AffiliationUniversity of Arizona
MetadataShow full item record
DescriptionPublished as Coyote Papers: Working Papers in Linguistics from A-Z
AbstractKeley-i, a Philippine language, has two rules assimilating consonants across a vowel. Such rules might be taken as evidence against the Morphemic Tier Hypothesis (MTH) and against the Locality Condition (LC). The MTH states (1) Morphemic Tier Hypothesis (MTH) If and only if two segments are members of separate morphemes are those two segments aligned in separate phonological tiers. The Keley-i data suggest that the MTH does not hold universally because consonants assimilate across vowels, which has been taken as evidence for two segmental planes in order to prevent the crossing of association lines. The data also create problems for the Locality Condition: (3) Locality Condition (LC) A phonological rule is applicable only if the target and trigger are adjacent. The consonant features assimilate across an intervening vowel: the target and trigger, being skeletal slots, are not adjacent. I suggest here that adopting the feature hierarchy as proposed in Archangeli and Pulleyblank (1986) (which is a modification of Clements 1985) combined with underspecification theory (Archangeli 1984, Pulleyblank 1986, Archangeli and Pulleyblank 1986) allows an analysis of the Keley -i data which permits maintaining the MTH and the LC. A further result is that the Spreading Hypothesis is maintained as well, thus supporting the hypothesis that phonological assimilation is formally expressed in one manner only, namely by insertion of association lines, and not by feature copy rules. (See Hayes 1986, Archangeli and Pulleyblank 1986.) (4) Spreading Hypothesis Phonological assimilation is expressed only by rules adding association lines. The discussion is organized as follows. First, the feature hierarchy and the theory of underspecification are briefly outlined. I then present a partial analysis of the Keley-i data. The analysis consists of a syncope rule and some rules of consonant assimilation. Finally, I return to the problems that Keley -i presents for the MTH and the LC and propose that the relevant Keley-i data are not only in accordance with the MTH and the LC but predicted by the interaction of the two sub -theories, the Feature Hierarchy and Underspecification.