Response Practices in Multilingual Interaction with an Older Persian Woman in a Swedish Residential Home

Abstract

In the present case study, a care encounter between an older multilingual (Farsi/Swedish/English) Persian woman and staff in an ordinary, Swedish residential home is investigated. The woman is perceived as suffering from dementia symptoms, but has not received any formal diagnosis of the disease. More specifically, the study focuses on how the woman's contributions in her mother tongue, Farsi, are responded to by a carer, who is also multilingual and speaks Swedish as a second language (L2), but has a very limited knowledge of Farsi. The data consists of recorded material from a mundane morning activity in the residential home, as the woman is undressed and prepared to go to the shower. The method employed is conversation analysis, and the study addresses the interactional outcome of this type of multilingual encounters, highlighting the way the establishment of mutual understanding is negatively affected by the fact that the participants do not or only to a limited extent share a common language. Analysis of the data shows that most of the woman's contributions in Farsi are responded to in L2-Swedish by the carer, primarily by means of seven different response practices: soothing talk, instrumental talk, minimal responses, explicit expressions of understanding, mitigating talk, questions, and appraisal. The findings are discussed in light of new demands on Swedish (and Western) care- and health care systems to adapt to the increasing number of multilingual, older people, who will become residents in care facilities and attend day centers within the coming years.

Introduction

In the present study, a care encounter between an older multilingual (Farsi/Swedish/English) Persian woman and staff in an ordinary, Swedish residential home is investigated. More specifically, the study focuses on how the woman's complaints produced through nonverbal cues, and verbal contributions in her mother tongue Farsi, are responded to by a carer, who has a very limited knowledge of Farsi. The handling of the resident's complaints becomes problematic, since the carer, most of the time, is unable to address the actual content of the client's complaints, due to the lack of a common language.

In Sweden, the proportion of older people with a background other than Swedish is rapidly growing (www.scb.se ¹). As a result, the number of multilingual residents in older people's care facilities is expected to increase in a steady fashion during coming years. A similar trend can be observed throughout the Western world (Nielsen 2011). This raises new demands on care and health care systems, which are not always prepared for the challenges associated with cultural diversity. Some of these older residents are given dementia diagnoses, but very many remain undiagnosed due to cultural norms and values, language barriers, problems associated with interpreter-mediated clinical assessments (Nielsen 2011), or as a result of difficulties for the authorities in spreading information to immigrant communities about what kind of care that they are entitled to. The role that language, culture, and interaction play for multilingual speakers with dementia is acknowledged by The Swedish National Board of Health and Welfare that recommends that multilingual persons with dementia be given the opportunity to perform their religion, eat food that is adapted to their culture, and to get access to staff who speak the same language as themselves (Socialstyrelsen 2010-5-1, pp. 20–22). Despite these recommendations, multilingual speakers with dementia or dementia symptoms often live in ordinary Swedish residential homes, where the possibilities for cultural and linguistic adaptations may be sparse or even lacking. This is naturally also the case for many older immigrants without any cognitive challenges. For practical and economical reasons, it is not always possible to find staff who speak the same language as residents of a certain cultural origin. Even if Sweden has a fairly large population of immigrants (15 % of the country's population²), and the majority of people of a foreign origin tend to gather in the larger urban areas (Stockholm, Gothenburg, Malmö), there are limitations in what cultural and linguistic adaptations can be made in relation to the individual needs of older immigrants, particularly in rural areas, but also in the larger cities.

The lack of “cultural and linguistic matching” between staff and residents in care contexts creates very complex situations for all participants involved. A prerequisite for interaction to be satisfactory to interlocutors is that they manage to achieve a sufficient degree of collaborative understanding for their present purposes and actions (Clark 1996). The mutual interpretation of the ongoing talk is accomplished by the participants through their mutual orientation towards parsing talk into smaller units, signaling understanding, or requesting repair (Schegloff et al. 1977). Shared understanding thus refers to how participants collaboratively contribute to the conversation in ways that demonstrate that they are satisfied with the progression of the talk. An essential ingredient in the joint negotiation of meaning, and the establishment and maintenance of shared understanding, is the use of different kinds of responses (Duranti 1986). The importance of responses is stressed by Bakhtin (1981), p. 282, who states that “To some extent primacy belongs to the response, as the activating principle: it creates the ground for understanding, it prepares the ground for an active and engaged understanding.” Response practices in interaction involving participants who do not share a language are therefore of great relevance, as they very clearly reveal how mutual understanding is made manifest, as well as how understanding may be jeopardized. Staff in residential homes are obliged to care for and make sure that the residents are treated in ways that uphold their autonomy, integrity, and well-being (Socialstyrelsen 2010-5-1, p. 69) alongside making sure that routine issues such as eating, personal hygiene, dressing, and so on are executed in a way that maintains the dignity of the person in need of care. Problems related to a lack of shared understanding due to multilingual circumstances may have severe negative effects on the pursuit of well-functioning and person-centered care.

Studies that address response practices in staff–resident interaction in long-term care settings highlight the role of the conversational partner for the implementation of person-centered care (Carpiac-Claver and Levy-Storms 2007; Coupland et al. 1988; Grainger 1993, 1995; Grainger et al. 1990; Jansson and Plejert 2014). Grainger et al. (1990) emphasize the carers' modes of responding to the older person's telling of personal troubles and discomforts, arguing that the degree of empathy and engagement will impact significantly on the older person's well-being. They show how the resident's troubles are constantly renegotiated to a less serious level or dismissed through compliance-gaining strategies and evasive techniques. In a similar vein, Carpiac-Claver and Levy-Storms (2007) stress the importance of nurse aides' communicative skills in providing emotional support for cognitively impaired residents. This is discussed in terms of institutional conflicts between residents' needs for emotional support and task-oriented goals (Backhaus 2008; Heinemann 2009). Nurse aides are shown to resort to a range of response strategies in the course of task completion, e.g., face compensation and minimization, that all function to close residents' painful self-disclosure. The authors argue that affective communication in responses to residents' troubles-telling, like soothing, comforting, and reassuring talk, in reality is task-oriented. The ultimate aim of such strategies is to reorient the resident to the legitimization of the task. This is also shown by Grainger (1993), in a study of carer–caree relationships where a bath is constructed as lovely and nice in opposition to the resident's expressed opinion. As a contrast, Jansson and Plejert (2014) show how carers, by making small, step-by-step adaptations to the needs of care recipients, in ways that make the recipient experience a sense of autonomy and involvement, are positive when carrying out a potentially imposing task, such as taking a shower.

