Translanguaging as methodology to study language cafés: implications for managing multilingual data

ABSTRACT

This article explores the affordances of translanguaging as methodology by reflecting upon the role of the multilingual repertoires of research participants (including the researcher) in shaping an ethnographic inquiry on ‘language cafés’ (LCs), understood as public events which provide a non-formal learning space for (foreign) language socialisation. Drawing on the ‘researching multilingually’ framework proposed by Holmes et al. (2013, 2016), we reflect on the affordances and complexities of using different languages in the research process. In particular, we focus on how the researcher’s fluid multilingual approach enabled her to co-construct translanguaging spaces with LC participants as part of a methodology to study multilingual socialisation for and through the lived experience of those involved in the research. We aim to inspire researchers to make visible the multilingual, collaborative, and relational processes that shape their research, and to problematise and be reflexive about their choices of transcription of multilingual data. We argue that applying translanguaging as methodology to study multilingual environments can challenge the monolingual ideologies that still prevail in research, while enabling research participants to perform and develop their multilingual social selves.

RESUMEN

Este artículo explora las affordances o posibilidades de acción que ofrece el translenguar como metodología a través de la reflexión sobre cómo los repertorios lingüísticos de los participantes de la investigación (incluida la investigadora) pueden moldear un estudio etnográfico sobre ‘language cafés’, entendidos como eventos públicos que ofrecen un espacio informal para la socialización en lenguas (extranjeras). Siguiendo el marco propuesto por Holmes et al. (2013, 2016) para ‘investigar de manera multilingüe’, reflexionamos sobre las posibilidades y complejidades que surgen al usar varias lenguas durante el proceso de investigación. En concreto, el artículo se centra en cómo un enfoque multilingüe fluido permitió a la investigadora co-construir junto a sus participantes espacios para translenguar como estrategia metodológica para estudiar la socialización multilingüe para y a través de la experiencia vivencial de los involucrados en la investigación. Invitamos a que se visibilicen los procesos multilingües, colaborativos y relacionales que dan forma a la investigación y proponemos la problematización y reflexión sobre la transcripción de datos multilingües. El translenguar como metodología para estudiar ambientes multilingües desafía las ideologías monolingües que prevalecen en investigación, a la vez que permite a los participantes vivir y desarrollar su ‘yo’ multilingüe en sociedad.

KEYWORDS:

• Language cafés
• researching multilingually
• translanguaging
• intelligent transcription

PALABRAS CLAVE:

• Language cafés
• investigar de manera multilingüe
• translenguar
• transcripción inteligente

Introduction

Language cafés (LCs) are a type of public event which provide a non-formal learning space for (foreign) language socialisation. They are non-formal because, although foreign language development may often be foregrounded, these events do not offer any formal instruction or belong to any institutional programme. LCs have received little research attention, and the scant literature published on them focuses on English as a foreign language (e.g. Balçıkanlı 2017; Gao 2007; Murray and Fujishima 2016; Mynard 2020). This article draws from an ethnographic study of two LCs focused on languages other than English. We are interested in the transformative potential of LCs as they offer a space for people to engage in intercultural communication and develop their idea of selves as multilingual social beings (Polo-Pérez and Holmes 2022). In this multilingual transformative space, participants draw upon and perform their multilingual repertoires, understood as the meaning-making heteroglossic resources that individuals have built up throughout their life trajectories as speaking subjects (Busch, 2015). How researchers’ multilingual repertoires impact their resourcefulness in managing the languages in play needs further attention, particularly concerning the handling and representation of multilingual data.

Translanguaging as methodology involves the researcher’s purposeful engagement in organic translanguaging practices (which we discuss later) as part of the methods to study multilingual environments. It might involve, for instance, the researcher’s intentional decision to participate in the field as a language learner in order to neutralise hierarchical relationships and experience LCs from within. A translanguaging orientation to research promotes ‘co-operative dispositions and performative competence for cosmopolitan relationships’ (Canagarajah 2013, 202). Thus, we shift the focus from doing research about language learning and socialisation, to doing research for and through language learning and socialisation. Such an approach prioritises relationality, collaboration, and human multilingual connections over the collection of ‘tidy’ data that might ignore linguistic complexity. This focus necessitates a commitment on the part of multilingual researchers to be reflexive about their linguistic resources, and the role these play in shaping the research process (Giampapa 2011; Holsapple 2022; Martin-Jones, Andrews, and Martin 2017).

Undertaking research in LCs likely requires the researcher to engage in multilingual practices leading to the generation of multimodal and multilingual data, which, in turn, has implications for how the researcher represents these data in the research report. Recourse to ‘intelligent transcription’,¹ which omits the elements of spontaneous speech that do not add meaning, may risk overlooking the human multilingual and multimodal practices through which meaning and understanding are co-constructed among research participants. However, verbatim transcription may also raise ethical issues, and may not be fit for purpose in the thematic analysis of data.

