The paradox from within: research participants doing-being-observed

This article analyses a collection of cases from video recordings of naturally occurring interaction in institutional settings, where members display an orientation to the presence of the recording equipment. Such instances have been treated elsewhere as evidence of contamination of the ecology of the setting. The findings suggest that participants do remain aware of the recording activity, but that they publicly display when they are attending to it. Indeed, it is used as one resource to occasion identity work as competent, knowledgeable members of a particular institutional community, displaying to one another their understanding of the research aims, and their knowledge of how these kinds of data are constituted. Investigating how observational research is oriented to and constituted by the observed allows for a better understanding of what at that moment and in that setting is deemed recording-appropriate or -inappropriate conduct, and offers a more nuanced perspective on how data are co-constituted.


Introduction
Drawing from a large corpus of audio-visual data collected at an international university in Denmark, 1 this article explores a recurrent feature that will be familiar to many engaged in this type of empirical endeavour, namely participants' orientations to recording equipment introduced into the settings for the purposes of the research. As the larger project was designed to generate data of interactions in their natural ecology, the question arises whether this represents a 'contamination' of the data, as has been suggested elsewhere (see Speer, 2002aSpeer, , 2002bSpeer and Hutchby, 2003). As large data sets are produced at quite considerable cost and effort, and involve commitments of both the research participants and members of the research team, it seems prescient to consider any resulting by-products of the research, which may point to a corruption of the quality or validity of the generated data.
The current investigation draws on Speer and Hutchby (2003) and explores how the recording activity is worked up and topicalized in the interactions, and how such orientations play out. Whereas earlier studies used audio recordings (Speer and Hutchby, 2003;Gordon, 2013) or ethnographic field notes (Monahan and Fisher, 2010), here we consider data produced with both video and audio recording devices. Although in principle there may not be any difference between 'tape-affected' interaction registered through audio or through video, it has been suggested that video per se may be oriented to as problematic. On the other hand, video recordings have the added benefit of allowing researchers access to embodied displays of attention to the equipment, and therefore a more nuanced understanding of how these research tools are treated (see Mondada, 2012). The examples presented here show how such visible displays allow for shifts in engagement framework (Goodwin, 1981), with participants able to visually demarcate attention paid to the alternative activity orientations, and between the different social identities that they engender in the process.
Rather than reducing the phenomenon to a methodological headache for the researcher, this article will argue that the participant displays themselves can offer insight into their understandings of what it is that is being studied. Furthermore, when participants utilize such orientations as resources in the enactment of their identities as members of a particular institutional community, here a research university, they are also available to the researcher interested in understanding how the particular institutionality of an encounter is worked up in situ. As such, the phenomenon, instantiated as a member's concern, may be regarded as part and parcel of an object of study, rather than a methodological blemish. This in turn will allow a better understanding of the impact that the recording activities have upon the interactions, which may allow for a more thoroughgoing reflexivity concerning the naturalistic methods in interaction research. As such, this article is situated within a broader discussion concerning the involvement of researcher and research participants in the research enterprise, with discussions elsewhere seeking to elucidate the social practices of carrying out qualitative research (see for example De Fina and Perrino, 2011;Potter and Hepburn, 2005;Richards, 2003;Talmy and Richards, 2011) The observer's paradox in empirical research William Labov (1972) coined the phrase which has since become synonymous with the particular methodological issue which has stalked the corridors of social science research. The 'observer's paradox' describes how the object of investigation is transformed in the process of being observed, such that the research intervention leads necessarily to a contamination of the setting and a modification of the researcher's target interest. He argued that '[a]ny systematic observation of a speaker defines a formal context where more than the minimum attention is paid to speech' (Labov, 1984), and as such, there is always influence on some variables when compared with other, non-research framed settings. For Labov, the presence of the researcher, the 'observer', was deemed to cause the 'natural' linguistic features to be subsumed by something else, having the 'paradoxical' effect of rendering invisible to observers that which they had trained their sights on.
The dilemma has enjoyed widespread discussion since, with researchers from a range of social scientific fields demonstrating an awareness of the importance of minimizing the impact of the Observer's Paradox, and furthermore, making explicit the elements of their research design which seek to address the impact one has on the setting and research participants. Along these lines, authors have included in their reports accounts for the impact on the setting, and have discussed research design features that were implemented so as to offset the undesired contamination. For example, researchers may offer assurances of the temporally limited impact as the participants become accustomed to the recording equipment (e.g., Jordan and Henderson, 1995). Researchers have argued that it is the presence of video-, rather than audio-recording equipment that may be intrusive per se, opting for the latter in order to safeguard to the best of their ability an unspoilt setting. Others again, scan their data sets subsequently for orientations to the recording equipment and argue on that basis that the limited presence of such participant displays is evidence of the unobtrusive nature of the research tools (Heath, 1986). As was the case also for the present study, a researcher may further opt to be physically absent from the setting while the recording takes place, allowing the recording equipment to become 'the proverbial "piece of furniture" that nobody pays much attention to ' (Jordan and Henderson, 1995: 56). Martin (2006) attempted to circumvent this type of corruption of data, by recording people surreptitiously as they engaged in talk, only asking for consent to use the recordings as data afterwards. Ethical issues aside, with the researcher being in (over)hearing distance of those recorded, it would still not meet Potter's (1996) test for naturalistic data, 'whether the interaction would have taken place, and would have taken place in the form that it did, had the researcher not been born' (135).
