The Voynich Ninja

Back in 2018 I had a brief obsession with the way Voynichese text seemingly effortlessly flows around the images. 
Post 1: You are not allowed to view links. Register or Login to view.
Post 2: You are not allowed to view links. Register or Login to view.

Looking for similar behavior in other plant manuscripts, I came across a few categories that are close in behavior, but still different.

1) Extreme economy of space. In these manuscript, like the Trinity Herbal regularly mentioned here on the forum, they try to fit as many images and their text as possible onto the page. The difference with the VM large-plant section is that its main concern does not appear to be saving space. Its defining characteristic is one large plant drawing per folio.
2) Text blocks around the plant. Some herbals do write all around the plant, but prefer to do so in somewhat justified and ruled text blocks.
3) "Soil text". A surprisingly large number of manuscripts prefer to fit the text around the root of the plant drawings. I called this "soil text", because it almost looks as if the text forms the soil in which the plant is growing. The VM herbal section does the exact opposite, avoiding text around the roots.

This left me with two manuscripts behaving just like the VM: the 6th century Juliana Anicia Codex (JAC) and the 14th century Padova, Biblioteca del Seminario, 194. The latter is a copy of the former.

Then I went down a rabbit hole to find out how these manuscripts obtained their current layout. It was a long search, but after reading all the articles I could find and contacting specialists, I was able to reconstruct the whole picture:

* The JAC was made in the early 6th century. Its original text is in Greek uncial and behaves mostly like "soil text".
* A monk named Neophytos copied the JAC in the 14th century. At this stage, people cannot read Greek uncial well anymore, so he transcribes it to Greek cursive. This is the first time (that I know of) we get true, consistent Voynich-style layout in a plant manuscript: Neophytos copying the JAC drawings very faithfully, then transcribing the text and adding this to the drawings. This MS is now in Padua.
* Another monk, Chortasmenos, restored the JAC in 1406. Since the Greek uncial was no longer preferred, he added a transcription in contemporary Greek cursive. This new, 15th century text added to the 6th century codex is often in Voynich-style layout. 


Now I thought it was a good idea to revisit this topic, since there was discussion in another thread about the Leiden Herbarium BPL 3103. A glance at You are not allowed to view links. Register or Login to view. reveals what might be an important detail, which I already suspected upon seeing the script:

"This herbal was written in South Germany [c.1460]. Originally the manuscript only contained paintings of the plants and their names. The extensive commentary in German, Latin, and Czech was added later, in the seventeenth century."

So in these three cases, the Voynich-layout style text is:

* Padua MS: transcribed from continuous Greek uncial, added to faithfully copied images.
* JAC: added to ancient codex
* Leiden: added to 200-year old herbarium

It feels like there is some similarity between these cases: a very strong primacy of the image, i.e. the image preexists and is found very important. And a certain "distance" between the image and the application of the text.

One additional piece of information which might be of interest (or not) is that the original text of the JAC (6th century) did not contain spaces. I am not sure if they respected word breaks. Here is one page which exceptionally has the uncial at the top and the later minuscule at the stem and root:

(30-10-2023, 01:28 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view."This herbal was written in South Germany [c.1460]. Originally the manuscript only contained paintings of the plants and their names. The extensive commentary in German, Latin, and Czech was added later, in the seventeenth century."

Thank you for pointing that out.

(30-10-2023, 01:28 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.One additional piece of information which might be of interest (or not) is that the original text of the JAC (6th century) did not contain spaces. I am not sure if they respected word breaks. 

I could find a partial transcription of the You are not allowed to view links. Register or Login to view. Anicia page in this pdf document:

You are not allowed to view links. Register or Login to view. (University of Ioannina)

Text from the document:

βολβὸς ἐδώδιμος εὐστόμαχος ὁ πυρρός, ἀπὸ Λιβύης κομιζόμενος, ὁ δὲ πικρὸς καὶ σκιλλώδης εὐστομαχώτερος, πεπτικός. πάντες δὲ δριμεῖς καὶ θερμαντικοί, συνουσίαν παρορμῶντες, τραχύνοντες γλῶσσαν καὶ παρίσθμια, πολύτροφοι καὶ σαρκοποιητικοί, ἐμπνευματοῦντες· ποιοῦσι δὲ καταπλασσόμενοι πρὸς στρέμματα καὶ θλάσματα καὶ σκόλοπας καὶ πρὸς τὰς τῶν ἄρθρων ὀδύνας καὶ ποδάγρας σὺν μέλιτι καὶ καθ ̓ ἑαυτούς, καὶ τὰ ἐπὶ τῶν ὑδρωπικῶν οἰδήματα καὶ κυνόδηκτα ὁμοίως σὺν μέλιτι· ἐπέχουσι δὲ καὶ ἱδρῶτας

This is the Uncial part, where I added +I/-I for image breaks that respect / break word boundaries. Similarly, +L/-L for line breaks.

