The structured narrative interview

In this study, Greimas’s work on narrative structure is used to improve a specific practice: the research interview. In the social sciences, narrative interviewing often consists of collecting data from which a narrative is then constructed through analysis afterwards. In the interview method presented here, the interviewer instead prompts the interviewee to construct a narrative. We introduce the method, contextualize it by comparing it to previous and contemporary interview methods, and illustrate it with a small, sociolinguistic study: students ( n =12) from a humanities faculty and a science and engineering faculty at a Dutch university were interviewed about experiences with the use of different languages than the language of instruction in an international learning environment. The method allowed for smooth data collection, due to its narratively structured questioning and consequent rich data. Moreover, using narrative structures to guide the interview also facilitated easy analysis and comparison of the stories.


Introduction
As argued throughout this special issue, when narrative work -the creation, circulation and interpretation of stories (Gubrium & Holstein, 2009) -is a significant aspect of a specific practice, practitioners and researchers may benefit from theories, concepts and methods developed within narratology or other forms of narrative studies.Applied narratology, creating impactful links between theories and practices of narrative, aims to improve such practices, or offer the opportunity to critically reflect on them (Moenandar 2018).In this paper, we contribute to this emerging field by presenting an interview method that adds to the arsenal of qualitative research methods for collecting narrative data, and has wider applicability in contexts as diverse as journalism, career counselling and institutional research.Informed by Algirdas Julien Greimas's work on narrative syntax, it provides an efficient way of generating similarly structured narratives recounting specific events, allowing for easy analysis and comparison between interviews.To show how this method works, we present a small-scale research project on linguistic diversity in higher education.Before introducing the method, we give a short history of the interview as a method in the social sciences, including a discussion of previous narrative interview methods, and discuss what our method adds to current practices of narrative interviewing.

The interview as research method
From neutral conduit to construction site Historically, within the social sciences, the interview was conceived as "a pipeline for transmitting knowledge" (Holstein & Gubrium, 1995, p. 3, italics ours) and "a conduit for transporting experiential knowledge from the respondent, on the one side, to the interviewer and sponsoring agents, on the other" (Gubrium & Holstein, 2003, p. 3, italics ours).In line with a dominant positivist paradigm, up until the late twentieth century, the assumption was that there was an external reality, for which the interviewee could be an informant (Platt, 2001).To avoid that the interview site itself would become a source of distortion of the data (Holstein and Gubrium, 1995), the main format was the epistemological interview, with standardized, pre-prepared questions designed to gather data that were objective and neutral (Denzin, 2001), unadultered (Holstein and Gubrium, 1995), accurate and unbiased (Platt, 2001).
Fear for data contamination shaped interview methods well into the 1980sand still does within some strains of the social sciences.Variation in answers due to multiple interviewers was seen as a problem, and any impact on the outcomes by the interviewer as an error.Interviewers were trained in accordance.An interviewee was to be approached as a vessel of sedimented, factual knowledge, and of stories that the interviewer could and should retrieve without disturbing the content.The interviewer should therefore refrain from interruptions and leading questions and remain, in the words of Holstein and Gubrium (1995, p. 8), "epistemologically passive".To increase standardization, strict interview protocols were designed in which scenarios for questions and answers were heavily scripted.
In the 1980s, the linguistic turn in the social sciences introduced an understanding of language as charged with multiple meanings, used in a particular order (grammar), context (social, cultural, political, economic, et cetera), and genre (sarcasm, irony, et cetera).Attempts to exclude variation were now seen as Sjoerd-Jeroen Moenandar et al. futile, because variation was considered to be inherent to communication.New methodological trends emerged, in which both interviewer and interviewee were regarded as active agents, bringing their biographies to the interview (Fontana 2001).Moreover, the power balance shifted from the privileged interviewer asking prying questions, to a natural conversation between equals (Taylor and Bogdan, 1984 as cited in Platt, 2001).Yet, occasionally the conduit and pipeline metaphors seeped back into descriptions of the interview and its purpose: to "tap into what is important to people" (Taylor andBogdan, 1984, in Platt, 2001, p. 40, italics ours) or gain "access to people's ideas, thoughts and memories in their own words" (Reinharz, 1992, cited in Platt, 2001, p. 41, italics ours).The data remained to be found through the interview, rather than constructed during it.
With the postmodern turn in the social sciences around the turn of the Millennium, the epistemological status of data was reconsidered: "Standardized representation has given way to representational intervention, where the dividing lines between fact and fiction is blurred to encourage richer understanding" (Gubrium & Holstein, 2003, p. 3).This conceptualization of the interview as a site of knowledge construction (Holstein and Gubrium, 1995;Gubrium & Holstein, 2003) where data is not collected, but produced, suited the new constructivist paradigm within the social sciences.

