Abstract
Recent research on multimodal communication in the material world shows how things matter in social contexts and make a traceable difference in the unfolding of interactions. Interestingly enough, the artifacts typically used as tools of inquiry (i.e. the recording devices) are rarely deemed worthy of similar analytical attention, as if they were irrelevant or inconsequential to the organization of social interaction taking place in the field. Adopting a theoretical framework on distributed and hybrid agency, this article discusses and empirically shows how these objects play a crucial role in defining the institutional goal of the interaction and, therefore, contribute to the crafting of the data. The analysis of examples from a collection of references made to the recording devices in different research interactions illustrates the circumstances in which these references occur and the activities accomplished by participants by referring to these ‘embarrassing’ objects. In the discussion we propose that the analytical underestimation of the role of things ‘talked-into being’ in the research setting is consistent with and can contribute to a vision of research practices as a mirror of the social reality out there.
Keywords
Guido – an epidemiologist, director of a research center and genuinely interested in qualitative research – is illustrating to a group of colleagues the aims and methodology of the ethnographic research that will be carried out in some Intensive Care Units: two ethnographers will observe the everyday life and activities of the wards for a an extended period of time. They will note and record the everyday practices of the nurses and physicians, focusing particularly, yet not exclusively, on non-clinical practices that hypothetically concern or impact the prevention and treatment of infectious diseases. The aim is to have a detailed description of how practitioners do things and cope with the practical problem of nosocomial infections. After Guido’s reference to the necessity of using the video camera and the audio recorder, a colleague asks: ‘but don’t you think that the video camera will affect people’s behavior? They will naturally pay more attention!’. Guido replies by stating that this problem concerns only the very beginning of the fieldwork: ‘after a while they will forget they are being recorded; we clearly noticed this in our research on end-of-life decisions’.
Any researcher who has had to present or justify to an external audience a record-based research project, invariably had to reply to questions concerning, in one way or another, the ‘observer’s paradox’ (Labov, 1972). Researchers dispose of sophisticated arguments on how to cope with this paradox. Reflexive approaches in particular, contend that the issue at stake is neither how to minimize a supposed contextual bias to gather uncontaminated data nor whether participants forget the recording devices and return to a presupposed ‘natural behavior’. Rather, the issue at stake is what participants do by referring, indexing or taking into account the recording devices (see Duranti, 1997; Speer, 2002; Speer and Hutchby, 2003). Despite this and similar constructivist arguments, pretending that participants will sooner or later forget they are being recorded generally remains the more persuasive claim. In this paper we advance the hypothesis that pretending that after a while people will forget they are being recorded is more of a ‘rhetorically useful’ claim (Speer, 2002: 521) than an evidence-based statement. It is based upon, and at the same time supports, an implicit vision of methodological tools as ‘“transparent windows” into social reality’ (Mondada, 2009: 69). This concealed epistemological assumption still pervades ‘much of the literature on methodology’ (Mondada, 2009: 69), although the interpretive turn in social sciences had profoundly questioned it decades ago (see among others, Rabinow and Sullivan, 1987; Steier, 1991). 1 In line with the reflexive approach to data collection and with Speer’s critical stance toward the sustainability of the ‘natural vs. contrived data’ distinction (Speer, 2002), this article analyzes some of the activities accomplished by participants in referring to the recording devices, namely the epistemic activities. By adopting a theoretical perspective centered on the ‘agency of things’ (Caronia and Cooren, 2013; Cooren, 2010; Cooren and Benchercki, 2010; Orlikowski and Scott, 2008), the paper adds to this line of inquiry a reference to the ‘passive side’ of any human activity and the ‘active role’ of things in shaping, organizing and constituting human practices, including scientific ones. After a review of relevant literature on distributed and hybrid agency (Caronia and Cooren, 2013; Cooren, 2010, 2012; Latour, 1996, 2005) and on multimodal interaction in the material world (see Streeck, 2010; Streeck et al. 2011a), we will analyze examples from a corpus of ‘references to the recording devices’ collected during different studies or recorded as second-hand material over several years. They provide empirical evidence of a very simple fact: participants do not necessarily nor always forget they are being recorded. They also illustrate some circumstances other than the canonical ‘beginning’ under which reference to the recording device is deployed: a) in a problematic interactional environment, and b) when the interactional scene splits into two and participants have to cope with multiple parallel deictic focuses.
Previous research as well as other data in the corpus, show that participants do many types of activities by referring to the recording devices: they can use these artifacts as a resource to stress the problematic nature of other participants’ behavior (Hutchby et al., 2012), to solve local practical problems as often happens when children are concerned (Fasulo and Pontecorvo, 1999), or to engage in playful interactions building on the surveillance-related affordances of the device (Speer and Hutchby, 2003). Here we will restrict the analysis to epistemic activities that formulate what is going on, assess its recordability (see Iversen, 2012) and, therefore, contribute to defining what can or should count as ‘data’. In particular, the examples illustrate that by referring to the recording devices the participants: a) ratify the ongoing behavior as designed for the institutional purpose of the event; b) assess the recordability of the just performed or imminent talk, and c) enlighten the (methodologically and epistemologically problematic) transitional zone between what is going on (the recordable) and its representation (the record).
Once transformed into an analytical focus, the practice of ‘referring to the recording devices’ reveals its indexical and reflexive properties (see Lynch, 2000). It is indexical as it is embedded in, points to and ultimately depends upon the circumstances, i.e. the physical presence of the artifacts here and now ostensibly available for participants. It is reflexive as it frames, shapes and ultimately constitutes what is going on. The indexical and reflexive properties of this practice make it relevant to argue in favor of the contribution of the artifacts participants refer to, in the unfolding of the interaction. Adopting an analytical framework on hybrid and distributed agency (Cooren, 2006, 2012; Latour, 1996, 2005; Mondada, 2009, 2011), we contend that the physical presence of the artifacts in the interactive scenario and their technical affordances (Gibson, 1977, 1979) are precisely what allows participants to accomplish the epistemic activities mentioned above. In short, these material objects have a traceable ‘agency’, i.e. they make a difference in the unfolding of the events that depends on their material features and recognizable functions. 2 As we will show, participants often allocate agency to things in ways that are visible and traceable. Acknowledging the ‘agency of things’ (e.g. what the recording devices actually do in the research setting) is therefore consistent with an analytical approach oriented to the members’ point of view.