Where researchers are physically present, they may account for which measures they felt would curtail their ratified presence as institutional researcher. Greer (2007), for example, mentions the wearing of casual clothing while he was undertaking fieldwork, and how he would refrain from speaking too much with the subjects. Käänta (2010) describes camera operators seeking to blend into a liminal background in the setting. In order to distract from the research activity, and to allow for the semblance of 'normality', researchers may incorporate into the research design contextual features of everyday life participants would be deemed familiar with. Hornsby (1999) allowed participants to choose environments of their own or places of regular association such as their home, and the participants in the Jarvella et al. (2001) study were invited to a domestic setting and given pasta and beer in order to create a relaxed atmosphere.
Such wide-ranging design choices, and the manner in which they feature within methodology discussions, give some indication as to the importance afforded the issue, and the manifold ways researchers attempt to control for the impact of those research tools that they introduce to the setting. Indeed, should scholars fail to include what is deemed sufficient enough a discussion with regards the impact of the research activity on the object of interest, they may be held to account by others in the field. Implicit in these accounts and others found elsewhere is that there are pristine objects of study which exist independent of the research activity, and that these become degraded by a corruption of the scene through the introduction of some or other research tool. An alternative to the aforementioned treatment of 'researcher effects' as problematic is offered by Speer (2002a) and Speer and Hutchby (2003), who describe instances where participants work together to treat the recording activity as an occasion for jocularity, or actively promoting the insignificance of the devices' impact on the interaction as desirable to the outcome of the data collection activity. Such accounts then serve to illustrate 'the presence, and possible interactional impact, of the recording device being treated as a participants' issue' (Speer and Hutchby, 2003: 329), its liminal status in the unfolding interaction not to be discounted by the analyst. Indeed, such participant displays allow for a fuller consideration of the participants' engagement in the conjointly occasioned activities, and the way in which the introduction of observers or recording devices may allow for these participation-or contextual configurations (Goodwin, 2000) to be drawn on as resources to display to one another their understandings of the activity. These epistemic displays, where participants display to one another their domains of knowledge, can furthermore be utilized as analytic resources, where they evidence the influence of the data collection activities on the settings, from the members' perspective (see Heritage, 2012).
The presence of particular artefacts is not a priori consequential to the way an interaction is conducted (e.g. Hazel, 2014). Spaces may include all manner of situated tools such as furniture, personal belongings and office paraphernalia, which may or may not be used to structure the unfolding engagement framework. They may include objects that allow for others external to the interaction to intrude into the interaction, for example windows, doorways, mobile phones and email accounts, all of which may afford channels through which externally located individuals may gain access to ongoing interaction. Yet the presence of these objects and the affordances they offer people, present or not, to impact upon the interaction is not a given. Rather, objects emerge as relevant to the interaction through an achieved orientation to them, at times vocalized, at other times brought into focus through visual displays such as gaze conduct and gesture (for example Goodwin, 2003bGoodwin, , 2007Streeck, 1996). Indeed, objects are constituted in the environment in ways relevant to the activities being carried out: as members' concerns (Suchman, 2005).
Where objects are mediating tools through which the 'absent presence' (Raffel, 1979, in Laurier andPhilo, 2009) of someone external to the ongoing encounter is represented, then this 'absent presence' itself needs to be reflexively constituted by the co-participants in situ. In doing so, research participants and the observer(s) enter into a social relationship -for all practical purposes: an asynchronous participation framework where the involvement of the researcher is temporally offset in relation to the recorded interaction. How the observed participants index the observer is the central concern here. An observer of an interaction is not simply that: the gaze is not only instantiated in the action it mediates, but may also be constituted as one of many kinds of looking, for instance absentminded contemplation, a lascivious leer, or professionally oriented scrutiny. It could then be suggested that an object of attention reflexively orients to a particular act of looking as determined by his or her understanding of the kind of observation that is taking place. In turn, this may be a determining factor with regards what said observer ultimately gets to see.

