Abstract

The notion of ‘mediation’ has been used in dialogue interpreting research to address those features of the interpreting work which were not easily explainable in terms of translation proper, including adaptation in language and cultural perspectives (Pöchhacker 2008). The word ‘mediation’ however comes from studies on conflict mediation and includes the idea of managing disputes, an activity that is not part of dialogue interpreting. The choice made by some public services, particularly in Italian health care, to employ ‘intercultural mediators’ for their interpreting activities has further contributed to give extra meaning to the notion of mediation in interpreting, now increasingly conveying that mediating means addressing different, and potentially conflictive, cultural features and perspectives. In this paper we get back to the idea of mediation, as intended in conflict mediation, and compare mediation practices found in interactional studies of dispute resolutions (by Garcia 1995; 2019) with practices used by intercultural mediators providing interpreting service in Italian healthcare institutions. We analyse a selection of 400 mediator-interpreted encounters in women’s health with North and West African women speaking respectively Maghrebin Arabic and English as a second language. Our findings suggest that practices used in dispute resolutions and healthcare interpreting do have features in common and that these features do not have to do with emerging conflict (cultural or otherwise), rather with facilitating talk between parties.

Keywords: healthcare, mediation, interaction, conversational practices, conversation analysis, cultural features, talk facilitation

1. Introduction

In Italian public healthcare, dialogue interpreting is provided by so-called intercultural mediators. The employment of intercultural mediators, rather than professional interpreters, is a response to the increasing demand of public services for migrants, consequent to large inward migration flows. The intercultural mediator’s job is defined in a document issued by the Italian National Council of Economy and Labour (CNEL 2009) and aims to enhance positive intercultural relations, favouring migrants’ integration, facilitating their use of public services and improving their participation in social life. While in the 2009 version of the CNEL document, required skills include interpreting competence explicitly, work on mediating intercultural relations clearly remains of primary importance. Despite the document’s focus on the importance of mediating intercultural relations, however, mediators are de facto called mainly, if not exclusively, when migrants do not speak the Italian language thus making interpreting service the main requirement of the mediators’ job (Baraldi and Gavioli 2012; Falbo 2013). In interpreting studies, the Italian choice has aroused considerable debate (see Pöchhacker 2008; Merlini 2009; Pittarello 2009 among others). As noted by Merlini (2009), interpreting does include work on mediating between cultures; moreover, public service interpreting (not mediation) is a well-established profession in many European and non-European countries. Currently however, and in line with the CNEL document, ‘intercultural mediation’ is the preferred choice in Italian public services.

The word ‘mediation’ comes from research on conflict mediation. In Italy, conflict mediation has been connected to ‘intercultural mediation’ in the first wave of studies on interpreting in public services (see for example Ceccatelli Gurrieri, 2003; see Luatti 2011 for a critical overview), a connection which might entail the assumption that conflict is implied when different habits or viewpoints come in contact. Thus, in a first phase, Italian studies on public service encounters with migrants might have considered these encounters as potentially conflictual and in need of (conflict) ‘mediation’. The words ‘mediation’/’mediating’ have also been used in reference to interpreting practices, highlighting the focus on the ‘intercultural’ component of interpreting. In a well-known discussion of mediation in interpreting, Pöchhacker (2008) has suggested that mediation actually accounts for practices of dialogue interpreting, both as a cognitive process, in making sense of the interlocutors’ talk, and as a linguistic-cultural component, when addressing different cultural features and viewpoints. Some recent studies, however, have increasingly regarded the ‘cultural’ component of the mediation concept as a debatable way of looking at dialogue interpreting. In particular, it was argued that treating migrant service users as culturally-specific leads to ‘othering’ migrant minorities (Felberg and Skaaden 2012) and assigns them to categories in which cultural features prevail and their being ‘persons’ is no longer a possibility (Baraldi 2012).

While reflection has thus been dedicated to the cultural issue, much less has been said about the actual significance of ‘mediation’ and ‘mediating’, a concept encompassing any type of modification and adaptation in both the activity of interpreters and in the activity of mediators. To the best of our knowledge, a systematic comparison between the interpreting work of professional interpreters and professional mediators in public services has not been carried out yet. Preliminary attempts to work out differences between the actual practices used in interactions with one or the other professional (Amato and Gavioli 2008; Pittarello 2009) were not further elaborated because of the lack of comparable data, a collection of which would still be strongly needed in interpreting research. Indeed, the interpreting work of mediators shows up as a different one related to what would be expected in ‘interpreting’, giving rise to difficulties in either making sense of which role the mediators are expected to cover in the varied contexts of public service interactions (Leanza 2005) and in defining the mediators’ professionalism (or non-professionalism) as interpreters (Baraldi and Gavioli 2016).

