Knowledge negotiation and interactional power: epistemic stances in Arabic–Swedish antenatal care consultations

1 Introduction

This article deals with the negotiation of knowledge in multilingual settings involving an interpreter. More specifically, it investigates how Swedish-speaking midwives, Arabic-speaking pregnant women and bilingual interpreters claim, challenge and formulate knowledge as part of antenatal care consultations. We see such knowledge negotiations as an aspect of power in interaction.

Existing research has demonstrated how medical encounters are typically characterised by power asymmetries between the participants, including with regard to knowledge (Gillespie et al. 2014; Wodak 1996). Interpreters may subtly or more forcefully influence such asymmetries (Angelelli 2004; Mason and Ren 2012), and miscommunication may increase the risk of medical problems during pregnancy and childbirth (Esscher et al. 2014).The multilingual and cross-cultural settings of interpreter-mediated encounters mean that participants – in our case midwives, pregnant women and interpreters – have to navigate complex aspects of information exchange, relationship formations and so on (cf. Angelelli 2004; Penn and Watermeyer 2018), all of which pertain to knowledge and power in various ways.

Firstly, midwives have access to a wealth of knowledge based on scientific research, on their own professional training and experiences, and on the long history of their profession. This is part of their ‘institutional power’ (Mason and Ren 2012). Secondly, pregnant women have first-hand knowledge of their own bodies, and may also have access to knowledge of pregnancy based on their own previous pregnancies, the experiences of others and information gleaned from other sources such as books, websites and social media (see e.g. Hanell 2017). Thirdly, interpreters have knowledge by virtue of their bilingual and bicultural competence, which they can use in various ways to influence the interaction (Mason and Ren 2012). All of these different kinds of knowledge may come into play in antenatal care consultations, as part of participants’ varying ‘interactional power’ (Mason and Ren 2012).

In this article, we analyse antenatal care consultations in Sweden as part of a larger research project on norms and interaction in antenatal care, focusing on Swedish-speaking midwives and Arabic-speaking pregnant women (Bitar and Oscarsson 2020). Swedish antenatal care follows a nationwide programme, with minor regional variations. Typically, the programme involves eight to ten visits to an antenatal care centre during pregnancy. The first one or two visits include enrolment and a health talk. The health talk concerns the pregnant woman’s health, as well as guidelines and practices regarding alcohol and tobacco, nutrition and physical exercise. Subsequent visits focus on information regarding breast-feeding, childbirth, contraception and so on. The visits are also used to monitor foetal development and the health of the pregnant woman, through a series of tests. Non-Swedish-speaking patients have the right to an interpreter in health care consultations.

The aim of this article is to investigate knowledge negotiations as an aspect of interactional power in three-way interaction between Arabic-speaking women, Swedish-speaking midwives and interpreters in Swedish antenatal care. The study specifically focuses on sequences in Arabic–Swedish conversations where an Arabic-speaking woman takes, or is given, the opportunity to question information, introduce a topic not addressed by the midwife, or put forth her own opinion or experience – in short, to make her ‘voice’ (Blommaert 2005) heard, possibly challenging the midwife or the Swedish antenatal care programme.

2 Interaction, knowledge and power in multilingual medical settings

Interaction is crucial in many health care contexts, from the medical professional’s determination of whether the patient has a medical problem at all (see e.g. Heritage’s 2009 notion of ‘doctorability’) and, if so, which treatment should be given, to the patient’s understanding of how they should subsequently act. In order to tease out the interactional and contextual conditions of the antenatal care consultations in our study, this section reviews previous studies on power and knowledge in medical settings, the role of the interpreter in interaction and factors influencing interpretation.

