RE: Voynich Reconsidered
kckluge > 28-01-2023, 02:53 AM
Robert,
I'm not saying this to pick on you, rather it seems an opportune time to make a significant point. In your latest Goodreads post you refer to ,"the two main strands of debate in Voynich research: namely, plaintext versus ciphertext, and meaning versus nonsense." With respect to that first strand:
From the point of view of bulk statistics of the text (entropy, etc.), any plaintext in some natural language written in some script is, in fact, a ciphertext -- it's a text that has been encrypted with a Caesar cipher with shift 0. If the Voynich text doesn't have the characteristics of something encrypted with a simple substitution cipher, then it also doesn't have the characteristics of a plaintext. The hypothesis that the Voynich text is a plaintext written in an unfamiliar script is nothing more than a special case of the hypothesis that it is a plaintext enciphered with a substitution cipher using an unfamiliar script. There is no "strand" to be had.
I don't say that to dismiss or ignore issues such as possible abbreviations or whether it's a phonetic rendering of an unfamiliar language, but to make the point -- taking such issues into account -- that from a statistical POV you can't meaningfully distinguish plaintext in an unfamiliar script from a substitution cipher using an unfamiliar script. They are equivalent hypotheses.