Discussion of "A possible generating algorithm of the Voynich manuscript"

[Torsten]

But then would a medieval scribe have the patience and/or the motivation to generate it?

My answer to your question is that it is the most effective way to generate some dummy text. Imagine that someone in medieval times wanted to write a mysterious book. His goal is to write a book nobody is able to read. What is the best way to generate manually large amounts of dummy text?

A) One possibility would be to copy some text using an exotic script and language. For doing so the writer has to find enough text using this exotic script. But the danger would be that someone knowing the script comes along and reads the text. If this happens the mystery is solved and the book of secrets would be no longer mysterious anymore.

B) Another possibility would be to invent gibberish by generating new words. But to invent something new is never an easy task. This is even true for generating gibberish out of the head. If we take the generation of passwords as an example we can find out that the most common passwords are easy to guess. Most commonly used are words like 'password', 'qwerty', or '12345678'. This passwords are used since it is far easier to copy or to repeat something already known than to invent something new.

C) A third possibility would be to copy some existing text and to obfuscate the copying process by adding various small changes. In a similar way someone writing a plagiarism might purposefully add some changes and mistakes while copying the source text.

The self citation method is a variant of method C). The main difference is that the source words are chosen from previously written lines of text. This method is not only very efficient. There are also further advantages. Since the idea of this method is to modify copied words it is impossible to misspell a word. Moreover it is also possible to generate words of a given length to fill out the lines to the margin. This is exactly what we observe in the case of the VMs. There are no corrections and the text perfectly fits into its margins. 

What is now the most effective way to fill a page with some text? The easiest way to test all three methods is to generate some sample texts. This type of experiment could be done by writing three pages full of text and measure the time needed to do so:

In case A) copy some text from an external source. Copying errors are not allowed.
In case B) use arbitrary letter sequences without copying a previously written word and without repeating the word generation method. Copying errors and repeated words are not allowed.
In case C) start with a given word or phrase like 'some repeated words' and copy previously written words while replacing at least one letter with a different one. Copying errors and repeated words are allowed in this case.