Scientists had been scratching their heads of half-a-millennium about the so-called Voynich manuscript, but no one was able to solve it.
And now a computer scientist claims to have cracked it using AI.The manuscript is written in ancient Hebrew and the code involves shuffling the order of letters in each word and dropping the vowels.
For decades codebreakers have poured over the 240-page relict, from which pages have been carbon dated back to the early 1400s. It has been suggested that the manuscript, which is housed at the Beinecke Library at Yale University and contains brightly colored images of plants and an unknown language, is a medieval medical journal for women.
The first stage of the research was working out the manuscript’s language. The experts used 400 different language translations from the “Universal Declaration of Human Rights” to identify the language used in the text.
“That was surprising,” said Kondrak, in a statement. “And just saying ‘this is Hebrew’ is the first step. The next step is how do we decipher it.”
Kondrak and Hauer worked out that Voynich manuscript was created using ‘alphagrams’ that use one phrase to define another so built an algorithm to unscramble the text.
The initial part of the text was then run through Google Translate. “It came up with a sentence that is grammatical, and you can interpret it.”
The sentence was: “She made recommendations to the priest, man of the house and me and people.”
The full meaning of the text will need the involvement of historians of ancient Hebrew. The vellum, or animal skin, on which the codex is written has been dated to the early 15th century.
The research study is published in Volume 4 ofTransactions of the Association for Computational Linguistics.
There have been multiple attempts to decode the Voynich manuscript. In 2014, for example, researchers argued that the illustrations of plants in the manuscript could help decode the text’s strange characters. In 2011, a self-proclaimed “prophet of God” claimed that he had decoded the book.
