The Voynich Ninja

(08-06-2026, 09:40 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.Try to imagine the situation.  The most pressing concern must have been to record what the Dictator was saying, without slowing him down too much.  Only in second place, the notation should let him later read what he wrote and (if he had the time) ask some doctor to please explain what "bèn tún" meant, so that he could put that in his glossary.

So, if I understand correctly, the SBJ was widely printed and distributed with the same chinese characters across a large area consisting of various dialects and languages. In each of these areas, people would read those characters in their own dialect and come to understand the contents.

If this is the case, and if the VMS creator wanted one day understand the contents by asking someone, why on earth would it be a more logical choice to create a written language that depicts a single dialect when read aloud? Surely the chinese symbols would have been hugely more valuable here, as they may have been understandable to far more people over a far larger area (in fact, by any potential learned asian person who may have arrived by sea in europe). On the other hand, the VMS script spoken aloud would have only been half-understandable to a single dialect. 

How many of the words in the SBJ do you expect the VMS author to have understood when spoken? Perhaps a couple of the nouns, perhaps some simple verbs? I cannot imagine that he would ever have expected to understand much of it. In that case, he would surely have expected to to be necessary to eventually ask someone.

This theory implies that he thought that he would later be able to understand the majority of what he wrote, or that he would at least be able to later seperate the nouns, verbs and adjectives. Otherwise, would he not eventually ask something like: "Do you have any 'commonly found'?" / "Where can I find 'grows'?" / "Do you sell any 'slowly'?".

(08-06-2026, 10:42 AM)eggyk Wrote: You are not allowed to view links. Register or Login to view.So, if I understand correctly, the SBJ was widely printed and distributed with the same chinese characters across a large area consisting of various dialects and languages. In each of these areas, people would read those characters in their own dialect and come to understand the contents.

If this is the case, and if the VMS creator wanted one day understand the contents by asking someone, why on earth would it be a more logical choice to create a written language that depicts a single dialect when read aloud? Surely the chinese symbols would have been hugely more valuable here, as they may have been understandable to far more people over a far larger area (in fact, by any potential learned asian person who may have arrived by sea in europe). On the other hand, the VMS script spoken aloud would have only been half-understandable to a single dialect. 

How many of the words in the SBJ do you expect the VMS author to have understood when spoken? Perhaps a couple of the nouns, perhaps some simple verbs? I cannot imagine that he would ever have expected to understand much of it. In that case, he would surely have expected to to be necessary to eventually ask someone.

Jumping back in briefly, this kind of recitation of a classical text was extremely unusual because it is incomprehensible. Language change meant it was a very different situation from Latin where national pronunciations were quite understandable and educated people were expected to be able to converse in them. A phonetic transcription would have been opaque to the reader even if he were proficient in the scribe's alphabet. Further, the gap in vocabulary was wider than the gap between Latin and it's vernacular descendents, not just because of sound change, but because of a fundamental shift from one character to two character vocabulary, so the scribe would have no point of entry for understanding it. This method would not create a useful record for the scribe, nor could someone trained in Classical Chinese understand it if he read it back to them, and this would have been apparent to everyone involved at the outset

(08-06-2026, 10:42 AM)eggyk Wrote: You are not allowed to view links. Register or Login to view.why on earth would it be a more logical choice to create a written language that depicts a single dialect when read aloud? Surely the chinese symbols would have been hugely more valuable here, as they may have been understandable to far more people over a far larger area (in fact, by any potential learned asian person who may have arrived by sea in europe).

An European who liveed for 4-5 years in "China" (China proper or any of the other candidate countries) would have become reasonably fluent in the local spoken language,but would be essentially illiterate in the written one.  Have you ever tried to learn Chinese characters? 

Thus copying the Chinese characters would be pointless: he would not be able to read them when he got back to Europe.  Whereas, if he  recorded the sounds of the characters as read by a local  doctor, back home he could  at least understand the common words, and could hope to learn the others by context etc.

Besides, if copying the characters was ok, it would have been simpler to buy a copy of that same medical encyclopedia.  Presumably he had to pay the Dictator for the service, so that would probably have been cheaper too.

Quote:How many of the words in the SBJ do you expect the VMS author to have understood when spoken? Perhaps a couple of the nouns, perhaps some simple verbs?

 

Just a guess from the blue, maybe 50%.  Here is a typical recipe:

  <b1.4.095>

  (A)    |Donkey-hide Gelatin:
  (A1)   |··[Nature] Sweet, neutral.
  (A3)   |··[Mainly for]
  (A31)  |····heart and abdomen
  (A311) |······internal hemorrhaging;
  (A32)  |····extreme exhaustion
  (A321) |······with shivering,
  (A322) |······resembling malaria;
  (A33)  |····lower back and abdomen
  (A331) |······pain;
  (A34)  |····four limbs
  (A341) |······aching soreness;
  (A35)  |··(In women)
  (A351) |····uterine bleeding,
  (A352) |····stabilizing pregnancy.
  (A4)   |··[Long-term Use]
  (A41)  |····lightens the body,
  (A42)  |····boosts vital energy.
  (A5)   |··[Another Name] Fùzhì-brand gelatin.

