Corps de l’article

I. Introduction

Imagine that you are an experienced community interpreter, working primarily in the field of healthcare. You are asked to interpret a medical appointment, and you have worked with both the physician and the patient before. The physician runs a busy family practice. He works on a fee-for-service model, and he pays hefty overhead and insurance. He is under intense pressure to move patients through his office quickly, in order to cover his costs. To further complicate matters, the physician sees a large number of patients who are HIV positive, and who have complex medical needs. Most times, these patients require more time from the physician than he can realistically give. He works long hours and weekends to keep up with the demand.

The patient is a relative newcomer to Canada. She moved here just under a year ago with her husband and her children. Like many women in her cultural community, she stayed at home to look after her children, and she has had little contact with broader Canadian society. She speaks almost no English. Her husband worked outside the home and took responsibility for communication with the community at large. Unfortunately, six months after arriving in Canada, the patient’s husband became seriously ill (he told his wife he had “leukemia”), and he passed away three weeks prior to the medical appointment you are interpreting. The patient was understandably devastated. A short time after her husband’s death, the patient had a meeting with another healthcare practitioner-the details are unclear-but the patient was told “horrible things” about her husband that she could not believe. That her husband had been having sex with men. That he had not died of leukemia, but of AIDS. During the meeting, a blood sample was taken, and an appointment was made for the patient to see the busy family doctor. She arrives for the appointment struggling to understand all that has happened to her. She is bewildered and afraid.

You are sitting in an examination room with the patient when the physician enters suddenly and brusquely. He sits down at his desk and shuffles some papers. He picks up a print out, turns to the patient, and says, “Yup, you’re positive.”

Both the physician and the patient turn to you for the interpretation. What do you do?

II. The Conduit Model

This scenario was presented by a colleague of mine during a volunteer training session.

Everyone in the session was a community interpreter. That is to say, we all had experience interpreting dialogues between service providers and clients in a variety of community settings, such as government offices and social service agencies. That day, we were focusing our attention on one specific setting within community interpreting – healthcare – because a particular healthcare agency in the city where we lived was recruiting interpreters, and the agency required us to attend a short training course before it would add our names to its roster. As my colleague told her story to the group, I could see a number of nodding heads and looks of recognition around the room, and this reaction suggested to me that most of the other interpreters had been witness to similar medical encounters. The experience of the bewildered female patient seemed to be a common one.

In the discussion that followed the anecdote, my colleague eventually shared the interpreting strategy she had used in the examination room that day. With the eyes of the patient and physician weighing heavily upon her, she quickly considered her options. If she conveyed the physician’s words exactly in the patient’s language, my colleague was afraid that patient would enter into a state of shock and would fail to take in the important and complex information about treatment options that the physician would no doubt give to her. Consequently, my colleague decided to alter the physician’s word choice, telling the patient that “the tests are positive”. In so doing, my colleague hoped to reduce the patient’s shock of learning about her seroconversion. My colleague felt strongly that she had not altered the meaning of the doctor’s statement-implication in context made it very clear that the tests in questions were the patient’s-but that she had merely given the patient “a little personal distance” in which to receive the news.

When she finished explaining her choice of strategy, my colleague was reprimanded by the trainer leading the session. Her interpretation, he argued, was simply not faithful. He maintained that the interpreter should never do anything other than repeat, in the other language, everything said that is said, exactly as it is said. To do otherwise, he continued, is not ethical. As a rule of thumb, he suggested that interpreters should routinely ask themselves, “What would happen if the patient were comfortable speaking English?” If the patient my colleague met had been a native English speaker, she would not have needed an interpreter, and there would not have been anyone in the room to cushion the blow of the physician’s words. Instead, she would have born the brunt of the doctor’s insensitivity alone. The trainer maintained that the actual medical encounter should have been the same. The interpreter should have refrained from intervening, leaving the patient to deal with the impact of the physician’s statement. Only by remaining a non-participant in the encounter could the interpreter do what was right.

These comments, and others made throughout the session, made the trainer’s position clear: the role of the interpreter is to act as a simple conduit, transferring information from practitioner to patient and back again with the utmost fidelity. And although I have singled out the trainer in my telling of the anecdote, he is not alone in adopting this position. The so-called “conduit model” of interpreting is very widely discussed in the field of community interpreting, although the terminology used to discuss it varies from author to author. For instance, in the key literature on community interpreting, interpreters are interchangeably categorized as “invisible” (Angelelli, 2004) or “neutral” (Metzger, 1999), as “message converters” (CHIA, 2002), “translation machines” (Bot, 2005), or “direct linguistic translators” (Kaufert & Koolage, 1984). The words may change, but the description of desired behaviour remains the same. Its hallmarks are features like “accuracy” and “completeness” (see CCHCP, 2005; CHIA, 2002), and it restricts interpreters from making additions, omissions, or changes/distortions (see CCHCP, 2005; Abraham, Cabral & Tancredi, 2004). The conduit portrays interpreting as an exercise carried out on linguistic forms, one in which even the smallest changes in perspective (e.g., exchanging “you” and “the tests”) are not permitted. As noted in the literature, the conduit has at times been called the traditional perception in interpreting (Metzger, 1999, p. 1), its central perspective (Dysart-Gale, 2005, p. 92), and even its ideal (Angelelli, 2004, p. 2).

