Shakespeare and his co-authors, as told by Penn engineers

Four hundred years after the death of dramatist William Shakespeare, enduring questions remain about whether the Bard of Avon had an uncredited co-writer on some of his world-famous plays. A team of Penn researchers has found an answer—in the School of Engineering and Applied Science, of all places.

Shakespeare and His Co-Authors, as Told by Penn Engineers

Four hundred years after the death of dramatist William Shakespeare, enduring questions remain about whether the Bard of Avon had an uncredited co-writer on some of his world-famous plays.

A team of Penn researchers has found an answer—in the School of Engineering and Applied Science, of all places.

Alejandro Ribeiro, the Rosenbluth Associate Professor in Penn Engineering’s Department of Electrical and Systems Engineering, along with lab members and network engineers Santiago Segarra and Mark Eisen, and Gabriel Egan, a professor of Shakespeare studies at De Montfort University in the United Kingdom, created an algorithm for author attribution that determined that Shakespeare did, in fact, have a co-author or collaborators on many of his plays.

“There is evidence that about a third of plays that Shakespeare wrote contain pieces that were not written by him,” Ribeiro says.

Research by Penn engineers has found that William Shakespeare was not the only author of the three “Henry VI” plays, which were most likely a collaboration with Christopher Marlowe.
Research by Penn engineers has found that William Shakespeare was not the only author of the three “Henry VI” plays, which were most likely a collaboration with Christopher Marlowe.
 

To reach this landmark conclusion, Ribeiro and his research team used “word adjacency networks,” a way of analyzing a piece of text to determine unique, hidden characteristics of its writer or writers. With enough text samples, the networks can be used to assign authorship to a text with an unknown author.

The word adjacency networks are based on function words such as “to,” “and,” “or,” “one,” and “the”—words writers must use to construct a sentence. The researchers’ algorithm counts word adjacencies, or how often and how closely different sets of function words appear to each other in a text. Since all writers must use these words no matter what they are writing about, they are a consistent way of comparing different authors’ writing styles.

“What we find is that actually, there is a lot of nuance in these variations of how these sort of sentences come out, and they’re very particular to an author,” Eisen says. “These can be used as sort of a stylistic fingerprint because the ways authors construct their texts is very unique.”

For a particular author, the researchers can take a set of works that they know were written by the author and combine all of the word adjacency networks for each text into one network, which they call the author profile.

Ribeiro and colleagues built word adjacency networks for all of Shakespeare’s plays, and combined them to create a Shakespeare profile. They fashioned similar profiles of the works of Shakespeare’s contemporaries, including Christopher Marlowe, Thomas Middleton, Ben Jonson, George Peele, and John Fletcher. The researchers then combined all of the texts from every candidate into a single profile to create, more or less, an average fingerprint for English-language authors of the Shakespearian era.

word adjacency networks
To reach their landmark conclusion, Ribeiro and his research team used “word adjacency networks,” a way of analyzing a piece of text to determine unique, hidden characteristics of its writer or writers.
 

Analysis of Shakespeare’s author profile suggests that he was not the only author of three “Henry VI” plays, which were most likely a collaboration between Shakespeare and Marlowe or Peele.

Egan says the evidence points to Marlowe, and credits him with penning the famous line, “First thing we do, let’s kill all the lawyers” in “Henry VI, Part 2,” along with other scenes featuring the character Jack Cade.

What we find is that actually, there is a lot of nuance in these variations of how these sort of sentences come out, and they’re very particular to an author. These can be used as sort of a stylistic fingerprint because the ways authors construct their texts is very unique.


PhD Student Mark Eisen, Department of Electrical & Systems Engineering

Ribeiro says the case of “Henry VI” and Marlowe is particularly important because it is one of Shakespeare’s earliest plays and Marlowe was one of the most established writers of his time prior to Shakespeare.

“This collaboration could be evidence of Shakespeare at least looking up to Marlowe,” he says. “We know now that Marlowe played an important role in Shakespeare becoming Shakespeare.”

The discoveries of Ribeiro and his team will be published in the journal Shakespeare Quarterly. Partly because of their research, the “New Oxford Shakespeare Complete Works,” one of the most authoritative scholarly resources on the playwright, will identify Marlowe as Shakespeare’s co-author for the three “Henry VI” plays.

“Having this knowledge will change how people learn about Shakespeare because we, as a species, are changing our perception of how Shakespeare himself worked,” Ribeiro says. “What we have discovered is that the environment of his time was much more collaborative than we thought it was.”