健康与遗传心理学研究室知识库

在语言文字的阅读过程中，情绪语境对词汇加工整合起到的作用是研究者们较为关注的问题。部分研究通过情绪性刺激诱导出内隐情绪状态观察其对语句中情绪词加工的影响，普遍证实与读者当前情绪一致的词汇加工显著快于不一致的词汇，这种效应被称为情绪一致性效应。而在以语言性情绪材料构建情绪语境的研究中也观察到了类似的效应，主要表现为与语境中情绪语义期待相符的目标情绪词加工快于不一致的目标词。但是在语句水平很少有研究观察到情绪语境对后继情绪词的一致性促进效应，部分研究者发现了相反的研究结果，即不一致情绪词晚期语义整合被跳过因而与一致情绪词加工无差异。现有考察情绪语境对情绪词加工影响的研究多数为脑电实验，采用快速序列视觉呈现范式，逐字呈现关键语句，观察到的加工机制可能不完全适用于自然阅读中清绪信息的加工。而采用眼动研究范式，考察自然阅读过程中言语性情绪语境对后继情绪词汇加工影响的研究目前并不多见，且具有一定的研究价值。

本研究采用眼动研究范式，考察自然阅读中语句水平的情绪语境如何影响后继情绪信息的加工整合。三个实验均选择低可预测性(低完型概率)词汇作为目标词，并匹配了各条件下的语句合理性。实验一采用2X2的设计，两个因素是语境效价、情绪词效价，各自包含正性、负性两个水平。在反映早期加工的凝视时间和总加工时间上都观察到了语句水平的情绪性确实对后继情绪信息加工整合起到影响，主要表现为语境效价和目标词效价的交互作用，即不一致情绪词汇整合加工得到了易化。实验二采用单因素实验设计，三种条件分别为情绪一致、情绪不一致和中性控制条件，在总时间上观察到一致条件加工时间显著长于不一致条件，差异与实验一一致，不一致条件与控制条件间同样存在显著差异，不一致条件加工时间更短，由此推断情绪语境对不一致效价词产生了加工促进作用。实验三采用实验二的设计，但把语义理解任务变为外显的效价评价任务，观察到显著的一致性效应，一致条件下目标词加工速度显著快于不一致目标词，说明语句中情绪语境的效应和任务相关。

本研究认为，在自然阅读条件下，情绪性语句中情绪词受到语境的调节使得在中性语句中观察到的自上而下的情绪加工受到了抑制。本研究通过降低目标词可预测性，并控制语义合理性来考察情绪语境在单纯的效价水平如何对后继情绪信息的加工做出调节。在语义理解任务中观察到的情绪语境对不一致效价词的加工促进作用证实了前人的情绪优先假设—情绪性语境使效价违反词汇的情绪表征成为注意的中心，而语义加工整合被跳过了。并在前人的研究基础上提出情绪语境对一致和不一致情绪词的调节存在不同的作用机制。研究认为，语句中情绪语境对后继情绪词加工存在两种不同的调节机制，一是较低水平的情绪价加工优先性激活，二是较高水平的基于期待的情绪一致性效应。当目标情绪词可预测性低时，效价线索只能做出较低水平的调节。

Emotional context is considered as an important factor during text reading and word processing. Many studies have found that the induced mood or the affection expectancy could facilitate lexical access of the target word embedded in the sentence with the congruous valence, namely emotional congruous effect. While some other studies suggested that the highly automatic attention-capture emotional words can sometimes override integration processes and semantic processing during discourse comprehension, because of their emotion salience. It is still unclear, whether on not the processing of the emotional words is mutually interrelated when they are embedded in the same sentences. However, most of those studies were based on the EHF' studies in which test, the sentence appeared on the screen one word at a time. Unlike those studies, we used eye tracking technique in the current study, which readers could read sentence naturally.

Our study examined how the emotional context influences the processing of the following emotional word in normal reading. Emotional words were embedded in emotionally negative or positive sentences context. Semantic predictability9 and plausibility were matched across every condition. In Experiment 1，the 2X2 design was carried out9 one factors was context valence ( positive versus negative levels and the other was target word valence (positive versus negative levels).In experiment 2 and 3, we used the same set of sentence materials, participants should read sentence of both emotionally congruent, incongruent or neutral target words embedded in the same scenario stems with low word CP(Close probability).We also manipulate the task conditions. The participants were required to do the semantic comprehension in experiment 1 and 2. In experiment 3 the task is on line evaluative categorization.

As a result, on neither fixation tune, gaze duration, nor total reading time, could we found any significance of context valence effect or word valence effect. there is only a significant interaction between these to factors, the total reading time of emotional incongruous words was shorter than the congruous ones. In the experiment 2, the ANOVA data of the total reading time on the target words shows the same significance as the previous study, moreover, the total reading time of emotional incongruous words was also shorter than the neutral ones. However, the opposite reading reactions was observed in the valence relative task compared to semantic task which showed significant facilitation to the processing of emotional congruous words compared with the incongruous emotional words, indicated that the facilitation effect on the processing of emotional incongruous words was strongly depended on task.

Based on these results, we concluded that in normal reading, the valence of the word wasn't integrated with the emotional context when emotion violation happens, and this incongruous caused deeper semantic processing of the target word to be bypassed. Moreover the emotional effect found in by the top-down processing caused by the valence relative task, suggesting that emotional congruous effect is a superior processing mechanism that strongly depend on the power of expectancy. As a summery, the valence cue of the emotional context will actually do the effect on the emotional words' on-line processing in two ways, the research would be manipulated during our normal sentence reading. The arousal on the valence level will only facilitate the processing of emotional incongurous words with low word predictability. Instead, as the contribution of semantic expectancy is strong enough, the emotional congruous effect obviously would be seen.