The popularity of interviewing as a method of data collection in the social sciences is a recognized fact. In their survey of qualitative research paradigms and methods, Denzin and Lincoln (Reference Denzin and Lincoln2004:353) declare that “the interview is the favorite methodological tool of the qualitative researcher.” And, describing data-collection techniques in sociolinguistics and dialectology, Fuller (Reference Fuller2000:388) argues that “much of the data in the field comes from interviews.” These assertions are hardly surprising given the central role that interviews have assumed as an essential part of the toolkit of the qualitative researcher since the early decades of the twentieth century (Fontana & Frei Reference Fontana, Frei, Denzin and Lincoln2004). Interviews are the most common cross-disciplinary research instruments since they are widely used by investigators in fields as diverse as education, anthropology, sociology, social psychology, and social history, where they serve as vital research methods alone or in combination with other techniques such as participant observation. Given the centrality of interpretive and qualitative research paradigms in sociolinguistics, ethnography, linguistic anthropology, and narrative studies, the interview has acquired an even more prominent place for investigation in these disciplines. However, this research method and tool for collecting data has been the object of extremes of confidence and criticism. On one side there are those who try to erase the interactional context of the interview, believing that it is both possible and desirable to make participants forget about the event so that interviewers can access their “natural” behavior. On the opposite side there are those who argue that interviews are “inauthentic” and “artificial” contexts for data collection and therefore it is best to avoid them completely. In both extremes, the interview ends up being a problem to overcome. One unfortunate result of these attitudes has been that the interview as a real communicative event has been understudied. Our objective with this special issue is to contribute to redressing this tendency by drawing attention to the need for, and advantages of, the research interview as a legitimate interactional encounter, and taking narrative as our focus. In doing this we build on a small but significant cross-disciplinary body of mostly recent scholarship that has analyzed a variety of issues related to the use of semi-structured and open-ended interviews in qualitative research, and that has recognized the crucial importance of placing interview data in context.

As mentioned, our primary focus in this special issue is on narratives in interviews, a choice motivated by the following theoretical considerations. First, as a consequence of the narrative turn (see Bruner Reference Bruner, Schiffrin, Fina and Nylund2010), most research in the social sciences has revolved around the elicitation of stories, biographical accounts, and narratives of personal experience told in interview contexts. As discussed below, however, such research has been heavily content-oriented and has devoted minimal attention to the contexts in which these narratives are produced.

Second, because of their complex discursive characteristics, narratives afford the researcher access to the delicate interactional management of different communicative levels. As argued by Young (Reference Young1987), narratives always involve the discursive production and coordination of distinct domains: (a) the domain of the communicative event (in this case, the interview), (b) the domain of the storytelling world, and (c) the domain of the storyworld. Other scholars have used two-fold distinctions, such as “narrated” versus “narrating” events (see Jakobson Reference Jakobson, Waugh and Monville-Burston1957) or “denotational text” versus “interactional text” (Silverstein Reference Silverstein and Sawyer1997). In any case, participants in storytelling events align towards all of these worlds, thus generating interactional positionings and often also meta-discursive practices whose analysis demonstrates how discourse shapes and is shaped by its context (Bauman Reference Bauman1986).

With these points in mind, we see storytelling in interviews as a fertile ground for theoretical and methodological reflection that is paramount for narrative analysis, but that also goes well beyond the limits of the field to address fundamental questions about the role of the research context in the collection and analysis of data, the interaction between researcher and “subject,” and the nature of the process of interpretation. By focusing our attention on the research interview as an interactional event and on narratives in this environment, we draw on a body of scholarship that has developed in relation to different issues and within different disciplines, but that as a whole has contributed to the emergence of a reflexively oriented research paradigm that all authors in this special issue share. Thus, before discussing the antecedents and implications of such a paradigm for the field of narrative studies, we review research that has had a fundamental role in putting the interview context on the map for qualitative research. In particular, we focus on:

1. 1. Contributions to the debate over the problem of bias in social science research, particularly ideas on reflexivity developed by ethnographers such as Cicourel (Reference Cicourel1964) and Briggs (Reference Briggs1986);

2. 2. Theoretical reflections related to the “Observer's Paradox” in sociolinguistics;

3. 3. Suggestions offered by conversation analysis and ethnomethodology-oriented researchers within debates over the opposition between interview-as-topic and interview-as-resource (Rapley Reference Rapley2001).

