[2] Anna Gil-Bardají Ethics, accuracy, and interpreting in social settings Assessing a non-professional interpreter profile Anna Gil-Bardají Universitat Autònoma de Barcelona Accuracy tops most of the codes of ethics for public service interpreting. This dominant position is not accidental, since the lack of truth in the trans- mission of the message may entail a violation of the users’ and providers’ fundamental right to truth. In Catalonia, where intercultural mediators have taken on the role of interpreters for many years, the issue of accuracy becomes even more complex due to their dual roles as professional media- tors and non-professional interpreters. Drawing on more than four hours of video recordings and post-task interviews with study participants, this paper analyzes the accuracy of five intercultural mediators working with Arabic, Spanish, and Catalan when in the role of interpreter. Keywords: intercultural mediation, non-professional interpreters, public service interpreting, ethics, accuracy Introduction The increasing presence of immigrant communities among users of public ser- vices in Spain – and the subsequent communicative difficulties that arise – has favored the appearance of new professional profiles capable of effectively guar- anteeing such communication. In Catalonia, a range of professionals, such as intercultural mediators, public health workers (agents de salut comunitària), interpreters in public services, or telephone interpreters, participate in this com- municative transfer from diverse perspectives and settings (MIRAS 2011; Ugarte Ballester and Vargas-Urpi 2018). Of these roles, the intercultural mediator is quite recognizable throughout Catalonia as a result of its being a common profession in the region (Desen- volupament Comunitari et al. 2009; Vargas-Urpi 2013). Acting as a bridge between languages and cultures, these professionals usually offer their services Ethics, accuracy, and interpreting in social settings [3] out of solidarity (MIRAS 2011: 42), for little or no compensation, in order to fill a professional gap of increasingly important human and social dimensions (Ugarte Ballester and Vargas-Urpi 2018). Unlike in other regions where intercultural mediators are present, in Catalo- nia the professional roles and profiles of the intercultural mediator and the pub- lic service interpreter often converge. Yet these two profiles diverge in many other aspects, the first and most important of which is their principal objective. While the main aims of the intercultural mediator are to actively bring the parties closer together, to perform information-sharing tasks, or to offer support or advice as necessary (Bermúdez et al. 2002), the role of the interpreter in public services is to facilitate communication when there are linguistic barriers by means of a range of specific competencies (Abril Martí 2006: 666–671; de Pedro Ricoy 2010; IoLET 2017). Thus, it is no surprise that translation and interpreting studies in Catalonia, and particularly research on public service interpreting and translation, have begun to examine how intercultural mediators perform specific interpreting tasks. In addition, of the three traditional areas of public service interpreting (PSI) – i.e., healthcare, socio-educational, and legal – intercultural mediators in Catalonia principally act in the healthcare and the socio-educational realms (Ugarte Ballester and Vargas-Urpi 2018: 48). In this research, the term for intercultural mediators in their function as inter- preters is “mediator-interpreter” (MI). Adopting García-Beyaert’s (2015) green- blue hat metaphor to describe their dual roles, this term refers to those cases in which the intercultural mediator takes off their green mediation hat (creative and interventionist) to put on their blue interpreting hat (vehicular and non- interventionist). Drawing on more than four hours of video recordings and in-depth personal interviews with each participant, this article examines the accuracy of the perfor- mance of five intercultural mediators working with Arabic, Spanish, and Catalan in three simulations of common situations in interpreter-mediated communica- tion in social and educational settings in Catalonia. To do so, the renditions of these intercultural mediators are examined to identify when, why, and under what circumstances accuracy has – or has not – been achieved. Theoretical background This research stands at the junction of four principal lines of investigation within PSI. The first line focuses on interpreting behavior in the specific setting of social services and education. This refers to studies that explore situations defined by the [4] Anna Gil-Bardají Figure 1. General descriptions of and overlaps between mediation and interpreting roles (García-Beyaert 2015) domains where the communicative transfer requirement develops; namely, citi- zens’ advice offices1 (e.g., MIRAS 2011; Ugarte Ballester and Vargas-Urpi 2018), primary and secondary schools (e.g., Arumí 2017 and 2018; Vargas-Urpi and Arumí 2014), and immigration and refugee services (e.g., Pöllabauer 2004; Merlini 2009; Baraldi 2014). This article contributes to this first line of investi- gation by empirically analyzing interpreters’ real behavior in three prototypical situations within these specific settings (educational and social) by means of screenplays carefully designed and monitored by a research team. The screen- plays take into account the main characteristics and most common interpreting challenges of these situations. The second line focuses on professional standards of practice and codes of ethics in PSI, including the following three types of contributions: (a) non- research based proposals from organizations of interpreters principally aimed at professionals (e.g., NAATI 2016; NCIHC 2005; NPRSI 2007); (b) scientific and academic research principally directed at PSI researchers or trainers (Angelelli 2006; Angelelli et al. 2007; Baixauli-Olmos 2014 and 2017; Baker and Maier 2010; 1. In Catalonia, the Oficines d’Atenció Ciutadana (OAC) are in-person local services that pro- vide information and counseling about specific and general administrative procedures. Ethics, accuracy, and interpreting in social settings [5] Kalina 2015; Pym 2001 and 2012); and (c) works dealing with the concept of “accuracy” in interpreting, in general and PSI in particular (Gile 2009; Setton 2015; Tiselius 2015). According to Tiselius (2015: 3), “there are few explicit defini- tions of the concept of accuracy, or consolidated descriptions of what accuracy in interpreting actually consists of.” Yet, as Setton (2015: 161–162) recalls, within the notion of fidelity and its equivalent