Implications of the continuity of Currier "languages"

[ReneZ]

The earlier post I wrote on this topic was just a very brief one, the tip of the iceberg so to speak, and not in any way systematic. That is in any case not possible in the frame of a blog post, but it is possible to expand a bit.

All possible options of what the Voynich MS text means, and how such a continuous variation of the language statistics could arise, can be subdivided into groups in several different ways. For example:

Choice 1: text meaningful or meaningless?
For the sake of the argument, let's assume that the option 'meaningful' means that there is a some source text which has been converted in some way to result into the Voynich MS text.
For the 'meaningless' case, either the text comes out of nowhere, or is somehow based on an existing text that was used as a 'model'. This second option is more interesting as this would explain the consistency and the language-like properties of the MS text.

Choice 2: the variation comes from a variation in the source text, or from a variation in the method of conversion.
We don't know this at all. This could also apply to the meaningless case, in case some model text was used.

Choice 3: the variation is in the meaningful part of the text, or is part of meaningless filler.
For this it is interesting to note that the B language tends to be more verbose, and also that frequent A language words exist also in B text, but this is not so much the case in the other direction
(Note that this is not really proven yet, but is an impression)

Choice 4: either A and B language words are different encodings of the same source text word, or, they are different source text words.
This is one of my favourite questions about this topic. I don't know the answer (of course).
This is also highly correlated with Choice 2.

Choice 5: the text is the product of a single person or of 2 (or more) persons
The main difference here is that for a single person, such a large variation would probably imply a longer creation time.

This gives rise to 2**5 = 32 combinations, but of course not all of them are meaningful.
About probabilities I don't even want to begin.....

Let's look at all this with an example of how the text could have been generated. This is the 'stroke encoding' that was once proposed by Elmar Vogt. Basically, it means that every component that makes up the written source text (horizontal lines, vertical lines, curves, circles, serifs, etc etc,) is reflected in one Voynich character.
(This may not seem a very realistic encoding method, but who knows: maybe the originator was very clever and wanted to devise a writing system which which all languages and scripts could be represented: Latin, Arabic, Hebrew etc.. Probably anachronistic, but it is just a model to be used as an example).

So, now the source text to be 'converted' is of course not a printed text, but a handwritten one.
This would mean that a text in a humanist hand could result in a different output than the same text written in a gothic hand. A source text with a variation in handwriting style could result in a variation of the output 'language'.  Even, a source text with a discontinuity in the handwriting style could cause a gradual (but fast) change in the encoding, as the encoder gradually adapts.
Different people might break up the source characters into components in different ways.

The last case would be an example where Currier-A words and Currier-B words map back to the same source text words, but perhaps not in a trivial manner.

Anyway, this can be continued ad libitum.