- Options
- Login
- Desktop Style
Options
The VMS as a possible chain encryption ( Mod 23 ).
- The VMS as a possible chain encryption ( Mod 23 ).
-
RE: The VMS as a possible chain encryption ( Mod 23 ).
MarcoP > 25-07-2025, 07:16 AM
(24-07-2025, 05:51 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.For me, the question arises as to whether certain text characteristics (e.g., word repetitions or repetitions with minor variations) depend heavily on the original nature of the text or not. For example, a poem should produce a different output than a medical free text.
A poem explains "line as a functional unit", since verses tend to coincide with sentences. But word repetitions and (more clearly) character entropy are typically not very different in poetry than in prose. A poem with interesting repetitions is the Finnish Kalevala.
You are not allowed to view links. Register or Login to view. -
RE: The VMS as a possible chain encryption ( Mod 23 ).
oshfdk > 25-07-2025, 12:25 PM
(24-07-2025, 05:51 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.What kind of encryption could possibly fulfill all these conditions in one go?
I don't think it's very hard to build one, unless it also has to be practical. Given a large enough mapping and using chain encryption, you can create very plausible Voynichese from any language with about 2.5 Voynichese glyphs per plaintext character.
To give you a very basic example, suppose we want to create a cipher that encodes "WHAT KIND OF ENCRYPTION COULD POSSIBLY FULFILL ALL THESE CONDITIONS IN ONE GO" as "fachys.ykal.ar.ataiin.shol.shory.cthres.y.kor.sholdy...".
We say that the first letter is arbitrary, and only affects the encoding of the following letters. Then we create the following rule set:
"ach" is "W" when it is preceded by "f",
"ys" is "H" when it's preceded by "h",
"yk" is "A" when it's preceded by "s",
"al" is "T" when it's preceded by "k",
"ar" is "K" when it's preceded by "l",
"at" is "I" when it's preceded by "r",
"ai" is "N" when it's preceded by "t",
"in" is "D" when it's preceded by "i" (so all "aiin"s will be ?D, and all "taiin"s will be ?ND),
"sh" is "O" when it's preceded by "n",
"ol" is "F" when it's preceded by "h",
"sh" is "E" when it's preceded by "l",
"or" is "N" when it's preceded by "h",
"ycth" is "C" when it's preceded by "r",
"r" is "R" when it's preceded by "h",
"es" is "Y" when it's preceded by "r",
"yk" should be "P", but this is the first instance where our encoding breaks down, because yk preceded by "s" was defined to be "A", and we need "P" here. So, instead of "yk" we should use something different.
But we have successfully encoded "WHAT KIND OF ENCRY" as "fachys.ykal.ar.ataiin.shol.shory.cthres". If needed we can create a mapping that would produce funny qokeedy cycles and chol repetitions from ordinary English text.
You can implement this encryption as a huge table, something like the small version below. By allowing multiple variants in each cell for "narrow width/lot of vertical space/short label" versions you can explain many LAAFU features. The table is built in such a way that no code starts with the same glyph as the title glyph of the column, hence no duplicated glyphs. Relatively small number of codes, compared to the size of the alphabet, can explain low entropy, increasing to normal levels for longer sequences. Also this particular version is space-sensitive, r (y) and ry are two different codes.
But I don't think this encryption is practical for a large book. To use the full Latin alphabet a table of many hundreds of cell should be used.
-
RE: The VMS as a possible chain encryption ( Mod 23 ).
bi3mw > 26-07-2025, 12:47 PM
It is interesting to see that your approach is also a chain cipher. Extensive mapping avoids the problems that inevitably arise with my approach (entropy, letter distribution, last letter in the previous word, etc.). So I have to conclude that my method does not stand up to scrutiny, although the idea that a simple process could work was very tempting. -
RE: The VMS as a possible chain encryption ( Mod 23 ).
oshfdk > 26-07-2025, 01:22 PM
(26-07-2025, 12:47 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.It is interesting to see that your approach is also a chain cipher. Extensive mapping avoids the problems that inevitably arise with my approach (entropy, letter distribution, last letter in the previous word, etc.). So I have to conclude that my method does not stand up to scrutiny, although the idea that a simple process could work was very tempting.
My idea of the ideal cipher system for the Voynich MS is something that has the right statistical properties and at the same time is very simple to read and write, ideally allowing one to read off the page with nearly normal reading speed. So, a simple process that could work would be very interesting. -
RE: The VMS as a possible chain encryption ( Mod 23 ).
bi3mw > 27-07-2025, 01:22 PM
I am already looking forward to @magnesium's contribution on Voynich Manuscript Day. He wants to present a substitution cipher that encrypts Latin and Italian as VMS-like ciphertext.
I would guess that the method is similar to that of @oshfdk (see post You are not allowed to view links. Register or Login to view. ). At least, extensive mapping will be almost impossible to avoid. -
RE: The VMS as a possible chain encryption ( Mod 23 ).
magnesium > 27-07-2025, 03:15 PM
(27-07-2025, 01:22 PM)bi3mw Wrote: You are not allowed to view links. Register or Login to view.I am already looking forward to @magnesium's contribution on Voynich Manuscript Day. He wants to present a substitution cipher that encrypts Latin and Italian as VMS-like ciphertext.
I would guess that the method is similar to that of @oshfdk (see post You are not allowed to view links. Register or Login to view. ). At least, extensive mapping will be almost impossible to avoid.
Thanks! Yes, extensive mapping is impossible to avoid, and I completely agree with @oshfdk that to reconcile the properties of the VMS with most natural languages, you need 1 plaintext letter often mapped to a string of 2-3 Voynichese glyphs. There is a lot to like about the chain cipher approach, but I didn't pursue that exact strategy for the cipher I will be presenting at VMD, in large part because I wanted to see how far a more "vanilla" approach to verbose homophonic substitution could be pushed.