Complex intersections of language and culture: the importance of an ethnographic lens for research within transnational communities

ABSTRACT

A growing body of work exploring transnational interaction has brought to light the importance of awareness of multilingualism in research contexts, yet little consideration has been given to researchers working in a later learned language (LX) or the process of investigating linguistically diverse communities. This research takes place within a multilingual social space, a transnational Spanish language group in London, against the backdrop of a globalized and mobile world. The work explores the complexities of conducting research in a multilingual, international, and transient context where sharing all language varieties and/or cultures with participants may not be practical, possible, or even preferable. The data, collected by myself (a participant-researcher), includes recordings of informal social conversations, ethnographic observations, and interviews with key participants. Their analysis highlights linguistic adaptation to facilitate transnational understanding, demonstrates the value of interviews in a common LX, and considers how language ideologies and norms affect transcription of multiple language varieties. The paper argues that utilizing an ethnographic lens, particularly when working in an LX, allows for a deeper understanding of localized multilingual interaction through closeness to the participants and advocates a slow, detailed approach to data analysis.

RESUMEN

Cada vez más estudios sobre interacción transnacional revelan la importancia de tomar conciencia del multilingüismo en contextos de investigación, pero se ha prestado poca atención a los investigadores que trabajan en lenguas aprendidas después de la infancia (LX) o al proceso de investigar comunidades lingüísticamente diversas. La investigación se realiza en un espacio multilingüe-un grupo transnacional de hablantes de español en Londres-en el contexto de un mundo global con alta movilidad. Este artículo explora las complejidades de investigar en un contexto multilingüe, internacional y transitorio donde compartir variedades lingüísticas y/o culturas con los participantes puede no ser práctico, posible, o incluso preferible. Los datos, recogidos como participante-investigadora, incluyen grabaciones de conversaciones sociales informales, observaciones etnográficas y entrevistas con participantes. El análisis destaca cómo la adaptación lingüística facilita el entendimiento transnacional, demuestra el valor de realizar entrevistas en una LX compartida y considera cómo las ideologías y normas lingüísticas afectan la transcripción de múltiples variedades lingüísticas. El artículo sostiene que una perspectiva etnográfica, particularmente al investigar en una LX, permite un entendimiento más profundo de la interacción multilingüe localizada a través de un mayor acercamiento a los participantes. Asimismo, defiende un enfoque lento y detallado en el análisis de datos.

KEYWORDS:

• (Socio)linguistic ethnography
• researching multilingually
• language groups/language cafés
• researching in an LX
• crosscultural research
• transnational communities

PALABRAS CLAVES:

• (Socio)lingüística etnográfica
• investigar de manera multilingüe
• grupos de idiomas/cafés de idiomas
• investigación en ‘LX’
• investigación transcultural
• comunidades transnacionales

Introduction

The growing body of work on transnational interaction in applied and sociolinguistics (see De Fina 2016) means that research increasingly crosses linguistic and national borders, bringing multilingualism, in its myriad forms, to the forefront of investigation. These linguistically and culturally heterogeneous groupings, increasingly common in social and professional realms, would benefit from consideration of how multiple and overlapping language repertoires and cultural understandings impact the research process. Referred to as linguistically diverse communities, these are heterogeneous groups of people with a range of first and learned languages, a variety of cultural backgrounds, differing or multiple nationalities, and often disparate competencies regarding language and culture. In essence, these groups are defined by their linguistic and cultural dynamism.

Recent work has begun to consider ‘researching multilingually’ as an object of research itself, foregrounding the importance of conscious reflection in multilingual contexts (Holmes et al. 2013, 2016). Yet, there is a lack of consideration given to researchers working in an LX¹ and limited discussion of linguistically diverse communities. While some prior work suggests that sharing a language or culture is preferable when researching across cultures (see Liamputtong 2008b and discussion on ‘insider’ status below), studies often fall short by not specifically considering situations in which language is shared, but cultures are multiple and varied. Yet, this is a typical configuration within transnational settings such as the Spanish language group (SLG) at the heart of this study and is further complicated for LX researchers who bring their own linguistic complexities.

Drawing on interactional, observational, and interview data, this article explores practical and methodological complications that may arise when conducting research in an LX and within linguistically diverse communities. Three ‘challenges’ I experienced conducting research within the heterogeneous SLG are analyzed: a concern that language used within the group would be changed ‘for me’ as an LX speaker of Spanish, questions around interview language choice as related to competency and preferences, and the possibility of crosscultural misunderstandings. Throughout, it is argued that making use of an ethnographic lens, particularly when working in an LX, allows for deeper understanding of localized, multilingual interaction, and that bringing awareness to the multilingualism of a research context and its researcher(s) can strengthen the investigation.

Transnationalism, researching multilingually, and the ethnographic lens

Complexity within transnational spaces

Research in transnational spaces, particularly in superdiverse (Vertovec 2007) urban centers, has resulted in wide acknowledgement of linguistic complexity related to increased mobility and high levels of interaction within diverse populations. This is evidenced by concepts such as translanguaging (Li 2018), metrolingualism (Pennycook and Otsuji 2015), and flexible bilingualism (Blackledge and Creese 2010) that theorize hybrid and creative language use in these diverse, often multilingual, settings. This type of language contact has resulted in a ‘plethora of linguistic varieties that cannot be easily accommodated within traditional sociolinguistic categories’ (De Fina 2016, 167). Communities may express transnational identities through flexible multilingual practices (e.g. Li and Zhu 2013) that vary from community to community and exchange to exchange, necessitating intimate knowledge of localized and contextualized language processes. If we acknowledge the ‘capacity of transnational individuals to mobilize their linguistic resources to (re)construct different relations and meanings within a specific social context’ (Li and Zhu 2013, 519), it is essential to ask how the researcher fits into that space and makes sense of the meaning-making processes within. These flexible and diverse linguistic practices, paired with complex participant and researcher identities, can lead to research challenges. For example, a researcher may lack full access to the significance of localized language use that incorporates multiple varieties and cultures. Such complications underline the need for ethnographic and reflexive approaches on the part of the researcher, where a consideration of one's own layered identities is essential.

