Differential weighting of acoustic cues to consonant voicing by Spanish and English listeners

Fernando Llanos, Alexander L. Francis, Amanda Shultz, Olga Dmitrieva, Rachel Chapman, Mengxi Lin

Linguistics, Purdue University

Acoustically, even linguistically simple phonetic features such as voicing are typically instantiated in terms of multiple acoustic cues. For example, Lisker (1986) described 16 distinct acoustic properties that consistently distinguish productions of the words “rabid” and “rapid.” Among these, voice onset time (VOT; the time between the burst release of the consonantal closure and the onset of voicing in the vowel) has been shown to be the dominant cue in English (Lisker 1978), with voiced stops exhibiting a short lag VOT (0-20 ms) and voiceless stops showing a longer lag (> 20 ms). However, secondary cues, especially onset f0 (the fundamental frequency at the onset of voicing) may also be used, especially when dominant cues are not available or are ambiguous (Haggard, et al. 1970). In Spanish, VOT is also a cue to voicing, but the 0 ms boundary between Spanish voiced (VOT < 0 ms) and voiceless (short lag) stops has been argued to be psychoacoustically more ambiguous than the 20 ms boundary in English (Holt, et al. 2004). Given that listeners appear to turn to secondary cues when primary cues are more ambiguous, it is possible that Spanish listeners may be more dependent than English listeners on secondary acoustic cues such as onset f0. In this talk we will present the results of an experiment demonstrating exactly this pattern: Spanish listeners showed a greater reliance on onset f0 than did English listeners in a stop consonant identification task. However, in the course of analyzing the results, a number of additional possible explanations were identified. Thus our presentation will focus not just on the initial results, but will also provide a discussion of these other possible interpretations, predictions derived from these interpretations, and proposed subsequent research designed to distinguish between these predictions.