Language choice when conducting interviews with multilinguals on multilingualism

When we conduct interviews with multilinguals on their multilingualism, what ironically remains invisible and not well discussed in published studies is the language of choice. In this article, Martin Cortazzi, Nick Pilcher, and Lixian Jin (the veteran duo of intercultural communication scholars Cortazzi and Jin) investigate what effect the language of choice has on interview findings. Their article is in two parts. First, they review interview-based studies in China and the Chinese diaspora, analyzing the language of choice and the pros and cons of various choices. Second, they conducted their own study in which seven participants were interviewed in English, then in Cantonese or Mandarin about 10 days later, with the same questions. After analyzing quantitative and qualitative differences in the data, they conclude with suggestions for dealing with the challenges of multilingual interviewing.

Cortazzi, M., Pilcher, N., & Jin, L. (2011). Language choices and ‘blind shadows’: Investigating interviews with Chinese participants. Qualitative Research11(5), 505-535.


We know languages develop in domain-specific ways, with disciplinary and professional discourses developed from the language(s) of higher education and professional experience, metalanguage for linguistic analysis developed from the language(s) of schooling, and resources for expressing emotions, complex thoughts, and delicate pragmatic/social positioning developed from the language(s) of primary interaction. This is already a big reason to draw on the whole language repertoire—to translanguage—during interviews, if interviewer and interviewees share languages, to optimize interview participants’ range of expression. Even if Cortazzi et al. don’t use the word “translanguaging” in their article, they discuss translanguaging as a solution to the problems that arise from language choice.

Cortazzi et al. begin their paper by explaining that there is increasing research in English involving Chinese L1 participants (that is, typically Mandarin or Cantonese as L1), partly because Chinese students form the largest group of international students in many countries in the world. They note that this has implications for researchers, including Chinese L1 researchers doing research in English, who conduct interviews in Chinese and/or English and write up the research in English. Cortazzi et al. begin with a discussion of three scenarios: (1) interviewing in English because the interviewer doesn’t speak Chinese, (2) interviewing in English, which is an L2 for both interviewer and interviewee, and (3) interviewing in Chinese, which is an L1 for both interviewer and interviewee.

In Scenario 1, the interviewer could be missing out on data due to not speaking Chinese. There are ways of getting around this, but they come with costs of their own: the cost of money and time training bilingual research assistants, the higher cost of employing interpreters, the emotional debt of recruiting family members or volunteers, the additional outsider/insider dynamics that come with bringing in another person, feasibility of time, issues of collaboration and power dynamics, degrees and commitment to the research project, care and concern for other people, and issues of disclosure—the more people are involved, the less privacy the interviewee has, and each person may affect what the interviewee says in a different way. This is why, in this article, positionality comes across as more important than language, but of course, language is a key tool for negotiating positionality.

Now, what if interviewer and interviewee speak Chinese as L1 and English as L2? If they interview in English (Scenario 2), in which the researcher is typically more proficient, the researcher may overestimate the participant’s ability to express themselves in English (i.e., not realize the extent to which the participant is holding back), which can also happen with a researcher who only speaks English. Conversely, if they interview in Chinese (Scenario 3), some professional or topic knowledge that is the subject of the interview may have been acquired in English. As a third option, if they translanguage, Cortazzi et al. state: “This should allow participants to choose languages according to ease of expression or familiarity with the topic. To offer participants the choice may help establish rapport and trust, and move the locus of control towards the participant. It may be beneficial for the interviewer to use both languages when explaining a question” (p. 510). Yet bilingual questioning is not just to help the participant better understand the question—translanguaging, if normalized throughout the interview, can build rapport. Even if the interviewer does not speak Chinese, a little bit of translanguaging (e.g., written cue cards as prompts) can help (p. 510).

Findings of the systematic literature review

Cortazzi and Jin next review 28 studies from 1996 to 2011 (the year of publication of their study) in Hong Kong, the U.K., China, Australia, or New Zealand, in which researchers’ first language was Cantonese, English, Mandarin, or Swedish. Some researchers identified as having two L1s, Cantonese/English or Mandarin/English. The interview language was Mandarin in 8 studies and English in 2; in the rest, it was unspecified. Sometimes, researchers seemed to make the language choice for participants; in other studies, they asked participants what language they would like to interview in. There is a collection of comments by researchers in each study justifying language choice on pp. 512-516.