There is also a rich body of research on communication strategies of care providers of people with dementia diseases, including practices employed by family caregivers (e.g., Small et al. 2000; 2003) and strategies used by carers in residential facilities (cf. Vasse et al. 2010; for a review, Lindholm 2010) as well as interactional devices used by clinicians in memory clinics (Saunders 1998; Saunders et al. 2011). Some such studies have an interventionist approach, providing caregivers training in using certain communication strategies in order to enhance interaction with the person with a dementia disease (Ripich et al. 1998, 1999).

It should be noted that all studies referred to above, apart from Nielsen (2011), concern monolingual and monocultural care encounters. It is therefore of relevance to also scrutinize less culturally and linguistically homogeneous care encounters.

Research on Multilingual Interaction in Older People's Care

Although the present study deals with an older multilingual woman without any formal diagnosis of dementia, research on multilingualism and dementia is of great relevance for the case at hand, since a reported observation is that, for this group, second languages are lost first, and the mother tongue is affected at a later stage (Mendez et al. 1999; Paradis 2008). However, the survey below is also relevant concerning older immigrant populations in a broad sense (those who are “healthy” as well as those with various cognitive/communicative impairments), as it illustrates the prospects and consequences of multilingualism and shared understanding in different care and health care settings.

Considering the fact that not only the number of multilingual people with dementia in need of older people's care (e.g., in the shape of residential homes and/or daily activities) is increasing throughout the world, and the fact that staff working with old adults in many countries are second language speakers, surprisingly little research has been carried out on interaction involving multilingual people with dementia. Some of the first studies date back to the late 1980s and mid-1990s, and mainly concern features of language choice and language separation (Hyltenstam and Stroud 1989; Hyltenstam 1993), i.e., abilities related to choosing a language appropriate for the current setting and participants (De Santi et al. 1990; Luderus 1995). Code-switching patterns in people with Alzheimer's disease have also been approached using conversation analytic methodology (Friedland and Miller 1999).

For research on interaction involving multilingual people with dementia in residential care settings, the work by Ekman (1993) and Ekman et al. (1993, 1994) must be considered pioneering. Ekman investigated communication between Swedish monolingual and bilingual (Swedish/Finnish) carers and Finnish-speaking people who had migrated to Sweden. The major findings by Ekman et al. concerned the benefits for bilingual Finnish people in Sweden to have access to carers who spoke their mother tongue, since the analysis of interaction with caregivers speaking Finnish showed that the bilingual people with dementia were not as badly off when interacting in their native tongue, as they were perceived when they communicated with a Swedish-only-speaking caregiver. In practice, the results of this work amounted in the development of ethnically profiled care facilities, directed towards the large population of Finnish-speaking people residing in Sweden.

Jansson (2012a, b, 2014) has written a number of articles based on her fieldwork in a residential home, combining observation and interviews with detailed analyses of audio- and video-recorded data. Jansson's work has primarily focused upon the perspective of multilingual staff and how they employ various strategies to cope with and get past the language barriers that arise in interaction with residents with whom they do not or only partly share a language. In Jansson's studies (Jansson 2012a, b, 2014), there are also care workers who share the language of some of the residents, which enables a comparison between care workers with no or limited fluency in the residents' native tongue, with care workers who have fluency. Jansson's findings are in agreement with that of Ekman (1993) and Ekman et al. (1993, 1994) in that the linguistic and cultural competence of care-giving staff is vital for creating a trustful relationship with multilingual residents, which in turn has a positive effect on their sense of identity and well-being. Another finding by Jansson (2012a, b, 2014) concerns the ways in which the care workers with no or limited language competence develop strategies that positively affect their interaction with residents. Such strategies comprised learning certain frequently occurring words in the language of the resident, recycling elements in the resident's prior turn, and singing phrases in the native tongue of the resident. This use of playful language enhanced the care workers' social relationship with residents, making them more compliant with the everyday activities that had to be carried out.

Research on multilingual interaction involving people with dementia outside of Scandinavia is surprisingly sparse. This is perhaps due to the fact that few studies limit their scope just to people with dementia. Instead, multilingualism and ethnicity are approached from a more broad and workplace-oriented perspective comprising staff working with older persons (where dementia is not in focus) in a range of care and health care settings (e.g., Hatton 1992). Multilingual issues are sometimes also noticed in passing, but not explicitly addressed (e.g., Carpiac-Claver and Levy-Storms 2007; de Bot and Makoni 2005). There is also some research within the vast field of interpreted-mediated interaction that acknowledges the specific problems of multilingual people with dementia and other neuropsychiatric disorders, for example, in being adequately examined and diagnosed (Nielsen 2011; Paradis 2008).

Participants, Data, and Methods

The focus of the present case study is a Farsi-speaking, Iranian woman in her 80s (here called Mehri) living in an ordinary residential home located in a suburban area in Sweden. The woman, who had immigrated from an urban area in Iran to her relatives in Sweden, was bedridden after a right-hemisphere stroke. She often complained of having pain in her left arm and shoulder that were impaired by the stroke. Dementia symptoms, such as agitated behavior, disorientation, and language loss had gradually begun to develop after the stroke, but she had never been formally diagnosed. This latter issue may be due to the fact that families of foreign origin, for example Persian, do not always conceive dementia as a disease (Mazaheri 2013), which may also contribute to few people of Persian background being referred to Swedish memory clinics. In addition, simply the issue of leaving a parent in a residential home rather than taking care of them at home is for many people of Persian and Arabic background associated with shame.³ In this particular case, staff gave accounts that suggested that Mehri suffered from dementia, based on their experience of residents with the disease, but Mehri's relatives opposed to having her moved to a dementia unit. Farsi was Mehri's dominant and preferred language, and she had never achieved a high level of proficiency in Swedish, a language that she had learned during old age in a basic language course for newly arrived immigrants. She had been a fluent speaker of English, something that was related to her former occupation in Iran as a teacher of that language. During care, she kept to her first language and only randomly responded to her carers in Swedish, although she seemed to understand the language to some extent. She was even more resistant towards using English.

