322 Thomas C. Purnell
Figure 13.1a is transformed into Figure 13.1b and 13.1c. Perceptual dialectol-
ogy presupposes the use of crucial percepts that are further assumed to derive
from a set of acoustic characteristics. What we see in this chapter is that these
assumptions, while not fatal to the progress made in investigating data at Fig-
ure 13.1a, 13.1b, and 13.1c, require a non-axiomatic approach. As such, future
research should investigate how experimentally manipulated cues relate to
the perceived dialect as well as to the intended dialect.
Notes
1 Compare, for example, the role of dialect as subordinate variety (Chambers and
Trudgill 1998: 3), or as a regional or historical derivative (Romaine 2002: 310;
Hock 1991: 380 –381).
2 Preston’s work is based on Hoenigswald (1966). See also the related conception of
this claim in Figure 1 in Bradac et al., 2001: 146.
3 An example of an important issue bearing on the perception of ethnically af¿ li-
ated dialects is how close the input is to the prototype or exemplar, e.g., Attitude
Representation Theory (Lord and Lepper 1999), and Structural Alignment Theory
(Markman and Gentner 1993).
4 See also Medin 1989; Yamauchi 2005.
5 The matched guise paradigm has been used across languages (Lambert et al.
1960) and dialects (Purnell et al. 1999).
6 The Wisconsin English data presented herein is the product of collaborative work
with Jennifer Mercer, Joe Salmons and Dilara Tepeli.
7 See Purnell et al. 2005a,b; Salmons et al. 2006; Tepeli et al. 2008.
8 For example, the pronunciation of “The Bears” used in a Saturday Night Live
skit about Chicago sport fans saying da Bears with a ¿ nal [s] is often, ironi-
cally enough, used disparagingly by Wisconsinites about the Chicago football
team.
9 The weight is approximated by the amount of variation a measure accounts
for and the correlation coef¿ cients and eigenvectors assigned by the statistical
model.
10 For a more detailed description of the methodology see Purnell et al. 2005a, b.
11 The change in F0 was transformed by
(–1* ̈F 0 )
100,where ̈F0 is the original change in F0. The change in F1 was transformed by
(–1* ̈F 1 )
100,