Cortazzi et al. then go on to discuss how the complexity of the ethnic and first language differences between an interviewer and participants is more marked in interviews with linguistic minorities in China. Among the 55 officially recognized minorities, 53 speak at least one other language apart from Mandarin; there are many other additional languages spoken by Chinese minorities that are not officially recognized, much to their detriment in education and during crises like COVID-19, plus English and other foreign languages. “This opens up trilingual choices for interviews: Mandarin, one or more minority language, and English” (p. 517). As I explained in last week’s blog post about the similar great linguistic diversity in Senegal, some people do not like to use the national language because it is the language of the dominant group; thus, English may be seen as a more neutral lingua franca between two people, e.g., a Mandarin L1 and Cantonese L1 speaker in Guangdong, or a Mandarin L1 and Min L1 speaker in Taiwan. The reduced cultural markedness of English in certain contexts, as colonially charged a language as it is, is discussed in Higgins’ (2009) research in Tanzania, when an employee greeted his boss with a relatively unremarkable salutation in English because he felt the Swahili greeting was too formal and subservient.

Cortazzi et al. next review five studies on ethnic minorities in China. Four of these studies did not go into language matters that much, even if there was “a highly detailed analysis of the participants’ use of language in daily life but no comment on the language(s) of the interviews” (p. 518). One book-length study, however, by Hansen (1999), described an interesting event in its prologue (p. xviii). Hansen was a Norwegian L1 speaker from Denmark studying Naxi and Dai (Tai) ethnic minorities in Yunnan. She wrote:

Usually only the interviewee, my assistant Peitong (an officially approved and obligatory assistant provided by my host organization) and I were present. When interviewees spoke Chinese, interviews were conducted in Chinese. Otherwise they were conducted in Tai or Naxi through the assistant’s interpretation. Some villagers would gather to listen, or students who shared the dormitory of an interviewee would be present. Sometimes I interviewed two or three students or teachers together. Occasionally the interview developed into an interesting discussion between students, sometimes of different ethnic groups, or between villagers. (as cited in Cortazzi et el., 2011, p. 518)

Once a translanguaging space has been established, it seems to result in an interaction order of mutual assistance. The question is how to create and maintain this translanguaging space in the first place, one that brings not only official languages but others into the discussion, especially when translanguaging spaces (see Li, 2011 and Li & Zhu, 2013) are typically ludic/playful arrangements between friends who share languages, not in official interview settings like this, and even when a translanguaging space appears in an interview (Li, 2011), the interviewees have rapport with the interviewer who also shares the languages.

Throughout this article, Cortazzi et al. imply that each additional participant brought into the conversation opens up and closes down opportunities for self-disclosure, as there are some things that we feel more comfortable disclosing to in-group members, or that in-group members will better understand, while there are also things that we can disclose more easily to outsiders because we will not be judged.

Findings of the double-interview study

The data from the second, empirical half of Cortazzi, Pilcher and Jin’s article comes from a larger study about Chinese doctoral students in the U.K. These doctoral students were interviewed in English four times during various stages of their Master’s dissertation writing, and their experiences were the subject of Pilcher’s PhD dissertation (Pilcher, 2007). There were 45 participants, but 7 were selected as special: these 7 people were interviewed by a Chinese speaker (possibly Jin) 1-2 weeks after each interview with the English speaker (likely Pilcher), and asked the same questions in Chinese. The Chinese interviews were called “blind shadow” interviews because the English-speaking interviewer did not know which 7 of the 45 participants had been interviewed in Chinese until all participants had completed their degrees, upon which point Cortazzi et el. examined the data that they would present in this article. How this repetition was justified to the 7 participants was not entirely clear, but at least, the English interviewer explained to them that if they were contacted by the Chinese interviewer, they must not reveal this to the English interviewer, and feel free to say what they felt in Chinese without considering the English interviewer, who would not know what they said until after they graduated.

Cortazzi et al. are honest about the many confounding variables, not just that Chinese always followed English but also the time gap. For example, in 10 days, something may have happened to make people feel differently about their Master’s dissertations, which would be the reason why they gave different answers in the English and Chinese interviews. While many of the interview questions were about states rather than progressions, Cortazzi et al. admit that some participants might feel the second interview was redundant, and it could also be influenced by the memory of the first (pp. 521-522). Nevertheless, they do come up with some interesting findings.

The first finding is that English interviews were, on average, longer than Chinese ones; while Chinese interviews typically lasted 20-30 minutes, English ones typically lasted 30-40 minutes (p. 523). This means that people need extra time answering in their L2.

The second finding is that there was considerable difference in the content of the interviews, with the most similar pair of interviews being 65% similar (i.e., same answers given, but different language). The least similar pair of interviews had a 43% overlap. On average, there was a 51% overlap (averaging the 7 participants across their English and Chinese interviews). Side note: I think that if I (the author of this blog) were interviewed with the same set of questions ten days apart, both times in English, my only fluent language, I would not be surprised if there was only a 50% overlap in what I said, as I cannot remember much after 10 days what I said in the previous interview. If there were a much higher overlap in the data, I would suspect the English-speaking interviewer and the interviewee were in cahoots, rehearsing what the interviewee would say to the Chinese-speaking interviewer.

What is more meaningful, then, are the content comparisons and what they show us about language and positionality (i.e., social positioning). Sometimes, the answers were the same in both languages.

