Interpreters/Translators and Cross-Language Research: Reflexivity and Border Crossings

Introduction

There is now a growing volume of research being conducted in English language societies with people who speak little or no English. Very few researchers working in this field, however, spell out in their methodology the effects of having to employ interpreters and translators in research projects. There is little written about interpeters and translators involvement in research interviews and even less written about language difference in focus group research with people who do not speak English (see Esposito, 2001, for an exception). There is some work on the role of interpreters and translators in relation to best practice and models of provision — for example, the relative merits of professional services, bilingual providers, community interpreters, interpreter advocates, or informal interpreters. In this vein, Thomson, Rogers, Honey, and King (1999) review bilingual support services in the North of England and provide a thorough analysis of the relative merits of employing bilingual workers in different roles. There is also a body of literature aimed at English speaking health and social welfare professionals giving them advice on how to work with interpreters (see Freed, 1988; Fuller & Toon, 1988; Karseras & Hopkins, 1987; and also Edwards, 1995, 1998, for discussion).

Beyond issues of ‘value free’ technique (for example, Jentsch, 1998), however, there is very little reflection on the implications for qualitative research of language difference and the use of third parties in communication across languages (see Edwards, 1995, 1998; Temple, 1997; Birbili, 2000). This is a strange omission given that qualitative approaches are steeped in a tradition that acknowledges the importance of reflexivity and context. Interpreters and translators seem to be excluded from such calls to investigate different perspectives. There are many useful lessons that can be learned from the translation literature, which suggests that there is no one correct translation and that the translator is like Aladdin in the enchanted vaults: spoiled for choice (Bassnet, 1994). Rather than there being an exact match, word for word, in different languages, the translator is faced with a dazzling array of possible word combinations that could be used to convey meaning. Issues concerning translation and interpretation are not always the same, not least because the latter usually involves face-to-face interaction rather than working with written texts. Nevertheless, in this article we are drawing on some of the lessons that can be learned from a reflexive examination of the literature on issues in translation in order to explore working with interpreters/translators in qualitative research. We use the term interpretation to include both written and oral communications across languages as we focus here on similarities between them. We begin with a description of our conceptual base and then go on to explore issues around identity and knowledge. As illustrative empirical material, we refer to two research projects that we have conducted and in which we worked with interpreters/translators. We consider the way in which we separately came to the same conclusion: that to conduct meaningful research with people who speak little or no English, English speaking researchers need to talk to the interpreters and translators they are working with about their perspectives on the issues being discussed. Finally, we draw out the advantages of our perspective on working with interpreters.

Conceptual Base

There is now a large and complex literature on different perspectives in relation to research generally and qualitative research specifically (see, for example, Hammersley, 1995; Stanley and Wise, 1983, 1993; Harding, 1987). We will only engage in this debate to illustrate the implications of adopting a perspective that acknowledges differences in the way the social world is seen. Theorists and researchers have elaborated a range of ways of understanding people as social actors, including interpretative or social constructionist views (see, for example, the influential work by Berger & Luckmann, 1991). Researchers who see the social world in these terms do not subscribe to the view that there is only one correct way in which to describe it. They argue that the researcher and the research participant are both producers of accounts. Their social location in the world influences how they come to experience and describe it. People have particular histories and occupy social positions, which means that they do not see the world from another's standpoint — although they may understand each other across difference through dialogue (Young, 1997).

The strengths of qualitative research lie in its attempt to carry out this dialogue, and to record and reconcile complexity, detail, and context. A critical appraisal of this can be the integration of reflexivity — the ability of researchers to take stock of their actions and their role in the research process, and to interrogate systematically research relations (see Hertz, 1997; Steier, 1991). Even within qualitative frameworks that are open to the influences of social location, however, it appears that qualitative researchers have not discussed the influence of translators and interpreters on their research where they have been part of their work. Much of the translation literature points to the impossibility of a literal movement of meaning from one language to another (see contributions to Hantrais & Mangen, 1999; Wilson & Revauger, 2001) and would therefore sit comfortably within social constructionist views of the world. If there is no one meaning to be gleaned from experiences of the social world, then there can be no one translation and it may be necessary to convey meaning using words that were not spoken by research participants. The theoretically informed, rather than technically focused, literature on translation demonstrates that communication across languages involves more than just a literal transfer of information (Bhabha, 1994; Simon, 1996; Spivak, 1992; Temple, 1997). Simon (1996) shows that the translator is involved in discussing concepts rather than just words, and that context is all important in deciding equivalence or difference in meaning. It is not a case of finding the meaning of a text from a culture. Simon describes the problems with such an approach:

