Undocumented migration, informal economic work and peripheral multilingualism: challenges to neoliberal regimes?

ABSTRACT This article investigates the intersections of spatial immobility and informal work among homeless Ghanaian migrants and how these interplay with their multilingual practices. By analysing personal-life narratives and conversations recorded over a two-year ethnography in a bench in Catalonia, it shows that informants practice immobility to gatekeep subsistence resources. They present themselves as dispossessed of welfare rights and belie unregistered economic tasks. They establish non-legitimised translinguistic normativities for intercultural communication yet engage in linguistic regimes demanding ‘integration’ through the nation-state language. This reveals how undocumented migrants challenge but simultaneously perpetuate the neoliberal work/legality conditions and sociolinguistic orders to which they are subjected, which positions them as ‘illegal’, ‘de-skilled’ and ‘languageless’ non-citizens.


Introduction
The geographic mobilities and displacements of individuals around the world have led to the emergence of very heterogeneous societies (Vertovec, 2009). These are now constituted by people who hold various citizenship statuses and have many work experiences and prospects, family projects, religious affiliations, and cultural and language backgrounds (Blommaert, 2013). These mobile populations are translocal, in the sense that they network across and beyond established geopolitical boundaries and are simultaneously locally and globally informed (Castells, 2004;Glick Schiller, 2010).
In this sense, they challenge the citizenship regimes of neoliberal governments which control populations via bureaucratic systems and technologies managed according to the principles of socioeconomic order and rationality (Pujolar, 2007;Rose & Miller, 2008).These citizenship-gatekeeping mechanisms include: restrictive regulations for accessing 'legality' statuses (Inda, 2006); 'late-capitalist' work orders and conditions (Duchêne & Heller, 2012) and job marketplace dynamics which defy what counts as 'labour' in exchange for a salary (Vigouroux, 2008); and, finally, 'linguistic regimes' (Kroskrity, 2000, p. 3) which have mainstreamed elite multilingual language policies but in practice regiment populations through monolingual 'integration-through-national-language' rationalities (Krzyzanowski & Wodak, 2011).
My aim in this paper is to contribute to current debates in critical sociolinguistics research (compiled, e.g. in Coupland, 2003;and Canagarajah, 2017) on how these nation-state orders enable, limit, or enforce the spatial and socioeconomic im/mobilities of transnational populations in connection to their multilingual resources and intercultural communication practices. I specifically focus on the extent to which the neoliberal citizenship regimes, work orders and language testing policies described above shape and lead to newer practices of social differentiation and inequality for migrants in contexts of precariousness. I do so by exploring how the intersections of transnational immobility and informal economic practices among a small group of homeless Ghanaians interplay with their peripheral multilingual practices in an alternative migrant-regulated socialisation space of a Catalan town.
I first present the informant-oriented ethnographic project and the research space where I conducted his study. I then show the findings (Section 2) and argue that informants invest in spatial immobility to gatekeep access to subsistence resources, including food and information about legality issues. I also suggest that they present themselves as having no work opportunities and as dispossessed of welfare rights and benefits. I posit that this is their way to account for the informal-work practices that they tend to leave unmentioned while self-ascribing law-abiding, materially-deprived migrant identities. Next, I report on how they establish non-legitimised language norms based on translinguistic practices, yet simultaneously engage in 'integration-through-national-language' orders which prescribe monolingualism and 'native like', standardised talk. I conclude that this illustrates how undocumented migrants challenge but simultaneously perpetuate the neoliberal work/legality conditions and linguistic orders to which they are subjected in resident societies (Section 3). I suggest that this contributes to understanding how the interplay between immobility, informal work and peripheral multilingualisms may generate new knowledge on migrants' alternative transnational organisation practices in under-explored contact zones such as the one described in Section 2.
