Research is constantly underway here.
That’s how I see it.
Lots of new researchers are signing up with their ideas. They don’t read up on the subject and insist their idea is new. Even though it’s been discussed 20 times already that it can’t work that way, they still stick to their guns.
I’m tired of explaining everything over and over again. And I’m not the only one who feels this way.
Now the nonsense is spreading faster than the truth.
It’s great to watch things moving backward. But a hurdle race right away?
I have my successes; maybe the little bit of cryptology experience I learned from Klaus Schmeh is helping me there.
Translated with DeepL.com (free version)
(25-05-2026, 05:21 PM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.Who here is still honestly and consistently working on a plausible solution? And in what direction is it heading?
I am still working on matching the Shennong Bencaojing (SBJ) entries to paragraphs of the Starred Paragraphs secton (SPS).
The criterion is, as I explained in the Chinese Theory thread, matching certain hanzi (Chinese characters) in the SBJ entry to their presumed EVA translation/transcriptions in the candidate SPS parag, and checking whether the spaces between the latter match the spaces between the former, assuming an average ratio of ~5 EVA letters per hanzi. A good match is when I have several such cribs in the entry and the actual gap lengths in the SPS parag are well below ±5 EVA (±1 hanzi) of the predictions.
I have a program that runs that test for one SBJ entry on all SPS parags. But preparing the input for it is not trivial. The source for the SBJ I am using now is an OCR scan of the Zenghe Bencao (ZHB), a large state-sponsored materia medica published in 1080 CE which was still current after 1700 CE. First I must clean each entry taken form the ZHB, removing various late additions that were not supposed to be in the SBJ text that the VMS Author would have copied from, or were systematically skipped by by him (like the location where plants were grown, and the classification of the remedy in traditional Chinese medical theory.
Since I can't read Chinese, I am relying heavily on Google AI (GAI) for this part. GAI has access to multiple translations of each entry and to the Himalaya of scholarly papers that debated every letter of the SBJ and its variants over the last 2000 years. In spite of that (or because of that) I must ask GAI repeatedly and separately about each field and each disease, and confront it with other SBJ transcriptions; because it keeps hallucinating and contradicting itself.
It is taking me about one hour to prepare each entry for input to my matching program. The program then runs in seconds. It takes me another 15 minutes to check its output and decide whether the best matches it found are good enough. I had previously done ~65 entries from the wrong digital source file, then re-did 30 of them from the (hopefully) correct source file. Thus I have ~330 entries still to cleanup and process.
Doing the math, I concluded that I should instead take a plane to Beijing, borrow a library fac-simile of a 1400 CE edition of the ZHB, and pay a starving medical student to read aloud the white-on-black large-font text from it, while I write it down using an alphabet and shorthand invented for the purpose. Then, back to Campinas, I would pay another starving student to transcribe my notes into a computer, after teaching him my alphabet. It would take less time overall.
All the best, --stolfi
(26-05-2026, 04:56 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view. (25-05-2026, 05:21 PM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.Who here is still honestly and consistently working on a plausible solution? And in what direction is it heading?
I am still working on matching the Shennong Bencaojing (SBJ) entries to paragraphs of the Starred Paragraphs secton (SPS).
The criterion is, as I explained in the Chinese Theory thread, matching certain hanzi (Chinese characters) in the SBJ entry to their presumed EVA translation/transcriptions in the candidate SPS parag, and checking whether the spaces between the latter match the spaces between the former, assuming an average ratio of ~5 EVA letters per hanzi. A good match is when I have several such cribs in the entry and the actual gap lengths in the SPS parag are well below ±5 EVA (±1 hanzi) of the predictions.
I have a program that runs that test for one SBJ entry on all SPS parags. But preparing the input for it is not trivial. The source for the SBJ I am using now is an OCR scan of the Zenghe Bencao (ZHB), a large state-sponsored materia medica published in 1080 CE which was still current after 1700 CE. First I must clean each entry taken form the ZHB, removing various late additions that were not supposed to be in the SBJ text that the VMS Author would have copied from, or were systematically skipped by by him (like the location where plants were grown, and the classification of the remedy in traditional Chinese medical theory.
