Workshop on Syllable Structure and Gesture Timing, Columbus,Ohio, September 16-19, 1998

语言话语在语音方面传统上被表示为音素片段的串联线性串。从 1950 年代初 Jakobson、Fant 和 Halle 的研究开始，独特特征的概念被引入作为语音描述的最小单位，根据该概念，音素是二值语音定义特征的同时集合。在这个概念中，音素是具有时间整合的最小语音单位。完整音素片段的识别在很大程度上得到了声学信号中观察到的谱图不连续性的支持，特别是声音的开始和偏移以及发音内爆和塞音辅音的爆发。然而，从一个音素到下一个音素的过渡通常在语音信号中表示为连续平滑的频谱变化，没有明显的不连续性。对流行的音素串联和协同发音理论的严峻挑战主要来自 X 射线微生物研究的发音数据。随着发音姿势的时间细节变得可用，很明显，无法维持表现给定音素的发音姿势的同时性概念。 从语音学的角度来看，从语音表征的线性生成理论（发音顺序和分段问题非常简单）到当前的非线性表征（分段和特征之间的映射相当宽松，并且经常假设高度清晰的超分段结构），学术上已经有了稳定的进展。当前的非线性表示模型允许单个“片段”承载给定特征的多个值，并且非常自由地允许单个特征与许多“片段”相关联，其结果是人们可以有意义地质疑离散“片段”作为形式对象的存在。事实上，特征几何理论并没有明确表示“线段”的概念，并且有人认为传统的线段概念不对应于当前模型中的任何特定节点或表示级别。非线性模型越来越倾向于设置复杂但内部永恒的单个“片段”来代替片段簇。例如，汉语端木提出，根本不存在起始辅音簇，并且这种明显的序列是单个次要铰接片段（(k) inter alii）。 本次研讨会的目标是关注对表征理论问题至关重要的特定领域，并通过研究与音节相关的问题以及作为语音结构的“片段”来评估该领域的竞争理论。关于片段及其顺序存在广泛的理论观点，使得检验关于可能的辅音簇的经验主张变得困难。本次研讨会将聚集相关领域的主要领导人，创建一个论坛，以考虑解决这些问题的不同方法进行最新讨论。

Linguistic utterances in their phonological aspect are traditionally represented as a con catenated linear string of phonemic segments. Beginning with research by Jakobson, Fant and Halle in the early 1950's, the concept of the distinctive feature was introduced as the minimal unit of phonological description, according to which a phoneme is a simultaneous bundle of bi narily valued phonetically defined features. In that conception, the phoneme is the minimal pho nological unit with temporal integration. The identification of the integral phonemic segment was largely supported by spectrographic discontinuities as observed in acoustic signals, in particular voice onset and offset along with articulatory implosions and explosion of stop consonants. However, the transition from one phoneme to the next is often represented in the speech signal as a continuous smooth spectral change with no apparent discontinuity. A serious challenge to the prevailing phoneme concatenation and coarticulation theory derives from articulatory data largely due to x ray microbe studies. As temporal details of articulatory gestures became avail able,.it was clear that the concept of simultaneity of articulatory gestures which manifest a given phoneme could not be maintained. From the phonological perspective, there has been a steady intellectual progression from the linear generative theory of phonological representation where issues of articulatory ordering and segmentation were quite simple, to current nonlinear representations where the mapping between segments and features is quite relaxed, and highly articulated suprasegmental structures have been often assumed. A consequence of current nonlinear representational models, which allow single 'segments' to bear multiple values of a given feature, and which quite freely allow a single feature to be associated with many 'segments', is that one can meaningfully question the existence of the discrete 'segment' as a formal object. Indeed, the theory of feature geometry does not explicitly represent the notion of 'segment', and it has been argued that the traditional notion of segment does not correspond to any specific node or level of representation in current models. Nonlinear models have increasingly tended towards positing complex but internally timeless single 'segments' in place of clusters of segments. For example, it has been proposed by Duanmu for Chinese that there are no onset consonant clusters at all, and that such apparent se quences are single secondarily articulated segments ((k) inter alii). The goal of this workshop is to focus on a particular domain which is central to the theo retical issues of representation, and seeks to evaluate competing theories of that domain, by looking at issues relating to the syllable, and to the 'segment' as a phonological construct. The wide range of theoretical perspectives which exists regarding segments and their ordering renders it difficult to test empirical claims about possible consonant clusters. This workshop will assemble major leaders in relevant areas, to create a forum for up to date dis cussion considering different approaches to these problems.