Interpretation, translation and intercultural communication in refugee status determination procedures in the UK and France

This article explores the interplay between language and intercultural communication within refugee status determination procedures in the UK and France, using material taken from ethnographic research that involved a combination of participant observation, semi-structured interviews and documentary analysis in both countries over a two-year period (2007–2009). It is concerned, in particular, to examine the role played by interpreters in facilitating intercultural communication between asylum applicants and the different administrative and legal actors responsible for assessing or defending their claims. The first section provides an overview of refugee status determination procedures in the UK and France, introducing the main administrative and legal contexts of the asylum process within which interpreters operate in the two countries. The second section compares the organisation of interpreting services, codes of conduct for interpreters and institutional expectations about the nature of interpreters' activity on the part of the relevant UK and French authorities. The third section then explores some of the practical dilemmas for interpreters and barriers to communication that exist in refugee status determination procedures in the two countries. The article concludes by emphasising the complex and active nature of the interpreter's role in UK and French refugee status determination procedures.


Introduction
Article 1A(2) of the 1951 United Nations Convention Relating to the Status of Refugees defines a 'refugee' as someone who has a 'well-founded fear of being persecuted for reasons of race, religion, nationality, membership of a particular social group or political opinion'. The administrative and legal procedures adopted by signatories to the convention for determining whether an individual satisfies this definition characteristically involve complex processes of cultural and linguistic translation and interpretation. In the following article, we explore the interplay between language and intercultural communication within refugee status determination procedures in the UK and France, comparing the ways in which linguistic and intercultural issues are addressed, in theory and in practice, in the two countries. We will be concerned, in particular, to examine the role played by interpreters in facilitating intercultural communication between, on the one hand, asylum applicants and, on the other, the different administrative and legal actors (e.g. civil servants, rapporteurs, lawyers and judges) responsible for assessing or defending their claims.
The United Nations High Commissioner for Refugees (UNHCR) Handbook recommends that '[t]he applicant should be given the necessary facilities, including the services of a competent interpreter, for submitting his [sic] case to the authorities concerned' (UNHCR 1992) [para. 192 (iv)]. Similarly, para. 13 of the Introduction to the European Council Directive 2005/85/EC of 1 December 2005 (the 'Procedures Directive') states that: the procedure in which an application for asylum is examined should normally provide an applicant at least with (…) access to the services of an interpreter for submitting his/her case if interviewed by the authorities, (…) and the right to be informed of his/her legal position at decisive moments in the course of the procedure, in a language he/she can reasonably be supposed to understand.
More specifically, Article 13.3(b) of the Procedures Directive asserts, with regard to the personal interview conducted with an asylum applicant, that it is incumbent on Member States to: select an interpreter who is able to ensure appropriate communication between the applicant and the person who conducts the interview. The communication need not necessarily take place in the language preferred by the applicant for asylum if there is another language which he/she may reasonably be supposed to understand and in which he/she is able to communicate. 1 Among the issues raised by the wording of the UNHCR recommendation, and by the Procedures Directive in both its original and recast forms, are what constitutes 'appropriate communication', how is the competence of an interpreter to be defined and assessed and how is the applicant's understanding of and ability to communicate in a language other than their preferred one to be defined and established. 2 This article investigates how such questions have been approached within refugee status determination procedures in the UK and France, using material taken from ethnographic research that involved a combination of participant observation, semi-structured interviews and documentary analysis in both countries over a two-year period (2007)(2008)(2009).
The first section of the article provides an overview of refugee status determination procedures in the UK and France, introducing the main administrative and legal contexts of the asylum process within which interpreters operate in the two countries. The second section compares the organisation of interpreting services, codes of conduct for interpreters and institutional expectations about the nature of interpreters' activity (particularly with regard to the extent to which this is defined as 'translation' or 'interpretation') on the part of the relevant UK and French authorities. The third section then explores some of the practical dilemmas for interpreters and barriers to communication that exist in refugee status determination procedures in the two countries. The article concludes by emphasising the interplay of language issues and intercultural communication and the complex nature of the interpreter's role in UK and French refugee status determination procedures.