Overcoming a linguistic reductionism (Erickson, 2010; Levinson, 2005; Streeck et al., 2011a) and a ‘restricted analytical geography’ (Goodwin, 2011: 184), we describe these epistemic activities as jointly produced by human and non-human participants. Sometimes at least, the recording devices ostensibly cooperate with humans in conferring a public nature to data, sustain informants’ situated construction of relevant differences in their recordable behavior (problematic vs. unproblematic, relevant vs. irrelevant; private and sensitive vs. public and displayable), and contribute to defining the recorded behavior as a document of the object of inquiry. We propose considering the recording devices as epistemic agents which make a difference in the making of the data in ways that, in some circumstances at least, can be traced and analyzed. They do so not as extrainteractional objects but as inter-actants having a part in a ‘network of social and material, linguistic and non-linguistic agencies which shapes the activity setting and the relevance and force of the linguistic performances occurring within it’ (Licoppe and Dumoulin, 2010: 213).
In the discussion we contend that the claim that participants forget the recording device is often an undemonstrated (and perhaps indemonstrable) statement. Even when the recording devices are not ‘demonstrably relevant for the parties’ (Schegloff, 1997: 183) we cannot infer upon this ‘absence’ that parties forget these not ‘so ordinary’ objects. We argue that assuming participants forget the recording device is a premise consistent with and functional to maintaining the illusion – sometimes implicit even among qualitative researchers (see the notion of ‘fiction’ advanced by Potter and Wetherell, 1995) – that with due caution ‘scientists can safely avoid disfiguring the picture of nature with their own fingerprints’ (Gergen and Gergen, 1991: 76). This unstated empiricist premise may orient the analyst’s selective attention to what happens during the research interaction and possibly causes her to not note what could infringe the ‘cargo-cult view’ of scientific practices’ neutrality (Rabinow and Sullivan, 1987: 5) that seems to be still present among some social researchers. Recursively, the premise is (re)instantiated and maintained for ‘another next first time’ (Garfinkel, 2002: 92) each time the researcher ignores (or minimizes) those small cues, words, gestures, gazes or stretch of talk through which participants – in some circumstances – refer or point to the recording devices. As Latour put it, science ideologies and scientific practices create each other.
In analyzing the practice of referring to the recording devices, this paper builds on the tradition of studies on scientific practices and ordinary action (Latour, 1996; Lynch, 1993) and adds to this line of inquiry by empirically showing the constitutive (although often ‘embarrassing’) role of things in the crafting of scientific data.
The agency of things: perspectives from social semiotics
The study of the material features of social places (i.e. workplaces, classrooms, households) and of things that inhabit these places has a long-standing tradition in the social sciences (see among others, Appadurai, 1986; Augè, 1995/1992; Baudrillard, 1996/1968; Cieraad, 1999). Social research ‘on the side of things’ underlines how and to what extent the material features of everyday life contexts are more than an inert background for social practice. Human artifacts are part of the environment and – as the affordance theory (Gibson, 1977, 1979; Hutchby, 2001; Norman, 1988) reminds us – any environment (or object) has cues that indicate possibilities for action. Disregarding any radical technological determinism, the sociology of techniques (see Akrich and Latour, 1992) and the semiotics of objects (see Landowski, 2002) underlined that things set constraints and create possibilities that work in favor of the genesis of certain meanings and courses of action rather than others. Once designed and introduced into the interactional scene by humans, texts, artifacts and objects of any kind make sense and have an agency on their own (Latour, 1996). They make a difference and have effects thanks to us but also despite us. We might use them, ‘but [we] also might be at their mercy’ (Brummans, 2007: 724).
Following Latour’s critical stance toward a sociology without objects (Latour, 1996), research on artifacts and texts as meaning-making devices (Aakhus, 2007; Barley et al., 2012; Brummans, 2007; Bruni, 2005) and studies on the communicative constitution of social environments (Cooren, 2010; Taylor and Van Every, 2011) stress and empirically show that the material disposition of a place, its spatial organization, texts and documents, the location of artifacts (e.g. the technical equipment of an office and the diffuser, Cooren et al., 2012) are consequential to the definition of what is taking place in an interaction and matter for the participants.
The very question is not whether they matter, but for whom and how we (as analysts) know that (Schegloff, 1997). Recent approaches in communication strongly advocate for never leaving the terra firma of interaction (Cooren and Bencherky, 2010) to grasp the agency of things at work (‘materiality in practice’, Pentland and Singh, 2012: 289). Following an ethnomethodological perspective on the constitution of social life (Garfinkel, 1967, 2002), these approaches assume that the material aspects of any given context are embodied in the ways people carry on and order their practices (not necessarily, nor always, discursive). The agency of things is assumed to be visible (observable, traceable and analyzable) in the design of people’s actions and interaction, whether members know and acknowledge it or not: as far as the agency of things is displayed (embodied, enacted) in actions, it need not be in the mind or the words (of participants). Acknowledging the agency of things in the unfolding of the interaction means recognizing and theorizing ‘the passive dimension of any action’ (Cooren and Bencherky, 2010: 46; emphasis in original), i.e. the role material things play in leading people to act as they act. New perspectives in social semiotics radically assume that there is no ontological primacy of humans over material reality in meaning-making; rather, they belong to the same ‘ecology of representation systems’ (Bateson, 1972; Gibson, 1979). Therefore, meaning and action are conceived as a joint accomplishment of both human actors’ (linguistic) practices and material things’ affordances.