3 Theorising power and knowledge

The interactional dynamics between pregnant women, midwives and interpreters can be theorised with the help of Foucault’s reflections on the interconnectedness of discourse and power/knowledge. As Foucault famously put it, discourse can be understood as ‘a practice that systematically forms the object of which it speaks’ (1972: 49). As such, discourse constitutes specific forms of knowledge, which, in turn, circulate and get invoked and contested by a variety of social actors. Thus, fundamental to discursive knowledge is power. Or, in Foucault’s words, ‘there is no power relation without the correlative constitution of a field of knowledge that does not presuppose and constitute at the same time power relations’ (1977: 27). Put differently, knowledge and power are co-constitutive: knowledge is a pre-requisite of power, but knowledge is also, in turn, shaped by power.

It is important to clarify that, according to a Foucauldian perspective, power is ‘not that which makes the difference between those who exclusively possess and retain it, and those who do not have it and submit to it’, but rather ‘as something which circulates, or … only functions in the form of a chain’, and where individuals ‘are always in the position of simultaneously undergoing and exercising this power’ (Foucault 1980: 98). This is reflected by Mason and Ren’s notion of ‘interactional power’, which they characterise as ‘power within the exchange’ (2012: 238). They state that the interpreter’s special interactional power arises from their bilingual and bicultural expertise (Mason and Ren 2012: 238).

Analytically, a Foucauldian approach to power/knowledge can be operationalised with the help of a variety of sociolinguistic approaches which include critical discourse analysis, feminist conversation analysis, ethnography of communication and many others. We believe that the notion of stance, in particular, is an important analytical concept which allows us to capture the push and pull of power/knowledge negotiations in the interactions between midwives, pregnant women and interpreters under investigation in this article, thus offering a granular analysis of how the participants both exercise and undergo power by the way in which they mobilise their respective knowledge.

Stance has been famously defined by Du Bois as ‘a public act by a social actor, achieved dialogically through overt communicative means […] through which social actors simultaneously evaluate objects, positions subjects (themselves and others) and align with other subjects, with respect to any salient dimension of the sociocultural field’ (2007: 163). The three facets of stance-taking – evaluation, positioning and alignment – have been captured visually through the ‘stance triangle’ (Figure 1). Of course, in the case of the interactions under investigation in this article, the stance triangle is complicated by the fact that besides Subject 1, the pregnant woman, and Subject 2, the midwife, there is a Subject 3, namely, the interpreter.

Figure 1:

The stance triangle (Du Bois 2007: 163).

A plethora of terms has been employed in the literature to describe various types of stance-taking. The most important for the purpose of this article is that of epistemic stance, which refers to dialogically achieved claims to knowledge. Epistemic stance has been investigated in several different domains, including medical encounters (Landmark et al. 2016; Lindström and Karlsson 2016). As Jaffe points out, such stances are always ‘culturally grounded, because claims to know are embedded in and index particular regimes of knowledge and authority’ (2009: 7), and she goes on to clarify that such acts of claim-staking serve ‘to establish the relative authority of interactants, and to situate the sources of that authority in a wider sociocultural field’ (2007: 7). Through the notion of epistemic stance, then, it is possible to paint a fine-grained picture of how power/knowledge is produced, negotiated and contested in the unfolding of interaction, without losing sight of the different cultural groundings and regimes of knowledge at play in tripartite encounters involving a midwife, a pregnant woman and an interpreter.

4 Data and method

The data in this article come from Expecting a Child in Arabic and Swedish, a cross-disciplinary research endeavour involving scholars from linguistics and health science (Bitar and Oscarsson 2020; Oscarsson and Stevenson-Ågren 2020), investigating Arabic–Swedish interactions in antenatal care. As part of this project, an Arabic–Swedish communicative aid in the form of an app, named Sadima, was developed. Sadima is based on the Swedish national antenatal care programme, and can be used to elicit information and show instructional films. The app is intended to be used in combination with interpreters.

The versions of the app that were used during the development phase were equipped with an audio recording function, as a way of producing data for research. These app versions also included a text log which stored information about users’ screen selections. Published versions of the app, in the App Store and as a webpage (https://sadima.lnu.se), do not record or log any user information.

