unknown script he was really out of his depth. He naturally saw what he was familiar with, not something quite outside his acquaintance.
Nick: I think we can certainly learn from others mistakes in constructing Voynich theories.
I think Bax’s assumption that we are dealing with a known or unknown language in an unknown script is a classic example of an assumption we cannot make and yet his whole theory hinges on it.
One thing that struck me about Gerard Cheshire’s “map” theory is that he starts from the assumption that the bottom left rosette is a volcano and treats the assumption as obvious with no need to justify it. You just can’t make that assumption. It is not even a widely held assumption. He then derives his whole “map” theory from that assumption.
I imagine that in forming many theories people make very cavalier and unjustifiable assumptions. However I would caution about assuming that is nothing of value in a theory even if one thinks it is fundamentally flawed, there may not be, but there may be if one is lucky.
I have serious reservations about aspects of your theory, not to compare it with those other theories, but I also think there are genuinely very valuable aspects to it. Theories don’t have to be all wrong or all right, we can be more discerning when studying them. Sure, theories can and probably often are nonsense from start to finish, but we should be open to finding gems amidst the muck. So in that sense I find Rene’s statement unduly negative.
The troubling problem of null words:
As I have mentioned before, I have thought the question of the generation of null words problematic and given further consideration of the question I find it more problematic. If there are this supposed large number of null words then how were they all generated. Yes, I believe they conform to, from what I have seen, a standard structure such as purely hypothetically, at this stage, a large group of them being less than or equal to 5 letters long and having the second letter as a gallows character. However even within this rigid frameword there is a lot of scope for variability and it appears that some substructures are more common in certain contexts. This would imply some degree of care in selecting null words in each instance. It would mean being creative and imaginative in choosing the next null word which requires a little effort. Alternatively one requires a complex mechanism for the generation of null words, but then one has to provide a plausible and evidentially consistent mechanism for so doing, which is a big problem in and of itself. Now I am sure that I could produce null words manually in a way consistent with what we observe without huge effort. However it would in total be quite a lot of effort in order to produce “nothing”, but still arguably it would have been worthwhile effort. This question still troubles me, but I don’t think poses a fundamental flaw yet to this theory. Actually other objections that have been presented by myself or others I find much less troubling.
Mark Knowles wrote: “Now I am sure that I could produce null words manually in a way consistent with what we observe without huge effort.”
A number of people have attempted to recreate VMS text, and often manage to recreate subsets of the text, but finding a method for generating the larger portion of the text may seem straightforward, but is rather deceptive.
Rugg’s method doesn’t really come close. He assumes a prefix/suffix system that doesn’t really reflect how VMS text is structured. Timm’s method shows better awareness of the dynamics of the text, but there are still levels of complexity in the text that haven’t been explained.
Now, you could cheat, and simply reproduce entire tokens and it would come out looking like… VMS tokens, but even doing it this way only tackles a small part of the problem. Tokens then have to be strung together in sequences that at least somewhat replicate the balance and form of the overall text.
Recently I have been reading a lot from Bax and I must say I completely agree so far.
Bax implies that the Voynich has been “written in a natural language encoded in an unknown script” is the simplest explanation”. And again I totally agree. Not that this has been proven, not that I am convinced, but no serious observations disprove it. And in my opinion all other theories are far more complicated or require unproven other theories.
Enter the dreaded Ockham… According to him Bax’s statement has to be the preferred theory.
Pointing to conspiracy theorists, telekinists, paragnostics, aliens,.. saying their theory is proven by Ockham’s razor, does not disqualify Ockham, it disqualifies the opinion that Ockham is useless, merely because idiots apply Ockham unjustified.
The theory that Voynich faked the VM might at some time have been the simplest theory, right now, knowing about the carbon dating for one, it is a “simple” theory based on an unproven theory that the carbon dating is an invalid argument. With all its complicated consequences.
Of course saying the VM is a cypher is a simple theory, but that requires another step. As long as there is no proof of it being encoded other than by an unknown alphabet the simplest theory above has to be preferred.
At first when people saw the Sun, Moon and Stars consciously their simplest theory was: they all revolve around the Earth. And Ockham would prefer that. Then people became aware of the complicated tracks of the planets. And Ockham struck again, simplifying that putting the Sun at the center. Then after Newton Ockham struck again making Sun and planets and the lot one system in which the Sun too has its (tiny) movement. And then Ockham strikes again when we have to realize that the whole Milky Way is one extremely complicated system of millions of stars, planets, etc., but still the simplest theory available to explain all. Since all along theories appeared to be wrong later on, Ockham did not prove anything – it prescribed the best way to look at things until proven otherwise.
When someone right now would say “the Earth is flat, it is the simplest way to explain my observations” from his personal view he may be right to call on Ockham, so be it, but loads of observations contrary to that will not convince many others to adhere to the Flat Earth Theory, because of that same Ockham.
Ger Hungerink: Occam’s Razor is a beautiful thing… but only in retrospect. For example, phlogiston theory was at least as appealing – Occam-wise – as everything else that had been suggested to explain combustion: but it was horribly, horribly wrong. From my perspective, then, the only thing that makes Occam’s Razor look good is hindsight.
My position is that if you have any way at all of assessing the relative likelihood of two competing theories, you should use that way in preference to Occam’s Razor.
In the context of the Voynich Manuscript, however, we have literally tens of thousands of statistical tests and results to work through: and it was clear – even to the Friedmans fifty years ago – that these results are flatly incompatible with a simple natural language, whether encoded in an unknown script or not.
The only way Bax was able to keep the foolish fiction of the correctness of his theories in the air was by rubbishing every single Voynich researcher who had run a statistics test. And that was wilful, shameful, feigned ignorance on his part (and I told him so). I believe that Bax knew better, but thought that it was better to try to find ways to pull everyone else’s pants down rather than reveal that it was him who was standing naked, not them.
Mark: it’s a tricky point to get across, but my hope was that my Averlino theory would turn out to be the “right type of wrong” (i.e. wrong, but built on good groundwork) rather than the “wrong type pf wrong” (i.e. wrong, and built on poor evidence and poor reasoning). It is entirely possible to be wrong and yet for the work that contributed to that wrong conclusion to have genuine (non-pathological) value for other researchers.
Nick: What I like about the Averlino theory is that it is a complete detailed theory, but yet plausible, whereas most other theories of the kind appear implausible. In many ways I think, despite the fact that I think it has flaws, it can serve as a model in some sense to other general theories. Overall your cipher ideas are clever and intriguing, but at times feels to me like a very interesting invention on your part rather a theory that immerges out of the evidence.
Ger Hungerink: I argue that the cipher theory is the simpler one as in the context of the time and place of the Voynich having a language known or unknown written in an unknown script was quite unusual. Are there other examples of Central European languages written in rare scripts from the period? Now, it can be argued that a one to one substitution cipher is the same thing. However one is a rare script used by a community as opposed to a script invented for the simplest of ciphers. So I think given the place and the time postulating a cipher in an invented script is the most natural and simplest. We can debate what we mean by simplicity here, but to reiterate when it comes to Occam’s Razor we are talking about the simplest theory GIVEN the evidence, not independent of it.
JKP: However one thing that makes it easier than in the case of Rugg and others is that I don’t need a means of generating all the text, only the subset that are supposedly null words. The words that I am inclined to think the most likely candidates to be null words are words that I have described as “monotonous” as opposed to “distinctive” i.e. words that fit a specific standard structure. Randomly manually producing specifically structured words is much easier I think than randomly producing all words.
A cluster of words like okol, otol, okor, otor plus many more with one letter difference between different members, if all null, require much less imagination on the part of the author to generate them. So whilst I have real concerns with this idea due to this issue I still find it the most plausible.
My take on Occam’s razor is the following.
It is the type of statement that seems to kind of make sense in a general way, but it is entirely qualitative. Almost like arm-waving. One cannot make any practical use of it. It certainly cannot be used as proof of anything.
The next problem is: who decides what is simpler?
As it is, Stephen Bax thought hat his solution was the simplest.
Rich Santacoloma recently stated at Koen Gheuens’ blog that his “Voynich faked it” is the simplest. Since I am intimately familiar with all its details, I can say that this a very complicated, contrived set of assumptions and arguments that is also full of contradictions.
Mark considers his “there are many meaningless words but not all words are meaningless” the simplest.
Clearly, the owner of a particular theory or hypothesis is the wrong person to decide if it is simplest or not.
Rene: all good points, thanks. I’d say that the number of people who are properly equipped to decide this is at most one, depending on whether or not you are an atheist. 🙂
Rene: It is not entirely qualitative. It has a rigorous foundation. As a simple example take a set of axioms that model a set of phenomena, one wants a minimal set of axioms i.e. there is no point adding axioms to a model that in no way improve the predictions in line with evidence even if they make the predictions no worse. Think of the axioms of Euclidean geometry there is no point adding more axioms where the current axioms serve their purpose(Non-Euclidean geometry is where one changes an axiom rather than add an axiom and the various kinds of it model something different in each instance.)
