The Voynich Ninja

Well, that's too unfortunate - I don't understand anything anymore - so I can only sit back and hope that you will eventually have a consistent model that translates larger sections. I can't even figure out if it's Eisigesis or not...

But is it supposed to be a native Arabic speaker who transcribed a Chinese text phonetically in China, and that then turned up in Europe?

(15-04-2026, 06:37 AM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.hope that you will eventually have a consistent model that translates larger sections.

Well, I already have a translation of the whole Starred Parags section.  But I know nothing about the language, other than those three cribs and that the words map to Chinese characters mostly one-to-one. And the parags seem to be jumbled, and finding the matching pairs is not easy.

The recipes were not numbered in the original Shennong Bencaojing, and the Chinese did not have the notion of "alphabetic order" as we have and use now.  It was not like Culpeper's Herbal, but more like the Alchemists' Herbal.  Thus different editions of the SBJ reordered the recipes as the editor felt was more logical. I got two digital versions from the internet, and they are in very different orders.

Moreover, I suspect that, when taking dictation of the original, the Author wrote down each recipe (or every few recipes) on a separate piece of paper, and did not bother to number those pages; so that they got scrambled before they were put to vellum.  

Then the bifolios of that VMS section were scrambled when the book was bound...

Quote: But is it supposed to be a native Arabic speaker who transcribed a Chinese text phonetically in China, and that then turned up in Europe?

The best and most comprehensive geography treatises of the antiquity were written by Arab travelers, and were well-known in Europe.  In 1400 a Muslim would be extremely unwelcome in France, England, or north of the Alps; but Spain and Portugal were part of the Islamic Empire, and Muslims would be tolerated in other parts of Europe, including some places in Italy (like Venice, maybe).  However, an Arab who converted to Christianity should be accepted anywhere in Europe. 

But there were also many great Jewish travelers, and Jews did live everywhere in Europe at the time.

When the Portuguese explorer Vasco da Gama reached India by sea for the first time, he was surprised to find a Polish Jew living there, with some status in the local court.  Eventually this guy agreed to go back to Portugal with Vasco, and along the way he converted to Christianity and was adopted by Vasco as his son, with the name Gaspar da Gama. Gaspar was valued in the Portuguese court for his knowledge of languages, including several of India.  

In 1500, when the crown sent a small fleet to retrace Vasco's route, Gaspar went along as interpreter.  The expedition went way off course (maybe intentionally) and ended up "discovering" Brazil.  Gaspar tried to talk to the "indians" who came to the shore, but they did not understand any of the "indian" languages he tried.  The ship's scribe noted in his report that it must have been because of the noise of the waves. 

But, anyway, note that the VMS illustrations only tell us that the Scribe was probably from Europe. They don't imply that he was in Europe when he penned the book.

All the best, --stolfi.

(15-04-2026, 07:44 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.Well, I already have a translation of the whole Starred Parags section.  But I know nothing about the language, other than those three cribs and that the words map to Chinese characters mostly one-to-one. And the parags seem to be jumbled, and finding the matching pairs is not easy.

Setting aside the controversy of whether or not the SPS sits in the Shennong Bencaojing textual tradition or not, I do not think you can claim to have translated the text. You do not know what most of the SPS says under the assumption that it is the Shennong Bencaojing tradition, nor have you identified any process by which it encodes the text that could result in a repeatable translation, and as you concede you haven't even identified a specific Shennong Bencaojing or translation (even a previously unattested one) to say you have translated the SPS to.

I still think the specific ways translation is failing cast significant doubts on the claim it sits in that particular herbal tradition, but I haven't had a chance to look closely at the new proposed cribs. I will do so in the next few days, but a few new cribs---especially with the caveats you've placed on them quoted above---does not a translation make

(15-04-2026, 05:49 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.And , by the way also, the match with f104v.22 would be even better if one accepted lsaiin as a spelling error for daiin. Note that ls and d are in fact rather close in "ink distance"...

Out of interest, what does a "spelling mistake" actually mean in this context? Like, what are the individual characters representing?

If voynichese is a sort of insruction set on how to write out words (hand strokes for chinese characters), a spelling mistake implies an incorrect drawing instruction. If they represent sounds, or instruction on tones, it represents an incorrect tone for the word. If either of these are the case, you would expect that similar looking (or sounding) chinese words would share features in spelling. 

The alternative is that the author chose random sets of voynichese characters for each individual word, and then maintained consistency across pages for each word. That however begs the question "Why bother?".

And it turns that closer look requires some more clarification. Can you elucidate how you arrived at You are not allowed to view links. Register or Login to view. cribs beyond frequency? This broad method has been attempted in many languages, and it is always possible to identify potential cribs by assuming high frequency words correspond. (Most recently, and illustrating how it often has sophisticated bells and whistles, You are not allowed to view links. Register or Login to view..) It seems like you matched specific parts of the text as well, and that might give more insight into the viability of those correspondences

As an aside, you may also want to check if the Shennong Bencaojing has a Zipfian distribution of characters. Usually Chinese characters are not distributed that way, which is another reason I have doubted a Chinese source

@ rikforto Thx 
Somehow, the fact that this is a viable way to decipher a text from the 15th century seems to be overlooked...

