The Voynich Ninja

There has been much discussion as to the essence of the "gallows" characters in the VMS. Basically, there are several types of gallows there:

• "plain" gallows - EVA p, f, k and t
  
• composite "benched" gallows - such as cph and the like - actually we don't know if they represent a single character or a sequence of "plain" characters
  
• "embellished" gallows - which well may be just actual embellishments of the "plain" gallows
  
• weird gallows with one leg in one "word" and other leg in another - notably, EVA t exhibits such behaviour.
  

Let us consider the "plain" gallows. They exhibit some interesting properties:

• They have not so many variations – just four (p, f, k and t).
  
• They are not rare in the corpus.
  
• They occur very frequently as the starting character of the paragraph. Sometimes many paragraphs in a row begin with the gallows.
  
• They seem to never occur in the end of any distinct high-level logical entity (paragraph or label). I was not able to confirm this 100% due to the absence of the respective query in any Voynich tool, but I found no occurrences offhand.
  

What set of elements could have such properties? It occurred to me that a set of articles is a corresponding match.

E.g. (modern) English has two articles – “a” and “the” (if we add the "an" word form, we will have three), German has five (der, die, das, ein, eine) etc.

Of course, the adoption of this idea would mean that Voynich spaces are not real spaces and that there are real spaces where we don't observe them in the MS.

One objection to this idea refers to short labels including gallows. E.g., consider otol (Voynich "star" in f68r). If t here stands for an article, then it is strange to have only one letter before the article in a given phrase. However, what if o does not stand for a single letter (as t does not, in our assumption)? What if o is a shorthand for some notion (like "star" or "stone")? OK, then we have the sequence <notion X> <article> <notion X> l. Not very promising, unless this is something like "star of the stars..." or "ol" is not the same as "o"+"l".

Well, although this article idea probably does not apply directly, I vaguely suspect that something may be developed out of here.

If they are articles, then as you say, spaces do not apply.


Latin has zero articles. However, many Romance languages will have five (male / female singular and plural plus a neuter).
So for Spanish, for example, you have el los la las and neutral lo.
French has more.

Remember there are only 25 glyphs, which aren't enough for a shorthand notation such as you envision. But they do fit with the notion of a European alphabet.

(29-12-2015, 12:02 PM)David Wrote: You are not allowed to view links. Register or Login to view.Remember there are only 25 glyphs, which aren't enough for a shorthand notation such as you envision.

That's a good point also.

Well, frequent occurrence in the beginning of a paragraph and no (if I am not mistaken) occurrences in the end of the paragraph is a peculiarity that makes us think of the specificity of the set of gallows (as opposed to all other characters).

However, let us consider standalone labels. The latter are a valuable object for observation, because most probably they stand for single plaint text words - thus the problem of "Voynich spaces" does not present itself in the case of labels.

What can we see in labels with respect to gallows? We can see that:

- a label may contain a gallows
- a label does not have to contain a gallows
- a label may contain more than one gallows
- a label may contain at least two different gallows
- a label may contain at least two similar gallows (with no other gallows in between)

The second and the fourth observations somewhat undermine the concept that gallows are some "keys" or "markers" (such as those introduced to mark the encryption scheme in use that may change from word to word).

Of course, this is not utterly decisive. One may argue that if a label does not contain a gallows, then the encryption scheme in use (in this word) is designated by a gallows situated somewhere else (e.g. in the previous label). Further, if a label contains two different gallows, then this might mean that the encryption scheme does change within a single word.

But the last point is decisively against the "marker hypothesis". There is no sense in marking the same encryption scheme twice in the same word.

What adds against the "marker" hypothesis is the presence of combinations like f69r, line 3, word 2. Here the t gallows is clearly not a standalone character, but part of a combination with e and a horisontal crossbar (taken together they will represent a c). Interestingly, the benched gallows are put down not as in this example. Rather, the legs of the gallows extend towards the very baseline (which makes us wonder whether they are single complex characters or sequences of plain characters).

Anyway, if not markers or shorthands, the gallows appear to be characters. But then we are back to this strange behaviour in respect to paragraphs.

Let us look at a random page - You are not allowed to view links. Register or Login to view. (and I promise this is a random example).
The outer ring, from the top, contains the same start sequence with different endings. It's an example of my evolving epizeuxis - the same prefix is repeated a number of times with distinct suffixes.
Now, that is clearly not a plaintext with the exception of medieval charms (or spells) which I have discussed in the past - repeating rhyming nonsense words designed to be spoken or chanted aloud.


Now, if we look at the word on the outer ring at 3 o'clock, we see the same gallows character is repeated no fewer than three times in the same word. Other words have repeated gallows. So it's unlikely the gallows character represents a concept. It appears too often.

Your You are not allowed to view links. Register or Login to view. example looks to me like a hiccup, the scribe miscalculated the writing as he went along. We've all done it as we write.

So I still vote for a glyph - letter correlation.

But as you say - why are gallows so common in labels, yet less common and more fixed within paragraphs? This just doesn't correspond to any natural text, as I discussed recently on my blog.

