Migration-related multilingualism has been increasing in many European countries over the past few decades (Eurostat, 2024). As a result, classrooms often consist of students who are proficient in multiple languages, including their home language(s) and the local language. The multilingualism of students related to migration contrasts with school systems dominated by majority language(s). However, it can be considered fundamental from various perspectives to include the family languages of multilingual students in the classroom. For example, it can be argued that this is crucial from a holistic understanding of language, which regards language as an integral part of identity (Cummins, 2001). From a cognitive or communication-oriented perspective that views all language-related competencies as part of an individual's entire linguistic repertoire, it is fundamental that all languages can be used flexibly for communication and learning (García 2009). Translanguaging involves seamlessly navigating between languages and treating diverse linguistic repertoires as an integrated system (Canagarajah, 2011; Creese, 2017). Educators' purposeful adoption of strategies to facilitate students' translanguaging, thereby augmenting their learning, is denoted as pedagogical translanguaging (Cenoz & Gorter, 2021; Prilutskaya, 2021). Pedagogical translanguaging aims to improve students' metalinguistic awareness, helping them to organise and elaborate on content for a deeper understanding of the subject matter (ibid.).

However, not all students use their family languages as a communicative resource in otherwise monolingual educational settings, even if they are explicitly invited to (Meyer & Prediger, 2011; Reitenbach et al. 2023; Schastak et al., 2017; Storch & Wigglesworth 2003). From Grosjean's (2008) theory on the language modes of bilingual speakers, three overarching factors can be derived for (not-)using the full language repertoire: individual, communication partners and the teaching context. Individual factors mainly relate to the linguistic repertoire encompassing vocabulary, literacy access, and language preferences. They may vary in multilingual individuals across languages and specific language domains such as academic language. In self-reports, students referred to individual factors by pointing to their language skills and perception of usefulness as well as their language affinity (Reitenbach et al., 2023; Schastak et al., 2017; Storch & Wigglesworth, 2003). According to Grosjean (2008), the multilingual mode is activated during interactions with other multilingual individuals who share more than one language and accept mixed-language interactions. Therefore, the chosen language mode is also influenced by the communication partners’ linguistic competencies, language preferences, linguistic habits, or power dynamics between the speakers (Grosjean 2008). Students mentioned language skills, perception of usefulness, and language affinity when referring to their communication partners as relevant factors for using or not using their entire linguistic repertoire (Reitenbach et al., 2023; Schastak et al., 2017; Storch & Wigglesworth, 2003). Grosjean (2008) argues that contextual factors at the meso-level (e.g., instruction) and macro-level (e.g., educational system and society) account for speakers’ language mode. Classroom interactions, which are influenced by structural elements such as time, space, goals, content, and methods (e.g. task types and materials), constitute meso-level contextual factors that influence (not-)using the full language repertoire. In self-reports, students refer to the teaching context as perceiving the language use being either an offer, obligation or prohibition (Reitenbach et al., 2023; Schastak et al., 2017; Storch & Wigglesworth, 2003).

Overall, there is little evidence in the teaching context. In particular, the relative importance of students’ reasons for using family languages has not been sufficiently empirically investigated. This article therefore uses data from an intervention study to investigate

(1) which reasons increase the probability of students using family languages and

(2) the extent to which these differ in their predictive power.

We used data from an intervention study on reading promotion in German lessons at elementary school. Prior to the intervention, the teachers attended three afternoons of training in small groups on the topic of Reciprocal Teaching (Rosenshine & Meister, 1994) and multilingualism in the classroom from a perspective of pedagogical translanguaging (Cenoz & Gorter, 2021; Prilutskaya, 2021). The teachers implemented the content in their lessons using a structured programme developed by the researchers. To encourage multilingual interaction, they employed three strategies: 1. creating a classroom environment that is welcoming to multilingualism (by using language portraits; Gogolin & Neumann, 1991); 2. using multilingual teaching materials (translations into > 30 languages); and 3. forming small groups based on shared family languages. In the 44 participating primary school classes, 69% of the fourth-graders were identified as multilingual. Those 499 students are included in the analyses. In the post-intervention survey, 62% of participants reported speaking a language other than German during the intervention. Data is available on language competence in the family language, operationalised as vocabulary, measured by the BVAT (adapted from Muñoz-Sandoval et al, 1998). Additionally, perceived academic benefits of multilingualism were measured using a 4-item scale with a Cronbach's α of .872 (e.g., “It helps me to work on tasks”). Attitudes towards multilingualism were also measured using a 4-item scale with a Cronbach's α of .694 (e.g., “I think it's cool if someone can speak more than one other language”). Data on the context were recorded at the student level, including availability of multilingual material (96% of students had access), availability of language partners (64% of students had language partners), and frequency of communication in languages other than German in the classroom prior to the intervention ("yes, very often" = 23.0%, "yes, but only sometimes = 56.2%; "no, never" = 14.2%). The analyses were carried out using binary logistic regression with SPSS 28 (cluster = small groups). Initially, bivariate models were calculated. Then, significant independent variables were tested in a joint model.