This contribution sets out as a preliminary attempt to discuss the concept of mediation and mediating, dis-associating it from the ‘intercultural’ component that has so strongly affected the debate on mediation in interpreting. We do so, by getting back to the function of mediation and looking at the interactional practices used by conflict mediators to help disputants speak to each other. While conflict is not a characterising feature of talk involving public service interpreting, studies on interpreter-mediated interaction recognise that interpreters coordinate a speech exchange system (Wadensjö 1998) and in particular that interpreters’ renditions may represent one participant’s position in the language of another participant (Wadensjö 2018). Studies of interaction in conflict mediation similarly show practices based on ‘relaying’ talk that facilitate the speech exchange and that similarly pose problems in enhancing the interlocutors’ participation while maintaining neutrality on the part of the mediator in the dispute (Garcia 1995; 2019). Relaying in conflict mediation has been addressed, in Heritage and Clayman (2010: 209), as ‘transformative’ in that it modifies and re-presents one disputant’s position so as to invite a responsive concession from the other disputant.

In this paper, we analyse data recorded in Italian women’s health services, where intercultural mediators provide interpreting in conversations between female patients and clinicians. In our data, mediators sometimes use transformative relaying practices to render providers’ talk by ‘re-presenting’ it not only as to make providers’ contributions more easily understandable or acceptable by the migrant patients, but also to invite a relevant response from the patient as a receiver. In this case, relaying is ‘transformative’ since it invites a response based on personal experience – unlike conflict mediation in which what is pursued is a responsive concession. Our concluding question is whether and how far transformation involved in these mediating practices is in line with the public services’ requirement for mediation and whether these mediating practices are oriented to facilitating talk through interpreting in public services.

2. Conflict mediation and transformative relaying

In conflict mediation, two parties ask for help from a third party to resolve a dispute. While conflict mediation clearly involves some form of conflict, it is non-adversarial and finding solutions is much more important than disputing. The third party, the mediator, is not the one who provides solutions, but participates in talk, so as to help the disputants talk to each other and find their own solutions through talk. Mediators thus participate in the interaction with a coordinating function facilitating the parties’ access to each other’s position or point of view. It is this coordinating and facilitating function of conflict mediation that we are interested in and would like to explore in this paper.

In ‘the speech exchange system of mediation’, as she calls it, Garcia (1995, 2019) shows that the main practice used by mediators is that of ‘representing’ the disputants’ positions. Mediators represent the disputants’ positions in three ways: a. paraphrasing disputants’ talk; b. extending or elaborating on a disputant’s stated position (revoicing); c. replacing the disputant. 

Paraphrasing consists in repeating a disputant's position, by summarising it slightly and making the point on that disputant’s position clear. Paraphrases are never exact repetitions of disputant's statements. Rather, there are small changes oriented to the interlocutor’s response, for instance paraphrases articulate the main point of a story or the main area of disagreement by highlighting common ground. Paraphrases show that the way in which a statement, or an offer, is represented may have an impact on how it will be responded to by the other party. They enable the mediator to serve as an intermediary both showing that disputants’ statements were heard and understood (Garcia 2019: 130) and in offering a paraphrased (thus ‘understandable’) statement to the other party. The mediating function of paraphrase thus mainly consists in reinterpreting one disputant’s statement as plausible, thus hedging the possibility that the respondent rejects it.

Extending or elaborating a disputant's stated position (Garcia 1995), also referred to as revoicing in Garcia’s later work (2019), includes instances where mediators elaborate or extend a position taken by that disputant in a previous utterance, involving the opposing disputant in direct exchange (Garcia 2019: 132), so that talk is not directed to both interlocutors, but to one of them mainly (the opponent). Revoicing differs from paraphrasing because the mediator goes beyond what was actually said by the disputant by explaining the disputant’s position and is based on conversational moves that facilitate the speech exchange by modifying and re-presenting one disputant’s concessionary position to invite a responsive concession from the other disputant, as Heritage and Clayman (2010) put it. In this sense, revoicing makes ‘changes that might bridge the gap between the two disputants and help move toward agreement’ (Garcia 2019: 133).

The third way of representing the disputants’ positions mentioned by Garcia (1995; 2019) is called replacing. When replacing one disputant, the mediators do not restrict themselves to re-presenting the disputant's expressed positions; rather, they go beyond what the disputant said and argue in place of the disputant. The mediator thus directly engages in negotiations with the opposing disputant, creating his or her own arguments, and acting as a principal, rather than re-presenting what the disputant said. Garcia notes that while paraphrasing and revoicing have positive consequences on conflict mediation because they facilitate the disputants’ understanding of each other position and autonomous elaboration of their conflict, the replacing modality puts one of the participants under strong pressure, diminishing the disputants’ autonomy and mediator’s impartiality or neutrality (Garcia 2019: 141, 148) and increases the disputants’ reluctance and reservations about possible agreements (Garcia 1995: 38). Replacing, according to Garcia, reduces the possibility of successful mediation, of disputants’ satisfaction and also compliance with agreements.

Garcia’s discussion of conflict mediation shows, in our view, two main points of interest concerning mediation through interpreting: the first is that replacing actions obstacle the disputants’ participation in the interaction in that one of them is substituted by the mediator. So while replacing is a form of conflict mediation it cannot be considered a form of mediation facilitating talk between the interlocutors and is seemingly not effective even in dispute resolutions. Work in interpreting studies too has warned against the risk of interpreters replacing one of the interlocutors, depriving these interlocutors of their possibility to participate in talk (see for instance Mason 2009). The second, definitely more interesting point, is the mediating function triggered by both paraphrasing and revoicing. Paraphrasing and revoicing show that interpreted (by the mediator) repetition of disputants’ positions and stories makes responses relevant by one or both interlocutors making their positions increasingly clear to each other. This interactional function of paraphrasing and revoicing in conflict mediation is referred to in Heritage and Clayman (2010: 210-11) as transformative relaying.

