Development and processing of non-canonical word orders in Mandarin-speaking children

Abstract

BACKGROUND: Cross-linguistically, syntactic structures bearing a word order different than the basic (canonical) word order in the specific language, i.e., non-canonical structures, have been shown to be more difficult than the canonical word order in the language, for children to acquire. This includes monolingual typically developing children (TD children), monolingual children with Developmental Language Disorder (DLD children), heritage children whose first language is the societal minority language, among others. Furthermore, research suggests that unlike in TD children who reach adult-level mastery of these structures with the increase of age, the development of these structures in DLD children and heritage children is not guaranteed given age. However, the underlying causes for the difficulties children, especially DLD children and heritage children, have with non-canonical structures remain to be understood.

AIM: This study investigated how Mandarin-speaking TD children, DLD children and Mandarin-English heritage children produce and comprehend (online and offline) Mandarin non-canonical structures. Specifically, the study aimed to understand the role of linguistic factors, e.g., word order and the presence or absence of morphosyntactic cue, in children’s production and comprehension of non-canonical structures. Furthermore, we also aimed to examine how child-level factors, i.e., chronological age, and input quantity (for heritage children), and their interplay, if any, modulated their non-canonical structures’ development.

METHODS: We adopted a comprehension-to-production priming task to examine the production and a self-paced listening task with picture verification to test participants’ online and offline comprehension of non-canonical structures. To shed light on the role of linguistic factors, we targeted three Mandarin non-canonical structures with differing word order and/or the presence or absence of morphosyntactic cue, i.e., BA-, BEI-, and OSV-constructions. A total of 132 participants took part in the study. Among these participants, 38 were monolingual adults, and the rest 136 participants were 5-to-9-year-old children (40 TD children, 22 DLD children, and 32 heritage children).

RESULTS: An interaction between group and structure types was found that typically-developing groups, i.e., adults, TD children and heritage children, had more difficulties with OSV-constructions than with BEI-constructions in production and comprehension (offline). Furthermore, heritage children had better performance in BA-constructions across tasks. For child-level factors, online comprehension was not modulated by chronological age across child groups, whereas it predicted TD children’s production and offline comprehension but not DLD children and heritage children. Other group-wise differences were summarised as follows: (1) Production: TD children were adult-like. DLD children were more likely to produce BA-constructions, while heritage children were more likely to produce SVO-constructions when their TD peers preferred BEI- and OSV-constructions. (2) Online comprehension: TD children were again adult-like, i.e., they made use of different cues immediately when these cues were available in the linguistic input. However, although DLD children had knowledge of these cues, they had difficulties using more than one cue at a time and relied on the most valid cue in the specific structure. As for heritage children, they used different cues immediately when they were available just like TD children, but they took a longer time reanalysing their initial misinterpretations. (3) Offline comprehension: although TD children were more likely to misinterpret non-canonical structures compared to adults, they outperformed their age-matched DLD children and heritage children.

CONCLUSION: The results suggested that TD children were qualitatively similar to adults in their production and comprehension of non-canonical structures from the age of five onwards. Compared with age-matched TD children, DLD children and heritage children had significantly more difficulties with non-canonical structures. Importantly, the current thesis highlighted the role of the presence or absence of morphosyntactic cues. Additionally, heritage children’s performance showed signs of cross-linguistic influence from the societal dominant language to the heritage language, which further interacted with word order. As for child-level factors, the development of non-canonical structures was modulated by chronological age only for the TD children, whereas language therapy status and input quantity were more important in modulating DLD children and heritage children’s acquisition of non-canonical structures respectively.