Occam’s razor, like so many tools of reasoning, is much easier to state than use correctly. Just, because someone says that they are applying logic and yet they seem to come to clearly wrong conclusions does not mean that one should conclude that logic is useless. Occam’s razor like so many tools of reasoning is very hard to apply in situations where we are having very complex discussions and assessing variables and data is difficult especially as there often is not a lot to work with, that’s why reaching the right conclusions is a real challenge when trying to get to the bottom of this Voynich mystery, So I think aspiring to apply Occam’s razor correctly just like logic in general is a worthwhile goal even if one can easily fall short.
Rene: I should say that assessing which option is the simplest given the known evidence is hard, but people have addressed this question and it is not arbitrary.
I am starting wonder if there are some “null drawings” i.e. pictures that have no reason meaning, but are just there to distract, some of images of naked women with weird contraptions come to mind. I just thought I would further aggravate those who are already aggravated by my speculations(that’s not really my intent).
Indeed like I said claiming a theory to be the simplest does not prove the theory. When two fulfill all observations one might prefer the most likely. If that is not part of the observations anyway… And indeed it depends on who is claiming the theory, like in my Flat Earth example.
The use of Ockham is valuable without imposing it to anyone in choosing it (for yourself, or pupils, or a wider audience, or most of the scientific community if they agree…) as the point of departure for research: this is what we got at least, now do you want to complicate things, prove it. One could call it the “Null Hypothesis” for scientific research. Start with the “Flat Earth”, look around, and reach to the Universe.
As with Bax’s “natural language” some would say “we passed that point, it is proven not to be a natural language, so now we depart from….”.
I have not read everything from Bax by far, but from him I got the impression he was often misunderstood e.g. in the sense that if he brought up possibilities he would later be “accused” of having claimed that as a proven theory. And I can quite imagine that happening in this community, like in many like this 🙂
Rene: I think when determining which is simpler between two premises it often is necessary to drill down to the specifics as is so often the case with these kinds of complex questions. Can we elaborate or flesh out our premises so that we can better assess them. This is often a very difficult and really challenging thing to do.
Mark: now you’re inventing a whole new genre of commenting, Ockspam’s Razor.
The day you or anyone else becomes omniscient enough to judge “simpler” in a rigorous and satisfactory way will be the day Skynet takes over the world. Which will, by my clock, be at half past never precisely.
Ger Hungerink: to be honest, I’d say that Bax’s biggest problems came much more from those who understood him than from those who misunderstood him. His plausibly-simple-sounding (and supposedly specific) claims about the Voynich Manuscript quickly proved to be both individually false and also mutually inconsistent: the way he responded by blaming an entire community of researchers (mischaracterising them as somehow hating on him) was shameful and wrong.
Your comment is a source of great sadness to me, because it means that the cheap fairground rhetorical tricks Bax used to bait both individuals (such as me) and indeed non-linguistic Voynich researchers live on long after his passing. His Voynich legacy would seem to be one of divisiveness and misrepresentation, which continues to demean us all.
Ger wrote: “I have not read everything from Bax by far, but from him I got the impression he was often misunderstood…”
I don’t think he was misunderstood. If you watch his video a couple of times, I think he honestly believed he was on to something, that he had discovered something others had overlooked. He was very excited about his discovery. Regardless of how he worded things (“provisional”, etc.), I think his initial impression was that he had opened a door to the mystery.
The reality is that he really didn’t understand VMS text well enough to describe it.
If you read all his papers a couple of times, AND you become intimately familiar with the VMS text, you will see that his reasoning was based on insufficient understanding of VMS structure. He consistently overlooked important possibilities that should have been factored into his analyses, and he overlooked or misunderstood the full context of the patterns he was trying to describe. Even when he tried to present context, it was too narrowly focused and based on faulty premises—the logic was flawed.
It takes time to become familiar with the VMS text, and to be able to describe it accurately.
Dear Nick & Pals (y[all know who is writing this short note) — please don’t ‘dump’ me ! If you really want to get a full translation of Fray Sahagun and his Nahuatl students enormous library Books One “Earthly Things” and Book Eleven “The
Gods”…
Please don’t let Mark what’s-his-name take over your entire Cipher Mysteries web -site. !
My wife’s favorite “item” is the goblet-shaped crocus blossom.
I was hinting at Bax’s “complaint” here: https://stephenbax.net/?p=1807
And although I think his suggestions are mostly too tentative, I still find them very amusing, very reasonable, and a good example of how to do proper research. He himself seems to be fully aware of the speculative nature of his remarks, citing him:
However, I feel it is important to emphasize that I am NOT arguing that the language of the VM is in fact a form of Romani. The evidence is simply too thin to make any such deductions at this stage.
I am sadly aware from past experience that some people in the Voynich community jump on the smallest statement, twist it, exaggerate it, and then misquote it to further their own agendas. (Previous experience has been compelling 🙂 in this respect.)
So let me state it loud and clear – I am not stating that the manuscript is written in Romani, merely that we should treat this as an interesting possibility and keep investigating it carefully and with an open mind to other possibilities.
[End citation]
By the way – this is not meant to rub something in: “I was not there”. As with Bax’s languages, I am equally impressed by Nick’s compelling cypher theories and explanations.
Although I still think the VM represents a meaningful natural (or invented) language in some way (even e.g. satirical nonsense) and that it is the theory to depart from, I am fully aware of many strange properties that hint to it being coded or meaningless. But I have not seen anything convincing enough in that direction to leave the simplest explanation of a language.
Ger Hungerink: if Stephen Bax had confined himself to making tentative suggestions about Voynichese, he would have merely been wrong. However, rather than accept that there might possibly have been something of substance to the (very basic and very obvious) comments he received, he went on the attack, trying to rubbish everyone who didn’t share precisely his view of how the Voynich ‘must hav’ been made. By doing this he formed an utterly unnecessary (and completely unhelpful) split between pure linguistic Voynich theories (which he liked) and cryptographic Voynich theories (which he loathed).
But this kind of divisiveness and confrontationalism was just shallow, facile thinking on his part: I, for one, see Voynichese as a text puzzle that is part linguistic, part cryptographic, and indeed part steganographic, all at the same time. I think we will need to solve at least parts of all these different aspects of Voynichese simultaneously to stand a chance of cracking it.
All in all, I think Bax’s posturing and fighting had the effect of putting Voynich studies back a decade.
Ger wrote: “And although I think his suggestions are mostly too tentative, I still find them very amusing, very reasonable, and a good example of how to do proper research.”
No, I’m sorry. Picking out a couple of words and figuring out a one-to-one substitution cipher that seems to work for those words, and then generalizing the system to ten other words and then declaring it to be a “provisional solution” and sending out news releases to all the major press agencies is not good research.
It’s very very poor research, and naive and presumptuous in the extreme.
Based on that alone, he published a video, even though he hadn’t even studied the structure of the text in the slightest detail! Even the logic in the video, the WAY in which he picked the tokens to “translate” and the way he did it, is very flawed and inconsistent. His paper on “dain” is equally flawed in numerous ways.
It’s very disturbing to hear someone refer to his work as proper research. It most assuredly is not. It worries me when people can’t see that.
If it were good research, I would be his strongest supporter. Even good researchers make mistakes, but if they are on the right track and capable of correcting their errors, I will be behind them, but as it is, I would never encourage anyone to follow Bax’s example. There are people on the Voynich forum who are doing much more enlightened and careful research.
If some labels in the Voynich are null then how does the reader work out what the labelled item is? This would seem to be a reason to doubt the idea that some labels are null. So if they are indeed null, then what is the answer to this question.
1) The null word labelled item is a red herring, maybe like one of the illustrations of naked women in weird contraptions i.e. the illustration like the label is meaningless.
2) The labelled item is obvious and so not in need of a label e.g. possibly in the case of some plants in the recipes section..
3) There is no need for the label as the illustration does not require one.
4) The reader is expected to guess the correct label given the adjoining real labels.
5) The labelled item is in excess or addition to the other similar items e.g. extra unnecessary stars on the star diagrams. So the item is surplus to requirements other than the null deception usage.
6) The item has 2 separate distinct labels attached to it, one null and the other genuine.
I have been revisiting the subject of null words as opposed to very abbreviated text. I always like to reconsider my perspective. A big problem I have with the notion of a verbose cipher of abbreviated text, a notion I think has some merit, is that I feel it would render the text of the Voynich virtually unreadable.
