认知与发展心理学研究室知识库

口语产生是将人的思想通过口头言语表达的过程，包括概念化、词汇化和发音。图-词干扰范式中，当干扰词和图片表征的事物归属于同一范畴并在语义上存在关联时，图片命名时间长；当干扰词和图片表征的事物分别归属于不同的范畴并在语义上不存在关联时，图片命名时间短，此即为图-词干扰范式中经典的语义干扰效应。在图片的范畴水平命名中，语义关联的干扰词促进了图片命名，出现了语义促进效应。对于图-词干扰范式中语义效应的发生阶段，Rahman 和Melinger用动态词汇网络假说中的语义概念促进与词汇干扰的权衡来解释； 概念选择模型认为语义效应发生在口语产生的概念和词汇阶段，重视前词汇水平；属于非竞争假说的反应选择假设则认为是在后词汇水平的语音形成阶段。 已有理论对于语义抑制效应和促进效应所发生阶段的假设各有不同， 且仅能解释部分实验结果。实验一通过设置不同的 SOA，变化图片和词语呈现之间的时间间隔，使得两者分别作用于语言产生系统中不同的阶段， 从而在材料不变的情况下使反应时表现出语义抑制效应和语义促进效应的反转， 同时利用高时间分辨率的事件相关电位指标考察反应时表现为抑制效应和促进效应时的语义效应发生阶段。在SOA0 条件下，发现了反应时的经典语义干扰效应。SOA-400 条件下出现个体差异，分为促进组、干扰组。反应时表现为语义促进效应时，在 50ms 开始的概念阶段以及 200-450ms 的词汇阶段波幅有差异，更多与概念激活相联系。反应时表现为语义干扰效应时，波幅差异主要发生在 300-450ms 的词汇阶段。实验二使用语义范畴大小作为自变量，在基本水平命名中，出现经典的语义干扰效应，且语义范畴加工在概念阶段即表现出不同。 本研究对建构汉语的言语产生模型提供了新的实验证据。

Speech production is the process of transforming thought into verbal language, which consists of three stages: conceptualization, formulation and articulation. The research of Heij La （1988） showed that in the picture word interference paradigm, there was a longer time when the interference words and pictures belonged to the same category, and there was a shorter time when the interference words and pictures belonged to the different categories. When the category of the picture was named, the semantic association of the interference word and picture promoted the picture naming, which was called the semantic facilitation effect （Glaser & Glaser, 1989）. For the mechanism of semantic effects, the Swinging Lexical Network Account considered it to be a trade-off between semantic conceptual promotion and lexical interference. The Conceptual Selection Model considered it happened in the concept and lexicon levels, especially the concept level. Response Selection Hypothesis emphasized the Articulation level. The existing theories are different from the assumption that the semantic inhibition effect and the facilitation effect, and can only explain some of the experimental results. By setting different SOAs, the time interval between images and words were different, and they were different in the stages of language production system, which could show the effect of semantic interference and semantic facilitation effect when the material was the same. In the SOA0ms condition, the classical semantic interference effect of the reaction time was found. Individual differences in SOA-400ms conditions were found, and the subjects were divided into facilitation group and interference group. When the reaction time was the semantic promotion effect, there were amplitude differences in the concept stage from 50ms and 200-450ms, and the relation of concepts were more active. When the response time showed semantic interference effect, the amplitude difference was mainly in the lexical stage of 300-450ms. Experiment two used semantic category size as independent variable in the basic level naming, the classical semantic interference effect was found, and semantic category processing was different from the concept stage. This study provided new experimental evidences for the construction of the speech production model.