The characterization of the conduit as an ideal is an interesting one. It hints at an association between models of interpreter behaviour and perceptions of moral correctness. In other words, proponents of the conduit promote it not because it is based on a strong foundation of empirical evidence, but rather because they see it as the right thing to do. For example, in establishing his rule of thumb for the training group in the scenario above, the trainer could have pointed to quantitative information to support his position. He might have argued that X percent of interpreters in the field follow his rule, or that Y healthcare practitioners surveyed preferred interpreters to give faithful translations. But instead, his strategy was to take the moral high ground, by arguing that remaining faithful to the language forms uttered by patient and practitioner allows the interpreter to act ethically.

III. Origins of the Conduit

The trainer’s recent discussion of the conduit, together with its treatment in the literature, suggest that it is currently a topic of importance in community interpreting. However, despite all the attention paid to it, descriptions of the model’s origin and of the rationale for using it are rare. As a result, there are two questions that need to be answered, if we are to understand the model more fully. Where does the conduit model come from, and how did it come to be applied to community interpreting?

The second question is perhaps the easier of the two to answer, and there are a couple of explanations to be found in the literature. The first, offered by Metzger (1999, p. 8), suggests that community interpreting has adopted the conduit model because of the example set by conference interpreting. In the conference hall, interpreters have limited opportunity to interact directly with the people for whom they interpret, and the speeches they work with are long, formal, text-like monologues (Wadensjö, 1998). When we view interpreting from this “monological” perspective, we tend to focus on meaning as it is created by the SL producer’s intention, and through the producer’s use of the tools of language (lexicon, syntax, etc.) (Wadensjö, 1998). The tendency to view conference interpreting in this way has likely also been emphasized by the information-processing models that have been used to map out the interpreting process. These models describe the process in mechanistic, machine-like terms (Metzger, 1999, p. 8), and they suggest that meaning is largely derived from language[2]. When we try to draw conclusions about interpreting as a whole based on our observations of conference interpreting and conference interpreting research, we are led to view discourse as emanating largely from a single speaker, the TL receiver as essentially passive, and the interpreter as someone with no “personal” involvement in the discourse event (Roy, 1993, p. 149)[3].

The second explanation focuses on the history of a subset of community interpreters, specifically those who work with signed languages. Over the course of their history, sign language interpreters have worked under a number of models (Witter Merithew, 1986), and the most influential of these is arguably the conduit. To explain how this came to be, scholars note that in the days before there were professional sign language interpreters, Deaf people relied on “helpers” to communicate with the hearing world. These helpers were frequently hearing friends and family who had some knowledge of both the signed and spoken language in question (Roy, 1993, p. 139; Metzger, 1999, p. 22). Helpers were free to act as they saw fit. Many offered advice, made decisions for the Deaf person, shared confidential information with authorities if they thought it was in the Deaf person’s best interest, and selected and edited the information they interpreted according to their perception of the Deaf person’s understanding. The helpers’ behaviour underscored an attitude that the Deaf were incapable of making decisions and taking care of themselves, and this attitude was sometimes internalized by the Deaf themselves, with obvious negative repercussions (Roy, 1993, pp. 139-140; Bar-Tzur, 1999, ¶ 2). However, as interpreting began to be recognized as a profession, there was a sharp move away from the helper model. Two key events are usually cited as the hallmarks of this transition: the founding in 1964 of the Registry of Interpreters for the Deaf, the principal professional organization representing sign language interpreters in the US, and the publication in 1965 of the Registry’s first manual for interpreters, Interpreting for Deaf People (Bar-Tzur, 1999, ¶ 3; Roy, 1993, p. 140). These two events signalled the advent of a new level of professionalization, one that was incompatible with the inequality underscoring the notion of the helper. The relationship between interpreter and client had to be a relationship between equals (Bar-Tzur, 1999, ¶ 2), and there were calls to reject the emotional and personal involvement of the helper model, and to strive instead to be neutral, invisible, and uninvolved. The interpreter was frequently described metaphorically as an inanimate device or a machine (Roy, 1993, p. 141), such as a “telephone wire that served as a conduit for information flow” (Bar-Tzur, 1999, ¶ 4). The profession as a whole has moved on to consider other models of professional behaviour, but it has been argued that many sign language interpreters still subscribe to the underlying notion that they are a channel through which messages are transmitted, and thus passive participants in the interpreted encounter (Roy, 1993, p 148). It is possible that this notion has spread widely across community interpreting as a whole, that is, to those community interpreters who work with spoken languages.

