In Japan, the impulse to reason about literature quantitatively goes back at least to Natsume Sōseki’s Theory of Literature (1907). Recently, computational techniques and digital corpora have taken this impulse further, promising new ways of reading literary history. Before diving into this future, however, it is worth reflecting on its past. In this talk I trace a genealogy of quantitative imagining from Sōseki’s theories of reading, to psycholinguistics and stylistic analysis from mid-century, and up through recent applications of computational methods like machine learning. I consider why scholars have previously turned to numbers to reason about literature, what they have gained from it, and what it means to do so now.

Hoyt Long is Associate Professor of Japanese Literature at the University of Chicago. His research interests include sociology of literature, media history, and cultural analytics. He co-directs the Textual Optics Lab and is currently working on a project that tells the history of modern Japanese literature through a quantitative lens.