1 This is a post-print version of the following article: Rubio Carbonero, Gema; Vargas- Urpí, Mireia; Raigal Aran, Judith (2021). Child Language Brokering and Multilingualism in Catalonia: Language Use and Attitudes in a Bilingual Region. Language and Intercultural Communication. https://doi.org/10.1080/14708477.2021.2005617 Child Language Brokering and Multilingualism in Catalonia: Language Use and Attitudes in a Bilingual Region Gema Rubio Carbonero (https://orcid.org/0000-0002-5009-8291) Gema.rubio@uab.cat Universitat Autònoma de Barcelona Mireia Vargas Urpí (http://orcid.org/0000-0001-6302-581X) Mireia.vargas@uab.cat Universitat Autònoma de Barcelona Judith Raigal Aran judith.raigal@urv.cat Universitat Rovira i Virgili Abstract Children and young people from migrated families often learn host languages faster than their parents might do, and from very young ages they help their parents, families or community members by translating or interpreting, known as child language brokering (CLB). Language brokers need to mediate with different languages in different contexts and are more likely to become aware of the different status of the languages they use or are in contact with than other non-brokers from a migrant background. This article studies the phenomenon of CLB in Catalonia, a bilingual region, from a sociolinguistic perspective. It seeks to understand the linguistic reality of these language brokers, and their attitudes towards the different languages they speak. The results show a complex sociolinguistic picture in which language brokers are clear archetypes of multilingualism, and the tensions and preferences towards each of the languages spoken. It also reflects on the key role CLB plays in the maintenance and revitalization of heritage language/s and culture/s. Keywords: Child Language Brokering, Multilingualism, Language attitudes, Language use, Catalonia Resum Els infants i adolescents de famílies migrades sovint aprenen les llengües de la societat receptora molt més de pressa que els seus pares i mares i, per aquest motiu, des d’edats molt primerenques ajuden la família o altres membres de la comunitat tot traduint o interpretant; tasques que es coneixen amb el nom d’intermediació lingüística per part de menors (ILM). Els intermediaris lingüístics sovint fan de pont entre llengües diverses en contextos diversos i això fa que sigui més probable que prenguin consciència dels diferents estatus de les llengües que empren o amb què estan en contacte, sobretot si els comparem amb d’altres joves de famílies migrades que no fan tasques d’ILM. Aquest article estudia el fenomen de 2 la ILM a Catalunya, una regió bilingüe, des d’una perspectiva sociolingüística. Té l’objectiu de comprendre la realitat lingüística d’aquests intermediaris lingüístics, així com les actituds que mostren vers les llengües que parlen. Els resultats dibuixen una situació sociolingüística complexa, en què els intermediaris lingüístics són clars exemples del plurilingüisme, i on també s’observen les tensions i preferències vers les llengües parlades. L’article també reflexiona sobre el paper clau que la ILM té en el manteniment i revitalització de les llengües i cultures d’herència. Paraules clau: Intermediació lingüística per part de menors, plurilingüisme, actituds lingüístiques, usos lingüístics, Catalunya 3 1. Introduction Children and young people play an important role in migratory processes. Schooling and linguistic immersion in the new society help children learn the official language(s) of the new country faster than their parents. In the Catalan schooling system, Catalan is officially the main language spoken by teachers to protect the language that is considered minoritized. There is a welcome class created for those students who cannot speak Catalan as a resource for them to achieve a basic communicative competence.1 Children also learn Spanish at school since they communicate on a daily basis with native teachers and peers in the two official languages: Catalan and Spanish; and there are also Spanish and English lessons in the school curriculum. However, their parents may not have such a linguistic immersion in their daily life, and this is why they are not usually as communicative competent in the host languages as their children attending school may be. Parents and other relatives with limited language proficiency in local languages often therefore rely on children and teenagers to translate and interpret for them when they need to communicate in the host society. One of the terms used by researchers to refer to this phenomenon is ‘child language brokering’. Antonini (2015:88) provides the following definition: Child language brokering (CLB) denotes interpreting and translation activities carried out by bilingual children who mediate linguistically and culturally in formal and informal contexts and domains for their family and friends as well as members of the linguistic community to which they belong. 1 For a more detailed description of the Catalan linguistic immersion programme, see The language model of the Catalan education system Language learning and use in a multilingual and multicultural educational environment, published by the Catalan Education Department and available online at: http://educacio.gencat.cat/web/.content/home/departament/publicacions/monografies/model- linguistic/model-linguistic-Catalunya-ENG.pdf 4 This article analyses CLB in the province of Barcelona (Catalonia, Spain) from a sociolinguistic perspective, considering language use and attitudes. Catalonia is a region with official bilingualism (Catalan and Spanish). 2 The two languages are in contact, sometimes even in tension, and have different social values and status within the territory, as often happens in regions with diglossia. This situation may generate various attitudes towards the host society’s languages. The article focuses on young people (aged 14-28) who come from linguistically diverse backgrounds and might already be fluent (to varying degrees) in several of their heritage languages, which are languages that are not the dominant ones in a given social context, but the speakers use them because of the existing family, ethnic, cultural or emotional connection with such languages (Polinsky, 2018). This results in a purely multilinguistic environment, where many children from migrated families have to learn not only one, but two host languages to adapt to the new society and succeed at school, while also trying to maintain their heritage languages. In this complex sociolinguistic context, language brokers need to mediate with different languages in different situations. Language brokering enhances (multi)linguistic awareness to a greater extent if compared to other non-broker young people from a migrant background (Valdés, 2014), and, therefore, brokers are more likely to become aware of the different status of the languages they use or are in contact with. Accordingly, this article seeks to understand how multilingualism takes place in an officially bilingual context by exploring the linguistic reality of these language brokers, the attitudes they have towards the different languages they speak (which in turn condition their use) and explore if and how CLB influences children’s acquisition and proficiency in each of the spoken languages. 2 In fact, Catalonia is a region with official trilingualism, also including Aranese (an Occitan dialect spoken in the Vall d’Aran county). 5 2. Previous Research In this section, we will first present the situation of language use and attitudes in Catalonia and, more specifically, among the migrant population, and will then focus on the previous studies that relate CLB and language acquisition and proficiency. 