The Voynich Ninja

(23-08-2026, 05:34 PM)Linda Wrote: You are not allowed to view links. Register or Login to view.You had said that you found a term in a recipe re red rooster, that matched a vms word which appears to occur at the same frequency, and that sounded very promising. Can you show this correspondence visually with whatever copy you are using and the starred paragraph you have linked with it

The Shennong Bencaojing (SBJ) entry <b1.4.096>,72 hanzi:

           3 　　 丹雄鸡
        1 12 主　 崩中漏下赤白沃补虚温中止丹雄鸡
        1  8 血　 通神杀毒辟不祥头
        1  3 主　 杀鬼肪
        1  3 主　 耳聋肠
        1  7 主　 遗溺肶胵裹黄皮
        1  4 主　 泄利屎白
        1  8 主　 消渴伤寒寒热翮羽
        1  1 主　 下
        2  2 血闭 鸡子
        1 10 主　 除热火疮痫痉可作虎魄

The entries in the third column are the "crib" keywords in the entry: 主 = "mainly for", 血 = "blood", 血闭 = "amenorrhea". The fourth column are the characters between those cribs.  So the whole entry is "丹雄鸡: 崩中，漏下，赤白沃。补虚，温中，止血，通神，杀毒，辟不祥。头：..."  The first two columns are the character counts in the other two columns.
 
The Starred Parags Section (SPS) parag <f105v.32>, 371(+6) EVA letters (excluding punctuation)

              8(-7) poarkeeo
  daiin      63(+2) qoaiinaracpheeyqoeedeodyqokaiinqotedaiinaporaiinapylsheodytaiin
  oteey      42(+1) oteeoolotaiinokeeyqokaiinoraiiinaldalsheeo
  daiin      17(+1) chsdqokeeeydaiino
  kaiin      22(+6) otaiinchedaiinolkallkl
  dain       34(-1) doeeokcheeoltaiinotcheedychoraiino
  daiin      19(-1) chedyotaiinalkaishd
  laiin      45(+4) sheodokeeodyqoaiinytaiinotaiinchdaldydaiiinch
  daiin       6(+0) ockhey
  ysheyckhy  10(+0) sheoqoeeol
  kaiin      52(+1) chsokoltchdysheeeyokaiinaraildycheodyoaiiinainokshey

The entries of the first column are the substrings of the VMS entry that are claimed to correspond to the cribs in the SBJ entry as parsed above.  That is, it is claimed that the first 主 corresponds to the "daiin" after "poarkeeo", the first 血 corresponds to the "oteey", the second 主 corresponds to the "daiin" after "dalsheeo", etc.  

The entries of the last column are the strings between those cribs. That is, "poarkeeo" would be the transcription of 丹雄鸡, "qoaiinaracpheeyqoeedeodyqokaiinqotedaiinaporaiinapylsheodytaiin" would be the transcription of 崩中漏下赤白沃补虚温中止丹雄鸡, and so on.

Here is a graphical presentation of the data above:
[attachment=17361]

The numbers in the middle column are the counts of EVA letters in the last column, and the error of that count compared to the prediction based on the chinese character counts.  Thus, for example, the 63(+2) on the second lines says that the string"qoaiinaracpheeyqoeedeodyqokaiinqotedaiinaporaiinapylsheodytaiin" is 62 EVA letters, which is 2 letters more than expected given that 崩中漏下赤白沃补虚温中止丹雄鸡 is 12 characters, and on average the VMS text has a little over 5 EVA letters for each Chinese character.

In this entry, the eight 主 characters correspond to daiin, daiin, kaiin, dain, laiin, daiin, and kaiin.  This variation is assumed to be due to inconsistent transcription of spoken "Chinese" syllables by the Author and/or confusion by the Scribe when reading the Author's draft.  I have explained before how an EVA d can be mistaken by an EVA k or l in sloppy handwriting.

The mere fact that the cribs (and variants) occur in the SPS text is not remarkable.  What is remarkable is that the gap lengths -- the distances between the cribs -- in the two texts are so close to proportional.  And not just in this recipe, but in many others. 

In this recipe, the largest gap length discrepancy is before the first 主, where the SBJ recipe has 丹雄鸡 = "red male chicken" (3 characters) while the SPS has just "poarkeeo" (8 letters, 7 short of the predicted 3 x ~5 = 15 rounded down).  

