| Welcome, Guest |
You have to register before you can post on our site.
|
| Latest Threads |
I am not convinced by the...
Forum: Provenance & history
Last Post: Koen G
9 minutes ago
» Replies: 25
» Views: 780
|
[split] Measure uncertain...
Forum: Analysis of the text
Last Post: Pointless..
1 hour ago
» Replies: 8
» Views: 282
|
Deconstructing the VMS : ...
Forum: Analysis of the text
Last Post: JoJo_Jost
1 hour ago
» Replies: 2
» Views: 58
|
116v
Forum: Marginalia
Last Post: Aga Tentakulus
2 hours ago
» Replies: 1,360
» Views: 959,010
|
The 'Chinese' Theory: Fo...
Forum: Theories & Solutions
Last Post: JoJo_Jost
4 hours ago
» Replies: 801
» Views: 212,742
|
15thc perception on swall...
Forum: Imagery
Last Post: Koen G
Yesterday, 08:13 PM
» Replies: 602
» Views: 335,308
|
No text, but a visual cod...
Forum: Theories & Solutions
Last Post: ololololo
Yesterday, 07:59 PM
» Replies: 1,869
» Views: 1,306,381
|
Voynichese is a numeric c...
Forum: Theories & Solutions
Last Post: ololololo
Yesterday, 07:55 PM
» Replies: 196
» Views: 20,808
|
New members: trouble regi...
Forum: News
Last Post: Koen G
Yesterday, 07:39 PM
» Replies: 0
» Views: 71
|
Voynich is encrypted ENOC...
Forum: Theories & Solutions
Last Post: Radim Dobeš
Yesterday, 05:50 PM
» Replies: 361
» Views: 52,452
|
|
|
| The Voynich as a rhythmic structure and height distribution in the Voynicl hypothesis |
|
Posted by: nintus - 11-05-2026, 05:35 PM - Forum: Analysis of the text
- Replies (12)
|
 |
Hello everyone.
Today I’d like to share one of my ideas about the Voynich script. Remembering my last experience here with you, I decided to approach it through a calligraphic analysis of the text, using an Excel spreadsheet.
After studying many EVA transcriptions and different interpretations of the Voynich manuscript, especially the work of Stephen Bax, I looked into phonetic studies and the attempts to match glyphs with specific sounds. In my opinion, those correspondences often seem sporadic and subjective. I also explored the idea of Gregorian chants derived from the manuscript.
The Voynich text has a remarkable lexical redundancy. With very few glyphs, a massive text was written, yet many pages almost feel like obsessive repetitions of nearly identical words. It gives the impression of something chant-like, melodic, repetitive — almost like a litany, a prayer, or something similar.
This aspect really caught my attention. How is it possible to generate a fully coherent text with so few letters?
I rule out the hypothesis that it was simply a fake or meaningless text created only for profit.
This is where medieval Gregorian chant comes into play. The Voynich text seems to share several similarities with it. For example, medieval Gregorian notation used four lines in the tetragram and only a few note heights. The notes, often written in square form and called neumes, were used to identify syllables. They could appear in different forms that modified pronunciation.
What interests me most is the number of lines and how, with only a few notes, an entire text could be written.
There are also compound neumes like podatus, clivis, torculus, and porrectus — groups of notes forming a single syllable — as well as simple neumes like punctum and virga, where a single note represents one syllable.
If we compare this idea with the Voynich text, the compound neumes could resemble compound glyphs such as ch, sh, ckh, cfh, cph, or cth.
Example of a neume with square annotations
While studying the Voynich script, what immediately stood out to me was the visual shape of the glyphs themselves. They reminded me of medieval Gregorian notation.
At that point, I started analyzing the glyphs and their heights, arranging them into a kind of scale. Of course, the text is written in cursive, so it isn’t as regular as Gregorian notes, but I tried to estimate the heights and see if this perspective could reveal something interesting.
Please ignore the heights marked with a “+” sign, which I used for compound glyphs with ligatures or connected forms. In Excel I couldn’t make the calculations work correctly with the “+” signs, so I had to leave them out. I suspect those glyphs may represent something additional compared to the others, but for now I don’t want to push the theory too far.
Now I’ll explain exactly what I did.
I took the manuscript and assigned heights to the glyphs: A high, M medium, B low, C short.
This is how I categorized the glyph heights.
