Complicating Voice in a Refusal to “Let Participants Speak for Themselves”

Refusing the Romance of Voice

Our article is a response to the editors’ call to challenge simplistic and mechanistic approaches to qualitative research that preclude dense and multilayered treatment of data. The editors assert that such practices can lead to (over)simplified knowledge claims, something especially risky when participant “voice” is presented as an expression of “experience” devoid of context. In this article, we specifically challenge simplistic treatments of voice in qualitative research that beckon voices to “speak for themselves” or that reduce complicated and conflicting voices to analytical “chunks” that can be interpreted free of context and circumstance.

It is our view that such resistance of methodological simplicity is a practice that qualitative researchers should be concerned with. We are not referring to a use of theory and/or jargon for the purpose of obfuscation and erudition. We are advocating such as a move to create a way of thinking methodologically and philosophically that gets us out of the trap of fixing meaning and instead opens up previously unthought questions. In other words, we are advocating that qualitative researchers think with theory ¹ as a guard against being seduced by the desire to create a coherent and interesting narrative that does little to challenge hegemonic discourses and (over)simplified knowledge claims. We use theory in our effort to refuse the romance of voice as we attempt to open up, rather than foreclose, meaning.

Conventional, interpretive, and critical approaches to qualitative inquiry frequently privilege voice because it has been assumed that voice can speak the truth of consciousness and experience. In these paradigms, voice lingers close to the true and the real, and because of this proximity, has become seen almost as a mirror of the soul, the essence of the self. Qualitative researchers have been trained to privilege this voice, to “free” the authentic voice from whatever restrains it from coming into being, from relating the truth about the self. This drive to make voices heard and understood, bringing meaning and self to consciousness and creating transcendental, universal truths, gestures toward the primacy of voice in conventional qualitative research.

To solve the problem of voice in conventional, interpretive, and critical qualitative research, methodologists have taken up various practices in attempts to “let voices speak for themselves,” to “give voice,” or to “make voices heard” (see Jackson, 2003, for an epistemological perspective and critique). As Guba and Lincoln (2005) explain, “As researchers became more conscious of the abstracted realities their texts created, they became simultaneously more conscious of having readers ‘hear’ their informants—permitting readers to hear the exact words (and, occasionally, the paralinguistic cues, the lapses, pauses, stops, starts, reformulations) of the informants” (p. 209). Along with this crisis of representation described by Guba and Lincoln, qualitative researchers have recognized the dangerous assumptions in trying to represent a single truth (seemingly articulated by a single voice) and have therefore pluralized voice, intending to highlight the polyvocal and multiple nature of voice within contexts that are themselves messy and constrained. This practice of “more is better” has indeed highlighted the ways in which voices are not singular, yet the obsession for more full voices side-steps what we view as a more salient feature of the problem: These practices remain attached to notions of voice inherited from metaphysics—voice as present, stable, authentic, and self-reflective. Voice is still “there” to search for, retrieve, and liberate.

A refusal to seek methodological practices that attempt to retrieve and liberate voice requires a decentering of the humanist voice that “knows who she is, says what she means and means what she says” (MacLure, 2009, p. 104). Such a refusal then challenges romanticized and oversimplified narratives that attempt to veil the fragmented subject that speaks with a voice that can never bear the burden of its weight. In writing of the “productive insufficiency of voice,” Maggie MacLure encourages a fostering of methodologies that dwell on the complications and troublesome voices that indeed do not ascribe to hegemonic discourses and knowledge claims. She continues by writing that these insufficiencies “are productive: they allow people to mean more than one thing at a time; to fashion mobile and nuanced readings of situations; to connect with others despite not knowing exactly ‘who’ they themselves are” (p. 98). In our move to complicate voice, we look for such inconsistencies and places of “meaning more than one thing at a time” as described in the sections that follow.

Do participants ever “speak for themselves?” Can we claim a higher ethical ground as qualitative researchers by taking the stance that we are presenting the “exact words” of our participants in an unadulterated form? We concur with Linda Alcoff’s (2009) delegitimizing of the authority, validity, and neutrality of the speaker and attempt to complicate voice by moving away from an “emphasis on the meaning of the discursive speech event to the importance of the movement of social location, possibilities of understanding, and the status/value/significance of truth” (Mazzei & Jackson, 2009).