Transformative relaying refers to transformation in two ways. The first are the changes that mediators produce in their relaying one participant’s contribution. The second, possibly less obvious but even more interesting, is the transformation that relaying produces towards the respondent, enhancing the respondent’s understanding of the opponent’s position and the relevance for the respondent to respond. Paraphrasing and revoicing do not ‘transform’ the disputants’ positions, but the interaction proper, by assisting the disputants in gradually finding a solution by themselves.

3. Mediating in dispute resolution and in PSI

To the best of our knowledge, there are no studies giving evidence that conflict characterizes PSI, particularly in healthcare services. According to Garcia (2019), on the other hand, conflict is not a characterising feature of dispute resolutions either. So mediation does not have to do with conflict mitigation, not even in dispute resolutions. Rather, mediation in dispute resolutions works by relaying one participant’s contribution both to make that contribution accessible and relevant for the interlocutor’s purposes and to invite an interlocutor’s relevant response. This is what mediation does in PSI too. The delicate point, in both cases, is how to mediate without pushing the interlocutors’ positions, an action which might, in both cases, be detrimental for the success of the interaction – finding a convincing solution or getting adequate service. In this sense, it seems that inquiring into how mediation is carried out in dispute resolutions may throw light on the significance of mediation practices in PSI too.

There is of course a difference between addressing disputes in monolingual conflict mediation and addressing understanding in bilingual PSI. While both conflict mediation and PSI aim to enhance interlocutors’ active participation, the difference is between a transformation concerning the participants’ management of conflicts and a transformation concerning the participants’ possibility to respond to other participants’ utterances when language is not shared.

Garcia (2019) shows that conflict mediators mainly (and more successfully) relay by paraphrasing and by revoicing. In both cases they add significance to the content of the previous turn by inserting or highlighting points of convergence. The difference between paraphrasing and revoicing consists in the degree of tentativeness used in relaying, as paraphrasing is proposed to both interlocutors, with a possibility by the speaker who produced the content first to confirm or adjust the new formulation, while revoicing takes it for granted that the content-producer agrees on the new formulation and what is instead pursued is a reaction by the other.

Our analysis shows that both paraphrasing and revoicing are used also by the intercultural mediators in our data (and in other data too, see Raymond 2014), and significance is added transforming content from one language to another. Since PSI involves two languages, however, participants do not have full control on how their contributions are relayed in the other language, thus making their possibility of amending less likely. What distinguishes paraphrasing from revoicing in PSI is thus that while transformation in paraphrasing only affects the participant’s contribution content, by for instance making explicit what is implicit, revoicing also affects the turn-taking system by projecting a turn-shift from one to the other interlocutor, inviting the latter to provide more or even new pieces of information. Revoicing then not only interprets what is said by one interlocutor, but invites replies by the other interlocutor, giving them floor to talk.

As for the third type of practice identified in Garcia’s studies, replacing, intercultural mediators in our data use replacing practices and in fact substitute one interlocutor. Space is not enough, in this paper, to provide an adequate account of mediators’ replacing practices, but we will focus on some cases in which the mediators say what might be expected by the clinician. We note, however, that replacing may be a blurred concept in PSI since sometimes intercultural mediators provide service whose responsibility is very hard to attribute and might fall in ‘no one’s responsibility’. This finding suggests that there are areas of communication in healthcare for migrants that still need exploration and recognition.

4. Data and methods

In the following sections, we analyse data recorded in public healthcare services in two provinces of Northern Italy. All the data are audio-data from a large corpus of interpreted interactions collected over the last 20 years and including a variety of settings and languages (the AIM corpus, see Corradini et al. 2024). Interpreting in the healthcare services represented in the corpus is provided by intercultural mediators. Although intercultural mediators may have received no or little training in interpreting techniques (Merlini 2009; Pittarello 2009), within the category of mediators, training, skill and expertise may vary a lot. The healthcare services we collaborated with are among the most advanced in Italy as for the organization of mediators’ services, permanently monitoring and training their staff (see Chiarenza 2020).

The data examined in this paper are a selection of 400 encounters collected in women’s health services – obstetrics and gynaecology, transcribed using the Jeffersonian transcript method recently elaborated in Hepburn and Bolden (2017). While the data are audio, one researcher was present during data-recording, providing notes for frequent sounds like typing, printing, and reading or handing out papers. The data selection includes only two languages besides Italian: English and Arabic. English is spoken as a second language in the Ghanaian, Nigerian and, to a smaller extent, Indian variety; Arabic is spoken in the Maghrebin variety of Morocco and Tunisia and transliterated to allow for an adequate representation of overlaps^[1]. All mediators have themselves a migration history.