Increasing use of the concept transnationalism invites critical reflection of a simplistic view of migrants as ‘uprooted,’ rejecting a non-binary view of those who cross or straddle national borders (De Fina 2016) and acknowledges a more global, connected economy (Vertovec 2009). Yet, describing transnational links as those between ‘country of origin’ and ‘country of settlement’ (Schiller, Basch, and Blanc 1995) does not address the myriad places people may encounter within a highly mobile world. Importantly, connections to places are made and maintained as people move, resulting in new identities and cultural practices (De Fina 2016). Thus, this important shift in the conceptualization of transnationalism as dynamic is just the beginning. With transnational communities employing a variety of linguistic and cultural resources in their communication and identity construction, it is essential to question which research practices are most suitable to explore language use in these contexts. Without this consideration, researchers may apply more traditional understandings of multilingualism, which may be ‘filtered through monolingual ideologies’ (Pennycook and Makoni 2019), to their understanding of their participants or research setting. Just as researchers increasingly accept, appreciate, and even celebrate multilingual practices in our participants, there needs to be space to reflect on and apply concepts and scholarship related to creative multilingual practices within the research process itself. Developing such reflexivity within research is particularly important in multilingual contexts (see Clark and Dervin 2014), enabling further examination of ‘the interrelationships between language, communication, culture and society’ (Pérez-Milans 2017, 1).

Since the research in this paper was conducted within a group of Spanish speakers, it is notable that the Spanish language is of particular interest transnationally, with global influence and an increasing number of speakers, including learners (see Paffey 2014). With 440 million L1 speakers in 21 officially Spanish-speaking countries (Perez Inofuentes et al. 2021), Spanish has ‘diversified considerably’ and ‘social and geographical variation is considerable’ (Lipski 2008, 1–2). The diverse Spanishes present within the SLG mirror this worldwide usage, making questions of L1-LX socialization and communication, within and outside research contexts, more widely relevant.

Having established the field's acknowledgement of complexity and heterogeneity, one may wonder why more explicit reflections on methodological and practical considerations of research in transnational spaces have not been published (an exception being Warriner and Bigelow 2019), particularly given the impossibility of being an ‘insider,’ linguistically and/or culturally, for all levels of belonging and connection. Gibson and Zhu (2016, 185) have looked at practical, multilingualism-related issues, considering ‘language barriers’ and interview concerns such as which language(s) to use and how to avoid misunderstanding in an LX. However, the work has limited discussion around possible solutions and does not specifically acknowledge potential consequences of ‘cultural barriers.’ One key researcher account is Knowles' (2019) reflection on the limitations of conducting interviews in her LX (Spanish). Yet, unlike this project and with differing ‘consequences,’ she positions herself as a learner who ‘would not be considered fully proficient in Spanish’ (95).

Ethnographic approaches within diverse communities

Ethnography has proven a fruitful and effective tool in multilingual, transnational, and similar contexts, whether through general ethnographic approaches (Zhu and Li 2021; Orellana 2015), (socio)linguistic ethnography (Madsen 2015; Zhu, Li, and Jankowicz-Pytel 2020), or critical sociolinguistic ethnography (Barakos and Selleck 2019), to name a few. In the present work, an ethnographic lens is broadly constructed, with ethnography viewed as a paradigm, with ethnographic fieldwork aimed at ‘finding out things that are often not seen as important but belong to the implicit structures of people's life’ (Blommaert and Jie 2020, 3). As Li Wei (2019, 154) states, ‘[e]thnography is an account of someone's observation of and experience with a community and their cultural practices in specific contexts.’ Throughout this article an ethnographic lens is suggested as a tool and an orientation to ensure that multilingual and multicultural researchers are attuned to their own positionality, the emic experiences of those in their research contexts, and the cultural and linguistic practices within.

Reflexivity and researching multilingually

The past decade has begun to see an important shift towards a recognition of multilingualism in research processes within applied and sociolinguistics. This has resulted in important contributions such as the ‘researching multilingually’ framework (Holmes et al. 2013, 2016), increased reflexivity around multilingual research processes (Bashiruddin, 2013; Ganassin and Holmes 2020; Giampapa and Lamoureux 2011; Warriner, Bigelow, and 2019), and political considerations in multilingual research (Holmes, Reynolds, and Ganassin 2022). In addition, at least one edited volume has focused on researching multilingualism as a topic of study (Martin-Jones and Martin 2017), while limited considerations of multilingualism in research contexts can be found in research method publications (e.g. Gibson and Zhu 2016). Even so, there has been a lack of practical and methodological consideration of the topic (see Rolland et al. 2023) and prior publications may not be entirely applicable to transnational spaces. The present work builds on growing methodological awareness of multilingualism in research, offering experiences of navigating the practical complexities faced as an LX researcher within a transnational context. This reflexivity highlights critically important, often overlooked, considerations in similar environments and argues that an ethnographic lens may allow researchers to more fully explore the depth and diversity of linguistic and cultural practices within these spaces.