Thus, the aims of this article are twofold: to explore how the researcher and research participants may use their fluid multilingual repertoires to co-construct translanguaging spaces, and to make visible the transcribing challenges and dilemmas that this presents for the researcher handling the emergent multilingual data. The study is underpinned by a researching multilingually perspective (Holmes et al. 2013, 2016; Andrews and Fay 2020) and the concepts of languaging (Phipps and Gonzalez 2004) and translanguaging (Canagarajah 2013; Garcia and Li 2014; Li 2018), discussed below.

After presenting our theoretical approach and the research questions guiding this paper, we outline the context of the study and methods. Then, we present our findings organised in three sections: (1) fluid multilingual fieldwork and researcher relationality, (2) co-constructing translingual interview data, and (3) problematising transcription of translingual data. In these three sections, we present an approach to multilingual research that challenges monolingual ideologies (Gramling 2016) and is committed to reinforcing participants’ understandings of selves as multilinguals in translanguaging spaces. To conclude, we answer our research questions, and discuss the limitations and directions for future research.

Languaging and translanguaging

In this study, we use languaging ontologically to mean that language is always in the making and intrinsically connected to humans acting and being in the world (Li 2011; Phipps 2011, 2019). This position resonates with an understanding of language learning as a lived and embodied experience (Ros i Solé 2016), and one which involves identity investment (Norton 2000). Phipps and Gonzalez (2004, 127) argue that ‘no learning of language can be conducted in isolation from living through it’. For them, languaging is not just trying to use the language one is learning, but ‘the effort of being a person in that language in the social and material world of everyday interactions’ (Phipps 2007, 7, our emphasis). Moreover, languagers are more than purely language learners in that they

move in the world in a way that allows the risk of stepping out of one’s habitual ways of speaking and attempt to develop different, more relational ways of interacting with the people and phenomena that one encounters in everyday life. (Phipps 2011, 365)

These perspectives capture the way we interpret the communicative practices between researcher and LC participants.

Translanguaging develops the concept of languaging, focusing on its multilingual, multimodal, and transformative dimensions (see Moore, Bradley, and Simpson 2020 on translanguaging as transformation). It refers to the practice of using one’s full linguistic repertoire in hybrid and creative ways in communication with others (Canagarajah 2013; Li 2018), allowing for the enactment of multilingual identities not necessarily available in monolingual spaces (Garcia and Li 2014). In that respect, Li (2011) argues that translanguaging

creates a social space for the multilingual language user by bringing together different dimensions of their personal history, experience and environment, their attitude, belief and ideology, their cognitive and physical capacity into one coordinated and meaningful performance, and making it into a lived experience. (Li 2011, 1223)

According to Li (2018), the purpose of adding the trans prefix is to argue that ‘[m]ultilinguals do not think unilingually in a politically named linguistic entity, even when they are in a “monolingual mode”’, and that ‘[h]uman beings think beyond language’ in multimodal ways (Li 2018, 18). In the context of LCs, this may appear troublesome. The monolingual ideology that prevails in many language classrooms still informs many language learners’ expectations regarding authentic interactions with target language speakers (Trentman 2021). Murray, Fujishima, and Uzuka (2017) found that some learners preferred not to attend the LC, believing that they had to ‘speak in perfect English’ (238). Mynard (2020) explored LC participants’ views on the use of their mother tongue (Japanese) and target language (English) during LC events, and concluded that views were conflicting: for some, implementing and monitoring an English-only policy was necessary, whereas for others this policy was a deterrent for less-confident learners. An underlying assumption of monolingual ideologies is that mixing languages interferes with language acquisition. Even if translanguaging is the norm in multilingual contexts, translanguaging practices are often interpreted with a monolingual mindset (Trentman 2021) and seen as indexing language deficiency and inability to sustain talk in one language, and therefore, a practice to avoid if one wants to be accepted as a legitimate speaker of that language. Zhu (2020) suggests promoting social change in this area, by reconceptualising language learning so that learners are empowered and prepared to maximise their language learning in multilingual environments (Trentman 2021). Similarly, we are interested in reconceptualising researchers’ multilingual repertoires as a methodological resource for transforming researcher praxis, and for handling multilingual and multimodal data.

Given the multilingual complexity of LCs as a field site, our study is underpinned by a researching multilingually perspective, which we discuss next.

