nGrams, Craters, and 19th Century Sectarianism

Late last year Science published an article, "Quantitative Analysis of Culture Using Millions of Digitized Books," which introduced the term "culturomics" to the digital humanities community. The Google nGrams viewer, which the team behind culturnomics developed and released, allows users to track the frequency of words across time within the Google Books corpus. The resulting graphs can allow users to spot linguistic trends, perhaps noting larger cultural shifts that the use of particular words might signal. Though the term "culturomics" was new, the ideas of text-mining, analysis, and "distant reading" were not. The application of text mining to such a large corpus—according to the Science paper, 4 percent of all books ever printed—did spark much conversation in the digital humanities community. In "Counting on Google Books," Geoffrey Nunberg looked at the potential implications of culturomics to humanities research. He notes the skepticism of some scholars who worry that "second-rate scholars will use the Google Books corpus to churn out gigabytes of uninformative graphs and insignificant conclusions." While conceding that point, Nunberg dismisses more apocalyptic predictions about the influence of these ideas on the field. Instead, he suspects that "the data will probably figure in the literature the way observations about origins and etymology do now." In other words, Nunberg doesn't see the Google nGrams viewer as the dawn of a new science or destroyer of humanities worlds, but as a useful interpretive tool among others.

This is precisely what I found this week while revising a traditional article about James Fenimore Coopers 1847 novel, The Crater. In my article, I argue that the denominational infighting on Cooper's fictional island colony matched a heightened concern among contemporary religious commentators about denominationalism in the United States in the 1840s and 50s. Essentially, many American Protestants were worried that disunion among Protestant groups in the United States was endangering the nation's millennial destiny—how could the United States become the seat of Christ's kingdom on earth, they worried, when American Christians were constantly bickering about minute points of doctrine? When commentators wrote disparagingly about these debates, they often used the term "sectarian": relating such disputes to the minute concerns of "sects" and not, say, to the grand ideals of faith. In Cooper's novel, this inderdenominational bickering contributes to a more dramatic result: the colony's apocalyptic destruction by volcano (if that doesn't make you want to read some Cooper, nothing will).

To demonstate this nineteenth-century worry about sectarianism, I refer to a number of newspaper and magazine articles from religious publications in the 1840s and 50s—most of which I found in online archives such as The Making of America Collections at Michigan and Cornell, 19th Century U.S. Newspapers, and, of course, Google Books. This latter fact made me curious whether the nGrams viewer might provide some "distant reading" evidence to back up the claims I was making through close readings of these periodicals.

When I searched for the terms "sectarian" and "sectarianism" within the American English corpus between the year 1800 and 1900, the results proved very interesting. Between 1800 and 1850, use of both terms increased by an order of magnitude. When Cooper wrote The Crater in 1847, in other words, his readers were ten times more likely to encounter the word "sectarian" in their books and magazines than readers in 1800. Such a dramatic increase in the use of the word "sectarian" does indeed signal a widespread preoccupation with the issue. When correlated with the textual examples I'd already gathered, this data gives me a powerful piece of evidence about how Cooper's novel contributed to a much larger discourse about denominational discord in the 1840s. Just like an "observation...about origins or etymology," this data supplemented my interpretation, but it didn't create it out of whole cloth.

The nGrams Viewer alone says nothing about how these words were used in 1800 or 1850. That's where humanists come in, to make sense of these results. And that is precisely the point that Nunberg comes around to in his article: the nGrams Viewer can't entirely replace the work that humanists do. But this tool (and others like it) can be useful within a suite of interpretive tools, both technical and non-technical. I plan to continue experimenting, to see if this early success holds up in future research projects.