Interestingly enough, a similar ‘material turn’ recently concerned fields of study traditionally and programmatically centered on language and talk-in-interaction.
Things in language and interaction studies: a renewed attention to the material world
From speech act theory (Austin, 1962; Searle, 1969, 1979) to conversation analysis (Heritage, 1984; Pomerantz and Fehr, 1997; Sacks, 1992; Schegloff, 2007) and interaction analysis (Fairhurst, 2004), scholars interested in the detailed study of human interaction theorize and analyze their object of study in terms of what people actually do (and how they do what they do) in talking to each other. Language use and its emergent property are conceived as the backbone of culture (Heritage, 2010). This so-called language centered approach (or even logocentric bias, Erickson, 2010) has been questioned as the analytical interest moved from telephone conversations in the early studies to highly contextualized and multimodal workplace talk and activities. In these contexts it is quite evident that if people do things with words, they also do (and say) things with things (Caron and Caronia, 2007; Streeck at al., 2011a).
In their overview of research on the local production of accountable activities in workplace studies, Streeck, Goodwin and LeBaron acknowledge that ‘one of the hallmarks of this research program was recognition of the paramount importance of physical objects in the conduct of work-related activities’ (2011b: 10). This research program officially (re)legitimizes the domain of things as both an analyzable and an analyzer of human conduct. According to Mondada, multimodal analysis typically focuses on (among other dimensions), ‘the investigation of spatial and material resources including artifacts and technologies manipulated by co-participants’ (2011: 207). Human interaction is conceived as accomplished through and mediated by heterogenic semiotic tools. Within this broadened analytical perspective, things and social interaction are conceived as components of the same hybrid ecology; they are intertwined and, therefore, shape each other.
A question arises as to what things the analyst should consider and pay attention to. After decades of critical theorizing on the notion of context (see Cicourel, 1980, 1992; and Schegloff, 1987, 1992), there is a ‘safe’ answer to this issue: we should pay attention to those ‘things that are visible for the analyst or made live by people in the scene themselves’ (Antaki, 2012: 93). If our analytical attention should be paid only or at least to these things is still a debatable affair (see Caronia and Mortari, forthcoming; Waring et al., 2012). However, there is general consensus among scholars interested in multimodal communication on the premise that sense-making is distributed on and accomplished by different ‘participants’ whether they are human or non-human entities. From this theoretical standpoint, things are just as ‘context-shaping and context-renewing’ (Heritage, 1984: 242) as human communicative actions are.
This research program concerned a wide range of workplace studies interested in analyzing the detailed crafting of naturally occurring social practices: from clinical work in the hospital (see Bruni, 2005) to the chair-organized dentist’s work (Marsciani, 1999); from the airport technicians (Goodwin, 1996) to courtroom distributed hearings (Licoppe and Dumoulin, 2010). Interestingly enough, in Streeck, Goodwin and LeBaron’s long and quite exhaustive list of workplaces and activities studied from this analytical perspective (Streeck et al., 2011b: 10–11), only a few references concern studies on the constitutive role of things that researchers use as methodological tools to investigate social reality: objects that are tools for inquiry appear to be less investigated than objects that are objects of inquiry. Some seminal exceptions should be noticed. Investigating laboratory life and natural sciences research, Latour shows that scientific data are ‘handmade’ constructions depending on the artifacts used by the researchers (Latour, 1999; Latour and Woolgar, 1986): maps, diagrams and the technical tools scientists use to investigate and represent reality (see the Topofil Chaix, Latour, 1999: 43) have their proper agency and cooperate in shaping data and, therefore, knowledge. A kind of similar attention to the constitutive role of artifacts in scientific research has been paid to contemporary hard equipment of research in social interaction (see Ochs et al., 2006; Mondada, 2009).
Illustrating the methods and techniques used in recent ethnographic research on working-family everyday life, Ochs et al. (2006) focus on ‘how to do’ video-recording, the suitability of some techniques over others with respect to the research aims and questions, and the overall unique potential of audio and video recording in research on language and human interaction. Although some epistemological alerts are given (i.e. procedures might alter the very process the research seeks to clarify, p. 407; video in no way assures observer objectivity, p. 389), the constitutive role of the researcher’s ‘material stuff’ is in some sense reduced to a potential bias that should be minimized (on the ‘theories of bias’ as a remedial perspective, see also Briggs, 1986). In quite a different perspective Mondada (2009) focuses on video-recording practices (and video-editing practices) as second order reflexive activities: these object-mediated scientific activities are constitutive practices as much as the ordinary activities they pretend to ‘just’ record. By comparing professionals’ and researchers’ video-practices, Mondada ultimately shows that there is no theoretical or empirical reason to assume that the artifacts we use as research tools are immune to the agency that we are quite ready to attribute to the artifacts that are the objects of our inquiry.
This article contributes to this line of inquiry by investigating the context-shaping role of softer and less intrusive recording devices: the audio-recorder (typically used in interviews) and the single fixed video-camera. These devices are different as to their presence in the perceptual field and – moreover – in terms of what they record. However, they share a common basic trait – the recording function – that allows them to be analyzed as members of the same category of objects. As it has been shown for the role of the camera in family therapy sessions, participants do use and refer to them in ways that cannot simply be ignored: ‘the assumption that those whose behavior is being recorded will ultimately “acclimatize” to and hence ignore [the device’s] presence’ (Hutchby et al., 2012: 677), is just another way to establish a correspondence between knowledge and reality that ignores what sciences ultimately do: ‘through successive stages they link us to an aligned, transformed, constructed world’ (Latour, 1999: 79).
