(24-04-2026, 09:14 AM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.@oeesordy:
On the entropy point: I've tested that against actual MHD corpora. "Ortloff von Baierland" shows 75.4% Levenshtein-1 connectivity, "Breslauer Arzneibuch" 72.8% — basically the same range Timm claimed was unique to artificial text.
So "low entropy / many word repetitions" isn't diagnostic of a hoax - it's what you get with german medieval medical texts in general.
As for the conlang-from-substitution idea: that would require an author inventing a language with consistent morphology and syntax. A medieval scribe around 1430? That seems rather unlikely.
And I don’t want to discuss modern forgery theory in this thread, because that’s a completely different topic and would distract too much from the actual subject… I ask for your understanding. 
I guess we shall see with all those translated German/Bavarian paragraphs from Michael Voynich's tomb.
I created a heatmap that shows the distribution of letters for each starting marker of a line. It is quite clear that there are certain glyphs that respond to the first letters, but also that other letters are distributed very evenly. The conclusion is the exciting question. For me (!), two theories are currently competing with each other:
- There is a second layer of variation in the lines; they are, so to speak, "partially patterned." The effect of this is actually very clearly visible in the following heatmap. Looking at it, one could roughly say that one-third to one-half of the glyphs are subject to at least a second layer of variation.
- Of course (and I know this has already been discussed), it could just as easily be that "p" and "f" (if they aren't actually just a single glyph) are something like a zero glyph, and thus cph as well. And the rest of the anomalies could be due to different types of lines: labels, descriptions, lists, etc. That is also conceivable, and at the moment I'm leaning very slightly in that direction.
------------
So, with that, I'm now concluding my investigation into the line-start markers. I did find a lot of other statistical anomalies in this context, but they aren't "spectacular" enough to publish here - the significance gets lost in the mass of data.
I apologize again if I've published something already known, but since threads here tend to fade into the background over time, it might still be interesting for newer researchers. Thank you for your attention. If I happen to find anything else related to this, I'll post it.
Well, so now I'm turning my attention back to my Bavarian thesis - my theory needs some adjustments after these findings. And I'm now going to think about how to weave what I've found into a masterful (haha)
body of work....
[
attachment=15326]
Note: the heatmap excludes the first token of each line from the count to avoid self-bias.
Not that it happens a lot, but shouldn’t cfh also be on this chart favouring follow initial lines?
@ Grove
The data set was too small. But in principle, if p and cph were zero glyphs and f = p, then cfh would also have to be a zero glyph. Maybe I should have included that.
oeesordy, please keep your theories out of other threads. You are free to discuss these in your own thread.
@ Koen
Are you talking about me? Because I mentioned that I’m focusing on Bavarian theory again? I didn’t discuss it or defend it, did I?
And I was the one who started this thread, right?
Sorry, I didn’t know I wasn’t even allowed to write that in my thread...