The approach we use is derived from Conversation Analysis, particularly applied to talk in institutional settings (Antaki 2011; Drew and Heritage 1992; Heritage and Clayman 2010) and is based on studies in interpreter-mediated interaction as elaborated in Wadensjö (1998) and Mason (2006) as well as in our own work (Baraldi and Gavioli 2012, 2016). More specifically, in this paper, we refer to the studies carried out by Garcia (2019) in conflict mediation, and we look at paraphrasing, revoicing and replacing as mediating features of the interpreting work of intercultural mediators. We argue that the extra-supplement of mediation in mediator-interpreted talk has very little to do with cultural management. Rather, the transformation that is visible in intercultural mediators’ work is oriented to a welcoming, hospitable idea of opening migrants’ access to services. Whether this transformation is facilitative or patronising or, in other words, whether, in so doing, intercultural mediators are access-openers or gatekeepers of the clinician agenda is a reflection that should probably be taken in serious consideration in both interpreting research and in studies on public service communication.

5. Paraphrasing and revoicing in mediator-interpreted interaction

An analysis of our selected corpus shows that intercultural mediators render talk by using paraphrasing and/or revoicing quite frequently. For the sake of our argument, we will focus only on renditions of talk from the clinician to the patient, since these two participants are those who represent the public institution. An analysis of renditions of talk from the patient to the clinician is, however, likewise possible and similarly relevant.

5.1 Paraphrasing

As mentioned in section 2 above, paraphrasing consists in repeating the contribution of one of the interlocutors by making the point on what is considered relevant in that contribution. Similarly to conflict mediators, intercultural mediators paraphrase clinicians’ talk by summarising, but more frequently their renditions contain forms of clarification either of single items mentioned in the clinician’s turn or highlighting what seems to be the main point and purpose in the clinician’s contribution. Let us look at two extracts. Extract 1 shows paraphrasing of a lexical item, while extract 2 shows paraphrasing by stressing the main point in the clinician’s contribution.

In extract 1, the clinician asks the patient, a refugee woman from Ghana, if she has a card for temporary residents, enabling her to be given extra health service, not just emergency care. The card is abbreviated STP, standing for ‘Straniero Temporaneamente Presente’ (temporary present foreigner), an abbreviation that is used by the clinician in her request in turn 71. As can be seen in the extract, the mediator starts rendering in turn 74 in overlap with the clinician and paraphrases the item mentioned by the clinician twice, in turn 77 as ‘hospital card’ and in turn 81 as ‘the temporary white one’.

Extract 1 (DOC: gynaecologist; MED: mediator; PAT: patient)^[2]

The mediator’s paraphrasing shows that using the abbreviation STP might have an impact on the patient’s response possibly leading her to ask for clarification or even non responding. The mediator thus serves as an intermediary interpreting the clinician’s talk as technical, but ‘paraphrasable’ for the other party into a more easily recognizable item: a white card given by the hospital. The patient shows her recognition of the item asked about by providing a no-answer twice: in turn 78, immediately after the first paraphrasing, and another one following the mediator’s second paraphrasing in turn 81. The no-answer (rendered to the clinician in turn 79) is easily provided by the patient and the floor is then open to either the clinician or the patient to take the next turn. After a 5-second pause, the clinician will take the turn with a new question (data not shown).

In Extract 2, the clinician provides two pieces of information: the first is that the patient will take blood tests in June, the second is that she has to go personally to collect her HIV test record of a blood test previously taken. The HIV record in Italy is a legally protected document that is not available to the clinician through the hospital intranet so, unless the patients are hospitalised, clinicians have no access to HIV records, only the patients. The clinician’s turn lasts 30 seconds overall (feedback and pauses included) a stretch of time usually manageable in interpreting. Within these 30 seconds, the clinician marks the second part of her contribution (opening turn 39) as the crucial one: ‘What is important’. Let us see the clinician’s contribution in Extract 2a: first part up to turn 37 prescribing next blood test and second part (turns 39- 44) recommending to collect the HIV record:

Extract 2a (OBS: midwife; MED: mediator)

The mediator provides ‘mhm’-feedback allowing the clinician to go on talking, then after her last ‘mmh’ in turn 45 and a following 0.7 pause, she starts rendering. Her paraphrasing inverts the textual order of the two parts of the clinician’s contribution starting with the second piece of information, the HIV one, presented by the clinician as the most important. Extract 2b shows the mediator’s paraphrasing of the clinician’s recommendation to collect the record: the specific test is introduced first, then the importance to collect the test record, then the necessity to collect it personally and finally the reason why the patient has to personally go to the desk.

Extract 2b (MED: mediator; PAT: patient)

The mediator’s paraphrasing is clearly not an exact repetition of the clinician’s contribution and not only in her switching the two pieces of information. There are, in the mediator’s version, small changes oriented to showing the patient that the report is needed and that she can help by collecting it ‘on her own’, personally, or otherwise it will not be possible for the clinician to read it. In this case, the mediator serves as an intermediary in both showing that she understood the main point of the clinician’s request and in offering a compliable invitation to the patient. It is interesting to note that the mediator does not push the patient to comply with the clinician’s request, at least not as much as the clinician herself does; rather she clearly gives the patient reasons why the record is not available unless she personally goes and gets it.

The patient produces continuation feedback during the mediator’s paraphrasing and shows she got the point by proposing a completion of the mediator’s contribution in turn 54: ‘they don’t give out’. The mediator accepts the patient’s conclusion partially and completes her explanation in turn 56. After the mediator’s conclusion, there is a pause in which none of the parties take the turn and the mediator will then go on and render the first piece of information given by the clinician.

