The Voynich Ninja

The general rule people have here is that reading of one or even twenty words is not enough. People were in the past able to give readings of many words in many languages. Problems appear when you have to read longer passages. Up till now nobody was able to give reading of a longer text that would be gramatical and actually make sense.

(27-12-2025, 09:58 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.f68r3 is one of the pages with faint tails on many y and q: it happens more on some pages with circular/radial texts than elsewhere it seems.

Welcome to the Massive Retracing Theory (MRT).  There is You are not allowed to view links. Register or Login to view. about it, already at You are not allowed to view links. Register or Login to view.. (For some reason is it hanging under "Talk" than "Imagery" or whatever.) 

All the best, --stolfi

Across 3 pages now, I were not able to find the elephant in this room...
maybe that's my fault.

But you claim not to be a linguist, or a statistics guy, just plain a sociologist.
So I wonder why you start to translate into Latin and work with probabilities, counts and matches here:

(27-12-2025, 09:29 PM)MHTamdgidi_(Behrooz) Wrote: You are not allowed to view links. Register or Login to view.[..]
I now begin with the “word” beside the seven-star image on f68r3. As you know it is transcribed wrongly on the VM browser, the last letter looking like a ‘9’ not an ‘o’. I render it in my preferred symbols that retain a sense of their visual appearance, 8oa29. It seems there has been a consensus that the image beside it represents Pleiades (the seven stars).

Botrus in Latin (and Greek) means cluster (of grapes, specifically). The 9-looking letter at the end as you know was a standard Latin abbreviation for -us (and alike). So the 8-looking letter is just B, the ‘o’ is just an ‘o’, and the letter looking like 2 is just an ‘r’ in the word, ending with ‘9’ for -us.
If this is what was meant by the “author” we have a clear indication of both a Latin writing present in abbreviations prevalent at that time. But such a reading, that perfectly matches the image, also hints at a ciphering effort. 
[..]
 Of the five letters, four perfectly correspond to Botrus, and the image is a perfect match. In my view, that word is meant to stand for Botrus, and can serve as a key to deciphering a lot in the manuscript, in text and in images. 

Problem with the flood of AI "solutions" is that these AI tools are used to "proove" a wrong guess, feed such a theory with masses of follow-up wrong guesses and produce lots of wellsounding sentences to "explain" this "solution".

Quote:Stephen Bax had considered the word to be (in my view, wrongly) Taurus, and others have either followed his lead, or rejected it. In my reading it is not Taurus, but Botrus.[..]

Yes, Bax made quite surely a wrong guess with Taurus.
You are overbidding him with another wrong guess, but thankfully do not try to base that with any AI solutions -- thanks for that.
Problem is: your guess is still unfounded and just based upon a feeling about these "stars".
Even if these are really images of grouped astronomical stars, commonly, but not everywhere known as Pleiades, there is no hint that VMS author(s) had anything in mind like
- noting their given names within the few label words
- writing down any links between these stars or giving a connection to the 3 label words

= this could mean anything, from "Pleiades" or any other specific cultural description for these stars, like "the sisters", "the seven", "the bright group" or even, "the grapes". 
Could mean any of this, or just nothing.

To me, the directly accompanying 2 words read like "kapelkaya"  and "keasa"/"keasya". I have reasons to be a bit more sure about this, but single words do not help very much at all with VMS.

Quote:Perhaps it is the first word we can safely claim to have translated in the VM. Is it?

Steep claim.
Plain answer: no.

Quote:I am not yet claiming anything about any other texts in the VM. I am focusing just on this word for now, and it hints at several findings. It is in Latin, uses standard abbreviations of its time, some letters are visually readable in a standard way (B and o). ‘9’ is the most widely used abbreviation in Medieval Latin, and 2 is a standard abbreviation for ‘r’ or phonetic ‘ur” [..]

Well, and there it is, the only elephant in the room: it is not Latin.

You will find some ~20 "solutions" in the graveyard of non-solutions ("the list") and several failed Greek theories, which all just give proof that it is not Latin/Greek at all.
(By the way, as much as "9" anka abbreviation -us appears in the manuscript, this would mean most of those words are nominatives -- you will find not many Latin texts were the load of nominatives is such overwhelming)

Overlooking and ignoring this is the true `elephanting´ with Voynichese...

@ nablator, yes those are good examples, and I am aware that some have not been as legible as others. I was just focusing in my last post on that word only. In any case, if something is found to be worthy of attention, it is likely to start with just a word or two.

@ oshfdk, you have a good point that in the picture you copied, that could be a ‘t’ but that particular image does not look like an ‘a’ to me, and in any case that is not the “word” I was referring to. In the text near Pleiades stars, it seems convincingly an ‘a’ or what looks like it. I really don’t know how to explain that seeming exception to the rest of word matching Botrus and the adjacent image. I just wanted to bring this to all’s attention, as I had not seen others suggest that particular reading of the “word”.

