Italian-Bangladeshis in London: Onward Migration and Its Effects on Their Linguistic Repertoire

: This article discussed language use and language maintenance among the Italian-Banglad eshi community in London, considering in particular the effects of onward migration on the reor-ganisation of their linguistic repertoire. Drawing on focus groups and interviews with the second-generation members of Italian-Bangladeshi families, initial ﬁndings revealed that Italian is maintained through communication with same-age friends and siblings, with older siblings acting as the main agents of language maintenance. English is considered the most important language and, together with a British education, functions as a pull-factor for onward migration to improve the second generation’s future prospects. Bengali, on the other hand, is spoken by parents among themselves and children are not always ﬂuent in the language. Bengali also represents a marker of identity for the Italian-Bangladeshi community as opposed to the larger Sylheti-speaking British-Bangladeshi community.


Introduction
During the last decade, thanks to the freedom of movement allowed within the European Union, a new migration phenomenon has been taking place, namely, that of third-country nationals. After residing for many years in a European country, certain individuals become citizens but then decide to migrate onward to another EU country. This onward migration often involves entire families with children who were born in other various European countries. A growing body of literature in the field of migration studies has investigated this phenomenon (Lindley and Hear 2007;Van Liempt 2011a, 2011bKelly and Hedman 2016;Mas Giralt 2017;Ramos 2018;Ahrens et al. 2016;Della Puppa 2018;McGarrigle and Ascensão 2018;Della Puppa 2018) but has only hinted at the second generation of these onward-migrant families (Ahrens et al. 2016;Della Puppa 2018), identifying the future of these families' children as one of the main pull factors influencing their decisions to migrate onward. Furthermore, no comprehensive studies, to the best of the author's knowledge, have examined this type of migration from a sociolinguistic perspective with the exception of some initial observations on language practices of Spanishspeaking migrants in London who arrived via Spain (Reiter and Rojo 2015;Mar-Molinero and Paffey 2018) and families who migrated onward from Italy (Goglia 2021).
In line with the emergent literature on this topic, I choose to use the term 'onward migration' to refer to this particular migration phenomenon, but other terms are also used in sociolinguistics, including step migration (Tovares and Kamwangamalu 2017), twice migration (Rathore-Nigsch and Schreier 2016), and secondary migration (Mar-Molinero and Paffey 2018). This article focused on a particular group of people, the Italian-Bangladeshi families, who migrated onward from Italy to London. I aimed to contribute to the emerging research on onward migration in the fields of geography, sociology, and immigration by adding a timely focus on the sociolinguistic implications of such migration; in particular, I examined how these families have reshaped their linguistic repertoire in London and their attitudes towards the languages in this repertoire.
Section 2 explores the phenomenon of onward migration in more detail and reviews the existing research in this field. Section 3 provides the theoretical framework of this study. Section 4 presents background information on the Bangladeshi community in Italy and their choice to migrate to London. Section 5 presents the study's methodology and data collection. The results are then presented and discussed in Section 6, and relevant issues introduced by the onward migration of the Italian-Bangladeshis in relation to language use and maintenance are highlighted.

EU Onward Migration to the UK
In the European Union, any person who is a citizen of a member state is also regarded as an EU citizen and is entitled to freedom of movement within the EU member states; this was also true for the UK until 31 December 2020. In 2016, 16 million of the people living in EU member states were citizens of another EU member state (Eurostat 2017). In addition to native-born EU citizens, there are naturalised third-country nationals. However, very little is known about this group of new EU citizens because official statistics often fail to provide information on this kind of mobility. The very first studies on onward migration focused on refugees who initially went to the EU countries that granted them asylum and then migrated to the UK to reach their preferred destination after obtaining citizenship there (Valentine et al. 2008;Valentine et al. 2009;Ahrens et al. 2016). However, this onward movement is characteristic not only of new European citizens with refugee backgrounds, but also of economic migrants who relocate to build a better future for themselves and their families. Initial qualitative studies on this topic reveal that several national groups have relocated to the UK since the end of the 1990s (e.g., Lindley and Hear (2007) on the Dutch-Somali and Sri Lankan Tamil Europeans; Van Liempt (2011a, 2011b) on the Dutch-Somalis; Kelly and Hedman (2016) on the Swedish-Iranians; Ahrens et al. (2016) on three case studies: the Dutch-Somali, Swedish-Iranian, and German-Nigerian migrants; Mas Giralt (2017) on the Spanish-Latin Americans; Ramos (2018) on the Spanish-Colombians and Spanish-Ecuadorians; Della Puppa (2018Puppa ( , 2021 on the Italian-Bangladeshis; Goglia (forthcoming) on the Italian-Nigerians). These studies regarding onward-migrant communities have highlighted several general socioeconomic push/pull factors. These factors include the desire to move to a country with a large co-ethnic and co-religious community, transnational links with friends and family, employment and housing access, cultural and religious maintenance, better educational opportunities for the next generation, the decreased bureaucracy in the UK in comparison with that of continental Europe, and the 2008 economic crisis.
