Presidential Address: Where Are the Dialects of American English At Anyhow?

THE QUESTION OF THIS PRESENTATION asks if we can locate regional speech areas, an apparently geographical quest. That's not all bad, so long as we remember that even physical geographers don't just locate things. Why things are where they are may be explained by where they came from. The "knobs" on the Ohio River where I grew up in southern Indiana are big piles of rubble pushed there by glaciers, a kind of Ice Age carpetbagging. Similar explanations exist in linguistic geography. When people in Milwaukee say by Aunt Mary's (while nearly everyone else in the United States says at Aunt Mary's), they borrow the semantics of the preposition by from German, a hardly surprising fact considering Milwaukee's dominant early German population.

Although the science has always concerned itself with historical explanations, dialectology has been predominantly not only of the ground (a geographical undertaking) but also of the mouth (a production-oriented undertaking) and, more recently, of the ear (a perception-oriented undertaking). In what follows, I want to say a few good words for the brain. In doing so, I will agree with Labov, Ash, and Boberg (forthcoming), who, among others, note the following criterion for dialect boundaries: "features should be systematic, rather than isolated, reflecting relations among two or more elements of the phonological system" (1). I believe this attention to organized linguistic features provides a more powerful motivation for boundary drawing than counting isolated features. I also believe that Labov, Ash, and Boberg will not be unhappy with me in extending this desideratum beyond the phonological, and I begin in that territory as I set out to capture the now-you-see-it-now-you-don't region of North American English—the U.S. Midland.

Benson (2003) looks at need and want plus particles (e.g., Jack wants in, Bill needs out) and shows that, in acceptability judgments, directional meanings of the particles (It's cold out, and Jack wants in) show a regional distribution (fig. 1). Midland respondents like these constructions, but Northerners and Southerners do not (although, oddly, there is no statistically significant difference among the regions for want in). But respondent judgments of these same constructions with nondirectional meanings (Bill's got a good deal, and Jack wants in) show no such distribution; there is moderate acceptability in all three areas (fig. 2).

There are two reasons why Benson's findings are interesting. The normal process of the extension of meaning goes from more concrete to less concrete (e.g., Traugott 1982), but non-Midlanders prefer the extended rather than the basic sense. That will require us to reexamine some theoretical foundations, as dialect data acquired from linguistically sophisticated fieldwork often do. I cannot explore this in detail here, but, in the spirit of Traugott (1982 and subsequent work), since pragmatic and even conversational factors are involved in establishing nonliteral meanings, attention may be directed toward those processes and away from the evaluation of the grammaticality of a construction. If so, nonusers of want/need plus particle require no extra pragmatic interpretation for the literal uses, which they evaluate on strictly grammatical terms. The nonliteral uses, however, may escape the nonusers' grammaticality detectors since their pragmatic reinterpreters are hard at work. This suggestion is obviously speculative, but, even if my analysis is wrong, we have some solid data to use in speculation.

The point here is that Benson's linguistically sophisticated research looks at systematic semantic-pragmatic differences and detects a Midland distinctiveness. This is a more significant finding than those of traditional Midland-haters (e.g., Carver 1981; Davis and Houck 1996), who count up unrelated lexical or lexical and phonological items and declare the Midland a figment of earlier dialectologists' imaginations.

Is there also systematic phonological support for a Midland? Deeds, I shirly din't—that is, dudes, I surely don'...