The data on which the current case study is based is drawn from a larger project on immigrant care workers' communicative practices in three residential homes in Sweden. Data collection methods included participant observation, audio/video recordings, and interviews with the staff. Ethnographic research was conducted by one of the authors several hours each week over the course of 1 year (in the unit that hosted the Farsi-speaking woman; over the course of 6 months). Participant observation included shadowing care workers during their shifts at work and spending meals and breaks with them.

The current study is based on one single audio recording made during morning care in the resident's room. The carers' interaction with Mehri constitutes the main data, but ethnographic field notes and interviews have also informed the analysis. The two carers who were in charge this morning were both second language (L2) speakers of Swedish. Neither spoke Farsi, and they had in fact no knowledge of the language except for a few words. At the time of the recording, Mehri had spent 6 months in the residential home. Staff reported that during that time, Mehri's status had become worse. Stress-related symptoms, like agitated behavior occurred frequently during care. This was not only due to her physical impairment, but also to the fact that she was generally resistant to being cared for in a residential home.

The analyses below are based on an episode of the undressing and initial cleaning of Mehri, before she is taken to the bathroom for a shower. There are two carers present. However, one of the carers comes and goes during the event, and the majority of talk in the episode is exchanged between Merhi and a carer called Sarah.

As mentioned in the introduction, the primary focus of the study is response practices employed by a carer when interacting with Merhi. Here, responses are viewed through a conversation analytic (Hutchby and Wooffitt 2008; Schegloff 2007) lens. Conversation analysis has its origin within sociology and ethnomethodology (Heritage 1984). It shares its focus on interaction with discourse analysis, however primarily focusing on micro- rather than macro-levels of social encounters. Although conversational analysis (CA) originally focused on mundane social interaction in typical speakers, it has subsequently also successfully been applied to analysis of interaction involving people with various communicative disabilities, e.g., aphasia and dementia (e.g., Goodwin 2003; Kitzinger and Jones 2007; Lindholm 2008, 2010).

From a CA perspective, utterances are linked to and form a response to something that someone else has previously said, at the same time as they shape the context for new contributions (Heritage 1984, p. 242). Thus, a response sets the agenda for what is to follow, at the same time as it is shaped by its prior context. A response may be treated as preferred or dispreferred by participants in an ongoing interactional event (Schegloff et al. 1977). This can be exemplified by the fact that a question is normally responded to by an answer, that a request for confirmation ordinarily is followed by a confirmation, and so on. If such patterns for some reason are violated (e.g., by problems in speaking, hearing, or understanding), participants engage in “repair,” i.e., methods to resolve interactional trouble (Schegloff et al. 1977). As the analytical section will show, responses that are treated as dispreferred by the care recipient represent a recurring problem throughout the care encounter.

Within CA, researchers approach data with no pre-conceived ideas on what to find (Heritage 1984). Instead, a bottom-up approach is employed, where recurring patterns and practices are noticed after careful and systematic viewing and transcription of video and audio recordings. In order to validate phenomena, transcriptions and recordings are presented in data sessions, i.e., seminars in which experienced and novice CA researchers gather to critically assess each other's categories and analyses. The response practices presented in the current article have been presented and assessed in data sessions at three occasions. At each session, between three and six experienced and highly skilled conversation analysts were present, and categories have been revised in accordance with their viewpoints.

Transcription conventions follow CA standards and are adapted from Ochs et al. (1996) with some modifications. A transcription key is found at the end of the article. Translation of material in Farsi was made by one of the authors, and translations are inserted just underneath each line of original language in the transcript.

The study was approved by the Regional Board for Ethical Vetting in Stockholm. The staff, the residents, and their relatives were informed by means of a letter about the aims of the study, and their rights as participants. To protect the participants' identities, all names are pseudonyms.

Analysis of Response Practices During Morning Care

Subsequent analyses present how the carer, Sarah, responds to the contributions to talk in Farsi that are produced by the resident, Mehri. The episode begins when Sarah starts undressing Merhi, in order to clean her. She then prepares her to make her go, by means of a lift in the ceiling, from her bed over to her wheelchair.

Below, Sarah's response practices have been categorized based on the interactional “job” that they perform, or in this case, rather the “job” they aspire to perform. Ordinarily, from a conversation analytical perspective, when establishing the function of a contribution to talk, it should be related to preceding as well as subsequent turns, in line with the next turn proof procedure (Sacks et al. 1974). However, this is only possible to a limited extent in the present context in which it is sometimes hard to establish to what extent an utterance is at all comprehended by the persons interacting. There are, of course, many other cues apart from speech that contribute to the common understanding of the participants, such as body movements, facial expressions and the like, as well as the fact that certain activities, such as being cleaned and taking a shower, are routinely carried out in the residential home. As there is no video for the present episode of interaction, the categorization is based on what can be perceived from the audio tape, complemented by information from ethnographic observations. Thus, the taxonomy of responses presented here, taking into account the complexities mentioned above, is based partly on how the response practices systematically recur as observable patterns of participants' orientations, and partly on interactional cues, such as form and intonation, e.g., prolongation of sounds, and fall and rise in intonation. Prosodic features of talk, including speech rate, were determined utilizing the Praat 5.3.56 software (http://www.fon.hum.uva.nl/praat/) in order to control perceptual analyses performed by the authors.

