How Not to Introduce Blues Prosody:

This article delivers a two-pronged intervention into blues prosody. First, it argues that scholars have repeatedly misidentified the metrical organization of blues poems by Langston Hughes and Sterling Brown. The dominant approach to these poems has sought to explain their rhythms with models of alternating stress, including both classical foot prosody and the beat prosody of Derek Attridge. The article shows that the systematic organization of blues structures originates in West African call-and-response patterning (not alternating stress), and is better explained by models of syntax and musical phrasing. Second, it argues that these misclassifications — far from being esoteric matters of taxonomy — lie at the heart of African American aesthetics and identity politics in the 1920s and 1930s. Whereas literary blues verse has long been oversimplified with conventional metrics like “free verse,” “accentual verse,” and “iambic pentameter,” the article suggests that its rhythms arise instead from a rich and complex vernacular style that cannot be explained by the constraints of Anglo-American versification.

The effect of emotion on voice production and speech acoustics

The study of emotional expression in the voice has typically relied on acted portrayals of emotions, with the majority of studies focussing on the perception of emotion in such portrayals. The acoustic characteristics of natural, often involuntary encoding of emotion in the voice, and the mechanisms responsible for such vocal modulation, have received little attention from researchers. The small number of studies on natural or induced emotional speech have failed to identify acoustic patterns specific to different emotions. Instead, most acoustic changes measured have been explainable as resulting from the level of physiological arousal characteristic of different emotions. Thus measurements of the acoustic properties of angry, happy and fearful speech have been similar, corresponding to their similar elevated arousal levels. An opposing view, the most elaborate description of which was given by Scherer (1986), is that emotions affect the acoustic characteristics of speech along a number of dimensions, not only arousal. The lack of empirical data supporting such a theory has been blamed on the lack of sophistication of acoustic analyses in the little research that has been done.By inducing real emotional states in the laboratory, using a variety of computer administered induction methods, this thesis aimed to test the two opposing accounts of how emotion affects the voice. The induction methods were designed to manipulate some of the principal dimensions along which, according to multidimensional theories, emotional speech is expected to vary. A set of acoustic parameters selected to capture temporal, fundamental frequency (F0), intensity and spectral vocal characteristics of the voice was extracted from speech recordings. In addition, electroglottal and physiological measurements were made in parallel with speech recordings, in an effort to determine the mechanisms underlying the measured acoustic changes.The results indicate that a single arousal dimension cannot adequately describe a range of emotional vocal changes, and lend weight to a theory of multidimensional emotional response patterning as suggested by Scherer and others. The correlations between physiological and acoustic measures, although small, indicate that variations in sympathetic autonomic arousal do correspond to changes to F0 level and vocal fold dynamics as indicated by electroglottography. Changes to spectral properties, speech fluency, and F0 dynamics, however, can not be fully explained in terms of sympathetic arousal, and are probably related as well to cognitive processes involved in speech planning.