Je publie aujourd'hui les résultats complets d'une analyse statistique et linguistique ressortent sur le manuscrit de Voynich (corpus EVA).
**Thèse :** Le manuscrit est un manuel de procédures encodé en base 60 babylonienne (sexagésimal) avec des racines sumériennes phonétiques. Il ne s'agit pas d'un récit, mais d'un ensemble de données.
**La méthode :** Une seule règle d'analyse syntaxique déterministe (« Maximal Munch ») a été appliquée à 12 079 mots (3 727 uniques). Aucune exception, aucun tri sélectif.
Règle : - **Préfixe** = Premier digramme reconnu (correspondance la plus longue) → Valeur entière (base 60) - **Suffixe** = Fin du mot de la liste fermée → Valeur fractionnaire (base 60) - **Corps** = Caractères restants → Contenu phonétique
**Principaux résultats numériques :** - Préfixe `ok-` = 31 → Position des Pléiades (environ 31° dans l'Antiquité, correspondant au calcul de précession pour l'an 0) - Préfixe `ot-` = 32 → Culmination solaire - Préfixe `qo-` = 30 → AN (ciel) - Préfixe `ch-` = 24 → G (grand) - Préfixe `sh-` = 25 → MUD (sang) - Préfixe `cth-` = 26 → KUR (montagne/terre) - Suffixes `-ol` = +2/60, `-al` = +1/60, `-ar` = +3/60, `-or` = +4/60
**Note statistique :** Coefficient de Zipf = 0,81 (plus plat que le langage naturel ~1,0, cohérent avec les données tabulaires/techniques).
**Ce que je demande :** Vérifiez les calculs. Examinez les corrélations sumériennes. Trouvez les erreurs. L'algorithme d'analyse syntaxique complet et le lexique complet sont ci-dessous.
According to recent results of Lisa Fagin Davis (which was also suggested earlier) Voynich Manuscript wasn't originally bound a a book but was a set of loose sheets.
You are not allowed to view links. Register or Login to view.
If I understand it correctly binding not only messed the original order of bifolios but also made you see only half of a bifolio together with a half of some other, random bifolio when you open a book which was never originally intended.
Take Quire 13:
Let's take an example. If I understand correctly page 78v is on the same same side of bifolio as 81r.
(Actually it shouldn't even be called be bifolio because such term suggests belonging to a book but lets leave it)
So the original author who worked with loose sheet of calfskin actually saw with his eyes something like this:
Is it possible somewhere online to browse Voynich Manuscript this way?
Perfectly I would like to have an update of the websites like You are not allowed to view links. Register or Login to view. and voynichese.com but don't know if it is going to happen
You are not allowed to view links. Register or Login to view. looks to me like a brussels sprouts plant with the leaf buds twisting around the stalk like a vines, and the foliage on top as well, the leaves on the side are also pretty common with brussels sprouts, the empty stalk on the bottom is also something you can see on the plant.
You might say that the leaves look more like those of an oak and that the leaves on a brussels sprouts plant are too smooth, But it might be a different cultivated breed from the 15th century seeing that there are many vegetables that have curlier leaves and that originate from the wild cabbage (savoy, collard greens, kale, kalettes and wild cabbage itself).
Hi, all, just wanted to let you know that the account book cited by Stefan Guzy that records the purchase of books from Widemann on behalf of Rudolf for 600 Fl. is now available online: You are not allowed to view links. Register or Login to view.
The relevant transaction is on line 2 of image 48 (direct link above).
These images are much more clear than the reproduction in Guzy's 2023 publication, in which this is plate 7 (“Rudolf II. und das väßl mitt allerlai selzamen büchern. Neue Erkenntnisse zur älteren Besitzgeschichte des Voynich-Manuskriptes,” Neuerscheinung: Herold-Jahrbuch, Neue Folge, 28 (2023), 65-143)
You are not allowed to view links. Register or Login to view.
