Sometime in the early 15th century, someone wrote a book. They filled it with 240 pages of flowing, confident text — a script that has never been identified, in a language that has never been decoded, accompanied by illustrations of plants that don't exist, astronomical diagrams that don't match any known system, and naked figures bathing in strange green liquid.
We don't know who wrote it. We don't know what language it's in. We don't know what it means. We don't even know if it means anything at all.
The Voynich Manuscript is the most studied, most debated, and most stubbornly unreadable document in the world. Cryptographers who cracked Nazi codes couldn't break it. AI systems trained on every known language have failed to decode it. The world's top linguists have spent careers on it and walked away empty-handed.
It has been sitting in a Yale library vault since 1969, still waiting for someone to read it.
What we actually know
The manuscript was carbon-dated in 2009. The vellum — the material the pages are made from — dates to between 1404 and 1438, placing its creation in the early 15th century, almost certainly in northern Italy or central Europe.
The ink has been tested and matches the period. This rules out one of the most persistent theories: that the manuscript is a modern hoax. Whatever it is, it is genuinely old.
It first surfaced in recorded history in 1639, when it was mentioned in a letter written by a Prague alchemist named Georg Baresch. He described it as a "Sphinx" — something he had spent years trying to understand without success. He believed it contained secret alchemical knowledge written in a deliberately obscure code.
The manuscript was eventually purchased in 1912 by a Polish-American book dealer named Wilfrid Voynich, from whom it gets its name. He spent the rest of his life trying to decode it. He failed. After passing through several owners, it was donated to Yale University's Beinecke Rare Book and Manuscript Library, where it remains today as MS 408.
What's inside
The manuscript is divided into sections that researchers have labeled by apparent subject matter — though since no one can read it, these labels are educated guesses based on the illustrations:
| Section | Content | Number of pages |
|---|---|---|
| Botanical | Illustrations of plants, most of which don't match any known species | ~130 pages |
| Astronomical | Circular diagrams resembling zodiac charts, sun and moon illustrations | ~20 pages |
| Biological | Small naked figures — mostly female — in pools of green liquid connected by pipe-like structures | ~20 pages |
| Cosmological | Large folding pages with complex circular diagrams, possibly maps | ~11 pages |
| Pharmaceutical | Illustrations of plant parts alongside containers, possibly a recipe section | ~34 pages |
| Recipes | Dense text with small star-like illustrations in the margins | ~23 pages |
The text itself has clear structural patterns. It flows left to right. It has consistent word lengths. Certain "words" appear with extremely high frequency, while others appear only once. These patterns are consistent with a natural language — or with a very sophisticated constructed one. The text does not look like random gibberish. Something is encoded in it. We just don't know what.
The theories — and why each one fails
Over the past century, dozens of serious proposed solutions have been published in peer-reviewed journals and presented at academic conferences. Every one of them has been challenged, disputed, or quietly abandoned.
| Theory | Core argument | Why it fails |
|---|---|---|
| Unknown natural language | The manuscript is written in a real but obscure language using a unique script | No known language family matches the statistical patterns of the text |
| Constructed language | Someone invented an entirely new language and writing system | Possible, but explains nothing about the content or purpose |
| Steganographic cipher | The text encodes a hidden message using a complex cipher system | Every proposed cipher has failed to produce coherent output when applied consistently |
| Glossolalia / meaningless text | The text is deliberately meaningless — a work of art or fraud | Statistical analysis shows patterns inconsistent with pure random generation |
| Hoax by the original author | The manuscript was created to deceive a wealthy buyer into paying for "secret knowledge" | The sheer volume and internal consistency makes a pure hoax extremely labor-intensive to sustain |
The most serious modern attempts
In 2019, a researcher named Gerard Cheshire published a paper claiming he had decoded the manuscript, identifying it as written in "proto-Romance" — a precursor to modern Romance languages. His solution was widely covered in the press. It was also swiftly dismantled by linguists, who called the methodology flawed and the translation unsupported. The journal that published it later issued a statement distancing itself from the conclusions.
In 2017, a team of computer scientists applied AI-based pattern recognition to the text, comparing it against 380 known languages. The closest match was Hebrew — but even this match was weak, and attempts to translate the manuscript using Hebrew produced nothing coherent.
A separate 2013 study by Marcelo Montemurro and Daoud Zanette applied information theory — the same mathematics used to analyze communication systems — and found that the text contains meaningful structure. It is not random. Something is being communicated. What that something is remains unknown.
The curious connection
The Voynich Manuscript raises a question that goes beyond medieval history: what happens when human knowledge becomes permanently inaccessible?
We tend to assume that given enough time, enough computing power, enough intelligence, any code can be broken. The Voynich Manuscript suggests that this assumption may be wrong. If the key to the cipher was never written down, if the person who created it died without passing it on, the information may be irretrievably lost — regardless of how sophisticated our tools become.
This is not an abstract problem. Linguists working with undocumented dying languages face the same crisis right now. There are estimated to be around 7,000 languages currently spoken on Earth. Roughly half are expected to go extinct by 2100. When a language dies with its last speaker, the knowledge encoded in that language — the specific words for concepts that other languages don't have, the oral histories, the ways of categorizing the world — disappears with it. Not encrypted. Gone.
The Voynich Manuscript, sitting in its climate-controlled vault at Yale, is a 600-year-old reminder of how fragile the transmission of knowledge really is.
FAQ
Has the Voynich Manuscript ever been decoded?
No. Despite numerous claims over the decades, no proposed solution has been accepted by the academic community. Every attempted decoding has either been shown to be methodologically flawed or has failed to produce consistent, coherent results when applied to the full text.
Where is the Voynich Manuscript now?
It is held at the Beinecke Rare Book and Manuscript Library at Yale University, catalogued as MS 408. A full digital scan is freely available online through Yale's digital collections.
How old is the Voynich Manuscript?
Carbon dating of the vellum places its creation between 1404 and 1438, making it over 600 years old. The ink has also been tested and is consistent with this date range.
Could the Voynich Manuscript be a hoax?
It is possible, but the manuscript's statistical text patterns are inconsistent with randomly generated nonsense. The internal structure of the text behaves like a natural or constructed language, which makes a pure hoax difficult to sustain as an explanation.
Can I see the Voynich Manuscript online?
Yes. Yale's Beinecke Library has digitized the entire manuscript and made it freely available. You can view every page at beinecke.library.yale.edu.