TheDoxa of Reading andrew goldstone

Reading Graphs, Maps, Trees as a late-stage graduate student in 2008 was invigorating.1 Here was an approach to literary history free from the pieties of “close reading,” committed to empiricism, seeking to fulfill the promise of the sociology of literature with its “materialist conception of form” (92). And, at the time, it seemed entirely natural that the way to follow the path laid out by Franco Moretti in Graphs and in the essays he had published over the previous decade was to go to my computer, polish my rusty programming skills, and start making graphs. Yet reconsidering Distant Reading now, it is striking how non-digital the book is. In fact, the term “distant reading” has undergone a rapid semantic transformation. In “Conjectures on World Literature,” originally published in 2000, Moretti introduces the phrase to describe “a patchwork of other people’s research, without a single direct textual reading” (Distant 48). Today, however, “distant reading” typically refers to computational studies of text. Introducing a 2016 cluster of essays on “Text Analysis at Scale,” Matthew Gold and Lauren Klein speak of “using digital tools to ‘read’ large swaths of text” as “ ‘distant reading’ ” (Introduction); in his contribution there, Ted Underwood embraces “distant reading” as a name for applying machine learning techniques to unstructured text (“Distant Reading”). Discussions of distant reading have become discussions of computation with text, even if no section of Distant Reading features the elaborate computations found in the Stanford Literary Lab pamphlets to whichMoretti has contributed. The semantic shift takes about ten years. At first—judging from the 190odd results for the phrase in a JSTOR search in early 2017—invocations of the term are concentrated in debates about world literature (e.g., Mufti 487). But soon enough, computers enter the picture: in 2007, AnaMitrić published