terms, accuracy has been traditionally seen as the primary criterion of quality and associated with notions of trust, integrity, and professional ethics. Setton (2015: 162–163) proposes the following definition of optimal fidelity: [“Optimal” fidelity encompasses] every aspect of the intent, content, expression, tone and style of the communication, that can be pursued in choices to optimise form (style, coherence, concision, packaging, recapping, prosody and body lan- guage), content (adding, explicating, clarifying, “localising” cultural references, filtering, toning down, correcting errors, etc.), or process (e.g. intervening in the exchange to correct misunderstandings). This definition provides a useful framework for the present study, since it consid- ers different levels and dimensions of the communication process. The third relevant line of research analyzes interpreters’ strategic behavior and capacity to resolve problems during the communicative interaction. This research, generally empirical, focuses on the problem-strategy relationship; in other words, what kinds of problems interpreters usually face when performing their task and what techniques or strategies they employ to solve them. Even though this research has garnered considerable interest in a range of aspects of translation and interpreting, PSI has yet to fully explore problem-solving strate- gies used by interpreters. Within this line, Abuín’s contributions (2004, 2007) as well as those of Valero-Garcés (2005), Napier (2004), Arumí (2012), Vargas-Urpi and Arumí-Ribas (2014), and Arumí and Vargas-Urpi (2018) are particularly rel- evant, since they shed light on specific interpreting problems such as terminol- ogy, interpreters’ own interventions, role shifting, and coping with non-verbal communication, as well as possible technical and strategic solutions. This arti- cle expands and enriches this specific line of inquiry by relating the “problem- strategy” binomial to the interpreter’s active search for accuracy. The problem- solving perspective will be used to observe and understand the inter- preter’s behavior when coping with specific challenges during the interaction, whether linguistic or not. Finally, the fourth line of research encompasses the role of non-professional interpreters in the PSI setting (Aguilar-Solano 2015; Boéri 2010; Martínez-Gómez 2015a, 2015b; Mayoral Asensio 2012; Olohan 2014; Pérez-González 2010; Susam- Sarajeva and Pérez-González 2012). Particularly relevant are those studies that [6] Anna Gil-Bardají relate, contrast, and reflect on the similarities and differences between the figure of the intercultural mediator and that of the interpreter in public services. In most cases, this type of research is conducted where these two professional pro- files co-exist, such as Italy and Spain. For example, several studies investigate decision-making and problem-solving behavior of intercultural mediators (e.g., Arumí 2018; Gil-Bardají 2016), while others explore the limits of mediation when interpreting in healthcare settings (Baraldi 2014) or more generally (Pöchhacker 2008). Still others are interested in power dynamics during interpreted events (Merlini and Favaron 2003) as well as the co-existence of multiple professional profiles in these contexts (MIRAS 2011). The approaches have varied, including corpus-based or observational studies (e.g., Vargas-Urpi and Arumí 2014; Vargas- Urpi (2013, 2016), while others have adopted an interactional approach to investi- gate interpreted renditions (Baraldi and Gavioli 2007, 2014, 2015). Objectives and participant profile The main objective of our study is to investigate the level of accuracy of ren- ditions produced by five Arabic-Spanish/Catalan mediator-interpreters (MIs) working in socio-educational settings in Catalonia. More specifically, the study analyzes the type and frequency of MI interpreted renditions.2 These observa- tional data provide potential indicators related to their decision-making process when interpreting. The sample consists of four women and one man, all natives of Morocco (four from northern cities and one from the capital, Rabat). Based on MIRAS database of active MIs, ten MIs were short-listed based on their experience with local inter- preting services in public institutions and the five MIs with the most extensive experience were selected. The MIs, who agreed to participate as volunteers in the study, have worked in the Catalan public services for nine to sixteen years. Three speak Moroccan Arabic as their native language, while the other two also speak Riffean Tamazight or Tarifit. All of them are equally fluent in Spanish and Catalan and they all have university education: two hold degrees in translation and inter- preting and three in linguistics. In addition, one of the interpreters (with a degree in linguistics) has had specific training in PSI. Finally, three of the five MIs stated 2. This study is a result of the project “Problemas y estrategias de traducción e interpretación en el ámbito socioeducativo. Estudio de situaciones de interpretación de chino, rumano y árabe” [Translation and interpreting problems and strategies in socio-educational settings. A study of interpreting situations from Chinese, Rumanian and Arabic] conducted by the MIRAS research group. Ethics, accuracy, and interpreting in social settings [7] that they were completely fluent in standard Arabic (Fusha), three in French, and two in English. Rich points (different types of consciously chosen “accuracy problems”) were intentionally introduced throughout the three scripts in order to observe how the interpreters would cope with them. Methodology This research project combines methodologies in order to investigate the research questions. The qualitative element was based on two parameters: (1) the clas- sification and analysis of more than one thousand renditions produced by the MIs, based on categories proposed by Wadensjö (1998); and (2) the observation and coding of the audiovisual recordings (particularly non-verbal language; style, tone and register; management of the conversation by the MIs; and contextual aspects). In addition, interviews that were conducted after each of three simula- tions with the MIs were coded and analyzed. During these interviews, the MIs commented and reflected on various aspects of their performance, and they were prompted by questions about their performance when interpreting. These ques- tions also focused on the identified rich points inserted in the scripts to better understand questions surrounding why they interpreted a sentence in a particular way or why they used a specific strategy. Wadensjö’s (1998: 106, emphasis original) classification covers eight types of renditions, which are understood to be “a stretch of text corresponding to an utterance voiced by an interpreter, relating in some way to an immediately pre- ceding original.” These eight rendition types are as follows: close renditions; expanded renditions; reduced renditions; substituted renditions; summarized renditions; two-part or multi-part renditions; non-renditions; and zero rendi- tions. In the present study a ninth category, contributed by Gavioli and Baraldi (2011), has been added: stand-by renditions, which are those left uninterpreted but recovered and translated at a later stage of the interpreting. The analysis draws on the transcription and classification of the 1,283 rendi- tions uttered by the Arabic-Spanish/Catalan MIs in each of the three simulations.3 This permitted the analysis of the accuracy of the translation that will be pre- sented in the next section. 