Shakespeare's Art Is Still Being Questioned, This Time by Machine Learning
Machine learning, a subset of artificial intelligence, is improving more and more every day to take over our roles and replace us in real life. It wouldn't be surprising to be treated by a robot doctor in the future or our kids to have a robot teacher or even robot classmates. But chill, we're not there yet.
We're witnessing simpler yet scary things done by machine learning and AI at the moment.
If you didn't grow up in a cave, you know who Williams Shakespeare and what he did. And if the last and the only book you read in your life wasn't the user manual of your TV, then you probably read at least one book of his. And surely, you're familiar with his writing style.
Until his death in 1616, Shakespeare worked for a playing company called King's Men. After his death, the company needed a replacement, and that's when John Fletcher became the new playwright for the company.
In 1850, James Spedding, an English author, and editor noticed a similarity between Fletcher's style in his plays and Shakespeare's Henry VIII. For example, Fletcher tended to use ye instead of you and 'em instead of them.
All that being noticed, Spedding and other literary analysts came to the conclusion of Fletcher's involvement in Henry VIII. But it wasn't known for sure which parts were written exactly by Fletcher.
And it's been one of the biggest concerns of the literature and history for the years; to find out which parts were written by whom in Henry VIII.
Petr Plecháč of the Czech Academy of Sciences in Prague came up with a solution for the problem by using machine learning. Plecháč said, "Our results highly support the canonical division of the play between William Shakespeare and John Fletcher proposed by James Spedding."
The machine uses the work of each writer to determine their writing styles; however, since an author's style can change throughout the years, all works tested by the machine should be written in a similar time. So, the machine used Shakespeare's other plays such as The Tragedy of Coriolanus, The Tragedy of Cymbeline, The Winter’s Tale, and The Tempest, written at the same time as Henry VIII.
Then the work of John Fletcher written at the same time was recognized by the machine, these works include Valentinian, Monsieur Thomas, The Woman’s Prize, and Bonduca
And in the end, the machine put forward the same conclusion with Spedding; that some of the scenes of the play were written by Fletcher.
Most of the change between authors occurs between the scenes; such as scenes 1.1 (Act 1, Scene 1) and 1.2 were written by Shakespeare while 1.3 was written by Fletcher.
According to the algorithm, the authors don't only change between the scenes, they also change towards the end of a scene. For example, scene 3.2 was written by both Shakespeare and Fletcher. After line 2081, there was mixed authorship and in line 2020, Shakespeare took over the rest of the scene.
This work shows that machine learning can be useful in many fields, and literary analysis is one of those fields. Surely in the future, there'll be more innovations regarding AI, but for now, we should be content (or scared) with what we have.