The difficulty with such statements is that they seem to presume a unified cultural field which the term inhibits; the translator must simply track down the precise location of the term within it and then investigate the corresponding cultural field for corresponding realities. What this image does not convey is the very difficulty of determining “cultural meaning”. This meaning is not located within the culture itself but in the process of negotiation which is part of its continual reactivation. The solutions to many of the translator's dilemmas are not to be found in dictionaries, but rather in an understanding of the way language is tied to local realities, to literary forms and to changing identities. Translators must constantly make decisions about the cultural meanings which language carries, and evaluate the degree to which the two different worlds they inhibit are “the same.” These are not technical difficulties, they are not the domain of specialists in obscure or quaint vocabularies. They demand the exercise of a wide range of intelligences. In fact the process of meaning transfer has less to do with finding the cultural inscription of a term than in reconstructing its value (1996, pp. 137–138).

Language is an important part of conceptualisation, incorporating values and beliefs, not just a tool or technical label for conveying concepts. It carries accumulated and particular cultural, social, and political meanings that cannot simply be read off through the process of translation, and organises and prepares the experience of its speakers. It speaks of a particular social reality that may not necessarily have a conceptual equivalence in the language into which it is to be translated (Bassnet, 1994). Language can define difference and commonality, exclude or include others; it is not a neutral medium. The same words can mean different things in different cultures and the words we choose matter. For example, Temple (1999) argues that ‘the Polish diaspora’ in England has been defined in research to date in a way that excludes people who do not belong to formal organisations led almost exclusively by the male intelligentsia.

Overing (1987) argues that we should not be over-anxious about this loss of ability to translate words literally but we should be concerned about the scope that then opens up for the use of perspectives that are alien to the people who actually used the words. She argues that “It is not the ‘word’ about which we should be anxious, we should be concerned, instead, about an ‘alien’ framework of thought which is based upon an ‘alien’ set of universal principles about the world” (1987, p. 76). Applying our own set of views about the world to other people who may hold alternative beliefs sets up an over-arching and supreme framework for understanding. Language is the medium for promoting claims to a dominant and correct perspective. The interaction between languages is part of the establishment and maintenance of hierarchical relations (Corson, 1990; Kalantzis, Cope, & Slade, 1989) with English often, usually by default, used as the yardstick for meaning in societies where it is the language of state and public participation.

The translator, and by extending these arguments, the interpreter, is pivotal to the final research product. Without talking to people who are communicating directly to others for us, how can we even begin to know if we are imposing our framework of understanding? Some researchers have begun to look at ways in which we could investigate perspectives in cross-language research (Edwards, 1998; Temple, 1997). The use of a particular language or form of language can be an important element of identity, and aspects of identity, such as gender, ethnicity, religion, sexuality, as well as moral status, are constructed and ascribed in the process of using language. ‘Speaking for’ others is a political issue (Alcoff, 1991; Back & Solomos, 1993; Wilkinson & Kitzinger, 1996). Spivak establishes links between language and identity in a way that does not fix or privilege either when she states that “Language is not everything. It is only a vital clue to where the self loses its boundaries. The ways in which rhetoric or figuration disrupt logic themselves point to the possibility of random contingency, beside language, around language” (1992, p. 178).

