The Voynich Ninja

Hey all, I'm just making this thread to organize my approaches to decoding the manuscript.

I've got two angles of approach:

Method 1) Figuring out the grammar of the manuscript. If I can figure out that the manuscript has conjugations or declension patterns, I can start to narrow down the underlying language.

Method 2) Figuring out what characters act like vowels and which act like consonants. Or: trying to figure out if some Voynich characters represent multiple sounds (like Latin abbreviations).

Hi Dolokhov,

Before starting decomposing the VMS in that way, you probably may wish to solidify the assumption that VMS is written in plain text of some kind of natural language - which is not an easy task at all.

There was a good discussion You are not allowed to view links. Register or Login to view., not to mention efforts of those before us

In the spirit of collaboration, here are two things I've noticed:

1. o can appear next to almost every other Voynich letter and frequent combinations (and it is the only Voynich letter that can do this):
oe, qo, ot, ok, or, oy, om, oi, oiin, of, op, os, oo, oa, ol, oSh

2. I made a chart of the qo- words. It is not a full chart, but it should be helpful to someone researching this topic:

(09-09-2016, 12:33 AM)ThomasCoon Wrote: You are not allowed to view links. Register or Login to view.In the spirit of collaboration, here are two things I've noticed:

1. o can appear next to almost every other Voynich letter and frequent combinations (and it is the only Voynich letter that can do this):
oe, qo, ot, ok, or, oy, om, oi, oiin, of, op, os, oo, oa, ol, oSh

2. I made a chart of the qo- words. It is not a full chart, but it should be helpful to someone researching this topic:

Thomas, I should probably explain what I mean when I say the spaces could be arbitrary or contrived (not saying they are, but that the possibility is there).

When I say arbitrary, I mean this:

  On ceup on atimet herew a sac ast leo nthem oun tain. (No particular pattern to the word breaks.)


When I say contrived, I mean this:

  On ce upon atime the re w as ac astle on the moun t ain. (Planning the word breaks so they fall before and after certain letters.)

Since contrived spacing is still relatively easy to read, add a couple of null characters, a common character like "o", and something that looks like a Latin prefix/suffix abbreviation like "9" and you get this:

On9 oce9 upon9 atime the9 ore9 ow as9 ac oastle ...


And then apply a simple substitution code using Voynichese characters, treating u/v/w as the same glyph, and you get something with many of the properties of the VMS—a lot of short words of approximately the same length with a large number of certain characters (characters that may be a combination of real letters and nulls, depending on position) showing up at the beginnings and ends of words by design.


I've said I lean fairly strongly toward the not-natural-language camp but the above process results in something rather like Voynichese, which is why I still think there's a possibility that it might be.


So, with reference to your chart, if it's a natural language with contrived spaces (especially if there are also null characters), lists of words with similar beginnings might not reveal as much as one would hope. If it's a constructed language, the lists are likely helpful.

You're definitely right that a constructed language could account for the changes (and patterns of changes) between similar-looking vords. That thought has also occurred to me, and I definitely understand why you believe it.

Just to mention, I theorized a different difficulty for encoding Voynichese with contrived spaces:

In my units chart from the other thread, let's say for example that /d/ is encoded in Voynichese as unit 22 (ke),  /e/ is unit 6 (ee) and /r/ is unit 17 (ey). If you want to write the word "der" in this setup, you would get keeeey. If you wanted to connect two <e>s with a bar, you could make kcheey (an actual VMS vord, by the way).

Now the difficulty is figuring out where one "unit" ends and the other begins. I believe this is why we have so many <eee> or <che> clusters in the VMS; I believe it is intentionally done to make the script harder to crack. Contrived spaces may help the scribe who knows the cipher system, but it doesn't help an outsider who is confused why <qo> is always at the front of a word and <aiin> is almost always at the end  .

I don't understand what the argument is for arbitrary spaces. Word structure is very strong and character position is often dependent on beginning, middle, and end.

Given that the writer of the manuscript was familiar with the concept of the word divider we must begin from the assumption that he was using it to divide words. There needs to be a strong argument contrary to this before we discard them as arbitrary.

Word structure is such a strong feature that it is one of the few problems in the text that we can easily think about. Starting by discarding it leaves us much worse off.

Going back many years, there used to be a statement about the word spaces along the following argument:

given that the label words should be individual words, and they primarily tend to appear in the main text separated by spaces, it seems that the spaces are real.

The problem is that I have never seen anyone really demonstrating this.
This would be a bit of work, but not too difficult to do and it might really tell us something.

All labels can be matched with a version of the main text from which all spaces have been removed.

One can then see how many labels are not found at all, and for the remainder whether there is indeed a preference for them to reoccur separated by spaces.


Depending on the result, the situation could of course not be entirely clear.

(09-09-2016, 10:22 AM)ReneZ Wrote: You are not allowed to view links. Register or Login to view.......

given that the label words should be individual words, and they primarily tend to appear in the main text separated by spaces, it seems that the spaces are real.

The problem is that I have never seen anyone really demonstrating this.
This would be a bit of work, but not too difficult to do and it might really tell us something.
.........

It's very easy to show this, perhaps I have even put it online somewhere. 
Not only me but also scholars that researched the word-length and word distribution showed that the words must be words as in a language.
If words are words (=vords are words) then spaces musty be spaces.
But, in order to understand and accept such analysis, one must be willing to look into the approach and understand the (simple) statistical theory and of course accept that theory. 
For what it is worth: my own conclusion: The VMS spaces are spaces.
(However there is one exception which i am not willing to disclose at this moment in my research.)

(09-09-2016, 09:57 AM)Emma May Smith Wrote: You are not allowed to view links. Register or Login to view.I don't understand what the argument is for arbitrary spaces. Word structure is very strong and character position is often dependent on beginning, middle, and end.

Given that the writer of the manuscript was familiar with the concept of the word divider we must begin from the assumption that he was using it to divide words. There needs to be a strong argument contrary to this before we discard them as arbitrary.

Word structure is such a strong feature that it is one of the few problems in the text that we can easily think about. Starting by discarding it leaves us much worse off.

Emma,
There are two main reasons why I believe spaces are arbitrary. Here is the bigger one:

In order to explain the "morphology" of a group of words that look similar, it seems you would need (at best) an agglutinative language like Turkish, or (at worst) a polysynthetic language like Cherokee, to create all the different forms:

qot
qoteo
qoteol
qotcho
qotchol
qotcheo
qotcheol
qotol
etc.

I have not heard a linguistic explanation that can account for all the variation of <qot> and <qok> in my above chart, but I would be open to hearing one from the people who support the "valid spaces" side.

Hi Thomas,
why should not an inflected agglutinative language be enough to explain this?

From Aeneides Book I (much shorter than the VMS):

qui
quibus
quicquid
quid
quidve
quies
quiescit
quietas
quietem
quietum
quin
quinquaginta
quippe
quirinus
quis
quisquam
quisquis


quo
quocirca
quod
quodcumque
quondam
quoque
quorum
quos
quove