Take as an example the following:
“Th Nk fr yr tm ad patience”
(Where the author has abbreviated all the words except the word patience(
This corresponds to:
“Thanks Nick for your time and patience”
Given the length of most or many Voynichese words with a kind of verbose cipher of abbreviations I would have though that each word would be abbreviated to 2 or 3 letters long. This also seem to imply that abbreviations are not unique. By I mean that more than 1 word would often have the same abbreviation.
So take the example of words repeated multiple times and consider the following
“I loe loe loe”
Corresponding to:
“I loathe lonely lorrydrivers”
(1 unique Voynichese word mapping to multiple real words)
The former is unintelligible and hard to imagine how the reader could determine that it corresponds to the latter. Now this may not be a good example. Nevertheless it does seem that when we have 3 or more Voynichese words deducing the corresponding real word in the case of each abbreviation would be well nigh impossible. It just seems to me that if the author had written the manuscript in this way then he/she would have realised that he/she was now incapable of rereading much of what he/she wrote on the first page by reaching the fifth page. I can’t see how the author could have written a 200 page manuscript for which he/she was incapable of reading size chunks.
Again I find this hard with repeated labels as it just feels in many cases the abbreviations would be so short that determining the specific plant or star from the label would be more than difficult.
However one thing I do like about that theory is that with the null words theory one is throwing out a lot of text as being devoid of information and so in the case of labels it feels like a very brief abbreviation still has more informational content in it than a null with zero informational content. It just feels like a waste for the label not to convey some meaning and given the different spellings there would be scope for including meaning even if greatly abbreviated, but I guess one could say the same of all filler text in ciphers.
On balance I still lean to the null word theory as I can’t see how with verbose abbreviations one could generate generally readable text. I suppose some hybrid of the two theories could be conceivable, though it is hard to imagine how that would work. The kind of phenomena that Torsten Timm and maybe Rugg, like myself, point to seem to imply that there are some null words; they of course, unlike myself, think all words are null.
Another thing, how would the author insure that he used a consistent way of abbreviations?
So Thanks could go to Ts -> First letter, Last letter
Or Th -> First 2 letters
Or Ths -> 3 letter abbreviation
It ought to be consistent to make it easier to identify.
I guess also the question is how many letters will the typical abbreviation has?
If 2 Voynich glyphs map to 1 real letter then the abbreviation length will 2 to 3 letters. Will that make for readable text?
My way of resolving the abbreviation problem was by allowing the inclusion of spaces in the verbose cipher mapping allowing for real words to be longer than Voynich words. However having studied labels, I determined that I thought that would not resolve this issue.
With an abbreviation system it could be that rare specifc words were unabbreviated or abbreviated to a much less extent, as I did with the word “Patience”.
Mark: the problem here is that you’re guessing at an abbreviation system in a modern language that you know, and then hoping that you can draw conclusions about abbreviating in Voynichese. This is like asking a long-jumper to reach a triple-jumper’s distance – you can’t jump over all the steps that you need to get you there.
Scribal abbreviations were widely used in the 15th century and 16th centuries (Tironian notae had started to fall out of favour in the 13th and particularly 14th century), particularly in Italy: but my many attempts to find excellent palaeographical sources on these have all come to nothing (so far). When we understand these specific systems better, hopefully we’ll get to an improved starting point for talking about them, rather than just guessing.
Mark wrote: “a verbose cipher of abbreviated text”.
This is a little bit confusing.
Verbose is the opposite of abbreviated. When people have considered in the past that the Voynich MS text might be the result of a verbose cipher, it means that the text would have been expanded, not abbreviated.
Rene: to be fair, that is basically what I proposed in Curse 13 years ago, i.e. that Voynichese is both abbreviated (plaintext words are shortened) and expanded (verbose cipher) at the same time.
The idea was that this would leave the overall size of words broadly the same, but impenetrable unless you could reconstruct both the abbreviation and the expansion stages simultaneously.
Rene & Nick: For me the difficulty with Nick’s model is that we have Voynichese words each mapping to multiple different real words. So an EVA-> okal abbreviation would correspond to a number of different real words(such as could be argued in the case of labels). So as any word would have multiple meanings then any sentence could have many different readings. The problem with this is it creates ambiguity, which is the enemy of comprehension. So one wonders if the level ambiguity created would logically becomes so great on the basis of what we know of Voynichese as to render the text often unreadable by the author.
The verbose cipher that Nick talks of resolves the problem of the perceived uniformity of Voynichese words as glyph pairs or more can result in single real letters. So for my example EVA-ok -> latin-g and EVA-ot -> latin-v and also EVA-al -> latin-g too. (This could also apply to standard word endings) So this cipher would be verbose and homophonic. The problem with this system is that it reduces the length of the resultant real words, which means that words must be abbreviations as they would frequently be too short to be full-length words.
The way I resolved that I described elsewhere was by using space characters, so that the real words could be longer than one individual Voynichese word. The problem comes when one studies labels, which contain no spaces, as the problem remains the same as with sentence text.
The appearance of the same label being applied to many different drawings in the Voynich in addition to other observations has made me think that the likely conclusion is that there are a ponderous of “null words”. When it comes to the manuscript as a whole “null words” means filler text. The presence of words with more “distinctive” spellings compared to the more monotonous words such as the words beginning EVA-ot and EVA-op and EVA-ok would seem to stand out as an indication that there is real meaning somewhere in the text.(I also have other reasons that I think nonsense text is unlikely.)
I would add the Torsten Timm/Rugg theories may explain the more repetitive text, but they don’t explain the more distinctive text, I think.
I think also if we have one Voynichese word mapping to multiple real words then it would seem normal for the more similar the spelling of two Voynichese words the more similar the spellings of the resultant words, overall, though maybe not in every case. Or instead there could in theory be a mapping such as:
EVA-okal -> fish or house or rose (Using the English language as an example)
where the resultant words have completely different spellings. This would require huge Voynichese word mapping lookup tables.
I think for me the question comes down to looking at the informational content of each Voynichese word and if there is really enough varied informational content in each word to be able to carry the whole text.
This kind of problem is exemplified by repeated words. Can EVA-“daiin daiin daiin” carry enough distinct information in it for 3 real words? I don’t think so.
Mark: the whole daiin daiin thing messes up everyone’s neat hypothetical Voynichese grammar, not just mine. Luckily I proposed a separate mechanism to explain this. 🙂
Nick: I gave that as an example of repeated words, which in general poses that problem. To be honest very early on when I became familiar with the fact that there are repeated words in Voynichese I thought them likely to be null strings of glyphs. The frequent proximity of very similarly spelled words in the Voynich is a related problem concerning informational content.
So anyway that is why my preferred hypothesis is “null words” which would mean that the manuscript as a whole is a mixture of real text and filler dummy text, where the more monotonous text is the filler.
So in that model “daiin daiin daiin” would be no problem as it can simply be ignored as filler.
If we take a hypothetical(non-Voynich) cipher text:
adfr adfp adfr adfp adfr adfp adfr adfp
i.e. the words “adfr” and “adfp” alternating 4 times
If that were to contain a message I would have thought the informational content quite low as there is very little variation in the text and so what message could it carry, I would think only a simple one not a complex 8 word message.
If one was to “compress” the Voynich text then wouldn’t the resultant size be much lower that “compressing” a comparable non-Voynich text of the same length. It just seems that there just isn’t enough information contained within the text of the Voynich as a whole produce a real text of the same length. Working on the basis that the true text is significantly shorter than the Voynichese text is a way to resolve this problem. And working on the basis of there being a large amount of filler text is one way this informational deficit can be handled.
(By “compress” what I mean is the way that on a computer one might compress an image or a document to reduce the amount of space that it takes in memory. In this case I would be talking about lossless compression. For those unfamiliar, who might be reading this, you can find more about this process online.)
Mark: having written plenty of commercial compression software myself, I’m well aware about how that works. 🙂
You have to be careful with null words. Null words are what you use to explain away all the awkward behaviours you run into, but they don’t help you explain what’s left behind. And worse still, you don’t have any guarantee that you didn’t filter out (or mistranscribe or miscategorize) meaningful stuff as null words.
In short, null words give a false sense of reassurance that you’re making progress, but without actually making any progress. 🙁
If you had any analytical reasoning to back up null words, things would be different, but you haven’t yet found any. It’s not helping, sorry.
Nick: I know perfectly well that you are familiar with compression software that is why I wrote “For those unfamiliar, who might be reading this”, so not to confuse people who are not from a software kind of background as to what I am referring to by “compression”.
Nick: I think you are absolutely right, once you introduce the concept of “null words” you have to start to think of what “non-null words” are. In fact if the theory is correct then “non-null words” are what is interesting. So how do you distinguish between the two classes or categories of Voynichese words? Well, to start with my notion is broadly that “null words” tend to be monotonous and “non-null words” distinctive. So what does that mean? I think null words tend to group in large spelling clusters, where every or almost every permutation of glyph sequence is a Voynichese word.
