This study proposes and tests loudness standard deviation (SD) of fictional sound events as an acoustically grounded proxy for detecting explicit content in romance fiction. Working with a subcorpus of novels from the Harlequin Men Made in America series, scenes were annotated for character and ambient sound with loudness levels. Additionally, the scenes were annotated on a ternary severity scale with two content advisory categories drawn from the PG-story taxonomy, SEX & NUDITY and VIOLENCE & SCARINESS (Tsai et al., 2024), and tested whether within-scene loudness SD of character and ambient sound correlates with either category. Loudness standard deviation analyses of character and ambient sounds in scenes featuring explicit content reveal that erotic scenes are acoustically marked by significantly higher variability in character-produced sounds, reflecting the dynamic range from whispered dialogue to vocalized arousal, while no significant correlation was found between high ambient sound loudness SD and scenes of elevated VIOLENCE & SCARINESS.
---
The paper makes three contributions: (1) it extends Guhr’s sound annotation framework to the US-English romance genre and to scene-level content analysis; (2) it proposes a scene-level explicit content assessment annotation schema for literary fiction; and (3) it provides preliminary empirical evidence that loudness standard deviation is a candidate acoustic signal for identifying erotic scenes in romance fiction, while the parallel hypothesis for violent and scary scenes remains inconclusive given the limited data. (369)
---
The corpus consists of 50 Harlequin romance novels from the Men Made in America (MMiA) series (1982–2002), each set in a different US-State and written exclusively by female authors. Despite their shared romance premise, the novels frequently blend detective fiction, Gothic suspense, and high-stakes drama, making the series a generically rich testing ground. (369)
---
in general romance fiction tend to provide fewer ambient sound representations, while focusing on soundful character interaction. (376)
Here's the abstract:
---
---
---