The Voynich Ninja

Marginal annotations in You are not allowed to view links. Register or Login to view. (top), You are not allowed to view links. Register or Login to view. (bottom) and You are not allowed to view links. Register or Login to view. seem to me particularly interesting because they mix Latin-alphabet and Voynichese words. Two of them (66r and 116v) also include illustrations of naked women. The shape of the characters in these marginalia seems to me largely consistent with the characters that Rene Zandbergen and Alain Touwaide consider possible colour annotations. Finally, the presence of an 8-shaped character in both scripts is noteworthy: Latin scripts often include 'd' and 's' with two loops, but not as symmetrical as EVA:d d.
In 1954, You are not allowed to view links. Register or Login to view. thought it possible that the whole manuscript (last page included) was written in a single hand:
"In my opinion the whole manuscript is by the same hand with the possible exception of the last page; but I am by no means sure of that".

While thinking of a single hand for the main body of text and these mixed-alphabet annotations may be too much, the author of the annotations seems to have been part of the environment that produced the manuscript, or least close enough to write Voynichese and draw nymph-like figures.

Here I would like to discuss the script used for these annotations, without going into the well-known problems of their interpretation. Since there are transcription ambiguities, not all characters can be identified with certainty. I have assembled a tentative (and obviously incomplete) 116v alphabet, also based on a character-by-character analysis by Vogt and Schwerdtfeger (You are not allowed to view links. Register or Login to view., 2010). I expect there will be different opinions about some of the characters, still there should be enough agreement to discuss the script as a whole.

[attachment=2218]

[attachment=2219]


In You are not allowed to view links. Register or Login to view., Nick Pelling discussed a parallel for f17r, which also compares very well with 116v: Basel University Library A X 132. In the same post, he provides digitally enhanced images of the 17r marginalia that make the presence of Voynichese clear. Unluckily, the available image from the Basel ms is quite small. Also, it consists of a single three-words sentence, so many characters are missing. This is an enlarged, cropped and manually equalized detail:
[attachment=2217]

I guess the sentence reads:
Vocabularii hebreic(us) et grec(us) - I cannot read the following two words: vi gr(ecis)??
Hebrew and Greek dictionaries

The first lines of the Hebrew-Latin dictionary are visible in the lower part of the image. They seem to me to be in a totally different script. For instance, the first line of the main text reads:
(A)lma v(ir)go abscondita

Compare the 'b's in the marginal sentence with that in abscondita, or the gs in grecus and virgo. Actually, comparing the first line with the main body of text makes clear why the script of the first line is special.
Basel A X 132 marginal sentence is a good parallel: finding better images of a longer text with a similar script would already be a success.
Some comparable features:

• the initial v matches the first letter in "umen"? (116v first line)
  
• the 'h' (with ascender and descender) is similar to that in 'anchiton' (116v second line)
  
• l and b are also good matches, but with less triangular loops
  
• lines are thin, with no outstanding bold strokes.
  

In his comment on Pelling's blog, JKP has pointed out some differences:

Quote:I notice greater connectivity between letters and a different style of “r” in both -lary and hebre- Also the stem of the g extends above the loop (which means the stroke order is different from the VMS g) and the descender is shorter and more curled, also the base of the ell has a serif and connector (and the first leading stem on the first letter is very long).

I think his observation about r is particularly important. This script uses two different kinds of 'r', one shaped like a '2', the other more like a 'v'. In the Voynich marginalia, only the second kind appears. This seems important: something specific to hunt for in other candidates. Of course, a perfectly 8-shaped final s would also be welcome.

I've mentioned this before, but I am pretty confident that the marginalia on 116v and on 17r are written by the same hand, and possibly also the ven mus mel.

I have a blog in draft form about this topic, I just have to add the images (which always takes longer than I expect it to). I'll try to do that this evening after work and get it posted.

Quote:MarcoP: Of course, a perfectly 8-shaped final s would also be welcome.

Haha! Yes, I completely agree. With diligent effort, it's possible to find handwriting similar to the VMS marginalia, but finding similar writing that also has the same figure-8 character (a symmetric one rather than the more common asymmetric style), a flat-bottomed "b" (note that it is not round), an "i" with a long leading-serif, and the same "g" (with an s-curve tail, and a lowered, separately added serif on the right) is a real challenge!

Thank you, Marco! I was planning to make a thread just like this (following your recent blog post and the discussions in the other thread), but you did so much more competently than I would have
Will check back later when I have some more time, I will be forced to watch football in a bit....

It may help to compile a bullet point list of features that stand out in this script. What are all the pointers we are looking for?

(28-06-2018, 06:35 PM)Koen Gh. Wrote: You are not allowed to view links. Register or Login to view.It may help to compile a bullet point list of features that stand out in this script. What are all the pointers we are looking for?

Hi Koen, the script is a "Cursiva" as defined You are not allowed to view links. Register or Login to view.:

Quote:The identifying features of Cursiva are a single-story a; loops on the ascenders of b h k l; and descenders on f and the straight or long (formerly “tall”) version of s.

A first filter could be checking for the presence of these features, but they will be frequent in XV Century manuscripts.

