Preserving the zoeng-construction in heritage Cantonese: Evidence from the Winston corpus
Other conference paper


Full Text

Other information
AbstractHeritage language acquisition is an emerging subfield of research in language acquisition and bilingualism (Benmamoun, Montrul & Polinsky, 2013; Kupisch & Rothman, 2016; Montrul, 2008, 2016). So far only a few studies have discussed the structural properties of Cantonese spoken by first and second-generation immigrants in English-speaking countries (e.g. Dennig & Leung, 2012, 2014; Mai, Kwan & Yip, 2016). Nonetheless, these studies unequivocally found that the zoeng-construction, a pretransitive construction in which the object appears before the relevant verb after a marker zoeng (Matthews & Yip, 2011), is produced to a significantly lesser extent in elicited production tasks by English-dominant heritage speakers of Cantonese, as compared to Cantonese-dominant bilingual baselines. A similar pattern is found in the Mandarin pretranstive (the ba-construction) by heritage speakers of Mandarin (Polinsky, Zhang & Gómez Gallo, 2010). It is proposed that the zoeng-construction is particularly vulnerable to cross-linguistic influence and input conditions due to the lack of structural equivalent in English, as well as its relatively low frequency in child-directed speech, namely, input (e.g. 0.009-0.16% in Hong Kong Cantonese Child Language Corpus (CANCORP), Lee & Wong, 1998; finding in Mai et al, 2016). This paper further addresses the combined roles of cross-linguistic influence and input frequency by examining the development of the zoeng-construction in a Cantonese-Mandarin-English trilingual child in the US, between the ages of 1;7 and 3;7, based on a newly constructed corpus consisting of longitudinal speech data collected from Winston and his adult interlocutors.

The Winston corpus recorded a relatively high percentage of zoeng-sentences by Winston (38 tokens, 1.225% of total Cantonese utterances). Among them only one token is structurally incomplete lacking a lexical verb; the other 37 tokens all appear in the target-like manner, depicting events with temporal boundaries with resultative or directional verb phrases, one of them illustrated in (1). To evaluate the roles of Mandarin influence and Cantonese input, we examined, respectively, the ba-sentences in Winston’s Mandarin and the zoeng-sentences in Winston’s mother, who is the main source of Cantonese input for Winston. Results show that Winston produced an even higher percentage of ba-sentences in Mandarin (51 tokens, 2.155% of total Mandarin utterances) and his mother, who is a highly proficient Cantonese-Mandarin bilingual, produced a significantly greater proportion of zoeng-constructions (49 tokens, 0.865% of total Cantonese utterances), compared with her adult peers in CANCORP. It is likely that the robustness of the zoeng-construction in Winston’s Cantonese is attributable to Mandarin influence and increased occurrence in the input, consistent with the proposal in Mai et al. (2016). Structural properties of the zoeng-construction in the child’s utterances will also be discussed.

(1) Zoeng-construction: Agent-zoeng-Patient-VP
我 想 你 將 嗰 個 蓋 放 翻 去 哩度
ngo5 soeng2 nei5 zoeng1 go2 go3 goi3 fong3 faan1 heoi3 le1dou6
1SG want you ZOENG that CL lid put back to here
‘I want you to put the lid back here.’ (Winston, 3;5)
Acceptance Date02/01/2018
All Author(s) ListZiyin Mai, Yuqi Wu, Kay H. Y. Wong, Tze Yan Law, Virginia Yip
Name of ConferenceThe 4th Workshop on Innovations in Cantonese Linguistics (WICL-4)
Start Date of Conference23/06/2018
End Date of Conference24/06/2018
Place of ConferenceUniversity of British Columbia, Vancouver
Country/Region of ConferenceCanada
Year2018
Month6
LanguagesEnglish-United States

Last updated on 2018-23-10 at 12:37