Abstract
This study explores how manners of mediation, and the use of embodiment in interpreter-mediated conversation have an impact on tests of cognitive functioning in a dementia evaluation. By a detailed analysis of video recordings, we show how participants—an occupational therapist, an interpreter, and a patient—use embodied practices to make the tasks of a test of cognitive functioning intelligible, and how participants collaboratively put the instructions of the tasks into practice. We demonstrate that both instructions and instructed actions—and the whole procedure of accomplishing the tasks—are shaped co-operatively by embodied practices of all three participants involved in the test situation. Consequently, the accomplishment of the tasks should be viewed as the outcome of a collaborative achievement of instructed actions, rather than an individual product. The result of the study calls attention to issues concerning interpretations of, and the reliability of interpreter-mediated tests and their bearings for diagnostic procedures in dementia evaluations.
Keywords
Introduction
This present qualitative study deals with an interpreter-mediated, face-to-face encounter between a Kurdish-speaking elderly woman, a Swedish-speaking occupational therapist, and an interpreter, in a memory clinic when they carry out a screening test of cognitive functioning as part of a larger dementia evaluation. The article aims to explore this particular activity and provide observations about the collaborative accomplishment of interpreter-mediated tests of cognitive functioning in dementia evaluations by drawing special attention to the impact of mediation and the use of embodied practices on the evaluation procedure. Studies on instructional and test situations, such as formal tests for children (Maynard & Marlaire, 1992; Plejert & Samuelsson, 2008), and instructional activities, e.g. surgical procedures (Mondada, 2014), and crocheting lessons (Lindwall & Ekström, 2012), have shown how embodied instructions play a significant role in the performance of the person who receives the instruction, and also how instruction gains its meaning in and through the course of its execution (Garfinkel, 2002). Previous research has also shown that the accomplishment of instruction in face-to-face interaction (e.g. Maynard & Marlaire, 1992) is a collaborative accomplishment of both the instructor and the instructee (see also Mondada, 2014). However, nothing of the sort has ever been explored in the study of the instruction of, and execution of a test of cognitive functioning in dementia evaluations (cf. Plejert, Antelius, Yazdanpanah, & Nielsen, 2015 on verbal instructional activities in dementia evaluation). By attending to the details of the collaborative achievement of three sub-tests in an interpreter-mediated dementia evaluation, we aim to explore the complexity of this interactional setting. We underscore the fact that providing instruction in situ and in vivo and performing it are interdependent sets of actions that not only one follows the other but that they are achieved co-operatively and collaboratively (see Garfinkel, 2002; Mondada, 2014; cf. Maynard & Marlaire, 1992). By so doing, we highlight the question whether the result of the tasks of the test is the product of a single individual, or the product of the collaboration among all parties involved in the test situation (cf. Maynard & Marlaire, 1992). Because, unlike other data sets for studying instructional activities, our data is comprised of a mediated talk through an interpreter for the purpose of dementia evaluation, we aim to draw attention to the sensitivity and delicacy of the procedure in relation to its clinical consequences. By means of demonstrating the complexity of the situation, we aim to raise awareness about a) the impact of mediation and the use of embodied resources on the testing procedure as a neglected constituent in relation to dementia evaluation activities, b) the embodied collaborative achievement of tests in dementia evaluations as consequential in terms of the validity of test results. We would also like to draw practitioners’ attention to the crucial details of accomplishing instructions and instructed actions through focusing on the interactional work of the interpreter, and the use of other communicative resources apart from talk (i.e. gaze, gesture, bodily orientations, etc.) in the administration of tests of cognitive functioning, the effect of which on the procedure may be decisive in the evaluation of the patient’s performances.
In order to detail how the test is conducted, i.e. the ways that instruction is presented, rendered, understood and followed in interaction, we draw on ethnomethodological conversation analysis (EM/CA, Garfinkel, 2002; Mondada, 2014; Sacks, 1992) and analyze the participants’ collaboration in real time. It should be stressed that we do not intend to evaluate the routines of test administration in the memory clinic. Neither will we evaluate the interpreter’s interpreting performance, or the occupational therapist’s clinical abilities. With a close description and analysis of the video recordings of a test situation, we aim to shed light on part of what is not very well known in dementia research, namely, the detailed quality of the procedure of clinical testing in interpreter-mediated situations (however, see Plejert et al., 2015).
Dementia evaluations and patients with linguistically and culturally diverse background
A dementia evaluation is a complex procedure, comprising different parts and sub-tests (Ardila, 2005; Parker & Philp, 2004), such as history-taking (medical interview), physical and psychiatric examinations, and ancillary investigations, blood tests, brain scans, lumbar punctures, and tests of cognitive functioning. A range of cognitive screening tests is available, among which the Montreal Cognitive Assessment (MoCA; Nasreddine, 2003–2014) is the test used in the context investigated here. The MoCA is a screening instrument that assesses short-term memory (immediate and delayed recall of five nouns); visuospatial abilities (clock-drawing and three-dimensional cube copy); executive functions (adapted Trail Making B, phonemic fluency and verbal abstraction); attention, concentration, and working memory (target detection, serial subtraction, digits forward and backward); language (confrontation naming, repetition of two syntactically complex sentences, and phonetic fluency); and orientation to time and place. As for all tests of a formal nature, it is supposed to be carried out in a specific and regulated way. In this process, the manners in which instructions are given, and how they are followed, are key for a patient’s scores on different tasks within the test.
