+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Theories & Solutions (https://www.voynich.ninja/forum-58.html)
+--- Thread: Voynichese is a numeric cipher? (/thread-5798.html)
RE: Voynichese is a numeric cipher? - ololololo - 21-06-2026
(21-06-2026, 08:34 AM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.That's true. If we accept my assumption that the balneological section is an anatomy, then there can be no "fire" or "earth" there, because this is not an alchemy textbook. It's most likely an allegory for humors, as you have noticed. However, I believe that it is more important to identify the meaning of the word than to accurately identify what the author drew.(20-06-2026, 11:07 PM)ololololo Wrote: You are not allowed to view links. Register or Login to view.Such liberties are possible, but they do not take place, and when you write a book about medicine, you rather sign the fire as fire, because you do not need to play the reader.
I can't agree, if this is a book about medicine and not a kindergarten primer, I'd say it makes little sense to draw fire and label it as "fire". Certainly the reader knows what fire looks like.
In a medical book of the time one example could be a text talking about how humors relate to elements, in which case labelling elements for their corresponding humors would be informative. Then fire would be labelled something like "bile", air as "blood", water as "phlegm", etc.
In reality I don't believe we can expect the fire image to be literally labelled as "fire". It is possible, but it's far far from certain.
RE: Voynichese is a numeric cipher? - oshfdk - 21-06-2026
(21-06-2026, 09:10 AM)ololololo Wrote: You are not allowed to view links. Register or Login to view.That's true. If we accept my assumption that the balneological section is an anatomy, then there can be no "fire" or "earth" there, because this is not an alchemy textbook. It's most likely an allegory for humors, as you have noticed. However, I believe that it is more important to identify the meaning of the word than to accurately identify what the author drew.
It is, I'm just cautioning against treating image cribs naively. If it shows a can of Coke, we can't simply assume that the label says 'a can of Coke'.
RE: Voynichese is a numeric cipher? - ololololo - 21-06-2026
I will share my speculations about what the cipher of the Roman numerals in the manuscript looked like. This is not a complete model, but rather a set of assumptions.
1). I believe that the cipher can be divided into three types: the original form and the Currier's languages, which may be the result of a change or "adjustment" of the cipher to the author's needs or scribes. I consider Language A to be the closest to the "original form," because it was the first language used in the manuscript.
Original form is a simplified nomenclature cipher, where Latin letters are encrypted with Roman numerals, which are in turn encrypted with an unknown alphabet. At the very least, this cipher had two tables, one with letters and the other with abbreviations. Unlike other nomenclatures, where unique words or frequent morphemes were encrypted with a distinct combination of numbers, the manuscript cipher used a single alphabet for both tables and did not have any unique words. I don't rule out the possibility that some letters could be encrypted using two unknown symbols, as this was a common and everyday practice at the time (for example, a symbol encrypted as ee could be encrypted as ch). There's another option - there were several tables for the letters, and in one, for example, they were numbered in ascending order (A = 1... Z = 26), while in another, they were numbered in descending order (A = 26... Z = 1).
However, two tables remain the minimum requirement.
The secondary encryption process (i.e., manipulation of the already encrypted text) involved creating nulls and splitting the word. Nulls should be symbols of the same alphabet, but placed in a way that allows them to be recognized and discarded. A rough example: YMOR YNIN YG (the word morning is divided into three parts, with an added null "y"). In terms of numbers, they could be aligned, and the principle was that the word C-X-VV-V-CI became CCXVVVI. Will you say that it will be impossible to decipher it in this form? I can tell you that this can be done by focusing on the word itself. In the unencrypted form, there was a combination of CI, which is considered a single letter, and when you tried to decrypt CCXVVVI, you realize that the unit doesn't fit, you'll immediately understand that there's a letter of the form x-I in the word. This will lead you to CI. Next, you'll realize that VVV doesn't work, and you'll immediately understand that there are either V and VV or three Vs in a row (you'll soon discover that this is also incorrect). You will also understand that C is the first letter in the word, and that the combination XV is not suitable... After a few iterations, you'll eventually guess the word, and all you have to do is check the table to confirm that you've correctly guessed the C-X-VV-V-CI. While this is possible, it is also expensive, so it is possible that the author provided hints for the syllables. For example, he divided the word into CX VV CVI, preserving the structure of the syllables to make them easier to recognize.
Currier A is a classic version of this cipher. In fact, it is not certain that it was the same as the original form (for example, the scribe may have encrypted the prefix not with a separate symbol from the second table, but with letters from the first table). The supposed feature of Currier A (I would call it the scribe's approach) is the infrequent alternation of tables. This means that when writing, the author tried to avoid confusion during decoding by using specific symbols for a single value rather than two. This means that the "value" of the letter will be stable, and in most cases it will produce the same result. But it could also be more complicated, using aspects that I didn't consider.
