Multilingual research practices in community research: The case of migrant/refugee women in North East England
Researching in communities where the research participants are multilingual raises questions about research design and practice. This paper illustrates post-research reflections of a study of multilingual women from ethnic minority migrant/refugee/asylum-seeking backgrounds living in the North-East of England. The study involved 68 women speaking more than 25 languages, and 17 researchers, including volunteers, most of whom were multilingual. The analysis revealed the importance of flexible multilingualism in the project design – from inception, to choice of research instruments, data generation processes, analysis and write-up – to accommodate participants' (and researchers') asymmetric multilingual practices. These multilingual research practices, often under-discussed in research methodology, have implications for the authenticity and trustworthiness of the research outcomes, especially among marginalized, vulnerable, and disadvantaged groups.
Condurre ricerca all'interno di comunità di minoranza-etnico-linguistica dove i partecipanti sono multilingue suscita diversi interrogativi in termini di pianificazione e pratica/implementazione della ricerca stessa. Il seguente articolo espone una serie di riflessioni post-ricerca basate su uno studio avente come oggetto donne con competenze multilingue residenti nel nord est dell'Inghilterra ma con storie personali di immmigrazione e asilo politico. Lo studio discusso ha coinvolto 68 donne parlanti oltre 25 lingue e 17 ricercatori, inclusi ricercatori-volontari, per la maggior parte aventi competenze multilingue. L'analisi proposta nell'articolo rivela l'importanza di un multilinguismo flessibile a livello di progettazione e implementazione-inclusi scelta di strumenti di ricerca, processo di generazione dei dati, analisi e produzione del rapporto di fine ricerca-funzionale ad accomodare le competenze linguistiche assimmetriche di partecipanti (e ricercatori). Tali pratiche di ricerca multilingue, spesso solo parzialmente discusse a livello di metodologia di ricerca, presentano implicazioni in termini di autenticità e attendibilità dei risultati della ricerca stessa, specialmente quando questa sia condotta all'interno di gruppi vulnerabili e disavvantaggiati.
The increasingly multilingual nature of communities often requires community researchers to engage in research where more than one language is in use. In this paper, we reflect on and discuss the multilingual nature of community research where there is an apparent tension between a monolingually conceived research design and multilingual practices among researchers and researched. To illustrate this tension we reflect on a one-year study (Hudson and Ganassin 2010), where community researchers engaged with multilingual participants in a larger project aimed at tackling cultural inequalities among vulnerable ethnic minority women's groups. Multilingualism – of both the researchers and the researched – was central to the investigation of these women's cultures, identities, and heritage and in the data generation. Although prominent and noticed by the research team, the need to use multiple languages was unproblematized and largely undiscussed in the planning, execution, and reporting of the study. Instead, the primary focus of the research was to gather evidence, represented monolingually via a report (in English), to inform the implementation of a range of community grassroots activities with ethnic minority women.
Our aim is not to present a research report, but to revisit aspects of the methodology to reflect on how a group of community researchers used multiple languages to negotiate and make meaning when researching multilingually – where more than one language was in use – and to discuss the methodological implications. While the local use of languages and their representations in communities are the subject of increasing investigation (e.g. Blackledge and Creese 2010; Pennycook 2010; Heller and Duchêne 2012), our aim here is to be reflexive and transparent about how researchers and researched use multiple languages to negotiate and make meaning in multilingual community research contexts, and the methodological implications that emerge. First, we present the multilingual context and research design, and the role of feminist approaches in researching multilingually and with vulnerable ethnic minority women. We then discuss multilingualism in focus groups, analysis and reporting, followed by three emergent methodological challenges that researching in multiple languages elicited: dealing with asymmetric linguistic competence; the need for flexible multilingualism; and questions of representation. Finally, we offer implications for researching multilingually that emerged from this post-reflective analysis.
The multilingual research context
The initial one-year study (Hudson and Ganassin 2010) was part of an overarching, three-year, regional community project with minority ethnic women, the Aspen Culture Project (ASP) (2009–2011) (http://aspenculture.org.uk/resources/publication). The project aimed to find creative ways of overcoming inequalities of migrant, asylum-seeking, and refugee women, and establish support programmes for them. The project built on the skills and interests of the North of England Refugee Service (NERS) staff and participants, enabling them to explore how these women could access, contribute to, and engage with local culture, arts, and heritage. As community research, the study was conducted largely outside of university academies. Sara Ganassin participated as a researcher and project officer in various projects under the umbrella of the ASP. Here we primarily revisit data collection processes from the study; however, insights from the larger ASP project are reported where they support our description and analysis of certain researching multilingually practices.