In total, the episode consists of 70 responses by Sarah. Analyses of these responses revealed seven different practices: soothing talk, instrumental talk, minimal responses, explicit expressions of understanding, mitigating talk, questions, and appraisal. These practices were not evenly distributed as some were more commonly used than others. Figure 1 above presents the distribution of each type of response practice. The distribution is calculated dividing each type of response with the total number of Sarah's responses (n = 70) that occurred during the investigated episode.

Fig. 1

The distribution of response practices calculated in percent of Sarah's total number of responses during the investigated episode: soothing talk (34 %), instrumental talk (30 %), minimal responses (22 %), explicit expression of understanding (3 %), mitigating talk (4 %), questions (3 %), and appraisal (4 %)

Soothing talk

The most frequently occurring way in which Sarah responded to Mehri was to try to calm her down (34 % of Sarah's total number of responses to Mehri). The most common form of attempting to calm Mehri down was to use the Swedish expression “så::: ja” (there) almost exclusively produced with a prolonged vowel and falling intonation. The expression was very often also ended by the addition of Mehri's name, i.e., “så::: ja Mehri.” (there Mehri.).

Excerpt 1 below comes from the initial phase of the recording. Mehri is at this point lying on her bed and the carer attends to her, attempting to undress her; an activity that Mehri objects to⁴:

In the beginning of the excerpt (line 8), Sarah is explicitly talking about her current physical action (cf. “instrumental” talk, below). This action is met with resistance from Mehri's part (line 9) as she refers to having booked someone else and that that person will turn up. According to ethnographic information, Sarah had no knowledge of Farsi; however, she seemed to understand a few words every now and then. One such word was the negation “na” (line 9), so it is likely here that she is able to perceive that Mehri is verbally (alongside other signs of resistance) objecting to the activity at hand, although she might not understand the very details of Mehri's contribution (the claim of having booked someone else). Her way to address this resistance is to try to sooth Mehri (lines 10, 15, and 18), producing prolonged vowels and falling intonation. At one point, she also gives praise to Mehri “a så rätt” (a that's right) (line 13).

The responses exemplified in excerpt 1 conforms to what Grainger (1993) has labeled as “nurturing discourse,” which puts nurses into a “motherly role” and patients in a “dependent role” (1993, p. 252). According to Grainger, this kind of talk is also often embedded in the management of routine tasks. This is also very much the case in the present data, as Sarah's soothing responses are used to calm Mehri down in order to make her more compliant to the tasks to be carried out. In a study by Carpier-Claver and Levy-Storms (2007), the positive effects of nurses' affective talk (including soothing talk and praise) to residents with dementia were brought forward. Interestingly, in their study, affective talk only sometimes coincided with more task-related activities. In the present episode, it appears that the affective talk to a great extent is used for instrumental purposes, i.e., making Mehri calm in order to be able to proceed with the task of undressing her. This difference is very likely due to the activity at hand. Carpier-Claver and Levy-Storms (2007) investigated meal time interaction, which presumably is a context which allows for communication that is more socially oriented than the undressing and cleaning situation currently described. Similarly, Chen et al. (2000) highlight the value of carers reducing their focus on routine tasks and instead engage in socially oriented talk; in that way providing care that not only is more person-centered but may also reduce residents' agitated behavior. However, in the present context, engaging in social talk is made very difficult due to the language barriers between interlocutors.

The cues for understanding that Sarah has access to are Mehri's prosody, her body movements, facial expressions, and a few words that Sarah might understand. All cues signal Merhi's unhappiness with and resistance towards the activity. It is understandable that under such premises, trying to calm seems to be a sensible option, as it at least displays some sympathy with the resident's agony. In research on responses to trouble's talk in monolingual contexts, deflective strategies may be used, i.e., attempts to decrease the seriousness of the trouble expressed by the patient (Grainger et al. 1990). Although far from ideal, as such strategies trivializes the concerns of the older residents, in comparison to the soothing responses, and as will be shown below, the use of minimal responses, deflective responses at least display that the carer has understood what the resident has said; a feature that is in question in the current multilingual context.

Instrumental talk

The second most frequent response practice by Sarah (30 % of the total amount of Sarah's responses) consisted of instrumental talk, i.e., telling Mehri what was going to happen, what they were doing, asking Mehri to move her body and limbs in certain ways, etc. The talk in excerpt 2 below, takes place as Sarah attempts to make Merhi more positive to go to the bathroom, producing a self-repair in which the word “shower” is exchanged with a term for bathing in Farsi, “hammam” (lines 96, 99):

Mehri is occupied with a recurring theme of her talk of having booked someone else to clean her (cf. line 9, excerpt 1). After a brief pause (line 95), the carer comes in and uses the term “hammam” (bath) instead of her previous use of the word “dusch” (shower). Here, the carer draws from the linguistic resources that she has, perhaps hoping that the word in Farsi might have a positive effect on Mehri's attitude towards being cleaned. It is perhaps also possible that the word “bath” has a more positive ring for Mehri, since there might be cultural aspects underlying the elderly woman's reaction to the idea of a shower (cf. Twigg 2000, on how elderly persons may associate baths with pleasure and showering with discomfort).

In relation to the self-repair conducted by Sarah in excerpt 2 (lines 96, 99), one might claim that she is in fact trying, despite her limited language understanding, to attend to Mehri's objection, using vocabulary that is recognizable for Mehri and perhaps puts the prospective shower in a brighter light. It does, however, not get the wished-for effect here, as Mehri continues to object to the activity (lines 98, 100, 102). It is noteworthy that Mehri, on her hand, is attentive to the fact that Sarah has just used both the word “dusch” (shower ) and “hammam” (bath), and is, in fact, conducting a similar replacement of terms as the carer has just done, as she also changes “dusch” into “hammam” in her counterargument against Sarah (line 102). This also reveals that Mehri's problem, at least the way she repeatedly puts it, is not so much the shower, or bath, but who is to assist her with it, i.e., someone that she herself has accepted. Thus, despite Sarah's attentive try, it nonetheless does not address the core of Merhi's troubles.