The following sections empirically illustrate the problematic nature of claiming that informants will ultimately ignore the recording devices and the amount of things they do by referring to these objects. For the aims of this analysis, data were transcribed to provide an intermediate level of granularity (Jenks, 2011; for variations in transcription, see among others Bucholtz, 2007; Duranti, 2006; Mondada, 2007; Ochs, 1979). They are analyzed using a CA-inspired approach combined with an interactionally-grounded analysis of the constitutive role of things in the unfolding of the event (see Caronia and Cooren, 2013; on eclectic approaches to discourse-in-interaction, see Kerbrat-Orecchioni, 2010).
Audio-recording as a stage-making activity: an illustration
The following two excerpts are from interviews on the policies and practices adopted in preprimary schools in regards to the presence of ethnic minority children. In both cases, the school coordinator introduced the research project to the teachers and managed the selection of the one that would be interviewed. The interviewer double-checked the teachers’ consent to be recorded before turning the audio-record on. The interviews took place in the school where the teachers worked.
Excerpt 1
The researcher has already explained the aims of the research and asked the first general questions concerning the school. She now introduces the core topic of the interview by asking how many foreign children are attending the school this year.
1 Res quanti bambini stranieri ci sono quest’anno qui da voi?
2 Tea cinque. 3 Res cinque. (.) e ↑di che nazionalità sono? (2.0) 4 Tea → u::m (0.2) non- (0.2) sta registrando? 5 Res si=vabé 6 Tea → si?=sta registrando? 7 Res si 8 Tea ci so:no: u:n un filippino, u:n (.) pola:cco mi sembra
The reference to the recording device (line 4) occurs after a question concerning the children’s nationality, 12 minutes after the beginning of the interviews, and is lodged within a problematic interactional environment. The question (line 3) is followed by a significant pause and when the teacher begins to respond (line 4) she shows some of the typical vocal markers signaling a problem in aligning with what is projected by the just-preceding turn (Schegloff and Lerner, 2009): a filler, a pause within the turn, a cut-off, another pause within the turn. After delivering this alert about the problematic nature of delivering the second pair part, the teacher breaks the question/answer sequence: instead of providing an answer she produces a question (on producing a non-answer response, see Stivers and Robinson, 2006) that opens an insertion sequence (Schegloff, 1972, 2007; see also Antaki, 2002) dedicated to checking if the recording device is working. After the completion of the insertion sequence, the answer is finally produced (line 8) although not in the expected second position: the utterance shows some hesitation marks and a modal verb as if the teacher is trying to remember the children’s nationalities. She remembers only two out of five and seems to be unsure.
The teacher uses plenty of ways of communicating a problem in aligning with the ‘project undertaken in the first pair part’ (Schegloff, 2007: 58), i.e. providing the requested information.
Note that the researcher’s question implies a strong presupposition (Bolden, 2009): that the teacher has this information. The question basically assumes that this teacher knows what the nationality of the children attending the school is. The question positions the teacher as the spokesperson of the school (and not, for instance, of her classroom): in the participant status projected by the question (Bolden, 2011), not knowing, not remembering or not being sure of this particularly sensitive data could be perceived as inappropriate by the interviewer. A typical ‘face problem’ (Goffman, 1967) is at stake here. In this problematic interactional environment, the teacher makes a reference to the recording device twice.
By checking if it is on, the teacher does a certain number of things. First, she shows an orientation to the artifact and to the institutional role it plays in the interaction (i.e. she is clearly aware of it), second, she uses these references for interactional purposes (i.e. delaying the answer) and, third, she exhibits her concern about whether her incoming answer will be ‘on record’ (i.e. textualized, available, and analyzable) or not.
The (received as) problematic question makes relevant the reference to the recording device that suddenly shifts from the background to the foreground of the perceptual field. Recursively, the (reference to) the recording device highlights (for the interviewer and for the analyst as well) the problematic nature of not knowing or not exactly knowing the information requested. Two entities with different ontologies (Latour, 1996; Streeck, 2011) are actively producing a face problem: the spoken question and the recording device. In the specific stage created by the research interpersonal and material setting, the teacher’s performance may be assessed as inadequate. As we will see, face problems occurring in research may have crucial epistemic consequences as they invariably reveal the constitutive disalignment or sliding zone between the reality under scrutiny and the version produced for scientific purposes.
Changing stage, changing footing: tracking the epistemic work of the recorder
The next excerpt illustrates another environment where the recording device becomes ostensibly relevant for the interviewee: the split of the interactive scene into two parallel scenes and the participants’ problem of orienting to multiple deictic focuses (shared speech situations, Givón, 1989: 74; see also, Bazzanella, 2004) in appropriate manners. The interview started 45 minutes before the following exchange occurs. The teacher is now illustrating an intercultural education project by showing the interviewers a picture book made with the children participating in the project. The story was about an imaginary little fish travelling around the world and visiting different countries. During this account, some children enter the room where the interview is taking place.
Excerpt 2
1 Teac poi inve:ce: c’era la guerra c’è stato il rifiuto, 2 il loro animale era il pesciolino 3 ↓adesso poi lasciam passare ‘sti ragazzi ((a group of children enter the room)) 4 Teac 5 6 Res1 vabè non c’è problema 7 Teac stan facendo un gioco particolare recentemente 8 per cui(.)bisogna battere molto le mani 9 Res1 [hahaha 10 Res2 [hahaha
Clearly enough, the fact that the teacher has been talking on the record for 45 minutes does not imply eo ipso that she forgot she was being recorded: something makes the artifact shift from the back to the foreground of her attention. When the children enter the room, the interactional scene splits in two: the new deictic focus makes it relevant for the speaker to act as a teacher while at the same time participating in the interview as a spokesperson of the school. This change of ‘footing’ (Goffman, 1981) is clearly visible in the conversational shifts of the speaker. The teacher changes the topic (from the educational project, line 2, to the children entering the room, line 3) and the verbal activity she is accomplishing (from an account, lines 1–2 to a directive, line 4); she changes the addressed recipient twice by designing two semantically, prosodically and grammatically different directives (‘Clap your hands’ (line 4 and ‘turn it off’, line 5; note that in Italian this shift is grammaticalized in the desinence of the imperative form). The speaker clearly orients to the new deictic focus and at the same time she maintains an orientation to the main interaction she is engaged in.
