문장 정서가와 이모티콘 정서가의 통합 처리에대한 ERP 증거

Abstract

Yunju Nam & Upyong Hong (2018). ERP evidence on the integrative processing of the emotional valences of sentence and emoticon. Language and Information 22.3, 45-65. The current study investigated the interaction between the emotional valence of a sentence and that of an emoticon in text processing. In text fragments including an emoticon followed by a declarative sentence, emotional valence of each text component (a sentence[S] or an emoticon[E]) was manipulated in terms of emotional valences (positive[p] or negative[n]) and their combination patterns (same or different). 40 target text fragments were distributed in 4 conditions (pS+pE, nS+nE, nS+pE, pS+nE), and 24 Korean speakers’ neural responses were recorded at the position of emoticon, while they read the text in word by word (plus a text-final emoticon) fashion. In addition, participants were required to categorize the emotional valence of each text ([p] or [n]) directly after the occurrence of the emoticon. Compared to the baseline conditions in which sentences and emoticons were of same emotional valence, the early N400 component which reflects the detection of semantic anomaly was elicited only in the nS+pE, and not in the pS+nE condition. This implies that the negative emotional valence of a sentence collides strongly with the conflicting emotional valence of the following emoticon, whereas the positive emotional valence of a sentence does not. Moreover, the emotional value of the whole text was categorized as largely identical to that of the sentence (86% [n] in the nS+pE condition and 76% [p] in the pS+nE condition). Taken together, the current study demonstrates that the emotional valence of a text including both language and image information tends to be determined by the language information, and the strength of the so-called negative bias effect varies depending on which text component (language or image) contains the negative emotional value. (Konkuk University)