BIAS IN SOCIAL SCIENCE RESEARCH AND REFLEXIVITY

Most social science research manuals insist on the neutrality and unobtrusiveness of the researcher as a guarantee against research bias. Within the positivist tradition, a premise of scientific investigation, and a good way of eliminating bias, is an allegiance to the principle that the researcher has to observe the world without influencing it and minimize her or his presence in the data-collection process. Thus, in the case of interviews and surveys, a great deal of effort has been devoted to develop techniques that ensure data comparability and to standardize the researcher's behavior in order to make sure that answers to questions are given under the same conditions by all respondents (see Hyman et al. Reference Hyman, Cobb, Feldman, Hart and Stember1975, Krosnick Reference Krosnick1999). Both Cicourel (Reference Cicourel1964) and Briggs (Reference Briggs1986) have offered insights on this illusion about researcher neutrality and on the inevitable contradictions that arise from the attempt to completely standardize a distinctive social event such as the interview. Both authors noted that since each interview is unique and the roles and relationships between each interviewer and each “informant” are different, trying to impose a uniform procedure is a violation of “naturalness.” Cicourel (Reference Cicourel1964:87) openly criticized the concept of comparability between data, which he deemed unattainable under the conditions of the interview as a real communicative encounter.

Twenty years later Briggs took up the same theme in focusing attention on the fact that most methodological debates on the interview do not go to the root of the problem because they do not ask what kind of interactional event the interview is. Briggs proposes the importance of examining the implicit communicative mechanisms that operate in such events. In his study, he addresses this issue in detail by focusing especially on the meta-communicative repertoires (i.e. the set of implicit shared norms that govern communication) underlying both the interview and the communicative practices of the target population—in his case, Mexicano speakers in northern New Mexico. Briggs emphasizes how meta-communicative norms between native “informants” and interviewers often clash, leading to communicative blunders, or worse, to the unilateral exercise of power by the interviewer. In both Cicourel and Briggs, there is an effort to make explicit the ideologies and presuppositions that lie behind so-called “objective” research methods and to denaturalize the practices that characterize modern scientific investigation. Another important contribution by both Cicourel and Briggs is their insistence on reflexivity, which they define as the ability of “speech, whether contained in interviews, myths, or ‘natural’ conversations,” to provide “an ongoing interpretation of its own significance” (Briggs Reference Briggs1986:106). In that sense the investigation of meta-communication strategies and principles at work in interviews is not only a tool for understanding this genre's properties as a communicative event, but also a source of insight into fundamental methodological issues in social-scientific research.

THE OBSERVER'S PARADOX

Related to the question of research bias is the issue of the so-called Observer's Paradox, first mentioned by Labov in connection with his studies on vernacular speech in New York and Philadelphia in the 1960s. Labov, who was concerned with the difficulties of obtaining natural speech in research situations, thus described the paradox:

Among the solutions devised to tackle the problem of the possible effect of the researcher's presence on the respondents' speech was the elicitation of narratives through the famous “danger of death” question, that is, a request to recount an event in which the interviewee had been close to dying. Some of the presuppositions implicit in the formulation of the Observer's Paradox and of solutions to it have been the object of critiques that are directly relevant to our discussion here. One presumption in Labov's work was that with minimal intervention of the researcher the interview could become as similar as possible to a “natural conversation.” In that sense, the interview was seen more as a context to be erased than as an interactional event whose specificity should be understood, and perhaps even exploited. Also, behind the attempt to neutralize the interviewer there was a focus on the respondent as the producer of talk that was supposed to be analyzable independently of the interaction with the interlocutor. Finally, Labov related variations in speech styles, particularly the alternation of “formal” and “informal” speech, mainly to the degree of attention and self-monitoring that subjects devote to their own speech; here, the role of the interlocutor in the unfolding interview was important only insofar as s/he helped regulate the use of these varieties. The relationship between the two interactants was not itself an object of empirical study.