We can take an example from that same page (Geneva, Bibliothèque de Genève, MS fr. 1/2, 1rb, ll. 9-10.)
[attachment=2220]


These are some of the features of the Voynich script I find relevant (but I am sure more can be added):

• Characters are "smooth" without the Gothic "angles" one can see in the Geneva ms (this is more a peculiarity of the Geneva ms than of the VMS).
  
• Rare use of abbreviation symbols in the VMS (a single abbreviated word, You are not allowed to view links. Register or Login to view. "luc(em)?", in more than 20 words).
  
• In the VMS, characters are typically not connected: "malhor" in You are not allowed to view links. Register or Login to view. is made of six glyphs that do not touch each other. If it is "mallier", li is written as a ligature and two characters touch each other. "anchiton" (f116v) is made of four unconnected bits: an-ch-i-ton. In the  Geneva manuscript, words mostly are fully connected, if by almost invisible hair-strokes.
  
• No particularly bold strokes in the VMS (see the heavy long s in the Geneva ms) but consistently thin strokes.
  
• Open-top p (see 116v first line)
  
• x with a loop on the right side
  
• Downward extending flourishes in the first / last characters of words (116v: umen - first line; ubren, gas and mich - line 4)
  
• Open g (no lower loop): again gas, 116v line 4
  
• Open d (no upper loop): 116v, line 2, "oladabas" and 66r
  

And the already mentioned:

• Markedly triangular loops on b, l, h
  
• Perfectly 8-shaped s
  
• No 2-shaped r (see pro and colleterali in the Geneva ms)
  

Marco: thanks! It looks like the broad, flag like ascenders are the most eye-catching property of the script (or hand?). These should be easy to scan for in other MSS, and then check the rest of your list to see how well it matches up.

JKP: do you have any pointers on how or where we'd be most likely to find matches for this script? Have any examples been found apart from Nick's MS? 

By the way, it's remarkable that these are also marginalia. Does this mean we may have to look later than the first half of the 15th century?

Koen, I've been collecting samples of scripts similar to the VMS marginalia and the VMS main text since 2008.

I have about 1,000 sampled texts that are similar to the marginalia (a total of about 35,000 individual character samples), rated on a mathematical scale for similarity, chosen from thousands of manuscripts. About 200 of them score high on the "very similar" scale. Many of them use the same conventions for Latin abbreviations as occur in the VMS main text.

I have fewer for the main text, because it's more difficult to find similarities when you only have a few glyphs that can be confidently compared (e.g., "o", EVA-y, and the swept-back tails). I have a feeling the back-leaning letters are deliberately leaned rather than being the scribe's normal handwriting. I only have about 100 or so that I feel reasonably confident about.


So, if you are looking for matches, all I can say is that you'll probably look through the same thousands of manuscripts I did and will probably come up with similar choices.


Can you see how it's difficult for me to post what I have? It's a MOUNTAIN of data. It's all searchable, sortable, I have it narrowed down to the best choices, but presenting it all is a bigger job than it might seem.


As for the time-frame for the marginalia, it is probably in the same century as the main text. I don't think there's a big gap between them. They use the same conventions as were common in the 15th century (from about 1390 to about 1520), and it's even possible that they are close in time. That's one of the reasons I don't think it's a modern hoax. You would need quite a bit of paleographic knowledge to create something this authentic-looking.

JKP: I'd just like to see some more examples of MSS that use these kinds of loops on L etc. Do you happen to have some that you could share easily?

I assume you mean a Gothic ell (not EVA-l). Yes, I have tons of them. They are common to most Gothic text (which is primarily 15th century).

I can tell you with complete certainty that the loop-ell in scripts that are similar to the marginalia usually have a little serif on the foot and the ell without that serif is uncommon (not rare, but not common). Here are some examples of the letter ell in manuscripts with writing that is somewhat similar to the marginalia in other respects:



ÖNB Cod. 15096 Han is interesting in a number of respects. It is more widely spaced than most scripts, has the looped l and b, has the rounded loop on the "h" (this is the less common form), a rounded e, a non-looped d. It's similar to the VMS script in a number of respects. Not the closest I have, but worth a look. It's from Wiener Neustadt (Latin/German text).


Edit [addition]: I should mention that these are from all over... England, Gerrmany, Alsace, Austria, Italy. Scholars carried their traditions with them.

Damn, I thought close matches for the almost triangular loop would be rarer, but some of those are spot on.

I can tell you exactly which marginalia letters are the less common forms so you have an idea what to look for. I've studied this intensively. It took me years and thousands of samples to be sure, but as of the last couple of years, I'm quite confident that I have a good understanding of which of the marginalia letters are individual to the scribe or unusual compared to other texts written in the same basic style. Unfortunately, I can't post it until after work. Sorry, duty calls.  


I can say don't look for the "p". It's common. Look for the "g". It's a Gothic Book style, not the usual Gothic Cursiva style, so one doesn't usually see it blended into a cursive text. I have a hunch the marginalia writer learned both Book and Cursive hands. There are other slight traces of Book hand. This is clearly not a professional scribe, but that doesn't mean the writer didn't have a good education.