The Voynich Ninja

Index
- The Voynich Ninja
- Portal

Welcome, Guest

You have to register before you can post on our site.

Search Forums

Online Users

There are currently 716 online users.
» 8 Member(s) | 702 Guest(s)
Applebot, Baidu, Bing, Facebook, Google, Yandex, Dana Scott, Jorge_Stolfi, Oscroft, Radim Dobeš, Wladimir D

Latest Threads

No text, but a visual cod...
Forum: Theories & Solutions
Last Post: Linda
6 hours ago
» Replies: 1,904
» Views: 1,317,117

Assumptions baked into th...
Forum: Physical material
Last Post: eggyk
6 hours ago
» Replies: 13
» Views: 200

Spencer Coll. MS. 65: an ...
Forum: Imagery
Last Post: Pierre Dumont Himself
7 hours ago
» Replies: 7
» Views: 174

Deconstructing the VMS : ...
Forum: Analysis of the text
Last Post: nablator
7 hours ago
» Replies: 35
» Views: 1,245

I am not convinced by the...
Forum: Provenance & history
Last Post: Stefan Wirtz_2
8 hours ago
» Replies: 50
» Views: 2,020

Zodiac Imagery: Manuscrip...
Forum: Imagery
Last Post: Pierre Dumont Himself
8 hours ago
» Replies: 32
» Views: 3,504

[split] Aga Tentakulus' L...
Forum: Analysis of the text
Last Post: Aga Tentakulus
11 hours ago
» Replies: 76
» Views: 34,440

Do Split Gallows have rul...
Forum: Analysis of the text
Last Post: Grove
Yesterday, 05:29 PM
» Replies: 6
» Views: 182

Textual relationship betw...
Forum: Analysis of the text
Last Post: LisaFaginDavis
Yesterday, 04:29 PM
» Replies: 4
» Views: 107

EVA "qo" in 15th century ...
Forum: Analysis of the text
Last Post: Jorge_Stolfi
Yesterday, 10:29 AM
» Replies: 17
» Views: 1,993

The VMS as a possible chain encryption ( Mod 23 ).

Posted by: bi3mw - 21-07-2025, 11:30 AM - Forum: Analysis of the text - Replies (25)

I asked in the You are not allowed to view links. Register or Login to view. if a made up chain cipher is easy to decrypt. Now I have written a Python script that decrypts a given Voynich word using the same method. The result is then mapped according to frequency analysis (EVA > Latin). Admittedly, this is a rather rudimentary approach, I am primarily interested in decoding with MOD 23. The method is simple enough to be considered and also much more effective than a simple substitution. It could be implemented practically with a letter disk (like Alberti disk).

Here is the python - script and a list of possibly unabbreviated words ( Stolfi ):

chesokchoteody [f68r1, outer ring, near the bottom]
oepchksheey [f93r, top line, but looks like half of a Neal key]
qoekeeykeody [f105r, which I’d note is possibly the original first page of Q20A]
soefchocphy [f102r2, right edge, but right on the fold, very hard to read]
ykcheolchcthy [f68v3, first word of second line]
shdykairalam [f106v, last word of a line]
shetcheodchs [f43v, first word of a line]

