The Voynich Manuscript

by Brian Dunning
April 5, 2011

Today we're going to look at the most famous undeciphered text of all time: A medieval book of science, full of beautiful illustrations and strange wisdom, and containing not a single word that anyone's been able to make heads or tails of: The Voynich manuscript.

So let's get the big question out of the way right up front. The Voynich manuscript is an unsolved mystery, at least so far. According to the best information we have now, we still don't know who wrote it, what it says, or what its purpose was. We do have some theories, but there will be no unveiling of a glorious answer today. However, the voyage of scientific exploration is always a fascinating one, and much of what we have learned is just as interesting as what we haven't.

Somewhere in Europe, probably northern Italy, sometime in the early 1400s, animals were slaughtered (either sheep, calves, or goats) and their skin turned into parchment. Probably very soon thereafter, someone, most likely two people, took a quill pen in hand and wrote a 38,000-word book using common ink, beginning to end, using an alphabet and language that have defied all identification. It's not a huge book, measuring about 16 by 23 cm (about 6 by 9 in), and about 5 cm (2 in) thick. There about 240 pages, most of them illustrated, the exact number depending on how you count pages that fold out into large diagrams or drawings, of which there are several. The alphabet has between 23 and 40 distinct characters, depending on how you classify some which may be decorative versions of others or two-character combinations.

The book has six sections, delineated by the types of illustrations. Section 1 is the largest at 130 pages, and contains detailed drawings of 113 plants and flowers that nobody has been able to identify. It's called the Botanical section. Section 2 is 26 pages of Astrological drawings; lots of circular and concentric diagrams, and some signs of the zodiac. The third section is called the Biological section and contains mainly drawings of nude women frolicking in intricately plumbed pools. Section 4 is the Cosmological section, featuring the most impressive foldouts that appear to be circular diagrams of some cosmic nature. The fifth section is Pharmaceutical, with over 100 drawings of herbs, roots, powders, tinctures, and potions of undecipherable contents or use. The last section, called Stars, is the most mysterious; it's 23 solid pages of text only, in short paragraphs, each marked with a star. Some of the illustrations show Eastern influence, including a probable map of the circular city of Baghdad, the center of Eastern knowledge.

A few hundred years later (we don't know exactly when), a cover was added, but unfortunately it's blank. Also at some later date, the illustrations were colored, by someone less careful than the original artist.

The book was owned by the English astrologer John Dee during the 16th century, who wrote his own page numbers in the upper right corner of each leaf. Dee sold it to Emperor Rudolph II of Germany, with the understanding that it was the original work of Roger Bacon, a 13th century friar widely regarded as one of the fathers of the scientific method. From there the book passed to one or two other owners, who wrote their names in it, and at one point it was presented to the scholar Athanasius Kircher in Rome along with a signed letter from a Johannes Marcus Marci expressing a hope that Kircher could translate it, in 1666. Marci's letter is still preserved with the book. The manuscript's history becomes unclear at that point, until it was discovered by antique book dealer Wilfrid Voynich in 1912 at the Jesuit college at Villa Mondragone in Italy. Voynich brought it to the attention of the world. After several owners, the book was eventually donated to its current home, the Beinecke Library at Yale University, under its official name of MS 408.

Since its discovery, hypotheses have abounded as to what the Voynich manuscript means. Many believe it's written in a type of code, but all efforts to find decodable patterns have failed. Some believe it may be what's called a constructed language, which is a language that's deliberately planned and invented rather than naturally evolved. Some have speculated it's to be used with a Cardan Grille, a paper with holes in it that you lay on top of the page and read only the revealed letters. Perhaps the most popular theory is that it's a hoax, written at practically any time since the parchment was made, and for just about any purpose ranging from financial gain to scholarly fraud to someone's weekend lark.

Guesses on its authorship are just as plentiful. Roger Bacon remains the usual suspect, but this is based only on the presumption of most of its owners and is not supported by any evidence. Roger Bacon never wrote anything else in the Voynich language so far as we know. Moreover, he died in 1294, more than 100 years before the book was written.

We can be certain of that, because we do know when the parchment was made, a fact that neither Voynich nor his predecessors could have known. Carbon dating of the parchment was performed at the University of Arizona in 2011 by Dr. Greg Hodgins, and nailed it down to the early 1400s. Dating the ink, however, is not something that we have any good way to do. Most ink can't be radiocarbon dated because it doesn't necessarily contain any organic matter; and even when it does, we don't have the technology to reliably separate carbon in the ink from carbon in the parchment. We've found that the pigments used are consistent with what is known to have been used in those years, but it could also be consistent with expert modern fakery.