The reference to the recording device (‘no, turn it off’, line 5) occurs right after the opening of a potential side-conversation (sideplay, Goffman, 1981) where the speaker engages in some subsidiary matter with respect to the main interaction (i.e. the directive addressed to the children, line 4). By referring to the recording device, the speaker opens a sequence embedded within the main conversation (Jefferson, 1972): the opening points to the main conversational activity (i.e. the interview, ‘turn it off’) and connects it to the side interaction with the children (‘otherwise I cannot say it’). By clarifying why she said what she just said (turn it off), the speaker frames what she projects to say as unsuitable for the institutional aims of the main interaction. The reference to the recording device works as a preface defining the imminent discourse as non-recordable, yet at the same time it works as an afterword for the previous one: it implicitly frames the already provided (and recorded) talk as recordable.
The fragmentation of the interactional scene (Licoppe and Dumoulin, 2010) and multiplication of the participants’ social identities (informant vs. teacher) makes the reference to the recording device relevant: what is appropriate for the ongoing interaction and characters (i.e. the recording device is on) may be inappropriate for the side interaction and character the speaker is engaging in. In this complex interactional circumstance, the artifact is used as a resource to define ‘what we are doing’ (i.e. an interview), the nature of the discourse participants are producing (i.e. a public document about the object of inquiry) and the speaker as the informant who authors this document and is in control of its representational nature. The artifact sustains the participant’s situated construction of relevant differences in her identities and recordable behavior: an informant’s public informative talk vs. a teacher’s educational and sensitive talk, relevant vs. irrelevant discourse with respect to the topic and institutional goals of the interview. The recorder hasn’t been turned off: interestingly enough the interviewers didn’t receive any further details about the ‘particular game’ the children were playing those days.
As similar analyses of participants’ iterative negotiation of things that should be ‘on or off record’ have documented (Hutchby et al., 2012; Speer and Hutchby, 2003), the public and documentary nature of the discourse collected so far is displayed by the informant during the unfolding of the interaction. What was negotiated at the beginning of the interaction (when asked for consent) is ratified once again by referring to the audio recorder’s presence and function. The object participates in this epistemic work of filtering the occurring behavior and framing it (or, more precisely, part of it) as ‘data’.
Assessing recordability: the sliding zone between ‘what is going on’ and its textual representation
The next excerpts report ‘references to the recording devices’ recorded during two different studies on ‘talk-at-work’. In both cases the reference occurred right after the production of a component that the speaker formulates as ‘out of key’ by referring to its ‘being on record’. In these cases, references to the recording devices are complex forms of self-repair (Schegloff et al., 1977; Schegloff, 2007): they signal that there is a problem in what has been already said and what the problem is.
The first study concerned informal talk in institutional settings. The participant observer wore the recording device that was on during selected informal moments and activities of the teachers’ daily work at school. The colleagues knew that Nicole was carrying the recorder for research purposes and gave their consent to be recorded.
The exchange reported here occurred approximately 10 minutes after the beginning of a conversation collected on the ninth day of fieldwork. The colleagues are in the hall of the school and are trying to solve a practical problem: they want to construct a tree using natural branches but they are not able to cut them.
Excerpt 3
1 Mar solo che si taglia qualche fibra. (0.3) 2 qualcosa sta cedendo eh, si sente il rumore. 3 °hu::ff ° sto sudando cinque camicie ma va bene. 4 Nic hahaha 5 Mar → °che due fiocchi° ↑non registra vero? 6 quando ho detto che due fiocchi. quel coso lì? 7 Nic hahahahah $non ti preoccupare$ hahah
In this example the reference to the recording device occurs immediately after the use of a mitigated variation of a (recognizable) coarse, although quite common, Italian expression indicating how boring something is (line 5). The speaker explicitly ascribes her concern (whether ‘that thing over there’ is recording, line 6) to the utterance of this formulaic expression by reporting her own speech (see the use of a saying verb plus a literal repetition of the repairable component, line 6). The reference to the audio-recorder has a meta-pragmatic function (Silverstein, 1993): it comments on what has just been said and frames it as unsuitable to be on the record. The speaker frames her words as inappropriate for the social encounter that takes place in that moment: it is not so much a naturally occurring interaction between colleagues but rather a recorded ‘naturally occurring interaction’. Although the study was introduced as an investigation on informal talk between colleagues, there is a level of informality that sounds problematic (i.e. repairable) at least for one of the participants. The audio-recorder helps in making a finer distinction within the fuzzy category of informal talk and between a naturally occurring social event and a recorded one.
As in Excerpt 1, the speaker refers to the audio-recorder by pointing to the peculiar activity it is performing (i.e. recording): the participant ostensibly perceives it as an agent doing something and having an agency on its own. 4 This agency consists specifically in transforming what is going on in a public document about what is going on (i.e. the record). Although subtle, there is a discrepancy between the so-called ‘natural occurring’ behavior and its document, between the recordable and the record, the living experience (the actual talk-in-interaction) and the data: according to the speaker at least, not every kind of discursive behavior is or should be considered as a specimen of ‘informal talk between colleagues at work’.
The next excerpt reports an exchange which occurred during a supervision working session where five home-based social educators discuss one of their problematic cases with a psychologist. The working session lasted 1.5 hours and was recorded by means of a fixed video-camera and an audio-recorder. The researcher was a member of the team. The episode occurred 50 minutes after the beginning of the meeting: it is the third time participants refer to the recording devices. The previous two playful instances typically occurred at the very beginning of the encounter (‘this is typical initial bullshit, you’ll cut it afterword’) and right after a technical problem that forced the researcher participant to intervene on the video-camera (‘and then? what if she gives details to journalists? No, don’t record it’).