2.1. Language Use and Attitudes among Migrant Population in Catalonia According to Fishman (1965/2000), language choice by speakers of bilingual regions is determined by factors as diverse as group membership, situation or topic of the conversation. In Catalonia, the differences in language use are clear in contexts such as the legal field, where Spanish clearly prevails over Catalan (Vargas-Urpí, 2020). The last language use survey conducted by the Catalan government reflects that Catalan is recovering in certain contexts (Generalitat de Catalunya, 2018), especially if compared to the survey published in 2013. However, it is still less spoken than Spanish in contexts such as healthcare or the Administration, which are also two of the contexts where CLB usually takes place. While bilingualism is official, Catalonia is de facto multilingual: according to the GELA,3 more than 300 languages are spoken in Catalonia daily. Bilingualism adds a layer of complexity to Esser’s (2006: 8) taxonomy of types of social integration and language proficiency (table 1). Integration to the host country Yes No Yes Multiple inclusion / Segmentation / competent monolingual bilingualism segmentation No Assimilation / Marginality / Integration to the monolingual limited Ethnic Group assimilation bilingualism Table 1. Types of social integration and language proficiency (Esser, 2006: 8) 3 Study Group of Threatened Languages (Grup d’Estudi de Llengües Amenaçades). URL: 6 Regarding language acquisition, a person that reaches ‘competent bilingualism’ is considered to have an equal proficiency in both L1 (here referring to the heritage language) and L2 (i.e., the host language). If they only reach a proficiency level in L2, Esser (2006) suggests talking about ‘monolingual assimilation’. On the contrary, ‘monolingual segmentation’ would be used to describe people who only reach a proficiency level in their L1. Finally, ‘limited bilingualism’ refers to those who cannot develop a proficient level in either of the two languages. In both Esser (2006) and Álvarez-Sotomayor & Gómez Parra (2020), a tendency towards monolinguistic assimilation is observed. In our study, though, the possibility of becoming a competent trilingual needs to be added, as well as the possibility of being bilingual in the host languages (Catalan and Spanish) with limited proficiency in the heritage language (L1), as reflected in the following adaptation of Esser’s taxonomy (table 2), which helps representing the cases in a taxonomic way. However, we are aware that there might be several cases in between and different degrees of competency, assimilation, segmentation and marginality. It could also include another row if more than one heritage language was considered. Integration to the host country Yes (L2a + L2b) Yes (L2a OR L2b) No Yes competent competent Monolingual trilingualism (L1 + bilingualism (L1 + segmentation L2a + L2b) L2a or L1 + L2b) (L1) No bilingual monolingual Marginality / Integration to assimilation assimilation (L2a, limited the Ethnic or L2b) bilingualism or (L2a + L2b) Group trilingualism Table 2. Adaptation of Esser’s (2006) proposal for bilingual regions (L2a = Catalan, L2b = Spanish) 7 An exhaustive review of the studies of language attitudes in Catalonia (see, for instance, Alarcón et al. 2013; Bernaus et al. 2007; Huguet et al. 2008; Ianos et al. 2017; Madariaga et al. 2016; Trenchs-Parera 2013) is beyond the scope of this article, but we would like to reflect on some of the key ideas found in Fabà Prats & Torres-Pla (2020), Estors Sastre (2004), Larrea Mendizabal (2016), Cortès-Colomé, et al (2016), Reyes & Carrasco (2018), Sanvincén Torné (2017) and Llompart Esbert (2016), since they can help us frame our results. Fabà Prats & Torres.Pla (2020) analysed the results of the last language use survey conducted by the Catalan government and conclude that the demographic group where Catalan is less spoken is that of the population born outside Spain, i.e. mostly migrants arrived during the first of 21st century (Fabà Prats & Torres-Pla, 2020: 227), who tend to adopt Spanish as language of identification. Estors Sastre (2014) surveyed students of Catalan from four groups of L1 (speakers of Arabic, Chinese, Indo-Iranian languages and Spanish) and 96% of them showed a positive attitude towards learning Catalan. Interestingly, though, 80% of Chinese learners received negative messages from their parents when they knew their children wanted to start learning Catalan. According to Estors Sastre (2014: 164), people from multilingual countries with the predominant ideology that one language is identified with one country tend to perceive Catalan as an inferior language to Spanish (a provincial language). Larrea Mendizabal (2016) analysed language attitudes among Punjabi adult migrants in Catalonia, trying to see if attitudes towards heritage language (e.g., preference for Punjabi or Hind-Urdu) are then transposed to their preferences towards local languages (i.e., Catalan or Spanish), which is also known as a ‘mirror effect’ and has been studied in other cases of minoritized languages (see Badosa Roldós, 2020, for an analysis of 8 Quechua and Amazigh in Catalonia). While some cases of ‘mirror effect’ were identified by Larrea Mendizabal (2016), and some of his informants confirmed a more positive attitude towards Catalan because they appreciated the revitalization effort shared by both Catalan and their mother tongue, Punjabi, in general he also identified cases of negative attitudes towards minoritized languages, as well as cases of acceptance of minoritized languages but preference for the dominant language. Furthermore, all his respondents showed a positive attitude towards dominant languages (Spanish and Hindi-Urdu). Similarly, Cortès-Colomé, et al. (2016) explored how the discovery of Catalan (a minority language) lead some individuals from diverse cultural background that were previously in contact with other minority languages to modify their attitudes and ideologies and develop an alliance with Catalan. They found out that even if at the beginning their participants were reluctant to learn Catalan and preferred the dominant language of the country (Spanish), they finally adopted it as the commonly used language. Regarding the specific studies that focus on the sociolinguistic situation of secondary schools in Catalonia, Reyes and Carrasco (2018), who analysed how Catalan and Spanish co-official status affects students’ progression into post-compulsory secondary education, described the example of a Pakistani student: during the first years of school in Catalonia she learnt Catalan because she attended the welcome class (aula d’acollida), but afterwards she discovered that Spanish was the most used language among peers. This reflects the dichotomy between official linguistic immersion in Catalan, but de facto widespread use of Spanish, especially among pupils. Reyes and Carrasco (2018) also pointed out a process of linguistic decapitalization suffered by students from migrated families, because they tend to worsen their proficiency in some of the languages they knew (e.g., heritage languages or English in 9 the case of Pakistani students), and because they cannot use their own language in any of the subjects. This contrasts with the plurilingualism fostered institutionally4 because, in the end, except for English and Spanish subjects, primary and secondary school is predominantly monolingual (Catalan). Sanvicén-Torné (2017) analysed language transmission in immigrant households and focused specifically on the role of women (mothers and daughters). Among the eight households analysed, only in two of them informants acknowledged using Catalan sometimes at home, while Spanish was more frequently spoken, alongside Arabic in the case of Moroccan families. She also noted that when mothers understood Catalan, but did not speak it, children did not use Catalan to talk to them (Sanvicén-Torné, 2017: 162). In a similar vein, Llompart Esbert (2016), who conducted research in a ‘superdiverse institute’, found out that children from migrant families often teach the host language to their parents, thus reflecting an inverse intergenerational language transmission. In all the cases Llompart Esbert analysed, Spanish was the language transmitted from children to parents. She also pointed out that children acting as CLB had to position themselves in roles that are contradictory to traditional family roles. Parents depend on their children, and this involves an inversion of responsibilities and knowledge. Further to this, her research showed that CLB activities make children develop sophisticated language skills and give them sociolinguistic competence to understand the social meaning of an ongoing activity and adapt linguistic varieties accordingly. 