The dynamic and multidimensional nature of ‘insider’ status

When researching across languages and cultures, there are levels of ‘in-group’ and ‘out-group’ relations (e.g. nested identities Bailey 2001 and intersections of identities Kiesling 2013). Thus, one relevant concern when researching linguistically diverse communities is the challenge of defining insider status, which encompasses linguistic and cultural dimensions. Without theorizing what a ‘linguistic insider’ is, it is unclear what it means to be on the inside in terms of language. This is seen in Andrews et al. (2019) who argue, importantly, for explicit reflection on the multilingualism in any given research project, yet still refer to researcher Bashiruddin as a ‘linguistic insider’ without expansion. In their own work, (Bashiruddin 2013, 359) claims to be ‘bilingual and fluent in both the languages’ [of the research context], which requires further unpacking and defining to determine how the researcher's specific relationship(s) to associated languages and cultures may impact the research process. To further complicate matters, an LX speaker may never be considered an ‘insider,’ or their categorization could change over time as competencies and experiences fluctuate (Merriam et al. 2001), highlighting the dynamism of ‘insider’/‘outsider’ relations. Additionally, with the frequent focus on language, the question of ‘cultural insider’ status can be neglected. Further reflection on who defines ‘insider’ status, and on what basis, is required to determine its role in the research process.

Beyond the complexities surrounding the term itself, some see epistemological advantages for ‘insiders,’ while others recognize the value of ‘outsider’ status. Research in crosscultural communication has long suggested that researchers should be ‘insiders’ (see Liamputtong 2008a, 7), ‘those who share social, cultural and linguistic characteristics [with] the research participants,’ in order to ‘reduce cultural and linguistic barriers.’ Others claim that ‘insider’ status results in fuller participant accounts, with interviewees providing their ‘best’ responses to those with overlapping characteristics and commonalities (see brief overview in Adamson and Donovan 2002). However, these approaches assume singular linguistic or cultural identities and/or a team of researchers with varying backgrounds, which is not always possible, particularly if ‘insiders’ are those ‘who share their participants’ ethnicity, culture and first language’ (Murray and Wynne 2001, 157).

In a competing view, some suggest that ‘insiders’ may be culturally too close to their participants, making them biased (e.g. Bhopal 2001). Irvine, Roberts, and Bradbury-Jones (2008) take a more practical approach, considering translation and interpretation, while acknowledging the blurred lines between ‘insider’ and ‘outsider,’ yet still rely on an overly simplistic dichotomy. Murray and Wynne (2001, 159) suggest ‘researchers have found that when interviewees speak in a second language they perceive themselves as less confident, happy and intelligent,’ unlike findings in challenge 2 of this article (presented below). These differing viewpoints result in uncertainty for researchers weighing their approach, although this can be mitigated by developing an ethnographic understanding of the research context. Since familiarity with one's participants and research context gives a researcher a better understanding of local norms and practices, these can be applied to practical and/or methodological decisions, especially regarding multilingualism. Of course, where possible, working in multilingual research teams may increase the diversity and range of linguistic and/or cultural knowledge available (see Martin-Jones, Andrews, and Martin 2017; Reilly et al. 2023).

Clarity on the researcher's relationship to the languages and cultures being studied, with further unpacking of these interconnections, would be particularly useful to build an understanding of best practice while researching multilingually. This is particularly important since many researchers work in an LX, whether with fellow researchers and participants, in their data, or during dissemination or publication. It is essential to be explicit about these connections and competencies (or lack thereof, see Lorette 2023) since ‘generally, researchers do not have the language skills necessary to communicate with a linguistically diverse population’ (Murray and Wynne 2001, 157). As argued below, a perceived lack of language or cultural skills is not necessarily a problem, but rather, when coupled with ethnography, can result in a deeper understanding of the meaning-making processes between researcher and participant.

The research context: complexities of the Spanish language group

This paper draws on data and experiences from an interactional, sociolinguistic study of a London-based group of transnational and multilingual Spanish speakers. The SLG (as it is referred to in this paper) meets in central London to socialize in Spanish and/or for Spanish language related cultural events. I² joined this group in January 2017, well before I started conducting research in the space, and have continued to attend throughout the data collection stage and beyond as a member of the group primarily, and a researcher secondarily.

Since ‘critical research calls for a more sustained and rigorous exploration of the ways the researcher's subjectivity influences the research process’ (Canagarajah 1996, 325), I will further outline my experiences with language groups and the SLG specifically. Over the years, I have made a habit of joining language groups when moving around the world. These transient communities have connected me with others who share interests in travel and languages and improved my French and Spanish. In London, I encountered the well-established and long running SLG, with ‘core’ members and an insistence on intermediate Spanish or above.

My involvement in the SLG has led me to consider myself a participant-researcher, someone who is involved in a group or community they are studying (and may have developed interpersonal relationships) prior to and/or alongside their role as a researcher in that space. I use the term to acknowledge and reflect on my relationship to the SLG and the participants within. Being a participant-researcher calls for reflexivity, ‘an on-going self-awareness’ that ‘refers to how you position yourself in the research context, and contemplate how one's own self could influence the actions one takes’ (Miyahara 2019, 88). This reflexive work acknowledges participant and researcher identities as ‘multiple, fluid, and negotiated’ (Sharma 2021, 234).

During my time in the SLG, a typical monthly gathering included approximately 15–40 participants and, unlike language groups where language learning is the primary focus, the SLG feels more social than pedagogical, attracting a relatively large number of L1 Spanish speakers. Research on language groups, until now (see also Polo-Pérez and Holmes, 2023 on language cafés) has been mostly limited to groups speaking the English language (e.g. Gao 2009; Balçıkanlı 2017) with a primarily pedagogical focus (e.g. Murray, Fujishima, and Uzuka 2017), while this project examines a group that prioritizes Spanish language socialization.

It is important to note that the SLG is transnational (in terms of their cross-border connections), multilingual, and cosmopolitan, attracting members who are generally affluent (as evidenced by their high levels of education and professional positions), gather for leisure, and are culturally interested (Hannerz 1990). Cosmopolitan individuals have been shown to make use of linguistic heterogeneity in a playful and creative way (Codó 2014), further necessitating an ethnographic understanding of the context. Further, it is a site of elite multilingualism (Barakos and Selleck 2019), characterized by the social capital and access to a ‘global perceived elite’ that knowledge of the Spanish language affords members of the SLG, with many attendees’ multilingualism based on free choice (Rydenvald 2015). Thus, awareness of cosmopolitan and/or elite orientations towards multilingualism aids in understanding language practices within this space.