The first version of the app was tested during antenatal care appointments by about ten midwives in their meetings with Arabic-speaking and Swedish-speaking women. The midwives were in control of the tablet app and the sound recording, and could switch the recording on and off whenever they wanted. Midwives frequently switched off the recordings, making most recordings quite short – typically a few minutes – and the data fragmented. In fact, in the whole dataset there is only one recording of a meeting from start to end. For a large number of files there are no sound recordings at all, only text logs. However, it is still possible to identify highly interesting interactional patterns in the audio data. In the end, the whole dataset for this article consists of sound recordings made by five midwives during 33 consultations across a time period of 10 months.

No researchers were present during the appointments. Several of the interactions included a telephone interpreter (no on-site interpreters are used) and some included the pregnant woman’s partner.

The audio recordings of midwives and pregnant Arabic-speaking women have been approached using interaction analysis (Sidnell and Stivers 2014; Wadensjö 1998). This, we argue, can be viewed as an instance of Foucault’s (1980: 99) ‘ascending analysis of power’ (emphasis in original), starting from the ‘infinitesimal mechanisms’ of power in the form of, in our case, micro-level phenomena in multilingual spoken interaction.

For the purpose of this article, all app files have been investigated and all app recordings have been listened to by the first author, who took notes of a broad range of potentially interesting interactional phenomena in the recordings. For the most part, the interactions consist of the midwife giving information or asking questions to which the pregnant woman gives a reply. Identifying the Arabic-speaking women’s interactional possibilities and initiatives as both empirically and theoretically interesting, the first author then built a collection of the sequences where:

• – the woman takes an initiative to express an opinion, a feeling, a question, etc., or

• – the midwife does something out of the ordinary to ensure the woman’s participation (as opposed to asking a question which is replied to in the next turn, upon which the conversation moves on to a new question or topic), or

• – the woman and the midwife have different views, interactional goals, ways of framing an issue, etc.

This resulted in a collection of six sequences. The first author listened to these sequences several times, made rough transcriptions and took notes. Three of these sequences are included in the present article, chosen to show different aspects of the Arabic-speaking women’s possibilities for being heard (these different aspects are reflected in the subheadings in the analysis below). The first and the second authors then collaborated on translating and analysing the utterances spoken in Arabic by the women and by the interpreters. The first author is a linguist trained in interaction analysis. The second author is a health scientist with no previous knowledge of linguistic interaction analysis but with experience in transcribing interview data. The selected sequences have also been analysed in sessions with other interactional researchers external to the project, and by the third author, who is a discourse analyst in the field of multilingualism.

The research project was ethically vetted by a Regional Ethics Board. Informed consent was obtained from all participants ahead of any data collection, using either Arabic or Swedish.

5 Analysis

The analysis is structured into three sections, each showing a different way in which the epistemic stances of the participants unfold interactionally: mutual knowledge adjustments, the pregnant woman’s unsuccessful attempts to claim knowledge, and joint assertion of the midwife’s knowledge.

6 Conclusion

In this article, we have used an interaction analytical approach (Sidnell and Stivers 2014; Wadensjö 1998) and theories of power and knowledge (Du Bois 2007; Foucault 1980; Jaffe 2009; Mason and Ren 2012) to analyse patients’ possibilities for making their voices heard in interpreter-mediated interactions between Arabic-speaking pregnant women and Swedish-speaking midwives. The data used in the study consist of excerpts from partial recordings of 33 antenatal care consultations involving pregnant women and five different midwives. Three cases of knowledge negotiations have been identified and analysed: (1) the midwife and the pregnant woman mutually adjusting their knowledge claims, (2) the pregnant woman unsuccessfully attempting to claim knowledge and (3) participants jointly asserting the midwife’s knowledge.