4 There is even the ‘General Sub-directorate of Language and Plurilingualism’ in the Catalan Department of Education. 10 Therefore, we argue that they might be more aware of the status of the different languages they use than other migrants who have not acted as language brokers. In this sense, our study will contribute with a detailed first account of the complex sociolinguistic reality of these child language brokers in a context of two co-official host languages. Concretely, it contributes with an understanding of the languages language brokers use in different contexts and the linguistic attitudes they have towards them, since this is something that has been so far approached in literature of multilingualism in Catalonia only as a secondary result. 2.2. CLB, Language Acquisition and Linguistic Proficiency The experience of language brokering provides the broker with a unique opportunity to develop and practise the necessary skills to fully comprehend, interpret, and translate, using more advanced vocabulary and cognitive ability than other children of the same age who do not broker (Buriel, et al., 1998). This is particularly the case if we consider that brokers are often confronted with situations in which they must deal with otherwise unfamiliar, adult-level language which will probably result in the acquisition of advanced vocabulary. In a study by Tse (1995), adult and adolescent brokers reported that their brokering experiences helped in the development of their first and second languages. Shannon (1990) found that while no formal measurement of acquisition was possible, brokering would very likely result in increased vocabulary knowledge in both languages for her participants. Similarly, almost all of the informants of a study by Tse & Mcquillan (1996) considered language development to be a benefit of brokering. More recent research shows that brokering experiences enhance first language maintenance and metalinguistic abilities (Buriel et al., 1998; Valdés, 2014). This is 11 precisely because language brokers translate and interpret written and oral information from their first language to the host language and these brokering activities offer different opportunities to revitalize their language skills (Valdés, 2014). The continuous contact between the different languages they broker may enrich child language brokers’ heritage language and improve the acquisition of their L2 (Flores et al, 2003; Angelelli, 2016), thus possibly strengthening their bilingualism and biculturalism (Love and Buriel, 2007). Further to this, language brokering may also enhance language awareness, since it may help the brokers to think about language in terms of pronunciation, lexical, and semantic similarities across languages while brokering. In this respect, language brokering experience also provides informal language training that is not available in school curriculums (Orellana & D’warte, 2010). All these previous studies have looked at the interaction of CLB and language development tangentially, while our study offers a wholistic sociolinguistic approach in which the focus is on multilingualism. In particular, we provide an account of how language brokers perceive and use the languages they speak and how acting as language brokers have an impact on language development. 3. Method This study has used both quantitative and qualitative research methods. The approach and data collection techniques have been approved by the Ethics Committee of the researchers’ university and all the informants have explicitly agreed to participate in the project after reading and signing an informed consent form. In the case of minors, their parents were also asked for permission. 12 3.1. Data Collection The following data collection methods have been used to collect data from informants: online questionnaires, interviews and a focus group. 3.1.1. Online Questionnaires In order to have a quantitative and representative general description of the reality of CLB in Catalonia, two online questionnaires were created, with similar questions: one addressed to secondary school teachers, and one addressed to social workers.5 While parents’ perceptions would also be very valuable, these two groups (teachers and social workers) were chosen for the data collection by means of questionnaires because their continued interaction with all kinds of families might have given them a more diverse and holistic perspective of the situation of CLB. Whereas parents’ views would be limited to the specific experience they might have had when their children were brokering. They were created using Google Forms and included six sections. A specific section covered questions regarding language use in CLB interactions and the results are shown in figures 3 to 5 below. To ensure the validity, after the questionnaire was created, a pilot study was carried out with eight secondary school teachers that were not going to take part in the study. These eight participants were asked to complete the questionnaire and to provide feedback about the questions they had been asked in case they noticed anything that was worth mentioning, they did not understand any of the questions or they were not sure about 5 A questionnaire addressing healthcare workers was also initially planned, but due to the complex situation produced by Covid-19 pandemics, this third version of the questionnaire has not been distributed. 13 what was being asked or how to answer. This process was very useful as it led us to improve the questionnaire in terms of accuracy and approach. To distribute the questionnaires addressing the teachers, we asked for the collaboration of the Catalan Department of Education, which provided us with the contact details of the secondary education schools in the province of Barcelona considered of ‘maximum complexity’ or ‘high complexity’, i.e., schools that often need more resources due to the high diversity of their students. Questionnaires were sent to 116 secondary schools. The period for data collection covered 20th January - 28th February 2020 and 60 responses were received from 7 different counties, all located in the province of Barcelona (see also Orozco-Jutorán and Vargas-Urpí, forthcoming). The questionnaire addressing social workers was sent to the main responsible for Social Services in Catalonia, who forwarded it to the 106 basic areas that supply social services for Catalonia. The period for data collection covered 4th February – 30th March 2021 and 96 questionnaires were received. 3.1.2 Retrospective Interviews In-depth qualitative semi-structured interviews were conducted with 19 ex-child language brokers (aged 18-28), i.e., young adults who arrived in Catalonia as children and used to broker for their parents or relatives as children or teenagers. The retrospective questions sought to extract information about interviewees’ experiences as child language brokers, the languages they spoke, as well as their attitudes towards each language, and how CLB affected their language acquisition, among other aspects. Ex-CLBs from four of the main migrant communities in the province of Barcelona were contacted through acquaintances of the members of the research group. The distribution of informants reflects some of the major migrant groups in the province of Barcelona: 14 • 4 interviewees from Amazigh cultural background • 2 interviewees from Amazigh + Arabic mixed cultural background • 2 interviewees from Moroccan Arabic cultural background • 7 interviewees from Chinese cultural background • 4 interviewees from Pakistani cultural background Regarding gender distribution, 2 informants were male and 17 female. Most interviews were individual, but there were two interviews that were conducted with two informants together. Four face-to-face interviews were conducted in places agreed with the informants. Due to Covid-19 pandemic, the other fifteen interviews were conducted online. Most interviews lasted around 45-50 minutes and were conducted in Catalan or Spanish (the interviewees could choose the language in which they felt most comfortable). 