<p>There has been a surprising dearth of research on language maintenance and shift in New Zealand over the last decade. This thesis addresses this gap by examining incipient patterns of language maintenance and shift in families in two refugee communities in Wellington. Earlier research suggests that immigrants may maintain their ethnic languages in spite of societal factors pressuring language shift for up to three generations. By then, however, language shift is often completed, with the third generation using the majority language only (Fishman 1991). In a largely monolingual country such as New Zealand, this shift may be accomplished in only two generations (Holmes et al. 1993). Understanding the language dynamics at the micro level that eventually lead to language maintenance or shift requires more research into actual language use among family members than traditional methods provide. This investigation therefore uses ethnographic observations, semi-structured interviews and recordings of naturally-occurring interactions between mothers and their children to highlight the challenges involved in transmitting a minority language. Using Spolsky’s (2004) tripartite model of language policy, I investigate family language beliefs, practices, and management in the refugee-background Ethiopian and Colombian communities. The Amharic-speaking Ethiopian community consists mostly of first and second generation members. They first settled in New Zealand in the 1990s and now display awareness of the challenges of maintaining their language. Most Ethiopian parents consider it their responsibility to teach their children Amharic in the home and many have introduced explicit language policies to promote Amharic use. These families exhibit an ‘impact belief’ (De Houwer 1999) which links their positive beliefs about Amharic with their management of family language practices. Nevertheless, in some cases children subvert and contest explicit language management and become primary agents of language shift. Supporting the parents’ efforts, the Ethiopian Orthodox Church provides a social space where Amharic may be used backed by an explicit policy which requires all members to use the language when at church. This policy provides valuable institutional support and cultural capital for Amharic and contributes to the vitality of the language in Wellington. The Colombian community has had a relatively shorter stay in Wellington, with the first members arriving as recently as 2008. Colombian mothers want to transmit Spanish and many seem confident that their children will maintain the language. In particular, they consider the Colombian variety of Spanish to be a source of pride and a core value (Smolicz 1992), as many participants closely link this variety to their Colombian identity. They further capitalise on the prestige of Spanish as a world language that motivates them to use it even outside their ethnic community. However, few families have put in place explicit language policies to use Spanish in the home; instead, many regard it as a more urgent concern that their children learn English. Overall, despite the community members’ positive attitudes towards their ethnic languages, their efforts to transmit these languages appear to be constrained by the fact that English is invested with considerable cultural capital (Bourdieu 1977) in New Zealand. English acquisition often takes priority, particularly for many newly arrived Colombian families. The participants’ refugee experiences, length of residence in New Zealand and the societal status of their ethnic languages seem influential factors on the degree of control they assume over their children’s language practices. Families also dynamically adapt their language policies to the circumstances, for example by introducing an explicit minority language policy after their children have acquired what they consider to be enough English. Despite a strong desire for their children to continue speaking the ethnic language, the parents have many other (non-linguistic) responsibilities and they frequently lack knowledge about “success strategies” for minority language transmission. Moreover, the children often take significant agency by introducing English into the home domain, in some cases even influencing other family members to use it, and thus initiating language shift. The detailed interactional data in this research provides insight into the different ways parents have instantiated their varying language policies and negotiated home language choice with their children. In sum, this research provides insight into language transmission efforts at the family level, and, using data from observations, interviews and recordings of mother-child interaction, describes in detail the unfolding of language maintenance dynamics. The thesis presents valuable insight into the underlying beliefs about Amharic and Spanish, the role of explicit language management strategies, parental socialisation and discourse styles and children’s agency. As the first such research covering two recent refugee communities it will hopefully assist the individual families to socialise their children in a way that enables them to become proficient minority language speakers. This will ensure a linguistically rich future for New Zealand.</p>