Abstract
Instead of viewing the ‘observer’s paradox’ or the presence of ‘tape-affected speech’ as a methodological problem that spoils ‘natural’ data, in this article, I advocate exploring the opportunities that a recorder offers to research study participants, and the insights tape-affected speech gives to researchers. In my discourse analysis of the week-long self-audio-recorded discourse of members of four American families, I draw on the notions of ‘frame’ and ‘footing’ to uncover the various ways the recorder is conceptualized and oriented to in interaction—as an object and a person, and in literal and playful ways—and demonstrate how participants use the recorder as a resource for identity work in interaction. Analysis reveals how recording shapes the interactions that constitute ‘data’ and how study participants manipulate the presence of the recorder to display identities.
Keywords
Introduction
The methodological notion of the ‘observer’s paradox’ (Labov, 1972) is insufficient for understanding how the presence of a researcher’s audiotape-recorder affects collected conversations. The idea that the presence of a fieldworker or recording equipment paradoxically inhibits researchers from exploring that which they seek to study—‘how people speak when they are not being observed’ (Labov, 1972: 97)—has greatly influenced the methods of data collection and analysis used by scholars interested in studying ‘naturally-occurring’ human interaction. Because, for ethical reasons, research participants must be informed that their communication is being recorded, the paradox is inescapable. 1
Influenced by the observer’s paradox and related concepts (like ‘researcher effects,’ ‘observer effects,’ and ‘tape-affected speech’), scholars across the social sciences have sought to obtain ‘relatively “naturalistic” discourse that is as close as possible to what it would have been like if it was not being taped’ (Johnstone, 2000: 104). Thus, the presence of a recorder is seen as a limitation, as inhibiting the collection of ‘natural’ speech. I suggest an alternative: An audio-recorder can serve as a resource for participants to accomplish identity work, and for researchers, who can glean insights into how participants conceptualize research study participation and the experience of being recorded.
In line with recent scholarship, I suggest that researchers need to consider ‘what it is that participants are doing when they orient to being recorded’ (Speer and Hutchby, 2003: 317, emphasis in the original; see also Monahan and Fisher, 2010). Furthermore, we should rethink how the notion of ‘naturalness’ has both helped and hindered analyses of social communication (Speer, 2002a, 2002b, 2008). In investigating these issues, in this article, I demonstrate how ‘observer’ inadequately describes an audio-recorder’s effect on the everyday conversations of participants in a family discourse study. As part of their participation, four American families audio-recorded their own conversations for 1 week as continually as possible (see Gordon, 2009; Tannen et al., 2007). Listening to, transcribing, and reading the data, I was struck that ‘observer’ only begins to capture how the study participants oriented to the recorder.
To flesh this out, I bring the notions of ‘frame’ and ‘footing’ to bear on moments of conversation in which participants talk into, to, through, or about the audiotape-recorders they used to capture their everyday talk. I demonstrate how they oriented to the recorder variously as an object and as standing in for a person (or persons), and how they did so both in literal and nonliteral (play) ways. I also show why consideration of these excerpts is valuable beyond revealing participants’ frames of understanding. Specifically, participants use the recorder to put forth distinctive identities: Members of one couple treat the recorder as an audience to entertain, while members of a second, in collusion with their coworkers, treat the recorder as part of their (imaginary) involvement in a ‘government spy operation.’ These different framings fit in with these participants’ wider patterns of linguistic identity construction (see Gordon, 2004, 2009). This suggests not only that participants extend identity display to moments when the recorder is in focus but also that they incorporate the recorder into their patterns of everyday talk.
I first summarize research that has investigated talk in which the recorder’s presence is highlighted for participants and provide theoretical grounding for the study. Next, I detail the design of the study providing the data for this analysis. I then move to analysis, identifying four basic understandings of the audiotape-recorder that surfaced in the data—the recorder as an object versus a stand-in for a person, and in literal versus nonliteral (play) ways—and demonstrating how participants used the recorder to portray distinctive identities. I end with a discussion of why it is important for scholars of interaction to examine data such as these.
Background
The observer’s paradox and research methods
As Erickson (2004) indicates, the work of conversation and discourse analysts depends on recording technology: ‘Because the behavioral phenomena of the real-time conduct of talk and listening [is] so complex and fleeting, it [is] necessary to capture them for purposes of analysis by means of machine recording’ (Erickson, 2004: 196). Other areas within the social sciences (e.g. ethnography) do not consider recording obligatory, but commonly collect and work with recordings. However, machine recording fundamentally affects—alters, and maybe even ‘damages’—interaction, an influential idea introduced in early sociolinguistic research by Labov (1972) that continues to shape how scholars in various fields collect and analyse conversational data.
