Doctoral Dissertation Research: Phonological Prediction in Spoken Language Processing

语言在时间上迅速展开，但母语人士的理解通常是准确和有效的。先前的研究表明，语言处理看起来如此轻松和强大的一个原因是语言使用者可以使用上下文来预测即将到来的语言输入。人们普遍认为，在这种预测性语言处理中，高期望词的语义和形态句法特征可以被预先激活。然而，一个可预测的单词的语音形式是否也可以在实时句子处理中常规地预先激活，这是一个正在进行的争论。这个项目通过利用预期的音调变化现象来解决这个争论。本研究的结果将有助于我们进一步了解预期音调变化在语音识别和语音处理中的作用，以及语音形式预激活在预测语言处理中的作用。这将反过来帮助我们理解语言感知的认知过程。本项目首先研究了预期音调变化线索如何利用眼动追踪技术促进语音感知中即将到来的单词的音调预测。然后，该研究将通过使用脑电图（EEG）技术测量听众对前一个词的大脑反应，来检验他们是否能预先激活一个可预测词的音调形式。至关重要的是，在某些实验条件下，预期词会被一个意想不到的词所取代，这个词与预期词的音调形式相同或不相同，因此，前一个词的预期音调变化线索要么与预期词的音调一致，要么不一致。本研究的指导假设是，如果预期词的声调形式能够被预激活，那么当意外词与预测词的声调不同时，事件相关电位（ERPs）对前一个词的调幅应该更大；然而，如果很少或没有语音形式的预激活，在遇到意想不到的单词之前，应该很少或没有ERP振幅调制。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。

Speech unfolds rapidly in time, yet native speakers’ comprehension is generally accurate and efficient. Previous research suggests that one reason why language processing seems to be so effortless and robust is that language users can use context to predict upcoming language input. It is generally agreed that semantic and morpho-syntactic features of a highly expected word can be pre-activated in such predictive language processing. However, there is an ongoing debate about whether the phonological form of a predictable word can also be routinely pre-activated in real-time sentence processing. This project addresses this debate by exploiting the phenomena of anticipatory tonal variations. The results of this study will advance our knowledge on the role of anticipatory tonal variations in spoken-word recognition and speech processing, and on phonological form pre-activation in predictive language processing. This will in turn inform our understanding of the cognitive process in language perception.This project first examines how anticipatory tonal variation cues can facilitate tonal prediction of an upcoming word in speech perception using eye-tracking technique. Then, the study will examine whether listeners can pre-activate the tonal form of a predictable word by measuring their brain responses to the preceding word using electroencephalography (EEG) technique. Crucially, the expected word in some experimental conditions will be substituted by an unexpected one that shares or does not share the tonal form with the expected word, so the anticipatory tonal variation cues in the preceding word is either concordant or incompatible with the tone of the expected word. The guiding hypothesis is that if the tonal form of the expected word can be pre-activated, there should be larger amplitude modulation of event-related potentials (ERPs) to the preceding word when the unexpected word and the predicted word differ in tone; however, if there is little or no phonological form pre-activation, there should be little or no ERP amplitude modulation prior to encountering the unexpected wordThis 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.