Sociophonetic perception of African American English in Minnesota

Abstract

Although it can be authentically spoken by people who don't share their lineage, African American English, a variety of American English, is primarily spoken by the descendants of forced immigrants from Africa to North America. An assumption underlying most work on African American English (AAE) is that the variety is not subject to regional variation. Despite this assumption, some studies have found regional variation in AAE (Hinton and Pollock, 2000; Thomas, 2007). This variation is typically explained as assimilation toward or away from local varieties spoken by European Americans. Some studies have suggested that it assimilates with other dialects in less segregated areas or where blacks have greater access to educational opportunity (Hinton and Pollock, 2000). Other studies show that AAE speakers are less likely to produce mainstream regional variants and even less likely in cases of greater racial segregation (Labov and Harris 1986; Bailey, 2001.) This dissertation studies listeners' associations between regional variation and ethnicity. The study focuses on the influence of the regional features of Minnesota English on the perception of talker ethnicity. Hinton and Pollock (2000) begin their study of regional AAE phonology with the understanding that that the Midwest is less segregated than the south, and consider that this may imply that AAE in the Midwest is more likely to assimilate with regional European American varieties. Hence, we would predict that listeners in Minnesota would expect some tendency on the part of African Americans to use Minnesotan English (MNE) features, and hence said listeners would have little hesitation labeling speech containing Minnesotan variants as having been produced by European Americans even if it were produced by an African American. This study examined this topic with a perception experiment. Previous research has shown that listeners can ascertain a speaker's race from audio-only samples of content-neutral speech (Buck, 1968; Roberts, 1966; Walton and Orlikoff, 1994; Plichta, 2001; Thomas and Reaser, 2004). We examined listeners' judgments of the likelihood of particular speaker-listener comparisons. We paired the speech of African Americans and European Americans from Minnesota with pictures of African Americans and European Americans. We were particularly interested in whether listeners would be less likely to judge the speaker-picture pairs to be a match when the tokens contained variants that were characteristic of the 'mainstream' regional variety spoken in Minnesota, and the pictures were of African Americans. Listeners were more likely to rate actual matches between voice and face ethnicity as matches than they were to rate them as mismatches for male voices, but not for female ones. The unwillingness to rate voices produced by European Americans with local Minnesotan features as matches to African American faces suggests that listeners do not believe the local variant of AAE to incorporate Minnesota English features, at least for male speakers. Implications for models of sociophonetic perception and for studies of variation in AAE are discussed.