Influence of affective context on prediction in L1 and L2 of Chinese-English bilinguals
Katherine Sendek1, Tamara Swaab1; 1University of California - Davis
Predictive processing models have proven useful in explaining first-language (L1) processing and how individuals learn from exposure to language, but we know much less about the ways predictive coding affects processing and learning of second languages. The focus of the current study is specifically on the anticipation and processing of emotional words in sentence context in bilinguals. Previous work in monolingual speakers of Chinese has shown that emotional words can allocate additional attentional resources to facilitate prediction (Ding et al., 2020). Emotional words showed anticipatory effects regardless of contextual constraint. These effects were seen in the presence of a sustained negativity for highly, but not weakly, constraining contexts with neutral verbs. However, it is unknown if these effects are present in bilinguals. While semantic prediction is similar in L2 as compared to L1, bilinguals report less emotional salience in L2 (Dewaele, 2015). Experimental work with ERPs has shown that rapid emotional language processing requires direct social experience— something that L2 speakers often lack (Sendek et al., 2021). I will examine if emotional language has a different effect on predictive processing for those operating in their L2 by manipulating contextual constraint and emotional content both participants’ L1 and L2. Additionally, I will investigate if these effects are influenced by proficiency and immersion. To do so, 48 Chinese/English bilingual participants with immersion experience will read 496 sentences in Chinese or English (translated from Ding et al., 2020), while their neural responses are recorded. ERP components (P1, N400, LPC) will be measured at the same critical nouns in all four conditions and at the verbs preceding the critical nouns, which will be manipulated for emotionality. Analysis will be done using ANOVA to compare ERP components across conditions. Additionally, regression analysis will be used to determine relationships between critical words and their preceding verbs to determine the influence of emotional words on predicted elements, as well as the influence of language proficiency and experience on these processes. If bilinguals process emotion and generate predictions similarly in L1, then results for L1 will replicate those found in Ding et al. (2020). Given that bilinguals show semantic prediction— but reduced emotional salience— in L2, then there are two potential outcomes for L2 processing. If bilingual experience within L2 is sufficient to both generate predictions and rapidly access emotional features, then processing of L2 will replicate findings in L1. If bilingual language experience in L2 is not sufficient to generate of predictions or rapidly access emotional features, then effects will only be seen in L1. Additionally, if emotion effects are the result of direct social experience in a language, then emotional processing effects will correlate with length of immersion in L2 context. This study will expand knowledge of the effects of emotional words on predictive processing to bilinguals, as well as investigate how differences in language experience may interact with these effects for L1 versus L2.
Topic Areas: Multilingualism, Meaning: Combinatorial Semantics