Any sufficiently advanced technology is indistinguishable from magic.

Arthur C. Clarke

Bridging together computational tools to analyze literary and cultural products, roughly speaking digital humanities, can be fraught for literary and cultural studies scholars trained in traditional methods such as close reading. On the one hand, using machine learning to analyze hundreds, perhaps thousands, of texts at once can seem magical. After all, it can often take years for one scholar to rigorously analyze one text or cultural phenomenon. The possibility of working with vast amounts of texts seems fantastic and fantastical.

On the other hand, most literary scholars and cultural critics are not trained in statistics and data science, important tools for the digital humanities (and working with big data generally). So, they are unfamiliar with the methods and the underlying concepts. Moreover, as Nan Z. Da has pointed out, for many scholars, the usefulness of such tools at best, isn’t clear, and at worst, seems like an empty exercise in either counting words or confirming what is already known about a particular field (601). Digitizing medieval texts or creating sophisticated computer models of ancient artifacts can certainly expose countless students and amateurs to cultural phenomena only a few people used to know about. But can digital tools really do the work of analysis (interpretation) that is the hallmark of humanities research?

Being a classically trained scholar, I attempted to use digital tools in a “proof of concept” project to try to become more familiar with these tools and, in so doing, challenge my own skepticism of distant reading. This attempt to uncover secrets in plain sight through distance reading aided by text mining tools have been mixed but generative. My immediate results, such as they are, are not necessarily promising, but they do not at all rule out the utility of computational methods. More importantly, the process of experimentation—a method that has traditionally been the territory of the physical and social sciences—has been, frankly, fascinating. It’s one thing to push at the boundaries of the meaning of one word or metaphor in Shakespeare or Cervantes, to engage in the cognitive pleasures of the Kantian sublime. It’s quite another to ask targeted questions of dozens of authors without having any idea of what the result might be. The moment prior to either discovery or dissolution is thrilling.

This proof of concept began as an exploration of the relationship between custom (social norms) and the law in 19th century U.S. fiction, but along with the findings, the project has allowed me to more distinctly appreciate how traditional literary training meshes (or not) with not only the use of computational tools but also thinking and approaching literature in radically different ways. In fact, noticing how the the ontologies and epistemologies upon which close reading and distant reading depend constitute a site of intellectual struggle has been a key metafinding. Indeed, much more than the findings, the insights I’ve gleaned have led me to a serious reconsideration, and at times reaffirmation, of method as one of the most important aspects of research. This confrontation with method is not new (neither for me nor other researchers in every field), but it suggests that too many of us go about doing (traditional) literary and cultural studies perhaps not questioning our ontological and epistemological assumptions as much and as often as we should.