Page 38 - 2016Summer

Acoustics of Regionally Accented Speech
ral patterns are shaped by regional variation. For example, Jacewicz et al. (2010) found that southern speakers have a significantly slower articulation rate than the northerners and that this difference is maintained across the life span. An important question is whether durations of individual segments (vowels and consonants) are globally reduced in the North because of the faster speech tempo and globally lengthened in the South given that the southerners speak slower. The mounting evidence suggests that the correspon- dence between segmental timing and speech tempo is not as straightforward and that temporal relationships are more complex. For example, the temporal distinction between long and short vowels (such as in “dad” and “kid,” respec- tively) is manifested differently in different dialects irrespec- tive of dialect-specific speech tempo (Fridland et al., 2014; Clopper and Smiljanic, 2015). A complicating factor is a dialect-specific use of pauses so that the temporal properties of the pauses, such as their frequency and duration and the resulting prosodic phrasing, may have a differential effect on the duration of vowels and consonants across dialects. Much more research needs to be done to better understand how complex temporal relationships are shaped by regional variation.
Naturally, acoustics of regional accents extends to their re- ception. Sociolinguists have studied dialect identification
and intelligibility of individual dialects at least since the 1950s (Dickens and Sawyer, 1952), but it is the modern work in the perception of regional variation, notably by Clopper and Pisoni (2004), that introduced experimental rigor and methodological advancement. This work has examined the salience of acoustic information and listeners’ strategies in perceptual categorization of dialects. For example, it was shown that untrained listeners have an explicit awareness of distinctive features of AE dialects and that “army brats” who lived in several dialect regions categorize talker dialect more accurately than “homebodies” who lived in only one place (Clopper and Pisoni, 2004). Also, intelligibility of re- gional dialects under difficult listening conditions such as in a background noise can vary as a function of dialect and talker gender, although General American, the more “stan- dard” midwestern variety, seems to be more intelligible than other dialects (Clopper and Bradlow, 2008).
The Changing Demographics in the United States and Their Influence on Regional Variation
Media reports and folk perception have increasingly sug- gested that long-standing regional distinctions and many regional variants have been receding among younger people in favor of more General American forms. Thus, what is the
36 | Acoustics Today | Summer 2016
Figure 4. Acoustic details cueing regional dialect. Left: Spectrograms of the vowel in “bad” with formant tracks spoken by a female speaker from southeastern WI (top) and western North Carolina (NC; bottom). The first two formants, F1 and F2, are emphasized in green for WI and in blue for NC. Right: Mean F1 and F2 (from multiple repetitions) plotted at 20, 35, 50, 65, and 80%-time points. Arrows are at the 80% point and depict the differential direction of formant movement in each dialect.