An ERP study of the effect of L1-L2 distance on word translation : evidence from Cantonese-English and German-English bilinguals

Abstract

The study aimed to investigate the effect of L1-L2 distance on the degree of semantic involvement during word translation using both behavioural and electrophysiological methodology. In this study, L1-L2 distance referred to the similarity in script between an individual’s L1 and L2. Cantonese-English bilinguals, whose scripts of L1 and L2 are two totally different writing system (large language distance), and German-English bilinguals, whose scripts of L1 and L2 share alphabetic script (short language distance), orally translated written words from their L1 to L2 and vice versa while event-related potentials (ERPs) were recorded. Behavioural results showed that reaction time (RT) of forward translation was longer than that of backward translation for both Cantonese and German participants, supporting the claims of Revised Hierarchical Model. However, the overall results indicated an inconsistency between RT data and ERP data, implying that reaction time might not be a comprehensive indicator of the extent of semantic processing. Furthermore, the difficulty in controlling L1-L2 distance (or script similarity) was revealed. The two bilingual population did not only differ in terms of the similarity of scripts between their two languages, but also the learning strategies of languages. The differential learning strategies of English between Cantonese and German participants provided a logical explanation to the larger N400 amplitude observed in German group during backward translation.