3.1.3. Focus Group Finally, we also conducted one focus group with minors aged 14-15 years old who currently perform CLB tasks for their parents. It took place in a secondary school in the province of Barcelona in October 2020. Six students participated in this focus group which was conducted during school time. Two of the participants were Moroccan Arabic, two were Pakistani, one was Chinese, and one was Amazigh. There were three boys and three girls. The focus group followed a semi-structured guide. A reference teacher was present during the focus group to make students feel more confident. The focus group sought to extract information about the kind of situations where children and teenagers are more frequently asked to interpret, the kind of situations they find more difficult or their 15 views towards languages used, among other aspects. Most of the focus group was conducted in Spanish, although some questions were asked in Catalan. 3.2. Data Analysis Percentages were calculated based on the answers collected through Google forms questionnaires and then were used to produce graphics using Microsoft Excel. All the interviews and focus group were audio-recorded and transcribed verbatim and then coded using Atlas.ti software for qualitative analysis. Although the general tone and topic of the conversations were set by the interviewer, the list of codes employed in the subsequent analysis was created inductively. After the codification of four of those interviews, the codes were revised by two members of the research team and adapted. This list was then applied to the following interviews and then new codes were created ad hoc to reflect specific aspects. The codes have then been used to extract quotations and compare responses regarding the same topic. 3.3. Triangulation Each set of data has been analysed separately, and results have been compared and triangulated according to the objectives of this paper. Data triangulation (Denzin, 1978) allows confirmation of the trustworthiness of the data obtained and thus ensures the reliability of the results. The next section reflects the data after this triangulation process. 4. Results Before presenting the results of the questions that analysed language uses and attitudes, we will first contextualize CLB in Catalonia based on the first questions that were included in the questionnaires addressed to teachers and social workers. 16 The first question was asked only to teachers to discover how they overcame language barriers when they had to communicate with adult users (parents) who did not speak Catalan or Spanish. According to teachers (fig. 1), 37% answered that they ask the student whose parents participate in the interview to language broker, while 18% said they look for another student to liaise, which means that in 55% of the situations, 6 young people are used as language brokers. I ask the parents Other to bring someone 5% I ask the student who can interpret to interpret during the during the interview. parent-teacher 24% interview 37% I talk to the direction of I search for another the centre to hire a student that speaks the professional interpreter same language as the or mediator parents or legal guardians 15% 19% Figure 1. Solutions resorted to by teachers when language barriers are encountered According to our participants, figures regarding the frequency of CLB in the educational and the social administration are similar: in both cases, most respondents (72% of teachers, 76% of social workers) affirmed that CLB is used ‘occasionally’ (fig. 2). It is perhaps a bit more regular in secondary schools: 18% of teachers selected ‘regularly’ while only 8% of social workers chose this option; 14% of social workers marked never, while only 3% of teachers picked up this option. 6 In fact, this figure can be increased to 57% if an open answer which explained that ‘it depends’ is included. 17 Never N/A Regularly Never N/A 3% 7% Regularly 18% 14% 2% 8% Occasionally Occasionally 72% 76% Fig. 2(a): Frequency of CLB in high schools Fig. 2(b): Frequency of CLB in social services As we can see, even if CLB seems to occur occasionally according to our informants, it is relevant to note that professional interpreting services are hardly ever used, and ad- hoc interpreting is the most frequently used tool in educational contexts. This might be because professional interpreting services are limited due to budget constraints in Catalonia nowadays, and it is easier and faster for teachers to ask children to intervene. 4.1. Languages Used by (Ex)-Child Language Brokers In this section our aim is to present in general terms the linguistic background of our participants. The table in Appendix I shows the details of each participant, regarding their heritage languages, their languages spoken at home and outside home, and also reveals the language/languages with which each participant feels more comfortable.7 As the table shows, 32% of the participants only speak one language at home (their heritage language), while 44% speak two languages at home, usually the heritage language and one of the host languages. Four participants have reported speaking three languages. There is only one case of a person speaking four languages and another speaking five languages at home. Therefore, the global picture shows that 68% of our informants speak two or more languages at home. In the families where more than one language is spoken at home, typically the participants (ex-CLBs) speak Catalan or 7 Participants with pseudonyms come from the in-depth interviews (former brokers). Participants numbered correspond to the focus group (current brokers). 18 Spanish with their siblings and the heritage language with their parents. Some participants state that their parents prefer to speak with them their heritage language to preserve it, as Farah explains: Extract 1 Farah: Amb els meus pares darija... clar, perquè ells també tenen aquest pensament de, clar, estan estudiant aquí a Espanya, “parlaré amb ells darija perquè així ells aprenen la nostra llengua.” On estudiarem darija si no és a casa? Llavors amb els meus pares darija, amb els meus germans espanyol, per costum... i per comoditat també. (With my parents Darija... well, they also think, well, the children go to school here in Spain so “I will speak with them in Darija, so that they learn our language”. Where are we supposed to learn Darija but at home? Then, I speak Darija with my parents and Spanish with my siblings, because we are used to it…and it is easier too.) It is important to remark that 40% of the participants speak at least one of the host languages at home, mostly among siblings, and in all cases but one the host language spoken at home is Spanish. There is also one person that speaks both host languages. Regarding the languages spoken outside home, we obtained 19 answers. 36.8% speak one of their heritage languages outside home. As for host languages, 47.3% speak both; seven people speak only Spanish and three people only Catalan. When it comes to the use of the host languages, it seems that Spanish is more often used both inside and outside home by former and current brokers, as also mentioned in Llompart Esbert (2016). In particular, P6 explained that ‘in the hospital you don’t need to speak Catalan, so I speak Spanish’, and P1 then added that ‘some doctors talk in Catalan if you choose to speak Catalan’, which might reflect the fact that sometimes service providers adapt to young language brokers’ language choice. 