The Voynich Manuscript Mystery Is Solved… But the Translation Is Unexpected

Ajouté : 2026-06-02

[00:00:00]Wilfried Voyage who discovered the manuscript in 1912.

[00:00:04]>> On the campus of Yale University sends the Bicki Rare book and manuscript library. This cryptic book is referred to as the Voinich manuscript.

[00:00:16]>> The Voinich manuscript has been sitting in a glass case at Yale University for over a century. 240 pages. An alphabet no one recognized. A language no one could place. 600 years of silence. In early 2025, a team of AI researchers cracked it. And the first thing they did was go quiet, not celebrate, not publish. Quiet. What they found inside those worn medieval pages did not fit any category of knowledge that exists in any academic tradition alive today. The institutions holding the full translation are still deciding how much of it the world is allowed to see. That pause from people who built careers on publishing tells you everything. Stay with us. We reveal what they found. The man who paid a king's ransom. Before the AI, before the codereers, before the academic papers and the press silence, there was a book dealer named Wilfred Voyage standing in a dusty storage room at a Jesuit college in Frascati, Italy in 1912.

[00:01:13]>> The book dealer, Wilfred Voyage, in the library of a Jesuit order in Italy.

[00:01:19]>> He was used to finding old things, forgotten manuscripts, overlooked texts.

[00:01:24]That was his business. He had spent decades moving through European libraries and private collections, pulling overlooked volumes from neglected shelves, recognizing value where other people saw clutter. He was good at it, patient, trained by years of handling history with his bare hands.

[00:01:42]But when Wilfred Voyage pulled this particular volume from the shelf, something stopped him. The pages were vellum animal skin, not paper. That alone told him something. Vellum was expensive in the 15th century. The Voyage manuscript is a book written by hand that is reliably dated to the early part of the 15th century.

[00:02:03]>> Deliberately chosen. The kind of material you selected when you needed words to survive, not just decades, but centuries. Someone had invested real money in making sure whatever was written here would last. And covering every inch of those pages was handwriting in an alphabet he had never seen in his life. not Latin, not Greek, not any shorthand system or coded notation he recognized from decades of handling rare manuscripts. The characters were consistent, deliberate, written with the steady hand of someone completely fluent in whatever this was.

[00:02:36]This was not someone experimenting. This was someone writing in a language they knew deeply and used naturally. Voy Nick bought it without hesitation. He spent the next 8 years until his death in 1930 trying to understand what he had found.

[00:02:51]He showed it to cryptographers. He brought in linguists. He contacted scholars across Europe. Nobody could give him an answer. The book that should have been the crowning discovery of his career became the obsession that followed him to his grave unsolved.

[00:03:07]Carbon dating would later confirm what the vellum had already suggested. The manuscript was not vaguely old. It was specifically old, created somewhere between 1404 and 1438. That places its composition in the early Italian Renaissance, a period when Europe was clawing itself back from the devastation of the black death. When onethird of the continent's population had been erased within a generation, when the survivors were rebuilding not just cities, but entire frameworks for understanding the world. It was a strange and electric moment in history. Scientists and mystics operated in the same rooms.

[00:03:44]Physicians studied astrology alongside anatomy. The boundary between medicine, religion, and what we would now call occult knowledge was not clear. It was a blurred, contested, constantly shifting space where serious scholars worked without embarrassment. Someone inside that world during that precise window sat down with ironrich ink and expensive prepared vellum and wrote 240 pages of something then encrypted it so completely that it would defeat every expert who touched it for the next six centuries. That choice to write and then to hide is the question the manuscript never stops asking. And here is the detail that reframes everything. Wilfred Voyage was not the first serious person to own this book. Not even close. Before it sat forgotten in that Jesuit storage room before Voyick's hands ever touched it, the manuscript passed through one of the most powerful courts in European history.

[00:04:39]>> The Western Roman Empire, Europe divided, and communication beyond its borders. Holy Roman Emperor Rudolph II, ruler of an empire stretching across central Europe, collector of the strange and the rare, patron of astronomers and alchemists and artists, had this book in his possession in the late 16th century.

[00:04:59]Rudolph was not a casual collector. He assembled one of the greatest cabinets of curiosities in the world. Taiko Brahe worked under his patronage. Johannes Kepler served as his imperial mathematician. Rudolph spent his reign surrounding himself with people who pushed the edges of what was knowable.

[00:05:16]And for this manuscript, he reportedly paid 600 gold ducats. That number needs context. 600 gold duckats in the late 1500s was not the price of a rare book.

[00:05:27]It was the price of a property, an estate, the kind of sum that represented years of income for a skilled craftsman.

