Research

Current Projects

Homunculus: A Cultural History

The homunculus has proven a useful fiction for contemporary cognitive science, personifying the many unconscious subsystems that contribute to functions like visual and auditory processing. This project tracks the history of the homunculus from its origins in seventeenth-century embryology, arguing that the homunculus of today’s neuroscience descends from fictional representations of the figure in the nineteenth and early twentieth centuries. Part history of science, part cognitive literary studies, the project’s readings focus on two case studies in what I call “homuncular form”: the second part of Goethe’s Faust, and Proust’s In Search of Lost Time.

Representations of Race and Ethnicity in American Fiction, 1789-1964
with Mark Algee-Hewitt and J.D. Porter

Using quantitative textual analysis, this project attempts to reconstruct the evolution of racial discourse in U.S. fiction, paying particular attention to the influence of immigration policy, demographics, and other sociological factors on the marked or unmarked status of characters. After defining the “background” racial discourse of the period by identifying significant collocates of racial and ethnic terms, the project turns to race as a component of literary character, using a combination of Named Entity Recognition and collocate analysis to sketch out the distinctive descriptive language associated with specific characters. Comparing the two registers allows us to see the differential operation of race as a system of generalization and as a mechanism of individuation at the level of character.

Suspense: Language, Narrative, Affect
with Mark Algee-Hewitt, Chelsea Davis, Abigail Droge, Tasha Eccles, Morgan Frank, J.D. Porter, and Andrew Shephard

The fruit of years of experimentation, this book project offers both an empirically grounded theory of narrative suspense and a series of historically specific applications of that theory. Drawing upon a wide range of quantitative techniques, from topic modeling to the extraction of most distinctive words, our collaborative research culminated with the generation of a neural net capable of automatically tagging passages of high and low suspense in novels. Our joint analysis frames suspense as a structure built from semantic components and occurring episodically in texts; my chapter examines one of our most unexpected findings, the discovery that “unsuspense” is itself a well-defined semantic field rather than a simple absence.