Labov (1972) introduced the concept of the observer’s paradox in suggesting that in order to most productively develop linguistic theory, scholars need to capture how people use language in ‘natural’ circumstances; he thus aimed to minimize possible effects of fieldworkers and recording devices on research participants’ discourse. For example, he found that eliciting highly involved stories seemed to cause speakers to ‘forget’ that they were talking to a linguist and in an audio-recorder’s presence (Labov, 1972). Thus, he was able to collect samples of the ‘vernacular’ language he sought. 2
The idea that a recorder might adversely affect the data’s quality and usefulness remains current and widespread. As Speer (2002a: 511) remarks, it is common for social scientists to attempt to ‘eliminate extraneous, research-induced “contaminants”.’ One such ‘contaminant’ is the presence of an ‘observer,’ be it a person, recording device, or both. Researchers, thus, aim to minimize the observer’s paradox, reduce ‘observer effects,’ and avoid ‘tape-affected speech.’ They might discard the first minutes of recording, or choose to focus analyses on conversational moments in which the recorder is not in focus. While it is widely accepted that the observer’s paradox cannot be ‘beat,’ the aforementioned measures attempt to gather the data that are as naturalistic as possible.
Participant orientation to recording devices
Despite a general emphasis on data naturalness, a small body of research, across a range of research traditions, has brought into focus moments in which recording is foregrounded in talk. Studies especially relevant to the analysis that follows are reviewed briefly here.
In a study whose context is quantitative analysis of sociolinguistic interview data, Schilling-Estes (1998) finds that the audiotape-recorder serves as an elicitor of ‘performance speech’ produced by participants to demonstrate a particular language variety. Although such ‘performances’ would normally be excluded from studies seeking to uncover sociolinguistic variable distribution in ‘natural’ speech, she suggests that they may be useful for sociolinguists; for instance, moments of performance can reveal ‘which aspects of linguistic production are most salient’ for speakers of a given dialect (Schilling-Estes, 1998: 64).
Wilson (1987) focuses on what he calls ‘tape-affected speech.’ Examining conversation among adolescents, he found participants making explicit reference to the device, treating it as a ratified participant, monitoring one another’s language use based on its presence, and treating the researcher as an audience. Wilson observes that the recording equipment’s effects fluctuate over the recording period for various reasons and suggests that tape-affected speech is worthy of further exploration because it reveals participants constructing context for their utterances (see also Wilson, 1994). He also identifies ‘participation framework’—Goffman’s (1981) identification of the diversity of audience roles available, including overhearer, ratified participant, and addressed recipient—as a useful way of approaching tape-affected speech.
In linguistic anthropology, Wertheim (2006) explores how study participants orient to being observed and recorded, finding inspiration in Goffman’s (1981) study and in Bell’s (1984) notion of audience design. Wertheim undertook ethnographic fieldwork in Tartarstan (an autonomous Republic in the Russian Federation). While primarily examining copresent fieldworker effects, Wertheim also discusses the audiotape-recorder: For her study participants, members of the intelligentsia of Tartarstan’s capital city, recording was evocative of KGB surveillance. Wertheim also found that when periodically self-audiotape-recording, participants communicated through the recorder to an end-listener who is conceptualized in various ways. Overall, she demonstrates how reconceptualizing the observer’s paradox within audience design reveals how different imaginings of audience correspond to speakers’ style shifting. This echoes Wilson’s (1994: 289) suggestion that attention to tape-affected speech reveals ‘how certain groups sociolinguistically adjust their speech in relation to perceived audience types.’
In summary, these studies point to the productivity of reexamining and reevaluating the observer’s paradox—for linguists, communication scholars, linguistic anthropologists, and, by extension, other social scientists. In a similar spirit, Monahan and Fisher (2010: 363) advise ethnographers to reconsider ‘observer effects’ because the performances ‘staged’ by participants for fieldworkers constitute ‘rich symbolic texts that lend themselves to multiple interpretations and provide critical insights into the cultures being studied.’ More broadly, Speer and Hutchby (2003) propose moving beyond a focus on how ‘natural’ the data are, suggesting social scientists should consider how participants ‘orient to being recorded’ and ask, how this might ‘play a part in the ongoing construction of specific situated interactions?’ (p. 317).
‘Frame’ and ‘Footing’
In reexamining the observer’s paradox, I draw on the concept of ‘frame’ as used in discourse analytic research stemming from its origins in the works of Bateson (1972) and Goffman (1974, 1981). Following Tannen and Wallat (1993), I conceptualize ‘frame’ as both a cognitive understanding and an interactional phenomenon, as well as a notion appropriate for capturing the fluidity that characterizes social interaction, in particular, in conjunction with the concept of ‘footing.’
In clarifying and demonstrating applicability of various understandings of ‘frame’ for discourse analysts, Tannen and Wallat (1993) examine an interaction involving a pediatrician, a child with cerebral palsy, and the child’s mother. They differentiate between the notion of ‘knowledge schemas,’ which are cognitive and refer to ‘participants’ expectations about people, objects, events and settings in the world’ (p. 60) and the concept of ‘interactive frame,’ which is the ‘definition of what is going on in interaction’ (p. 59). They demonstrate how the mother’s and pediatrician’s unshared knowledge schemas lead the pediatrician to continually need to shift the interactive frame, that is, to (re)define the situation, and to swiftly change between multiple footings (Goffman, 1981), or alignments, vis-à-vis her co-interlocutors. For example, when the mother misinterprets the child’s ‘noisy’ breathing—which is normal for a cerebral palsied child—as gasping for air, the doctor switches from examining the child (examination frame) to addressing the mother’s concerns (consultation frame). In addition, as she shifts between interacting with the mother and child, as well as with the medical residents who will later view the videotape of the encounter for educational purposes, the pediatrician manipulates language, creating both playful and nonplayful frames. For example, she uses a clipped style and technical vocabulary when reporting her examination techniques and findings to the residents; she tends to use high-pitched speech, teasing, and simple vocabulary (like ‘tummy’) when talking to the child.