Minimal responses

Minimal responses (e.g., Hutchby and Wooffitt 2008), or back channels (Yngve 1970), permeate ordinary talk-in-interaction and primarily function as a display of understanding, alignment, and of active listenership. Often, they are accompanied by nonverbal cues such as nodding and/or an attentive gaze by the person who produces them. In the present data, there are two types of minimal responses, distinguished by their prosody. The most frequent kinds of minimal responses (14 % of the total amount of Sarah's responses) that the carer uses are “mm,” “jo jo” (yea yea), “aa” or “ja ja ja” (yea yea yea) with falling or level intonation. The other type (8 % of the total amount of Sarah's responses) is a “jaa↑” (yea↑) or “mm↑” with a fairly strong rise. Excerpt 3 below displays the use of the first type of minimal responses:

Excerpt 3

Excerpt 3 demonstrates how a complaint may be responded to simply by means of minimal responses. The minimal responses indeed show that the carer is listening, but they do not display much more than that, and they do not attend to the problems or objections expressed. Of course, a complaint need not always be explicitly addressed, and Sarah's response (line 52) “jo jo” (yea yea) with a slight fall works reasonably well as a sign of understanding and recognizing Mehri's objection towards being undressed (lines 50, 51). However, the response is not accompanied by any real compliance with Mehri's objection, since the carer is obliged to continue with her task. The second minimal response “a a” (line 56) with level intonation appears disengaged, particularly considering the seriousness of the trouble expressed by Mehri in the preceding turn, feeling filthy when being washed by the carers in the nursing home (lines 53,54,55). Again, as Jansson (2012a, b, 2014) as well as Grainger et al. (1990) have shown, carers who do share a language with the resident seem more prone to attend also to the actual reason for their distress, although perhaps attempting to minimize and decrease the seriousness of the trouble. For Sarah, this is not possible due to her limited understanding of Farsi, and she tries to make the best of the situation, providing Mehri with tokens that show that she is listening. A problem that arises from this action, however, is that feedback of this kind, with falling or level intonation, may also be perceived as if what has been said has also been understood and not only “listened to.” Thus, there appears to be a mismatch here in what would be a preferred response in relation to a complaint or expression of distress. One might also reflect upon whether it may even be the case that this kind of minimal response, as it does not attend to Mehri's expressions of distress, contributes to her many repetitions. Alternatively, it should not be forgotten that Mehri is perceived as suffering from various dementia symptoms, and her repetitious verbal behavior (Guendozi and Müller 2005) could thus be a sign of the disease. However, as Mehri's recurring phrases in the particular episode analyzed here, most of the time appear relevant, both in relation to the activity taking place, and to Sarah's talk, it is very hard to say anything more precise about the possibly pathological nature of her repetitions. This issue will, however, be returned to briefly in the discussion part of this article.

Minimal responses with level or falling intonation were primarily found in the initial phases of the undressing and cleaning episode. As the carer and Mehri get closer to the shower, some minimal responses with a rise occur:

Excerpt 4

In Swedish, the use and interpretation of a minimal response with a rise is highly dependent on the context in which it occurs. The rise may signal a strong engagement in what someone is just saying, be an answer to someone seeking a person's attention, and it could also be used in situations where a speaker signals listenership, but in fact is more focused on something else. At a first glance, the minimal response in line 103 looks like Sarah is displaying her engagement in what Mehri is saying in her prior turn (lines 101,102). However, ethnographic field notes and observations state that Sarah is at this point managing the lift, used to assist in moving Merhi from her bed onto her wheelchair, so it is perhaps more apt to assume that the minimal response is used as a sign of listenership, rather than a case of strong engagement in the content of Mehri's talk (and again, Sarah does not understand the Farsi).

The response in line 106 occurs in relation to a turn in which Mehri has addressed Sarah, using her name (do not fake the thing (0.6) Sarah) (line 105), so Sarah's response might be an answer to Mehri's seeking her attention. Sarah's name occurs after a short pause and is produced with a slight rise as for calling upon her. Although Sarah is attending to the practical matter of the lift, she does acknowledge this summons. That she interprets Mehri's use of her name as a call for attention is also evident in the way that she explicitly asks Mehri to repeat what she was saying (sorry what did you say Mehri) (line 108). This way of acknowledging that Mehri is calling for her attention thus displays Sarah's ability to uphold the basic structure of the turn-taking, although she is unable to decipher the rest of the content of Mehri's talk.

The last case of a minimal response produced with a rise (line 111) appears similar to the first instance (line 103) that displayed Sarah's listenership and perhaps also its relation to her engagement in the current physical task of the lift.

Explicit expressions of understanding

Despite the fact that Sarah did not understand but a few words of Farsi, a few times she explicitly phrased that she understood Mehri (3 % of the total amount of Sarah's responses). In excerpt 5, the reality of the shower is coming closer, and Mehri is objecting to the upcoming activity, continuously repeating I am not gonna allow anyone bathing me (lines 130, 133, 136):

The first instance of objection from Merhi (line 130) is not verbally responded to, apart from the audible sound of an inbreath from Sarah (line 131). The inbreath is followed by a pause of 1.0 second, and Merhi's word-by-word repetition (line 133) of her previous contribution might be due to the fact that she does not receive any response that attends to her concerns. However, as she begins recycling her previous utterance, right after the pause, Sarah comes in on overlap, saying “sådär Mehri” (line 134). The term “sådär” (there you go) occurs twice in excerpt 5 (lines 134, 137) and may serve two purposes. On the one hand, it can be interpreted as soothing and synonymous to “så” and “så:::ja” that were presented in excerpt 1. On the other hand, “sådär” may also be interpreted as more explicitly commenting on the progression of some task and specifically to signal its termination (Andrén 2010). Here, “sådär” is produced with a fall, perhaps being both calming at the same time as it addresses the progression of the task at hand, since the carer is now occupied by moving Mehri over to her wheelchair. “Sådär” in both line 134 and 137 might thus be employed as a comment on every part of the move that the carer perceives as finished.

