Effects of first formant onset properties on voicing judgments result from processes not specific to humans

Abstract

Both first formant (F1) transition duration and F1 onset frequency have been proposed to be perceptually significant in categorization of voiced and voiceless syllable-initial stops. Transition duration per se may not, however, explain the fact that, for longer transitions, longer F1 cutback is required in order to perceive a stop as voiceless. Longer transitions result in lower F1 onsets at any duration of cutback greater than zero, and it is possible that the major effect of F1 is determined by its frequency at onset. In this study, F1-transition duration, onset frequency, and slope were varied across four types of F1 transition in which one of the three variables (onset frequency, duration, slope) was held constant while the other two were allowed to vary. Each of the four F1 types was used in syllables with higher formants appropriate for labial, alveolar, and velar places of articulation. By far, the best predictor of identification of these stimuli by human listeners was F1 onset frequency. F1 duration, F1 slope, and place of articulation had little or no effect on labeling boundaries. In a second experiment using Japanese quail (Coturnix coturnix japonica), birds were trained to respond differentially to voiced versus voiceless stops. The differential effects of F1 onset frequency on the "labeling" behavior of these birds was strikingly similar to that of humans listening to the same stimuli. These results are taken to provide strong evidence that F1 onset frequency is the primary determinant of shifts in voicing boundaries across place of articulation, and that general mechanisms not unique to humans appear adequate to account for the effects of F1 onset frequency on perception of voicing for syllable-initial stops.