Run · Voynich Manuscript
T40 exploratory: Herbal A decoded with learned sign boundaries, against its own words shuffled
The same kind of gap was found for the session-8 mixed key (0.18-0.27 nats per element), and content-free generators gave 0.00-0.02 there. Shuffling removes the edge dependency and line layer as well as any content, so the gap can come from structure without meaning. The sharper null for next time is the fitted generator run on shuffled Latin words, which keeps the cipher's structure and removes the plaintext's order.
Result
| Run | Measured | Result | Against | Decision | Date |
|---|---|---|---|---|---|
| Run | Held-out element-trigram score (nats per element; Latin's own -2.418) / dictionary coverage of the decoded held-out text | Herbal A -2.563 / 0.296; shuffled -2.824 / 0.216 | Planted control under the same rung-B pipeline: -2.767 / 0.271 (true plaintext 0.697) | Exploratory, not a reading: the real word order beats its shuffle by 0.26 nats per element, but the decoded text is not Latin ('iumu ae s io s e i nea n un a li d uee u uei i u n tu e n ti oru u') and the pipeline fails its control at this rung |
What it used and found
Latin genre texts
DATDataset
ZL3b transcription
DATDataset
Latin as a re-spaced homophonic stream (period practices)
ANSAnswer it tests
Parameters
null = the section's words shuffled across its pages (seed 5), setting = best Herbal A setting (as rung A), iterations = 300000, segmentation = {'V': 200, 'rounds': 3, 'max_len': 3}Cite it
Mysteries Department. (2026). T40 exploratory: Herbal A decoded with learned sign boundaries, against its own words shuffled [Run RUN0DCBGWKQ4KGM]. https://mysteriesdept.com/r/RUN0DCBGWKQ4KGM
@misc{mysteriesdept_run0dcbgwkq4kgm,
title = {T40 exploratory: Herbal A decoded with learned sign boundaries, against its own words shuffled},
author = {{Mysteries Department}},
year = {2026},
note = {Run, catalog key RUN0DCBGWKQ4KGM},
url = {https://mysteriesdept.com/r/RUN0DCBGWKQ4KGM}
}