I did a run where i grouped the text based on the image type, ignoring all other parameters,
This may be a mix of language A and languageB
$I (see below) Illustration type on this page (see below)
A Astronomical (excluding zodiac)
B Biological
C Cosmological
H Herbal
P Pharmaceutical
S Marginal stars only
T Text-only page (no illustrations)
Z Zodiac
You are not allowed to view links.
Register or
Login to view.
made a run splitting the imagetypes in A and B groups based on language
You are not allowed to view links.
Register or
Login to view.
Davidd, do you show anywhere which pages are included in the sections (Astronomical, Biological, Cosmological...) that you use in your analysis?
For me the surprising section is
quireT, which seems different from the rest, having the quality score much higher and similar to real texts.
You know, we often ask question if Voynich is meaningful or meaningless. Actually in theory the truth may be more suble - some sections may be meaningful while other not
And people in their research usually focused on the parts with pictures as they give some clues and didn't work much on plain walls of text.
(10-06-2025, 10:46 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.For me the surprising section is quireT, which seems different from the rest, having the quality score much higher and similar to real texts.
QuireT has about 10000 vords while the others about 1500, I wonder if this impacts the score. I must have also missed where he said that but it would be good to get an actual folio split for the quires/sections analysed.
I do think the fact that the 10000 vord quire scores so closely to the vulgata text and so far from the randomised genesis is a good indicator of some sort of meaning or at least a structured relationship between vords. I think the paper "gibberish after all" showed that Voynich behaviour could be coherent with a sort of asemic writing rather than a purely randomised text. I guess now we only need 38000 words of asemic writing to compare
(10-06-2025, 10:46 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.Davidd, do you show anywhere which pages are included in the sections (Astronomical, Biological, Cosmological...) that you use in your analysis?
For me the surprising section is quireT, which seems different from the rest, having the quality score much higher and similar to real texts.
You know, we often ask question if Voynich is meaningful or meaningless. Actually in theory the truth may be more suble - some sections may be meaningful while other not
And people in their research usually focused on the parts with pictures as they give some clues and didn't work much on plain walls of text.
Included the pages that were used in each analysis, noticed the index link doestnt work yet for the imagetype selections.
You are not allowed to view links.
Register or
Login to view.
(10-06-2025, 11:36 PM)davidma Wrote: You are not allowed to view links. Register or Login to view. (10-06-2025, 10:46 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.For me the surprising section is quireT, which seems different from the rest, having the quality score much higher and similar to real texts.
QuireT has about 10000 vords while the others about 1500, I wonder if this impacts the score. I must have also missed where he said that but it would be good to get an actual folio split for the quires/sections analysed.
I do think the fact that the 10000 vord quire scores so closely to the vulgata text and so far from the randomised genesis is a good indicator of some sort of meaning or at least a structured relationship between vords. I think the paper "gibberish after all" showed that Voynich behaviour could be coherent with a sort of asemic writing rather than a purely randomised text. I guess now we only need 38000 words of asemic writing to compare
It might also be because the quire analysis only uses 7 groups, where I think based on the analusis run on genesis shows that 18 is closer to the reality of indo european languages.
I can do a run with 18 groups tomorrow, see what the score is then.
….so is replacing m for iin showing that it helps or worsens or doesn’t change voynich …. Lack of entropy ? Strict positional preference based on these glyphs? That’s what the challenge was right? Let’s see if we can mess with m as shorthand to see what happens. What happened? Not good at math here you might need to um, paint a description? ??
(10-06-2025, 10:46 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.the surprising section is quireT
The reason why quire 20 gives a higher value might lie in that fact that the pages of that quire make up two language clusters. The words on pages 103, 107, 108, 111, 112, 116 seem to be like the text in BioB2. You are not allowed to view links.
Register or
Login to view. has some further information on this.
(11-06-2025, 02:22 AM)anyasophira Wrote: You are not allowed to view links. Register or Login to view.….so is replacing m for iin showing that it helps or worsens or doesn’t change voynich …. Lack of entropy ? Strict positional preference based on these glyphs? That’s what the challenge was right? Let’s see if we can mess with m as shorthand to see what happens. What happened? Not good at math here you might need to um, paint a description? ??
In the tests it doesnt seem to change much about the score, which is a good sign that it may be true. If those are equivalent ways to write the same thing, the software should already be more likely to put those equivalent vords in the same group.
(11-06-2025, 09:37 AM)davidd Wrote: You are not allowed to view links. Register or Login to view. (11-06-2025, 02:22 AM)anyasophira Wrote: You are not allowed to view links. Register or Login to view.….so is replacing m for iin showing that it helps or worsens or doesn’t change voynich …. Lack of entropy ? Strict positional preference based on these glyphs? That’s what the challenge was right? Let’s see if we can mess with m as shorthand to see what happens. What happened? Not good at math here you might need to um, paint a description? ??
In the tests it doesnt seem to change much about the score, which is a good sign that it may be true. If those are equivalent ways to write the same thing, the software should already be more likely to put those equivalent vords in the same group.
Interesting. Same or similar ? Or would similar not be good enough to see this sort of pattern?