The third reference reported here occurred within the third macro-sequence of the encounter: the analysis and assessment of the case. The supervisor is holding the floor, elaborating on what the main teller has said thus far and providing her with some interpretation and guidelines on how to cope with the problematic case. The reference to the recording device occurs within a long turn where the supervisor basically confirms the educator in her function and valorizes her role: she is not there to engage in any psychological counseling or support but to provide the mother with a concrete scaffolding for doing what she is supposed to do for the well-being of her daughter.
Excerpt 4
1 Sup allora >se l’obiettivo è portare la bambina dalla logopedista< 2 → e <tiro fuori tu:tti i miei contro coglioni> da educatrice. 3 → registrato. vabè 4 → ↓i supervisori che dicono le brutte: parole. nei convegni verrà il bip. 5 eh. (.) allora è fond- é=é probabilmente la tua funzione è molto importante
As in the previous case, the reference to the recorded quality of the speech occurs right after the production of an instance of coarse language: the Italian metaphor in line 2 indicates someone who suddenly stands up, shows her hidden resources and efficiently acts in a given situation. The reference breaks the discursive trajectory and opens an insertion where the speaker expands on the inserted topic. After downgrading the relevance of the fact that her words are on record (line 3), the speaker raises an impression management issue by referring to her institutional identity and formulating the already spoken words as ‘bad words’: what is inappropriate is not so much the ‘bad words’ as such, but rather their being used by a supervisor (line 4). Note that she generalizes her concern using the plural form of a membership categorization device (‘supervisors that say bad words’, line 4) as if her idiosyncratic utterance was (or might be considered) representative of the supervisors’ talk. In so doing the speaker orients to the documentary nature of what is ‘on record’ and to the typifying process taking place in the research. This orientation to the research purposes is displayed also in the closing of the insertion. The way the speaker formulates the place (Schegloff, 1972) where the recorded behavior will possibly be presented (‘in the conferences’, line 4) and the reference to the presumed-obvious editing practices (Mondada, 2009) necessary for a public presentation (‘there will be the bip’) indicate the speaker’s epistemic status as to the nature of what is going on: the social interaction is being recorded for research purposes and it will be submitted to a series of successive transformations that produce ‘an aligned, transformed, constructed world’ (Latour, 1999: 79). Once recorded, the interaction becomes the data; data are used as a specimen of something more general and they are not (i.e. do not correspond/overlap with) what they stand for. Interestingly enough, both the typifying process (from the local idiosyncratic talk to the membership category’s talk) and editing practices are mentioned as if they were taken for granted or perceived as typical of the researcher’s job.
The examples above present some common features. In both cases the speaker accomplishes a certain number of things by referring to the recording device or its consequences: they assess the recordability of what is going on by differently distributing this property to the behavior under scrutiny. Thus they establish an epistemic difference between reality (the recordable) and data (the record), and position themselves as the epistemic authority in charge of such a decision. In both cases the reference occurs right after an instance of coarse language.
Establishing if it is the linguistic behavior that makes the presence of the recording device relevant or if it is the recording devices that lead the speaker to pay attention to her own words is problematic and, we suggest, nonsensical. Within a hybrid agency paradigm, things and social interaction are conceived as components of the same ecology, shaping each other and sharing the agency of making some dimensions relevant over the others: the recorder gives relevance to some spoken words that, in turn, make the presence of the recording device relevant.
Both actants (the recorder and the speaker who refers to it) cooperate in splitting the ongoing interaction into two sliding interactions as if participants were simultaneously performing on two different stages that do not necessarily nor always overlap or should overlap. Far from being granted a priori, the correspondence between the ongoing interaction and the recorded one (i.e. textualized, represented) is negotiable and negotiated. Although this negotiation may be interpreted as simply addressing a typical impression management problem (Goffman, 1959), it has huge epistemic consequences: negotiating the correspondence between the occurring behavior and its recorded version means establishing what counts as specimen of what. This epistemic work is distributed between human and non-human participants. As the data show, informants ostensibly use these devices as an epistemic resource, build upon them to establish some crucial differences in their recordable behavior and to clearly point to the sliding zone between ‘reality’ and the version crafted for research purposes. Yet, the artifacts have a traceable agency: their presence, affordances and recognizable function (i.e. recording) define the situation as submitted to a textualization process; they sustain, channel and ultimately make relevant for the informants to assess the recordable nature of what is going on and, therefore, contribute to defining what is considered as a document or data of what.
As we contended (see above), no ontological nor pragmatic primacy should be given a priori to one component of the interactional scenery over another: linguistic and non-linguistic entities cooperate in shaping the ongoing interaction, formulate what is going on and accomplish the epistemic work of distinguishing reality and its representation, i.e. the data.
Downgrading the constitutive role of the recording devices: the researchers’ replies
In this section we propose some interpretations on the epistemic stance of the researchers, building on the analysis of how they receive the informants’ references to the recording device and behave in the insertion sequences opened by these references.
In Excerpt 1, the researcher answers the speaker’s first question about whether the recording device is on, by confirming it is functioning and downgrading the relevance of this information (yes=that’s ok, line 5). She answers the second question by a simple yes-answer (line 7). In both cases she provides the information requested by limiting her contribution to a type-conforming reply to a yes/no question (Raymond, 2003), adding only a downgrading component the first time.
In Excerpt 2 we notice a similar, although more elaborated pattern: one of the researchers replies to the informant’s request to turn the recording device off (‘turn it off’, line 5) by not aligning to the request (he does not turn the recording device off) and providing a justification for keeping it on. The justification is a form of downgrading (‘that’s ok no problem’, line 6).