The second-largest gap length discrepancy is between the 3rd and 4th 主, where the SBJ has  耳聋肠 or "耳聋。肠：" = "ear deafness. Intestines:" (3 characters)  and the SPS has "otaiinchedaiinolkallkl" (22 characters, 6 more than the expected 3 x 5 = 16 rounded up.)  Those errors are equivalent to about -1.4 and +1.3 Chinese characters.  

All the other gap length discrepancies are less than 5 EVA letters, that is, less than 1 Chinese character.

All the best, --stolfi

(24-08-2026, 04:38 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.And not just in this recipe, but in many others. 

Could you show us some other examples like that, then?

Maybe, if I manage to find a comparison with a conventional Tibetan herbal, will I be able to draw any conclusions? As I understand it, the Shennong Bencao is simply a guide to remedies, describing what needs to be done and what ingredients are involved.
Can we use the recognized hieroglyphic words in relation to Botany/Pharmacy?
And why are you only providing this recipe and not any others?

Ololololo, I think Jorge provided this one because I asked about it specifically. 

But yes I would like to see more correlations too, it seems to me that if this were correct, that the rest can be figured out through comparing the various possibilities. Like poarkeeo would seem to mean red male chicken... I think I have seen poar before in a spot where red might actually make sense... 

I just checked Voynichese and it looks like there are only 2 more instances of poar, one at the beginning of another starred paragraph, and I think also (Voynichese does not count it, I think because of scribbling, but it still looks like a P) at the beginning of what would be the Asia or water section of the inverted TO, depending on which style it is supposed to emulate. Red Sea? Qualifies both for Asia boundary, in the correct place, at the border of Africa and Asia, and also as water that is connected to a greater sea to the south. Does red come up again in the SBJ?

Also it would seem to me that if one was transcribing things, it would be in the order found in the original. Does it work out that the paragraphs before the red rooster paragraph also align, all the way back to the beginning of the recto of that page? (I think the red rooster paragraph is at the bottom of a verso, so going forward, or backward past the recto, is a little harder if the pages are out of order like they seem to be in several quires, but it seems like you could figure it out if the matches were found).

Just looking for previous info to remind me...

Quote:68712-07-2026, 05:54 AM (This post was last modified: 12-07-2026, 05:54 AM by Jorge_Stolfi.)
The evidence for SPS≈SBJ, specifically, includes the crib position matches (主 <-> daiin, 气 <-> Chedy, etc) in about 35 recipe-parag pairs (~50%) of 67 that I have tried to find.  Considering that 4 pages are missing and a few dozen SPS parags were merged by the Scribe into mega-parags with unknown boundaries, the success rate could not be more than ~67%.  Some of those 35 matches are somewhat strained or uncertain, especially when they have only one or two cribs, or the SPS keywords are "spelling variants" like kaiin or dain for daiin. Any one of these, by itself, could argued to be mere coincidence.  But others just cannot be coincidences.  The latter include the 7-crib match between the longest SBJ recipe ("red rooster") and the longest non-mashed-up SPS parag (f105v.32).

Quote:58812-06-2026, 11:19 PM (This post was last modified: 12-06-2026, 11:38 PM by Jorge_Stolfi.)
(12-06-2026, 10:18 PM)Linda Wrote:
What are [your cribs]?

For sure: 主 = "mainly for" --> daiin (but sometimes dain, kaiin, laiin, dair; or abbreviated as dam).

Almost sure: 气 = "vital energy"  --> chedy (but sometimes cheda, cheky, shedy, ...)

The former (zhǔ in modern Mandarin) almost always introduces the list of diseases that the drug is good for. The latter (qì in modern Mandarin) is a fundamental concept of Chinese medical theory: some sort of energy or impalpable fluid that IIUC)  carries the energy of food from the stomach and spleen to the lower abdomen and maybe other parts.

The other cribs are still to uncertain to list.  Please wait until I can confirm them...

...

Quote:I am now going through the ~60 recipes that I had already tried to match, updating their Chinese text to the ZHB version, and then looking again for matching parags.  I am about halfway through those 60. Then there are another ~300 that I have not tried yet...