Cortazzi, Pilcher, & Jin (2011), p. 524

In this case, the question is fairly neutral to the interviewee, and the answer is matter-of-fact. Other times, there was greater personal divulgence in English. This occurred when the participant showed a stance that would not be taken well by a fellow Chinese.

Cortazzi, Pilcher, & Jin (2011), p. 525

Here, the participant expresses in English that he likes to laze around and that this is the reason he is not happy with his progress. In other words, he may “feel it is more acceptable to admit to laziness to the UK ‘outsider’” (p. 525).

The same thing occurred on the other side: there was greater personal divulgence in Chinese when the participant showed a stance that would not be taken well by the English-speaking interviewer, due to their investment in the project or their institutional role.

Cortazzi, Pilcher, & Jin (2011), p. 526

Here, the participant expresses in Chinese that he is more relaxed interviewing in Chinese, but he said he was relaxed during the English interview. (Of course, there are degrees of relaxation, but the participant is not likely to tell the English-speaking interviewer they would feel more comfortable if the interview is in Chinese, especially if the interviewer doesn’t speak Chinese.) See also this example where more was disclosed in Chinese:

Cortazzi, Pilcher, & Jin (2011), p. 528

Here, the participant bluntly expresses in Chinese what he thinks about the mercenary attitude of the department towards making people write dissertations (though I do think this was hinted at when he talked to the English interviewer).


The article doesn’t really answer the question “What is the effect of interviewing in Language X?” or “What are the differences between interviewing in L1 and L2?” (apart from fluency and time length). Those questions cannot be definitively answered due to a vast array of individual and contextual differences. What the article does accomplish is to outline the individual and contextual factors that researchers should consider when choosing language of choice during interviews. Translanguaging is also discussed, even though the authors don’t evoke that particular term. Cortazzi, Pilcher, and Jin have some suggestions about how to conduct the interviews in light of the individual differences and contextual factors:

  1. Their systematic literature review shows: Use all language resources to the greatest extent possible. If there is little linguistic overlap between interviewer and interviewee, make pre-made question cards in the interviewee’s language (p. 510). (If an interpreter is present, I think it’s best if the interpreter and interviewer already know each other well and can work together to make the interviewee comfortable.)
  2. Their study with 7 participants shows: If a question potentially involves threats to face, frame it in the third person. e.g., “Reflexive questions about a project can be better framed through greater researcher/project distance. For example, instead of asking whether participants had felt relaxed or nervous, the question could be framed How could people doing projects such as this one help participants to feel more relaxed? This allows participants to express thoughts without any potential face-threatening comments” (p. 527).
  3. The study with 7 participants also shows: If a question could be affected by cultural insider/outsiderness, frame it as a statement that participants could agree or disagree with (pp. 528). For example, in the case of “For the department what do you think the aims of the dissertation are—why do they want you to do it?”, it might have been better phrased as, “For the department, the purpose of making students write a dissertation is…” (open-ended question), or “The purpose of making students write a dissertation is just to earn money. Agree or Disagree?” (yes/no question). I think this wording positions both the researcher and the participant as temporarily removed from the culture, as if looking down on it from above.

In their conclusion, the researchers also have suggestions for designing and framing your own study:

We recommend the following steps. At the design stage, to consider the possibilities and implications of different language choices available (with the above scenarios in mind). In interviews, first, to talk about likely face threatening topics by distancing them from the interviewer and the participant (e.g. through third person reference); second, to allocate greater time when interviewing participants in their second language; third, to encourage and integrate the use of translation devices such as bilingual dictionaries in interviews; fourth, to acknowledge [e.g., in presentations and publications] that the data collected through second language interviews may be qualitatively different from first language interviews. … [Fifth,] researchers need to consider and discuss explicitly language choices and their likely effects—not only in their research reports but also with participants, who as we have shown do not always take the assumed choice. Discussing the influence of language choice at the end of an interview may be a useful validation check for researchers on participants’ perceptions and feelings of these issues. (p. 529)


Higgins, C. (2009). English as a local language: Post-colonial identities and multilingual practices. Multilingual Matters.

Li Wei (2011). Moment analysis and translanguaging space: Discursive construction of identities by multilingual Chinese youth in Britain. Journal of Pragmatics43(5), 1222-1235.

Li Wei & Zhu Hua. (2013). Translanguaging identities and ideologies: Creating transnational space through flexible multilingual practices amongst Chinese university students in the U.K. Applied Linguistics34(5), 516-535.

Pilcher, N. (2007). Mainland Chinese learners and U.K. supervisors: Perceptions and experiences of completing Masters dissertations. PhD dissertation. Heriot Watt University, Edinburgh.

Published by annamend

Assistant Professor in the Department of Linguistics, University of Illinois at Urbana-Champaign

%d bloggers like this: