Voynich Manuscript: Comparing Notes

[milton63]

I am not sure if this thread violates the guideline of this forum or not, but I am starting this thread as an amateur enthusiast who has been interested in the research on the Voynich Manuscript.

Luckily, I was able to publish an article on Voynich Manuscript several year ago.
https://journal.equinoxpub.com/JRDS/article/view/9474
(I have to say that it is not free to download.)

I know that there are many enthusiasts like me out there trying to solve the secret of manuscript and reveal the messages hidden behind its writing.
In my article, I explained methods that I used to translate the first paragraph of Folio 2r.
The majority of VMS researchers are oriented to identify the matching language to the manuscript, as I understand.
On the contrary, the methods I used do not assume or need to know what language the manuscript was written in.

My intention of starting this thread is not to advocate or solicit anything, but hoping to introduce some new dynamics to this subject by asking people to consider it as a reference that they can use to compare their notes with.

Even though, to be honest, I can not claim that my methods are a solution, but comparing notes is simple.
As I used my methods to translate the first paragraph of Folio 2r, I would like to invite anyone to present his/her translation results of the first paragraph of Folio 2r and explain how they got there.
I am eager to know what ideas and methods others might come up with and what all of us could achieve out of collected wisdoms.

As a reference for others, the following is the translation result of the 1st paragraph of Folio 2r presented in the article. Differential abstraction method does not assume what language the Voynich Manuscript was written in and it can produce results with no need to know the language.

"The flower head is connected to the root through the middle stem and two branches and the plant grew three flower heads. The middle stem changes into a flower head. The flower grows at the end of the seed bump. The third leaf on the left branch splits into stems. The leaf breaks and transforms into a flower at the extremity of the plant, the seed breaks from the flower and falls down onto the ground by the plant. A leaf is divided into three stems. Each stem is divided into parts. A flower emerges on the top of a stem. That is how a flower grows. The flower head and its petal emerge at the end of the three stems. After the growth of a leaf has ended, then it comes to the end with the growth of flower. The emergence or transition ends after the leaf has turned into three branches. The second part or flower head of a stem comes next after the growth of the petal part ended. The flower head emerges or breaks from the rest on the top of its petal. The growing continues after the growth of the petal ended. The growth of flower starts from the top of its petal on the stem branched from a leaf. After the extension of stem branched from a leaf ended, the flower head and its petal at the end of stem from the extension of a branch starts branching three ways."

[Daniel]

Welcome to the forum. I don't know how much discussion on this topic will happen here, just because it's a niche interest. But you're welcome to ask!

I did happen upon this recent article that might be of interest, which seems to be a good summary of the situation from a linguistic perspective:
https://doi.org/10.1146/annurev-linguistics-011619-030613

(One other user here discussed this topic a while ago, but was either unwilling or unable to provide any translations, despite having claimed to have solved it in what seemed to be a relatively trivial way, so it was pretty clear that it was just guessing rather than systematic translation, and the linguistic terminology was off, e.g. "Proto-Romance" used in a way that didn't mean anything standard. [The linked article above briefly mentions that point but nothing else about that research.] Many linguists were, I think, surprised to see that "Proto-Romance" research actually published and wondered if there was any peer review at all, if for nothing else than the completely non-standard, bordering on meaningless, terminology. In the few years since it doesn't seem to have caught on or really made an impact, beyond a few casual news articles saying "it's been solved" and not diving deeper.)

[milton63]

Thank you for the info. Since my educational background is very remote from linguistics, sometimes it is hard to get used to the linguistic terminology, but I think the article is written so well to understand.

[FlatAssembler]

Welcome to the forum. I don't know how much discussion on this topic will happen here, just because it's a niche interest. But you're welcome to ask!

I did happen upon this recent article that might be of interest, which seems to be a good summary of the situation from a linguistic perspective:
https://doi.org/10.1146/annurev-linguistics-011619-030613

(One other user here discussed this topic a while ago, but was either unwilling or unable to provide any translations, despite having claimed to have solved it in what seemed to be a relatively trivial way, so it was pretty clear that it was just guessing rather than systematic translation, and the linguistic terminology was off, e.g. "Proto-Romance" used in a way that didn't mean anything standard. [The linked article above briefly mentions that point but nothing else about that research.] Many linguists were, I think, surprised to see that "Proto-Romance" research actually published and wondered if there was any peer review at all, if for nothing else than the completely non-standard, bordering on meaningless, terminology. In the few years since it doesn't seem to have caught on or really made an impact, beyond a few casual news articles saying "it's been solved" and not diving deeper.)

I am interested, what do you think about the Stephen Bax'es ideas about Voynich Manuscript? They seem rather compelling to me. Especially the idea that Voynichese word for Hellebore plant was "kaur": that turned out to be unexpected confirmation (as it is the actual name of Hellebore in some languages, such as Kashmir and Punjabi).
Until I was acquainted with the Bax'es work, I used to think Voynich Manuscript was a hoax produced using a table of prefixes and suffixes chosen randomly, which would indeed have many statistical properties of a natural language (but would, obviously, not be decypherable). But if some words can be so convincingly decyphered... then it is probably meaningful.