Data and method
Data were recorded featuring a wide range of university settings and activities, which included student project meetings, tutorials, language classrooms, administrative service encounters and study guidance counselling, informal social settings, and lectures. 2 The spaces varied in size from small meeting and seminar rooms to large open plan office areas and lecture halls, and recording equipment was installed in each with a view to generate data at an appropriate level of technical quality, while striving at the same time not to disrupt the activities in the settings too dramatically (see also Rendle-Short, 2006). For example, whereas research interviews were recorded using a single video camera and audio recorder, the larger lecture theatres could include as many as three video cameras, a multi-track digital audio-recorder with a sufficient number of external microphones to cover the relevant areas where interactions occur. Participants consented to the recordings being used for research, and nothing was done to conceal the recording equipment during the recordings. Aside from the research interviews, none of the recorded interactions were conducted at the behest of the researchers. Rather, the data sets were designed to include recordings of interaction which would have taken place regardless of any intervention on the part of the researcher, conventionally referred to as 'naturalistic data' of 'naturally occurring interactions' (for discussion, see Lynch, 2002;Potter, 2002;Speer, 2002b;Ten Have, 2002).
The data extracts discussed in this paper were recorded using pocket-sized digital video cameras. These visually discrete devices were chosen above the more visually conspicuous camcorder devices in an attempt to reduce the impact on the settings and the interactions. They were supplemented with audio-recording equipment placed in more proximal positions to the interactions. CLAN transcription software was utilized to process the data. This tool allows for close integration between transcript and digital media files, which in turn enables the researcher to remain alert to both the sequential organization of the unfolding talk as well as the embodied features that co-constitute the interaction (MacWhinney and Wagner, 2010). Transcripts of the vocal production were produced using transcription conventions modified from those common in Conversation Analytic research (hereafter CA; Sacks et al., 1974) and attributed to Gail Jefferson (see Appendix for conventions used here). Where visual features judged relevant to the activities are included in the analysis, supplementary video-stills are provided. As such, readers will be in some, albeit limited, position to reference the visual features described, and to judge the strength of the claims made.

Analysis
The analysis draws on CA methodology to explicate the moment-by-moment displays of participant understanding of ongoing activities. A small number of illustrative examples of an exploratory nature allow for a fuller explication of the sequences in their momentby-moment unfolding. The examples represent orientations to the recording devices at different stages of the recording activity and ongoing interaction(s). In the first two examples, participants orient to the recording equipment at the start of the recording, in one case after the researcher has switched it on, in another where the research participants operate the recorders themselves. A third example concerns participants displaying an orientation to a recording device that was activated prior to the participants arriving in the setting. A fourth sequence shows a subsequent re-orientation to the recording by participants who have earlier in the meeting already touched upon the research activity.

'Abstract yourself from it'
In the first example (Figure 1), taken from the opening stages of a study guidance counselling meeting between a student (Sara) and a counsellor (Adam), 3 the researcher has just left the room, having activated the cameras and placed an audio recorder in the centre of the table between the participants.
As the researcher leaves the room, the counsellor Adam arranges a notepad on the table in front of him, in a position conducive to writing, and brings a pen to the top left hand part of the page. Sara at this point is sitting with her body slightly at an angle from the table, and is adjusting her jacket and hair, with a notebook on her lap. The moment the door closes, Sara, whose facial expression has been somewhat neutral to this point, produces a broad smile and orients her gaze in the direction of Adam's hands, then down at the centre of the table, producing a number of laughter tokens (line 24). As she does, Adam makes a horizontal line at the top of the page, which he follows with a reorientation of gaze towards her and a stretched 'yes::' (fig. a). Elsewhere in data of the same type of activity, such co-ordinated actions on the part of the counsellor are treated as prompts for the student to formulate the reason for arranging a meeting (see Hazel and Mortensen, 2014). This is also the position where participants, if this has not yet been established already, adjust their postural orientations into an aligned F-Formation (Kendon, 1990) and stabilize this . Here, however, Sara does not provide any such next action or uptake to the prompt. What follows is a pause (line 26), during which she suspends all preparatory activities, drops her hands to her lap and onto her notepad, and turns her head to orient her gaze at the audio-recorder (fig. b). This torqued body configuration (Schegloff, 1998) is maintained throughout the pause, with Adam's gaze oriented to her. Sara then turns back to the notepad, which she now brings up to the table-top, turning her body toward Adam in the process. As she does so, she says, with a smiley voice, 'ah it's really quite strange ↑heh heh' (line 27), and returns to the activities concerned with adjusting her hair and jacket.
Having displayed her orientation to centre of the table, her comment appears to reference the recording activity, and provides an account for her lack of uptake of his elicitation, which has occasioned a suspension of the move into the counselling meeting. Adam, in his subsequent turn in lines 29 and 30, displays his understanding of the account as relating to the recording, and both offers a formulation of what she is meant to do ('I hope you can abstract yourself from it'), and reassurance that the recording activity can be suspended at any point if she so decides (line 30). As he does so, he moves his hands away from the notepad to either side of his torso and produces two 'rejection' gestures, sweeping his open hands to the sides (fig. c). He then passes the pen to his left hand, and picks up a glass of water and takes a sip. She declines his offer of suspending the activity in lines 31-33 ( fig. d), giving a further account for her reaction to the recording activity by offering that 'people just have to get used to their being recorded' (fig. e), followed by more laughter particles, which may act as post turn-completion stance markers (Schegloff, 1996), here displaying the interactional environment as delicate (cf. Haakana, 2001). Sara then gazes down, handles her notepad on the table in front of her, and produces a number of discourse markers ('so but er yeah'), then formulates the reason for her visit. At the point she moves into this topic, she raises her gaze to Adam, and as she does so, Adam switches the pen back to his right hand, and places this on the notepad ( fig. f). This restart of the move into the counselling meeting activity is further marked by a switch in facial configuration to a less animated expression on Sara's part.