Much work produced within sociolinguistics challenged these fundamental assumptions by emphasizing the lasting effects of aspects of the interlocutor's identity on the speech style of the interviewee and the emergent qualities of the speech of both interlocutors. Proponents of Accommodation Theory (Bell Reference Bell1984, Giles & Coupland Reference Giles and Coupland1991, Niedzielski & Giles Reference Niedzielski, Giles, Goebl, Nelde, Starý and Wolck1996) argued, for example, that speakers may radically change their accent and other prosodic, syntactic, and discursive aspects of their language in response to their interlocutors' identity and role and that phenomena of speech accommodation or, conversely, differentiation, are common in both naturally occurring contexts and research interviews. Indeed, sociolinguists in the last decades set out to prove that social identities of interviewers such as gender (Eisikovits Reference Eisikovits and Pauwels1987) and race (Rickford & McNair-Knox Reference Rickford, McNair-Knox, Biber and Finnegan1994, Cuckor-Avila & Bailey Reference Cukor-Avila and Bailey2001) could have a significant impact on the use of linguistic variables by respondents. This work has provided a foundation for more recent analyses of the interaction between interviewer and interviewee where the focus is not on psychological perceptions of the interlocutor or on the impact of social identities only defined in terms of portable attributes, but on how interactional moves and negotiations indexing identities are consequential for the linguistic and discursive choices of participants in an interview (see Schilling-Estes Reference Schilling-Estes2004). This shift in focus is well expressed by Giles, Coupland, and Coupland (Reference Giles, Coupland, Coupland, Giles, Coupland and Coupland1991:1), when they state that

INTERVIEW-AS-TOPIC VERSUS INTERVIEW-AS-RESOURCE

The role of the interview as a research method and source of data has been at the center of many debates among researchers who come from Conversation Analysis (CA), Ethnomethodology, or traditions close to these. Taking up a distinction first proposed by Seale (Reference Seale and Seale1998), Rapley summarizes the debate as involving an opposition between approaches that regard interview data as “resource” and approaches that regard it as “topic.” In the interview-data-as-resource approach, “the interview data collected are seen as (more or less) reflecting the interviewees' reality outside the interview,” while in interview-data-as-topic, “the interview data collected are seen as (more or less) reflecting a reality jointly constructed by the interviewee and interviewer” (Reference Rapley2001:304).

The interview-data-as-resource approach has been strongly criticized by CA-inspired researchers, who oppose the idea of generating data for the specific purpose of analyzing them. Amongst the most frequently quoted statements of such a position is Potter's assertion that research should be able to pass a (conceptual) “dead scientist test,” that is, that the data used should be based on an interaction that “would have taken place in the form that it did had the researcher not been born or if the researcher had got run over on the way to the university that morning” (Potter Reference Potter and Richardson1996:135). Similar ideas have animated the rejection of interview data by many other CA-oriented scholars (Heritage & Atkinson Reference Heritage, Atkinson, Atkinson and Heritage1984, Potter & Whetherell Reference Potter and Wetherell1995, ten Have Reference ten Have1999) in general and in connection with narratives in particular (see Goodwin Reference Goodwin1997, Schegloff Reference Schegloff1997). The basic argument against interviews as a method of data collection is that they produce “unnatural” data since the interviewer influences their production (through questions, interruptions, silences, etc.), and offers ad hoc interpretations through the use of etic (i.e. non participant generated) and not emic categories of analysis. For CA-inspired researchers, interviews are interesting not as windows into a reality beyond them, but as real interactional events in their own right. A thorough discussion of these positions would go beyond the scope of this introduction, but it may be noted that many have rejected both the idea that research interviews cannot be used to gain understandings about social phenomena beyond the local event, and the notion that “natural” and “unnatural data” can be opposed in such black and white terms. In particular, they have argued against the notion that the researcher's presence in the data can be avoided just by the investigator “not being there” (see Speer Reference Speer2002).

NARRATIVES IN INTERVIEWS

The debate over interview data has only marginally touched narrative studies. Indeed since the publication of Mishler's influential book, Research interviewing: Context and narrative, more than thirty years ago, relatively little progress appears to have been made. There are, however, as we will argue below, signs of change.