While multilingualism emerged as a constitutive feature of this study, it was already deeply rooted in the everyday multilingual community work in which the researchers had engaged. In order to avoid power issues between researchers and participants (Wilkinson 1999), the researchers supported the creation of a research context where women were “looking at themselves and understanding each other in the act of cultural translation” (Hudson and Ganassin 2010: 7). Multilingualism was at the heart of this endeavour. Therefore, we ask the following research questions:
- R.Q. 1: How do community researchers and researched use multiple languages to negotiate and make meaning when researching in multilingual contexts?
- R.Q. 2: What are the methodological implications when multiple languages are present in migrant community research?
The project was designed and led by the youth and community programme manager (employed at NERS). Four community practitioners, including Ganassin, were also involved in the project design and research. Ganassin, whose first language is Italian, shared some of the languages of the 11 female volunteers described below (Chinese, Spanish and French), and therefore, was already proficient in working in languages where speakers have differing levels of competence. The other three female staff included a British-English speaker whose main role was to design a training package by drawing on the research findings; a staff member from an Italo-Latin-American-Czech background who spoke Spanish, English and Czech; and a British consultant photographer, who then trained the women research participants in participatory photography (a part of the project's activities).
In addition to the above four community workers/researchers, 11 volunteer researchers, some of whom were ethnic minority women and migrants themselves, were recruited from the research community via websites relevant to the voluntary sector. They received training on qualitative research via two tailored seminars delivered by Teeside University. Among them they spoke English and a range of other languages (e.g. Hindi and other regional Indian languages, Mandarin, Italian, Finnish, Russian, Armenian, Farsi, and regional languages of Zimbabwe). Some had done prior voluntary work for NERS, and/or had previous work experience – as teachers, journalists, researchers – in other countries.
Sixty-eight migrant and refugee/asylum seeking women from 17 countries (e.g. India, Nepal, Iran, Afghanistan, China, Ivory Coast, Zimbabwe, Congo, Ethiopia, Poland, Finland, Armenia, Italy, Bosnia, the Czech Republic, Latvia and the UK), all living in North-East England , participated in the study. Collectively they spoke over 25 different languages. Four had English as a first language (being British English) while another five (three from Zimbabwe and two from India ) had English native-speaker proficiency. Their ages ranged from 16 to 80 years and their average age range was from 25 to 35. Participants were recruited through NERS networks.
The research design: A feminist approach
As the research explored sensitive social issues around migrant women's understandings of and attitudes towards diversity and foreignness, feminist perspectives were adopted (Hesse-Biber 2007; 2012). By placing participants at the centre of the research process, and in a safe space where their opinions are valued and they can actively get involved, the women, instead of being seen as ‘recipients’ of community projects and on the margins, became centre stage, active in gathering evidence in the community research from which they would benefit (Hesse-Biber 2007). In this way, the methodology opposed privileging dominant perspectives and forms of knowledge. In order to resist critiques of feminist stand point theory, such as a lack of recognition for the women's individual experiences or their class, race, and political status, all of which risk homogenization of the group and loss of individual voices (Hesse-Biber 2012), the researchers sought to maintain the individuality of the participants. However, this approach was challenged by the large number of participants and the short time scale of the research. The researchers tried to address such issues by focusing on the “strategy of immersion” (Devault and Gloss 2012: 214), that is, engaging with participants actively in the research site, but also enabling the women themselves to make new friendships and create bonds with one another within and beyond the research context.
Cultural spaces as the context for data collection
Researchers also wanted to engage participants in cultural spaces where participants felt they belonged so they could develop a sense of ownership over the research and a feeling that their voices were being listened to (Wilkinson 1999). To achieve this goal, researchers organized participants to meet in safe spaces that allowed the establishment of “sustained contacts over long periods of time” and as a matter of free choice (Hesse-Biber 2012: 214) in order to create an ideal ground for “truly collaborative encounters”. Thus, researchers organized focus groups in four main cultural spaces – fine arts, cuisine, landscape and heritage, and performing arts – which were assumed to trigger and ground participants' experiences and ideas about cultural engagement in the community.
