The thrill of discovery

With the help of computer analysis, a team of Wheaton scholars has uncovered a 1,200-year-old secret about an Old English poem.

English Professor Michael Drout’s summer research students discovered in June that the author of the 10th century poem “Christ III” created the work not out of whole cloth, as previously believed, but by dividing an older poem into two pieces and then inserting new material in the middle. This discovery led the scholars to further insights about the nature of the historic text.

“Using computer programs written by Wheaton students and techniques developed at Wheaton over the past four summers, the team was able to figure out what was sitting on an anonymous author’s desk over a millennium ago,” Drout said. “We determined that the unknown author had a written source, that it was in the form of poetry rather than prose, and that the source was already old when our author adapted it.”

The team reached these conclusions by using an original methodology they call Lexomics, which refers to the use of computer programs to conduct statistical analyses of text, employing cluster analysis and other techniques, and in a sense treating a written text as a genome. Wheaton students helped build the cutting-edge software that enables them to analyze the frequency, distribution and arrangement of words in patterns. The data they gather gives the scholars insight into factors such as where and when a piece was written, and what kinds of sources the author used.

Drout and Mark LeBlanc, professor of computer science, joined forces four years ago to find ways to apply computer analysis to literary research. The pair began to lead student researchers in the dissection of many ancient texts, including some of Shakespeare’s works as well as the epic poem Beowulf.

After the spring semester ended in May, the team started analyzing a variety of Anglo-Saxon poems. When studying long poems, the researchers divide the piece into segments before running their analyses—in the case of “Christ III,” five sections of 1,000 words each. Looking at the data, Drout and rising senior Elie Chauvet ’14 noticed an odd concentration of the alphabetical symbol called “eth” (which denoted the “th” sound) in the first and fifth segments. Other parts of the poem, however, favored the letter “thorn,” which denoted the same sound but was used more frequently at a later time period.

Chauvet, who is an English major, wrote a computer program that visually graphed the ratio of thorns to eths throughout the poem. The researchers concluded that parts of the piece had been written by someone else at a much earlier time and were borrowed by the “Christ III” poet.

“Our team now knows that the author of ‘Christ III’ had a text that doesn’t exist anymore, that was lost hundreds and hundreds of years ago. We know that it was on his desk, and that he put together this more complex poem based on this material,” Drout said.

Chauvet, now in his second summer of working on Drout’s team, aspires to become a professor of Medieval literature someday. “This research is not only good for my career plans,” Chauvet said, “it’s also just really fascinating.”

The Wheaton Lexomics team has published a number of scholarly articles on their research. Later this summer, Drout, LeBlanc and four students will travel to Dublin to present a hands-on Lexomics workshop to graduate students and elite Anglo-Saxon scholars at the Biennial Conference of the International Society of Anglo-Saxonists.
Professor Michael Drout explains the Lexomic analysis of Beowulf, while Natasha Piirainen ’14 works on generating a dendrogram—a tree diagram that illustrates similarities and differences between sections of a text.