The Voynich Ninja

I looked through the paper quickly, tomorrow I'll try to read more carefully. But the first impression is that our works are similar in many interpretations of the glyphs,  and, I think, errors or oddities of Paul's works are also similar to my ones, while the interpretation differs.

@Paul
Thanks, in German really makes it much easier for me.

My suggestion, if you want a full discussion, is that you do a step by step translation that is open to comment by the Ninja community as you go through the process. If you can do that, then there is an interesting possibility that you might have something.

Having translated the middle band, the other example of White Aries text designated by the paired Stolfi "Start Here" makers, should give you a leg up.

I haven't read your paper and probably won't. There are many others, more eager and far better suited to comment on it than I am.

In over 10 years of interest in the VMs, I have seen a host of such proposals come and go, go, go.

In fact, I have come up with a deeply encrypted, vernacular code word to describe this phenomenon: QUOILS

On the one hand I feel like I should read this paper, because everyone deserves a hearing. On the other I would much rather spend time on my own Voynich research than reading the seemingly endless stream of theories. But how can I then expect others to read my theory, when or if I present one, when I can't be bothered to read theirs?

I have produced my own translation test, so that might help me be persuaded if someone can pass it.

However in general I find the problem is that people tend to provide scraps of translation and not the full translation of the text. I think someone who claims to have produced a translation of a large proportion of the manuscript is more likely to attract my interest. Also someone who can answer interesting questions about the manuscript from their translation is also likely to intrigue me; the answers to these questions hopefully then can be examined.

---

One thing which would be unlikely, but very persuasive is if two people independently produce the same decipherment; that does mean it is correct, but would certainly make me wonder if it is.

Paul,  you devote a lot of space to explaining how Latin has "general ambiguity" and how it can be difficult to translate certain words. This might be true for someone unfamiliar with Latin, but it is not a problem for those who know Latin. There are many Latin scholars.

Personally I don't find Latin to be especially ambiguous. Try Japanese. It is very nuanced and much of it is understood through knowing the culture rather than it being said outright. A Japanese-born Japanese teacher said to me, "Well, this could mean three things... it just depends... it could mean... [insert different interpretations]." This happens quite frequently with Japanese. Much more so than with Latin. And yet people still manage to translate Japanese texts.

As long as you have a big enough corpus, you don't have to get every word. You just have to get enough of the sentence to work out the word.


.
It's confusing for readers when you mention Capricorn as one of the "respective star signs" in the context of a "series of 12 drawings" and then a few sentences later state that Capricorn is missing. It would be better to list respective star signs that actually exist in the manuscript.

The month name is not "roughly decipherable" as "abiril". It says "abiril" in 15th-century Gothic script. And saying it "could mean April" is an odd way to put it, as though you are unsure. It does translate as April. There is no ambiguity about the month names. They are abbreviated Gothic script with spellings that are specific to a particular region.


As for the argument that Frederick III might be involved, I'm not going to argue against that. I have a list of "persons of interest" with regard to the VMS (about 35 names, but some are higher on the list than others) and he is one of the people on the list, for the very reasons you mentioned... he was interested in astrology, plants, and ciphers.

You need to read the information on binding on Rene Zandbergen's site. What you have written is speculative and does not fit the facts. The rebinding happened later. A number of the manuscripts that had been bound within boards were suffering from worm infestations and were rebound.


.
For those who have not read Paul's paper, it essentially says that the positional characteristics of the VMS text can be overcome by treating the "o" character as a wildcard vowel, by assigning two or more letters of the alphabet to each Voynich character, and by treating it as abbreviated Latin.

You devoted about 50 pages to the same Latin abbreviation conventions that I outlined in my blogs between 2013 and 2019, so the information about these conventions is already out there. The main difference is that you depend mostly on Cappelli, while I have posted examples from dozens of different manuscripts so that readers can see how these abbreviations were actually used in the context of medieval writings.

On page 107, you note that EVA-a and EVA-e, which you transliterate as "a" and "e", can be in any position in a Voynich "word". This does happen, but a and e are not very common in the end-position in the VMS, whereas they are very commonly at the end in Latin (often several times per line). This creates a big discrepancy in letter position between your proposed translation-system and actual Latin texts.


You have written about the letter o: "We always encounter this written character at the beginning of a word only in very short words."

This is simply not true. The "o" character is also at the beginning of longer words like otaldam, otchody, otolom, okolShol, olSheol, okechoked, otchorol otardam olchckhy, olchdaiin, oochockhy, Shocthol, and others.


You list an interpretation for d at the start or middle of a word, but it is also quite frequently at the ends of words, in both Currier A and Currier B (several hundred times).

