Researchers have used artificial intelligence to decipher the centuries-old Voynich Manuscript.
For over half a millennium, the Voynich Manuscript has been one of the most perplexing enigmas that researchers have failed to solve. However, the mysterious codex may soon be decoded-all thanks to recent advancements in AI technology.
Two computer scientists from the University of Alberta in Canada have allegedly made considerable progress uncovering the book’s hidden meaning. The key to their success? They used artificial intelligence.
More specifically, with the aid of an algorithmic decipherment technique, researchers Greg Kondrak and Bradley Hauer determined the underlying language used to write the book.@UAlberta researchers use #ArtificialIntelligence system to decode the Voynich Manuscript!Click To Tweet
Their study was published in the journal Transaction of the Association for Computational Linguistics. In this study, the duo describe how they used samples from nearly 400 different languages to identify which language the codex was closely correlated with. After several trials, it turned out that the manuscript was written in Hebrew.
“That was surprising,” Kondrak said. “And just saying ‘this is Hebrew’ is the first step. The next step is how do we decipher it.”
The Voynich Manuscript
The Voynich Manuscript is an illustrated codex that was handwritten in an unknown language. Aside from the language, the author of the manuscript is also a mystery.
Radiocarbon dating of some parts of the codex performed at the University of Arizona in 2009 revealed that it may have been composed between 1404 and 1438.
In 1912, Polish book dealer Wilfrid Voynich, after whom the manuscript is now named, obtained the book. It has a total of 240 pages, but many researchers believed that it initially had 272 pages. Some evidence also suggests that the book’s bifolios were reordered at different times in its history, making its original page order different from what it is today.
The book’s text consists of over 170 thousand characters. 8,114 of the words are deemed unique word types, following some kind of phonological or orthographic laws. It was first assumed that the manuscript was written in Latin language and was created to serve as a pharmacopeia or a book of medicine.
Using Artificial Intelligence to Decipher the Voynich Manuscript
While researchers in the past failed to uncover the mystery of the Voynich Manuscript, Kondrak and Hauer’s efforts yielded surprising results.
The first test, which was to figure out the language of the manuscript’s text, identified Hebrew as the book’s language. That was after the AI system analyzed the words and showed that 80 percent of them were found in the Hebrew language.
Using a Hebrew translator, the duo now believes that the first sentence of the codex translates to: “She made recommendations to the priest, man of the house and me and people.”
“It came up with a sentence that is grammatical, and you can interpret it. It’s a kind of strange sentence to start a manuscript but it definitely makes sense,” Kondrak was quoted as saying.
The researchers believe that the words were alphagrams, or anagrams where the letters of every word were ordered in an alphabetical manner. The two were also able to identify the words “farmer,” “light,” and “air,” saying that an expert in the Hebrew language may decode the manuscript further.
Despite their progress, the pair said that their algorithm is limited, and may perhaps be wrong. Right now, it only picks up individual words in a text. This, according to the researchers, is still too “noisy” to provide “a fluent output.”
Kondrak and Hauer concluded that the results of their study “could be interpreted either as tantalizing clues for Hebrew as the source language of the VMS, or simply as artifacts of the combinatorial power of anagramming and language models.”
The two are planning to refine their algorithm, hoping to apply it in other ancient manuscripts in the future.