3. S1: Simulation of a meeting between a mother and her son’s tutor at a public secondary school; S2: Simulation of an interview between a mother and an official at a Municipal Center for Education; S3: Simulation of an interview between a mother and a social worker and social educator (“educador social” in Spanish) at an immigration office. [8] Anna Gil-Bardají Results Before analyzing the data, it is necessary to explain that in some cases, a particular rendition may correspond to more than one type in Wadensjö’s classification. For example, a rendition might be both expanded and substituted, or multi-part and reduced. As a result, in some cases the sum of the percentages exceeds 100%, as the total number of classifications exceeds the number of renditions. Rendition typology and frequency of use in terms of accuracy The transcription and classification of the 1,283 renditions produced by the MIs in the sample reveal two particularly significant results. The first is that more than a quarter (27.51%) of the renditions produced are non-renditions; they are interven- tions made up by the MIs that do not correspond with any of the utterances in the source language. The second is that only 17.77% of the renditions are reproduced in a more or less precise way with regard to the content or form of the source lan- guage utterance, while 82.23% are modified in some way (also included here are the 27.51% of the MIs’ non-renditions). The modifications to the original can be divided into two types: (a) modifica- tion in content and (b) modification in form. The first, namely content, comprises 38.56% of the total and are formed by zero, reduced, substituted, and expanded renditions. Renditions that show a change in form constitute 22.51% of the total and are represented by stand-by, multi-part, and summarized renditions. Those remaining consist of non-renditions (27.51%) and close renditions (17.77%). With respect to these non-renditions (27.51% of the total), a little more than a third (34.26%) are justified interventions from the point of view of interpreting as interaction (Arumí 2017; Wadensjö 1998). These MI interventions are (a) to request a break (1.98%); (b) to request clarification or repetitions (17.84%); or (c) to verify information following an issue of comprehension (14.44%). These three types of interventions are exemplified as follows: a. A request for a break, verbally or by means of a gesture: USER: =‫هو سمييتو داودي عادل‬ [Her name is Dawdi Adi] MI: ‫بالتي )*( الداودي عادل‬ [= Wait (x) Dawdi Adil] b. To request clarification or repetition: USER: =‫أأنا كنسكن حدا السوق‬ [I live next to the market] MI: ‫[=أأش من سوق؟‬Which market?] Ethics, accuracy, and interpreting in social settings [9] c. To verify information: USER: ‫ ما خدامة والو‬،‫=ال أأختي‬ [No, sister, I don’t work=] MI: =‫والو والو؟‬ [Not at all?=] The remaining 23.79% of the non-renditions offer additional information to the user; that is, they add information that was not stated by the provider, but that the MI considers necessary to include. This additional information takes the form of cultural or thematic explanations or information about how the service works. These types of interventions correspond to the MIs’ personal quality standards, as they expect to be active in the communication event, in addition to carrying out informative, supportive, and advisory tasks for the users, as in the following example: Offering extra information, performing the role of helper to the provider: PROVIDER: Tell them to bring these documents (points at the pages) Alright?=] MI: ‫ واخا؟‬.‫( داك البطاقة وتعملي ليها فوتوكوبي‬.)‫ جيبي يعني‬.‫=واخا‬ [OK. Bring, or::: (.) this card and make a photocopy. Similarly, 18.13% are interventions in which the MIs express conciliation, complic- ity, empathy, or cooperation with the user (e.g., “don’t worry, you will see, every- thing will be alright”). These interventions also belong to the field of intercultural mediation, as one of the principal objectives of this role is to form a close bond with the parties and achieve mutual understanding and trust. Expressions of conciliation, complicity, empathy, or cooperation with the user: PROVIDER: OK, we will do this as quickly as we can, shall we? MI: ‫( يبدا دﻏية‬.) ‫قالت ليك هما ﻏا يزربو و يديرو اللي ف جهدهم باش يبدا‬ [They said that they would be quick and will do everything they can so that it can start (.) start very soon] USER: ‫إإن شاء اهلل إإن شاء اهلل‬ [ Yes Please God, yes] MI: ‫ما يكون ﻏير الخير اختي‬ [Everything will be alright, sister] Another type of non-rendition is referred to as reactive tokens, discursive mark- ers, and repetitions of confirmation habitually used in dialogic and interactive communication (e.g., uh huh, OK, or I understand). In the current study, these reactive tokens form 10.88% of the total of non-renditions. Some examples are pre- sented below. [10] Anna Gil-Bardají Reactive tokens, discursive markers, and repetitions of confirmation: USER: ‫واه أأنا باﻏة نعرف النقط ديال ولدي فهاد الدورة وكيف ﻏادي مع القراية ديالو‬. [ Yes, I would like to see my son’s grades for this trimester and find out how he is doing at school] MI: ‫ آآه‬،‫…واخا‬ [(to the USER) Yes, OK] [(to the PROVIDER) A::: she wants to know::: her son’s grades for this::: trimester how he is doing and::: if she can, she wants to know if you can tell her this] However, in only a few occasions did the MIs respond in the place of the user (7.93% of the non-renditions) or offer advice, warning, or personal opinion (1.98% of the non-renditions), demonstrating an awareness of some professional ethical standards of PSI. Concerning those few cases in which MIs respond in the place of the user, they mainly occur when “translating” a facial expression or other form of non-verbal language, as in the following example: Responds to a question from the provider on behalf of the user: PROVIDER: Is that clear? USER: – (nods lightly). MI: Yes, she understood. In the case of giving advice, warning or opinion – although observed very rarely (1.98%) – it is evident that the MI exceeds the limits of professional ethical stan- dards for PSI, which clearly prohibit the expression of personal opinions under any circumstances. The following example illustrates this type of intervention: Warnings, advice, expressing a personal opinion: ‫للبالد‬.USER: ‫ خاصني نجمع باش شوي د الفلوس نسرد عاود تاني‬.