Academic Writing

Do We Really Need Data?

In his article, “The Bechdel Test and the Social Form of Character Networks,” professor Scott Selisker discusses the benefits of applying the Bechdel Test, in the form of network creation, to literary analysis. A professor of English at Arizona State University, Selisker situates his DH scholarship not only with the DH realm, but also within traditional literary studies; he puts the two in conversation with one another (as good literary DH does!). 

Selisker’s main argument is that applying the Bechdel Test to the formation of character networks can provide quantitative data that will help empirically legitimize — and thus impart political importance upon —  otherwise overlooked literary analyses that prove women and other minority groups lack recognition in literature beyond their flat, intermediary purposes. I found the argument that networks can help impart legitimacy upon critical, academic scholarship to be well… annoying — not because Selisker was wrong, but rather because he was (annoyingly) right. Why does there need to be empirical data in literary studies in order for it be a piece legitimate evidence for political action? Selisker states that rigorous, academic scholarship is often overlooked when it comes to be used as evidence in political situations because it lacks the ability to be quantified. He then goes on to argue that perhaps including visual, numeric data, like character networks, might make the (already present) evidence in literary studies more digestible and thus be a way to incite “larger reading publics to acknowledge the validity of literary scholarship” (Selisker 518). To an extent I do see where numbers might be more accessible to some audiences. There is no doubt that academic scholarship can be dense and downright hard to process, whereas as numbers and visualizations are less so. However, there should be no reason why that would, to an extent, invalidate literary, academic writing, or what I’d like to call here “wordy data.” 

My real qualm is this: If we (literary scholars) begin to turn away from close reading and to surface, distant reading in order to create these visualizations and compile quantitative data, to what extent are we then betraying our field and playing into the hegemony of tech? I’m going to return to this question, but first I want to talk about the academic scholars that Selisker invokes. 

In “The Bechdel Test and the Social Form of Character Networks” Selisker dialogues with feminist scholars Judith Butler and Eve Sedgwick, both of whom have published rigorous feminist scholarship on the representation of women in fiction. If you take one read of Butler’s “Gender Trouble,” you’ll notice a few things: 1) yes, it is difficult to digest, 2) it is artfully, extensively, historically, and politically researched, and 3) it could be used as evidence in political engagement and debate. Does it provide numerical data? No, but that does not, in any way shape or form, negate the strength of its argument. While the addition of data to Butler’s scholarship might make it more digestible to some, I’d argue it also takes away what makes it inherently impressive: It takes away that literary, academic, “wordy data.” 

The risk we, as literature scholars, run in shifting our mode of thought from reading to counting is a loss of what makes our scholarship our scholarship. In my opinion, if people wanted numbers, then they should’ve engaged with a mathematician, not a reader. I do not want to have to convert my scholarship into data in order for the public to want to engage with it. Why are numbers privileged over words? Why aren’t we asking how we can make these arguments more accessible in their native form rather than converting them into networks? I think it is in literature’s best interest to seriously debate the need for data in any argument before inputting it. I think, also, the more data we input in our scholarship, the more it will then be expected of us, and the more it will strip our words of agency. Our arguments will then become intermediaries between our data, and just as the Bechdel tests whether women talk about something other than men, we will have to be testing whether or not our scholarship talks about something other than our connected data (when it should be taking an active role — just as the women — and talking about the text.) 

Works Cited:

Butler, Judith. “Gender Trouble.” 2002, 

Scott Selisker, “The Bechdel Test and the Social Form of Character Networks (Links to an external site.),” New Literary History 46.3 (2015).

One Comment

  • Karalee

    Hey Emma. I think you made a lot of interesting points here. I particularly like your suggestion towards the end — that if we want to make our scholarship have a greater reach, we should find a way to do so without necessarily transforming our “wordy data” into numbers. From a personal angle, I generally like engaging with words more than visual data. I would seriously appreciate it if some (many) of the articles I deal with in literary discourse were more accessible…but I wouldn’t necessarily love it if they achieved that level of accessibility by being transformed into numeric data. I definitely think data visualization has its place in literary studies but more as a support to the text, rather than a replacement for it!

Leave a Reply

Your email address will not be published.