(24-06-2026, 11:27 AM)ololololo Wrote: You are not allowed to view links. Register or Login to view.I assume that the author had basic intelligence and was also aware of this problem, but he could have solved it differently...
There is no shortage of ways to implement a scheme superficially similar to Voynichese. One of these scheme could be the right one. The ultimate test is of course if Voynichese under this scheme decodes to anything sensible. But cracking even a more or less known method can be hard. A good intermediate result would be just showing how the system reproduces the statistics of Voynichese.
(24-06-2026, 01:04 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view. (24-06-2026, 11:27 AM)ololololo Wrote: You are not allowed to view links. Register or Login to view.I assume that the author had basic intelligence and was also aware of this problem, but he could have solved it differently...
There is no shortage of ways to implement a scheme superficially similar to Voynichese. One of these scheme could be the right one. The ultimate test is of course if Voynichese under this scheme decodes to anything sensible. But cracking even a more or less known method can be hard. A good intermediate result would be just showing how the system reproduces the statistics of Voynichese.
It is currently impossible to decipher Voynichese, as we do not have any indirect data (the weight of the characters helps us estimate their approximate meaning, but it does not provide much information). Additionally, we have very limited information about the structure of the cipher. Maybe author used multiple alphabets or other techniques?
About statistics... Well, it's potentially possible, but I don't know how to do it at the moment
But at the moment, I want to pay more attention to these mini-codes that reproduce the text of the manuscript. Perhaps something will come of it.
(24-06-2026, 01:04 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.A good intermediate result would be just showing how the system reproduces the statistics of Voynichese.
Just to be clear. When people say "a system that reproduces Voynichese statistics," they usually mean an algorithm that tries to follow ALL the characteristics of Voynichese (repeats, word similarities, low entropy, etc.), because they are understood as something composite.
From the point of view of my theory, all these features are mostly the consequences of secondary encryption. In any case, I'll have to change the text, which in its pure form will not be so convincing in appearance.
Of course, I can try (and plan to do so), but it seems to be useless.
(24-06-2026, 09:06 PM)ololololo Wrote: You are not allowed to view links. Register or Login to view.Just to be clear. When people say "a system that reproduces Voynichese statistics," they usually mean an algorithm that tries to follow ALL the characteristics of Voynichese (repeats, word similarities, low entropy, etc.), because they are understood as something composite.
From the point of view of my theory, all these features are mostly the consequences of secondary encryption. In any case, I'll have to change the text, which in its pure form will not be so convincing in appearance.
Of course, I can try (and plan to do so), but it seems to be useless.
I think the statistics part is mainly for your benefit as the solver. I'll be quite satisfied with just the plain old simple full text decoding producing a sensible plaintext. However, this requires a lot of effort, if at all possible, so the statistics thing may work as a good pre filter to avoid spending time on theories and solutions that simply won't work. I don't mind if you spend the next 3-5-30 years refining and building the numeric cipher case only to find that this is a dead end, it's not my life after all. But in my investigation I prefer to reject the things as quickly as possible.
(24-06-2026, 09:50 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I think the statistics part is mainly for your benefit as the solver.
This won't give us much of a clue. The model that fits the encryption method I described is incredibly simple. Even if we try to adjust it endlessly, it won't make much of a difference.
(24-06-2026, 09:50 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I don't mind if you spend the next 3-5-30 years refining and building the numeric cipher case only to find that this is a dead end
I'm not an expert, and my conclusions may not be accurate, and it's likely that if I take on this job, I won't get anything. However, this does not mean that the nomenclator theory is futile or unsuitable.
We can try to model the numbers used using the derived character weights, we can check the interchangeability of the characters to determine the two-digit numbers (we can also check their two-digit nature to see if they replace a letter or an abbreviated part of a word), and then we can determine the boundaries between the numbers... This will certainly enrich our knowledge of the manuscript.
(24-06-2026, 11:03 AM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.There is no way to tell 'read' from 'dear' or 'news' from 'Swen'
Let me add a little bit of speculation here.
Let's say I sent you a letter that I've encrypted by turning the words into anagrams based on a specific characteristic (for example, the weight of the letters).