Linguistic ethnographic data were collected, using an inductive approach, primarily between 2018 and 2020. They include 11 hours of recorded conversations, taken mostly during monthly social gatherings and resulting in nearly 100,000 words of transcription, three semi-structured interviews with core participants (approximately two hours each), and over 60 pages of typed field notes from ethnographic observations. This article will examine interview data, field notes, recorded conversations, and personal communications in its consideration of LX research in a heterogenous context. The project received ethical approval from the author's institution with individual consent given by the participants.

Considerations for LX research in a transnational context

During the process of data collection for this project, I encountered some challenges I had expected and others which came as a surprise. Although they are entitled ‘challenges’ below, I invite the reader, alongside myself, to question the negative connotations and explore the ways in which perceived difficulties may result in richer, thick data (Geertz 1973). Through an examination and interrogation of challenges I faced while conducting LX research in a transnational context, I consider how an ethnographic lens helped me to ‘overcome’ potential or perceived shortcomings and look at practical solutions which emerged within the context of the SLG. Three main challenges are presented, paired with two sub-challenges illustrating each point and drawing on a range of data types collected within the wider project.

Challenge 1: are participants changing their language for me?

In a heterogeneous group of culturally, linguistically, and nationally diverse L1 and LX speakers, many language varieties are present. For me, an LX speaker of Spanish, collecting data within this group made me wonder if the participants might be changing their language ‘for me.’ Research shows that people adapt speech to their interlocutors in many contexts, whether to benefit language learners in the classroom (Pappamihiel and Lynn 2016), facilitate LX users’ understanding (Giles, Taylor, and Bourhis 1973; Ylänne-McEwen and Coupland 2000), index identities across national borders (Watt, Llamas, and Ezra Johnson 2010), to avoid or negotiate misunderstandings (Cogo and House 2017), through convergence and divergence within an interactional episode (Auer 2007) or for other reasons. However, a consideration of how this adaptation might play out in a particularly diverse community and its impact on the research process is essential. Two scenarios below illustrate how linguistic adaptation in the SLG may facilitate transnational understanding and show that members of the group are often aware of their language modification.

Venezuelan slang: ‘You won't understand anything we say’

Early in the data collection process, some SLG members asked about my research project. Speaking to myself and Emma³ (a British LX speaker of Spanish), a Venezuelan participant, Pedro, commented on the difficulty of understanding Venezuelan slang.

• Field note⁴ 1: 13 May 2018

  Talking about slang, Pedro told us that we wouldn't understand any of the Venezuelan slang (jerga) or if two Venezuelans talked together we wouldn't understand either.

Here, by conceptualizing ‘insider’ knowledge (jerga) as based on nationality, Emma and I were positioned as ‘outsiders’ by Pedro. Since the comment was directed at two L1 English speakers, one reading could be a perceived lack of Spanish language skills. However, knowing that Emma lived in Spain for three years and having witnessed her ability to understand most conversations in Spanish, I interpreted this comment more generally. In making a distinction, ‘the identity relation of differentiation’ (Bucholtz and Hall 2005, 600), between himself, a Venezuelan, and us, LX speakers – or perhaps more importantly non-Venezuelans, he constructs us as lacking culturally specific language knowledge, local slang. This comment illustrates the potential challenge of understanding multiple language varieties (for both LX and L1 Spanish speakers) while demonstrating participants’ awareness of linguistic and cultural differences and their ability to interactionally construct difference through distinction.

Although I was not required to interpret a conversation between two Venezuelans in the SLG, I did encounter difficulties in understanding, transcribing, and translating a conversation between participants with shared linguistic and cultural knowledge. One such recording, featuring two Latinas (Guadalupe, from Mexico and Ana, from Peru), ultimately resulted in me enlisting help (see challenge 3 on transcriber ideologies). This provided me with some evidence that the mixed L1-LX interactions typical of the SLG involved a good amount of language adaptation and that L1-L1 interaction may prove challenging for me.

Indeed, the members of this group do appear to be ‘changing’ their language, not just ‘for me,’ but for the group. It is well established that ‘people modify their communication according to situational, personal or even interactional variables’ (Williams 1999, 152; see also Gallois, Ogay, and Giles 2005), with similar communication in English as a lingua franca (ELF) featuring negotiation and ‘joint achievement’ (Cogo and House 2017). Yet, in transnational contexts, with a shared language (L1 or LX) and many cultures at play, interlocutors appear to modify their language to be more easily understood. It follows that a researcher with intimate knowledge of the social space and communication within may be better positioned to interpret local conversations. Additionally, if speakers are conversing differently than they might with co-nationals, an L1 Spanish researcher might mistakenly perceive language used within the SLG as ‘incorrect’ through their L1 linguistic or cultural lens since L1 status may carry conscious or subconscious language ideologies that prioritize prescriptive language use. Overall, linguistic adaptation in heterogeneous groups appears to facilitate transnational understanding and is perhaps the norm in social spaces like language groups where in one way or another, everyone is an ‘outsider.’

‘Intentas adaptarte un poco⁵’: metacommentary on language in the SLG

Members of the SLG show awareness of linguistic and cultural diversity within their group through acknowledgement of the adaptation required for localized communication, which has implications for researcher and participant understanding and meaning-making. Gabriel, a Chilean member of the group, makes this type of claim below (as reported in my field notes).

• Field note 2: 4 Feb 2018

  [P]eople bring all different dialects to the table and they approach one another/negotiate the way they talk [. . .] this type of conversation comes up every time they meet. Specifically, talk about how each person says a certain word or phrase.

