Abstract
This thesis is aimed at describing the phonology of Mayrinax Atayal, one of Formosan languages, and moves a step forward to observe and analyze the phonology as well as its interaction with morphology within Optimality Theory (OT), set forth by Prince and Smolensky (1993) and McCarthy and Prince (1993b). Special reference of UM affixation, a genetic morpheme in Austronesian languages, will be given relating to OCP effect on labial. It is shown that /b, r, g/, unlike in other dialects, are allowed in coda position in Mayrinax, and that there are only three underlying vowels /i, u, a/ in the inventory with [e] and [o] being found in a restricted number of borrowings. No consonant cluster is found in this language, in which I use *COMPLEX to regulate it. Stress systematically falls on the ultimate syllable, and an iambic foot construction sensitive to weight is proposed in this thesis based on vowel harmony. Apart from the phonology of Mayrinax Atayal, I have gone a step forward to discuss the interface of morphology and phonology. Morphemes that occur in different places, which are formalized by morpheme-specific alignment constraints, interact with phonological constraints that require the prosodic well-formedness. Morpheme dislocation, coalescence, glide formation and resyllabification are found to repair the possible prosodic ill-formedness caused by morpheme concatenations. The order of the prefixes, representing focus and aspect, is also discussed as an argumentation of the privilege of OT, the approach I use in the thesis. Within all the morphemes, two VC prefixal infixes are picked out for further discussion. A phonetic measurement is carried out to show that the consonants in these two VC infixes are measured significantly longer than other identical consonants in the same place in intervocalic position. The interpretation of these longer consonants relates to the OCP effect on labial, which causes the deletion of the first syllable when infixing UM to stems with labial-initials. I have offered four possible interpretations of these longer consonants. These explanations all lead to the same conclusion that these two VC infixes, UM and -in-, must be parsed tautosyllabically. The OCP effect is thus proposed to hold in the same syllable in Mayrinax. The OCP effect is also found in other Austronesian languages with different repairs, including Feature changing (Paiwan), Null Parse (Tagalog), Metathesis (Chamorro) and Assimilated prefix (Toba Batak). Thus, this thesis has moved a step forward to incorporate some of those languages into our analysis. A typological analysis with the re-ranking of three families of constraints, OCP, ALIGN and MAX-MP, is proposed but with some gaps needed to be filled in or re-examined to see if the typology would over-predict or under-predict.