+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Theories & Solutions (https://www.voynich.ninja/forum-58.html)
+--- Thread: Help with Arabic/Syriac Reading of F1R (/thread-5516.html)
Pages:
1
2
RE: Help with Arabic/Syriac Reading of F1R - 008348dc760f858fd668476b75fb6f - 09-04-2026
(09-04-2026, 04:09 PM)rikforto wrote: You are not allowed to view links. Register or Login to view.(09-04-2026, 07:53 AM)008348dc760f858fd668476b75fb6f wrote: You are not allowed to view links. Register or Login to view.I think I can fairly convincingly demonstrate "readability" of the VMS under the context of a pharmacological reference document in the vein of the Kitab al-Azmina and I hope to have that with the community in the next few days. The You are not allowed to view links. Register or Login to view. reading is a side quest, it's a request to the community for falsification while I also work with Arabic scholars to do the same.
One of the many problems with AI "solutions" is that LLMs don't understand Popper, so it guides people to do stuff like ask the audience to prove a hypothesis wrong rather than the person putting it forward to nullify the null hypothesis.
While not super formally, several people have now given you what you need to piece together why substitution based on the EVA alphabet is not likely to lead people to rejecting the null hypothesis, and the knowledge to know attempting this style of solution *very* well-tread ground. In short, the statistical properties of Voynichese letter ordering are very different from most natural languages, including Semitic languages, and at any rate, the EVA alphabet was not chosen with any insight into the underlying letter values.
Further recourse to an LLM is not going to help you overcome these problems. Because the answer is neither in the LLM's training set nor based on known emergent phenomena, whatever ML process you are using is not going to result in a convincing solution
But that's not what I've said or what I'm claiming, I understand the reservations on the forum about ML use, about the underlying properties of the text and why it doesn't align with "language". To restate my point:
The remainder of the text doesn't function as a "text" it's separate from f1r. If any folio were to contain a readable text then it would be f1r, it behaves quite differently to the rest of the MS. The remainder of the text is (under this theory) a compression of reference tokens encoding plant components and their Galenic qualities, the methods for preparation, the agents of preparation, the produced compounds, the treatments, the ailments they treat... but that's a whole separate thread, for this post I'm focused on You are not allowed to view links. Register or Login to view. as it doesn't encode tokens in the same way.
The statistical properties we talk about are a property of the whole MS, this thread is working on the proposition that You are not allowed to view links. Register or Login to view. behaves differently to the rest of the MS and **might** contain something closer to a narrative form. I don't think anyone will agree to that proposition until I get a wholesale explanation out as to how and why I arrived at that conclusion, but that is not what I am trying to do here. I'm literally just asking people who have experience in Arabic and Syriac to cast their eyes on a rendering and either thumbs up or thumbs down it. If it's broadly a thumbs down then I don't need to write up the tortured path to how I separated the tokens into raw ingredients, preparatory compounds, preparatory agents, treatments, ailments and why that allowed me to isolate You are not allowed to view links. Register or Login to view. as "different".
If you want to have a debate about the statistical properties of the MS then fine, there's a place for that, if we want to debate whether LLM/ML tech is every useful in any context then sure, lets have that talk. But my sole aim was to catch the eye of a rare kind of person with experience in Syriac who could look at the text and say "nah it's a giant pile of garbage" or "kinda looks ok".
RE: Help with Arabic/Syriac Reading of F1R - DG97EEB - 09-04-2026
RE: Help with Arabic/Syriac Reading of F1R - tavie - 09-04-2026
RE: Help with Arabic/Syriac Reading of F1R - rikforto - 09-04-2026
RE: Help with Arabic/Syriac Reading of F1R - 008348dc760f858fd668476b75fb6f - 09-04-2026