The economic crisis has been particularly severe in southern Europe over the last decade, pushing many Italians, Portuguese, and Spaniards, including families of onward migrants, to migrate to the UK. Some very recent studies focus on the onward migration that was triggered by the economic crisis in southern Europe, concentrating particularly on Latin Americans onward migrating from Spain (Mas Giralt 2017; Ramos 2018). After residing in Spain for a considerable amount of time and acquiring Spanish citizenship, they made the decision to migrate to the UK after the 2008 crisis. This group was part of a larger migration to the UK comprised of native-born Spaniards, as the UK offered more job opportunities during the post-crisis period. For these Latin Americans migrating onward to the UK, the relevant push factor was solely the job market, as they had no historical or linguistic ties with the UK, in contrast to other onward migrants such as the Italian-Bangladeshis and Italian-Nigerians.
Additionally, the first destination EU country of many migrant people is not their preferred choice. For example, many economic migrants settled in Italy during the 1990s because the country offered job opportunities and had less strict rules on immigration, not because Italy was their first or preferred destination. In a recent study on Nigerian migrants in Italy, nearly all the participants stated that their preferred choice of migration destination was the UK and that they did not know anything about Italy before migrating (Goglia 2010(Goglia , 2015. Furthermore, the UK is very appealing as a second destination for migrants from both northern and southern Europe because of its long tradition of multiculturalism. In Italy and the Netherlands, for example, immigrants are often scattered across rural areas containing small towns (Ahrens et al. 2016). This situation also is the case in northern Italy, where the highest number of immigrants reside, and where immigrants are scattered across the many small towns with factories (Goglia 2015). Moving to the UK means that, in most cases, migrant people settle in large or medium-sized urban centres such as London, Birmingham, or Leicester where there are vibrant, established ethnic communities and where elements of their cultures of origin can easily be found. This is certainly the case for the Dutch-Somali communities in Birmigham andBristol (Van Liempt 2011a, 2011b), the Italian-Indians in Leicester and Birmigham (Goglia 2021), and the Italian-Bangladeshis in London (Winstanley 2016;Della Puppa 2018;Goglia 2021).
Many migrants who choose to move onward are part of a wider network of crossborder immigration that includes friends and family who inform and assist one other in the process of re-migrating. This is the case for large transnational immigrant communities such as Moroccans, Bangladeshis, and Indians, and particularly for smaller and less 'visible' immigrant subgroups such as Igbo-Nigerians (Goglia 2010(Goglia , 2015 and East Timorese (Goglia and Afonso 2012).

Theoretical Framework and Key Concepts
The literature on multilingualism and language maintenance in immigrant communities is vast, but it tends to focus on the maintenance of the languages of the country of origin in the country of arrival. Relatively recent sociolinguistic research has emphasised the transnational nature of contemporary migration and the importance of examining individual migrants' trajectories to highlight the effects of these trajectories on migrants' linguistic repertoires and identities in their host destinations (Blommaert 2010;Blommaert and Dong 2010;Pauwels 2016;Tovares and Kamwangamalu 2017). Because onward migration within the EU is a new phenomenon, there are no comprehensive studies on multilingualism and language maintenance among migrants who have migrated onward.
Drawing on Blommaert and Backus (2013), I use the term linguistic repertoire to refer to the totality of the linguistic-semiotic resources that individuals gather throughout their lives. Lists of the languages spoken by individual migrants do not provide the full picture of their multilingualism, as individuals may not have the same competence level in all the languages that they speak, or they may have a linguistic competence within specific domains or activities (Blommaert et al. 2005). From a sociolinguistic perspective, onwardmigrating families have an enriched linguistic repertoire that includes the language(s) of their parents' countries of origin and the language(s) of the previous EU countries in which they lived. Life in a third country requires them to reorganise their linguistic repertoires and their language usage both within and outside the family domain. These individuals' experiences of migration and linguistic repertoires are more complex than those of migrants arriving directly from outside Europe in a one-step migration process. Furthermore, migrants must reshape their linguistic resources according to the host country they are in, and their usage of languages and perceptions of people's attitudes towards them are indexical to the language ideologies in their host countries.