Tannen and Wallat’s (1993) dual understanding of frame as social/situational (interactive frame) and as cognitive (knowledge schema) is important for my analysis, as is Goffman’s (1981) related concept of footing: I suggest that the schema a participant uses in interpreting the recorder—in contextualizing it in interaction and using it to contextualize the talk produced—is interconnected to the interactive frames and footings created in the discourse. In other words, the participants use schemas to interpret the recorder at any given conversational moment, and in doing this, they take up particular footings vis-à-vis the recorder within particular frames, which contributes to identity display.
The foundation of the argument I make regarding identity lies in a broad swath of sociocultural/interactional research suggesting that people create and display identities in conversation via the construction of alignments (alternatively conceptualized as footings, stances, or positions). Bucholtz and Hall (2005) provide an integrated and useful overview of this research while also highlighting the role of indexicality in identity construction (e.g. as discussed by Ochs, 1993). They describe how speakers use language to mark their ‘orientation to ongoing talk’—stances (or footings) that in turn create identities (Bucholtz and Hall, 2005: 595). As Ochs (1993: 301) explains, a woman’s use of a simplified register when speaking to her infant does not directly map onto a ‘competent mother’ identity. Instead, identity work occurs via the stances or footings created (mother as accommodating to child); these index different social meanings in different sociocultural contexts and thereby construct identity. This idea underlies my analysis suggesting that footings taken up vis-à-vis the audio-recorder (and the imagined listeners to the recordings) create identities.
Data and methods
My data are drawn from a study for which members of four dual-income American couples with at least one child recorded their own conversations over the course of 1 week. The study was designed by Deborah Tannen and Shari Kendall to investigate how women and men talk at work and at home, and how they use language to balance work and family life (see Tannen et al., 2007). It was conducted with informed consent. The four participating families were all White, middle class, and lived in the Washington, DC, metropolitan area. 3 Data were collected in the year 2000. Each mother and father carried a recorder, which was relatively small (handheld). Participants often carried the recorder in a pocket or set it down on a table. They used lavaliere microphones at their discretion. The recorders used digital audiotapes that recorded for 4 continuous hours, so participants changed tapes several times per day.
The relatively long-term recording by these families, with no researcher present, can be viewed as an attempt not only to capture family interactions across a range of contexts but also to reduce observer effects on the data, as habituation may lead people to ‘get used to the machine after a while and sometimes even forget it is there’ (Johnstone, 2000: 106). However (and perhaps a bit paradoxically), this methodology also brought into relief the various ways in which participants oriented to their recorders.
The families collectively recorded approximately 460 hours, which resulted in over 1 million words of transcription. Although all families were instructed to record as continuously as possible from morning to night, for various reasons (relating to practicality, motivation, and ability to record at work), they recorded different amounts of time: Janet 4 and Steve (parents of daughter Natalie, almost 3 years old) recorded for 165 hours; Kathy and Sam (parents of daughter Kira, just over 2 years old) recorded for 134 hours (over 14 days rather than the 7 required); Clara and Neil (parents of son Jason, almost 5 years old) recorded for 89 hours; and Nora and Greg (parents of 3 children: John (age 22), Susan (20), and Jeremy (17)) recorded for 75 hours. As a member of the research team, I listened to the complete recordings of two of the participants (Janet and Clara), and to a large portion of the others. I transcribed many hours of the data. I also read through all the transcripts, manually searching for conversational moments in which the audio-recorder was in focus.
Although all four families did seem to become acclimated to the recorders (‘embarrassing’ events like family arguments were captured on tape), they did intermittently pay attention to them, a finding that echoes previous studies. Identifying moments of so-called ‘tape-affected speech,’ however, was not a straightforward task, and was necessarily interpretive (as Wilson (1987) also notes). While in many cases it was obvious that participants were orienting to the recorder, such as when they spoke directly into it or addressed a researcher by name, in other cases, it was less clear how or whether the recorder was affecting their communication. I examined both the content of what was said and its form in identifying participants’ talk that seemed oriented to the recorder. I relied on participants’ orientation to the others’ talk as evidenced in the transcripts, my own understanding of the conversations, other transcriptionists’ notes, and paralinguistic information that ‘sounded like’ it was aimed at the recorder or at the researchers who would later listen to the tape. Through this method, I believe I captured most tape-focused moments of conversation in the data. I also presented my findings to two different interdisciplinary research groups as a means of checking my interpretations (see ‘Acknowledgements’).