These two explanations-which point to conference interpreting and sign language interpreting respectively-provide an account of how the conduit came to be applied in community interpreting. However, they have not offered a satisfying description of the model’s ultimate origin, and, as a result, the first question asked above has gone unanswered. Yet there are a number of helpful clues in the literature that shed light on the problem. For example, Dysart-Gale (2005, p. 98) links the conduit model in interpreting to early research in the broader field of communication. She points to a set of older communication theories, which have viewed the act of communication essentially as the movement of information from one party to another. These theories advocate enhancing communication by making optimal use of technology and techniques. She refers to this set of theories as the “technical family”, and she notes that one of its “members” is still the dominant perspective in American scholarship[4]. What is interesting in Dysart-Gale’s account is the parallel between her description of the technical family and the hallmarks of the conduit model in interpreting. Most notably, the theories seem to suggest that making improvements is an exercise carried out on the technical stuff of communication, and they leave out a consideration of wider contextual factors. So too does the conduit model focus on the linguistic fidelity and the actual technical make-up of the interpreted utterance, neglecting any larger concerns. If Dysart-Gale is correct in drawing the link between interpreting and communication, the conduit may simply be a reflection of an older theoretical perspective and its continued influence.

Another interesting clue can be found in Reddy’s (1979) discussion of the metaphors used to describe communication. He examines colloquial speech and takes note of statements that people use when faced with a communication barrier. A receiver who has not understood a producer’s utterance might instruct her interlocutor to “try to get your thoughts across better”, or note that “you still haven’t given me any idea of what you mean.” Reddy argues that these kinds of statements imply that thoughts and ideas are communicated directly, essentially given by the producer to the receiver. In popular perception, communication is a straightforward affair, where mental processes are simply handed back and forth between participants, without any sort of intermediate representation. Reddy explores this perception further. Admonitions like “put your ideas into words carefully” and claims such as “your words are hollow-you don’t mean them” provide a clearer picture of our beliefs about the transfer of ideas, for they imply that words are a sort of empty container. Producers fill words with meaning before handing them over to receivers. Under this system, poor communication is the fault of the producer, who has neglected to choose the proper containers, and clearing up miscommunication is a simple matter of making better word choices. Reddy’s discussion is enlightening, because it links the conduit to common perception. In other words, our desire to have the interpreter act as a conduit may simply be the result of the way laypeople think about communication generally. They see it as direct and uncomplicated, and tend to view interpreting in the same manner. The discussion also provides a potential warning for interpreters: they should be ready to be faulted for miscommunication by both of the other parties in an interpreted encounter. After all, interpreters spend their time choosing “containers” for the thoughts of both parties, and under the logic of the conduit, they are twice as likely to make “poor choices” and be the “cause” of misunderstandings.

IV. Four Ethical Approaches

Returning now to the rule of thumb cited above, the trainer maintained that a faithful interpreter is an ethical interpreter. In order to behave in a morally correct manner, the interpreter must remain passive, faithfully relaying the linguistic forms uttered by patient and practitioner into the other’s language. This connection between ethical conduct and fidelity is an interesting one, because it suggests that there is only one justifiable way for interpreters to behave. Yet we must ask whether this view is accurate and informed. Is fidelity really the only pathway to ethical behaviour? As a pathway, does it present a complete understanding of the interpreting process? If interpreters adopt fidelity as an ethical guide, do they, for instance, have the full benefit of current knowledge in the field of Translation Studies? In order to better understand fidelity as an ethical principle, I found it helpful to lay it against the backdrop of the evolution of thinking in the discipline.

To do this, I turned to an outline of the different ethical approaches that have characterized Translation Studies. The outline was provided by Pym (2001) and based on Chesterman (2001). In it, Pym maps out four separate approaches.

  1. Representation

    Under the ethics of representation, the primary objective of the translator is to faithfully represent the source text. The key term in this approach is faithfulness or fidelity. To demonstrate fidelity, the translator is frequently instructed not to add, omit, or change anything in the text. (For this reason, words like accuracy and equivalence also figure prominently in discussions of representation.) A lack of fidelity-demonstrated by a feature present in the source text but not in the target (or vice versa)-is a sign that the translator has been unethical. Initially, the understanding of fidelity was discussed in terms of language forms (words, syntax, etc.). Later it was expanded to include first notions of textuality (as described by Meschonnic, 1973) and then considerations of the potential of the source text and its future translation (a nod here to Benjamin, 1923).

  2. Service

    The translator’s objective in this approach is to provide loyal service to the client. The key term here is loyalty, which is assessed primarily by determining whether the translator has met the requirements as outlined by the client. However, translators also need to take into consideration other communication partners, such as the ST author and the TT readership. A translator who does this is ethically correct. Clearly the ethics of service is inspired in large part by the proponents of the Skopostheorie (see Vermeer, 1989).

  3. Communication

    In this approach, the translator’s main goal is focus on the Other, not for the purposes of representing that otherness per se, but rather in order to better understand it. Developing this understanding is key, and it requires the translator to shift attention away from texts and onto the people that use them. Perhaps not surprisingly, Pym suggests that the ethics of communication are often discussed in the realm of community interpreting. This is doubtless because the impact of the act of translation on actual people can be seen in real time. It seems logical to conclude that ethical behaviour in this setting must surely rest on the interpreter’s awareness of the participants who act in it, and of their differing and often unequal access to power.

