This presentation takes the field of heritage languages as a starting point and approaches it from a diachronic perspective. Since 2011, when Polinsky described heritage languages as “a relatively new field in linguistics [that] draws on research on first language and language acquisition, bilingualism (as heritage speakers are a subset of bilinguals), and language attrition”, the field has been thriving with a focus on different language scenarios, theoretical perspectives and theories, data sets and approaches. A lot of existing  studies on heritage languages, which can serve as the basis for relevant linguistic theories,  focus on present-day data that were collected with the help of different methods  like interviews, questionnaires, experimental and judgement tasks (see Aalberse et al. 2019: 87-109). Many of these approaches are however not applicable to heritage language data in the  past. The current presentation therefore sheds light on the  extent to which different methods and theories can be applied to diachronic heritage data. More precisely, the  focus will be on the “Age of Mass Migration” (1850-1920s) when different nations left Europe for North America and other continents. Based on case studies of Swiss migrants (see Auer 2022), the main  focus of the talk will be on (a) language maintenance and shift, (b) dialect and language contact scenarios, as well as (c) identity construction over time. I will illustrate how a range and combination of different sources  allows us to reconstruct heritage language scenarios in the fairly  recent past.  

References 

• Aalberse,  Suzanne, Ad Backus & Pieter Muysken. 2019. Heritage Languages: A Language Contact Approach.  Amsterdam: John Benjamins. 

• Auer, Anita. 2022. The Verticalization of Language Shift: A Historical (Socio)Linguistic Model. In Joshua R. Brown (ed.), The Verticalization Model of Language Shift: The Great Change in American Communities. Oxford: Oxford University Press, 195-214. 

• Polinsky, Maria. 2011.  Heritage Languages. In Mark Aronoff (ed.), Oxford Bibliographies Online: Linguistics. New York: Oxford University Press. 

This contribution addresses the process of destination-language acquisition among individuals of various ethnic backgrounds and different types of recent immigrants who have arrived in Germany, one of Europe’s prominent destinations, within a similar period. The focus is on new immigrants since the early years following their arrival are pivotal for second-language acquisition. During this initial period, the learning curve tends to be steeper than it is in later years, making it a dynamic phase that significantly influences future prospects. The theoretical account builds upon a general and widely applied model of language learning according to which destination-language proficiency is a function of the efficiency with which immigrants learn a new language, the incentives (or motivation) for investing in its acquisition and the degree of exposure to this language. The model will be used to elaborate on various conditions that shape proficiency development after immigrants establish residency in the destination country. The account combines general processes of language learning with conditions that are specific to immigrants of diverse ethnic origins and different types of immigrants, such as refugees and labor migrants. The empirical part encompasses various longitudinal findings derived from multiple data sources that provide insights into the dynamics of language acquisition. These results consistently indicate that exposure to the destination language serves as the primary driver of proficiency, followed by the efficiency of language learning. Incentives, while still relevant, are less important for language acquisition. Overall, the findings suggest that language acquisition follows a general process that exhibits similarities across different types and groups of immigrants.

To research policy positions and public opinion formation on issues such as migration, a crucial step is analysing preferences expressed in e.g. party manifestos, news items, or social media. In general, extracting structured data such as topics or stances from unstructured text is a core task in computational social science. Although Supervised Machine Learning has been part of our toolkit for at least two decades, it has traditionally suffered from data scarcity problems as the complex and shifting nature of social science concepts are unsuitable for large standardized data sets such as common in computer vision and NLP. BERT and other pre-trained models (so called large language models, LLM) may well be a game changer here, as they can offer valid results even with relatively small data sets. We show empirically how BERT and BERT-NLI can be used for valid measurement of political communication concepts in identifying topics and policy positions on migration, and we show how a deeper understanding of the 'downstream' task helps identify and alleviate bias/fairness issues.

Linguistic diversity has long been a common feature of most human societies. However, deep-seated changes in some of the key characteristics of this diversity raise new challenges regarding its management through public policies. Language policy is one of the increasingly necessary tools that can be used in this endeavour. This talk starts out from a selection of key questions in order to review how a language policy perspective addresses these challenges. More specifically, it proposes to (i) position the issue of migration with respect to diversity in general; (ii) assess the claim that the learning of the local language by immigrants is a necessity; (iii) revisit the notion of “integration” in relation with local language acquisition by immigrants; (iv) emphasize the role of non-rights based considerations in relation with the question of heritage language maintenance; (v) warn against certain analytical pitfalls in the examination of the preceding points. The talk concludes with three main recommendations for the selection and design of language policies linked to migration flows.