Access of semantic knowledge during speech perception

描述（由申请人提供）：大多数理解语音感知机制和过程的努力都没有研究语义知识是如何在语音感知过程中获得的。本研究旨在利用行为和眼动追踪的方法来探讨在口语单词理解过程中意义是如何被获取的，并发展一种包含语义表征激活的言语感知机制理论。提出的行为实验测试了语义邻域密度对单词识别的影响，并比较了罕见和频繁语义特征在确定语义相似度方面的重要性。提出的眼动追踪实验提供了细粒度的收敛数据，将进一步阐明如何从语音输入中获取语义知识。提出的计算模型将整合语音处理中的语义知识，以解释行为和眼动追踪结果，并做出新的预测。

DESCRIPTION (provided by applicant): Most efforts to understand the mechanisms and processes underlying speech perception have not examined how semantic knowledge is accessed during speech perception. The proposed research aims to investigate how meaning is accessed during spoken word comprehension using behavioral and eye-tracking methods and to develop a mechanistic theory of speech perception that incorporates activation of semantic representations. Proposed behavioral experiments test the effect of semantic neighborhood density on word recognition and compare the importance of rare and frequent semantic features in determining semantic similarity. Proposed eye tracking experiments provide fine-grained convergent data that will further elucidate how semantic knowledge is accessed from spoken input. A proposed computational model will integrate semantic knowledge in speech processing to account for the behavioral and eye tracking results and to make novel predictions.