19 In order to triangulate this information, we asked teachers and social workers which host language is most frequently used in CLB interactions. The answers from both were, again, similar, as figure 3 shows: 53-54% perceive that young people tend to interpret into Spanish, while 29-32% believe both languages are used equally. Catalan as a brokering language is marked by a smaller proportion: 10% of teachers and 15% of social workers. Catalan N/A Catalan N/A 10% 5% 15% 2% Spanish Spanish Spanish 53% Spanish 54% and and Catalan Catalan equally equally 32% 29% Fig. 3(a): Host language used in high schools Fig. 3(b): Host language used in social services According to the former and current brokers and the service providers’ answers, there seems to be a majority tendency towards the use of Spanish, compared to the use of Catalan, but we cannot talk about a clear monolingual assimilation tendency in Esser’s terms. What seems to be clear is that if we take a look at our adaptation of Esser (2006) in table 2 above, language use as reported by our participants inside and outside home seems to be in line with that of competent bilingualism, trilingualism and even quadrilingualism. Accordingly, competent multilingualism (to different degrees) characterizes all our participants, but it is difficult to assess the real level of use of each of the languages because, depending on the different contexts, one language will be more dominant than others, as Amira explains: 20 Extract 2 Amira: El castellà, sí perquè començant per al mòbil que és el tema de les xarxes socials, totes són en castellà que són com una cosa del dia a dia. Llavors per escriure, la universitat, tot és en català, castellà i tal. Llavors, l’únic moment en el que utilitzo àrab és l’assignatura d’àrab que faig a la universitat, que és la assignatura d’Àrab, i llavors parlo el darija a casa, tot i que tinc algunes companyes també amb les que a vegades anem barrejant. (Spanish, yes, because starting with the mobile and the social media, everything’s in Spanish. Then, to write, at university, everything’s in Catalan and Spanish. So the only time I use Arabic is the Arabic course at university and I speak Darija at home, and also with a few classmates.) We can see in this example how language use is fully determined by the context in which the persons are embedded, particularly where the person is and who s/he is talking to, which is one of the reasons for language choice, according to Fishman (1965/2000). Therefore, it is not an easy task to fully comprehend the use of each language, since it depends on different contextual variables. However, looking at the attitudes former brokers have towards each language they speak may help understand this complex multilinguistic framework, since such attitudes eventually condition the use of each language. 4.2. Attitudes towards the Spoken Languages by (Ex)-Child Language Brokers We received seventeen answers from the in-depth interviews to the question regarding attitudes towards the different languages the participants speak. Seven people identified only Spanish as the language they felt more comfortable with and three selected only both host languages. Two people named their heritage language and Catalan as their preference, two people selected their own heritage language, and one person chose only Catalan as the one they would preferably use. We must recall that we did not evaluate 21 their proficiency in these languages but rather asked for their language use and preferences. In Esser’s (2006) terms, we could say that, regarding attitudes, there is a majority tendency towards monolingual assimilation (Spanish), followed by a bilingual assimilation (Spanish and Catalan). Only three people reported feeling more comfortable with Catalan than with Spanish. In particular, two of them preferred their heritage language and Catalan and one person preferred only Catalan. Precisely, the very same three people that mostly speak Catalan at home and outside were the same who chose Catalan over Spanish. Such a preference in these three cases might be connected with the geographical zones where these participants live, since they are zones in which Catalan is usually the working language, as Karim puts it: Extract 3 Karim: Et diré que jo parlo més català que un altre idioma. De fet, des de que he vingut, és un idioma que m’agrada molt i el parlo en el meu dia a dia... a totes les feines que he anat treballant han sigut per Osona8 o amb gent de Girona, i ja comprendràs que sí o sí has de parlar català. Allavorens és un idioma que el parlo més que els altres. (I speak more Catalan than any other language. Since I arrived, I’ve always liked this language and I speak it in my everyday life… All the jobs I’ve had have been in Osona or with people from Girona, so you have to speak Catalan. So I speak it more than other languages.) Nevertheless, it is worth mentioning that the three persons that selected Catalan, which is a minoritized language, as the one they spoke more often and felt more comfortable with, also had a minoritized language as mother tongue (one Punjabi and two Amazigh, respectively). Therefore, since they understand what a minoritized language is, they 8 Osona is a county located at a distance of 50 km from Barcelona. It belongs to the province of Barcelona. 22 might be more sensitive to using and promoting other minoritized languages the ‘mirror effect’ already mentioned by Mendizabal (2016), Cortès-Colomé et al. (2016) and Badosa Roldós (2020), among others. However, we do not have enough data to prove this is the case, precisely because there is a minority of participants that selected only Catalan (and not Spanish) as their preferred host language. Regarding the greater use of Spanish, Dounia wonders: ‘¿el catalán para qué lo utilizamos? Pues para las clases, no más’ (Why do we use Catalan? Just for the classes). Similarly, Zhousi also agreed that ‘you need to know Catalan if you are attending university in Catalonia’, as this is the language most frequently used in tertiary education according to her. Jameela’s answer is similar, because she relates Catalan to studies and to the work done in the local administration: Extract 4 Jameela: Entre los profesores hablamos catalán; cuando hacía prácticas en el ayuntamiento hablaba solamente en catalán, porque creo que allí no les gusta hablar castellano, y entonces yo llevaba un año hablando catalán con ellos. (We speak Catalan among the teachers; when I did a placement with the town hall I only spoke in Catalan, because I don’t think they like to speak Spanish there, and I’d been speaking Catalan with them for a year.) In this example, Jameela reports that there is an expectation in the local administration to speak Catalan and this is why she adjusted her language use to what she thought was expected of her. These previous examples also reflect the limited and selective use of Catalan by some of the informants. This is the case to such an extent that they might not even consider Catalan as one of the host languages. For instance, when asked if she believed CLB 23 could help improve languages, Zhousi answered positively, but she only referred to Chinese and Spanish, and spontaneously excluded Catalan from her answer. One reason for this tendency towards a broader use of Spanish and a limited use of Catalan might be connected to what people from a migrant background think the host society expects from them. For instance, when we asked Li which were the languages she used regularly, she only answered Chinese, English and Spanish, in that order, even though she admitted that she used to speak more Catalan when she first arrived in Catalonia, but she no longer used it that much to avoid uncomfortable questions, as reflected in extract 4: Extract 4 Li: Pero los primeros años casi que utilizaba más el catalán. El catalán fue el primer idioma que aprendí en el Aula de Acollida y no tenía muchos amigos españoles en ese momento… siempre estaba con mis profesores, pues con mis profesores siempre les hablaba catalán. Fue después en la universidad que conocí a gente de toda España y de todo el mundo, pues allí empecé a usar más el castellano que el catalán (…). No sé, cuando… si hablas catalán con tu cara asiática la gente se sorprende [riu]… y la gente empieza a preguntarte muchas cosas. Pues muchas veces quiero evitar que me pregunten y por eso les hablo castellano… (The first years I used Catalan more. Catalan was the first language I learnt in the Welcome Class and I didn’t have many Spanish friends at that time… I was always with my teachers and I always spoke Catalan with them. Then, at university, I met people from all over Spain and the world, and began to use more Spanish than Catalan (…). People are surprised when you speak Catalan with an Asian face [laughs]… And people ask you many