[00:05:34]You do not pay that for a curiosity. You pay that when you believe what you are acquiring contains knowledge that exists nowhere else on Earth and cannot be reconstructed if lost. Rudolfph II with access to the finest scientific minds of his era with resources no private collector could match never cracked it either. The book took his money, gave him nothing back and moved on. The pictures nobody could explain. Most people who hear about the Voinish manuscript assume the text is the strange part. They have not looked at the pictures. The illustrations cover roughly half the manuscript's 240 pages.

[00:06:10]They are not decorative. They are not filler. Every drawings appears to be doing deliberate, specific work, communicating something the surrounding text was describing in detail. Whoever created this book was not just a writer.

[00:06:24]They were a trained illustrator working with clear technical purpose. The lines are confident. The compositions are structured. This is not the sketch work of someone experimenting. This is documentation which makes what the pictures actually show considerably more unsettling. The botanical sections contain dozens of plant illustrations rendered with the kind of precise technical accuracy common to 15th century medical herbals and pharmacopyas. These were serious reference documents. About half of it is pictures, drawings that take up the entire page of a plant, an herb or a flower used by physicians and apothecaries to identify plants, prepare medicines, and treat patients. The illustrators who produced them were trained professionals. Their drawings were meant to be accurate enough that a practitioner in a different city could look at the image and recognize the plan in the field. By that standard, the Voinich botanical illustrations are extraordinary. The draftsmanship is confident. The detail is careful. The technical execution matches the quality of any legitimate herbal from the same period. But not one of those plants exists. Not in any botanical record compiled across six centuries of systematic global research. Not as a known species, not as a regional variant of a known species, not as an extinct ancestor of anything currently documented. Professional botonists, people who have spent careers cataloging plant life across every continent, have examined these drawings and come back with nothing. No matches, no near matches, no educated guesses that hold up under scrutiny. Some of the illustrations appear to be hybrids. The root system of one identifiable plant fused with the leaf structure of a second, topped with the flowerhead of a third, assembled into a single organism that follows no evolutionary logic recorded anywhere on Earth. Not a mistake, not artistic license. The combinations are too specific and too consistent for that. It reads like someone was documenting plants that existed somewhere in a location, a time, or a layer of reality that no existing catalog has ever reached. The astronomical sections offer no more comfort. Circular diagrams fill these pages. Careful, deliberate illustrations of star formations and celestial relationships produced with the precision of a trained scholar working from direct observation or serious reference material. Medieval astronomical illustration was a sophisticated tradition.

[00:08:56]>> Nine medieval manuscript books from a rare book firm called Laser Luminere.

[00:09:02]They started an initiative called manuscripts in the curriculum.

[00:09:06]>> Islamic scholars had mapped the sky with extraordinary accuracy. Greek astronomical knowledge had been recovered and was actively circulating through Renaissance intellectual networks. The visual language of serious astronomical documentation was well established. The Voyic astronomical diagrams follow that visual language perfectly in form and then place every star in the wrong position. Researchers have spent decades attempting to align these diagrams with known star charts, medieval Europeans, Islamic, recovered Greek traditions from every culture that produced serious astronomical work before 1438. Nothing aligns. The star formations do not correspond to any known constellation tradition from any era. The relationships between the celestial bodies shown do not match any sky that has ever been documented by any civilization on record. The stars are simply wrong. Consistently, precisely, deliberately wrong. And then there are the bathing scenes. Large sections of the manuscript are devoted entirely to illustrations of women in elaborate pools and bathing structures connected by intricate systems of pipes, channels, and tubes that feed into and out of each other in complex layered patterns. The infrastructure shown is not simple. It is engineered. A system with inputs, outputs, and deliberate flow logic. The women inside it are calm, purposeful.

[00:10:30]Their expressions and postures carry none of the surprise or discomfort. You would expect from someone encountering something strange. They are completely at ease with whatever is happening around them. They have done this before.

[00:10:43]This is familiar to them. Medieval bathing for medical purposes was well documented and entirely unremarkable.

[00:10:50]Physicians prescribed therapeutic baths routinely. Illustrations of medicinal bathing appear in dozens of legitimate manuscripts from the same period.

