Lexical Change and Variation in the Southeastern United States, 1930-1990 (review)

1 98Reviews Lexical Change and Variation in the Southeastern United States, 1930-1990. EllenJohnson. Tuscaloosa: University ofAlabama Press, 1996. Pp. xvi + 319. $19.95. Dialectology is a discipline on which lexicography must depend for information about certain aspects of word usage: geographical, social, and historical limitations. Unlike Gaul, dialectology is divided into just two parts: a traditional, largely rural- and historical-oriented, geographical dialectology and a more recent, largely urban- and contemporary-oriented, social dialectology . Ellen Johnson's book attempts to merge the two; and, in the process, she documents some trends in the vocabulary that are of lexicographical interest. Johnson chose 150 semantic items investigated during the 1930s in Georgia, South Carolina, and North Carolina by Guy S. Lowman for what was to become the Linguhtic Atlas ofthe Middle and South Atlantic States (LAMSAS) . In 1990, she matched Lowman's informants as closely as possible by the criteria of geographical region (mountain, piedmont, coastal), ruralness (rural, urban), age (over 77, 62-74, 31-59), race (black, white), education (college, high school, elementary school, less than full elementary school) , and sex (female , male). Johnson got close matches for all except two of the criteria: education and ruralness. The difficulty she had in matching those two criteria suggests some important changes in American life during the past 60 years. Education is the more striking divergence. Lowman had no college graduates among his informants.Johnson had no informants with less than a full elementary-school education. The educational level ofthe population has risen sharply. In the 1930s, 76 percent of the population in the Southeast region had less than seven years schooling, and only 8 percent had completed high school. In 1990, only 22 percent had not attended high school, and 78 percent had completed high school or gone beyond. Thus Lowman underrepresented the lowest educational level and overrepresented those with an elementary and high-school education . Johnson overrepresented informants with less than a high-school education and those with a college degree, while underrepresenting those with only a high-school diploma. The other unmatched criterion was ruralness. Of Lowman's informants , 77 percent were rural and 23 percent urban. OfJohnson's informants, 64 percent were rural and 36 percent were urban. The actual percentage of shift of the region's population was from 29 percent urban in the 1930s to 56 percent in 1990, so whereas Lowman underrepresented the urban element slightly,Johnson, in attempting to match him, did so to a greater extent. These two criteria in which Johnson does not closely match Lowman simply reflect demographic changes in America. In extrapolating from dialect studies like these to the interests of lexicography , several points need to be kept in mind. Regional geography tends to oversample that portion of the population that is white, old, and rural. The Reviews1 99 age factor especially needs to be kept in mind. Whereas 1 percent of the region 's population was over 75 years old in the 1930s, and 4.5 percent in 1990, in Lowman's sample and in Johnson's the percentages were respectively 30 and 28. Older generation usage is therefore overrepresented in both studies. Although that conservative bent does not invalidate the historical trends that are revealed, it must be remembered that the usage reported is old-fashioned. Several lexical changes shown by Johnson's study are especially noteworthy . The 150 questions in the LAMSAS worksheets investigating the semantic items with whichJohnson deals elicited 1,007 different lexical responses in the 1930s. Johnson's 1990 interviews produced 1,402 responses for the same items: close to a 40 percent increase. When we consider that the semantic items were not changed and that they were predominantly universal (terms for one's mother and father) or agricultural (calls for cows and sheep), this striking increase in the number of responses might suggest that the English vocabulary , even in its more conservative ranges, has grown during the past several generations, as conventional wisdom has assumed. Another striking change is that, whereas regional variation was the most significant type of dialect difference in die 1930s, it was the least significant in 1990, having been superseded by social variation. This too confirms conventional wisdom. To...