Data collection methods: Focus groups
While data collection methods included focus groups, interviews, participant observation, and participatory photography,1 participants drew on multilingual resources most noticeably in the focus groups. Each focus group occurred immediately after a cultural visit or activity, for example, after participants had watched a theatre performance, visited an art gallery, or participated in a textile workshop. These recreational activities provided an ethical way for participants to be compensated for their time and collaboration in the research. After the cultural activity, participants formed random focus groups of 6–10, depending on how many women were present (four groups for the fine arts visit, three for cuisine and landscape and heritage, and one for performing arts). Focus groups took place simultaneously, lasting 40–50 minutes, and were moderated by a researcher. Another researcher operated the voice recorder and took notes. The focus groups were recorded and transcribed.
In revisiting the methodology of this study, we noted in our post-reflection the emergence of three key themes: researchers' and participants' asymmetric linguistic competence; the importance of flexible multilingualism; and the representation of voice and language in the writing up. Finally, we discuss the methodological implications when researching multilingually in the context of migrant community research.2
Dealing with researchers' and participants' asymmetric linguistic competence
As this study shows, transnational flows of people result in translingual flows, and therefore, the construction of multiple language communities (Risager 2012). Many women, either because of their background or their personal history of migration across different countries, were fluent in more than one language; thus, asymmetric competencies were apparent. English was not necessarily their main or preferred language; nor did their nationality imply linguistic competence in the national language of that country. For example, a participant of Russian-Armenian origin, who had migrated to the United Kingdom from Uzbekistan and despite having lived there permanently, never spoke Uzbek. Born in Siberia from an Armenian family, she had Russian as a mother tongue. Only during her adulthood, after her marriage and the collapse of the Soviet Union, did she reconnect to her Armenian language identity; she now uses Armenian, together with Russian, to communicate with her husband and children.
A second feature concerns the participants' multilingual competence. Many defined themselves as bi or trilingual without expressing a preference for one language over another. For example, a Zimbabwean woman, although having learnt English at school, preferred to use Ndebele and Shona, the official languages of her county, or her tribal language in her daily life. Women from African backgrounds often referred to their tribal languages – languages of specific tribes or ethnic groups (e.g. in Ivory Coast, Zimbabwe, Congo, Ethiopia). Indian women spoke of ‘regional’ languages – local languages spoken in a certain region (e.g. Bengali or Punjabi). Others spoke of colonial languages imposed during colonial regimes (e.g. English, French, Russian). In particular, translations from/into French occurred frequently in the focus groups as the women from French speaking countries (e.g. Congo and Ivory Coast) had different levels of English and helped each other. The English volunteer-researcher who had studied in Quebec also supported these participants.
This linguistic diversity and asymmetry had consequences for the research design. As participants' language choice was often not determined by their geographic origin or the places where they had lived, this linguistic complexity was difficult for researchers to plan for.3 Second, this diversity of languages and competence appeared to point to a lack of confidence of some of the women to engage in discussion in the focus groups, particularly, those who were non-native speakers of English. Ganassin recalls inner questioning about why participants were not responsive: Had they not understood the question? Were they unable to express themselves in English or in one of the languages in the group? Or were they just non-responsive, as may happen in any research context:
And what about you, I mean, what are your thoughts looking at this exhibition? It's not about what you like or what you don't like.
No, no, no, I understand. I can't speak very well, but I understand that. No, no, because I'm not (from) here, I'm not from English country, but I understand you.
These same issues emerged in the case of the other languages spoken within the group, for instance, languages that the women might understand as they had learnt them as children, but were no longer fluent in them as they had migrated elsewhere. For example, participants who came from the Horn of Africa and who had fled their countries (e.g. Somalia, Eritrea) or who had lived in or migrated through Arabic speaking countries (e.g. Sudan) preferred to speak Arabic.
These asymmetries raised several issues for researchers. First, in terms of ethics, researchers had to manage the confidentiality of the participants, many of whom were vulnerable. They needed to ensure that participants' identities were protected in the writing up of the research.
Second, the authenticity of the emergent data was at stake. To engage participants, researchers tried to use simple but meaningful language in designing and asking questions, and rephrased sentences when they felt the meaning was unclear. The researchers questioned to what extent they were constructing the data themselves through their language support. In these situations, they used available multilingual resources present in the group, drawing on the women's relationships with one another to provide peer support, and joining in the interpretation as the women spoke for one another in multiple languages. Ganassin recalls that participants would ‘whisper’ words/phrases in one language to another participant who would translate. However, she recalls conversations in Dari and Farsi that were not translated, and thus, absent in the data. The challenge of capturing all of the women's voices was also confounded by the focus group context. The focus groups took place simultaneously and adjacent to one another in a shared public space. The recordings were of poor quality, noisy and ‘crowded’, with participants and researchers talking over one another, commenting and translating.
