The Sociolinguistic Significance of /ð/ Assimilation

Session 15 – Phonology-Consonants Sociolinguists have occasionally mentioned regressive assimilation of the voiced interdental fricative /ð/ to a preceding consonant as a feature of some vernacular dialects or at least " colloquial " speech. For example, Wolfram and Christian (1976) found tokens of an' nen for " and then " and all le for " all the. " However, there has been very little quantitative research on the topic and virtually no instrumental analysis, with the exception of Thomas's (2011a) study of durational patterns of assimilated /ð/ clusters. However, close examination reveals several types of assimilation: nasal (e.g. an' nen for " and then "), voicing (e.g. guess sat for " guess that "), manner (often indistinguishable from eth-stopping, e.g. had de for " had the "), and place (e.g. was ze for " was the "). In this study, I examine /ð/ assimilation and its connections with three sociocultural variables: sex, ethnicity, and geographic location. For sex and ethnicity, interviews with men and women from Robeson County, NC, were used with three ethnic groups: African American, Native American, and European American. For geographic comparison, I also used interviews with White Anglo and Hispanic men and women from Pearsall, TX. Tokens of phonemic /ð/ and preceding consonants were examined for each interview using consonant evaluation techniques described in Thomas (2011b) to determine the characteristics of each consonant. Post-consonantal tokens were coded for the social variables sex and ethnicity. Proportions of different variant tokens were compared between social categories. A logistic regression model was run to determine which social factors are significant in their effects on realization of /ð/. In order to compare rates of /ð/ assimilation in two separate Southern communities, the Pearsall and Robeson County data sets were also compared. Another logistic regression that included the Pearsall data in addition to the Robeson County data, using the independent variable of location, tested the significance of regional effects on the assimilation process. For Pearsall, another regression tested the significance of ethnicity within this community. Finally, European Americans in Robeson County were compared to those in Pearsall using a linear regression to determine the significance of location for this particular ethnic group. The results show that /ð/ assimilation is robust and widespread in the Southern U.S. It occurs among speakers from two distinct communities with varying frequency. Sex is a significant determiner of /ð/ assimilation, as men are more likely than …