However, we can still make some strong educated guesses. Parchment was commonly washed and re-used (it's a good way for modern forgers to create a document that legitimately radiocarbon dates to an ancient time), but doing so leaves chemical footprints. We do know that the Voynich manuscript was the first application of ink to its parchment. And from history, we know that parchment was always in good demand; it would have been virtually inconceivable for perfectly good parchment to sit unused for decades or centuries waiting for someone to come along and make a forgery on it. Combined with the fact that we have no reason to doubt the history of the book's ownership as given in Marci's letter, we can be pretty confident that the book was written about the same time as the parchment was made.

So let's look at the book's other properties to see what we can learn.

Here's an important one. There are no corrections in the book. There are also no places where the text has been squeezed to fit onto the page. This would be highly improbable if it were an original manuscript; we would absolutely expect there to be such minor errors in a first edition. So how do we explain this? There are a number of possible explanations, but two of them are most likely.

The first is that the book is a copy, perhaps even of something written by Roger Bacon. If a scribe has an original to work from, he can see how many words there are and properly plan everything to fit onto the page. And if he copies carefully, there will be no corrections. The copy theory is also consistent with other characteristics, such as its appearance of having been written straight through by only one or two people. If it is a copy, this alone doesn't tell us much that's useful in trying to decipher it. But it does leave us wondering why anyone would go to the trouble of making a nice copy of a book that doesn't say anything.

The second theory to explain the book's neat appearance is perhaps more revealing. The text could be complete nonsense, made up as the scribe went. There would be no need for corrections. There would be no need to compress the writing as space ran out.

The "complete nonsense" theory has one thing working against it. If it is nonsense, it's very good nonsense. It's almost too good to expect of an amateur. Computational analysis of the text has been run, exhaustively, many times by many different researchers, using many different techniques. This allows us not only to try and translate it (at which all attempts have met utter failure), but also to compare its metrics to those of actual languages. The letter frequency, word length, and word frequency are very similar to what we see in real languages. But they don't quite match those of any real languages. It's speculation, but I can imagine a monk or professional scribe who does this all the time being well aware of such things and deliberately giving the book a realistic appearance, but it seems less likely that an amateur, just a Joe Blow or professional from a different field, would happen to write gibberish that's such good gibberish.

But clues indicating that there is meaning within the text don't end there. Patterns of word usage and word relationships are also found within the six different sections, as if the sections are actually about different subjects. The pages within each section are more similar to each other than they are to pages in other sections.

However, a broader analysis of this leads to another interesting point, which we in the brotherhood often describe as "The plot thickens."

A famous analysis done in the 1970's by US Navy cryptographer Prescott Currier found that the Voynich manuscript is written in two distinct languages. He used the term languages, but also cautioned that they're also consistent with different subject matter, different encryption schemes, or possibly just different dialects. He called them Voynich-A and Voynich-B. Interestingly, Voynich-A and Voynich-B are in two different handwritings, though both use the same alphabet and script. Every page of the book is written entirely in either A or B. The Biology and Star sections of the book are written in Voynich-B; the others are written in Voynich-A. The exception is the first and largest section, Botanical, which contains some of each. But they're not simply interspersed. The way the book is bound uses bifolios, which are groups of pages folded together, which are then stacked on top of one another to be bound. Each bifolio in the Voynich manuscript is written entirely in one language or the other.

So let's wrap this up with my favorite theory, and the one that is perhaps best supported by all that's known. In the early 1400s, some professional, perhaps a physician or astrologer or alchemist, wanted to create some marketing material that demonstrated he had rare knowledge from the East. He engaged a monk or other scribe to produce a book filled with wondrous and curious illustrations from multiple sciences, and a text that nobody could read, which he could tell his customers was the source of whatever great Eastern wisdom he wanted. The monk had a colleague assist, and the two devised an alphabet and used their own multilingual familiarity with written language to devise a convincing nonsense text. It was well done enough that its owner could even use it to impress his colleagues. Thus, this anonymous professional ended up with impressive marketing collateral that's conceptually identical to the labcoat worn by a naturopath, the energy diagram on the wall of a yoga guru, and the purchased-online title of "doctor" sported by alternative practitioners of every variety.

This remains the leading theory. Not quite a hoax, and very deliberately and carefully created; yet full of nothing but pure nonsense. Perhaps one day the Voynich manuscript will reveal a different purpose, but for now, this theory is as good as any.

By Brian Dunning
Follow @BrianDunning