- Options
- Login
- Desktop Style
Options
Why and how the text could be Bavarian
- Why and how the text could be Bavarian
-
RE: Why and how the text could be Bavarian
JoJo_Jost > 07-06-2026, 07:01 AM
The Vowel Bridge Model fits remarkably well into the statistical peculiarities of the VMS - an overview
I would like to take a step back and summarize why the Vowel Bridge Model (VBM) is, in my view, a serious candidate for a solution to the VMS.
First of all, an important point: the VBM does not invent everything from scratch. It builds on observations and work that are already well known in Voynich research: Stolfi's slot theory, Currier's A/B observations, the work on LAAFU, the studies on line starts and line endings, and observations by Elmar Vogt, tavie, Emma May Smith, and Patrick Feaster. Feaster's rules on word boundaries are especially important here.
The VBM takes these findings seriously and is based on these observations, but it places them in a different context.
First, here’s a brief overview of how the model works:
The decisive difference from many other solutions and statistical approaches is very simple:
EVA spaces are no longer treated as secure plaintext word boundaries.
Instead, the text is read as a continuous consonant/vowel stream:
V1 | C1 | C2 | V2
Here, "V" is not simply a single letter, but a bigram, a vowel bridge across the EVA space:
VL.VR
VL = the left part of the vowel bridge, that is, the final glyph of an EVA token.
VR = the right part of the vowel bridge, that is, the first glyph of the next EVA token.
So the visible EVA word boundary lies in the middle of the vowel bridge. The actual boundary of a word or syllable, however, often lies rather between:
C1 || C2
C1 = would then be the final side, that is, the final consonant or final consonant cluster of an element.
C2 = would be the initial side, that is, the initial consonant or initial consonant cluster of the next element.
This very simple change of perspective suddenly explains a whole series of Voynich peculiarities and statistical anomalies:
1. The repetitions are not necessarily word repetitions
One of the strongest arguments against language has been this: Voynichese contains seemingly absurd repetitions, and their frequency is far above the statistical rate of word repetitions in normal texts:
qokedy qokedy qokedy qokedy
chol chol chol
qokeedy qokeedy qokeedy
daiin daiin daiin
On the EVA word level, this does indeed look almost impossible for a natural language. But in the VBM, these do not necessarily have to be repeated words. They can be repeated stream segments, as I have shown in this thread with several examples.
Example:
qokedy qokedy qokedy qokedy
= e-nd-e-nd-e-nd-e...
This is not absurd in German or Middle High German. Such chains can arise through normal word formation and inflection, for example in structures such as:
"im Elend endenden, den ..." (who end in misery, for)
The point is: the VBM turns a seemingly nonsensical word repetition into a language-like stream. This is because repetitions of glyph sequences within sentences are completely normal.
2. Seven rules for word boundaries become one principle
With seven or eight simple rules, as described by Patrick Feaster, one can account for a very large part of the EVA word boundaries. That would hardly be expected in a normal European language. This is exactly where the VBM comes in.
The VBM reduces these rules, in essence, to a simpler principle: word boundaries are separations of vowel bigrams, VL.VR , where the left side consists of only a small number of glyphs.
And suddenly this fits the picture very well. Vowels are very frequent, and especially "y.qo" as a word-boundary bigram, as an "e", makes a lot of sense.
The actual word boundary often lies here:
C1 || C2 that is, between final consonant and initial consonant.
This also explains why EVA words look so stable and artificial, even though the underlying stream may be language-like.
3. Low entropy becomes less mysterious
Voynichese is unusually predictable on the glyph level. In a normal alphabetic text, this is a problem.
In a slot system, however, this is exactly what one would expect.
If:
C1 constrains C2
C2 constrains VL
VL constrains VR
then the next glyph is often strongly constrained.
That means: the low entropy does not have to speak against language. It can be the result of a mechanical encoding of language.
If vowels are encoded by bigrams, for example:
"y.qo" = "e" then "qo" after y becomes extremely predictable.
If C2 constrains the left part of the vowel bridge, predictability increases further.
If C2 itself works homophonically or positionally, that is, if different consonants or consonant clusters are encoded with similar visible forms, predictability increases again, especially in the e-chains.
The VMS then does not look repetitive because there is no text underneath. It looks repetitive because a natural language stream runs through a very narrow, mechanical slot system.
4. C2 is not a simple letter - the "dy" phenomenon
A central point is the strong relationship between C2 and the final glyph of the token, that is, VL.
