Macroanalsys: Digital Methods & Literary History
Matthew L. Jockers
As a scholar of Irish literature, Jockers has a particular interest in how nationality and place influence theme and attitude. He can, for example, "look at the way that Ireland gets expressed in literature. And I can break that down further to how a male Irish author writes about Ireland compared to the way a female author writes about Ireland. … [Or] when writers have set their fiction in Rome and they're talking bout Catholicism, what is the attitude being expressed toward that, and how does that change when the setting is no longer Rome but Dublin, Ireland?"
He's adamant that macroanalysis does not endeavor, nor is it equipped, to identify quality in a work, only trends and similarity, such as the ways that themes and subject matter, whaling for instance, are propagated through the corpora. "Moby Dick has a lot in common with Edgar Allan Poe's The Narrative of Arthur Gordon Pym of Nantucket," says Jockers, for instance. "It seems likely that Melville borrowed some of his material from Poe."
Jockers is hopeful that a new partnership with Book Lamp, home of the Book Genome Project (think Pandora.com), will unlock the 20th Century corpus, and that work being done at the Stanford Literary Lab to make the 18th Century corpus available will enable him to extend the trend arcs that he has identified in the 19th Century.
"What if we could write an algorithm that could ID the major characters in a book?" asks Jockers almost giddily. "How many are there in 1809 on average vs. 1904? How are they portrayed? Is there a relationship between male and female characters and the author's gender?" The possibilities seem endless.