A pair of Canadian codebreakers may have deciphered a 600-year-old book that has been baffling cryptologists for centuries. But, more likely, they probably haven't.
In a study published in the journal Transactions of the Association of Computational Linguistics, computing scientists from the University of Alberta used an algorithm to try to decode parts of the Voynich Manuscript, a medieval book written in an undecipherable code with an unknown language.
But other scholars are skeptical, and the manuscript remains a document very much shrouded in mystery.
What's the Voynich Manuscript?
The Voynich Manuscript is likely what cryptologists call a cipher, or a coded pattern of letters. Written in Central Europe in the 15th century, the book is slightly larger than a modern paperback and contains 246 fragile pages of bound vellum, or script-ready animal skin. It doesn't include an index but likely had foldouts that have long since gone missing. There are gaps in the page numbers and evidence that it could have been rebound at some point, so the order of the pages today may be different than they were when the book was published.
THE UNBREAKABLE CODE Even the best cryptographers cannot decode the mysterious Voynich manuscript.
n elegant, looping script of 25 to 30 characters runs from left to right in short paragraphs down the pages, interspersed with detailed illustrations. The renderings show doodles of castles and dragons along with diagrams of plants, planets, naked figures, and astronomical symbols, all detailed in green, brown, yellow, blue, and red ink. One particularly curious passage shows dozens of naked women bathing in pools of interconnected green liquid.
The manuscript has been housed at the Beinecke Rare Book & Manuscript Library at Yale University since 1969. It's named for Wilfrid Michael Voynich, the Polish book dealer who purchased it from a Jesuit library in Italy in 1912. He tried to interest people in translating it, but alas, none have succeeded.
Do we have any idea what the manuscript might be about?
Based on the illustrations, scholars believe the book is divided into six sections: herbal, astronomical, biological, cosmological, pharmaceutical, and recipes. It's possible the manuscript is of magical or scientific nature.
Historical records show the text has fallen into the hands of alchemists and emperors alike. In the late 16th century, a German emperor purchased the manuscript from an English astrologer for 600 Venetian ducats, thinking it was a work of medieval friar Roger Bacon. From there, it eventually traded hands to a Bohemian pharmacist.
What's in this new study?
The study authors write that the Voynich Manuscript is "the most challenging type of a decipherment problem," since we don't know its secret code but—perhaps more importantly—don't know what language it's in, either.
They approached the text armed with a computer program of their own design. Originally, the scientists suspected the manuscript to be made out of a type of vowel-less alphagram, or an anagram in which letters in a word are rewritten alphabetically. (For example, "manuscript" would be alphagramized as "acimnprstu.") So, they trained an algorithm to decipher 380 different-language versions of the UN "Universal Declaration of Human Rights."
Once the AI turned up a 97 percent success rate in matching anagrams to modern words, the researchers fed text from the first ten pages of the Voynich Manuscript into it. The algorithm found that 80 percent of the encoded words appeared to be written in Hebrew.
Now, supposedly, the researchers had a language. Next, they needed to figure out the code it was in. So they handed off the opening sentence to a colleague who was a native Hebrew speaker. When he wasn't able to translate the text into coherent English, they turned to Google Translate because no other scholars were available. After the researchers corrected some spelling errors, the first sentence read: "She made recommendations to the priest, man of the house and me and people." It's a strange sentence but it does make some sense.
ENCODED MEDIEVAL MEDICINE Restorator Paula Zyats, with Voynitch expert René Zandbergen, makes educated guesses as to what the manuscript is about.
The scientists also translated a 72-word section—known as the "herbal" chapter—and were able to decipher the words "farmer," "light," "air," and "fire" with their new code.
Wait … Google Translate?
Yes, Google Translate. The machine translator works by analysing hundreds of millions of documents that have been translated by humans. Then, using statistics, the tool spits out a translation based on those documents. Although the tool works to translate words in groups, rather than word-by-word translation, it's still not nearly as effective as human translators.
Now, back to the manuscript.
What are some other problems with the study?
For starters, the AI program was trained by translating different modern-day languages into English, as opposed to languages from the 15th century. Although the Voynich Manuscript may have been written in Hebrew, it would have been medieval Hebrew and not the modern equivalent Google Translate uses.
Although the manmade algorithm matched up 80 percent of the text as Hebrew, that leaves another 20 percent matched with different languages. According to the study, other languages that could have been used in the manuscript are Malay, Arabic, and Amharic, which are vastly different from Hebrew.
To be fair, the researchers are not claiming to have unlocked the secrets of the entire Voynich Manuscript. Rather, they're saying they've determined the language and coding scheme of the text. The next step is to find a scholar well-versed in Hebrew and alphagrams, and they're excited to apply this code-breaking technique to other ancient manuscripts.
But people have been wrong in the past, and other theories have been quickly debunked by scholars. Even Alan Turing of Nazi Enigma code fame wasn't able to unscramble the Voynich Manuscript.
We're still not sure if the text is written in coded or constructed language. And, still, it could be entirely meaningless.
What are some other theories on the manuscript?
In addition to the Canadian team, other researchers have proposed the manuscript is written in Hebrew. Dozens of other languages have also been brought into question, including Latin and a language derived from the Sino-Tibetan family.
Some suggest the book contains early discoveries and inventions by Roger Bacon. But it could also be a pidgin prayer book from a heretical Christian sect, or a meaningless collection of gibberish sold by an occult philosopher for monetary gain.
The Voynich Manuscript remains one of the biggest unsolved problems in the history of cryptology. Many translations are proposed each year, but a definitive code has yet to be determined.
Lead Image: The Voynich Manuscript, a small unassuming book stored in a Yale University vault, is one of the most mysterious books in the world. The precious document containing elegant writing and strange drawings is believed to have been written six centuries ago in an unknown or coded language that has never been cracked. PHOTOGRAPH BY CESAR MANSO, AFP, GETTY