We take then as our starting point unproblematized notions and practices of voice that posit the presentation of unadulterated participant voices or the inclusion of a multiplicity of voices as attempts to solve the problem of voice. We are convinced, for instance, that innovative practices that attempt to provide voice data that are more authentic, spontaneous, or realistic, whether that be a script for an ethnodrama, a raw transcript, or a photograph, may do little to engage the epistemological and methodological limits of voice. Letting readers “hear” participant voices and presenting their “exact words” as if they are transparent is a move that fails to consider how as researchers we are always already shaping those “exact words” through the unequal power relationships present and by our own exploitative research agendas and timelines. Even those accounts of voice that are more critical and that attempt to equalize and democratize the research process may do little to make transparent how decisions are made to “give voice”: Who decides what “exact words” should be used in the accounts? Who was listened to, and how were they listened to? How might voices be necessarily complicated, distorted, and fictionalized in the process of reinscription? And indeed, how are those voices necessarily complicated, distorted, and fictionalized in the process of both transcription and reinscription? The task would then be to examine whose interests are served by particular reinscriptions and whose are further marginalized. Extending the work by various authors in the text Voice in Qualitative Inquiry (Jackson & Mazzei, 2009), we will present examples from our own research that resist simplified and mechanistic treatment of voice and that refuse naïve claims of “letting participants speak for themselves.”

Maggie MacLure writes of being “interested in how the voices that are ‘heard’ in research texts carry so few of the qualities associated with the spoken voice” (2009, p. 98). What might it mean to seek those discarded and discounted qualities that not only complicate but also enrich meaning and the research text? That this spoken voice is not easily “captured” or “contained” is not cause for despair. As we wrote, “The fact that an easy truth cannot be spoken does not mean that truths, or narratives, or experiences cannot be represented” (Jackson & Mazzei, 2008, p. 307). It means that in a refusal to let voices speak for themselves, a more nuanced, complicated, and productive story may be told.

In an effort to tell such complicated stories, we go to examples from our research that have caused us problems. Problems in the sense that simplistic and mechanistic treatments of voice were insufficient as we attempted to produce knowledge. The participants were refusing our attempts to make them speak for themselves and were using other means beyond our knowing to communicate. In our refusal to tell an easy story and in succumbing to the siren call of the anxiety that such refusal portends, we began to “hear” these other voices and refusals. As we surrendered to the possibility of not knowing, we began to know and ask in ways that were previously unthought.

“Plugging in” Voices

To refuse to let participants speak for themselves, we offer a methodological move that is a practice that not only pushes against (over)simplified treatments of voice-as-data but also the traditional, interpretive analytic strategies that tend to take for granted what voice “does” in qualitative research. That is, the interpretive processes we were using (i.e., coding) were insufficient and we could no longer ignore what we had put up with before (see Deleuze & Parnet, 1987/2002, p. 126). In the context of qualitative research, specifically qualitative research that concerns itself with an analysis of speech and conversations, good methodologists are taught to organize what they have seen, heard, and read in order to make sense of and represent what they have learned. Well-trained methodologists are carefully taught to be attentive to their field notes and transcription data in order to sort and sift and identify the codes and categories that emerge from the data. Perhaps, we realized, we were no longer good methodologists. Perhaps we were becoming postmetho-dologists in the way that Patti Lather (2007) and Elizabeth St. Pierre (2009) have described.

We have moved away from thematic coding to a reading of data (of voices) that we have coined “plugging-in” (Jackson & Mazzei, 2012). The little phrase “plugging in” is one that we picked up from Deleuze and Guattari to capture our methodology of complicating voice. We first encountered “plugging-in” on page 4 of Delueze and Guattari’s (1987) A Thousand Plateaus: “When one writes, the only question is which other machine the literary machine can be plugged into, must be plugged into in order to work”. In our method of complicating voice, we were confronted with multiple texts—or literary machines: interview data, tomes of theory, conventional qualitative research methods books that we were working against, things we had previously written, traces of data, reviewer comments, and so on ad infinitum. That is, we had a sense of the ceaseless variations possible in having coauthored texts that relied on a plugging-in of ideas, fragments, theory, selves, sensations. And so we moved to engage “plugging-in” as a process rather than a concept, something we could put to work, for as Rosi Braidotti urges in this time of change, “the challenge lies in thinking about processes, rather than concepts” (2002, p. 1).

So what did we plug in? We used voices of participants, our own voices, theoretical voices, voices of our teachers/mentors, voices of other scholars, and so on. Rather than succumbing to the primacy of ONE voice in qualitative research (that of the participant), we plug in voices to produce something new: a constant, continuous process of making and unmaking. An assemblage isn’t a thing—it is the process of making and unmaking the thing, a process of arranging, organizing, fitting together. We characterize this plugging-in, of entering the assemblage, of making new connectives as making a different relationship among texts: They constitute one another and in doing so resist (over)simplification. Voices (of data, of theory) make each other in the plugging-in and create new ways of thinking about both theory and data. Articulation is about making new combinations to create new identities. In what follows, we show how we decenter both theory and data to see what emerges in the middle of plugging-in, to show the suppleness and overabundance of all voice, and to diffract rather than foreclose thought.