Researching multilingually

‘Researching multilingually’ is defined as:

the process and practice of using, or accounting for the use of, more than one language in the research process, e.g. from the initial design of the project, to engaging with different literatures, to developing the methodology and considering all possible ethical issues, to generating and analyzing the data, to issues of representation and reflexivity when writing up and publishing. (Holmes et al. 2016, 101)

The researching multilingually framework comprises three phases: (1) the researcher’s realisation of the role of languages in their project; (2) consideration of these multilingual aspects vis-à-vis the spatial and relational dimensions of the research; and (3) informed and purposeful decision-making regarding language use in all research stages, from planning to writing up. These phases should be seen as an iterative reflexive cycle rather than a linear process. Rather than developing an increasing awareness of the multilingual possibilities and complexities of their research as it progresses, researchers might develop a translingual mindset and reflect upon the role of languages in their research from the outset (Andrews, Fay, and White 2018). This reflexive stance enables researchers to make informed and purposeful decisions, that is, to develop intentionality, regarding how their multilingual practices may shape their research: ‘[b]ecoming aware of the potential diversity of linguistic possibilities in research with multilingual dimensions seems both prerequisite and integral for developing researcher intentionality’ (Stelma, Fay, and Zhou 2013, 313).

To increase critical awareness of the multilingual complexities and possibilities of the research and develop intentionality, researchers need to consider the spatial and relational aspects of the research (Holmes et al. 2016). The spatiality dimension includes four spaces, described here in relation to this study: the researched space or phenomenon (e.g. informal language learning and socialisation); the research space where the research is conducted (two multilingual LCs in the United Kingdom [UK]); the researcher space which includes the researcher’s multilingual resources (Spanish, English, Italian, Portuguese, some French and some Arabic); and the research (re)presentation space (e.g. a doctoral thesis in an English-medium UK university) and the possibilities of disseminating the research in one or more languages, thus acknowledging the likelihood of a multilingual readership (Ganassin and Holmes 2020).

The reflexive stance that is needed in researching multilingually praxis (Martin-Jones, Andrews, and Martin 2017) can be enhanced by Byrd Clark and Dervin’s (2014) notion of reflexivity, which encompasses three dimensions: critical reflection, awareness, and hyper-reflexivity. Through critical reflection, researchers reveal power structures in their research and reflect upon the ways in which their own identities play a role and are represented. This involves developing an awareness of self, in order to identify the researcher’s own biases and blind spots. Hyper-reflexivity refers to the researcher’s proactiveness in trying to identify those blind spots, best achieved as a collaborative, dialogic, and interactive endeavour involving the constant ‘meshing of self and other’ (Byrd Clark and Dervin 2014, 23).

Intentional decisions may be revisited based on the researcher’s new realisations and ongoing reflexive considerations as the research develops. Nemouchi – in her study of students’ developing intercultural competence in English language programmes in two Algerian universities – found that when interviewing teachers, adopting the norm of translanguaging between Algerian-Arabic and French and including some English (instead of using English only in interviews) resulted in the teacher gaining more power and confidence. Nemouchi concluded that her strategy developed a relationship of trust, and enabled her to elicit more credible data, hence ensuring the trustworthiness of the research (Nemouchi and Holmes 2022). Similarly, Cortazzi, Pilcher, and Jin (2011) were concerned about the influence of language choice on the quality of interview data. They examined qualitative studies published in English where Chinese participants had been interviewed, and compared the data generated in participants’ first or second language. They concluded that ‘data collected through second language interviews may be qualitatively different from first language interviews’ (Cortazzi, Pilcher, and Jin 2011, 529).

Ganassin and Holmes (2013) encourage researchers to practice ‘flexible multilingualism’, that is, to draw upon, or make strategic use of, the multilingual resources naturally available in the research context in order to accommodate asymmetric multilingual practices. Likewise, Costley and Reilly (2021) contend that multilingual research needs to adopt a flexible and responsive approach, and Andrews and Fay (2020, 77) argue that a translingual mindset can support researchers to be ‘prepared for the unexpected, dynamic, or even playful uses of language in their research contexts rather than predictable and unchanging uses of language’. Thus, intentionality needs to be complemented with a flexible approach which allows the researcher to sense a ‘feel for the game’ (Bourdieu 1990, 56) and adapt appropriately to the emergent multilingual aspects of relational work in the field, which can be difficult to predict. However, undertaking research through languaging and translanguaging involves generating complex multimodal and multilingual data. As the previous studies suggest, such practices are important for the trustworthiness of the research; they also require researchers to be transparent about the challenges and dilemmas faced in generating and handling their multilingual data. In this article, we investigate these challenges and dilemmas, guided by the following research questions:

RQ1: How can the researcher co-construct with research participants fieldwork encounters as translanguaging spaces when researching multilingual environments?

RQ2: What are the implications of translanguaging as methodology for the management of emergent multilingual data?

