Nobody has translated the Voynich manuscript because nobody who has announced a solution has produced a passage that other specialists can check against the page and confirm. Three people have made that announcement since 2017, and none has produced a translation other experts accept.
more
The manuscript itself is why the puzzle holds. It runs through 240 vellum pages of handwriting in a script nobody recognises. Illustrations run alongside it: plants nobody can name, zodiac wheels that match no known calendar, women bathing in green pools, and pages of recipes. Carbon dating carried out in 2009 put the vellum at 1404 to 1438. Wilfrid Voynich, a Polish book dealer, bought it in Italy in 1912. It had earlier belonged to the Holy Roman Emperor Rudolf II, who is said to have paid 600 gold ducats for it believing the medieval English scientist Roger Bacon wrote it. Bookseller H.P. Kraus later bought the manuscript from Voynich’s widow and gave it to Yale University’s Beinecke Rare Book and Manuscript Library in 1969. It stays there still, in a vault, fully digitised for anyone to study.
Why a partial decode never counts as solved
Kate Wiles, a medievalist and senior editor at History Today, says a new reading of the manuscript turns up roughly twice a year. None of them has done what earlier scripts such as Egyptian hieroglyphics and Linear B needed before anyone accepted they were solved. That took decades of work drawing on many kinds of scholarly expertise until, as Wiles put it, “the field, on the whole” agreed. A decoder who translates a word here and a line there, or reads meaning into the drawings, has not cleared that bar. A translation of the Voynich manuscript has so far escaped scholars, code breakers and computer programs alike.
The 2017 claim: a copied medical guide
Nicholas Gibbs, a history researcher and television writer, wrote in the Times Literary Supplement that the manuscript was a women’s health guide. He said it was built mostly out of two commonly copied medieval Latin medical texts, the Trotula and De Balneis Puteolanis. Working from Yale’s digitised pages, Gibbs said the script was not a code at all but a set of standard medieval Latin abbreviations used in herbal medicine: “aq” for aqua, water, and “dq” for decoctio, a decoction. On that reading, the bathing women marked the book as a gynaecological manual, since doctors of the period prescribed baths as medicine, and the zodiac maps recorded which astrological sign suited which cure.
The claim did not hold for long. Lisa Fagin Davis directs the Medieval Academy of America, and she told the Atlantic’s Sarah Zhang that Gibbs’s decoding “doesn’t result in Latin that makes sense.” Other researchers added that much of what he presented as new had already been found by earlier work.
The 2019 claim: a nuns’ handbook in a lost language
Two years later, Gerard Cheshire, a research associate at the University of Bristol, said he had identified the manuscript’s language in two weeks, “using a combination of lateral thinking and ingenuity.” His paper, peer-reviewed and published in the journal Romance Studies, argued the text was written in proto-Romance, an early spoken form of Latin that never appeared in writing and faded as it split into the Romance languages spoken today. He went further, arguing the book was compiled by Dominican nuns as a reference for Maria of Castile, Queen of Aragon, on the island castle of Castello Aragonese off Ischia. Maria’s great-niece, he added, was Catherine of Aragon, Henry VIII’s first wife. He offered no translation of the text itself, saying that would take further time and money.
The University of Bristol withdrew its press release the next day, saying the research was “entirely the author’s own work” and not affiliated with the university, after linguists and medieval historians raised concerns. Lisa Fagin Davis was blunt. “Sorry, folks,” she wrote on social media, “‘proto-Romance language’ is not a thing.” Wiles was unconvinced too: Cheshire’s language, she said, drew words from too many places and periods to work as anything a real community could have spoken.
The 2020 claim: an Egyptologist and Hebrew
In 2020, Rainer Hannig, a German Egyptologist at the Roemer- und Pelizaeus Museum in Hildesheim, said three years of analysis had convinced him the manuscript’s language was Hebrew. He argued the word structure ruled out any Indo-European language, and that the European style of the illustrations narrowed the field to Arabic, Aramaic, or Hebrew before he settled on Hebrew. Hannig said he had translated individual words, and then full sentences, after matching certain Voynich characters to Hebrew letters. A complete translation, he cautioned, would need years more work from specialists in medieval Hebrew. He was not the first to suspect it: a 2016 study using a computer algorithm reached the same conclusion, though other researchers questioned its method, and it produced no translation either.
What over a century of failure has not ruled out
William Friedman, the American cryptologist who broke Japan’s wartime cipher, spent years on the Voynich manuscript and concluded only that it looked like an early attempt at an artificial, universal language. A 2014 study by Brazilian researchers used network modelling to show the text has the statistical shape of a real language, without producing a translation of it. Other researchers have argued the manuscript is a forgery or a deliberate hoax.
The puzzle reportedly reached Alan Turing and, during the Cold War, the FBI. Neither is known to have solved it.