Referring generally to ‘people’ and ‘all different dialects,’ Gabriel includes anyone who may come to the SLG, using frequency (‘every time they meet’) to indicate the commonness of this behavior. His description of ‘negotiation’ of talk connects with theories of accommodation and ELF, where intercultural interlocutors adapt to their counterparts.

Speaking about LX speakers from an L1 point of view, José Antonio describes his experience of adaptation as unidirectional.

• Excerpt 1⁶: Interview 2, José Antonio, 22 Mar 2019

  I have been told that when I talk to non-native speakers I speak more slowly . . . most of the time I don't realize it.

Through this comment José Antonio shows some awareness of his own convergence, adaptation to another's ‘communicative behaviors’ (Giles, Coupland, and Coupland 1991, 7), but only as reported to him. Rather than claiming agency, he presents his adaptation as a subconscious response to the LX speaker, presumably to facilitate understanding for those he characterized as having limited competence in Spanish (i.e. requiring slower speech). He later acknowledges many attendees’ ‘excellent level of Spanish,’ but when they are ‘a little more limited in their level, you try to adapt a little.’ Here, adaptation is constructed as a more active choice (‘you try’) for accommodating LX speakers.

Adaptation, however, may not always take the same form. For Guadalupe, there is a ‘common’ way of speaking within the group.

• Excerpt 2: Interview 1, Guadalupe, 3 Feb 2019

  I feel that, but maybe I'm also speaking from my experience, that it becomes a more monotonous or more common accent for everyone, I think.

Hedging (‘but’, ‘maybe’, ‘I think’), hesitation markers, and the possessive pronoun ‘my’ indicate that Guadalupe is reluctant to speak for all members of the SLG and takes ownership of her subjective experience. Yet, her comment substantiates those of José Antonio and Gabriel, pointing to noticeable adaptation of the Spanishes used in this transnational group, which she presents as converging (‘more common’) and inclusive (‘for everyone’).

LX speakers are also attuned to the many language varieties and potential for limited comprehension in the group. Emma speaks to this below.

• Excerpt 3: Interview 3, Emma, 17 Feb 2019

  We have many Spanish speakers, from many countries as well, and there is a huge variety . . . with many accents . . . [shared laugher] including those from Andalusia, haha.

Emma first acknowledges the multiplicity of Spanishes within the SLG, described as nationally (‘many countries’) and linguistically (‘huge variety,’ ‘many accents’) diverse. Through use of the personal pronoun ‘we,’ she constructs herself as part of the group, lending authority to her epistemic claim. However, the laughter surrounding her mention of a specific accent may be perceived as making fun of a language variety. An ethnographic point of view, however, draws on previous conversations where Emma and I discussed the challenges we had of understanding a specific group member's Andalusian accent. Thus, since Emma is referencing our prior discussion, we can interpret the laughter as shared, which serves to connect us and points to our own awareness of our limitations as LX speakers.

Within this challenge, we see language adaptation and accommodation as part of the SLG experience. If there is, indeed, a more ‘common’ Spanish spoken in the group, this should be considered when interpreting interaction and is worth reflecting on, particularly as an LX participant-researcher. Familiarity with a sociolinguistic space and ethnographic knowledge can help researchers interpret interaction across language varieties. In the SLG, directly acknowledging and making adjustments for linguistic variation can be seen as a feature of local interaction and one that may be better understood by an LX researcher who may be ideologically or otherwise distanced from L1 norms.

Challenge 2: in which language should I conduct my interviews?

Although some research indicates that conducting interviews is most effective in the participant's L1 (Cortazzi, Pilcher, and Jin 2011), further questions arise when a researcher may feel disadvantaged by or uncomfortable with their own LX status when interviewing L1 (or LX) speakers. Furthermore, with complex levels of ‘insider’/‘outsider’ relations, the ways in which these relationships overlap and intersect can have significant impact on how an interview unfolds and the resulting data (see Merriam et al. 2001). This isn't to say the effects are necessarily adverse. In fact, an ethnographic lens could help draw interesting ideas from data that may not emerge with an L1 researcher. Within this second challenge, two examples will be examined, showing how my LX status affected the planning and execution of interviews and considering which language might be more ‘suitable’ in transnational interviews.

Preparing to interview in the researcher’s LX: what am I really asking?

Although interested in conducting semi-structured, ethnographic interviews, my own linguistic insecurities impacted the interview preparation and execution. Even though the questions were intended for Spanish-speaking interviewees in a Spanish language context, I first prepared my questions in English. On reflection, the English version was mostly for me, since I didn't use the English questions during the interviews. On the positive side, it helped ensure I knew what I wanted to ask and made me feel more comfortable, but it also required translation when preparing the Spanish language version of my bilingual interview guide (as in Rolland, Dewaele, and Costa 2020). This could have resulted in questions that were less ‘natural’ in Spanish or put distance between myself and the interview process.

As an LX researcher, my own perceived linguistic weakness (particularly regarding Spanish in a research context) at times impacted the flexibility with which I had hoped to conduct my interviews. I noticed this when data analysis revealed a few moments of interaction that stuck particularly closely to script, with some opportunities for expansion missed. Likely, if I had been conducting research in my L1, I would have ‘gone with the flow’ a bit more, which may have encouraged my participants to do the same. Of course, my experience is not universal for all LX researchers, however some do feel uncertainty around their language resources (Ganassin and Holmes 2020). Importantly, LX limitations don't necessarily prevent meaningful engagement (Knowles 2019) and Phipps (2013) argues that openness about ‘linguistic incompetence’ could actually lead to a more compelling interaction. Variations on this theme could affect all stages of research, making further consideration and reflection crucial.

An example to illustrate this point comes from the interview with Emma when she responds to the question, Where did you go to university and what did you study? by talking about how she has always wanted to learn more about other cultures.