Context of the study

This study draws from the first author’s (Nuria’s) doctoral ethnographic project, which investigated the affordances of LCs as sites for multilingual and intercultural socialisation. The doctoral research was guided by the following three research questions: (1) How do participants co-construct language learning and multilingual socialisation in the ecologies of LCs? (2) How do participants experience their multilingual identities in the LCs? (3) How do participants experience interculturality in the LCs?

The research sites were two LCs in North England (UK). Language Café One (LC1) was organised by university language teachers to offer university students an extracurricular event to practise and socialise in their languages, e.g. Spanish, French, Italian, and German, although occasionally there were also speakers of Chinese, Japanese, Catalan, Arabic, Russian, and Portuguese. Although open to the public, it took place in the university’s student union bar – a space considered as ‘student territory’ – and most attendees were university students (aged 18-23). Nuria taught Spanish at this university at the time, and it was common for her to see some of her own students during the LC events, which had implications for relational work and her positioning in the field. Language Café Two (LC2), on the other hand, was a public event advertised via Facebook for anyone interested in speaking French. Located in a pub, it attracted people of diverse ages and professional backgrounds. Eventually, LC2 became a rather private event, as the core regular attendees became friends and started to arrange their rendez-vous via a private Whatsapp group. Nevertheless, the group always welcomed newcomers.

Methods

To answer the research questions underpinning this paper, we draw on the data set from Nuria’s ethnographically-inspired study. Nuria participated in LC1 and LC2 for over three years mainly as a learner of French, although she also drew on her other languages for different purposes. Ethical approval was received from the university where the project was based prior to fieldwork. The data sources include Nuria’s researcher-as-participant reflective journal (approximately 35,000 words); audio-recordings of naturally-occurring conversations in the LCs (7 h); 17 audio-recorded, semi-structured interviews and focus groups, and short written reflections from eight participants. Nuria analysed these ethnographic data to reflect on how translanguaging played out in her ethnographic study. In particular, she looked at instances where the researcher and research participants used their fluid multilingual repertoires to co-construct translanguaging spaces during fieldwork. All the participants’ names have been anonymised.

The researching multilingually framework (Holmes et al. 2013, 2016) and Byrd Clark and Dervin’s (2014) notion of reflexivity are the theoretical tools which guided the analysis. Nuria’s critical reflection involved revealing the language spaces she occupied in the research and the power structures underlying multilingual relational work in the field. Critical reflection increased her awareness of her own biases and multilingual positionings, while hyper-reflexivity entailed the two authors engaging in collaborative dialogue to make sense of Nuria’s reflexivity and identify any blind spots.

From the outset of the research project, we recognised that languages would play an important part in LCs (the research space). Nuria’s research approach was informed by her complex positionality as an avid language learner, a Spanish teacher, and a transnational individual whose daily life takes place between and beyond linguistic, national and cultural boundaries (Garcia and Li 2014). She was motivated to seek opportunities to practise her French, and being a genuine learner of this language would allow her to experience the LCs from within and to forge symmetrical relationships with other language learners. Thus, rather than an unspoken requirement prior to fieldwork (Gibb, Tremlett, and Danero Iglesias 2020), language learning – and revealing the researcher’s linguistic incompetence – was a conscious strategy to approach this ethnographic research (Garrett and Young 2009), and an opportunity to grow personally as a ‘whole-person-who-researches’ (Attia and Edge 2017).

Next, we explore the fluid multilingual practices of LCs, and how the researcher’s awareness of these practices shaped her decisions regarding data collection. Then, we address the issue of transcription of data generated collaboratively through translanguaging.

Fluid multilingual fieldwork and researcher relationality

The two LCs in this study ran without any ‘teacher-facilitators’: participants were unsupervised and free to self-initiate and self-manage their own conversations and groupings. Both LC1 and LC2 demonstrated fluid multilingual dynamics due to the natural presence of other languages. LC1 took place in a large venue around different ‘language tables’. Participants could move among tables easily to practise different languages. However, the data from naturally-occurring conversations show that participants in these language groups did not just speak one language, and translanguaging was natural.

I can hear Ella, Milan, and Luke (at the French table) talking about the languages they speak. Ella explains that she was born in Switzerland, but grew up in Norway, but at home and school she spoke English. She also speaks some Spanish and Arabic. Ella code-switches a lot between French and English. One of the guys says he also learned Spanish, and there is an interesting short exchange between them in Spanish (‘Ah, me gusta l’España!’ / ‘Oui, mais olvido’ / ‘Hablas un poco … un poquito’). The guys ask Ella to say something in Arabic, and she says ‘yalla, yalla, habibi!’. One of the guys says (in English) ‘that’s like “come on!”, right?’ (Researcher’s notes on audiorecorded LC1 conversations, 07/11/2017)

A common conversation topic in the LC concerned participants’ language repertoires and learning trajectories. Participants often became aware of their shared multilingual repertoires in interaction with one another through what can be referred to as metalanguaging: languaging when the discussion is about language learning and languaging (Polo-Pérez and Holmes 2022). In metalanguaging exchanges, individuals did not perform one particular language identity at a time, but all their multilingual repertoires played a role in how they presented themselves to others. Like Ella, Nuria did not participate in LCs as an Arabic speaker, and yet she sometimes drew on her experiences with this language to engage in metalanguaging, which evoked memories and emotions about languaging in Arabic. Thus, Arabic, too, played a role in how she performed her multilingual identities in the LCs.