We note how an orientation to the recording device is consequential to the progression of the activity at hand, here the counselling meeting. However, the way this is worked up provides further insight into how the recording activity is treated by the participants. First, Adam makes explicit that there are two separate activities that they are engaged in, and articulates an order of priority, with one activity (the research) being able to be suspended for the sake of the other (the counselling). This order is also displayed by Sara, whose body torque displays different levels of engagement in the two divergent activities that constitute the business-at-hand. Schegloff (1998) has shown how in composite engagement frameworks where more than one activity is being attended to, torqued body configurations such as this display a participant's orientation to a 'main' activity through lower body orientation, with secondary activities marked through upper body and head orientation. As such, Sara is able to display how she understands the order of import regarding the activities at hand.
A second point is that it is Adam who offers the suspension of the recording activity. Although informed consent was obtained from both participants, it appears that he has the institutional upper hand in this interactional setting. As such, the asymmetric interactional rights which characterize the main activity, the counselling meeting, seem to be consequential for the research activity too. Adam is then able to work up his social identity as institutionally ratified staff member. Not only does he position himself as stakeholder in the research activities connected to this, his workplace, but also as gate-keeper to the research object, i.e. the interactions which take place in the setting.
Finally, both participants display an understanding of the type of data they are expected to produce. Both Adam's statement that he hoped Sara could abstract herself from the recording points to an understanding that the recording activity should not become a feature of the counselling meeting, and Sara confirms this understanding, stating that one just needs to get used to one being recorded also seeming to indicate that not being used to being recorded, and possibly orienting to it, is dispreferred in the activity they are about to move into. These accounts in turn provide to one another epistemic displays of participants' understandings of research activities, positioning them as informed members of a research community.
The following excerpt is drawn from a similar counselling meeting. Here, however, the participants themselves activate the recording equipment.
'Now we just pretend they're not here' The segment below is taken from a 25-minute meeting between two participants, both students at a university in Denmark. 4 The study-counsellor, Tod, is a non-Danish student at the university. The student-client in this encounter, Mari, is from an East-Asian country. On Mari agreeing to the meeting being recorded, Tod activates the two video cameras, and then the external audio recorder. The first sequence (Figure 2) shows the participants in the process of switching on the audio recorder. Although the actual manual operation of the various pieces of equipment is undertaken by Tod, the activity is attended to by both. Mari monitors Tod's progress as he first switches on the cameras and then activates the audio recorder on the table, and they both mutually elaborate on the equipment, with Tod providing an online commentary, and Mari displaying a stance of an interested party. As can be seen in the transcribed data, talk pertaining to the projected counselling meeting is still absent from this pre-meeting sequence.
In lines 19-21, Tod switches on the audio recorder and checks the digital display to ascertain that it is working, offering a positive evaluation in line 19, which is acknowledged by Mari's 'okay' in line 20 (Figure 2). Tod subsequently expands his turn with a further qualification in 'it's measuring the voice', which he produces with a deictic pointing gesture, initially directed to the digital display on the recorder, and then, with a wrist rotation, repositioning the gesture in the direction of the table-top microphone at the centre of the table on 'the voice'. Tod then initiates a move of the hand to the central area of the table as he produces the deictic reference 'this' in line 21 and taps the table-top in the vicinity of the microphone with his index finger. For her part, Mari, directs her gaze in the direction of Tod's pointing hand gesture, then produces a change-of-state token (Heritage, 1984), 'oh' (line 22) 5 . She further displays her attentiveness to the ongoing activity by leaning in to visibly watch Tod's demonstration of the objects of attention, initially in the direction of the audio recorder and subsequently toward the microphone. Aligning with Tod's own postural orientations in relation to the two components of the recording instrument allows the participants to mutually monitor a shared focus during the emergent activities, while displaying to one another an understanding of the task at hand.