Third, the transcriptions were all in English. Ganassin recalls the transcribing process: she translated into English words/phrases that the participants had translated into French during the focus groups. As she was involved in the analysis, she did not recognize any need to include the French words in the transcription. The mulitilingual complexity of the data was an unrecognized aspect of the process of data collection, transcription, translation, and analysis.
The importance of flexible multilingualism in intercultural communication in focus groups
The research focus challenged participants' abilities to express feelings and emotions in English about complex and personal and culturally sensitive experiences, and researchers' capacities to elicit such information. Where focus groups were concerned some women confided to Ganassin a preference for an English non-native speaker moderator whose English was slower, and whose accent, pronunciation, and vocabulary were more understandable. In these circumstances, participants and researchers used different languages and communication strategies flexibly to co-construct meaning.
The participants used multilingual communication strategies to seek understanding and translate for one another. Their life experiences of displacement as migrants, refugees and asylum seekers in different countries and social contexts, where they had little, if any, linguistic knowledge or capital, had already exposed them to experiences of either being translated or translating for someone else. This awareness encouraged them to make efforts to support the other participants beyond the act of translation. For example, a woman from Afghanistan, Farsi speaking and reasonably fluent in English, expressed demonstrable pride in helping the researchers by explaining concepts and translating when necessary for an Iranian participant struggling with English (and that no one else could understand). These flexible language skills and communicative processes sustained the fluidity of the conversations and created a climate where all participants could feel comfortable in communicating.
This relationship of support was more akin to that of mentor-mentee than interpreter-client. Temple (1997), in discussing the role of interpreters in research, notes how communication across languages involves more than just literal transfer of information. The person performing a translation act does not simply translate words, but has the choice of different word combinations. There is not just one correct translation (Bassnet 1994), and the translator faces the challenge of choosing the best way to convey meaning rather than just transferring words (Temple and Edwards 2002). While engaging at a personal level with the other participants, the women who supported their less linguistically-flexible peers not only translated words, but also looked at how concepts could be transferred and understood, thereby making choices about the cultural meanings that language carries (Temple 1997).
The researchers were not always aware of the women's linguistic competences, and therefore, who might be an appropriate ‘translator’. Further, in line with feminist approaches and to minimize power relationships, they avoided asking the women to support one another with translations or suggesting that they sit next to or help particular participants. Instead the women themselves often came forward to translate and transfer concepts and ideas, helping the researchers in their work and supporting the other participants in a way that was still informal. Ganassin observed how the women offered support and explanation using a variety of linguistic and communication strategies – here, synonyms and paraphrasing when discussing art work, albeit unclearly:
- Participant 1
They're pleasant to look at, but it's pleasant for us to look at them, but for them, it might not be pleasant to be looked at and just to have the life that they had and like what often happened with the negation.
- Participant 2
- Participant 1
Negate, yeah, and the opposite like to what's the standard and traditional.
This example shows how multilingual and non-native English speakers tend to paraphrase, using words and concepts from different languages with a confidence and richness that is not only linguistic but also conceptual (Bradby 2002). Words, concepts, and ideas were conveyed, particularly when participants tried to find the exact idea, image or object. This role of paraphrasis (Bradby 2002), where non-native language and multilingual speakers use their skills and knowledge to make meaning, was critically important in generating the data.
Researchers themselves drew on their own linguistic resources. Recognizing the important role of linguistic families (e.g. Romance and Slavonic languages) in bridging concepts and conversations, two researchers – having Spanish and Italian as mother tongues – realized the importance of privileging Latin verbs instead of phrasal verbs (e.g. ‘to continue’, ‘to enter’ rather than ‘to go on’ or ‘to get in’) and Latin nouns instead of those of Anglo-Saxon origin (e.g. ‘apartment’ instead of ‘flat’, and ‘dessert’ instead of ‘pudding’). These strategies helped participants from French-speaking backgrounds to understand questions and contributed to making conversations more fluid.
