Doctoral Dissertation Research: The role of listener experience in perception of conditioned dialect variation

尽管人类常常想当然地认为自己有能力听懂别人说话，但任何使用过Siri或Alexa等语音助手的人都会意识到，理解语音实际上是非常困难的。造成这种困难的一个原因是不同的说话者有不同的口音，所以两个说话者可能会对同一个单词发出不同的发音(例如：‘car’在美式英语和英式英语中的发音相同)，而不同的单词发音相同(例如：美式英语中的gloss听起来和伦敦人的glass一模一样)。虽然人类通常能够适应这些口音差异，但人们对这种适应是如何起作用的知之甚少。另一个谜题是人类如何适应不同的口音：听者不仅可以提高他们对新口音的理解程度，还可以提高他们如何利用这种口音的微妙之处来捕捉说话者的社交特征。例如，许多美国人无法区分英格兰北部和南部的口音，也无法区分英国上层阶级和工人阶级的口音，但在搬到英国后，他们可以学会做出这些区分。同样，人们对这种适应是如何起作用的知之甚少。这个项目调查了听者是如何随着时间的推移适应口音的——既包括他们对口音的理解程度，也包括他们对这种口音的社会知识——并探讨了这两种适应之间的关系。通过比较生活在美国的美国人和生活在英国的美国人如何学习英国南部英语，研究了对口音的长期适应。该项目主要研究像“bath”、“grass”和“pass”这样的单词，这些单词在美式英语中发音为“short-a”（“cat”中的元音），而在英国南部英语中发音为“long-ah”（“father”中的元音）。它调查了：(1)美国人对英国南部英语中哪些单词含有“long-ah”的了解程度如何？(2)随着他们掌握了更多的英式英语，这些知识是如何变化的？(3)这告诉我们他们对英国口音的心理表征是什么？研究人员将招募美国和英国的美国人，以及一组英国参与者作为对照组，进行几项感知实验。关键项目将包括与“长ah”类似，但在英式英语中发音为“short-a”的单词。例如，“grass”和“gas”在美式英语中押韵，所以如果“grass”在英式南部英语中发long-ah，人们可能会认为“gas”也会发long-ah，但事实上，英国南部的人发“gas”时发short-a。美国人知道short-a是正确的英式发音吗，还是他们错误地把grass概括为gas ？如果美国人听到“gas”的发音带“long-ah”，他们会认为这听起来很优雅，很英式吗？还是他们知道这听起来像有人在装口音？比较不同程度接触英式英语的听者对这些问题的回答，可以告诉我们听者在适应时做出了什么样的语言概括，这将为理解人类如何如此容易地适应新口音提供重要的一步，并可能为语言处理的技术进步提供信息。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Although humans often take for granted our ability to understand other people when they talk, anyone who has ever used a voice assistant like Siri or Alexa will be aware that understanding speech is in fact enormously difficult. One reason for this difficulty is that different speakers have different accents, so two speakers might pronounce the same word differently (e.g. 'car' in American versus British English), while pronouncing different words identically (e.g. 'gloss' in American English can sound identical to a Londoner's 'glass'). While humans are typically able to adapt to these accent differences, little is known about how this adaptation works. A further puzzle is how humans adapt to accents socially: listeners can improve not only in how well they understand a new accent, but also in how well they use the subtleties of that accent to pick up on a speaker's social characteristics. For example, many Americans cannot distinguish between accents from the north versus south of England, or between upper-class and working-class British accents, but upon moving to the UK can learn to make such distinctions. Again, little is understood about how this adaptation works. This project investigates how listeners adapt to accents over time--both in terms of how well they can understand an accent and in their social knowledge about that accent--and explores the relationship between these two types of adaptation.Long-term adaptation to accents is investigated by comparing how Americans in the US versus Americans living in the UK process southern British English. The project focuses on words like 'bath', 'grass', and 'pass' which are pronounced with "short-a" (the vowel in 'cat') in American English but with "long-ah" (the vowel in 'father') in southern British English. It investigates: (1) how well do Americans know which words have "long-ah" in southern British English? (2) how does this knowledge change as they gain more experience with British English? and (3) what does this tell us about their mental representations of British accents? Americans in the US and UK, along with a control group of British participants, will be recruited for several perception experiments. Crucial items will include words that are similar to long-ah words, but pronounced with "short-a" in British English. For example, 'grass' and 'gas' rhyme in American English, so if 'grass' is pronounced with long-ah in southern British English, one might expect 'gas' to also be pronounced with long-ah--but, in fact, southern British speakers pronounce 'gas' with short-a. Do Americans know that short-a is the correct British pronunciation, or do they incorrectly generalize from 'grass' to 'gas'? If Americans hear 'gas' with long-ah, do they think it sounds refined and British--or do they know that it sounds like someone faking an accent? Comparing the answers to these questions for listeners with different amounts of exposure to British English will tell us what kinds of linguistic generalizations listeners make as they adapt, which will provide an important step in understanding how humans are so easily able to adapt to new accents, and may inform technological advances in language processing.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.