The emergence of serial verb constructions in child Cantonese

Abstract

This study examines the emergence and development of the Cantonese serial verb

construction (SVC) in children from 1;03 to 4;06 by investigating the naturalistic data

from two longitudinal corpora. This study presents a descriptive account of the

emergence and development of SVCs in early child Cantonese, seeks to explain the

developmental facts from a constructionist usage-based perspective, and compares the

development of SVCs in Cantonese-English bilinguals with that in Cantonese

monolinguals.

It is found that children start to produce SVCs spontaneously at an early age of 1;10

and that the overall frequency of occurrence is low during the developmental period

studied. The early emergence of SVCs is attributed to children’s preference for iconic

structures.

Four surface forms are identified and shown to emerge with a consistent order:

two-verb contiguous (1;10-11) < two-verb non-contiguous (2;00-01) < multi-verb

contiguous (2;02) < multi-verb non-contiguous (2;03-06). Structural and conceptual

complexities are suggested to be the possible factors that influence the order. The

earlier emergence of contiguous forms than non-contiguous forms is explained by the

hypothesis that cross-linguistically unmarked structures tend to be acquired earlier

than the marked ones (O’Grady 2000). Such a generalization is compatible with

constructionist approaches in suggesting cross-linguistic cognitive functional

preferences for language processing.

Children tend to use certain component verbs that express eight main semantic

notions. The study interprets children’s SVCs as concrete instantiations of eight

sub-constructions, which are subsumed by a more abstract high level SVC schema. It

is observed that sub-constructions develop asynchronously, as the developmental

paths of the four more frequently used SVCs (directional, dative, purpose and

resultative SVCs) are more advanced than the four less often produced SVCs

(instrumental, benefactive, comitative and locative SVCs). Developmental paths of

the former are shown to be consistent with Tomasello’s (2003) usage-based account of

language development: from concrete expressions, to pivot schemas, then to

item-based constructions. However, this study does not have enough data to suggest

the emergence of an abstract schema for the high level SVC.

It is found that children imitate adults’ previous SVCs and repeat their own

spontaneous productions frequently. These highlight the roles of the ambient language

and linguistic use to children’s language development.

The overall error rates of SVCs are found to be low. The reasons proposed for error

production, that are, adult input, generalization from item-based constructions and

complexity of target constructions, are considered as evidence to support the

constructionist usage-based approach.

This study shows that Cantonese-dominant bilinguals resemble Cantonese

monolinguals in developing SVCs. Language dominance is invoked to account for the

developmental similarities observed. Only a few code-mixed instances are recorded,

suggesting limited English influence on Cantonese SVCs. It is argued that SVCs are

not a vulnerable domain as they emerge early, and are produced with low error rates,

and are not susceptible to English influence.

By delineating the patterns of emergence and development of SVCs in child

Cantonese and explaining the findings with the constructionist usage-based

framework, it is hoped that this study would contribute to our understanding of child

language development.