The study confirms that interpreters should be considered participants in their own right (see e.g. Angelelli 2004; Mason and Ren 2012; Wadensjö 1998), as their actions influence the ongoing activity of health care consultations in various ways. Mason and Ren (2012) identify three ways in which interpreters enact interactional power: as co-interlocutors, to empower a disadvantaged party and through non-neutrality. The examples that they investigate all involve the interpreter taking very active stances themselves, with highly ‘divergent renditions’ (Wadensjö 1998), typically including whole contributions that have no correspondence at all to contributions made by the primary parties. The interpreters’ interactional power in the present study is enacted in a subtler way, typically consisting of brief expansions (e.g., ‘many’, ‘only’, a short laugh) of otherwise close renditions. However, as we have shown, such small modifications may have an impact on the unfolding of the interaction with regards to knowledge claims.

The knowledge claims of the pregnant women and the midwives in the data are highly dependent on the interpreters’ competence and performance. In investigating asymmetry of knowledge in monolingual doctor–patient consultations, Ariss (2009) identifies several different ways in which patients’ claims to knowledge are met by doctors, from non-engagement, through elaboration and disagreement, to agreement and symmetrical knowledge claims. With regards to the bilingual, tripartite consultations in the present study, it is very clear that the patients’ possibilities for making knowledge claims, and the midwives’ possibilities for meeting those claims, are highly dependent on what they hear from the interpreter. The skilled performance of the interpreters in Excerpts 1 and 3 means that the Arabic-speaking women can claim knowledge and be heard by the midwives as doing so, and the two can then reach a joint stance that either involves certain mutual adjustments (Excerpt 1) or agreement on the midwife’s stance (Excerpt 3). By contrast, the faulty interpretation in Excerpt 2 means that the Arabic-speaking woman’s knowledge claims and attempted rapport building fail. This means that the midwife is not given a reasonable possibility of meeting the woman’s knowledge claims at all, no joint stance is reached or even negotiated and the pregnant woman is left with inadequate knowledge.

Across the excerpts, we see that all three participants wield their ‘interactional power’ (Mason and Ren 2012) through various ways of negotiating knowledge, including requesting, obtaining and questioning knowledge. Additionally, the midwives wield their ‘institutional power’ (Mason and Ren 2012) by referring to scientific knowledge, their knowledge of Swedish antenatal care, etc. Thus, power and knowledge are achieved collaboratively, by all three participants, and they all both undergo and exercise power through interactionally achieved claims to knowledge (see Foucault 1980).

A limitation of the data used in this study is their fragmented nature. The midwives made frequent use of their ability to turn the recording off and on, and we have no knowledge of the parts of the consultations that were not recorded. Even so, this study has taken a qualitative approach, investigating how pregnant women, midwives and interpreters negotiate knowledge in antenatal care consultations. The three cases – mutual adjustment, patient’s unsuccessful attempts and joint assertion of the midwife’s knowledge – are not to be seen as representative of interpreter-mediated antenatal care consultations generally. Rather, the cases have been identified and analysed in order to show some of the ways in which power and knowledge may unfold and play a meaningful part in such interactions.

However, as a general remark it may be noted that across the whole dataset, there are very few cases of Arabic-speaking pregnant women making their voices heard, in the sense of creating or being given interactional possibilities and initiatives as explained in Section 4 above. All in all, only six such instances have been identified (of which three have been closely examined in the current study). Of course, this may mean anything from the Arabic-speaking women not receiving sufficient information (in line with previous studies of multilingual miscommunication and higher risks during childbirth and pregnancy, e.g. Esscher et al. 2014) to their potential questions and concerns mostly already being met by the midwives and the antenatal care programme. One possible way of receiving a partial answer to this would be to compare the current findings to the monolingual Swedish antenatal care consultations that we have also recorded as part of the project.

Transcription conventions

wo-

interrupted word

de=e

two words pronounced as one

word

emphasis

WORD

strong emphasis

o:

prolonged sound

°word°

weak voice

>word<

faster pace

<word>

slower pace

(0.6)

pause in seconds

≋

continued turn

⌈words

overlapping speech (one speaker)

⌊words

overlapping speech (the other speaker)

(words)

uncertain transcription

((comment))

meta comment

@

laugh