汉语双字词听觉加工中字、词信息激活的动力模型研究

What is going on during spoken word recognition? Psycholinguists are always interested in such an issue and yet debates are not ended. Most of the researchers believe that spoken word recognition system is a dynamic system that different levels of mental representations are involved in it. But what are these representations? More than a decade ago, Stone, Vanhoy, and Van Orden (1997) put forward a very intriguing and counterintuitive hypothesis. According to this hypothesis, a recurrent network was proposed that the flow of activation of sound and spelling is inherently bidirectional. Consistent symmetrical relations (phonology is spelled consistently) result in stable and fast activation, whereas inconsistent and asymmetrical relations slow down the system on its way to equilibrium. This hypothesis acquired stable off-line and on-line evidence from research on alphabetic languages (Perre, & Ziegler, 2008; Ziegler, Ferrand, & Montant, 2004; Ziegler, Petrova, & Ferrand, 2008). However, lacking of the evidence from non-alphabetic languages, such as Chinese, the catholicness of this theory is still not sure. In addition, research on visual word recognition showed that sublexical and lexical information has great impact of accessing Chinese (Ding, & Peng, 2006; Peng, Li, & Liu, 1994; Wang, & Peng, 1999; Zhang & Peng, 1992; Zhou & Marslen-Wilson, 2000; Zhou, Marslen-Wilson, Taft, & Shu, 1999). But people still don’t know how this information affects the auditory part. Therefore, three parts of experiments will be arranged in the present project: in the first part, different kind morphemic representations will be controlled and activation of phonological, orthographical, and semantic information of morphemes in accessing spoken Chinese disyllable compound words will be discussed. Homophone effects under the disyllable context is expected to be found so that the other assumptions of the present research can be further investigated; the second part will mainly discuss whether the homophone effects found in the first part is reliable in different experimental paradigms. Orthography effect is expected to be weakened in the visual world paradigm; in the final part, feedback from lexical representation of Chinese disyllables will be studied and the processing mechanism of spoken word recognition will be explored. According to the foreseeable results, an edited version of dynamic model applying to Chinese spoken word recognition is expected.

对词汇通达中各水平表征的激活共振研究是当前语言认知研究的热点。词汇通达的动力模型认为，词汇通达过程中语音、词形和语义之间存在着相互作用。然而，现有的研究结果仍缺少来自意音文字词汇通达的证据，尤其是听觉词汇通达中词形激活的普遍性亟待研究。本项目将结合行为、电生理(ERPs)和眼动指标，揭示汉语听觉词汇加工的心理机制。项目包括三项研究：研究一考察听觉词汇通达早期音、形、义的激活研究，探讨汉字表征对汉语听觉词汇通达的影响；研究二通过改变实验任务和材料，考察汉字表征对词汇通达的影响在不同的任务中是否具有稳定性；研究三进一步考察汉字表征之间、汉字与整词表征之间的关系对听觉词汇通达的影响，探讨各因素之间的相互作用。最终，结合三项研究的结果，拟构建汉语双字词听觉加工的动力模型，以丰富和扩展现有对汉语听觉词汇加工过程的认识。

项目通过三项研究，探讨了语言加工中，各水平信息激活与通达的动态过程。. 研究一结合行为和电生理 (ERPs) 指标，探讨言语理解和产生过程中，语音、词形和语义之间的相互作用，揭示汉语词汇加工的心理机制。结果表明，在不同实验任务和材料条件下，汉字表征对词汇通达的影响具有一定的稳定性。语音、字形和语义之间存在动态的激活过程，三者之间相互影响，共同实现词汇通达的顺利进行。. 研究二以行为研究为主，探讨词汇通达中各类信息的提取。结果表明，已有的经验 (具身经验和世界知识) 对词汇通达具有重要的影响。但是，这些影响因素具有任务依赖性。即，词汇通达时对不同信息的选择受任务导向的影响。. 研究三在句子水平下，采用眼动技术进一步考察在汉语词汇语义提取过程中，世界知识和即时信息的交互作用。结果表明，世界知识和即时信息均影响句子中词汇的语义提取速度。但是，两者在整个通达过程中起作用的阶段不同。两者亦以动态激活的形式相互影响。. 结合三项研究的结果表明，词汇语义提取的过程中会激活多种信息。这些信息的激活强度在不同阶段不同，且相互之间会存在相互反馈作用，从而进一步促进各项信息的提取，最终完成词汇语义的通达。基于此，我们构建汉语词汇加工的动力模型，以更全面地展示汉语词汇通达的心理机制。