(15-04-2026, 01:49 PM)rikforto Wrote: You are not allowed to view links. Register or Login to view.I do not think you can claim to have translated the text. You do not know what most of the SPS says under the assumption that it is the Shennong Bencaojing tradition

Actually I do know the complete translation of the SPS into Chinese (including that of the four missing VMS pages!), and translations of that Chinese text into English are widely available.  

So I can tell you right away that the SPS says that roasted dung beetle is a good remedy for mental disorders, that pig testicles will stop your vital energy from running around your body like a scared piglet, that the white part of chicken droppings is a cure for diabetes, and that if you eat 1 kg of jade powder just before you die your body will not decay for three years after that.  Seriously, not kidding.  And a couple thousand other fascinating pearls of 2000-year-old medical wisdom like those. I have them all on file on my laptop.

What I still don't know is;

The matching of individual SBJ recipes to specific SPS parags. I have only a couple dozen tentative pairings so far.  I have code that scans the whole SPS to find the best match for a given SBJ entry, but it takes half an hour or more to prepare the data for it.  Because I must first get a translation of the entry in order to decide what specifically to search for, and then look at the output to decide whether the best matches are correct or just coincidences.  

This will get better as I get more cribs, like the two new ones above.  There may be strings of several consecutive  SBJ recipes that occur in the SPS in the same order; but I cannot tell that with confidence yet.

Precisely which VMS words correspond to which Chinese characters.  The transcription of the SBJ into SPS seems to be fairly literal, with one Chinese character corresponding to one Voynichese word (a bit more if commas in the SPS file are considered word spaces, a bit less if they are ignored).  Also one Chinese character corresponds to almost precisely 5 EVA characters.  So the SPS parag that is the translation of a 50-character SBJ entry will have very close to 50 Voyniches words or 250 EVA characters.  Indeed the way I look for matching parags is by comparing the distance between occurrences of known keywords, scaled by those factors. 

But that is on average.  Character 23 of that entry may correspond to any of the words 20 to 25 of that parag.  Or maybe to half of one of those words, or to a pair of words in that range.  And then there are many spelling variations, possibly spelling errors.

Which parts of the SBJ entries were omitted by the VMS author.  The VMS author apparently omitted systematically certain fields of each entry, like the "flavor" and "warmth" of the remedy (that only make sense within traditional Chinese medical theory), alternative names, and places whether the remedy grows or is collected (which obviously are useless outside China, and even within it.)   So the SPS will not tell you that the penis of a white horse is salty (dried, of course), nor that iron grows in low-lying marshes and plains --- which the SBJ will clearly tell you.  

But it seems that the VMS also omitted some parts of some entries that presumably he did not care about. Like the fact that the white grubs that grow in chicken manure are good for fattening pigs.  Or that reddish vaginal discharge is a condition specific to women, not of people in general.  And that aconite, a famous poisonous plant, is toxic.  And that is why I must check the translation of each entry before searching for its match: to remove the parts that were certainly omitted, and repeat the search with and without those that may have been omitted.

What is the language.  The fact that the Chinese characters are mapped to Voynichese words almost one to one implies that the text is either a phonetic transcription of the original Chinese SBJ, read by a native in any of 50 Chinese "dialects"; or is a reading of a literal translation of the same into any of a few other languages, like Vietnamese or Tibetan.   That is, the words may be Tibetan, but with the original Chinese word order.  (That type of "translation" from Chinese was in fact common in those places and times.) 

If the meaning of q is indeed "and", as suggested in my previous post, that makes this second alternative more likely -- because Chinese itself, written or spoken, does not have a (commonly used) word for "and", it just runs the various items of a list together.  But the q may alsoy be an "and" sign that the author added on his own, during the dictation or afterwards, because he found that Chinese "implicit and" syntax too confusing.

What are the letters of the alphabet and their sounds.  Basically a consequence of the above.  I still cannot tell whether CTh is one phoneme or four, whether Che is a modified way of pronoucing Ch, or a Ch sound followed by an e sound, etc.