The tests that are used for dementia evaluation in memory clinics in Sweden commonly originate in and are only validated in certain European and North American countries and languages (see Yeo & Gallagher-Thompson, 2006, part II for an overview over cognitive assessments and culturally and linguistically diverse patients in the US; Nielsen, Vogel, Riepe, et al., 2011). This is potentially problematic when tests are implemented on people with ethnic minority language and culture, especially those with little or no formal education, who do not share with clinicians the same language and culture (see e.g. Nielsen & Jörgensen, 2013). Among different types of tests, there are, however, some that do not require advanced degrees of literacy and might be implemented more easily with the aid of an interpreter (Naqvi, Halder, Tomlinson, & Alibhai, 2015). Sub-parts of these tests, such as recalling common objects, clock reading or drawing, etc. seem to be less linguistically and culturally biased in comparison to other tests, such as word-list learning tests, or tests of mathematics and language skills (cf. Nielsen, Vogel, & Waldemar, 2012). Nonetheless, clinicians who do not share the language with their patients may find it very challenging to implement cognitive tests, and at the same time maintain the professional quality of the diagnostic evaluation (Plejert et al., 2015). The study of the results of neuropsychological tests in culturally and linguistically diverse patients shows a disparity of quality in diagnosis evaluation of these groups compared to the general population (Nielsen, Andersen, Kastrup, Phung, & Waldemar, 2011). This concern is also shared in other studies. For instance, a study in Denmark (Nielsen, Vogel, Phung, Gade, & Waldemar, 2011) shows an over- or under-diagnosis of dementia in ethnic minority groups, which are incongruent compared to the nationwide register of the distribution of dementia in the general population. The authors point out that this incongruity may have to do with the quality of dementia evaluation influenced by the linguistic and cultural differences of the patients and clinicians in test situations (Nielsen, Vogel, Phung, et al., 2011). What is common among these reports is that little is known about how these results come about (i.e. the quality of the procedure of testing) even if there is a consensus that the test results are not quite reliable when the tests are administered via interpreters (see e.g. Casas et al., 2012, p. 98).
The impact of mediation and embodied actions on clinical encounters
In this section, we report on studies that deal with interpreter-mediated interaction, and particularly research that emphasizes the role of embodied conduct in interaction, specifically interpreter-mediated clinical talk. By focusing on embodied conduct, here, we simply mean that we attend to verbal as well as non-verbal resources used by participants in interaction. Our analytical unit is thus human action that we consider key to this study as well as to social interaction generally, and to be configured by verbal and non-verbal conduct (see Goodwin, 2013).
Interpretation as interaction: The impact of mediation on clinical encounters
Based on studies on interpreter-mediated interaction in the last decades, we know that mediation is far from ‘passive’ or ‘neutral’; the interpreter’s role is way beyond just the transferral of information between two parties at talk, but is rather contextual and interactional, and thus dialogical (e.g. Angelelli, 2004; Bolden, 2000; Davidson, 2000; Hsieh, 2008; Li, 2013, Linell, 1997; Ray, 2000; Raymond, 2014; Roy, 2000; Wadensjö, 1998, 2001, 2008; Warnicke & Plejert, 2012; inter alia). Wadensjö, for instance, shows how interpreter-mediated encounters are triadic situated systems of interaction for different types of ‘coordinations’ between an interpreter and other parties in ‘rendition’ or ‘non-rendition’ activities (Wadensjö, 1998: ch. 6). Like other social situations, interactional order in interpreter-mediated interaction is organized in and through participation frameworks (Goffman, 1981), participants’ multiple role performances, their cooperating, coordinating and remedial work, etc. Such interaction is not only co-constructed at the level of content but it is also coordinated in a special way at the level of its formal organization, such as turn-taking (see also Warnicke & Plejert, 2012).
Studies in institutional settings also elsewhere point out that the interpreter’s contribution in mediation is not limited to the ‘faithful’ rendition of utterances, but it is extended to the institutional goals, habits and routines through which the interpreters may also become gatekeepers (see Davidson, 2000), advocates for either parties, cultural brokers (Kaufert & Koolage, 1984; Kaufert & Putsch, 1997), or facilitators regarding building trusts, brokering knowledge, and providing direct information, etc. (see Hsieh, 2008; Meeuwesen, Twilt, Thije, & Harmsen, 2010; Raymond, 2014). As an interactional practice, mediation can consequently influence other parties’ participatory statuses as to how—and how much—to receive information (Keselman, Cederborg, & Linell, 2010), how to receive the rendered utterances with some ‘perspectives’ (Wadensjö, 2004), or how to fuse linguistically and culturally bound assumptions and contextual bearings into the understanding of what is interpreted (Mason, 2006). Besides indicating the situational and contextual specifics of mediation, there are studies on the issue of control in the flow of interaction, especially in medical settings, which point out various problems arisen in medical procedures because of the mediatory practices (e.g. Hsieh & Kramer, 2012). There are also studies proposing models of interaction, and also solutions as to how to predict and solve possible recurrent problems in mediatory practices (e.g. Raymond, 2014). In dementia research studies, however, the study of cross-linguistic interactions which necessitates a third party to mediate the talk is not very abundant (see e.g. Plejert et al., 2015).
The impact of bodily conduct on interpreter-mediated dementia evaluations
There is an established tradition of examining embodied actions in interaction analysis in everyday or institutional social encounters (Goodwin, 1981, 2013; Heath, 1986; Mondada, 2007a, 2007b, 2014; Streeck, 1993, 2008, etc.). There are, however, not many studies on the use of bodily acts in interpreter-mediated interactions. In her paper on therapeutic encounters, Wadensjö (2001) shows the significance of the configuration of participants in interpreter-mediated interaction. She suggests that the interpreter’s bodily position in relation to the other parties at talk may impact the quality of interaction and the work of interpretation.