Therefore, a salient feature of these encounters was the availability of multiple linguae francae with the conscious, agreed decision among interlocutors to foreground one of them. LC2 was set up as a ‘French only’ conversation group; however, most participants were also fluent in Spanish and had different biographical and emotional attachments with this language (e.g. sojourns in Spain, being Spanish language teachers, or having a Spanish-speaking partner and friends). Some participants also knew other languages (e.g. Portuguese, Arabic, and Greek). Still, what brought them together was their interest in speaking French, as participants hinted in the focus group:

Nuria: Pourquoi vous avez décidé de commencer à venir dans le groupe? <Why did you decide to start attending the group?>

Joanne: D’abord c’était pour … comme l’a déjà dit Mike, c’était pour ne pas perdre la langue […] Mais, après ça, c’était pour l’amitié, c’était pour les discussions qu’on a eues, et c’était à part de la langue, parce que beaucoup de fois on a commencé à parler en anglais, ou en espagnol, ou quoi que ce soit, c’était … c’est beaucoup plus important que la langue maintenant. Mais ça commence avec … <At the beginning it was for … like Mike said, it was not to lose the language […] But after that it was for our friendship, for the discussions we’ve had, and it was apart from the language, because many times we started speaking in English, or in Spanish, or whatever, it was … it’s much more important than the language now. But that begins with … >

Nuria: avec l’intéresse de la langue <with an interest in the language>

Joanne: Oui oui oui <yes, yes, yes>

(LC2 focus group)

Participants had different motivations for choosing to speak a particular language in the LCs (e.g. ‘not to lose it’). Although they engaged in translanguaging, and often described translanguaging practices in conversation with the researcher, Nuria found that monolingual ideologies often surfaced in their views on language learning (Trentman 2021). For instance, many participants mentioned the need to interact with native speakers of the target language as the only way to learn properly. Translanguaging also involved the natural ‘letting go’ of linguistic accuracy during conversations in favour of meaning making and mutual understanding (Canagarajah 2013). This is illustrated in the above excerpt by the erroneous construction that Nuria uses, ‘avec l’intéresse de la langue’ (instead of ‘avec un intérêt pour la langue’), which Joanne had no problem to understand, even if it was not grammatically correct in French.

Nuria’s intentional decision was to use French in the field, being unable to foresee how her multiple languages and identities would play out in the environment. However, she ended up drawing on her Spanish self more than expected. She sensed that her background as a Spanish migrant in the UK was valued as linguistic and cultural capital which added diversity to the conversations, and an affordance to forge cosmopolitan relationships in the field (Canagarajah 2013). Given the flexible multilingual environment in LC1, where people moved among languages easily, she also saw her use of other languages from a reciprocity point of view:

One of the most relevant things about this night is that I was determined to speak only French, and I ended up speaking a lot of Spanish, as well as Italian, English, and even a bit of Portuguese at the end! I guess this is how it works: people see each other as an affordance to practise their target language, especially when their target language is your mother tongue. Nevertheless, that made me feel a sense of reciprocity, in that I was not only being helped all the time with my French, but I was also helping others. (Researcher’s reflective journal, LC1, 07/11/2017)

The realisation that many LC participants were language enthusiasts informed Nuria’s decision to give participants the opportunity to choose the language of the interviews. Researchers ought to give something back to the field (Ladegaard and Phipps 2020; Smith 2012). In this case, the interviews were aimed at offering a translanguaging space that would be true to the spirit of LCs, instead of prioritising participants’ language fluency and accuracy for data collection purposes. Given that LC2 was a small group of regular attendees who had become friends, Nuria decided to conduct a focus group instead of individual interviews with them and, since the raison-d’être of the group was socialising in French, that was the main language also used in the focus group. In LC1, six out of nine participants who spoke Spanish as an additional language chose to do the interview in Spanish. The opportunity of having an extended conversation in Spanish with a ‘native speaker’ was motivating for them, consistent with the findings by Cortazzi, Pilcher, and Jin (2011). Three participants who were Erasmus Italian students chose to be interviewed in English, even if they knew Nuria could speak Italian, showing their willingness to practise the local language (English) while abroad.