We observe that both participants are attentive to the setting up of a particular interactional activity framework, where the recording of their subsequent utterances and activities is to be occasioned. The target activity for research purposes is oriented to as being the subsequent talk-in-interaction, witnessed in Tod's assertion in line 19, 'okay (.) now we're good now we're recording'. The participants display their mutual involvement, with Tod's use of the pronoun 'we' here indexing the two participants as undertaking the research activity in collaborative partnership, rather than being solely the interlocutors who happen to produce the interaction that constitutes the object of study. In addition, his objectification of their talk through the use of a definite article 'the' in 'yep it's measuring the voice' (line 21) isolates the vocal product from the personal sources, which is further instantiated through a deictic gesture that situates said 'voice' in the area of the microphone. As such, Tod depersonalizes the objects of research and their constituent features, the vocal production and the projected meeting, while organizing the current common project, the collaborative activity of registration.
As in the previous example, the co-participants here display not only an appreciation of how the activity is to be carried out, but also the type of data that is being generated, when they make explicit an interpretation of the research activity as being one of assimilated covertness (Figure 3).
Both participants display an understanding of the data-collecting framework and more specifically an understanding of a particular type of research data, namely that pertaining to naturally occurring interaction. In lines 23-24, each follows Mari's aforementioned change-of-state token with a jointly occasioned postural re-orientation, where mutual gaze is established, and Mari produces the comment 'just like mission impossible' (line 23). This can be heard as referring to the recording activity, and only makes sense as a reference to the US television series / movie franchise of the same name, featuring a CIA operative engaged in covert operations. The utterance can then be heard as indexing the recording equipment as some form of surreptitious bugging device. Tod, in overlap with Mari's utterance, produces what can be heard as a different meta-comment on the projected activity, where he and Mari must feign unawareness of the presence of recording equipment, accompanying this with a metaphoric gesture depicting 'dismissal', waving the recorder away. The comments are subsequently followed by collaborative laughter, which parallels the laughter sequence observed in the previous example.
Although Mari and Tod have different takes on the presence of the recording equipment, one treating the recording devices as covert, the other treating the participants as complicit in a pretence, these divergent interpretations are not treated as mutually problematic. Indeed, both instantiations of recipient design in the respective formulations work to achieve local affiliation between the participants, as they each treat the conversational partner as a competent member of a particular institutional culture, one who is able to pick up on the in-joke nature of the comments. We note again that Tod uses the personal pronoun 'we', which not only serves to classify the participants as jointly represented in the activity, but does this at a point when the recording devices are already operational, and an alternative group, the researchers, are therefore present by proxy.
In this and the preceding example, we have discussed participants' orientations to the recording activity at the start of their meetings. In the following section, we have examples of similar orientations to recording devices, but with participants here making the recording relevant during activities that are already in progress.

Orientations to recording activity in progress 'Living in nineteen eighty-four'
The following sequence (Figure 4) is from recordings of an informal kitchen area used by a cohort of students. In the extract, one participant expresses surprise at the presence of one of the cameras, when a companion points it out to him.
Simon draws Antony's attention to a small video camera that had been attached to an adjacent wall, asking what 'the camera is for today' (line 11). As he does so, he enters into a dramatic embrace with Antony, positioned in a way for both partners to be visually accessible to the camera ( fig. a). Following a gap during which they disentangle themselves, Antony appears to express some confusion as to what Simon is referring to. He orients his gaze to two areas adjacent to the video camera, and then fixates his gaze on the camera. He lets out what can be heard as an exclamation of surprise with a marked shift to high pitch (fig. b). He then produces a conditional formulation, one which is left syntactically incomplete ('I mean what if', line 18), followed by another incomplete formulation (line 19), 'ah yeah had to fill out the' which he co-occasions with a pantomimic gestural flourish, seemingly depicting the using of a writing utensil. Simon at this point turns to Portia, another student present, and they co-produce a jocular sequence following his suggestion that it was 'time to make my sex tape' (lines 20-29), a sequence marked by co-produced laughter. Antony then produces an assessment of the recording with reference to the futuristic dystopia of George Orwell's 1984, where one's every move was subject to camera surveillance.
One observation concerning the above sequence is that prior to this, Antony has already noticed the recording device.
Earlier, Antony had entered the kitchen area at the far end of the space and had approached a group of students sitting near where the camera was positioned (see fig. c). As he approaches, he fixes his gaze on the camera (fig. d) and keeps it there for a full second ( fig. e), before initiating talk with the others (fig. f). This is the only object in the room he orients to in this way, which is not unsurprising as it is not an object that is normally part of the setting. In addition, on an even earlier occasion Antony entered the space when the research team were in the process of installing the recording equipment. In the data we see that Antony does not treat this as unexpected. An explanation for this may be that he, and all others who use the kitchen area had earlier been briefed on the intended research activity, and that he had signed a document giving his consent. This latter point appears to provide the basis for his statement in line 19, 'ah yeah had to fill out the' which was accompanied with a pantomimic gesture of some writing flourish.
