A new algorithm from Penn Engineering implies that Shakespeare may not have written all of his plays.

Credit: Morgan Rees / The Daily Pennsylvanian

The debate over whether Shakespeare wrote all 38 of his plays has been ongoing for millennia — and Penn Engineering, of all things, just proved it’s unlikely he did.

In a recent study led by Alejandro Ribeiro, a Rosenbluth Associate Professor in the Department of Electrical and Systems Engineering, researchers sought to determine whether William Shakespeare actually wrote all of the plays that he is claimed to have written.

The team used computer algorithms to figure out whether the syntax and style of writing in various plays attributed to Shakespeare matched up with each other.

“We generated a network that matched up an author with his text based on the language he used,” Santiago Segarra said. Segarra is a graduate student at Massachusetts Institute of Technology, who was a member of Ribeiro’s research team.

The machine learning program that the Penn Engineering team used is extremely accurate in detecting differences in authorship.

“If we trained this system on a play by me and a play by [someone else], and then gave it another play written by one of us, it could tell which one wrote it 98 percent of the time,” Professor Ribeiro said.

Ribeiro’s research built upon previous findings about the use of language to determine authorship by actively trying to eliminate biases based on word choice. Segarra explained that the researchers looked at how the various plays used generic words, such as “who,” “we” and “at,” instead of using specific words like “university” or “person.” The frequency of these general words in writing allowed for more data to be collected on their usage in a wide array of contexts, thereby improving the program’s accuracy.

By examining the similarities and differences in the usage of these words in Shakespeare’s canon, the researchers found that there was a significant enough difference between some plays to determine that Shakespeare did not write all of them.

“I think our findings are a testament to the creative tactics we used,” said Mark Eisen, an Engineering graduate and doctoral candidate at Penn, who worked on the project.

Although Penn’s research team for this project did not announce a specific author who they can confirm to have written some of these works, the researchers strongly suspect that Christopher Marlowe played a large role in the writing of some of the plays. Marlowe, a household name among English professors, has been central to this Shakespearean debate for decades, and many scholars already believe that Marlowe wrote the three Henry VI plays attributed to Shakespeare. The language in these three plays is extremely similar to Marlowe’s works, as confirmed by these computer algorithms.

These findings have had a real impact. New Oxford Shakespeare Complete Works, the premier scholarly resource on all things Shakespeare, will give credit to Marlowe as a co-author for the Henry VI plays. Marlowe’s credit will likely play a huge role in modern Shakespearean studies.

Ribeiro began this project in 2012 with a general interest in improving authorship attribution. With the debate over Shakespeare’s authorship so prominent in academia, it was only natural that their research quickly drifted to focus on the English playwright. But since the program that the researchers created can work with all kinds of literature, there’s much more potential for this project, he said.

“We used very unique methods and came up with an interesting conclusion. Hopefully we can build upon our success in the future,” Segarra added.

All comments eligible for publication in Daily Pennsylvanian, Inc. publications.