Theoretical underpinnings I approach mobilities as interplaying with various situations and practices of spatial and socioeconomic demobilisation and deterritorialisation (Hannam, Sheller, & Urry, 2006). I understand 'formal' and 'informal' economy as a continuum in the range of economic tasks that individuals conduct in order to sustain a living and to access resources in the neoliberal marketplace. Thus, I go beyond a narrow understanding of informal economy as solely including 'illegal', underground economic activities and I posit that informal work, like formal work, is socioeconomically and linguistically regimented and requires particular employability 'skills' and language resources (Vigouroux, 2013). Finally, I adopt and adapt the term 'peripheral multilingualism' (Pietikainen & Kelly-Holmes, 2013), originally not applied to migration contexts, to describe individuals' non-standard, inextricable amalgamations of allochthonous and autochthonous languages from local and distant contexts (Jacquemet, 2010). I argue that the spaces where peripheral multilingualism is crucial for socialisation may become 'novel and revealing spaces to examine contemporary complexities in multilingualism' (Pietikainen & Kelly-Holmes, 2013, p. 5); that is, research spaces where migrants foster newer normativities based on delegitimised communicative practices and establish core 'alternate models of the social world' (Gal, 2001, p. 425) in resident societies.

Context
The study took place in Catalonia, a community of about 7.5 million inhabitants located in Spain whose legality system managing transnational populations (ultimately controlled by the Spanish government) is based on controversial migration policies which may allow foreigners to be granted reunification or temporary visa permits without any work authorisations. This is one of the reasons why it has the highest estimated rates of undocumented people in Europewith a debateable percentage of 30% of undocumented migrants among all foreigners in Catalonia (Sànchez, 2008, p. 251).
In terms of language policies, Catalonia is officially bilingual in a majority and a minority language: Spanish and Catalan. Spanish is the dominant language of the Spanish nation-state, fuelled by a political and socioeconomic project of 'recentralisation' geared toward Spanish-only policies (Pujolar, 2015). This recentralisation today includes a language test demanding a basic level of Spanish as a requirement for foreigners' 'naturalisation' (BOE, 2015, p. 105524). Catalan is a minority language which has been historically persecuted and 'minorised' in the political, economic and sociocultural arenas (Bastardas, 1996). Today it is not recognised as an official language by the European Union, and it is regarded as the co-official 'vernacular' language of Catalonia, with the Catalan government making attempts to present it as the depoliticised cohesive 'language for everybody' (Generalitat de Catalunya, 2016). Linguistic diversity is recognised, though via 'pragmatic accommodations' (Kymlicka & Patten, 2003, p. 5) rather than specific language policies protecting migrants' language rights (see Generalitat de Catalunya, 2008). Foreign populations are systematically addressed in, and tend to choose to learn, the majority language first (Pujolar, 2010), to the detriment of Catalan, and to the detriment of the migrants' access to local powerful networks in Catalonia, fostering a local Catalan/non-Catalan ethnolinguistic boundary (Woolard, 2016).
The research project took place in a town located an hour from Barcelona City which is illustrative of the sociolinguistic configurations of Catalan-speaking medium-sized towns in the metropolitan area of the Catalan capital. It was called Igualada and had about 40 thousand inhabitants, 14.7% of whom consisting of foreigners (the percentage of foreign residents for Catalonia was then 15.7%). The first largest migrant group consisted of people born in Africa, the Ghanaian populations presented below being the second largest subgroup after the Moroccans, with 112 single men aged 35-44 (Ajuntament d'Igualada, 2012).

The participants
The three informants, Alfred, Benedito and Paul (pseudonyms), were, respectively, an English teacher, an accountant, and a schooled cocoa farmer in their forties who were born in an urban town and two villages near Sunyani, the capital of Brong Ahafo, the second largest province in Ghana (West Africa), undergoing mass emigration (Pierre, 2012). Between 2000 and 2001, trying to find job opportunities in Europe and to protect their transnational family income from religious violence in their region (described in Tsikata & Seini, 2004), the three moved to Southern Spain and started working in agriculture, frequently visiting their relatives in Ghana, and in Italy and the Netherlands. Later on, they moved to Catalonia, pursuing socioeconomic improvement, informed by other Ghanaian acquaintances that had followed similar work-related migratory paths. Benedito and Paul settled in Barcelona City, and Alfred moved to Lleida (Northern Catalonia) to pick fruit. Between 2004 and 2007, they decided to move to a smaller yet still well-connected town, Igualada, where they expected to become workers in the industry sector, and where they met each other for the first time. There they employed the most widely spoken, prestigious variety of the Akan language, Ashanti, as a lingua franca among themselves. They also spoke other African languages such as Akyem, apart from English and Spanish, and had a command of Catalan and Arabic.
