Linguistic Security, Ideology, and Vowel Perception 259considered “incorrect” or nonstandard, and the perceived users of such forms.
Such studies attempt to collect what given sets of subjects “know” about how
language patterns in their speech communities, and, as most studies reveal,
much of the work involves prescriptive ideology.
Sociophonetic work on the perception of vowel tokens in the Detroit
area (Niedizelski 1999) shows that subjects react in perceptual tests if they
“know” even low-level phonetic features of a speaker’s dialect, even if
that knowledge is inaccurate. In a number of studies (e.g., Preston 1989,
Niedzielski and Preston 1999), Anglo middle-class residents of southeastern
Michigan have demonstrated high degrees con¿ dence in the correctness of
their own variety of American English, or what Labov 1966 called linguistic
security. Perhaps the clearest demonstration of this is Figure 11.5. In this
¿ gure, results of a ranking task are presented: subjects from southeastern
Michigan were asked to label the correctness of the dialects spoken in each
state, with the darker shades representing the highest degrees of “correct-
ness.” As is apparent from an examination of the ¿ gure, Michigan—and
only Michigan—receives the highest correctness ranking. This phenom-
enon of ranking their own speech as the most correct is con¿ rmed by the
interviews presented in Niedzielski and Preston (1999). Michigan residents
unequivocally offer their variety as an example of correct American Eng-
lish, often referring to theirs as the variety closest to what dictionaries or
grammar books prescribe.
Figure 11.5 Correctness rankings of Michigan residents (Preston 1989).