‫ واهلل إإال‬،‫ عندك الصح‬،‫آه‬ [Ah, you’re right, it’s true. I have to save a bit of money to send to my country] ‫ بالحق جاو دابا هما‬MI: ،‫عاود تسردي للبالد! فكري بعدا ﻏير فرأأسك حتى تفكي رأأسك ومن بعد فكري ف البالد‬ ‫يساعدوك ولكن‬ [Send money home? First think of yourself, until you find a solution to your problems and then you can think of your country. Now they are here to help you, and you…] Finally, the last type of the non-renditions are comments about interpreting itself, answers to direct questions to the MI, or the initial introduction of the MI, as shown in the following example: [10] Anna Gil-Bardají Comments about the interpreting itself, answers to direct questions, initial intro- duction of the MI: PROVIDER: Well, as on this occasion the meeting has been called by the parents, I would like to know if there is anything concrete they want to talk about, anything that is worrying them. ‫ هاد‬،‫ لينتريبيستا‬MI: ‫( هاد‬.) ‫( هاد‬.) ‫ اليوم الهدف ديال هاد‬،‫ مترجم د البلدية‬،‫ انا محمد‬،‫السالم عليكم‬ ‫))مع اأألستاذة ))ﻏير مسموع‬. ‫( باش تهدري‬.) ‫( عالش طلبت هاد اأألخر‬.) ‫( هو يعرفوا يعني‬.) ‫المحادثة‬ [Hello, I am Mohamed, the council’s translator, today the objective of this (.) this (.) interview. This interview (.) so that they know (.) why you have requested this (.) to speak to the teacher ((inaudible))] The analysis of the non-renditions demonstrates that a large proportion of such interventions, despite not corresponding with any original, are justified either from the point of view of communication as interaction, or from the perspective of intercultural mediation. The discussion of results will deal with this question in depth. Renditions with changes in the form of the message This type of rendition, represented by stand-by, multi-part, and summarized ren- ditions, make up almost a quarter (22.51%) of the total renditions analyzed. A multi-part rendition is understood to be formed by two or more renditions cor- responding to a single original intervention. This segmentation of the original is usually produced when the segment presents a significant extension or when the original includes several questions that the MI prefers to formulate separately to be in better control over the information. Summarized renditions are those in which the MI combines, in a single intervention, the information contained in two or more original interventions, at times corresponding to a single speech and other times to more than one. Finally, stand-by renditions are those not inter- preted at a given moment but which the MI recovers later. None of these three types of renditions entail a loss of information contained in the original, although they do imply a formal restructuring of said message. Likewise, the MIs’ use of them indicates a certain mastery of dialogue interpreting strategies. These strategies, frequently unconscious and automatic, are oriented toward resolving specific communication problems, such as an excessive accumu- lation of information in a single turn (which may result in a loss of informative content on the part of the MI or the message receiver); a disordered presentation of priority and secondary information (which may confuse the receiver); or the presentation of related or complementary information in distinct turns (which may not cohere in the final message). Ethics, accuracy, and interpreting in social settings [11] In all cases we find a behavioral strategy oriented toward resolving informa- tion management problems. Renditions with changes in the content of the message As mentioned, changes in the content of the message make up 38.56% of the total number of renditions produced by the MIs in the sample and are represented by zero, reduced, substituted, and expanded renditions. A rendition is zero when an original intervention remains untranslated. In the present study, zero renditions account for 13.09% of the total number of inter- ventions by the MIs and correspond in the vast majority to brief oral or ges- tural expressions of affirmation, gratitude, or agreement with one of the parties whose meaning is not translated by the MI, considering its translation unneces- sary. Some examples of these expressions are interjections, clicks of the tongue, sighs, expressions of affirmation such as “m-huh,” “eh,” “uh huh,” or non-verbal language. Most PSI codes of ethics and professional standards of practice stipulate that interpreters should translate these verbal or non-verbal expressions, given that they reinforce both the communicative flow between the two parties and the role of the interpreter as guarantor of informative transfer. Only in a few cases do the MIs in this study not transmit a whole turn for other reasons, such as when they consider the word or sentence sufficiently clear by context. Reduced, substituted, and expanded renditions are those in which the MI has modified the content of the message by incorporating additional information, omitting relevant information, or expressing different information than that con- tained in the original segment. Significantly, in the present study these three types of renditions add up to 25.47% of the total, meaning that a quarter of the rendi- tions do not conform to the original meaning. According to data obtained from the observation and the post-task comments of the MIs themselves, this lack of accuracy with respect to the original message can be attributed to three principal reasons, corresponding to each type of rendition. With respect to the expanded renditions, the MIs usually cite the need to add certain information in the interest of transmitting the message more clearly or in order to reinforce certain ideas. In the following example, “the name of the town” is added for the service provider to realize that information is required by the user. USER: ‫ما عارفة شي ما عارفة شي حتى المدارس فين هما ال والو‬ [I don’t know anything, I don’t even know where the schools are, or anything] MI: She doesn’t even know where the schools are or the name of the town. [12] Anna Gil-Bardají In the reduced renditions, however, the