So for example combinations like okal, oral, otol, opol (You could retain positionality and then generate every permutation.) I noticed that Torsten Timm has a good illustration of this kind of thing in his paper.
In languages like English it is normal for words to be in small spelling clusters such as:
calf -> calm, half, call etc.
However spelling clusters where all or almost all combinations of letters are distinct words are rare. This phenomena seems much more pronounced with Voynichese albeit positional.
So I would say that words in large spelling clusters would seem to either always or almost always be “null words”. Behaviours like words being very high frequency, repeated or adjacent to words with very similar spelling, whilst not guaranteed to be indicative of being null I would think make them far more likely to be null.
So what does this mean about identifying null and non-null words. Well Voynichese words that are more than one letter change from another Voynichese word are much more likely to be distinctive words and these words can be listed. So one can start to construct a set of prospective distinctive words.
The key is having a sufficiently large set of words that are extremely likely to be distinctive, so that one has a large enough dataset to analyse.
I certainly think the better I can identify the structures of “null word” spelling clusters then I can more easily differentiate these words from non-null words. Again words that start EVA-ot or EVA-ok or EVA-op amongst others look extremely likely to be null words. If I can confidently identify the null words then I can consequently identify the non-null words that I am really interested in.
An analogy comes to mind, think of an audio recording of a conversation taking place while a vacuum cleaner is running, now one wants to get at the conversation and remove the noise of the vacuum cleaner. However really the sound of the vacuum cleaner is the repetitive audio pattern and the conversation is like the audio noise. I think that “null words” are like the sound of the vacuum cleaner and “non-null words” are like the conversation. So to separate the audio I really need to familiarise myself with the sound of the vacuum cleaner i.e. I need to understand the null words to understand the non-null words.
Nick: You make a good point, which I had thought about, which is what if I can’t separate the two sets of words correctly then this will screw up my analysis. I think I need to generate a set of distinctive/non-null words and if in the process of rejecting null words I throw out some non-null words then that could be OK. So I don’t necessarily need to identify all non-null words just a large number enough to do reasonable analysis. However I must do my best to make sure that no null words slip into my non-null set.
I think I will study labels even more as this is where I think the distinction between the 2 types of words is most stark.
I am obviously persuing this hypothesis as it seems to me to best fit the data that I am aware of. Clearly, if I thought another hypothesis a better fit then I would explore that.
Anyway, Merry Christmas!
Hi Nick, I wasn’t too much aware that this was your theory/hypothesis. Its being both verbose and abbreviated remains a bit of a contradiction, which I consider problematic, but then there doesn’t exist any Voynich theory without major problems.
The word length is an issue for the ‘verbose cipher’ theory (as is the fact that there aren’t to many examples of such). However, the word length could be explained in different ways.
Needless to say, I am not easily convinced, and I am still waiting for the first time that I read a proposed Voynich solution of which I get a feeling that there it has ‘something’ .
Rene,
Of course, the problem is that if one engages in the hard work of research, one is always inclined to be most persituaded by its results. On the other hand, if one has to sit back and wait – as I do where efforts concern Voynichese – the lack of my own ability in that area and lack of any original efforts in it, as a result – mean that I’m hardly qualified to form a judgement, myself, and tend to believe half-a-dozen impossible linguistic and cryptological theories before breakfast.
Happily, the ones who work are best qualified to judge work done by others, so I – for one – don’t try to pronounce judgement, but leave that to the best able. It’s very interesting, though, to read the informed debate and discussion which is so freely engaged here.
Rene: You say that “the word length could be explained in different ways”. Can you describe these different ways? I haven’t seen arguments that satisfactorily resolve this issue. I have proposed my own verbose cipher that could resolve the problem, but I find it doesn’t resolve the word length issue when it comes to labels to the extent I hope it would.
I don’t think being verbose and abbreviated is a contradiction, just that I think widespread use of abbreviations in the Voynich creates serious readability concerns.
Diane: Nobody has yet developed a theory of the text to the point where they have rendered it readable satisfactorily. So all these theories remain speculative. This means that one has to form a hypothesis or hypotheses that best fit the facts as one knows them. I really like the verbose cipher hypothesis, but I don’t see how can it create a model for the text that fits in all significant cases.
I can see that any hypothesis of null text can seem like a cop out, whether it is the full on Torsten Timm/Rugg style “all text is nonsense” perspective or what I have proposed that “a large proportion of text is nonsense”. One can say that it is easy if one doesn’t understand something to say it is nonsense, which can appear to be really a kind of laziness. Although, the appearance of laziness in dismissing certain words as null doesn’t mean they are not null.
However, whilst I don’t have the kind of familiarity that Nick has with historical ciphers, I believe that having large filler texts is common with ciphers, so my own hypothesis is in keeping with common forms of cipher.
I think there is a disappointment that some people share in the idea that a significant proportion of text or all text is null. The idea that many of the patterns that people observe in Voynichese are more a function of nullity than meaning could be a little dispiriting; I just think at the moment that is where the evidence points.
Mark: null text could easily be a good adjunct to a theory that specifically explained the rest of the Voynichese text. (As you know, many fifteenth century diplomatic ciphers included null letter tokens and/or null word tokens and/or null bracket tokens (within which everything was to be considered null.)
But on its own (and without specific hard evidence to support it), null text is an antitheory, i.e. something that initially looks and sounds like a theory, but that in fact is neither testable (disprovable) nor useful (making testable predictions about the artifact). Gordon Rugg’s theory was (just about) testable, though he never actually put in any work to try to test it: and it was certainly not useful. Torsten Timm’s antitheory is not useful, and is not testable insofar as it seeks to explain away everything that he deems testable (and hence render them void), while weaselling around all the other text behaviours he’s not interested in.
If you have any actual evidence that you can bring to the party that demonstrates even slightly that any of Voynichese’s highly structured text patterns are null words, then that’s great. But every one of the specific examples you have given so far seems to fall into the category of “look, here’s a local textual behaviour that doesn’t fit the particular model of text behaviour I believe Voynichese follows, therefore it must be a null word“: which is, as I hope you can work out, almost certain to be fallacious reasoning. 🙁
Rene: my suggestion that the Voynichese text is simultaneously verbose and abbreviated (that is, the text was first abbreviated and then enciphered using a verbose cipher) is closely tied up with trying to answer the question of why Voynichese letter adjacency rules are so rigid (i.e. why letters are so predictable).
The first sideways step is to see that the reason common Voynichese letters pairs (e.g. qo, or, ol, ar, al, aiin, air, am, ee, eee, etc) are so tightly bound together is because they are verbose cipher pairs (and if you have heard a better explanation for this, please say).
The second sideways step is to see (as the late Mark Perakh proposed many years ago) abbreviation as an active mechanism within Voynichese: and so it is a balancing combination of (non-Tironian) scribal abbreviation and verbose cipher that I have been talking about for some years.
Mark Knowles wrote: “I believe that having large filler texts is common with ciphers, so my own hypothesis is in keeping with common forms of cipher….”
Large filler texts in ciphers? I have not seen evidence of this. Most of the medieval ciphers I’ve seen have no filler text. Some medieval ciphers have nulls, but they do not comprise the majority of the text and would not be considered “large” filler texts.
Parchment was expensive. Writing with a quill was arduous. Most ciphers were not sophisticated (they were easy to break). Secret texts were easier to keep secret if they were small and easy to hide (filler text would make them bigger). There was not much incentive to create large filler texts.
JKP: Firstly, I didn’t use the word “medieval”. Secondly, that is what Trithemius does essentially.
Parchment was expensive, yes. Although how expensive depended on how much money you had. If you were very rich then parchment was really quite cheap. Writing a 200 page manuscript like the Voynich with a quill or even a modern pen would be quite arduous, that’s true. Most ciphers were not sophisticated they were easy to break, so what do conclude from this that the Voynich is not a cipher? This is a situation where the Voynich’s uniqueness makes it hard to pigeonhole as comparison with any other contemporary document cipher or not falls flat. Secret texts were easier to keep secret if they were small and easy to hide, though I doubt that is a serious consideration, as you seem to imagine that the author expected to be travelling around a lot with the manuscript trying to keep it hidden from prying eyes transmitting it from one person to another, that may not be how the manuscript was intended to be used, so size may have not been a major consideration. There could be plenty of incentive to create larger filler texts if it made the manuscript that much harder to decipher, again look at Trithemius. Ask yourself, why people ever create filler texts. So there is more effort and space involved, but the greater overall costs may be outweighed by the benefits.
Mark: to be fair, there is no sign that anybody prior to Trithemius (and indeed hardly anybody after Trithemius, truth be told) used nulls on such an industrial scale.