Like any other interactional situations, participants, including the interpreter, are sensitive to the gaze, posture and the body movements of their co-participants (see also Wadensjö, 2004, 2008). Apfelbaum (1998), for instance, shows that interpreters may use non-verbal signals, among others, during synchronization of interaction with other parties at talk. The use of non-verbal signals in the interpreter’s behavior has also been observed in other studies (e.g. Pasquandrea, 2011; Ticca, 2008). In medical settings, Pasquandrea (2011) shows how the relevance place for initiating interpretation, or the request for interpretation, can be signaled by participants through non-verbal means such as mutual gazes or body orientations. In comparison to Ticca’s (2008) data from non-professional interpreting among non-acquainted participants where there is a competition of turn taking, in Pasquandrea’s (2011) data there is a long-term acquaintanceship between the doctors and the professional interpreters. Such social relations may allow a different set of multimodal resources used in interaction (Pasquandrea, 2011, p. 476).
Non-interpreter-mediated clinical encounters have been the topic of many studies in the EM/CA literature (Heath, 1986; Heritage & Maynard, 2006; Modaff, 2003; Mondada, 2007a, 2007b, 2014; Robinson, 1998; Ruusuvuori, 2001; Stivers & Robinson, 2006, among others). In some of these studies, the role of embodied behavior, such as body movements or gaze in doctor-patient talk has been analyzed, e.g. in the management of recipiency, or the distribution of speakership in the management of turns in the progression of talk (Heath, 1986; Robinson, 1998; Ruusuvuori, 2001; Stivers & Robinson, 2006; cf. Goodwin, 1981 for the use of gaze in the organization of everyday talk). Similarly, in medical and clinical consultations, embodied resources may be used for shaping and reshaping the participation framework (Heath, 1986; Robinson, 1998), for the display of active participation in interaction (Laakso, 2012), the transition between the tasks (Modaff, 2003), etc. However, the mobilization of embodied resources has been only sparsely studied in the context of interpreter-mediated interaction (cf. Pasquandrea, 2011; Wadensjö, 2001; inter alia), and it seems to be entirely overlooked in the setting of dementia evaluation.
Theoretical and methodological approach
To analyze the details of embodied practices, and the examination of the resources used in the participants’ coordinations in our data, we use conversation analysis and its original tenets in ethnomethodology (see e.g. Mondada, 2014) as our theoretical framework and methodology of research. Ethnomethodology and conversation analytic (EM/CA) perspectives are based on an empirical and emic approach to naturally occurring activities (see e.g. Rasmussen, Brouwer, & Day, 2014). Such an approach focuses on analyzing the organization of studied activities from within. It aims for understanding the views of participants in an activity through investigating their sense-making procedures and methods. It thus attends to members’ displayed understanding of the organization of actions/activities/tasks/instructions, and their display of skills and competences in the activity and how skills are shared and developed within embodied practices (e.g. Deppermann 2013; Koschmann, LeBaron, Goodwin, & Feltovich, 2011; Mondada, 2014). What comes to be the basic tenets of ethnomethodology of instruction can be originally found in Garfinkel’s study of different instructional settings (2002; see ch. 6). We draw on his three uncovered properties of the instructional activity, namely its indexicality, reflexivity and embodied achievement. What is meant by these properties are accordingly a) what an instruction topically has (its followability, its meaning, its completeness) is only achievable in a course of actions when doing it. b) Instruction is made up of the practical, local, and occasioned work that turns the instruction into an achieved embodied work of implementing it, or rather following it. c) Instruction and instructed actions are therefore reflexively tied together. If instruction is made to come out as something, how that something is going to come out is situatedly occasioned and cannot be objectively predictable once and for all situations (Garfinkel, 2002, p. 204). This has a methodological consequence as well; embedded in the instructed actions, examining any part of instructions should be done from within the procedure of its implementation. It thus requires that we attend to the sequential and temporal organization of instructional activities and the embodied actions of participants in interaction in real time (cf. Mondada, 2014: 132). This has been done in a series of EM/CA studies on instructional activities in various settings, for instance, in classroom interaction (e.g. Macbeth, 2011; Majlesi, 2014), crocheting lessons (Lindwall & Ekström, 2012), driving lessons (e.g. de Stefani & Gazin, 2014), surgical instructions (e.g. Koschmann et al., 2011; Mondada, 2014), etc. While these studies (e.g. Lindwall & Ekström, 2012) show that an instructing activity is more a matter of compliance than competence in providing and following an instruction, Mondada (2014), demonstrates that the instructional activity is organized and achieved also through collaborative embodied work of participants (cf. Maynard & Marlaire, 1992). Such collaboration however gets a new dimension when the instructor and the instructee are not equally competent in their cognitive and communicative skills, especially when they do not share the same language and, thus, require an interpreter to mediate their talk. In a clinical setting, such as a dementia evaluation, the mediation of interpreters and the organization of the setting become even more sensitive when it deals with evaluating and testing the cognitive skill of the patient whose communication with the clinician is achieved through a third party. This requires close scrutiny with regard to the sequential, temporal and embodied organization of interaction among the participants. That is why we use audio- and video-recordings in our study of the test of cognitive functioning during the dementia evaluation. The recordings give us the unique opportunity to meticulously examine and reexamine the sequence of events as it takes place in real time (see Sacks, 1984). Audio- and video-recording is part of a research practice to produce ‘intelligible and interpretable data’ (Mondada, 2007a: 811). In such a research practice, the videos are closely viewed, various social phenomena are uncovered, and usually episodes containing a particular phenomenon are extracted, multimodally transcribed if necessary, annotated and analyzed. Moreover, the procedure of collecting the data for this study has been approved by a regional board for ethical vetting. The participants’ identities were anonymized, and collecting data is carried out after receiving informed consent from the participating clinical staff, patients, and their companions.
