Good timing: place-dependent voice onset time in ejective stops

Abstract

Abstract Voice onset time after voiceless unaspirated stopsdemonstrates a dependence on place of articulation, mostreliably being shorter for labial and coronal than for velarstops. Some of the proposed explanations for this patternsuggest that a parallel dependence is not be expected foraspirated or ejective stops. However, similar patterns dooccur with both aspirated and unaspirated stops. Cho andLadefoged [1] have suggested that ejectives do not follow thesame trend, but they had little data on bilabial ejectives tocompare with more plentiful data on velars. This papercontributes more material to this debate with expanded dataon Yapese and the first published material on ejective VOT inNez Perce. The results suggest that ejectives have a similarpattern to plosives and that therefore a unified explanation forall three types of stops should be sought. 1. Introduction It is well known that after the release of a prevocalic voicelessunaspirated stop the time that elapses before voicing beginsfor the vowel shows dependence on the place of articulationof the stop. The voice onset time (VOT) is quite reliablyshorter after a bilabial ([p]) than after a velar stop ([k]), withcoronal stops often being intermediate and almost alwaysshorter than velars (see, for example, [2, 3]). One proposedexplanation for the patterns seen is that the rate of apertureincrease in the releasing gesture differs for differentarticulators and locations. For example, Stevens [4] estimatesthat at the release of a labial stop the cross-sectional area isincreasing at about 100 cm