The difficulty of Currier A is compensated by the fact that, for example, you don't have to read the entire botanical section. All you need to do is open a specific page, decipher it, close the book, and forget about it. To see if the page has the plant you need, you could decipher the first word (which could be the entire name of the plant, or a part of it) and get your bearings quickly. Briefly, the complexity of the cipher is compensated by its practicality. The practicality could also be expressed in the fact that Currier A encrypted small amounts of text, so it wouldn't take too long to work on the encryption and decryption.
Currier B, it will be derived more from Currier A than from the Original form, being a simplified and adapted form of the former. The peculiarity of Currier B is that it contains large amounts of text. If you try to encrypt a large amount of text at once using the methods described above, you will probably not want to continue working halfway through.
Therefore, it is likely that the scribe deliberately simplified the encryption method, which resulted in a more monotonous text than Currier A.
How could these simplifications be expressed? I can't say for sure, but I can assume that
a). The word could have been encrypted in its entirety or lazily (i.e., using a pure pattern)
b). The author may have made changes to the tables, one of which I believe was the addition of the Quincunx symbol (EVA x). Maybe the scribe didn't want to write 150 in Roman numerals for a long time, so he remembered that the quincunx is equal to 150 degrees and wrote it down. It is also possible that he assigned higher values to some characters than necessary (maybe this characters is ch/sh?).
In addition, the reason for the monotony of the text may have been that the author often alternated between tables, writing qo as a prefix in some places and as a letter in others.
In the case of Currier B, we need to decode it completely in order to obtain information.
So, here's my opinion on all of this: If we want to study the structure of the cipher, we need to focus on Currier A. If we want to study the flexibility of the cipher, we need to focus on Currier B.
2). What will the output text look like? In any case, we will have difficulties, because we will have to solve the anagrams of words that inevitably arise when arranging numbers. In the case of Currier B, this will be really problematic.
3). Which symbols are most commonly used as abbreviations? It may be f and p (they do not form stable combinations, unlike their counterparts k and t, and are often used at the beginning of a paragraph/page), and qo (in this form. If it is a prefix, the author could have given it a large value in advance to make it always stand at the beginning of the word).
4). The reasons for repetitions? There can be several of them: splitting a word with the addition of garbage elements, using the same word in different meanings (for example, the pair qokedy qokedy can differ in the meaning of the q symbol) or, as already mentioned, clues for decryption.
5). As for the zodiac section and astronomy in general, it may turn out that nothing is encrypted there, and the numbers themselves are written directly. We'll have to work on that, too.
And again, regarding the complexity: yes, since it is a nomenclature, the cipher will be complex, which is why we should expect it to be simplified during the writing process becoming readable.
RE: Voynichese is a numeric cipher? - ololololo - 23-06-2026
The second attempt to determine the weights of the symbols.
I measured it using the same methods, i.e. according to statistics, "is the symbol x more often placed before or after the symbol y?". But this time, more characters have already been taken.
If we divide the symbols into groups and summarize the readings of their weights, we get the following (group names are conditional):
Some remarkable ratios:
1). ch's more than a dozens of them, but fewer gallows. Also, ch's more than o.
2). x is more than a dozen (if a = ei)
3). Although y is often found at the beginning of words, voynichese.com shows that y stands at the beginning in 2203 cases, and at the end - 15496. I think the difference is sufficient.
4). There are several words in the manuscript containing two gallows, and I deduced the relationships within their group thanks to them. But I could be wrong.
5). In the category with the symbol q, you can enter c, which has the property of standing in front of gallows (I mean cases like ct, not cth)
6). It is difficult to determine the weight of s, but I assume that it is the largest in the fozens group, as well as more than ch.
7). The symbols in the unit group are conditionally equal to each other.
8). There are cases when the letter y comes after n in a word, which suggests to me that it may be the "second unit", i.e. it may also be equal to 1, but in the case of rows of units, for some reason, it stands only after them. Maybe that's 0,5?
RE: Voynichese is a numeric cipher? - ololololo - 23-06-2026
It seems that using character weights it is possible to distinguish between nulls.
Perhaps, for example, a great quote of the author: cho chol chok choty chotey (f1r). All words start with ch, all contain the letter o, and the number of letters in words seems to increase with each new word.
But! ch stands at the beginning of words, but in theory it should stand after the gallows. It turns out that ch is a null here. We remove it and get: cho ol ok oty otey. Of course, there are doubts about the recurring o, but perhaps we'll leave it at that.
RE: Voynichese is a numeric cipher? - ololololo - 23-06-2026
It's time to share my latest attempt to reproduce the cipher. At the moment, this option seems to me quite worthy of publishing it here, since it fully meets the criteria of the intended source cipher.
We will turn the letters into numbers by adding 20 to their ordinal number in the alphabet. For example, A = 21 (XXI).