The interview as a narrative environment
Any qualitative interview includes narrative work.A regular so-called semistructured interview, for instance, is geared towards eliciting interviewees' attitudes, experiences, and opinions, with an interview protocol where the interviewer encourages them to elaborate on topics with prompts from a theorydriven list (Dörnyei 2007).The expectation is that the interviewees, selected for having some stake in the research topic, will provide a significant perspective (Collins, 2010;Given, 2008).This perspective will often be conveyed through stories, but these normally do not interest the interviewer as such: the focus lies on the content, rather than the narrative form as a specific way of negotiating that content.However, there has been a growing interest in interview data as narrative data: the stories that interviewees construct during the interview do not reflect their social world as it is, nor how they experience it, but rather reveal how, through narration, they make sense of those experiences and ascribe certain meanings and values to them (Ricoeur, 1992;Korthals Altes, 2014), either consciously or unconsciously (Murray & Sools 2015, p. 138).To gain such insights, stories shared during qualitative interviews can be subjected to a narrative analysis (Moenandar et al., 2022).
The structured narrative interview [3] As said, any qualitative interview is also a "narrative environment" -a space where narrative work takes place (Gubrium & Holstein, 2009, p. xvii) -but researchers specifically interested in the narrative working through of experience may try to enhance this aspect of the interview setting and "actively seek to engage participants in narrative work" (Baboulene et al., 2018, p. 33), inviting them to "organize their account in a story format" (Murray & Sools 2015, p. 137).The common assumption is that such interview formats do not follow "a definite protocol" (ibid., 138), as they are, typically, not guided by topic lists.However, there certainly is some protocol to most narrative interviews, given that the interviewer is looking for something specific: stories.
That protocol will, first of all, be informed by whether or not the interviewer assumes those stories pre-exist the interview.If they do, this implies the aforementioned conduit metaphor and a positivist conception of what stories are: the interviewer's task is to unlock the stories the interviewee has stored somewhere in their mind, and they will follow a protocol geared toward this.If, instead, the interviewer sees a narrative as "the telling of a story by someone to someone on some occasion for some purpose" (Phelan, 1998, p. 800), the protocol will be designed to provide the interviewee full opportunity to engage in such "story construction" (Gubrium & Holstein, 2009, p. xvii).This protocol may be very open ended (Riessman 2008), with the interviewer providing little to no instruction, relinquishing "control and [allowing] the participant to develop the direction of the narrative" (Murray & Sools, 2015, p. 138).Such an approach is very much what Mishler (1986), an early developer of the form, envisioned a narrative interview to be.This is not necessarily the case, though, and instead, the protocol may also be more instructed, following an interview guide that consists of specific narrative questions designed to establish narrative details such as events and chronology (Jovchelevich & Bauer, 2000;Wengraf, 2001).If we place these two types of poles that inform the design choices for the protocols of narrative interviews on crossing axes, we can visualize four ideal-typical forms (Figure 1).
Often, interviews will contain elements of two or more of these types.However, given that these axes make explicit often implicit assumptions about the ontological status of the research object (vertical axis) and one's definition of a narrative (horizontal axis), this model can help to make more conscious choices when designing narrative interview guides.It falls beyond the scope of this paper to discuss examples of each type.Suffice it to say that our method falls firmly under type III, which is why we speak of a structured narrative interview. [4] Sjoerd-Jeroen Moenandar et al.

