Browse by author
Lookup NU author(s): Dr Daniel DuncanORCiD
This is the authors' accepted manuscript of an article that has been published in its final definitive form by Duke University Press, 2021.
For re-use rights please refer to the publisher's terms and conditions.
Corpus studies of regional variation using raw data from the internet focus predominantly onlexical variables in written language. However, online repositories such as YouTube offer thepossibility of investigating regional patterns using phonological variables, as well. This paperdemonstrates the viability of constructing a naturalistic speech corpus for sociophonetic researchby analyzing hundreds of recitations of Abraham Lincoln’s “Gettysburg Address.” We firstreplicate a known result of phonetic research, namely that English vowels are longer in durationbefore voiced obstruents than before voiceless ones. We then compare /æ/-tensing in recitationsfrom the Inland North and New York City dialect regions. Results indicate that there aresignificant regional differences in the formant trajectory of the vowel, even in identical phoneticenvironments (e.g., before nasal codas). This calls into question the uniformity of “/æ/-tensing”as a cross-dialectal phenomenon in American English. We contend that the analysis of spokendata from online social media can and should supplement traditional methods in socialdialectology to generate new hypotheses about socially conditioned variation.
Author(s): Bleaman IL, Duncan D
Publication type: Article
Publication status: Published
Journal: American Speech
Year: 2021
Volume: 96
Issue: 2
Pages: 161-191
Print publication date: 01/05/2021
Online publication date: 26/06/2020
Acceptance date: 22/04/2020
Date deposited: 23/04/2020
ISSN (print): 0003-1283
ISSN (electronic): 1527-2133
Publisher: Duke University Press
URL: https://doi.org/10.1215/00031283-8620511
DOI: 10.1215/00031283-8620511
Altmetrics provided by Altmetric
