| 其他摘要 | During discourse comprehension, people can infer the emotional state of the protagonist and keep track of the emotional changes. In the research of emotional discourse comprehension, most studies adopted prediction or violation paradigm to investigate the effect of emotional context on critical words, in which both valence contrast and semantic violation were included. What we wanted to know was, how were emotional words were processed in the coherent emotional context, and how was influenced by the congruent emotional context or the contextual arousal? In addition, current research on emotional discourses paid less attention to memory representation, so we wanted to know, how the semantic memory representation of emotional words was influenced by emotional context? In this paper, three studies were designed to examine these problems.
In this paper, Study 1 (Experiment 1) was used to detect the influence of congruent emotional context on the integration of subsequent emotional words. In Exp 1, two-sentence discourses were used as experimental material. The first sentences expressed negative or neutral context, and the keywords in the second sentences were negative words. ERP results found that, compared with neutral context, the negative context elicited smaller N400 and larger LPC on negative words, indicating that the congruent emotional context reduced the semantic analysis of subsequent emotional words and led to more continuous evaluation of emotional information in late stage.
On the basis of Exp 1, Study 2 (Exp 2+ Exp 3) explored the influence of contextual arousal on the integration of negative and positive words respectively. Two-sentence discourses were used as experimental materials. In Study 2, two-sentence discourses were used as experimental material. The first sentences expressed a high-arousal positive event, a low-arousal positive event, or a low-arousal negative event, which were consistent or inconsistent in valence with the critical words in the second sentences. We observed the valence shift effect in the P200 and P300 in the high arousal context, while the valence shift effect in the low arousal context only appeared in P300, and the distribution of P300 effect in the low arousal context was more limited, which indicated that the increase of contextual arousal promoted the integration of negative words. Exp 3 adopted the same experimental design to detect the influence of contextual arousal on the integration of positive words. ERP results showed that the High-arousal inconsistent condition induced smaller N400 and larger LPC than the Low-arousal consistent condition. These results indicated that high arousal context might weaken the semantic processing of positive words, while lead to more elaborated emotional evaluation in the later stage.
Both study 1 and study 2 found that emotional context brought more elaboration during the late evaluation stage on subsequent emotional words, but might weaken semantic processing. Based on the results of Study 1 and 2, we wonder whether emotional context could also affect semantic memory representation of emotional word. We examined this problem in Study 3 (Exp 4+ Exp 5). In Exp 4, eye movement technology and text-change paradigm were adopted. The discourses composed of four sentences were used as material, and the first three sentences served as negative context or neutral context, and the fourth sentence contained one negative word as the critical word, which was identical among different conditions. Similarly, Exp 5 adopted the same experimental design and paradigm to explore the influence of emotional context on the semantic representation of positive words. The results of Exp 4 showed that the accuracy of highly relevant words was lower than that of low relevant words, and the total reading time and number of fixations of highly relevant words in the emotional context were larger than that in the neutral context. The similar result pattern was also found in Experiment 5. These results indicated that emotional context inhibited the semantic processing of emotional words and resulted in worse semantic memory representation compared with neutral context.
In conclusion, this paper investigated the influence of emotional context on the online processing and memory representation of emotional words, by means of three studies and five experiments. The results showed that congruent emotional context reduced semantic analysis of emotional words, but enhanced the late elaborated emotional evaluation. In addition, the increase of contextual arousal exerted different effects on the integration of positive and negative words. Furthermore, emotional context reduced the semantic processing of positive and negative words and weakened the semantic memory representation. Therefore, this study revealed the effect of emotional context on the processing of emotional words from two aspects of online processing and memory representation, and further illustrated the different forms of positive bias and negative bias during discourse comprehension. |
修改评论