I don't know whether the specific Chinese characters for "lower back and abdomen" in that recipe (腰腹, yāo fù in modern Mandarin) were the ones commonly used at the time, or were rare synonyms used only in the SBJ and other old medical texts.  That is, whether yāo fù means "back and belly" or "lumbar and abdominal region".  Assuming the former, he probably understood those words right away -- and would have understood them later, back in Europe, reading his transcript.

Quote:In that case, he would surely have expected to to be necessary to eventually ask someone.

Indeed.  So he probably did build a glossary on the side. And not just for this "VMS project", but probably also for new words that he encountered outside of it.

All the best, --stolfi

(08-06-2026, 05:21 PM)rikforto Wrote: You are not allowed to view links. Register or Login to view.Jumping back in briefly, this kind of recitation of a classical text was extremely unusual because it is incomprehensible. Language change meant it was a very different situation from Latin where national pronunciations were quite understandable and educated people were expected to be able to converse in them. A phonetic transcription would have been opaque to the reader even if he were proficient in the scribe's alphabet.

Sorry, I don't understand this objection.  What "reader"? 

Again, the premise of this scenario is that the Author had lived in "China" long enough to have learned the local spoken language.  That was the language that the Dictator used when reading, and the language that he wrote down.  The existence of other languages elsewhere, that would also read the same text but with different sounds, was not relevant at all.

Maybe you are still hung on the idea that  the Voynichese script was created by some community, to communicate among themselves?  That idea is definitely excluded by the COT.

All the best, --stolfi

reader = Dictator above

I perhaps misunderstood something important here. Are you arguing you've used the Classical Chinese text to identify a translation? Or are you arguing that it is in Classical Chinese, phonetically rendered with vernacular pronunciation? I took from our conversation about Kanbun that it was latter, but you seem to be implying the former now

(09-06-2026, 02:01 AM)rikforto Wrote: You are not allowed to view links. Register or Login to view.reader = Dictator above

Ah, sorry.  Indeed, in that scenario the Dictator did not know or care about the Voynichese script.  He just read the Chinese characters aloud, and the Author just wrote down the sounds.  

Quote:I perhaps misunderstood something important here. Are you arguing you've used the Classical Chinese text to identify a translation? Or are you arguing that it is in Classical Chinese, phonetically rendered with vernacular pronunciation? I took from our conversation about Kanbun that it was latter, but you seem to be implying the former now.

All I can say is that the Starred Parags text quite accurately matches the Chinese character text of the Shennong Bencao, as extracted from the Zhenghe  Bencao (ZHB) file, in the sense that the cribs I have identified so far occur in both texts with proportional spacings (at the ratio of ~5 EVA letters to 1 hanzi) -- with accuracy that is very unlikely to be a mere coincidence.  

That is the case for the longest SBJ recipe, the "red rooster" (ROOS) and the longest single-star SPS parag (f105v.32), and for several other pairs as well.  

The cribs are not categorical, though: of the eight 主 in ROOS, five match a daiin in the SPS parag, while the other three match a dair, a laiin, and a kaiin.  All but three of the nine distances before, between, and after these cribs on the SPS -- 8, 105, 11, 15, 46, 19, 43, 25, and 51 EVA letters -- are all less than  ±5 letters away from the expected distances based on their distances in the ROOS recipe (15, 115, 15, 15, 35, 20, 40, 25, and 50).  Those errors correspond ot less than one hanzi in the recipe.  The three exceptions are the gap before the first matched daiin, which is 8 EVA letters instead of the predicted 15 (an error of -7 letters, which is ~1.4 hanzi); the first gap between the cribs, which is 105 letters instead of the expected ~115 (an error of -10, or ~2 hanzi); and the 4th gap, which is 46 EVA letters long instead of the expected ~35 (an error of +11 or ~2.2 hanzi).  There are no other daiin or laiin in that parag; there is only one other dair, that does not match any 主, and four other kaiin. 

Those position matches (and the match between the histograms of recipe and parag lengths) are the facts.  The scenario with the Dictator and phonetic writing is my best explanation for how those facts came to be.  And my theory that the Scribe sometimes misread iin in the Author's draft as ir, and d as k or l, is my best explanation for the variant translations of 主.  There are other possible scenarios, but they seem a lot less likely.  That scenario has at least many historical quasi-precedent: Europeans who lived for years in "China" and learned the local spoken language. And in a few cases, even recorded it phonetically.