Rolland, Dewaele, and Costa (2019, 283) argue that researchers need to be cognizant of the fact that ‘participants may be at a disadvantage by using a foreign language’ in research interviews. However, in this study, offering a space for languaging, and in another language, was a motivating factor for avid multilinguals to participate in the study, and an affordance for the researcher to make the interviews mutually beneficial.

Thus, fluid multilingual practices were naturally present in our study’s research space, and flowing multilingually in interaction with others shaped researcher relationality in the field. Researcher reflexivity – becoming aware of these multilingual and relational processes in the LCs, and critically reflecting on the complex multilingual subjectivities of those involved in the research – enabled Nuria to be purposeful in co-constructing translanguaging spaces (Li 2011) during the interview data collection.

Co-constructing translingual interview data

Nuria conducted semi-structured interviews with a protocol that promoted natural, flowing, and collaborative interaction (Brinkmann and Kvale 2015) between research participants and researcher-as-language-learner. There was also a voluntary pre-interview activity in which participants were invited to write, in the language(s) of their choice, a short reflective piece about their LC experiences, and those reflections then informed the discussion during the interview. Therefore, the interview topics were emergent and co-constructed in interaction with participants.

As some LC1 participants were university students learning Spanish at the time, they regarded the interview as an opportunity to engage in extended conversations with a native speaker. Out of the 19 participants interviewed in LC1, nine of them spoke Spanish as an additional language. Five of them were Nuria’s students, and the other four were unknown to her before the LC. Out of these four, one was an Italian student who chose to do the focus group in English with two Italian peers, and the other three chose to take the interview in Spanish. These three were confident fluent speakers, with extensive socialisation experiences in the language, who took the interview as an opportunity to improve their already advanced skills in the language (for instance, they explicitly asked Nuria for feedback). Among the five participants who were Nuria’s students, two of them decided they could provide more informed responses if they spoke English (their first language), while the other three were keen to use the interview as an opportunity to speak Spanish. At A2-B1 level (CEFR), they often needed scaffolding to express their ideas (Garcia and Li 2014). Although translanguaging was not limited to these three participants, as we view translanguaging as an embedded practice in all multilingual contexts, these interviews triggered Nuria to question her approach to representing translingual data.

Nuria framed the interviews as translingual from the beginning by telling participants that they could mesh English and Spanish during the conversation as they pleased. Consistent with a translingual and collaborative orientation to communication, Nuria used different strategies to co-construct meaning during the interviews, for example: repeating and paraphrasing information to confirm understanding; responding to non-verbal cues which indexed communication issues (such as laughter, raising intonation, or eye gaze) by filling language gaps when necessary; and using the phatic function of language to express acknowledgment (i.e. ‘aha, I’m following’). The example in Table 1 illustrates some of these strategies (English translation provided in Table 2).

Table 1. Verbatim transcript (English translation provided in Table 2).

Rebecca: Ehm … Me interesa cuando los estudiantes de otros países hablan de la cultura inglés … inglesa Nuria: inglesa, aha Rebecca: y los opin … no, is that right? ¿Opiniones? Nuria: las opiniones, sí Rebecca: opiniones de esa, porque para mí es normal y no tengo que pensar Nuria: Sí Rebecca: y cuando otra persona dice algo me … I realise … ¿di cuenta? Nuria: me doy cuenta Rebecca: me doy cuenta de que tiene razón Nuria: aha Rebecca: así que es una manera de reflexionar sobre mi propia cultura. Nuria: sí, sí, sí, sí … vale (Rebecca’s interview, LC1)

Table 2. Intelligent transcript.

Me interesa cuando los estudiantes de otros países hablan de la cultura inglesa y las opiniones de esa, porque para mí es normal y no tengo que pensar, y cuando otra persona dice algo me doy cuenta de que tiene razón, así que es una manera de reflexionar sobre mi propia cultura. <It’s interesting when students from other countries talk about English culture and their opinions about it, because for me it’s normal and I don’t have to think, and when someone else says something I realise they are right, so it’s a way of reflecting upon my own culture> (Rebecca’s interview, LC1)

