Abstract

Straniero Sergio was the first scholar who looked at the functions that repetitions deployed by interpreters can have in interpreter-mediated interactions, specifically in the context of interviews during TV talk shows. Taking inspiration from his work, the present paper examines repetitions deployed by intercultural mediators working with migrants in the public sector in interactions where English is used as a lingua franca.  By analysing a dataset comprising interactions from the healthcare and educational settings and using a methodology grounded in Conversation Analysis, repetitions have been classified and investigated in order to understand their functions in interaction and their implications for public service interpreting. Results show that repetitions both of other speaker’s words by the mediators and of the mediators’ own words can have the function of eliciting the migrant interlocutor’s confirmation of the information being gathered by the mediator. Repetitions are also shown to be deployed particularly when migrant speakers show low levels of language proficiency while speaking English as a lingua franca. In this context, mutual comprehension is often potentially at stake, and repetitions are shown to function as an effective way of double-checking migrant service users’ understanding, since they do not introduce new linguistic complexities in the exchange. Overall, the aim of this paper is to show how a conversational mechanism such as repetitions can also be used effectively to achieve interactional success in mediator-interpreted exchanges in the public sector.

Keywords: intercultural mediation, conversation analysis, institutional interaction, repetitions, migrants

1. Introduction

Of the many innovative scholarly contributions that Francesco Straniero Sergio made to research on interpreting, one which, to the best of my knowledge, was unique when it was first published in 2012, is relative to the functions and interactional implications of repetitions in dialogue interpreting, within the context of TV talk shows. In interviews during talk shows, when a professional interpreter is involved, interpreting can often take place both in the whispering and consecutive modes: the TV host’s utterances are translated via whispered interpreting for the TV guest, while the TV guest’s answers or comments are rendered consecutively. This allows the audience to listen to the TV guest’s own voice and native language, and to have a direct experience of their non-verbal language. By scrutinising the strategies used by interpreters in this context, Straniero Sergio (2012) showed that interpreters may repeat some lexical elements uttered by the TV host, when rendering the TV guest’s words, so as to create lexical cohesion. The following excerpt will exemplify the phenomenon:

Excerpt 1. (adapted from Straniero Sergio 2012) (H = tv host, G = tv guest, I = interpreter)

As we can see in the example, when providing their rendition of the TV guest’s utterance, the interpreter (line 3) repeats some of the words which had been previously uttered by the TV host when asking the question (line 1), ‘thereby strengthening the topical continuity between the question and the answer’ (Straniero Sergio 2012: 41). In concluding his paper Straniero Sergio provided a threefold classification of the repetitions in his data: comprehension-oriented, as they help the audience better understand the interaction, production-oriented, as they facilitate the interpreter in choosing translation equivalents, and finally interaction-oriented, as they favour mutual, intersubjective, understanding (Straniero Sergio 2012: 51).

Drawing inspiration directly from Straniero Sergio’s paper, by using a methodology grounded in Conversation Analysis (CA), the present study aims to further investigate the functions of repetitions in dialogue interpreting in different institutional settings, specifically my dataset comprises institutional encounters in the healthcare and educational settings. These two settings differ from the one examined by Straniero Sergio  first, for the absence of a silent audience (which the practice highlighted by Straniero Sergio makes reference to), and second, because they are characterised by less strict time constraints; while a TV interview is necessarily limited by the duration and the running order of the TV show, during medical consultations and parent-teacher meetings primary speakers have greater possibilities of modifying the duration of the interaction taking place. This allows for the initiation of dyadic sequences, between one primary speaker and the interpreter/mediator, to clarify potential misunderstandings and better negotiate the content of the interaction (cf. Section 2 for further details).

In line with Wadensjö’s observations about the ‘situatedness’ of renditions in interpreter mediated interaction (cf. Wadensjö 1998: 154), also the repetitional phenomenon found in my data is different: while Straniero Sergio investigated cases of repetitions of words uttered by one primary speaker and inserted as a repetition within the interpreter’s rendition of the other primary speaker’s words, the repetitions investigated in this paper, on the one hand, occur within dyadic monolingual sequences where questions are being asked and answered, and, on the other hand, are not limited to repetitions of the primary speaker’s words, but can also be repetitions of a mediator’s own words. Following a quite established distinction within the literature on CA, all instances of repetitions have been first classified according to two main categories, the first being other-repetitions (ORs), regarding instances of a mediator’s repetition of the primary speaker’s words, the second being self-repetitions (SRs), occurring when a mediator repeats their own words from a previous turn-at-talk. Each of these two types of repetitions has been examined in detail and, as for SRs, three specific patterns have been singled out and independently described, in order to understand their function in mediator-interpreted interaction. What has been observed is that by deploying these repetitions, mediators not only acknowledge the receipt of what their interlocutor has said, but can also elicit further confirmation of the information that they are gathering from their interlocutor, even in cases in which a clear answer to the question being asked has already been provided by the primary interlocutor. Mediators seem to orient to the deployment of repetitions with such a function when the interaction involves primary interlocutors with difficulties in expressing themselves in English, used as a lingua franca, thus in scenarios where mutual understanding is potentially at stake, therefore confirming Straniero Sergio’s classification of repetitions in interpreter-mediated interaction as an interaction-oriented practice.