Figure 1. Ideal-typical forms of narrative interviews
The structured narrative interview Positioning oneself on the horizontal axis of Figure 1 when designing a narrative interview protocol, may depend on whether one works with a clear definition of narrative.Riessman (2008), for instance, argues that because it is difficult to define narrative, the narrative interviewer should refrain from all too explicit instructions.If, instead, the interviewer demarcates narrative more clearly, this often means their protocol will include at least some instructions that will help to elicit specific elements or structures implied by that demarcation.This choice may also be pragmatic: a clear working definition of narrative clarifies what the interview is supposed to yield, allowing for a more effective interview design.Since not all research projects come with the time and financial resources needed for the lengthy fieldwork that proponents of leaving narrative undefined often argue for (Riessman, 2008;Murray & Sools, 2015), this may be useful for at least some narrative interviewers.
For this, we turned to Greimas's structuralist work on narrative syntax, which was an attempt to map an underlying grammar regulating all narratives.Whether one accepts the universalist claims of this work or not, it is still, arguably, one of the most concrete attempts to conceptualize the structure that underpins any narrative.In an interview setting, such a structure can help the interviewer to prompt The structured narrative interview [5] interviewees to share narratives.We were also attracted to its simplicity and intuitive nature.This has given it an afterlife beyond the heyday of structuralist semiotics as a tool for the analysis of narrative, not only within the study of literature (Hoek, 1985;Korthals Altes, 1992), which was the central concern of Greimas's own work and the structuralist narratology inspired by it, but also within organizational and educational studies (e.g.Robichaud, 2003;Moenandar & Huisman 2019).We will briefly sketch some of the main aspects of Greimas's narrative syntax, before discussing how we used these to design a protocol for narrative interviews.Fundamental for any narrative, according to Greimas, is that its beginning and endpoint are never the same.Between these, a transformation -a change of state -takes place.This change is set in motion because a subject desires an object and attempts to acquire it, or, if the subject already possesses the object, hold on to it.Every subject comes with a narrative program, which is the trajectory the subject covers throughout the narrative.It consists of four stages: (i) the manipulation stage: the events that lead to the subject of a narrative desiring a certain object; (ii) the competence stage: the events through which the subject acquires the ability to act in order to achieve this object; (iii) the performance stage: the subject either achieves the object, or not; and (iv) the sanctions stage: the subject's actions are evaluated and sanctioned.Longer, or more complex narratives may consist of any number of narrative programs, but tend to have an overarching program with which the narrative as a whole can be summarized (Greimas & Courtés, 1982;Hoek, 1985;Greimas, 1991).
Any narrative revolves around the relation of desire between (1) a subject and (2) an object.These are the main actants, or roles, of a narrative: "beings or things that participate in processes in any form whatsoever, be it only a walk-on part and in the most passive way" (Greimas & Courtés, 1982: 75).There are four more actants: (3) a sender, instigating a subject's desire for the object; (4) a receiver, receiving the object after the subject has acquired it; (5) a helper, aiding the subject with obtaining or maintaining the object; (6) an opponent, obstructing the subject in their attempts to obtain or maintain the object.Together, these actants form a narrative's actantial model.They should not be confused with a narrative's characters: in one narrative, a character may function as several actants and several characters can function as one actant.Actants are not necessarily human: e.g.animals, phenomena or mindsets can all function as actants (Greimas & Courtés, 1982;Robichaud, 2003).
Stage i of the narrative program clarifies "the incentive, desire, or imperative to act", while stage iv assigns "meaning and value" to the actions performed in stages ii and iii (Korthals Altes, 2014, p. 259n3).The actantial model, meanwhile, maps actants' significant relations to each other as they act or are acted upon (Ricoeur, 1984, p. 55).As such, inviting interviewees to use the narrative program to turn their experience into a "pattern of transformations", and the actantial model to establish a "relational system" (Robichaud, 2003, pp. 39-41), will give insight into individual "intentions and values", as well as the "sociocultural webs of meanings" within which these are embedded (Meretoja, 2018, p. 28;Jovchelevich & Bauer 2000, p. 67).This is an important use value of the data that this method yields.
We designed a protocol that allows the interviewer to 'fill in' stages i-iv and actants 1-6 while keeping the conversational and personal quality that characterizes most, if not all forms of narrative interviewing.The interview starts with an invitation to the interviewee to recount what to them is the most significant life event that they are willing to share in relation to the research topic.This event may span many years, or just be a short moment in time.An additional invitation such as 'can you take me with you back to that event and walk me through it step by step?' may help to engage the interviewee in narrative work.
As the interviewee recounts their experience without being interrupted (Wengraf, 2001;McAdams, 2007), the interviewer takes notes, using a fill-in sheet based on the narrative program and actantial model (see Figure 2).Once the interviewee has finished their initial account (Jovkolevich & Bauer 2000, pp. 63-64), the interviewer gives additional prompts in order to come to a full narrative.Since narratives may be told in haphazard, non-chronological or fragmented ways, any of these cells may remain empty until late in the interview.Sometimes, fitting content may be only implicit in the interviewee's answers, but The structured narrative interview [7] where possible the interviewer will check with the interviewee what goes into each cell, thus facilitating the co-construction of the final narrative.Note that the interviewee will often be the subject in accounts of their own life events, but this is not necessarily so.It is important to stress that this is not meant to be a list of standard questions.These are prototypical questions, to be adapted to the context of the interview, phrased as much as possible in the interviewee's own words, and fine-tuned through improvised follow-up questions.So what does this protocol look like in practice?Often, a structured narrative interview starts with the interviewee bringing up any number of experiences that could be turned into a narrative, but which are then abandoned.For example, in the transcript of an interview by Taran, from which excerpts are given in Appendix 1, this takes more than 16 minutes.Several times, the interviewer asked the interviewee questions such as to think of "one specific situation that you then kind of describe to me like a storyteller", and reassured him repeatedly that any experience, no matter how small, would suffice.A number of experiences were either dismissed because the interviewee thought they are not significant enough, because the interviewer decided they lay outside the scope of the current research, or, in one instance, because the interviewee was uncomfortable with further exploring the experience.Although sixteen minutes is exceptionally long, such 'false starts' are not uncommon.
Finally, an experience is found that both interviewer and interviewee are willing to invest in, and transform into a narrative.After one more prompt, the interviewee, quite suddenly, switches to a storytelling mode, saying, "So I was at school…" The interviewer leads him on with an encouraging "Yeah?" and then lets him recount what he remembers, facilitating him doing so by asking him, "what were you working on?What was […] the context?Where were you?All of these things".After that she helps him to further construct his narrative.At times, rather than just asking the questions as they appear in the protocol, she formulates them using the interviewee's own phrases from earlier in the interview.Sometimes she asks a straight question ("So in that situation, what was […] your goal?"), at other times she paraphrases him, or checks whether he agrees with something being, for instance, a helper or an opponent.Prompted by the interviewer to consider what enabled him to tell a friend to stop talking Dutch in the presence of foreign students, for instance, the interviewee reaches the insight that his own experience as an exchange student may have played a role.And later on, the interviewer checks whether he acted the way he did because he wanted to avoid friction, but the interviewee rejects this as an important motivator in his narrative.
Thus, the structured narrative interview asks for the ability -and sensibilityto take quick decisions about whether to be persistent, or whether to prod more gently.Narrative literacy, i.e. the ability to recognize how experiences can be turned into narratives (Romero-Ivanova et al., 2019), is also needed when deciding whether to continue to explore an experience offered by the interviewee or not.Throughout the interview, there will be continuous co-construction, starting with the decision which experience can be transformed into a narrative, and especially apparent when interviewer and interviewee decide on the different actants and their relations, and how the experience can be mapped according to the narrative stages.
The structured narrative interview [9] Pilot study