A concept which is very useful in the case of onward migration is linguistic capital, which is part of Bourdieu's cultural capital (Bordieu 1986). For Bourdieu, knowledge of languages and education represent cultural capital that can provide power and access to resources to the citizens in a given society. The value of cultural capital varies according to societies and ideologies. A very simple example is the value of the official national language of any EU country in terms of gaining access to the job market. In Italy, for example, immigrants are required to speak Italian, and a knowledge of English is not necessary. Even when there is a need for English, immigrants from English-speaking countries are very often not hired because their education credentials are not recognised in Italy; additionally, if their variety of English diverges from British English or American English, their proficiency is looked down upon. For example, many English-speaking immigrants from West Africa, despite holding high school certificates or university degrees from their countries and being fluent in their varieties of English, cannot use this cultural capital in Italy (Goglia 2015).

Italian-Bangladeshis and Onward Migration to London
Bangladeshis started migrating to Italy as economic migrants in the 1990s. Today, the Bangladeshi community is the eighth-largest immigrant community in Italy, and it comprised 139,409 regular residents in 2018 (Giacomello et al. 2018). Until 2000, the majority of these Bangladeshis were in Rome, but during the following years, they moved to other parts of Italy, particularly to the regions of Lombardy and Veneto in the northeast, which were offering job opportunities in their factories. In the meantime, a process of family reunion took place. Since 2010, many Bangladeshis have acquired Italian citizenship along with their children. Some of these new Italian-Bangladeshis have decided to make use of their newly acquired Italian citizenship, which grants freedom of movement within the EU, and have moved to London.
England, particularly London, has been a major destination of this new Italian migration wave of the last decade, which has included newly naturalised Italians. According to the Office for National Statistics, in December 2019, the Italian residents in London represented the fourth-largest foreign community after the communities of Romanian, Polish, and Indian nationals (Office for National Statistics). Records from the Italian Consulate in London reveal that, as of July 2020, 430,000 Italians were registered with the AIRE (Register for Italians living abroad). The information on these individuals' places of birth reveals a more refined picture: 51% were born in Italy, 25% were born in the UK as the descendants of Italian migrants, and 24% were born in third countries (Consolato Generale d'Italia di Londra 2020). The latter group is composed of either descendants of Italians in other countries who obtained Italian citizenship by ancestry or immigrants who obtained Italian citizenship through naturalisation in Italy. Among those who were born in a third country, the largest groups are composed of Brazilians (30%), Bangladeshis (11.7%), Pakistanis (6.36%), Argentinians (4.44%), Indians (5.6%), Ghanaians (4.8%), South Africans (2.97%), and Nigerians (2.9%) (Consolato Generale d'Italia di Londra 2020).
Recently, the new Italian migration wave to London was studied by scholars in the field of migration studies who stressed the differences from previous mass migrations: the new migrants are young, relatively globalised, and searching for job opportunities in the flexible London job market. Indeed, they struggled in the Italian job market where the older generations hold life-long contracts and do not allow generational changes; they also encountered a lack of meritocracy and an inability to learn English, which is a skill required both in Italy and internationally (Tirabassi and Prà 2014;Scotto 2015;Licata 2019).
The study of this new Italian migration wave from a sociolinguistic perspective is even more recent, and very limited research has been conducted so far (Cacciatore and Della Puppa (2018) and Della Puppa and King (2019) have researched the Italian-Bangladeshi community in London from a sociological perspective. One of the themes they discussed, the role of English and British education as a pull factor of migration to London, is particularly relevant to the present study. Winstanley (2016) discussed language ideologies shared within the Italian-Bangladeshi community in London. She found that English was regarded as the most prestigious language in the repertoire of her participants and influenced the choice of migrating onward. Most participants accepted their children having a minimal knowledge of Bengali and considered the knowledge of Italian not useful in London. Marchese (2016) presented the sociocultural features of the Italian-Bangladeshi community based on interviews with key members of this community.
All these studies, however, have focused on these groups' first arrivals to London and only on adults, i.e., the parents of the families who engaged in onward migration. The experience and perspectives of the second generation have not yet been studied. This article filled this gap by concentrating on interviews and focus groups with the children of Italian-Bangladeshi families and shedding light on the way they reorganise their linguistic repertoire. The research questions guiding my analysis are as follows: 1.
Is Italian maintained by Italian-Bangladeshis in London? 2.
What is the role of English in the onward-migration process? 3.
What attitudes do Italian-Bangladeshis have towards Bengali?

Methodology and Data
The data discussed in this article were drawn from a larger ongoing project funded by the Leverhulme Trust, which investigates multilingualism, language maintenance and shift, and language attitudes among families who migrate onward from Italy to the UK. Unlike the previous studies on onward migration, this data collection did not focus on any particular ethnic community, but rather on young individuals who belong to families who have migrated onward from Italy to the UK. All the included families had migrated within the last 8 years and were from the following groups: Italian-Nigerians, Italian-Ghanaians, Italian-Indians, Italian-Bangladeshis, and Italian-Sri Lankans.
