About halfway through the Chicago Colloquium on Digital Humanities and Computer Science last week, the always witty and insightful Martin Mueller humorously interjected: “I will go away from this conference with the knowledge that intelligence analysts and literary scholars are exactly the same.” As the chuckles from the audience died down, the core truth of the joke settled in—for those interested in advancing the still-nascent field of the digital humanities, are academic researchers indeed becoming clones of intelligence analysts by picking up the latter’s digital tools? What exactly is the difference between an intelligence analyst and a scholar who is scanning, sorting, and aggregating information from massive electronic corpora?
Mueller’s remark prods those of us exploring the frontiers of the digital humanities to do a better job describing how our pursuit differs from other fields making use of similar computational means. A good start would be to highlight that while the intelligence analyst sifts through mountains of data looking for patterns, anomalies, and connections that might be (in the euphemistic argot of the military) “actionable” (when policy makers piece together bits of intelligence and decide to take action), the digital humanities scholar should be looking for patterns, anomalies, and connections that strengthen or weaken existing theories in their field, or produce new theories. In other words, we not only uncover evidence, but come to overarching conclusions and make value judgments; we are at once the FBI, the district attorney, the judge, and the jury. (Perhaps the “National Intelligence Estimates” that are the highest form of synthesis in the intelligence community come closest to what academics do.)
The gentle criticism I gave to the Chicago audience at the end of the colloquium was that too many presentations seemed one (important) piece away from completing this interpretive whole. Through extraordinary guile, a series of panelists showed how digital methods can determine the gender of Shakespeare’s interlocutors, show more clearly the repetition of key phrases in Gertrude Stein’s prose, or more clearly map the ideology and interactions of FDR’s advisors during and after Pearl Harbor. But of course the real questions that need to be answered—answers that will make other humanities scholars stand up and take notice of digital methods—are, of course, how the identification of gender reshapes (or reinforces) our views of Shakespeare’s plays, how the use of repetition changes our perspectives on Gertrude Stein’s writings, or how a better understanding of presidential advisors alters our historical narrative of America’s entry into the second World War.
In Chicago, I tried to give this critical, final moment of insight reached through digital means a name—the “John Snow moment”—in honor of the Victorian pharmacist who discovered the cause of cholera by using a novel research tool unfamiliar to traditional medical science. Rather than looking at symptoms or other patient information on a case-by-case basis as a cholera outbreak killed and sickened hundreds of people in London in 1854, Snow instead mapped all incidences of the disease by the street addresses of the patients, thus quickly discovering that the cases clustered around a Soho water pump. The city council removed the water pump’s handle, quickly curtailing the disease and inaugurating a new era of epidemiology. Snow proved that cholera was a waterborne disease. Now that’s actionable intelligence.
What can digital scholars do to reach this level of insight? A key first step, reinforced by my experience in Chicago, is that academics interested in the power of computational methods must work to forge tools that satisfy their interpretive needs rather than simply accepting the tools that are currently available from other domains of knowledge, like intelligence. Ostensibly the Chicago Colloquium was about bringing together computer scientists and humanities scholars to see how we might learn from each other and enable new forms of research in an age of millions of digitized books. But as I noted in my remarks on the closing panel, too often this interaction seemed like a one-way street, with humanities scholars applying existing computer science tools rather than engaging the computer scientists (or programming themselves) to create new tools that would be better suited to their own needs. Hopefully such new tools will lead to more John Snow moments in the humanities in the near future.