Reflection by Tara K Menon
Jonathan Farina’s Everyday Words examines the “external and visible character” (the style) of the work of a small set of canonical authors (Dickens, Eliot, Austen, Trollope) by examining the recurrent use of a series of common words or phrases: “turn” in Martin Chuzzlewit, “particular” and “general” in Jane Austen’s novels, “something” in Middlemarch among others. To strengthen his argument about the prevalence of these words and phrases, Farina uses quantitative findings. Statistics appear periodically throughout Everyday Words, for example: “variants of ‘turn’ turn up 310 times in Martin Chuzzlewit” (6); “attention occurs most frequently, including a high 69 times…in Pride and Prejudice”; “Charles Dickens’s prose contains an extraordinary number of ‘as ifs’: 411 in Dombey and Son, 393 in David Copperfield…” (94). This data is wonderfully suggestive.
Despite using some digital methods, however, Farina disavows Digital Humanities in general. DH, he suggests, “reduc[es] literary scholarship to data and graphs, as if these numerical and visual representations were the only respectable forms of knowledge.” This is, I think, a criticism of a straw man—proponents of DH approaches no longer, if they ever did, claim that graphs and statistics are the only respectable form of knowledge. He declares that he, on the other hand, “merely make[s] cursory use of searchable digital texts here to confirm the prevalence of words whose extraordinary frequency I noted the old-fashioned way, by slow reading” (xix). Yet, I wonder if Farina’s use of quantitative findings could have been greatly improved by leaning on some of the DH methods he otherwise dismisses. Given the raw data he uses throughout Everyday Words, and which I cite above, Farina doesn’t do enough to put these raw numbers in context. What percentage of the novel is 69 “attentions”? How much more frequently does it occur than other “everyday words” or any other words? Does “turn” appear more or less frequently in Martin Chuzzlewit than it does in other mid-century novels? Occasionally, Farina gestures to some of these answers. For example, in footnote 4 in Chapter 3, about the use of “as if” in Dickens’s fiction, he provides percentages in addition to raw totals and states his findings are statistically significant when compared other authors. By contextualizing all of his findings as he does here, Farina could greatly strengthen his quantitative claims, and therefore also reinforce his qualitative.
Fortunately, Farina’s argument doesn’t depend on the data—the study is replete with textual examples and with these he convinces us of the stakes of paying attention to these seemingly innocuous words that have so far gone unnoticed. Yet, his stated method—using data to “confirm” his “old-fashioned” slow/close readings—has tremendous potential. Digital tools can be invaluable resource for the study of the novel: they can furnish evidence to corroborate either long-held assumptions or newly made observations. By doing so, the data occasion considered explanations of what we might take for granted. Further, if we ask the right questions, statistics can do more than just uphold expectations, they can also subvert or surprise. By revealing unexpected information, data can prompt questions about, and renewed scrutiny of, what we might otherwise overlook rather than what we have already noticed. Quantitative findings can thus both redirect and supplement historically and narratologically attuned close readings, or, as Farina puts it, “technologically savvy data mining might serve interpretive ends” (xix). However, if literary critics choose to use quantitative methods we must use statistics appropriate to our questions. Here, we would do well to look to disciplines that have long used numbers as evidence or even to the models within our own such as “A Quantitative Literary History of 2,958 Nineteenth-Century British Novels: The Semantic Cohort Method” by Ryan Heuser and Long Le-Khac of the Stanford Literary Lab, which uses word frequencies with the necessary diligence. If we combine this statistical rigour (demonstrated not only by Literary Lab but also by many scholars who use digital methods such as Ted Underwood and Andrew Piper) with the complex close readings of the kind Farina performs so adeptly in Everyday Words, we can arguments that are authoritative and testable alike.