omission of certain information may be due to conscious or unconscious reasons. In some cases, for example, the MIs deliberately omit a certain idea or a comment made by one of the parties, as illus- trated in the next example. A dismissive and generalizing comment on Moroccan population is omitted. When asked about the omission in the post-task interview, the MI explains that she omitted information to tone down the openly discrimi- natory message of the provider. PROVIDER: And tell her above all to remember to bring everything, let’s see if just for once a Moroccan brings me everything the first time. ،‫تجيبيهم كاملين‬. MI: ‫ قالت لك يعني إإال تقدري‬،‫وهما مع تعودو المغاربة ما كا يجيبو شي الكواﻏط كاملين‬ ‫الكواﻏط كاملين‬ [As they are used to Moroccans not bringing all the papers, they ask if you could bring them all, try and bring them all!] The interview with MI3 shows her awareness as to the omission when she ratio- nalizes her decision: Do you think you Well, the social worker’s way of speaking was sometimes a have had difficult little “harsh.” […] It was also delicate for me, the part when or delicate the provider commented: “let’s see if just for once a moments during Moroccan brings me all the documentation the first time.” I the interpreting? didn’t transmit this fragment to the user as it was said, but I told them that they always make generalizations about Moroccans. […] I toned down the comment about “the child has to come to school clean” because I think it’s racist. I explained it to the user in a more acceptable way … In other cases, the MIs recognize that they neglected to translate certain infor- mation due to, for example, the accumulation of content in one intervention or when they believe the message would be perfectly understood in context, as in this example: USER: ‫ جا من المغرب هادي شهرين وبغيت نسجلو ف للمدرسة‬،‫جيت على ود ولدي‬ [My son arrived from Morocco two months ago and I am looking for a school for him.] MI: Er her son just::: just arrived, two months ago and they want to::: sign him up at::: the school. [OMISSION: “from Morocco”] As MI4 explained in the post-task interview, this was not due to experienced dif- ficulties and can be understood as the MI deciding the information is irrelevant: Ethics, accuracy, and interpreting in social settings [13] Were there any Nothing. I think that both the terminology and the other aspects that you aspects were normal in the the simulations. found particularly But what was hardest for me to remember were the proper difficult in the nouns. For example, I didn’t say the names of the towns three simulations? because I couldn’t remember them, I forgot them. Finally, it is possible to see that the substituted renditions frequently respond to poor understanding on the part of the MI, as shown here: ‫ بعد قرا ف‬USER: ‫ ف المغرب قرا ف الجامع ومن‬.‫ عندو سبع سنين‬.‫االسم هو اسميتو الداودي عادل‬ ‫التحضيري‬ [His name is Dawid Adil. He is 7 years old. In Morocco he went to the Mosque school and then he went to first year of primary school] MI: Er, her son, her son is called Dawdi Adil, er he is 7 years old, and::: well, in Morocco they hadn’t sent him to school yet, only::: he only went to the Mosque school. Some of the MIs rationalize the substituted renditions and errors of interpreting as solutions based on intercultural differences. This extract of the post-task inter- view with MI2 may serve as an example: Do you remember I usually have problems with the social workers. I usually having corrected have problems of terminology. Like, for example how to your interpreting explain the term “psycho-pedagogical” to an illiterate at any point? When woman? I have to explain every time that it is a woman who and how? helps children with difficulties, etc. Translation of the style, tone, and register The duty of the interpreter to reproduce the same style, tone, and register used by the two parties is usually summarized explicitly as “accuracy” in the majority of PSI professional ethical codes. In general, style in oral language is understood as the level of clarity, concision, coherence, simplicity, and naturalness of the speaker; tone is the mode of articulating the voice, reflecting a state of mind, or a certain intention (e.g., anger, irony, or inquiry); and register is the combination of contextual, sociolinguistic, and other variables that condition the mode in which a language is used in context (e.g., formal or informal, direct or indirect). After detailed observation of the MIs’ performance in relation to these three factors, none of the five MIs reproduce them systematically and consciously; rather, in most cases, they opt to either ignore them (as if the whole communi- cation was established outside of the constrictions of professional linguistic stan- dards) or adapt them to the communicative situation. [14] Anna Gil-Bardají In the case of style, on no occasion did the MIs in the sample opt to reproduce the incoherent style of the teacher in simulation 1; rather, they ordered the infor- mation and presented it in a coherent way to the user. In the interviews, they all explained that they had done this consciously, so the information would reach the user in a clear manner. Extract of the post-task interview with MI4 Did you interrupt Yes, I interrupted the provider so that the user could understand. the provider to be The most important thing is that the user is clear about the able to interpret? information and the things they have to do. Sometimes, we ask them at the end “did you understand everything?” and they say yes, and we say to them “repeat it for me please!” We do this with some people in particular. Something similar happens when translating tone. There are practically no occa- sions in which the MIs in the sample reproduce the tone of the user or that of the providers of the service. It is curious to observe that in those cases in which the tone carries a significant semantic charge (for example, uncertainty), the MIs opt to add verbal information that reinforces the expression (for example, adding “but they are not sure”). Where register is concerned, a significant tendency to adapt the register of the providers to the user and vice versa was observed. On some occasions, however, the MIs opt to maintain the high register used by the providers, as MI5 explains in a post-task interview. Extract of a post-task interview with MI5 Language/register. It is very important to know where the user comes from to work We have noticed out how to talk to them. To see if they are from the city or the that you tend to country, their level of education, their level of language, if they are use a high register reserved or more open… we