Quantitative assessment of English-American speech relationships

This study explores synchronic and diachronic aspects of English phonetics and phonology by applying a set of complementary quantitative tools to English and American dialect data. The tools yield measures of aggregate variation in phonetic usage, identify dialect regions as clusters of speakers or localities with relatively similar patterns of usage, distinguish regions of extensive variation from others of comparative uniformity, and isolate groups of linguistic features that distinguish specific regions. Applied to data from England, the results largely corroborate the standard view but differ in the identification of regional boundaries and regionally distinctive features, as well as in placing those elements in a broader context of largely continuous and random variation. Applied to English and American dialect data, the analysis provides strong support for a model of new-dialect formation during the colonial period involving competition within and selection from a pool of linguistic features introduced by speakers from different regions of England. Consistent with the historical record of migration patterns, the analysis reveals American speech to be a relatively uniform amalgam of variants brought mainly from the English southeast, but also identifies distinct English regional sources for important elements of geographic and social variation in twentieth-century American speech. The study thus illustrates the broad applicability of the quantiative toolkit to linguistic research as well as some of its significant limitations, and demonstrates how the tools can be used in concert to yield important insights into historical linguistic variation and development.