Since I can't read Chinese, I am relying heavily on Google AI (GAI) for this part. GAI has access to multiple translations of each entry and to the Himalaya of scholarly papers that debated every letter of the SBJ and its variants over the last 2000 years. In spite of that (or because of that) I must ask GAI repeatedly and separately about each field and each disease, and confront it with other SBJ transcriptions; because it keeps hallucinating and contradicting itself.
It is taking me about one hour to prepare each entry for input to my matching program. The program then runs in seconds. It takes me another 15 minutes to check its output and decide whether the best matches it found are good enough. I had previously done ~65 entries from the wrong digital source file, then re-did 30 of them from the (hopefully) correct source file. Thus I have ~330 entries still to cleanup and process.
Doing the math, I concluded that I should instead take a plane to Beijing, borrow a library fac-simile of a 1400 CE edition of the ZHB, and pay a starving medical student to read aloud the white-on-black large-font text from it, while I write it down using an alphabet and shorthand invented for the purpose. Then, back to Campinas, I would pay another starving student to transcribe my notes into a computer, after teaching him my alphabet. It would take less time overall.
All the best, --stolfi
Silly question, Jorge, but given your academic network, couldn't you find someone from a Chinese university to work with on it?
(25-05-2026, 05:21 PM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.When I look around here, there are very few people who are committed to working on a solution over the long term (e.g. Stolfi, Rubin Novacna, myself, u.a), and many of the “one-hit wonders” who pop in, propose a completely off-the-wall solution, and then disappear again.
That’s why I’m interested in the question: Who here is still honestly and consistently working on a plausible solution? And in what direction is it heading?
I believe that there is currently not enough information available to "solve" the manuscript, let alone prove the solution. I would try, but I fear that I probably lack the ability to make a genuine breakthrough in text analysis beyond some kind of initial insight, and even that is probably wishful thinking.
So I have been trying to expand the surrounding knowledge and context of the manuscript instead, such as the ink/parchment properties, new image analysis, and comparisons to other manuscripts for the imagery, paleography, the month names, and other marginalia in general. Hopefully, whatever small amount I can contribute will one day be a small part of evidence in favour of a found solution. Alternatively, I hope that those findings may help others to have new insights and leads, which could one day lead them to a solution.
(26-05-2026, 07:23 AM)DG97EEB Wrote: You are not allowed to view links. Register or Login to view.Silly question, Jorge, but given your academic network, couldn't you find someone from a Chinese university to work with on it?
You over-estimate the amount of interest that the VMS can generate among serious academics.
I personally know only one Chinese-speaking (and Cantonese-speaking) academic. He was two years my senior as engineering student at the University of São Paulo (USP), was my mentor when I was a student intern at the university's computing center, and my colleague for the five years I taught there. Since then he became Dean of the Math School at USP and president at a federal univ near São Paulo, and he is now retired. He knows about the Voynich Manuscript and about my Chinese Origin theory (COT) since ~2000. I once asked him about You are not allowed to view links.
Register or
Login to view., and he agreed that it sounded plausible. He finds the VMS and the COT intriguing -- but not nearly enough to spend any time on either.
Back sometime around 2002, after I became convinced of the COT, I sent a query by email to someone I picked almost at random at the Chinese Academy of Sciences (thus a Real Academic!). He had a quick look at the VMS and replied that he did not see anything that looked Chinese in it; it all looked European. Well, thank you, but I knew that...
I gave up on the VMS shortly after that because I concluded that, in order to make any further progress, one would need to be fluent on the right "Chinese" language and know how it sounded in ~1400 CE; which we know it was very different from how it sound now. Knowledge that I did not have and did not expect to ever gain. I had hoped that some "Chinese" linguist or paleographer would pick up the challenge, but apparently no one outside Europe and the Americas finds the VMS puzzle worth a second look.
That was before I knew about the SBJ and found the uncanny statistical coincidences between it and the SPS. I do hope that eventually the claim SPS = SBJ will be proved well enough that some "Chinese"-speaking scholar will take over.