The researchers found, through participant feedback at the end of the study, that the linguistic support the women gave to one another enabled them to value their own linguistic and cultural diversity (Hudson and Ganassin 2010). This was evident among participants who came from regions where different languages coexist, (e.g. Zimbabwe), and from places where colonial languages are still diffused (e.g. the former Soviet Union). These participants informally described using different languages as ‘normal’: monolingual competencies were surprising to them. By contrast, the British-English monolinguals (two researchers and four participants) felt disadvantaged and less skilled than the multilingual participants, despite English being the official language of the research.
Thus, language can be considered a resource and people who are able to speak different languages are often able to use them strategically and according to the context (Creese, Bhatt, Bhojani, and Martin 2006), rather than according to some specific national and/or ethnic identity. Similar to Mestheneos's (2011) findings, participants in this study, in fulfilling important ‘interpreting’ roles, felt valued for their multilingual contributions, and for the role they played in generating the data. Women from migrant, asylum seeking, and refugee backgrounds tend to be used as interpreters, gatekeepers, and/or points of access by community researchers. The flexible multilingualism apparent in this study shifted that utilitarian role to one where participants were co-constructors of the research by being involved in the redefinition of concepts across languages (Temple and Edwards 2011).
Further, prior research (e.g. Martin, Creese, Bhatt, and Bhojani 2006 ) has demonstrated how different languages are not compartmentalized formally, but naturally dealt with by their users as a matter of choice. This flexibility allows the users to address different dimensions of their lived experience, as this study illustrates. Therefore, flexible multilingualism plays a key role as part of a communication strategy (Blackledge and Creese 2010) and encourages participation and continued conversation among people who have different linguistic preferences (Nussbaum 1990). In community grassroots projects, as this study exemplifies, flexible multilingualism is an important resource for data generation.
Questions of representation: Whose voice? Whose language?
In their genuine interest to try and represent the women, the researchers aimed to foreground the participants' voices and perspectives in all their linguistic messiness and confusion as participants translated and ‘voiced’ one another across multiple languages. Researchers discussed and planned the data collection and focus group processes carefully among themselves in accordance with the feminist approach of the study. The researchers were also advocates for the women, voicing concerns about the experiences participants raised during data collection, and proposing suggestions on behalf of women in the North East of England (Hudson and Ganassin 2010).
In this role the researchers had to ensure effective participation of and communication with the participants, maintain an ethical representation of the women involved, and demonstrate a commitment to cultural and – we add, linguistic – justice to avoid cultural and linguistic domination, non-recognition, and misrecognition (O'Neill 2010). They also had to ensure that internal divisions and perspectives within the same community were not glossed over or normalized to avoid essentializing or misrepresenting the women and their roles in the groups to which they belonged (Temple and Edwards 2011). Acknowledging the multiple languages at play in the research site was important in addressing questions of representation: who speaks for whom, when, where, and for what purposes (Krog 2011).
However, in the writing up of the report, local government end users needed to be considered, and English was the chosen language of reporting. A tension emerged between the aims of the research funders and leaders and those of the researchers who were committed to the multilingual dimension of diversity and equal representation of it in their professional practice. Two questions emerged: What English was to be used in writing up the report? And what English is the appropriate variety of English?
Initially, in the first phase of the write up, Ganassin had been in charge of analysing and writing up the ‘findings’ section of the report. Although the original design had intended to involve the 11 volunteer researchers in this phase, time constraints prohibited this. Ganassin's co-author, a British-English researcher (who had lived in Libya in his childhood and had spoken Arabic), had responsibility for the other sections, including the introduction and conclusion. The co-author also proof-read and polished Ganassin's writing up of the findings. A third writer, and researcher supervising the participant photography section of the larger ASP project, was also drawn into the report writing. As a former English language teacher, she was considered the most competent to write the report. While the three authors acknowledged the importance of producing a final report in polished English (the report was part of a government-funded stream of work), a further question arose: Should the co-author, a British English-speaking person, who had been a university academic, be cross-checked by someone else – the English language teacher?
To address this question the researchers agreed that amendments to the report should be aimed at form rather than content and delivered in a professional style in order to satisfy local government requirements as an official publication, and enable recommendations concerning cultural inclusion of minority women into community events. They believed that having involved the minority women in the data generation performed the function of including their collective points of view. In this sense, the feminist concern and commitment to “make strong arguments for dissemination practices that benefit the least advantaged social groups” (Harding 2012: 59) had been addressed. The issue of whether participants' voices were present – either in multilingual textual quotations as they drew on local, regional, tribal, and colonial languages, or in the multilingual messiness of their co-constructed accounts – went undiscussed. These complex issues were glossed over in the report as “the cultural translation of women translating each other around language, culture and faith” (Hudson and Ganassin 2010: 6).