You have tried to debunk the analogy with Roman numerals, but those of us who have pointed out the similarity of some of the sequences were not specifically saying they ARE Roman numerals. In my case, at least, I was pointing out a positional prioritization similar to Roman numerals. It can be the same concept and still be interpreted in a way specific to the VMS.

As a counter-interpretation, you proposed that eee is an abbreviation for "esse", but you don't explain 1) why this would be so frequent in the middle of a word, 2) why it is so often preceded by "o" which you say is a wildcard vowel (as in oeeeos, oeees, oeeesoy, oeeey, etc.), and 3) how we are to interpret eeee when there are four in a row (you suggest "sloppiness" as a possible explanation, but I do not think this is the case, I think it is deliberate, just as an, ain, aiin, aiiin sequences are deliberate).

Regarding the r character, you write, 

Quote:"If the first stroke before the loop is bent strongly to the left, then the suffix -cis is present."

This is certainly not how it works in Latin.

It is the shape of the stem that distinguishes between -cis and -ris (not the angle of the stem). Or are you saying the stem-angle distinction is specific to the VMS? If so you will have to give examples because I am skeptical. The VMS already makes a distinction based on shape. It includes both g and m. I notice that g is not included in your transliteration tables.

In your example translation for token number 2 (page 123), you expand the Latin abbreviation y into "con"... but then you further expand the transliterated word "contam" into percontari or contaminare because "contam" is not actually a word. You have done this with other words, as well.

So, if a transliterated word in Voynichese does not match a Latin word, are you saying that it's okay to subjectively expand it further until it becomes a Latin word?


This, in addition to the numerous transliteration options for each of the gallows characters gives you a lot of degrees of freedom. Even with all this freedom, it's not really Latin...


On folio 1r, sentence 3, your translation sororem/soter cicutas oret/orexis tantum/tent humanum careat animae is translated as Sister/rescuer The Hemlocks requests desires merely to try a human to dispense/be free from spirit/soul.

This is not how it would be expressed in Latin. It is missing common key words like "ad" "et" and "ex", along with appropriate grammatical endings.

Hi Paul,
I only had a short look at part 3, so I might have missed something important (but I doubt this is the  case).
You might want to check the recent paper by Lisa Fagin Davis "What will it take to solve the Voynich manuscript?" (see You are not allowed to view links. Register or Login to view.).


These are the three criteria for evaluating solutions she discusses:

1) An acceptable proposal must be consistent with the realities of the object itself.

For instance, one could wonder how frequently Latin manuscripts arbitrarily mess-up/drop all word endings. The VMS is a XV Century ms. If it is a XV Century Latin manuscript, it must be compared with other XV Century Latin manuscripts. In all the other cases I am aware of (cipher or clear-text) suffixes are not written arbitrarily: what you propose does not seem to be consistent with the realities of a XV Century Latin manuscript.

2) An acceptable solution must be the result of a sound and explicable methodology that is logical and repeatable.

I think you have great margins for improvement here.
Your method is highly ambiguous, so much so that you cannot translate anything and are left with listing multiple Latin words for each Voynichese word. For instance,  for each of the first four words in You are not allowed to view links. Register or Login to view. you list 2 possibilities, meaning that you are undecided between 2*2*2*2=16 different readings.
But of course the possibilities are many more. You write that ar can be area (area) or arvum (field), but it could also be ara (altar), aro (a plant of the genus "arum"), arem (I plow)... In most cases, you cannot even tell if a word is a noun or a verb or some other part of speech. With tens of options for just four words, the translation process cannot be repeated.

In my opinion, an even bigger problem is how you split the text into sentences. Latin manuscripts often had very little or no punctuation because you could understand the structure of the text on the basis of grammar. You decide to throw away grammar and suffixes and by doing so you give up any hope of understanding where a sentence ends. Your sentence-splitting appears to be totally arbitrary and unrepeatable. Why should others think that the first sentence is 4 words long and the second 6 words long?

Finally, suffixes are not absent in the readings you propose, you just treat them as irrelevant. Why would someone write "cicutas" (accusative plural) if he meant "cicuta" (nominative singular) or "cicutae" (nominative plural)? This does not seem logical to me.

3) An acceptable proposal must result in a reading that makes sense semantically, chronologically, and logically.

You stop at word lists, without even trying to create a semantically coherent text, so your attempt cannot satisfy this requirement. Others used similar methods to produce something still arbitrary but more or less meaningful (e.g. You are not allowed to view links. Register or Login to view.), nothing satisfactory anyway. Your choice not to push things that far seems quite understandable to me.