Historically, attempts to intepret Voynich have been aimed at discovering the procedure with which one could map Voynichese glyph sequences to meaningful letter sequences in one of existing languages. This approach still prevails up to date. Simple substitution (with slight variations) attempts constitute the vast majority of the proposed solutions, and we are accustomed to hear of new solutions on a more or less regular basis.
Considering Voynich a more complex cipher basically falls into the same broad vein of investigation, - alas, with no success so far.
Of lately, I've been thinking if such approach is efficient after all. The issue is that there are some indications that the Voynichese text, while conveying pretty meaningful message, may not be what we are believing it to be.
One such indication follows from the work by Wladimir which suggests that no plant names are contained within botanical folios. The imagery which is manifesting its mnemonics supports this thesis. In a (supposedly enciphered) text, what reason would there be to exclude plant names? Nothing. The situation is quite different for the representation that relies on a nomenclator. If your nomenclator does not contain plant names, you won't be able to include them.
Another strange thing is the high degree of morphological similarity between vords being members of homogenous sets - such as my favourite "Voynich stars" (f68r1, r2). Of 53 Voynich star labels in total, 39 (or 74%) start with "o". Of those 39, 15 (or 28% of the total) start with "ot", and 9 (or 17% of the total) start with "ok". Those two subsets constitute 45% of all Voynich stars. In other words, notions homogenous in nature are designated by vords similar in morphology. This does not very much look like what we find in natural languages. This could be explained, however, by vords encoding positions in a nomenclator. Homogenous notions may have been grouped in a nomenclator. Encodings of their positions (close to each other) would then appear morphologically similar.
If there is no mapping between Voynichese and plain text on the glyph level but, instead, mapping exists only on the vord-to-word level, then all attempts at "deciphering" would be vain. What one should do instead is to shift from "decrypting" to "translating". Suppose extraterrestrials land and we are presented wtih their writings. We would not try to invent a procedure to decrypt their writings into English or Russian, that would be waste of time. We would seek a way to translate those instead, based on our understanding of what words of theirs map to what notions known to us. This is the direction that might prove fruitful for Voynichese. The problem is with the methodology, as always...
Hi Anton,
the perspective you suggest can certainly provide a different angle. We do have a number of hints, e.g. the words that occur as labels are more likely to be nouns.
A couple of years ago, David suggested You are not allowed to view links.
Register or
Login to view.: a task that could be a subset of your proposal.
In any case, exact reduplication of words does not disappear even if one considers a nomenclator. And then You are not allowed to view links.
Register or
Login to view.n often appears as similar to You are not allowed to view links.
Register or
Login to view. (in particular, the q- prefix is frequent in both). Also, there are elements that suggest that the ending of a word has an influence on the prefix of the following word (see You are not allowed to view links.
Register or
Login to view.). For these reasons, I find it difficult to think of a purely-nomenclator solution. But thinking of ways to approach translation more directly seems useful to me: similarly to David's idea, a first step could be identifying groups of words that behave similarly (there are several algorithms that do this kind of stuff).
Hi Marco,
Quote:In any case, exact reduplication of words does not disappear even if one considers a nomenclator.
Yes, that's a good point, I've been thinking of that. One possible case that I can bring forward when repetitive sequences could occur (in a "nomenclator solution") is when one is in need of composite numerals, such as e.g. 333. In the absence of such large numerical in a nomenclator, one has two options to represent that. The first one is to say "three hundred and thirty three", the other one is say "three three three". The latter option would produce a repetitive sequence. Not that I can point the exact sequence in the VMS where I believe this be the case, but this is simply for the sake of argument. Other potential cases may probably be imagined as well.
I would not stricly advocate a "purely nomenclator" solution; with 8114 (as per voynich.nu) different vords comprising the vocabulary of Voynichese, it's quite an open question whether it would have been both efficient and possible to base that entirely on a nomenclator. That's quite a number; for example, Shakespeare's vocabulary embraces (as I just googled out) roughly 12000 words.
What I'm bringing forward is, in the first place, this "word-based" paradigm, I wonder if an unknown text can be efficiently approached from this viewpoint, given a) the extensive imagery, b) the statistical tools that are at hand for modern researchers and c) our rough understanding of the contexts in which the vords such as labels do occur. E.g. if labels occur near stars, then they would designate either star names or some associations with stars - like e.g. stones governed by stars etc. This paradigm is solution-invariant (with the meaningless text hypothesis set aside, of course).
(17-03-2019, 06:24 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.I would not stricly advocate a "purely nomenclator" solution; with 8114 (as per voynich.nu) different vords comprising the vocabulary of Voynichese, it's quite an open question whether it would have been both efficient and possible to base that entirely on a nomenclator. That's quite a number; for example, Shakespeare's vocabulary embraces (as I just googled out) roughly 12000 words.