This drawing looks a lot like the drawing in You are not allowed to view links. Register or Login to view.r.
You are not allowed to view links. Register or Login to view.
On the left side it has been numbered just like You are not allowed to view links. Register or Login to view..
You are not allowed to view links. Register or Login to view.
This drawing looks similar to 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. (again).
I hope this can be of use, although it isn't from the same time period as the voynich manuscript.
(This is my first post, i would've used pictures to make it easier to look at but they didn't work so i just put link to the pages)
Desde mi punto de vista el voynich no se traduce porque los medios vistos hasta ahora están mal, el modelo Eva por ejemplo no está del todo bien aparte de que solo intentan traducirlo no comprenderlo.
Good day everyone. This is my first post here. I am an independent reader of the Voynich Manuscript, approaching it from a structural and statistical perspective rather than a decipherment-oriented approach. I would like to share a theory paper. This paper does not propose a decipherment, language identification, or hidden plaintext for the Voynich Manuscript. Instead, it advances a functional hypothesis intended to account for several well-documented statistical and structural anomalies—most notably position-dependent entropy, low conditional entropy, and extreme positional rigidity. The framework presented here treats the text as a structured visual system rather than a semantic narrative. It is offered as a provisional, falsifiable model, not a conclusion. The goal is to test whether known quantitative properties of the manuscript can be coherently explained without invoking linguistic encoding. I am presenting this analysis for critical review and welcome corrections, counterexamples, or prior work that directly contradicts the proposed mechanism.
The Voynich Codex as a Structured Visual System with Position-Dependent Entropy
Abstract
This paper proposes a novel functional framework for the Voynich Manuscript (VMS), positing that the codex is not a linguistic vessel for information storage, but a structured visual system compatible with sustained attentional regulation under controlled visual scanning conditions. This analysis focuses on the unique statistical anomalies of the script, particularly position-dependent entropy (Lindemann 2020; Schinner 2007). Building on these analyses, the paper proposes that the text functions as a visual scaffold rather than a semantic narrative. It introduces the Gate–Scaffolding Model, in which line-initial tokens (“Gates”) calibrate local cognitive load, while subsequent low-entropy text (“Scaffolding”) facilitates a non-narrative, rhythmic boustrophedon scanning pattern. Furthermore, illustrations are proposed to function as high-contrast interference stimuli that engage parafoveal processing and spatial mapping, modulating spatial attention across foveal and parafoveal visual fields. This framework is exploratory and intended for discussion; it does not assert authorship or deliberate purpose.
References
Lindemann (2020), Statistical Patterns in the Voynich ManuscriptYou are not allowed to view links. Register or Login to view.
Schinner (2007), The Voynich Manuscript: Evidence of the Hoax HypothesisYou are not allowed to view links. Register or Login to view.
Scaffolding – Text following the Gate, highly predictable, supporting hypothesized visual-attentional flow.
Boustrophedon – A style of text layout in which lines are read alternately left-to-right and right-to-left, mimicking the way an ox plows a field. In this paper, it refers to the hypothesized scanning pattern of the Voynich text, where readers’ eye movements alternate direction line by line, increasing spatial remapping demands and sustaining visual-attentional engagement.
Terminals – Line-final tokens providing vertical anchors during scanning.
Lens Protocol – A conceptual framework introduced in this paper for analyzing the Voynich Manuscript as a structured visual system rather than a linguistic text. It models how line layout, token positioning, and visual features interact with foveal and parafoveal attention. “Tactical” refers to foveal engagement with the script, while “Strategic” refers to parafoveal processing of surrounding text and diagrams. The protocol draws on empirical observations of position-dependent entropy (Schinner 2007) and employs simulated boustrophedon scanning to examine how these visual patterns sustain attentional flow.
Foveal Anchor – The operational role of the Scaffolding during active scanning; the low-entropy text region that stabilizes foveal eye movements.