Before analysing specific examples of ORs and SRs (cf. Section 5), the following sections will provide an overview of previous research on the function of dyadic sequences in dialogue interpreting (cf. Section 2), on repetitions and their functions in interaction (cf. Section 3) and of the data and methods used for this study (cf. Section 4).

2. Dyadic sequences in mediator-interpreted institutional interaction

 The distinction between the monologic approach to interpreting, typical of conference interpreting, and the dialogic nature of interpreting taking place in public services was first clearly presented in the seminal work by Cecilia Wadensjö (1998), Interpreting as Interaction. Here the author highlights that, within the triadic setting of dialogue interpreting, interpreters are not mere conduits of translated messages, but they function as coordinators of the interaction, for example in allocating turns-at-talk, disambiguating potential trouble sources, preventing and/or correcting misunderstandings, and so on. This coordinating role entails the possibility and sometimes the necessity for interpreters/mediators to take communicative initiatives, in order to safeguard the positive and effective outcome of the triadic encounter. Further studies have confirmed this function, for example by showing how there might be circumstances in which interpreters need to address and answer their interlocutors’ questions or doubts directly (Metzger 1999). This fosters, as already noted by Wadensjö herself (1998: 11), the development of dyadic exchanges within the triadic encounter, namely monolingual sequences in which the interpreter/mediator only speaks either with the service provider or service user in their own language (cf. also Bolden 2000 and even more explicitly Davidson 2002). While some studies have foregrounded some potential downsides of dyadic sequences, such as the fact that they may enhance the patient’s sense of exclusion (Baraldi 2009), cause the potential exclusion of one participant (Gavioli and Baraldi 2011), or hinder direct contact within primary speakers (Cirillo 2012), other studies have focused on their positive effects in interaction. Dyadic sequences have been shown to support the patient from an affective point of view (Baraldi and Gavioli 2007), they have been found to be the locus of affective alignment while triadic sequences are more prone to cognitive alignment (Zorzi and Gavioli 2009), they can also be the locus where misunderstandings are cleared (Baraldi and Gavioli 2016), and they can offer the opportunity for mediators to negotiate what needs to be rendered (Baraldi 2023). Since service-users show an orientation to communicating more effectively with the interlocutors who speak their own language, dyadic sequences are a decisive factor for the success or failure of triadic encounters (Gavioli and Merlini 2023). Finally, in one of her studies of the AIM Corpus (cf. Section 4), Gavioli (2015) has also shown that service providers, namely the clinicians, display an orientation to clarifying the information they want to gather from or communicate to the patient in dyadic sequences, which is coupled with the orientation to authorising the mediator to manage the communicative process with a certain degree of freedom, via a subsequent dyadic sequence with the patient. A similar orientation has also been described in the educational setting by Davitti who wrote that teachers ‘seem to pass the ‘baton’ of interaction to interpreters’ (2015: 18).

3. Repetitions and their functions in social interaction as described in the literature

As repetitions are such a pervasive element in talk-in-interaction both their definition and their many functions have been extensively discussed in the literature. Repetitions are a recurrent instance in human communication, since they regularly occur both in spoken and written language (Johnstone 1987: 206), although they appear to be more frequently found in oral discourse (Tannen 1987: 226). From a strictly linguistic point of view, definitions of repetitions are somehow blurred. Johnstone (1994: 3) notes that what can be considered as repetition ‘will have to do with how easily we can identify something as being “another one of those”’. She also highlights how the idea of repeated pattern is not generated only by the reoccurrence of the same lexical items but can be identified in the reiterated presence of the same sound(s) in different words (such as in the case of alliterations), or of the same prosodic patterns in different phrases or sentences, an idea which was also explored in talk-in-interaction by Jefferson (1996). As regards the scope of this study, following Schegloff (1997), with the word repetition (or repeat) I refer to the reproduction of a preceding word, phrase, or sentence in the conversational continuum. More precisely, in Schegloff’s words:

