Effects of Consonantal Context on the Pronunciation of /æ/ 73from a wordlist. The main difference in procedure from Stevens and House
was that every combination of initial and ¿ nal consonantal context was tested
for each vowel instead of being limited to symmetrical syllables. Therefore,
although this study eliminated the variable of manner by using only stop
consonants, the researchers were able to comment on differences in effect
between initial and ¿ nal consonants. Overall, conditioning environment had
minimal effect on vowel pronunciation in the Hillenbrand et al. study, but
this may be because natural speech was not elicited due to the extremely
controlled nature of the experiment. The effects they did see largely agreed
with Stevens and House, with the additional ¿ nding that preceding conso-
nants showed larger effects, in general, than following consonants, especially
for F2.^1 With regard to social factors, Hillenbrand et al. found the same effects
in both men’s and women’s speech.
Although the vowel patterns found in these two studies suggest different
dialects, the similarity of results suggests universal tendencies in the coartic-
ulation of speci¿ c consonants and adjacent vowels. The issue of whether coar-
ticulatory effects may, in fact, differ across dialects is not directly addressed
in either article. As mentioned previously, this analysis is concerned with the
same general Northern Cities Shifted dialect area investigated by Hillenbrand
et al. (2001), but it takes a more sociolinguistic approach. For example, in the
Hillenbrand et al. study, speakers who were familiar with linguistics were
asked to read phonetic transcriptions that were blocked by vowel (not scat-
tered) from a word list.^2 In the current study, however, naïve speakers were
asked to read actual words, written in regular orthography, and presented
one at a time on a computer screen to avoid the intonational effects that often
accompany reading a list.
Several other sociolinguistic studies on the effects of phonetic environ-
ment on vowel production have been done in this dialect region. Ito (1999)
analyzed Anglo speakers in rural Lower Michigan; Evans (2001) analyzed
speakers in Ypsilanti, near Detroit, who were originally from Appalachia;
and Jones (2003) analyzed African Americans in Lansing. The methodology
of the current study is very similar to that used in these three earlier Michi-
gan studies, enabling a comparison of results. Labov (1994: 100) discusses
the fronting and raising of /æ/ together as “advancement” of the NCS. He
provides ordered lists of environments that promote this advancement, based
on the results of a few major studies, including his own, in several Northern
Cities Shifted regions, including Detroit. The two environments on which he
comments are following manner of articulation^3 and point of articulation (not
speci¿ ed as preceding or following).^4 The combined results of these studies
are discussed in the following section.