In the forum, this has often been described in various forms as the "dy" or "edy" phenomenon, and it has also been described in slot analyses.
Examples: (Only unique C2 clusters were counted)
C2 = eed -> VL = y in 730 out of 732 cases
C2 = ed -> VL = y in 1,332 out of 1,339 cases
Other C2 types behave very differently:
C2 = a -> mainly VL = r or l
C2 = o -> mainly VL = l or r
This is not a free distribution.
C2 and VL apparently belong to the same mechanical system.
This means: C2 cannot be read in isolation.
The unit is more likely:
C2 + VL = route for an initial consonant or initial consonant cluster
VR = vowel value, vowel variant, or selection within the route
This explains why C2 is so difficult to crack. It is probably not a simple substitution table, but an almost polyphonic encryption. This fits perfectly into the period in which the transition between monoalphabetic and polyalphabetic encryption took place.
So one should not only ask: What does eed mean?
but rather:
What does eed.y.qo mean?
What does eed.y.o mean?
What does eed.y.ch mean?
This is more complicated, but it fits the structure of the VMS very well.
5. The connection between C1 and C2 is expected
The connection between C1 and C2 is not a problem for the VBM either. It is exactly what one would expect.
If the real word or syllable boundary lies between C1 and C2,
... V | C1 || C2 | V ...
then the final consonant of one element stands on the left, and the initial consonant of the next element stands on the right.
In real language, such transitions are not completely free. Certain final consonants combine more often with certain initial classes. Other combinations are rare or almost impossible.
If, on top of that, there is also a cipher logic, this coupling becomes even more visible. The C1-C2 dependency is therefore not an argument against the VBM. It follows from the logical consequences of language.
6. e, ee, and eee are not free repetitions
Another well-known finding concerns the e-chains.
A single "e", "ee", "eee", and longer chains do not behave like free repetitions of the same sign.
The preceding anchors change systematically:
single e -> often after ch/sh
ee / eee -> much more strongly after k/t, that is, after Gallows
This is not a free counter.
It looks more like anchor-bound composite forms.
So not:
e = 1
ee = 2
eee = 3
but rather:
che
kee
keee
she
tee
...
as bound slot signs or composite signs.
This fits the VBM because e-runs are almost always token-internal and hardly ever run across EVA spaces. They therefore do not belong to the vowel bridge, but to the internal C2/composite layer.
This separates two levels:
C2/composite layer: token-internal
vowel-bridge layer: across the EVA space
This too fits well into the model.
7. Near-variants become expected
Voynichese is full of very similar forms:
qokedy
qokeedy
qokeey
qokey
okeedy
okeey
On the word level, this looks like an artificially dense vocabulary.
In the VBM, it is expected.
If EVA tokens are only visible cuts through a slot stream, many near-variants automatically arise. This is especially true because it mostly affects C2; the connecting vowel bridges are simply the most frequent plaintext letters, such as "e", while the C1 slots are also frequent plaintext consonants, such as "n".
A different vowel-bridge element, a different C2 block, a shifted cut - and a new EVA token appears that looks very similar to another one.
So this is not an artificial lexicon. It is the result of the VMS using a false word segmentation as camouflage.
8. Hapaxes become less problematic
The VMS has many hapaxes. And many similar words.
This is not surprising in the VBM either.
If EVA words are not plaintext words, but excerpts from a running syllable and morpheme stream, many visible tokens can occur only once, even though their building blocks are very frequent.
German and Middle High German constantly work with short elements such as:
-en
-er
-es / -ez
-end
-de
-den
-der
-se / -ze
etc.
This is where the frequent tokens lie in EVA. But of course there are also many words in which these frequent syllable parts do not occur. And since the system cuts between consonants, many hapaxes must inevitably arise. The distribution between hapaxes and the seemingly similar words of the known families therefore also appears to be a logical consequence of the VBM.
9. LAAFU, LSM, and LEM get a function, and their statistical peculiarities become at least theoretically logical
Line beginnings and line endings are among the most striking problems of the VMS.
In the VBM, line beginnings and line endings are not neutral. And this fits a problem the VMS has:
The running stream has to begin, continue, adapt, or end at a line. Depending on whether the following plaintext begins with a vowel, a consonant, or a cluster, the stream has to be started differently - while still remaining inside the stream.
LSM, meaning Line Start Marker, and LEM, meaning Line End Marker, therefore do not have to be normal words.
They could be phase markers, continuation markers, or boundary markers inside the encoding, indicating for each line how the stream is started in this particular case.
