ABSTRACT: How Southern is it, really?: What Linguistic Atlas Project data tells us about Southern English features

Dialect map-making is a “theory-driven enterprise” (Kretzschmar 2003, p. 130) and when Hans Kurath created his well-known map, published for the (1949) A Word Geography of the Eastern United States, he used not just data from the Linguistic Atlas Project’s two earliest regional surveys, New England (LANE) and the Middle and South Atlantic States (LAMSAS), but also his knowledge of colonial settlement patterns and migration routes to solidify what concurrently running isoglosses suggested in terms of where the dialect boundary lines should be drawn. Based on lexical data, Kurath’s map demarcates three main dialect regions of the eastern US: the North, the Midland, and the South and these demarcations have (more or less) withstood the test of time, even as more recent maps have used additional data and data of different kinds, as in Carver (1987), Labov, Ash and Boberg (2006), and Yuan, Guo, Kasakoff and Grieve (2016). Dialect maps such as these provide useful context, historical background, and (sometimes) even an ‘explanation’ as to why certain features appear where they do (such as the use of thundersquall in coastal areas) -- all this, despite the reality that in terms of the actual distribution of individual features, there are always exceptions and outliers. In other words, most of us are okay with the idea that dialects have fuzzy boundaries, and even if not always stated explicitly, contemporary cartographic representations of language use are usually thought of as well-informed generalizations made by dialectologists, who themselves are usually keenly aware that they are smudging the details (Kretzschmar 2003, Upton 2006). Kurath himself expressed such a sentiment in his Word Geography:

In most cases, the word lines are chosen to exhibit the focus or ‘core’ of an area as well as its periphery or margin. One must, of course, keep in mind that this procedure does not tell the whole story. The full complexity of regional and local usage is displayed in the Linguistic Atlas itself. (1949, p. 11)

In addition to making a case for the re-introduction of ‘word lines’ to our dialectological vocabulary, this talk will attempt to tell a (more) whole story of Southern English, a story that starts with perceptions.

Perceptual dialectology has shown that the South is the “touchstone” in terms of laypersons’ beliefs about how people in the US talk (Preston 1997, Bounds, Cramer & Tamasi 2021). Recent work that I have been involved in (Burkette & Antieau 2023 and Cramer & Burkette forthcoming) suggests that linguists, too, have perceptions of Southern English that also bear scrutiny. When you look at the LAP speakers’ production of features traditionally hailed as “Southern” (e.g. poke for paper bag, a-prefixing, was-leveling, intrusive /t/ and /r/, [aɪ] ungliding, etc.), you find that their distributions stretch well beyond the confines of a geographic area that could be termed ‘the South’, which begs the question: Is there an empirical basis for drawing a southern dialect region? Ideas about and attitudes toward the South (good and bad) are powerful cultural forces, to be sure, but is Southern English more than an idea?

This talk attempts to answer these questions by looking to the actual distribution of linguistic features used by Kurath to delineate ‘the South’ and ‘the South and South Midland’, expanding his inquiry by using data not just from LAMSAS and LANE, but also data from the neighboring projects of the Linguistic Atlas of the Gulf States (LAGS) and the Linguistic Atlas of the North Central States (LANCS). Casting a wider geographic net to compile LAP data, this talk will re-assess the distributions of low (for ‘moo’), lightwood, co-wench! (to call a cow), (corn)shucks, light bread (for wheat bread), and you-all. While this re-assessment is still not “the whole story”, it does underscore the need for an adjustment of expectations in terms of how ‘Southern’ these features really are.