But at the same time, relatively few ciphertexts have survived to the present day, so it’s hard to be sure what the practical culture of using nulls was.
One thing I have found continually in many areas of Voynich research which I think is mistaken is the consistent expectation that everything in and to do with the Voynich should have a contemporary precedent. However this seems to fly in the face of the fact that to the best of our knowledge there is no present for the Voynich. So as an example the assertion that the Voynich cannot be written in a polyalphabetic cipher, because polyalphabetic ciphers were not yet invented rules out the possibility that the author of the Voynich may have invented his own polyalphabetic cipher. Now I give that as an example, as I don’t myself believe it is a polyalphabetic cipher, but this could apply to innumerable aspects of the manuscript.
So clearly the author of the Voynich had influences, but we have no parallel document from that time, the work of Fontana being the closest in some ways I guess.
Another couple of examples:
I remember Nick saying, something like the use of symbols to represent spaces didn’t happen until the end of the 15th century, so this can’t be the case with the Voynich. But the author of the Voynich could have invented this technique independently, it is hardly beyond the wit of man to do so.
The idea that the 9 rosette foldout can’t be a map as nobody has seen a contemporary map quite like that, isn’t for me a reason to preclude it being a map.
I think there must be aspects to the Voynich cipher that do not have a close precedent and which may not have been transferred to subsequent ciphers.
I think we surely recognise, if we have a cipher, that there was some degree of invention by the author involved in it.
I mention this as JKP talks about the precedent of lack of it for filler texts.
I think to some extent the Voynich must be a dead end. By that I mean, probably due to its secrecy, some of the ideas and techniques introduced by it were not disseminated to subsequent people. In other words I doubt the Voynich influenced later 15th century texts or ciphers, though of course in significantly later periods I think it has been of influence right up to the Codex Seraphinianus.
It seems to be a stretch to view the Voynich as the only surviving examples of a number of complex cipher manuscripts from the time. I don’t see it as the sole survivor of a tradition, but rather as it already seems to us something unique, in so far as anything is unique. Clearly there were influences, which it would be foolish to deny. If Leon Battista Alberti, Abbot Trithemius, and Vigenère as Nick lists above were capable of invention, why not the author of the Voynich? The author could possibly have been more inventive than any of those, I don’t think that should necessarily be discounted, though I wouldn’t go that far myself yet.
Mark: there is no real question (in my opinion) that the creator of the Voynich Manuscript was indeed “capable of invention”, as you put it. Where this falls short is where people try to (unreasonably) extend it to suggest that the sole explanation for everything we can’t directly and easily understand is that same all-pervasive inventiveness – this is just lazy, defeatist, empty-minded thinking. Rather, the default case should always be the assumption that whatever behaviour we are looking at has some kind of precedent, unless we have some very strong specific reason otherwise.
As an aside, the #1 reason we can rule out normal classes of polyalphabetic ciphers (e.g. Alberti, Vigenere) is simply that the text we see is strongly structured rather than unstructured (the main point of polyalpha is to eliminate structure).
Mark, Nick, while I can think of some options, I certainly have no satisfactory answer. Everything I can come up with immediately triggers a list of issues.
Just to give an example.
A verbose cipher (much along the lines suggested by Nick) should increase word length. This problem could also be resolved by assuming that the Voynich words are not complete words, or: one cannot trust the word spaces.
This then raises the problem of the labels. It would not make sense (at least at first sight) that these are only syllables or even less.
This could be resolved by the hypothesis that the labels are just copies of items in the text, and refer to that text.
This then raises the problem that not all label words can be found back in the text. (Many can).
This could be resolved in several ways, e.g.:
– we are missing about half the manuscript
– a degree of freedom in the encoding, such that (e.g.) ‘cheol’ and ‘chedy’ are actually the same thing, and we just don’t know it.
One can entertain such ‘trains of thought’ on many different aspects of the MS. They tend not to lead very far, as long as one stays objective, and doesn’t get carried away in fantasies.
Rene: at the same time, you’d surely be quick to admit that labelese presents quite a few differences from the rest of the text, and so there is good reason to suspect that something quite different may be going on there – it’s certainly not sustained dense text in the style of Q13 or Q20, let’s say. And if that is the case, I think we should be wary of drawing inferences (as you just did) about the rest of the text from labelese. 😐
Nick: On the face of it, one would expect labelese to be different. The default expectation would be to consider labels to be nouns or possibly adjectives. So we wouldn’t expect labels to be pronouns. It seems harder to imagine that labels would be verbs, adverbs, prepositions or conjunctions. It is possible that the odd label is not a noun or adjective, but it is seems unlikely to be common.
(This, of course, assumes that labels are independent entities i.e. are not to be strung together somehow to form sentences or other interrelationships. This assumes that labels are not references “Diagram 1” etc.)
One’s natural expectation could be that labels would tend to be rarer specific words than one might find elsewhere in the text.
So I think one’s default position would be to expect that labelese presents quite a few differences from the rest of the text. I suppose one could compare the labels to the rest of the text for any book or manuscript to get a sense of the kind of differences to be expected.
What concerns ‘labelese’, this is different in the sense that this seems to present individual words rather than running text.
However, these individual words are for the most part standard Voynich words.
They just tend to exclude certain families of Voynich words (no chol, no daiin and hardly any starting with qo).
Any explanation of the Voynich MS text has to accommodate this. In this case I am sure that you would agree.
I would have thought that single word labels would tend on average to be longer than normal words as pronouns and conjunctions tend to be shorter than other words on average in most languages. I don’t know how much the statistical properties of labels differ from those sentence text in other texts either contemporarily or now. My guess is that labels might be more likely to be unusual or foreign words than one finds in normal text and so more likely to have a different frequency of certain letters than in normal text.
Whether these differences tally with the differences we see with labelese and Voynichese as a whole is another question.
Rene: I wouldn’t advise presuming even that zodiac labelese is consistent throughout the zodiac pages. The labelese text you see on the 15+15 dark/light pages (e.g. Aries and Taurus) is quite different from what we see on the other pages (such as Sagittarius). The lack or paucity of chol / daiin / qo- is an intriguing start, but I think we have so far barely scratched the surface of labelese, analytically.
Rene: When say “However, these individual words are for the most part standard Voynich words.”
And that is one example of the crux of the problem for me, because I don’t think they ought to be “standard words”, though from my recollection there are quite a few non-standard words. So why do we have “standard words” amongst labels? If we were to look at labels on other manuscripts of the period or now would one find many standard words?
Mark: all the while you apply “ought” to Voynichese behaviour (whether labels or elsewhere), you give every impression of back-projecting your normal modern ideas about how normal modern languages work in normal modern contexts to something that is neither normal nor modern. I can’t help but feel that you’re really missing the boat here, sorry. 🙁
If you could give yourself a bit of a slap every time you find yourself about to type the word “ought”, there’s a reasonable chance you’ll start to approach Voynichese sensibly. 😉
Nick: I used the word “ought” recognising that this use of this word could be viewed as being laden with lots of assumptions.
I think the nature of label text versus other text is a subject that should be independent of the underlying language. Regardless of how Voynichese glyphs map to real words the question of how real words are distributed amongst labels and other text seems language independent.
Now there are assumptions such as there not being some kind of magic changing key that transforms words that are spelled the same into different words, that spaces in sentence text are spaces and that labels aren’t references.
I suppose one could see how sentence text and labels are related from similar texts of the same period. This would remove the question of it being a modern bias.
One can ask whether it is common or rare for texts of the kind and period of the Voynich to have the same labels for different and unrelated items as we see in the Voynich, which could be addressed by looking at surviving manuscripts. If this is very unusual as I suspect that we could potentially infer that some strange is going on when it comes to labels in the Voynich.
Mark: you’re still making plenty of assumptions about labels and language (and how they interact, how they ought to work, etc), even though we’re still a long way from having what I would consider a basic workable knowledge about labelese. That’s not your fault, it’s our (collective) fault: and it’s a symptom of a wider set of lacks that I’ve been thinking about over this Christmas period.
The core question I keep coming back to is: what should constitute a ‘problematique’ (i.e. a well-defined scope plus a set of distinctive behaviours to explain) for each difficult Voynichese area? Not just A vs B (and all points between) and labelese, but also line-initial words, line-final words, span-final words? Given each particular scope, what are the right questions to be asked for each of these areas? And – given all the difficulties to do with parsing – how can we hope to make progress with these questions if we still aren’t even remotely sure how to parse the words?
This is not a simple question at all.
Evaluate VM characters.
History: In cryptology, different characters were supposed to simulate an alien language. Here are some examples.
Deception of language and deception in the writing ?
You have to ask yourself, why does VM text look so monotonous ? Is there a possibility of deception here as well ? I have already listed some of the possibilities. But how does it look like here ?
