The Voynich Ninja

Thank you, Stolfi, for your detailed and constructive feedback. You raise several valid points—the separation of the VMS sections, more careful filtering of labels, normalization of Hw for the sample size, and the question of p and f. All of these are improvements that make sense.

But I would like to point out that entropy analysis is really not my main area of expertise—I simply wanted to show that Bavarian is well suited as the basis for a possible absorption cipher to approximate Voynicheese.

So this was intended as a first rough comparison to see if the numbers were roughly correct, not as definitive statistical proof.

The fact that you describe the Hw result as a “match” is actually quite encouraging for my hypothesis.

As far as “unencrypted European natural language” is concerned

As I said, I'm not yet convinced by the Chinese theory, but if more evidence emerges, I may have to abandon my Bavarian theory with a smile and a tear in my eye. Until then, I'll continue a little further. Purely based on gut feeling, Bairish simply fits much better into the context of the VMS – but feeling is no proof.

4. First steps towards a solution

I have now expanded the cipher – basically, I have structured it according to frequency in the original Bavarian comparison texts. A lot of tests prove that this is the form that best matches the frequency of the letters and the cipher.

The amazing thing is that the many tests I have now carried out tend to support the system rather than destroy it. This is a good indication, even if it is not proof.

A very impressive example is the EVA ‘s’, which is my search term ‘und’ (and) – one of the most common words in German.

s standing alone: 302×, at the beginning: 908×, at the end: 825×
Total s: 5.87% vs MHD and/vnd: 5.9% — almost identical
201 of 281 (72%) s-initial words also exist without s in the VMS
Result: s = and
I think it's a Brevigraph


The problem remains that the cipher leaves too many possibilities open; what I am now showing is only one of many possibilities. In addition, we still do not know the exact underlying dialect. That will be a lot of work.

Nevertheless, individual words can be recognised (bold) and a sentence structure that makes sense.

So this is what the first deciphered text (Bavarian Theorie) of the VMS looks like:


CHIFFRE
=======

ch = n/m
e = e/i
y = e / y
d = s/z/tz/ss/ß
o = o/u
a = a/o
l = l (oder r?)
r = r (oder l?)
sh = s/sch/ch
tsh = tsch
ee = eu/au/ou/ui
ai = ai/ei
t = t/d
k = w
p = f/v
f = ?
cth = b
ckh = g
cph = h
cfh = f?
o- = Artikel
qo- = Präp+Art
y- = ge/ver/be/zer
s = und
aiin/ain = en/ain/ein
m=? (here m)

You are not allowed to view links. Register or Login to view. ====

L1  tsholdchy qotchy opchear ypchedy
    tscholsne in die tne (tinne) die fnear und gefnese

L2  dcheodaiin ckheody ckhody chody
    sneosen (schnäuzen) geose gose nutze

L3  ydchody ckhey oty ckheodar qoky
    gesnose gee ote geosar in die we

L4  daiinces okol cheody ckhy cheeey
    sence und die wol (wohl) neose ge neuee

L5  tcheody qodaiin okeey qockhey taiin
    tneose von dem sein die weue in die gee den

L6  ycho daiin chokaiin ykchodaiin ykald
    verno sein nowen gegossen gewals und

L7  ychos ar eeckhy kcho qokchy qotal
    geno und ar euge (auge) wno in die wne von dem tal

L8  oshol odaiin ckhey ckheody qokey otydy
    die schol (soll) das sein gee geose von dem wee (schmerz) die tese und

L9  tol daiin daim qchodal dal qody qoetam
    tol sein sam und von dem nosal sal und in die se in die etam

L10 ykchol dor shey qokeol kchey shol okam
    gegol sor sey von dem weol wnee schol (soll) die wam

L11 tchodaiinoeody qokol oteodaiin kol otag
    tnosenoeose und von dem wol (wohl) die teosen und wol die tag

L12 yoees ckheey kol cheeal okeor qockhey pchodal
    geoeu und geue wol maual die weor in die gee fnosal und

L13 oaiin ckhol ykieol otchey cpheo daiin ykeoldy
    ein gol gewieol die tnee heo sein geweolse und

L14 daiiithy qodaiin kaiiidal cphodal s al dam
    senithe von dem sein wensal und hosal(hose?) und al sam und

L15 qokol cheor ckhal s or aldy otal
    von dem wol neor gal und or (oder) alse die tal

a good starting point...

Addendum, I forgot to mention: What becomes clear is that the structure of the sentences, articles, conjunctions and prepositions shows that VMS can actually represent a real language with this code....

Here it is in English for better understanding, and then with additional options that you wouldn't otherwise see if you don't speak German.

L1  tsholdchy qotchy opchear ypchedy
    tscholsne (keyflower?) into the tinne (vessel) the fnear and gefnese

L2  dcheodaiin ckheody ckhody chody
    sneosen (blow nose) geose gose nutze (apply)

L3  ydchody ckhey oty ckheodar qoky
    gesnose (blown out) gee ote geosar into the we

L4  daiinces okol cheody ckhy cheeey
    sence and the wol (well, probably) neose ge neuee

L5  tcheody qodaiin okeey qockhey taiin
    tneose of the sein (his/its | be) the weue into the gee the

L6  ycho daiin chokaiin ykchodaiin ykald
    verno sein (his/its | be) nowen gegossen (poured) gewaltz (rolled/kneaded) and

L7  ychos ar eeckhy kcho qokchy qotal
    geno and ar euge (eye) wno into the wne of the tal (part)

L8  oshol odaiin ckhey ckheody qokey otydy
    the schol (shall) the sein (his/its | be) gee geose of the wee (pain) the tese and