1. Introduction: Why Linguistic Models Remain Inconclusive
Traditional cryptanalytic approaches to the Voynich Manuscript rely on the assumption that the script encodes a natural language. However, decades of statistical analysis reveal a paradox: while the script adheres to Zipf’s Law—suggesting linguistic structure—it exhibits extraordinarily low conditional entropy (h₂ ≈ 2.0 bits per character), substantially below the range observed in natural human languages (approximately 3.0–4.0 bits). Critically, unlike shorthand or repetitive litany, this low entropy is tightly linked to each word's position on the line. While entropy alone does not falsify linguistic hypotheses, the combination of low conditional entropy with extreme positional rigidity places the script outside known natural language families. This paper argues that this low-entropy characteristic is not necessarily an artifact of encryption, but may instead reflect a functional property of the system. The Lens Protocol is proposed as a model in which the text is compatible with reading conditions associated with reduced narrative self-referential processing, sometimes discussed in relation to Default Mode Network (DMN) activity. While word-level distributions roughly follow Zipfian curves, the conditional character entropy suggests a system where 'words' are emergent properties of a character-generation algorithm rather than semantic units. This framework suggests these 'words' function as visual pulses rather than lexical entries.
2. The Gate–Scaffolding Architecture
This analysis identifies a rigid structural law governing the text, here defined as the Gate–Scaffolding Architecture. This model provides a functional account of the position-dependent entropy observed in recent computational studies.
2.1 The Gate (Contextual Calibration)
The Gate is defined as the initial sequence of high-entropy glyphs, often characterized by gallows or distinct morphology, persisting until the first instance of high-frequency repetition (Scaffolding onset).
Statistical Evidence: Lindemann (2020) demonstrated that Voynich “vocabulary” is strongly constrained by line position, with certain morphological clusters (e.g., qo-, y-) appearing almost exclusively at line-initial positions.
Functional Role: The Gate exhibits the highest local entropy relative to the remainder of the line. Its role can be modeled as contextual calibration, signaling the structural frequency profile of the line to the visual system. With repeated exposure, decoding effort diminishes, shifting processing from explicit interpretation toward implicit visual tracking.
This model interprets the Currier A/B distinction as differences in Calibration Tuning within the Gate–Scaffolding Architecture. While both Hands operate within the Goldilocks Zone of entropy, variations in line-initial Gate patterns and Scaffolding complexity reflect alternative implementations of the same attentional flow mechanism, rather than a fundamental change in functional design.
2.2The Script Scaffolding (Foveal Anchor)
The Script Scaffolding comprises the text following the Gate through to the line terminus.
Predictive Decay: Following the Gate, conditional entropy decreases sharply. Schinner (2007) showed that character transitions become highly redundant, producing a constrained “random walk” with minimal informational gain.
Redundancy as Mechanism: This high predictability is critical for Attentional Flow (AF), defined here as sustained, low-interruption foveal scanning under reduced semantic load. By reducing decoding demands, the Scaffolding suppresses interpretive effort and promotes continuous visual scanning.
Vertical Anchoring: Line-final tokens (“Terminals”) frequently display identical or highly similar suffixes (e.g., -m, -g). These are interpreted not as grammatical markers, but as vertical anchors consistent with a role in stabilizing return sweeps during line transitions.
The rigid adherence to line boundaries (lack of hyphenation and line-spanning words) is often cited as a linguistic mystery. Under the Lens Protocol, the line is the fundamental unit of the Foveal Anchor. The line break is not a grammatical pause, but a mechanical reset for the eye's scanning sweep.