Please note (and this is addressed to all, so to Rafal’s reply as well), I never claimed and I am not claiming that what is happening in that word is anything to be generalized (yet) as a cipher rule. I am just saying that for that one, single, word, there is nearly a 90% perfect matching of a Latin word (with standard abbreviations) with the image near it, and the exception of the ‘a’ may be hinting at something for your and other expertise to consider. That is all I was saying.

It may turned to be an extremely rare 90% coincidence, but for me it is very interesting.

@ Rafal, beside the point above, I wish to ask you and others, ReneZ as well, a question that has come up in my mind when reading your posts in rejecting some solution lacking grammatical expectation, i.e., being a “salad.” While this may make sense for a text we may suspect is not ciphered at all, why a “salad” itself could not be a ciphering method?

A poem, for example, may not meet grammar expectations, but a choice of words can convey a meaning, and even the multiple meaning may be intentional sometimes, as in metaphors, or as not giving away the author’s intent too easily. Why have we taken for granted that the author had to follow a grammatical order, rather than sharing key words to convey a meaning? Can a “word salad” be a way of keeping a writing private, inventing one’s own abbreviation rules and symbols, especially the text was intended as a personal notebook?

@ Jorge_Stofli, yes I am aware of your notes about the evidence of retracing, and it is very much possible they may have been retraced. The idea that sometimes re-inking the pen could introduce variations, as other researchers have point out, may explain some of them. I am not really sure. By the way, I am inclined to agree with you on the fruitfulness of an “origin theory” entry into solving the VM puzzle, as one of the possible venues, as I heard you present your talk on the panel. But, that approach itself can take a variety of forms.

@ Stefan Wirtz_2, of course you are entitled to your opinion, and if you don’t find value in what I wrote, you can just ignore it. I wish you the best in finding better ways of understanding that word, and I will be the first to congratulate you if you do so. Regarding your first point, it does not take a linguist or statistician (or even a sociologist) to notice a 90% matching of a word with a Latin translation and adjacent image. That’s all I was trying to say.

With apologies to all, I am not quoting passages from replies, in the interest of not making this longer.

(28-12-2025, 01:08 AM)MHTamdgidi_(Behrooz) Wrote: You are not allowed to view links. Register or Login to view.... a question that has come up in my mind when reading your posts in rejecting some solution lacking grammatical expectation, i.e., being a “salad.” While this may make sense for a text we may suspect is not ciphered at all, why a “salad” itself could not be a ciphering method?

A poem, for example, may not meet grammar expectations, but a choice of words can convey a meaning, and even the multiple meaning may be intentional sometimes, as in metaphors, or as not giving away the author’s intent too easily. Why have we taken for granted that the author had to follow a grammatical order, rather than sharing key words to convey a meaning? Can a “word salad” be a way of keeping a writing private, inventing one’s own abbreviation rules and symbols, especially the text was intended as a personal notebook?

I'm not sure "word salad" is only about the grammar. I understand "word salad" as an incoherent sequence of words that requires some elaborate interpretation to convert it to an at least somewhat meaningful phrase. "See red star sunset Vesper blood sign bad omen" is agrammatical, but I wouldn't call it "word salad", more like "word chaining". On the other hand, "star water bring grass tuna poem larva thread blood giraffe" is a word salad.

Also, I don't think there are any prohibited assumptions. We cannot interrogate the creator of the manuscript and get a definite answer as to what was the idea and the method by which the manuscript was created. So, the only reasonable way to evaluate a theory is by how well it is accepted by other researchers. There are no prohibited assumptions, it may be possible to produce a good theory with agrammatical text, but then it should be very persuasive in some other aspect and probably provide a good explanation of why there is no proper grammar in the manuscript.

@ oshfdk, yes I agree with you regarding the distinction you are making, and that we cannot (yet) have prohibited assumptions re. the VM text.

But the problem is how to tell if a text is of one or another type if each word has multiple meanings, especially as found in poetry contexts when use of metaphors and double meanings are expected.

Let’s say we have a statement A B C D E, where each has a variety of meanings and connotations corresponding to it, as 1-5, from most apparent to more hidden. The thread A1-B3-C2-D5-E4 may sound incoherent and readily dismissable, but the thread A3-B1-C4-D2-E5 may be meaningful.

It will take an effort to figure it out, of course, but then that’s the point I was making. A grammatically “free licensed” poem or poetic narrative can provide an effective way of hiding hidden meanings.

Regarding your previous reply, sorry that I missed the last line. Yes, I would be interested in seeing your set of labels which may come of use later. At this point, I was not claiming anything beyond that one word, just a larva thread for the elephant’s silk!

(29-12-2025, 05:38 AM)MHTamdgidi_(Behrooz) Wrote: You are not allowed to view links. Register or Login to view.But the problem is how to tell if a text is of one or another type if each word has multiple meanings, especially as found in poetry contexts when use of metaphors and double meanings are expected.

