The Voynich Ninja

Thanks, Marco!
Yes, you are right. Actually I am mixing two problems. My main motivation for making this thread was your point (2). I rather wonder if the [a] followed by [i] is not the form VM minims take in certain contexts. So [ai] could be like "m" or "n". If the first minim takes a leading serif, it would explain why there is always a (pseudo?) [a] before [i].

I think it's really important in all of this to keep making the distinctions between an alphabet, a scribal inventory, a linguistic inventory, and even a phonemic inventory. Many of these differences in transcribing [iin] (for example) can be attributed to different goals of the transcriber. If you want to know how large the Voynichese alphabet is, then you have to make a decision about whether [iin] is a letter or a combination, but if you're going to use that transcription for, say, frequency analysis, you have to take into account that your count of [i] might be way too high. The number of glyphs in a scribe's graphic inventory (which is where my own work is headed) is almost always going to be larger than the number of "letters" in the "alphabet." For example, a medieval scribe might use three very different types of [r] depending on the context (word-end, word-initial, and in ligature with [i], for example), and someone looking at the script who didn't know that might think these represented three different "letters" and therefore end up with a larger number of letters in the proposed alphabet than there should be. A linguistic inventory might be closer to an alphabet, although it might also include bigraphs like our [qu]. A phonemic inventory, on the other hand, might be smaller OR larger than the alphabet, depending on the language. Not sure if this if helpful to others, but those distinctions have been helpful to me as I think about what the actual questions are that I'm trying to answer.

(06-04-2020, 03:21 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.
...

This said, your idea of treating aiin as a single entity makes perfect sense: if the writing system is phonetical (meaningful or not) iin could be like an accent, that in a few languages (e.g. Italian) is only used for vowels at the end of words.
...

Marco, thanks for the graph. Data visualization is always appreciated.

Your comment made me wonder... Have we ever compiled a list of possible interpretations for "ain"?

• I know that many people have suggested "and" mainly because it is such a common glyph sequence and maybe also because it begins with the "a" shape + minims ("n"?) plus a glyph that sometimes looks like a "d" but it seems to me quite a few of them were suggesting "daiin" and not "aiin" might be "and" and that seems less likely to me. I think it's important to make a distinction between "daiin" and "aiin".
  

But to get back the various possibilities, in addition to the one you noted...

• Another possibility is a grammatical ending (like an ending that creates a noun or adjective). It seems too frequent to be the ending on a specific word but if it were something that could modify numerous words maybe that could account for the frequency.
  

• One possibility I may have mentioned on my blog when I was talking about Roman numerals, is that perhaps it's a value. What comes to mind (just as an example) is the way medieval texts classified various things according to their elemental properties. For example, it might be hot in the 3rd degree or cold in the 2nd degree. There are instances of av, aiv, aiiv, and aiiiv (I use "v" rather than "n" in my transcripts because it looks more like the tail that occurs on the VMS chars). In other words, potential values from 1 to 4. In this scenario, I'm not sure why the "a" would be there. If it were hot/cold/wet/dry in 1 to 4 degrees (grado) then one might expect the "a" shape to have 4 variations but I'm not sure it has 4 consistently-discernible variations. Or, if the "a" is actually a curve followed by a ligatured minim, then potentially it's 1 to 5. If it is numerals, maybe it's not elements, maybe it's numbers for something else.
  It's only one idea, not necessarily the right one, but the concept is that maybe the minims indicate values... for something. The "c" shapes only go up to 4, but if the minims are maxed out at 4, then the two groups would have the same max.
  
• Maybe it's a connector or stop-point. If so, it's a horribly verbose one, but we don't know what's going on in the VMS, so I thought I'd add it to the list.
  

Koen's examples are moderately variable, especially considering they were sampled on the same folio. It's very hard to say if the variations are meaningful.

I had some other ideas for what ain could signify but I don't want to make the post too long and I'm sure others have ideas too.

(06-04-2020, 05:27 PM)-JKP- Wrote: You are not allowed to view links. Register or Login to view.

(06-04-2020, 03:21 PM)MarcoP Wrote: You are not allowed to view links. Register or Login to view....