βολβὸς ἐδώδιμος εὐστόμαχος ὁ πυρ <-I> ρός, ἀπὸ Λιβύης κομιζόμενος, ὁ δὲ πικρὸς καὶ <+I> σκιλλώδης <+L> εὐστομαχώτερος, πεπτικός. πάντες δὲ <+I> δριμεῖς καὶ θερ <-L> μαντικοί, συνουσίαν παρορμῶν <-I> τες, τραχύνο <-L> ντες γλῶσσαν καὶ παρίσθμι <-I> α, πολύτροφοι <+L> καὶ σαρκοποιητικοί, ἐμπνευμα <-I> τοῦντες· <+L> ποιοῦσι δὲ καταπλασσόμενοι <+I> πρὸς στρέμματα <+L> καὶ θλάσματα καὶ σκόλοπας καὶ πρὸς τὰς τῶν ἄρθρων ὀδύνας <+I> καὶ ποδάγρας <+L> σὺν μέλιτι καὶ καθ ̓ ἑαυ <-I> τούς, καὶ τὰ ἐ <-L> πὶ τῶν ὑδρωπικῶν οἰδήματα <+I> καὶ κυνόδηκτα <+L> ὁμοίως σὺν μέλιτι· ἐπέχουσι δὲ <+I> καὶ ἱδρῶτας

The same for the cursive at the bottom:

βολβὸς ἐδώδιμος εὐστόμαχος ὁ πυρρός, ἀπὸ Λιβύης κομιζόμενος, ὁ δὲ <+I> πικρὸς καὶ σκιλλώδης εὐστομαχώ <-L> τερος, πεπτικός. πάντες δὲ δριμεῖς καὶ θερμαντικοί, συνουσίαν <+I> παρορμῶντες, τραχύνοντες <+L> γλῶσσαν καὶ παρίσθμια, πολύτροφοι καὶ σαρκοποιη <-I> τικοί, ἐμπνευματοῦντες· ποιοῦσι δὲ <+L> καταπλασσόμενοι πρὸς στρέμματα καὶ θλάσματα καὶ σκόλοπας <+I> καὶ πρὸς τὰς τῶν ἄρθρων ὀδύνας καὶ ποδάγρας σὺν μέλιτι καὶ <+I> καθ ̓ ἑαυτούς, καὶ τὰ ἐπὶ τῶν <+I> ὑδρωπικῶν οἰ <-I> δήματα καὶ κυνόδηκτα <+I> ὁμοίως σὺν μέλιτι· <+I> ἐπέχουσι δὲ καὶ ἱδρῶτας <+I>

Apparently, both scribes tried to respect word boundaries, the uncial script also shows smaller characters used near the image or the right-side border in order to squeeze words into a tight space.
καταπλασσόμενοι with smaller 'oi' near the image break:
[attachment=7821]

The uncial scribe manages not to break words in 60% of cases (12/20); the cursive scribe scores 77% (10/13). (Of course, my count may not be totally correct)

Thanks for the amazing work, Marco! It's interesting to see that both tendencies occur together: they strive to safeguard word boundaries, but at the same time they split words when it's really necessary.

I feel that this image has some alignment and disalignment properties of VMS.

That's the Trento Herbal, Erbario di Trento, ms 1591 Museo Provinciale d'Arte, a copy of the Alchemical Herbal made in Veneto at the end of the 15th century. Though it may not be clear at first sight, the text is arranged in two distinct columns.
See other pages from gettyimages
You are not allowed to view links. Register or Login to view.

Yeah so this manuscript, though certainly interesting for its images, is not a really good match layout-wise. It is part of categories (2) and (3) I outlined in the first post. 

Marco explained (2) already. The MS is clearly going for a certain layout esthetic, whereby you have two text blocks on both sides of the root. Sometimes it even looks like they put some thought into this while drawing the root, for example in the following image, where the root forms a nice text frame:



It also follows a different fill-up strategy than the VM: as soon as the text runs into the image, it stops. The VM will often allow the text to fill up gaps in the image.

As for (3), this is more about where the text starts. In the Trento Herbal, like many others, the text prefers to be around the root. With relatively narrow roots, it starts around the level where the ground would be. When the root has a broader upper part, it starts lower. One exception I see in the images Marco linked is the Chamomile of f43r, where the text sits around the lower stem. This is something we can also find in the VM. In all other folios I can see here, the Trento herbal uses soil text.

If you feel like this is all a bit arbitrary, please do the exercise for yourself, and you will notice the difference. Go through the VM large plant folios, and see where the text starts. 

The rules are generally as follows:

- Allow for a top margin; the image may break through this margin, but the text does not like to start all the way at the top of the page.
- Having taken the margin into account, start as high as possible, unless...
- If the top part of the plant is particularly wide, prefer to start lower instead.
- Avoid the roots. There are only a few semi-exceptions to this rule. Of course, it may be that this rule is just a consequence of the preceding ones, not an active desire to avoid the roots.

When there are multiple text blocks, they may leave a gap between them, preferring to avoid wide parts of the drawing. Text blocks like to be around relatively narrow parts of the drawing.

