Syntactic transfer of right dislocation in Cantonese-English bilingual children

Abstract

Right Dislocations (RD) in Cantonese and English are different despite superficial similarity. English RD consists of a complete sentence, immediately followed by an additional NP (John in (1)) that is coindexed with an argument of the sentence: (1) Hei is happy, Johni. Cantonese RD, in contrast, comes in two flavors. The first resembles English RD, except that the dislocated element could be an NP, adverb, etc, as in (2). The second is shown in (3) where part of the sentence occurs after the sentence-final particle (SFP) (Cheung 2009). (2) Keoi jinggoi zau-zo laa, jinggoi he should leave-PFV SFP should ‘He should have left.’ (3) Zau-zo laa, keoi jinggoi leave-PFV SFP he should ‘He should have left.’ At least on the surface level, then, English RD is a subset of Cantonese RD. This comparison sets the stage for the study of bilingual development where structural overlap is considered a condition for cross-linguistic influence to occur. Inspired by studies on the bilingual acquisition of RD which have shown cross-linguistic influence in French-English and French-Dutch bilingual children (Notley, van der Linden & Hulk 2007), we investigate syntactic transfer of RD in Cantonese-English bilingual children and compare the acquisition of English RD by 3 English monolingual and 5 bilingual children, using longitudinal data from CHILDES (MacWhinney 2000; Yip & Matthews 2007). The results show qualitative and quantitative differences between bilingual and monolingual children in the development of English RD. Bilingual children produce more RD than monolingual children. The five bilingual children, especially the Cantonese-dominant children, consistently produce non-target RD in English like (4) and (5), which are structurally more like Cantonese RD. Such patterns are not found in monolingual English children. (4) So green, this one. (Alicia 3;04;12) (5) Come back, you can. (Alicia 2;06;28) Moreover, Cantonese-dominant children have a longer period of non-target RD production as well as a late emerge time of target RD than non-Cantonese-dominant bilingual children. Our preliminary findings indicate that there is cross-linguistic influence in this domain: the non-target RD utterances in bilingual children exhibit transfer effects from Cantonese. Language dominance patterns influence the direction of transfer in this construction, as Cantonese-dominant children produce this non-target structure more frequently. In addition, structural overlap and input ambiguity are factors accounting for the difference between bilinguals and monolinguals.