Context, aims and research set-up
The aim of the pilot study was to assess experiences with, and attitudes toward the use of other languages than the language of instruction in an international learning environment, more specifically in the form of translanguaging: "dynamic and fully integrated" flexible language practices where speakers may switch at any moment from one language or dialect to another, and back again (Mazak, 2016).This was part of a larger research project on multilingualism in university education, for which the starting point was what Cenoz & Gorter (2021) call "pedagogical translanguaging": an "integrated approach" (p. 2) to the education of multilingual students, which "is learner centered and endorses the support and development of all the languages used by learners" (p. 1).The concept of translanguaging builds on theories of cross-linguistic independence, which point to the transfer of knowledge between multilinguals' different languages, including partial competences (García and Li, 2014, p. 13), and conceptualizes speakers' competences in various languages as one linguistic repertoire, rather than relating to clearly delineated different linguistic codes (Mueller, 2018).Translanguaging can happen at various communicative levels, and with different goals, such as ideational (i.e., providing explanations, elaborations, and examples), textual (i.e., highlighting shifts between what speakers are doing, or its importance), and interpersonal (i.e., signaling changing relationships, identities, and social distance or highlighting common values or norms) (Lin, 2013).It comes with academic and social benefits, such as approachability and easier interpersonal connections (Duarte & van der Ploeg, 2019).The larger research project, steered by Moenandar and Duarte, was aimed at creating a knowledge base for more informed language policies for higher education institutions.When developing such policies, it is crucial to assess the motivations behind and attitudes toward translanguaging among students and staff to gain insight into how students negotiate experiences with translanguaging.This is why the structured narrative interview, developed by Moenandar and Basten, was deemed a suitable method for the entire project and three BA students were trained in using the structured narrative interview method.Furthermore, we wanted to assess whether there were differences between (attitudes towards) translanguaging across academic fields.This remains an underexplored issue (Ferguson, 2009), while such differences can certainly be expected: academic fields have been shaped by very different traditions (Cohen 2001).Take, for instance, the humanities and the natural sciences -often seen as two opposing points (Bod & Kursell 2015;Geertz, 1973).Traditionally, the humanities empha-size socially-constructed realities, while the natural sciences tend to focus on the physical world and on discovering general, independent truths, such as in physics (Bod & Kursell, 2015).This has led to two, fundamentally different, academic cultures (Snow 1959).While historically (Philips, 2010), humanities scholars focus on the contextualization of knowledge, scientists aim to increase the stock of knowledge, in order to come to ever more accurate descriptions of reality (Snow, 1959;Stueber, 2012).If attitudes toward language use, including translanguaging, in international classrooms are shaped by these cultural differences, this also needs to inform universities' language policies.The pilot study was a first attempt to explore these possible differences, following the main question: what are the attitudes toward translanguaging among students in higher education?As sub-questions, it aimed to find out (1) what strategies of translanguaging could be found in students' narratives about switching languages in learning environments; and (2) whether there were significant differences between the attitudes among students in a humanities faculty and a science faculty.
For the pilot study, students from a humanities faculty (n = 6), as well as from a science and engineering faculty (n = 6) at a large Dutch university with a highly international body of staff and students were interviewed about their experiences with translanguaging.All twelve interviews were conducted by the same interviewer (Taran).The interviewees were bachelor and master students who responded to an open call that was spread through formal and informal networks of both faculties.The aim was to have the same number of participants from each faculty.This yielded 514 minutes of recordings and 12 sheets (see Figure 2) with extensive notes.The native languages of the interviewees were: Dutch (6), English (1), German (1), Romanian (1), Polish (1), Azerbijani (1), and Lithuanian (1).Due to Covid-19 restrictions, the interviews were mostly conducted via the university's video platform.The average interview length was 43 minutes.The protocol described above was followed, and the interview started with the question to describe a moment in which another language than the language of instruction was used.If further prompting was needed -which was especially the case for the science students -this was followed by the encouragement to think of a particularly significant moment, whether positive or negative.

