Activation of non-current language in bilinguals facilitates the implicit learning of novel form-meaning connections

Abstract

Ample evidence suggests that bilinguals are unable to switch off their non-current language(s) in real-time processing, even in contexts when only one of the two languages is relevant. Previous psycholinguistic studies have shown that such language co-activation may result in various cross-linguistic facilitation and interference effects, stemming from bilinguals’ existing linguistic knowledge (Marian et al., 2003; Mishra & Singh, 2016). However, what remains an empirical question is whether the activation of the non-current language(s) also contributes to the development of new linguistic knowledge at the implicit level. Implicit learning, often defined as learning without conscious awareness and intention (Williams, 2009), has been found to be susceptible to cross-linguistic influence (e.g., Leung & Williams, 2014), but the cognitive basis of such influence has not been investigated. This dissertation seeks to bridge the gap between bilingual language activation and implicit learning literatures by investigating [1] whether the activation level of bilinguals’ non-current L1 would influence implicit learning of novel form-meaning connections, and [2] if the covert L1 activation is the cognitive basis of cross-linguistic influence in implicit language learning. Cantonese-English bilinguals and native English speakers participated in an experiment that involved learning the mappings between novel articles (gi, ro, ul, ne) and fire/water semantic categories. Fire and water semantic categories were chosen as the learning target because of the difference in how these are encoded in English and written Chinese. Written Chinese frequently employs phono-semantic compounds, which contain a semantic radical that denotes whether a character is generally related to fire or water (e.g., 河 ‘river’ features the semantic radical 氵 which denotes “water” whereas 炸彈 ‘bomb’ uses 火 to denote ‘fire’), while English rarely explicitly encodes fire and water semantic categories in its spelling except in a few cases of compound nouns (e.g., waterfall, fireball). It was predicted that Cantonese-English bilinguals would outperform native English speakers when the learning target involved associations between novel articles and fire/water semantic categories. Using an incidental learning task and retrospective verbal reports adapted from Leung and Williams (2014), Experiment 1 did not confirm my initial predictions, as both the bilingual and the native English groups learned the target form-meaning connections, and the knowledge they developed may be implicit. Experiment 2 was conducted to test whether the learning effects obtained in the bilingual group were facilitated by the activation level of their L1 Cantonese. A new group of Cantonese-English bilinguals watched a 30-minute English video before the experiment to help them ‘zoom in’ to their L2 and suppress the activation of L1 Cantonese. With the same experimental procedure and materials as Experiment 1, the bilingual group in Experiment 2 did not exhibit implicit learning effects of the same form-meaning connections, which may be attributed to the suppression of their L1 Cantonese. The present study provided

preliminary evidence that is consistent with the assumption that bilinguals’ non-current L1 could influence implicit language learning, and covert activation of non-current L1 may be the cognitive basis of cross-linguistic influence and L1 transfer in SLA.