Refugee status determination processes in the UK and France
In the UK, asylum claims are administered by a branch of the Home Office: the UK Border Agency (UKBA). As soon as UKBA receives a claim, it conducts a screening interview with the applicant in order to establish their identity and collect basic personal information. A far more detailed interview with a UKBA case-owner takes place a few weeks later. This second interview usually lasts for several hours and follows a rigid question and answer format. The case-owner focuses here on establishing the basic chronology of the applicant's narrative and on testing its internal credibility. At both interviews interpreters hired by UKBA are present. 3 After the interview, the case-owner must decide whether the asylum claim should be granted or refused. In the former case, the applicant is granted refugee status or another form of international protection, and notified of this decision without any specific reasons being given. More frequently, however, the claim is refused and the case-owner writes a Reasons for Refusal Letter (RFRL) explaining and justifying this decision. Most RFRLs claim that the applicant's story lacks credibility as a result of alleged inconsistencies in their answers or on the grounds that aspects of their narrative are inconsistent with the cited Country of Origin Information (COI) or because their story is deemed inherently unlikely.
An appeal to an Immigration Judge (IJ) in the First-tier Tribunal (Immigration and Asylum Chamber) is possible following most refusals of asylum by UKBA. 4 A solicitor assembles the documents for the appeal, including the appellant's own witness statement as taken down by the solicitor with the aid of their own interpreter, and any relevant COI, but (in England though generally not in Scotland) the advocate who actually represents the appellant in court is normally a barrister (counsel). UKBA is generally represented in court by a Home Office Presenting Officer (HOPO), who is a civil servant rather than a professional lawyer.
An appeal hearing begins with the appellant's 'examination-in-chief'. 5 Usually very short, this involves an appellant being asked by their counsel simply to confirm that the contents of their asylum interview transcript and witness statement are true (clarifying any points that are claimed to be inaccurate) and that they wish to submit these documents as evidence. The HOPO then cross-examines the appellant. Typically lasting one or two hours, this is the longest part of the hearing. If the barrister wishes, they may then re-examine their client, and very occasionally other witnesses may be called to corroborate the appellant's story or give expert evidence.
The final part of the hearing begins with closing submissions to the IJ by the HOPO, who argues that the refusal of asylum should be upheld. The HOPO's submissions generally involve attacks on the credibility of the appellant's narrative, but 'objective evidence' in the form of COI about the situation in the appellant's country of origin is also cited, with the claim that it supports the UKBA's original decision. The appellant's barrister then attempts to rebut the credibility points and offers rival interpretations of the COI. After the hearing, the IJ produces a written determination announcing the decision and indicating how much weight has been given to each piece of evidence, including COI. 6 In France, asylum applications are examined in the first instance by the French Office for the Protection of Refugees and Stateless Persons [Office français de protection des réfugiés et apatrides (OFPRA)]. 7 Established in 1952, the OFPRA is a public institution endowed with legal personality as well as financial and administrative autonomy. It currently works under the authority (tutelle) of the French Ministry of the Interior (Ministère de l'intérieur).
After being received by the OFPRA, an asylum application is assigned to a caseworker (officier de protection) who, in most cases, subsequently interviews the applicant. Asylum interviews can vary considerably in length, but usually last between an hour and an hour and a half. They tend to be divided into two parts: in the first, the caseworker seeks to establish the applicant's identity and to collect other basic personal information; in the second, the focus is on the applicant's narrative and reasons for applying for asylum. With non-francophone applicants, either an interpreter is provided by the OFPRA or the caseworker conducts the interview directly in a language other than French. After the interview, the caseworker forwards a proposal to accept or reject the application to the head of their section (or division), the person responsible for signing the final decision. The applicant is then sent a letter informing them of the outcome of their application.
Appeals against the OFPRA's decisions can be made to an administrative court: the National Asylum Court [Cour nationale du droit d'asile (CNDA), known until 1 January 2008 as the Commission des recours des réfugiés (CRR)/Refugee Appeals Board]. Appellants are entitled to be assisted by a barrister (conseil) and an interpreter when their case comes before the CNDA. In advance of the actual hearing, a CNDA official (known as a rapporteur) writes a report that concludes with an opinion (avis) as to whether the appeal should be accepted or rejected, based on the current state of the case file.