• Excerpt 4: Interview 3, Emma, 17 Feb 2019

  I studied history in school and after I wanted to learn something about the world. And then, Spanish I don't know because, well, I don't know, it's that, it's interesting, it's like, it was to learn more about other cultures and all that.

This was an interesting topic for me at the time both personally and professionally, yet I responded in a way that, looking back, may have closed the conversation down. I said, ‘well, I have more questions about Spanish later,’ and moved to the next planned question. Through reflection, it could be that my own discomfort around improvising in Spanish resulted in an overly ‘stick-to-the-script’ attitude. Although I did practice my interview questions with a bilingual friend, had I written these questions in Spanish and explored possible responses and sub-questions in my LX, I may have been more prepared, ironically, to improvise. Ultimately, since it sometimes felt challenging to follow up on ideas in the moment, treating the interview data as interactional (Talmy 2011; Wooffitt and Widdicombe 2006; Gibson and Zhu 2016) has allowed for an exploration of communicative and interactional ‘consequences.’ Such as whether those moments were treated as ‘unnatural’ (or not), since ‘speakers recognize choices as either unmarked or marked’ (Myers-Scotton 1983, 115).

Unwittingly, my LX status led me to conduct more structured interviews than intended. Had I anticipated this, I may have looked to alternative data collection tools, such as ‘ethnographic chats’ (Selleck 2017), which are well suited for an LX participant-researcher as they could help alleviate power differentials or lead to more relaxed data collection. However, as Blommaert and Dong Jie say (2020, 56), there is ‘[n]o such thing as a bad interview’ as any discourse reveals potential positionings and encourages the researcher to ask ‘why it went that way’ (58). Therefore, the use of ethnography, and linguistic ethnography in particular, provides tools for the researcher to treat and analyze the interview as an interactional event (De Fina 2019; Talmy 2011), reflecting on the researcher's role in knowledge production. In doing so, there is scope for uncovering and analyzing the interactional significance and impact of one's own contributions, ideologies, or perceived mishaps throughout the interview process.

English or Spanish? Language choice in interviews

There is some academic advice on language choice in research. For example, Ganassin and Holmes (2013) and Phipps (2013) argue for the possibility of researching in a neutral language to address inevitable power imbalances in research relationships (Woodin 2016), however multilingualism is ‘hierarchical and ideologically invested’ (Barakos and Selleck 2019, 364). As such, ‘power-based dynamics [are] inherent in any and all research,’ making power not only something to be aware of during the research context, but also something to negotiate (Merriam et al. 2001, 413). Following research that points to a lack of clear guidance on language choice in interviews, but encourages participant agency where possible (Rolland, Dewaele, and Costa 2020), I attempted to ensure Emma felt she could interview in her preferred language by writing the following.

• Personal communication⁷, Hannah to Emma, 4 Feb 2019

  I wanted to ask if you are going [to the next meeting], if we can do an interview? :) [. . .] and of course if you would like to do it in English that is always possible.

Here, the English language option was framed as a given (‘of course’) or an afterthought, following the request. Thus, Emma's response that it would be great to see me and to speak Spanish after so much time was not surprising. My indirect question and the subsequent lack of a direct response about language choice indicate that the English option was not expected to be taken up (as in fieldnote 3). Instead, Emma makes explicit her desire and intention to speak Spanish, mirroring research where it was difficult to elicit an explicit language choice (Madoc-Jones and Parry 2012). Considering ethnographic knowledge such as our Spanish-language friendship, it appears that Emma treated our interview the way she treated gatherings of the SLG, as an opportunity to speak and practice Spanish. My field notes further illustrate this reading.

• Field note 3: 17 Feb 2019 (following interview with Emma)

  [We had] a conversation, about, the idea that we have relationships, friendships or whatever, in various languages, and how it can be quite weird to change the language. [W]e [. . .] talked about [. . .] coming there, talking to each other, we did that in Spanish, you know. Our friendship is in Spanish and therefore, it wasn't really a question.

Reflecting on this LX-LX interview, the most ‘natural’ choice was Spanish, bearing in mind that language choice is not a one-off decision, but is open to shifts and discussion (Martin-Jones, Andrews, and Martin 2017) and that our extended engagement (Martin-Jones, Andrews, and Martin 2017) requires consideration, negotiation, and evolution of researcher/participant dynamics. Further, Emma and I displayed a desire to stick to our LX Spanish throughout the interview. Perhaps our shared LX was ‘a resource for the creation of a “third space” in the interview’ (Rolland, Dewaele, and Costa 2020, 285), with a more equal balance of linguistic power leading us to feel more comfortable. This idea is supported by Emma's second recorded conversation, in which she speaks of a conversation between herself and two L1 Spanish speakers. Oliver (a British male) had just asked if she completed a recording before.

• Excerpt 5⁸: Recording 10, 14 Oct 2018

Emma confirms that she did a recording, but didn't talk much (l.1 & l. 3) followed by Oliver's genuine request for knowledge (‘why?,’ l.4). This results in Emma pointing out that the interlocutors spoke ‘more naturally’ (l. 5), citing nationality (Española, l. 9) and region (América Latina, l. 11). In doing so, she constructs her interlocutors as having linguistic authority and expertise due to their nationalities and L1 status, positioning them as idealized native speakers (Leung, Harris, and Rampton 1997) and invoking the problematic ideology of native speakerism, that ‘so-called “native speakers” are the best models and teachers’ (Holliday 2006, 6).

By speaking about her previous interaction in this way, Emma constructs herself as lacking Spanish language skills, in contrast to her L1 interlocutors. This sheds light on her choice to interview in Spanish, suggesting she may have felt more comfortable recording an LX-LX conversation, particularly since invoking shared ‘foreigner’ status can neutralize linguistic power between researcher and researched (Ganassin and Holmes 2020). As with Emma and I, choosing to conduct interviews in a common LX could be more appropriate in transnational contexts and may even reduce linguistic power dynamics.

