A key to understand the VMS

[Torsten]

I never said that there was no word network.  The question is whether such a network demonstrates that the VMS is not a natural language and if not, then how to account for it.  I have proposed both that it could be due simply to phonotactic rigidity, or due to a language in which most or all words are compounds of two or three basic elements.  Obviously we don't know enough about how the language works to decide which if either of these possibilities is correct (and it could be a combination of factors).  I am, however, fairly satisfied that there's nothing about the word network that suggests we're looking at something other than a natural language text.

My main argument against language is still that that beside the repetition of similar words no word order exists for the VMS. As long as you didn't know if the text of the VMS represents language or not a question like "how the language works" is build on something you didn't know.
Anyway the topic of this thread is "A key to understand the VMS" and not "the VMS is not a natural language". Therefore the question is: What does the network of similar words mean for the VMS? And if you assume that the VMS represents language the network of similar words means at least that you can't "come up with similar examples by using noun declensions or verb conjugations in many languages (including German)" [see You are not allowed to view links. Register or Login to view.].

What's more likely is that how the words are structured has influenced the design of the script.

No. For instance 'e' is common after 'k' and 't' but not after 'p' and 'f'. Words like 'peShol' and 'feeedy' are unusually rare whereas words like 'keeedy' and 'teeedy' are common for the VMS.

A second example is  that a gallow glyph following 'l' is most likely a glyph 'k' or 'f' but not a 't' or 'p'. An example for a word using 'lk' is 'olkchedy' and an example for a word using 'lp' is 'olpchedy'.

With other words the shape of a glyph has some influence to the selection of the next glyph.

I don't see how any of these points contradict the idea that the structure of the language has influenced the design of the script.  Actually this looks like evidence in favor of the idea.

See the sequence [oltar olkaiin] in line f78v.P.27. Since the [l] is larger then an [o] or [e] there is less space between [l] and [t].
   

In my eyes it makes sense to prefer [k] over [t]  if there is less space available. Therefore it makes also sense in my eyes to prefer [t] and [p] if enough space is available as for instance in line initial position.

A similar observation is that of Currier: "'Words' ending in the [y] sort of symbol, which is very frequent, are followed about four times as often by 'words' beginning with [qo]" [see You are not allowed to view links. Register or Login to view.].

Now, how many of these properties that you have mentioned can be found in sample texts generated by your auto-copying code?  My guess is: none.

It is not necessary to guess here. I have published an App to simulate the auto-copying hypothesis. See for instance my post from 08-02-2017, 07:23 PM in the thread. In fact all of the properties I know can be explained with the auto-copying hypothesis. 

Yes, as I've already stated I've tried your app and found the text it produces to be nothing like the actual VMS text.

The purpose of the app is to simulate the auto-copying hypotheses and to generate a text with statistical ratios (entropy, random walk etc.) comparable to the VMS.

I'll ask once more: where are the word grids produced from texts generated using your auto-copying code?  Can you show that they are complete with no gaps, like the ones you have made for the actual VMS text, or not?  If you can't make such grids then I don't see how your auto-copying method can account for even what you seem to think is the most important property of the text, let alone the many other properties that I and others have pointed out.

Since the app is generating new words by modifying existing ones no other outcome is possible then a network of similar words:

   


  VoynichGenerated.gephi.zip (Size: 115.51 KB / Downloads: 59)

This is the text I have used as source for the graph:

  voynich_generated.f103r.P.18.txt (Size: 62.21 KB / Downloads: 45)

The first line of the text [polchedy qokeol okain checThy oteey lShedy okain qokain] is used ad seed and is copied from line f103r.P.18 in the VMS.