In the next section, we present two examples of what might enact a methodological plugging-in and, therefore, produces a retelling that displaces many of the normalizing features of (voice) in qualitative inquiry, that relishes in the promise of uncertainty and the thrill that such ambiguity might indeed produce. The first example is from Lisa’s research with white teachers and the problems encountered when attempts to speak about race are silently articulated. The second example is from Alecia’s interview with a Southern woman who belongs in a particular way to her present and speaks both directly and indirectly particular truths about her present (Foucault, 1990, p. 88). Let us reassert that all voices are ambiguous. Our resistance to simplistic treatments of voice acknowledges and exploits this ambiguity in our process of plugging-in.

“Seeing” Voice

In our efforts to allow expression of a complicated voice, we plug in Deleuze’s concept of the “image” of the speech-act in cinema (Deleuze, 1983/1986, 1985/1989) to some previously collected interview data. We begin this section with an explication of what a Deleuzian “image” prompts in our thinking, then we move to the practice of plugging-in with the data.

In his writings in Cinema I and Cinema II, we are reminded by Deleuze that in silent pictures, the voice is not contained by a speaking subject because subjects speak only indirectly through the use of intertitles (printed dialogue or narration that appears between scenes in a silent film) to transmit dialogue, visual text in the form of written documents, and visually constituted speech-acts (e.g., gestures, facial expressions, movements). The voices of the actors are communicated through the use of a “seen” image and an “intertitle” that is read by the viewer of the film. The intertitles are thus used to convey, in addition to other elements, “speech-acts.” Deleuze wrote that the silent film did not just call for the talkie but “already implied it” (Deleuze, 1985/1989, p. 216). Prompted by Deleuze, we might consider how our participants give voice, not in ways that are deemed absent as silent, but in ways that are meaningful as noiseless. By so doing, we begin to consider the intertitles and images used by our participants that function to convey voice. To consider the voices both performed and projected through these intertitles and images is to consider what is missed if we only rely on one or the other in the viewing of film (or encounter with research participants) as silent rather than noiseless.

A reading of voice from a multidimensional perspective as described above seeks a retelling that displaces simplistic notions of voice via a reimaging of voice in the context of our research projects. One such example is to think about “viewing” voice in qualitative research and how such viewing might make it possible to think voice as enhanced and multidimensional, much like 3D technology enhances the visual elements of film. Thinking about film, and how voice was conveyed in “silent” pictures, what might it mean to “see” a speech-act according to Deleuze? How does this inform methodological thinking that discards the truth/fabrication, speech/silence, or simplistic/complex binaries? ²

In a preservice course that Lisa taught for K-8 teachers, “Diversity and the Learner,” Cassidy was one such student who relied on all of the above-mentioned strategies. A traditional-age white student who had attended public school in an affluent suburb and who had previous field placements in a Catholic school and a suburban school, she wrote the following in her journal for the class: “This was my first experience in this type of school setting.” What is she saying when she describes a field placement experience at an elementary school in the large urban district where the university is located as her “first experience in this type of school setting [emphasis ours]?” We could interpret Cassidy’s assertion as “evidence” that she is concerned with the seeming lack of resources and state of the physical facilities in contrast to an earlier placement in an affluent suburban community. Or we might engage a more complicated listening that allows preservice teachers to talk about race without talking about it and that relies on a continuum of sound elements that reject a singular speaking subject. Perhaps she is concerned with the inequities that exist in schools, but perhaps she is also stating that “this type of school” is filled with low-income, minority students. If we only rely on the spoken or written words, “This was my first experience in this type of school setting,” we miss the out-of-field voices that may both contribute to and complicate meaning.

How might we begin to account for those voices both performed and projected? What are the implications for a reconceptualization of data as a result? First, consider the literal words that Cassidy speaks when she describes “this type of school setting.” Next, put those literal words alongside the out-of-field voices present in this particular context: (a) journal reflections by her and her classmates that make sweeping assumptions in reference to working with “at-risk” students in their “urban” field placements, (b) friends who are not education majors who voice fear of venturing away from the confines of the campus into the surrounding “urban” neighborhood, (c) local newspaper articles that perpetuate an image of crime and low achievement of “urban” youth, and (d) class discussions in which “urban” is used as a code word for Black, poor, crime-ridden cities. ³ Thus, we can read the out-of-field voice by substituting Black and poor in place of at-risk and urban. Such a viewing of voice forces us to rethink how voice and power are reasserted through such practices.