Are all 3 characters the same or are they all different ? Or are there only 2 ?
Is number 1 and 3 the same, just to hide the height ? In any case, he starts with the spring differently. ( middle and top ). In this case the deception would be given.
https://www.facebook.com/photo.php?fbid=2449103088645664&set=gm.2499194766856976&type=3&theater&ifg=1
An aside. I have just learnt that in their script the Mayans had many different homophones for a letter/sound. The purpose of this appears not to conceal the contents of their writing, but rather artistic. I very much doubt this artistic purpose is the case with the Voynich. As a note I was surprised to come to the idea that homophones were used in the Voynich before I learnt that homophones were features of medieval ciphers.
I was having a thought and wondering if we could find something like the following:
$%&£ -> horse
But also
$%&£ -> layer
In addition
+&#!# -> horse
And similarly
@*£%# -> layer
In other words the same word can be spelled more than 1 way with completely different spellings. Also more than 1 completely different word can have the same spelling.
If true then we would have a high level of homophonicity going on. It seems that in this instance it would be necessary to identify the underlying homophonic components i.e. are individual letters homophones? are pairs or sequences of letters homophone blocks? or is there something more complicated going on? With possibly so much phonics going on then it begs the question about problems of ambiguity that could occur for the author.
This is just a side thought and not based on in depth thought or careful study,though it could be in the future. I have wondered about the presence of nulls and homophones for some time, core features of the ciphers of that time.
I been thinking about the question of whether Voynichese is some kind of cipher. This is something I have quite a strong conviction of. But it seems to me that this is probably a minority position in Voynich research; perhaps looking at this blog has lead me subconsciously to view this as the standard position.
The most popular idea seems to be that we are dealing with a known or unknown language in an unknown script. However this seems to have huge problems with it as other have observed assuming it isn’t some extremely weird language from some place far from Europe. In fact I don’t quite understand why this idea persists. Glyph ordering, repeated words, first glyph paragraph gallows and this just scratches the surface.
I have argued about the question of latin abbreviations ad nauseam, so I won’t touch on that here.
Then there is the subject of an artificial language. I can see that in theory this is possible as with the notion of an artificial language you really have free rein to explain away any weirdness as a function of the language. So you could attribute a meaning and interpretation to repeated words, first paragraph glyphs and any oddity one finds in the Voynich or that is not in the Voynich, but can be imagined. In fact I suppose cipher features like nulls or homophones or glossary could all be features of an artificial language. In fact this idea could be a catch all. However lets take the question of labels, how do we address label words that are the same for quite different drawings? How do we address labels the occur only once in the whole manuscript? More importantly an artificial language is from what I know quite out of keeping with the early 15th century. Is there some kind of contemporary precedent for this?
It is nonsense; I have discussed this elsewhere.
Now I daresay I have missed other options.
I think we can all come at this question with a bias. Those that are linguists or are not familiar with the alternatives possibilities a natural language in an unknown script is the default assumption. Some could argue that there are reasons why I am biased towards it being a cipher e.g. my rosettes analysis.
I have been a little worried about the question of carelessness by the author. It has been pointed out that the botanical/pharma small plant drawings don’t exactly correspond to the large plant drawings; my explanation is that this is just a function of imprecision or errors in reproduction of drawings.
Evidence of carelessness or laziness on the part of the author impacts on their seriousness and professionalism. This does raise questions then if this is also the case for Voynichese i.e. if it is some unprofessional disorganised construction. However it still remains that the author is unlikely to have written 200 pages of text that he/she could not read back. So even if it is a loose or sloppy cipher, it must still be a functional one, unless as others claim it is all nonsense. I guess I am saying that we should be aware that the cipher may not be neat, but could be disorganised, though just about workable.
Mark: fine, but be aware that when I read the word “explanation”, the phrase “bullshit narrative” flashes up in my mental subtitles. 😐
Nick: I have become well aware of your perspective, so I won’t be offended by your scepticism. I guess brainstorming these kinds of issues is something I think valuable and may be of interest to some viewing my comments.
Mark Knowles: “I have been a little worried about the question of carelessness by the author. It has been pointed out that the botanical/pharma small plant drawings don’t exactly correspond to the large plant drawings; my explanation is that this is just a function of imprecision or errors in reproduction of drawings.”
Nothing about the VMS seems careless to me except the sloppiness of the less-careful painter (who is probably a different person from whoever did the drawings and is certainly a different person from the more-careful painter).
Based on the way they are organized, I think the big-plant sections and small-plant sections may serve different purposes, in which case there doesn’t have to be perfect overlap.
JKP: No, I agree there doesn’t have to be a perfect overlap, but I think there is a large overlap, however there is a spectrum of similarity of drawing and some small plants, whilst they are I think the same plant as the corresponding large plant, have a greater divergence than others from their associated large plant drawing.
Now it is possible that colouring is a part of this, which is hard to assess when I can’t see what is under the colouring.
I hope you are right that the author was not careless as if he/she was this could create a whole world of work deciphering the text.
JKP: My concern is that we could have inconsistencies in the enciphering method. However I am inclined to reject that as presumably then the author would have as much difficulty in deciphering what they wrote and you don’t write 200 pages and then realise you can’t read what you have written.
One line that has come back to me again is this idea that an early 15th century cipher would have been cracked already. I think, as perhaps this website demonstrates it is easy to produce a very hard to crack cipher. I think producing a very difficult to crack cipher becomes much much easier if one uses a lot of filler text. Filler or null text is a very simple concept, but at the same time an incredibly effective one. In fact the larger the proportion of filler text to real text one has the likely hard it is to decipher to the point where if one has only 3 words amongst 1000 words that are not filler then it becomes virtually impossible to decipher.
However I wonder to what extent other simple techniques can be used to render something close to undecipherable. Certainly the larger the lookup table one has such as the number of discrete complex homophones the harder it gets.
The more filler text and the larger the lookup table the more time consuming it is for the reader to process the text and for the writer to generate the text.
Still, I would hazard, not that I have tried this, that I could produce a very difficult to decipher code based around simple techniques broadly similar to those used in the early 15th century. So I think the argument that I hear that it could not be a cipher as we could crack it seems weak to me as there does not appear to be a big leap from a cipher of that period to uncrackable one except possibly the greater level of effort to implement it.
Mark: it really doesn’t take a very large step sideways from ciphers we know to get to a cipher we don’t recognize at all.
If the resulting cipher is objectively easy to crack but is protected only by lack of knowledge of the system, this falls into the modern crypto category of “security by obscurity”.
This scenario seems overwhelmingly likely to me for the Beale Ciphers and even for the Rohonc Codex. It’s unlikely to be far from the truth for the Voynich, too.
Nick: I quite agree. I like the phrase “security by obscurity”.
Nick: I was thinking about trying to chase up scans of some more enciphered letters this summer, though it may not be easy without being on the ground.
In addition, I was also thinking about how I might reconstruct as best as possible, the pre-1447/Filippo Maria Visconti cipher ledger.
There is a cipher key in Meister labelled “In Milano” from Modena, now it is possible that that is the cipher key used by Modena to communicate directly with Milan.
As I noticed before there is a cipher key in the Tranchedino which is the same as a cipher key in the Codex Urbinate albeit in very different handwriting.
So I speculate that for Modena the Milanese cipher ledger would have the same key, albeit with different handwriting, as we find in the Meister “in Milano” cipher key from Modena.
Likewise I would hope to find similar cipher keys amongst the archives of Mantua, Florence and other city states. Now these cipher keys might be headed “Milano” or “Visconti” or “Filippo Maria” or some other heading associated with the Duchy of Milan.
It seems clear, from what I can tell, that Meister chose a selection of cipher keys from each city state for his book and so therefore there are a number of cipher keys not included there. One might hope to find amongst these other cipher keys the Milanese cipher keys that I am interested in.
If I remember correctly you pointed out that the cipher keys in the Milanese ledger could be for communication directly with their ambassador in the different city states and so they would not appear in their archives for those city states.
In the case of Urbino it appears the Milan cipher key was used for direct communication not via an ambassador. Now Urbino was small and so maybe didn’t merit having an ambassador from Milan. Similarly Modena was small and again there might not of been a Milanese ambassador. However I wonder if in some other city states such as those in Meister they may not also have had a Milanese ambassador pre-1447. If so then this does make it possible that part of the Milanese cipher ledger from the time of the Visconti is recoverable from the archives of these other city states.
Meister lists where in each archive he obtained a given cipher key that he included, which, provided the archives have not completely been rearranged in the last 100 years, should indicate where the other cipher keys from that archive are likely to be found.
Now of course getting hold of scans of these may be a hassle, but it does sound like something I might try to do, though the enciphered letters are the priority.
Nick: I should explain that my motivation for reconstructing the pre-1447 Milanese cipher ledger is to see to what extent there are common symbols to the Voynich.
