| Abstract: | This paper presents a constraint-based analysis of a process that in New Mexico Spanish (NMS) generates a syllabic consonant from a sonorant consonant + stressed high vowel combination. Despite the fact that vowels make better prosodic heads than consonants, it is demonstrated that a positional markedness constraint that bars the marked value on the place-of-articulation scale (e.g. Dorsal) from the position of foot DTE favors consonants over vowels in the role of foot and syllable head. Contrary to previous approaches that assume that vowel deletion is a condition for the syllabization of the consonant, it is argued that the vowel that disappears in the process of syllabizing a consonant is not deleted but either absorbed by that consonant or assimilated to it. It is shown that syllabic consonants are subject to two universal alignment constraints that govern their distribution by forcing them to be coarticulated with another consonant. Although languages may vary as to whether the syllabic consonant is coarticulated with a preceding or with a following consonant, there are no languages where syllabic consonants appear between two vowels or between a vowel and a pause, precisely because in such environments there is not an adjacent consonant available for coarticulation. The requirement that the syllabic consonants of NMS be coarticulated with a following consonant is the reason why they are never found before a vowel or a pause. Moreover, because they are also subject to sonority-related constraints, the syllabic consonants of NMS are more likely to occur after a vowel or a pause than after another consonant. |