The bivariate analyses showed that language competence in the family language (individual score and group average), perceived academic benefits of multilingualism (individual score and group average) as well as availability of material and language partnerships were significantly related to the use of family languages. The individual score of perceived academic benefits of multilingualism had the greatest predictive power (coefficient = .536; p < .001; odd's ratio = 1.710), followed by the availability of language partners (coefficient = .527; p < .001; odd's ratio = 1.694), and the individual score of language competence in the family language (coefficient = .505; p < .001; odd's ratio = 1.657). In the joint prediction model, perceived academic benefit of multilingualism at the individual level (coefficient = .429; p = .016; odd's ratio = 1.536) and language partners (coefficient = .413; p = .002; odd's ratio = 1.511) continued to make significant explanatory contributions. The availability of a language partner increases the likelihood of family language use by 55.1%. Additionally, an increase of 1 in the perceived academic benefit of multilingualism results in a 53.6% increase in the probability of family language use. The joint model accurately predicted whether the family language was used or not in 78.3% of cases. The analyses indicate that teachers can encourage the use of family languages in their classes by starting at a low threshold. An essential step in this regard is to establish small group work with language partners who speak the same languages. It is equally important for students to perceive their family languages as useful in a predominantly monolingual school context. In the classroom, teachers can demonstrate, enable, and motivate this by using subject-specific methods (e.g. Oomen-Welke, 2020 for German classes) or subject overarching approaches such as Linguistically Responsive Teaching (Lucas & Villegas, 2013).

Canagarajah, S. (2011). Codemeshing in academic writing: Identifying teachable strategies of translanguaging. The Modern Language Journal, 95(3), 401–417 Cenoz, J., & Gorter, D. (2021). Pedagogical translanguaging. Cambridge University Press. Creese, A. (2017). Translanguaging as an Everyday Practice. In B. Paulsrud, J. Rosén, B. Straszer, & Å. Wedin (Ed.), New Perspectives on Translanguaging and Education (1-9). Multilingual Matters. Cummins, J. (2001). Language, Power and Pedagogy: Bilingual Children in the Crossfire. Multilingual Matters. Eurostat (2024). Migration and migrant population statistics. https://ec.europa.eu/eurostat/statistics-explained/index.php?title=Migration_and_migrant_population_statistics [18.01.2024]. Gogolin, I. & Neumann, U. (1991). Sprachliches Handeln in der Grundschule [Linguistic practice in primary school]. Die Grundschulzeitschrift, 5, 6–13. García, O. (2009). Bilingual education in the 21st century: A global perspective. Wiley-Blackwell. Grosjean, F. (2008). Studying bilinguals. Oxford University Press. Lucas, T. & Villegas, A.M. (2013). Preparing linguistically responsive teachers: Laying the foundation in preservice teacher education. Theory into Practice, 52(2), 98–109. Meyer, M. & Prediger, S. (2011). The use of first language Turkish as a resource. A German case study on chances and limits for building conceptual understanding. In M. Setati, T. Nkambule & L. Goosen (Ed.), Proceedings of the ICMI Study 21 Mathematics and language diversity (225–234). São Paulo University Press. Muñoz-Sandoval, A.F., Cummins, J., Alvarado, C.G., & Ruef, M.L. (1998). Bilingual verbal ability tests: Comprehensive manual. Riverside Publishing. Oomen-Welke, I. (2020). Mehrsprachigkeit im Deutschunterricht [Multilingualism in German lessons]. In I. Gogolin, A. Hansen, S. McMonagle, & D. Rauch (Ed.), Handbuch Mehrsprachigkeit und Bildung (181-188). Springer VS. Prilutskaya, M. (2021). Examining pedagogical translanguaging: A systematic review of the literature. Languages, 6(4), 180. Reitenbach, V., Decristan, J., Rauch, D., Bertram, V., & Schneider, K.M. (2023). Selbstberichtete Gründe für die (Nicht‑)Nutzung von Familiensprachen beim mehrsprachigkeitssensiblen Reziproken Lehren [Students’ reasons for (not) using their home languages during linguistically responsive Reciprocal Teaching]. Unterrichtswissenschaft, 51, 221–243. Rosenshine, B. & Meister, C. (1994). Reciprocal teaching: A review of the research. Review of Educa-tional Research, 64, 479–530. Schastak, M., Reitenbach, V., Rauch, D., & Decristan, J. (2017). Türkisch-deutsch bilinguale Interaktion beim Peer-Learning in der Grundschule: Selbstberichtete Gründe für die Nutzung oder Nicht-Nutzung bilingualer Interaktionsangebote [Turkish-German bilingual interactions during peer-learning in elementary school. Self-reported reasons for acceptance or rejection of bilingual interaction]. Zeitschrift für Erziehungswissenschaft, 20(2), 213–235. Storch, N. & Wigglesworth, G. (2003). Is there a role for the use of the L1 in an L2 setting? Teachers of English to Speakers of Other Languages, 37(4), 760–770.