At the CNDA, most appeals are heard by panels of three judges. In any one morning or afternoon session, a panel of judges may hear up to 13 different appeals. How long the examination of an individual case lasts varies, depending on its complexity and a number of other factors. However, the Cimade (Comité inter-mouvements auprès des evacuées), a French association that provides legal advice and other support to asylum seekers and refugees, observed 203 cases at the CNDA over a three-month period in 2009 and found that the average time taken to hear an individual appealincluding the report, which the rapporteur reads out at the start, and the barrister's statementwas 33 minutes (Cimade, 2010, p. 47). After the hearing, the three judges discuss all the appeals that have just come before them, deciding in each case whether to annul OFPRA's original decision (and therefore grant refugee status or subsidiary protection) or to reject the appellant's appeal against this decision. A letter is subsequently sent to the appellant, informing them of the outcome of their appeal.
Interpreting services, codes of conduct and institutional expectations A distinction is usually drawn between 'interpretation', i.e. 'the oral transfer of meaning between languages', and 'translation', the 'process of transferring meaning from a written text in one language to a written text in another' (Colin & Morris, 1996, p. 16). As indicated in the previous section, refugee status determination procedures in the UK and France usually require both the translation of written documents and the interpretation of oral exchanges in asylum interviews and appeals hearings. However, only the second of these forms of intercultural communication will be examined in the rest of this article. It is important to note before doing so that legal professionals in both countries also refer to 'interpretation', in the context of their own activities, as the process of determining the 'true meaning' of a legal text or document (Colin & Morris, 1996, p. 16;Cornu, 2007, p. 510). As Colin and Morris (1996, p. 17) explain, in their study of interpreters in the legal process in England and Wales, this frequently leads lawyers to insist in court that interpreters 'translate' rather than 'interpret' a speaker's utterances, by which they mean provide a literal or verbatim 'translation' (as opposed to an 'interpretation', in the legal sense). Colin and Morris (1996, p. 17) point out, however, that 'word-for-word or literal translation often produces distorted communication', due partly to the fact that words depend for their meaning on how they are combined with other words within a given utterance; an understanding of this context is required for accurate translation and interpretation to be possible. Against this background, the remainder of the present section will describe and compare the provision of interpreting services and the expectations surrounding the interpreter's work (as reflected in codes of conduct) in the different administrative and legal settings associated with the UK and French asylum processes.
In England and Wales, following a 1993 recommendation by the Runciman Royal Commission on Criminal Justice, the National Register for Public Service Interpreters (NRPSI) was set up, together with a qualifying diploma examination in Public Service Interpreting. From 1997 onwards, a National Agreement required every interpreter working in courts to be registered with one of four professional bodies, all of which had codes of professional conduct 8 : similar arrangements governed UKBA procedures. However, a January 2010 audit found that booking and payment arrangements were inefficient, quality assurance systems were inadequate, and some courts used outdated NRPSI lists or unregistered interpreters. The costs of the system were not clear. In August 2011 the Ministry of Justice (MoJ) therefore signed a four-year contract with a private company, Applied Language Solutions (ALS), following a competitive tendering process, and ALS began supplying interpreters to courts and tribunals from 30 January 2012. The initial months of this contract were dogged by complaints and controversy, however. Many experienced interpreters refused to work under the greatly reduced pay scales offered by ALS, and the MoJ itself admitted to an 'unacceptable' number of problems, such as failures to provide interpreters or provision of incompetent interpreters, leading to repeated adjournments (House of Commons Justice Committee [HCJC], 2013, pp. 5-11, 41n). By the end of February 2012, the MoJ was forced to allow courts to revert to the old system as an interim measure to avoid further interpreter-related delays. In October 2012, ALS was taken over and rebranded as Capita Translation and Interpreting, but complaints about pay rates and about the reduction or removal of travelling expenses have continued.