Further, in this British government-funded project, English appeared to be central at all levels: in the research design (e.g. in consent forms), as the assumed language for generating data (e.g. in focus group discussions, in the recording of audio files, in transcriptions), and as the language of the written report. British-English culture was also conceptually central in the research, that is, the research peripherized these women's positions vis-à-vis the essentialized dominant ‘English’ culture of the North-East; yet, the data generation processes indicate how multilingualism was pervasive among the researchers and researched.
Our purpose was to revisit a study which focused on the inclusion of migrant women into the cultural practices of their local community to explore two questions: How do community researchers and researched use multiple languages to negotiate and make meaning when researching in multilingual contexts? And what are the methodological implications when multiple languages are present?
Where multilingualism is concerned, this study reveals two contradictory layers of language use in the research design. The first concerns the displacement of English as a lingua franca by the multiple languages the participants brought into play, and the important role these languages played in participants' and researchers' multilingual communicative practices as they co-constructed knowledge. The second layer of language use concerns the writing up of the report in the English of local government, where, by contrast, multilingualism was displaced. Prioritizing English ignores the importance of linguacultures, the places where people (re)construct their language communities and where language practices and processes merge with others in the face of global flows of people (Risager 2006). These layers of language use were present in this study, but went unproblematized and undiscussed.
Concerning the first research question, this study demonstrated how the linguistic diversity and multilingual abilities of researchers and participants were a research resource. Feminist research methodology enabled researchers and participants to accommodate one another's asymmetric linguistic competence to construct data in focus groups through processes of flexible multilingualism. They drew on flexible language skills, linguistic resources and families, language strategies (e.g. paraphrasis), and supportive relationships to co-construct meaning, encourage participation, and thus, value linguistic and cultural diversity. These processes empowered these potentially vulnerable and marginalized women, giving them a voice and enabling them to draw on their funds of knowledge (González 2005).
Regarding the second question, this post-research reflection highlights the consequences of leaving multilingualism unproblematized and undiscussed in planning the study, and generating and writing up data in multilingual contexts. All the researchers were well intentioned and aware throughout the project in ensuring the women's participation and sought ways to co-construct meaning. However, because the planners and funders prioritized English in the final report, the women's voices – present in their use of multiple languages in the initial focus groups – were erased through processes of recording, translation, transcription, analysis, and writing up. This lack of awareness of the possibilities of researching multilingually raises questions regarding voice and representation and concerns about trustworthiness and authenticity in the data (Lincoln and Guba 1985).
Developing researcher awareness, through education and training, is critical in rendering authentic the research outcomes on which policy and investment are often based. It may also empower researchers to educate the research drivers (funding bodies and project leaders) to resist normative multilingualism – where the marginalized remain marginalized, voiceless, misrepresented and misunderstood. Krog (2011: 384) argues that, for the voices of the marginalized (e.g. the women in this research) to be heard, change is required through “adaptation and incorporation on the side of the standard or the majority”; only then can the marginalized enter “our own discourses in their own genres and their own terms”. First steps might be to recognize the possibilities of including the languages used by the women, and to be more transparent about processes of translation (by researchers and participants), representation, and reporting. They might also include the women producing their own reports/newsletters in the languages of their communities.
Future research needs to explore further how multilingualism is incorporated into research design and researcher practices, the possibilities engendered by the inclusion of multiple languages and their translations in research outputs, and the role of flexible multilingualism in research praxis. These processes need to be considered in order to give voice to the voiceless, thus rendering authentic the research outcomes on which policy and investment are often based.
We thank the reviewers for their helpful and constructive comments. We also thank AHRC-funded Researching Multilingually project (AH/J005037/1) which generated our interest in writing this paper.
Further details on the methodology can be found in the published report (http://www.aspenculture.org.uk/resources/publication).
Our examples come primarily from the published report. Ethical issues prevent us from drawing on transcripts for further data to support our claims as the project is completed and these data are no longer accessible; nor do we wish to use these data for purposes that were not originally explained to participants.
Funding precluded the employment of interpreters.