The use of the term ‘repetition’ or ‘repeat’ here is more or less strict; that is, it allows for transformations geared to deixis, tense shift, speaker change, etc., as well as changes of prosody; it excludes paraphrase and other substantial rewording of its target. (Schegloff 1997: 525)

Such definition will be applied to cases of other-repetition (OR), while instances of self-repetition (SR) may also be defined, in line with the literature (Schegloff 1987), as instances of recycling. In the CA literature, there is an established practice of examining repetitions according to two specific characteristics (Schegloff 1996), namely their intonational contour and their sequential position. As for the former, the aspect which will be considered here is the pitch movement at the end of the turn (rising or falling contour), since prosody ‘furnishes important, if not crucial cues as to what a speaker is doing when repeating the words of another in subsequent position’ (Couper-Khulen 2020: 521). As for the latter, repetitions can generally both occur in first position, for example when the words of a speaker are repeated by another to ask a clarification question, and in second position, for instance when answering a question or responding to a greeting. In addition, whenever a speaker reacts to an utterance in second position by producing a new utterance, said reaction will occupy a third position, in which also repetitions have been observed to occur. As all instances of ORs considered in the present study are examples of repetitions of answers provided by the mediators’ interlocutor they fall under the definition of third-position repetitions; they all also happen to be uttered with final falling intonation contour. Conversely, all the SRs analysed below are examples of first-position repetitions with rising intonation, since they are the redoing of questions previously asked by the mediators.

As for the functions of repetitions in interaction, the CA literature has shown that speakers deploy ORs so as to connect their turns-at-talk to the preceding ones. For instance, Drew (2013) shows that repetition is one of the four principal ways in which speakers design their turns and display them to be connected to and coherent with prior turns. Besides, repetition of words by one speaker can represent an opportunity for another interlocutor to enter the speaker’s turn-space and therefore can constitute a case of conditional entry (Hayashi 2013: 184), but they can also be deployed by next speakers as a means to (re)take the floor (Simpson 1994), or to compete for it (Schegloff 2004). ORs can also be deployed to make an epistemic claim when answering questions (Raymond 2003; Stivers 2005; Heritage and Raymond 2012), to initiate or close repair sequences (Schegloff, Jefferson and Sacks 1977; Kitzinger 2013), to introduce rejection, corrections, disalignments and other dispreferred actions (Pomerantz 1984), to confirm allusions (Schegloff 1996). With regards to the correlation between sequential position and intonation contour, among the early studies that looked at the interactional functions of these two aspects combined, Schegloff (1997) noted that, while ORs with rising intonation tend to be associated with the initiation of repair, while third-position ORs with falling intonation contour can also have the function of registering and acknowledging the receipt of information. However, Schegloff also highlighted that third-position ORs with falling intonation can lead to an ambiguity about whether they register receipt or initiate repair, since some speakers orient to proffer some sort of confirmation after the repetition. Rossi (2020: 514) suggests that one way of solving this ambiguity is by looking at whether the repeated talk ‘is “taken in” rather than problematised’. Another study has confirmed that ORs with falling intonation can be uttered to register receipt but also to present a claim of understanding (Svennevig 2004) while, in a study of interactions between palliative care doctors and patients, third-position ORs with falling intonation have been shown to prompt the patients to add further details to what they have already said in their answer to previous questions (Jenkins et al. 2021).

As for SRs, fewer studies have examined their function in interaction. However, the CA literature has shown that SRs are often associated with initiating or solving instances of repair (Schegloff 2004) or with the phenomenon defined as framing, in which a speaker repeats their own words so as to highlight the relevance of one single word which has been changed in the repetition (Schegloff 2013). Besides, a few studies have found other functions of SRs, such as displaying that the speaker finds ‘the prior speaker’s course of action problematic and proposes that that course of action be halted’ (Stivers 2004: 288), closing sequences of talk (Curl, Local and Walker 2006), delaying the uttering of other sequentially adjacent words (Lerner 1996).

4. Data and Methods

The dataset analysed for the present study comprises audio-recorded interpreted interactions from two different institutional settings within the provinces of Modena and Reggio Emilia, in the North of Italy. The first subset of data contains seven interactions from the healthcare setting; these are gynaecological visits involving Italian clinicians, and patients and mediators from Nigeria or Ghana speaking English, for a total of approximately three hours of recordings. The second subset of data is completed by six more interactions from the educational setting, adding approximately two hours of recordings to the dataset. These are parent-teacher meetings recorded in local schools: two of them were recorded in a nursery school with Nigerian parents and mediator, speaking English, while the remaining four interactions involved Ghanaian mediators and parents speaking Twi, a Ghanaian language commonly used as a lingua franca in Ghana (Huber 2013) and containing loanwords from English.