In both examples, the researchers engage in a work of reducing the relevance of the device presence, try to close the insertion sequence ‘as soon as possible’ and to restore the main interaction as if the recording device was not there.
Excerpt 3 shows another way of receiving references to the recording devices: laughter (line 7).
Since the studies by Gail Jefferson (Jefferson, 1979, 1985; Jefferson et al., 1987), laughter has been carefully investigated as a relevant multimodal and multifunctional component of talk, which plays a crucial role in establishing meaning (not necessarily nor always humor, see Potter and Hepburn, 2010) and shapes the social organization of the interaction (see Glenn, 2003). It may be used as a resource by participants to cope with delicate aspects of the interaction (Haakana, 2001), as a marker signaling or masking troubles in talk (e.g. obscenity, Jefferson, 1985) or to manage affiliation between co-participants (Glenn, 2003; see Excerpt 2, lines 9 and 10, where the researchers’ laugh seems occasioned by the projected-as-embarrassing topic). It is commonly assumed that the laugh is highly indexical and ‘locally responsive’ (Sacks, 1974: 348): it refers and defines the element immediately preceding the laugh as laughable (see Jefferson et al., 1987; Schenkein, 1972).
In Excerpt 3, laughing is produced by the next speaker (the researcher) immediately after the current speaker’s reference to the recording device. In the researcher’s turn, a laugh precedes and follows a component where the researcher mitigates the informant’s concern about whether the recording device is on (laugh + don’t worry + laugh). The laugh defines the preceding reference to the recording device as laughable and sustains the researcher’s activity of minimizing the relevance of the artifact presence pointed to by the other participant.
In short, while the artifact is ‘talked into being’ (Heritage, 1984: 290) by informants, it is (tentatively) talked ‘out of being’ by the researcher. How can sense be made of this struggle in which an artifact ‘is evoked and made actionable in and through talk’ (Heritage, 1997: 222) in so many different ways by participants?
We advance the hypothesis that the accountable and understandable character of these activities a) implies a form of material agency (Pickering, 1995) and b) reveals and, at the same time, (re)instantiates the researcher’s epistemic stance toward her data.
The object being referred to is not just any kind of object: it is a physically present recording device. It is only by taking this into account that we can make sense of the participants’ actions and reactions to its peculiar contribution (suitable or unsuitable) to what is going on. Simply put, if it wasn’t there doing something (i.e. recording) none of the above analyzed activities would probably have taken place. The artifact contributes to producing the context as a context of a certain kind, in defining the institutional goal of the encounter and – as we have seen – it sustains the informants’ pointing to the epistemic difference between the recordable (i.e. what is going on) and the record (its textual representation). Far from being presupposed by the analyst, this material agency is traceable in the ways participants design their activities: they actively cope with its presence and functions as they skillfully react and respond to what it indicates. Clearly enough, this agency does not determine the participants’ behavior: the informants make the artifact relevant by talking it into being; the researcher responds by locally constructing its irrelevance. Although in different ways, they both have to cope with it.
We propose that the researcher’s way to cope with the agency of the artifact (minimizing the relevance of the recording device) is coherent with a positivistic epistemic stance: data that exhibit cues (fingerprints, Gergen and Gergen, 1991) of their indexical nature (i.e. their being locally responsive to the interactional and material environment) indicate the reflexive nature of the research process, i.e. its constitutive role in producing, shaping or renewing what the researcher is supposed to ‘record’. Although accomplished through minimal interactive details and visible in ‘small phenomena’ (Sacks, 1984), purifying data from those cues and selectively establishing what part of the interactional context counts or doesn’t counts (‘that’s ok, no problem’) are crucial epistemic practices: they serve the purpose of nourishing the ideology of methodological tools as ‘“transparent windows” into social reality’ (Mondada, 2009: 69).
Making some features of the context irrelevant (i.e. the artifact), does not mean that these features are irrelevant for participants, nor that the talk is context-independent; on the contrary, it is a practical way to construct the talk (i.e. data) as independent or produced as if the artifact wasn’t there. The crucial role this artifact plays in the conduct of participants’ activities is – paradoxically – enlightened by the researcher’s doing ‘ignoring it’.
Building upon Garfinkel’s words, we contend that in the researchers’ workplace ‘there exists a locally produced order of the work’s things; they make up a massive domain of organizational phenomena’ and the researchers’ classical work ‘depends upon the existence of these phenomena, make use of the domain, and ignore it’ (Garfinkel, 1986: vi, quoted in Streeck et al., 2011b: 10).
The epistemic work of things: who does forget it? Concluding remarks
Since the pioneering work of Charles Briggs (1986), researchers interested in the indexical and reflexive dimension of scientific data have begun to pay attention to the conversational interaction taking place between researchers and participants: these epistemologically and methodologically sophisticated studies show how informant’s talk is occasioned by, dependent on and unavoidably intertwined with the researcher’s talk (see among others, Houtkoop-Steenstra, 2000, Kvale and Brinkmann, 2009; Maynard and Schaeffer, 2006; Potter and Hepburn, 2010; Steier, 1991). Studies on social research methods strongly benefited from studies on talk-in-interaction that ultimately proved the local and co-constructed nature of scientific data, i.e. the inherent and unavoidable reflexivity that characterizes research interaction as much as any social interaction (Lynch, 2000).
Interestingly enough, things and artifacts that inhabit the great majority of research scenarios are rarely deemed worth similar analytical attention (but see Mondada, 2009; Hutchby et al., 2012), as if they were irrelevant or inconsequential to the organization of social interaction taking place in the field. This analytical indifference is tricky as a great amount of contemporary research shows how and to what extend things matter in social contexts (e.g. workplaces) and make a traceable difference on the unfolding of interactions taking place in such contexts (see Caronia and Cooren, 2013; Caronia and Mortari, forthcoming; Cooren, 2010, 2012; Streeck et al., 2011a).