They chose Igualada because it then was one of the biggest centres of the Catalan textile industry and the first tanning market of the Iberian Peninsula. They found employment upon arrival and obtained/ renewed their temporary residence visas. In 2010, however, the local industry collapsed and most factories were dismantled. Informants became unemployed and started participating in informal economic practices. They had received some severance payments, though when I met them they no longer had any formal source of income. Their transnational mobilities became very limited, too (none had visited Ghana since 2008). By the end of the fieldwork, they could no longer rent a room and became homeless. They took refuge on the bench of an open-air public transport area located on the outskirts of the town; a socialisation space where they networked daily, under precarious conditions.

Methods and data
The data were gathered between July 2012 and January 2013 by means of a network ethnography. This included not only active participant observation of social life on their bench but also 'co-ethnographic visits' (Convey & O'Brien, 2012, p. 339) (i.e. journeys on foot with the informants) to the places of socialisation that informants mentioned during the research project (e.g. the mosque), all located at a 20-minute walk from one another (for the details on this ethnographic method see Sabaté i Dalmau, 2018).
I had observed the informants for a year on my way to the bus station, where we had short conversations. I introduced myself as a Catalan English 'teacher' wanting to investigate migrants' languages in town. They were unimpressed by the university certificates with the project information, and verbal informed consent to participate in the study was not granted until they saw that I did not work for the town hall or for any local NGO, because they mistrusted both. 1 I did not command their African languages, and I introduced myself in Catalan, and then in English and in Spanish. This was a marked sociolinguistic behaviour, for, as stated above, migrants expect locals to use Spanish with them, providing evidence that it is Spanish which is conceived of as the 'language of integration' indexing a 'right to naturalisation' and 'proper citizenship behaviour' (Pujolar, 2007). This explains why informants' preferred language choice for interviews was Spanish (and, less frequently, English, with extensive codeswitching).
The data collection was as follows. Over six months, I audio-recorded narrative interviews, here understood as self-reflective, transformative communicative events (De Fina & Perrino, 2011), on the following narrative themes: geographic im/mobility; un/employment; non-legality; social relationships and identities and multilingualism. I approached the personal autobiographical narrative moments which emerged in interviews as social acts revealing how informants negotiated who they were, at a given time and space (Bamberg, De Fina, & Schiffrin, 2007); i.e. as ways into how they apprehended spatial fixity, unemployment and linguistic delegitimisation.

Immobilities
In interviews, the most salient narrative theme on which migrants topicalised was spatial immobility. They presented their routines in Igualada as if being limited to the habitation of the bench, constructed as a protective 'safe mooring space' (Hannam et al., 2006, p. 2). This is illustrated in Excerpt (1), taken from Sabaté i Dalmau (2015) (see transcription conventions in the Appendix).
Excerpt ( In Excerpt (1), Benedito grounds his local sociospatial orientations on the bench by using: deictically anchored (durative/iterative) proximal verbs like 'sit down' (line 6) and 'come back' (line 8); proximal locative adverbs like 'here' (in lines 6, 8, and 13); and temporal frequency adverbs that mean 'every time'/'all the time', like 'always' (lines 5 and 13), used repeatedly and emphatically in interaction, as a way to give coherence to his relocation story.
Benedito presented himself as a passive, victimised narrator and as a compliant, 'unproblematic' migrant in front of the researcher. He thematically links his enacted immobility to three narrative focuses of concern which were also taken up by the other informants. Firstly, he presents the bench as the place where to overcome precariousness after long-term unemployment (line 2) and where to secure transnational living under the protective umbrella of the group (line 9). Secondly, he focuses on his temporary 'legal' citizenship status, and presents the bench as a place where to avoid 'problems' with registration authorities like the police (lines 3, 5, 10, and 11) or other unnamed migrants (lines 5 and 13).
