Determined by carbon dating to be written in the early 14th century in Central Europe, the manuscript’s written context has remained a mystery due to a cipher, a coded pattern of letters. Often called the world’s most mysterious book and rightfully so, the 240 page of text is written by an unknown author/authors in an unknown language is filled with various diagrams and illustrations. Determined to finally crack the code, computing scientists from the University of Alberta have used an AI algorithm to decode the mysterious manuscript once in for all.
Since its conception, the manuscript has been bought and traded by alchemists, emperors, and collectors until being housed at the Beinecke Rare Book & Manuscript Library at Yale University since 1969. The Manuscript is named after Wilfrid Michael Voynich, a Polish book dealer who purchased it from a Jesuit library in Italy in 1912 and attempted to interest scholars in deciphering the script.
Even with modern technology and expert cryptographers, the manuscript has yet to be deciphered. Wishing to change this, Bradley Hauer and Grzegorz Kondrak of the Department of Computing Science at the University of Alberta, Canada, have created an algorithm to decipher vowel-less alphagrams and anagrams in which letters in a word are rewritten alphabetically.
After testing the algorithm on 380 different language versions of the UN “Universal Declaration of Human Rights“, the AI reported a 97 percent success rate in matching anagrams to modern words. The researchers proceeded with feeding texts from the first ten pages of the Voynich Manuscript to the AI algorithm and found that 80 percent of the encoded words appeared to be written in Hebrew.
Now with a base language, the researchers attempted to figure out the code. First handing off the opening sentence to a colleague native Hebrew speaker, who ultimately wasn’t able to translate the text into English, the researchers turned to Google Translate. Their findings found that the first sentence read: “She made recommendations to the priest, man of the house and me and people.” The researchers were also able to translate a 72-word section, known as the herbal chapter and were able to decipher the words such as farmer, light, air, and fire with their new code.
These findings are, however, flawed due to the AI program being trained to translate various modern-day languages into English, as opposed to languages from the 14th century. The Voynich Manuscript matched 80 percent of the texts to Hebrew, but the AI would not have accounted for medieval Hebrew and instead the modern equivalent of Hebrew. This also leaves another 20 percent that was not determined to be Hebrew. According to the study, other languages that could have been used in the manuscript are Malay, Arabic, and Amharic. Lastly, the researchers used google translate to translate out the code, nowhere near as accurate compared to a human translator.
This, however, does not discredit their research and findings, as the researchers have determined the most prevalent language in the text and the coding. The next step in deciphering the manuscript codes would be to find a scholar well-versed in Hebrew and alphagrams to finally translate the Voynich Manuscript.
To learn more about the research conducted by the computing scientists from the University of Alberta, click the link below.
http://www.aclweb.org/anthology/Q16-1006
To view the scanned pages of Voynich Manuscript, click the link below.
http://www.bibliotecapleyades.net/ciencia/esp_ciencia_manuscrito07a.htm