At the center of Mishler's work was “the proposition that an interview is a form of discourse” (1986:vii) and that narratives are co-constructed between the interviewer and the interviewee. He asked:

Many of these questions are still pertinent today and have not been answered in recent research. The field continues to be dominated by the study of narratives elicited in interviews and not narratives in interviews. Labov based his original model of narrative analysis (Labov & Waletzky Reference Labov, Waletzky and Helm1967, Labov Reference Labov1972) on stories told in response to his “danger of death” question and since then many of the studies focused on narratives have come from similar contexts. As noted above, one reaction to this situation has been the rejection of elicited narratives as data, particularly among conversation analysts, ethnomethodologists and interactionists in general (see e.g. Goodwin Reference Goodwin1997, Keller-Cohen & Dyer Reference Keller-Cohen and Dyer1997, Schegloff Reference Schegloff1997, Ochs & Capps Reference Ochs and Capps2001). They have argued that Labov practically erased from his model any reference to the context in which the narratives that he analyzed emerged and pointed to the risks of analyses that do not recognize the context-embeddedness of talk. They have therefore tended to favor the study of “naturally occurring narratives” as an antidote to the problems related to work with elicited data.

Those who do choose to study narratives produced in interviews have often neglected any consideration of narrative's embeddedness in the interactional context. Narrative analysis has become one of the most widespread research tools in the social sciences as a result of the “narrative turn,” but most of the social-scientific work spurred by this new interest focuses on the content of stories and regards them as unmediated expressions of the narrator's self and as accounts that offer information about the interviewees' experiences. As noted by Atkinson and Delamont (Reference Atkinson and Delamont2006), the unproblematic equation of narrative talk with self expression and of narrative content with facts and events derives from the neglect of the performance aspects of narrative and of the interview as a communicative context (see also Wortham Reference Wortham2001). Such neglect is widespread in the analysis of interview narratives, though some recent research has started to redress this tendency. Work produced in the last two decades has focused on the interaction between interviewers and interviewees in accounts and other kinds of narratives (Baker & Johnson Reference Baker and Johnson2000, Perrino Reference Perrino2005, Johnson Reference Johnson2008, De Fina Reference De Fina2009). Narrative analysts have also seriously questioned the role of researchers in the selection and interpretation of data (see Riessman Reference Riessman2008) and their co-constructive role in the production of narratives of personal experience (see Lucius-Hoene & Depperman Reference Lucius-Hoene and Depperman2000, Bell Reference Bell, Fina, Schiffrin and Bamberg2006). However, studies that analyze the interactional and contextual aspects of different types of elicited stories are still very limited in comparison with studies of conversational narratives. This lack of attention to the interview context is partly the consequence of a treatment of the interview as a somewhat unnatural event, of an excessive focus on genres such as the canonical story or the life story within interview narratives, and of a tendency to treat narratives in interview as a much more homogeneous genre than they really are.

The objective of this special issue is to put the interview context firmly on the map of narrative studies. The contributions assembled here offer fine-grained analyses of narrative data produced in interviews while addressing theoretical and methodological questions that pertain not only to narrative, but to research based on interviews in general. The questions that authors in this special issue try to answer can be summarized as follows:

1. 1. How accurate is the sharp opposition between interviews and “naturally occurring” events? Some essays interrogate this point, showing how interviews are as interactionally complex and as worthwhile studying as other communicative events. In that sense, the idea that narratives told in interviews are unperformed recapitulations of past experience is a gross oversimplification.

2. 2. What does it imply to analyze interviews as interactional events? Authors stress the need to go beyond the idea of co-construction and joint participation between interviewer and interviewee and to start engaging with finer analyses of interactional roles at different levels. At the same time, researchers should also shift their focus from the strategies enacted by interviewers to get their data (to elicit “natural” speech and have subjects forget about the interview) to ways in which interviewees negotiate their relationships with interviewers.

3. 3. What further methodological consequences derive from taking the interview as a communicative event seriously? Some authors point to the need for a reflexive examination of the process of data interpretation in order to critically assess how conclusions were reached by the researcher. They also argue for the need to look at insider/outsider status and its impact on data collection, assessing first of all how these roles emerge in given interactions and, secondly, what the advantages and disadvantages of these positions are in concrete research situations.