In line with the Italian public system, interpreting services in the data are provided by intercultural mediators (CNEL 2009), also known as linguistic and cultural mediators (Falbo 2013; Baraldi and Gavioli 2019). Like the users they help, all linguistic and cultural mediators have a personal history of migration and the main language of mediation analysed in the data is English as a lingua franca (EFL). Various studies have focused on interpreting as provided by intercultural mediators from a professional perspective (Merlini 2009b; Pittarello 2009; Pöchhacker 2008; Gavioli and Wadensjö 2023; among others), and while some studies have highlighted some critical aspects of linguistic and cultural mediation (Bolden 2000; Davidson 2000, 2001; Merlini 2009a; Baraldi and Gavioli 2019), others have shown its effectiveness in coordinating various aspects of interactions between migrant patients and clinicians (Luatti 2011; Baraldi and Gavioli 2012; Urlotti 2023). For an analysis of the role of intercultural mediation in interpreting studies within the Italian context and in the interaction please also refer to Gavioli and Baraldi (this issue).

Both subsets come from a much larger corpora of mediator-interpreted interactions, the AIM Corpus (Corradini et al. 2024), collecting around 700 mediator-interpreted interactions in multiple settings and with different languages, the most represented being English, French, Mandarin Chinese, and Arabic. This corpus is the result of a long-lasting research cooperation between the University of Modena and Reggio Emilia and the public services of the area, the major one being with public healthcare. The interactions in the educational setting come from a recent collection, gathered for a Horizon 2020 project (CHILD-UP Corpus, cf. Baraldi 2021) and now them too included in the AIM corpus. All data were transcribed using ELAN 6.2, an open-source software which allows for great accuracy in the synchronization of the recordings with their transcripts^[1]. The transcription conventions (cf. the details at the end of this paper) follow the Jeffersonian style which is usually adopted for CA, and all backtranslations of the data in English are by the author.

This paper develops a recent study scrutinizing the role of repetitions in combination with multi-part renditions as a practice to maintain mutual understanding between interlocutors in mediator-interpreted institutional encounters (Urlotti 2024). The analysis, there and here, has been carried out using the analytical tools offered by CA. Within the CA theoretical framework some of the concepts which are particularly significant for this study are the general idea that speakers orient to designing their turns-at-talk by considering their recipient’s characteristics, a practice also called recipient design (Sacks, Schegloff and Jefferson 1974). Another type of speaker orientation which is paramount for this study is the orientation to the interlocutors’ knowledge about the topics at hand, which falls within the domain of epistemics (Heritage 2012a, 2012b) and, finally, the mechanism of preference in relation to responses (Pomerantz 1984; Schegloff 2007), the idea that certain elements in responses, such as inter-turn delays, prefaces, mitigations or accounts, indicate that the given response may show misalignment with the questioner.

Before moving to the analysis of the data, an explanation is required on the selection of the extracts shown below and their relationship with the phenomenon at hand. As previously stated, this paper develops a wider study of repetitions within multi-part renditions, which were found in both settings included in my dataset, the healthcare and educational ones. Nevertheless, the only instances of the types of repetitions described in this paper were found in the data coming from the healthcare setting. As the analysis below will show, such repetitions seem to be deployed by mediators when interlocutors show a lack of linguistic competence which might put mutual understanding at stake. As a matter of fact, the interactions in the educational subset involve either speakers of Twi, who do not seem to have problems in using the language, or speakers of English as a lingua franca who show an above-average linguistic competence. Conversely, all the migrants from the healthcare subset speak English as a lingua franca at notably different levels of proficiency, with some cases of speakers at times struggling to express themselves, with whom mediators deploy the types of repetitions here examined.