From the sum of these parts, we may infer that Antony was already aware of the camera and the recording activity when he expressed surprise at spotting the camera following Simon's comment. His difficulty at locating the camera on the wall in front of him when it was brought to his attention, scanning the area around the camera before fixating on it and producing the exclamation token, appears then to be an elaborate performance. As Simon at this point is looking at the camera and not at Antony's face, we can surmise that this act of 'visible looking' (Goodwin, 2003a), scanning the area for an object he has already registered minutes prior to this, is performed for a different observer. The subsequent exclamation of surprise is then played out not only to his colleagues, but also for the benefit of the recording, and to the researchers for whom the data is being generated.
As in the earlier examples, the recording activity is oriented to as a bounded secondary-activity that can be demarcated from the ongoing 'primary' business in the setting. It is treated as something that participants in the setting should be able to 'abstract' themselves from, and that this is desirable for the research purposes. Indeed, participants may create some pretence of not being conscious of the recording activity, acting out an appearance of being unaware of the recorders. Lastly, it affords participants a resource for performing identity work (Antaki and Widdicombe, 1998), in being able to produce epistemic displays of their understanding of the recording activity ('we're living in nineteen eighty-four huh (0.7) big brother'). What is different in this example is that this is a liminal institutional setting (Hazel and Mortensen, 2013) where informal interactions predominate. As such, we do not see any orientation to one participant having more rights than others to determine whether a recording is made or not; here, it is the absent observer, 'big brother', who is the one afforded such rights.
In the above examples, we have looked at single instances of participants working up an orientation to the recording devices. The first examples looked at sequences at the start of recording and prior to the counselling activity commencing, while the third was drawn from later comings and goings in an informal setting at a university. In the final section, we will observe how an initial indexing of the recording activity described in the second section is revisited later in a meeting.

'For the record I'm not a hack writer'
We first encountered Mari and Tod at the start of their meeting (Figure 2). Although the subsequent talk proceeds with the study guidance activity, an orientation to the audiovisual equipment and the 'absent presence' (Raffel, 1979 in Laurier andPhilo, 2009) of the researcher-observer is explicitly occasioned at one other point, and immediately aligned with by the co-participant. In the following sequence (Figure 6), Mari provides a comment in the direction of the recorder, briefly changing the participant framework, and positioning Tod as an overhearing spectator.
In line 63, while discussing certain university-related requirements, Tod makes reference to Mari's non-university activities as a 'writer by trade'. Mari ratifies this categorization, confirming it with a synonymous term, 'hack', a term which is both less formal, as well as being notably less affirmative, derogatory even, than the one offered by Tod. Mari extends this assessment with 'I'm a professional' (line 67), produced with prepositioned and within-speech laughter particles, and followed by 'and I write all sorts of things' (line 69) which has further laughter tokens in turn-final position. Tod produces laughter in overlap with her 'I'm a professional', before offering an extended reformulation in the form of 'professional writer' (line 68). He then resumes the topic he was developing prior to this insertion sequence (line 71).
The initial assessment produced by Tod is then treated by Mari as a compliment, responding with a downgraded second assessment. Pomerantz (1978Pomerantz ( , 1984 has described the preference organization for compliment-receipt trajectories, and the constraints that the preference for self-praise avoidance places on the receipting of a compliment. Although this second assessment is formatted as an evaluation shift, it nevertheless conveys the idea of being able to make a living from one's writing abilities, and as such is not wholeheartedly dismissive. Mari is thereby able to receipt a compliment with a 'praise downgrade' (Pomerantz, 1978), while ratifying the assessment produced by Tod. However, by doing this, Mari has introduced one reading of her alternative career which may not be in accordance with the social identity construction she wants to display. This obviously remains a relevant issue to her, as we will see that she is unwilling to let it go unremarked.
Tod returns to the topic he was developing prior to the compliment sequence (Figure 7), although he only manages what can be heard as a pre-sequence projecting further topic expansion, as Mari suddenly aligns her gaze with a deictic pointing gesture to the audiorecorder on the table (line 79), and says 'for the record I'm not a hack writ(huh)er', before looking back to Tod. The within-speech laughter particle in 'writer' provides Tod with what Jefferson (1979) has called a laugh specific-recognition point, 'a locus for recipient laughter ' (1979: 82) which Tod duly ratifies with laughter tokens in line 81.
What Mari occasions with her account to the recorder is a juncture point into a side sequence (Jefferson, 1972) where the participants attend to the concurrent data-  (2008) have suggested that one characteristic of side sequences is that they can index and articulate contextual features at the periphery of the main business-at-hand, and furthermore permit the participants to engage in non-serious, playful activity, without this impacting on the main proceedings. The implication is not that such side sequences are trivial or inconsequential. As can be observed in the case at hand: on one level they may not appear to contribute to the advancing of the main institutional activity in progress, but as affiliative actions, they allow the participants interactional space or reaffirm their co-participation in the event. Taking a moment to revisit the activity they conjointly initiated prior to the counselling meeting commencing, they share some meta-commentary on it. Yet by marking the digression as a side sequence, they are able to isolate their mindfulness of the data-gathering activity, briefly making it relevant to the interaction, but without this shift in focus 'contaminating' the main body of the encounter.
