An investigation of the dynamics of vowel nasalization in Arabana using machine learning of acoustic features

This paper presents exploratory research on temporally dynamic patterns of vowel nasalization from two speakers Arabana. To derive a measure nasality, we use gradient tree boosting algorithms to statistically learn the mapping between acoustics and nasality in speaker-specific manner. Three primary findings emerge: (1) NVN contexts exhibit throughout entirety interval, propose that similar co-articulatory realization previously acted resist diachronic change this environment; (2) anticipatory is nearly as extensive carryover nasalization, which contrary previous claims; (3) degree word-initial relatively high, even #_C environment, suggesting sound *#Na > #a has involved loss oral constriction associated with N but not complete velum gesture.