In this article, I focused on Italian-Bangladeshi families. The data were collected in various ways among the Italian-Bangladeshi community: a sociolinguistic survey; focus groups with Italian-Bangladeshi students and parents in a secondary school in Newham (East London), a core area where Italian-Bangladeshis have settled; interviews with Italian-Bangladeshi university students recruited through personal contacts and the snowball technique; and observations made in key places (Italian-Bangladeshi cafes and association gatherings in London). In this article, I presented excerpts from the interviews and the focus groups. Both the focus groups and the interviews focused on the participants' experiences of onward migration from Italy to the UK, biographies, their language use and that of their families in Italy and London, their attitudes towards these languages, their language use with different interlocutors, and their language maintenance. In both the focus groups and the interviews, I used a list of questions to trigger discussions and focused on engaging in spontaneous informal conversations on the participants' experiences of migration and language use. The interviews and the focus groups allowed the participants to reflect on important issues such as their experiences and views on language maintenance, their identities, and the challenges they might have had to face when they migrated to the UK. The focus groups with the students lasted approximately 20 min, were conducted in Italian, were audio-recorded and took place in the school with 9-10 10th-year Italian-Bangladeshi students. The focus groups with 15 Italian-Bangladeshi parents took place during a meeting with Italian-Bangladeshi parents organised by the school and were conducted in Italian; the interviews with six Italian-Bangladeshi university students were conducted either in English or Italian according to the participants' preferences, lasted approximately 30 min, were audio-recorded, and took place mostly in public places such as cafes or university canteens.
The data used in this article were collected between September 2018 and January 2020. The project received ethical clearance from the Ethics Committee of the University of Exeter, and the author also obtained a DBS clearance to conduct data collection in the school. In the excerpts in the following sections, all the participant names are pseudonyms.
The approach taken in this study was to view each interview as a type of communicative interaction in which the interview content was locally constructed by both the interviewer and the interviewee (Pavlenko 2007;De Fina 2011;De Fina and Perrino 2011). With this approach, the presence of the interviewer was not seen as a disturbance to obtaining natural data; rather, the interviewer participated with the interviewees in communicative interactions. The object of this study was indeed the way the participants presented their use of languages and reflected on this issue during their interactions with the interviewer. The participants of this study were generally keen on taking part in the study. Not only were they often pleasantly surprised by a researcher taking an interest in their experiences of migration and multilingualism, but they were also very willing to share their ideas and the challenges that they had faced in terms of language learning and identity. My role as an Italian interviewer in this study actually triggered reflections and storytelling that otherwise would have remained undetected (cf. De Fina 2011).

Results
The Italian-Bangladeshis in London have a very complex linguistic repertoire. The parents in this community usually speak Bengali, and they may have some knowledge of the variety of English spoken in Bangladesh. Upon their arrival in Italy, they had to learn Italian, and depending on the region in which they resided, they may have some knowledge of an Italo-Romance dialect. During their stay in Italy, there was little or no opportunity to learn or improve their English, as their new life and work in Italy required only Italian or Italian dialects in virtually all domains of language use except for in-group interactions. Occasionally, some knowledge of English may have been used to communicate with Italians who spoke it or with other English-speaking immigrants (cf. Goglia 2021).
As the second generation was born in Italy or arrived at a very young age, they speak Italian as their first language. They may have learned Bengali from their parents and may have some knowledge of Italian dialects, which are often used by Italian youngsters in the context of code-switching during conversations with peers. They may have also learned English or other foreign languages at school. When these families migrated onward to London, English was added to their linguistic repertoire. The parents learned it informally as they did with Italian in Italy, while their children joined the educational system and soon replaced Italian with English as the language used in the education and public domains of language use. For example, the European Somalis in the UK add English to their already complex linguistic repertoires, which include, for example, Dutch or Danish (depending on where they were migrating from), Somali, Arabic, and any other language that they may have learned (Valentine et al. 2009).
Studies on the EU onward migration, although not focusing on sociolinguistic issues, touch upon the English language and British education as pull factors for the choice of the UK as a preferred destination. In her study on Dutch-Somalis (of refugee origin) migrating onward from the Netherlands to the UK, Van Liempt (2011b) reported that the parents in these groups wanted their children to be fluent in English and obtain British educational degrees, which they regarded as more recognised. A similar attitude towards English and a British education was echoed in the work of Ahrens et al. (2016), who found that several German-Nigerian and Swedish-Iranian parents had even sent their children to British international schools to study English so that they could take advantage of transnational ties in other countries in the future. In their research on Italian-Bangladeshis who migrated to London, Della Puppa and King (2019) found that the parents in these communities were concerned that their children could not be educated in English in the Italian educational system, as they regarded English as key capital for the future careers of their children both in Bangladesh and anywhere else in the world.