Abbreviations that we use: dif- = 15 (XV); -a = 61 (LXI), I decided to distinguish between a as a letter and a as a morpheme; -ica = 62 (LXII); -is = 70 (LXX)
The original text: Hepatica in silvis diffusa est
1). H-E-P-A-T-ICA = 28-25-36-21-40-62. Sort in ascending order: 21-25-28-36-40-62.
2). I-N = 29-34. We will not sort it.
3). S-I-L-V-IS = 39-29-32-42-70. Sort in ascending order: 29-32-39-42-70.
4). DIF-F-U-S-A = 15-26-41-39-61. Sort in ascending order: 15-26-39-41-61
5). E-S-T = 25-39-40.
Our super-enigmatic alphabet will be:
e = X
d = V
i = n = I
y = 1 (
ch = XX
sh = XL/XV
ei = IX (a little trick, without it, the text will be plain. If anything ei = a)
p = XXX
k = XX
T = L
Encrypting the text: Kchy chd diin pdn et tan
ka pdn s ka pin pa etin tee
ed kdn pa etn ta sh pa et
s - whitespace
Only the first word is recognizable, the result hurts the eye. But let's try to improve this with the help of nulls...
Kchy cheody daiin cphdain ethy (cthy) otain kain pdaiin s kain ethain (cthain) sh pain etham (ctham)
You can introduce a trick with ch as a null and dissect the words, and then, for example, pain would become cphy cha. It will be clearer and more practical for the decryptor. Or, for example, turn chd into chey dy. But I didn't think about it at first and decided to go head-on...
What can I say? As we can see, it is not difficult to turn a series of Roman numerals into a Voynich-like text. Secondary encryption seems to reproduce well even with such a simple example. This proves not only that it is possible, but also that in general it is not as difficult as it seems.
RE: Voynichese is a numeric cipher? - oshfdk - 23-06-2026
As I said elsewhere it's trivial to convert a short string to a semblance of Voynichese, especially so if using nulls and sorting letters in words. What does this really show?
RE: Voynichese is a numeric cipher? - ololololo - 24-06-2026
(23-06-2026, 11:55 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.As I said elsewhere it's trivial to convert a short string to a semblance of Voynichese, especially so if using nulls and sorting letters in words. What does this really show?Well, first of all, this is my most successful example so far (that's why I posted it at all). Secondly, yes, it's trivial
Using the methods I have described, you can quickly transform such a text into a Voynich-like form, and this will look logical with Roman numerals.About what it shows: this example shows that it is possible to easily reproduce the proposed encryption algorithm for the manuscript (based on my descriptions), without the need for complex models. We can also easily reproduce other properties of the text, such as making rows of words of the same length or making similar words (like chol chor), without experiencing any difficulties.
The ease and historicity of this nomenclator means that it could have been accessible to a medieval author.
Also, keep in mind that this is just an example of how the cipher works, not a reconstruction of it.
RE: Voynichese is a numeric cipher? - oshfdk - 24-06-2026
I don't think it has even been a problem to find a way to convert plaintext snippets into Voynichese snippets without any difficulties. And if nulls are involved it has never been a problem to explain the how of 'chol chol chor' (not the why though - why make the nulls obvious this way). The problem as I understand it is to match the large scale statistics of the manuscript.
Also, if your method involves sorting letters in each word, I think it's not encryption, because the underlying text is mostly destroyed. There is no way to tell 'read' from 'dear' or 'news' from 'Swen', it's just a way of generating garbage from a plaintext.
RE: Voynichese is a numeric cipher? - ololololo - 24-06-2026
(24-06-2026, 11:03 AM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I don't think it has even been a problem to find a way to convert plaintext snippets into Voynichese snippets without any difficulties. And if nulls are involved it has never been a problem to explain the how of 'chol chol chor' (not the why though - why make the nulls obvious this way). The problem as I understand it is to match the large scale statistics of the manuscript.You're right! It really is. Therefore, I assume that the author provided clues for decoding, expressed in the form of repetitions and similar words.
Also, if your method involves sorting letters in each word, I think it's not encryption, because the underlying text is mostly destroyed. There is no way to tell 'read' from 'dear' or 'news' from 'Swen', it's just a way of generating garbage from a plaintext.
The problem with READ and DEAR is actually solvable. Suppose A = I, E = II, D = III, R = IV. R-E-A-D = IV-II-I-III, from more to less: IV-III-II-I (RDEA); D-E-A-R = III-II-I-IV, from more to less: IV-III-II-I (RDEA). It's really unclear, and the correct word can only be guessed from the context. But I can do this: III-III IV-II-I (DD REA), or III-III II-I-IV (DD EAR), and it won't be so difficult to distinguish DEAR from READ. We can easily mark the first letter or simply leave it in place.
I assume that the author had basic intelligence and was also aware of this problem, but he could have solved it differently...