A problem of macrosociolinguistics:

Uniformity of the North vs. diversity of the North Midland

William Labov, Sharon Ash & Charles Boberg

Sat. 4:00-5:40 D

Sociolinguistics has made a certain amount of progress in determining the relative degrees of heterogeneity and homogeneity to be expected in a given speech community. While all urban communities studied so far show stratification by gender, socioeconomic class and sometimes age, it has been fairly well established that a metropolis of 1,000,000 or more will show a uniform structural base and a uniform set of social norms without geographic differentiation (with the additional proviso that in the United States, speech communities are defined within racial boundaries).

We are now engaged in a larger study of the distributions of phonological systems of English, through a telephone survey of North America. The sampling design of this study is sensitive to both population and area. It obtains data from 2 speakers from all urbanized areas over 50,000, with at least 4 from the largest areas of over 1,000,000. Our current results for the Northern and North Midland areas show remarkable patterns of homogeneity and heterogeneity that pose a problem for macrosociolinguistic explanation.

The largest perturbation in the phonetic realization of English vowels is the Northern Cities Shift, which rotates the five vowels /æ, e, [[radical]], ø, o/ in a chain shift triggered by the tensing and raising of /æ/ as a whole. Our present findings confirm the indications of exploratory work beginning with Labov, Yaeger and Steiner 1972 that this shift is correlated with the size of the speech community in a pattern that is geographically uniform throughout the entire Northern speech area from upper New York State to eastern Wisconsin. Geographic differences found within the NCS area have to do with which element is most advanced, not the direction or mechanism of the change.

This delineation of a homogeneous Northern area confirms the definition of the Northern area established by lexical and grammatical distributions, and simultaneously defines the northern limits of the North Midland area. But instead of a uniform North Midland pattern, we were surprised to discover a heterogeneous region where almost all the major speech communities show diverse phonological systems and directions of sound change. While the Northern Rochester, Syracuse, Buffalo, Cleveland, Detroit, Grand Rapids, Gary and Chicago all follow the same pattern, North Midland Philadelphia, Pittsburgh, Columbus, Cincinnati, Indianapolis, Evansville, Springfield and St. Louis are diverse communities that have little in common aside from their joint differentiation from the Northern and Southern chain shift patterns. In addition to diverse combinations of conditioned mergers, these cities show unconditioned mergers and reversals of long-standing phonetic shifts that are not found elsewhere.

The macrosociolinguistic problem is to account for this opposition of the larger Northern and North Midland patterns by settlement histories and modern patterns of influence and communication, as well as the internal logic of the linguistic changes involved. The present paper is designed to present and define the problem, and indicate the directions of research in which a solution may be found.