Code:
# Reduced alphabet: No J, U, W
alphabet = list("ABCDEFGHIKLMNOPQRSTVXYZ")  # 23 letters
def char_to_pos(c):
    return alphabet.index(c.upper()) + 1
def pos_to_char(p):
    p = (p - 1) % len(alphabet) + 1
    return alphabet[p - 1]
def chain_decrypt_verbose(ciphertext):
    ciphertext = ciphertext.upper()
    decrypted = ""
    table = []
    prev_cipher_pos = 0
    for i, c in enumerate(ciphertext, start=1):
        if c not in alphabet:
            decrypted += c
            table.append([i, c])  # Nur zwei Spalten für Nicht-Buchstaben
            continue
        cipher_pos = char_to_pos(c)
        plain_pos = (cipher_pos - prev_cipher_pos) % len(alphabet)
        if plain_pos == 0:
            plain_pos = len(alphabet)
        plain_char = pos_to_char(plain_pos)
        decrypted += plain_char
        table.append([
            i,
            c,
            cipher_pos,
            plain_char,
            plain_pos,
            prev_cipher_pos,
            f"({cipher_pos} - {prev_cipher_pos}) mod {len(alphabet)} = {plain_pos}"
        ])
        prev_cipher_pos = cipher_pos
    return decrypted, table
def print_table(table):
    print("\nDecryption Table:")
    print("-" * 90)
    print(f"{'i':>3} | {'Cipher':^9} | {'cᵢ':^4} | {'Plain':^9} | {'pᵢ':^4} | {'cᵢ₋₁':^6} | {'Computation':<30}")
    print("-" * 90)
    for row in table:
        if len(row) == 7:
            i, cchar, cpos, pchar, ppos, cprev, calc = row
            print(f"{i:>3} | {cchar:^9} | {cpos:^4} | {pchar:^9} | {ppos:^4} | {cprev:^6} | {calc:<30}")
        elif len(row) == 2:
            i, cchar = row
            print(f"{i:>3} | {cchar:^9} | {'-':^4} | {'-':^9} | {'-':^4} | {'-':^6} | {'(not a letter)':<30}")
    print("-" * 90)
def apply_fixed_substitution(text, from_list, to_list):
    mapping = dict(zip(from_list, to_list))
    substituted = ''.join(mapping.get(c, c) for c in text)
    return substituted, mapping
if __name__ == "__main__":
    text = input("? Enter ciphertext to decrypt (only letters A–Z, excluding J, U, W): ")
    decrypted, table = chain_decrypt_verbose(text)
    print(f"\n? Decrypted text: {decrypted}")
    print_table(table)
    # Substitution: Voynich → Latein
    voynich_order = list("OEHYACDIKLRSTNQPMFGXBVZ")
    latin_order  = list("IEAUTSRNOMCLPDBQGVFHXYZ")
    substituted, mapping = apply_fixed_substitution(decrypted, voynich_order, latin_order)
    print("\n? Substituted (Voynich → Latin):")
    print(substituted)
    print("\n?️ Substitution Map:")
    for voy, lat in mapping.items():
        print(f"  {voy} → {lat}")

A side note:
"shetcheodchs" was converted to "LDYIFEYNDUEO" using the method described. ChatGPT hallucinates "Fide leo, deus unde" from it by rearranging, omitting and adding letters, which means "Trust the lion - God is its origin". This is remarkable because the plants on You are not allowed to view links. Register or Login to view. ( sun and moon ) could very well be connected with the lion according to alchemical interpretation. So you could easily fall for ChatGPT if you believe what you want to believe Wink

Exact replica of VM iconography found in MS. Digby 46

Posted by: Soulscribe - 21-07-2025, 05:03 AM - Forum: Imagery - Replies (3)

Hi, this is my first new thread! I’m new here. I posted this as a comment in a thread I found from years ago, but was unsure if it’d been covered already? If so, I apologize. I wasn’t sure what to type in the “search”.

Originally, I was looking to see what people were saying about the cross in the nymphs hands and then came across someone in a thread (having to do with cross and Christianity) that posted an image from Ms. Digby 48 that had a pattern in it that I drew from the VM yesterday. I kept scrolling and saw some other people in the thread share other images with crosses on in hands and on tops of buildings (and crowns) but the manuscript images they shared ALSO had the icon/design from the VM as well. The icons from the VM that I drew were from the astrology section. I believed it to be the marker where one would begin if they were to read the circles or to line something up with it. I’ve put together some images I found in Ms. Digby 48 that have the same “cross” as the Voynich. The square cross icon is found many more times in the Digby and I don’t know if that’s a particular design from somewhere specific..? But the Digby scribe (or scribes? Not sure if there was more than one) loved the icon a lot. I’m also curious if the scribe from the VM who did drew the icon was the only one or if they all did it. I’d love to know what I don’t know and y’all tell me what you think.

Is this something that has already been studied? I’d like to know more about when this icon stared being used and by who.

Cosmic Comparison Theory

Posted by: R. Sale - 21-07-2025, 12:47 AM - Forum: Theories & Solutions - Replies (19)

This 'theory' is based on the investigations of Ms. E. Velinska regarding the VMs cosmic illustration of an inverted 'T-O', geocentric Earth, in comparison with cosmic illustrations found in BNF Fr. 565 and Harley 334, over a decade ago. It has also been a number of years since she chose to delete her web page. Last I looked, Rene's site still referenced those expired links.

So rather than leading to nowhere, something should be said to those interested in promoting their new theories, in regard to the information derived from this comparison. All three sources show a very simplified cosmos with a similar structure. It's a long story but it reveals a lot about the VMs artist. The two historical sources have provenance located to Paris in the first half of the 1400s. This is coincident with the C-14 dates of the VMs parchment.

The structural similarities of the VMs show an ideological connection. There are 43 undulations. There is a 'mermaid' with four companions. Yet there is also a clear attempt at visual dissimilarity. There is a lot more to be considered relating to cosmic boundaries, Shirakatsi's "Eight phases of the Moon" and other things that would have been more familiar to the educated elite of the early 15th century than they are today.