While the use of interpreters in French refugee status determination procedures is not a new phenomenon, a specific 'interpreting service' (service d'interprétariat) was only created at the CRR/CNDA in 1994 and at the OFPRA in 2001. In both cases, this was largely a response to an increase in the number of asylum applications and appeals received and of the languages spoken by asylum applicants. Since 2003, the provision of interpreters has been organised through a system of 'competitive tendering' (marchés publics), covering both the OFPRA and the CRR/CNDA. While 46% of asylum interviews at the OFPRA were conducted in the presence of an interpreter in 2003, this had risen to 82% in 2011 (OFPRA, 2011(OFPRA, , p. 106, 2012. Not surprisingly, the cost of providing interpreting services at the OFPRA has also increased, from 1.1M€ in 2005 to 2.4M€ in 2011 (OFPRA, 2007(OFPRA, , p. 38, 2012. At the CNDA, 95% of appellants are not French speakers (CRR, 2007, p. 19), and the cost of interpreting services was 1.1M€ in 2008 although this is likely to have risen since then, as has been the case at the OFPRA (2009, pp. 44, 50). In 2011, the OFPRA received 40,464 first applications, an increase of 9.6% compared to the previous year, while 31,983 appeals were lodged at the CNDA, a rise of 16.5% in one year (Conseil d'État, 2012, p. 33;OFPRA, 2012, p. 74).
Both UKBA's Central Interpreters Unit and the Tribunals Service have codes of conduct for the interpreters they use. While mostly concerned with general professional behaviour, the UKBA code does include a section on 'Accurate and Precise Interpretation'. Typically, this begins by stressing that interpreters must 'retain every single element of information that was contained in the original message, and interpret in as close verbatim form as English style, syntax and grammar will allow' (italics added). 9 Similarly, the Tribunals Service's Handbook for Freelance Interpreters (2011) instructs interpreters to 'use the witness's exact words. If you cannot make a direct or exact interpretation, interpret it as accurately as possible in the witness's own words and then inform the Judiciary what the phrase means'; later, it re-emphasises 'Please do not … use an English expression or phrase which is not an exact translation of the witness's own words' (Henderson & Pickup, 2012, para. 34.24). What constitutes an 'exact translation' is left unclear.
In France, there is a code of conduct for interpreters at the CNDA but not at the OFPRA. The CNDA code (a copy of which is displayed on the wall of the interpreters' room at the Court) sets out the five 'duties' (devoirs) of interpreters working there: punctuality and diligence, impartiality and independence, neutrality, accuracy (la justesse de l'interprétation) and confidentiality (le secret professionnel). Although an equivalent document does not exist at the OFPRA, a number of 'basic rules' (règles élémentaires) for interpreters, such as neutrality, are written into the specification or tender documents (cahier des charges) used in the competitive tendering procedure through which the provision of interpreting services there (and at the CNDA) is organised (interview with Head of the Interpreting Service, OFPRA). Although a single system of competitive tendering covers interpreting services at the OFPRA and the CNDA, and the same interpreter may therefore be contracted to work at both institutions, what is expected of the interpreter in practice is not the same in the two settings. Semi-structured interviews conducted in 2008 and 2009 with the Head of the Interpreting Service at the OFPRA, her counterpart at the CNDA, interpreters working at the OFPRA and/or CNDA, OFPRA caseworkers and CNDA rapporteurs, as well as observation of asylum interviews and appeal hearings, highlighted a number of important differences. 10 One of these relates back to the distinction between 'translation' and 'interpretation' introduced at the beginning of this section. Thus, the Head of the Interpreting Service at the OFPRA commented in the following way on the interpreter's role at the administrative institution: The interpreterwhat I always say to interpretersfor me, interpreters here [i.e., at the OFPRA]and it is not pejorative, on the contraryare instruments of communication. And that is no small matter. (…) I explain to interpreters that their role is to translate 11 (traduire), and I'm using this term translate (traduire) deliberately. Because you know that in French interpret (interpréter) can be something else. (…) So, the interpreters should translate. It's an essential role and it's not as simple as it might appear.
Thus, the expectation at the OFPRAand this was confirmed by interpreters interviewed for the research, as will be shown belowis that interpreters will 'translate' rather than 'interpret' the speaker's words, in the sense of providing a word-for-word or verbatim translation. Although the OFPRA is an administrative institution and not a court, the perspective on the interpreter's role here appears similar to the one held by many lawyers in English and Welsh courts, as discussed above.