The concept of boundaries or ‘borders’ provides a useful tool for understanding the multiple and simultaneous positionings around ‘us’ and ‘them‘, and various authors have used the concept as a way of discussing belonging and otherness that is situated in the politics of location (Anzaldua, 1987; Simon, 1996; Brah, 1996; Temple, 1999). Brah (1996), for example, defines borders as simultaneously physical and metaphors for “the psychological, sexual, spiritual, cultural, class and racialised boundaries” (1996, p. 198). She builds on feminist debates on the politics of location, and argues against a universal essentialist position at a border and for “multiple semiotic spaces at diasporic borders, and the probability of certain forms of consciousness emerging [that] are subject to the play of political power and psychic investments in the maintenance or erosion of the status quo” (1996, p. 208).

Researchers, research participants, interpreters and translators will all present constructions of their own identity borders during interactions. Significantly for this paper, interpreters/translators are also involved in producing identity borders for those whose words they work with. For those who need to use language intermediaries, this raises questions around how they construct, and understand, others' ascriptions of their identity borders in specific circumstances. In this article we discuss the advantages of including translators and interpreters as active in research, rather than as existing in the background and treated as irrelevant other than as transmitters of messages, and the ways in which we have put this approach into practice.

Our Research Bases

In order to explore some of the issues raised by recognising that interpreters or translators involved in research are constructing and producing identity borders for themselves and others, we draw on two projects — one conducted by each of us. Here, and throughout this article, when we write about our particular research projects and intellectual biographies, we will first introduce which one of us is concerned (Bogusia Temple: BT, or Rosalind Edwards: RE) and then continue in the first person. The focus here is on our reflexive consideration of working with interpreters in the research process, so we do not address other aspects of the studies (interested readers may follow this up through the references provided).

The first ‘project’ started life as research for BT's PhD thesis (Temple, 1993) and has continued for over ten years. It is unfunded and grew out of an interest in White ethnicity and Polish communities in particular. It has been the focal point of my intellectual journey over the length of the research. I have now carried out indepth interviews with over 40 people who sometimes describe themselves as Polish. Some people have been interviewed two or three times, with interviews across ‘families’ (see Temple, 2001, for an example of this work). The interviews have covered a range of issues, including identity, ethnicity, gender and family. At one point, I decided to employ an interpreter/translator, as cross-language interviews are very time consuming to transcribe. The transcription involved both interpretation and translation: listening to the taped interview and then providing a written English version. The transcription of the interview threw into sharp relief the differences between our perspectives on what a woman interviewee speaking on tape was saying.

The second project concerned an indepth evaluation of services aiming to facilitate homeless families' use of child health services and other facilities, carried out by RE (Edwards, 1993). As well as indigenous and English speaking families, a substantial proportion of homeless families are immigrants or refugees who speak little or no English. Of the 20 people I interviewed for the study, 9 could not communicate fully in English and I could not speak their languages: 4 mothers from Bangladesh, 2 from Somalia, 2 from Eritrea and 1 from Nepal. I thus worked with 3 interpreters in order to carry out repeated semi-structured interviews with these women. Two of the interpreters were paid staff members, and the other an unpaid volunteer, from local agency projects working with the relevant minority ethnic groups. This method of accessing interviewee's accounts of their lives made me very aware of the way in which my own, my interpreters' and my interviewees' identities were invested in and constructed by the particular research process.

We conducted these projects separately, and were not aware of each other's work at the time of first thinking about the issues involved in working with interpreters. When we did eventually make contact and discuss our experiences, however, we found that we had both come to question the dominant model of working with interpreters and developed an alternative. This was because both of us were steeped in feminist understandings of reflexive research practice.

Intellectual Journeys

The implicit model behind the consensus of advice about working with interpreters is a version of the traditional, supposedly detached and free from criteria imposed by subjective values. Presentations in this vein write the researcher out of the research, ignoring their role in collecting data (or seeking to minimise it at least) in an attempt to eliminate ‘bias’ and maintain ‘objectivity’. The aim is for interviews to take place ‘through’, rather than with, the interpreter. The interpreter is posed as a neutral mouthpiece, faithfully and passively translating back and forth between languages: “The interpreter is a conduit linking the interviewer with the interviewee and ideally is a neutral party who should not add or subtract from what the primary parties communicate to each other” (Freed, 1988, p. 316).