I should note that some ‘moments’ in which the recorder is in focus are short; for instance, a speaker might announce the date into it (as researchers requested at each tape’s beginning), but then set off on another task, seeming to pay no more attention to the recorder for the next several hours. On other occasions, multiminute conversations among multiple participants revolved around the recorder; I counted this as one extended ‘conversational moment.’ While I do not believe numbers are very meaningful, given the different lengths of recording per family and the difficulty in exhaustively identifying instances of speech affected in some way by the recorder’s presence, it seems worth noting that I found 82 moments during which Janet and Steve verbally oriented to the recorder, 82 for Kathy and Sam, 55 for Clara and Neil, and 23 for Greg and Nora (this count includes any conversational participant—not necessarily the parents—paying attention to the recorder; for example, Clara and Neil’s coworkers took an especially large interest in the recorders, which generated quite a bit of tape-focused speech at work.) In other words, although how many times each participant oriented to the recorder varied, all participants intermittently did so. This reinforces findings of previous research that participants’ attention to the recorder fluctuates over time; it also provides many instances for analysis.
Analysis
I sorted all instances of tape-focused speech into data-driven categories. These categories sometimes intersect; nonetheless, they capture all identified examples. The categories were driven by two overarching observations I made in considering the data from the perspective of framing. First, the research study participants treated the recorder as standing in for a person (or persons) or as a nonhuman object; in other words, they oriented to the recorder using either the schema—or cognitive frame—of recorder as ‘stand-in person’ or ‘object/machine.’ Second, the participants treated the recorder in both nonliteral and literal ways; in other words, they oriented to it in different interactive frames (defined as ‘this is play’ (Bateson, 1972) or ‘this is not play’). There were also nuances that emerged within these categories where they intersect, which I identify as footings (shown in the shaded area of Table 1).
Frames, schemas, and footings for the audio-recorder.
In the analysis that follows, I first briefly address the literal interactive frames: The recorder is treated in a nonplayful way, sometimes conceptualized as an object and sometimes as a person (or people), but always working to reveal underlying knowledge schemas and to project certain general kinds of identities. Then, I turn to the nonliteral (play) frames, which members of two of the couples studied and their conversational partners exploited to construct more nuanced identities primarily through constructing particular footings for the imagined end-listeners.
The audio-recorder in literal frames
In literal frames of interaction, three primary footings were taken up vis-à-vis the recorder. When approached using the cognitive frame (schema) of ‘object,’ the recorder is interactively treated as a burden or as a data collector; when it is conceptualized as standing in for a person, it is treated as a conduit to the researcher(s). These footings collectively illustrate the recording’s impact on the families (it is somewhat taxing), help create their identities as compliant study participants, and ultimately show the participants to be concerned with self-presentation as sociable, good people.
Recorder as burden
‘Recorder as burden’ was a common footing created when the recorder was approached from the schema of ‘recorder as machine or object.’ When oriented to in literal frames, the object interferes with everyday life. The burden footing emerges, for example, when participants complain about the recorder, struggle to carry it, and try to balance recording with other activities (like child care). For example, one mother, Kathy, who was taking a graduate course at the time of recording, talked about the recorder with her classmates, as shown in extract (1). The students have just broken into groups for discussion, and Kathy has explained that she is carrying an audio-recorder because she is participating in a study of family conversations. A classmate then asks about the need to replace the tapes 5 :
(1) Classmate: You have to keep on changing the tape in there?
Kathy: <sighs>
Every four hours.
The batteries are dying every four hours too,
→ so … it’s really been … it’s really been a chore.
Through her labeling of changing the digital audiotapes (and the batteries) as ‘a chore’ and sighing, Kathy takes up a footing toward the recorder as a burden. She orients to it in a literal, nonplayful way as she complains (this is not picked up; Kathy’s group proceeds to course-related discussion). (Note that another participant in the study, Janet, also complained about short battery life; replacing rechargeable batteries with regular batteries seemed to resolve this issue.) In a similar example, Kathy complains to her husband Sam that it is difficult to carry the recorder and notes that earlier in the week she became frustrated and felt like quitting; Sam replied, ‘Yeah, it’s hard.’ (Then Kathy actually backed off from her initial assessment, saying ‘You get used to it.’) Nevertheless, it is clear that managing the recorder, for these two working parents of a young child, constituted at least a minor burden. In fact, elsewhere in this family’s recordings, we hear Sam struggle to label a tape he has just finished recording (something he was requested to do by the researchers) while his daughter (Kira, age 2 years 3 months) vies for his attention.
Members of all four families take up alignments toward the recorder as a burdensome object. This is perhaps not surprising, given that the couples agreed to participate in a study exploring the balancing of work and family life—the recorder was one more thing to manage in their already busy lives. However, the finding also suggests that the recorder affects the participants in consequential ways beyond speech (like causing stress or affecting the unfolding of certain daily activities).
Recorder as data collector
Study participants take up a second footing when they demonstrate an understanding of the recorder as a data collector and orient to it in literal interactive frames. In treating it as an object required for the research study, they reveal themselves to be cooperative participants. For example, participants do this when they state certain information at the beginning of each tape (at the researchers’ request); Nora, the mother of three grown children, does this in extract (2).
(2) Nora: Nora Marsh,
uh … six twe:nty, tape one.
Here, Nora treats the recorder as a machine used it to capture information requested by the researchers (her name, the date, and the tape number). She then moves on to other tasks.