All of these questions are addressed and exemplified through narrative data. But there are also some reflections that apply specifically to research on narrative. Indeed, the articles show that narrative genres not only are intricately interconnected with interactional roles but also subject to continuous redefinition as interviews develop. In what follows, we provide an overview of the contributions to this special issue.

Gabriella Modan and Amy Shuman argue that most of the previous work on sociolinguistic interviews has tended to focus on how these events can hinder the researchers' goals. In contrast, they show that interviews may be useful for interviewees in meeting their own discursive goals. The focus of their analysis is a small community of stone carvers in the town of Pietrasanta, Tuscany, where Shuman conducted research. In the analysis of one of the interviews, the authors illustrate how the interviewee uses embedded orientation to position participants as ratified/unratified and insiders/outsiders, and also to build an interpretation of history that includes information not normally discussed among community members.

Anna De Fina's article compares interviewer/informant roles and their relation to storytelling in two interactions between Italian researchers and members of Italian-Australian families in Australia. De Fina shows how very different identities and role-relations are enacted and negotiated between interviewers and informants in the two cases and how such differences affect both the kinds of stories that are told and the way they are managed. Like other authors in this special issue, De Fina notes that the effect of the Observer's Paradox, that is, the highlighting of the presence of the researcher in the data collection event, is not always a problem, but may become a resource for the analyst.

In their contribution, Stanton Wortham, Katherine Mortimer, Kathy Lee, Elaine Allard, and Kimberly Daniel White argue that interviews are events in which people often reproduce social action and beliefs that are typical of their community. In that sense, for example, they may repeat stories that they tell to others in their community making the same kinds of points that they would make with interlocutors other than researchers. The case of “pay day mugging stories” told to interviewers by members of different ethnic groups of a new migration community in the Philadelphia region is exemplary because these stories get reproduced throughout research contexts. Thus, it is an interesting site to study positionings and constructions of social reality that are widespread in a community.

The focus of Stef Slembrouck's article is the importance of what he calls “analysis of data histories” for the correct understanding of research results. Data histories imply an extension of ethnographic reflexivity beyond the examination of interviewer/interviewee behavior in individual interviews to a consideration of the processes that led to the data collection and interpretation. Such an enterprise also involves looking carefully at how presuppositions and stances expressed by interviewers and interviewees affect both the interviews and the analysis of the data. The particular case studied is that of interviews with parents who produce narratives about children in foster care. Slembrouck illustrates how class differences played a key role in the way parents constructed those narratives.

In his article, Mike Baynham examines the use of breaks into performance within different kinds of narrative genres by ESOL teachers in interviews conducted by other teachers. Baynham notices how performances can be “fleeting moments” rather than sustained deliveries and how these moments not only take place in canonical narratives but in all kinds of narratives. These performance shifts within storytelling are used by teachers to index a variety of stances vis-à-vis their students, their professional experiences, and their interlocutors. Again, the reciprocal roles of researcher and subject turn out to be not only central to properly contextualize the data, but also subject to continuous adjustments and repositionings.

Michele Koven rejects the pressure to “analyze narratives either from interviews or spontaneous conversation” and instead compares the two. Focusing on French-Portuguese bilinguals, and using qualitative and quantitative methods, she compares a story that occurs in both conversational and interview contexts. She looks especially at the way narrators take on different roles when they narrate (e.g. whether they quote speech directly and adopt the voice of characters, or whether they stand back and narrate from afar). While some may expect the interview version to be less interlocutory and less “involved,” she demonstrates how both events show a good deal of involvement, only in different ways.

Sabina Perrino's article focuses on narratives taken from two sets of data: interviews conducted with Italian residents of the Veneto region on their reaction to migration in the area, and interviews with Senegalese on their relation with traditional medical practices. She argues that conventional conceptions of narrative posit a sharp separation between the spatio-temporal dimensions (chronotopes) of the story and the storytelling event, thus also contributing to an approach to analysis of narratives in which the “denotational text” of the story is abstracted from the interaction in which it takes shape. Perrino argues that coeval alignment phenomena (i.e. resources through which narrators shatter the separation between the spatiotemporal there-and-then of the story and of the here-and-now of the storytelling world) are an important site where the embedding of stories into interactional contexts can be studied.

We hope that the analyses and reflections presented in this issue will advance debates about the interview as a site for the study of narrative and provide useful directions for further research.