The next sections present some initial results on how Italian-Bangladeshis reshape their linguistic repertoire in London based on excerpts of interviews and focus groups with participants.

Italian Language Maintenance in London
The Italian-Bangladeshi families arrived in London after spending many years in Italy. The parents migrated to Italy during the 1990s and had to learn Italian, whilst their children were born in Italy and speak Italian as their L1 with Bengali. In London, the parents continue to speak Bengali among themselves and, in some cases, with their children, but Italian is also used.
The use of Italian varies according to several factors, including the family language policy that was used to prioritise Italian in Italy, the access of the parents to English classes and jobs that require contact with English speakers, and the ages of the children when they arrived in London.
Italian is mainly maintained by the children, particularly the elder siblings as they migrated to London in their teenage years and are therefore more fluent in the language. The younger siblings may be too young to actively maintain Italian, or they may have a basic knowledge of Italian but are learning English at school. If the siblings in a family are all fluent in Italian, they may speak Italian among themselves together with English and speak either Italian or Bengali with their parents, as exemplified in excerpt 1. In this excerpt, Rakash, a 20-year-old male, explained that at home, he and his siblings speak in Italian whilst Bengali is used to communicate with their parents.
Rakash: A casa, parliamo bengalese, ma con i miei fratelli parlo sempre italiano. Int: Parlate italiano con i vostri genitori? Rakash: No con i miei genitori bengalese, con i miei fratelli solo italiano, non parliamo mai in bengalese Int: Quindi non parli in bengalese con i tuoi fratelli What Rakash is describing is a situation in which the second generation speaks Bengali mainly for intra-generation communication within the family while prioritising the use of Italian (at home) and English in other communicative domains in London. It is worth noting that Rakash reported the use of Bengali as a perceived secret language with his siblings in public, since Italian is more widely spoken in London and may not fulfil this function.
The maintenance of the Italian language in Italian families who migrated abroad is mainly due to parents or grandparents (Rubino 2014;Pauwels 2016). However, in Italian-Bangladeshi families, two heritage languages compete for maintenance. Parents actively contribute to the maintenance of Bengali since Italian is their L2, while the elder siblings act as agents of Italian language maintenance in the family domain, particularly for their younger siblings. This has also been observed in other families with different ethnic backgrounds who have migrated onward from Italy: Italian-Indian, Italian-Ghanaian, and Italian-Nigerian families (cf. Goglia 2021). The second generation of Italian-Bangladeshis that participated in the study may maintain their Italian through their interactions with siblings as well as with their friends and classmates who also migrated onward from Italy. Ratul, a 14-year-old participant of a focus group, explained that in his school, there are many students who speak Italian as well.
Ratul: Sì io c'ho tanti amici Int: Nell'intervallo parlate italiano? The friendship domain, which is often regarded as less important than the family domain in the literature on heritage language maintenance, becomes key to language maintenance among the second generation of families who have migrated onward (cf. Winter and Pauwels 2006). Pupils in schools can form friendship groups based on shared languages and identities, as Valentine et al. (2008) showed in their study on Somalis from the Netherlands and Denmark in Sheffield. In the school where the focus groups were conducted, there are several other groups of children from onward-migrating families who socialise in their shared languages during break times. For example, pupils from Angola, Brazil, and Portugal speak in Portuguese, while Spanish-Nigerian and Spanish-Bangladeshis speak in Spanish (p.c. with the teachers in the school).
The children of the Bangladeshi families, both in the interviews and in the focus groups, also recall that having the chance to meet Italian speakers when they first arrived in London and joined the school was very helpful as Rihan, a 14-year old student who took part in the focus group, stated in 3.
(3) Rihan: All'inizio, per uno nuovo che non parla bene l'inglese un aiutino in italiano può sempre far comodo Rihan: At the beginning, for a new one who cannot speak English-well, a little help in Italian may be useful.
The main reason why the second-generation participants are keen on maintaining Italian is that they feel that it is their main language and a strong marker of their identity; this is particularly the case among those who came to the UK as teenagers and have already completed several years of schooling in Italy. Aziz, a 21-year-old male who arrived in London 6 years prior to the interview, speaks both Bengali and Italian but indicated that Italian corresponds to his preferred identity, calling it his 'mother tongue'. The choice of this term has a very important connotation among Bangladeshis, as it is linked to the struggle for independence of Bangladesh from Pakistan in 1971.