normally do this before the social of spoken Arabic. workers, or on the telephone when calling the user to remind them about the appointment they have with us and we take advantage of the call to work out their profile a bit. Discussion and conclusions As shown, the participants exhibit a significant lack of accuracy, given that only 17.77% of the renditions produced have an exact or almost exact correspondence to the original, while the remaining 82.23% are subject to modification, whether in form or content, including a high average (27.51%) of free interventions (non- renditions). Ethics, accuracy, and interpreting in social settings [15] These results connect to the debate around what degree of intervention is appropriate for the PSI interpreter (Arumí 2017: 24–26). In the nineties, a vision of the interpreter as a strict conduit between the two parties, obliged to stick closely to all aspects of the message, prompted some to advocate for a more “human” and genuine role for the social interpreter. This vision considered the interpreter to be a party involved in the co-construction of meaning, reflecting the language theo- ries of Mikhail Bakhtin (1981), for example. Cecilia Wadensjö was one of the first to explore this issue and, in her 1998 study, to demonstrate empirically that inter- preter performance in social settings should be understood in both textual terms (talk as text) and interactional terms (talk as activity). The current study contributes two more factors for consideration in the debate. The first is the profile of the study subjects themselves, who in this investi- gation (in contrast to that of Wadensjö, among others) are characterized by a dual nature, halfway between two professions – intercultural mediation and PSI – with distinct objectives, codes of conduct, and competencies. Both profiles share cer- tain characteristics: meeting a social need for communication between a linguis- tic minority and the administration; competence in two languages and cultures; knowledge of the area of activity in which they work; belonging to professions being consolidated; and a lack of job security and regulated training. But their main objectives diverge, as shown at the beginning of this article. The second factor for consideration comes from the inevitable reflection on the ethics underlying the communicative transfer, specifically related to the right to truth for both the user and service provider when communicating through an interpreter. According to the vast majority of ethical and deontological PSI princi- ples, the interpreter must translate everything without alterations in form or sub- stance. Yet in practice this reflects the idea of the interpreter as a conduit, leading us to ask ourselves how best to combine interaction (talk as activity) with such an ethical dimension; that is, how to participate interactively in mediated commu- nication without violating core ethical codes regarding the accuracy and faithful- ness of the message. The results of the present study can be interpreted from multiple standpoints: (a) the restrictive models of conduit interpreting and core codes of conduct for PSI; (b) intercultural mediation and its core codes of conduct; or (c) social inter- preting understood in terms of its contrasting facets of equivalence (talk as text) and interaction with its shared construction of meaning (talk as activity). With respect to the unjustifiable renditions from the interactional perspective, the problem appears to reside, more than in normative or regulatory issues, in the non-consensual form of integrating such renditions in the three-way communica- tion or, in other words, in the lack of consensus about what the interpreter should or should not say. For example, interventions such as extra information provided [16] Anna Gil-Bardají to the user by the MIs (cultural or thematic explanations or those about the func- tioning of the service) can result in a lack of transparency and cause errors linked to the delimitation of competences, by not clarifying whether the information is coming from the service provider or the MI. The following extract reflects this process. Extract from the post-task interview with MI4 Introductions. At work I don’t usually introduce myself because in the office they You didn’t give us a form with the user’s details, we call them before the introduce yourself. interview, but we don’t tell them that we are interpreters. We tell Do you normally? them our names and we inform the user that we will be with them. Sometimes, it is the provider who introduces us to the user. Assuring confidentiality and other things will have already been done before starting the session. A simple solution to this situation could be an explanatory marker on the part of the MI for each non-rendition, for example, making a hand gesture to announce that what follows is not a translation but rather their own intervention. This type of code can be decided upon and explained at the beginning of the session, so that the three parties understand and share its meaning. Some unjustified renditions, from an interactional perspective, are the result of limited knowledge of liaison or bilateral interpreting techniques in general, as observed in the almost total absence of note-taking, management of turns, or memory aids and information restructuring, leading the MIs to more frequently request repetition and clarification. Other aspects that were lacking included: (a) sight translation; (b) professional strategies specific to PSI; (c) knowledge and application of professional core codes of ethics; and (d) standards and best prac- tice guidelines for PSI. Knowledge of these PSI techniques on the part of the intercultural mediators would not only improve the quality of their interpreting but also, above all, facil- itate their performance when faced with the frequent problems that arise when interpreting. A lack of familiarity, for example, with the codes of ethics and profes- sional standards of PSI limits the strategic competence of the MIs, despite having to face situations of significant professional complexity (Gil-Bardají 2016: 39). Therefore, specialized training for intercultural mediators would result in greater accuracy in their role as interpreters by providing specific tools to solve semantic and terminological problems, as well as by helping them define profes- sional limits and codes of conduct. This ad hoc training for intercultural medi- ators as interpreters should also focus on competencies, codes of conduct, and professional standards of practice for PSI. Ethics, accuracy, and interpreting in social settings [17] Training for intercultural mediators, however, would benefit from contrasting the results of this study with results obtained from observing the performance of trained PSI interpreters. It would also be beneficial to compare these results with those obtained from an observational study of at least one real-life interaction. Although access to data in these contexts might be difficult, such data would add insight to the current results. References Abril Martí, María Isabel. 