How are your investigations going? You were saying that your copy was not matching, that was the same thing I had noted with the one I found, although perhaps i was not accounting for spelling variations as you seem to allow for, and i am afraid i don't know which one i was looking at, at the time. Which one are you using now? Do you have a link you can provide?

(24-08-2026, 03:09 PM)Linda Wrote: You are not allowed to view links. Register or Login to view.I would like to see more [SBJ-SPS] correlations too.

I am still hacking at the automatic plotting of the matches.  Please wait a bit more.

But responding to your query about "qo" on another thread: one of the reasons why I think that qo is a prefix meaning "and" is this SBJ recipe

青石赤石黄石白石黑石脂等味甘平主黄疸泄利肠澼脓血阴蚀下血赤白邪气痈肿疽痔恶疮头疡
疥瘙久服补髓益气肥健不饥轻身延年五石脂各随五色补五脏生山谷

which apparently is to be parsed as

• (A01) | 青石，　　　　　 | qīng shí,              | Blue stone,                           
  
• (A02) | 赤石，　　　　　 | chì shí,               | red stone,                           
  
• (A03) | 黄石，　　　　　 | huáng shí,             | yellow stone,                         
  
• (A04) | 白石，　　　　　 | bái shí,               | white stone,                         
  
• (A05) | 黑石脂，　　　　 | hēi shí zhī,           | black stone fat,                     
  
• (A06) | 等：　　　　　　| děng:                  | together:                             
  
• (A1)  | 　［味］甘，平。 | ··[wèi] gān, píng.     | ··[nature] sweet, neutral.           
  
• (A3)  | 　［主］　　　　  | ··[zhǔ]                | ··Mainly for]                         
  
• (A31) | 　　黄疸，　　　 | ····huáng dǎn,         | ····jaundice,                         
  
• (A32) | 　　泄利，　　　 | ····xiè lì,            | ····diarrhea and dysentery,           
  
• (A33) | 　　肠澼脓血，　 | ····cháng pì nóng xuè, | ····gut dysentery with pus and blood,'>
  
• (A34) | 　　阴蚀，　　　 | ····yīn shí,           | ····genital erosion/ulceration,       
  
• (A35) | 　　下血赤白，　 | ····xià xiè chì bái,   | ····blood and white pus in stool,     
  
• (A36) | 　　邪气，　　　 | ····xié qì,            | ····pathogenic energy,               
  
• (A37) | 　　痈肿，　　　 | ····yōng zhǒng,        | ····inflamed skin abscesses,         
  
• (A38) | 　　疽痔，　　　 | ····jū zhì,            | ····anorectal abscesses and fistulas, 
  
• (A39) | 　　恶疮，　　　 | ····è chuāng,          | ····festering sores,                 
  
• (A3a) | 　　头疡，　　　 | ····tóu yáng,          | ····scalp ulcers,                     
  
• (A3b) | 　　疥瘙。　　　 | ····jiè sào.           | ····scabies and itching.             
  
• (A4)  | 　［久服］　　　  | ··[jiǔ fú]             | ··[Prolonged consumption]             
  
• (A41) | 　　补髓，　　　 | ····bǔ suǐ,            | ····tonifies the marrow,             
  
• (A42) | 　　益气，　　　 | ····yì qì,             | ····boosts the vital eergy,           
  
• (A43) | 　　肥健，　　　 | ····féi jiàn,          | ····makes one robust and strong,     
  
• (A44) | 　　不饥，　　　 | ····bù jī,             | ····prevents hunger,                 
  
• (A45) | 　　轻身，　　　 | ····qīng shēn,         | ····lightens the body,               
  
• (A46) | 　　延年。　　　 | ····yán nián.          | ····prolongs life.                   
  
• (A5)  | 　（五石脂，　　  | ··(wǔ shí zhī,         | ··(The five clays,                   
  
• (A51) | 　　各随五色　　| ····gè suí wǔ sè       | ····each according to the five colors 
  
• (A52) | 　　补五脏。）　  | ····bǔ wǔ zàng.)       | ····tonify the five solid organs.)   
  
• (A6)  | 　［生］山谷。　  | ··[shēng] shān gǔ.     | ··[Provenance] mountain valleys.
  