(4) Parlo bengalese, ma penso che l'italiano reppresenta meglio la mia identità perchésono cresciuto in Italia, è la mia madrelingua. I speak Bengali, but I think Italian best represents my identity because I grew up in Italy, it is my mother tongue.
Most of the children from these families have friends in Italy and are still in contact with them. Italian is their preferred mode of communication with these friends. In example 5, Ahmed explained that the main reason Italians do not speak English well is because of the way it is taught in schools.
Ahmed: Italiano, l'inglese chissà se lo sanno Int: Non lo sanno bene gli italiani l'inglese? Ahmed: Non è nemmeno colpa loro perché l'istruzione per fare l'inglese è fatta male, gli fanno insegnare le robe a memoria, cioè e non funziona, tipo, oggi imparo un verbo e domani me lo scordo Int: With your friends on WhatsApp, do you communicate in Italian or English? Ahmed: Italian-English, who knows if they know it. Int: Italians do not know English well? Ahmed: It is not even their fault because the teaching of English is not done well; they make them teach things by heart, so it does not work. Like, today I learn a verb and tomorrow I will forget it.
Onward migration does not mean a complete rupture with Italy. Besides maintaining friendship ties in Italy via online communication, participants also stated that they often visit friends and relatives in Italy.
The maintenance of Italian is also linked to the fact that, in the UK, Italian is studied as a foreign language and may be useful for the future career prospects of the children of the Italian-Bangladeshi families. Both the parents and children are aware of this. The majority of the young participants have either obtained or are planning to obtain official certificates in Italian, such as the GCSE and A-Level in Italian. The participants provided three main reasons for choosing to obtain recognised language qualifications: (1) the opportunity to receive formal teaching in a language they feel they may lose and feel insecure about, particularly in terms of its formal usage; (2) the desire to obtain a formal qualification that acknowledges their knowledge of the language for their future careers; and (3) the opportunity to get a high mark on an exam in a language that they have already mastered.
(6) Ho fatto A-level in italiano, così ho preso più punti per andare all'università I did the A-level in Italian, so I got more points to go to the university.
Winstanley (2016) found that her Italian-Bangladeshi participants (first generation) considered Italian not useful in London and geographically restricted to Italy. A more positive attitude towards Italian in this study was due to the fact that participants belong to the second generation and regarded Italian as their main language.

The Role of English in the Onward Migration Process
All the participants of this study mentioned the English language and the British education system as the main pull factors for their families' decisions to leave Italy; among them, Tariq, a male Italian-Bangladeshi, migrated onward to London with his family eight years prior to the interview. In excerpt 7, he explained that his family decided to move to allow him to be educated in England.
(7) La ragione principale per venire a Londra è stata la mia educazione perché onestamente, io direi che l'educazione in Inghilterra è molto migliore dell'Italia [. Tariq's statements demonstrate the ideology of the prestige of a British education. His positive attitude towards a British education is based on his family's ideology and the generally positive attitude in Bangladesh towards British education and institutions (cf. Della Puppa 2018). He arrived in the UK when he was twelve without any knowledge of English. In the above extract, Tariq seemed to acknowledge that leaving Italy at a young age was vital for him to acquire knowledge of the English language and to be able to access the British university system.
Studies on onward migration from EU countries to the UK, although not focusing on issues of language, have identified the role of English and a British education as important pull factors for the decision to migrate onward. In her study on Dutch-Somalis migrating onward from the Netherlands to the UK, Van Liempt (2011b) found that parents wanted their children to become fluent in English and get British degrees because they are perceived more internationally recognised. Ahrens et al. (2016) also found that several German-Nigerian and Swedish-Iranian parents had sent their children to British international schools to prepare them for a future onward migration to pursue an international higher education and career. Mapril (2021) reported that Portuguese-Bangladeshi families in Portugal sent their children to schools that can provide English curricula and the opportunity to access British universities. Italian-Bangladeshis of this study do not report having studied English in international or British schools prior their onward migration (cf. Della Puppa 2018).
In excerpt 8, while recalling his family's decision to move to London, Tariq explained the opportunities that proficiently speaking English as an international language can offer (molte porte si aprono; 'many doors will be opened') in contrast to the limitations imposed by a local language such as Italian. Winstanley (2016) also found that English is not viewed as being linked to a specific geographical place, but rather to the opportunities that it can provide as linguistic capital. Tariq's reference to English as linguistic capital reproduces a language ideology held by Tariq's family prior to their migration and points clearly to the opportunities available to young students such as Tariq (the second generation) but not to their parents.
