University of Pennsylvania Working Papers in Linguistics


Conventional wisdom suggests that the dialects of American English are converging due to the influence of mass media and improved communication. Yet in fact, American regional dialects appear to be diverging from one another and these diverging regional dialects are expanding at the expense of smaller, distinctive speech islands within each region. One major, apparent exception to this pattern seems to be the American Midland, a region whose three largest urban centers ? St. Louis, Pittsburgh and Cincinnati ? have been shown to exhibit unique dialect patterns which distinguish them from the more general pattern of the surrounding region. But another major Midland center, Indianapolis, Indiana, has been largely ignored by previous research. This paper examines the state of Indianapolis English with regards to three key Midland identifiers: the fronting of back vowels /ow/, /uw/ and /aw/; the transitional merger of the low-back vowels /o/ and /oh/; and the monophthongization of /ay/ before resonants. The results of this study suggest that Indianapolis does indeed follow the Midland regional pattern. First, the Indianapolis speakers all show back vowel fronting. Second, while Indianapolis does not have an unconditioned low-back merger, all of the youngest speakers showed a merger in some environment (before /l/, before /n/ or both) and only half of the oldest speakers did. Finally, while there is no overall /ay/ monophthongization in any environment in Indianapolis, /ay/ glides show significant reduction before resonants as compared to non-resonants across all age groups. Thus, Indianapolis is a Midland speech prototype representing the target of convergence for the larger urban centers.