The data
The data investigated in this study consist of sequences of interaction extracted from video recordings of an interpreter-mediated dementia evaluation at the point when cognitive abilities of a patient are tested. The whole evaluation took approximately one hour and 54 minutes, and was done in two consecutive parts, where history taking preceded the test. A range of cognitive screening tests is available, among which the MoCA (Nasreddine, 2003–2014), described in “Dementia evaluations and patients with linguistically and culturally diverse background” section, is the test used in the context investigated here. We focus on the execution of three sub-tests in MoCA, which are the drawing of a clock and a three-dimensional cube, and a picture-naming task (three animals). These are considered simpler tasks, and less culturally and educationally biased in comparison to tasks that require certain degrees of literacy and mathematical skills, etc. (cf. Nielsen et al., 2012).
The videos are recorded in a Swedish memory clinic. The patient is an elderly woman from a Kurdish speaking minority in Sweden, but does not speak Swedish. However, apart from Kurdish, she also knows some Arabic, and occasionally also uses Turkish words. During data collection, she did not articulate any Swedish words; neither did she show any sign of understanding it. The clinician is a Swedish speaking occupational therapist who does not speak any Kurdish. The mediation between the patient and the occupational therapist is assigned to a professional Kurdish-speaking interpreter who speaks Swedish as a second language.
In the next section, the transcriptions of three sequences are presented and analyzed. The transcripts present, besides language, other modalities in interaction, which are analytically significant for the purpose of this study, such as the movements of bodies, gestures, gaze, etc. Transcription conventions are found in the Appendix.
Embodied practices in interpreter-mediated interaction: A multimodal analysis
In a triadic interaction, as shown in the picture (Fig. 01), the participants are sitting at a rectangular table with the patient in the middle, and the clinician and the interpreter, facing each other. The analysis focuses on the three abovementioned tasks of the MoCA. First, we present how instruction is rendered and modified through mediation. Second, we demonstrate the use of embodied actions in instruction and its rendition. Third, we will exhibit the impact of embodied mediation on the instructed actions and how instruction is understood and followed by the patient. At the end, certain implications will be concluded as regard the effects of interpreter-mediated talk on the dementia evaluation procedure.
Embodied, collaborative directives to render the instruction
In our data, ‘the clock drawing test’ is a step-wise task in which the patient is first instructed to draw a numbered clock face. After finishing this first part, she is then asked to draw the clock hands at ten past eleven. The achievement of the second part of the task naturally hinges upon the accomplishment of the first part. It should be mentioned that the ways in which each part of the test is accomplished is assessed by means of different scores. For example, to be able to draw a closed circle receives a high score in comparison to not being able to do so, etc. Our data is delimited to the first part of the task, and how the instruction is rendered, understood and begun. As Excerpt 1(a) shows, the occupational therapist asks the patient to draw a clock in two disjunct turns: first she asks the patient to draw a clock (line 01), and then as an increment to that, she clarifies that the patient is required to draw “a round circle with the numbers in” (line 17). The transcript of the interaction reveals that it takes an exchange of a few turns to accomplish each step. These turns are embodied turns, i.e. they are designed by verbal and non-verbal resources. It is, for instance, evident that the occupational therapist uses her body to depict the instruction, or to show the locus where the task must be carried out (line 03). Moreover, immediately after the occupational therapist begins her instruction (line 01), the patient, who seems to reciprocate her recipiency, at least through the mutual gaze and an articulation of a sound (‘a’, line 02), redirects her gaze to the interpreter toward the end of the turn (line 02, Fig. 02). The shift of the gaze seems to be, and is treated as, soliciting help from the interpreter, which is done by a head turn and an orientation of her torso toward the interpreter (cf. Pasquandrea, 2011; Wadensjö, 2008).
The body movements and the gaze which are discussed in the analysis are illustrated in pictures or drawings, and their exact places in turns at talk are shown in the transcripts by hash symbols “#”.

From the extensive research on body movements and gaze, we can claim that the patient’s body and gaze orientations in talk signal the mutual engagement of participants in conversation (Goodwin, 1981). This is the rule of involvement in social encounters (Goffman, 1967) which also allows for a “talk-relevant-look-away” (Goodwin, 1981: 104–105; cf. Rossano, 2013). Accordingly, the patient’s body and gaze redirection toward the interpreter may be interpreted as a relevant move to display recipiency to the interpreter. This is also accounted for by the interpreter’s rendition of the occupational therapist’s utterance (line 05).
The shift of gaze toward the interpreter by the patient (line 02) however takes a long pause of a second to be responded to. This may be explained by attending to two facts. First, the occupational therapist’s turn with rising intonation projects that her turn has not yet come to an end (line 01). Second, she keeps her gaze toward the patient and does not turn it to the interpreter. The lack of redirecting gaze by the clinicians toward the interpreters is also reported elsewhere as problematic for the interpreter to recognize if rendition is sanctioned (cf. Pasquandrea, 2011, p. 465). At the same time, there are other signals, which eventually elicit a response from the interpreter. First, the patient shifts her gaze and sustains it toward the interpreter until the interpreter starts rendering (line 02). Second, the occupational therapist also stops at the first possible transition-relevance place for turn-taking (Sacks, Schegloff, & Jefferson, 1974; line 01), and provides an interactional space for the interpreter to make his contribution to the interaction.
The interpreter’s rendition (line 05) is a reduced and a summarized version of the occupational therapist’s turn (see different types of renditions in Wadensjö, 1998). He changes an indirect request to an outright, straightforward directive in the imperative mood, “draw a clock in here. a clock.” (line 05). He simultaneously points to the sheet of paper in front of the patient (line 06, Fig. 04).
