This study concerns the simultaneous acquisition of conditional constructions in Singapore Chinese-English bilingual preschool children, with a focus on the structural aspects of acquisition. Cross-sectional data are obtained from subjects ranging from 2;10 to 6;06. The target languages are Singapore Colloquial English (SCE) and Singapore Colloquial Mandarin (SCM), both of which are contact varieties that are representative of the Low variety in a diglossic situation.  In the acquisition process, these low varieties have become indigenised as native languages that deserve even more to be studied in their own right.

This study adopts the Head-marking and Dependent-marking typological framework in the investigation of the structural features of bilingual acquisition.  Under such a classification, the canonical marking form for conditional constructions in SCE and SCM differs: SCE conditionals are canonically marked by the DM (If) structure; whereas for SCM, the HM (jiu) type is preferred. Conditionals of various marking types are also allowed in both SCE and SCM, some of which are attributable to a large extent to substrate influences.

Two elicited imitation experiment sets (EI 1 and EI 2) were designed and conducted.  EI 1 aimed at examining if the convergence between SCE and SCM that is evident on a societal level is also represented in the developing linguistic repertoires of Singaporean bilingual preschool children.  The results obtained suggest that convergence between the two developing linguistic systems is indeed apparent, reflecting the linguistic sophistication of a bilingual child even at a preschool age.  For both EI 1.1 (SCE) and EI 1.2 (SCM), the accuracy scores in general demonstrate the best performance in response to Absent-marked (AM) constructions, followed by Dependent-marked (DM) ones, then Head-Marked (HM) ones. In terms of substitutions produced, a preference for AM and HM structures is evident. Even though the percentage accuracy for HM (then) constructions are lower than DM ones, the frequency of HM (then) substitutions are high, indicating that HM structures may be acquired with less difficulty than DM ones since they require no subordination.  This is further substantiated by the lack of substitution of DM structures for HM ones.
 
On the individual level, however, a general lack of convergence between the two developing grammars within a single subject is observed in EI 2. In most cases, the marking preference for SCE and SCM tend not to coincide for a particular subject, although interaction between the two developing grammars is detectable, demonstrating that the grammar of one language converges to accommodate to the grammatical constraints of the other.

The diachronic grammaticalization pathway is also established for the innovative conditionals marked by afterward in SCE and dengyixia in SCM. I argue that substrate influence from Hokkien, Teochew and even Cantonese cast a strong impact on the emergence of these innovative conditionals, where markers of temporal relations grammaticalized into markers of conditional relations. Ontogenetic grammaticalization data obtained in EI 1 is used to further substantiate the established diachronic pathway, and to argue for the important role played by the child both in accelerating the developmental process and in nativising the use of such innovative forms.