Paul linked his homelessness to work constraints as well (note that he could then still share a flat with four other Ghanaians, intermittently for every other month, before finally becoming totally homeless). He inhabited a 'docile' migrant identity by presenting himself, seemingly apologetically, as an individual without the 'transcultural capital' (Kiwan & Meinhof, 2011)the knowledge/experiencerequired to follow migratory paths successfully; almost as a self-delegitimised, 'non-competent' Self who in the past made 'wrong' transnational-life choices. This is seen in Excerpt (2), where he defines himself as 'not knowing anything' about 'Europe', distortedly imagined, he claims, as the provider of 'much money' (line 2), through repetitions (lines 4 and 6). He explains that because he did not pay taxes in 2008 (when he was employed) he was not eligible for social benefit payments (lines 7, 8 and 11) (apparently he still had to pay his pending taxes). Tax non-payment is emphatically constructed as 'foolishness' (lines 15, 17 and 19 Benedito and Paul present immobility as an enforced way-out to 'macro' socioeconomic constraints, and claim to have made 'wrong' use of (misunderstood) host-society economic rationalities. This seems to have led to some sort of acceptance, in self-blame acts, of a socioeconomic/geographic fixity position on the periphery. The temporary, quiet habitation of the bench is a way for informants to self-attribute 'non-disturbing' migrant identities deserving social aid. All this reinforces an ethnocentric, moralising narrative on how migrant personas from the imagined 'underdeveloped' South end up entangled in the North because 'they do not know better' (see Smith & Sparkes, 2008). Evidence of this is the fact that informants presented themselves as itinerant populations with no reterritorialisation projects (and with no projects to 'use up' resident-society resources), as when Paul stated 'I want to go and stay in my country for good I go I don't want to come here again because Europe is not like before' (interview 18 July 2012).
Despite these claims, though, the bench had become an alternative 'shelter' where emplacement practices unfolded. These included, for example, 'tactics' to gain access to a bed when being ill; to cheap international calls, and to media information, as when a local resident allowed them to sleep on the floor of his shop, when the Pakistani call-shop owner offered special offers to call Africa, or when the bar owner of the station allowed them to watch TV with no purchase obligation. Thus, the bench was actually a space whereby resources for transnational life were gatekept and redistributed not only among locals and migrants but also across different migrants groups. All in all, this shows that immobilities, as discourse and as practice, are a way to strategize with precariousness in order to secure transnational living and to counteract circulating images of homeless migrants as 'wrong-doing wanderers'.

Informal work
Informants topicalised, too, on their unemployment conditions and socioeconomic stagnation. They mobilised a narrative linked to their work trajectories constructed discursively in three stages, including a successful past period of work, followed by unemployment, and leading to the engagement with informal work (see compilations in Baynham & De Fina, 2005, for workrelated mobilities, too).
In 2008, Alfred started working in a tannery; Benedito, in a foundry; and Paul, in the construction sector. Alfred explains how this added to his previous positive work experience in Catalonia and allowed him to send remittances to his wife in Ghana, in Excerpt (3), illustrative of the first stage of 'success'.  (3), Alfred explains how he, like Benedito, was hired in the local fruit sector in Lleida (lines 1, 4 and 5, and 25-26). His job tasks took place inside a warehouse organised in a taylorised, factory-like, manner (lines 5, 8 and 10), in a protected building (line 28), narrated as being much better than outdoors farming (lines12 and14) and other jobs in the construction sector (line 7). Alfred recalls this work experience as a turning point in life; as a moment of relocalisation. He states that this work-motivated trajectory allowed him to get away from a burdensome life in Barcelona (line 21), and to gain networking capital to establish contacts with other Ghanaian acquaintances, who, in turn, provided him with more job prospects (line 23)bearing witness to the importance of the informal networks of support to access employment (Ajenjo et al., 2008). Finally, he highlights that with this first contract he became temporarily documented for the first time (line 30).
Alfred's 'there-and-then' narrated work experiences allow him to self-attribute a resourceful, determined migrant identity capable of navigating neoliberal economic rationalities, because he presents himself as a person who in the past made the 'right' transnational-life choices. This serves to show that his 'hereand-now' immobilities are mostly the result of structural economic realities, since he had already 'shown' his employability 'skills' and work abilities (counteracting the self-blame acts previously discussed).