things. Often I want to avoid the questions so I talk in Spanish…) Accordingly, Li reports here that people in the host society tend to expect people from a migrant background to speak Spanish and that may cause them to adjust to such expectations in order to avoid uncomfortable questions or having to provide too many 24 explanations. This is also reflected in Solana, et al. (2018: 124), who explain that this tendency to use Spanish to address people from a migrant background (especially if this is reflected by their phenotype) results in a feeling that swings between reject and confusion, because they feel they are treated as outsiders and aliens in the society where they are living in. Finally, regarding attitudes towards languages, it is remarkable that all our participants seem to agree that speaking different languages is an advantage. In particular, 70.6% of our informants see it as a professional asset that can be taken advantage of, as Aisha points out: Extract 6 Aisha: Mai sabria que l’urdú em pogués donar feina, saps? [Riu] he treballat més, no pel fet de tenir el títol d’Humanitats, però pel fet de saber llengües... o de saber urdú... llegir i escriure a part de parlar. Perquè també he fet traduccions, és una manera de estar també vinculada amb la llengua. (I never thought that Urdu could give me work [laughs]. I’ve worked more not because I studied humanities but because I know languages, or Urdu, reading and writing as well as speaking. I’ve also done translations; it’s another way of being linked to the language.) Aisha understands that her heritage language is something valued in the professional world and, in particular, remarks that having acted as a language broker has allowed her to be more connected to her language. This leads us to the next question we would like to explore: how having acted as a language broker may have an impact on the acquisition and proficiency of the different languages the (ex)broker speaks. 4.3. CLB Impact on the Spoken Languages Regarding host languages, when asking teachers if they thought CLB has a positive effect on the acquisition of Catalan, as figure 4 shows, 45% thought this happened 25 sometimes, 27% did not believe so and only 8% agreed. In the case of social workers, 51% did not think there was a positive effect on language acquisition, 31% answered ‘sometimes’ and, again, the smallest proportion is for the ‘yes’ option, with only 6%. N/A Yes, significantly N/A Yes, significantly 20% 8% 12% 6% Sometimes 31% Sometimes No 45% 27% No 51% Fig. 4(a) and 4(b): Positive impact on Catalan acquisition (left: teachers’ answers; right: social workers’ answers) In the equivalent question, but regarding Spanish, answers are more positive, which is also reasonable because Spanish is used more frequently as a brokering language according to respondents (see fig. 3). Again, teachers tend to believe there is a higher positive impact in general: 23% of teachers agreed with this, and 42% marked ‘sometimes’, while only 13% said no (fig. 5A). In the case of social workers, ‘no’ still occupies the highest proportion (35%), but 34% thought there is ‘sometimes’ a positive impact, and 19% answered ‘yes, significantly’ (fig. 5B). N/A N/A Yes, Yes, 22% 12% significantly significantly 23% 19% No No 13% 35% Sometimes Sometimes 34% 42% Fig. 5(a) and 5(b): Positive impact on Spanish acquisition (left: teachers’ answers; right: social workers’ answers) 26 Participants in the focus group also felt that CLB helped them improve in Spanish, but not so much in Catalan, because they tend to use Spanish to communicate with public services, as already discussed above. Regarding young adults’ responses, extracted from the interviews, 26% felt that CLB had helped them improve in Catalan and Spanish. In the rest of the interviews, this was not mentioned. Among informants who thought it had had this positive effect, some mentioned that CLB helped them learn new words, as Zhousi explains in extract 7. Extract 7 Zhousi: Sí, sí, tanto uno como el otro, las dos lenguas. Porque si no lo haces nunca te toca ampliar ese vocabulario, porque no sabría cómo se nombran esos documentos, ni en chino ni en español, porque es un vocabulario muy técnico que tienes que ir a aprender al momento. (Yes, both languages. Because otherwise you don’t expand your vocabulary, I wouldn’t know how to call these documents in Chinese or in Spanish, because it’s very technical vocabulary that you have to learn at the time.) However, this extract is an example of the only reported linguistic challenge by our participants: the difficulty of finding the appropriate specialised or technical word in any of the spoken languages when they needed to broker in contexts where these words were likely to appear (mostly in medical consultations or when dealing with the administration). Therefore, vocabulary acquisition was a challenge that ended up having a positive effect on the linguistic skills of our participants. Together with that, Yue mentioned an increased sense of confidence thanks to the CLB experience: Extract 8 Yue: Yo creo que sí, porque cada vez aprendes algo más. Quieras o no aprendes algo más… en términos de vocabulario. Yo creo que la parte buena, la parte positiva es que también te abres más… yo me veía como el cambio en este… porque yo era muy tímida, me daba miedo hablar al público y con esto pues como que me obligaban. (I think you learn 27 more each time, whether you want to or not, in terms of vocabulary. The positive part is that you open up more. I saw this change, because I was very shy; I was frightened to talk in public and this forced me to.) This increased sense of confidence might well be related to a greater development of pragmatic skills, which is mentioned by Buriel et al. (1998), especially when compared to other children or young people of the same age. However, not all the informants have a clear perception of these positive effects, as we see in Rashmir’s opinion: Extract 9 Rashmir: Crec que no t’ajuden, o sigui, a aprendre... bueno, sí que és cert que vas agafant paraules, vocabulari, sempre quan vas a un lloc així, no sé, acabes aprenent una paraula nova. Però no t’ajuda a aprendre així... no t’ajuden a que tinguis un vocabulari més ampli a mesura que vas treballant... (I don’t think they help you to learn, although it’s true that you always end up learning new words when you go to a place like this. But it doesn’t help you to learn, or to have a broader vocabulary as you work…) To sum up, despite the fact that most informants share the feeling of having learnt something due to their experience in CLB, none of them acknowledges a great positive impact of this practice regarding a major development of competencies and skills in the host languages; rather, they limit it to vocabulary acquisition. However, the picture is very different when it comes to heritage languages. 78.9% of our informants acknowledged that thanks to acting as language brokers they improved their communicative skills in their heritage language. As Farah puts it: Extract 10 Farah: En el meu cas sí que ha influenciat una miqueta el fet que hagi estat intermediadora, eh… perquè he après més àrab i m’he interessat en aprendre’n més encara [...] perquè jo l’àrab que parlava era d’haver acabat la primària només, saps? I he anat actualitzant el meu àrab aquí també a través d’aquestes feines que he fet. (In my case being an 28 intermediary influenced me a little, because I learnt more Arabic and became more interested in learning it, because the Arabic I spoke was only from the end of primary school. Doing these jobs helped me to update my Arabic.) Only one participant explicitly said that she found no improvement in her heritage language skills and the rest did not talk about this issue at all. We can see in the example that Farah talks not only about improving her language skills, but also about having more interest in her heritage language. Precisely, 36.8% of our participants also acknowledged that acting as language brokers awoke their interest in their heritage language and that allowed them to be more connected not only to the language, but also to their culture, as Hana explains in this extract: Extract 