The value of English as linguistic capital is also enhanced due to the use of English in the high domains of Bangladesh as a result of British colonisation; as Tariq says, quindi l'inglese è rispettato lì'; 'so English is well respected there'.
(8) ...una volta che sai l'inglese, molte porte si aprono, è una delle lingue più parlate al mondo, moulto più che l'italiano. Poi, il Bangladesh era una colonia inglese, quindi l'inglese è rispettato lì ...once you know English, many doors will be open; it is one of the most spoken languages in the world, much more than Italian. Additionally, Bangladesh was a British colony, so English is well respected there.
It must be noted that this language ideology that views international English as highly prestigious is shared by other groups who migrate onward to London, such as the Italian-Nigerians (Goglia forthcoming). This ideology was dormant in these families while they were living in Italy because Italian was the dominant language for their children's future careers. While living in Italy, the Italian-Bangladeshi families underwent a language shift towards Italian, and English added no major extra value to their daily lives in Italy. Obtaining Italian citizenship provided an opportunity for these families to choose to migrate onward and, by choosing to move to London, reactivate their language shift towards English at least at the level of the second generation.
For the parents of the Italian-Bangladeshi families, at the time of remigration, they may have had some knowledge of English, although in most cases, they knew very little. During the time that they spent in Italy, English was not used, as a knowledge of English was not useful for everyday life and their interactions with local people. Furthermore, the parents speak Bengali among themselves as their preferred code. In the interviews, the second-generation participants stressed that their parents have a poor knowledge of English and, in London, they had to face the challenge of learning it. Atanu, a 14-year-old male in a focus group, explained the following: (9) Io c'ho mio padre che parla inglese perché l'ha imparato e poi c'è mia mamma che capisce, peró non sa parlare quindi c'ho anche dei familiari qui che l'aiutano an imparare I have my father, who can speak English because he learned it, and then there is my mother who understands but cannot speak; so, I have also some family here who help her to learn.
This struggle in learning English echoes a struggle of beginning a life in a new country. Mapril (2021) stressed that the access to educational capital in English for the children results from "the sacrifice" of the parents.

Bengali
Italian-Bangladeshis are a sub-community of both the large Italian community in London and the equally large community of British-Bangladeshis. In addition to having a different migratory trajectory from that of the Italian-Bangladeshis, British-Bangladeshis come from the district of Sylhet and speak the local vernacular of the district, Sylheti (Rasinger 2007). The Italian-Bangladeshis are Bengali speakers. The participants were aware of this difference and used it to define their Italian-Bangladeshi identity. Ashek, a 22-year-old male, defined Bengali as quello più normale 'the more normal one', implying that it is the standard.
Ashek: No, perché la mia famiglia non è da quelle parti Ashek: There are two big distinctions, Sylheti and that Dhaka, Bengali, and we speak the more normal one. Int: You do not speak Sylheti?
Ashek: No, because my family is not from there.
Ashek's comment reproduced a post-independence language ideology that regards Bengali as the prestigious national language of Bangladesh. In this extract, Ashek stated that he does not speak Sylheti because he is not from the district of Sylhet, but in other interviews participants explicitly admitted a negative attitude towards Sylheti. In excerpt 11, a parent from the focus group with parents in a secondary school explained the family language policy that he implemented in Italy and now implements in London, which gives preference to standard Italian and standard Bengali and ensures that his children do not learn the dialects of these languages.
(11) Parlavamo solamente italiano con i nostri figli, non volevo che imparassero i dialetti italiani. Allo stesso modo, volevo che parlassero il bengalese buono, non i dialetti. Non va bene che parlano Sylheti come gli altri bengalesi in Inghilterra. Non voglio che la gente giudica i miei figli per come parlano. We only spoke in Italian with our children; I did not want them to learn Italian dialects.
In the same way, I wanted them to speak good Bengali, not dialects. It is not good to speak Sylheti as the other Bangladeshi do in England. I do not want other people to judge my children because of the way they speak.
This parent wants his children to speak il bengalese buono, non i dialetti 'good Bengali, not the dialects', referring to the way British-Bangladeshis speak. This negative attitude towards Sylheti mirrors the stigmatization of Sylheti in Bangladesh as a language associated with low social status (Rasinger 2007). Della Puppa (2018Puppa ( , 2021 pointed out that the dichotomy between the British-Bangladeshis and the Italian-Bangladeshis in London is not only at the level of languages, but also at the level of social classes. The first generation of Bangladeshis who migrated to Italy and later re-migrated to London was made up of young middle-class individuals who were often from urban centres and educated in standard Bengali and English. All the second-generation participants stated they have some knowledge of Bengali: some speak it with their parents, whereas others have a receptive knowledge of it; they understand their parents speaking in Bengali, but answer in English or Italian. In (12), Tariq, a 21-year-old male, explained that the reason that he prefers to use Bengali with his parents is because they do not speak Italian well.