  4. Respect for Norms

    The final ethical approach advocates following the accepted practices of a given time and place. It suggests that ethical conduct cannot be determined first in isolation and later applied to real-life settings, for this course of action inevitably brings the translator into conflict. Instead, the approach instructs the translator to follow norms and behave in predictable ways. This kind of conduct is likely to earn the translator the trust of others, and trust is the hallmark of this approach. The respect for norms is influenced by the work of Toury (1978) and by that of other proponents of descriptive translation studies.

When we reflect upon these four ethical approaches, and when we consider where the conduit model should be placed among them, the answer seems clear. The model is based on an ethics of representation. To be sure, it bears all the hallmarks: an emphasis on fidelity, accuracy, and equivalence; and an injunction against additions, omissions, and changes. What is more, the conduit seems to apply the ethics of representation in only a limited way. It restricts itself to language forms (hence the critique of my colleague’s decision to change “you” to “the tests”), leaving to the side considerations of textuality or potential. In other words, proponents of the model do concern themselves with the actual technical make-up of the SL and TL utterances, but they do not generally concern themselves with remaining faithful to the textuality of the interpreted conversation as a whole, or to potential forms of that conversation.

Any doubts about the conduit model’s connection with representation can quickly be cleared up by juxtaposing the conduit against the backdrop of the other approaches. To begin, the model is at odds with the ethics of service. I noted above that the conduit is often described in terms of the interpreter’s neutrality (see Metzger, 1999). Indeed, several codes of ethics build on this notion, stating that the interpreter should remain impartial (CCHCP, 2005) and show no preference or bias towards the other participants in the interpreted encounter (Abraham, Cabral & Tancredi, 2004). Yet the interpreter cannot remain steadfastly neutral and still display the loyalty to the client required by the service approach. Similarly, the conduit is inconsistent with the ethics of communication. The trainer’s rule of thumb instructs the interpreter to consider what would happen to the patient if she were a native speaker of English, and it argues that the patient should be left to fend for herself the way a native speaker would. Yet a native speaker would likely have the benefits of education and familiarity with the healthcare system that many immigrant women do not. To pretend that the patient has the same access to power as a native speaker is to fail to understand her otherness. This failure is at odds with the communication approach. Finally, the conduit is also not compatible with the ethics derived from a respect for norms. To see that this is so, we need only consider an example cited by Pym (2001, p. 135), that of an interpreter who passes German Red Cross workers off as members of the “German Red Crescent”, in order to facilitate their access to victims in a Turkish earthquake. From a norm-based perspective, this behaviour is ethical, because it meets the expectations of the people in the situation. Turkish locals understand the function of the German workers in a manner that is meaningful to them, and the German workers gain access to the people who need their attention. However, the interpreter has by no stretch of the imagination provided a faithful repetition of anything the German workers have said. From the perspective of the conduit, the interpreter’s actions are unacceptable.

V. The Persistent Conduit

At this point, it seems fair to say that the conduit model presents a number of drawbacks. First, contrary to what the interpreter trainer in the opening scenario claimed, the conduit is by no means the only way for interpreters to ensure that their professional behaviour is ethical. As the four approaches outlined above make clear, interpreters can choose a number of alternative courses of action and still be able to justify themselves from an ethical standpoint. Indeed, Pym points out that the ethics of communication are often raised in connection with community interpreting, suggesting that the more appropriate ethical concern in the community may not be fidelity, but rather understanding. When my colleague interpreted for the busy family physician and the immigrant woman, she was acutely aware of the difference in power between the two. The physician was an educated man in a prestigious profession who enjoyed social recognition. The woman was a widowed mother with no language skills and few contacts outside her home. The physician knew a great deal about HIV disease. The woman was struggling to make sense of what had happened to her. When she interpreted between the two, my colleague used her understanding of Otherness to map out a course of action that was at least as ethically justifiable as the fidelity imposed by the conduit.

Second, the conduit as a model is not representative of the full breadth of scientific knowledge that might inform interpreted communication. From the perspective of Translation Studies, the conduit is linked to the ethics of representation, and it is typical of the thinking that characterized early studies of translation and interpreting. It focuses on the actual concrete make-up of the SL utterance and its TL interpretation, and pushes to the side any of the larger considerations that have marked translation scholarship since that time. As a result, the conduit model has little in common with current research on translation or interpreting, which tends to look beyond units of translation to examine how different people use translated communication for specific purposes, in specific times and places (see Pym, 2001, p. 137). From the perspective of communication, Dysart-Gale (2005) points out that the conduit is related to early theories that focus on techniques and technology, and that it is disconnected from more recent scholarship, which examines larger contextual factors. From the perspective of metaphor, Reddy’s (1979) work suggests that the conduit may be nothing more than the result of the lay public’s uninformed beliefs about the nature of language in communication.

