In 1912, a Polish-born rare book dealer named Wilfrid Voynich was picking through chests of manuscripts at the Villa Mondragone, a Jesuit college outside Rome, when he found a small vellum book unlike anything he had ever handled. Its pages were covered in a flowing, confident script in no known alphabet, wrapped around paintings of plants that matched no known species, naked women bathing in networks of green tubs and tubes, elaborate circular diagrams of stars, and rows of what looked like pharmaceutical jars. Voynich bought it, spent the rest of his life trying to have it deciphered, and gave the mystery his name.
More than a century later, the Voynich Manuscript — now catalogued as MS 408 in Yale University's Beinecke Rare Book and Manuscript Library — remains unread. Not for lack of trying: it has defeated medieval scholars, linguists, statisticians, and some of the finest professional codebreakers of the twentieth century, including William Friedman, the legendary American cryptologist whose team broke Japan's PURPLE diplomatic cipher during the Second World War. Friedman studied the manuscript on and off for decades and got nowhere.
What the book actually is
The physical object, at least, has given up its secrets. Radiocarbon dating of the vellum, carried out at the University of Arizona in 2009, placed the animal skins in the early fifteenth century — between about 1404 and 1438. That single result demolished a century of theories, including Voynich's own favourite: that the book was the work of Roger Bacon, the thirteenth-century English friar and proto-scientist. The manuscript runs to about 240 surviving pages, with evidence that some leaves are missing, and its text flows left to right in neat lines with no obvious corrections — around 170,000 characters forming some 35,000 'words' in an alphabet of roughly two dozen core symbols, dubbed 'Voynichese.'
By its illustrations, the book organises itself into sections that look tantalisingly conventional: a large herbal section of plant drawings; astronomical and astrological diagrams with zodiac symbols; a 'balneological' section of bathing women and plumbing; a pharmaceutical section of plant parts and jars; and pages of dense text with starred paragraphs, like recipes. It resembles, in structure, a perfectly ordinary medieval compendium of natural knowledge. Except that the plants are mostly unidentifiable, and the words mean nothing.
The documented paper trail is thin but real. A 1665/66 letter found with the manuscript, from Johannes Marcus Marci, rector of the University of Prague, accompanied the book when Marci sent it to the polymath Jesuit Athanasius Kircher in Rome, begging him to decipher it. Marci passed on a rumour that the Holy Roman Emperor Rudolf II — the famously occult-obsessed Habsburg who kept court in Prague — had once bought the book for 600 ducats. Earlier in the seventeenth century it had belonged to a Prague alchemist named Georg Baresch, who also pestered Kircher about it, apparently without success. From Kircher's Jesuit collections it eventually reached Villa Mondragone, and Voynich's hands.
Language, cipher or hoax?
Every explanation on offer has a serious problem. Is it a cipher — real text, encrypted? Then it has resisted attacks that broke the ciphers of empires; simple substitution was ruled out long ago, and known medieval encryption methods don't produce text with Voynichese's properties. Is it an unknown natural language in an invented script? The text does behave eerily like language in some statistical respects — its word-frequency distribution follows the Zipf-like patterns of real tongues — but it also does un-language-like things: words repeat two or three times in a row, the same word appears in slightly mutated forms with strange regularity, and the character structure within words is unusually rigid.
Or is it a hoax — meaningless glossolalia manufactured to look profound, perhaps to extract ducats from a credulous emperor? The scholar Gordon Rugg demonstrated in the 2000s that a low-tech device known to Renaissance cryptography, a Cardan grille sliding over a table of syllables, could generate Voynich-like gibberish quickly. But hoax theories struggle to explain why a forger would bother with the text's subtle statistical structure — and the carbon dating sits awkwardly early for the usual suspects.
Meanwhile, claimed decipherments arrive every few years — it's been declared Hebrew, Latin abbreviation, a Turkic dialect, an extinct Mexican language, and 'proto-Romance' — and every one has collapsed under expert scrutiny, usually within weeks. The Beinecke Library, which receives a steady stream of correspondence from would-be solvers, has made high-resolution scans of every page freely available online, effectively crowd-sourcing the siege.
And so the stalemate holds. Somewhere behind the looping script is one of three things: a lost language, an unbroken code, or the most patient practical joke of the Middle Ages. Whichever it is, a small book of calfskin from the early 1400s is still, quietly, undefeated.
Quiz nuggets
- The Voynich Manuscript is named after book dealer Wilfrid Voynich, who bought it in 1912 at the Villa Mondragone, a Jesuit college near Rome.
- Radiocarbon dating in 2009 placed its vellum in the early 15th century (c. 1404–1438), ruling out Roger Bacon as author.
- It is held at Yale's Beinecke Library as MS 408 and contains roughly 240 pages in an undeciphered script dubbed 'Voynichese.'
- A 17th-century letter claims Emperor Rudolf II once paid 600 ducats for the book.
- Even William Friedman, the US cryptologist behind breaking Japan's PURPLE cipher, failed to crack it after decades of study.