Using the methodological move of “plugging in” a Deleuzian visual speech-act to the range of both audible and silent “voice” data as presented above both contributes to and complicates meaning. If in our work as researchers we seek data and meaning in the form of a text that is directly communicated by participants, in other words, basing what we know on what we hear (literally), then we fail to also consider how what we know and subsequently hear might be based on what we see—not in a literal sense of what we see, although this can be the case, but in the sense that we narrowly define (or simplify) voice and thereby consider only one aspect employed by our research participants to convey meaning. Put differently, if we focus only on the scripted, spoken words in our strategies to capture data and make meaning, then we limit our understandings of what our research participants are saying, or trying to say. The limit arrives when we gather and produce “evidence” of these voiced encounters in the form of transcripts that reproduce and classify direct speech-acts.

The “plugging-in” above shows how Cassidy often communicated not via a “scripted text” but via silent dialogues not represented (or representable) according to simplistic notions of what “counted” as data. Cassidy instead projected voices only knowable to the researcher in the form of “transgressive data” (St. Pierre, 1997) to “speak” in ways that were outside the bounds of an easy knowing. Such a thinking of “seeing” voice as informed by Deleuze permitted a listening that did not simply rely on spoken words as the sole source of voice but on the out-of-field (what her friends say), the intertitles (journal reflections), and the silent speech-acts (the use of urban to describe particular schools). In other words, it relied on a complicated voice that refused to let participants speak for themselves.

Expanding the Frame

Deleuze refers to sound that does not have a relation to what is seen in the visual element of the film as having an out-of-field source, for example, the sound of boots when marching soldiers are not seen in the frame. In research with teachers, we could consider songs with homophobic lyrics on a teacher’s iPod, or the slamming of a door that interrupts our conversation with a research participant, as such out-of-field sources. Describing the effect of sound in cinema, Deleuze discusses the continuum of the use of sound elements in film including music and silence. “The voice is not separable from noises,” and noises can not only mask voice, they can also be used as elements of this continuum that is voice in film. For example, “sounds of doors, sounds of the sea . . . revolver-shots . . . the ‘attack’ of music and the ‘attack’ in the bank, the correspondences between these elements . . . form the power of one and the same sound continuum” (Deleuze, 1985/1989, p. 225). This continuum of sound fills both what is in the frame (the visual image) and the out-of-field, that sound which has no relation to the visual image presented in that moment. Sound both dwells in the out-of-field and “fills the visual not-seen with a specific presence” (p. 226). This continuum of sound and a consideration of what is in the frame and heard, in addition to what is out of the frame and heard with perhaps the same degree of impact and intensity, prompt us to consider what might be opened as we further explore voice in film and how it helps us rethink voice in qualitative research.

In our use of Deleuze to transgress the boundaries of data and meaning, we expand this notion of the out-of-field to include those images and “voices” that may serve to both guide and constrain our participants, some of which we have access to as researchers, but many which we may not (e.g., conversations in the teachers lounge, an argument with a spouse). Furthermore, to engage with Deleuze and cinema is to think the “speech-act” as an “image” in keeping with the visual, because as he states, “The heard speech-act, as component of the visual image, makes something visible in that image” (Deleuze, 1985/1989, p. 223). Deleuze compares the components of the silent image with the talking image and, in so doing, makes it possible to question what is made “visible” in the image of voice, or the speech-act broadly defined.

In this section, we plug in the image of voice in a Deleuzian visual sense of the word to read interview data from a multidimensional perspective. We use data from Alecia’s interview with Amelia, a Southern woman who participated in a qualitative study about adult women living out their adult lives in their childhood hometowns. When we plug in this “voice” data and Deleuzian indirect images, we see how voice is projected through intertitles, indirect images, and out-of-field “sounds.” In the particular data excerpt that follows, the intertitles and images that are produced project ways in which Amelia’s voice is bound to power. Amelia and Alecia were seated at Amelia’s dining table as she spoke of her church life and the new preacher, Steven, in her Baptist church when she said,

What is in the frame (the visual image) at the start of the excerpt is an interview situation between Alecia and Amelia, sitting across from each other at the dining table with the tape recorder between them—children and husband out of the home. Amelia had prepared sweetened iced tea and had taken Alecia on a tour of her new home before the interview began. Yet during a moment in the interview excerpt, the out-of-field (that sound which has no relationship to the visual image presented in that moment) appeared. Amelia’s husband enters the kitchen, preceded by a faint sound of the automatic garage door opening and followed by the sounds of a turning door knob, a squeaky opening of a door, and a shutting sound. Will’s footsteps interrupt the interview and “fills the visual not-seen with a specific presence” (Deleuze, 1985/1989, p. 226). Here we have a continuum of sound that forces us to assume all of these variations of voice as we consider “what is voice doing?” that is in the frame and heard, in addition to what is out of the frame and heard.

