Best Catherine T, Hallé Pierre A
MARCS Auditory Laboratories, University of Western Sydney, Penrith NSW 1797, Australia and Haskins Laboratories, 300 George St., New Haven CT 06511, U.S.A.
J Phon. 2010 Jan 1;38(1):109-126. doi: 10.1016/j.wocn.2009.09.001.
Cross-language differences in phonetic settings for phonological contrasts of stop voicing have posed a challenge for attempts to relate specific phonological features to specific phonetic details. We probe the phonetic-phonological relationship for voicing contrasts more broadly, analyzing in particular their relevance to nonnative speech perception, from two theoretical perspectives: feature geometry and articulatory phonology. Because these perspectives differ in assumptions about temporal/phasing relationships among features/gestures within syllable onsets, we undertook a cross-language investigation on perception of obstruent (stop, fricative) voicing contrasts in three nonnative onsets that use a common set of features/gestures but with differing time-coupling. Listeners of English and French, which differ in their phonetic settings for word-initial stop voicing distinctions, were tested on perception of three onset types, all nonnative to both English and French, that differ in how initial obstruent voicing is coordinated with a lateral feature/gesture and additional obstruent features/gestures. The targets, listed from least complex to most complex onsets, were: a lateral fricative voicing distinction (Zulu /ɬ/-ɮ/), a laterally-released affricate voicing distinction (Tlingit /tɬ/-/dɮ/), and a coronal stop voicing distinction in stop+/l/ clusters (Hebrew /tl/-/dl/). English and French listeners' performance reflected the differences in their native languages' stop voicing distinctions, compatible with prior perceptual studies on singleton consonant onsets. However, both groups' abilities to perceive voicing as a separable parameter also varied systematically with the structure of the target onsets, supporting the notion that the gestural organization of syllable onsets systematically affects perception of initial voicing distinctions.
塞音浊音的音系对立在语音设置上的跨语言差异,给将特定音系特征与特定语音细节联系起来的尝试带来了挑战。我们更广泛地探究浊音对立的语音-音系关系,特别是从特征几何和发音音系学这两个理论视角,分析它们与非母语语音感知的相关性。由于这些视角在关于音节起始部分内特征/手势的时间/相位关系的假设上存在差异,我们对三种非母语起始部分中塞音(塞音、擦音)浊音对立的感知进行了跨语言研究,这三种起始部分使用了一组共同的特征/手势,但时间耦合不同。以英语和法语的词首塞音浊音区分在语音设置上存在差异,对这两种语言的听众进行了三种起始类型的感知测试,这三种起始类型对英语和法语来说都是非母语的,它们在初始塞音浊音与一个边音特征/手势以及其他塞音特征/手势的协调方式上有所不同。从最不复杂到最复杂列出的目标起始类型分别是:一个擦音边音浊音区分(祖鲁语 /ɬ/-ɮ/),一个边音释放的塞擦音浊音区分(特林吉特语 /tɬ/-/dɮ/),以及塞音+/l/组合中的舌尖塞音浊音区分(希伯来语 /tl/-/dl/)。英语和法语听众的表现反映了他们母语中塞音浊音区分的差异,这与之前关于单辅音起始部分的感知研究一致。然而,两组听众将浊音感知为一个可分离参数的能力也随目标起始部分的结构而系统变化,这支持了音节起始部分的手势组织系统地影响初始浊音区分感知的观点。