Understanding the perception and recognition of spoken words: Effects of phonetics, phonological variation, and speech mode

每天，人们都要面对语言的变化。读者看到的是用不同字体和字体打印的单词，通常是静态的。听众听到的语音信号充满了变化。在不同的情况下，不同的说话者用不同的口音以不同的速度，带着不同的情绪，以不同的速度发音，听众以每秒5到7个音节的速度流。没有一个单词的两个发音是相同的，但这种变化通常是不被注意到的。尽管存在这种差异，但听众很快就能熟练地理解这些不同的发音，把它们当作一个词而不是另一个词；考虑到许多单词的发音都是一样的，这一点很值得注意。语言理论的一个核心问题是听者如何能够接受这些可变的语音串并理解它。对听众如何完成这个任务的解释包含了对问题的纯语言解决方案。萨姆纳博士的研究项目采用了一种新的方法，以声学-语音值传达的现实世界信息为基础。从言语的任何窗口，听者都能得到关于声音模式的线索，但他们也能得到关于特定言语情境的社会特征的线索。这项工作背后的推动力是，理解对口语的感知和识别的进步将来自于研究听者如何将这些无处不在的线索结合在一起。萨姆纳博士将通过研究听众对不同语言和社会维度的不同词汇的反应方式来研究这个问题。她将用实验方法告诉我们，一个单词的发音如何提高对心理词汇中相关单词的识别能力，以及听众如何根据所发出的语言和社会背景对不同的声音进行分类。这项工作调查了听者在理解语音时依赖于附近声音和单词的社会信息声学值的说法，以及语言信息存储在影响语音感知的社会表征中，其水平远低于曾经认为的水平。这项工作为语言感知理论增加了语言和社会经验之间的新联系，并应促使该领域更广泛地重新考虑声学模式的作用及其在理论中的地位。结合语言和社会信息，本研究可能为改进交互式口语技术提供一条途径。

Everyday, people face variation in language. Readers see words printed in different fonts and typefaces, typically static on a page. Listeners hear a speech signal that is riddled with variation. Words are pronounced at different speeds, with different emotions, by different speakers with various accents in different situations, streaming by listeners at a rate of about 5 to 7 syllables per second. No two utterances of a single word are identical, yet this variation typically goes unnoticed. Despite this variation, listeners quickly and adeptly understand these different pronunciations as one word and not another; this is remarkable, considering that many words sound the same. An issue central to linguistic theory is how listeners are able to take this variable speech string and understand it. Explanations of how listeners accomplish this task incorporate purely linguistic solutions to the problem. Dr. Sumner's research program takes a new approach, grounded by the real-world information conveyed by acoustic-phonetic values. From any window of speech, listeners encounter cues about sound patterns, but they also encounter cues about the social characteristics of a particular speech situation. The impetus behind this work is that advancement in understanding the perception and recognition of spoken words will come from examining the ways in which listeners use these ever-present cues together. Dr. Sumner will examine this question by examining the ways in which listeners respond to words that vary along different linguistic and social dimensions. She will use experimental methods that inform us about how a word pronunciation improves recognition to related words in the mental lexicon, and about how listeners categorize different sounds depending on the linguistic and social context in which they are uttered.This work investigates the claims that listeners rely on socially-informative acoustic values of nearby sounds and words when understanding speech, and that linguistic information is stored with social representations that influence speech perception at a level much lower than once thought. This work adds a new link between linguistic and social experience to theories of speech perception and should prompt the field to reconsider the role of acoustic patterns and their place in theory more broadly. Integrating linguistic and social information, this research may provide a path to improve interactive spoken language technologies.