The recorder-as-data-collector footing is also created when participants explain the recorder’s presence to their co-interlocutors, and when they list activities about their day for the benefit of the research project. In all such examples they orient to the recorder as something that collects data relevant for the research study; these orientations help create literal interactive frames.
An example of this occurs when Natalie (age 2 years 11 months) asks her father, Steve, about his recorder when he comes downstairs one morning to join her and her mother, Janet, for breakfast.
(3) Natalie: (Where’s) your tape-recorder.
(Why) you bring your tape recorder down there.
Steve: Because Mommy and Daddy are doing a study.
We’re taking part in a study.
Natalie: Me too?
Steve: Yeah.
Janet: [<laughs>]
Steve: [Well yes you too.]
Janet: Come babe.((to Natalie))
Natalie: Me too Mommy.
Janet: Yes you too.
Come sit down and get your breakfast.
Here, young Natalie brings up the audiotape-recorder, and Steve treats it literally, as a required component of a study, a data collector. (Note also that it builds family togetherness: Everyone is participating in the study.) Across the families, examples similar to extract (3) appear as adult family members explain the recorder’s presence to various people. This footing, while not terribly interesting on the surface, reveals participants orienting to the recorder in ways they were instructed to: They were to make its presence known to others (for ethical reasons), and they were asked to state some information at the start of every tape for the researchers’ benefit. (During most of the recording period, however, participants seem to ignore the recorders—something the researchers also were seeking.) These behaviors can be viewed as participants creating ‘cooperative study-participant’ identities (which contrasts somewhat to some of the identities created in the nonliteral frames, as will be shown).
Recorder as stand-in for researcher
Beyond conceptualizing the recorder as an object, participants understand it as a stand-in for a human being; in literal frames this person is a researcher. For example, the participants sometimes speak through the recorder to a specific researcher. Consider how Kathy begins a new tape in extract (4); here she addresses her utterance to ‘Shari’ (referring to Shari Kendall, project coprincipal investigator):
(4) Kathy: Good morning Shari,
It is about ten after nine on Friday morning,
and we thought we were recording for the last hour and a half,
but I don’t think we were because,
Sam: (?) for an hour and a half.
Kathy: I didn’t turn the microphone on.
This extract is similar to extract (2) in that the speaker is stating information into the recorder; it is different, however, in that here Kathy speaks to the recorder not (only) as an object (a machine) but as if to a person: She greets a specific researcher, by name, with the salutation ‘good morning.’ In this way, Kathy takes up a sociable footing toward the person who she (correctly) believes will later listen to the tape. While the study participants do not always address researchers by name, this kind of sociable orientation occurs across all four families.
In addition to acknowledging that the recorder stands in for a sociable human being, and is not just a mechanical object, participants take up this footing in acknowledging concern that those who listen to the tapes may judge their behaviors. This not only fits in with prior research suggesting that participants may monitor their speech when they know they are being recorded (e.g. Labov, 1972) but also extends to behaviors beyond speech style, such as family arguments (see Taylor, 1995).
Participants sometimes retrospectively (on tape) expressed embarrassment when they realized certain things (such as bodily noises or arguments) were captured by the recorders. They expressed dismay when an event not intended to be recorded inadvertently was (these accidental recordings were erased). Parents also worried that their children’s behaviors (such as tantrums) or their own parenting behaviors would be negatively evaluated. Extract (5) is an example of the latter. Here, Kathy has been complaining to her coworkers about her husband Sam’s insistence on recording everything: One night when Kira woke up crying at 3:30 in the morning, to calm her, Kathy watched a video with her (‘Elmo,’ a character from the children’s television show Sesame Street). Sam told her, ‘This is an example of something that you should be taping’:
(5) Kathy: I just looked at him. ((Sam))
And I said,
‘we are watching Elmo. ((a children’s video))
And we are not talking.’
Coworkers: <laugh>
Kathy: ‘And I don’t think—
I think she’s heard enough of Elmo.’
Coworkers: <laugh>
Kathy: There’s a lot of Elmo and stuff on the tape.
→ Probably gonna say,
→ ‘these are such bad parents,
→ look how often they turn the TV on.’
In this extract, Kathy complains about Sam, but also references the listener of the recordings using ‘she,’ likely referring to coprincipal investigator Deborah Tannen, who was mentioned moments earlier in the conversation as being in charge of the study. Kathy shows concern that her parenting behavior might be judged negatively; in using ‘she,’ Kathy treats the recorder as representing a person who will later not only hear Kira repeatedly watching television but also Kathy’s attempt here to save face. (The topic of parenting continues in this conversation as a coworker ‘confesses’ to allowing her child to watch Star Wars.) Extract (5) is but one of numerous instances of saving face that occur in the recordings of all four families. These acknowledge the humanity of the researcher—the listener of the tapes, after all, is not a machine. These instances also do identity work, and especially save face, for the self and other family members.
