The Research Methods of the Arkansas Language Survey

Abstract

SINCE OCTOBER 1970, when plans for a preliminary survey of Arkansas dialects were first proposed (Underwood 1973b), the research design has been considerably modified.1 The early plan was for a wide-meshed survey similar to traditional linguistic atlas projects; however, the desirability of beginning the study of Arkansas dialects as one would for a linguistic atlas has been reconsidered in light of data from three sources: from tape recordings of Arkansans interviewed for the DARE (Dictionary of American Regional English) project, made available to us by Frederic Cassidy of the University of Wisconsin, from eighteen pilot interviews conducted in Arkansas during 1970-71, and from published observations on Arkansas dialects. These data suggest that Arkansas English is much more uniform than previously assumed. While there may be numerous localized differences in vocabulary and incidental pronunciations, they are of minor importance in dialect differentiation-a point that will be treated in more detail below. Systematic differences in syntax and phonology (that is, differences in the ordered generative and transformational rules constituting native speaker competence) appear to be few. Moreover, the Arkansas survey ought to incorporate, wherever possible, the innovations and improvements in methodology and theory provided by investigators of urban dialects, such as Labov (1966) and Shuy, Wolfram, and Riley (1968). Also, the arguments set forth by critics of American linguistic geography, for example, Pickford (1956), Bailey (1968), and Dillard (1969a and 1969b), have been given serious consideration. As a result, an alternative research design for the survey was drawn up in consultation with sociologists at the University of Arkansas at Fayetteville. Revised plans for the first coherent state-wide survey of Arkansas dialects, which has been titled the Arkansas Language Survey, were put into finished form in 1972. The general aims of the study are-