11 Hana: Sí, por supuesto, y a parte, yo creo que también tiene que ver mucho con… una vez conoces la lengua, tiene mucho que ver con la identidad, con tus orígenes, quién eres… despierta en ti un cambio ¿no? También despierta en ti cierto sentimiento de decir “soy de aquí, pero también soy de otro sitio” a nivel cultural, lingüístico, de entendimiento… es un proceso bastante amplio. Yo precisamente el amazigh no lo hablaba y empecé a aprender hace tres años y me ha generado bastante… o sea, te permite tener tu comunidad y conectar con ella, que el idioma es prácticamente todo, también a leer ciertas cosas que… yo creo que generas un vínculo… es una carencia que tienes y empiezas a tener un rencuentro con tu lengua. (Yes, of course, and I think once you know the language it’s related to your identity, your origins, who you are… It awakens a change in you and also a certain feeling of saying ‘I’m here, but I’m also from somewhere else’. It’s quite a broad process on a cultural, linguistic and understanding level. I didn’t speak Amazigh and I began to learn it three years ago. It allows you to have your community and connect with it. Language also helps you to read certain things. You create a link … It’s a shortcoming you have and you begin to rediscover your language.) 29 Accordingly, because she needs to use her heritage language when she acts as a language broker, she acknowledges an increasing linguistic awareness that goes hand in hand with a cultural awareness, which ultimately facilitates the connection with the home culture and community. 5. Conclusions The results obtained through different data collection strategies shed light on different realities. First, current and former child language brokers reflect real multilingual practices, with no examples of monolingual segmentation according to Esser’s (2006) taxonomy. At home, it is interesting to note that 68% of informants (17) speak two or more languages. Globally, 9 informants seem to reflect competent trilingualism (with competency in Spanish, Catalan and their heritage language) or competent bilingualism (10 informants), with a clear tendency towards a major use of Spanish to communicate with the host society and as a brokering language —which is in turn related to external factors, such as Spanish being a more widespread language in urban areas (where most research was conducted). In three cases, a preference for Catalan could be related to the mirror effect suggested in previous studies (Cortès-Colomé et al., 2016; Mendizabal, 2016; Badosa Roldós, 2020), but more research is needed to confirm this. Answers from questionnaires also confirmed a major use of Spanish as a brokering language, in both education and social services. In the case of education, this contrasts with some comments made by ex-CLBs, as they often related Catalan to the most frequently used language in education and academic environments, as well as in the administration. In this situation, we might well think of Catalan as the language for instruction, but Spanish as the language for interactions with the teachers outside the classroom, although this is only a hypothesis that should be confirmed in further research. 30 This apparent contradiction shows that the Catalan education system seems to oscillate between two pressing needs: on the one hand, the need to protect Catalan as a minoritized language; on the other hand, the need to promote multilingualism. While Catalan immersion might be the reason for most of our informants acknowledging that they understand Catalan to various degrees (despite feeling more comfortable in Spanish), CLB situations might be one of the few instances where real multilingualism has a place in schools. In this respect, current and former child language brokers in our study are clear archetypes of multilingualism, and this is rendered visible through CLB practices. Regarding the impact of CLB on the acquisition of host languages (either Catalan or Spanish), the general feeling among teachers and social workers was that it is only positive sometimes. In particular, they thought that CLB was more helpful to improve Spanish than Catalan, which is reasonable if we consider that Spanish prevails over Catalan in CLB situations. When asking ex-CLBs about this, many of them mentioned very specific aspects, such as vocabulary learning, especially regarding certain topics, something that has already been put forward by Buriel and DeMent (1993). Only one of the interviewees mentioned a major sense of confidence in L2, which could be related to an improvement in pragmatic skills identified by Buriel et al. (1998). If we consider the impact of CLB on L1 acquisition and proficiency, our results are in line with those of Buriel et al. (1998) and Valdés (2014), in the sense that, as reported by our participants, CLB had a positive impact on the maintenance and improvement of skills in their heritage language/s. Some also acknowledged that thanks to these tasks as language brokers they increased their interest not only in their heritage language/s, but also in the culture. In this respect, we could speak of a revitalised biculturalism (Acoach & Webb, 2004), since language brokers have an advantage compared with other 31 immigrant teenagers, because they experience increased exposure to the shared patterns and belief systems of the host culture, while maintaining connection with their native culture. This article sheds light on one topic that has only been explored tangentially in the literature of multilingualism in Catalonia: it confirms the connection between language brokering and language awareness and contributes with a detailed account of the language use and attitudes of these brokers towards the heritage and the host languages in a context of two co-official host languages, which, as we have seen, complicates the sociolinguistic picture. Another key contribution of this paper is that it reflects on the importance of CLB as a revitalizer of the heritage language and culture. It is also worth noting the value of qualitative research on addressing these complex issues and how quantitative approaches have proved to be also valuable to triangulate our results. Our results have implications for the understanding of multilingualism and how different languages are used when there are several tensions interacting such as: the protection of minoritized language(s), the preservation of heritage language(s) and the dominance of the language(s) that are spoken by a majority in a given context. We are aware that both positive and negative emotions can emerge from performing CLB tasks (Arumí and Rubio, forthcoming) and there a number of challenges that derive from it. However, it is out of the scope of this paper to reflect on such issues. Rather, from the sociolinguistic perspective, when it comes to the impact of CLB on the languages spoken, this study has only revealed positive effects for multilingualism, which goes in line with what Flores et al (2003) and Angelelli (2016) had already pointed out. 32 Finally, as a part of a work in progress of the project in which this study is embedded, we have also carried out interviews with the families of current and former language brokers. We believe this paper will be nicely complemented by the forthcoming study on the sociolinguistic attitudes of families, which could further ascertain our main findings. Acknowledgements This article is part of the project Young natural interpreters: Child Language Brokering in education, social services and healtchare (Grant No. RTI2018-098566-A-I00) funded by MCIU/AEI/FEDER, UE. References Acoach, C. L., & Webb, L. M. (2004). The Influence of Language Brokering on Hispanic Teenagers’ Acculturation, Academic Performance, and Nonverbal Decoding Skills: A Preliminary Study. Howard Journal of Communications, 15(1), 1–19. https://doi.org/10.1080/10646170490275459 Alarcón, A. & Garzón, L. (2013). Children of immigrants and social mobility in officially bilingual societies. The case of Catalonia. Spanish in Context, 10(1). 