I should point out again that the Voynich text is written in cursive. Someone might notice, for example, that in the word “ykal” I classified the glyph “l” as C, even though visually it could seem closer to B. But after checking most instances of that glyph, I noticed that, like the “y” glyph, part of it often extends below the line. In other cases, the “y” glyph appears shorter than expected.
Since I obviously cannot manually retranscribe the entire manuscript using these rules, I decided to catalog the glyphs this way and use the full EVA transcription by Takeshi, despite the errors I found in it. That’s another discussion entirely. What matters here are the calculations I’m about to explain. A few errors are acceptable
Method
What I did was import the entire EVA transcription of the manuscript into an Excel file:
You are not allowed to view links. Register or Login to view.
After that, I needed a way to understand whether this method could reveal any interesting patterns in the text. For example, I wanted to measure how frequently glyph heights changed across entire pages, to see whether there was some regularity or complete chaos — in other words, whether the glyph heights had meaning or were simply decorative or random.
Procedure
I took each EVA page and arranged it line by line into columns.
Then I created a formula that transformed the entire text into the height categories I had chosen, returning each line as a sequence of height letters.
After that, I created additional columns containing the number of A, M, B, and C heights, so I could count how often each height appeared on every line.
I also created another column to calculate the Runs for each line. By “Runs,” I mean the number of transitions from one height to another between consecutive letters.
Then I created another column for the average Runs per page, combining all the extracted data. This was the result I was most interested in: the average Runs.
I also generated columns showing the percentage distribution of each height on every page, to see how much the patterns changed from one page to another.
I even tried to create a column for entropy calculations, but I couldn’t manage to make it work.
The Excel file I generated contains all these results. At the bottom right there are the average Runs for each page. In the center there are the percentages of each height category across the entire page. On the left side, I also made a comparison using a Gregorian chant. I took the notes written on four lines and converted them into my height categories. I wasn’t interested in identifying the actual notes themselves — I simply took them in sequence. As you can see, the distribution is very balanced, and the average Runs are surprisingly close to those found in the Voynich manuscript. It is impossible to get similar results with a natural language text, even with stylized texts that visually resemble Voynich. Those tend to produce much more chaotic and irregular patterns. Going back to the Voynich data: if you look at the average Runs on each page, they are all remarkably similar. The percentages of the different height categories also remain very consistent from page to page.
Conclusion
In my opinion, these results suggest that the glyph heights are not random, and that the encoding method could perhaps be related in some way to medieval Gregorian chant notation — though I can’t say exactly how. Still, these are the patterns I found. I’d love to hear your thoughts.
|
|
|
| I think the Voynich Manuscript was written after 1500 |
|
Posted by: 99b86aeea - 09-05-2026, 10:52 PM - Forum: Theories & Solutions
- Replies (10)
|
|
Hi everyone, it is a pleasure to meet you all. I am an independent researcher, and I believe that the Voynich Manuscript was written after 1500. My claim goes against the carbon dating of the parchment, which puts the manuscript almost a century earlier, but I think that the parchment was used later for writing this famous manuscript.
I hope that my arguments will convince, or at least provoke some debate. And of course, I am really happy to have the opportunity to do so, in a forum of such renowned researchers.
So I will begin by saying that I am enthousiastic about the Renaissance period, I am not a historian but my father was one. And so, it struck me that when looking at the Voynich manuscript, it has many elements that look like the famous engineer's book that was published between 1500 and 1520, by Francesco di Giorgio Martini. I will present some visual comparisons, sorry for the quality of the pictures, which I made with the Windows snipping feature.
1. The tubes
These tubes are the same type:
These barrels have the same kind of bad perspective:
This water flow looks similar:
These pipes look like the Voynich manuscript:
2. The containers
Several column designs look like the Voynich manuscript containers:
And this:
And also this:
3. The architecture diagrams
Please take a look at the extremities of these circular diagrams:
In fact, many have the same details as the Voynich manuscript:
4. Top-down, or bottom-up view or buildings
Such diagrams as below look like Voynich manuscript illustrations, if we interpret tham as top-down or bottom-up views of some building's interior, such as a church or a basilica:
5. Green coloured water illustrations
In the following images, you will see water illustrations that are coloured green, like in the Voynich manuscript, for some reason:
I would like to share with you the online facsimile of this manuscript, which is here:
You are not allowed to view links. Register or Login to view.
Interestingly, it is also in the Beineke rare books collection. An amusing coincidence, I will add.