5. Repetitions in mediator-interpreted interaction

This section is dedicated to presenting extracts from the data which can effectively exemplify some of the functions that repetitions can have in mediator-interpreted interaction. In my analysis I am only going to focus on repetitions uttered by mediators after a patient’s response. In uttering a repetition, mediators initiate a short dyadic sequence which may produce a specific effect on the exchange, thus clarifying what function the repetition has had in the interaction. As previously specified, repetitions will be divided into two different groups, according to whose words the mediator is repeating: in those cases, in which she is repeating the patient’s words, I will use the definition of other-repetition (OR), while if she repeats (or recycles) her own words, I will opt for the definition of self-repetition (SR). One issue that should be given some consideration before getting into the detail of the analysis is that the patients involved in interactions where instances of ORs or SRs are found display lower levels of language proficiency in English as a lingua franca (ELF). Despite their coming from countries where ELF is commonly spoken by most of the population, such as Nigeria or Ghana, not all migrants speak it at the same level of competence, creating discrepant levels of proficiency, which can be the cause of serious misunderstandings or other communicative problems when accessing public services. Therefore, there seems to be a correlation in the examples below between the use of repetitions and lower levels of linguistic competence in ELF, which would characterize my occurrences as particularly apt to accomplish clarification functions. I will now proceed with the analysis of the two groups of sequences, starting from the examples of ORs. Section 5.1 will be entirely dedicated to the main type of ORs I have in my data, while Section 5.2 will present three main types of SRs and their functions.

5.1 Other-repetitions

The pattern of ORs which is presented in this study can be summarised as follows: first there is a wh-question asked by the mediator (first position), which is followed by a response with which the patient provides some information (second position), and then a repetition of the patient’s answer is produced by the mediator (third position) with falling intonation, hence all the ORs here analysed belong to the category of third position ORs with falling intonation. As previously described (cf. Section 3), the existing literature on third position ORs with falling intonation has already highlighted that this type of repetition displays that the utterance produced in second position has been received and/or registered (Schegloff 1996: 179); in other words, it might be said that such a repeat is useful to acknowledge the relevance of a speaker’s previous talk. The analyses below will show that acknowledging the receipt of the interlocutor’s utterances can have a fundamental role in guaranteeing and safeguarding mutual understanding in a context where the use of a lingua franca may put comprehension at stake. This idea is further corroborated by the fact that these repetitions registering the information provided by the patient systematically occur before the mediators’ renditions into Italian, thus directly linking their function to the mediating process.

Another element which completes the description of the pattern here presented, and is fundamental for the analysis, is the presence of elements which a mediator may interpret as a sign of the patient’s hesitation in between the mediator’s question in first position and the patient’s answer in second position. Usually, such hesitations take the form of an interturn gap that is longer than 0.2 seconds. On the length of inter-turn gaps, it should be specified that the CA literature has shown how English speakers orient to a 0.25 second silence as a possible transition relevance place (TRP), namely a moment in the interaction when another speaker takes the floor and produces an utterance (Hepburn and Bolden 2017). Thus, any longer gap may be interpreted as a sign that there is some interactional dynamic which is preventing the interlocutor from taking the floor, therefore also as a sign of hesitation. The following extract showcases a clear example of this pattern, completed by a third position OR with falling intonation:

Extract 3. (M = mediator; P = patient)

As the extract shows, after asking the question (lines 1 and 2), the patient’s response ‘eight’ (line 4) is preceded by a 0.3 gap (line 3) and by a stretched ‘a:h’, another element which may be interpreted as a sign of hesitation or doubt, and it is promptly repeated by the mediator with falling intonation (line 5). The mediator then proceeds to translate the information for the clinician (line 7) after leaving a 0.3 gap (line 6), one which is potentially long enough for the patient to be able to add something. In doing so the mediator shows an orientation to leaving space for potential expansions, and to making sure that the patient does not want to add anything to the previous answer; the same applies to the idea that, by repeating what the patient has said, the mediator offers the patient a possibility to rectify or correct the piece of information that is being discussed. Consequently, this extract also seems to be in line with the function of third position ORs with falling intonation as sequence closing elements (Schegloff 1997).

However, as previously said (cf. Section 3), third position ORs with falling intonation can sometimes also be interpreted by the speakers engaged in talk-in-interaction as initiating repair or requesting confirmation (Schegloff 1997), prompting the interlocutor to utter a reaction to such repetition. The following extract illustrates an instance in which such reaction is produced:

Extract 4. (M = mediator; P = patient)

Here we can see that the mediator’s question (line 1) is followed by a long inter-turn gap (line 3). Then, like in the previous extract, the patient’s answer (line 4) is preceded first by a stretched token which announces that the patient is going to answer and then by a micropause; these three elements together account for a patient’s display of potential uncertainty and hesitation. After the patient’s answer, the mediator immediately produces a repetition of the given information (line 5) which, unlike in the previous extract, is here followed first by a 0.4 gap (line 6) and then by a confirmation token uttered by the patient (line 7). In line with Schegloff (1997), the patient orients to interpreting the mediator’s repetition as a repair initiation or as a request for confirmation. Moreover, like at the end of the previous extract, after the patient’s confirmation token has been uttered (line 7), the mediator seems to leave the necessary space for the patient to add extra information by waiting 1.1 seconds (line 8) before producing her translation of the information provided by the patient (line 9).

