Belles lettres Meets Big Data
Quantitative analysis of poetry and prose has roots deep in the 19th century.
The Lit Lab
Sherman performed some prodigious feats of counting. One summer he set aside three weeks to tally the words in all 40,000+ sentences of Macaulay’s five-volume History of England. But he didn’t do all the counting by himself. After all, he was a professor. He had students!
Sherman presents his analytic method as a pedagogic tool as much as a research program. Looking around at other university departments, he applauds the transformation then underway in the teaching of physics, chemistry, and biology, where memorization and classroom recitation gave way first to lab-bench demonstrations by the lecturer and then to hands-on experiments by the students. In a similar way he aimed to make English a laboratory course, in which students would dissect poetry and prose to identify the vital organs.
Not all of his students greeted this innovation with enthusiasm. One of the skeptics was Willa Cather, a novelist-to-be who was already writing professionally when she was an undergraduate. She and Sherman dueled as rival columnists and critics for the Lincoln newspapers, and she wrote satiric verses about the analytics course for the campus literary magazine. Years later, Cather mockingly recalled her time in Sherman’s class as “trying to find the least common multiple of Hamlet and the greatest common divisor of Macbeth.”
Beyond Nebraska, Analytics of Literature did not go entirely unnoticed—a review in Science called it “epoch-making”—and yet it clearly failed in its mission to transform literary criticism into a laboratory science.
Today the book seems almost entirely forgotten. (I learned of it from Mark Liberman of the University of Pennsylvania, writing on the Language Log website.) Whereas Mendenhall’s work is still cited by statisticians, Sherman is seldom mentioned outside of a few specialized realms: Nebraska history, biographies of Willa Cather, and studies of “readability.”