+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Analysis of the text (https://www.voynich.ninja/forum-41.html)
+--- Thread: Bifolio as a functional unit? (/thread-5131.html)
RE: Bifolio as a functional unit? - Aga Tentakulus - 15-12-2025
@Lisa
I have no idea why the translator wrote ‘leaflets’; I mean the foldouts.
But to explain my train of thought.
I'm starting from the rosette side (black).
Folded sideways (brown). Double book height and triple book width.
If I were to divide the sheet in the middle, I would get two identical sheets with 3 identical compartments (orange). (2x3).
If I cut two normal pages (green), I get one remaining page (red). (3x2). However, this is longer than page (green) but shorter than page (orange).
I think he picked up the wrong sheet here. (red) instead of the whole sheet (orange).
That's probably why he pushed the zodiac circles so close together. Too close for three rings.
Why not two rings and cut off the rest?
Was it a coincidence, or did he plan for all the fields to be the same size?
Just a thought experiment.
Translated with DeepL.com (free version)
RE: Bifolio as a functional unit? - dashstofsk - 15-12-2025
(13-12-2025, 02:22 PM)LisaFaginDavis Wrote: You are not allowed to view links. Register or Login to view.104v and 115r (conjoint) are more closely related than, say, 104v and 105r (consecutive).
If it is of help to anyone here are some of my plots of correlations between pages in quires 13 and 20. Some plots show closely related pages. Others do not. One thing that stands out is f105v. This page seems odd. It is poorly correlated with all other pages in all plots.
RE: Bifolio as a functional unit? - Bernd - 15-12-2025
Could you elaborate on what we're seeing in the plots? I can read the filenames but what are the axes and where is You are not allowed to view links. Register or Login to view. ?
RE: Bifolio as a functional unit? - dashstofsk - 15-12-2025
From the top you have the 20 pages of quire 13 followed by the 23 pages of quire 20. The black line separates the quires. You are not allowed to view links. Register or Login to view. is the 6th line in quire 20. Brighter colours show where there is better correlation. Darker colours show where it is worse. You can ignore the columns on the right.
RE: Bifolio as a functional unit? - nablator - 15-12-2025
(15-12-2025, 11:49 AM)dashstofsk Wrote: You are not allowed to view links. Register or Login to view.Brighter colours show where there is better correlation.
How do you define correlation? I tried one method as far as possible from LSA similarity to see what happens.
The idea is to test various methods to find out what makes the studied property appear (a "very strong textual correlation across conjoint bifolia" in Q13 and Q20). LSA similarity is a black box, it does not help understand what matters exactly and what does not: word order or proximity? Sub-word tokens (BPE segmentation or prefix-stem-suffix decomposition by some grammar)? Glyph bigram/trigram statistics?
RE: Bifolio as a functional unit? - dashstofsk - 15-12-2025
I searched for agreement in a number of parameters: whole words, prefixes, suffices, character pairs, character ends ( first character of word + last character ), bridge pairs ( last character of word + first character of following word ), 3 characters in a row, and used a correlation method similar to what was given in You are not allowed to view links. Register or Login to view.
RE: Bifolio as a functional unit? - DG97EEB - 23-12-2025
Hello
Fascinating analysis.. Have you looked at Cosine similarity based purely on token commonality? Somewhere I read that within the herbal pages there are some natural clusters outside the current bifolio ordering.. I'm not a statistician so want to stay in my lane
Thanks
Ed
RE: Bifolio as a functional unit? - dashstofsk - 23-12-2025
(23-12-2025, 12:21 PM)DG97EEB Wrote: You are not allowed to view links. Register or Login to view.Have you looked at Cosine similarity based purely on token commonality?
Me? No I have not done this. I am not quite sure what this is or what it would demonstrate.
RE: Bifolio as a functional unit? - nablator - 23-12-2025
(23-12-2025, 12:21 PM)DG97EEB Wrote: You are not allowed to view links. Register or Login to view.Fascinating analysis.. Have you looked at Cosine similarity based purely on token commonality?
I'll try it with word frequency vectors, it should be interesting. I'm trying to figure out what makes You are not allowed to view links. Register or Login to view. similarity special. The usefulness of the SVD is totally unclear to me, maybe all this linear algebra is overkill for the purpose of comparing pages. Note: frequencies may need to be adjusted with tf-idf because the most frequent words are usually less meaningful than rare words.
RE: Bifolio as a functional unit? - DG97EEB - 23-12-2025
(14-12-2025, 03:37 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.(13-12-2025, 02:22 PM)LisaFaginDavis Wrote: You are not allowed to view links. Register or Login to view.In other words, 104v and 115r (conjoint) are more closely related than, say, 104v and 105r (consecutive).
Yes, but there are 4 exceptions in Q20 (4/10 versos are not more closely related to the conjoint recto than to the consecutive recto), not a "very" strong textual correlation across conjoint bifolia. Probably there is a better metric than the one I'm using. And, of course, the result would be different if the bifolios were reordered, so there may be a way to make consecutive pages more closely related.
Comparison of the vocabulary (set of words) of Q20 pages using the RF1b-er transliteration (without '?' words, with all spaces kept, i.e. all '.' and all ','). The % is the You are not allowed to view links. Register or Login to view. coefficient.
Apparently this method is even more precise than the Cosine similarity, though my Cosine similarity does find natural groupings in herbal