This raises the question why the stream has to be restarted in every line. The answer is simple: because the stream is essentially a forced description of natural language. It often fits, but not always. And unusual words, among other things, disturb it considerably. So it is easier to restart it in every line.
Depending on how it is started, the lines become longer or shorter. And this is one of the statistical peculiarities of LSMs: they lead to longer and shorter lines, depending on what insertions are needed to start the stream.
The change in word length over the course of a line may also be connected with this. If the stream begins at the start of a line in a relatively clean VCCV-like structure, it may become increasingly complicated later in the line through inflection, clusters, endings, and function words. This may explain why EVA word lengths change over the course of a line.
10. Currier A/B becomes less mysterious
Currier A and Currier B differ clearly on the surface. But the underlying slot logic can still remain the same.
That is exactly what one would expect from a table system. Different sections, hands, or conventions may use different routes, while the basic mechanism remains stable. So the VBM does not have to explain Currier A/B away. It can treat Currier A/B as different manifestations of the same mechanical system.
11. The model produces language-like streams
On the EVA word level, the VMS looks alien.
Under VBM segmentation, however, the text begins to look like a consonant/vowel stream:
V - C - C - V - C - C - V ...
That is exactly what one would expect if a natural language were encoded through a position-dependent mechanism.
The perspective shifts:
Not:
EVA word = word
but:
EVA token = visible segment of a running stream
This turns stubborn Voynichese into a system that produces very clear language-like structures, as can be seen perfectly in the translations of the word repetitions.
12. The clue on f8v
On f8v, the plant shows a striking similarity to hazel or hazelnut. On the same page, there is the unusual chain:
okcholksh chol chol chol cthaiin
The strong inner part, under the current VBM working values, produce a stream (probably the only option—I haven't been able to find any others) like:
"nuzze zezzen"
This is remarkable because a German / Middle High German pharmaceutical text contains a very similar formulation:
"hasel-nuzze zezzen"
meaning roughly: to give hazelnuts to eat.
I do not claim that You are not allowed to view links. Register or Login to view. is solved by this. But image, structure, and pharmaceutical parallel point in the same direction. This is a strong indication that the VBM works. It is not yet proof - i know.
13. What the VBM cannot yet do
The VBM does not yet provide a complete plaintext.
It does not yet finally clarify which language was encoded, although MHG / NHG are very likely because of their structure.
It does not yet completely clarify how C2 works.
It cannot yet explain why P-lines contain more Gallows than other lines.
But it offers a coherent explanatory framework for many central Voynich peculiarities:
- apparent word repetitions
- strange EVA spaces
- low entropy
- slot-like word structure
- C1-C2 dependency
- C2-VL dependency
- e/ee/eee as a composite layer
- near-variants
- hapaxes
- LAAFU effects
- line-start and line-end effects
- Currier variation
- language-like V/C streams beneath the surface
- possible botanical and textual cribs
The VBM should not simply be seen as just another reading attempt!
Rather, it is a clear structural model that, with very simple and few assumptions, explains a large part of the statistical peculiarities and gives them a meaning.
To my knowledge, it is one of the strongest structural models for capturing the VMS as a system and a possible language.
Jost -
RE: Why and how the text could be Bavarian
anyasophira > 07-06-2026, 08:15 AM
On the EVA word level, the VMS looks alien.
Under VBM segmentation, however, the text begins to look like a consonant/vowel stream:
V - C - C - V - C - C - V ...
That is exactly what one would expect if a natural language were encoded through a position-dependent mechanism.
I am trying to understand how
This works.
Please walk me through this - you state that we expect this in natural language. Why? When and where is it demonstrated in language and when and where is it not demonstrated in other examples.
Also
Are you saying that chol means the sound ze? -
RE: Why and how the text could be Bavarian
oshfdk > 07-06-2026, 08:47 AM
(07-06-2026, 07:01 AM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.The Vowel Bridge Model fits remarkably well into the statistical peculiarities of the VMS - an overview
I think there is a simple way of providing evidence for this claim, one can take some text and encode it using this model. If the result is indeed statistically similar to Voynichese, this to me will be a good argument. -
RE: Why and how the text could be Bavarian
JoJo_Jost > 07-06-2026, 10:49 AM
(07-06-2026, 08:15 AM)anyasophira Wrote: You are not allowed to view links. Register or Login to view.I am trying to understand how This works. Please walk me through this - you state that we expect this in natural language. Why? When and where is it demonstrated in language and when and where is it not demonstrated in other examples.