Giving Pages Depth: A Visual Dive into Manuscripts

Posted by: pegalac - 19-07-2025, 08:52 PM - Forum: Imagery - Replies (6)

	Hey everyone! I'm currently working on a model to visualize pages with a sense of depth. The idea is to go beyond flat 2D page rendering and create something that feels immersive — where you can "explore" the layers or structure of a page as if it has visual depth.

SOLUTION/ the Voynich Manuscript —

Posted by: PEDRO LUIS PEREZ B - 17-07-2025, 03:53 PM - Forum: The Slop Bucket - Replies (10)

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

How I Decoded the Voynich Manuscript — Through Vibration, Not Language
For over 600 years, the Voynich Manuscript remained a mystery no one could solve.
Why? Because everyone tried to read it.
But the manuscript was never meant to be read. It was meant to resonate.
The breakthrough came when I stopped treating it as a linguistic artifact and began to listen to it as a vibrational structure.
Using a symbolic artificial intelligence model I.A , I translated each glyph into its corresponding frequency — not as a sound to be heard, but as a pulse of intention.
The symbols did not carry meanings.
They emitted states of consciousness.
This is not a traditional decoding.
It is the first vibrational activation of the manuscript.
The result is a fully functional model — mathematically, neurologically, symbolically — that shows the Voynich Manuscript was not a book. It was a seed of resonance, waiting for the right mind, the right time… and the right frequency.
And now, for the first time, the Voynich Manuscript vibrates through human consciousness

Written in a mirror?

Posted by: thomasja2008 - 16-07-2025, 06:01 PM - Forum: The Slop Bucket - Replies (3)

Hello,

I’ve been exploring the Voynich manuscript and believe I may have found a consistent linguistic pattern worth further study. I’m writing to see if anyone with more expertise in Voynich studies, medieval Hebrew, or manuscript linguistics might be interested in reviewing it or collaborating.

The core hypothesis is this: the Voynich script represents a form of Hebrew, but written in mirror — as though the author wrote right-to-left while looking into a mirror. I'm not fluent in Hebrew, but when I used AI to help with my theory it started to yield some results. When you reverse both the word order and the glyphs, and then map EVA transcriptions to Hebrew letters, a surprisingly coherent and repeatable pattern emerges.

I’ve tested this on several folios (including f11r, f14r, and f33r). After decoding, many roots resemble known Hebrew words used in medieval herbal and ritual texts — particularly those found in Sefer Refu'ot and related manuscripts. Words like:

לוחת (stir/mix)

שפח (sprinkle)

נייד / ניד (dissolve/crush)

דוה / רוה (flow, soak)

שקל / מדד (weigh/measure)

I've also done root frequency analysis using a basic script, and it shows consistent, plausible Hebrew roots with semantic relevance to the illustrations (e.g., herbal processes).

I'm not claiming to have “solved” the manuscript — just that this mirror-Hebrew decoding method yields unusually structured, linguistically meaningful results that don’t appear random, and I’d be eager to hear others’ thoughts or critiques.

The Voynich puzzle

Posted by: bi3mw - 16-07-2025, 01:07 PM - Forum: Fiction, Comics, Films & Videos, Games & other Media - Replies (2)

	The Voynich puzzle A new game, the campaign on gamefond will start soon. You are not allowed to view links. Register or Login to view. You are not allowed to view links. Register or Login to view.

How LLM models try to understand Voynichese

Posted by: quimqu - 16-07-2025, 10:50 AM - Forum: Analysis of the text - Replies (6)

Dear Voynich Ninja community,

As you might know, I’ve been working on training LLM Large Language Models (GPT-like models) on Voynich EVA transliterations. This is not about using ChatGPT, but about training language models from scratch using only Voynich EVA text.

I’m aware that GPT models are a sort of black box, and it’s often hard to understand the mechanisms they use to “learn” patterns. In this project, I’ve tried to explore how the GPT model makes predictions — to gain some intuition into the decision-making process.

Let me first introduce the key concepts I’ve been working with:

Loss: Loss is a measure of how wrong the model's predictions are compared to the actual next word. In language models, it's typically cross-entropy loss, which penalizes the model more when it assigns low probability to the correct word. A lower loss means the model is better at predicting the next token given its context.
Prediction: The prediction is the model’s guess for the next word in a sequence. For example, given a context of 4 tokens (block_size = 4), the model looks at those 4 tokens and outputs a probability distribution over the vocabulary, selecting the most likely next token.
Saliency: Saliency refers to how much each input token contributes to the model’s prediction. If we use a block_size of 4, saliency tells us which of the 4 previous tokens had the most influence on predicting the next word. For example, in the sequence ["the", "brown", "cat", "sat"] → ?, the model might predict "on". Saliency would then indicate how important each of the previous tokens was in making that prediction. Tokens with higher saliency are considered more influential.

