Morpheme-level Features: Chaoyang Syllable Structure and Nasalization

In Chaoyang, the features [constricted glottis] ([c.g.]) and [nasal] are specified at the level of the morpheme, and distributed within the syllable in conformity with a set of ranked and violable output constraints. Coda consonants consist of Place features alone, and surface as voiceless glottalized stops if the syllable carries [c.g.], and as nasals if the syllable is not [c.g.]. The constraints governing [nasal] require that [nasal] be realized, that rhymes and syllables harmonize for [nasal], but that [nasal] may not associate to segments unmarked for [voice]. The interaction of these constraints means that in certain syllable types nasality surfaces on the entire syllable, in others on the rhyme only, in others on the onset only, and that in a final class nasality does not surface at all. The analysis of [c..g.] and [nasal] as morpheme-level features explains a pattern of segment loss in reduplication, where it is shown that the retained features of tone, [c.g.], and [nasal] are exactly the morpheme-level features of the language. The analysis is worked out in Optimality Theory (OT), and it is argued that an OT analysis avoids the familiar "conspiracy" problem, where MSC's and rules converge on a single output in both underlying and derived forms. It is also argued that MSC's or rules would be forced to state rules non-locally, to explain an interaction between onset nasalization and the presence vs. absence of a coda, but that OT offers a straightforward local constraint interaction account of these facts. The paper is organized as follows. After a brief summary of Chaoyang syllable structure, section 2 offer an analysis of glottalization which has as a surprising consequence the result that 'nasal' codas are not phonologically [nasal] (or [voice]), but allophones of final stops found in the absence of [c.g.]. This conclusion plays a major role in the analysis of [nasal] that follows. Section 3 provides a description of the [nasal] facts, and a preliminary analysis. Section 4 summarizes the essentials of Optimality Theory. and section 5 gives a formal OT analysis. Section 6 discusses the advantages of the OT analysis over an MSC or rule-based account. Section 7 introduces the reduplication data, which offers striking support for the morpheme-level nature of [c.g.] and [nasal]. The data is drawn from several papers by Zhang, also Zhu (1982) and Chiang (1991).