We need to be careful and precise about things such as the "number of words in Shakespeare's vocabulary". I imagine that in that count of 12,000 words, they did not include the numerous spelling variations that actually appear in all of the quartos, folios, etc., of all of the texts of Shakespeare's plays. If they did include all of the spelling variations, I bet the number would be much higher than 12,000. And that was 150 years *after* the invention of the printing press! I can only imagine that spelling variations were even far greater in manuscripts prior to the invention of the printing press, such as at the time of the Voynich ms.
Good point. Of course that's without spelling variations, and, more important, without inflexions.
But one who is heading to encipher his opus by means of a nomenclator would, most certainly, exclude spelling variations (you have one word form in a nomenclator, and you consistently refer to it), and, most probably, minimize inflexions. E.g., there may be different tenses (so as to distinguish "are" vs "were" and "will"), but probably one would not be in need of genders. But I'm speculating, I don't know if there are examples of large texts enciphered solely with a nomenclator.
It is of great interest whether there are inflexions in the VMS or there are none. I wonder if any method can be proposed to investigate that.
(17-03-2019, 11:29 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.It is of great interest whether there are inflexions in the VMS or there are none. I wonder if any method can be proposed to investigate that.
I strongly suspect that there are. For example, the suffix [-in] is very common in the ms, but it is almost entirely absent in any of the labels in the ms. This suggests to me that the suffix [-in] represents some type of inflexion that does not tend to occur in simple nominative case noun forms that tend to occur in labels of names of things.
At a glance, f68r1 has "or daiin" star, and f68r2 has "odaiin" star. But that's interesting... although purely morphological guesses will leave us with too much uncertainty. E.g., one could think of "in" as a tense selector, say, to turn a verb into the past tense, the "in" is appended. That would automatically mean that all vords ending with "in" are verbs. That's just an illustration, "odaiin" most certainly precludes that. ("odaiin" is the second most frequent Voynich star, not some outsider).
I'm thinking of some broader tests. One that I can think of is collecting a pool of texts and calculating the ratio of "inflections vocabulary size" to "vocabulary size". The former shall count all word forms, the latter - just unique words, flexion-invariant. This ratio would be greater than unity. Given a collection of such ratios for texts of supposedly comparable time and topics (herbals, medical books, astrology), we could then calculate the supposed vocabulary of the VMS in the assumption that the text is inflected. And see if the resulting figure looks reasonable.
Calculating the "inflexions vocabulary" of known texts and comparing that with the 8114 figure of Voynich would be the other way around, only we need to deal with normalization, because the overall volume of different texts differs.
By the way, about the "verb test". What about beginnings of paragraphs? E.g. in Russian it is not very common to have a verb starting a sentence in a narrative, unless in poetic or colloquial, and also unless the sentence is not interrogative or imperative. Although there are cornercases. In English, same thing, I think - unless to express subjunctive (like "Were I a boy,..."), a verb is not common in the beginning of a sentence, is it? What about other languages?
(17-03-2019, 11:34 PM)geoffreycaveney Wrote: You are not allowed to view links. Register or Login to view. (17-03-2019, 11:29 PM)Anton Wrote: You are not allowed to view links. Register or Login to view.It is of great interest whether there are inflexions in the VMS or there are none. I wonder if any method can be proposed to investigate that.
I strongly suspect that there are. For example, the suffix [-in] is very common in the ms, but it is almost entirely absent in any of the labels in the ms. This suggests to me that the suffix [-in] represents some type of inflexion that does not tend to occur in simple nominative case noun forms that tend to occur in labels of names of things.
That might seem like good logic if the labels are names, nouns, but if they were I suspect someone would have figured out at least a few of them, they've probably been studied more intensely than any other part of the text. I haven't been able to resolve them into names, at least not in a consistent way that works for a large portion of them.
I know the assumption is almost always that they are probably nouns, and being a fan of Occam's razor, it's the first thing I tried, years ago, but I really don't see why they
have to be nouns. They could be instructions (verbs like pour, freeze, steep, bathe, crush, dry, etc.), or they could be quantities or references. A reference doesn't necessarily have to match anything in the text (be repeated), it could be a location in the text (e.g., folio 5 section 2). There are quite a few ways references could be formatted. If you look at Greek herbal manuscripts, they are often annotated this way, with [cross-]reference text next to each drawing.
It's also possible that VMS "labels" are part of a longer narrative, that they stretch over a section of images (more like a phrase or sentence rather than a list of nouns).
I try not to assume they are nouns, it's only one of a number of possibilities.