In contrast, what is expected of the interpreter at the CNDA, an administrative court, is not a literal or verbatim translation but instead the transmission of an equivalent message or meaning from one language to another. The following comment from the Head of the Interpreting Service at the CNDA makes this clear: I am not sure that I ask them [i.e. interpreters] to do the same job (le même métier). At the Office [OFPRA], they want word-for-word, a word-for-word translation. What I ask them to do is to transcribe the meaning (sens) of what is said. (…) That is a fundamental difference.
A 25-page document entitled 'Interpreting at the Refugee Appeals Board: Code of Conduct and Organisation' (CRR, 2007), written by the Head of the Interpreting Service at the CNDA and distributed to all interpreters working at the Court, appears to make a similar distinction when it states that: 'The role of the translator 12 is not to transcode (n'est pas de faire du transcodage) but rather to render the meaning (sens) contained in the appellant's discourse and in the questions asked by the panel of judges' (p. 14). This sentence is immediately followed by a paragraph in which it is emphasised that interpreters should not 'limit themselves to a mechanical translation (une traduction machinale) of the questions and answers' uttered by the various participants in the hearing, but must also attempt to reflect the 'nuances' contained in the questions asked by the judges when interpreting these for the appellant (p. 14).
Interpreters with experience of working at both the OFPRA and the CNDA confirmed that expectations with regard to their role were not the same at the two institutions. As one interpreter stated: In fact, there is a huge difference. At the OFPRA, what we are asked for is more a translation, in the literal sense of the term, that is to say an almost literal translation of the asylum seekers' words. What we are asked for here [at the OFPRA] is really a translation of the words, practically an oral translation. They call it here -I like the expression very muchbrut de décoffrage. Décoffrage is when you make cement, and as soon as you remove [the framework or coffrage], you see exactly what is there. That is to say, they want to have the unprocessed material (la matière brute). That's the first big difference. At the CNDA, it is more the case that we do oral interpretation. 13 In other words, the interpreter at the OFPRA is generally called upon to provide a word-forword or verbatim translation of the asylum applicant's words that leaves them in a 'raw' or 'untreated' state, available for subsequent 'interpretation' by the caseworker. The situation is different at the CNDA, where accurate and precise interpretation is not equated with a literal translation of words from one language to another. There is thus a striking contrast between the two French institutions in this respect, whereas the expectations of their British counterparts appear broadly similar to each other.
From the users' perspective, it is worth noting, finally, that the MoJ in the UK has general guidance for criminal court staff on working with interpreters, but the Civil Procedure Rules say nothing whatever about this. 14 Similarly, the Tribunal's Practice Directions and Practice Notes (www.justice.gov.uk/tribunals/practice; accessed 6 March 2013) do not discuss interpreters or how to work with them, even though judges are expected to assess their performance after every hearing. 15 However, its Guidance Notes do provide a script for IJs to follow when explaining the interpreter's role to the appellant and establishing that the appellant and the interpreter understand one another. Rather oddly, the (very basic) checking of mutual comprehension is left right to the end of this script, after the judge has explained the structure of the hearing and the role of the interpreter, all of which have to be translated even though it has yet to be confirmed that the appellant understands it. 16 Another guideline states that if applicants are unrepresented in court, which is increasingly common following cuts to legal aid, requirements of fairness may entail the interpreter reading out the asylum interview transcript, RFRL, and other case documentation to them, possibly in full, while 'Summaries of the objective evidence may be prepared in advance to be translated to the appellant in Court'. 17 The guidance does not say whose responsibility it is to prepare such summaries. Guidance for lawyers on working with court interpreters in asylum hearings is given in the Immigration Law Practitioners Association's Best Practice Guide (Henderson & Pickup, 2012, chap. 34). Unlike the other UK documents mentioned, this discussion does reveal an awareness of the limitations of even the most competent interpretation, especially when, as is invariably the case, different interpreters have been used at each stage of the refugee status determination process (Henderson & Pickup, 2012, paras. 34.4-34.8).