There are also two other drawbacks inherent in the conduit model that have not yet been raised in this discussion. Chief among these is the fact that the conduit does not accurately describe actual interpreter behaviour. Put simply, interpreters–even the best trained ones–are not passive or neutral non-participants in the communication process. This has been shown repeatedly in some of the seminal writings in community interpreting, specifically those that describe observational studies of community interpreters. For example, Kaufert and Koolage’s (1984) 18-month study of Cree- and Saulteau-speaking interpreters in two Canadian hospitals found that the interpreters took on a number of different roles in any given assignment–from “direct linguistic translator” to cultural informant to patient advocate. Wadensjö’s (1998) observations of 20 Swedish-Russian interpreted encounters in both court-like and community settings showed that interpreters were actively engaged in both translation and co-ordinating activities. Similarly, Metzger (1999) observed eight hours of medical interviews that were interpreted between English and ASL, and she concludes that interpreters’ supposed neutrality is a myth. In her study, interpreters exerted a very real influence on the interactions between patient and practitioner. In another study of ASL-English interpreting, Roy (2000) conducts an in-depth analysis of a videotaped meeting between a student and a university professor, and concludes that the interpreter she observed showed evidence of linguistic competence and communicative competence. In other words, the interpreter took on responsibility for mediating between the participants’ differing approaches to structuring conversation. In addition, Angelelli’s (2004) 22-month study of Spanish-English interpreting in a California hospital involved intense data collection: the author gathered documentation, interviewed 11 interpreter informants, had 14 interpreter informants complete a survey, and audio-recorded 392 interpreted encounters. Her data revealed that interpreters frequently move beyond the confines of the invisible conduit–for instance by controlling the flow of traffic, by exercising power over another participant, or by replacing a participant–thereby influencing communication in consequential ways. Finally, Bot’s (2005) analysis of three therapist-interpreter-patient triads suggests that interpreters who attempt to follow the dictates of the conduit model may be more likely to produce divergent renditions of patients’ or providers’ utterances, to increase the likelihood of communication breakdown, and to complicate providers’ attempts to repair communication breakdown (a technique she refers to as “recycling”). Viewed together, these different studies raise two important points. First, the conduit is not an adequate model for predicting or describing what interpreters actually do. Each of the studies cited here has shown that the interpreter is actively engaged in operations that go beyond simply matching SL forms with corresponding TL ones. Second, the conduit is not an effective training tool, because it is not realistic to expect interpreters to follow its dictates in actual interpreted encounters. Indeed, if Bot’s study is any indication, the conduit may actually lead interpreters into difficulty. Interpreters who only “repeat exactly what is said” will not enable anyone to communicate, because true communication requires interpreters to attend to features that are beyond the linguistic.

The final drawback that needs to be raised in this discussion is that the conduit model is simplistic. It fails to incorporate more up-to-date information about the complexity of interpreting, and it consequently does not provide a comprehensive account. To be certain, part of the interpreter’s task is to work with language forms, and so the conduit does portray a portion of the interpreting process adequately. But it needs to combine this portrayal with other features of the process to form a more sophisticated model. There have been attempts in the literature to provide such a model. For instance, Solomon (1997) takes a step in this direction when she states that the aim of an interpreter working in healthcare should not be to achieve neutrality, but to build shared meaning between patient and practitioner. To illustrate her point, she describes cases where a practitioner’s notion of individual decision-making conflicts with values of filial piety held by patients’ families. This situation leads to an ethical impasse, because Western doctors are trained to be ethical by giving all available information to the patient and allowing the patient to make decisions about care, but many families consider it their duty to protect their members from bad news. To avoid these kinds of dilemmas, Solomon suggests that practitioners sit down with their patients at the onset of treatment, before conflict arises, and ask about the degree of truth-telling the patient desires. Does the patient want to receive information from the practitioner alone? Does the patient want to receive information in the presence of family members? Or does the patient want the practitioner to transmit information directly to family members? Solomon advises practitioners to rely on interpreters to help them conduct these kinds of discussions with their patients, but she notes that to derive the full benefit from the interpreters’ expertise, practitioners have to permit them to move beyond simple repetition in the other language and allow interpreters to fully interpret. Of course, giving interpreters more latitude does increase risk. For instance, an interpreter may want to delete a practitioner’s request if it violates a cultural taboo, but the practitioner may consider the information gained from the request critical for treatment. Solomon therefore argues that interpreters should institute a “transparency rule”: whenever making a faithful translation is problematic, interpreters should explain this to the practitioner and identify the information they would like to add, delete, or change. The rule does what the conduit model does not–it recognizes the complexity of the interpreter’s role, and it gives the interpreter credit for having the expertise to navigate the role effectively. It also provides a safeguard for the practitioner, and it encourages more active communication between the practitioner and the interpreter.

