I finished data collection for an experimental study in which I look at English influence in Russian speech of Russian-English bilingual children living in Toronto, Canada. Below, I share my findings.
In particular, I looked at embedded yes-no questions in Russian (e.g. Ya sprosil drakona, lyubit li on letat ‘I asked the dragon if/whether he likes to fly’). The syntactic structure for such questions is more complex in Russian: the focused constituent (what you are asking about; here ‘likes’) has to move to the beginning of the sentence, and li ‘whether’ has to come after it. There is also a colloquial version, without li (Ya sprosil drakona, on lybit letat’?), but it has limited use. Adult heritage speakers of Russian (people who partially forgot Russian or did not learn it fully in their childhood) are known to say ungrammatical sentences such as *Ya sprosil drakona, yesli on lyubit letat‘, using yesli ‘if’ (and the structure associated with it), which in Russian is used only in conditional clauses (e.g. Yesli priletit drakon, ja udivlyus’ ‘If a dragon comes, I will be surprized’). The goal of the study was to find out whether bilingual children growing in Canada, in an English-speaking city, make the same error, and why it happens.
I tested 24 Russian-English bilingual children in Canada, age from 5 to 8, and two control groups: 24 monolingual Russian children living in Russia, and 8 Russian adults living in Russia. Tasks included: 1) elicited speech production, in which children were given non-embedded questions and had to convert them to embedded ones; 2) elicited imitation, in which children had to repeat sentences if correct and change any incorrect sentences; 3) telling a story in Russian based on a wordless picture book; 4) telling a story in English based on another wordless picture book. Parents also filled out a questionnaire about their children’s language learning history, exposure to the languages, language use, and proficiency in each language.
Adults produced the li-structure most of the time, and corrected the colloquial structure and the English-like yesli-structure, changing it to the li-structure. However, children in Russia preferred the colloquial structure in the first task, and sometimes omitted li in the second task. Therefore, the li-structure is learned in Russian very late, after the age of 8. Bilingual children produced and imitated all kinds of structures, but almost half of the time they used the colloquial structure, and they also used the English-like yesli-structure about as often. Adults and children in Russia never used yesli, therefore, its use is clearly due to English influence. In the first task, the bilingual children could be divided into three groups: those never using yesli, those always using yesli, and those using both yesli and other structures. In the second task, all bilingual children accepted at least some sentences with yesli as if they were correct; monolingual children rarely accepted them, and adults never accepted them.
In the second task, I added linguistic variables related to those, acquisition of which is required for the li-structure acquisition: the subjunctive clitic by was included to test clitic placement; multiple wh-fronting, to test focus movement; case and agreement, to test general grammatical knowledge in Russian. Bilinguals showed a delay in Russian overall, but improved with age. Mastery of clitic insertion and focus movement predicted development of li, but not use of yesli. Cross-linguistic interference in this case is not due to any grammatical deficits that interfere with use of li.