In 2010, Igualada was struck by the Spanish economic crisis. The leather and tanning industries, as well as the construction sector, collapsed, and the region experienced the highest percentage of employment loss in Catalonia, the most affected by it being foreigners, whose unemployment rate reached 37.1% (GalíIzard & Vallès, 2010), at a time when, in Catalonia, it was 22% (Comissió Obrera, 2011, p. 16 Benedito addresses unemployment by making extensive use of simple negative sentences with the adverbial negative particle 'no' in Spanish (e.g. 'I don't get'; line 1), and with indefinite negative pronoun meaning 'emptiness' ('nothing'; in line 3). His repetition of simplistic negative forms is an interactional device to give coherence to, and strengthen, his narrative of socioeconomic stagnation and his presentation of the self as a pauperised narrator. Later, Paul, in an overlap, shows alignment with Benedito by presenting himself as a declassed worker living under similar critical conditions, voicing that their troubling experiences were linked to the economic crisis which struck the town (line 6) via simple negative statements and repetitions of 'anything' (lines 8, 10 and 12), too. When I asked whether they received unemployment severance payments, to which both informants replied negatively, Benedito explains that social benefits are no longer available (particularly the 'PIRMI', a temporary minimum 'incorporation' payment of about 420€/month), due to financial cuts (lines 13-15). He names Càritas, the official confederation of charities of the Spanish Catholic Church, as the manager of the bureaucracy required to gain access to PIRMI; i.e. as the social agent participating in the regulation of migrants in neoliberalism. This bears witness to the dismantling not only of the Spanish/Catalan welfare systems but also of the ancillary branches to which their social-service administrations were outsourced (see 'special issue contributor', this volume). Informants kept their former workmates' telephone numbers and were in touch with four temporary-work agencies. They were offered the possibility to plough fruit in Lleida over a summer, which they refused mainly due to health issues (one of them required hospitalisation due to a heart condition, in 2013).
Between 2009 and 2011, transnational incoming migration to Catalonia stalled, and the numbers of new family reunification petitions and temporary residence permits each decreased by more than 500,000. Besides, 15,581 foreign residents left Catalonia to return to their countries or move to another place (INE, 2011, p. 3). This was not the case of the informants, who could no longer travel freely around Europe with their non-renewed temporary visas. In this third stage, narrated as a spiral of socioeconomic downwardness, unprotected informants now lacking all 'safety nets' in society engaged in informal economic activities.
Data on transnational informal work and on its impact on migrant populations tend to be controversial and difficult to access (Vigouroux, 2013). There seems to be agreement, though, that local populations also engage in it, and that migrants organise unregistered economic activities in socialisation spaces where the 'grey market' already existed (Sànchez, 2008), contravening discourses presenting migrants as the promoters of enclaves of 'illegality'. In terms of numbers, at the time of fieldwork it was estimated that approximately half a million migrants worked in the Spanish informal economy (Carrasco Carpio & García Serrano, 2012, p. 15), which translated into about 18% of the foreign population in Catalonia (Sànchez, 2008, p. 252). I found no data for Igualada, and I could only observe the informants' engagement in it partially, since they warned me to stay away from their 'unregistered' work. I could see (and they did not hide) that they got involved in scrap-selling, by first collecting it from garbage containers, individually, in rotating turns. They also begged in the car park of the supermarket located in front of the bench. Supermarket clients frequently gave them bread and tuna, and the 50-cent coins of their shopping carts.
