Doctoral Dissertation Research: Articulatory Dynamics of Sibilant Convergence & Change

本论文的研究目标是有助于科学地理解声音变化，重点是一些声音变化是不对称分布的方式。也就是说，既然一个语音变化的语音动机可能存在于多个语音环境（相邻的语音）中，为什么在一个环境中观察到变化，而不是其他环境？这个项目将通过对/s/-retraction的研究来解决这个问题，这是美国英语中一种持续的声音变化，其中/s/在/str/单词中发音为/sh/（例如“street”发音为“shtreet”），但很少在/spr/和/skr/单词中发音（“scream”不发音为“shcream”，“spree”不发音为“shpree”）。该项目的结果将最直接地影响对/s/-收缩的科学理解，并更广泛地影响声音变化及其启动和传播，但可能有助于相关领域和行业的发展。通过研究个体差异，这些发现可以提供对第一和/或第二语言习得的见解，并可能有助于语音识别系统的开发，更快，更准确地为更多的用户，特别是那些与非标准accents.This项目将检查参与者的生产和感知/s/在这些环境中，以更好地了解观察到的不对称/s/-收回。在这个项目中，参与者将听到与视觉呈现的图像相对应的不同语音环境中包含/s/的听觉提示，并反过来为计算机提供口头指令以选择正确的图像。这个设计有两个重要的组成部分：首先，它是一个词汇决策任务，这意味着参与者必须选择最好的意义（由图像表示）给出可能模糊或意外的声学信号（听觉指令）。其次，这项任务鼓励语音调节，通过这个过程，说话者开始听起来更像他们的对话者（给出指示的模范说话者），这可以深入了解声音变化如何在言语社区传播。在整个实验过程中，将同时收集三个测量结果：音频记录、舌头的超声成像和眼睛注视跟踪。超声成像将阐明/s/在不同的语音环境中的不同发音，也将为语音调节的发音现实提供新的证据。眼睛注视测量将提供关于听众在真实的时间中使用/s/变化的信息。这些测量结果将同时收集，以检查参与者的生产和感知/s/在这些环境中之间的关系。

The goal of this dissertation research is to contribute to the scientific understanding of sound change, focusing on the way in which some sound changes are asymmetrically distributed. That is, given that the phonetic motivations for a sound change may be present in multiple phonological environments (the adjacent sounds), why is the change observed in one environment and not the others? This project will address this question through an examination of /s/-retraction, an ongoing sound change in American English in which /s/ is pronounced as /sh/ in /str/ words (such that 'street' is pronounced as 'shtreet') but rarely in /spr/ and /skr/ words ('scream' is not pronounced as 'shcream' and 'spree' is not pronounced as 'shpree'). The results of this project will most directly influence the scientific understanding of /s/-retraction narrowly and sound change and its initiation and propagation more broadly, but can potentially contribute to developments in related fields and industries. By examining individual variation, these findings may offer insights into first and/or second language acquisition and may aid in the development of speech recognition systems that are faster and more accurate for a greater number of users, especially those with non-standard accents.This project will examine participants' production and perception of /s/ in these environments in order to better understand the observed asymmetries in /s/-retraction. For this project, participants will listen to auditory prompts containing /s/ in different phonological environments corresponding to visually presented images, and in turn give oral instructions for the computer to select the correct image. This design has two important components: firstly, it is a lexical decision task, meaning that the participant has to select the best meaning (represented by the images) given the possibly ambiguous or unexpected acoustic signal (auditory instructions). Secondly, the task encourages phonetic accommodation, the process by which a speaker begins to sound more like their interlocutor (the model talker giving instructions), which can provide insight into how sound changes propagate across a speech community. Throughout the experiment, three simultaneous measurements will be collected: audio recording, ultrasound imaging of the tongue, and eye gaze tracking. Ultrasound imaging will elucidate the varying articulations of /s/ in different phonological environments and will also provide novel evidence for the articulatory reality of phonetic accommodation. Eye gaze measurements will provide information about listeners' use of /s/ variation in real time. These measurements will be collected simultaneously in order to examine the relationship between participants' production and perception of /s/ in these environments.