- Options
- Login
- Desktop Style
Options
The geomantic theory
- The geomantic theory
-
The geomantic theory
JustAnotherTheory > 19-06-2026, 09:09 PM
In this post I am proposing a system that generates Voynichese using a method available in the early 1400s.
The resulting text is meaningless, in the sense that it does not encode a particular language, concept or idea. On the other hand, it is consistent, to some degree, with several features of the text in the VMS. These include local features such as common prefixes, suffixes, combined glyphs (such as bench-gallows), as well as more global feautres like positive word autocorrelation, repeating consecutive words, words that differ by only a few glyphs, and page-wide (or indeed section-wide) topics.
The method of my theory is based on geomancy. Geomancy was a widely used divination technique in the late middle ages, and even much before that. It was not only used in Europe, but in the Arabic world as well. There is a very good explanation on this divination system on the website of Princeton University (You are not allowed to view links. Register or Login to view.).
In a few words (and simplifying a lot) geomancy is a way to gain knowledge about future events by making a number of random points on a sheet of paper (or tossing pebbles, etc…) and then counting the result. If the result is an even number, you translate that to 2 points. If the result is odd, you get 1 point.
You do this 4 times, so that you get a "figure". Here is an example of a figure:
In geomancy, you start the divniation process by constructing 4 figures in such a way, these are called Mothers. Then 4 more, the Daughters. Then 4 more, the Nieces, then 2 Wtinesses, and then 1 Judge.
What is interesting about this is that when you go from one layer to the next (f. ex. from the Mothers to the Nieces), you use information from the Mothers to create the Nieces using a combination. This makes the Nieces a "transformation" of the Mothers and Daughters. In other words, each subsequent layer is dependent on the one above.
The exact way of making these transformations is quite simple and is basically binary arithmetic. Please see the Princeton page for the full details (otherwise I wouldn’t have any space left in this post, sorry).
The final step then consists of assigning a House to each figure. There are 12 houses, but some systems add 4 more so that all 16 combinations of the figures can be "housed".
Each house is a way of telling a particular aspect of the future. For example, the house "Vita" will tell you about life. The house "Carcer" about prison or illegal activites. Here is an example of a House diagram:
In my system of generating Voynichese, I make use of this geomantic process:- The figures are Voynichese glyphs,
- Each layer of figures is a combination of glyphs above,
- Each house is a prefix- and/or suffix-rule, or a NULL rule.
Thus, when casting the figures and placing them on the table, we get Voynichese.
Since the different layers of the figures are dependent on the layers above, this creates a "structural" effect in the text whereby it appears that words in each other’s vicinity are related, or connected in some way. Repetitions might also occur.
The rules of each house are chosen in a unique way, which therefore creates the "signature" of the page, or its "topic". Finally, going through the houses in a given order will generate the text. I will give a simple example (and then a more complex one).
Example 1. Simple illustrative example
Let us assume that we have cast our Mothers, which gives (all Voynichese is written in EVA):- Mother 1: d
- Mother 2: k
- Mother 3: c
- Mother 4: h
Therefore, using the aforementioned combination rule, the 2 Nieces are:- Niece 1: dk
- Niece 2: ch
Having the Houses as follows:- House 1 (Vita): X + "aiin"
- House 2 (Lucrum): "qo" + X + "aiin"
- House 9 (Itineris): X + "ol"
- House 10 (Regnum): X + "o"
Then we would get the text:
daiin qokaiin dkol cho
Remember, whoever is casting the figures, does so randomly. It is, in some sort, in today’s terminology, a "random seed". All other glyphs and words are therefore related to these first seeds, which makes a curious structure emerge in larger texts. I will illustrate this with a more complex example.
Example 2: A more complete example
Now let us assume we have these figures:- Mother 1: ain
- Mother 2: d
- Mother 3: y
- Mother 4: k
- Daughter 1: h
- Daughter 2: c
- Daughter 3: p
- Daughter 4: e
Then the Nieces are:- Niece 1: dain
- Niece 2: ky
- Niece 3: ch
- Niece 4: ep
And the Witnesses:- Witness 1: kydain
- Witness 2: epch
And if we have these Houses:- House 9 (Itineris): NULL
- House 11 (Benefacta): ot + X + al
- House 13 (Witness house 1): X + y
- House 14 (Witness house 2): y + X + ol
Thus, the result is (reading in a particular order):
kydainy yepchol dain otchal
Discussion
I will briefly discuss a few points about this method:- The order in which the houses are read, in traditional geomancy, is sequential (ie, 1, 2, ... ). In this method, no such rule is made. Therefore, a scribe, having cast the figures and placed them into houses, might decide to follow an alternative path, which can be row-by-row, column-by-column, zig-zag, etc. This will make the resulting text none the less structured.