(12) Con i miei genitori parlo bengalese, solo qualche parola in italiano, ma non una conversazione in italiano, perché loro non lo parlano bene, è normale secondo me perché sono cresciuti in Bangladesh. With my parents, I speak Bengali; only some words in Italian, but not a conversation in Italian because they do not speak it well. It is normal to me because they grew up in Bangladesh.
Bengali plays a significant role in daily interactions with parents, other adult members of the Italian-Bangladeshi community, and communication with family in Bangladesh. Some families choose to spend one or more years in Bangladesh to ensure that their children learn Bengali and English through schooling. Rezza, who is 14, explained that his family spent 4 years in Bangladesh when he was younger and they were still residing in Italy, so that he and his siblings could learn Bengali (poi l'ho imparato e sono tornato in Italia; 'then I learned it and I came back to Italy'). Although Bengali is maintained within the family domain, the second generation is undergoing a language shift towards English and those who are fluent in Italian maintain Italian, a language that offers more linguistic capital than Bengali.

Conclusions
Onward migration has been studied so far from a sociological and anthropological perspective and most studies have focused on first-generation adult migrants. Furthermore, the effects of onward migration on language use and language maintenance have not been studied yet. This article focused on the second generation of Italian-Bangladeshis in London, providing their perspective on how they use and maintain the main languages of their linguistic repertoire: Italian, English, and Bengali.
Our findings showed that Italian is still used at home, particularly in interactions with siblings and, in some cases, with parents. The friendship domain also appeared to be very important, as Italian is still widely spoken among the participants' friends and classmates whose families migrated onward from Italy in a similar fashion. Being able to speak in Italian with other Italian-Bangladeshis is perceived as a positive factor that alleviates the culture shock that the participants experienced when they arrived in London. Italian is also widely studied as a foreign language in the UK, so many participants took exams in Italian, as they wanted to have official qualifications proving their knowledge, improve their written and formal uses of the language, and receive higher marks for university admission. Depending on their ages at the time of arrival, some participants also continued to maintain transnational links with their friends in Italy. It is worth noting that most of the participants indicated that they have contacts with other naturalised Italians but not with the larger Italian community in London. This is explained by the fact that the Italian-Bangladeshis mainly live in East London and that their migration project was already an in-group matter; the families joined family and friends who were already part of the Italian-Bangladeshi community in London. All the participants in this study considered Italian to be a key marker of their identity and their 'mother tongue'. What this study found was that, in the case of onward migration, the elder siblings were the main agents of maintenance, differently from other kinds of migrations (cf. Pauwels 2016).
All participants mentioned English and a British education as the main pull factors prompting them to migrate to London. English is viewed as an international language that can be used everywhere, including Bangladesh. This is in line with previous studies on onward migration regardless of the migrant community studied. Italian-Bangladeshis do not send their children to international schools or British schools in Italy, but some participants reported having spent some years in Bangladesh while living in Italy to study in English-medium schools (cf. Ahrens et al. 2016;Mapril 2021). The most important sociolinguistic outcome of onward migration for the second generation is the re-activation of language shift towards English that a long stay in Italy had blocked.
The declared use of Bengali by the participants varied according to the language policies of the individual families. Bengali is mainly maintained within the family domain in communications with parents, but some participants admitted having only a receptive knowledge of the language. Bengali represents a marker of identity for the community. The Italian-Bangladeshi community in East London is large and clearly represents a subgroup that is both culturally and linguistically distinct from the larger Italian community, but it is also different from the British-Bangladeshi community who are speakers of Sylheti as well as other onward-migrating groups such as the Spanish-Bangladeshis or the Portuguese-Bangladeshis.
This study provided the very first insights into the still under-studied sociolinguistic issues linked to the process of onward migration and, in particular, to the second generation of such migrating people. Further research needs to be done to investigate the actual language practices and aspects of language mixing in this context beyond the participants' declared linguistic uses and attitudes.
Funding: This research was funded by the Research Fellowship "Onward migration from Italy to the UK: sociolinguistic implications" Leverhulme Trust.

Institutional Review Board Statement:
This study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of the University of Exeter.
Informed Consent Statement: Informed consent was obtained from all subjects involved in this study.

Data Availability Statement:
The data presented in this study are not publicly available due to restrictions in data collection policy and to protect participant anonymity.

Conflicts of Interest:
The author declares no conflict of interest.