2.3Spatial Locking (Peripheral Tethering)
Paragraphs in the VMS are often arranged in stable, non-random spatial relationships with adjacent illustrations. Under the Lens Protocol, the illustration is hypothesized to function as a global peripheral anchor, providing a constant, non-linguistic stimulus to the parafoveal field. While the fovea executes the local Gate–Scaffolding scan of the text line, the parafoveal system remains anchored to consistent spatial coordinates associated with the diagram. This dual-channel engagement is proposed to reduce spatial drift and mitigate perceptual habituation across multi-line sequences. Under this model, paragraph boundaries do not function as semantic divisions, but as synchronization points that re-lock foveal scanning to a stable peripheral reference, maintaining attentional stability over extended viewing.
Testable Predictions:
Line-initial entropy should peak at paragraph onsets, reflecting re-calibration of the foveal anchor following spatial re-locking.
Conditional character entropy should decrease in regions where text is most tightly constrained by illustration geometry, consistent with increased reliance on predictive fluency under elevated spatial remapping load.
Paragraphs lacking adjacent illustration anchoring should exhibit weaker or absent entropy modulation relative to text blocks embedded within dense diagrammatic contexts.
Failure to observe these effects would weaken the Spatial Locking sub model without, by itself, falsifying the broader Gate–Scaffolding architecture.
3.Proposed Boustrophedon Scanning and Spatial Remapping Load
The alignment of line-initial Gates and line-final Terminals is compatible with a hypothesized boustrophedon scanning pattern, in which a reader’s eye alternates direction line by line. This is not a claim about the scribal hand’s writing order, but rather a proposed mechanism for how visual attention could engage with the text, increasing spatial remapping demands during reading.
3.1 Bi-Directional Entropy Smoothing
While Scaffolding text remains highly predictable in either direction, the relative position of Gates and Terminals changes with scan direction.
Mechanism: A strictly left-to-right scan of highly redundant text would be expected to lead to rapid habituation. Alternating scan direction (left-to-right, then right-to-left) requires repeated spatial remapping of the relationship between Gates and Terminals.
Cognitive Effect: This alternating directional scanning imposes greater spatial remapping demands than unidirectional reading, sustaining active visual engagement despite low informational content.
This hypothesized scanning pattern is a reader-imposed cognitive strategy rather than a claim about the manuscript’s writing direction or paleographic execution.
4.Visual–Spatial Dynamics: The Diagram as Peripheral Anchor and Controlled Interference
A core component of the Lens Protocol is the interaction between foveal (central) and parafoveal (peripheral) vision. Under this model, the illustrations are not treated as objects of semantic study, but as interference stimuli to be managed. Within this framework, the diagram simultaneously functions as a stabilizing peripheral reference and a source of controlled interference.
4.1 Parafoveal Interference and Failure Modes
Reading Condition: Attentional stability depends on maintaining the boustrophedon sweep of the text (tactical anchor) while retaining the diagram within the parafoveal field.
Failure Mode: If foveal attention shifts toward the diagram (for example, attempting detailed identification), attentional stability degrades. The diagram functions as visual noise.
Proposed Effect: Sustained scanning in the presence of high-contrast distractors engages inhibitory control mechanisms associated with prolonged attentional regulation.
4.2 Bleed-Through as Parafoveal Priming
Without asserting authorial intent, ink bleed-through may produce secondary perceptual effects.
Observation: Many folios exhibit substantial bleed-through from the reverse side of the parchment.
Hypothesis: Incidental faint reversed forms may serve as parafoveal priming, allowing the visual system to pre-map spatial boundaries of the subsequent page.
Mental Rotation: Processing reversed shapes requires unconscious mental rotation, further engaging visual–spatial resources and limiting the emergence of internally generated narrative processing.
5.Discussion: The Goldilocks Zone of Predictability
The functional viability of the system depends on operating within a narrow entropy range, termed the Cognitive Goldilocks Zone.
Excessive Predictability (h₂ < 1.0): Habituation and disengagement dominate.
Excessive Complexity (h₂ > 3.0): Narrative interpretation and problem-solving are triggered, disrupting flow.
Observed Voynich Range (h₂ ≈ 2.1): The script occupies a range that demands continuous saccadic engagement while denying semantic resolution, sustaining task-oriented attentional states.