Nick: Having looked into it, it does appear that some cipher keys are for ambassadors to certain city states. However it also appears that certain cipher keys are for given states. As an example if we look it the Tranchedino we can see a cipher key headed “Marquis of Mantua” which like the Urbino cipher key would seem to imply direct communication with Mantua and therefore a good chance this cipher key is held in the Mantua archive.(It is conceivable that sometimes city states had one cipher key for the ambassador and another for direct communication with the other state. Maybe certain situations merited one and one the other e.g. when the ambassador is away.)
It is strange that when city states communicate directly with one another using the same cipher key they would be sharing their techniques and the symbols they tend to use.
Nick: Just been rereading Meister “Die Anfange..” and thinking. I would reassert that the sophistication of Milanese diplomatic cipher keys did not increase from 1447 to near the end of the 15th century. Up until 1447 the Visconti ruled Milan. From 1447 to 1450 the Ambrosian Republic governed Milan. From 1450 overwards the Sforza governed Milan.
If we look at the first 2 examples for Milan in Meister, the first is dated 1447 and the second dated 1483. Between these two there appears to be no increase in complexity and sophisticated.
I have seen cipher keys from pre-1450 that were generated in the 19th century from deciphering letters from that period, so there is no possibility that the cipher keys were added to at a later date and certainly no reason to believe the letters were modified some time after they were written.
I find it difficult not to conclude that the sophistication of Milanese diplomatic ciphers did not increase in Milan from the era of Filippo Maria Visconti through the era of Francesco Sforza. And so Cicco Simonetta did not preside over an era of increased sophistication of diplomatic ciphers, but rather in Milan that era of significant increase in diplomatic ciphers, almost certainly took place during first half of the 15th century and most probably under the reign of Filippo Maria Visconti. So I would argue that the 4 wars in Lombardy were the backdrop to the development of more complex cipher keys in Milan rather the peaceful era after the pact of Lodi. In fact the development of more complex diplomatic cipher keys would seem possibly to coincide with the resumption of war at the end of the 15th century. This is consistent with what we know from our own time that advances in cryptography tend to coincide more with wartime than peacetime. (Note the work of Alberti and others were largely separate at that time from the ciphers used in diplomacy, so I am not including them in these statements.)
Nick: In Meister “Die Anfange…” he writes that the cipher key used by Modena for communication with Milan is much more sophisticated than the other cipher keys that he saw in the Modena archive from the same time, which indicates that the Milanese cipher secretaries were quite advanced around 1430.
In the Mantua archive according to Meister there is a cipher key dated between 1401 and 1416 and headed “Dux Mediolani”. This same key almost certainly had a twin residing in the Milanese cipher ledger of the time and headed something like “Marquess of Mantua” or “Gonzaga”. There is an equivalent much later cipher key in the Tranchedino for communication with Mantua.
So obtaining a scan of this could provide insight into the state of Milanese cryptography of the time. This could also be a step towards reassembling the Milanese cipher ledger through the Visconti era.
There ought to be a cipher key headed “Dux of Mediolani” or “Filippo Visconti” or something of the kind in the cipher ledger of Paolo Guinigi, Lord of Lucca.
So far that means I can think of the following Visconti cipher keys:
Modena – “Die Anfange…” by Meister – “In Milano”
Mantua – Number 27 – “Dux Mediolani”
Lucca – Paolo Guinigi Cipher Ledger
That leaves possible cipher keys for direct communication with Milan for Venice, Florence, Sienna. As far as other archives go I have ideas, but they are not yet fully formed.
I forgot to mention, of course, the 1447 cipher key generated from a letter from Filippo Maria Visconti.
I have some evidence that says that Martino Ghisolfi after working for the Gonzaga then went to work for the Visconti. The extent to which there was transfer of knowledge or whether he just went join a larger operation is unclear.
I have obtained scans of a few cipher keys from the Gonzaga cipher ledger. One is headed with the name of Gian Galeazzo Visconti’s Chancellor. This is very likely dated to before 1403 when the disastrous reign of Giovanni Maria Visconti began.
I would be surprised if an identical cipher key was not in the Visconti cipher ledger.
Recently I have seen cipher symbols one which is a clockwise swastika and also a counterclockwise swastika; the Nazis for ever ruined that symbol.
Also I have recently seen another cipher symbol that arguably looks almost like a part of the Male anatomy as well as the symbol for used by the “Zodiac”, as in the serial killer; though I think I have seen that used before in other ciphers.
Just came across the following:
https://www.lanuovapadania.it/cultura/ecco-il-codice-segreto-degli-sforza-il-regalo-domenicale-della-biblioteca-di-cremona-su-fb/amp/
Someone might find it interesting.
I am not sure yet if there is anything to interest me there.
Looking over the above blog post it seems to me that it makes the central point that I am convinced of based on the evidence, namely that we see a big transition in the complexity of diplomatic ciphers from 1401 to 1450, the like of which we don’t see before or for some time afterwards. Looking at the 3 cipher keys for Milan in Meister “Die Anfange…” we see that from the first cipher from 1448 to the last cipher from 1530 there is relatively little change. All the evidence seems to indicate that the Milanese ciphers from around 1400 were much the same as that from Mantua that is shown above. I think the question has to be why did these changes start over this period and end when they did. My go to reason has been the start and end of military conflict, in particular the Wars in Lombardy and the Pact of Lodi. In modern times cipher development has been greatly influenced by military considerations, particularly the possibility of war. There is an argument that by 1450 there was really no scope for further complexity within substitution ciphers without rendering them so complex to render them unmanageable; this being too much work to draw up a new cipher key, too much work to encode a message and too much work to decipher a message, in addition they may have been the opinion that the ciphers were now so advanced that they were almost uncrackable.
On the arms race point in the blog.
I think the goal was simple to produce a cipher key that the enemy or rival or interested party couldn’t crack if they intercepted an enciphered letter. But I am sure that similarly the best was done to crack the intercepted letters of others.
Over this period there were significant examples of intercepted letters and in some instances they were deciphered resulting in significant problems for the city state from which the letter emanated. I am sure these kinds of mishaps were strong incentives to make significant enhancements to ciphers by city states. Likewise I am sure that those breaking ciphers were improving and refining their techniques.
So is that an arms race? I don’t know.
One point that fits with the post above is that one can see cipher keys with a design which is greatly hampered by choosing poor substitution symbols. So for example each item of nomenclature starts with a “b” shape. Another example is having homophone symbols for a given letter that look very similar to each other relative to other letters. Obviously a random distribution of arbitrary symbols throughout the cipher key is the most secure option, however too often this is not what is done.
Mark: the Pact of Lodi brought about the much wider (and day to day) use of ciphers, it’s true. But I’ve never really bought into the idea that it necessarily was also behind the complexification of ciphers.
In fact, as you point out here, the rapid evolution of ciphers seems to have preceded 1450 (and perhaps even 1425, truth be told).
Mark: randomisation was something that was not really understood until centuries later, so any argument about fifteenth century ciphers based around the presence or absence of randomness is likely to be somewhat suspect. Just saying.
Nick: I think there is reason to believe it probably preceded 1425. Given the implosion of the Duchy of Milan under Giovanni Maria Visconti and the Duchy returning to safer hands in 1412 when Filippo Maria Visconti and important figures in Gian Galeazzo Visconti’s administration returned to power there could have been an effort to improve the security of their ciphers as Milan regained control over its former territories. I have doubts that any such advances in cipher techniques occurred under Giovanni Maria.
Nick: Whilst I think it is certainly true that in the 15th century they had no acquaintance with the ideas of modern probability I would suspect some designers of cipher keys had an instinct as to what would make a cipher more or less breakable born out of intuition or more likely practice breaking ciphers themselves. We do see some cipher keys where the symbols appear to have been chosen in a much more “randomised” way than in the case of others. Now this could have been due a conscious decision of the designer to construct the key that way or merely “chance” that they produced such a less breakable cipher.
If it is your job to design a cipher key to write letters that if intercepted would not be deciphered and you live in the 15th century then your back is against the wall. If you screw up the consequences could be severe not just for the state you serve, but also for yourself; it doesn’t just mean signing on and looking for other employment. So if you are tasked with improving the current cipher design used by your chancellery then what do you do? Well, it has to be a design that is easily reapplied for each different correspondent, easily implemented when writing a letter, easily usable for reading a letter and sufficiently hard for someone without the cipher key to break. One would imagine that there might have been some experimentation with different designs that no longer survived before the standard template that we see in the 2nd half of the 15th century was arrived at. If such different designs were used the chance of there being any surviving examples of them in the sparce diplomatic cipher record may be slim. It is probably the case that a “Voynich” style of cipher would have been completely impractical for diplomatic use and similarly a standard diplomatic cipher even easier to break with a 240 page manuscript than with a page long letter. Nevertheless I maintain that despite the statistical differences that we see they could very well have a common root. However it remains worth keeping an eye for such “dead end” ciphers as they may just have an echo of something we see in the Voynich. It is a lot to ask for given how hard it is to find any cipher keys for say the late 1430s. We don’t know when the final template cipher key design that we see from 1447 was arrived at, it could be as early as the late 1430s.