The extracts showcase two rather different examples of paraphrasing, one in which what is paraphrased is a technical concept rendered in two different versions oriented to an easier recognition of which card the clinician is seeking; the other is a more elaborated, textually re-organized paraphrase in which the main issue highlighted by the clinician is further highlighted and explained by the mediator. What the two have in common is: a. that the patient receives considerable help in identifying both the object and the objective of the clinician’s request; b. that they make the understanding reaction of the patient visible; c. that nothing else is visible besides the patient’s confirmation of understanding, thus making it clear to the other interlocutor, the clinician, that her contribution has been passed over and received. So the floor gets open to either of the two parties to initiate a new series of actions.

It may of course be argued that, differently from monolingual conflict mediation, the clinicians here do not have control on how their contributions are paraphrased thus making it very unlikely for them to provide amendments or adjustments. The mediators’ paraphrase however is represented as a ‘what the clinician wants to know’ making the clinicians’ statements or requests respondable, and responded, by the patient. Mediators’ relaying through paraphrasing modifies the clinicians’ contributions on a mainly textual basis, transforming them from a not easily respondable, or possibly not respondable at all, request or statement into a respondable (and responded) one. The response normally closes the current sequence and leaves the floor open for a new one.

5.2 Revoicing

Garcia’s conceptualization refers to revoicing as relaying one party’s contribution explicitly addressing it to the other one. By explicitly addressing their relaying to one party, the mediators take it for granted that the relayed party agrees with the mediator’s version of that party’s position, while the opponent is the invited responder. In interaction interpreted by mediators, revoicing goes beyond paraphrasing because the mediator’s version does not only modify the original version textually, but there is something extra added by the mediator that is explicitly addressed to the other interlocutor and elicits that interlocutor’s response. By revoicing, the mediator ‘gives voice’ to what s/he has interpreted to be the relayed speaker’s orientation and collaborates in getting additional relevant information from the interlocutor. Examples of revoicing in mediators’ rendition of clinicians’ contributions for the patients are visible in rendering clinicians’ questioning and explanations, recommendations or reassurance. Let us look at two examples revoicing questions.

Extract 3 shows a sequence with a pregnant woman, speaking Arabic. The problem is again that of collecting test records from the hospital desk. In turn 1, the midwife inquiries about the records producing a statement probably deriving from the data available to her: ‘they haven’t gone to take them’. In the mediators’ rendition, the clinician’s statement is paraphrased into a question (turn 2) in which the mentioned records are made explicit (‘those that were available from 15^th March’) and the statement of the clinician reported (‘you haven’t gone to take the record’). The mediator’s rendition, however, also includes something extra, a hypothesis about a plausible reason why the couple might not have gone to collect the record. The plausible reason (‘your husband was at work’) is a mediator’s addition to what the clinician said and is explicitly addressed to the patient, inviting her to confirm or to provide another reason for not going. The patient’s answer in turn 4 in fact shows that the patient has taken up the mediator’s invitation.

Extract 3 (OBS: midwife; MED: mediator; PAT: patient)

In turn 131, the patient mentions her daughter; turn 132 shows the mediator’s interpretation of the patient’s contribution as the invited, plausible reason for not collecting the record: ‘kanit marida’, ‘was she sick?’. This time, the mediator’s provision of the reason for not going is confirmed by the patient (turn 133) and rendered by the mediator to the clinician in Italian (turn 134). Turn 135 shows the clinician’s acceptance and acknowledgment.

The mediator’s rendition of the clinician’s question is thus paraphrased by re-shaping it into an explicit question form and by mentioning which test records the clinician is referring to. Differently from the previous cases of paraphrasing that we have seen in section 4.1 however, it also includes a supplement, revoicing the question by suggesting that there might be good reasons why the patient has not (yet) complied with the clinician’s requirement. In her revoicing, the mediator renders the same content, but transforms the embedding interaction: on the one hand she hedges the possibility that pressure is put on the patient for her non-compliance, and, on the other, she invites the patient to provide the good-reason-she-surely-has, thus reassuring the clinician that the test record will eventually be there.

Let us pass to the next extract. Here the clinician, in turn 41, addresses the patient by using her first name Betty (fabricated in the transcript). The clinician asks two questions: where the patient works and if she works. The mediator in turn 42 renders the second question first (if the patient works), then almost in overlap with the clinician’s repetition of her own first question she renders ‘where?’. It should probably be noted that, in Italian, ‘dove lavori’ (where do you work?) is often used to mean ‘what do you do’, which is probably what the clinician aims to get since the type of work may have an impact on the patients’ health. In turn 45 the patient answers the ‘where’-question and provides the name of a small town (fabricated in the transcript) in the nearby of the hospital main town, which is acknowledged by the clinician. The mediator’s revoicing starts in turn 47 in which she adds an extra that is explicitly addressed to the patient and was not in the clinician’s question: ‘you don’t work on the road’. The patient in fact works on the road, an information she partially gives in turn 48 and that gets completed by the mediator in turn 49