Also Are you saying that chol means the sound ze?
Yes, it’s confusing yet simple—you just have to think outside the box. I’ll try to describe it as clearly as possible:
In most of the natural languages, vowels (V) and consonants (C ) follow each other in a very strict sequence.
V = a,o,e,i and diphthongs: ei, au, eu, ou (in German)
C = Consonants an C cluster (e.g. "th" in englich)
You can see the pattern in it, roughly: CVCVCVCVC = CVC and, when words end with C and one begins with C also CCVC or maybe CVVC.
I noticed that word boundaries in VMS could be a break in a vowel bigram. "y.qo" (the dot is a space) would then be a bigram for “e” opch(y.qo)kedy the vowel ist in the middle.
But that means the actual word boundary would have to occur somewhere in qokeedy and qokedy
Many tokens in the VMS have a 4-slot structure. We use qokeedy: If we apply this to it, the first slot would be the right part of this vowel bigram VR = qo. Then comes the consonant that ends the word, C1 here "k". That’s where the space would have to go. And the consonant that begins the next word, C2 here "eed", followed by the left part of the next vowel bigram, VL = y
This results in the following structure for the token qokedy:
qo k ed y
VR C1 (Space) C2 VL = VCCV
With the right word boundaries (VV=V):
CVC CVC
qokeedy.opchy.qokedy.dypchy with the "correct" word boundaries:
VCCV would be the basic structure of the VMS. Of course, language doesn’t flow that smoothly; for one thing, not all words have only a CVC structure—they are often longer. But for Bavarian, this word length is actually quite typical. There are many simple, short words with this CVC structure—the language is almost monosyllabic. Of course, there are also words that start with vowels, and some end with a vowel. When words are longer, there is often not a pair of consonants in the middle, but just one. So the flow isn’t exactly VCCV.
That is why there are two- and three-glyph words to account for this inconsistency, as well as longer words that can also accommodate it.
No, “Chol” isn't “ze”— the “o” in “Chol” is the “z” (s/zz/ss)
In “chol.chol” the “l.ch” is the “e” and ‘o’ are the “z,” = ez
@ oshfdk
Yes, that’s on the agenda anyway: I prefer it.... -
RE: Why and how the text could be Bavarian
anyasophira > 07-06-2026, 06:30 PM
JoJo_Jost dateline='[url=tel:1780825771' Wrote: You are not allowed to view links. Register or Login to view.1780825771[/url]']
anyasophira dateline='[url=tel:1780816509' Wrote: You are not allowed to view links. Register or Login to view.1780816509[/url]']
I am trying to understand how This works. Please walk me through this - you state that we expect this in natural language. Why? When and where is it demonstrated in language and when and where is it not demonstrated in other examples.
Also Are you saying that chol means the sound ze?
Yes, it’s confusing yet simple—you just have to think outside the box. I’ll try to describe it as clearly as possible:
In most of the natural languages, vowels (V) and consonants (C ) follow each other in a very strict sequence.
V = a,o,e,i and diphthongs: ei, au, eu, ou (in German)
C = Consonants an C cluster (e.g. "th" in englich)
You can see the pattern in it, roughly: CVCVCVCVC = CVC and, when words end with C and one begins with C also CCVC or maybe CVVC.
I noticed that word boundaries in VMS could be a break in a vowel bigram. "y.qo" (the dot is a space) would then be a bigram for “e” opch(y.qo)kedy the vowel ist in the middle.
But that means the actual word boundary would have to occur somewhere in qokeedy and qokedy
Many tokens in the VMS have a 4-slot structure. We use qokeedy: If we apply this to it, the first slot would be the right part of this vowel bigram VR = qo. Then comes the consonant that ends the word, C1 here "k". That’s where the space would have to go. And the consonant that begins the next word, C2 here "eed", followed by the left part of the next vowel bigram, VL = y
This results in the following structure for the token qokedy:
qo k ed y
VR C1 (Space) C2 VL = VCCV
With the right word boundaries (VV=V):
CVC CVC
qokeedy.opchy.qokedy.dypchy with the "correct" word boundaries:
VCCV would be the basic structure of the VMS. Of course, language doesn’t flow that smoothly; for one thing, not all words have only a CVC structure—they are often longer. But for Bavarian, this word length is actually quite typical. There are many simple, short words with this CVC structure—the language is almost monosyllabic. Of course, there are also words that start with vowels, and some end with a vowel. When words are longer, there is often not a pair of consonants in the middle, but just one. So the flow isn’t exactly VCCV.