Fragmented narratives, interpreting dilemmas and barriers to communication
At all the different stages described in the previous sections the interposition of interpreters creates barriers to communication, irrespective of their competence. For example, in the UK each of the various codes and guidance stresses the need for asylum applicants to answer questions in short phrases or sentences, so that these can be fully translated. While it is of course important that everything an applicant says is communicated to their interlocutor, this fragmentation of the narrative introduces limitations of its own, as interpreters themselves are fully aware; they also know that questioners turn these limitations to their own advantage. As one interpreter commented: People are discouraged from talking, and the interpreter is always made the excuse for that; you know, that you need to give short answers so that the interpreter can translate? That is another constraint brought on by the interpretation process itself … and it works in the favour of the Home Office because people do not speak like that naturally, and they will lose track; they will say less than they mean to say simply because they have to break it down. I've seen it on their faces; they just give up, you know? They try to do it for a bit and then they give up. And let's not forget, the interpreter has practice at doing this, so do the Presenting Officers, so do the judges, so do the lawyers, but the asylum applicant doesn't. He's the only one who comes to this all fresh and raw and natural, and the flow of speech, when it's constrained like that, very often dries up.
While this applies to the entire asylum process, its different stages do vary in practice in terms of the restrictions that interpreters themselves are under. When comparing screening and substantive interviews, one interpreter with more than a decade of experience in Home Office work stated: With the substantive interview … the interpreter isn't supposed to intervene all that much, whereas in a screening interview it's okay to intervene because it's just collecting the data. It's less formal. I think the interviewing officers are taking the substantive interview more seriously in the sense that there's more important stuff to talk about. They know the screening interview is just provisional, and they often stop people from talking, 'No, not now, not now, you're going to tell me later, you're going to tell me later'. So that is done in quite a hurry.
Whereas those assignments both involve almost entirely the consecutive interpretation of questions and answers, at hearings themselves the interpreter is also supposed to provide the appellant with whispered simultaneous interpretation of dialogue involving other participants. Often this does not happen, however, not least because it is extremely demanding for interpreters to have to work continuously for such long periods, especially when required to switch to and fro between consecutive and simultaneous modes. Even when they do attempt this judges sometimes ask them to stop, as they find it distracting.
The various UK codes and guidelines also specify that interpreters should, as the Tribunal's Handbook puts it, 'endeavour to reflect the type of language that is being used, whether it is simple, formal, colloquial etc.' (Henderson & Pickup, 2012, para. 34.24). This may place the interpreter in a quandary when appellants speak ungrammatically, rudely or colloquially 18 : That's a dilemma. I find myself actually (laughs) brushing up [their] speech because I certainly don't want to come across as a bad interpreter and it can be believed that it is me who is making the grammatical errors. 19 And also, for me, it is difficult to interpret in a different register to the one I usually speak in. I realised that for a long time I did that, when interpreting for [one particular minority group] at the Home Office. I would use polysyllabic words … and then I thought, hang on, these people are going to go to appeal and the way they speak is going to come across so, so disjointed from how they came across in the asylum interview. What's that going to do to their credibility?
In addition to having possible implications for perceptions of the credibility of an appellant's narrative, differences in the register of speech employed by participants in the hearings constitute potential barriers to communication and pose further challenges for interpreters. The need for interpreters sometimes to adapt the wording of questions in order to ensure effective communication is explicitly acknowledged in the code of conduct for interpreters at the CNDA in France. Under the heading 'The Accuracy of Interpretation', it is stated that: [The Interpreter] must behave in a useful and positive manner: to the extent that is strictly necessary for the expression of justice, the interpreter is authorized to adapt their language and to reformulate the questions when the appellant has a level of language that is less sophisticated (moins élaboré) or in the case of incomprehension.
The following extract from an interview with an interpreter who had experience of working at both the OFPRA and the CNDA provides some examples of this kind of adaptation: Interpretation at the Court is more difficult, technically speaking. Why? Because you must remain faithful to the translation, to the appellant's words, but at the same time you must make a dialogue possible. Now, on the one hand I have judges who have a string of qualifications (qui ont fait bac plus 36) and on the other peasants, fishermen, farmers, people who don't know anything about the system in their country and even less about the system in France. So, for example, if I'm asked to translate 'Were you arrested and kept in police custody (arrêté en garde à vue)? Were you brought before an examining magistrate (déféré devant un juge d'instruction)?' I can translate that to the person, absolutely, but they won't understand a thing. Do you see? Therefore, while remaining faithful to what was meant and to the words, I have to adapt my language in such a way that it is understood, because if the legal dialogue (le dialogue judiciaire) is not possible, I'm not being of use. (…) [At the OFPRA] they don't want this adaptation; they want the raw material (ils veulent du brut).
