+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Voynich Talk (https://www.voynich.ninja/forum-6.html)
+--- Thread: How many pages? (/thread-3666.html)
How many pages? - Pardis Motiee - 13-10-2021
Talking about Voynich manuscript, how many pages should be translated/decoded? And what features are required for a correct solution?
RE: How many pages? - Koen G - 13-10-2021
I don't think this can be expressed in a number of pages alone. A better way to put it is that the ideal method combines repeatability and scalability. If those requirements are met, even a single sentence is sufficient.
- Can the translator explain his system in a way that someone else can use it? If not, this probably means that there is too much interpretation going on. Many proposed translations fail this step.
- Can someone who has not read your translation obtain a similar result by using your method on the same sentence? Almost all proposed translations I can think of fail this step.
- Finally, can an independent researcher use your method to translate a different sentence from the same page? This has never happened so far.
So you see, it is not a matter of quantity. If one's system is flexible enough, it will be possible to "translate" many pages, but nobody will pay much attention to this creative writing exercise. However, if your method is independently repeatable and scalable, presenting a single sentence is enough.
RE: How many pages? - nablator - 13-10-2021
(13-10-2021, 04:26 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.So you see, it is not a matter of quantity.
Yes. Quality matters, but quality is subjective. If the method produces a word salad, it will not be clear to someone who is not proficient in the language(s) involved. This is why I don't comment on Persian, Turkish, etc. theories. Some amount of ambiguity and interpretation must be allowed in translation. All sorts of special pleading can then be invoked to allow almost anything: the author used words from several languages, words have a lot of undocumented variants, grammar is bad, etc. All possible, but where is the limit?
An important matter is to protect against cherry-picking and confirmation bias:
What does your theory predict, specifically, that can be checked objectively? (Like grammar or external validation of meaning) If the answer is nothing, your theory is too vague or not failible "not even wrong", it is not a strong theory.
What does your theory not predict that actually happens? If there is too much of that (like no explanation for any of the many known properties of Voynichese), your theory is incomplete at best.
RE: How many pages? - Koen G - 13-10-2021
(13-10-2021, 06:08 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.Yes. Quality matters, but quality is subjective. If the method produces a word salad, it will not be clear to someone who is not proficient in the language(s) involved. This is why I don't comment on Persian, Turkish, etc. theories. Some amount of ambiguity and interpretation must be allowed in translation. All sorts of special pleading can then be invoked to allow almost anything: the author used words from several languages, words have a lot of undocumented variants, grammar is bad, etc. All possible, but where is the limit?
I think the criteria I described account for this situation. A translation of Voynichese consists of two big steps: convert Voynichese to plaintext, and then translate the resulting plaintext to something a modern audience can understand (i.e. modern English). We (you and I) may not know any medieval Russian for example, but if someone were to propose a Russian theory, he could still explain to us how to convert Voynichese to the plaintext. You and I will then be unqualified to translate this plaintext to modern English, but we should be able to assess if Voynichese can be converted to a linguistically feasable plaintext using the proposed method.
What many proposed translations do is to circumvent part of this process by squeezing Voynichese towards existing vocabulary making use of the dreaded interpretative step. This is why their method is hard to explain to others and near impossible to repeat, and even more impossible to expand.
To put it differently: if your method actually works, I should be able to use it to convert Voynichese to a real language. Whether or not I am well versed in this language should not matter.
RE: How many pages? - nablator - 13-10-2021
Often the conversion step and interpretation steps are not separable, because several readings are allowed for the same Voynichese glyph, or spaces can be moved, or some redundant parts can be optionally ignored, if it helps interpret the result. This is all plausible to some extent, I'm not criticizing. The method consists in cherry-picking the best possible conversion among many.
Ideally there should be a way to assess the interpretation step by comparing translations made by different people. But several experts in some language can make nearly the same translation by using the same knowledge, the same dictionaries, and then there is often a reasonable way of interpreting a word salad that will result in nearly the same translation by several people (one that makes some sense) even if it is highly subjective. So it's possible to say "all linguists agree with my interpretation" in the end. Conversely differences in interpretations are perfectly normal for some ancient languages, even among experts.
RE: How many pages? - Koen G - 14-10-2021
I agree that translating an ancient text correctly can be difficult, in some cases impossible. A new Dutch translation of the Bible has just been released where they corrected thousands of issues. And how much has been written about Homer's colour terms.
When checking a Voynich solution though, there is a lot we should be able to confirm before this difficult step. For example I cannot translate medieval Greek, and even using dictionaries it would take me ages to translate a page directly. But before this difficult step, I must somehow convert Voynichese to Greek. And I am very capable of judging whether the result of the conversion is an actual language.
I cannot judge the accuracy of your translation, but I can certainly judge the viability of your conversion method. So far, nobody has been able to present a repeatable method that can convert Voynichese into anything that comes close to an existing language.
What happens is often, as you say, that the conversion step becomes inseparable from some type of interpretation step. Even when we don't master the language, we can deal with some ambiguity there. Does the system conflate a and o? Fine, I'll write @, and I will still be able to see if something like language comes out. But there must be a system, otherwise it is just a creative writing exercise.
RE: How many pages? - MarcoP - 14-10-2021
I am not sure this has been mentioned yet. Though I can't find the original post at the moment, I think it was Rene who pointed out that solvers should focus on the encoding process, rather than decoding. They should define a repeatable encoding process and encode a couple of short historical plain-texts in the source language; it is then possible to check that the results feature all the properties of Voynichese, e.g.:
- low character entropy
- zipfian distribution of words
- frequent repetition of identical words
- even more frequent consecutive similar words
- binomial distribution of dictionary word lengths
- constrained paragraph distributions for p f
- paragraph initial words with a preference for starting with t k p f
- constrained line distributions for m g
- words that mostly adhere to Stolfi's grammar
- a tendency for final y to be followed by the q prefix
- 'dialects' with largely different bigram distributions
- ...add you own favourite Voynicese properties
I think that doing this on two 500 word texts should be enough for an initial evaluation. This could possibly be equivalent to Nablator's point about the predictive power of a theory.
This method totally removes the problem of the meaningfulness of the source text. One encodes historical texts which are known to be meaningful and grammatical.
RE: How many pages? - zamolxe - 14-10-2021
(13-10-2021, 04:26 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.However, if your method is independently repeatable and scalable, presenting a single sentence is enough.
In VM case, with a relatively weird "letter" distribution, maybe a sentence will be not enough to cover all the glyphs. But it will represent for sure a good indication related to the encoding system and a huge step forward in the decoding process.
RE: How many pages? - nablator - 14-10-2021
(14-10-2021, 09:03 AM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.This method totally removes the problem of the meaningfulness of the source text. One encodes historical texts which are known to be meaningful and grammatical.
Yes. A good theory should be a "theory of everything". But strict adherence is not necessary, only the possibility to get similar results. We don't know whether unexplained statistical results are caused by an emerging property of the cipher/transformation/encoding/obfuscation/whatever or the underlying language or simply variable preferences.