As in the earlier example from the opening stages of the encounter, the concurrent recording activity is treated by the participants as a source for jocularity. The element of humour here, marked by the subsequent laughter and meta-commentary, is sparked by Mari's offer of supplementary information to the 'known-to-be-upcoming analysts' (McHoul and Rapley, 2005). Stokoe (2009) has noted how a participant introducing a stock-phrase from police discourse in another, non-institutional setting may introduce an ironic stance into the talk. It is possible that the 'for the record', a phrase more common in interactions where accounts of facts are consequential to a particular outcome, being employed here where a 'by the way' would suffice, acts as a similar type of humour device.
Further elements that precipitate a humorous orientation in the side sequence concern to whom the information is addressed, for whose benefit, and what the item is that comprises the additional information offered 'for the record'. Stokoe (2009)  about the use in audio-recorded police interviews of the phrase 'for (the benefit of) the tape', showing how the details produced at these points provide ancillary information for the benefit of recipients of the recordings who do not share the same epistemic status with the physically co-present participants. In the current data, where the encounter is being video-recorded, discrepancies in access to visual aspects are not treated by the participants as significant. With the aside to the absent researcher, however, Mari does undertake a similar action of providing additional information for how a previous spate of talk should be heard. The supplementary item of information is produced demonstrably in the direction of the audio-recorder, which here acts as a proxy for the absent observer. It is noteworthy that although Mari's speech direction, gaze and body orientation shift away from her co-participant, Tod's gaze remains focused on her. Rather than this switch in orientation having reconfigured his co-participation to constitute a shared attention to the recorder, as observed in the pre-meeting setting-up phase earlier, the new participation framework involves his having become a spectator to her aside. For her part, Mari does nothing, either on a linguistic level, or on a visual level for the benefit of the video, to ensure that the prospective recipients of the recording are able to grasp that the utterance is for their benefit. She neither marks her vocal production with any discernible adjustment in intonation, nor does she seek to affirm the intended recipient visually, by displaying any visual orientation to the video cameras present.

writes
Mari does not elaborate on the aside with any form of account for its inclusion in the talk. What she produces as additional information for the record is designed as a potentially corrective reading of her earlier formulation (line 64), marking it as not to be taken at face value. Indeed, what she does is clarify the significance of the sequential position in which it was produced, i.e. as a type of compliment receipt. Neither Mari nor Tod produce any subsequent commentary on the content of the interjection, and even though it acts as an interruption to Tod's line of talk, no further account is elicited or offered for the necessity for it at this point in time.
With an absence of any kind of orientation to an instrumental import of the interjection, there seems then to be something self-referential about its insertion at this point, a 'staging' of the act, rather than a simple 'doing'. It appears to be more a performance of a comedic routine for the benefit of Tod, 'mocking up a scene ' (McHoul et al., 2008) for amusement, rather than an act of actually supplying additional information for the benefit of the recording and subsequent analysis. In the absence of any cues to signpost the transition into an alternative participation framework, and the subsequent ease by which her co-participant picks up on this, Mari and Tod are able, however, to 'mutually and publicly display that they have supra-local (con)text(s) in common as members of a cultural order' (McHoul et al., 2008), membership knowledge that allows them to dynamically switch between alternate operational frameworks without causing interactional trouble to occur.
Once they re-establish mutual gaze (Figure 8), the participants enter into the coproduced laughter, with Mari paraphrasing the claim ('liter(huh)ary h(huh)a(huh)ck •he:', line 80), this time to Tod and in overlap with his chortling. Marked by both participants with a subsequent extended sequence of collaboratively occasioned laughter, Tod then produces what can be heard as a tongue-in-cheek reprimand'you had to put that in didn't ya' in line 83, produced with smiley voice and an emphatic hand-slap on the table on 'had', and reformulated in 85 without tag question. Although space precludes us from attempting a full analysis of this sequence, a brief gloss of Tod's referents may still provide some insight into the shifting contextual configuration of this side sequence and its implications for the general flow of activity. The deictic term 'that' in his turn could, syntactically speaking, refer to the very information Mari provides 'for the record', and indeed he not only reproduces her 'I am not a hack writer', but produces this quote with an exaggerated re-enactment of her earlier embodied orientation to the audio-recorder. In 87, Tod produces pointing gestures with both hands indicating the microphone, and then in 89 leans over in the direction of the audio-recorder when he produces the section of reported speech. Volosinov (1929Volosinov ( /1973) has written of how quoting a unit of another person's speech not only reproduces it, but offers a personal commentary on it too (see also Holt, 1996). Here, Tod's repeat of Mari's prior talk demonstrates appreciation, with Mari responding to his re-enactment with further laughter. We can see then that the 'you had to put it in' is unpacked by Tod as referring not to the repair, but to Mari having introduced an action into the ongoing institutional activity that violates the methodological considerations touched upon prior to the meeting having commenced. She has inserted, as it were, a direct orientation to the 'absent' researchers by making the recording apparatus relevant, thereby contravening the earlier suggestion of pretence. This breach is not, however, treated as critical, but 'something easily see-able on the record that for observational documentary would be consigned to the digital trash bin for deletion'" (Laurier and Philo, 2009).