Results: Four types of translanguaging stories
As said before, using a clear narrative structure like Greimas's in the interview protocol, allows for easy comparison.Each interview was recorded and transcribed to engage in coding and ultimately identify themes.Thereby, roughly four types of narratives of translanguaging were discerned in these interviews.

Prohibiting translanguaging leads to discomfort
A first type -always presented by humanities students -is set in language classes.In this narrative, a prohibition of translanguaging creates discomfort, even if students subscribe to this ban.One interviewee (R5), for example, discusses a Spanish class where no other languages than Spanish are allowed.This leaves him tongue-tied: "I was ashamed because all of my classmates were listening".Thus, being the narrative's subject and his own sender as he wants to improve his Spanish, he also becomes his own opponent: his performance anxiety keeps him from achieving his object.This can be explained somewhat by the "unhealthy amount of value" he says he attaches "to academic performance".Beyond that, however, we can also glean how his perfectionism has been fostered by an educational system that places great value on high achievement.
Another interviewee (R1) shared a similar narrative, also set in Spanish class.However, the interviewee is much more relaxed, saying, "I can't do it perfectly as you would with a native language, it's okay to make mistakes".Although she does not cast herself as that actant, the opponent is the same as in the previous narrative: "people got embarrassed or were afraid to speak up in front of the whole class".At a certain point, they even started to speak Dutch or English.To the chagrin of the interviewee, the teacher did not correct them.Motivated by her conviction that the "point of the course is to develop better proficiency", and also because she feels speaking Dutch excludes those who do not understand that language, her object becomes to maintain classroom order.Overcoming her shyness, she called out those not speaking Spanish.She constructs a sanction stage in which her actions may have annoyed fellow students but everybody gained a better command of Spanish.
In another narrative, this is turned upside down, when translanguaging is used to restore classroom order by the subject -not the interviewee, but his lecturer -rather than disturb it.The student recounts how, during a French class, his lecturer told him several times in French -the only language allowed -to put down his phone.She then suddenly switched to Dutch and sternly said, "do I need to translate for you to pay attention?Because I need you to pay attention to me!" The shock made him realize, "I need to be serious and she's going to kick me out if I don't." Here, it is especially the exceptional nature of the translanguaging that makes it such a forceful intervention -a punishment really -as he stresses that it had never happened before, and did not happen again.It was also immediately successful: he put away his phone (R6).