Did anyone study the similarity of the 15th century herbal manuscript Erbario - Simplicibus Benedicti Rini from Northern Italy, which has a lot of similarity with the Voynich manuscript, particularly the shapes of the roots. It also contains Slovenian words.
Out of 458 names of the plants in Latin and Greek, there are also 370 Slavic names and only 65 German. The debate is still going on whether the Slavic names are Slovenian or Croatian, however I believe the language (like the language of the VM is still close to both, since it comes from the region ruled by the Republic of Venice. The codex had been studied by Italian botanist E.De Toni, but not enough research has been done by Italian experts on the proper transcription and translation of the Slovenian words. There are several articles written about that in Slovenian and Croatian (only recently available on internet).
It was believed that the Code was written in 1415 by medical doctor Benedicto Rini, but later it was proposed that Nicola Roccabonelly had written it in 1449.
Italian researcher Mini believed the names of the plants are Czech, which was disputed by Croatian and Slovenian linguists and also by Italian naturalist of the 16th century, Ulisse Aldrovandi, who copied the Slavic names of the plants on another manuscript (I suppose the one mentioned). He refers to the names as Illyrian, which was the term used for Slovenian, and Croatian Kajkavian (which is closer to Slovenian than to Croatian).
Most pictures in the Rini Codex were made by Venetian artist Andrea Amaeo, except for some towards the end, that look different. Among them is the image of Mandragona, where text is not written over the picture.[attachment=7825][attachment=7826][attachment=7827]

I was only able to obtain some of the images from the Rini codex which is stored in the museum in Venice. Since the words mentioned in Slovenian and Croatian articles have similar properties as the words in the Voynich manuscript (dropped letters, occasional improper spaces, interchangeable use of c/ch, s, z, and long s,  etc.), it would be worth to examine all 360 words. 

Unfortunatelly, I do not understand Italian or Latin, so I cannot compare these studies with the original or with the Italian sources.

In the Slovenian/Croatian articles, the names of the plants are discussed only from the linguistic point of view to study how the 15th century Slovenian and Croatian written and how the vocabulary was preserved. 
Some examples: Viole bele, Beli tarn, cugel beli  (beli means white); ochyun, Lochyun, lan, telsti koren, lepuh, volcge iabolcha, posgi iasila (now pasji jezik), Drivo od basama, drivo od naranca (drivo - tree); orhic (nut), triputec, Mah cerlyeni (Papaver rhoeas), cerlena  meta /origanum vulgare (cerlen  - red), cemericha cerna /Helleborus viridis. Rini used the word 'selje'  to mean 'herb', and he spelled it in various ways: zeglie, selje, selie, selye, since Z, S and long S were often used for the same sound.

Cvetka

This MS can be viewed here: You are not allowed to view links. Register or Login to view.

Since this thread is about the layout, I will focus on that.

First of all, it is clear that the images existed first and a later person added the text (a frequent feature of the MSS in this thread).

* Many plants are still without text
* A particular feature of this writer is that he is not afraid to have his text overlap with the image. In fact, this is the case in all the folios with added text I looked at so far.
* Preference to place text at the roots or in any free space under the plant.
* Several exceptions, e.g. You are not allowed to view links. Register or Login to view. has really bulky roots, so here the text is written over the more forgiving leaves:

[attachment=7828]

* The main exception is when a plant has a long, slender stem, then the text is written over the stem.

In summary, the text insertion strategy of this writer appears to be as follows: he prefers not to have his text interrupted, and attempts to find the best place to lay his text block over the image. This will often be near the roots, presumably when these are not too important and leave a lot of empty space to their sides and underneath. 

Sometimes this results in similar layouts to certain VM pages, but both writers use different strategies. They can be generalized as follows:

VM: I want to start as high on the page as possible, as long as I can leave a top margin and the image is not too broad on top. When I encounter an image, I will fill up any gaps and continue on the other side of the image.
This herbal: I would like my text block to be as continuous as possible, therefore I will select a spot where it won't interfere with the image too much and I can just write over any image parts I encounter. This is often the roots or the stem.

Of course there are exceptions to all of this, but I think these general strategies best explain the MSS respective layouts.

I recently stumbled upon an illustrated herbal I don't remember seeing before: Biblioteca dell'Orto Botanico di Padova, Ar.26, Veneto, last quarter of the 15th century. You are not allowed to view links. Register or Login to view.

It contains a copy of (part of?) the Pseudo-Apuleius collection and possibly some other work, but many of the illustrations are of good quality, similar to Tractatus De Herbis. Typically, each page either contains a single plant illustration (with plant name in red) or text. But in the first part there are a few pages in which text appears together with the plant. In most cases, the text that appears with the plant represents the lists of names typical of Pseudo-Apuleius, but sometimes there are paragraphs of actual text. All text appears to have been written by the same hand.

The bottom-left image exemplifies another interesting feature of Voynich layout: plants are often shifted to the right or (less frequently) left of the page, as if to make room for the text (I am sure we discussed this at some point).