In this paper we advanced the hypothesis that the ‘vanishing of technology’ (Button, 1993) from many analytical accounts is consistent with and maintains an enduring positivistic vision of research practices as mirroring (‘recording, capturing, acquiring, gathering’, Mondada, 2009: 69) social reality as it is. This vision originates the methodological adagio that many researchers have heard or even used when confronted with the observer’s paradox: participants ultimately would acclimatize to and, hence, ignore the presence of the devices. As previous studies suggest (Hutchby et al., 2012; Mondada, 2009) this assumption seems to reflect a rhetorically useful claim more than an evidence-based statement. It serves the purpose of cultivating the illusion that the researcher’s tools do not contaminate data that, therefore, can be sustainably considered as natural and uncontrived (Speer, 2002).
Adding to this line of inquiry, in this paper we empirically have shown that a) participants do pay attention to the recording devices beyond the canonical ‘beginning’ of the research interaction references, b) they refer to the recording devices in some peculiar circumstances (i.e. in interactionally problematic environments and when the interactional scene splits and participants have to cope with multiple parallel deictic focuses), and c) they accomplish some epistemic activities by referring to these artifacts: assessing the ‘recordability’ of the just performed or imminent talk; ratifying the recorded behavior as designed for the institutional purpose of the event; and establishing an epistemic difference between what is going on (the recordable) and its textual representation (the record).
By analyzing how the researcher replies to the informants’ references to the recording devices, we have also shown her work of minimizing, as far as possible, the relevance of the recording devices talked into being by participants. In a few words, participants ostensibly craft data for research purposes and they monitor what counts or should count as a document of the object of inquiry, either building on the presence of the recording devices or minimizing it.
All these activities are clearly performed through talk (i.e. they are discursive activities) and – if noticed and transformed into an analytical focus – show the reflexive and indexical nature of research data. Is there any contribution of the artifacts in such a process?
Adopting an analytical framework on hybrid and distributed agency and overcoming any ‘logocentric bias’ (Streeck et al., 2011b: 11) we have contended that the physical presence and affordances of the artifacts are precisely what allows or leads participants to accomplish the activities mentioned above. None of them would be accountable, describable and even understandable without taking (as participants do) the presence of the recording devices into account. Even the researcher’s counter activities (i.e. downgrading the relevance of the artifacts in the unfolding of the event) are oriented to the presence of these objects made relevant by the informant. From this standpoint, we propose considering the recording devices as epistemic agents: they make a difference in the making of the data and contribute to defining the recorded behavior as a document of the object of inquiry. They do this not as extrainteractional inert objects but as nonhuman agents shaping the activity setting, the relevance and force of the linguistic performances occurring within it (Licoppe and Dumoulin, 2010) in ways that are traceable in people’s conduct.
Demonstrating the constitutive (and therefore often embarrassing) role of things in the making of the data is relatively easy when the analyst can track this role in participants’ talk-in-interaction as in the examples analyzed in this paper. In these cases, the analyst can focus on what is ostensibly relevant and procedurally consequent for members (Schegloff, 1987) and she is able to make a case for the agency of the recording devices from the participants’ point of view. When participants refer to these artifacts or use them as a resource to account for their conduct, the analysis respects the ‘unique adequacy requirement of methods’ (Garfinkel, 2002: 175) and the analyst’s interpretation is oriented to the members’ position.
Similar, although less easy to analyze, is the case where participants show an orientation to the recording devices in and through their gestures, gazes and movements (see Hutchby et al., 2012). In these cases, although it is clear enough that participants are aware of the recording devices and that these things matter for them, it is analytically problematic to demonstrate how they matter and in precisely what way they make a difference.
A far more problematic case is when participants do not show any visible orientation to the recording devices in their words, gestures, gaze or movements. Although less frequently than we suppose (see Hutchby, 2001; Speer and Hutchby, 2003), in these cases the interaction progresses smoothly and the artifactual equipment of the research setting seems to go unnoticed by participants: the artifacts are ‘simply there’ with their affordances, size and functions.
Are we legitimated to presume that, in these cases, things are irrelevant for participants or to treat this ‘absence’ of references as evidence that ‘they forget they are being recorded’?
We contend that even when careful and epistemologically sophisticated analysis of the research interaction does not find any trace of reference to the recording devices, this is not eo ipso evidence that they do not do their ‘epistemic work’ and are irrelevant for participants. As Cicourel (1980, 1992) reminds us, the constitutive role of a given context (and, we can add, of things and other affairs) in shaping members’ practices and interactions can be so deeply constitutive of the practices that it does not need to be marked in any specific and ostensible way by the members. We contend that even when the recording devices are not ‘demonstrably relevant for the parties’ (Schegloff, 1997: 183) or procedurally consequent in a traceable way, we cannot infer upon this ‘absence’ that parties forget these objects: we simply do not know (for a discussion on the methodological consequences of taking into account the ‘agency of things’, see Caronia and Mortari, forthcoming). We suggest that, in these cases, researchers should be agnostic as to whether recording devices are irrelevant components of the context or simply are not referred to in a traceable way: the latter does not necessarily indicate the former.
We advance the hypothesis that, even when unmarked, artifacts project certain meanings over others (e.g. the recorded behavior as a public document), indicate possible courses of action (e.g. recording) and encode a given simulacrum of the event (e.g. a research interview). Artifacts convey a particular definition of the situation and even an institutional order through their mere presence and recognizable function. Assuming their irrelevance for participants is as problematic as presuming their relevance. Reversing Schegloff’s classical perspective on the analyst’s misuse of context (Schegloff, 1987) we contend that when assuming that participants sooner or later forget they are being recorded, it is the irrelevance of the material context that is presupposed more than shown by the analyst. In these cases at least, ‘the absence of evidence is not evidence of absence’.
Footnotes
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.