Let’s say we have a statement A B C D E, where each has a variety of meanings and connotations corresponding to it, as 1-5, from most apparent to more hidden. The thread A1-B3-C2-D5-E4 may sound incoherent and readily dismissable, but the thread A3-B1-C4-D2-E5 may be meaningful.

After a theory has enough substance it's possible to use simple mathematical methods to estimate how likely it is to create a reproducible (unique well defined) solution. There are many people on the forum who are familiar with Bayesian analysis (or whatever it's called) who probably can compute some scores in science friendly way. I can compute some simple naive probabilities, they are good enough for me.

For example, let's take doary = botrus match. How likely is it to be accidental? The below is not an exact calculation, but more of a template. If we consider all the labels in the manuscript and assume Greek or Latin source:

1) I think there are at least 30 words in Latin (or Greek) that one would consider semantically relevant for this label (used to describe the Pleiades), words like: group, pile, seven, sister, gathering, star, proper names like Taurus, Orion, Pleiades, etc. So let's assume for an average label there are about 30 candidates that one would consider meaningful.
2) There is one letter mismatch, if we allow a one letter mistake and match all other letters reasonably by shape, with the alphabet size of 20 characters there will be 19*5 = 95 possible different readings for a five letter label. If we also use scribal abbreviations, this will expand the set of readings even further, let's say 400 combinations.
3) Now we need to get the total size of the corpus, suppose we limit ourselves to 20000 most common words. Given 400 sequences, how many of these would end up among valid words for the language? We can try using entropy bits per character or run a simulation on this, so this one need some further research, but I don't think it's unreasonable to assume that 400 letter sequences will on average produce 10 valid words (given that o and a look like vowels and appear in most words, this doesn't look very unreasonable, most sequences are quite word-like).
4) Given the corpus of 20000 words of which 30 words would be semantically relevant if we select N words at random, how likely are we to get one of the relevant words? To simplify a bit, we have the probability of 30/20000 or 0.0015 at each try. We perform 10 tries per label in the manuscript. There are at least 40 labels attached to images, where one can claim semantical relevance. So, we have 400 tries in total:

1 - (1 - 30/20000) ** 400 = ~0.45

So, this back of the envelope computation gives 45% probability of a single coincidental match as plausible as doary = botrus somewhere in the manuscript.

It's most certain that this number can be challenged, but in general a spurious match for one label is certainly not impossible. However, if we add more and more labels matched according to the same principle, the probability of a chance match will drop dramatically. For example, if you demonstrated that your method produces 5 plausible matches for 5 different labels, for me personally this would move your result significance from "likely curious coincidence" to something that certainly requires a proper explanation. As it stands now doary = botrus on its own doesn't require any explanation at all for me, pure coincidence works just fine.

For any proposed deciphering method, including your A3-B1-C4-D2-E5 example above, one can roughly compute the likelihood in a similar fashion.

(29-12-2025, 05:38 AM)MHTamdgidi_(Behrooz) Wrote: You are not allowed to view links. Register or Login to view.Regarding your previous reply, sorry that I missed the last line. Yes, I would be interested in seeing your set of labels which may come of use later.

Here they are:

[attachment=13205]

@ oshfdk, thanks a lot for sending the labels and offering your thoughts on how to evaluate probabilities. The main issue I see with the labels you provided is that in these cases we really don't know yet what the images are about, so they can't function in the same way Pleiades did in our example. But, I see your point that if any textual findings are to be tested, it can be tested in a reverse way on them at some point.

I will study them and your testing procedure suggestions and will get back to you later. Like your earlier suggestion about the 'a' looking like 't' at times, even though it may not work for this word, and what you also sent may help more or less, your spirit of trying to help out is most appreciated and it is this kind of spirit widely spread on this forum that will eventually solve the puzzle, I hope. I sincerely believe it will emerge not by way of competing separate theories, but by efforts that see and integrate grains of truth in all the past and newly offered theories.

(29-12-2025, 02:56 PM)MHTamdgidi_(Behrooz) Wrote: You are not allowed to view links. Register or Login to view.The main issue I see with the labels you provided is that in these cases we really don't know yet what the images are about, so they can't function in the same way Pleiades did in our example.

I think this is not an issue with the labels, but with the method. If there truly was a method to reliably read Voynichese, I think there would be no problem to find out what the images are about. If the very fact that the images are vague makes it impossible to apply the method, this hints that the method is more about guessing than reading/decoding/deciphering.

The images for other labels show things that look like clouds, tubes, rainbows, liquids, flows, vapors, pools, droplets. They could mean many things, but if the images and the text are related at all, after reading the text the images should make sense.

About labels that appear to match some known word...

I know EVA, Serbian and French so this perfect phonetic match between koreni (roots) and korainy appeals to me, a label in front of several roots: You are not allowed to view links. Register or Login to view.

What are the odds of such a coincidence? Very good.

If you don't believe in coincidences you end up with a Cheshiresque multilingual mess or a Bax-like nothing-burger.