All the best, --stolfi
I’m not working on any Voynich solution. Since the beginning of my study, mostly driven by curiosity, I’ve been trying to analyse the peculiarities of the text, look for patterns, and things like that, but today it’s impossible for me to propose any solution that really fits with what I found. Most of the things I found were already known before, I just arrived to them through diferent methods and experiments.
Right now I honestly can’t even say if I believe the text has meaning or not, or if it is a cipher or not. And I think it’s a waste of time to confront any theorical solution before we really understand the limits and characteristics of the text itself.
For me, the correct direction is this one: understand the text as much as possible first, and then maybe one day, with some inspiration (well, it may take more than one day), think about a possible solution that could explain everything.
I work tthe text in the search of a candidate set of ngrams (group of adjacents glyphs within a word) to conform the key of a cipher.
The most difficult thing is to determine is a word is a ngram or it can be split in 2 or more ngrams.
If a candidate set is not possible or it is not possible that they form a key then the meaningless hypothesis will be the right one.
To be clear: I still believe that the VMS consists of a meaningful text of some kind. Some of the text analyses I’ve conducted so far (most recently at the end of last year) clearly point in this direction. This also confirms the simple assumption that a work that was, after all, created with a certain amount of effort does not have completely meaningless content. For me, the main goal in terms of successful decryption is to recognize and replicate the system behind the text’s creation. On the one hand, one can approach this task step by step, using statistical analyses. On the other hand, it might also be possible by looking at the big picture—who knows.
Anyone reading this thread who has never seen the Voynich will think it's a work consisting solely of text and no images. Such is the almost exclusive obsession of researchers with the script. This exclusivity is not new. In the 20th century, the goal was also to decipher the supposed text. And even in the 17th century, the concern was the strange characters of the Voynich.
I wonder, as I have before, if it wouldn't be more fruitful to change our approach. Why not give more importance to the images we see and investigate their relationships? Images were a powerful means of communication in the Middle Ages.
No solution, per se, but I am still working on comparing and finding more coincidences with seeing quire 13 imagery as forming a periplus which can be reconstructed into something akin to a nautical chart of the old ecumene, through coordination with some visual information in quire 14.
However, things appear to be rotated, changed in comparative sizing or perspective, and seem to have been purposefully drawn both true and untrue to life, just enough of the latter to be unrecognizable to most in the way i see it, but just enough of the former to be convincing to me; and from my perspective, too many coincidences would have to exist for it not to be so. (i.e. I think the coincidences exist because that's what they were going for) Of course, it also has naked nymphs standing in strange poses nearby, distracting from the true size scales and topics involved. Yet as nymphs they embody the very things they stand for, and what I think the quire is about: water. Most convincing to me, is that in the right order, each page starts and stops with something adjacent to the previous and next pages.
It seems to be going well insofar as my own impression of my discoveries, it has always surprised me how I can look again and see more than I saw the last time, and have lately found some places I thought had been missing, and found more coincidences, that continue to amaze me to have been the case, if I am misinterpreting something that is not as I am describing. It is taking forever to put together the comparisons though, and I don't know that it will convince anyone else, but I want to try, in case it helps anyone else to see it the way I do, since I think it is amazing.
My current project is clipping nautical chart parts in ways that correspond to the vms imagery, creating a version of each page of quire 13, or sets of facing pages where applicable, wherein the nymphs are mnemonics, and each item is replaced with the items I think they stand for, arranged so they match the drawings in terms of positioning and sizing, such that they can be compared in this way. Not so much to look alike but to show what I think it is meant to show. Also I hope to show that there is a clear sequence that is followed which does not cherry pick but includes all the major components one would expect in a map of the world, at least in its shorelines, since again, it is water oriented.
I don't know what the deal is with the text, but for the amount of information I see in the images, I can only hope there is also meaning to be gleaned from the glyphs, makes no sense to me that anyone would spend their time filling it with nonsense, although I guess it does distract from the imagery, as mentioned in Antonio's post. I still hold hope of meaning in the text, but the drawings have already shown me the world.