The Voynich Manuscript: The Book Nobody Can Read

In a climate-controlled room at Yale University’s Beinecke Rare Book and Manuscript Library sits a small, unassuming codex, about two hundred and forty pages of soft vellum, catalogued as MS 408. Open it and you fall straight through the floor of ordinary understanding. The pages are covered in a flowing, confident script — clearly written by a practised hand, running left to right in neat lines, with what look like paragraphs and headings — and not one word of it can be read. No one knows what language it is. No one knows what alphabet it uses. Around the text sprawl hundreds of illustrations that only deepen the vertigo: plants that resemble no species that has ever grown, elaborate astronomical diagrams, and page after page of small naked women bathing in green pools connected by a plumbing of tubes and channels. It is beautiful, meticulous, and utterly, defiantly mute. For more than a century the finest codebreakers on Earth have tried to make it speak, and it has said nothing at all.

The Voynich Manuscript is the most famous unreadable book in the world, and its true subject is not really the script. It is us. A blank refusal on this scale acts on the human mind like a wound, and for a hundred years people have thrown themselves at it — professional cryptologists, amateur linguists, computer scientists, cranks and geniuses often indistinguishable at first glance — each certain that the book must yield, must contain a secret, must in the end make sense. The manuscript is a near-perfect laboratory for watching what happens when the human hunger for meaning meets a surface that gives it nothing to hold. To read the story of the Voynich Manuscript is to read the story of that hunger.

What we actually know

For a document so shrouded, a surprising amount is solid, and the solid part is worth laying down first. In 2009, researchers radiocarbon-dated samples of the vellum, and the results were clean and consistent: the animal skins were prepared in the early fifteenth century, most likely between about 1404 and 1438. The parchment, at least, is genuinely medieval; this is no Victorian forgery painted on fresh calfskin. Analysis of the inks and pigments is broadly consistent with that period as well. Whatever the Voynich Manuscript is, it was physically made in the early 1400s by someone with real materials and real skill.

Its modern history begins in 1912, when a Polish-born book dealer named Wilfrid Voynich bought it, along with other volumes, from the Jesuit-owned Villa Mondragone near Frascati in Italy. Voynich, whose name the manuscript now bears, spent the rest of his life trying to establish its origin and meaning and failed. Tucked in with the book was a letter that gives the trail its most tantalising thread: a 1665 note from Johannes Marcus Marci, a scholar in Prague, sending the manuscript to the Jesuit polymath Athanasius Kircher and repeating a story that Emperor Rudolf II — the occult-minded Habsburg ruler in Prague — had once paid the enormous sum of six hundred gold ducats for it, and that it was believed to be the work of the thirteenth-century English friar Roger Bacon. The Bacon attribution is almost certainly wrong; the carbon dating alone rules him out by a century and a half. But the Marci letter fixes the manuscript in the collecting world of Rudolfine Prague, a court that drank deeply of alchemy, astrology and secret wisdom, and that context has coloured every reading since.

The pages themselves fall into loose sections that scholars have named by their pictures: a “herbal” section of impossible plants, an “astronomical” or “cosmological” section of stars and circular diagrams, a “balneological” section of the bathing women in their green channels, a “pharmaceutical” section of jars and plant parts, and a run of dense text with star-shaped markers that reads like recipes. That is the honest inventory. Old vellum, skilled hand, coherent internal structure, a documented journey through Prague and Rome to Yale — and a script that has resisted every attempt to turn it into language.

A century of brilliant failure

The roll-call of people the Voynich has defeated is what gives its silence such authority. During the twentieth century some of the best cryptanalysts alive took a run at it. William Friedman, the towering American codebreaker whose team broke the Japanese “Purple” cipher in the Second World War, studied the manuscript for years with a group of colleagues and never cracked it; his considered guess, left in an anagram, was that it might be an early attempt at an artificial or universal language rather than an enciphered natural one. His British counterparts and other wartime cryptologists tried and failed too. If the men and women who broke the ciphers of great powers could not break this, the reasoning goes, then surely something profound is buried here.

The theories multiplied to fill the silence, and they sort into a few families. The first holds that the text is a natural language written in an invented alphabet — perhaps a lost dialect, or a European tongue in a private script — a view that has to explain why the statistical fingerprint of “Voynichese” matches no known language cleanly. The second holds that it is a cipher, a genuine encryption of a real message, which is the romantic favourite and also the one that a century of expert cryptanalysis has most thoroughly failed to substantiate. The third holds that it is a constructed or synthetic language, Friedman’s guess, built from scratch. And the fourth, the coldest, holds that it is meaningless — an elaborate hoax, gibberish designed only to look like text, perhaps produced to defraud a wealthy collector like Rudolf II out of his six hundred ducats.

