The changing social and linguistic orientation of the African American middle class.

In this dissertation, I quantify the use of five phonological features of African American English (hereafter, AAE) by African Americans in Detroit, including middle-class and female speakers, who are often overlooked in analyses of AAE. I also include both contemporary and 1966 recordings, which allows me to examine speakers' changing linguistic orientation. Results from this project address the heterogeneity of AAE, challenging traditional assumptions about language variation and posing questions for aspects of the divergence hypothesis. My results show that African American speakers cannot be labeled as simply being deficient or inauthentic speakers of two opposing varieties; in fact, middle-class African American speakers create a unique linguistic position. In this analysis, I include twenty-four speakers, twelve each from the 1966 and contemporary recordings. Each sample is equally divided by gender. Further, each speaker is assigned a score on a social status index, and speakers with a wide range of social status scores are included here. I analyze each speaker's use of non-initial (r), syllable-final (d), (ai) before voiced consonants, (ai) before voiceless consonants, and (upsilon). Results of these analyses show that two variables, both contexts of (ai), have remained stable over time; the amount of glide reduction has not changed between the 1966 and contemporary recordings. Both consonantal variables, (r) and (d), have changed significantly over time, but in opposite directions. The use of [r]-less variant has decreased dramatically, while the use of multiple variants for (d) has rapidly increased. Finally, the fronting of /upsilon/ has increased over time as well, but these changes have been closely tied to following context in phonetically predictable ways. The cross-time comparison in this analysis allows me to compare my real-time results with age-stratified results from 1966 in order to assess the accuracy of predictions based on apparent-time evidence. Further, by combining results for variables' use over time and the social embedding of the variables in each recording period, I trace changes in variables' social embedding and conclude that such an examination provides vital evidence for understanding how changes are instantiated among groups of speakers.Ph.D.American studiesBlack studiesLanguage, Literature and LinguisticsLinguisticsSocial SciencesSocial structureUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/125725/2/3208524.pd