Translanguaging makes things less formal
The notion that translanguaging is disruptive, and therefore something one normally ought to refrain from, also permeates the second type: stories in which translanguaging threatens to turn a formal learning environment into an informal, less effective gathering.One science student (R10, see also Appendix 1), for instance, tells of a classmate being so hungover that he switched from English to Dutch during a group project.Although told as an entertaining story, it is informed by strong convictions, such as "English is now the science language", and the interviewee looks back with satisfaction on how he strictly -and quite rudely -corrected his friend.As in most previous stories, translanguaging threatens the subject's success: he desires a good grade, wants to go home early, and for this, speaking English only is important.Otherwise, the collaboration would not go smoothly: non-Dutch groupmates may feel excluded or embarrassed to speak English, and "people who can't express themselves can't come up with good ideas" (note the paradox between the conviction that people must be able to express themselves, and excluding Dutch).Another science student (R8) constructs a similar narrative, in which he was speaking Polish to a fellow Pole while working on the same AI project.This distracted him and when he subsequently switched back to English out of politeness to others, this also restored focus, forcing everybody to go "back to working on the project again".
In this last student's narrative, returning to English was necessary, but somewhat troublesome, and he casts himself as his own opponent: speaking Polish brought "comfort" and switching to English and thereby to serious work meant he had to overcome a certain hurdle.This is phrased much stronger in a humanities student's (R3) narrative, who feels she needs to be polite, but also thinks, "I don't want to speak English now, I was already in the mood to joke around casually with my Lithuanian friends".During the interview, she chastises herself, judging her attitude as "lazy" and "selfish".In another interview, it is not the interviewee (R4), a humanities student, who needs to switch languages out of politeness or a desire for success, but her fellow students.She tells how she found herself in a classroom where everybody was speaking Dutch, which she experienced as a lack of "respect".Once again, an anxiety narrative is constructed: the student did not want to explicitly demand everybody to speak English, fearing the others would look "down upon" her.Instead, she attempted to subtly signal her not understanding Dutch.When someone finally noticed this, everybody immediately switched to English, to her relief.

Suddenly, it's okay to translanguage
In all the narratives above, translanguaging is frowned upon because it is seen as (1) impolite, and (2) being informal.This explains the plot of a third type: stories about situations in which the strict bans on translanguaging typical for formal learning environments suddenly no longer apply.One science student (R12), for instance, tells how he strictly refrains from speaking German with other Germans in class, out of politeness.He then recounts how he discovered that things are different during leisure time when he went climbing in a gym with classmates, mostly Germans.In his narrative, his object is sticking to English, but failing to achieve this actually ensured that he had a good time doing something he loves.Another science student (R11) tells how, when ordering two wheels from a woodshop for a project, he spoke English because this language was established when his non-Dutch professor introduced him to the Dutch carpenter.Midconversation, as he struggled to make clear the specifics of the wheel he needed, he realized it would be much more effective to continue in Dutch.Similarly, a humanities student (R7) tells how she went with a group of Dutch classmates to a restaurant and continued to speak the language of instruction, English.This caused difficulties for a waitress -the narrative's subject -who struggled with English.The student herself switches from being the waitress's opponent to her helper by switching from English to Dutch, and while constructing her narrative, she muses about how something that is seen as polite in the formal learning environment was impolite, and really rather odd behavior, in this informal setting.

Translanguaging can be good
Only in the fourth type -comprised of just two narratives -we find an unconditionally positive attitude toward translanguaging in formal settings.A humanities student (R9) recounts a "fun exercise" where the lecturer instructed students in an international classroom to describe something in their own language.The student constructs a collective narrative of growth (Moenandar & Huisman 2019), where she and her classmates enjoyed hearing others' native languages, and took comfort in speaking their own.In this narrative's sanction stage the group members have strengthened their mutual bond and become aware of how easy it is to understand somebody speaking a language you have no competence in.Emphatically also placing herself as a receiver in this narrative, the exercise also gave her "a more positive attitude toward hearing other languages", and made her critical of English as the default language in international classrooms.We see similar attitudes in a Serbian sciences student's (R2) narrative, although not made as explicit.He tells a witty story with himself as a subject driven by a strong desire to be successful in his field -AI studies -and thereby achieving a "better life".For this rea-son, he wants to establish contact with a "smart guy", a teaching assistant (TA) known for his knowledge of an obscure, but important aspect of AI.One day in class, he sees this TA talking with some friends in Bulgarian.Stealthily, he moves closer to their table.As he does not want them to feel intruded upon, he decides to use "humor".Employing his rudimentary knowledge of Bulgarian, he drops the Bulgarian word for potato when their conversation pauses, surprising and amusing the TA.Consequently, he gets to "pick the TA's brain".