In the SBJ, 石脂 "stone fat" means "clay". This entry is about clays of five colors, treated "together". The SBJ author factored out the 脂 "fat" from the first four clays, repeating only the 石 = "stone".  A free and more correct translation of the title of the entry is therefore

• "Blue clay, red clay, yellow clay, white clay, and black clay, together". 
  

It is possible that the matching SPS parag was lost or is in one of the blocks of several parags mashed-up without breaks.  Otherwise, my best candidate at the moment is  f104v.22:

• <f104v.22> <%>=ytchedy.qokchedy.qotchy.qokchedy.qokchd.lsaiin.?or.sheos.ytaiin.chey.tal=
  
• <f104v.23> =tchey.qokchey.qochey.qodeey.qodaiin.chodaiin.chckhdy.dairar.otar.qotain.ol=
  
• <f104v.24> =ykaiin.cheor.cheeey.daiiin.cheo,dal,aiin.chockhedy.chedaiin.otor.qokar,ary=
  
• <f104v.25> =dsheey.qoykeey.lchedy.qokedaiin.ar,chcthy.daiin.cheey.sair.ol.aiin.chedy=
  
• <f104v.26> =ysheor.cheey.qodaiin.chodar.chochs«<$>
  

Based on all the matches I have found, I am assuming that the fields (A1) [味] = "nature" (甘,平 = gān, píng = "sweet, neutral") and (A6) [生] = "provenance" (山谷 = "mountain valleys") were omitted by the VMS Author from all recipes.  The first only makes sense in the context of the theory of traditional Chinese medicine (TCM), and the second would be useless even in China, more so in Europe.  

In this recipe, I also believe that the VMS Author omitted also the comment at the end 五石脂各随五色补五脏 = "The five clays, each according to the five colors, tonify the five solid organs".  This comment too only makes sense in the theory of TCM, where the "five solid organs" are defined as the liver, heart, spleen, lungs, and kidneys; and each is associated with one of those five colors -- so that a remedy of a certain color would mostly act on the organ with the same assigned color. 

With the "nature" field omitted, in the SPS parag the title should be followed by the translation of the keyword 主 = "mainly for" (note the difference between 主 and 生!).  As seen in the "Red rooster" recipe and many others the "proper" translation of 主 seems to be daiin (daiin). However, the translation can often vary: in the "Rooster" case, four of the eight 主 match a "daiin", but the other four have "kaiin", "dain", "laiin", and "kaiin" in the corresponding positions.   I have discussed elsewhere why the Scribe could have misread a sloppy handwritten d in the draft as k or l.  On the other hand, not every "daiin" is a translation of a 主.

Anyway, on the f104v.22 parag there is a "daiin" on line 23 (part of the token "qodaiin"), which is rather too far to match the 主 of the SBJ entry.  But there is a "saiin" (in "lsaiin") at about the correct spot, and a "taiin" (in "ytaiin") just a bit further ahead. 

If we accept that "lsaiin" as the translation of the 主, then the title part of f104v.22 would be

• ytchedy.qokchedy.qotchy.qokchedy.qokchd
  

If we take the "ytain" instead, the title would be

• ytchedy.qokchedy.qotchy.qokchedy.qokchd.lsaiin.?or.sheos.
  

The "?" in "?or" is a weirdo glyph that could be transcribed as "qch" with the "q" ligated to the middle of the "c".  (I would guess that the token was actually "ychor", but it does not matter for this post.)

The SPS parag f104v.22 was identified as a match to the "Five clays" SBJ entry only by its total length and the relative position of my current cribs for  主, 气 (qi , sort of "vital energy"; twice in the recipe), and 久服 ("prolonged use").  There were no other parags that came close to it by those criteria.  The other parts of the Chinese and EVA texts, including the entry "titles", were not considered (except for their character/letter counts). While f104v.22 may still be a false match, it is notable that the uncommonly long presumed Voynichese title has five similar tokens, matching the five similar words of the SPS recipe -- four of which are prefixed with "qo".  

Hence my guess that "qo" means "and", and was inserted by the Author because he found the Chinese way of listing things -- with no "and" word -- too confusing.   Besides the title of that recipe, he presumably also inserted "qo"s in the lists of diseases and benefits. 

