Penn engineers use big data to show Shakespeare had coauthor on 'Henry VI' plays

combined
Based on a statistical analysis of the text, Penn engineers say Christopher Marlowe, right, was a co-author of Shakespeare's Henry VI plays.

Editors of a new edition of Shakespeare's complete works have caused a tempest by listing Christopher Marlowe as co-author of the Henry VI plays, the trilogy perhaps best known for the "kill all the lawyers" line.

A key reason for the change was a study at the University of Pennsylvania — conducted not by lawyers, but engineers.

Alejandro Ribeiro, an associate professor of electrical and systems engineering, is no Shakespeare buff. But a few years ago he realized that his specialty, the study of networks such as those in wireless communications, could be applied to questions of literary authorship.

Ribeiro learned that Henry VI was among the works for which scholars thought Shakespeare might have had a co-author, so he and lab members Santiago Segarra and Mark Eisen tackled the question with the tools of big data.

Working with Shakespeare expert Gabriel Egan of De Montfort University in Leicester, England, they analyzed the proximity of certain target words in the playwright's works, developing a statistical fingerprint that could be compared with those of other authors from his era.

Word-adjacency networks produced these stylistic "fingerprints" for Shakespeare, left, and Marlowe, right. (University of Pennsylvania)

Their identification of Marlowe as co-author impressed Gary Taylor, the lead general editor of the New Oxford Shakespeare compendium. Two other research groups had arrived at the same conclusion using different analytical techniques, including one that analyzed phrases rather than words. The third study from Penn sealed the deal, he said.

"This seemed to us to be decisive, a sort of tipping point," said Taylor, an English professor at Florida State University.

Taylor and his fellow editors of the new edition, including Egan, announced their decision this week to list Marlowe as a co-author of Henry VI. They also credited co-authors for six other Shakespeare works, based on additional research.

The Penn study is to be published in December in Shakespeare Quarterly.

The effort began when Ribeiro, who came to Penn in 2008, was looking for different kinds of networks to study. Most networks, such as patterns of interactions among hundreds of people on social media, are symmetric, meaning that connections go both ways. Person A is a friend of person B, and vice versa.

Ribeiro sought out examples of asymmetric networks, ones in which the order of the relationship matters — such as one word following another in a text.

He happened to see the 2011 movie Anonymous, a fanciful tale in which Shakespeare's plays were written entirely by someone else. Mainstream Shakespeare scholars dismiss such theories as rubbish, but Ribeiro learned that there were plenty of instances in which scholars thought the playwright had a collaborator.

What, ho! A research idea was born.

Though Ribeiro was not a big Shakespeare fan, he was encouraged to pursue the topic by someone who adores the Bard: his wife, Gabriela, a computer scientist. (Their children are Miranda and Ariel, after characters in The Tempest, and Guillermo — William in Spanish, as the couple are from Uruguay.)

Ribeiro's early efforts at analyzing Shakespeare were enough to get the attention of Egan, who told him of the debate on Henry VI, and offered to collaborate. They were joined by Eisen, then a Penn undergraduate, and Segarra, then a Penn graduate student who is now a postdoctoral fellow at the Massachusetts Institute of Technology.

Marlowe, a noted playwright of the day and a reputed government spy, had been proposed as a co-author by various scholars for years. But the field sought more evidence.

By analyzing the relative proximity between pairs of target words, Ribeiro's team created "word-adjacency networks" for plays known to be written by Shakespeare and five other dramatists of his time, including Marlowe.

The researchers used more than 200 common "function" words, such as against, despite, therefore, and whereas. They chose not to analyze unusual vocabulary that might depend more on the subject matter of the play rather than the author's underlying style.

Then they compared the individual scenes of Henry VI with Shakespeare's fingerprint, and found multiple scenes that were written in a distinctly non-Shakespearean style.

Of the five other playwrights the team studied, the style of those mystery scenes was closest to Marlowe's, Ribeiro said.

"It is fascinating, just the simple fact that there is a clue left to something that happened 400 years ago," Ribeiro said.

Not everyone is persuaded. Columbia University professor James Shapiro, who has studied the authorship of plays attributed to Shakespeare, declined to comment on the Marlowe attribution until he could see the other evidence cited by the New Oxford editors.

In the Penn paper, the authors wrote that "the presence of Marlowe in these plays...is now undeniable."

In an interview, Ribeiro was more circumspect. He said he was very confident that Shakespeare had a co-author for Henry VI, but added that it theoretically could be some other playwright for whom the team did not have enough text to create a statistical fingerprint.

"If you tell me, `You have to make a call,' I will say it is Marlowe," Ribeiro said.

Engineers do not normally traffic in the study of Elizabethan literature, though Shakespeare gives a nod to the profession in a few instances, said Florida State's Taylor. Hamlet speaks of an engineer "hoist with his own petard," meaning he was blown up by his own bomb.

In Troilus and Cressida, there is a happier reference, in which Achilles is described as a "rare engineer."

Ribeiro is developing quite an appreciation for all of it.

The divide between the sciences and the humanities is even stronger in his native Uruguay than on many American campuses, said Ribeiro, noting that he was not required to take any writing classes in college.

"When I was young, I thought I was lucky," he said. "Now that I am older, I lean the other way. I like reading."