Extract 4 (OBS: midwife; MED: mediator; PAT: patient)

Let us first consider the mediator’s revoicing and then the rest of the sequence. The rendition choices of the mediator in turns 42 and 44 show that she has interpreted the second clinician’s question as prior to the first one: if the patient does not work, she cannot say where she works. In turn 44 the mediator might have paraphrased the clinician’s first question into ‘what do you do’, but her rendition maintains the clinician’s question form, allowing the patient to provide a broad range of options, from a general area, like a town, to more specific locations like names of companies, institutions, shops or associations. The patient provides the name of a town, which might suggest the mediator that the place mentioned is an area of prostitution, or that the patient hesitates to tell about her job. By adding a negatively-framed question (see Heritage 2009) ‘you don’t work on the road’, the mediator explicitly invites the patient to confirm or disconfirm and shows her orientation to get the piece of information the clinician is presumably seeking: the patient’s job. The mediator and the clinician both provide understanding feedback, in overlap (turns 49 and 50) and the clinician’s next question in turn 51, relating to the patient’s safety on work, shows both that the clinician understood what was said by the patient and the mediator in English and that she considers the detail relevant for her course of actions.

In both extract 3 and 4, the mediator’s revoicing is produced by adding items to the clinician’s questions, which specifically invite the patient to provide more information, a plausible reason for not going to collect a test record or the job the patient currently has. By revoicing, in these cases, the mediator interprets that a supplement of information is involved in the clinician’s question and proceeds autonomously to elicit it from the patient. The mediator’s interpreting in this case works both on the orientation of the clinician towards getting a relevant piece of information and of the patient in providing what is required. Relaying is thus transformative in producing a relevant context for that (missing) piece of information to be produced.

In the examples seen in this section, then, revoicing provides evidence of how the mediator interpreted the orientation of the clinician to the patient. The mediator’s rendition in these cases is not just addressed to the patient as this-is-what-the-clinician-said, but elicits a patient’s specific response, orienting the patient to what is relevant to say (and do) next. The mediator’s relaying, giving voice to what the mediator interprets as relevant missing items, make these items explicit in talk and offered to be specifically addressed in the patients’ response. The interaction thus transforms the interlocutors’ participation by actively re-orienting it to what was implicit before: the relevance of providing a good reason for not collecting test reports or of saying which job the patient has.

6. Replacing

As mentioned in previous sections of this work, Garcia’s discussion includes three ways of mediating, replacing being the third after paraphrasing and revoicing. She notes that the replacing mode is not functional in conflict mediation in that replacing one of the interlocutors in the interaction means in fact preventing that disputant to participate autonomously. Interpreting studies have similarly warned against the risk that the interpreter replaces one of the interlocutors, who would be left aside and deprived of the possibility to participate – even more so in an interaction in which the interlocutors’ language is not shared (see for example Davidson 2000; Hsieh 2007, 2016).

In our data, cases of replacing basically confirm the literature’s findings. When the mediator substitutes one interlocutor, the clinician in our case, that interlocutor is deprived of the possibility of participating. This is not only a problem per se, but also leads to losing control of the interaction, making it difficult to re-invite the ‘left-out’ participant to participate again. There are however cases in which the mediator participates in the place of the clinician, without causing apparent trouble in the interaction, which smoothly reprises in its triadic form after the mediator’s contribution. The greatest part of these cases can probably be divided into two types: ‘direction-giving’ basically offering instructions about how to reach a surgery, a desk, a chemist’s and the like, or ‘clinician’s mandate’ that is to say explicit invitations by the clinician to the mediator to act in the clinician’s place. Below, we will see an example of each of these cases with details showing that the clinician is not excluded and indeed participates in the course of actions. There are however also more nuanced cases in which the mediator replaces the clinicians spontaneously, possibly treating the clinician’s contributions as mandates. The third example in this section will illustrate one such case.

Let us start from Extract 5 below showing an example of direction-giving. Here the patient has complained of white vaginal discharges, which seem to her unusual. Since it is the second time the patient comes and complains of the same problem, the clinician decides to prescribe a vaginal swab. The clinician’s advice is given in turns 412-414 and the mediator starts rendering in turn 415, while showing the patient the written prescription (as can be understood from the sound of paper in the recording, from what the clinician said a few turns before the extract and from our field notes). The patient starts responding in turn 416 (‘okay’) and then asks two questions, both of which are not rendered and are instead answered by the mediator.

Extract 5 (DOC: gynaecologist; MED: mediator; PAT: patient)

Let us look at the mediator’s answers. The first takes just one turn (turn 418) and is a response to whether the patient will keep the paper with her; the second, answering the patient’s question in turn 419 (‘where is the place’) is a full provision of information about where to go, starting in turn 420 and ending with patient’s ‘okay’ in turn 429. While the mediator is clearly replacing the clinician in giving these instructions, the clinician is not excluded and in fact she collaborates in the instruction-giving sequence, by handing in more written information: in turn 423 she passes a paper to the mediator (‘to’ to’ to’’, meaning ‘do take this’) and in turn 424 she points to the laboratory address (‘qua’). The mediator then renders the meaning of the clinician’s gesture, which is acknowledged by the patient in turn 430. In what follows the mediator will render the reason for taking a vaginal swab, which was left untranslated before and the clinician takes her turn immediately after the mediator, to greet the patient.