Discussion
Judging from their narratives, students' attitudes towards translanguaging are generally negative.In international classrooms, translanguaging is typically experienced as a rupture, something that happens suddenly, mostly performed by opponents and accompanied by feelings of discomfort, to the point where it can be used as punishment.The narratives show how these attitudes are embedded within personal histories and wider educational and social contexts.It is clear, for instance, that the learning environments in which these narratives are set, do not encourage -and often actively discourage -translanguaging.Among the students, there is a clear preference for monolingualism, both for pragmatic and ethical reasons: translanguaging is seen as an obstacle to academic success and as impolite.In most narratives, translanguaging is avoided or relegated to informal settings.Although this is true for most interviewees, in this small cohort there is a difference between humanities and science students, with the latter being more pragmatic in their aversion to translanguaging: they dislike it because they see it as being less serious, but readily employ it when it can be used as a means for success.Among humanities students, a more reflective stance toward translanguaging and its social impact can be discerned, although the outcome is more or less the same.There is also a difference in how strongly students feel affected by translanguaging.Humanities students express stronger emotions when narrating their experiences, science students seem relatively more indifferent towards the phenomenon -also illustrated by them finding it harder to think of an experience.
Institutions of higher education that want to employ the advantages of translanguaging mentioned at the beginning of this pilot study, will have to take into account these negative attitudes and negotiation strategies.The motivations that drive these narratives, however, can be summarized as a desire for orderly learning environments and study success.Educational language policies designed to foster translanguaging could therefore be accompanied by raising students' awareness that these desires can be met even better by no longer keeping translanguaging out of their learning environments -as the final two narratives illustrate.It must be noted that, because of limits of space, we did not discuss how language The structured narrative interview [15] ideologies (in which certain languages are often seen as more valuable than others) may have influenced students' attitudes towards specific instances of translanguaging (cf.Moenandar et al., 2022), which also needs to be taken into account when designing and implementing language policies.

Concluding remarks
The structured narrative interview introduced in this paper is meant to be an effective means for generating narrative data.Its interview protocol, based on the transformation pattern and relational system that, according to Greimas, underlie all narrative, allows for a naturally flowing conversation which, by means of co-construction between interviewer and interviewee, leads to fully rounded narratives in which an experience is reconstructed and recontextualized.These narratives provided rich and useful data, giving insight into how interviewees negotiate the value and meaning of certain actions and the relations between actants, show how certain attitudes lead to certain behavior, and how this behavior becomes meaningful and valuable in certain contexts.This kind of data enables a more nuanced assessment of attitudes.In our pilot study, for instance, a regular semi-structured interview according to a theory-driven topic list would probably have yielded certain attitudes from students, but would not have shown this negotiation of value and meaning, thus missing opportunities to understand what shapes these attitudes, how they lead to certain linguistic behavior among students, and how this can possibly be changed.While such richer data could no doubt have also been collected with less structured methods of narrative interviewing, those would have taken more time, as argued before.To be fair, this is also a limitation, since one's (remembered) behavior does not always corroborate with one's attitudes and ideologies, and the explicit expression of the latter is less naturally facilitated by this method.For researchers interested in this (lack of ) corroboration, combining methods may be an option.
There are further limitations to the method.Firstly, the method seems less suited for collecting entire life stories, or assessing how longer or more complex episodes were experienced, as its protocol almost automatically leads to a focus on a single event.Although in a longer interview, several narratives around an equal number of experiences could be constructed, such a broader range is not naturally facilitated here.Secondly, the method only works when people have a story to tell -that is to say, people may have a significant perspective on a given research topic that could be teased out by a regular semi-structured interview, but if they do not have actual experiences with it, the method is unlikely to yield useful data.
A final aspect that deserves pointing out, is that all interviews for the broader study of which this pilot study was a part were conducted by undergraduate students who learned the method during a three hour online workshop.Despite this short training, they then managed to conduct narrative interviews without much need for further coaching.This possibility to teach the method within a few hours is another attractive quality.Furthermore, as an easily teachable, timesaving method that yields rich narrative data, the structured narrative interview could be useful in other practices as well.An early version of this method was developed by one author (Moenandar) in the context of teaching narrative writing to journalist trainees.Since then, the method has been taught in settings as diverse as business schools, methodology courses for anthropologists, and workshops for study career counsellors.As such, this method illustrates how fruitful an applied narratology that reaches beyond disciplines can be.
I: You can say whatever you want, just recording it so that if you say a specific situation.No, just like one event.Nothing major.So, for instance, it could be just.Just an example, just to kind of… R: Of course.But I…' I: It could be me, being in a supermarket and I'm asking the cashier something in English, and she replies in Dutch, even though she sees that I'm speaking in English.R: I can give examples, but I'm not sure whether I can give…anecdotes?I: Yeah, anecdotes.But we just want to kind of delve into the situation a bit more.So if there's one situation that you can remember more.So try to think of the most memorable, or most… R: Well, most memorable.I mean it's also hard because of COVID, so I've never been really at class.I: So I mean you… Can also be something that's been a while back now.So, of course.R: Yeah, we were working on it.And we were just chilling.And the dude was hungover.And, he started speaking Dutch, to me, because he wanted to tell something, and I was like, 'dude, fuck it, we'll talk about it later, not now, we need to focus and just get this shit done.And also, don't speak Dutch.Just talk English so everybody can hear!Doesn't matter anyway.' And then he was like, 'oh, sorry, I'm so hungover!'And then he spoke English.And then we worked together in a group, and we finished the model.Well, not that day, but a few days after.
[20] Sjoerd-Jeroen Moenandar et al.I: And by 'viable' , you mean?R: I mean, there were two choices: just ignore him, or tell him to change the language.Or speak Dutch to him, but then we' d exclude everyone else.

