How Baldwin's voice moved Cambridge: Activation contours, mimesis, and a computational approach to rhetoric's sensorium

Rhetoric scholars have long recognized that vocal performance transmits affect through what Hawhee (2015) calls rhetoric's "sensorium,” the sensory, embodied dimensions of persuasion that operate beyond semantic content, but methodological limitations have constrained empirical analysis of these patterns. This exploratory case study examines how Large Language Models might enable a new computational approach to rhetorical analysis by synchronizing continuous acoustic data (pitch, intensity, timing) with discrete data from stylistic schemes, making what Keith and Whittenberger-Keith (1988) termed the "notoriously hard to capture" musical dimension of public oratory more empirically tractable. We introduce a toolkit that integrates phonological, prosodic, acoustic, stylistic, and musical data layers, providing analytical access to what psychologist Daniel Stern calls activation contours (temporal patterns of intensity that create conditions for mimetic response). Applied to James Baldwin's 1965 speech at the Cambridge Union, when he debated William F. Buckley and received the first standing ovation in the Union's 150-year history, the toolkit reveals strategic patterning invisible to transcript-based analysis: the phrase "it comes as a great shock" repeats with controlled pitch consistency, creating rhythmic entrainment, while the pronoun "you" erupts with a 2836-Δcent pitch leap (nearly two and a half octaves) at the moment of identificatory rupture. These measurements suggest empirical mechanisms through which vocal performance may amplify rhetorical force. By enabling computational analysis of affect in speech performance, this approach extends rhetorical scholarship beyond symbolic/semantic content toward the temporal, embodied dimensions through which voice transmits energy or affective force.