Voicing contrast in Ruwund

Abstract

Ruwund has a set of voiceless and voiced stops in its phonemic inventory. Length measurements reveal that voiceless stops are realized with duration comparable to voiced stops. There is also a positive VOT varying between bilabial, alveolar, and velar places of articulation. Voiced stops are characterized by a negative VOT and have greater voicing intensity. What is striking in Ruwund is that voiceless stops are also produced with voicing of weak intensity. These sounds should therefore be described as voiced stops. The presence and amount of aspiration during the delay between the first burst and the onset of voicing is the cue that contributes to these sound’s identification as voiceless. Repp [1] observed that the increase in the amplitude of aspiration noise relative to the following periodic vocalic portion increases the salience of this cue and helps the probability to classify these consonants into the voiceless category. A perceptual test m confirms this hypothesis. Removing the part lying between the first burst and the following voiced part accounting for the beginning of the following vowel, show that listeners do not recognize these stops as voiceless anymore but as voiced. Ruwund reveal therefore new subtleties of voicing distinction and VOT. [1] B. Repp, “Relative amplitude of aspiration noise as a voicing cue for syllable-initial stop consonants,” Lang. Speech, 23, 173–189 (1979).