What the occupational therapist and the interpreter do is an embodied practice of collaborative instructing in which the occupational therapist and the interpreter make use of different resources to have the instruction communicated to the patient (cf. Maynard & Marlaire, 1992). This collaboration involves the selective rendition of the information in order to achieve the ongoing action (cf. Garfinkel, 1967: 21 on ad hoc considerations). The occupational therapist’s display of the task, her embodied exhibition of the request (Fig. 03), the interpreter’s change of request into a directive, the incremented information through embodied redisplay of where to draw (Fig. 04), and the repetition of the word ‘clock’ are mutually informed resources to calibrate and recalibrate the patient’s understanding of the task and prompt her to begin with the task. By calibration, we mean that the occupational therapist and the interpreter collaboratively adjust the instruction by providing extra information to clarify instruction, acting out what is expected to be done by the patient, changing the form of instruction, cueing the action in its every step of implementation, and leading the patient with how to begin and proceed with the task.
Now in response to this collaborative instruction, the noun of the utterances in the prior turns is taken up by the patient as she asks “a clock?” (line 08). This is replied to by the interpreter as he confirms and reassures that mutual understanding is accomplished (line 10). He now has his wrist watch in the patient’s line of vision and points to it (Fig. 05), while saying “clock clock yeah” (line 11).
In the continuation of the interaction, what is interesting is the way the patient turns the event into a funny story as she chuckles saying “I swear (0.3) $ihhhn at home is not even one clock working he$” (line 14). Her chuckle is aligned with by the short laughter-like out-breath “hh” of the occupational therapist (line 15):

The patient’s funny comment as shown in the transcript above (line 14) was neither translated nor verbally responded to by other parties. The interpreter’s choosing to ignore the turn in spite of the mutual gaze with the patient may be due to the fact that the occupational therapist has already started to resume and develop the instruction (Fig. 06). Both the occupational therapist and the interpreter choose to stay with the line of instruction and orient toward the implementation of the task, rather than reacting toward the patient’s comment. The occupational therapist, already in the midst of the patient’s turn, has shaped her hand in the form of a circle (line 16). This pre-turn gesture could be inferred as projecting self-selection to talk (cf. Mondada, 2007b), and of course be a sign showing that the instruction is not yet complete. She, thereafter, says “a round circle, (0.2) with all numbers in” (line 17). When saying “with all numbers in”, she draws a circle in the air as well (line 18, Fig. 07). Her turn now gets immediately translated into Kurdish. This time again the rendered utterance gets the imperative mood: “do it like a circle” (line 19; Fig. 08). The clues, such as ‘drawing’, ‘a clock’ and ‘a round circle’ seem to be adequate for the practical purpose of starting the task. Exactly after hearing the words “do it like a circle”, the patient withdraws her gaze from the interpreter onto the paper in front of her, and starts drawing a round circle. Even if the interpreter continues to render the rest of the instruction, which is to put numbers in the circle, the patient adheres to the first part of the instruction. She practically takes the task as being accomplished step-by-step as it should naturally unfold, i.e. she first finishes drawing the circle before she begins the task of writing the numbers in it. When she is done drawing the circle (Fig. 09), she publicly announces her accomplishment with a vocal particle “a” (line 25), which may be marking the closing of the first part of the task. Immediately after, she, once more, gets reminded of the rest of the instruction:

The excerpt above exhibits how mediation is built not only by the rendition of talk but also by the rendition of embodied actions. The interpreter modifies both the directives in talk and in its affiliated gestures as he recycles the occupational therapist’s gestures with some modifications (cf. Anward, 2004 on recycling with différance; cf. Majlesi, 2015). The interpreter seems to be a foreman for the occupational therapist in a way that he is not just recapitulating the instruction (see Wadensjö, 1998: 92), he is also cooperating beyond interpreting. He independently responds to the patient (cf. Wadensjö, 1998), retells and recalibrates the instructed actions. He follows the patient’s performance, as clinicians do, and coordinates his delivery of instruction in Kurdish with the patient’s ongoing course of actions (lines 19, 22, 26 and 29). Together, the interpreter and the occupational therapist, underpin what to be done in the patient’s performance by using various resources. The interpreter, moreover, makes it even clearer what is expected of the patient by repeating the essential constituents of the directives, or directly responding to the patient without rendering back the patient’s turns to the occupations therapist: “write also the numbers. for example” (line 26). When the patient asks “numbers?” (line 27), it is the interpreter who directly responds to it: “yeah. (0.2) just a clock a clock. what is a clock like?” (line 29). Now with this confirmation, the patient gets into the task and writes down the numbers (Fig. 10).
Of particular interest is the use of gaze, and body movements, and especially gestures, e.g. pointing as crucial to foregrounding the significant parts of the directives and where actions are hinged or predicated upon. This is observable especially in the production of action as regard three linguistic constituents, ‘clock’, ‘a round circle’ and ‘here’ when a speaker uses gestures such as shaping the roundness of the clock in the air, pointing to, or beating on an artifact, such as a wrist watch or a sheet of paper to help visibly display these constituents. The co-operative work of interaction through embodied practices also seems to have a crucial bearing on the performance of the patient. Such collaborative embodied achievement of the instruction and proceeding with the instructed action has implications as regard the overall production of the test as to become the product of the participants together rather than an individual one. This point is further explored in the following examples.