Quote:nor have you identified any process by which it encodes the text that could result in a repeatable translation

I have said several times previously that the reason why no progress has been made towards a solution in the past 100 years is that every minimally serious researcher makes the same chain of logical mistakes, (1) to (3) below.  But (1) is wrong logic and the conclusion is false, thus everything that follows is false too.  And (3) has a number of other pernicious consequences:

1. "The material, layout, and drawings are European, therefore the language must be European"
   
2. "But the script is not related to any know script, therefore it must be some cipher"
   
3. "But it is not just an original alphabet for any European language, so it must be some non-trivial cipher"
   
4. "being a cipher, it must be invertible, with little if any ambiguity"
   
5. "there must be few if any errors, because one error can ruin the decription"
   
6. "finding the solution will be easy, as cracking any cipher: just guess the right algorithm"
   
7. "once the algorithm is identified, apply its inverse and immediately you will have the whole plaintext"
   

Well, the sorry truth is that

1. The language is not European.
   
2. It is not in cipher.
   
3. The language (even if it was Mandarin) is mostly unknown because of language change.
   
4. There are lots of spelling errors and other errors.
   
5. Deciphering it will be hard and slow work, like deciphering any text in a dead language.
   

It was a big stroke of luck that the SPS could be identified as a specific surviving Chinese book.  We may not have the same luck with the other sections.  I am confident that they are in the same language and spelling as the SPS, but, unless their source books can be identified too, deciphering them will be very hard and slow work.

All the best, --stolfi

Quote:It is not in cipher.

How do you know it?

If I understand correctly at this moment you have just supposed matches for some words.
You don't know what language they are in and how they should be pronounced.
And you don't know how the script works - is it phonetic, syllabic, logographic/made of codes, describing shape of Chinese letters or yet something different.

So how do you know it is not a cipher?

The difficulties you are describing largely stem from the fact that your method does not, at present, produce a translation per se. This is true even if we take it as a given that you have identified the textual tradition. (You have not identified the source text; in fact you repeatedly brought up the issue that there is no singular Shennong Bencaojing as an explantion for some of the difficulties you are having.) On the terms of your claim here, you have only identified a correspondence between 12 of the paragraphs. I have two problems with this:

1. Unless there is a major advancement I have not seen, the method is not a translation, or, depending on how you want to split this hair, there are significant gaps in the translation. In the "rooster paragraph", for instance, only a handful of characters are identified. This identification does not yield certain translation of other characters in the paragraph, at least as of yet.
   
2. Outside of those 12 paragraphs, the SPS remains opaque. Your position is that you cannot presently identify them, so even if I allow that identification as in the previous point is "translation", most of the text remains opaque
   

I have made the case elsewhere on this thread why I think the identifications in the rooster paragraph are unconvincing, and I don't see a need to recapitulate those arguments here. My point in my most recent post was simply that what you've put forward is, on its own terms, much more limited than a translation. I will, if you provide them, look at the 11 new paragraphs you claim to have identified and update my beliefs based on that evidence, but I certainly don't see a translation at present

(15-04-2026, 02:50 PM)eggyk Wrote: You are not allowed to view links. Register or Login to view.Out of interest, what does a "spelling mistake" actually mean in this context? Like, what are the individual characters representing?

My best guess, as I have said before, is that the Author was a foreigner (probably European, but may have been Arab, Persian, Turk, Jew, etc.) who wanted to bring home some of those Chinese books that were highly regarded by the locals.  He may have been more or less fluent in the spoken language, but could not read the script. (I suspect that he could not even tell whether some Chinese text was upright or upside down, and had not even noticed that Chinese books have the binding on the right side.)  

So he did the only thing he could: he recruited a native to read the book aloud, while he wrote it down using a phonetic alphabet that he had devised to record the weird sounds of the language, as efficiently as he could -- meaning, with few strokes per phoneme.  As we discussed before, he could not write down a translation because most of the words were names of plants and diseased that he did not know.  He must have thought that he could build a glossary later. (And maybe he did, but it got lost.)

Thus there is in theory a "correct" version of every word in the VMS, namely the phonemes (with tones and everything) that someone truly fluent in the language would hear when the original Chinese character was properly pronounced.   An "error" is when the sequence of glyphs that we get from the transcription file does not match that ideal transcription.

There are many possible causes of those errors:

• The Dictator did not know the reading of some Chinese characters, and said a random syllable instead.
  
• The Dictator's pronunciation was sloppy, so his zhǔ often sounded like zhū, his qì came out as xì, etc. 
  
• The Author was partly tone deaf, and/or found it difficult to tell shēn from shēng, etc.
  
• He often wrote the wrong glyph but could not go back and fix it because he had to  keep up with the Dictator
  
• His handwriting was not very clear, so 20 years later not even he himself could tell whether the glyph he wrote was d or k, ee or Ch
  
• The Scribe found reading his handwriting even more difficult, and also made many errors from distraction (like swapping t for k, or ain for aiin
  
• BEEP
  
• The modern transcribers often had to guess whether something as Ch or ee, r or s, y or a - and guessed wrong.
  

In fact I think I see clues of most of these events having occurred in various places.

Moreover the script devised by the Author may have been intended to be a stenography system, not a proper spelling.  In stenography one often omits details of pronunciation in order to reduce the size of the alphabet.  (For example, in one system that I learned in my teens - but never used -- the voiced/unvoiced distinction was to be ignored, so that the same squiggles were used for p and b, t and d, s and z, k and g, sh and zh, etc.  When reading that script back, one usually had no trouble resolving those ambiguities.  As happens when whispering, which is speech with voicing permanently turned off.)

All the best, -- stolfi