L9  tol daiin daim qchodal dal qody qoetam
    tol sein (his/its | be) sam (together?) and of the nosal sal and into the se into the etam

L10 ykchol dor shey qokeol kchey shol okam
    gegol sor (sour) sey (be | sei) of the weol wnee schol (shall) the wam

L11 tchodaiinoeody qokol oteodaiin kol otag
    tnosenoeose and of the wol (well) the teosen and (wol, probably) the tag (day)

L12 yoees ckheey kol cheeal okeor qockhey pchodal
    geoeu and geue wol (well) neual the weor into the gee fnosal and

L13 oaiin ckhol ykieol otchey cpheo daiin ykeoldy
    ein (a/one) gol gewieol the tnee heo sein (his/its | be) geweolse and

L14 daiiithy qodaiin kaiiidal cphodal s al dam
    senithe of the sein (his/its | be) wensal and hosal and al sam (all together) and

L15 qokol cheor ckhal s or aldy otal
    of the wol (well) neor gal and or (or) alse (also) the tal (part)

Exciting, I'm currently testing a new page, and logically I'm now at 52r.

The first word “tdokchcfhy” is another hapax, and we'll leave out the t again, probably a ‘t’ = “der die das” remains dokchcfhy = tzownpfe = Zaun Pfeffer = fence pepper

There is a plant called "fence garlic", which tastes like garlic and whose seeds taste like pepper. It was the poor man's pepper. It was probably also called fence pepper. Today we know it as “garlic mustard.”

And indeed, the plant has a lot of similarities. The small flowers, the slightly frayed leaves, and rhizomes.
So can we identify the plants with this cipher? Really?

And in the deciphered text, oil, vinegar (sour), and seeds seem (!) to appear, which are exactly the ingredients and instructions you would expect for a garlic mustard recipe...

I think I'll translate several texts over the next few weeks and look for more plants... I'll get back to you when I have more reliable results....

It's strange, could the VMS be so easy to solve? I still have my doubts...

Hey JoJo,

When I look at the translation, there seems to be some wiggle room in the mappings. Have you tried doing a robust 1:1 mapping and then running it across the folio? That would really strengthen the results..

(28-02-2026, 10:22 PM)DG97EEB Wrote: You are not allowed to view links. Register or Login to view.Have you tried doing a robust 1:1 mapping

That's not how the game is played. You're allowed to cheat in every way possible.

(28-02-2026, 10:26 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.

(28-02-2026, 10:22 PM)DG97EEB Wrote: You are not allowed to view links. Register or Login to view.Have you tried doing a robust 1:1 mapping

That's not how the game is played. You're allowed to cheat in every way possible.

Unfortunately true, but that's exactly what I described from the beginning. Of course, the code could be like that, but then it wouldn't be my fault

Still, strange coincidences...

---

(28-02-2026, 10:22 PM)DG97EEB Wrote: You are not allowed to view links. Register or Login to view.Hey JoJo,

When I look at the translation, there seems to be some wiggle room in the mappings. Have you tried doing a robust 1:1 mapping and then running it across the folio? That would really strengthen the results..

Unfortunately, the code isn't that simple...there are many variations.


But the code is described in detail—in all its variants! There are no other variants beyond these.

So, after two weeks’ holiday spent searching for sunshine – which meant there was plenty of sun here in Germany, whereas where I went, Portugal, there was no sun, just rain... well, let’s get back to the Bavarian idea.

I have tried to translate a sentence; to do so, I’ve chosen one that already contains many of the possible translations of the words I’ve found in other context. I would like to point out that this is, of course, still much Eisegesis at the moment. But it is also intended to check whether one can actually form reasonably meaningful sentences from the cipher presented here and the VMS.

Please bear with me, as it is, of course, still very rough. The actual work of identifying more and more words that gradually make sense in various sentences will take years. But I’ll simply write my progress – and also my setbacks – here; as I’ve already said, this thread is a work in progress...

Possible translations using my cipher that I have come across in other contexts so far
sar=sauer (sour)
or/ar = er (he / one)
aiin=ein (a)
shol=soll
okain=der Wein(o=articel the vine)
dal=sal(be) (salve)
dy=sy / sein (be)


EVA You are not allowed to view links. Register or Login to view. 9: sar or aiin chor or shkair shol or chckhy ar aiin okain dal dy 


new found possible words:
shkair=swair/ schwer (dickflüssig) (heavy/strong,(thick))
chckhy=mege = menge (mix/stir) (ckh= g / ng)
chor=mor / mache (make)


Bavarian: sar er ein mor er schwer soll er menge er ein den Wein sal sy

sauer er ein  / mache er schwer / soll er menge er ein den Wein / salbe sy


It ist difficult, to translate the dialect in english.... I try it: 

"Sour it in [pickle/steep it], make it thick, one shall mix in the wine — salve it be [a salve results]."

"The impersonal er (he = 'one') is characteristic of medieval Bavarian recipe language, equivalent to modern German man. The verb-subject inversion (mache er, soll er, menge er) is standard MHD word order for instructions. 


What’s interesting here is that this is quite clearly a recipe section. And I haven’t changed a single thing about it; the sentence was suggested to me by Claude (not translated by claude!!!). Claude simply helped me find a sentence in which the words mentioned above appear frequently - here were no further instructions, and of course Claude didn’t know either that it was supposed to be a recipe! In that sense, it’s a bit of a coincidence… which is quite intriguing.

But I realise that this is just a starting point, a beginning – nothing more. We’ll just see how things develop...
The search continues...