Shakespeare is suspected of having written the play "Henry VIII" with one or two co-authors. Thanks to a so-called machine learning algorithm, a Czech literature researcher has just unraveled the mystery. He machine-learned the style of each of the three authors by providing them with hundreds of words.
Thanks to artificial intelligence, it is possible to formally identify the author of a text. It could upset the little world of authors.
Take Shakespeare's play "Henri VIII", an extract that has fascinated researchers in particular since 1850. We suspect the playwright to have written it with one or even two co-authors. Thanks to a so-called machine learning algorithm, a Czech literature researcher has just unraveled the mystery. He machine-learned the style of each of the three authors by providing them with hundreds of words. Then, by comparing each style to whole passages of the piece in question, artificial intelligence, arrived at this conclusion: Shakespeare has well written passages with two hands with his friend Flechter but never with the other, Massinger. End of suspense.
Another enigma divided researchers in literature: Did Corneille write Molière's plays?
Florian Cafiero, a math engineer, has co-signed a study to dismantle this tenacious century-old rumor. No, Corneille did not write Molière's plays for him. To achieve this result, he relied on Artificial Intelligence to measure, among two million words, the quantity of proper nouns, adverbs, adjectives and conjugated verbs used by several authors of comedy in verse of the time. Titanic task impossible to perform by humans without the computing power of robots. Then, piece by piece, he compared these results to each person's style based on six very revealing criteria including rhymes, lexicon, tics of language, grammatical sequences. On each criterion, Molière leads far ahead of Corneille, according to relentless statistics. Corneille was therefore not the negro of Molière.
Does all this point to other applications?
Why not ! To identify the author of medieval texts whose origin is doubtful? Closer to us, to put a name on the editor of anonymous messages on social networks. To track down followers of plagiarism, this sport that regularly enlivens the small world of literature. In return, this raises ethical questions. What puzzle to defend the authors? How to protect a writer who publishes a book under a pseudonym? A whistleblower or a journalist who wishes to report a scandal in an anonymous forum? With the rise of these algorithms, they risk being quickly exposed. Tomorrow, there will no longer be Romain Gary who, thanks to his nickname, Emile Ajar, is the only writer to have won two Goncourt prizes. Jean-Pierre Montanay also wishes good luck to the next Elena Ferrante to maintain the mystery for as long.