In conclusion, I will allow myself to enjoy a small conjecture. In particular, if we assume that the author(s) of the Voynich manuscript were indeed readers of Martini's book, then the timing of the Voynich manuscript creation is much later than 1400. In fact, it would be dated after the invention of Alberti's cipher disk. And therefore, my conjecture is that the writing in the Voynich manuscript is an Alberti disk type of substitution. I will go a step further. I will say that that the so-called "gallows" characters are giving the offsets of the disk, if such a disk exists, in the "mobilis" position. The way that these gallows characters do this, is, in my opinion, by encoding the number of crossings in the character itself. So if there is one crossing, then we shift the disk by "1" position. If there are two crossings, then we shift it by "2" positions. I have described these kind of crossings in the classification below:
And so forth. It would then be easy for the writer, and the reader, to translate the text to ciphertext and vice versa, if they knew the "stabilis" portion of the disk.
Anyway, I am very happy to present my small personal theory to this distinguished forum of researchers, and I hope that you will find it somewhat interesting.
Have a great day
|
|
|
|
| Could this monastery on the Greek island of Patmos be relevant to the VMS? |
|
Posted by: JustAnotherTheory - 08-05-2026, 09:03 AM - Forum: Theories & Solutions
- Replies (19)
|
|
So this is not a full "theory" in the sense of the word, but more of a suggestion, to see what others think about it.
Basically, I recently noticed that the monastery on the Greek island of Patmos bears some resemblance to the architectural drawings in the VMS, especially the Rosettes section. It is a bit out of place within the context of the manuscript, because it is not anywhere near Northern Italy. Yet it has several interesting features.
The first is its size:
Here is another view from above:
Keen eyed observers will notice that this monastery, quite strikingly, has ghibelline merlons all around:
The entire complex was built with a concept of "sunken" or multi-storey buildings, which also reminds one of the VMS :
In the VMS rosettes, we see that the different structures are surrounded by what appears to be large bodies of water. Well, Patmos is a small island, so in that sense, it "fits" as well. Furthermore, the island of Patmos was under Ottoman rule for a long time, and there would have been many minarets all around. Just like in the VMS Rosettes section.
The interior of the monastery is lush with frescoes. A few examples remind one of the VMS (on the left hand side images from the VMS folio f1v, on the right, typical fresco patterns on Patmos):
The monastery is called "Monastery of St George the Theologian". Apparently, Saint Georrge, who wrote the Revelations part of the Bible, did so in this exact monastery, hence ist name. You can find some info here:
You are not allowed to view links. Register or Login to view.
The monastery housed a huge and varied library. In fact, it has been called the "one of the most sgnificant monastic libraries of the world". Already in 1080, thousands of books were kept there and there was a professional scriptorium.
So based on the information above, I ask myself (and the dear readers of this forum), whether Patmos could have been a place of relevance for the VMS. Could the VMS have been written here? Are the depictions in the Rosettes section describing this monastery? Could the vast library of Patmos have been an influence for the apparent multi-cultural drawings in the VMS?
What do you think?
|
|
|
|
| Structural Translation of the Voynich Manuscript: A Mathematical and Cross-Modal Appr |
|
Posted by: Keishi Oi - 08-05-2026, 01:07 AM - Forum: The Slop Bucket
- Replies (1)
|
|
Greetings, researchers and experts of the Voynich community.
My name is Keishi Oi. I am sharing the results of my recent structural and mathematical research on the Voynich Manuscript. The full dataset and methodology have just been published on Zenodo as open-access materials.
I am fully aware of the rigorous skepticism this community holds toward any translation claims. Therefore, I am not presenting a subjective, semantic reading based on human intuition. Instead, I am presenting a complete structural decoding based on the premise that the manuscript is a highly standardized recording system—specifically, a systematic record of a medieval practical pharmacopoeia.
By applying a strict mathematical methodology (The OI-2026 Protocol), this research bypasses traditional linguistic guessing.
Key Methodologies and Findings:
1.Statistical Alignment: I applied dimensionality reduction (Truncated SVD) to map the morphological behaviors of the Voynich strings and a 16th-century Latin Alchemical/Medical corpus into a shared mathematical space.
2.Objective Constraints: Using strict statistical thresholds (Z-score >= 2.0) and geometrically classified grammatical roles, 99.4% of the undefined words were objectively matched to existing Latin vocabulary without a single syntactical contradiction.
3.Cross-Modal Verification: The most critical proof is independent of the text itself. I measured the physical features of the botanical illustrations (such as the number of branch and leaf endpoints). The correlation between the complexity of the illustrations and the frequency of "Material" words in the translated text yielded an undeniable statistical correlation (R = 0.7080, P = 7.70×10⁻²⁹).
