DISCRIMINABILITY ALONG THE VOICING CONTINUUM: CROSS—LANGUAGE TESTS*
DISCRIMINABILITY ALONG THE VOICING CONTINUUM: CROSS—LANGUAGE TESTS*
复制标题
发声连续体的可辨别性：跨语言测试*
DOI：
--
复制
发表时间：
2019
期刊：
影响因子：
--
通讯作者：
A. Abramson
中科院分区：
文献类型：
--
作者：
A. Abramson
在语音和类语音刺激的心理声学实验中，人们发现听者在物理连续变化中检测差异的能力取决于他们对语言类别的刺激分配。关于合成的停顿元音音节的停顿数据表明，只有当它们被分配到不同的音素时，刺激才会被区分出来；然而，这样的研究仅限于英语。在进行跨语言测试时，2我们提出了两个问题：(1)在音素标签分布的边缘处，辨别是否变得更加尖锐？(2)可辨别性是由特定的语言经验形成的，还是暗示着一般的语音类别？语音起始时间（VOT）在这里很有趣，因为许多语言以相似但不完全相同的方式使用VOT来区分两个或三个顶级类别。3 ． 在这项研究中，我们使用了一组为随附论文中描述的识别实验准备的刺激。使用平行共振合成器，我们制作了31个音节，由唇部停顿和[a]类型的元音组成。VOT在十分钟内变化。从150毫秒开始。（- 150）爆发前，到150毫秒。爆发后（+150）为了便于区分，我们以三和弦的形式呈现了这些变体。在每个三联中，两个刺激是相同的，一个是不同的。任务是判断这个奇怪的人是排在第一、第二还是第三位。三联音是通过在VOT连续体上以两步、三步和四步的间隔配对刺激来形成的，从而比较VOT在20、30或40毫秒之间的差异。每个比较被安排在六种排列
In psychoacoustic experiments with speech and speech-like stimuli it has been found that listeners’ ability to detect differences along a physical continuum of variation depends on their assignment of the stimuli to linguistic categories. Data on stops in synthesized stop-vowel syllables show that the stimuli are distinguished only if they are assigned to different phonemes;l however, such studies have been confined to English. In going into cross-language testing,2 we asked two questions: (1) Is discrimination sharpened at the margins of the labelling distribution for a phoneme? (2) Is discriminability shaped by specific language experience, or does it imply general phonetic categories? Voice onset time (VOT) is of interest here because many languages use VOT, in similar but not identical ways, to distinguish either two or three 6top categories.3 . For this study we used a set of the stimuli prepared for the identification experiments described in the accompanying paper. Using a parallel resonance synthesizer, we made 31 syllables consisting of labial stops plus a vowel of the type [a]. VOT varied in ten msec. steps from 150 msec. (-—150) before the burst, to 150 msec. (+150) after the burst. We presented these variants in triads for discrimination. In each triad two stimuli were identical and one was different. The task was to decide whether the odd one was in first, second or third position. The triads were made by pairing stimuli at two-, threeand four-step intervals along the VOT continuum, thus comparing VOT differences of 20, 30 or 40 msec. Each comparison was arranged in six permutations