To summarize then, there are a number of factors that have an impact on the usefulness of the conduit model. First, there are approaches other than the conduit that interpreters can use to justify their professional behaviour as ethical. Second, theorizing in the fields of translation and communication suggest that interpreting is more complex than the conduit will allow. Third, observational evidence indicates that interpreters do not restrict themselves to the narrow role defined by the conduit. Finally, there are tools, in the form of existing scholarship, that can be used to help create more sensitive models of interpreter conduct. In the light of these factors, why does the conduit continue to be promoted as a viable and necessary model for the role of the interpreter?

To answer this question, it may be useful to consider the views of those who are involved first hand. Recently, I participated in a series of qualitative interviews that were part of a larger project entitled Health Care Interpreter Services: Strengthening Access to Primary Care. One of the goals of the project was to better understand the ways in which healthcare services are currently being provided to people in Canada who have limited knowledge of English or French. To generate this understanding, I travelled with a small group of colleagues to meet with important stakeholders in healthcare interpreting. In total, we spoke with over 150 different informants in a number of urban centres in Canada. The individuals we interviewed were either healthcare practitioners, like physicians, nurses, and social workers, or people involved in providing interpreting, such as managers of interpreting services, interpreter trainers, and interpreters themselves. Although the overall number of informants was somewhat large, our intent was not to assemble a representative sample from which we could draw inferences, but rather to develop an in-depth understanding of our informants’ experiences.

To conduct our in-depth interviews, my colleagues and I used an emergent approach. If we had established a list of questions beforehand based on our perceptions, this “questionnaire” would have led us back to a more quantitative research methodology, something that would have been both invalid in this instance and counter to our goals. To avoid this pitfall, we simply asked a basic question to launch discussions with our informants (“What can you tell me about your clinical work with patients who don’t speak English or French?” or “What can you tell me about the interpreting you’ve done in primary healthcare?”), and then we allowed our informants to steer the conversations to topics that were important for them. This approach enabled us to gain a deeper understanding of our informants’ experience of interpreting in healthcare. We analyzed each interview in a manner inspired by Grounded Theory (Glaser, 1992), extracting a number of themes that effectively represented our informants’ experience and then assembling the themes in such a way as to offer an overall account of the research situation. Space constraints mean that it is not possible here to give a thorough description of that overall account, but we can focus on some of the more pertinent details. In the course of our interviews with managers, trainers, and interpreters, three themes emerged that might help to explain the persistent conduit.

The first of these is the low status of community interpreters. Most of the interpreters interviewed complained that the value of their work was not generally recognized. Some of the interpreters could recall instances when they had been in a hospital or clinic and heard a page over the public address system asking for anyone who spoke a particular language to report to a certain department. The interpreters told us that many of the healthcare practitioners they worked with seemed to believe that anyone with knowledge of two languages could do the job, even those with little general education or specific training. Indeed, one of the interpreters told us of one occasion where she was asked to clean out a closet after she was finished speaking to a patient. The practitioner who made the request apparently thought she was some sort of “bilingual janitor”. The stories shared by our informants dovetail nicely with Reddy’s (1979) discussion of the conduit as a metaphor. People with no specialized understanding of general communication tend to view it as a simple activity, devoid of complexity, and a similar thing appears to take place with interpreting. People with no specialized understanding of interpreting view it as such a simple affair that anyone can do it, even those with little general education and no specific training. When those people happen to be the practitioner and patient in an interpreted encounter, interpreters are placed in a difficult position. Their co-participants in the encounter may possess very little understanding of interpreting, but they do have very strong expectations about it. “All the interpreter has to do,” they seem to think, “is to take the other person’s thoughts and put them into the words of my language.” Faced with this kind of misperception, it is difficult for the interpreter to move beyond it. In other words, interpreters become fenced into the conduit model, because uninformed people expect a conduit-like performance from them.

The second theme raised by our informants is the state of community interpreting as a profession. Many of the interpreters we met spoke about their inability to earn a living. They cited low hourly wages, high incidental costs (parking, travel, etc.), sporadic hours, and unreliable clients (i.e., clients who cancelled bookings). We spoke with the manager of an interpreting service who noted that one or two of the interpreters who worked for her earned around $30,000 per year. The rest earned substantially less, and often had to take outside work to make ends meet. These kinds of financial difficulties mean that there is very little incentive for interpreters to undergo extensive training. It does not make sense, for example, for an interpreter to pay tuition fees and attend postsecondary level training, if at the end of that training the interpreter will still have difficulty earning a subsistence wage. And without postsecondary training, interpreters will have little opportunity to study the intricacies of interpreted communication in detail, and to adequately understand alternative, more complex models of professional conduct.

