Why Molière most likely did write his plays

Abstract

Quantitative linguistics contradicts the much publicized theory naming Corneille as the author of Molière’s masterpieces.

Authorship Studies and the Dark Side of Social Media Analytics

“…Authorship attribution [Juola, 2006,Stamatatos, 2009,Grieve, 2005, Jockers and Witten, 2010, strictly defined, is the task of analyzing a document to determine, not its meaning, but the identity of the person who created it. For example, a recent paper [Cafiero and Camps, 2019] addressed the question of Molière's writings with a large-scale study of various attributes of his work, including lexicon, rhyme, morphology (including a separate study of affixes), morphosyntax, and function words, comparing writings traditionally attributed to Molière with some of his contemporaries. They found that Molière's writings displayed both the internal stylistic consistency as well as the systematic differences from others that would be expected if he had been writing his own work.…”

“…2006] have proposed using sentence length, vocabulary complexity, vocabulary overlap, synonym pairs, parts of speech, and many others. One of the most powerful and commonly used feature sets are character clusters (formally called N-grams) [Cavnar and Trenkle, 1994, Stamatatos, 2013, Mikros and Perifanos, 2013, Cafiero and Camps, 2019, groups of N adjacent characters without regard to word boundaries-a n-gram can be at the beginning, end, or the middle of a word, or even incorporate the end of one word, a separating space, and the beginning of another. Rudman [Rudman, 1998] has estimated that more than a thousand feature sets have been used successfully in authorship attribution studies.…”

“…Large scale testing [Vescovi, 2011,Juola, 2012 has failed to find any consistent "magic bullet" that significantly outperforms other methods in accuracy. Ensemble methods using multiple datasets and classifiers are practical [Juola, 2008,Juola, 2015,Cafiero and Camps, 2019 and are becoming standard practice as a way of maximizing the accuracy and reliability of authorship judgments.…”

“…The anti-Stratfordian controversy about Shakespeare's canon is well-known [Friedman and Friedman, 1957], even if considered a fringe theory. Perhaps less well known (at least in Anglophone scholarship) is the question of whether Molière wrote his work, 1 or whether it was written by Pierre Corneille, perhaps as part of a "wide system of ghostwriting" [Cafiero and Camps, 2019] in 17th century French Theater.…”

“…Although this type of scholarly controversy goes back millennia, and attempts to resolve this sort of question on the basis of empirical analysis and data date to at least the 19th century [de Morgan, 1882, Mendenhall, 1887, recent advances in artificial intelligence, machine learning, and natural language processing have made this type of analysis much easier, faster, and more reliable. Applications of authorship analysis include not only the literary [Collins, 2013,Cafiero andCamps, 2019], but also the historic Wallace, 1963, Mosteller andWallace, 1964], the journalistic [Brooks and Flyn, 2013, Brooks, 2013, Herper, 2014, Pesca, 2018, and even the legal [Leonard, 2006, McMenamin, 2011, Ainsworth and Juola, 2019. This technology has proven to be a useful scholarly tool across the academy.…”