This traditional, value-free representation of the research process has been criticised by those following what has been termed the ‘reflexive turn’ in social research (for example, Atkinson, 1990; Holstein & Gubrium, 1995; Stanley & Wise, 1983, 1993; Van Maanen, 1988). While methodological reflexivity is not necessarily uniquely feminist, our own concerns in addressing issues around interpreters in social research were informed by this perspective.

Both our approaches were rooted in feminist and other critiques of textbook ‘recipes’ for interviewing as a one-way process in which the interviewer attempts to disappear, and the accompanying questioning of the exercise of power and control in research (for example, DeVault, 1990; Hammersley & Atkinson, 1989: Holstein & Gubrium, 1995; Oakley, 1981). Rather, such critiques regard social involvement on the part of the interviewer as integral to the qualitative research process. They argue that researchers should acknowledge that they are part of the social world that they study and part of its production through research accounts. Moreover, it is possible for researchers to seek to implement more democratic, reciprocal, non-hierarchical and co-operative processes while still acknowledging responsibility. Thus, in this view, in order to be as rigorous as possible, researchers need to reflect on the ways in which they, as individuals with social identities and particular perspectives, have an impact on the interpersonal relations of fieldwork. They need to place these perspectives and relations in their wider historical, political, economic, social and intellectual context, and consider the consequences for the production of research accounts. For example, Edwards (1990) has explored the ways that institutional and structurally-based divisions between researcher and researched on the basis of ‘race’ infuse themselves into the research process and the interview situation in particular.

This intellectual, methodological, critique and perspective led RE to feel that she should work towards giving the interpreters who she was working with personhood and visibility in her research and to consider the impact they had on the research process and the data collected from interviewees. BT came to the same conclusion through realising that, although she had accepted the influence of researcher and research participant on her research, she had not initially extended the same consideration to the influence of the interpreter. Only when she felt the transcript she received was not what she had heard in the interview did she come to realise that she had done what many cross-language researchers do: assumed the interpreter was not a part of the production of the research account. Like researchers, interpreters bring their own assumptions and concerns to the interview and the research process. The research thus becomes subject to ‘triple subjectivity’ (the interactions between research participant, researcher and interpreter), and this needs to be made explicit. Rigorous reflexivity in research where researchers are working with interpreters requires an exploration of the social location of the interpreter.

The reflexive model RE developed for working with interpreters treated them as a form of ‘key informant‘. Key informants have been written about and utilised in qualitative research in varying ways, notably with researchers relying on professional and/or lay informants to provide a source of introduction to, information and discussion on the social world under investigation (see Lofland & Lofland, 1995; Bulmer, 1984; Whyte, 1955). RE's use of the term ‘key informant’ encompassed a reflexive evaluation of the interpreters' social location, their values and beliefs, and their understanding of their relationship to the researcher and the interviewee. This involved interviewing the interpreters I worked with, asking them about aspects of their own life experiences, their relationship to the ethnic groups they worked with, and what issues they regarded as important in relation to the topics being addressed in the interviews and the subject of the research project. This did not mean that interpreters became, as a consequence, experts whose insights were necessarily privileged over those of the interviewees (or indeed the researcher). Rather, the key informant interviews were part of the process of making interpreters visible, and to some extent ‘accountable‘, in the same way as researchers may seek to be explicit about their own social and political position, and seek to explore this for research participants as part of the study they are undertaking.

BT has used Stanley's (1990) concept of ‘intellectual autobiography’ as a way to describe and analyse her research. Stanley argues that research accounts are products of both the autobiographies of researchers and the biographies of research participants. She describes intellectual autobiography as “an analytic (not just descriptive) concern with the specifics of how we come to understand what we do, by locating acts of understanding in an explication of the grounded contexts these are located in and arise from” (1990, p. 62). Stanley argues that the labour process involved in producing accounts of lives is often ignored, and that both the subject of an account and the writer work at constructing texts. The perspectives of both are woven into this labour process. This concept is similar to that of key informant used by RE in that it acknowledges the active participation of all involved in research.