Summary: literal frames
In sum, when participants orient to the recorder in literal frames, two different knowledge schemas are at play: the recorder is an object or a stand-in for a person (or people). Participants take up three different footings in interaction toward it: It is a burdensome presence in everyday life, it is a collector of data for a family discourse study, and it is a stand-in for the researcher, someone who will have human reactions (feelings of solidarity and judgmental responses) to the recordings. I suggest that these footings collectively accomplish important, if subtle, identity work for members of participating families: They reveal themselves to be busy yet also cooperative, to be friendly yet to take recording seriously, and to be concerned with the impressions they are creating.
The audio-recorder in nonliteral (play) frames
The recorder is also oriented to in nonliteral frames, where identity work is even more prominent. While members of all families treated the recorder playfully to some degree, due to space limitations, I focus on the discourse collected by the two couples who did it the most frequently, consistently, and strikingly: Janet and Steve orient to the recorder as standing in for an audience to amuse and entertain, and use it to construct for themselves ‘performer’ identities in play frames. Clara and Neil, along with many of the coworkers at their respective workplaces, playfully orient to the recorder as a particular kind of object (a ‘wire’ for eavesdropping), and to the end-listeners as members of a government spy organization. In this way, they show themselves to be both humorous and oriented toward political life. In what follows, I demonstrate how these couples—as well as their collaborating co-interlocutors—create identities in interaction using the recorder as a resource. I first consider the discourse of Janet and Steve.
Janet and Steve: recorder as audience
Janet and Steve frequently treated the recorder nonliterally. In orienting to it within frames of play, they took up footings toward the imagined end-listeners as members of an audience who would be amused and entertained by their behaviors. These are examples of ‘staged performances’ that are ‘deeply revealing of how individuals perceive themselves and would like to be perceived’ (Monahan and Fisher, 2010: 363). It should be noted that Janet and Steve were both members of a children’s theater troupe at the time of recording and that they generally use language and paralanguage in playful, animated ways, for instance, taking on play voices and stylized accents (see Gordon, 2009). While it is perhaps not surprising that two actors would orient to the recorder as if it were an audience, it is interesting how Janet and Steve incorporate the recorder into their everyday lives and use it to do identity work.
Extract (6) shows Steve playfully reporting information into the recorder. He is not reporting information that has been requested by the researchers; he is joking around. Here, Janet, Steve, and Natalie are at home, and a song by the band Talking Heads is playing on the radio. The song includes the line ‘I will be bitter.’
(6) Janet: What’s the matter with (the DJ) today?
This is the Talking Head song he’s gonna play?
Natalie: Is this Talking Head song?
Steve: Yes!
Janet: Yes!
The worst one!
Steve: <burps>
Excuse me.
Janet: I’m bitter … just for the record. ((re: song))
→ Steve: <directly into recorder> In case you did not get that,
→ the wife is bitter.>
Janet: <laughs>
Natalie What happened.
((Janet sings along with the music))
In this extract, Janet playfully echoes a word she just heard in a song on the radio to comment on that song (she does not like it; she is ‘bitter’). Steve ‘reports’ this information directly into the audio-recorder, not because it is an important piece of information for the researchers, but seemingly in an attempt to be entertaining (to both those who will listen to the recordings, and to his wife, who laughs). Steve is playing for an audience. In fact, he succeeded, here and elsewhere, in amusing members of the research team who listened to and transcribed his tapes.
In the next extract, Janet and Steve are brushing their teeth in the bathroom just before bedtime (this is indicative of their commitment to recording).
(7) → ((Steve brushes his teeth directly into recorder))
Janet: <laughs>
All right.
Ha ha!
→ ((Steve brushes teeth directly into recorder))
<laughs>
((short pause, sound of running water))
<stage voice, high-pitched> Don’t be ridiculous, dear!>
→ Steve: I just thought I’d give them a little (brush,)
→ at the very end.
→ Janet: Oh boy.
((short pause, sound of running water))
→ Steve: Well transcriber,
→ it’s been fun, eh?
Janet: <chuckles>
Steve: I- I’ve said a few little things to the transcriber over the past week.
Janet: <laughs>
<laughing> Oh God!>
Steve first brushes his teeth directly into the recorder in an attempt to entertain the researchers as if they, along with Janet, are his audience. When Steve brushes his teeth near the recorder’s microphone a second time, Janet playfully chastises him, using a high-pitched and clearly ‘dramatic’ tone of voice, likely not only for Steve’s benefit but also for the listeners. Following this, Steve reports that he has made comments to the transcriber over the course of the week (which he did); while some of these were in a literal frame (e.g. describing workplace activities that took place while recorder was stopped), many were amusing comments such as ‘it’s been fun, eh?’
These are but two examples of many in which these participants alter tone of voice and perform amusing dialogues as they speak directly into the recorder or through it to their audience. In an especially common pattern, such as in excerpt (7), Steve ‘acts out’ in some way and Janet playfully rebukes him. Through routines like this, the participants take up nonliteral orientations to the recorder and indicate ‘this is play’. They thus create (and enact) ‘performer’ identities, while also contributing to the construction of themselves as part of a happy family.
