Bennett and Rose (to appear): Moro voicelessness dissimilation and binary [voice]

Direct link: http://roa.rutgers.edu/content/article/files/1716_will_bennett_1.pdf

ROA: 1334
Title: Moro voicelessness dissimilation and binary [voice]
Authors: Will Bennett, Sharon Rose
Comment: pre-print, appearing in Phonology
Length: 41pp
Abstract: This paper reports on a heretofore undescribed pattern of voicelessness dissimilation in the Kordofanian language Moro. Voiceless obstruents become voiced when preceding another voiceless obstruent in a transvocalic (?CVC) configuration. This pattern is robust and productive across multiple morphological contexts. The phonetic facts of voicing in Moro show it to be a difference between prevoiced and short lag VOT. This points to [voice] as the most realistic featural characterization of the voicing contrast; the pattern cannot be explained as dissimilation of another feature like [spread glottis]. The voiceless dissimilation pattern is strong evidence that [voice] is binary–and that [-voice] may be phonologically active, despite being ‘unmarked’. We show that when reference to [-voice] is allowed, the Moro pattern can be straightforwardly analyzed as [-voice] dissimilation. Our formal analysis uses the theory of SURFACE CORRESPONDENCE, which carries no assumptions about markedness as a prerequisite for dissimilation. An appendix compares the proposed analysis to alternatives based on other approaches to dissimilation.
Type: Paper/tech report
Area/Keywords: phonology, correspondence, dissimilation, voice, abc