Of specific interest in this excerpt, however, is the carer's use of the phrase “I understand” (lines 137, 139). At this point of the talk, Mehri is repeating the same phrase concerning not allowing anyone bathing her (lines 130, 133, 136, 140, 141, 142) over and over again. Her speech rate is higher than it has been up until this point of the interaction, and overlap is hard to avoid. Only once is the carer able to speak almost, but not entirely, in the clear (line 137). Sarah's response (line 137) starts with “sådär” and is immediately followed by “ja förstår de” (I understand that), Merhi's name, and a request for her to wait a little. Mehri, however, continues repeating that she does not allow anyone bathing her, in overlap with Sarah's second claim that she understands Mehri (line 139). Sarah's use of the pronoun “de” (that) is particularly interesting, since that must refer to some antecedent, plausibly the content of what Mehri is saying. In her second use of the phrase, I understand (line 139), it is also preceded by the feedback token “ja” (yea) which presumably also is a sign of addressing what Mehri is saying, together with the that. As Sarah does not understand Farsi, it is unlikely here however that she is really addressing the contents of Mehri's complaint. Rather, the display of understanding might be used as a strategy to respond in a more preferred way to a person's display of dissatisfaction and objection than, for example, just using soothing talk, or a minimal response. This plausible “pretended” understanding does, however, not result in any change in Mehri's repetition of not allowing anyone to bathe her. The understanding explicitly expressed by Sarah may of course also be used as a way of signaling her general understanding and sympathy with the way Mehri is feeling.

Mitigating talk

Sarah's response to Mehri's resistance displayed in excerpt 2 did not only deal with the activity at hand, i.e., telling Mehri that she was to have a bath. It could also be viewed as an attempt from the carer to mitigate an activity that Mehri finds unpleasant. In the data, there were a few other such instances of mitigation (4 % of the total amount of Sarah's responses). The carer could, for example, claim that the shower must be done “lite snabbt” (a bit quickly) (lines 152,153):

Apart from mitigating the upcoming shower in terms of describing it as something to be done a bit quickly (line 152), and thus soon to be over and done with, a persuasive strategy can also be observed as Sarah tries to make Mehri more positive to having the shower by telling her that her son, Ali, will be visiting that morning (lines 152,153). Like all other attempts from Sarah to make Mehri more compliant to the shower, these practices also fail. It might perhaps even be argued that the mitigation, and the argument about Ali, reinforce Mehri's objections, since, as has previously been shown, she is questioning who is to clean her, not the cleaning in itself. Relating to this fact, and presuming that Mehri understands what Sarah is saying in lines 152 and 153, her challenging expression well then why do you put me here (line 155) makes perfect sense. This challenge is not picked up by Sarah, however, who responds using the calming “så ja Mehri” (line 156) that she so frequently employs in order to sooth Merhi.

Questions

In the episode analyzed here, there were two instances (3 % of the total amount of Sarah's responses) where Sarah turned to Mehri, asking her something. One of these cases can be observed in excerpt 4, where she (line 108) makes an apology and asks Mehri what she just said (sorry what did you say Mehri), as a response to Mehri's calling for her attention. Mehri, however, does not provide a repetition of the troublesome turn, which otherwise would be the most common way of responding to someone who signals difficulties hearing (Couper-Kuhlen 1992; Drew 1997), but continues adding to her contribution until it is completed (line 110). It should be noted that Mehri conducts a self-repair, exchanging touching me with bathing me (lines 107, 109, 110), but this repair does not seem related to Sarah's request for repetition, since her question overlaps with the repairable touching me.

The second question occurs a bit further into the talk, when Mehri objects to leaving her room:

Excerpt 7

It can be observed that Mehri here again is occupied with repeating the phrase of refusing to let anyone bathe her (line 144). Sarah responds to this utterance by means of a contribution that is mitigating, as she uses the term “bara” (just) in we will just put you on the chair attempting to minimize the action of moving Mehri. The mitigating utterance, however, is met with continued resistance from Mehri's part (line 146). Mehri's objection (line 146) is responded to by Sarah, by means of a question that could be interpreted as either a request for clarification, or confirmation of Mehri's prior turn. Interactionally, Sarah's question is “well-formed” in terms of what one would expect from a prototypical confirmation check or clarification request, as it recycles an element (come) from Mehri's contribution (no I don't come) and acknowledges the fact that Mehri does not want to come, which in turn projects either a confirmation or a clarification from Mehri. As can be observed (line 149), Mehri confirms Sarah's query with a simple “na” (no).⁵

Excerpt 7 demonstrates one of very few occasions during the whole episode analyzed, where participants seem to share a “satisfactory” degree of understanding, although they are not in agreement. They are, in fact, orienting towards a mutual communicative project (Linell 1998) of sorting out an issue of understanding (by means of a confirmation-check sequence). In contrast to the other types of response practices that Sarah employs, her confirmation-check entails elements from Mehri's prior utterance, thus displaying to Mehri that she has actually picked up in greater detail what her objection is really about. Since Mehri understands some Swedish, although she very rarely produces any spoken output of the language, in this particular sequence her confirmation of Sarah's check contributes to the picture of the interlocutors' now sharing a mutual communicative goal. It is also noteworthy that in contrast to most of the talk that precedes this particular sequence, and most of what follows it, Mehri here, for a moment, discards her persistent repetition of the phrase about not letting anyone else bathe her.

Appraisal

The last analytical category that emerged in the data was appraisal from the carer (4 % of the total amount of Sarah's responses), and in all cases it had to do with Mehri's succeeding in doing something related to the progression of getting to the shower, e.g., moving from the bed into the wheelchair:

Excerpt 8

As can be observed in excerpt 8, which occurs towards the end of the episode scrutinized, Mehri is again strongly objecting to being washed, and she has returned to repeating the fact that she does not allow “anyone” bathing her (lines 173, 174). As she finally reaches a point where Sarah is able to assist her over to the wheelchair, Sarah expresses her appraisal with increased loudness (line 177). The appraisal is responded to in a way that displays a prosodic adaptation as concerns loudness, and Mehri is clearly getting increasingly upset. Sarah continues with emphasis (perhaps trying to encourage Mehri), as she enthusiastically says “DÅ KÖR VI” (THEN LET'S GO), with increased loudness and pitch (line 179). This last utterance is strictly oriented to the instrumental task of heading towards the shower room, and Mehri's strong objections are now ignored (Sarah at this point also has the advantage of Mehri finally sitting in the wheelchair, which puts her into a more favorable position to eventually become able to carry out the task that she is required to, i.e., to get Mehri showered).