- The fact that all Nieces are combinations of Mothers and Daughters, and that Witnesses are again, combinations of the above, makes the resulting generated text very much inter-dependent. It results in repetitions, small differences in words, and positive word autocorrelation, depending on the path taken through the houses.
- Combining benches and gallows via Mothers/Daughters/Nieces naturally gives rise to complex bench-gallow glyphs.
- Some glyphs or glyph pairs are only used in the houses. This results in words that have a specific prefix or suffix all throughout the text.
- Each page has its own house arrangement, and possibly rules. This gives rise to topics that emerge at each page.
I will not make any assumption as to why one would generate such text, nor do I prove anything with this method. I am only showing a possible way to generate Voynichese.
Thank you for reading - The figures are Voynichese glyphs,
-
RE: The geomantic theory
oshfdk > 19-06-2026, 09:40 PM
(19-06-2026, 09:09 PM)JustAnotherTheory Wrote: You are not allowed to view links. Register or Login to view.Thus, the result is (reading in a particular order):
kydainy yepchol dain otchal
How similar is this to Voynichese? Your examples look okeyish (other than yep combination, which I don't think appears in the MS), but it would take a much larger sample to actually compare this to the Voynichese statistically. -
RE: The geomantic theory
Rafal > 19-06-2026, 11:04 PM
How much time would it take to generate one page with this method ? Isn't it too tedious?
Maybe you could do a practical test ? Check your time for generating a line or two and multiply enough times to get the page time. -
RE: The geomantic theory
ololololo > 19-06-2026, 11:09 PM
The result is quite beautiful and visually flawless, but it lacks statistics.
But I don't think the reproducibility of Voynichese even in this form tells us anything.
But apparently, from this point on, geomantia is more interesting to us than anything else.
-
RE: The geomantic theory
oshfdk > 19-06-2026, 11:23 PM
(19-06-2026, 11:09 PM)ololololo Wrote: You are not allowed to view links. Register or Login to view.But apparently, from this point on, geomantia is more interesting to us than anything else.
Geomancy has been discussed a few times since 2015 on this forum. The most recent I can find is:
You are not allowed to view links. Register or Login to view.
It's trivial to invent a method to generate a short snippet of plausible Voynichese. Until the method in this thread shows plausible statistics for at least a few pages' worth of text, I don't think it makes sense to study geomancy
-
RE: The geomantic theory
JoJo_Jost > 20-06-2026, 05:09 AM
Assuming I understand everything correctly, this method offers a great deal of flexibility. This will, of course, potentially enable it to generate some VMS token sequences. See oshfdk post.
On the other hand, however, it does not explain some of the deeper structures of the VMS.
-
RE: The geomantic theory
JustAnotherTheory > 20-06-2026, 09:32 AM
(19-06-2026, 11:04 PM)Rafal Wrote: You are not allowed to view links. Register or Login to view.How much time would it take to generate one page with this method ? Isn't it too tedious?
Maybe you could do a practical test ? Check your time for generating a line or two and multiply enough times to get the page time.
The whole advantage of this method is that once you fill the houses, you can just "go through" them in any order. There are so many combinations, you can easily generate a full page without much effort.
(20-06-2026, 05:09 AM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.Assuming I understand everything correctly, this method offers a great deal of flexibility. This will, of course, potentially enable it to generate some VMS token sequences. See oshfdk post.
On the other hand, however, it does not explain some of the deeper structures of the VMS.
Absolutely. There is a lot of flexibility: you can add even more house rules, such as "repeat glyph/word". There are simply so many possibilities, that a single house + figures combo can offer.
(19-06-2026, 11:23 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.It's trivial to invent a method to generate a short snippet of plausible Voynichese. Until the method in this thread shows plausible statistics for at least a few pages' worth of text, I don't think it makes sense to study geomancy
Thanks, I didn't know about that. Perhaps someone can code the algorithm and put it on github. I might even do it myself if there is enough interest in this. -
RE: The geomantic theory
JustAnotherTheory > 20-06-2026, 09:37 AM
(19-06-2026, 09:40 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.Your examples look okeyish (other than yep combination, which I don't think appears in the MS), but it would take a much larger sample to actually compare this to the Voynichese statistically.
Thank you. I would say that my method at this stage is a proof of concept, to illustrate an idea. If the VMS was indeed written like this, then there would have probably been different rules and rule sets, that would make the text look EXACTLY Voynichese.
But as you can see, something similar can be produced without much effort. The only missing link is now a fine-tuning of the house rules and/or figure combination rules.