Both participants respectively reconfigure the contextual configuration through their talk to the recording device (in lines 79 and 89), and in doing so modify relations to one another. The configuration of the 'aside' is reminiscent of Kang's (1998) 'triadic exchange' arrangement, where in multiparty talk, speech may be produced to a specific ratified addressee ('the mediator'), but for the benefit of another co-present addressee (the 'target'). Here, Mari actively selects the third party recipient as addressee, but the talk appears to be for the benefit of her interaction with Tod. 6 Likewise, Tod demonstrates his understanding of this participation framework, by reproducing the configuration in a subsequent mimicking display, which mirrors the original action of Mari.
This sequence allows us to consider how the ease with which the co-participants pick up on a reference to the parallel research activity provides some indication of a maintained awareness of the concurrent data collection activity. We cannot claim, in this case at least, that participants simply forget about the peripheral activity of the recording. A sudden switch in orientation to the recording activity does not, here at least, elicit either surprise or confusion on the part of the interlocutor. Yet, neither does how the recording is actively worked up as a members' issue point to participants not being able to compartmentalize the different activities, assigning the secondary activity to a backstage position.

Discussion and conclusions
The current study aims to contribute to the exploration of the social practices that constitute research activity within qualitative research, adding to discussions concerning reflexivity and 'quality' within qualitative approaches (for example De Fina and Perrino, 2011;Potter and Hepburn, 2005;Talmy and Richards, 2011). This article has looked at instances of how participants who are being recorded for research purposes orient to the presence of the recording device, standing in as proxy for the observer. By introducing the recording equipment into the ecology of the encounter, the researchers allowed for the occasioning of certain sequences in the conversation, which, admittedly, would not have occurred 'if the researcher got run over on the way to the university that morning' (Potter, 2004: 612). However, rather than these sequences representing 'unnatural' or 'contrived' sections of data, they are arguably quite the opposite. It is entirely appropriate for participants involved in such a situated activity to orient to this parallel, albeit liminal, feature, even when doing so by pretending that it is not happening. This does not then imply that the recording activity in progress renders the entire interaction as performed for 'the benefit of the tape', as the sequences discussed in this analysis would fail to make much sense if that were the case. Neither, however, can we claim that the participants, over time, forget that the social ecology of the encounter is constituted in the way that it is, with the addition of recording equipment. The recording instruments that have been introduced into the setting are utilized as resources to publicly demarcate between concurrently unfolding activities, switching between the various contextual configurations involved in each interactional space. Participants are able to steer conjoint orientations between the activities without causing interactional trouble. In turn, the resources the participants use in occasioning these switches in orientation are then also publicly available for analysts seeking to understand the impact the observation has on the interaction. Rather than going to every length to negate the influence of the observation activity on the interactional event, or treating any resulting effect as a corruption of the data, researchers should seek to understand exactly how these additions are constituted as members' concerns, those of the research participants as well as of the researchers.
In the current setting, we have seen that the research activity is utilized by the participants as a resource to negotiate their social relations and identities as competent, knowledgeable members of a particular community, here a university community. They display to one another their understanding of the research aims connected to this type of research, as well as their knowledge of how these kinds of data are constituted. The data included in the current article were generated, of course, within a university community, where the carrying out of research enjoys a privileged position. In other settings where this isn't the case, other types of epistemic display and social identity construction may emerge as more relevant.

Transcription conventions
The transcription conventions (Figure 9) are based on those developed by Gail Jefferson (e.g. 2004). Some are used in modified form for use in the CLAN software tool (MacWhinney and Wagner, 2010).

Funding
This research was funded by the Danish Council for Independent Research | Humanities (FKK).

Notes
1. Research Centre for Cultural and Linguistic Practices in the International University, Roskilde University, Denmark; www.calpiu.dk 2. The CALPIU research centre was established to investigate processes of internationalization in various types of university setting, employing a variety of methodologies, including ethnographic approaches, sociolinguistic interviews, and interaction analytic methodologies such as conversation analysis. 3. Names and all references to the participants have been changed to ensure anonymity. 4. The final example in this article is also drawn from this meeting. 5. Although Tod actually calls this piece of equipment 'the little speaker', I will assume that this is a linguistic slip, as he demonstrates he does actually understand that it is a microphone. 6. With the addressee(s) being both temporally and distally absent, however, the participant framework here would need to be characterized as asynchronous triadic exchange, and one configured on account of the presence of the recording device being made relevant to the ongoing interaction.