语音信息在中文听觉词汇识别中的作用机制:来自眼动研究的证据

Spoken word recognition is the basis of oral comprehension. Prior studies found that the phonological information plays a vital role during spoken word recognition. However, the role of phonological information still remains less clear. Compared to the alphabetical languages, Chinese has its own unique characteristics which may result in the role of phonological information in spoken word recognition being more complex. Using the printed-word paradigm, this study aims to examine the role of the phonological information during spoken word recognition in Chinese. In this study, we will investigate the following questions: (1) how the initial syllable frequency influences the spoken word recognition; (2) how the tail information and its similarity influence the spoken word recognition; (3) how the tonal information and its similarity influences the spoken word recognition; (4) how the context predictability influences the role of phonological information during the spoken word recognition and its time course. This study is important in understanding the cognitive mechanism underlying spoken word recognition in Chinese and help to build a more complete theoretical model of spoken word recognition. In addition, it also strengthens the understanding of relationship between language comprehension and attention, providing the theoretical basis for the development of artificial intelligence voice system.

听觉词汇识别是口语理解的基础。语音信息在听觉词汇识别中发挥着重要的作用，然而语音信息在听觉词汇识别中的作用机制尚不清楚。与拼音文字相比，中文词汇具有自身的语言独特性，这些特性将可能导致语音信息在中文听觉词汇识别中的作用机制更为复杂。本研究将采用词汇情境范式结合眼动追踪技术，系统、深入地研究语音信息在中文听觉词汇识别中的作用机制。通过5个系列实验，我们拟回答:（1）首音音节频率如何影响听觉词汇识别过程；（2）尾音信息及其相似性如何影响听觉词汇识别过程；（3）声调信息及其相似性如何影响听觉词汇识别过程；（4）语境预测性如何影响语音在听觉词汇识别中的作用和时间进程。该研究将推动对中文听觉词汇识别背后认知机制的理解，进一步完善中文听觉词汇识别的理论模型。同时结合词汇情境范式还能深入理解注意与语言理解之间的关系，对人工智能语音系统的开发提供借鉴，具有重要的理论意义和实践价值。

语音信息在中文听觉词汇识别中起着重要的作用，但是语音信息具体的的作用机制尚不清楚。本项目采用词汇情境范式结合眼动追踪技术，从尾音相似性、声调相似性、语境预测性以及音段相似性几个方面进一步考查语音信息在听觉词汇识别中的作用机制。通过以上四个研究，我们发现：（1）尾音信息在听觉词汇识别过程中会被激活，且尾音相似性会影响听觉候选词汇的激活程度；（2）声调信息在听觉词汇识别中起着部分限制作用，声调相似性会影响听觉候选词的激活程度；（3）高预测性的语境信息能增强声调信息在听觉词汇识别中的限制作用；（4）声调信息可以在仅有部分音段信息重叠条件下对词汇识别产生影响。这些研究成果加深了我们对语音信息在听觉词汇识别中认知机制的理解，对构建和完善中文听觉词汇识别模型提供了重要的理论和实践支持。