This said, your idea of treating aiin as a single entity makes perfect sense: if the writing system is phonetical (meaningful or not) iin could be like an accent, that in a few languages (e.g. Italian) is only used for vowels at the end of words.
...

Marco, thanks for the graph. Data visualization is always appreciated.

Your comment made me wonder... Have we ever compiled a list of possible interpretations for "ain"?

Hi JKP,
from my point of view, the list of possible interpretations for the suffix -ain is infinite: compiling it would be an endless task.
The graph was my attempt to address Koen's question. When I say that a could be like 'a'  in Italian, and ain / aiin like the Italian 'à' I am just using an analogy for the kind of close relationship I see between a-followed-by-i and a-followed-by-anything-else.

Here is a slightly more sophisticated version of the graph, where the bars are expressed as % of the number of sequences starting with each  character. I also added o and y, so that we can compare ai and a-Not(i) with two symbols that are known to look and behave similarly to a (a o y are the three "circles" discussed by Stolfi). Since we are separting the two cases on the bases of the right context (i/not-i) the left context is all we have to check if the two behave differently

[attachment=4183]

What can be seen is that the hypothetical two versions of 'a'  behave very similarly: I would say that this graph is best interpreted as showing 3 different characters, not 4: 'a', 'o' and 'y'. The left context does not seem to support the idea that a-i and a-Not(i) represent something very different.
Differences in this kind of histogram are a necessary but not sufficient condition to conclude that two symbols have a different function: Lisa mentioned the context-sensitivity for the shape of [r]. If one was to create histograms for round-r and straight-r in a Gothic script, the left-context would appear to be largely different, since each r-shape is chosen depending on the "roundness" of the previous character.

If the Voynichese writing system is phonetic, 'a' appears to follow sounds in a uniform way, independently from the presence of the following i-sequence.

If the system follows some kind of visual "glyph harmony" (as discussed by Timm and Schinner, after Schwerdtfeger), what we observe is visual equivalence: nothing suggesting a c+minimSequence structure that requires a different  harmony from (say) -al.
Unless I messed up something, according to this specific side of the evidence, [a] appears to always be [a].

There is the interesting fact mentioned by Emma that i-sequences and bench-e-sequences (the two clearest candidates as minim-structures) tend not to coexist within words: this is not easy to explain, but I don't think it suggests that there are two different types of 'a'.

Great thread on a well-loved but not frequently well-articulated topic, Koen. The [aiin] series have always been my favorite element of Voynichese, and I suspect I'm not alone in this. They're unique in a number of ways that are highly interesting, particularly:

• Occurrence almost exclusively at the end of a vord, without a corresponding strong preference for the end of a line
  
• A variable number of middle elements, with the beginning and ending elements typically uniform
  
• Strong resistance to being parsed as a definite number of independent glyphs, making them the bane of transcription and statistical analysis attempts.
  

Ideas I can get behind, given what we know currently:

• [aiin] is one character. So are [ain] and [aiiin]. The occurrence of [air] and [aiin] in strings of otherwise isolated glyphs on You are not allowed to view links. Register or Login to view. and f57v, respectively, is good evidence of this.
  
• [ain], [aiin], and [aiiin] are highly similar but distinct glyphs from each other. Assuming the text is meaningful, there is some commonality between what these three glyphs encode, but the number of [i]s is indeed meaningful, and the three are not exactly the same.
  
• That said, there are probably nevertheless places where a scribe meant to write [aiin] but wrote [ain] or [aiiin] instead, and vice versa. Therefore, the [a...n] series should probably be seen by statistical analysts as introducing a certain amount of error in the system — noise in the signal, if you will — that cannot be eliminated until/unless a good deal of meaning can be extracted from the text through other approaches.
  
• There is some commonality between what [a] encodes, and what the [ai(ii)n] series encode.
  
• [an] and [aiiiin] are probably always scribal errors
  
• How far to the left the flourish of the [n] reaches is an idiosyncratic property of the scribe's handwriting, and is not orthographically meaningful. That is, it tells us ea lot about which scribal hand wrote it, but little to nothing about what information the vord encodes.
  
