This talk demonstrates how large-scale text analysis can provide new evidence for old questions in the cultural historiography of modern Japan and China.
Namely, the question of how exposure to Western intellectual and socio-economic systems gave rise to new forms of self-expression and new circuits of cultural exchange. We build on the idea that narrative form can be modeled by measurable linguistic traits, and that such models help to understand discursive relationality at scale. Beginning with the Japanese shishōsetsu (or I-novel), we test whether the genre exhibits a consistent narrative logic across hundreds of examples, and in comparison to other popular genres.
Specifically, we test a model of narrative form that classifies texts based on how much their lexical content shifts. Do they follow an arc of self conversion or crisis such that the language of the self changes over time? Or do the words remain the same, suggesting stasis or immobility? In part two, we test this model on modern Chinese fiction, in particular Romanticist fiction from the 1920s that was heavily influenced by developments in Japan. Many of these writers spent their formative years there at the height of the shishōsetsu.
Computation provides a way to study this complex cultural relation at the macroscale to see where Chinese writers did and did not align with their Japanese counterparts. More radically, it provides insight into the evolution and differentiation of archetypal narrative forms within and across national contexts.
IMPORTANT NOTE: Although focused on Chinese and Japanese cases, the analytical approaches examined here are valuable for scholars working across Asia, on all time periods.