Chapter 12
Identi¿ cation of African American Speech
Erik R. Thomas, North Carolina State University;
Norman J. Lass, West Virginia University;
Jeannine Carpenter, Duke University
- Introduction
Among the most heavily researched topics in the sociolinguistics of speech
perception is the issue of the discriminability of African American and Euro-
pean American voices. This issue involves at least four research questions.
First, can listeners distinguish African Americans and European Americans
on the basis of voice alone? Second, do demographic differences among
speakers and listeners affect discriminability? Third, what features are lis-
teners capable of accessing in order to make ethnic identi¿ cations of voices?
Finally, what features do listeners access in real-life situations where they
have opportunities to identify speakers as African American or European
American? The ¿ rst three questions appear to be more or less resolved. It is
the last question that has proved most intractable.
A large number of studies have addressed the ¿ rst question, that of
whether listeners can identify voices as African American or European
American (Dickens and Sawyer 1952; Stroud 1956; Hibler 1960; Larsen and
Larsen 1966; Br yden 1968a, 1968b; Buck 1968; Tucker and Lamber t 1969;
Shuy et al. 1969; Shuy 1970; Alvarenga 1971; Koutstaal and Jackson 1971;
Abrams 1973; Irwin 1977; Lass et al. 1978; Lass et al. 1979; Lass et al. 1980;
Bailey and Maynor 1989; Haley 1990; Hawkins 1992; Walton and Orlikoff
1994; Trent 1995; Baugh 1996; Purnell, Idsardi, and Baugh 1999; and Fore-
man 2000). The earliest of these studies, Dickens and Sawyer (1952), is over
50 years old, which demonstrates the sustained interest in discriminability
of ethnicity. These studies utilized a wide variety of research techniques
that are summarized in Thomas and Reaser (2004a). For example, the length
of the stimuli varied from utterances consisting of several sentences to, in
the case of Walton and Orlikoff (1994), utterances of a single vowel. Nev-
ertheless, the answer to the question of whether listeners can distinguish