The third theme that helps explain the continued influence of the conduit is the issue of bringing practitioners on board. Some of the interpreters and managers of interpreting services that we spoke to felt that practitioners–physicians in particular–were reticent to use the services of an interpreter. They cited instances where practitioners believed they could “muddle through” with a patient who spoke another language, where practitioners dismissed the contribution an interpreter could make, and where practitioners expressed outright hostility towards the interpreter. Faced with situations like these, the interpreters and managers felt they had to do their best to get the practitioners on board and allow interpreters into the examination room. However, the interpreters and managers felt that they could only be successful in convincing reluctant practitioners if the interpreter was as unobtrusive as possible, giving the practitioner full control over the dialogue with the patient. Some of our informants saw a move away from the conduit model as taking control or responsibility away from the practitioner. Other informants assumed that an interpreter who was no longer following the conduit model would necessarily act as a kind of patient advocate, and would therefore enter into a more adversarial relationship with the practitioner. These developments, they argued, would make practitioners even less willing to collaborate with an interpreter, with the result that fewer interpreters would find their way into medical encounters.

VI. Relationship Building

The managers’ and interpreters’ concerns about practitioner collaboration caught our attention, and as we entered into our interviews with healthcare practitioners, we were careful to make note of the practitioners’ thoughts on the role of the interpreter. To be certain, we did encounter practitioners who wanted faithful, conduit-like interpreting. For example, one informant insisted on being given a “word-for-word” translation. Another felt he had to give his interpreters the following instructions: “interpret everything exactly as I say it–don’t change one iota”. A third told us that “accurate translation” was the most important thing for him. When speaking about the role of the interpreter, these informants expressed their expectations in terms that are recognized hallmarks of the conduit model.

At the same time, we also met practitioners who had different expectations of the interpreter. A number of practitioners explained that they worked with other members of a treatment team–a group of professionals with expertise in different areas of healthcare–and that they expected interpreters to function as members of that team. At the end of a medical consultation, these practitioners made it a habit to seek input from the interpreter, just as they might seek input from other professionals. Another practitioner told us of a relationship she has with a particular interpreter who regularly provides important advice. For example, if the practitioner is working with the interpreter to give a patient treatment information, the interpreter might lean over to the practitioner and note, “the patient isn’t getting it.” The practitioner then knows she must try and give her explanation in a different way. Still other practitioners, notably those who work in mental health, rely on interpreters to help them understand the cultural aspects of pragmatic communication. For instance, one informant told us of a time when an interpreter made reference to a patient’s manner of speaking and told him, “when our people talk like that, it’s usually a sign that they are depressed.” The informant told us that he would not have picked up on this clue on his own. More importantly, when he followed up on the interpreter’s suggestion with a subsequent line of questioning, he discovered that depression was indeed an issue for the patient. Another practitioner told us that important information in a psychiatric assessment often comes from the way a patient manages a conversation. If a patient does not structure question-answer pairs or take turns in a manner that is consistent with the pragmatic rules of conversation, then this may be a first sign of delusion. But because pragmatic rules are language- and culture-specific, the practitioner may not always know there is a problem. He therefore relies on the interpreter to signal to him when patients’ responses “don’t make sense.”

When we looked at our interviews with healthcare practitioners more globally, a confusing picture began to emerge. We seemed to be getting different responses from different informants, and these responses were often contradictory. “Train interpreters to just repeat what I say and what the patient says,” one informant might tell us, while another might indicate that she is “counting on the interpreter to help her understand the patient better.” It was difficult to understand these differences, as they did not appear to be related to any easily identifiable variable, such as the informant’s age, background, or place of professional practice. We were presented with an interesting dilemma. On the one hand, we were faced with a model of professional behaviour that did not adequately describe what interpreters do in their work, and that was consequently not a good tool for guiding interpreter conduct. On the other, there are the healthcare practitioners–a group whose collaboration is critical if interpreters are to work from an alternate model of behaviour–who as a group appear to be very inconsistent in their expectations.

Towards the end of our interviews, we met with one practitioner who was able to help us make sense of what we had heard from our other informants. At an early point in the interview, the practitioner spoke about the role of the interpreter in conduit-like terms, stating that he expected “direct translation”, and that he wanted to know “exactly what the patient had said.” However, at a later point in the interview, the practitioner noted that he was happy leaving the interpreter to give some kinds of information to the patient, such as what will happen when the patient is referred to a specialist for secondary healthcare. When we brought this apparent contradiction to our informant’s attention, he explained it by describing the way in which working relationships develop between healthcare professionals.

When two professionals meet for the first time in a job setting, there is no guarantee that they will work together well. But over time, the two have the opportunity to observe one another’s professional behaviour as they work side by side. Eventually, if they have seen evidence of the other’s clinical competence, they will begin to develop a trusting relationship (for more on competence and trust, see Hallas, Butz & Gitterman, 2004). Our informant suggested that collaboration between a healthcare practitioner and an interpreter would likely follow the same path. More specifically, he broke the development of a working relationship into three phases.

  1. Control over the Situation

    When practitioners do not have direct access to the patient, they need to have great confidence in the interpreter who is the intermediary. If they are working with the interpreter for the first time, they will not have this confidence. As a result, they will need to direct the interpreter, and to have more control over the situation.

  2. Additional Information

    As the working relationship between the interpreter and the practitioner grows, the practitioner will come to rely on the interpreter to “make sure he’s not falling down”. If, for example, the practitioner unknowingly prescribes inappropriate foods to someone who has religious dietary restrictions, or if the practitioner tries to begin a treatment on a cultural holiday, the interpreter should step in to point this out to the practitioner. In many instances, patients will not do this for themselves.