In line with this consideration, apart from being interpreted as a request for confirmation, in my data patients seem to orient to another function of mediators’ ORs, that of eliciting expansions, like in the following extract:

Extract 5. (M = mediator; P = patient)

In this extract we can see that the mediator is enquiring the patient about her period (line 1) and specifically about when she last had one (line 3). Like extract 3 and extract 4, also in this extract the patient’s answer (line 5) is preceded by a relatively long gap (0.6) (line 4). Unlike the previous extract, extract 5 presents a new element which introduces the actual answer (line 5); before uttering ‘I saw it last month’ the patient prefaces it with ‘m:h I don’t remember’, a sentence which on the one hand seems like a disclaimer, while on the other it seems to introduce a caveat, not to be overlooked by the mediator, that the information provided by the patient might not be accurate. This type of patient expressed caveat, having to do with the patient’s ability to recall the required information, represents a type of potential trouble source in mediator-interpreted interaction which is not strictly connected to the patient’s linguistic competences but is relative to their cognitive abilities, therefore prompting the mediator to double-check the gathered information. The mediator indeed acknowledges her receipt of the information with a third position OR with falling intonation (line 6), which, like in the previous extract, is interpreted by the patient as a request for confirmation, signalled by the ‘yeah’ uttered at the beginning of line 7. However, the patient doesn’t limit herself to a confirmation token but goes on to add extra information ‘but I saw it (0.3) thrice’ which completes the answer to the mediator’s original question (line 3). In this case the OR has elicited not only a confirmation but also an expansion.

The analysis of the previous three extracts shows that third position ORs with falling intonation in mediator-interpreted interaction can be deployed as a way for the mediator to acknowledge the receipt of the information provided, in an interactional context in which hesitations or potential trouble sources are also apparent. Such acknowledgment can give the patient the possibility to rectify or amend the repeated information. In other cases, the data show that ORs are interpreted by the patient as a request for confirmation, such confirmation providing the mediator with an extra possibility to verify whether the information is correct. Finally, third position ORs with falling intonation are also observed to elicit expansions which add details to the patient’s answer. These repetitions, uttered during the dyadic exchange between mediator and patient and preceding the mediator’s rendition of the gathered information into Italian, signal the mediators’ orientation to not taking the reception of such information for granted, thus decreasing the possibilities for potential misunderstandings.

5.2 Self-repetitions

While SRs have been found to be mostly associated with repair initiatives in the literature (cf. Section 3), in my data SRs seem to be often deployed as a confirmation seeking device. A clear pattern which emerges is that the mediator asks a yes/no question to the patient, which is followed by the patient’s ‘yes’ as an answer; however, despite having received a clear positive answer, the mediator repeats the same question with a little variation so as to elicit a second positive confirmation. The CA literature is also familiar with the notion that repetitions that are identical in all aspects, including prosody, tend to be treated by other speakers as showing a problem with hearing (Schegloff 2004). In line with this consideration, another interesting feature of the types of SRs presented in the following three subsections is that mediators design their confirmation-seeking SRs so as to include some slight variations; by doing so, they avoid making their SRs too linguistically complicated, as such complexity might be the cause of misunderstanding with migrants showing poor levels of language proficiency. For reasons of space only one exemplar per practice is provided as an example.

5.2.1   Self-repetitions with epistemic upgrade

The first pattern of modified SRs includes the repetition of the same question from an epistemically upgraded stance. In his theorisation of epistemics in interaction, Heritage (2012a, 2012b) has shown how speakers can modify the design of their utterances so as to display their positioning on an epistemic gradient, moving from a less knowledgeable [K-] stance to a more knowledgeable [K+] one. An interrogative polar question with rising intonation, for instance, shows a K- epistemic stance compared to a K+ declarative question (Couper-Khulen and Selting 2018), also known as a B-event statement (Labov and Fanshel 1977), which is a turn designed as a statement eliciting confirmation or information. The following extract will present a clear example of change in epistemic stance through an SR:

Extract 6. (M = mediator; P = patient)