Concluding Discussion

The present article has investigated different response practices employed by a carer in interaction with a Persian woman in a Swedish residential home, as the mundane morning routine of undressing and cleaning is to be carried out. The episode was scrutinized with a particular focus on the fact that the setting was multilingual, and that there was a lack of linguistic matching between participants, as the carer only had a very limited understanding of the woman's mother tongue, Farsi, whereas the Persian woman appeared to understand some Swedish, but preferred not to use it verbally in interaction with the carer. The woman was also perceived by staff and relatives, as suffering from dementia symptoms, which might affect the interaction in different ways. The woman had, however, never received any formal dementia diagnosis.

Analyses revealed seven different response practices that the carer, Sarah, used in her interaction with the Persian woman, Mehri: soothing talk (34 %), instrumental talk (30 %), minimal responses (22 %), explicit expressions of understanding (3 %), mitigating talk (4 %), questions (3 %), and appraisal (4 %). Although the data analyzed only comprised one rather short episode of interaction, it is argued here that the frequency of occurrence of the different response practices nonetheless is of interest, as it says something about the potential response preferences of the carer in “managing” this linguistically asymmetrical situation.

The management of routine tasks in care facilities, e.g., helping an older person to get clean and tidy, often entails elements of “imposition” to the old person, and it is a great challenge for staff to be able to conduct such tasks in ways that maintain the elderly person's feeling of autonomy, sense of “self,” and dignity (e.g., Kitwood 1997). This does not only go for care for people suffering from dementia but for any situation involving an older person in need of assistance with everyday tasks and activities, e.g., due to limitations in physical abilities. When dementia strikes, the situation increases in complexity, since the disease hampers the person affected cognitively and linguistically in ways that might make it very hard for carers to adapt to the potential wishes and viewpoints of the person with dementia or to carry out tasks against the will of the person with dementia in a way where direct argumentation can be avoided, i.e., not “walking all over,” or simply ignoring the person (Jansson and Plejert 2014). In this article, we have shed light on a third context that adds on to this complexity; a context that is becoming increasingly common in Europe, i.e., multilingual care encounters where the lack of linguistic knowledge adds a further burden onto the already heavy load of carers as well as the multilingual persons being cared for, in establishing a decent quality-of-life where values of human rights are maintained. The importance of this latter issue is also explicitly expressed and stressed in the guidelines for dementia care as stated by the Swedish National Board for Health and Welfare (Socialstyrelsen 2010-5-1).

Already 20 years ago, Ekman (1993), and Ekman et al. (1993; 1994), and more recently Jansson (2012a, b, 2014) stressed how the language competence of care-giving staff is vital for creating a trustful relationship between carers and residents, which positively affects the identity and well-being of the elderly being cared for. The research by Ekman et al. (1993; 1994) even suggested that there was a risk that older bilingual people might be perceived as more heavily affected by their dementia disease than they actually were, when they interacted with people who did not speak their mother tongue (Finnish). In the present article, there is ample proof of the ways in which the lack of linguistic matching of the participants contributes to a situation that might be experienced by Mehri as highly imposing. As she repeatedly expressed her objection to the shower by means of the phrase “man nemizaram kesi mano beshore” (I am not gonna allow anyone bathing me), it was continuously responded to in ways that did not explicitly address the more precise nature of her troubles: In the majority of cases, soothing talk was employed. This strategy naturally addressed Mehri's overall sense of distress, and it did prosodically unveil an emphatic stance on the part of Sarah, but it did not deal with the more precise details of the actual problem expressed by Mehri. It did not even “deflect” (Grainger et al. 1990) Mehri's problem, since the carer did not have access to the linguistic means for doing so. Similar things can be said about the carer's use of minimal responses. Although minimal responses ordinarily are used to display an interlocutor's active listenership and engagement with what someone is saying (Hutchby and Wooffitt 2008), in the present context, and in relation to the actual content of the objections expressed by Mehri, they appeared out of place and might be considered as a dispreferred response to a complaint. As previous research shows (e.g., Grainger et al. 1990), this kind of troubles talk in monolingual interaction between carers and older persons is ordinarily handled by trying to minimize the seriousness of the trouble. Although far from ideal, what the deflective strategies do, in contrast to the minimal responses, is that at least they display the carer's acknowledgment of the trouble, as it is verbally dealt with “somehow.” The minimal responses, again, displayed that Sarah was an attentive listener, but the actual trouble was never addressed. Some minimal responses were prosodically produced with a rise. Most such responses seemed related to the fact that Sarah was busy with a physically demanding task that required her attention. In these cases, the responses also displayed listenership, but did not deal with Mehri's actual trouble.⁶

Instrumental talk, i.e., when the carer explains what she is doing, or mentions the upcoming shower, is of course necessary in order to inform the resident about what is going on. This kind of talk was also fairly frequent in the episode analyzed. It may of course be discussed to what degree instrumental talk may be viewed as “responses,” and perhaps they should rather be viewed as “initiatives” from Sarah's part. In this article, however, they have been described in terms of responses, since from Mehri's point of view, they occur as a next turn in relation to a turn of trouble that she has just uttered. From Sarah's point of view, however, it is perhaps slightly misleading to call them responses.⁷ However, what is of relevance is that when Mehri's objection is followed by a piece of instrumental talk, her trouble is not acknowledged. It may perhaps even be suggested that her distress increases, as Sarah's instrumental talk concerns exactly what Mehri resists; the fact that she is being undressed and taken to the shower by Sarah. Something similar happens in cases of “appraisal.” As was displayed in the analytical section (excerpt 8), despite expressing her objections, and becoming increasingly upset (as shown by increased loudness and pitch), as Mehri cooperates physically so that she finally ends up in the wheelchair, she is praised, despite her continuous verbal protests.

