Cross-Language Study of Perception of Coarticulated Speech

摘要过去25年来收集的关于语音产生和感知的跨语言数据表明，语言在协同发音模式和感知组织方面可能存在系统差异。本文提出的核心问题是语言感知的语言差异是否与协同发音结构的语言差异有关。需要测试的具体假设是，语言特定的协同发音模式会引起听者对协同发音的声学后果的期望；这样的期望反过来又导致了语言特定的感知适应模式。这里收集的跨语言声学和感知数据与语音感知和语言习得的理论问题有关。当前的语言知觉模型在知觉的原素是听觉事件还是声道手势方面存在分歧。补偿性语境效应是特定于语言的（反过来又与特定于语言的协同发音模式有关），这一发现将限制可能的模型。反对这些影响的纯粹听觉来源，但也反对发音手势方法，该方法提出声学信号为听者提供足够的信息来解开共同发音的影响。言语知觉的发展模型更集中地关注于初始知觉能力的语言特异性重组。即使有关协同发音-感知联系的具体假设不成立，这些数据也将拓宽关于感知能力可以通过语言经验和相关经验的性质进行修改的知识。＊＊＊

9319597 Beddor ABSTRACT Cross-language data on speech production and perception gathered over the past 25 years show that languages may systematically differ in both coarticulatory patterns and perceptual organization. The central question addressed by this proposal is whether language differences in speech perception are linked to language differences in coarticulatory structure. The specific hypothesis to be tested is that language-specific patterns of coarticulation give rise to listener expectations concerning the acoustic consequences of coarticulation; such expectations in turn lead to language-specific patterns of perceptual accommodation to coarticulated speech. The cross-language acoustic and perceptual data to be collected here are relevant to theoretical issues in speech perception and language acquisition. Current models of speech perception differ in whether the primatives of perception are auditory events or vocal- tract gestures. A finding that compensatory context effects are language-specific (and in turn linked to language-specific coarticulatory patterns) would constrain possible models. It would be argued against a purely auditory source for these effects, but also against articulatory-gesture approaches which propose that the acoustic signal provides sufficient information for the listener to disentangle coarticulatory influences. Developmental models of speech perception are more centrally concerned with the language- specific reorganization of initial perceptual capabilities. Even if the specific hypotheses concerning a coarticulation-perception link are not upheld, the data will broaden knowledge of which perceptual capabilities can be modified through linguistic experience and the nature of the relevant experience. ***