In lieu of an abstract, here is a brief excerpt of the content:

  • On Not Already Knowing
  • J. D. Porter (bio)
Distant Horizons: Digital Evidence and Literary Change by Ted Underwood . Chicago: University of Chicago Press , 2019. 200 pp. $83.00 hardback, $27.50 cloth.

Distant Horizons collects about a decade of cutting-edge literary digital humanities (DH) work into a concise, accessible volume. The five chapters work equally well as standalone experiments or in the service of Ted Underwood's overarching argument that DH reconfigures our understanding of literary history. It's a clear must read for anyone working in literary DH (especially text mining) or in literary history more generally, and when it is inevitably added to dozens of syllabi, both students and teachers will find a lot to admire. Yet the most profound achievement of the book is its demonstration of a genuinely new kind of literary critical knowledge.

This is not quite how Underwood casts it, though. He focuses on two methodological interventions made possible by the DH approach. First, there is the familiar issue of scale—the capacity of DH to tackle thousands of texts spread across centuries of production. Underwood argues that this new scale of attention fundamentally changes our understanding of literary history, writing that "we have narrated literary history as a sequence of discrete movements and periods because chunks of that size are about as much of the past as a single person could remember and discuss at one time" (ix). The digital approach, he says, enables us to consider changes that are too long term, slow moving, or widely dispersed to have been visible [End Page 445] to traditional methods. Chapter 1, for instance, shows that fiction grew increasingly distinct from nonfiction (especially biography and autobiography) over the course of 1800–2000, mostly on the basis of a rise in concrete language. Underwood approaches the problem from multiple angles, always with approachable, clear technical explanations, and arrives at the conclusion that literary history contains a large-scale pattern of change that has so far gone unnoticed by scholars. Yet the use of a long timescale may not entirely capture Underwood's achievement. It is not clear that people really do struggle to think about patterns of subtle and complex change on a 200-year scale (quick: Was a dollar worth more in 1818 or 2018?), and literary histories in particular often extend far past that—think of Erich Auerbach's Mimesis (1946), which stretches from Homer to Woolf. DH handles scale well and often, but it has forebears in traditional methods.

The second major intervention Underwood emphasizes is modeling, which is a hot topic in the DH world. To him, a model "defines a relationship between variables" as a way to study those relationships rather than "isolated facts" (ix). He views modeling, especially predictive and perspectival models, as a way to connect social and textual evidence, to mediate between theories and measurements. In chapter 2, Underwood demonstrates the historical stability of genre categories by showing how well certain models predict genre membership for novels in different periods, and by using models trained in one period to examine novels from another. The proof is in the pudding—in this as in every chapter, the results are both persuasive and interesting, as when Underwood shows that detective fiction has been remarkably cohesive as a genre since (but not before) Poe's Dupin stories. But again, Underwood may actually be underselling how radical his methods are. The recent discussion of models has never quite explained what isn't a model. Take Mimesis again: Isn't Auerbach's distinction between Hebrew and Greek literature in essence a model of historical literary thought, a way of relating variables (textual, social, theoretical) in order to provide an explanatory apparatus for complex literary data? Don't most critical approaches do that on some level?

What distinguishes the new empirical models is not that they are models, but that they are empirical. This is implicit in Underwood's emphasis on the rigorous comparison that his models enable, where rigor stands in for the many detailed and repeatable operations that numeric data allow—we cannot add and subtract the Hebrew or Greek approach to interiority, but we can do that and more with...

pdf