Home > Research > Publications & Outputs > The acquisition and use of Mandarin relative cl...

Electronic data

Text available via DOI:

View graph of relations

The acquisition and use of Mandarin relative clauses by monolingual and bilingual children and adults

Research output: ThesisDoctoral Thesis

Published
Publication date2022
Number of pages240
QualificationPhD
Awarding Institution
Supervisors/Advisors
Publisher
  • Lancaster University
<mark>Original language</mark>English

Abstract

Children have been found to understand and use relative clauses (RCs) at an early age. However, not all types of RCs are acquired at the same time, and are used with the same frequency (e.g., Diessel & Tomasello, 2000, 2005). Using corpus-based and experimental methodologies, the three studies presented in this thesis investigate the acquisition and processing of different types of RCs in Mandarin, aiming to understand the mechanisms involved in the acquisition and processing of RC involving varying degrees of complexity.

The first study (Chapter 3) presents a corpus analysis examining the naturalistic
production of Mandarin RCs by Mandarin-speaking monolingual and heritage MandarinEnglish bilingual children (1;00-5;00). The results show that both monolingual and bilingual children produce more object RCs than subject RCs in Mandarin. This is because Mandarin object RCs resemble simple Subject-Verb-Object (SVO) sentences the children had previously acquired, and occur more frequently than subject RCs in their input. Compared to monolingual children, bilingual children produce more object RCs, suggesting that the acquisition of Mandarin RCs is not only facilitated by SVO transitives in Mandarin, but also
SVO transitives in English.

In contrast to the first study, the second study (Chapter 4) reports a subject RC
advantage by looking at the comprehension of Mandarin subject and object RCs in heritage Mandarin-English bilingual children (4;00-10;11) and their vocabulary-matched monolingual peers (4;00-5;09). Using a character-sentence matching task, the results reveal that simple SVO transitives hinder children’s comprehension of Mandarin object RCs by misleading them to interpret the noun phrase occurring first as the head noun. Compared to monolingual
children, bilingual children who are more English dominant make this type of error more frequently for Mandarin object RCs, suggesting that both English SVO transitives and language dominance contribute to cross-linguistic influence.

However, unlike either the subject or object RC advantage shown in children, mixed results are found in the writing of adult Mandarin native speakers (L1) and advanced second language learners (L2) in the third study (Chapter 5). Using conditional inference trees and random forests, the results show that both adult Mandarin L1 and L2 speakers’ selection of subject and object RCs heavily depends on the discourse context that RCs are situated in.

The first and second studies (Chapters 3 and 4) are novel in taking Mandarin RCs
with omitted head nouns into account. In spontaneous speech (Chapter 3), the results indicate that monolingual and bilingual children as young as two can produce Mandarin RCs with omitted head nouns, and the omission of a head noun does not influence the subject-object asymmetry. Similarly, the absence of a head noun does not influence monolingual and bilingual children’s comprehension of Mandarin RCs (Chapter 4), suggesting that they are able to recover omitted head nouns from the context provided.

In addition, the first and third studies (Chapters 3 and 5) also examine the matrixclause positions in which Mandarin RCs tend to occur. RCs that occur in the non-centreembedded matrix-clause position (e.g., The goat saw the horse [that hugged the pig]) are expected to be easier to process than RCs in the centre-embedded matrix-clause position (e.g., The horse [that hugged the pig] saw the goat), as they require lower working memory load (e.g., Gibson, 1998, 2000). Supporting this assumption, in adult Mandarin L1 and L2 speakers’ writing (Chapter 5), non-centre-embedded RCs occur more often than centreembedded RCs. Moreover, the longer the RCs, the higher the possibility they are placed in the non-centre-embedded matrix-clause position. However, in children’s spontaneous speech (Chapter 3), both monolingual and bilingual children do not show a tendency to prefer noncentre-embedded over centre-embedded RCs, which may relate to the short length of the RCs they produce. The shorter the RCs, the less memory load is needed to process
centre-embedded RCs, and therefore the disadvantage of centre-embedded RCs may diminish.

The three studies of this thesis present mixed findings regarding Mandarin RC
processing, but consistently provide evidence to support the usage-based account. That is, the processing of RCs is shaped by an individual’s age and language experience, including input frequency, the related structures that have been acquired, language dominance and the discourse contexts that RCs tend to appear in.