This morning I was reading the wiki about the You are not allowed to view links.
Register or
Login to view.. Like some other alphabets, it has interesting properties, like Voynich-resembling glyphs and a non-decimal system for writing numbers - but that is not the subject of this thread. It was the following sentence that made me wonder:
Quote:Examples of Palmyrene inscriptions were printed as far back as 1616 but accurate copies of Palmyrene/Greek bilingual inscriptions were not available until 1756. The Palmyrene alphabet was deciphered in the 1750s, literally overnight, by Abbé Jean-Jacques Barthélemy using these new, accurate copies of bilingual inscriptions.
If Voynichese is the result of enciphering in the strictest sense, it is just a matter of finding the key. Like "oh the vowels have just been left out and then you can convert it to Latin". This can be found, if studied long enough.
However, what if Voynichese is a historically developed writing system that is only known from this one source? For example, Byblos script is known from a dozen of inscriptions, and as yet it remains undeciphered. So if Voynichese is
not a "find the key" cipher - is it possible to ever understand it without external sources like a bilingual document?
Well, according to You are not allowed to view links.
Register or
Login to view., Abbé Barthélemy was able to decipher it by comparing direct examples of Palmyrene with Greek, which quickly led him to the key. So he did have a "Rosetta stone".
I would venture to suggest that a completely unknown language which has no direct correlation for us - no proper names, no obvious link to a known reference - would be impossible to translate.
We might be able to draw up its syntax and grammar, and even correlate that framework to our own (ngram x is a proposition, ngram y when a suffix is a verbal modifier, etc) but actually saying that word a equates to "Paris" and word b equates to "David" would be completely subjective and impossible to verify.
In parallel of research of documents containing the key, we can already check all alphabets proposals made to date. If I understood among amateurs of Vm there are people who have mastered programming. They could create a program that would take into account all the variants proposed of alphabets and then compare with the proposed languages. Already we would be fixed if it's from the Latin, Greek, Arabic, Slavic, from Nahuatl or not.
But maybe that's not a good idea, as this may break the dreams of some one go (I mean me).
David - Indeed, that is what I mean. It seems like in such cases, a bilingual document or inscription is always required
Linear B didn't require a bilingual inscription.
I don't have enough time to explain here, but I believe that, so long as we're smart and clear about what we're doing, figuring out the Voynich script should be possible.
I think it's important to make a distinction between whether it's the script that's unknown, the language that's unknown, or both.
In the case of Palmyrene, it sounds like the language itself was pretty close to other forms of Aramaic and the script was pretty similar to other Semitic scripts, so presumably it would have been cracked eventually even without a bilingual inscription. Basically it would be the same as solving as a simple substitution cipher.
In the case of Linear B, it was mainly the script that was a mystery - the language was just an early form of Greek. But because the script was a syllabary, the fact that it was written in Greek was not immediately obvious. A lot of work had to be done on internal comparisons to establish which glyphs shared the same consonant but had a different vowel and vice versa before the identification of proper names in the text allowed researchers to assign sound values to the glyphs and read the text as Greek.
In the case of the VMS, the problem seems to be that the language is simply one that has not left any other known record. The script is not completely understood either, but I think it's clear that it's an alphabetic script, and also certain things like the distinction between consonant and vowel seem mostly clear, and we could probably make reasonable guesses about some other things. It's true that we don't know exactly which sound values to assign to the letters. However, the structure of words and the rules governing how letters may be arranged in sequence to form words in the VMS are very distinctive. Because these rules do not depend on the sound values assigned to the glyphs, we can simply look for the presence or absence of the same or similar rules in a given language to determine whether or not it can be the language in which the VMS is written. There is certainly no known match, and I've personally checked many languages. It certainly can't be any "major" language though obviously not every language in the world has been tried. But my guess is that we don't have any other record of this language.
If it's the language that's unknown, then without a bilingual text I'm pretty sure it's never been done. It's difficult to imagine how it could be done even in theory. The best-known example where a language unrelated to any other has been deciphered is Sumerian, and that was done through bilingual texts. So the question then becomes whether or not the illustrations can function as a bi-text of sorts, allowing us to infer the meanings of individual words and entire blocks of text based on context. I'm guardedly optimistic that this is actually possible, and this is my main approach to trying to understand the text.