Interestingly, this identity is not only created in moments where the recorder is acknowledged. In previous research (Gordon, 2009), I demonstrate how across moments that seem not to be affected by the recorder, these participants manipulate speech styles and prosody and use lexical and syntactic repetition to enact playful routines with one another. I thus suggest that Janet and Steve integrate the recorder into their everyday family culture by conceptualizing it in a way that fits in with what they regularly do: Perform. They thereby use it as a resource to construct and reinforce shared theatrical identities and experiences; they are performers not only for one another and for other copresent individuals (like friends) but also for a nonpresent audience.
Clara and Neil: recorder as ‘wire’; listener as government spy organization member
Clara and Neil, like Janet and Steve, often acknowledge the recorder in nonliteral frames, but they do so differently. First, they playfully oriented toward it as both an object and as a person, and they do this in a way that reinforces their shared identities as people interested in government and politics. Second, their coworkers contribute significantly to conversations that orient to the recorder in play frames.
Clara and Neil, among the four couples, were the most politically active. Clara worked for the US government; her direct boss was a Presidential (Clinton) appointee. Neil worked for a very politically oriented nonprofit organization. In addition, this couple recorded during the week of the 2000 US presidential election, which no doubt highlighted their shared political identity (see Gordon, 2004). In this section, I demonstrate how the recorder helps create light-hearted politically oriented identities for these participants.
Words used by Clara and Neil, as well as by their co-interlocutors (mostly coworkers), when talking about the recorder and the recording process, are illustrative. They repeatedly mention Linda Tripp (who secretly recorded Monica Lewinsky, the intern with whom Bill Clinton had an illicit relationship; the existence of these tapes was made public about 2 years prior to the taping period and seems to provide one context for Clara and Neil’s tape-focused talk). Clara and Neil use the words ‘spy’ and ‘surreptitious,’ and they joke about wearing a ‘bug.’ Both members of the couple mention looking or feeling like they are in the Central Intelligence Agency (CIA). In one conversation, Clara’s coworker brings up an earlier recording-related political scandal in likening Clara to ‘Rose Mary,’ former President Nixon’s secretary; the coworker suggests that Clara can erase data, ‘if anybody comes up and offers [her] a large bribe.’ (This refers to the infamous 18.5 minute gap in the Watergate tapes, which Nixon’s secretary Rose Mary Woods claimed to have inadvertently erased, although many people do not believe this.) In addition, when playfully orienting to the recorder as an object, Clara and Neil frequently refer to it as ‘a wire’ and talk about ‘being wired,’ which implies that they are recording secretly (although they literally are not). In addition, both participants (and their coworkers) joke that the recordings will get back to someone in the government. For instance, in the following interaction between Neil and his coworker Amanda. Amanda jokes that what Neil is recording will be played for (then-First Lady) Hillary Clinton:
(8) → Amanda: <laughs> (??) get back to Hillary and—
Neil: No. No.
It’s just going over to Georgetown.
→ Which is where Bill Clinton →
went to college you know.
→ Amanda: But my voice is disguised right?
<laughs>
→ Neil: <directly into recorder> Amanda Papas.>
Amanda: No, this is Jen Brown talking.
<laughs>
Neil: No one’s gonna know who it is.
Neil’s coworker Amanda introduces an understanding of the recorder as standing in for a person—Hillary Clinton (she is interrupted by Neil but may have been also going to add ‘Bill’). After denying this and saying that the data will be sent to Georgetown University, Neil takes up the joke as well, hinting that perhaps the Clintons will get the data because Bill Clinton attended Georgetown. When Amanda playfully verifies that her identity will be protected, Neil announces her real name (here, a pseudonym, of course) directly into the recorder. Amanda then claims to be another coworker and laughs. Then, the conversation shifts to a more serious frame, with Neil assuring Amanda that her identity will remain confidential.
Through examples of this type, Neil and Clara, aided by their coworkers in some conversations, playfully conceptualize the recorder as standing in for someone in government. (However, note that this also could simultaneously be viewed as performing for the researchers; they rarely do this elsewhere in the data.) Clara and Neil thus take up footings (and are assigned footings) of spies who are ‘wired’ and are covertly recording. Here, approaching the recorder and the end-listeners within the ‘spy’ schema occurs within play frames and serves to highlight interlocutors’ political (and political scandal) vocabulary and knowledge. Like Janet and Steve, members of this couple use the recorders to construct a shared aspect of identity: Neil, Clara, and their co-interlocutors are politically oriented people, who also have a sense of humor about recording and the follies of the government.
Summary: nonliteral frames
In summary, participants differently oriented to the recorder while using it as a resource in nonliteral interactive frames. Most important, the analysis showed how participants conceptualized and thereby interacted with the recorder (and its end-listeners), thereby accomplishing identity work. This suggests that tape-focused speech is productive (not only a burden), and that the ‘observer’ can be imagined to be an end-listener convenient for the speaker’s knowledge, skills, and identity projects. In addition, it identifies tape-focused speech as a rich context for identity work; the recorders were used to create identities that parallel those displayed by family members elsewhere during their respective recording periods. 6
Discussion and conclusion
Analysis of conversational moments in which the presence of a recording device is highlighted demonstrates that the recorder is more than an intrusive ‘observer’ that might lead to careful monitoring of one’s speech. This study offers a framework for understanding how different cognitive frames, or schemas, for the recorder intersect with interactive frames in which family members take up footings. It thus explores how the recorder is used as a resource in interaction. While much previous research has focused on limitations and adverse effects of audiotape-recorders, my research highlights the identity-work opportunities that they provide.
