The Voynich manuscript: the book nobody can read

Every page is written. Not one has been read.

For a book, the night of is the moment of discovery, and this one has a date. In 1912, according to the Yale catalogue record[yale-catalog], the book dealer Wilfrid Voynich acquired the manuscript from the Jesuits at Villa Mondragone, in Frascati, near Rome.

Scene · reconstructed from the paperworkPicture it. A dealer opens a small book and finds what the Beinecke Library describes[beinecke-highlight] as an illustrated codex in an unknown script. Page after page of careful, flowing writing that matches no known alphabet, and around the writing, pictures. The file doesn't say what Voynich thought when he first turned the leaves, and I won't pretend to know. It says what he found, and that he bought it.

What he found is still there to see. The book now sits in New Haven as Beinecke MS 408[yale-catalog], and Yale has put a full digital facsimile[yale-catalog] online. Anyone can open a page today, like the public-domain scan of one leaf[commons-voynich-32] in this file, and meet the same problem Voynich met: plants, circles, figures and diagrams, all carefully labelled, and none of it readable.

The pictures are part of why the book is so hard to leave alone. They look as if they should explain the text. Pages of plants suggest a herbal. Circles of stars and suns suggest astronomy or astrology. Other pages show small figures in pools and long diagrams that seem to map something. The Beinecke's own presentation of the manuscript[beinecke-yt], with the manuscript scholar Lisa Fagin Davis, walks through the book in exactly this way, by its pictures, because the words give nothing back.

I keep coming back to that imbalance. In most old books, the text tells you what the pictures mean. Here the pictures are the only way in, and they may be leading us somewhere the text never meant to go.

So the book is real, and it is old. The first question any archivist asks of a book like this is simple: where has it been? The answer starts long before 1912.

Most mysteries have a gap in the paperwork. This one has a paper trail with a long hole in the middle.

The Yale catalogue[yale-catalog] gives the provenance in order: the library of Emperor Rudolf II, Marci in 1666, Kircher, Voynich in 1912, the dealer H. P. Kraus, and a gift to Yale in 1969. Let me take those one at a time.

• Rudolf II. The earliest link is to the library of the Holy Roman Emperor in Prague. The source for it is a later letter, so the catalogue's link to Rudolf[yale-catalog] rests on what one owner believed rather than on a receipt. A name inscribed in the book, connected to Rudolf's court, is still on the file's list to verify[yale-catalog], along with a Prague owner before Marci.
• Marci to Kircher, 1666. On firmer ground: Johannes Marcus Marci sent the book to Athanasius Kircher in Rome[yale-catalog], and the accompanying letter is what links it to Rudolf's library. Marci evidently hoped the great Jesuit scholar could read it.
• The quiet centuries. After Kircher, the book stayed in Jesuit hands until Voynich acquired it at Villa Mondragone, Frascati, in 1912[yale-catalog]. The file records nothing in between.
• Kraus and Yale. After Voynich, the manuscript passed to the dealer H. P. Kraus, who listed it in his Catalogue 100[yale-catalog], and in 1969 it was given to Yale's Beinecke Library as MS 408[yale-catalog]. The year of Kraus's purchase is on the to-verify list[yale-catalog].

Then there is a second kind of paper, one written by a laboratory. On 17 February 2009, the NSF-Arizona AMS laboratory sent a letter reporting radiocarbon dates on four vellum samples[arizona-c14-letter]. Combined and calibrated, they gave a range of 1404 to 1438, at 95.4% probability[arizona-c14-letter]. On 9 February 2011 the University of Arizona announced the early fifteenth-century date[uanews-2011] to the public.

The file is careful about the quality of that second trail. The result comes from a lab letter and a university announcement, not a peer-reviewed paper with a DOI[arizona-c14-letter]. That does not make it wrong. It means the Archive cites the letter itself, and says so.

The provenance tells us who held the book and wondered. The laboratory told us how old the skin is. Neither told anyone what it says, and that is where the codebreakers come in.

For a book nobody can read, the turn comes in two parts: the claims of solution, and the one hard fact that cut many of them off.

The first part is long. From Marci's hopeful letter to Kircher in 1666[yale-catalog] to the present, people have tried to read this book. Cryptographers, linguists, historians, mathematicians and countless amateurs have taken a turn. Many have announced a solution: a lost language, a shorthand, a coded herbal. The analysis in this file states the result of all that effort in one line: there is no decipherment accepted by specialists[yale-catalog].

The second part is the laboratory. The radiocarbon letter[arizona-c14-letter] placed the vellum between 1404 and 1438. The university's announcement[uanews-2011] put it plainly as early fifteenth century. That matters because a good deal of the book's legend had tied it to much earlier or much later authors. A date on the animal skin narrows the field.

But note what the test measured, because the file is careful about it and so should we be. It dated the parchment, not the ink[arizona-c14-letter]. In principle, the text could have been written later on old vellum. The analysis adds, fairly, that nothing points to that[arizona-c14-letter]. The likeliest reading is that the book was made not long after the skins were prepared, but likeliest is not proven.

Now the other side, and it deserves its full weight. Some researchers argue the text could be a meaningless hoax made to sell[yale-catalog], careful nonsense written to impress a rich collector. A book that looks profound, is beautifully illustrated, and cannot be read would have been a fine thing to put in front of an emperor who collected wonders.

Against that, the dossier notes that statistical studies show word patterns resembling natural language[yale-catalog]. Words repeat and cluster in ways that look more like a real language than random scribbling. That has not produced a reading, but it has kept the hoax answer from closing the case.

So the turn is not a solution. It is a narrowing. We know roughly when the vellum was made. We know who held the book. We still do not know whether there is anything inside it to read.

Here is what the record leaves open.

• What it is. The question at the head of this file remains: is the text a language, a cipher or meaningless[yale-catalog]? Each answer has its defenders, and none has a reading that specialists accept.
• When the ink went on. The radiocarbon test dated the parchment, not the ink[arizona-c14-letter]. A direct test of the ink or a closer study of the writing would answer a different question from the one already answered.
• The radiocarbon paperwork. The date rests on a lab letter[arizona-c14-letter] and a university announcement[uanews-2011], not a peer-reviewed paper with a DOI. The file looked for one and did not find it.
• The early owners. The Rudolf II link[yale-catalog] comes from a later letter, and the names before Marci are still to verify[yale-catalog]. The two centuries after Kircher are a blank in the file.

What the record does show is solid, and worth saying plainly. The manuscript is real. Its vellum dates to 1404–1438[arizona-c14-letter]. Its later ownership, from Marci in 1666 to Yale in 1969[yale-catalog], is well documented. And anyone can see every page, because Yale has put the whole book online[yale-catalog].

What the record does not show is that the book holds secret knowledge, or that any announced decipherment is correct. I have not made it more mysterious than it is. It does not need the help.

The finding and grade stay as the file has them. A real book, an early date, and a text that has outlasted every claim of solution so far. The catalogue, the lab letter and the timeline are below, and so is the book itself, one click away.