The only indirect comments that I recorded concerning informal work seemed to indicate that the informants' access to unregistered economic tasks was a source of competition and contempt among them. When I inquired about the other informants' economic positions individually, they tended to use reproachful comments to position themselves as being more disadvantaged than the others. For example, when I asked Paul about Alfred, he stated 'él bebiendo un poco # él tiene vino para beber' ('he drinking a bit # he's got wine to drink'), seemingly categorising whom he called his 'paisano' ('compatriot') as a drinker not totally ready for formal work (interview 18 July 2012)in turn positioning himself as more morally righteous than him. And when I asked him about Benedito, Paul stated that he did not need to engage in informal work so frequently because he was already 'better-off': 'éste tiene dinero! […] él cobra ayuda!' ('this one's got money! […] he gets social benefit!'). On the grounds of this, I argue that the network under study functioned as an economic unit based on simultaneous comradeship/competition and trust/mistrust; that is, as a 'togetherness of loners' (Bauman, 2001, p. 68) where the ultimate goal was individual transnational survival and socioeconomic security. Informal-work practices were part of interested chains of exchanges in which money was not always involved, since food-for-cigarettes exchanges, for instance, were as important as getting a 'turn' to beg in the supermarket. All in all, the informal economy, as discourse and as practice, is the informants' way to claim that structural economic constraints precluded access to formal jobs, despite their proved predisposition and employment 'skills' for them. It also serves to claim right to social benefits on the grounds of this former contribution to the local economy, and it reinforces the image of the dispossessed migrant Self (with the informants' compliance to their 'deskilling' as only labour-work candidates). This unveils the socioeconomic benefits of engaging with spatial immobility on the bench that were left unmentioned when narrating immobility.

Peripheral multilingualism
On the bench, informants managed intercultural communication within and across migrant groups through the mobilisation of hybrid, truncated multilingual resources and translinguistic practices, establishing alternative language hierarchies. These were grounded mostly upon Spanish as a lingua franca, and upon unconventional English/Spanish codeswitching, as illustrated in statements like 'Ok-, that one eh español es muy difícil eh!' ('Ok-, that one [language] eh Spanish is very difficult eh!') (interview with Paul, 20 July 2012). → 19 *ALF: yes jo entén entén sí. %tra: yes I understand understand yes.
In Excerpt (5), Alfred, the English teacher, states that local populations do not command English (lines 2, 4 and 5). Paul, in an overlap, shows alignment with Alfred and answers that people in town speak 'small English' (line 3), positioning himself as a 'better' English speaker than them, too. Alfred clarifies, via repetition, that it is just a few who command this language (lines 5 and 7), adding that he does not use it with them (repeating 'no'; lines 9 and 11). And yet, on other occasions when they made reflexive statements about their English competence in public, they contradictorily downplayed their command of it and ended up denying their native speakerhood condition, in self-decapitalisation acts like the one presented in Excerpt (6).
Excerpt (6) @Location: 20 July 2012. Bench. @Bck: Paul (PAU) presents himself first as a non-speaker and then as a non-fully competent speaker of English. In Excerpt 6, Paul presents himself as speaking 'no English' (lines 2 and 4). I was surprised, because we were actually conversing in this language. After my overlap (line 5), he explains that he speaks 'small English' (line 8), with emphatic repetition of quantifying determiners ('no' or 'small'), following the aforementioned dominant linguistic ideology conceiving of non-monoglossic Englishes as faulty. This suggests that informants, in fact, endorsed established resident-society sociolinguistic orders when regimenting their own talk, which became more evident with the analysis of their practices and ideologies revolving around local bilingualism. Informants followed an integration-through-national-language ideology and made every attempt to use Spanish with locals and foreign populations (see Excerpt 5, lines 11, 13 and 17). They presented it as the legitimate language of reterritorialisation, and they inserted it in their other languages, as shown in statements dealing with work or legality issues, like 'almacén' and 'fábrica' (Excerpt 3, line 8)most Catalan towns were referred to in Spanish, too, as in the case of 'Lérida' (for 'Lleida'; Excerpt 3, line 28). The Spanish language was also a barometer of 'integration' to be used among themselves and with other migrant groups in linguistic competitions whose aim was to 'assess' who deserved linguistic legitimacy (and recognition) in the resident-society language.
Catalan was employed as well, because informants engaged in Catalan/Spanish bilingualism, as seen with their systematised use of the greeting 'déu' (equivalent to Spanish 'adiós' -'bye'). This was so despite their claims to the contrary, for they did not conceive of Catalan as belonging to their networking spaces (or to their linguistic repertoires), even when they were using it to try to access the local marketplace (for example, with their CV translated into Catalan by the temporary-work agency workers). In Excerpt 5 Alfred claims a 'receptive' command of it (line 19), in positivising stances where he makes it clear that the language is 'ok' (line 15) in front of a 'Catalan' researcher.
