A team of researchers from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) announced last Wednesday (21) the cre...
A team of researchers from MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) announced last Wednesday (21) the creation of a deciphering algorithm capable of automatically providing the meaning of languages lost long ago, even if they have no relation with other languages.
The project is based on an article written last year by Professor Regina Barsilay, from MIT, and PhD student Jiaming Luo, from MIT , who deciphered two dead languages: Ugarit and Linear B, which took many years to be deciphered by humans. However, these languages were related to the first forms of Hebrew and Greek.
With the new system, the relationship between languages is inferred directly by the algorithm. This issue was one of the greatest challenges of deciphering. In the case of Linear B, a syllabary used by Mycenaean Civilization between the 15th and 12th centuries B.C., it took decades of study to decipher it.
But there are languages, such as Iberian, of which no other related languages are known, with some defending Basque, while others categorically state that Iberian is not related to any type of known language.
Prepared to evaluate the proximity between two languages, the algorithm was tested in Iberian, comparing it to Basque and other less likely candidates, such as Romance, Germanic, Turkish and Uralic families. None of the languages were considered related, not even Basque or Latin.
To reach this conclusion, the algorithm had to categorize words in an ancient language and link them to their equivalents in other related languages. In other words, the process cannot be a Google Translator, translating "dead" languages directly into English, but it can identify the roots of several ancient languages.
For future work, the team intends to expand the deciphering beyond the act of connecting texts to words related to an already known language. A new approach would involve identifying the semantic meaning of words, even without knowing how to read them.