It is unlikely that the gallows mark a change in the encryption method in the labels - they are just too common. But then again, I can't see any evidence of an encryption process in the text, and nor could the great codebreakers of the XX century

So, off the top of my head:

- It's not a natural language.
- Gallows aren't indicative of a change in a hypothetical encryption method.
- Gallows don't represent a "concept" but correlate to a "letter".
- Ergo, glyphs correlate to letters in an alphabet.
- So there is a base alphabet writing letters combined into words (which are not natural language).
- So it's either encrypted into a very strange code which forces glyphs to appear at certain locations within words (but no-one has ever found a pattern) or it's.... nonsense?

"There is the important question whether the underlay-to-overlay (plain text to Voynichese) translation is sequential. In other words, shall we read it left-to-right, top-to bottom, and not in some interleaved fashion. If yes (and some points, including here the very existence of such objects as standalone labels, suggest this), then the question for gallows (especially for non-concept gallows) is:

What is it that is very often found in the beginning of paragraphs, but never in the end?

Articles is one ready suggestion. As we see, it does not withstand critique, unless perhaps, a given article is a frequent n-gram per se. E.g., in English "the" is an article, and also we have "the" in "they", "then", "breathe" etc.

What else?

Capital letters?
Digits?

---

There is also a curious peculiarity of which I'm not sure if it was ever discussed or not.

I call it the gallows' "coverage", and it is observed both in paragraphs (in the beginning and within) and in labels. It is observed with both plain and benched gallows, and sometimes in a mixed way.

What is it?

Imagine, for instance, a p. Then have a look at e.g. the beginning of f10v, which is either paiin.daiin or paiindaiin (does not matter). The leading p is not a simple p, but a p with its loop curiously extended rightwards, thus extending over (or "covering") aiin.d (or at least aiin). One may argue that this is simply an embellishment. OK. Have a look at f10r. The leading p covers chocthy.shor; the next two p's do not seem to cover anything, or probably they cover only the subsequent c. The fourth p, however, clearly covers two characters - ch. Is it embellishment, or we have an additional layer of complexity here?

Next, recall those strange t's with their legs in different places (denoted "weird" in my title post). An example would be seen in You are not allowed to view links. Register or Login to view. which begins (presumably) with otchal.chchsty, but there are no two t's, but one t with its left leg in the first word and its right leg in the second. This can be considered as the t covering chal.chchs.

Consider You are not allowed to view links. Register or Login to view. which begins with foar. Is not a loop in f  restrictedly narrow? Well, it looks so - so as not to cover the following oar (which suggests that the loop may have been added to the vertical line after oar was already in place).

Note that very often there is a "semi-space" after the plain p and f gallows, which is hard to decide upon whether it is a Voynichese space or not. For example, see the same f6r, last p in line 1. There is a somewhat increased spacing after chop and before chol, which makes us wonder if it is chopchol or chop.chol. The idea of "coverage" suggests that it is chopchol, only the increased spacing after p is intentional to show that p does not cover the following c.


There are still more complicated examples of "coverage", such as those with additional loops (see e.g. f68r1 or the beginning of f95v1). In the terminology of my title post, those are gallows both "embellished" and "weird".

So it looks like that both additional loops and "coverage" matter. So are we back to the "concept"? For what except a "concept" can embrace or "cover" a sequence of subsequent (sorry!) characters?

That is a very interesting idea and one that deserves to be more closely examined. I shall dedicate a bit more time to this later on.

OTTOMH, your "coverage" idea would link the successive glyphs with the gallows, thus clearly indicating a "glyph group" within a word.

Possibilities:

• Reference markers (Pelling's book reference idea?) where the gallows is the book, the underlying glyphs the marker (and the rest of the uncovered word talking about the purpose of the marker)
  

• Some sort of artificial language marker where the gallows indicates say, tense or grammatical function out of tune with the rest of the sentence, and the underlying glyphs the changed glyphs. 
  

• It could be something mad to our eyes, like "holy words", or pronouns.
  
• etc
  

Using mathematical terminology, some "function" or "operator" is what comes first to my mind. However the fact that the gallows with "no coverage" are often met with within words stands against this theory.

Using mathematical analogy again, we can have some function f(x), f(x*y*z) and so forth, but we cannot have a functon of nothing.

Further, suppose there is some word p**** (where * stands for any character), transcribed irrespectively of the "coverage" (as all current transcriptions do). This sequence then well may be found with p "covering" different number of subsequent characters or even "covering" no characters at all. Currently, they all would be considered as one and the same Voinichese word. Should they be considered different instead, and all the stats re-calculated?

I fear I don't follow the function example, which would be an anachronism for the time era in any case. 

Possibly it's a bit early to be suggesting a new transcription effort

I use the term "function" not applied to the era and mathematical developments of the era, but as the general notion.

Basically, function is a rule that associates or "maps" inputs (arguments) to outputs (values). If you have no arguments, you have no function values. Likewise, if the presence of a "covering" gallows character indicates that some rule should be applied to the subsequent characters being "covered" (e.g. Voynichese "pABC" means plain text "XYZ"), then there is no answer what does a gallows character mean or indicate when it does not cover anything.

OK, I thought you meant some sort of mathematical function as we understand it. I understand the basics behind the concept, I do have a degree in maths.
There is an answer to what a gallows means when it doesn't cover anything - it means something else.
Stop thinking logically and modern with these things, it doesn't fit into a medieval / Renaissance mindset.
Let us assume p means "I" or "Us".
So short p means "I hold this stone"; long p means "We (God) hold / power this stone" which is a completely different scenario under the same number of words.