So far I have obtained cipher scans from the following state archives:
Milan(though letters needed as well)
Modena
Mantua
Florence
Siena
Lucca
I waiting on scans from:
Pisa Archive(Hopefully should not be too difficult to find in the archive)
Vatican Archive (Should get soon, though it is likely there is more of interest in the Vatican archives somewhere.)
Barcelona Cathedral Archive (May be hard for them to locate as the details I have may not be specific enough.)
Genoa Diocesan Archive (1 year wait for the relocation and opening of the archive)
I need to investigate the following:
Asking Professor Senatore if he has any ideas where else to look and whether there is anything of interest in the Naples archive
Working out where the MB enciphered letters are in the Milan archive and then request scans of them.
I haven’t investigated the Bologna State Archives to see if there is anything beyond what Meister includes in his book on Papal ciphers.
I haven’t conducted a serious investigation into Italian (or other) ciphers from the early 15th housed in archives outside of Italy; so far I only know the Ekaterina Domina enciphered letter in the Moscow archive and the Tranchedino cipher ledger in Vienna.
Need to develop a better strategy possibly with the assistance of *Nick* for locating other ciphers.
I haven’t spotted any 4o symbols(EVA-qo) in the Lucca ledger, but I have spotted some 4p symbols(EVA-t).
I have been thinking a little about the origins of homophone substitution in diplomatic ciphers.
Homophonic substitution is not present to any extent in the cipher ledger of Gabriel de Lavinde de Parma for Antipope Clement VII which is dates to 1379. It can be seen to the cipher key of 1395 in the Mantua archive. On the basis of this I think the obvious conclusion would be that homophone substitution was an innovation of Martino Ghisolfi, Francesco Gonzaga’s cipher secretary. However it is clear that we see homophone substitution in the Milanese ciphers in Sercambi’s book and they have been dated to around 1397, I believe. So can we really be confident that homophone substitution originated in Mantua and not Milan. In this regard it could be worthwhile if someone were to decipher the 1395 enciphered letter in the Modena archive to see if it employs this technique. It is noteworthy that the early Milanese ciphers seemed to have not used null symbols and don’t seem to be particularly keen on having a glossary, but rather relied on other techniques to achieve encryption such as homophone substitution. So it is unclear who introduced the idea of homophone substitution, but it does appears that it gained popularity as a technique very quickly.
Not my area, but I want to know, if that’s possible, whether the Bank of St. George , founded 1404 in Genoa, used ciphers and if so how they line up with Milanese ciphers to c.1450.
An article from Cambridge core looks promising, but before parting, I thought I’d ask if anyone more knowledgeable on the subject had read it, if and if so if they think it’s worth a cipher-history novice getting.
Vincent Ilardi, ‘Fifteenth-Century Diplomatic Documents in Western European Archives and Libraries (1450-1494)’ first published in the journal
Studies in the Renaissance , Volume 9 , 1962 , pp. 64 – 112.
published online via Cambridge Core in 2017.
I see today that M.R. Knowles (February 17, 2018 at 1:33) is ahead of me in tracking Giorgio Costamagna and the Bank of St.George.
Milan…Genoa… accounting … ciphers. Let’s hope we meet somewhere more pleasant than Philippi. 🙂
Diane: I would advice you to look for:
“Un cifrario contabile del Trecento” by Frangioni, Luciana . (1987) – In: Nuova rivista storica Vol. 71 (1987) pp. 623-640
This, as the name says, is an accounting cipher; I have not seen it. If you find it can you send me a copy as I think the guys at the DECODE database would be interested. I know of some early 15th Genoese ciphers.
Diane: Track down that paper about the accounting cipher and I will share with you what I know of the ilardi microfilms. I haven’t done so myself as it is not really my area of interest. However I am sure if I pass it on to Beata Megyesi and the people at the DECODE database they will be interested.
Nick: Did I share with you the Ilardi microfilm listing? It is possible that you might see something in that that I have missed.
@ Kmowles. You can see that you have a lot of trouble with the manuscript 408. No boy from the DECODE manuscript can ever decipher. I hope this is clear to you.
Josef: My name is not spelled “Kmowles”. It is spelled “Knowles”.
Mark – thanks for the references. I’m flat out with commissions (non-Voynich) at present but maybe in a month or two I’ll get to them. Come to thik of it the University libraries may be open by then. Our librarians are wonderful.
Most grateful for the references and for your taking time to respond.
btw – Do you subscribe to Cryptologia?
Given that I have now finally and for the first time in public stated explicitly on Voynich Ninja that I associate the authorship of the Voynich manuscript with the Barbavara family I thought I ought to state it here too. I haven’t provided many specifics behind this theory there for reasons I have explained. It is probably better to read my statement on Ninja than here.
Comment #900 for your post!
I find it said everywhere that polyalphabetic ciphers were invented in Europe by Leon Battista Alberti.
On the other hand I’ve found examples of mixed-alphabet writings evidently made to render text obscure/occult from long… long… before Alberti’s time, and even within Europe, by thirteenth century, Roger Bacon cites the example of someone who lived during the 7th or 8thC when he says:
[quote]
[4] Fourthly, … by the mixture of several sorts of Letters, for so the Ethnick Astronomer hid his knowledge, writing it in Hebrew, Greek and Latine Letters altogether.
[unquote]
I sort of assume that there’s some recognised distinction between these which eludes me and that “polyalphabetic ciphers” is used in some stricter sense. but would be grateful for clarification. In what sense is Alberti credited with inventing polyalphabetic text-as-cipher?
M R Knowles,
In case you haven’t yet got hold of a copy of the article you mentioned last year – I apologise for having quite forgotten to follow it up until now – I see that that Journal is published by the Political Science Dept at La Sapienza.
The volume for 1987 doesn’t appear to have been digitised, but if you write to the editors, they may be kind enough to email you a copy.
The website for the journal’s ‘Vintage archive’ is:
https://www.nuovarivistastorica.it/category/archivio-annate/
Editorial contact:
https://www.nuovarivistastorica.it/contatti/
If they’re not helpful, and if you’re a current member of voynich.ninja, Marco Ponzi or some other member may be in a position to get a copy for you.
“Un cifrario contabile del Trecento” by Frangioni, Luciana . (1987) – In: Nuova rivista storica Vol. 71 (1987) pp. 623-640.
My uni library doesn’t seem to keep it; doesn’t keep many hard-copy journals now, sad to say.
Diane: grabbing glyphs from different types of alphabet doesn’t make it polyalphabetic, it just makes it a visual mess. 🤣
Alberti’s big brain crypto flex was to put one alphabet on a rotor and another on a stator, and to rotate the rotor alphabet every few words, which is quite different. Never caught on for centuries, though. 🙁
Nick
🙂
Strange, isn’t it, that never caught on when its just a variation on defining locus by adjusting the inner- and outer-rings of various instruments that were common enough in his day.
Diane
It isn’t a priority at the moment, given my research, to track down this article. If it was I would either contact the publisher and ask to scan and email me a photoreproduction or check out which libraries have a copy of the journal and maybe contact one of them to get a photoreproduction. I think it very probable that the Bodleian Library in Oxford has a copy of the Journal, so given I live in Oxford that might be the easier angle to pursue.
I would like to see a copy of the article as, for general interest, I am happy to see all 14th and 15th century ciphers that survive. Although it my own thinking I cannot see a connection with the Voynich.
I believe you expressed interest in ciphers used in banking or accounting, which is why I referred you to that particular cipher.
If you are still interested in it.
If you can track it down then it will save me doing it at some stage. If there is some fee for the photoreproduction then I would be happy to reimburse you, although it should be at most a small amount.
You may want to see if you can persuade Professor Beata Megyesi, who administers the DECRYPT cipher database to see if she can track it down. If you wish you can tell her that I referred her to you.
To be honest I wouldn’t think it would be much work to track it down.
Diane
There are two email addresses that I can see for the Journal:
[email protected]
[email protected]
Just send them an email requesting the cipher and I would have thought they could email you a scan back, probably for free, without too much work on their part.
Mark –
Crossed wires. I thought you wanted the article for your own research; you thought I did. 🙂
Diane
I am just going by what you wrote earlier in the comments to this page. I would be interested in that cipher as part of my general interest in 14th and 15th century ciphers. However I am not really interested in it as part of Voynich research, so it is not a priority of mine.