Apart from the "qo"s and the initial "y", the five tokens in the title part of f104v.22 would be "tchedy", "kchedy", "tchy", "kchedy", and "kchd".  I don't know how they would be parsed into colored clays.  iIt is problematic that "kchedy" occurs twice (where "red clay" and "white clay" should be).  But that could be a Scribal error too.

(By the way, the bifolio f104+f115 is quite distinct from all the others in a couple of ways, which could have a bearing on this analysis.  But that discussion belongs in another thread.) 

All the best, --stolfi

Well, but that's just another assumption—that he inserted the “and.” And why? Because he didn't know how to make simple lists in his language? I think that's a rather weak argument. It would be easier to say that “qo” is a comma.... But then you have to ask why he went to so much trouble to write it that way...

(02-09-2026, 06:34 AM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.Well, but that's just another assumption—that he inserted the “and.” And why? Because he didn't know how to make simple lists in his language?

Because in Chinese to say "red stone, beet, and sesame seed are wart drugs and spleen tonics" one would write just "red stone beet sesame seed wart drug spleen tonic".   

A native speaker used to read the classics would have been able to parse such phrases, and when speaking them he perhaps used pauses as hints.  Like "red stone, beet, sesame seed;  wart drug, spleen tonic"  But even of the Author had a working knowledge of the spoken language, and could get the sense from those pauses, he would find his written transcription as confusing as we do.

That would justify his adding of extra words to make his transcription easier to parse.

Quote:It would be easier to say that “qo” is a comma....

In Voynichese, the qo is written always as a word prefix; not as a separator, like the European period and comma were usually written.  

Quote:But then you have to ask why he went to so much trouble to write it that way...

If the VMS Author's native language was Arabic, he would have naturally followed the Arabic grammar, where "A, B, and C" is written "A waB waC".

Anyway, the "why" is secondary; what matters is whether "qo" indeed has that meaning, as that example seems to suggest.

All the best, --stolfi

(02-09-2026, 09:13 AM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.If the VMS Author's native language was Arabic, he would have naturally followed the Arabic grammar, where "A, B, and C" is written "A waB waC".

This does not explain the distribution of <qo> in quire 20, which is frequently paragraph initial and cannot be attaching to the right elements of serials. I realize it may be polysemic, but that must be accounted for.

But also, if The Author is inserting it, it is because he could already parse the list, so there is no problem to solve here. Parsing juxtaposition is a very basic piece of grammar in most languages, but Sinitic languages in particular, so under the hypothesis that this is a language which is grammatically identical to a language he could speak, why would he need to insert information he could understand on the fly?

(02-09-2026, 09:07 PM)ololololo Wrote: You are not allowed to view links. Register or Login to view.This experiment was inspired by Chinese theory, and in particular by Stolfi’s theses about the scribe’s mistake and writing by ear.  ... I tried to create such a “phonetic” script for the Ingush language. I used not only letters (but not all of them) as symbols, but also very common syllables in words in the text.

That's a very interesting experiment.  I should try a similar one: post an audio clip in some language and ask for a volunteer who does not know how to write that language to try to write down the sounds they hear, using whatever spelling system they chooses, and then see how consistent the result was.  Namely, if they always wrote the same word in the same way.

To better approximate the dictation scenario of the Chinese Theory, the volunteer ideally should be somehow familiar with the spoken language, even partly understand it, but they have no idea of its official spelling or even the nuances of its phonetics.  Maybe playing Portuguese to an Italian speaker, Dutch to an English speaker, Czech to a Polish speaker...

All the best, --stolfi

(03-09-2026, 04:36 PM)Jorge_Stolfi Wrote: You are not allowed to view links. Register or Login to view.I should try a similar one: post an audio clip in some language and ask for a volunteer who does not know how to write that language to try to write down the sounds they hear, using whatever spelling system they chooses, and then see how consistent the result was. Namely, if they always wrote the same word in the same way.

There is no limit to how garbled the text can get if the volunteer is unfamiliar with the language. I heard tourists, probably from Southeast Asia (my guess), talk yesterday and I thought: I have no idea how I would write (phonetically) what I just heard.

The hypothesis that a text is written in an awful semi-phonetic, inconsistent version of some Asian language is therefore unfalsifiable.