This highlights the fact that while there is often an institutional requirement or expectation that the interpreters will be 'invisible', 20 the latter's role in practice is often that of 'an active verbal participant' (Berk-Seligson, 2002, p. 64) in the interactions that take place in UK and French asylum interviews and appeal hearings.
The issue of the interpreter's role can be explored further by returning to the UK guidelines and codes of conduct discussed earlier. These stress that the interpreter's task is to provide verbatim or 'exact' translation and that they must not, as the UKBA code of conduct puts it: ask the interviewee what they mean by a particular answer … try to anticipate what the interviewer or interviewee is trying to say or give an answer other than what is being said (or) let your own experience or views get in the way of how you interpret the evidence.
The assumption here, very clearly, is that the interpreter's own understanding of what the interviewee means to say is to be suppressed because it will 'get in the way' of the desired verbatim translation.
Such naivety about the translation process is of course not confined to legal contexts. However, its stress here seems partly also to reflect the centrality of language to the entire legal process. As Wadensjö argues, building upon Morris's work cited earlier (and also Morris, 1995): established legal systems show little or no readiness to acknowledge the interpreter-mediated situation as essentially different from the ordinary, monolingual one, and the court interpreter's task as truly interpretive. Instead, the court interpreter is defined as a disembodied mechanical device. (Wadensjö, 1998, p. 74) This, she argues, is partly because law itself, as performed in court, depends so heavily on the skilful manipulation of language by lawyers and its incompetent or untrained use by those under cross-examination (which will almost certainly be portrayed in asylum contexts as damaging their credibility). As Atkinson and Drew (1979) pointed out in their seminal study of courtroom dialogue, cross-examination seeks to 'to challenge or blame the witness' by getting them to agree to the 'facts' progressively brought out during the questioning (pp. 105, 106). Both questions and answers are moulded by expectations over what the interlocutor will say next. For example, barristers expect that their accusations will produce denials, and try to turn that expectation to advantage by choosing forms of words such that witnesses damage their standing or credibility whatever answers they give ('when did you last beat your wife?'). Witnesses themselves may respond by 'hedging' to mitigate any potential admission; instead of a simple 'yes', they reply 'I suppose so'. These processes become far more difficult for lawyers when they are forced to work through interpreters, and legal efforts to limit and 'mechanise' the interpreter's role can be seen as attempts to maintain as far as possible the hegemony of the examining lawyer. Thus, as Wadensjö (1998) notes: it would obviously be a challenge to the court if interpreters were … allowed to clarify an attorney's deliberately ambiguous question. It would be a threat to the system if interpreters were allowed to improve the image of witnesses … by rendering eloquently and precisely statements which were originally voiced carelessly and imprecisely. (p. 75) On the other hand, the UK Tribunal's Handbook for Freelance Interpreters also provides potentially contradictory guidance: You may intervene at the hearing for the following reasons: to seek clarification if you have not fully understood what [has] been asked to interpret; to alert the Judiciary that although the interpretation was correct, the question or statement may not have been understood; to alert the court to a possible missed cultural inferencesuch as when an item of information has not been stated but knowledge of which has been assumed. (in Henderson & Pickup, 2012, para. 34.24;their gloss) Clearly then, interpreters constantly have to use their judgement over whether to 'intervene' and explain that a misunderstanding or missed inference may have occurred. Not surprisingly, they differ greatly in the extent to which they do this in practice. Their behaviour is likely to reflect not only their own personal shyness or ebullience but also the professional stress upon self-effacement. They also know very well that if they intervene in this way too often, this is likely to be attributed to their incompetence rather than their alertness and sensitivity.