  3. Assistance from a Team Member

    As the relationship between the interpreter and the practitioner continues to grow, the practitioner will come to place greater trust in the interpreter. As a result, the practitioner may simply ask the interpreter to have a particular conversation with the patient. “Just point out to the patient that she’s getting involved in A, B, and C,” the practitioner might say. It was of this third level of relationship that our informant was thinking when he suggested the interpreters help patients make transitions to other types of care.

When we compared these three phases to the rest of our interview data, they seemed to have a lot of explanatory power. For example, most of the practitioners who spoke about direct translation, word-for-word interpreting, and exact repetition were people who either admitted that they had little or no experience working with an interpreter, or who did not work in a team environment. They consequently had never had the opportunity to build up a working relationship with an interpreter, nor were they in the habit of making collaborative decisions with other healthcare professionals. Conversely, the informants who had spoken about treatment teams or relying on interpreters’ input were people who made regular use of interpreters or who worked in a team environment. As a result, they had had the opportunity to develop trust in a relationship with an interpreter, or they were familiar with relationship building through their decision-making work with other colleagues.

In fact, as our interviews drew to a close, we encountered other informants who echoed our practitioner’s thoughts about relationship-building. For example, one of our last interviewees provided helpful confirmation when he noted that, for him, the determining factor in working with an interpreter is the interpreter’s familiarity with medical practice. If the interpreter is uneducated and unfamiliar with what is going on, our informant will be much more demanding, and will want to know that what he is saying is being translated verbatim. On the other hand, if he has the sense that the interpreter knows where he is going with a particular line of questioning, that the interpreter understands his reasoning, he will be much more comfortable. He noted that interpreters may think they understand what “pain”, “fever”, and “vomiting” mean, but they need to know what they mean to the practitioner (and to the line of further questioning in the history taking) to be most effective. Clearly, the interpreter develops this knowledge by working alongside the practitioner, seeing clinical competence in action, and developing a professional relationship.

Perhaps the most interesting point to note about our informant’s three phases is that they are linked to the ethical principles outlined earlier in this article. In the first phase, practitioners are looking for fidelity. Until a working relationship is established with the interpreter, they want faithful representations of what the patient has said to them, and of what they say to the patient. In the second phase, practitioners are looking for understanding. They want the interpreter to provide them with the insight they need to help the patient reach a better outcome. The patient may not feel empowered to provide practitioners with information about cultural variables like diet and holidays, and so practitioners count on the interpreter to help them navigate this Otherness. In the final phase, practitioners are willing to work from a position of trust. The interpreter has demonstrated an ability to abide by the norms of the medical setting. For example, the interpreter may show an understanding of the goal of history taking, or of the importance of displaying competence when working with team members. Viewed in this light, the establishment of a relationship between practitioner and interpreter is not only a move towards greater acceptance, it is also a progression from the ethics of representation, through the ethics of communication, to the ethics of respect for norms.

VII. Conclusion

What our informants’ perspectives on relationship building suggest is that practitioners may be willing to give interpreters the latitude to go beyond the simple conduit model and take on a wider role, under certain circumstances. They also suggest that interpreters may have to engage in several different roles as their relationships with practitioners grow, but that each stage of growth can be guided by recognized ethical principles, such as fidelity, understanding, and trust.

However, we must consider these suggestions cautiously, because they come from a qualitative study where the main goal was to develop a deeper understanding of the realities experienced by our informants. Before we can draw any wider inferences from the three-phase outline, it needs to be put to the test empirically. Future work in this area might consult with a representative sample of healthcare practitioners, in order to discover if practitioners in the sample are generally in agreement with the three-phase outline. With this confirmation, we could then move confidently towards implementing an alternative model of interpreter conduct, knowing that we would likely have the support of healthcare practitioners.

But, in the meantime, we are able to make a number of observations about the way models of professional behaviour are currently being used in community interpreting. The conduit model has its place. It draws the interpreter’s attention to the linguistic make-up of the SL and TL utterances, and it is appropriate for those circumstances where fidelity must be the guiding principle. However, the conduit also has significant limitations that cannot be overlooked. It misrepresents community interpreting as a simple task, masking what is in actuality a complex activity. It makes claims about its exclusive relationship with moral correctness, when other approaches are also ethically viable. The conduit is not consistent with important research in the fields of communication and interpreting, research that points to the importance of larger, contextual factors and to their effect on the interpretive act. Finally, the conduit is often promoted with the insistence that healthcare practitioners will not have it any other way, yet there are indications that practitioners will collaborate with interpreters under other models. What the bulk of this evidence suggests is that we need to loosen the conduit model’s unrelenting grip on fidelity as the only acceptable ethical principle, and recognize that it can be complemented by justifiable alternatives. There may be a linguistic component in the role that interpreters play, but that role is inherently a social one. The ethical principles we present to interpreters must reflect this reality.”