Challenge 3: dealing with issues of linguistic and/or cultural (mis)understanding

As outlined above, an assumption is often made that when doing research in a shared language, cultural aspects are also shared or understood. Even if that were the case in some contexts, within transnational social spaces, such as the linguistically and culturally heterogeneous SLG, it would be impossible to share linguistic and cultural knowledge with all participants. This brings up potential issues with understanding, which can become apparent when transcribing conversations for analysis or translating data for dissemination. Below, two areas are discussed: how ideologies affect transcription or translation and questions about communicative norms in local contexts.

Transcriber and translator ideologies: 'Her grammar wasn't good at all and the speech was messy

When doing multilingual research, it is essential to account for sociolinguistic context during transcription (and translation) processes (see Hepburn and Bolden 2017; Vakser 2016). While transcription can be challenging in any language, as an LX researcher lacking expertise in some of the Spanishes present in my data, at times I sought help from L1 speakers, who I first viewed (problematically) as ‘experts.’

One conversation that I struggled to transcribe took place between two Latinas (as mentioned in challenge 1) and was affected by their overlapping linguistic and cultural repertoires and friendship outside of the SLG. To overcome this, I sought help from a friend of a friend who did professional translation work. I initially paid little attention to the fact that she was from Spain and enlisted her help with transcribing a small chunk of conversation. Yet, L1 speakers (like all people) are not without language ideologies and ‘[w]ho is translating whom, when and for what purposes are all political decisions’ (Kalocsányiová and Shatnawi 2022, 212), which extends to transcribing, especially across language varieties. This was highlighted when the translator asked about the country and nature of the conversation, saying that the ‘woman's grammar wasn't good at all and the speech was messy.’ In contrast, a Chilean friend later helped with the same data, without notable difficulty.

These responses to interaction, in which language ideologies are visible, underline the importance of seeking culturally aligned transcription or translation help where necessary, considering reflexivity (Vakser 2016), and acknowledging power imbalances (Bucholtz 2000). Although not always possible, transcription may be best when done with a degree of closeness to the participants, where ethnographic knowledge of the context heightens one's ability to see and understand what is typical in the local space. Andrews (2013) argues that interpreters should be an integral part of a research team, however I would add that a consideration of one's own cultural knowledge and language ideologies must also be part of the conversation.

In a sense, my status as an LX Spanish speaker and lack of a single cultural connection allowed me to step back and take a broader view on the language used in the space (a positive aspect of outsider status, as indicated by Merriam et al. 2001). Thus, I may have avoided certain language ideologies and stereotypes related to language varieties that may be inherent for some L1 speakers (a critique of insider researchers, see Merriam et al. 2001). I also advocate a view of transcription as a step in the analytic process (Bucholtz 2000), allowing the researcher to become more familiar with their data, and enabling relationship building and engagement within the project (Ganassin and Holmes 2013). As Bashiruddin (2013) notes, even a transcript that someone else does ‘verbatim,’ might make little sense to a researcher who has not listened to the data themselves.

What is 'normal' in a transnational social space?

Something that feels unnatural to an L1 speaker of Spanish due to their potentially monolingual and/or monocultural upbringing, may be unmarked in a transnational social space where Spanish (in this case) is essentially used as a lingua franca. In linguistically diverse communities, a multitude of features emerge from myriad varieties and competencies, complexifying interpretations of when differences are locally relevant or notable (or not).

An example comes from the same conversation in excerpt 5 above, where three LX speakers are discussing how to say ‘stage fright’ in Spanish. I am also present, having recently started the recording, and this short excerpt features just Oliver and myself.

• Excerpt 6: Recording 10, 14 Oct 2018

Ethnographically, I know it is common within the group to ask the nearest L1 speaker of Spanish for help when language questions arise, treating them as ‘experts.’ However, without L1 speakers present, there is no default Spanish language ‘authority.’ Instead, in this excerpt, Oliver offers some translations of ‘stage fright’ (miedo o pánico escénico, ls. 1-2). However, when I start to repeat the unfamiliar word (escénico, l.5), I pronounce it esˈteniko, indicating I heard Oliver's θ as a t sound. Oliver repeats his previous pronunciation (esˈθeniko), with a regional Spanish use of the ceceo (l.6). Another LX participant, who shared my pronunciation (eˈseniko), steps in soon after to clarify it is a ‘c,’ rather than a ‘t.’ Knowing that Oliver studied in Spain and does not habitually correct pronunciation within the group, I interpreted this comment as his own pronunciation of the word, rather than a ‘correction’ of my variety. Yet, without ethnographic knowledge or with the influence of L1 language ideologies, the same line could be interpreted as a correction of my differing pronunciation or a push for a preferred variety. As such, LX status, paired with local knowledge, may result in researcher advantages for interpreting, transcribing, or translating transnational interaction.

Discussion

This article has explored three challenges faced as an LX researcher in a transnational context. First, language modification for multiple L1 and LX varieties was seen as a salient feature of the SLG, while (linguistic) ethnographic methods and tools may enhance the ‘detection’ of normative or non-normative language (Bucholtz and Hall 2004). Furthermore, within a transnational space, an LX researcher may be particularly well placed to interpret relevant interactional events and avoid readings of certain divergences from L1 usage as errors. Although speech adaptation is common, it is potentially more prevalent in transnational social spaces where it may have differing interactional consequences and facilitate transnational understanding.

The second challenge highlighted the idea that conducting research in an LX and a linguistically and culturally heterogeneous group, can result in different questions, both for researcher and participant, such as which language(s) to use in which contexts and how to prepare for interviews. Ethnography helped mitigate possible misunderstandings and provided a lens through which to interpret interviews. Furthermore, language choices in (potentially) multilingual interviews are shown to be an essential part of complex and intersecting cross-linguistic and cross-cultural research. Preparation should include awareness of languages and varieties operating in and around the research context (Andrews, Fay, and White 2018) and a consideration of the most adequate tools for LX data collection.