From rendering instruction to eliciting instructed actions
Excerpt 2(a) starts with the occupational therapist presenting a picture of a cube to the patient. This time, she presents the whole task at once: “this is a cu:be I want you just to (.) draw it down here” (line 02). As it is clear from the produced turn, the indexical deictics, such as “this, I, you, it, down here”, may only be understood in the physical environment through embodied actions (see Figs. 11–13):

Similar to the previous example, the rendition of the turn from Swedish to Kurdish consists of both verbal translations of the words—albeit rather selectively in the interpreter’s rendering—and also recycling of prior gestures (see Majlesi, 2015). If the turn at talk and the pointing gestures affiliated to them produced by the occupational therapist (lines 01-2, Figs. 11–13) are compared to the corresponding talk and gestures produced by the interpreter (lines 04-5, Figs. 14–15), it can be observed that the interpreter recycles both the turn and the gestures. Even if there are also significant differences in the design of these two consecutive interactional turns, it seems that the interpreter adheres to the gist of the occupational therapist’s directive, though redesigning it in his own action. This redesigning goes beyond recapitulating the directive and includes the change in the directness of the request and transforming the instructional formats to be more imperative, reduced in size and summarized in content. He also recycles the occupational therapist’s body movement showing two loci on the paper although with a difference in the manner of display: he does it by pointing fingers and not by a pen, and not exactly in the same way as the occupational therapist does. In the interpreter’s turn, the request “I want you just to (.) draw it down here” made by the occupational therapist (line 02) turns into a louder and more concise turn: “IN HERE, (0.2) MAKE A COPY OF THIS” (line 04). While saying that, the interpreter also turns the prior gesture, which was done by making a circle around the cube, into a more emphatic gesture of pointing to and beating on the surface of the paper.
Despite this co-operative move by the occupational therapist and the interpreter, and all apparent clarity in the produced embodied turns, the patient responds to the directive by an open repair initiation “wh:at?” (Drew, 1997), asking for the repetition or reformulation of what was just said (lines 06). This may be due to the fact that the recipiency of the patient was not fully secured. The close observation of the unfolding interaction shows that the patient has had her gaze on the paper since the first turn (line 01, cf. line 07), and only by the time the interpreter points to the empty space on the paper (line 05, Fig. 14), she lifts her gaze toward the interpreter (line 07) and immediately asks “wha:t?” (line 06) in an overlap with the interpreter’s utterance (line 04). The interpreter however finishes his turn without immediately responding to the repair initiation. After the interpreter completes his turn, the patient now gets a chance to repeat the question “what?” (line 09):

The patient’s question “what?” (line 09), which is now posed twice, without being translated back to the occupational therapist, gets directly answered by the interpreter himself. Again, the interpreter is not just recapitulating the instruction, he is rather teaming up with the occupational therapist to reinforce the directive by repeating the verbal instruction (“draw this in here”—line 12) and its affiliated gestures as he beats on the illustrated cube and the empty space where the patient should draw the copy of the cube (line 13). After repeating the directive, the interpreter still keeps his fingers on the surface of the paper, and sustains his gaze toward the patient. The patient, with the interpreter’s directive, shifts her gaze from the interpreter toward the place where it is pointed at (line 14). The interpreter’s action, including his sustaining pointing and gaze, seems to treat the instruction incomplete unless his action is responded to, and shown receipted as understood, by the patient (cf. Garfinkel, 2002; Mondada, 2014). He still keeps his finger on the sheet during the following pause (line 16). So, the interpreter’s action is designed to elicit a response from the patient. It is proved in the subsequent turns that this response would not be counted as sufficient just by a verbal acknowledgement of understanding but setting off the task in practice. This shows how the interpreter is monitoring and walking the patient through the task. With the fifth of a second pause when the interpreter does not receive any token of understanding (line 15), he repeats the verbal directive and the gestures once more without consulting back with the occupational therapist (line 17). This time, he receives an immediate ‘aha’ receipt token. The interpreter does not treat this signal of “change-of-state” from the patient as the completion of the instruction either (cf. Heritage, 1984: 301). He, however, pushes for the practical action by saying “draw it” (line 21). The patient, in response, seeks confirmation as regard the theme: “draw it?” (line 23), which is directed to the interpreter and responded to by him: “yes this” (line 24), as he again points to the cube (line 25, Fig. 16). These instructional turns eventually get a more substantial reply when the patient informs the interpreter that “I see what you MEAN” (line 29) and starts drawing the cube (line 30).
Excerpt 2, like Ex.#1, has shown that the process of instructing a patient to do a task in interpreter-mediated interaction may entail the modification in the form and the content of directives in mediation, and the use of embodied practices to do so. This also entails maintaining understanding, not just in the form of verbal acknowledgment, but putting the directives into embodied practices in instructed actions. That is, the patient’s display of understanding is pursued not just through talk, but also through observable and visible hands-on practices when she shows her understanding in practice. This shows that during the execution of the test of cognitive functioning, the completion of instruction is not separated from instructed actions but intertwined with it (cf. Mondada, 2014). This practice includes the instructors (the occupational therapist and the interpreter) becoming part of the process of having the task begun by asking for practical actions in step-by-step directive-response sequences. They do it through actively monitoring and recalibrating the directives to have the instructed action set off. This will be pursued even in the continuation of the test where the patient is to carry out a picture-naming task.
Pursuing response
The use of objects and artifacts to test the ability of the patients to identify, recall, and produce their names is part of the evaluation process. This is illustrated in Excerpt 3. When introducing the task, the occupational therapist begins to show the pictures of three animals in order from the left to the right (Fig. 17). The first picture is a lion, which is pointed at by the occupational therapist when she asks the patient “do you see what kind of animal is
So far, it seems that the participants are running the task in coordination. Like previous examples, the activity of instruction and following the instruction are not two separate events but entangled in a way that the accomplishment of the whole task is the outcome of joint efforts. The occupational therapist foregrounds a picture with a steady pointing gesture; this pointing is kept still on the picture until the task of recognizing and naming the animal is complete (cf. Majlesi, 2014, about teaching grammar on a worksheet). The interpreter, instead of just replacing the occupational therapist’s action with his own, exploits her pointing gesture and builds his rendition upon it. When he asks the question: “what’s this?” (line 03), the occupational therapist’s pointing is already linking the deictic to the picture. The verbal and non-verbal resources are used in coordination for the practical purpose of accomplishing the task. The significance of these embodied coordinations is more transparent when this routine is not kept intact in the next step of the test.