The text looks like this:
Dear Joe, please read book and listen silent forest
And I'll encrypt it like this:
DREA EJO EPL EAS DREA KBOO DNA ENLSIT ENLSIT ESRFOT
Knowing other words, you can guess anagrams based on their meaning. However, if I encrypted Listen and silent differently, I would reveal the fact that they are different words. Without knowing anything, you could sit there and measure the statistics, and after 600 years, you would conclude that this text was dictated by a Chinese man to a medieval merchant.
In the case of Voynichese, the same - if we guessed one word, then we mean that we need to either change the order of letters in it, or look at other words, or doubt our conclusion. I do not see the complexity anywhere in the listed.
In fact, anagrams are not a serious problem, unless the encryption algorithm itself is complex, in which case it will be garbage.
I'm not sure I understand what you want to say here. I'm a bit repeating myself, but again, my main issue with this approach is not that it can't work. It's always possible to adapt an encoding scheme and overcome challenges and get something that would produce a semblance of Voynichese from some English or Latin plaintext.
However, I'm of the opinion that these exercises don't lead anywhere. I said the same about Naibbe cipher. Building a system that produces something similar to Voynichese is not the same as identifying a way to decode Voynichese and it's possible and even likely that any particular way of creating Voynichese from plaintext has nothing to do with the way the manuscript was actually created. As far as I understand, many people think differently and hope that building the right kind of a generator will bring us closer to understanding what Voynichese is, maybe your approach will be of interest to them.
(26-06-2026, 12:02 AM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.I'm not sure I understand what you want to say here.
I just added that anagrams are not a serious problem, and encryption in this way does not make the text "garbage", as you yourself said.
(26-06-2026, 12:02 AM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.many people think differently and hope that building the right kind of a generator will bring us closer to understanding what Voynichese is
And they're wrong. The very fact that we can create such a cipher does not tell us anything about Voynichese. The only thing we can prove using such models is that
they are possible in principle. But we can't take any of it.
In my case, it makes more sense because such a model will just be entertainment, where we will just play with the encrypted text.
Unexpectedly, I found an essentially perfect example of what I described in previous posts in the manuscript itself. This is f3r:
schey chor chal cham cham cho
This line is exactly what I described as adding a null (which in this case is the initial ch) and splitting the word into its component numbers.
The repetition of cham cham is also elegant in its own way. If am is a number and it appears twice in the encrypted word, then it will appear twice in any ordering (if I arrange the letters in the word morning in alphabetical order, it will become GIMNNOR, with the N duplicated). It seems too, too suspicious to me...
Also, with the help of this page, I can reasonably assume that m = iin. In addition to the fact that it matches a and o, like iin, it appears closer to the end of the line, and it seems to be in cases where iin would not fit. For example, if cham cham were written as chaiin chaiin, cho would not fit.
This may be the trace of the original cipher before it was transformed into Currier A.
It is also worth noting that there is a high probability that EVA a is indeed ei, as @Rafal suggested. I believe this is evidenced by the glyphs when a has "swings" like n or r. In EVA, this case is represented by the letter u. The question of this "wave" can be solved by representing these glyphs as en and er.
Can the "names of plants" give hints in the context of a numeric cipher?
Actually, not as much as we would like. Apart from additional information about the weights of the symbols, they do not give us anything. However, I can say that they generally follow the same rules as all other words and the text as a whole.
1). Most likely, the gallows in them are not a marker of the alphabet, table or something like that, but are a full-fledged part of the word. Perhaps this is some kind of root or part of a word found in many plant names.
By the way, in some cases, you may see a case where the letters are not placed in the expected order. Example: okeodof, where the gallows f stands at the end. I assume that here we see a "hint" from the author so that the decryptor can see where the letter f should be in the original word.
2). The names of plants, apparently, are also sometimes dissected. The first word may be followed by a series of hapaks, in some cases interrupted by a single common word like daiin.
Notable examples of dismemberment: pchor oeeockhy ofychey (f21r). pchor occurs 12 times throughout the manuscript. The word oeeockhy is suspicious...
At a minimum, it looks unusual compared to all the other words in the manuscript. Perhaps the symbols here are also "out of order";
And tshodpy oracphy koees ypsholy shy otoloaiin (f44r). There are a lot of non-standard words.
In general, at the text level, it is quite common to find alleged anagrams and "dissected" words.