In this extract the mediator is asking the patient about a medicine that she is supposed to take throughout her pregnancy, and she needs to know whether the patient still has enough of it. In lines 3 and 4 the mediator designs her first question ‘eh you still have? (.) enough of it?’ syntactically as a statement, but with rising intonation. After a long gap (line 5) the patient produces an answer ‘yes’, preceded by a false start ‘y-’ (line 6). As in the extracts regarding ORs, here we see the combination of a long gap with a false start, showing hesitation before providing the answer. In the following line the mediator produces an acknowledgment token ‘mh’ followed by a 0.5 pause and then deploys an SR of the same question ‘you still have enough.’ (line 7). Unlike the first-time proffering (line 3 and 4) now the question is designed as a statement, with falling intonation. This shift in intonation, from rising (line 4) to falling (line 7), accounts for the epistemic upgrading of the question, which in its second version is seeking confirmation. The patient utters her confirmation ‘yeah’ (line 8) and the mediator acknowledges it with a ‘mh okay’ (line 9). If we consider a recipient with poor language skills, this type of modified SR serves the purpose of asking for confirmation without introducing new linguistic elements which might put the patient’s comprehension at stake.

5.2.2   Inverted-order self-repetitions

I have called the second type of modified SR inverted-order self-repetitions by reason of the kind of modification in the design of the repetition of the original question; such modification, indeed, consists of a change in the word order of the phrases constituting the question itself. The following extract will serve as an exemplification of this mechanism:

Extract 7. (M = mediator; P = patient; D = doctor)

In this extract the doctor asks a question about the regularity of the patient’s period (line 1 and 2). The design of this question looks somewhat complicated as the micropause, combined with the use of two semantically equivalent expressions ‘every twenty-eight thirty days’ and ‘all the months’, seem to indicate (line 1 and 2). The question design is also complicated by the repetition of the verb ‘vengono’ (literally ‘they come’ referring to the patient’s periods). However, the question is asked in one turn-at-talk with an orientation to selecting the right words to design the question, which seems to rule out the possible argument that it is the doctor in the first place who repeats the same question by inverting the order of the phrases. The orientation to deploying a SR will instead become apparent when analysing the mediator’s turns. The mediator first asks the question in line 4. This is followed by a 0.4 inter-turn gap (line 5) and by a ‘yes’ uttered by the patient (line 6). Nevertheless, despite having received a clear answer to her question, the mediator repeats the same question by inverting the position of the two phrases ‘they come’ and ‘every month’. As for the previous extract, this way of rephrasing the same question without introducing new linguistic elements seems to be particularly functional in the case of an interlocutor with poor language skills. Moreover, this extract can also be taken as an example of SR with epistemic upgrading; as a matter of fact, the intonation of the two questions shifts from slightly rising (line 4) to falling intonation (line 6), therefore accounting for a case of epistemic upgrading.

5.2.3   Self-repetitions with expansions

The distinctive characteristic of the third type of modified self-repetition is the addition of a new element to the repeated question. The CA literature has for long been familiar with the phenomenon usually referred to as increments, in which words or phrases are added to a previous TCU (turn constructional unit) ‘not as a new TCU, but as a continuation of the preceding TCU’ (Schegloff 2016: 241). However, since the SRs in my data, being the repetition of a previous question, are by definition new TCUs, the element which is added to them cannot be defined as an increment proper and the term cannot be applied. I have therefore chosen to adopt the broader term expansion, following Auer (2007: 651) in his attempt at approaching the topic of unit expansions in conversational speech in a more general way than it has been described in the CA literature about increments. The following extract will showcase an example of SR followed by an expansion:

Extract 8. (M = mediator; P = patient; D = doctor)

After the doctor has asked a question about whether the patient is able to eat (line 1), the mediator first asks the question once (line 3) and, after a 0.6 second gap (line 4) and a ‘ye:s’ uttered by the patient (line 5), she deploys an identical repeat of the same question to which she adds the word ‘fine’ (line 6). The patient understands the question and confirms the request (line 7), and finally the mediator acknowledges the confirmation with a ‘m:h’ (line 8). A few aspects of this expansion are worth commenting on. First of all, while the added word ‘fine’ is semantically compatible with the question it is added to, its addition does not drastically change the meaning of said question. Secondly, if the previous comment is accepted as true, the expansion seems to have the purpose of signalling that the same question has been asked one more time, more than highlighting that something new has been added to the sentence. Lastly, although this pattern differs from the previous two, since it introduces a new linguistic component to the question design, the word added as an expansion does not introduce an element of complexity and therefore is suitable to achieve effective communication with speakers whose language skills are poor.

6. Concluding remarks

The analysis has shown different functions of various types of repetitions in mediator-interpreted interaction. To facilitate an orderly classification of the phenomena at hand, repetitions have been first categorised according to whose words were being repeated by the mediators: in case these were the words of one of the primary speakers, the repetition has been defined as other-repetition (OR), while in those cases in which mediators repeat their own words, the repetition has been defined as self-repetition (SR).