When the first sequence is done (lines 01–06), with a short pause, the occupational therapist moves her pen and points to the picture of a rhino (Fig. 18), and asks: “’n do you see what kind of (.) animal is


In the example above, a pointing gesture proves to be much more than making something on the paper vividly visible, relevant or topicalized. In the first sequence (lines 01–06), a pointing gesture is kept on the picture from the beginning to the end. As the participants move on to the next picture, the pointing gesture—along with the verbal questions—draws the focus of the task onto the next picture: a rhino (line 09). That is, the pointing is used as one of the cues signaling both the main object of inquiry and the beginning and the ending of the task as the pointing is coordinated with the length of the sequence by landing on the picture in the beginning of the sequence and being lifted in the end of the sequence. It also signals—along with the question—that with the move onto the next picture, the same practice of inquiring about the pointed picture is upheld in the current sequence as it did in the prior one. Therefore, as long as the pointing is held on the object, the focus of the interaction seems to be still on that one. This, however, is practiced differently as regard questioning about the rhino. When the patient is asked to recognize the picture of a rhino (lines 09 and 11), the pointing gesture, which was aimed at the rhino (line 10, Fig. 18), is lifted in the middle of the patient’s response (lines 12–13, Fig. 19). The occupational therapist moves away both her hands from the worksheet. The left hand is not directly pointing to the picture of the rhino any longer and the right hand is not concealing the next picture, which is a picture of a camel. This makes the next picture a projectable topic of the upcoming inquiry, and gives opportunity to the patient to observe the next task already now. The patient first makes a guess (in a question form) on the picture of the rhino: “it, (1) elephant?” and as soon as the occupational therapist’s hand is away from the camel, she continues “no (0.3) no (1.5) e::: it is a camel.” (line 12), and she points to the drawing of the camel at the same time (Fig. 20).
The interpreter, in turn, instead of translating the patient’s turn, just explains in Swedish “this she says camel but”, and immediately turns to the patient and points to the rhino and says in Kurdish: “this this” (line 17). This time both the occupational therapist and the interpreter point to the rhino (lines 18–19, Fig. 21). The interpreter once more co-operates on the occupational therapist’s action and coordinates with her, refocusing the instruction, asking the patient to complete her task as regard the rhino. The interpreter’s turn, which begins with Swedish and ends with Kurdish seems not to be captured by the patient. After a pause, the patient initiates a repair: “WHAT?” (line 21). This repair initiation is directed toward the interpreter and answered directly by the interpreter who repeats the question again: “what’s this?” (line 22). This is followed by another guess proposed by the patient and it leads the sequence to an intense negotiation over the name of the animal:

Besides the significance of every tiny move in interaction, which may have an effect on the patient’s performance, the way that the occupational therapist and the interpreter coordinate with each other also probably has an impact on the overall result of the test. One potentially influential issue is the amount of, and the ways of coordination and co-operation between the interpreter and the occupational therapist in rendering the instruction and pursuing its understanding by the patient. Although the interpreter’s action is positively oriented toward accomplishing the instruction and having the patient proceed with the task, the selective rendition or non-rendition of the patient’s verbal contributions, and the provision of response to the patient and guiding her in carrying out the task without coordinating with the occupational therapist have certain implications for evaluating the result of the task.
There are certain things that the occupational therapist is missing by not being informed through rendition of the patient’s turns back into Swedish. In the negotiation over the picture of the rhino, there is an obvious sign that the patient was attempting to distinguish the trunk from the horn, and moving from the first guess “elephant” (line 12), to the “boar” (line 24) which is reflected in the patient’s turn (line 24). This was not verbally responded to (line 26, Fig. 22), nor rendered into Swedish to the occupational therapist. Also the word “boar” was rendered as “pig” (line 29). Irrespective of the significance of these details for the evaluation procedure, the word “boar” in the patient’s answers (lines 24 and 27) has obviously a common feature with a rhino, be it, their horns. This is not reflected in the word “pig” (line 29) which was rendered instead of “boar” to the occupational therapist. After the occupational therapist is informed that the patient mistook the picture of the rhino for the picture of a pig (line 29), the occupational therapist now only seeks information if the rhino is an animal familiar to the patient: “if i say r
Conclusion
In this study, we have shown how mediation in talk with a patient and the use of bodily acts, e.g. the use of gaze, gestures and bodily orientations, etc. by the clinician and the interpreter calibrate the patient’s understanding of the instruction and influence her way of performing it. We have also shown that the instruction’s completeness, intelligibility and meaning are contextualized and organized through coordinated and co-operated actions among parties involved in the instructional activity (cf. Garfinkel, 2002; Mondada, 2014; cf. also Goodwin, 2013). By focusing on a part of an interpreter-mediated dementia evaluation as a social-interactional phenomenon, this article has drawn attention to some of the complexities of embodied practices in the heterogeneous social context of dementia evaluation of patients with diverse linguistic and cultural backgrounds. The roles of bodily movements and gaze have been extensively explored in non-mediated interactional settings (see e.g. Goodwin, 1981, 2013; Heath, 1986; Mondada, 2007a, 2007b, 2014; Ruusuvuori, 2001; Streeck, 1993, 2008, inter alia; see also Kendon, 1990). This study contributes to the handful number of studies on the significance and impact of coordinating bodily actions in interpreter-mediated interaction (e.g. Apfelbaum 1998; Pasquandrea, 2011; Ticca, 2008; Wadensjö, 1998, 2001, 2008; Warnicke & Plejert, 2012). With this study, we emphasize the crucial role of mediation and embodied actions in the organization of the execution of the test of cognitive functioning not just in the process of understanding the instructions, but rather more importantly in their implementation and in their outcome.
