+- The Voynich Ninja (https://www.voynich.ninja)
+-- Forum: Voynich Research (https://www.voynich.ninja/forum-27.html)
+--- Forum: Theories & Solutions (https://www.voynich.ninja/forum-58.html)
+---- Forum: The Slop Bucket (https://www.voynich.ninja/forum-59.html)
+---- Thread: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript (/thread-5563.html)
Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - gimmy328 - 11-04-2026
I am not claiming to have deciphered the Voynich manuscript, but I want to share an observation I made while analyzing the botanical illustrations. This is not an attempt to “solve” the manuscript - just an interesting pattern I noticed and decided to check more carefully.
I selected about 50 plants from the botanical section of the manuscript whose roots share a similar visual structure: branching, subdividing, and repeating graphical elements. The remaining plants were excluded because their roots differ fundamentally - some have thorns, some have tubers, some have bulb‑like bases, and some are not true branching roots at all.
So from roughly 110 plants, only 49 formed a clean, comparable dataset.
Within this group, the roots fall into three structural types: straight, intertwined, and what I call the “X‑format”
The X‑format consists of roots that contain crossings and repeated graphic motifs such as xx|x|xx|x|х... , ||x , |xx|x|x|x|x|x|x|x|x and other long, patterned sequences. Only six plants in the entire set show this X‑format, but these six display the strongest and most consistent internal pattern.
When I counted how often a plant has a flower depending on its root type, the results were:
straight roots - a flower appears in roughly 62% of cases
intertwined roots - a flower appears in roughly 68% of cases
X‑format roots - a flower appears in 100% of cases, without a single exception
And this is only the beginning.
All six X‑format plants not only have flowers - the flowers are always colored, almost always blue, and their shapes are strikingly similar (cup‑shaped or petal‑shaped). These pages also cluster together in the manuscript rather than appearing randomly. Straight and intertwined roots do not show this level of consistency: flowers may or may not appear, colors vary unpredictably, and shapes differ widely.
To test whether this pattern was real, I ran a small experiment: I used a structural‑recognition tool to see whether the characteristics of the flower could be predicted from the root alone. Based solely on the root type, the tool correctly predicted that the plant has a flower, that the flower is cup‑shaped, and that it is blue. This suggests that the connection between root type and flower characteristics is strong enough to be detected independently.
Further analysis revealed that the same pattern appears in the text.
I counted how many words on each page contain the sequence qot, by which I mean the EVA‑transliterated glyphs corresponding to that pattern (EVA - the Extensible Voynich Alphabet system used to transcribe the manuscript’s script). The distribution of EVA‑qot words mirrors the three root types perfectly:
straight roots - lowest qot counts (typically 1–7)
intertwined roots - medium qot counts (3–11)
X‑format roots - the highest and most stable qot counts (4–13), even on pages with very little text
In other words, the frequency of EVA‑qot words increases together with the structural complexity of the root and with the likelihood of a flower being present.
This creates a three‑level correlation:
root type --- presence of flower --- high density of EVA‑qot words.
Text length and paragraph structure do not explain this pattern: some pages with very little text have extremely high qot density, while long pages with straight roots have almost none. This suggests that qot‑words are not distributed randomly and do not depend on text volume, but instead correlate with the visual class of the plant.
I am not claiming this is the key to the manuscript.
But within this group of 49 structurally comparable plants, the roots behave more like encoded graphical symbols than like botanical drawings. The X‑format appears to be a distinct, highly structured class that predicts the presence of a flower, its shape, its color, and even the textual formula of the page. It may represent part of a visual marking system embedded in the illustrations.
I share this simply as an observation. Perhaps someone will find it interesting or useful for further analysis.
Pages with X‑format roots: 45, 51, 85, 95, 99, 106.
Here are the sources I relied on while developing my hypothesis:
You are not allowed to view links. Register or Login to view.
You are not allowed to view links. Register or Login to view.
RE: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - JoJo_Jost - 11-04-2026
But how is that different from counting the number of flowers, the number of leaves, or the number of stems, and correlating them with the text?
I don't really ur point
RE: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - gimmy328 - 11-04-2026
(11-04-2026, 02:53 PM)JoJo_Jost Wrote: You are not allowed to view links. Register or Login to view.But how is that different from counting the number of flowers, the number of leaves, or the number of stems, and correlating them with the text?Counting flowers or leaves would measure visual quantity, not structural logic.
I don't really ur point
My dataset isn’t about how many elements appear, but how the root geometry itself repeats and crosses - the “X‑format” is a specific structural motif, not a countable feature.
In other words, I’m not correlating decoration density with text; I’m testing whether a distinct morphological pattern (crossed roots) systematically aligns with textual anomalies.
That’s a completely different level of correlation - it’s about form and function, not number and ornament.
So yes, both approaches involve counting, but mine isolates a structural variable - the presence of a repeated geometric logic - rather than a visual abundance like petals or leaves.
TYSM for your question
RE: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - Tanay_alias_CG - 11-04-2026
“If root structure correlates with text patterns, could these structures represent categories like types of ingredients or functions?”
i am a new theorist so take this with a grain of salt but i made a theory which is slightly related to yours but i try to explain the whole picture
my post:
You are not allowed to view links. Register or Login to view.
RE: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - nablator - 11-04-2026
Replacing em-dash with minus does not hide AI slop.
Page 45 = You are not allowed to view links. Register or Login to view. has 0 qot
Page 51 = You are not allowed to view links. Register or Login to view. has 1 qot
RE: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - eggyk - 11-04-2026
(11-04-2026, 03:04 PM)gimmy328 Wrote: You are not allowed to view links. Register or Login to view.Counting flowers or leaves would measure visual quantity, not structural logic.
My dataset isn’t about how many elements appear, but how the root geometry itself repeats and crosses - the “X‑format” is a specific structural motif, not a countable feature.
In other words, I’m not correlating decoration density with text; I’m testing whether a distinct morphological pattern (crossed roots) systematically aligns with textual anomalies.
That’s a completely different level of correlation - it’s about form and function, not number and ornament.
So yes, both approaches involve counting, but mine isolates a structural variable - the presence of a repeated geometric logic - rather than a visual abundance like petals or leaves.
(11-04-2026, 04:37 PM)nablator Wrote: You are not allowed to view links. Register or Login to view.Replacing em-dash with minus does not hide AI slop.
The title was enough honestly.
RE: Root Typology as a Predictor of Floral and Textual Features in the Voynich Manuscript - tavie - 11-04-2026
We prohibit theories that have been assisted/developed by an AI Large Language model like ChatGPT, Claude, Gemini, etc. We move posts like this to The Slop Bucket where they will be automatically locked.