Methodological orientation and design.
The overall approach for eliciting language acquisition data is to use natural and quasi-natural communicative tasks such as narratives and interactive role-playing situations. Identity and attitudinal data are elicited indirectly from the speech patterns of interviews and more directly through interview content, person perception experiments and social network procedures.
We adopt a combined longitudinal/cross-sectional, within-subject and between-subjects design, examining age and language-defined cohorts at two time periods separated by 9-12 months and involving three age groups. Comparisons and analyses are to be conducted for individuals, age groups and for change over time.
Materials, Tasks, Procedures, and Analyses.
Linguistic measures, measures of bilingualism, and assessment of identity and attitudes are outlined in this section. They will be supplemented by parent questionnaires, sociolinguistic interviews and data available from kindergartens and pre-school programs.
I. Linguistic Measures. We will collect 60 minutes of controlled, naturalistic data every two weeks over a 3-month period and transcribe at least 200 intelligible, self-initiated utterances. The following year, this same protocol will be replicated, allowing longitudinal development to be monitored over a 9-12-month period. Individual sessions will include: free play with a uniform set of toys and stimuli designed to elicit information about friends and leisure activities; frog story protocol (Berman & Slobin 1994); narration of a familiar story using picture book stimuli, e.g. the Cat- and Fox-story, already used with 300 Russian- and German-speaking monolinguals (Gagarina 2006). Experimental tasks will target specific forms and functions and will include: a storybook sentence completion task, an elicited imitation task, and an elicited production task (role-playing enactment), etc.
II. Identity and attitude measures. Personal and group identity data gathered both by experimental procedures and the naturalistic data sources described above («linguistic measures.»)
- Language as a window to identity. Sociolinguistic interviews with children will provide data about language choice/preference as a function of setting (home, neighborhood, school), listener (parents, siblings, peers), and topic (language learning experiences, television, computer, literacy practices, sports, leisure activities, favorite possessions). Parents will be queried about reasons for emigration from Russia, initial immigration experiences, extent of integration in the target community, ethnic, national and religious group names and preferences (Walters et al. 2003).
- Identity and attitudes. Two experimental procedures are proposed for investigating identity and social integration: Allard and Landry’s (1986) Beliefs on Ethnolinguistic Vitality Questionnaire (BEVQ) and Anderson’s (1996) person perception procedure. The BEVQ is based on Tajfel’s (1974) theory of social identity and group relations and will be adapted for use with language minority children in the project.
The person perception paradigm (Anderson 1996) has been used widely in social and cognitive psychology. In terms of language attitudes, we have investigated attitudes to non-native accent, names and language proficiency in the social integration of Russian and Ethiopian immigrants in Israel (Walters, 2005). The following core description involves a 3 X 3 manipulation of information about «foreign accent» and «name use» and the child is asked how likely this child will make friends with native born Israeli peers.
In your kindergarten there is a boy named Alex/Avi. He speaks both Hebrew and Russian at home with his parents. He speaks Hebrew with (no Russian accent/a mild Russian accent/a heavy Russian accent), and he (uses his Hebrew name exclusively/alternates use of his Russian and Hebrew names/uses his Russian name exclusively). How likely is it that Alex/Avi will have Israeli friends?
- Social network procedures. Social network research in bilingual settings has made use of both actual language behavior collected via ethnographic observation (Li Wei, 1994) and reported language use via questionnaires (Milroy & Pong, 2000). The social network paradigm has provided scalable data for looking at language choice and social variables including age, generation, and gender. Data on social networks and language use offer a way to scaffold social/demographic information onto interactional data about the relationship between child and interlocutors