In our data, we have several direction-giving sequences in which it is the mediator, not the clinician, who provides the details. None of these cases show conversational trouble, like pauses, overlaps or hesitations, rather the clinician either collaborates with the mediator in giving the directions, as in Extract 6, or they just take the turn immediately after the direction-sequence continuing their course of actions, for instance opening a new sequence or closing the encounter. It can be noted that direction-giving sequences are rather transparent from a language point of view, in that the mentioned places and addresses are Italian institutional places which the clinician is most familiar with. Possibly for this reason, it seems that mediators replacing clinicians in giving direction is a rather standardised practice in mediator-interpreted interaction in our data.

Extract 6 shows what we have called clinician’s mandate. Clinician’s mandates are a complex category because there are many ways in which clinicians project expectations for the meditators to do ‘more than render’, for instance expanding explanations or insisting on recommendations, as shown in Gavioli (2015). Clinician’s explicit mandates are however not very frequent in our data and normally initiate a standard series of history taking questions produced by the mediator. In Extract 6, the clinician asks the mediator whether the patient said something more than what the mediator rendered previously and the mediator in turn 699 (first turn in the extract) confirms that there is nothing more to render. In turn 700, the clinician acknowledges the mediator’s response and in turn 701 asks her to investigate further (‘re indaga’). The mediator in turn 702 produces a couple of history-taking questions, the first (turn 702) checking that there is nothing else excluding the thyroid disease that the patient mentioned before, the second (turn 706) is a really replacing question, possibly formulated on the basis of the mediator’s experience in the service, in which the mediator acts in the place of the clinician. As can be seen the clinician acknowledges the mediator’s rendition of the patient’s answers in turn 708 and briefly passes on to complete the patient’s record including only the thyroid detail, nothing else (‘punto’/’full-stop’, in turn 709, meaning ‘that’s all’)

Extract 6 (DOC: gynaecologist; MED: mediator; PAT: patient)

As can be seen in Extract 6, the mediator’s investigation is largely a repetition either of what was previously asked (‘anything else excluding thyroidism?’) or of typical history-taking questions like ‘any operations?’. The patient’s no-answers are immediately rendered to the clinician and followed by her confirmation of understanding and acknowledgment. The mediator-patient exchange lasts 16 seconds, making it very unlikely for the clinician to be deprived of her possibility to participate. What is more is that the mediator’s replacing in this case is ‘governed’ by the clinician, who never loses control of the interaction, asks explicitly to be replaced and takes her place back immediately after obtaining the answer she wants.

Unlike the case of direction-giving, however, in the case of clinician’s mandate, the meditator replaces the clinician not only by becoming the main interlocutor of the patient but also by expressing content and competencies that are ‘the clinician’s’. While on the one hand, mandates in all cases regard highly repetitive procedures, like typical questions in history-taking or typical recommendations regarding for example personal hygiene or diet, on the other the mediator is not only the one who asks the questions or gives recommendations, but also the one who receives the patient’s responses first and needs to react to them promptly. It is probably in the management of this reaction that the risk of excluding the clinician is higher. In the case in question, the mediator collects the answers from the patient and renders them immediately back to the clinician, thus minimising such risk.

Let us now look at our last example of replacing. Here we have neither an explicit or implicit request for directions, nor an explicit mandate. The midwife’s contribution in turn 254 arrives after her consideration of giving the patient glucose screening, which is normally taken at week 28 in pregnancy. Her comment (‘she’s too late’) makes reference to the patient being at week 31 in her pregnancy, so too late to have a regular glucose screening. The mediator renders the clinician’s comment first by paraphrasing it in turn 255 and then engaging in a thorough explanation of the reason why the patient is not given the screening.

Extract 7 (OBS: midwife; MED: mediator; PAT: patient)

The mediator’s explanation makes sense of the clinician’s comment by clarifying why it is too late to have glucose screening at the patient’s stage of pregnancy and the details she gives are well-known to her, a long-time employee at the service. The patient and her husband show understanding and appreciation, and ask no questions. The exchange of the mediator with the patient and her husband lasts 18 seconds, so, again, a time short enough to allow the clinician to keep control of the interaction and re-enter it in turn 264 to eventually prescribe more tests. So, there is actually no evidence of disfluency or trouble in this sequence, which is opened and quickly closed by the mediator as if she had rendered a longer stretch of talk. In fact, the explanation is helpful for the clinician who, in the meanwhile, is filling the patient’s record and typing the prescriptions.

Replacing in our data is thus not so uncommon. For some actions, falling neither in the clinicians’ or the mediators’ most direct tasks, like giving directions, replacing is actually the standard practice. The types of replacing shown here do not cause interactional trouble and do not exclude the clinician from the interaction, rather in the case of mandates, the clinician keeps high control on what is going on and re-intervenes promptly when she deems relevant. In a way, replacing optimises the interaction, possibly reducing the time of the encounter.