Analysis of footings taken up in literal frames demonstrates how participants use the recorder, in subtle ways, to portray themselves as cooperative (yet burdened) research participants who are also friendly, good people. In acknowledging end-listeners as researchers who have human reactions to their talk, study participants are sociable and also work to save face, both key aspects of identity work.
Nonliteral frames provide an even richer site for interaction and identity construction: Participants incorporated the recorder into their everyday lives, treating it as standing in for an imaginary audience to entertain or as part of a spy operation. Thus, one couple performed (para)linguistically playful routines for the recorder, while members of another (often in conversation with their coworkers) used covert-recording vocabulary and highlighted their political knowledge and humor about government scandals. Importantly, these correspond to identities revealed elsewhere in these participants’ discourse, further suggesting that talk that foregrounds the recorder can be of analytical importance, and is worthy of further study.
Much remains to be investigated regarding the ways in which recording devices affect interaction, including how, for instance, participants’ talk to the recorder differs from participants talk about it. My research, in focusing on frames, necessarily obscures such other interesting issues. However, the findings have meaningful implications for our understanding of the observer’s paradox, for how to investigate tape-focused speech, and for what constitutes ‘good’ data in the context of social scientific paradigms that favor so-called ‘natural’ discourse.
First, the observer’s paradox shapes our methods, often invisibly; researchers have greatly favored conversational moments when the recorder is ‘forgotten.’ This study demonstrates that moments when the recorder’s presence is highlighted are not ‘contaminated,’ but multilayered: At these times, participants reveal their schemas about the recorder, and use it to create both playful and nonplayful interactional alignments, as well as identities. Thus, like Speer and Hutchby (2003) and Monahan and Fisher (2010), I advocate studying these conversational moments, rather than discarding them as evidence of methodological weakness.
Second, this analysis offers a new framework for investigating what has been called tape-affected speech. It demonstrates how the concepts of frame and footing—concepts that together constitute a theory of meaning-making and identity construction that has both cognitive and social components (see Gordon, 2009; Tannen and Wallat, 1993)—lend insight into how participants orient to audio-recorders and to what effect. Identifying knowledge schemas through which participants interpret the recorder and the kinds of social situations they construct around it give us a rich understanding of the complex ways participants use a recorder in interaction and connect it to their knowledge (e.g. of recent political events) and experience (e.g. with acting). Furthermore, analysis of footings revealed that while all four couples oriented to the recorder using the same basic parameters (object-machine/person, literal/nonliteral), they also did so differently. This shows how different conceptualizations of a methodological research tool—the audiotape-recorder—help create the conversational moments that revolve around it, and afford participants an interactional resource to create unique identities.
Third, my findings intersect with ongoing social scientific discussions regarding what constitutes ‘good’ data. In a series of papers, Speer (2002a, 2002b, 2008; Speer and Hutchby, 2003) strongly argues that there is no clear-cut distinction between ‘natural’ and ‘contrived’ data. More useful than treating the presence of a researcher and/or a recorder as contaminating the data, or of criticizing interview and focus-group data as ‘unnatural,’ Speer argues, is paying greater attention to the interconnections between methodology, context, and data. Thus, the potential usefulness of various kinds of data needs to be considered, keeping in mind the various ways that it has been ‘affected’ by the researcher/observer/recorder.
This is perhaps especially relevant in the context of present-day entertainment media and capture-and-share technology. When the data for this study were collected, a recent US political scandal involving recording provided a point of reference, at least for some participants (hence mentions of Linda Tripp, Bill Clinton, etc.). However, since that time, ‘reality TV’—including shows built on the premise of constant surveillance—has come to dominate television programming in the United States and elsewhere. 7 And, YouTube, Twitter, and Facebook are now omnipresent for many people. How participants orient to recording technology and the idea of being observed thus merits further attention from scholars of social interaction. As Scollon and Scollon (2004) point out, the presence of a researcher or recording technology always affects communities under study; we have to take that into account, within participants’ cultural contexts, instead of trying to escape the inescapable observer’s paradox.
Much remains to be investigated regarding the ways in which recording devices (both audio- and video-recorders) affect interaction, and how participants might draw on them as conversational resources. Doing this means moving beyond seeing the observer’s paradox as a methodological limitation to investigating the opportunities it might offer researchers and study participants alike.
Footnotes
Acknowledgements
I am grateful to Shari Kendall and Deborah Tannen, who codirected the larger study from which these data were drawn, as well as to the generous families who participated. The article has benefited from comments and suggestions given by members of Syracuse University’s Interdisciplinary Research Group (IRG) (College of Visual and Performing Arts) and by researchers at Emory University’s Center for Myth and Ritual in American Life. It has also benefited from useful feedback from Najma Al Zidjaly and Deborah Tannen.
Funding
The Alfred P Sloan Foundation funded the larger project from which the data for this analysis were drawn.