Yes, the You are not allowed to view links.
Register or
Login to view.
of Aramaic. These days any dialect should be quite easy to recognize.
The script is the bottle-neck indeed: if you understand how to transcribe the text, you are almost there.
Because we now can compare text with a computer it is simply a matter of comparison.
What you can do is to compare all possible transcriptions that you can think of with any known written transcribed language. That is what I am doing.
(01-10-2016, 09:49 AM)davidjackson Wrote: You are not allowed to view links. Register or Login to view....
I would venture to suggest that a completely unknown language which has no direct correlation for us - no proper names, no obvious link to a known reference - would be impossible to translate.
We might be able to draw up its syntax and grammar, and even correlate that framework to our own (ngram x is a proposition, ngram y when a suffix is a verbal modifier, etc) but actually saying that word a equates to "Paris" and word b equates to "David" would be completely subjective and impossible to verify.
I disagree with that. With hard work and serious research a lot is possible.
Based on solely the text, you can make a SVO or VSO model, which then can lead to a proper translation and the rough idea of the text can be "guessed" based on the correct approach. That is the same "typing/ language predictions" work.
As Sam G stated, there are three types of decipherments:
1) Known language in an unknown script (ex. Mayan Hieroglyphics written in Ch'olan language)
2) Unknown language in a known script (ex. Etruscan language written in Roman alphabet)
3) Unknown language in an unknown script (ex: Linear B tablets)
Right now, Voynichese is Type #3.
Type 1 is usually decipherable (think Mayan, Egyptian Hieroglyphics, etc.)
Type 2 is problematic but can be workable. For example, Etruscan is written in an alphabet similar to Latin, so we could "read" the words for centuries, we didn't know what many of them meant. There are two things we have done:
a) Use bilingual texts where Etruscan text is written alongside Greek, Latin, or Phoenician, and start to make educated guesses about what certain Etruscan words mean. Etruscan texts next to images could also be considered "bilingual". In addition:
b) Use the You are not allowed to view links.
Register or
Login to view. to analyze the content of many different inscriptions (especially bilingual ones), and piece together which words may be nouns or verbs, how grammar works, and ultimately what certain words mean. This is a slow and lengthy process but it has worked on Etruscan, aided mostly by bilingual inscriptions.
Type 3 texts have been deciphered (Linear B) but only if there is a large amount of writing in the unknown script. The length of the VMS would be sufficient, but in other unknown scripts where we only have writing on a few stone tablets, we can never solve the language. Usually "Type 3" decipherments have to start with statistics: how many signs are in the script, what combinations appear, etc. As Sam G said, we determined that Linear B was both a syllabary and an inflected language before we determined that it was Greek.
Nice overview, Thomas. I guess it would also make sense to further divide the third type:
3A: We don't know which language the inscription is in. But it will turn out to be a known language like Greek.
3B: We don't know which language the inscription is in. Additionally, this language does not exist anymore and has not been recorded anywhere else.
In the second case, I don't see how we could move much further than some deductions based on statistics.
There are two aspects which makes Voynichese even worse though.
1) We don't know what it is. If you find inscriptions on remains of buildings or stone tablets or what have you, you can generally assume with some certainty that it is a natural language. With Voynichese there's the whole "possibly hiding information" thing going on. The "cipher or not" question is not as resolved as one may believe.
2) We don't know where its from. We don't know its cultural background. It could have been made by immigrants, copied from whatever document... This is different than the Palmyra inscriptions or the Egyptian hieroglyphs.
Those are all good points, Koen. I wonder if a 3B situation would be solvable - I agree that it doesn't look good.
About the "cipher or not" question: Mary D'Imperio said that some "plants" have the leaves of one plant, the body of a second plant, and the roots of a third plant - so the entire herbal section may be a deception. That would make sense if Voynichese is a cipher (which it obviously could be).