The verbatim transcript in Table 1 captures the scaffolding provided by the interviewer (Nuria) to help the interviewee (Rebecca) articulate her thoughts. Interviewer and interviewee met halfway to co-construct a hybrid way of communicating (Canagarajah 2013). At the time, Rebecca was studying Spanish at A2 level (CEFR). It would be easy to frame this interaction as an example of lack of competence to sustain talk in Spanish, informed by monolingual ideologies and a deficit model of language learning (intrinsic to interlanguage theory, for instance). Instead, by focusing on voice, flowing multilingually, and the affordances of translanguaging, both researcher and researched performed their multilingual social selves by drawing on their shared repertoires. These were relevant affordances in this research project considering that the aim was to study (foreign) language socialisation. We agree with Zhu (2020) in that research should ‘make a stance’ and promote social action. In the field of language education research, ‘social action can be achieved through reconceptualising language learning’, for example, by ‘shift[ing] the focus away from language learning to doing language and translanguaging’ (Zhu 2020, 209). In response to Cortazzi et al.’s (2011) concerns about the quality of data generated by ‘non-native speakers’, we argue that the affordance of offering participants a comfortable space for translanguaging, resembling the LC context, outweighed those concerns. Further, language choices and (trans)languaging practices during the interviews were an integral part of the data and offered a live picture of participants’ subjective relationship with languages. At the risk of generating less polished (or even unquotable) data, this approach resonates with the idea of humanising fieldwork by prioritising relationality over data collection (Attia and Edge 2017), and with the pledge to decolonise multilingual research (Phipps 2019) by seeing participants as partners in social action, and not just ‘suppliers of data’ (Zhu 2020). Thus, translanguaging processes are important in subverting normative approaches to generating (multilingual) interview data.

Next, we discuss an implication of this stance for handling the emergent multilingual data, and in particular, the fourth space in the researching multilingually framework: the research (re)presentation space (Holmes et al. 2016).

Problematising transcription of translingual data

How the researcher co-constructs translingual data has implications for how these data are represented in the research report. Transcribing interaction has long been acknowledged as a complex process which requires purposeful decisions and is theory-laden (Lapadat and Lindsay 1999). However, when interactional data is co-constructed through translanguaging practices, there is an extra layer of complexity, which should be accounted for in the research report.

Verbatim transcription is extremely time-consuming, increases substantially the word count (even more so when translations are needed), and might not fit the purpose of the research analysis – e.g. if the research questions require a thematic analysis of content, as was the case in this doctoral study. Furthermore, Lapadat (2000) points out that ‘[s]poken language is structured and accomplished differently than written text, so when talk is re-presented as written text, it is not surprising that readers draw on their knowledge of written language to evaluate it’. This problem raises ethical issues. Kvale (1996) says that ‘the verbatim transcribed oral language may appear as incoherent and confused speech, even indicating a lower level of intellectual functioning’, and ‘may involve an unethical stigmatisation of specific persons or groups of persons’ (172–173). When speakers use a language in which they are not fluent, there is a greater danger of misrepresenting them as inarticulate; therefore, the question of how best to transcribe their speech becomes an ethical matter.

‘Intelligent transcription’ preserves the essence of what is said, but leaves the inclusion or exclusion of the characteristic features of spoken language (the ‘ers’, ‘uhms’, false starts, or fillers) at researchers’ discretion. Intelligent transcription, thus, prioritises the readability of the content by omitting the parts of the speech that are not considered meaningful. Table 2 shows how Nuria transcribed the excerpt in Table 1 as she engaged in the data analysis:

Braun and Clarke (2006) consider the transcription of verbal data as an integral part of the first steps in thematic analysis (i.e. familiarising oneself with the data and searching for initial codes). They also recognise that thematic analysis ‘does not require the same level of detail in the transcript as conversation, discourse or even narrative analysis’ (Braun and Clarke 2006, 17). As Nuria was interested in the content of participants’ remarks, she first transcribed the text sense-for-sense rather than word-for-word, omitting the interactive process that shaped many of the participants’ statements – such as Rebecca’s statement above, which spoke closely to the theme of interculturality in Nuria’s doctoral study. Like removing the scaffolding from a building on completion of its construction, removing the interactive-performative parts of interview data – as prescribed by Braun and Clarke (2006) – is considered to make the content more readable, and easier to code and analyse. Yet, in so doing, all the translanguaging and co-construction of meaning was lost in these transcripts.

On the one hand, researchers need to decide the level of detail to include in transcripts from spoken data depending on the purpose of their analysis. They may even create different versions of transcripts for different purposes (Lapadat 2000). When the focus is on content, intelligent transcription can make excerpts more readable, both for the researcher doing the analysis and for the reader of the research report. Furthermore, as mentioned above, there is a danger that verbatim quotations might present participants as inarticulate or deficient due to the hesitations, repetitions, or false starts that are natural in spontaneous speaking. On the other hand, intelligent transcription removed Nuria and Rebecca’s human and pedagogical experience of languaging and co-constructing meaning and talk together. The original complex interaction was translated into a polished, quotable excerpt – an example of how ‘often relational processes go unwritten, with research write-ups focusing on results, analysis and outputs’ (Moore, Bradley, and Simpson 2020, 5).

