Abstract:

Kluender et al. [J. Acoust. Soc. Am. 97, 2552--2567 (1995)] propose that voicelessness in a continuum from short to long lag VOTs is signaled by abrupt synchrony capture of F2--F3 tuned fibers by F1 when voicing begins. Their hypothesis also predicts more capture at higher presentation levels, producing crossover to voiceless judgments at shorter VOTs, an effect observed in their experiments. In an attempt to replicate Kluender et al.'s findings, listeners were presented with a /ba/-/pa/ continuum at 48, 60, and 72-dB SPL: VOT crossovers did not differ significantly between the two highest levels, but were significantly longer for the lowest compared to the two highest levels. A second experiment tested manipulations of relative aspiration level by combining aspiration levels of 42, 48, 54, 60, and 66 dB with a 72-dB vowel; VOT crossovers increased with decreasing aspiration level. The results of these two experiments are compatible with the alternative hypothesis that reducing aspiration level below a critical level makes aspiration effectively inaudible. In any case, the synchrony capture hypothesis fails on two counts: Sinex and MacDonald [J. Acoust. Soc. Am. 85, 1995-2004 (1989)] observe that capture of mid- and high-CF neurons by F1 does not vary with VOT for VOTs less than 50-60 ms. In addition, Lotto and Kluender [J. Acoust. Soc. Am. 99, 2591(A) (1996)] demonstrate that the level effect still obtains when F1 is presented to a different ear than the higher formants.