A Reader in Sociophonetics

Identi¿ cation of African American Speech 271

the accuracy by which they were identi¿ ed. In some cases, control stimuli

were identi¿ ed more accurately than corresponding /æ/ or /o/ stimuli. This

result seems to contradict the notion that vowel quality is a crucial cue, and

suggests that other cues not controlled for, such as timing, amplitude, or cer-

tain features of voice quality, may play important roles. The similarity in

accuracy levels for /æ/ and /o/ stimuli makes it dif¿ cult to determine which of

those two vowels is a better cue.

Table 12.2 Accuracy Levels by Sentence Type

Treatment

North Carolina

African American

listeners (n=11)

North Carolina

European American

listeners (n=33)

West Vi rg i n ia

European American

listeners (n=39)

male

speakers

female

speakers

male

speakers

female

speakers

male

speakers

female

speakers

unmodi¿ ed /æ/ 100% 96.6% 98.9% 93.9% 84.1% 76.8%

unmodi¿ ed /o/ 92.0% 96.6% 96.2% 93.9% 85.7% 83.2%

unmodi¿ e d c o nt r ol 98.9 % 93. 2% 97. 3% 96.6% 87. 2% 82 .0 %

monotonized /æ/ 94.3% 94.3% 96.2% 91.7% 81.4% 82.3%

monotonized /o/ 86.4% 92.0% 92.0% 90.5% 79.3% 72.3%

monotonized control 93.2% 80.7% 93.9% 90.9% 76.2% 76.8%

schwa- converted /æ/ 84.1% 63.6% 84.9% 68.2% 79.3% 63.7%

schwa- converted /o/ 70.5% 61.4% 76.0% 65.3% 68.9% 54.9%

schwa- converted

control

79.5% 62.0% 84.9% 59.4% 72.9% 61.9%

Several phonetic measurements were obtained on the stimuli using the

Praat acoustic analysis software in order to facilitate statistical analysis. The

mean F0 for each stimulus was measured because some previous studies

have reported that African Americans show a lower overall F0 than Euro-

pean Americans (Hollien and Malcik 1962; Hudson and Holbrook 1981,

1982; Hawkins 1993; Walton and Orlikoff 1994). F0 was determined using

the autocorrelation method described in Boersma (1993). The maximum

F0 and the standard deviation of F0 were measured for each stimulus in

an attempt to gauge intonational differences. Some European Americans,

mainly females, show a wide range of F0 values within an utterance, largely

because they produce an especially high peak, usually near the beginning of

the utterance. Other studies, however, have found that African Americans

show a wider pitch range (Tarone,1973; Loman 1975; Hudson and Holbrook