As regards ORs, I have shown how they are often deployed during information gathering sequences to acknowledge the receipt of the elicited information, leading to the closure of the dyadic sequence so as to move to the rendition into Italian (cf. extract 3). However, ORs may also be interpreted by the patient as a request for confirmation (extract 4 and 5), sometimes leading to further expansions (extract 5). This last function seems to confirm the results of another study which has shown how third position other-repetitions with falling intonation deployed by clinicians may elicit information from patients (Jenkins et al. 2021). More generally my results regarding ORs and their functions are in line with what was already known in the literature (Schegloff 1997) but had not been applied to nor verified in mediator-interpreted interaction.

Conversely, as regards SRs and their functions in interaction, the analysis has foregrounded a function of SRs which has not been described in the literature before, namely that of seeking confirmation after a yes/no question which has received a ‘yes’ as a clear answer. With this practice mediators seem to treat the first ‘yes’ as not sufficient for the information gathering process and orient to making sure that a redoing of the same question is met with the same type of answer. The analysis has also shown that, in order to carry out this confirmation-seeking function, SRs are designed by mediators with slight variations of different types: on the one hand, through such variations, mediators avoid deploying an identical repeat, which could be interpreted as signalling a problem of hearing, while on the other hand they avoid rephrasing, and thus maintain a level of complexity similar to that of the first question, which had already received a plausible answer. This element seems to be directly linked to another aspect which became apparent through the analysis of both ORs and SRs: in my data repetitions in mediator-interpreted interactions are deployed after a patient’s answer which displays some sort of hesitation, such as longer inter-turn gaps, false starts, stretched sounds within words, or when the patients introduce a caveat such as stating ‘I don’t remember’ before providing an answer. In other words, repetitions seem to be deployed as a confirmation-seeking device after perturbations or caveats in talk-in-interaction which may display that mutual understanding is at stake.

When looking at the analysis of ORs and SRs together, one might legitimately ask whether the two types of repetition are a different realisation of the same practice. The only SRs which were found in my data, apart all cases of self-repair which were excluded from the analysis, were cases of confirmation-seeking SRs. As regards ORs, although I have shown that, in line with the existing literature, some of the ORs in my data are interpreted by the other speaker as a request for confirmation, there are still various cases of ORs which can be ascribed to the simple function of registering receipt of the information, a function which has already been acknowledged as fundamental in the context of institutional interactions (Rossi 2020: 514). Thus, one of the aims of this paper is to show that despite the various functions of different types of repetitions in my data, what can be found as a common element among all of them is the mediators’ orientation to ascertaining that the information they are gathering is correct, therefore fulfilling their institutional task efficiently and responsibly.

As a matter of fact, the purpose of this paper is not only to look at the presence of repetitions in mediator-interpreted interaction from a strictly conversation-analytical point of view, but also to provide insights into some relevant professional implications for linguistic and cultural mediators working with migrants. The fact that in my data repetitions have been shown to be deployed in circumstances when mutual understanding seems to be potentially at stake once again highlights the paramount importance of the coordinating function of intercultural mediators. First of all, migrants as public service users often find themselves in a position of special vulnerability due to their condition of displaced people accessing public services in a foreign country. Secondly, when English is used as a lingua franca, various issues are raised concerning both how a migrant’s native language influences their way of speaking English (Guido 2018), and how their specific lack of linguistic competence in general can hinder effective communication. These are factors which should not be overlooked by linguistic and cultural mediators.

Therefore, from the point of view of research on interpreting, it can be argued that this study contributes to shedding light on how repetitions can be deployed by mediators as an effective means to double-check the service users’ and their own understanding of the information that is being discussed. Repetitions have been shown to have multiple advantages: first, they do not significantly modify what has already been said in the interaction, therefore constituting only a minor variation to the interaction; second, by using repetitions mediators can avoid more complex linguistic formulations which might hinder comprehension for service users with poor linguistic competences; finally, repetitions are not only useful to verify the information that has already been provided but can potentially leave space for expansions providing further information. In conclusion, this study shows how some communicative practices involving repetitions can be deployed by mediators in order to safeguard the information gathering process and ultimately contribute to the successful outcome of the institutional encounter.

About the author(s)

Daniele Urlotti is a postdoctoral research fellow and adjunct professor in the Department of Linguistic and Cultural Studies at the University of Modena and Reggio Emilia, where he completed a PhD in Human Sciences. His research centres institutional communication in healthcare and educational settings studied through Conversation Analysis, with a special focus on interactions involving migrant service users who require assistance from interpreters or intercultural mediators.

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

•  Previous article
• Next article