The Voynich Ninja

Koen: Yes, that's the kind of idea.

Imagine someone invents a cipher which they believe exhibits Voynichese like properties, now if they expressed this cipher using the Voynich symbols people will naturally be inclined to conclude that they are saying that this is the precise cipher.

I think the purpose is to, in broad general terms, construct a cipher which exhibits the overall general features we see in Voynichese without being tied to the very specific details of Voynichese.

For example one could invent a cipher which shares the general entropy and other statistical properties we see in the Voynich without being forced to demonstrate a one to one correspondence of symbols with the Voynich text. It creates a much better framework in which to experiment with general ideas without getting stuck in the minutiae of the precise Voynichese symbol combinations etc.

Given at this stage there is a huge degree of disagreement in the basic fundamentals of what Voynichese is it doesn't feel like the time for me and maybe others to get stuck in the minutiae, but rather try to establish the general nature of the beast.

Koen: I have my own ideas of what Voynichese is and I think for me working with a non-Voynichese script is a better way to analyse and develop ideas. At this stage I don't feel ready to attempt a decryption.

Alright, fair enough. In that case I would just use Latin letters, but whatever works the best for you.

Koen: There are definitely advantages of using latin letters, but I think also disadvantages. As discussed before it is perfectly possible to use an invented font which maps to a standard font, giving one the best of both worlds.

Quote:Mark Knowles: "Koen: There are definitely advantages of using latin letters, but I think also disadvantages. As discussed before it is perfectly possible to use an invented font which maps to a standard font, giving one the best of both worlds."

So you want something like what I uploaded in Post #14?

It is an invented font and when I inserted text from the Takahashi transcription, this is what resulted (it maps to a standard character set, in this case the EVA Latin letters).

JKP: Yes, I think that is more like the kind of thing I am talking about.

(18-04-2019, 10:44 PM)Koen G Wrote: You are not allowed to view links. Register or Login to view.Alright, fair enough. In that case I would just use Latin letters, but whatever works the best for you.

Yes, I do not know what speaks against using Latin letters. If I understood correctly, it is about understanding how to construct a cipher that has the known textual properties of the VMS.The easiest way to do that is using Latin letters.

In my opinion, the result at an early stage of development does not even have to look like VMS text. To see if there is a chance of success, the basics must be right first. -It's about to grope one's way.

The disadvantage of using latin characters is similar to the disadvantage of using Voynichese characters. Latin characters already have a specific meaning in the context of the cipher, which could lead to some confusion.

If one constructs a cipher system that is intended to emulate the Voynich MS text, but it isn't really *that* close, then, when using the Voynich alphabet it will be obvious that it isn't close. When using an invented alphabet, it won't be immediately obvious.
I wonder if this is the perceived advantage.

One of the big disadvantages of creating an invented alphabet to represent Voynich text is that we don't know which characters might be ligatures and which might be abbreviations. These forms were normal in the middle ages, so there's a strong possibility that they exist in Voynichese.

Is a gallows character one character? Two? or three? Does a tail represent an abbreviation?

If you create the character so it has some of the properties of the actual character, then you're back to basic Voynichese. If you don't include these possibilities, then you may be losing information during the transliteration process.