Our study can be concluded and discussed in three main senses. First, interpreters’ role and their participatory statuses are not limited to the fluctuation between a ‘reporter’, a ’responder’, and a ‘recapitulator’ (see Wadensjö, 1998: 91–92). The role can also be extended to a ‘(re-)calibrator’ role, where in coordination with the clinicians, they monitor, and help the patients through carrying out their tasks.
Second, the interpreter and the clinician coordinate their actions and thus coauthor, not only the instruction, but also the instructed actions. In our data, there is a constant orientation toward the goal of ‘getting the task done’ (see Plejert et al., 2015) in which the interpreter and the clinician work hand-in-hand to make the instruction understood and put into practice. There is a finely grained moment-by-moment assessment of the situation in and through which clinicians and interpreters coauthor the instruction. In such a co-authorship, both the clinician and the interpreter calibrate the patients’ understanding of instruction and her performance through their coordinated actions and collaborative interaction. That is, the sets of instructions presented by clinicians get reworked and redesigned by the interpreter and rendered to the patient with some differences in form and content. The understanding of instructions and their implementation are hence the result of these interactional co-operations.
Third, unfolding interaction, moreover, proves that bodily actions of participants—using pointing gestures, positioning body parts, their orientation, their gaze directions vis-à-vis one another, play a significant role in the way in which instructed actions are produced and tasks are accomplished. We have shown, for instance, that the relevance place for rendering a turn may be shaped by other means than verbal resources only (cf. Wadensjö, 2008), how such resources may result in belated renditions or the reduced and summarized ones, and how they consequently influence the management of the whole activity (e.g. pointing gestures are used to clarify the indexical expressions in the directives, or are mobilized to mark the beginning and ending of a task).
As a result, certain implications can be drawn from the study that may be of particular interest to practitioners and interpreters. First, instructing and implementing that instruction in interpreter-mediated tests of cognitive functioning in the dementia evaluation are by no means two separate events emerging in separable turns of asking and responding actions (i.e. adjacency pairs, see Schegloff, 1973, also see Bolden, 2000; cf. Mondada, 2014). In the actual handling of the test, however, the clinician’s request and the patient’s response are intertwined actions mediated and influenced by the coordination and co-operation between the clinician and the interpreter. The instruction in its mediation and actual implementation gets reformulated and redesigned by the clinician and the interpreter. It also gets modified through the patient’s repair initiations, reconfirming her understanding, soliciting help, seeking confirmation, negotiating meaning, etc. So, it is of importance that interpreters have some ideas about the significance of their role and that both clinicians and interpreters know of the impact of their collaboration. In addition, not uncommonly, interpreters as well as professionals who rely on interpreters orient towards language in a surprisingly lay manner, despite their training, professionalism and high proficiency in many areas; Interpreters, in the sense that they adhere to interpreters’ code of conduct, which is based on the conduit-model of interpreting (i.e. the interpreter as a “channel”), and professionals, in terms of not taking into account all aspects of the interdependence between language and culture; thus sometimes believing that an expression in one language has an absolute equivalence in another one. This latter issue sometimes affects clinicians in the sense that they put too much trust, and assign too great a responsibility on interpreters during the clinical encounter. We are not making the claim that this is the case in our study. However, the point is relevant in relation to the issue of raising professionals’ awareness of the impact on their collaboration on a clinical activity.Second, as in the course of testing, the normative sequential unfolding of instruction and its following seems not simply and unproblematically to be proceeded by telling/asking and doing/carrying out the task, it would also be of relevance to practitioners to know that however strictly their guidelines prevent them from meddling with the patient’s performance, testing appears to be an intertwined course of instructing and instructed actions. Patients and the clinicians are in a dialogical relation and in our case, when this relation is interfered with an interpreter, the pair of instruction and the performance become more entangled. For instance, the patient’s repair requests induce more instructions not provided just by the clinician but also by the interpreter as well during the patient’s performance, and thus have an impact on the execution and the results of the tests of cognitive functioning.
Third, interpreter-mediated dementia evaluation is conducted through the collaboration and coordination of embodied practices. Although some of the test instruction is only to be presented through articulated speech without giving any other clues to the patient (see the guideline for MoCA in Nasreddine, 2003), the use of the body is indispensible to face-to-face interaction (see Goodwin, 2013; Mondada, 2014). Together with talk, the body is also a semiotic resource; a vehicle for sense-making. Embodied practices will not only impact on the quality of instructions, but also influence the management of the test, and, also unequivocally, its outcomes.
As a result of our observations, we conclude that the outcome of each task—whatever it means to the participants—is therefore not the product of either of the interactants’ contributions alone, nor is it accounted for by the instruction itself. It is, rather, the procedural achievement of embodied and locally administered instructed actions. Therefore, the following of instruction, i.e. the implementation of the task and the outcome of the tests seem to be produced not by an individual patient, but co-produced by the procedural collaboration between all parties in interaction. We believe that this finding needs to be taken quite seriously, both in relation to interpreter and clinician training, in relation to interpretations of patient performance and test results, and not the least, in the design of formal clinical tests; particularly when they are to be used in mediated encounters.
Footnotes
Acknowledgements
We are deeply indebted to the helpful comments we received for an early presentation of this study at Nordisco Conference (Jyväskylä, 2014), and the insightful comments on an earlier version of the paper from our colleagues at the Center for Dementia Research (CEDER) at Linköping University. We owe a great debt to Charles Antaki and Leelo Keevallik for the comments on our last version of the analysis. We are also grateful for the comments we received from the anonymous reviewers. Finally, we acknowledge our enormous debt also to the participants in our study, the staff of the memory clinic, the patient and the interpreter for allowing actual events in the evaluation situation to be recorded and used for this study.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study was conducted as part of a program Dementia: Agency, Personhood and Everyday life at the Center for Dementia Research (CEDER) funded by the Bank of Sweden Tercentenary Foundation with the Grant Number: M10-0187:1.
