The intended purpose of this blog is to provide a space for the other fellows and I to relate our research interests to a larger audience. My research currently entails applying computer science methods to investigate the broad literary trends of what we now call passive voice. Many of the terms in that statement merit some unpacking and my hope for this introductory post is to anecdotally parse some of those terms.
There are many fabulous blogs detailing the various text mining approaches to digitized literary collections (Alan Liu, Matthew Jockers, Ted Underwood, Steve Ramsay, too many to list here). They all go to wonderful lengths to detail some of the advantages and problems that come with digital text analysis, so I will refrain from being redundant here. It might be sufficient to say that literary scholars are very gifted interpreters when operating on the scale of a single, tens, or a hundred texts, but have yet to formulate a set of questions to properly investigate digital collections that exceed this scale.
This is not to say that projects operating at the former scale of literary analysis are somehow not doing enough. Some of the most cogent literary interventions result from noticing subtle applications of a stylistic trope. A literary scholar might, for example, take note of how castles are depicted from Ann Radcliffe’s Mysteries of Udolpho (1798) to Jane Austen’s Northanger Abbey (1817). They then might derive a great deal of insight regarding Britain’s changing perceptions of property from the late eighteenth to early nineteenth century.
That sort of historical insight is possible due to the kind of interpretive questions that emerge when reading two texts for stylistic contrast. One might, and rightly so, raise questions such as, “What is the significance of Radcliffe’s crumbling abbey as the setting of the protagonist’s cultivation? Furthermore, what is the historical significance given that Austen’s novel contrastingly depicts a satirically modest and decidedly boring estate?” Literary studies, as a discipline, is familiar with asking questions on this scale in order to accentuate various humanistic subtleties.
But if one were to ask me the historically sensible follow up questions, “How long did Austen’s depiction of castles remain in vogue? Did Radcliffe’s depictions of Gothic castles remain more prominent in popular fiction despite Austen’s satire? Which setting-style does contemporary fiction inherit and when does it start doing that?” I would have given some vague hypotheses because I do not have a good answer for these kinds of questions.
Answering these sorts of historical questions is generally a challenge for literary scholars to answer. It would require reading the hundreds of thousands of fictional works that comprise the particular literary trajectory (generic, historical, gendered, etc.) before feeling a remote sense of hermeneutic confidence. An amount of fiction worth several lifetimes of reading. These kinds of questions have, for quite some time now, been considered trivially unapproachable and so left, on paper at least, unanswered.
I suspect though that literary scholars have operated on internalized and untested heuristics to these sorts of questions. It is worth asking how that internalization is detrimental to the discipline’s practices and historical formulations, but that will be pursued in a later post. What I wish to express is simply that the discipline has a fundamental, yet fruitful, problem in approaching a set of historical problems precisely because they operate at a scale too robust for close reading.
Furthermore, I would like to posit that text mining methods have already provided a productive avenue worth pursuing. Because digitized collections, such as HathiTrust, provide access to approximately 700,000 eighteenth and nineteenth century volumes, there is an opportunity to render those broader social-linguistic trends while acknowledging the gaps in that corpus. This still requires formulating the right set of questions and then further requires developing programming procedures to extract literary trends, but, at the very least, it is a place to start.
This is only half the story of my research interests though. The latter portion largely concerns the string of events and curiosities that lead to my interest in the passive voice. My interest in this style of writing stems from a curiously sacred notion, though perhaps one undergoing a hallowed burial, that sentences written in passive voice are stylistically gauche. This can largely be seen in sentences that take the object of the sentence and position it as the subject. (E.g. Active: “Tom painted the house.” Passive: “The house was painted by Tom.”)
Users of early word processing applications may recall the ambiguous green underlining and being advised “Passive Voice (consider revising)”. It is not categorically wrong per se, otherwise it would be underlined in red with the typos, but there is something potentially off. The degree to which one is off is largely a rhetorical problem. In a bureaucratic situation, passive voice might be taken as a very useful tool for being vague about who is responsible.
“Mistakes were made.” (Yes, but by who?)
This attempt to defer blame is commonplace, but you will hardly find someone refer to it as a stylistic faux pas. However, as I read over my seminar papers and see “p.v.” in the margins, I am reminded of my own tendency for verbiage, become preoccupied with the “assertiveness” of my writing, and consequently considered revising. On a pedagogical register, this annotation makes sense. It is much easier for instructors to assess whether a student derived some literary sensitivities from a lecture or seminar if they express them through precise interventions.Yet, as I look at both the passive and active versions of my sentences, it is unclear about how such language structures become so concretely configured with a sense of agency.
It would be shortsighted to label this as solely an issue of clarity or efficiency. Indulging in “Famous Movie Quotes Ruined by the Passive Voice”, each quote makes sense, but clearly there is something playfully awkward about each of these examples. It is out of this awkwardness that I began to wonder when and how passive voice got such a bad rap.
Having read my fair share of Sherlock Holmes, I have to wonder if this was always the case. Holmes is, after all, a detective constantly guided by passive objects — I mean clues. People do not help Holmes figure out the puzzle behind the mystery. “No, your boy had been observed, and that gave me a guide where to look.” (The Hound of the Baskervilles). Literary scholars are not so far removed from these statements, yet somehow we have inherited this notion that this kind of writing counts as “passive” of all things.
In a discipline where we regularly study verbose and passive texts, how do stylistic values form in such a way where passivity is connoted with verboseness and assertiveness with the pithy? Furthermore, can we track a broad historical trajectory that leads to these stylistic values? If so, what historical conditions motivated such preferences?