Missed social and cultural inferences are rendered more likely by what Rycroft labels the 'silent actors' in asylum hearings, namely, the legal elements of refugee law that motivate particular questions but of which the appellant is usually unaware. For example, a person is only recognised as a genuine refugee if they have first exhausted all avenues of domestic protection, so they are almost always asked about relocating within their own country: Thus the ubiquitous question is: 'Have you tried to move to a different part of Romania?' Many applicants say, 'No, because the police are hand in hand.' The fact that they did not attempt to relocate will count against them, although, had the applicants known where the question was aiming, they may have explained that in Romania the police keep centralised records, and that in order to move away one has to request a residence visa from the police. (Rycroft, 2005, p. 241) As an experienced legal interpreter who is herself Romanian, Rycroft is fully aware of this unspoken contextual background, both the legal reasons for the question and the relevant information missing from the appellant's answer. The dilemma for her is whether to interpose her own explanation. Some judges will accept such behaviour by an interpreter, at least occasionally, but they are more likely to rebuke her for exceeding her duties by giving evidence as though she were herself a witness.
Elsewhere, one of us gave a lengthy example involving an exceptionally proactive Farsi interpreter, showing how his own interpolations frequently helped the court by clarifying the appellant's answers, but also sometimes disrupted it, to the extent even of 'correcting' lawyers questions andon one occasionactually subverting the normal order of proceedings (Good, 2007, pp. 167-169, 177-178). The excerpts cited there also illustrated that there are certain matters, such as dates in non-Western calendars, or kin terms where kinship itself is structured very differently, that are inherently impossible to translate exactly or verbatim (see also Kalin, 1986). Different interpreters, or the same interpreter on different occasions, may offer differing but equally valid English (or French) equivalents, creating, purely as an artefact of interpretation, the impression that the appellant is confused and possibly untruthful.

Conclusion
Interpreters play a crucial role in facilitating intercultural communication in the context of asylum interviews and appeal hearings in countries such as the UK and France that are signatories to the 1951 Refugee Convention. As this article has highlighted, however, their task is a complex one. First, the institutions in which they work sometimes have differing expectations as to the nature of their activity (e.g. providing a literal/verbatim translation or transmitting the 'meaning' of messages from one language to another), and these shape the definition of 'appropriate communication' in the particular contexts concerned. Second, institutional codes of conduct or guidance for interpreters can provide potentially contradictory advice, confronting them with dilemmas and requiring them to exercise their judgement on a range of matters. The latter include whether or not to intervene to explain a cultural misunderstanding, how to negotiate different registers of speech without potentially damaging the perceived credibility of an applicant's or appellant's narrative and how to 'adapt' or reformulate questions (and answers) in order to ensure effective communication between the parties involved. Third, the fragmentation of applicants' or appellants' narratives that results from their having to answer questions in short phrases or sentences introduces barriers to communication and imposes restrictions on even the most competent of interpreters (as well as on the other parties involved in the process). Despite these constraints, interpreters are active participants in the often multilingual and intercultural exchanges that take place in asylum interviews and appeal hearings, although the mechanistic views of interpretation and ideas about the 'invisibility' of the interpreter that are sometimes (but not always, as has been emphasised) found in these legal and administrative contexts can obscure this fact.

Notes on contributors
Robert Gibb is a Lecturer in the School of Social and Political Sciences at the University of Glasgow. He has conducted research on the antiracist movement in France and, more recently, a comparative study of asylum processes in the UK and France (with Anthony Good). Recent publications include: 'Do the facts speak for themselves? Country of origin information in French and British refugee status determination procedures'. International Journal of Refugee Law, 25, 291-322 (2013, co-authored with Anthony Good). Anthony Good is Emeritus Professor of Social Anthropology at the University of Edinburgh, and was formerly head of its School of Social and Political Science. His research interests cover South India and Sri Lanka, focusing on kinship and religion. He frequently acts as an expert witness in asylum appeals involving Sri Lankan Tamils. His recent research concerns uses of expert evidence in British asylum courts and (with Robert Gibb) a comparative study of asylum processes in the UK and France. Recent publications include Worship and the Ceremonial Economy of a Royal South Indian Temple (2004) and Anthropology and Expertise in the Asylum Courts (2007).