Underspecified vowels

In words with an initial non-prominent syllable possessing a vowel in the environment /s/ _ Nasal Consonant, the vowel may predictably be either /i/ or /u/ depending on the phonological effects of the environment. In these words, the vowel is specified for the feature [high]. It is not, however, specified for the feature [back]. The reason is that the [back] feature, giving rise to /u/, in contrast to /i/, is predictable from the phonological environment. The [back] feature assigned to the vowel comes from the immediate environment. It may be assigned from the immediately following nasal when this is bilabial, i.e. /m/. The underspecified vowel derives its [back] feature here through the transfer of labiality, since back vowels in JamC are labial, i.e. produced with lip rounding. Otherwise, the back feature may be derived from the vowel of the immediately following syllable when such a vowel itself has the feature [back]. These items are all lexically specified as having an initial /sV/ sequence where V stands for the underspecified vowel, i.e. specified for [high] but not for [back] as demonstrated by the examples in the first two columns below.

Cassidy and Le Page (1980: lxii) note that the initial syllables in examples such as those above may be produced as a syllabic . Meade (1995: 33) refers to Akers (1981) as making a similar observation. We would argue that this is a case of the underspecified vowel in the /sV/ sequence becoming optionally devoiced under the influence of the preceding voiceless fricative, producing phonetically [s] and a voiceless vowel, i.e.  or . These forms are phonetically indistinguishable from the syllabic form,  proposed by Cassidy and Le Page.  and  are merely optional forms of  when the following consonant is a sonorant, as represented in the third column of the table. Where the following consonant is a voiceless stop, as in /sVp/, /sVt/ and /sVk/,  and  are the only possible manifestations of the underspecified vowel in an entirely voiceless environment. In such sequences, the underspecified vowel is obligatorily devoiced.