The Voynich Ninja

Thank you, Rene!
I agree that all the bullets in your list are well worth exploring. A good thing about this hobby is that it appears to be  inexhaustible...

About this specific point:

(28-04-2020, 04:37 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.- the impact of spelling variants could be looked at, but such variants should only be expected for the vernaculars (a linguist should be able to say more about that).

I think that palaeographers could be more qualified than linguists about this particular issue. As I see it, spelling variants correspond to identical linguistic entities.
If I understand correctly what you write about the phenomenon only occurring in vernaculars, you imply that Latin manuscripts had no spelling  variants. Do you consider abbreviations something entirely different from spelling variants?

We know that in Latin manuscripts the same word type could be expressed in different  ways  because most (all?) manuscripts were abbreviated and abbreviations were often used inconsistently.

A while ago  I carefully studied ten pages from a Swiss copy of a Latin text by Bonaventura (discussed You are not allowed to view links. Register or Login to view.). By comparing the handwritten text with a printed edition, I found out that the handwritten text has many more apparent word types than the printed edition (1371 vs 1107). You are not allowed to view links. Register or Login to view. I pointed out three different variants of "secundum".

This is an example of two different variants of "omnis" (each occurring twice):
[attachment=4260]

Another phenomenon that I tend to assimilate to "spelling variation" is arbitrary spacing. This of course contributes to creating spurious word types. 
[attachment=4261]
For instance, I transcribed the first fragment  as "ineam", while the second looks like "in ea". Instead of the two common types "in" and "eam", the first fragment results in an illusory new word type.
One of the great merits of the transliteration you produced together with Landini is making clear how extensive this phenomenon is in the VMS (8% of the spaces between words are marked as uncertain).

(28-04-2020, 10:56 AM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.
...

We know that in Latin manuscripts the same word type could be expressed in different  ways  because most (all?) manuscripts were abbreviated and abbreviations were often used inconsistently. 

...

Yes, in fact, it was perfectly acceptable to abbreviate the same word in numerous ways, even in the same manuscript. Since abbreviation conventions were conceptual, they could be applied differently to the same words. "Sanctus" is abbreviated in several different ways.

Common words tended to be abbreviated more consistently because common words had their own set of abbreviations (like the common "q" words in Latin, and syllables like "pro" and "per" that were used in many languages), but many words were at the discretion of the scribe.


It wasn't chaotic. For example, scribes would sometimes write the whole word the first time it was used. The next time it might be partially abbreviated. The next time, maybe more tersely abbreviated. Thus it was clear what was being written. I see this most often when the word is frequently repeated or if the repetitions are close together, or if the word is used numerous times at the left side of the line (e.g., words like "Item" in Latin and "peri" in Greek).

Not everyone was this organized, but many were.

(28-04-2020, 10:56 AM)MarcoP Wrote: You are not allowed to view links. Register or Login to view.Do you consider abbreviations something entirely different from spelling variants?

I had not thought about abbreviations, so that is a good point.

Let's think of the Voynich MS text as a 'derivation' of a source text. This could be an encryption or some other modification that would not normally be classified as a cipher.

Some properties of the Voynich MS could be considered a consequence of spelling variations in the source text. My point was that some languages in the 15th century had more spelling variations than others. Now if the source text was in an already written (older) document, this could, in addition, include abbreviations.

However, if the text was original to the MS, it seems unlikely (to me) that it was first written down with abbreviations and then 'encrypted'.

Many 'if's already, none of which we can be sure of, but it seems a good idea to keep all of it in mind.

I've never had the impression that the VMS has spelling variations. The text is so remarkably consistent from beginning to end (even taking into consideration the Currier "dialects"). It is almost too consistent. If there are spelling variations, I suspect they are rare.

There might be abbreviations (the 9 abbreviation for endings and sometimes beginnings is one of the most common abbreviations in medieval texts, everyone knew it and most of them used it, so if there are abbreviations in the VMS, perhaps this is one of them).

But if there are abbreviations, they would have to be applied fairly consistently, rather than being written in different ways, or the VMS text would be more varied in terms of glyph position, and the overall makeup of the tokens.

A random thought about reduplication: if you interleaved two texts in the same language, might this reproduce some of the reduplication properties we see in the VM? What I mean is: first word of text A, first word of text B, second word of text A, and so on. Both texts would remain perfectly legible. Probably large-window TTR would not change much either. But it might mess with narrow-window phenomena (i.e. reduplication, a.b.a sequences...).

(18-07-2020, 06:29 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.A random thought about reduplication: if you interleaved two texts in the same language, might this reproduce some of the reduplication properties we see in the VM? What I mean is: first word of text A, first word of text B, second word of text A, and so on. Both texts would remain perfectly legible. Probably large-window TTR would not change much either. But it might mess with narrow-window phenomena (i.e. reduplication, a.b.a sequences...).

Hi Koen,
if I understand correctly, that would result in a reduplication rate comparable with a scrambled text: in the mixed texts, reduplication would be casual, as in a scrambled text. On the other hand, reduplication in the VMS is about twice more frequent than if words are randomly re-arranged (see graph You are not allowed to view links. Register or Login to view.). Also, Voynich reduplication is selective, e.g. 'daiin' reduplicates and 'aiin' does not: it does not seem to be random.

Hmm yes, I think you are right, I didn't think of it that way. The ratio of high-frequency words should remain similar and there is no mechanism that would create more reduplication than what you'd get by chance.