[milton63]

I am not familiar with Prof. Bax's works. We will never know what his next step would have been on this subject. Since I also don't have enough knowledges of related fields, I am afraid I can not make any comment on this. If Hellebore plant really is one of the drawings in the manuscript and it was truly identified as in "kaur", then in my opinion deciphering the Voynich text is like playing Hangman the word puzzle with the advantage of four known letters, 'k', 'a', 'u', and 'r'. Since there are plenty drawings in the manuscript with captions, I guess someone who knows Kashmir and/or Punjabi very well might give it a try.

[FlatAssembler]

I am not familiar with Prof. Bax's works. We will never know what his next step would have been on this subject. Since I also don't have enough knowledges of related fields, I am afraid I can not make any comment on this. If Hellebore plant really is one of the drawings in the manuscript and it was truly identified as in "kaur", then in my opinion deciphering the Voynich text is like playing Hangman the word puzzle with the advantage of four known letters, 'k', 'a', 'u', and 'r'. Since there are plenty drawings in the manuscript with captions, I guess someone who knows Kashmir and/or Punjabi very well might give it a try.

I would be very surprised if it were a language that exists today, such as Punjabi. In fact, I would be surprised if it was a natural language in the first place. The first-order entropy of the text fallls within normal range for natural languages, but the second-order entropy is significantly lower, even lower than for the Hawaiian language. It is much more likely to be an artificial language, don't you think?

[milton63]

I am not familiar with what kind of entropy measures are adopted in linguistics.
As long as the Voynich manuscript is concerned, in my personal opinion, it might need to be determined how to handle the text with the context modifiers before applying any statistical measures.
I mean the "context modifiers" by referring to those content bearing words introduced by Montemurro and Zanette(2013) from the University of Manchester.
I called those content bearing words as skeletons in my article and non-content bearing words as fleshes.
By definition, a skeleton is Voynich letter(s) that repeat themselves in the trailing text.
In the Voynich manuscript, those skeletons mostly are identified as prefixes.
Skeletons consists of one or more Voynich letters.
If Voynich words are categorized, they can be grouped into several categories.

Category 1. A word is consisted of skeleton and flesh.
Category 2. Skeletons. The skeleton words separated from the words in the Category 1.
Category 2a. Cascaded skeleton. With two consecutive words in the form of AB BC where B is a skeleton, B is called a cascaded skeleton.
Category 2b. Delimiting skeleton. When two words are used to comprise a text segment in the form of AB ~ A where A is a skeleton, A is called a delimiting skeleton.
Category 3. Fleshes. The flesh words separated from the words in the Category 1.
Category 4. Standalone flesh word. A word does not contain any skeleton in it.

In the viewpoint of the context change or the shift in topics, skeletons are regarded as context modifiers.
Context modifiers have their meaning and can be used standalone in the text.
When they are attached to fleshes, it does not make up new words, but just modifies the context of flesh words.

For example, I divided the f80r text, except those captions on the top, into 13 segments and the first segment S1 consists of 14 words.

[Category 1. The text in S1 in its original form (as it is, from the reference site)]

pdol fshedy q*polkain octhor okchdy qokeedy qo*cheol ol*oiin y darshey dykshy olotchedy qokai*n qotain

[The text in S1 in a formatted style. (I mean "formatted" by using line breaks and indentations. The purpose of formatting is to align skeletons at the same indentation position.)]

L1: pdol fshedy q*polkain octhor
L2:                       okchdy
L3:             qokeedy
L4:             qo*cheol ol*oiin y darshey
L5:                                dykshy
L6:                      olotchedy
L7:             qokai*n
L8:             qotain

At the first indentation depth, the skeleton 'qo' can be identified from L1, L3, L4, L7, L8. (I assumed that 'q*' in L1 would be 'qo'.)
At the second indentation depth, two skeletons are identified and they are 'o-' from L1 & 2 and 'ol' or 'olo' if 'ol*' is actually 'olo'. If not, the skeleton will be just 'ol'.
At the third indentation depth, one skeleton 'd' is identified.
I categorize the leading two words 'pdol fshedy' as a title to S1 and both words are flesh words without skeleton.
As it can be seen, the categorization of skeleton and flesh is purely theoretical, produced purely based on a pattern match.

[Category 2: Skeletons only, 4 in total in S1]

'qo', 'o', 'olo', 'd'

[Category 3: Fleshes attached to skeletons, 12 in total]

L1: polkain cthor (without 'qo' and 'o')
L2: kchdy (w/o 'o')
L3: keedy (w/o 'o')
L4: cheol oiin y arshey (w/o 'qo', 'olo', 'd')
L5: ykshy (w/o 'd')
L6: tchedy (w/o 'olo')
L7: kai*n (w/o 'qo')
L8: tain (w/o 'qo')

[Category 4: Fleshes with no skeletons, 2 in total]

L1: pdol fshedy

When the original Voynich words are grouped into Categories 2 to 4, the categorization will change the data set of input words.
The statistical characteristics of those words after categorization definitely are different from the statistical characteristics of original Voynich words.
It means that any entropy calculation taken on categorized words may produce different results than those taken on original words.