Editing in the Age of Automation

At a recent conference,¹ Gregory Crane entreated: "We need editors—lots of them." Crane, the digital classicist who heads the Perseus Project, regards editing as a foundational scholarly act, as "the sustained process of making primary sources intellectually accessible." Crane sees many values in editing—including democratizing access and scholarship and rejuvenating the humanities—and contends that a new wave of digital editing will centrally involve two tasks. "As we confront the challenge of editing billions of words," he notes, "we need editors who can apply automated methods." At the same time, "the real effort lies, or should lie, in thinking about every word in the text." On the surface these injunctions seem contradictory: how can editing staggering quantities of words depend upon hands-off automated methods and thinking about every word in the text? Crane gets around this with a sensible explanation: editions should become living creations composed of contributions from multiple editors, and editors should create the editions carefully with later automated study in mind. I find Crane's appeal for editors refreshing, much needed, and well supported. In this essay I look at two scholarly values, one new and one old, both of which Crane alludes to here, that complicate the work of digital editing: automation and authenticity. Together, these different aims limit the development of digital editions that do not conform to the aims of distant reading or of single-author study.

The landscape of digital humanities projects is populated with well-developed examples of tools, digital libraries and mass digitization projects, and archives focused on single authors. There are not as many examples of digital projects that study multiple authors and that afford such texts the kind of attention that most single-author archives do. Methods for building and using digital texts seem today to be fairly polarized: on the one had we can invoke distant-reading methods, typically by searching or analyzing large quantities of texts; on the other hand we can pore over deeply edited collections of (usually) canonical authors. These methods work well for many users, but are not optimal for scholars who are interested in what I might call "medium-distance" scholarship—scholarship that studies multiple texts or even multiple authors, though not thousands or even hundreds of them, but still looks at textual or historical trends that are constrained compared to mass digitization projects. For medium-distance interests, the scope of single-author archives is too narrow, and much of their rich encoding is lost upon aggregation. The scope of mass digitization projects is too vast, and frequently lacks the metadata needed to find relevant resources. The literary scholar must often go digital dumpster diving² in hopes of finding relevant subsets of texts from among the millions offered by mass digitization projects—projects that are rarely primed for literary research. Many of us want to perform research on modest sets of text that are not necessarily representative of large-scale trends in the publishing industry—we are not looking at only novels or all novels, at only periodicals or all periodicals, and so on—which is what these tools seem to "recommend," to use John Unsworth's term.

In this paper I want to suggest that developed archival methods and mass digitization methods are not optimal for much of literary historical research because they are premised on single-author scholarship in the former case and an overly wide and long historical view in the latter. In other words, the field of digital scholarship seems polarized between two scholarly values: authenticity and automation, neither of which attunes us to the concerns of medium-distance scholarship, which examines textual phenomena occurring anywhere on a spectrum between the works of a single author and a large corpus. A digital project that I co-edit provides a concrete example of a project attempting to perform medium-distance work. The goals we had when we began this project—which is still tellingly incomplete and as beset by difficulties as it is encouraging—are ones that I hope will seem analogous in their breadth and approach to what many literary scholars are interested...