The researching multilingually framework does not prescribe how languages should be used in research, but encourages researchers to make visible the complex dilemmas and decision-making processes when researching in different languages. Thus, while using intelligent transcription might be appropriate for certain types of analysis, we contend that researchers should find ways to make visible in their research reports the enactive-performative nature of fieldwork which enabled the construction of the final product, that is, to acknowledge the multilingual methodological dynamics that shaped the generation of research data and findings. Offering this transparency regarding methodological choices may improve the trustworthiness of research that has been conducted multilingually.

Conclusions

In this study we have presented translanguaging as a research approach to investigate multilingual environments, and its implications for researchers handling the emergent multilingual data. We drew from the experience of the first author (Nuria) in conducting ethnographic research in two LCs in North England (UK).

Regarding our first research question (How can the researcher co-construct with research participants fieldwork encounters as translanguaging spaces when researching multilingual environments?), our findings refer to translanguaging as applied to researcher-participant communication. We illustrated the translanguaging practices in the LCs with data from naturally-occurring conversations, participants’ focus groups, and the researcher’s reflective journal. Recognising the importance of the naturally occurring multilingual dynamics at play in the field, Nuria kept a translingual mindset to participate accordingly (Andrews, Fay, and White 2018). She participated as a learner of French, but also used her other languages to forge relationships in the field. LC participants did not perform one language identity at a time; rather, they performed their fluid multilingual social selves through languaging (Phipps 2011) and translanguaging (Garcia and Li 2014) – two concepts that we see as intertwined.

Given Nuria’s awareness of the translanguaging practices in the field, and the multilingual subjectivities of participants, she intentionally decided to offer similar lived experiences in her research interviews. Participants could choose the language(s) in which they wanted to contribute to the study. By framing the interviews as collaborative conversations where both participants and researcher used their shared multilingual repertoires in fluid ways, Nuria facilitated the co-construction of translanguaging spaces which enabled those involved to perform their multilingual social selves. Rather than detaching herself from the generation of interview data, Nuria contributed to the conversations, through processes of languaging and translanguaging, to co-construct meaning and understanding about the LCs with her participants.

To answer our second question (What are the implications of translanguaging as methodology for the management of emergent multilingual data?), we used interview data which illustrated the hybrid communicative practices between Nuria and a participant who saw the interview as an opportunity to practise her Spanish (an additional language for her, and Nuria’s first language). Our analysis showed that both verbatim and intelligent transcription are problematic when dealing with complex multilingual data, as both types have implications for how participants’ voices are represented. We concluded that researchers should be reflexive about their decision-making, and find ways of making visible and explicit in their research report the multilingual processes that shaped the generation of their research data and outcomes.

These multilingual processes inform what we call translanguaging as methodology. In this methodological approach, research participants are encouraged to use their full multilingual resources, which is important in reinforcing their sense of selves as multilingual social beings and in generating trustworthy findings. We recognise the study’s limitations due to the specificities of our research context, where foreign language socialisation was foregrounded and both researcher and participants were multilingual speakers learning new languages. Translanguaging as methodology may be framed differently in research contexts where engaging in multilingual practices is seen as integral to communication, but not the purpose of it.

Through this article, we contribute to researching multilingually praxis, or the ‘methodological multilingual turn’ (Costley and Reilly 2021), by demonstrating the affordances of translanguaging as methodology. First, it challenges monolingual ideologies that risk denying the linguistic, affective, and performative aspects of communication. Second, it allows for more complex representations of those involved in communication as multilingual selves. Third, it encourages an ethical stance whereby researchers humanise fieldwork by moving away from a centred view of research participants merely as suppliers of data (Ladegaard and Phipps 2020; Phipps 2019; Zhu 2020). Fourth, it encourages researchers to prioritise the relational and affective aspects of research, showing awareness and sensitivity to the multilingual subjectivities of the research participants. Zhu (2020, 207) contends that ‘[s]eeing research as social action implies that our research embeds, not leads to, impact; and equally importantly, that it is a process of connections and conversations’. In multilingual research, these ‘connections and conversations’ happen multilingually, which has implications for how complex multilingual data are represented. In that respect, fifthly, translanguaging as methodology invites researchers to acknowledge what risks being lost in intelligent transcription while making fully explicit the basis of their transcription choices, which will depend on the research questions motivating their study. In conclusion, it urges researchers to include a researching multilingually perspective (Holmes et al. 2016) and researcher reflexivity (e.g. Byrd Clark and Dervin 2014) to add transparency to the shaping role of multilingual practices in their research, which, in turn, contributes to the trustworthiness of research outcomes.

Acknowledgements

We would like to thank the reviewers for their time and effort to review the manuscript. Their valuable comments and suggestions helped us in improving the quality of the manuscript.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The term ‘intelligent transcription’ is commonly used in websites offering professional transcribing services. It is, therefore, a term that is already in use and not one we coined ourselves. We would like to clarify that by using this collocation we do not imply that other transcription options are less intelligent.