[00:10:59]Ordinary, practical, carrying no particular weight. This is not that. The pools are too elaborate. The pipe systems are too specific. The sequential nature of the illustrations. The way they appear to show stages of a process rather than a single static activity points towards something with a precise intended outcome. Not hygiene, not therapy in any sense that medieval medicine recorded. a procedure, a sequence documented with the same careful technical purpose as the botanical and astronomical sections surrounding it. When the AI began translating the text directly alongside these illustrations in 2025, the research team stopped responding to press inquiries within days. They did not issue a statement explaining the silence. They simply went quiet. That detail is not a footnote. It is the most important data point in this entire story. These were academics, people professionally trained and institutionally rewarded for publishing, for communicating findings, for moving knowledge into the public record as quickly as possible. Going quiet is not their default setting. Something in those pages change that calculation entirely. Remember that when the translation arrives, the machine that say Hebrew for 600 years, the failure was consistent. World War II codereakers who had dismantled the Nazi enigma machine turned their full attention to the Voinich manuscript and came back with nothing. Linguists who could reconstruct a dead language from 50 surviving words could not find a single foothold in the text. Cryptographers who had broken every known encryption system ever built by human intelligence ran out of ideas inside a year. The book simply did not respond. That consistency of failure is worth sitting with for a moment. This was not a case of the wrong people trying. The 20th century produced the most sophisticated coderebreaking operation in human history. A coordinated, government-funded, mathematically rigorous effort that cracked German military encryption during World War II and shortened the war by years. The people behind that work were not amateurs. They were the best analytical minds of their generation, working at the absolute edge of what human intelligence could do with pattern recognition and cipher theory.

[00:13:17]They looked at the Voyic manuscript.

[00:13:19]They looked carefully. They left with nothing. After them came computational linguists armed with statistical tools that could extract grammatical structure from a language with fewer than 100 surviving words. Then cryptographers with access to early computing power.

[00:13:36]than teams with increasingly sophisticated algorithms. Each generation brought better tools than the last. Each generation failed in the same place at the very first step, the one that should have been simplest, identifying what language or language family or encryption tradition the text belonged to. Every attempt started with a guess. Pick a candidate, test for matches, adjust, repeat. The problem was the search space. Human researchers could realistically test dozens of candidate languages in a career, maybe hundreds across an entire field over decades. The Voinich text gave nothing away. No borrowed words, no recognizable grammatical markers, no structural clues that pointed toward any known linguistic tradition. Without a starting point, every guess was equally likely and equally wrong. The Voinage manuscript must be as well. At the same time, it exhibits so many peculiarities that don't confirm that theory.

[00:14:30]>> What changed in 2025 was not effort.

[00:14:34]Researchers had never lacked effort.

[00:14:36]What changed was the approach entirely.

[00:14:38]At the University of Alberta in Edmonton, Canada, researchers Greg Condrach and Bradley Hower built something different. Their AI system did not pick a candidate language and test it. It did not guess and check. It fed simultaneously on samples from over 380 languages at once. The entire documented breadth of human linguistic history searched in a single analytical pass.

[00:15:00]Every language family, every known grammatical tradition, every statistical fingerprint of every communication system humanity had ever recorded and preserved. The machine was not looking for a match. It was asking a simpler, broader question. Does this text feel statistically, structurally, rhythmically like anything in the entire history of human language? It came back with one answer, Hebrew, not modern Hebrew. Ancient Hebrew, and not straightforwardly encoded, either. The AI identified a second encryption layer built directly on top of the language itself. Every single word in the manuscript had its letters rearranged alphabetically before being written down, a technique researchers call alphags. Think of it as a systematic anagram, not random scrambling, but a strict consistent rule applied uniformly across all 240 pages. The word on the page is not the word. It is the word's letters sorted into alphabetical order written down in that sorted form, then transcribed into a constructed alphabet that exists nowhere else on Earth. Two layers, a built from scratch writing system concealing a scrambled ancient language. The outer layer defeated pattern recognition for six centuries because it had no ancestral connection to any known script.

[00:16:16]>> The Voinich manuscript thought to have been written some 600 years ago. Its author is unknown and the writing remains undeciphered.

[00:16:25]>> The inner layer defeated linguistic analysis because every word was prescrambled before encoding. Either layer alone would have been a serious obstacle. Together they created something that held for 600 years the against every tool humanity could bring to bear. Once researchers applied the full decoding logic, reversing the alphagram scrambling, mapping the constructed characters back to Hebrew letters, over 80% of the resulting words appeared in standard Hebrew dictionaries. 80%. In a text that had defeated every expert since the Italian Renaissance, the first full sentence the AI translated read, "She made recommendations to the priest, man of the house and me and people. Simple, domestic, almost shockingly ordinary, the kind of line that could belong to a household account or a personal letter."

[00:17:14]Not the revelation anyone had been bracing for after six centuries of mystery. But that sentence was a door and once it opened, everything behind it began to move. What was inside the bathing scenes? Of that first sentence.