In the third challenge, having LX status and being slightly ‘removed’ from certain ideologies or expectations provided possible advantages. Since it is impossible to share language(s) and culture(s) with all participants in heterogeneous groups, arguments that researchers should be ‘insiders,’ appear too simplistic (also critiqued in Ganassin and Holmes 2020). Indeed, researchers can ‘experience moments of being both insider and outsider’ (Merriam et al. 2001, 416), while LX researcher status may actually bridge the ‘insider’/‘outsider’ gap (Knowles 2019). To study a transnational group, where people may connect on a global level more readily than a national one, international ties, and/or sharing a multilingual and multicultural orientation with the participants, may increase a researcher's understanding and belonging within a linguistically diverse group. Ethnographic tools and close attention to how varieties are used and made relevant in interaction can help untangle underlying complexities of language and culture in transnational social spaces.

These insights, demonstrated through data collected by an LX researcher within a transnational social space, are applicable and relevant for researchers encountering multilingualism at any stage of the research process. Within a sociolinguistics of globalization (Blommaert 2010), communication crosses national, linguistic, and cultural lines, particularly in heterogeneous contexts where a researcher's language(s) may complement or impede understandings and interpretations. ‘[T]he building of researcher–researched relations [. . .] is a matter of ongoing negotiation and possible shifts in the nature of the relationship over time’ (Martin-Jones, Andrews, and Martin 2017, 191–192), furthering the need for extended ethnographic engagement when researching across languages and cultures. As researchers increasingly encounter transnational participants and colleagues, accommodations must be made for asymmetric multilingual practices (Ganassin and Holmes 2013). The reflexive questions, experiences, and data presented here will help researchers prepare for dynamic and complex intersections of language and culture in research settings.

Conclusion

This article argues, through consideration of three challenges of researching multilingually within the SLG, that spending time in a research context to build an ethnographic understanding of the sociolinguistic space helps overcome potential challenges for an LX researcher and in transnational contexts. Firstly, language adaptation is, in itself, a key feature of transnational social spaces and worthy of study in its own right, rather than a researcher barrier. An ethnographic lens can allow the investigator to focus on moments when language adaption is meaningful to the community, rather than taking an outside view on what is ‘correct’ or ‘incorrect.’ LX status may also aid in the interpretation of transnational language use. Secondly, an LX-LX interview might be more ‘natural,’ even when two interlocutors share an L1. This highlights the need for a more nuanced approach to complexity and deliberate decision making in regard to language policy in one's research context (e.g. Rolland 2023). Thirdly, linguistic and/or cultural misunderstanding can (and likely will) occur at a researcher, transcriber, and/or translator level, but can be mitigated through ethnographic knowledge of local language practices.

By prioritizing familiarization with participants in a research project, a more context-sensitive rendering of the data can be produced, giving priority to an emic understanding within a given community. Taking the view of ethnography as a ‘full intellectual programme’ with a particular ‘perspective on language and communication’ (Blommaert and Jie 2020, 5), the researcher can approach the research process holistically, embracing its multilingual complexities. Therefore, researchers who are conducting investigations in languages they may not feel fully comfortable with, can maximize their own abilities to understand the interactional context. The proposed reflexive and ethnographic approach will be particularly fruitful for those investigating linguistically diverse communities and transnational contexts, but has the potential to enhance any research with a multilingual element.

Acknowledgments

I would like to thank, first and foremost, my colleagues and co-editors Louise Rolland and Pernelle Lorette for their support and guidance on this paper. Many thanks to the reviewers for detailed and thought-provoking comments that greatly strengthened this piece of work (any and all mistakes remain my own). I am also appreciative of comments and support from my doctoral supervisor, Jean-Marc Dewaele. Thank you to Nuria Polo-Pérez for her translation of my abstract, title and keywords and to Gonzalo Pérez Andrade for further translation suggestions. Many thanks to the participants of my study: without you, this research would not exist.

Disclosure statement

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

Additional information

Funding

This work was supported by Birkbeck, University of London in the form of a School of Social Sciences, History, and Philosophy Research Scholarship.

Notes

1 In response to long standing critiques of the ‘native’/‘non-native’ dichotomy (e.g. Rampton 1990; Kamhi-Stein 2016) and in an attempt to avoid value-driven terms that imply inferiority, Dewaele (2018) has proposed the use of L1 vs. LX user, with LX referring to any language(s) learned after age 3, to any proficiency level. The term prioritizes order of acquisition, yet leaves room for overlapping linguistic repertoires, such as multiple L1s and inclusion of all learned languages. Although any distinction as such is inevitably an oversimplification, one's relationship(s) to language(s) certainly has an impact on their language use, making it relevant to discuss L1 and LX differences in this paper and in similar heterogenous contexts.

2 While personal pronouns have traditionally been avoided in academia to project ‘objectivity,’ ethnographic approaches (e.g. Hammersley and Atkinson 2007) and researcher reflexivity (Patiño-Santos 2019) encourage acknowledgment of researcher roles and positionality. Among other ways that the personal is performed in academic work, ‘the reflexive I in research and writing has become, at least as far as research that has made the linguistic turn is concerned, standard practice’ (Pennycook 2005, 299).

3 Pseudonyms are used for all informants and the group studied throughout the paper.

4 Fieldnotes are primarily in English, with some recorded as audio notes, so hesitations markers may be included.

5 ‘You try to adapt a little.’

6 All interview and conversation data were originally recorded in Spanish, all translations are my own. Please note that only the English language version of the interview data is presented in the text due to space constraints.

7 Original messages were in Spanish.

8 The transcription conventions of conversation data are adapted from Jefferson (2004).