38 Chapter 2 Phonology
the analysis of recordings of one male and one female speaker from East Madu-
ra indicate the vowel space (based on first and second formants) occupied by [ɨ]
is distinct from that of [ə]. At the same time, the difference in the vowel space
that they report for [ə] and [ɨ] is much less dramatic than the differences for the
other pairs, and indeed there was some overlap in their distribution. The space
occupied by each of the vowels in the other pairs was essentially completely
distinct from its partner. Cohn & Lockwood found little difference in the aver-
age F2s of the pair of vowels but found the difference between the average F1s
to be comparable for [i]/[ɛ], [u]/[ɔ] and [ɨ]/[ə] (the difference in the F1 averages
for [ɤ]/[a] were dramatically higher). In attempting to replicate these results
with a single male speaker from Jember, East Java, Bortscheller (2007) reports a
somewhat more dramatic difference in the vowel space occupied by [ə] and [ɨ].
He also reports a slightly greater difference between the average F1s for [ə] and
[ɨ] than found by Cohn & Lockwood for their male speaker. But again, the dif-
ference in average F1 is comparable to that found with the other pairs of vo-
wels. While not a matter of great urgency, given the instrumental evidence (al-
beit slim), it seems that the distinction should be recognized.
4.2 Other vowel processes
Finally, there are three low-level phonetic processes that have been proposed for
some apparent tense/lax allophonic variation and a vowel nasalization process.
The first is a rule that alters the quality of a high front or high back vowel in
closed syllables, accounting for perceived instances of [] and []. Stevens
(1968) cites this rule, and the reflexes ] and [] are included in Safioedin
(1977) and Oka et al. (1988/89). Potential examples include:
(34) [] ~ [n] ‘star’
[kippɔn] ~ [kppɔn] ‘confused’
[buntɔ] ~ [bʊntɔ] ‘tail’
Cohn & Lockwood (1994) report that for the two speakers they studied, high
front and high back vowels were significantly shorter in duration in closed syl-
lables than in open syllables. They reported no significant differences in the F1
and F2 values in the same environments. I have been able to confirm these re-
sults to some degree through acoustic analysis of data collected from four
speakers.^22 There were no significant differences in the F1 and F2 values in
open and closed syllables, but there was some difference in vowel length. These
(^22) Three speakers were recorded on Madura and one speaker originally from Situbundo
(East Java) was recorded in Iowa City. Thanks to Michael Bortscheller, Craig Dresser,
and Jeff Press for assistance with the acoustic analysis.