92– 113. https://doi.org/10.1075/sic.10.1.04ala Álvarez-Sotomayor, A. & Gómez-Parra, M. E. (2020). Modos de integración lingüística en el alumnado de padres inmigrantes en España: un estudio de caso. Lengua y migración, 12(1), 77-99. Angelelli, C. V. (2016). Looking back: A study of (ad-hoc) family interpreters. European Journal of Applied Linguistics, 4(1), 5-31. Antonini, R. (2015). Child language brokering. In: Pochhacker F. (Ed.), Routledge 33 Encyclopedia of Interpreting Studies (48). Routledge. Arumí, M. & Rubio, G. (forthcoming). Reflecting on past language brokering experiences: how they affected children’s and teenagers’ emotions and relationships. Badosa Roldós, A. (2020). Ideologías lingüísticas de las comunidades quechua y amazigh en Cataluña: el efecto espejo. Lengua y migración, 12(1), 127-173. Bernaus, M., Moore E. & Cordeiro Azevedo, A. (2007). Affective Factors Influencing Plurilingual Students’ Acquisition of Catalan in a Catalan-Spanish Bilingual Context. The Modern Language Journal, 91(2). 235–246. https://doi.org/10.1111/j.1540-4781.2007.00542.x Buriel, R., Perez, W., de Ment, T. L., Chavez, D. V., & Moran, V. R. (1998). The Relationship of Language Brokering to Academic Performance, Biculturalism, and Self-Efficacy among Latino Adolescents. Hispanic Journal of Behavioral Sciences, 20(3), 283–297. https://doi.org/10.1177/07399863980203001 Cortès-Colomé, M., Barrieras, M., & Comellas, P. (2016). Changes in immigrant individualsʼ language attitudes through contact with Catalan: the mirror effect. Language Awareness, 25(4), 272-289. Denzin, N. K. (1978) Sociological methods: A sourcebook. New York, NY: McGraw- Hill. Esser, H. (2006). Migration, Language and Integration. Berlin: WZB. Estors Sastre, L. (2014). Les actituds lingüístiques segons l’origen dels aprenents de català com a llengua d’acollida. Treballs de Sociolongüística catalana, 24, 153- 171. 34 Fabà Prats, A. & Torres-Pla, J. (2020). De la llengua inicial a la llengua d’identificació a Catalunya: qui adopta el català i qui se n’allunya. Treballs de Sociolongüística catalana, 30, 212-232. Doi: 10.2436/20.2504.01.168 Fishman, Joshua, 1965/2000. Who speaks what language to whom and when?. In Li Wei (ed.). The Bilingualism Reader. London & New York: Routledge, 89-106. Flores, G., Laws, M. B., Mayo, S. J., Zuckerman, B., Abreu, M., Medina, L., & Hardt, E. J. (2003). Errors in medical interpretation and their potential clinical consequences in pediatric encounters. Pediatrics, 111(1), 6-14Generalitat de Catalunya (2018). Els usos lingüístics de la població de Catalunya. Resultats de l’Enquesta d’usos lingüístics de la població. 2018. Departament de Cultura. URL: https://llengua.gencat.cat/ca/serveis/dades_i_estudis/poblacio/Enquesta- EULP/Enquesta-dusos-lingueistics-de-la-poblacio-2018/ Huguet, A. & Janés, J. (2008). Mother tongue as a determining variable in language attitudes. The case of immigrant Latin American students in Spain. Language and Intercultural Communication, 8(4). https://doi.org/10.1080/14708470802303082 Ianos, M. A., Huguet, A. & Lapresta-Rey, C. (2017). Attitudinal patterns of secondary education students in Catalonia. The direct and moderator effects of origin. Journal of Multilingual and Multicultural Development, 38(2). 113–129. https://doi.org/10.1080/01434632.2016.1177060 Larrea Mendizabal, I. (2016). Les actituds lingüístiques dels immigrants penjabis adults a Catalunya. PhD dissertation. Universitat Pompeu Fabra. URL: http://hdl.handle.net/10803/402437 Llompart Esbert, J. (2016). Pràctiques plurilingües d'escolars d'un institut superdivers: de la recerca a l'acció educativa. PhD dissertation. Universitat Autònoma de 35 Barcelona. URL: https://www.tdx.cat/handle/10803/399835 Love, J. A., & Buriel, R. (2007). Language brokering, autonomy, parent-child bonding, biculturalism, and depression: A study of Mexican American adolescents from immigrant families. Hispanic journal of behavioral sciences, 29(4), 472-491. Madariaga, J.M., Huguet, A. & Janés, J. (2016). Language attitudes in Catalan multilingual classrooms: Educational implications. Language and Intercultural Communication, 16(2). 216–234. https://doi.org/10.1080/14708477.2015.1118111 Orellana, M. F., & D’warte, J. (2010). Recognizing different kinds of ‘head starts.’ Educational Researcher, 39(4), 295–300. https://doi.org/10.3102/0013189X10369829 Orozco-Jutorán, M. & Vargas-Urpí, M. (forthcoming). Children and teenagers acting as language brokers: the perception of teachers at secondary schools. Across Languages and Cultures. Polinsky, M. (2018). Heritage languages and their speakers (Vol. 159). Cambridge University Press. Reyes, Ch.; Carrasco, S. (2018). Unintended effects of language policy on the transition of immigrant students to upper secondary education in Catalonia. European Journal of Education, 53 (4), 514-527. doi.org/10.1111/ejed.12304 Sanvicén-Torné, P. (2017). Dones i transmissió lingüística en llars d’origen immigrant. Impactes, reptes i oportunitats. Treballs de Sociolingüística Catalana, 27, 149-174. doi: 10.2436/20.2504.01.130 Shannon, S. M. (1990). English in the Barrio: The Quality of Contact among Immigrant Children. Hispanic Journal of Behavioral Sciences, 12(3), 256–276. 36 https://doi.org/10.1177/07399863900123002 Trenchs-Parera, M. (2013). Les vivències sociolingüístiques i el multilingüisme dels joves d’origen xinès a Catalunya. Revista de didàctica de la llengua i de la literatura,60. 28–39. Tse, L. (1995). Language Brokering among Latino Adolescents: Prevalence, Attitudes, and School Performance. Hispanic Journal of Behavioral Sciences, 17(2), 180– 193. https://doi.org/10.1177/07399863950172003 Tse, L., & Mcquillan, J. (1996). Language Broker Culture, Language, and Literacy: The Effects of Child Brokering on Language Minority Education. Paper presented at the Annual Meeting of the American Education Research Association. URL: https://files.eric.ed.gov/fulltext/ED394357.pdf Valdés, G. (2014). Expanding definitions of giftedness: The case of young interpreters from immigrant communities. Routledge. Vargas-Urpí, M. (2020). Juicios multilingües en Barcelona desde la perspectiva de la sociología del lenguaje: lenguas dominantes, lenguas minorizadas y lenguas invisibles. Onomázein, 47, 206-224. 37 Appendix I: Participants, Languages Spoken and Languages they Feel more Comfortable with Participant Heritage Language(s) spoken at Language(s) spoken Language(s) language(s) home outside home they feel more comfortable with Amira Darija Darija Spanish/Catalan/Arabic Catalan and Spanish Dounia Amazigh Amazigh/Spanish Spanish Spanish Salima Amazigh Amazigh/Spanish Spanish/Amazigh Spanish Yassira Darija/Amazigh Darija/Spanish Doesn’t say Spanish Farah Darija Darija/Spanish Doesn’t say Spanish Adila Amazigh Amazigh/Catalan Amazigh/Catalan Amazigh and Catalan Aisha Punjabi/Urdu Punjabi Catalan, Spanish, Urdu, Spanish Punjabi, Urdu and Catalan Meiling Qingtian Hua/ Standard Spanish Spanish Standard Chinese Chinese/Spanish Nazia Saraiki/Urdu Saraiki/Urdu Catalan, Urdu, English, All of them Spanish Xue Qingtian Hua / Qingtian Hua/Standard Spanish, Standard Spanish Standard Chinese Chinese/Spanish Chinese, Catalan Hana Darija/Amazigh/S Darija/Amazigh/ Spanish, Catalan, Catalan and panish Spanish English Spanish Jameela Urdu/Shindi/ Urdu/Shindi/English Catalan, Spanish Catalan and English Spanish Leizi Wenzhou Hua/ Wenzhou Doesn’t say Doesn’t say Standard Chinese Hua/Standard Chinese Zhousi Qingtian Hua/ Qingtian Hua/Standard Spanish /Catalan Standard Standard Chinese Chinese/Spanish Chinese Li Hua Wenzhou Hua/ Wenzhou Doesn’t say Standard Standard Chinese Hua/Standard Chinese Chinese Yue Qingtian Hua/ Wenzhou Doesn’t say Doesn’t say Standard Chinese Hua/Standard Chinese Huang Shanghai Hua/ Shanghai Doesn’t say Spanish Qingtian Hua/ Hua/Qingtian Standard Chinese Hua/Standard Chinese/Spanish/ Catalan Karim Amazigh/Darija Amazigh/Darija Catalan Catalan Rashmir Punjabi/Hindi Punjabi Catalan Punjabi and Catalan P1 Punjabi/Hindi/ Punjabi Spanish or Catalan Doesn’t say English P2 Standard Darija Darija and Spanish Doesn’t say Arabic/Darija P3 Standard Darija Darija and Spanish Doesn’t say Arabic/Darija/ P4 Amazigh Amazigh Spanish Doesn’t say P5 Standard Chinese/ Standard Chinese Spanish Doesn’t say Qingtian Hua P6 Hindko/Urdu Hindko/Urdu/Spanish Catalan and Spanish Doesn’t say