Whether the local language was a Chinese dialect or not, the text that the Dictator was reading was almost certainly the text, in Chinese characters, that was quoted in the ZHB or some similar treatise available at the time.  If the local language was a Chinese dialect, the Dictator would surely have read the hanzi in that language.  If the local language was not a Chinese dialect (like Vietnamese, Burmese, Tibetan, ...), the Dictator may have read each hanzi in some pseudo-Chinese (like the Japanese "on" readings) or in the local language (like the Japanese "kun" readings).  At this point I cannot tell which of all these possibilities is the case.

All the best, --stolfi

(09-06-2026, 07:48 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.All I can say is that the Starred Parags text quite accurately matches the Chinese character text of the Shennong Bencao, as extracted from the Zhenghe  Bencao (ZHB) file, in the sense that the cribs I have identified so far occur in both texts with proportional spacings (at the ratio of ~5 EVA letters to 1 hanzi) -- with accuracy that is very unlikely to be a mere coincidence.  

Insofar as you think the match is accurate, there is no ambiguity about the language. The Shennong Bencao Jing text you identify with the SPS is written in Classical Chinese. My point is that this has implications for the intelligibility of dictation separate from the problem that the Scribe would not know Classical Chinese, though I also don't think it's all that plausible he would extensively record a language he did not know if it were theoretically readable

(09-06-2026, 04:02 PM)rikforto Wrote: You are not allowed to view links. Register or Login to view.Insofar as you think the match is accurate, there is no ambiguity about the language. The Shennong Bencao Jing text you identify with the SPS is written in Classical Chinese.

The written language was Classical (written) Chinese, allright.  But, if the local language was any of the 50 dialects of Chinese, the Dictator would read each character as a syllable in that language.  

That is, the same 止血 characters would be read as zhǐ xuè by a Mandarin speaker, as zi2 hyut3 by a Cantonese speaker, as chṳ́ hiat by a Hakka speaker, etc.  Except of course for the 600 years of language change.  That is how the Chinese script works.  There is no such thing as "spoken Classical Chinese". 

A Cantonese Dictator would not read the characters in (1400's) Mandarin.  First, in his view the SBJ was just a written text, not a text in Mandarin or in any other specific dialect.  Second, he quite possibly did not know spoken Mandarin: he would not have needed to. And, third, he was dictating to a gringo who had become somewhat fluent in Cantonese, not in Mandarin.

The hanzi:EVA::1:5 proportionality seen in ROOS and other recipes would not depend on the Chinese dialect that was used by the Dictator.  In any of them each hanzi would almost always be read a single syllable.  The occasional exceptions may explain the gaps that are off by 1 or 2 hanzi from that proportonality.

For example, the name of the recipe is 丹雄鸡, literally "red male chicken".  The corresponding EVA text in parag f105v.32 is poar.keeo -- two words and 8 EVA letters instead of three words with ~15 letters. That may be because (say) poar is actually two words po.ar, and in the local language 丹 and 雄 read as two very short syllables. But it may also be because, in the local language, there was a single-syllable term for "rooster", and people would normally use it instead of "male chicken".  (Like in English people would say "rooster", never "male hen".)  Maybe.

The proportionality should also hold -- but with more variance -- if the local language was not a Chinese dialect (like Vietnamese or Tibetan). The Dictator would still read each character as a single syllable in the local language; but then the grammar would have been quite strained.  Still, given the very simple grammatical structure of the SBJ recipes, the stilted grammar would not have been a big problem for the Author.

Quote:the Scribe would not know Classical Chinese

I am using "Author" for the person who wrote the original draft, and "Scribe" for the person who clean-copied that draft to vellum.  Neither could read Classical (written) Chinese.  The Author would have been somewhat fluent in the local language and could pronounce what he wrote, although he probably understood only part if it.  The Scribe had learned to write the Voynichese alphabet but probably did not know the sound value of any glyph.

All the best, --stolfi

(10-06-2026, 02:21 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.But, if the local language was any of the 50 dialects of Chinese, the Dictator would read each character as a syllable in that language.

I understand this. The problems I have raised are a direct consequence of taking a Classical Chinese text and dictating it this way

(10-06-2026, 06:42 AM)rikforto Wrote: You are not allowed to view links. Register or Login to view.

(10-06-2026, 02:21 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.But, if the local language was any of the 50 dialects of Chinese, the Dictator would read each character as a syllable in that language.

I understand this. The problems I have raised are a direct consequence of taking a Classical Chinese text and dictating it this way

Sorry, I don't understand what problems you are referring to.

• The Dictator reads the SBJ aloud in Cantonese.
  
• The Author writes down the Cantonese (which he speaks) in a phonetic notation.
  
• Author goes back to Europe.
  
• Author reads the Cantonese and understand it (minus technical terms).
  

What is wrong here?

Do you mean the syntax?  You have seen examples of recipes: there is hardly any syntax in them.  The most complicated sentences are like "lengthens life", "releases blocked blood", "cough from reversed qi", ...

All the best, --stolfi