As in dispute resolutions, replacing has however some consequences on the parties’ relationships that may deserve attention. The first is that there is no relaying. This means that there is no mediator’s work on connecting the two parties, so no possibility of transforming their non-talk with each other into talk. While replacing may supply relevant and helpful details without really excluding one of the parties and making it clear that what is said is in agreement with that party, it is not ‘relaying’ proper, not the voice of that party transformed into something acceptable and familiar. As seen in the examples above, moreover, replacing occurs within ‘helpful sequences’, filling the gap of a ‘no-one territory’ like giving directions, or of a clinician doing something for the patient in parallel, or explaining procedures that were left unexplained for some reason. These are all beneficial actions making the interaction possibly more welcoming for the patient, but they deprive the clinician from the opportunity of enacting them and being the one who is helpful and welcoming. So while replacing practices in interpreted interaction may not be as disruptive as in dispute resolutions, they may suggest that mediators are filling gaps in what is not, after all, such a migrant-friendly service. By filling these gaps, mediators do not relay on an existing gap by making it visible, rather they cover it, thus preventing clinicians from the trouble of developing their own welcoming practices.

7. Conclusions

In this paper we have shown that conversational practices of mediation found in dispute resolution (Garcia 2019) can be found also in interpreting provided by intercultural mediators in Italian healthcare services. Relaying practices in particular, like paraphrasing and revoicing, seem to account for the types of changes that intercultural mediators carry out when rendering, explicating clinicians’ contents and interactional expectations for patients’ responses.

This finding casts a rather new light on the idea of ‘intercultural mediation’. While ‘cultural’ issues like wearing a veil, eating Nigerian food or differences in ways of calculating time and age are sometimes visible in our data, and explicated when necessary, they account for no more than five per cent of the 400 total encounters. There is thus very little ‘culture’ in the process of mediation as it appears analysing our occurrences of mediator-interpreted encounters. The bulk of the ‘mediating’ work of mediators in interpreted interaction, instead, consists in making clinicians’ contents explicit and easily recognisable for the patient and in interpreting the purposes of clinicians’ contributions, inviting the patients to provide relevant responses.

Similarly to what occurs in dispute resolutions, then, relaying through paraphrasing and revoicing transforms the interaction by showing a ‘more accessible’ version of the clinicians’ contributions and legitimising ‘non-knowledge’ on the part of the patients. An interesting difference between paraphrasing and revoicing in mediator-interpreted interaction is their orientation towards the interlocutors, the patients in our case. While paraphrasing slightly modifies the clinicians’ utterances, it makes the understanding reaction of patients visible, also making it clear to the clinicians that their contribution has been translated and received. Revoicing affects the modification of clinicians’ utterances more extensively, not only making implicit contents of these utterances explicit, but also adding content triggering patients’ responses. Revoicing has a clear impact on turn-taking in actively orienting the patients to what is relevant to say and do next.

As for replacing, we have seen that in our PSI encounters, mediators take initiative in supplying ‘further help’, for instance by giving directions, or are explicitly asked to take the clinicians’ place temporarily, by asking routine questions or ensuring that the patient has nothing else to add. We have observed that replacing practices do not hinder and indeed might improve efficiency in providing service, a finding that is suggested also in Mason (2009). Similarly to what Mason (2009) noted however, replacing does not favour the relationship between the clinician and the patient, the mediator taking the place of the ‘welcoming’, kind participant in the service encounter, thus depriving clinicians from the opportunity of providing autonomous support to migrant’s participation. There thus seems to be a paradoxical relation between mediators’ support to patients’ participation and their relaying between clinicians and patients: the more the latter is hindered, the more patients’ participation is supported. A crucial open question for the healthcare services would thus concern the extent to which they may find it useful or necessary to separate transformative relaying involving the clinician in patients’ supportive, welcoming service from access to services. Mediating activities like replacing can indeed favour patients’ participation and facilitate it, but clinicians may, after all, be prevented from being the facilitators. This, we believe, may be a fruitful point of reflection for services’ achievement of friendliness for migrants.

About the author(s)

Laura Gavioli is Professor of English language and translation at the University of Modena and Reggio Emilia, Italy, where she teaches in BA, MA and PhD courses. Her research work includes the study of spoken language in institutional settings and cross-cultural pragmatics in intercultural/multicultural settings. She has been engaged in research exploring authentic data of interpreter-mediated conversations involving speakers (not necessarily native-speakers) of English and Italian, mainly in healthcare services. Together with Claudio Baraldi, she edited the volume Coordinating Participation in Dialogue Interpreting (Benjamins, 2012) and a special issue of Health Communication (36/9 2021). With Cecilia Wadensjö, she has co-edited the Routledge Handbook of Public Service Interpreting published in 2023.

Claudio Baraldi is Professor of Sociology at the University of Modena and Reggio Emilia, Italy. Together with Laura Gavioli, he carried out lengthy exploration of intercultural and interlinguistic mediation in institutional settings -- mainly healthcare and more recently parent-teacher talk in educational settings. Another important research interest concerns the analysis of methods and techniques for dialogic facilitation of participation in education and health care. He has written several papers in international books and journals, he has published volumes for Palgrave and Springer and he has edited and co-edited international volumes for Bloomsbury, John Benjamins, Palgrave, Routledge and Sage.

Email: [please login or register to view author's email address]

•  Previous article
• Next article