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

[Torsten]

Since the mapping matched word frequencies in the two texts (Machiavelli and VMS) we basically imported the VMS network of similar words into Machiavelli, which already started with a very similar TTR. For instance, one of the properties mentioned by Torsten is that in Voynichese word types that occur frequently also have the highest number of similar word-types: this mapping makes so that the encoded Machiavelli shares this property.
In the original Italian, there are 2 words with edit distance 1 from "che" (the second most frequent word); in the encoded version, "che" is mapped to "ol" and there are 38 similar word types.

Indeed, to search for similar words in the Voynich manuscript is like searching for trees in the forest. There are far too many of them. It would be much more interesting to count the number of similar word types.

In 2014 I have executed a test for rarely used words: "In this paper all words occurring seven times and all words occurring eight times are used as two separate control samples. ..." (Timm 2015, p. 12ff).

For executing this test it was necessary to define what 'similar' means and also what 'near to each other' means:
"To consider the peculiarities of the VMS script, the edit distance is defined as follows: If a glyph is deleted, added or replaced by a similar glyph, this is counted as one change. Also, the change from 'ee' into 'ch' or from 'eke' or 'kch' into 'ckh' is counted as one change. If a glyph is replaced by a non-similar glyph, this is treated as deleting one glyph and adding another glyph. This is counted as two changes." (Timm 2015, p. 6).

"Near to each other means within a range of three lines before and after a word, and similar means that it is possible to transform them into each other by changing three or fewer glyphs (edit distance <=3)." (Timm 2015, p. 13).

Could shapchedyfeey in You are not allowed to view links. Register or Login to view. result from sho.pcheey.pchey in f8r?

For testing a rarely used word I find it useful to check the transcription first. In this case only Takahashi transcribes this token as <shapchedyfeey>. Currier and the first study group have transcribed it as two tokens <shapchedy> and <fchy> (see You are not allowed to view links. Register or Login to view.). In my eyes 'ee' is indeed a 'ch'. I would therefore transcribe this token as <shapchedyfchy>. 

The most similar word to 'shapchedy' on page You are not allowed to view links. Register or Login to view. is <chypchey> in f26r.P.1. Therefore I find it more likely that <chypchey> was used as source word for 'shapchedy'. But maybe <sho.pcheey.pchey> on page You are not allowed to view links. Register or Login to view. was used as source for generating <chypchey>.

The sequence 'pchey' in <chypchey> is also similar to 'fchy'. Therefore <chypchey> is also a possible source for the sequence 'fchy'. It is quite common that a source word was used twice. See for instance the tokens <pcheey> and <pchey> in <sho.pcheey.pchey>.