All in all, informants' peripheral multilingual practices and ideologies speak of how they navigate translocal language hierarchies, which include: the institutional language testing regimes of the nation-state gatekeeping access to citizenship, like the Spanish language test for naturalisation; the language policies of minority-language societies, like the requirement to hold an intermediate level of Catalan to work in certain areas of the Catalan administration (as in schools); the prevailing sociolinguistic behaviours of local residents granting who counts as linguistically legitimated in local networks of support; and, finally, the (informal) sociolinguistic normativities and orders negotiated among transnational populations themselves which are necessary to access crucial transnational-life resources, too. This allows for the visibilisation of the complexities of socialising and networking in the different linguistic markets that migrants face throughout their linguistic incorporation into their resident societies. This entails the management of linguistic practices in spaces where there are local linguistic conflicts concerning the legitimacies of Spanish and Catalan and of competitions of linguistic capitals dealing with migrants' languages (i.e. allochtonous languages), as in migrant-regulated spaces such as cybercafés or public benches. Finally, the informants' peripheral multilingual practices also reveal the language resources, language choices and language investments which are relevant for migrants themselves, and facilitate a deeper understanding of their ambivalent social meanings, at the local and global level.

Conclusion
In this paper I have tried to show that nation-states continue operating as the ultimate custodians of citizenship, in front of globalisation phenomena like unexpected transnational flows of 'unfitting' people that challenge the tenets of the territorial, ethnic and linguistic homogeneity upon which their sovereignty resides (see Heller, 2011;Park & Wee, 2017).
More specifically, I have focused on the Spanish nation-state and on how it is modernising its bureaucratic machinery and governance technologies in order to police 'unmanageable', mobile populations in late capitalism. I have centred in Catalonia, involved in the political machinery of the Spanish nation-state in a very complex manner, and I have claimed that it is illustrative of how this policing unfolds in Southern European bilingual societies where incoming migration occurred relatively late, and where migrants soon established alternative socialisation spaces catering for subsistence resources (e.g. food, informal job tasks) off the radars of governmental authorities (see Sabaté i Dalmau, 2014).
I have shown that in Catalonia migrants face a series of legality requirements, work regimes, and language regulations which together, like a governance block, make it very difficult for them to access citizenship statuses and rights. For instance, at present, foreigners in the Spanish state as well as in Catalonia need to have a work permit in order to access legalisation and 'naturalisation', and this permit is granted only after having passed an official language test in the Spanish languageshowing that citizenship, as well as personhood legitimacy, ultimately revolves around hegemonic nation-state languages.
I have focused on a small group of migrants who met none of the legality, employment conditions, and language requirements needed for resident-society incorporation and who were, thereby, immersed in a downward spiral of spatial immobility and socioeconomic stagnation (including health precariousness) as well as linguistic dispossession and silencing. These were three single Ghanaian men who socialised in a global research space, a peripheral public-transport bench, which I have presented as one of the many migrant-regulated alternative institutions of transnational survival that remain under-studied within transnational migration studies.
I have followed a narrative-in-ethnography critical approach to how homelessness and marginalisation are voiced and managed from a self-reflexive, informant-oriented perspective, uncovering ways in which forms of social difference and social inequality get re-produced and materialise in situated communicative events, both locally and transnationally. In this sense, I have shown that neoliberal regimes and interactional orders shape and are shaped by the informants' experiences of precariousness.
Thus, the results have provided evidence that migrants' peripheral multilingualism constitutes a 'record' (Blackledge & Creese, 2017, p. 34) or repository of life trajectories in which transnational im/ mobility and work experiences get inscribed and negotiated in intercultural communication, making it evident that non-legality, informal work and translinguistic talk are mutually constitutive.
All in all, homeless migrants' personal-life narratives around legality 'limbos', unemployment, non-welfare protection, and language normativities reveal how they challenge, but simultaneously perpetuate, the restrictive work/legality conditions and linguistic regimes to which they are subjected. This paradox opens up a new venue for the investigation and problematisation of 'illegalisation', 'de-skilling' and 'de-languaging' among and within migrant networks, in resident societies with exclusionary borders. Note 1. The confidentiality of the data as well as the protection of the informants' identities were ensured by the Ethics Committee at the Autonomous University of Barcelona (file 1818Barcelona (file , 2012.