• The origin of this series of glyphs is in Latin scribal abbreviations and shorthand. In this case, for a word ending made of Roman letters that were written entirely with minims (i, m, n, u, v, and w). This says nothing, necessarily, about what meanings (if any) this series of glyphs encode in the VMS.
  

Still unsettled, for me, is the relationship between [ai(i)r] and [ai(i)n]. Are these equivalent? They certainly behave similarly. But [ar] and [an] do not. Again the dilemma: do we treat this distinction as meaningful, treat it as meaningless and assume that scribal errors account for it, or throw the Rubik's cube at the wall in frustration and conclude all of the damn book is meaningless?

@Rene
I'm thinking more of the number crunch. I have a key with 26 possibilities. If I have filled out the key correctly, all meaningful words result in the riddle.
But now I have 162 VM characters, so I have 162 possibilities. If I write with "112 = ver" then each character 112 automatically becomes a "ver" in the book. If I have enough syllables or single letters in the key, also single words in the book result.
In case of mistakes I can change quickly. A small correction then affects the whole book.
At some point I get a feeling for the text. So that I can say, it is not called "vergehen sondern zergehen".
So the string "teol cheol otchey ??? cheor cheol ctheol cholaiin chol qkar" will eventually make sense.


Ich denke eher an das Zahlenrätzel. Da habe ich einen Schlüssel mit 26 Möglichkeiten. Habe ich den Schlüssel richtig ausgefüllt, so ergibt sich im Rätzel alles sinnvolle Wörter.
Jetzt habe ich aber 162 VM-Zeichen, somit auch 162 Möglichkeiten. Schreibe ich bei "112 = ver" so wird jedes Zeichen 112 automatisch ein "ver" im Buch. Habe ich genügend Silben oder Einzelbuchstaben im Schlüssel, ergeben sich auch einzelne Wörter im Buch.
Bei Fehlern kann ich schnell wechseln. Eine kleine Korrektur wirkt sich dann auf das ganze Buch aus.
Irgend wann bekomme ich ein Gefühl für den Text. So das ich sagen kann, es heisst nicht "vergehen sondern zergehen".
So wird die Zeichenfolge "teol cheol otchey ??? cheor cheol ctheol cholaiin chol qkar" irgendwann einen Sinn ergeben.

(06-04-2020, 11:52 PM)RenegadeHealer Wrote: You are not allowed to view links. Register or Login to view.Still unsettled, for me, is the relationship between [ai(i)r] and [ai(i)n]. Are these equivalent?

By pure coincidence, just a few days ago, I create this small overview:

You are not allowed to view links. Register or Login to view.

If I use logic, it's the endings, " -tis, -ris, -sis, -vis "
But logic does not always have to be right, but there is an example.

Question: Can your key be linked to the text?

Cuva alphabet table
with

You are not allowed to view links. Register or Login to view.

(07-04-2020, 05:42 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.

(06-04-2020, 11:52 PM)RenegadeHealer Wrote: You are not allowed to view links. Register or Login to view.Still unsettled, for me, is the relationship between [ai(i)r] and [ai(i)n]. Are these equivalent?

By pure coincidence, just a few days ago, I create this small overview:

You are not allowed to view links. Register or Login to view.

Thank you, Rene!
It really seems that i almost always appears after the diagonal stroke of a and before another diagonal-stroke character (n r l m).

Similarly, e almost always appears before another c-like stroke (d y). It often follows the c-like chape of a bench or bench-gallows, but there is the exception that it can also follow the vertical shape of a gallows....
Since, as Koen wrote in the initial post, one way to look at a is as c + i, it can connect a c-sequence with an i-sequence.

You point out a "continuum" in the position of the curl in n and r, which is another intriguing phenomenon: The strings ending in "r" also visually resemble the strings ending in "n" and there are numerous places in the MS where some intermediate form exists and the reading is very difficult.

Hi Marco, yes this is all true, but of course there are exceptions to just about everything.

Strings of i are not rarely preceded by o instead of a.

Also, I have seen cases where the r in iir looks more like s, but has still been transcribed as r.

Edit:

By the way, I really like the graph in You are not allowed to view links. Register or Login to view.

It has a lot of interesting information, e.g. it seems to say that k and t tend to have a different context, while that of r is more similar to t.