That is why there are two- and three-glyph words to account for this inconsistency, as well as longer words that can also accommodate it.
No, “Chol” isn't “ze”— the “o” in “Chol” is the “z” (s/zz/ss)
In “chol.chol” the “l.ch” is the “e” and ‘o’ are the “z,” = ez
@ oshfdk
Yes, that’s on the agenda anyway: I prefer it....
So the theory is that word boundaries don’t show
Where words are, and if you eliminate word boundaries, then you can encode slots with vowels and consonants. Do you encode the same vowels and consonants every time? If so then it’s still substitution
Just adding an extra few symbols- the space between the words? Or am I wrong about that? And do the vowels and consents follow
Slot grammar? If so , why? How does
That encode? And does the theory necessitate Bavarian language. -
RE: Why and how the text could be Bavarian
JoJo_Jost > 07-06-2026, 06:58 PM
So the theory is that word boundaries don’t show where words are (YES), and if you eliminate word boundaries, then you can encode slots with vowels and consonants. Do you encode the same vowels and consonants every time? If so then it’s still substitution.
No: C1 is a substitution, the Vokal Bridge is homophone, C2 ist nearly Polyphone, something between homophon and polyphon - every slot has its one Code.
Just adding an extra few symbols- the space between the words? Or am I wrong about that?
No, it's a completely different system, with bigrams, trigrams, and single glyphs. LSM / LEM, etc.
And do the vowels and consents follow Slot grammar? No, it has it one grammar.
How does that encode? i still dont know yet
And does the theory necessitate Bavarian language. No, but it seems to be Bavarian/FHND -
RE: Why and how the text could be Bavarian
JoJo_Jost > 07-06-2026, 07:42 PM
A small forward test - actually too early, but it was requested
(oshfdk)
The test sentence comes from a medical recipe context:
"Der nem minzen und mull die und trauff den saft in das or"
Modern German
Man nehme Minze und zermahle sie und träufle den Saft in das Ohr.
English: Take mint, grind it, and drip the juice into the ear.
The goal was not to prove a finished reading. The test was only meant to see whether a normal medical sentence can be transformed into a plausible Voynich-like surface using the current VBM rules.
However, many assumptions were made here. At this point I would say: this is only a first and theoretical forward test. Most of the assumptions are still very much assumptions .
Working table used
C1 values, mostly based on the frequency test:
German / VMS
n -> k
r -> t
m -> ch
s/z -> o
l/ll -> sh
d -> l
C2 values / C2 routes:
Assumptions:
d(C2)+e -> ed.y.qo
n(C2)+e -> ed.y.o
(So there are two cases using ed with different following bridge values: d+vowel e and n+vowel e. The letter is defined through the following vowel bridge. This is the phenomenon I described above.
From the crib:
z/s(C2) -> o
Further assumptions:
m(C2) -> eod
tr(C2) -> cho
Vowel bridges:
Except for e, these are assumptions - partly plausible, but not yet secure.
e -> y.qo / y.o / l.ch
i -> y.ch
ie -> l.o
u -> sh.ch
au -> r.o
a -> r.ch / r.o
o -> l.o
Ending block:
-en -> aiin
aiin is used here as the ending block for the German ending -en. There are further rules around this, but that would go too far here.
LAM / start mode:
Assumption, simply set for this test:
ch -> LAM / start mode
Collapse forms:
und mull die -> und = ol = olshedy
und trauff -> und = ol = olchor op
in das or -> ar edrol oty (Y = LEM)
"Und" (and) and other short words are probably represented by such 1/2/3-glyph words or connection forms. This is still very uncertain!
Extremely provisional value, only for this test:
ff(C1) -> p
because this is simpler than a bench gallows.
Application to the sentence
Source sentence:
der nem minzen und mull die und trauff den saft in das or
Generated Voynich-like surface:
chedy qotedy ocheody chkoaiin olshedy olchor opedy qokor chpar edrol oty
Short segmentation:
der nem minzen
-> chedy qotedy ocheody chkoaiin
und mull die
-> olshedy
und trauff den saft
-> olchor opedy qokor chp
in das or
-> ar edrol oty (Y = LEM)
---
The most important point of this small test is not the specific generated line itself. The point is that a normal medical sentence can be transformed into a clearly Voynich-like surface with the current VBM rules, if three levels are kept apart:
regular V/C stream
C1/C2 routes
collapse forms for short function words and connection groups
The result is not a proof. It is only a first forward test of the model. But in fact, with the VBM model, it is possible to generate a Voynich-like surface from a German sentence. -
RE: Why and how the text could be Bavarian
oshfdk > 07-06-2026, 10:14 PM
Thank you for this demonstration, however for a statistical test it will take a reasonable amount of encoded text (ideally the size of the actual manuscript) to be able to compare the results to Voynichese.