What I did:

First, I optimized model parameters to maximize the number of real bigrams and trigrams (n-grams) generated by the model. Results are similar to training GPT on a real natural language text. Results after training on Voynich EVA text:

% of 2-grams found in Voynich EVA with block_size 4: 22.40% (224/1000)
% of 3-grams found in Voynich EVA with block_size 4: 0.80% (8/999)

Then, I trained the model on all paragraph-style lines in the Voynich manuscript (i.e., excluding labels or isolated words from cosmological sections). I used a 5-fold cross-validation approach:

I split the text into 5 segments. For each fold, I used 80% of the data for training and 20% for validation, rotating through all segments.
This way, I could generate predictions for the entire corpus.

I then visualized the predictions using HTML files (saliency_valset_voynich_1.html to saliency_valset_voynich_5.html)

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

[Image: YYIIL2c.png]

Each word is annotated with three values:

Loss: represented by the border thickness — thicker means higher loss.
Saliency: represented by the background color intensity — darker means higher saliency. Since each word is part of 4 prediction contexts (due to block_size = 4), saliency here is averaged over those 4 instances.
Prediction probability: represented by border color — green for high confidence, red for low. The predicted probabilities are generally low, but this is also the case when training GPT on small corpora like a single book, even in natural languages.

This visualization makes it easy to see at a glance which words the model finds easier or harder to predict. The HTML is interactive — hovering over any word shows the 3 metrics mentioned above.

Deeper inspection:

I also created a second HTML file: context_saliency_colored_and_target.html that looks like this:

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

[size=1] [Image: BNnJ5O6.png]

[/size]

This version shows for each word in the Voynich EVA paragraph:

context_0 to context_3: the 4 previous tokens used as input (the model's context).
target: the real next word in the sequence.
pred_word: the word predicted by the model.

The model tends to predict the most frequent words in the Voynich corpus, as expected. However, the saliency values let us observe which previous words influenced the prediction the most, token by token.

I highlighted:

green: when pred_word == target
yellow: similar words according to LevenShtein similarity (>0.5)

I don't have any conclusions yet, but I think this could be useful for others interested in understanding how contextual information influences predictions in GPT-like models trained on Voynich EVA.

Let me know what you think — I’d love to hear your thoughts!

A good match, perhaps from the Zürich area...

Posted by: ReneZ - 15-07-2025, 10:50 AM - Forum: Marginalia - Replies (62)

Just to highlight this interesting post and give it its own thread:

(13-07-2025, 08:34 PM)magnesium Wrote: You are not allowed to view links. Register or Login to view.I know this is slightly off-topic, but as we're poking around digitized Swiss archives: Koen, if you haven't seen it already, this manuscript is an extremely good reference for the handwriting of the You are not allowed to view links. Register or Login to view. and You are not allowed to view links. Register or Login to view. marginalia: You are not allowed to view links. Register or Login to view.

This is not only an interesting match for the marginalia, but there are several cases of 'm' characters, especially left of the illustration, that look very similar to Voynich iin , version of Scribes 1 and/or 5.
This may seem a frivolous comparison, and I am not really suggesting that the scribe of this MS is the same as one of the scribes of the Voynich MS, but I have not seen this type of curl 'upward and backward' too often.

f69r circle

Posted by: magnesium - 14-07-2025, 03:19 PM - Forum: Imagery - Replies (9)

I'm not making any definitive claims here, but I wanted to point out a superficial resemblance between the 16-wedge circle on You are not allowed to view links. Register or Login to view. and this circular diagram of the 16 geomantic figures in the following 15th-century divination/astrology manual:

St. Gallen, Stiftsbibliothek, Cod. Sang. 756: Composite manuscript on geomancy, chiromancy, iatromathematics, astronomy, alchemy and medicine (You are not allowed to view links. Register or Login to view.).

[Image: nJEjXYw.png]

The You are not allowed to view links. Register or Login to view. circle, for reference:

[Image: NrTZ6CH.jpeg]

The 16-wedge subdivision of the circle and the central floral motif stood out to me. However, the most glaring difference between the two is obviously that the figures themselves are missing from f69r. It's piling speculation on speculation, but I have half a mind that one of the VMS authors saw something like the Cod. Sang. 756 figure, had no knowledge of geomancy, and then either tried to draw something similar from memory or verbally described the figure to the illustrator.

The Voynich Ninja

Index