The high level of 'self-citation' (repetition of similar words in close proximity) is often viewed as anomalous under the hoax hypothesis. However, in a bio-feedback model, this repetition is a feature: it facilitates perceptual fluency. It allows the foveal anchor to move with minimal cognitive friction, maintaining the reader in a continuous attentional flow without start-stop disruption.
6. Conclusion
The Voynich Manuscript constitutes a highly constrained visual–textual system whose properties are compatible with effects on attentional networks and cognitive processing. Through its Gate–Scaffolding text architecture and the presence of parafoveal interference diagrams, the manuscript strongly biases reading behavior toward a non-narrative attentional flow state under sustained engagement. Future work should focus on controlled behavioral and neurophysiological studies—such as eye-tracking and EEG coherence measurements during boustrophedon scanning—to empirically test these compatibility claims.
Falsifiability Conditions:
This model would be substantially weakened or falsified under any of the following conditions:
Demonstration that comparable levels of position-dependent entropy and low conditional entropy arise naturally in known human languages with similar positional rigidity.
Evidence that the proposed Gate–Scaffolding structure does not significantly contribute to predictive redundancy when evaluated against suitable control corpora.
Experimental or computational results showing no measurable difference between Voynich text and shuffled or synthetically generated corpora under equivalent visual-scanning conditions.
Evidence that paragraph boundaries and illustration adjacency show no statistically systematic relationship to entropy modulation, positional variation, or scanning load, indicating that layout features are epiphenomenal rather than functional.
Identification of prior models that already explain these statistical properties more parsimoniously.
I am particularly interested in negative results, counter-models, and empirical objections. If this framework fails, understanding why it fails is more valuable than defending it. This model acknowledges the Currier Language split and the Gordon Rugg Hoax Hypothesis, but proposes a third path: that the manuscript is a functional artifact whose 'meaning' is found in its physiological and attentional effects on the observer, not in a hidden plaintext.
this is not a claim, but a question about a work in progress.
Is this a legit strategy to find out something about the text or if it is just not helpful at all?
Algo:
- cut out random 10 pages off the full text (voynich)
- train a base llm with the text (10 pages missing).
- give one of the cut out pages to the llm with one EVA-Token missing. the llm guesses a token for the missing one.
- repeat this 20 times with different tokens missing.
- do the same with a random text
- do the same with a generated text (from one of supposed solutions for example).
- compare the probability.
The result will show if the llm can predict the real text better than the other texts.
The problem might be, that the result doesn't say anything about the text but only about the model. But it might say that the text is not random, for example.
I have made the llm and tested it, but not scientifically. with mixed results. I used chatgpt mostly for python code generation and getting a new perspective. Most of what it wrote were guesses that I dismissed.
I got significant better results for predicting the real cut out text against a random text from EVA-Vocab. I used Qwen/Qwen3-1.7B-Base as llm.
I had against random something like this:
Δ avg log P/subtoken (REAL − RANDOM): 4.1635
However this is of course not scientific, so take it with a grain of salt. Especially because I am new to this and do this as a hobby holiday project.
A variation is: Can the llm predict the token from both sides better than only the left side? You just have to hide all right tokens for the left side test.
It might however be possible to make this scientific. What do you think?
Happy Christmas everyone. After the flood of LLM slop theories this year, I really wanted my own one. Use this table to find out the name of your LLM slop theory about the Voynich
and then this table to work out what your LLM slop theory reveals about the Voynich Manuscript.
Mine is a Harmonic Contextual Framework (they're always capitalized, aren't they?) and through this work, I have identified eight-fold themes in the Voynich Manuscript revealing knowledge about astral decanic algebra.
What's yours?
(I would like to say I made up all of the possibilities but many are taken either from here or the Facebook Group)
![[Image: schema13.gif]](https://voynich.nu/q13/schema13.gif)