The manuscript records highly structured operational procedures (e.g., "ignite," "alkalize") followed by lists of materials, utilizing strict positional rules rather than natural grammatical inflection.
Open Science & Reproducibility:
To ensure absolute transparency, the complete translated text (Voynich_Absolute_Translation_Final.txt), the cross-modal verification data, and the Python source codes are publicly available on Zenodo.
Zenodo Repository:
You are not allowed to view links. Register or Login to view.
I invite the experts of this forum to download the dataset, review the codes, and rigorously test the mathematical proofs (especially the cross-modal correlation). I welcome all objective, data-driven critiques.
Best regards,
Keishi Oi
Exaｍple
<1r.1>
①EVA : fa19s 9 hae ay Akam 2oe !oy9 scs 9 hoy 2oe89
②LAT : <MOUENTE> <PERFECTIONEM> <EUADERET> <ADGLUTINATIS> <MARKASITA> <VARIARETUR> <LOCANDO> <PHOR> <PERFECTIONEM> <TACITUR> <INUENISTIS>
<1r.2>
①EVA : soy9 Hay oy 9 hacy 1kam 2ay Ais Kay Kay 8aN
②LAT : <CUSCUTA> <UERTENTES> <DETINERE> <PERFECTIONEM> <LUXATIS> <EUANESCENT> <PRAEUIDUARE> <EUPHRAGIAE> <PASTORIS> <PASTORIS> <SURREPTUM>
<1r.3>
①EVA : s9aIy 2ch9 oy 9ham o8 Ko ay9 Kcs 8ay am s9
②LAT : <CUSCUTA> <SUCCERNITUR> <DETINERE> <IGNEITATIS> <FERMENTATIONES> <REUOCA> <CIRCUNDATA> <SEPARABI> <CONTRAHERE> <LACERTA> <PROPINENTUR>
<1r.4>
①EVA : 8om okcc9 okcoy y oeok9 Aay 8am oham oy ohaN
②LAT : <OPERIATUR> <COGITAUERIT> <DESIDERARE> <DETINERE> <MARITE> <FECTE> <PERE> <COMITIS> <DETINERE> <CIEBUS>
<1r.5>
①EVA : saz9 1cay Kam Jay Fam 98ayai29
②LAT : <IUSQUIAMUS> <IMPERTIUNT> <DESIDERES> <SISSIMA> <CONIURATIO> <NATES>
<1r.6>
①EVA : o8ay !oe Go9 o98ay ! s Foam o8ay9
②LAT : <EXPLICATIS> <APTARE> <applicantur> <UADAT> <ANAGALLIDE> <LIFICANTE> <ANAGALLIDE> <ANAGALLIDE> <NATES>
|
|
|
|
| update to Zattera's slot machine |
|
Posted by: Labyrinthinesecurity - 07-05-2026, 10:36 AM - Forum: News
- Replies (18)
|
|
Using switchable templates, it looks like we can improve Zattera's slot machine performance:
Metric Before After Change
Generated 3,110 5,524 +2,414
TP types 1,226 1,680 +454
TP tokens 20,454 24,662 +4,208
Precision 0.3942 0.3041 -0.0901
Recall 0.1468 0.2011 +0.0543
F1 0.2139 0.2421 +0.0282
TokCov 0.5491 0.6620 +0.1129
Note that we worked on the IVTFF corpus, not Zattera's homebrew Slot Alphabet.
The improved grammar (F1=0.242) outperforms Zattera's grammar (F1=0.214) when both are scored on our corpus. However, Zattera reports F1=0.270 on his own filtered corpus (~5,105 types), and our grammar was F1-trained on the test corpus while his was not.
|
|
|
|
| Currier A/B split is not what we thought it was! |
|
Posted by: Labyrinthinesecurity - 04-05-2026, 03:09 PM - Forum: Analysis of the text
- Replies (49)
|
|
Hi all,
Some of you may have seen my earlier work confirming the Currier A/B distinction quantitatively. That paper showed the distinction is real, recoverable without labels, and predictive. But it also left a puzzle on the table that I could not explain at the time. I now have an explanation, and it leads somewhere unexpected.
Of the eleven character pairs I tested, one behaved paradoxically. The e/ch pair had essentially zero global correlation with the A/B split, yet it produced the strongest signal of all pairs at folio boundaries. And when included in clustering, it actively destroyed the A/B partition: removing it doubled the clustering accuracy.