I:
Okay.All right.Do you think there was anything that helped you in preparing yourself to act or in, in actually acting?R: When I was born, after a few months, I learned to talk.Well, not after a few months.After… I don't know, when do babies talk?At seven months?

Figure 2 .
Figure 2. Fill-in sheet for the structured narrative interview based on Greimas 1991.Original design by Moenandar; further development by Lucaci (in Moenandar et al., 2022).See Appendix 2 for an example of a filled-in sheet Moenandar et al.
mean, it doesn't have to be an extraordinary situation.Can also just be something ordinary, as you said.So for instance when you're… Someone switched to Dutch, you tell him, 'please go back to English' .Any sort of situation like that also works.R:Yeah, okay.So I was at school… I: Yeah?R: And a friend of mine spoke Dutch while we're working in a group, and I… I: With whom?R: With two other people, just a group of four.I: Where were those people from?R: One guy was from India, and the other was, from, what was his face again?Oh, yeah, [name of classmate]!He was from, I think, England.An Indian immigrant from England.Yeah, maybe we can start… So what were you working on?What was, kind of, the context?Where were you?All of these things.R: [Ponderously:] We were working on [respondent gives short (five words) description of his project].I: Yeah.

I:
And, what about another way?R: I should have hit him with a baseball bat or? I: I mean, I guess?I guess that's a possible strategy, although that might be a bit more brash.R: No, I mean.No, just tell him to shut up or tell him to change the language.Other than that… I: So you acted like this because it would cause the least friction.R: No, but… I guess.I: I don't know.Just… R: No, no, but I'm not saying I had in my mind, like, 'oh, we'll create the least friction' .I was just like, 'bro!' I mean, I wasn't like, 'oh, I'm going to ignore him because, you know' .I: Yeah.R: But it wasn't like, 'oh, I need to create less friction.' No, just like, you know, this is the choice.I'm going to make it, because it's the only viable option.
Any way.Anything, really.Physical.It can be anything in your mind.In others' minds.About people and things.But still.I mean, like, what else did I need to?I mean, it would just.Mm.Maybe.Maybe what I needed was the time in [country where student went on exchange] to make sure that I understood how horrible it is for people to be excluded when they're not able to speak a different language.Because then I had to chug a fucking beer.And that was awful.And at one point, it became vodka… Never mind… But then we were just fucking punished.And after like, one or two days, we were also like, of course.So that just helped.And that ingrained it.So… I: Okay.No, that's great.Then what I also wanted to know, why did you act in that way and not in another way?So what were other possible ways for you to act in that situation, to achieve the same goal?R: Well, I could have ignored them.But why the fuck would I do that?That would have given him a bad feeling.Just speak English, that way we can still communicate.I mean, I guess ignoring is also a form of communication, but still.I mean, it still leaves… How you say it?I don't even know how you say it in Dutch, let alone… No, but I mean, it still leaves room for communication.For talking to each other.Like, if you start ignoring them… Well, fuck it.