AI reveals how much of a Shakespeare play was written by him and how much by someone else

AI reveals how much of a Shakespeare play was written by him and how much by someone else

Published in: Artificial intelligence

by Pisana Ferrari – cApStAn Ambassador to the Global Village

 

Academics in Prague have recently used machine learning to investigate how much of the play Henri VII was written by William Shakespeare and how much was written by John Fletcher, another famous and prolific playwright of the time. Machine-learning algorithms have been used for some time to identify idiosincrasies and patterns in the way authors write, including for Shakespeare. For example, the results of a previous research on the three plays of Henry VI suggest that Christopher Marlowe and George Peele were the two most likely “suspects”.

Petr Plecháč, from the Czech Academy of Sciences in Prague, first trained an algorithm to recognize Shakespeare’s style using other plays written at the same time as Henry VIIIHe then trained the algorithm to recognize the work of John Fletcher using plays he wrote, and he also investigated the possible influence of Philip Massinger, another playwright of the time. Altogether there were 53 training samples for Shakespeare, 90 training samples for Fletcher and 46 training samples for Massinger. Finally, Plecháč “let the algorithm loose” on Henry VIII and asked it to determine the author of the text.

The analysis revealed that the probability that the text of Henri VII was a collaboration between Shakespeare and Fletcher was very high, and extremely low for Massinger: for 7 scenes all the 30 models agreed upon Shakespeare’s authorship, for 5 scenes all the 30 models agreed upon Fletcher’s authorship (in other words, Fletcher may have written over 40% of the play). The algorithm also allowed a more fine-grained approach that revealed how the authorship sometimes changed not just for new scenes, but also towards the end of previous ones. 

It is interesting to note that the machine learning findings corroborate the conclusions drawn by literary analyst James Spedding in 1850 – long before the advent of machine learning.

 

Sources

“Machine learning has revealed exactly how much of a Shakespeare play was written by someone else”, MIT Technology Review, Nov 22, 2019 – https://bit.ly/2QaKVNl

“Computer analysis reveals Shakespeare’s collaborators”, Michael Irving, New Altlas, October 26, 2016 – https://bit.ly/2MK3PIP

“Relative contributions of Shakespeare and Fletcher in Henry VIII: An Analysis Based on Most Frequent Words and Most Frequent Rhythmic Patterns”, Petr Plecháč, https://arxiv.org/abs/1911.05652v1 – pdf version https://arxiv.org/pdf/1911.05652v1.pdf

Photo credit: Jessica Pamp @ Unsplash