Over the last 50 years, a great deal of sociolinguistic research has focused on phonetic/phonological variables in African American Language (AAL), with postvocalic r-lessness often treated as one of the canonical features of AAL. Using data from the Corpus of Regional African American Language (CORAAL), Support Vector Machines (SVMs) were used to assess how vowel-like or/r/-like postvocalic/r/tokens are in CORAAL. The decision value, an output from the SVM, is used to examine postvocalic/r/-lessness in terms of social class and gender. Findings indicate that this r-lessness shows no statistically significant differences for gender in CORAAL but does pattern significantly differently across social class groups. This methodology gives continuous values for a feature that scholars know is gradient, giving researchers a tool to approach postvocalic r-lessness in a more time efficient and replicable manner, while also providing insight into the distribution of a classic sociolinguistic feature in Washington, D.C., AAL.

The text of this article is only available as a PDF.
You do not currently have access to this content.