This paper tries to confirm what David and Hammond (1995) assert in the analysis of on-glides [y] and [w] in American English. They argue on the basis of phonotactics of the two glides that [w] in a CwV sequence belongs to the onset, while [y] in CyV belongs to the nucleus as the first part of a diphthong. It is shown in this study that a constraint-based account can give a reasonable explanation to two observations: one is that [y] as well as [w] behaves like ordinary consonants in a GV sequence and the other is that [y] can precede vowel [u] only, not any other vowel in a CyV sequence such that only [Cyu-] is possible while there is no such a restriction in a CwV sequence. This phonotactic difference is found to come from adopting [yu] as a unit of one mora on the one hand and the interaction of various constraints of Optimality Theory on the other hand. It is asserted that labio-velar glide [w] is always in the onset in [wV-] and [CwV-] as well, while palatal glide [y] is in the onset in [yV-] but in the nucleus in [CyV-].
1. Introduction
2. On-glides in American English
3. Constraint-Based Analysis of On-glides
4. Discussion
5. Summary and Conclusion
(0)
(0)