2006. La interpretación en los servicios públicos: caracterización como género, contextualización y modelos de formación. Hacia unas bases para el diseño curricular [Public Service Interpreting: Genre features, contextual background, and training models. Toward a basis for course design]. Unpublished Ph.D. dissertation. Universidad de Granada. https://hera.ugr.es/tesisugr/16235320.pdf Abuín, Marta. 2004. “Problemas y estrategias en la interpretación consecutiva: consideraciones metodológicas” [Problems and strategies in consecutive interpreting: Methodological considerations]. Puentes 3: 19–28. Abuín, Marta. 2007. El proceso de interpretación consecutiva. Un estudio del binomio problema/estrategia [The process of consecutive interpreting. A study of the problem/strategy relationship]. Granada: Editorial Comares. Aguilar-Solano, María. 2015. “Non-professional volunteer interpreting as an institutionalized practice in healthcare: A study on interpreters’ personal narratives.” Translation & Interpreting 7 (3): 132–148. Angelelli, Claudia V. 2006. “Validating professional standards and codes: Challenges and opportunities.” Interpreting 8 (2): 175–193. https://doi.org/10.1075/intp.8.2.04ang Angelelli, Claudia V., et al. 2007. “The California standards for healthcare interpreters: Ethical principles, protocols and guidance on role and intervention.” In The Critical Link 4: Professionalisation of Interpreting in the Community, ed. by Cecilia Wadensjö, Birgitta Englund Dimitrova, and Anna-Lena Nilsson, 167–177. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.70.19ang Arumí, Marta. 2012. “Problems and strategies in consecutive interpreting: A pilot study at two different stages of interpreter training.” Meta 57 (3): 118–137. Arumí, Marta. 2017. “The fuzzy boundary between the roles of interpreter and mediator in public services in Catalonia: Analysis of interviews and interpreter-mediated interactions in health and educational settings.” Across Languages and Cultures 18 (2): 195–218. https://doi.org/10.1556/084.2017.18.2.2 Arumí, Marta. 2018. “La interpretación dialógica como práctica estratégica. Análisis de la toma de decisiones de cinco intérpretes en los servicios públicos” [Dialogue interpreting as strategic practice. Analysis of decision making by five interpreters in the public services]. Meta 63 (1): 117–133. Arumí, Marta and Mireia Vargas-Urpi. 2018. “Annotation of interpreters’ conversation management problems and strategies in a corpus of criminal trials in Spain: The case of non-renditions.” Translation and Interpreting Studies 13 (3): 421–441. https://doi.org/10.1075/tis.00023.aru [18] Anna Gil-Bardají ASTM International. 2007. ASTM F 2089-01 (2007), Standard Guide for Language Interpretation Services. West Conshohocken, PA: ASTM International. Baixauli-Olmos, Lluís. 2014. “Redescubriendo los códigos de conducta para la PSI médica: cultura, dilemas y papel profesional” [Rediscovering codes of conduct for medical PSI: Culture, dilemmas, and professional roles]. Panace@: Revista de Medicina, Lenguaje y Traducción 15 (40): 198–217. Baixauli-Olmos, Lluís. 2017. “Ethical codes as tools for change in public service interpreting: Symbolic, social and cultural dimensions.” Journal of Specialised Translation 28: 250–272. Baker, Mona and Carol Maier (eds). 2010. “Ethics and the curriculum: Critical perspectives.” Special issue of The Interpreter and Translator Trainer 5 (1). https://doi.org/10.1080/13556509.2011.10798809 Bakhtin, Mikhail Mikhailovich. 1981. The Dialogic Imagination: Four Essays by M.M. Bakhtin. Trans. by Caryl Emerson and Michael Holquist. Introduction and notes by Michael Holquist. Austin: University of Texas Press. Baraldi, Claudio. 2014. “An interactional perspective on interpreting as mediation.” Lingue Culture Mediazioni 1: 17–36. Baraldi, Claudio and Laura Gavioli (eds). 2007. Dialogue and Culture. Amsterdam: John Benjamins. Baraldi, Claudio and Laura Gavioli. 2014. “Are close renditions the golden standard? Some thoughts on translating accurately in healthcare interpreter-mediated interaction.” The Interpreter and Translator Trainer 8 (3): 336–353. https://doi.org/10.1080/1750399X.2014.972029 Baraldi, Claudio and Laura Gavioli (eds). 2015. Coordinating Participation in Dialogue Interpreting. Amsterdam: John Benjamins. Bermúdez Anderson, Kira, et al. 2002. Mediación intercultural. Una propuesta para la formación [Intercultural mediation. A proposal for training]. Madrid: AEP. Boéri, Julie. 2010. “Emerging narratives of conference interpreters’ training: A case study of ad hoc training in Babels and the Social Forum.” Puentes. Hacia nuevas investigaciones en la mediación intercultural 9: 61–70. CHIA, California Healthcare Interpreting Association. 2002. California Standards for Healthcare Interpreters. Ethical Principles, Protocols and Guidance on Roles and Intervention. Sacramento, CA: California Endowment. De Pedro Ricoy, Raquel. 2010. “Training public service interpreters in the UK: A fine balancing act.” Journal of Specialised Translation 14: 100–120. Desenvolupament Comunitari and Andalucía Acoge. 2009. Mediación intercultural en el ámbito de la salud [Intercultural mediation in health care]. Barcelona: Fundació La Caixa. García-Beyaert, Sofía. 2015. “Interpreting and mediation.” In The Community Interpreter®: An International Textbook, ed. by M. A. Bancroft et al., 371–392. Columbia, MA: Culture and Language Press. Gavioli, Laura and Claudio Baraldi. 2011. “Interpreter-mediated interaction in healthcare and legal settings. Talk organization, context and the achievement of intercultural communication.” Interpreting 13 (2): 205–233. https://doi.org/10.1075/intp.13.2.03gav Gil-Bardají, Anna. 2016. “Interpretar del árabe en los servicios socioeducativos de Cataluña. Resultados de un estudio experimental en torno a las nociones de problema y estrategia” [Interpreting from Arabic in socio-educational settings in Catalonia. The results of an experimental study into the notions of problem and strategy]. In Traducción, interpretación y estudios culturales, ed. by Lucía Molina and Laura Santamaría, 17–41. Granada, Spain: Comares. Ethics, accuracy, and interpreting in social settings [19] Gile, Daniel. 2009. Basic Concepts and Models for Interpreter and Translator Training. Revised edition. