Once More Beyond the Text

Abstract

Beyond CTRL+F: Text Mining Across the Disciplines Conference — Opening Keynote: In literary studies, scholars have spent the last half-century going beyond the text, insisting on the inextricability of textual meaning from the historical environment. Literary applications of text mining, by contrast, have tended to scant context in favor of scale. In recent "distant reading" work, the free-standing text returns through the back door of the corpus, which is "read" for intrinsic patterns. This process is often defended precisely for its "humanistic" quality. But if this is humanistic, so much the worse for humanism: it has limited prospects for reliably and validly relating texts to contexts. I'll suggest some alternative possibilities by retracing an alternative text-analysis tradition in the sociology of culture. Then I'll exemplify the challenges of both alternatives in my own quantitative work on the history of literary scholarship, taking as my case study the institutionalization of "Theory" from 1980 onwards. I show how the choice of source texts and the availability of metadata both enable and limit the ways in which a quantitative analysis can go, once more, beyond the text.