Articulatory cross-language study of initial consonant clusters in varying prosodic conditions

韵律结构不仅以音长和语调方式为标志，而且以所谓的发音强化为标志。本研究的目的是通过分析不同韵律条件下控制辅音群内部组织的发音原则，更好地理解韵律层次中音段层和更高层次之间的相互作用。韵律变化在这个项目中有效地发挥了双重作用：首先，它本身就是一个有趣的话题，可以更好地理解说话者的发音行为如何向听者提供韵律结构的信息。其次，韵律变异可以作为一种探针来揭示语簇的内部结构：单个手势的稳定性如何，不同手势之间的衔接性如何？集群是特别有趣的几个原因：（1）目前还不知道一个集群的哪一部分或属性将在韵律突出的位置，如重音，重音，或边界-最初，（2）分析集群在不同的韵律条件下将揭示新的光的具体条款“发音加强”，其中包括各种部分相反的现象。特别是考虑到三种语言德语，法语和英语将提供更多的经验知识，以证实统一理论的尝试，如特征增强，声音扩展，手势和超发音。(3)传统的音系学方法未能解释某些辅音序列的高频率，如初始的咝音+停止集群。成功的集群可以是那些表现出高度显著的声学特性同时仍然允许相当大的手势重叠（即，实现信息的有效并行传输）的集群。通过分析几个发音器官（舌头，下巴，软腭，嘴唇，声门外展）的运动学和集群内的手势协调，我们希望找到证据，发音限制在不太喜欢的集群，例如停止鼻，或策略，以避免掩盖相关的线索，如较少的手势重叠。辅音序列的可感知性的时间调整的相关性将在基于发音合成的感知测试中进行评估。简而言之，我们的目标是更好地理解人类语言中一些最复杂的发音的认知表征，无论是从语言学还是从运动控制的角度来看，以及潜在地揭示声音变化的过程。

Prosodic structure is not only marked by durational and intonational means but also by socalled articulatory strengthening. This project aims at a better understanding of the interplay between the segmental tier and higher levels in the prosodic hierarchy by analysing articulatory principles governing the internal organization of initial consonant clusters under varying prosodic conditions. Prosodic variation effectively has a dual role in this project: Firstly, it is interesting topic in its own right to better understand how speakers¿ articulatory behaviour provides information on prosodic structure to the listener. Secondly, prosodic variation acts as a probe to reveal the internal structure of clusters: how stable are individual gestures, and how strong is the cohesion between different gestures? Clusters are particularly interesting for several reasons: (1) It is not known yet which part or property of a cluster will be enhanced in prosodically prominent positions such as under accent, stress, or boundary-initially, (2) Analysing clusters in different prosodic conditions will shed new light on the specifics of the term `articulatory strengthening¿ which subsumes a variety of partly opposing phenomena. Especially taking the three languages German, French and English into account will provide more empirical knowledge for corroborating attempts at unifying theories such as Feature Enhancement, Sonority Expansion, ¿-Gesture and Hyperarticulation. (3) Traditional phonological approaches have failed to account for the high frequency of certain consonant sequences, such as initial sibilant + stop clusters. ¿Successful¿ clusters may be those that exhibit highly salient acoustic properties while still permitting considerable gestural overlap (i.e achieving efficient parallel transmission of information). By analysing the kinematics of several articulators (tongue, jaw, velum, lips, glottal abduction) and the gestural coordination within clusters we hope to find evidence for articulatory constraints in less preferred clusters, e.g. stop ¿ nasal, or strategies for avoiding the masking of relevant cues such as less gestural overlap. The relevance of temporal adjustments for the perceptibility of consonant sequences will be evaluated in perception tests based on articulatory synthesis. In short, we aim for a better understanding of the cognitive representation of some of the most complex articulations in human speech, both from a linguistic and a motor control perspective, as well as potentially shedding light on processes underlying sound change.