Border Crossings — Illustrations From Our Research

Difference, perspective and identity are implicitly acknowledged in the literature on working with interpreters, but as a technical rather than reflexive matter. The interpreter-interviewee match in terms of social characteristics is regarded as an important factor, with several writers providing a hierarchy of suitability (Freed, 1988; Fuller & Toon, 1988; Karseras & Hopkins, 1987). Particular stress is laid on interpreter and interviewee being of the same sex, but culture, religion and age are also seen as important within the hierarchy of suitability. Within these characteristics, trained interpreters or professionals are generally felt to be most suitable. This matching of characteristics is with the aim of producing ‘accurate’ and ‘truthful’ data — both on the part of the interviewee to the interpreter, and on the part of the interpreter to the researcher. A more reflexive approach reveals the narrowness and implausibility of assumptions that communication and interpretation are necessarily and unvaryingly ‘better’ on the basis of social characteristic correspondence.

Concluding Discussion

Though coming from different research experiences, we have both come to see the concept of ‘borders’ as a useful way of approaching the complex question of identity, perspective and who can represent others when translating or interpreting. Differences of ‘race‘/ethnicity, gender, class and so on are gathered around borders, and the concept allows us to acknowledge the cultural space in which ‘difference’ becomes the point at which identity and knowledge constructions and contentions surface and shift around language. Difference is experienced and lived as central to identity and ideas. The interpretation or translation of the research interview is revealed as a site of interface between different identity and knowledge claims. There is no simple or distinct separation between ‘us’ and ‘them’ — negotiations and disputes occur to manage contradictions of identity and knowledge. There is a multiplicity of borders and border-crossings.

Simon (1996) discusses the concept of borders in her work and describes translation and writing as “forms of ‘border writing’ in the ‘contact zone’. This is the place where cultures, previously separated, come together and establish ongoing relations” (p. 161). As we argue above, debate on perspective is crucial in cross-language research. The examples above, however, also demonstrate that there is no single axis of similarity or difference on which to make decisions about who is representing whose views in interpreting words. The influence of particular social characteristics and points of view may be important, or not, according to context — as is the case in all qualitative research. As Simon asserts, “rather than reconfirming the borders which separate nations, cultures, languages or subjectivities, translation shows them to be blurred. It is the very economy of translation as a system of regulating differences which has become problematic” (1996, p. 165).

This problematising of translation, identity and border writing has useful lessons for researchers working with interpreters. Spivak (quoted in Simon, 1996, p. 142) considers “the task of the feminist translator is to consider language as a clue to the workings of gendered agency”. Identity is produced and not merely described in language. Gender, ethnicity, and other social divisions are important aspects of both identity and language. How interpreters produce borders between cultures and identities in qualitative research should become the focus of the kind of scrutiny that such issues have received in translation studies. Without this, we cannot justify claims that qualitative research can shed light on different perspectives since we may have shut out one person who could enrich and challenge our understandings.

In this article, we have attempted to show that applying some of the concepts developed in translation studies and the politics of location to research with interpreters has many advantages. First, it extends debates over perspective in research to interpreters. As argued above, to exclude them from debates on reflexivity and context is inconsistent. Second, when interpreters are included in such debates, valuable insights into the politics of location and identity can be reaped. Language can be a significant barrier to research with people who are not like the researcher in various ways. To assume that there is no problem in interpreting concepts across languages is to assume that there is only one baseline, and that is the researcher's own. Third, trying to make the interpreter as well as the researcher visible highlights the tensions in asking anyone to represent other people's views. That the researcher is positioned along multiple axes of belonging and not belonging during research, and that these border locations are not fixed, has been well recognised in qualitative research (see, for example, Song & Parker, 1995). Extending these insights to the choice of interpreter highlights the problematic nature of existing debates limited to the technical role of the interpreter and their language proficiency. Finally, using existing developments in cross-language research with translators provides a way of linking social context to individual situations, and situating debate of concepts, without assuming any view is linked unproblematically to social location. Much cross-language research is an attempt at border writing, and the identities and definitions of concepts used to cross borders and discuss difference and commonality form part of, and shape, the final research product.