That last theory has real intellectual muscle behind it. The British academic Gordon Rugg argued in the early 2000s that Voynichese could be generated as convincing-looking nonsense using a low-tech method available in the sixteenth century: a “Cardan grille,” a table of syllables and a masking card that, moved across the table, spits out streams of word-like gibberish with plausible structure. More recently the German researcher Torsten Timm proposed that the text was self-copied, each “word” generated by slightly altering nearby words on the page, which would neatly explain some of Voynichese’s oddest statistical quirks without any underlying message at all. Neither theory is proven. But they establish something the romantics resist: it is entirely possible to make a book that looks exactly this meaningful and means nothing.

The tell in the “solutions”

Every few years, with the reliability of a season, someone announces that they have decoded the Voynich Manuscript. The announcements make headlines, ripple through the press, and then quietly die, and the pattern of their dying is the most instructive thing in the whole affair. In 2017 a television writer and researcher declared it a women’s health manual in a Latin abbreviation; specialists found the reading did not hold up across the text. In 2019 an academic claimed it was written in a “proto-Romance” language and could be read almost fluently; the University of Bristol, whose press office had promoted the paper, distanced itself after linguists dismantled the method. Others have “solved” it as Hebrew run through anagrams, as a Turkic language, as Nahuatl from the New World, as any number of things.

Look at what these solutions have in common, because it is the same tell that haunts every under-constrained mystery. Each works, more or less, for the person proposing it, and for no one else. Each requires the solver to grant themselves a great deal of interpretive freedom — to add vowels the script omits, to treat any letter as several possible letters, to read a “word” as an abbreviation or an anagram or a phonetic hint as convenience demands. Once you allow yourself that much latitude, a text this long will let you find almost anything you came looking for, and different seekers duly find different things: the Latin reader finds Latin, the Hebrew reader finds Hebrew, the Romance reader finds Romance. A genuine decipherment, like the ones that finally cracked the Zodiac’s ciphers, behaves in the opposite way: it produces the same fluent, continuous, reproducible plaintext for anyone who applies the recovered key, on page after page, without special pleading. Every Voynich “solution” so far has failed that basic test. They are mirrors, and what each one shows is the face of the person holding it.

Why we cannot let it be blank

Strip the case to its core and the deepest question is not “what does the Voynich Manuscript say?” It is “why can we not bear the possibility that it says nothing, or that whatever it said is simply lost?” Both of those are live, respectable answers. It may be sophisticated hoax-gibberish, made to look precious. It may be a real text in a system whose key died with its maker six hundred years ago, gone as completely as a spoken language with no last speaker. Either way the honest position is a shrug, and the shrug is the one thing almost no one who falls under the manuscript’s spell can accept.

The reason is not stupidity; it is a feature of how minds are built. We are pattern-finding animals, tuned by evolution to see signal in noise because the cost of missing a real pattern once was death, while the cost of imagining a false one was usually nothing. Confronted with something that has every surface mark of meaning — the neat lines, the paragraphs, the confident hand, the careful illustrations — the mind simply refuses to file it under “meaningless.” It insists there must be a message, because everything about the object promises one. The manuscript exploits that reflex more purely than almost any other artefact, precisely because it is so competently made. A crude fake would not tempt us. This one does, because it wears all the costume of sense.

That refusal to accept a blank is the same appetite at work across this desk. It is why a nation would rather believe a hijacker survived his jump and vanished clean than accept a body lost in the woods, and why an unsolved murder or an unfinished investigation breeds elaborate certainty where the plain answer is only we do not know. The Voynich Manuscript is the pattern in its purest form, stripped of blood or politics: a beautiful surface that promises meaning and withholds it, and a species that cannot stop reaching. Perhaps one day a genuinely reproducible key will emerge and the book will finally speak, and that would be a wonder. But if it never does — if the six-hundred-year-old script is a lost language, or an ingenious nothing — the manuscript will have taught us something almost as valuable as its contents. It will have shown us, page by unreadable page, exactly how badly we need the world to make sense, and how much of that sense we are prepared to invent when it declines to supply its own.