Note that even your short example is not very Voynich-like, with sequences like edrol, chkoaiin, opedy, that don't appear in the manuscript, even though they seem to correspond to normal frequent words and combinations of characters, as far as I can see. So far I'm under impression that this method might collapse very quickly when applied to a longer text. I'm not sure I understand how you derive the claim that "The Vowel Bridge Model fits remarkably well into the statistical peculiarities of the VMS". -
RE: Why and how the text could be Bavarian
JoJo_Jost > 08-06-2026, 05:09 AM
(07-06-2026, 10:14 PM)oshfdk Wrote: You are not allowed to view links. Register or Login to view.Thank you for this demonstration, however for a statistical test it will take a reasonable amount of encoded text (ideally the size of the actual manuscript) to be able to compare the results to Voynichese.
Note that even your short example is not very Voynich-like, with sequences like edrol, chkoaiin, opedy, that don't appear in the manuscript, even though they seem to correspond to normal frequent words and combinations of characters, as far as I can see. So far I'm under impression that this method might collapse very quickly when applied to a longer text. I'm not sure I understand how you derive the claim that "The Vowel Bridge Model fits remarkably well into the statistical peculiarities of the VMS".
You're not serious, are you?
1. If you ask for a preliminary forward test at such an early stage, and then immediately argue that only a manuscript-length statistical test would count, why ask for the preliminary test in the first place?
2. A forward test at this stage obviously requires provisional assumptions. If one of them is wrong, imperfect segments are exactly what one should expect. Treating that as a refutation of the model is a very weak and unfair argument.
The ff(C1) -> p assumption illustrates exactly this point. I explicitly marked it as extremely risky because it does not fit the frequency analysis well. So if non-Voynich-like structures appear around p in this example, that is not surprising. It shows the weakness of that particular provisional value, not of the model as a whole.
Frankly, I was more surprised by how much of the generated line looked structurally close to Voynichese than by the fact that a few segments did not.
3. Even the plaintext spelling is not fixed in MHD / FNHD. If the scribe wrote “or” (ear) as “ohr,” then the word “or” would not exist in the underlying text and therefore cannot be expected to have a direct VMS counterpart. The same applies to “minzen,” which could have been written as “minze.” So criticizing the generated output as if every normalized plaintext word must have a direct visible VMS equivalent is methodologically unsafe.
4. One point you have clearly overlooked is this: at the level of argument you are using, even producing 200 pages would not prove the model. The VMS is clearly based on a slot-like system, as Stolfi has shown. If one simply maps frequent plaintext elements onto frequent Voynich glyphs, bigrams, and trigrams within that slot system, a Voynich-like surface will inevitably emerge. Treating that as proof would be circular.
----
For me, the forward test did exactly what it was supposed to do: it showed that the proposed structure is operational. In other words, you can actually use the VBM rules to transform a German medical sentence into a Voynich-like surface.
That does not prove the model. I never claimed that it did. Anything beyond that would be an exaggeration.
But using a few imperfect segments from a first provisional forward test as a refutation of the claim that the VBM fits the statistical peculiarities of the VMS is simply bad methodology! It does not engage with the actual statistical argument; it only shows that parts of the current forward table are still incomplete.
Sorry, oshfdk, but please stay fair.
-
RE: Why and how the text could be Bavarian
oshfdk > 08-06-2026, 07:27 AM
(08-06-2026, 05:09 AM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.Sorry, oshfdk, but please stay fair.
I think I'm being fair. What I suggested was to
Quote:...take some text and encode it using this model. If the result is indeed statistically similar to Voynichese, this to me will be a good argument.
One sentence doesn't provide enough data for any statistical assessment, so what you did was just a demonstration of how the encoding works, not in any way the test I suggested.
But then since you stated that this one sentence also looked like Voynichese, I just commented on parts of it that don't look like Voynichese.
Note that if you for some reason don't really want to discuss your method with me, I'm ok not discussing it. I was just suggesting a way to actually showcase your work.