Analyzing Vowel Variation by the Numbers

An Acoustic Analysis of Vowel Variation in New World English By Erik Thomas Publication of the American Dialect Society 85 Durham, N.C.: Duke University Press, 2001. Pp. ix + 230.

An Acoustic Analysis of Vowel Variation in New World English by Erik Thomas is a significant contribution to our growing body of knowledge of the vowel systems of North American English.

Synthesizing the results of major studies in dialectology in the last century, it presents a excellent overview of regional variation in the pronunciation of American English. In addition, it presents an impressive body of original research, comprising 192 vowel plots, including the largest set of plots of speakers of African American and Mexican English to date. This is an important innovation in the description of vowel variation—the use of techniques of acoustic measurement and analysis in place of impressionist transcription. Primarily descriptive, this important reference volume offers an extensive acoustic treatment of distinctive elements of Southern English that have been largely unanalyzed in existing literature. Moreover, this work is exceptionally valuable for its analysis of archival recordings of speakers born before the Civil War, providing real-time data on diachronic changes in specific regional varieties of American English.

Beginning with an overview of current work focusing on the acoustic analysis of dialectal variation of vowels in American English, chapter 1, "Introduction and Methods," outlines the scope and limitations of the study. While the study makes no attempt to account for variation attributable to socioeconomic status and gender, as well as individual variation attributable to style and register, it is designed to highlight the "most salient dialectal divisions that occur among the speakers ... included" (4), organized by ethnicity, geography, and age. The bulk of the chapter explains the nature of the vowel formant plots and the techniques of acoustic analysis, providing useful information to the reader on how to use them. Thomas begins with a detailed explanation of the symbols used in the text and in the plots, following principally the 1993 revision of the International Phonetic Alphabet, which is reproduced in appendix B of the text. While detailed phonetic representations are included in the discussion, phonemic representations are used in the plots. The phonemic representations are arbitrary symbols, notes the author, which represent a set of historically related variants and should not be construed as phonological representations connected to a particular phonological theory. Rather than indicating all tokens for a particular speaker, the vowel plots indicate the mean value of tokens measured for each vowel by a filled square. Representing mean values rather than all individual tokens, Thomas argues, makes it possible for the glide trajectory of diphthongs to be clearly indicated in the plot. As he notes, detailed information on glides of diphthongs, "neglected in past sociophonetic studies," is "vital to understanding dialectal variation and sound change" (8). Information on standard deviation for each vowel is not included. The author's rationale for these decisions, and its drawbacks, is addressed below. Approximately 150-200 vowels, selected from reading passages and casual conversation, were measured for each speaker, using median linear predictive (LPC) values and harmonic estimation methods on a variety of equipment.

Chapter 2, "Vowel Variants in New World English," is, from my perspective, the most compelling and significant aspect of the monograph. It covers the diaphones for each stressed vowel, drawing not only on the vowel plots but also on an impressive range of research by others. In many respects, this chapter represents a significant update to chapters 3-5 of The Pronunciation of English in the Atlantic States (Kurath and McDavid 1961), based on all of the Linguistic Atlas projects, as well as contemporary sociophonetic research in the Labovian framework. In addition to a wider range of regional coverage than Kurath and McDavid (1961), Thomas also discusses the diaphones of each vowel in general, instead of in specific words, as is the case in the Kurath and McDavid study. The first section of the chapter presents...