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.8 HIN, Healthcare Interpreting Network. 2007. National Standard Guide for Community Interpreting Services. Toronto: Healthcare Interpreting Network. IoLET (IoL Educational Trust). 2017. Diploma of Public Service Interpreting. Handbook for Candidates. London: https://ciol.org.uk/sites/default/files/Handbook-DPSI.pdf. Last accessed 25 October 2019. Kalina, Sylvia. 2015. “Ethical challenges in different interpreting settings.” In MonTI: Monografías de Traduccción e Interpretación. Special issue 2: 63–86. https://doi.org/10.6035/MonTI.2015.ne2.2 Martínez-Gómez, Aída. 2015a. “Bibliometrics as a tool to map uncharted territory: A study on non-professional interpreting.” Perspectives: Studies in Translatology 23 (2): 205–222. https://doi.org/10.1080/0907676X.2015.1010550 Martínez-Gómez, Aída. 2015b. “Invisible, visible or everywhere in between? Perceptions and actual behaviours of non-professional interpreters and interpreting users.” The Interpreters’ Newsletter 20: 175–194. Mayoral Asensio, Roberto. 2012. “El nuevo traductor en Europa tiende a ser cada vez má s versá til y menos enfocado en la traducció n o la interpretació n puras” [The new translator in Europe tends to be increasingly versatile and less focused on pure translating and interpreting]. Revista cTPcba 111: 7–8. Merlini, Raffaela. 2009. “Seeking asylum and seeking identity in a mediated encounter.” Interpreting 11 (1): 57–93. https://doi.org/10.1075/intp.11.1.05mer Merlini, Raffaela and Roberta Favaron. 2003. “Community interpreting: Re-conciliation through power management.” The Interpreters’ Newsletter 12: 205–229. MIRAS. 2011. Comunicar en la diversitat. Intèrprets, traductors i mediators als serveis públics [Communicating in diversity. Interpreters, translators, and mediators in the public services]. Barcelona: Linguamon Casa de les Llengües. Departament de Cultura de la Generalitat de Catalunya. NAATI, National Accreditation Authority for Translators and Interpreters. 2016. The Ethics of Translating and Interpreting. A Guide to Obtaining NAATI Credentials. Canberra, Australia: National Accreditation Authority for Translators and Interpreters. NAJIT, National Association of Judiciary Interpreters and Translators. n.d. Code of Ethics and Professional Responsibilities. https://najit.org/wp-content/uploads/2016/09 /NAJITCodeofEthicsFINAL.pdf. Last accessed 25 October 2019. Napier, Jemina. 2004. “Interpreting omissions: A new perspective.” Interpreting 6 (2): 117–142. https://doi.org/10.1075/intp.6.2.02nap NCIHC, National Council on Interpreting in Health Care. 2005. National Standards of Practice for Interpreters in Health Care. USA: National Council on Interpreting in Health Care. NRPSI, National Register for Public Service Interpreters. 2007. Code of Conduct. United Kingdom: National Register for Public Service Interpreters. Olohan, Maeve. 2014. “Why do you translate? Motivation to volunteer and TED translation.” Translation Studies 7 (1): 17–33. https://doi.org/10.1080/14781700.2013.781952 Pé rez-Gonzá lez, Luis. 2010. “‘Ad-hocracies’ of translation activism in the blogosphere.” In Text and Context: Essays on Translating and Interpreting in Honour of Ian Mason, ed. by Mona Baker, Maeve Olohan, and María Calzada Pé rez, 259–288. Manchester: St. Jerome. [20] Anna Gil-Bardají Pöchhacker, Franz. 2008. “Interpreting as mediation.” In Crossing Borders in Community Interpreting: Definitions and Dilemmas, ed. by Carmen Valero-Garcés and Anne Martin, 9–26. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.76.02poc Pöllabauer, Sonja. 2004. “Interpreting in asylum hearings: Issues of role, responsibility and power.” Interpreting 6 (2): 143–180. https://doi.org/10.1075/intp.6.2.03pol Pym, Anthony (ed). 2001. “The return to ethics.” Special issue of The Translator 7 (2). https://doi.org/10.1080/13556509.2001.10799096 Pym, Anthony. 2012. On Translator Ethics: Principles for Mediation between Cultures. Amsterdam: John Benjamins. https://doi.org/10.1075/btl.104 Setton, Robin. 2015. “Fidelity.” In Routledge Encyclopedia of Interpreting Studies, ed. by Franz Pöchhacker, 161–163. London: Routledge. Susam-Sarajeva, Şebnem and Luis Pérez-González. 2012. “Non-professionals translating and interpreting: Participatory and engaged perspectives.” The Translator 18 (2): 149–165. https://doi.org/10.1080/13556509.2012.10799506 Tiselius, Elisabet. 2015. “Accuracy.” In Routledge Encyclopedia of Interpreting Studies, ed. by Franz Pöchhacker, 3–4. London: Routledge. Ugarte Ballester, Xus and Mireia Vargas-Urpi. 2018. “La interpretación en los servicios públicos en Cataluña y en les Illes Balears” [Public service interpreting in Catalonia and the Balearic Islands]. In Panorama de la traducción y la interpretación en los servicios públicos españoles: una década de cambios, retos y oportunidades, ed. by Ana Isabel Foulquié Rubio, Mireia Vargas-Urpi, and Magdalena Fernández Pérez, 46–64. Granada: Comares. Valero-Garcés, Carmen. 2005. “Terminology and ad hoc interpreters in public services. An empirical study.” Journal of Specialised Translation 3: 75–96. Vargas-Urpi, Mireia. 2013. “ISP y/o mediación intercultural: la realidad de los profesionales que trabajan en el contexto catalán” [PSI and/or intercultural mediation: The reality of the professionals working in the Catalan context]. Cuadernos de ALDEEU 25: 131–164. Vargas-Urpi, Mireia. 2016. “Problems and strategies in public service interpreting as perceived by a sample of Chinese-Catalan/Spanish interpreters.” Perspectives: Studies in Translatology 24 (4): 666–678. https://doi.org/10.1080/0907676X.2015.1069861 Vargas-Urpi, Mireia and Marta Arumí-Ribas. 2014. “Estrategias de interpretación en los servicios públicos en el ámbito educativo: estudio de caso en la combinación chino- catalán” [Interpreting strategies in public services in the educational setting: Case study of the Chinese-Catalan combination]. Intralinea 16. http://intralinea.org/archive/article /estrategias_de_interpretacion_en_los_servicios_publicos_en_el_ambito_edu. Last accessed 25 October 2019. Wadensjö, Cecilia. 1998. Interpreting as Interaction. London: Longman. Address for correspondence Anna Gil-Bardají Facultat de Traducció i Interpretació Universitat Autònoma de Barcelona 08193 Bellaterra Spain anna.gil.bardaji@uab.cat Ethics, accuracy, and interpreting in social settings [21] Biographical notes Anna Gil-Bardají is assistant professor in translation studies at Universitat Autònoma de Barcelona, Spain, and director of the Master of Contemporary Arabic Studies. She lectures Ara- bic language, translation from Arabic into Spanish and Catalan, and Public Service Interpret- ing.