Sounds of Passion: Character Sound and Scene Segmentation in the 20th-Century US Romance Fiction

What sounds do literary characters produce, and what can these sounds reveal about fictional intimacy? This talk explores the sonic dimensions of desire in 20th-century US romance fiction, with particular attention to character sounds such as groans, gasps, and whispers. Combining close literary analysis with computational approaches, such as automated sound annotation and scene segmentation, it investigates how erotic intensity is expressed through vocalization in the Harlequin “Men Made in America” series.

Drawing on a corpus of 50 romance novels enriched with annotations for character sounds and their corresponding loudness levels, the analysis examines sound as both a narrative and emotional device. Do erotic scenes erupt in cries of pleasure, or are they shaped by quiet, agonized whispers? What do the frequency and loudness of vocalizations reveal about gendered performance and the shaping of intimacy in romance fiction?

The study applies methods of fictional sound recognition and classification to scene-segmented texts to assess whether erotic scenes can be computationally identified based on the distribution and intensity of character sounds, offering new insight into how sound contributes to the affective architecture of popular romance narratives.