How can a pair be simultaneously invisible globally, maximally informative locally, and destructive to classification? That combination is not possible under a simple two-language model. Something more structured is going on.
The answer turns out to be surprisingly clean. If you look at the vowel that follows the digraphs CH and SH across the manuscript, you find that folios split into two sharply separated groups. The gap between these two states is enormous, and a two-state binomial mixture model fits with a 2,549-point AIC improvement over a single state. Of 197 folios, 195 are assigned unambiguously.
This is not the same thing as the Currier A/B split, although it correlates with it. It is sharper, it operates at the individual folio level rather than at section boundaries, and it persists within the Herbal section alone (where the A/B boundary is supposed to be clean).
I call it a boolean switch: a single binary parameter, set once per folio.
Here is where it gets interesting. If the switch were just replacing graphemes uniformly, every word containing those graphemes would respond the same way. They do not.
When you group words into templates, you find three classes:
- Fixed O templates: these are locked to O in both switch states.
- Fixed E templates: these are locked to E in both switches states.
- Switchable templates: these respond strongly to the switch.
Template identity accounts for 93.5% of the variance. The folio switch accounts for only 7.9%.
So the system has two components: a template structure that determines which contexts are switchable, and a boolean parameter that modulates the switchable ones. The Currier A/B distinction is a blurred projection of this system, not the system itself.
The e/ch pair is paradoxical because it responds to the boolean switch, but only in switchable template contexts. In clustering, the e/ch ratio injects variance along a dimension that does not align with the primary A/B axis. Mystery solved.
Now for the part that surprised me the most
Everything above is derived purely from text statistics. I had no reason to expect it would connect to anything visual. But then I found Koen's morphometric study of the Herbal plant illustrations (You are not allowed to view links. Register or Login to view.), which classifies plants as A-type or B-type based on twelve visual features: stem-root lines, flower morphology, daisy-type flowers, grass elements, root platforms, leaf venation, and so on. This classification was done entirely from the drawings, with no reference to text statistics.
I cross-validated my boolean switch against the morphometric classification on 101 Herbal folios (excluding quire 8). The results:
Boolean switch vs. morphometrics: 96.0% agreement, Cohen's kappa = 0.870, Fisher's exact p = 3.5 x 10^-15.
Currier vs. morphometrics: 78.2% agreement, Cohen's kappa = 0.106.
Read that kappa for Currier again: 0.106. Once you correct for base rates, Currier's section-level labels have almost no predictive power for plant morphology. The boolean switch, derived from a single text ratio, predicts the visual classification of the plant drawings with near-perfect accuracy.
Every Currier discordance is resolved by the switch
Of the 27 folios where my switch disagrees with Currier's label, 18 have morphometric data. In all 18 cases, the plant illustration sides with my switch, not with Currier. The probability of that under the null is 3.8 x 10-6.
These are not marginal cases. Folios like f31r, f34v, f39r, f43r, f46r, and You are not allowed to view links. Register or Login to view. are all traditionally classified as Currier A because they fall in the f1-f57 range. But their text is E-dominant, and their plant illustrations show B-type features (daisies, grass, root platforms, unidirectional leaves). Conversely, f87r, f90r, f93v, and You are not allowed to view links. Register or Login to view. are traditionally Currier B, but their text is O-dominant and their plants show A-type features (stem-root lines, A-type flowers and calyxes).
The switch is not just a better statistical classifier. It is detecting the same organizational principle that the illustrator(s) was(were) following.
|
|
|
|
| Upcoming lecture and article |
|
Posted by: LisaFaginDavis - 04-05-2026, 02:19 PM - Forum: News
- Replies (1)
|
 |
If you're going to be at next week's You are not allowed to view links. Register or Login to view. at Western Michigan University (Kalamazoo, MI), I'll be presenting some of the work Colin and I have done applying Latent Semantic Analytics to the VMS in a virtual session on Thursday at 1:30 PM Eastern: "Resequencing the Voynich Manuscript: Linguistic, Paleographic, and Material Evidence" (Session 96). Abstract: "Combining material evidence with novel linguistic analytics undertaken in collaboration with You are not allowed to view links. Register or Login to view., You are not allowed to view links. Register or Login to view. and others, this paper will present a radical and potentially game-changing hypothesis about the original structure of the Voynich Manuscript." The session will be recorded for ICMS registrants.
In other news, the first of two articles on this topic by Colin and myself will be published online by Digital Humanities Quarterly later this week. I'll post the link once it's available.
|
|
|
|
|
).