[00:17:27]The AI cascaded outward. Grammatical patterns locked in. Botanical references in the plant sections began resolving into readable text. The astronomical notation decoded. And then the system reached the sections surrounding the bathing women. The illustrations that had unsettled researchers for a century without anyone being able to say exactly why. Now they could say why. The decoded text alongside those illustrations did not describe medicinal bathing. It did not describe hygiene. What the ancient Hebrew passages described once the alphagram scrambling was reversed and the constructed alphabet was mapped was ritual pharmacology. Precise formulations, plant-derived compounds, sequential procedures administered in a specific order for a specific intended outcome documented with the precision of someone who had performed these procedures personally. The botanical sections confirmed it. Plant mixtures decoded into formulas that combined known medieval herbal remedies with compounds that have no counterpart in any documented medical tradition.

[00:18:31]Researchers who had spent entire careers studying medieval medicine looked at the decoded formulas and could not place them. Not as known's remedies, not as extinct variants, not as anything with a name in any living academic tradition.

[00:18:45]And then there were the ritual passages.

[00:18:48]Among the decoded text were references to angelic hierarchies, sequential incantations and operational instructions that read less like medieval theology and more like a procedural manual, not the comfortable metaphorical spiritual language common in medieval religious manuscripts where difficult things soften into allegory and symbol. These passages were specific, step by step, written by someone describing procedures they had carried out. One researcher speaking privately described the decoded material as knowledge that feels like it should not exist. Here is the part that made the press briefings stop entirely. The ritual sections were not scattered randomly through the manuscript. They were positioned deliberately alongside the botanical formulas and the astronomical charts as if all three were meant to be used together, not separately, as a unified system. a system for doing something the author clearly understood and clearly believed should never reach the wrong hands. The ink that never forgets. While the AI was working on the language, a completely separate team was doing something different entirely. They were shining light on the pages that human eyes cannot see. Multisspectral imaging uses ultraviolet and infrared wavelengths to reveal what is hidden beneath centuries of fading and deliberate eraser.

[00:20:08]Medieval ink contains iron. Iron bonds with vellum permanently at the molecular level during the drying process. Even if ink fades completely over centuries, even if someone deliberately scrapes it away, the ghost of every letter remains embedded in the parchment. You cannot truly erase medieval iron ink. The parchment holds the memory of it permanently, visible only under the right wavelengths of light. Yale's Bayi Rare Book and Manuscript Library had allowed a team from the Lazarus Project to scan selected pages back in 2014.

[00:20:41]Those scans sat unprocessed on servers for nearly a decade. No one was prioritizing them. When researcher Roger Eastston of the Rochester Institute of Technology reprocessed them in 2024 using updated image enhancement tools, he passed the results to scholar Lisa Fagan Davis, a paleographer and the executive director of the Medieval Academy of America. Davis was at her workstation late one evening running through the enhanced scans page by page.

[00:21:08]She reached the first page and stopped.

[00:21:10]I'm moving. In the lower corner, invisible under normal light, invisible to every researcher who had examined that page for over a century, were three slim columns of faint writing. One column showed the Roman alphabet written out completely A to Z. Another displayed Voinich characters alongside corresponding Roman letters, a direct mapping of the two writing systems. The third repeated the Roman alphabet shifted by exactly one letter position, the classic structure of a substitution cipher key. She recognized the handwriting immediately. It belonged to Johannes Marcus Marchie, a 17th century Czech physician, scientist, and imperial diplomat who had owned the manuscript between 1662 and 1665 before sending it to the Jesuit scholar Athanasius Kercher in Rome. Marchie had been working on the cipher himself, alone, in secret, scribbling his attempts in the corner of the first page where no one would think to look. He never succeeded. But what he left behind confirmed something skeptics had questioned for years. This manuscript was not a modern forgery. Its chain of custody ran unbroken and verifiable from at least the early 17th century backward. And the question his hidden notes raised nobody has fully answered is not whether he was trying to solve it in secret. It is whether he was hiding the fact that he was getting close. The Voyic manuscript is not fully decoded. The AI opened the door. It has not walked through it. What has been translated so far, the first sentence, the botanical formulas, the early ritual passages, the decoded astronomical notation represents a fraction of 240 pages. The institutions holding the full translation are not rushing. The Bicki Alberta, the University of Chicago, they are reviewing, cross-checking, and in some cases quietly setting material aside while they decide what responsible disclosure actually looks like when what is being disclosed defies every academic category that currently exists. That pause from people who built careers on publishing is the most telling detail of this entire story. Someone six centuries ago wrote something so dangerous they invented an alphabet, wrote in a second language, and scrambled every single word before committing it to Vellum.

[00:23:24]They were right for 600 years. They were not counting on what came next. The translations are still coming, and we will be here when they do.