The fact that Merhi is perceived as having dementia symptoms is also of relevance here. In the episode analyzed, it is very hard to determine whether or not any such symptoms are displayed. Mehri's agitated behavior could naturally be considered as one, but being anxious about being assisted with intimate tasks (cleaning, showering, undressing, etc.) has also been acknowledged in residential care involving elderly persons without dementia or other cognitive impairments (e.g., Twigg 2000). Mehri's repetitious behavior could be another symptom of dementia (Guendozi and Müller 2005) but need not be. In the current episode, Merhi's repetitions make sense in relation to the activities at hand and also in relation to the fact that the essence of her complaints is never thoroughly addressed. And even if the repetitions were related to her potential dementia disease, the use of minimal responses, soothing talk, etc. is not beneficial for “breaking” the repetitious pattern. In a few studies, it has been shown that repetitious talk by speakers with Alzheimer's disease may be positively or negatively affected by the ways in which it is responded to by interlocutors (e.g., Hydén and Örulv 2009, 2010; Hydén et al. 2012). Returning to the work by Ekman (1993) and Ekman et al. (1993, 1994), it is thought-provoking to consider that the linguistic asymmetry, and non-comprehension of Mehri's mother tongue, is perhaps reinforcing the perception of her as suffering from dementia symptoms, or, alternatively, increasing the speed of the progression of her disease, as she does not receive linguistic input that is cognitively challenging.

How to deal with the oppositions to tasks of daily living, expressed by older people in care facilities, is an ethical dilemma, irrespective of whether the context is multilingual or monolingual. The conflict between patients' needs for emotional support, and task-oriented goals is somehow inevitably inherent in the institutional setting (Backhaus 2008; Heinemann 2009), and carers struggle with this conflict every day. There seem, however, to exist methods and strategies that carers use that may be viewed as “less” imposing and “less” face-threatening (Heinemann 2009; Jansson and Plejert 2013). This also shows in research on communication strategies of caregivers of people with dementia (Lindholm 2010; Ripich et al. 1998, 1999; Small et al. 2000, 2003). When it comes to the kind of multilingual, linguistically asymmetrical context as is described here, however, it seems as if these kinds of settings are in acute need of further scrutiny, in order to see what can be done both in terms of supporting carers who do not share the language of their clients, and for supporting the person being cared for, in order to achieve as little imposition as possible. In the beginning of this article, the successful introduction of “ethnically profiled” care facilities for the Finnish and Finland-Swedish population in Sweden was referred to as one way of providing person-centered care for older immigrants. In Sweden, such profiled residential homes have very recently (as late as in 2012) started to emerge for a few other ethnical and cultural groups (e.g., older persons with a Persian or an Arabic background), and private home-help companies offer services that are carried out by staff with specific linguistic and cultural competencies.⁸ Such solutions are perhaps possible alternatives to “ordinary,” Swedish residential homes, where cultural and linguistic matching of staff and residents cannot always be obtained. However, a variety of factors govern how and to what extent immigrants in Sweden access, or even want to access the care facilities available to them (both ordinary, Swedish facilities, and ethnically profiled ones), and when it comes to older immigrants with potential dementia, there also appears to exist a range of issues associated with cultural differences in perspectives on, and knowledge of the disease (Mazaheri 2013). Such factors affect how and to what extent, multilingual older people receive care that is adapted to their needs.

The aim of the present article was to shed light on a setting which will become increasingly common in Sweden and the rest of Europe, in a near future, and to demonstrate the response practices that a carer employs in interaction with a resident, whose language she does not understand, and where the resident's language abilities might also be affected by a dementia disease. Despite the fact that the carer's response practices have been problematized as potentially imposing to the resident, the aim has not been to say that the carer has performed her job badly. An opposite claim could rather be made. Considering Sarah's limited linguistic skills in Mehri's language, she has nonetheless been able to resort to emphatic moves, e.g., using prosodic cues in soothing talk, expressed her sympathy in terms of claiming that she “understands,” acknowledging that Mehri is saying “something” (but most of the time not understanding what) by means of minimal responses and questions, and offered praise when Mehri is acting cooperatively. In many ways, she puts a lot of effort into acting in a respectful and sympathetic manner towards Mehri. The drawbacks, however, for both carer and resident seem to override this attempt to depict the encounter in a brighter light. As more and more multilingual older persons will be in need of residential care in many countries in a very close future, the issues raised in this article cannot await being further addressed.

Transcription Conventions

The following conventions have been used in the present article. They are adapted with some modifications from Ochs et al. (1996).

::::

Colons are used to indicate the prolongation of the sound just preceding them. The more colons, the greater the elongation.

-:

A hyphen after a word part or a word indicates a cutoff or self-interruption.

=:

The equal sign indicates that utterances follow immediately to each other with no discernible silence between them.

┌ ┐:

Brackets indicate where overlap begins and where it ends.

((nods)):

Double parentheses mark the transcriber's comment on how something is said or what happens in the context.

(1.6):

Numbers in parentheses indicate silence, approximately represented in tenths of a second.

(xxx):

x in parentheses indicates something being said, but no hearing can be achieved.

°°:

Degree signs indicate talk markedly softer or quieter than the adjacent talk.

Yes :

Underlining is used to indicate some form of stress or emphasis either by increased loudness or higher pitch.

NO:

Indicates especially loud sounds relative to the surrounding talk

.:

The period indicates a falling or final intonation contour.

,:

A comma indicates “continuing” intonation, not necessarily a clause boundary.

?:

A question mark indicates rising intonation, not necessarily a question.

↑:

Indicates marked shift into higher pitch in the utterance-part immediately following the arrow.

< >:

Right/left carets bracketing an utterance or utterance-part indicates speeding up

.hh:

A period before the letter “h” indicates inbreath. The more hs, the longer the inbreath.