The sudden shift in Amelia’s voice from critique to admiration can also be heard/seen using Deleuze’s “noiseless” frame. If we consider Amelia’s spoken words along with intertitles of discourse and power, then her words can be considered more than simply contradictory or multiple versions of truths. Quantifying truth does little to complicate truth and keeps us bound to relationships of meaning rather than relationships of power (Foucault, 1990). In Amelia’s excerpt above, the out-of-field, noiseless “voice” occurs in the intersection of the discursive fields of patriarchy and Southern Baptist religion, which (as discourses) regulate certain acceptable rules and values that sanction what should be said, and what should not be said, about the institution of the church. Amelia’s church/religion is one of the discursive structures in which she constructs her subjectivity; many times in our interview, she returned to the topic of the importance of her church to her personal life and to her community’s ethos.

Returning to Deleuze’s idea that the voice is not contained by a speaking subject, then we can further extend our hearing to the out-of-field voices that our participants “hear,” and that we can access, should we seek the intertitles and silent speech-acts in ways that allow us to reconsider what constitutes voice. If we consider both Amelia’s spoken words and the discursive structures that speak silently but forcefully, we can consider how this served as a noiseless text that enabled Amelia to produce a truth, to inhabit a structure in a different way.

If we also think of the data, in this case Amelia’s transcript, and include in our analysis the intertitles and out-of-field noises that intersect, then what gets produced is a voice that is on a continuum in a Deleuzian description of the “speech-act.” Envision Amelia and Alecia in conversation with different intertitles running along the screen—a noiseless speech that conveys Amelia’s desire for recognizability as an acceptable version of a Southern Baptist woman. You might see/hear noiseless “voices” and images that permeate Amelia’s mind as she speaks: a wife of a deacon; a life-long community member who honors and practices the discursive values of her hometown; a loyal, submissive ambassador-ess who is supposed to admire the town’s preacher—not critique him. These noiseless voices/images function as out-of-field “sounds”—signifying a shift in power relationships and pulling Amelia into a different grid of intelligibility. In other words, when the noiseless, discursive frame shifted, so did Amelia’s truth. So the unspoken, noiseless out-of-field voices enabled Amelia to position herself among various subject positions as she spoke about her present. Such nomadism allows Amelia to “wander away from [herself], allowing something else to be produced” (Goodchild, 1996, p. 3). This new production of herself and of her truths, in their many forms—as direct speech, as intertitles, as indirect images, as out-of-field voices—surfaced as she spoke of her present.

How Does It Work?

In keeping with our project of “plugging-in” and “thinking with theory,” we take up the question as prompted by Deleuze: How does it work? In other words, in our analysis of data and our use of Deleuzian theory (or any theory for that matter) we give up on a search for meaning and, instead, look for the places of rupture that signal the partial, incomplete, and always-in-process tellings. As Cixous and Calle-Gruber (1997) wrote, “All narratives tell one story in place of another story.” Complicating voice by considering the out-of-field and the intertitles spoken is a move that allows the multiple tellings and privileged accounts to be present all-at-once-and-at-the-same-time. These accounts and complicated voices are always present. The task then for researchers is to try to allow them to be part of the data that we listen to.

To read/hear/listen to Cassidy and Amelia as communicating with voices both performed and projected through intertitles and images is to complicate voice in a refusal to let participants speak for themselves. Certainly, truth-tellers construct many things as they represent their selves to others (and to themselves). And this truth-telling is often plentiful with much that confuse researchers and that do not make “easy sense” (Mazzei, 2009). However, to consider the function and effects of voice does not take it at face value but as a not-so-innocent product of “giving an account of oneself” (Butler, 2005). It is to demand rigorous readings of data that require an intimacy and thoroughness not possible with mechanistic and simplistic approaches to data analysis.

We believe that to complicate/refuse voice in qualitative inquiry is to be more concerned about what voice “does” than what it “is.” What voice does—or its function, its effect—is neither simple nor conclusive. When participants tell stories, they do not innocently enact a narrative voice. As Judith Butler (2005) wrote, “In speaking the ‘I,’ I undergo something of what cannot be captured or assimilated by the ‘I,’ since I always arrive too late to myself” (p. 79). If we arrive too late to tell the truth(s) about ourselves—or arrive too late to capture a holistic, authentic narrative that tells it all—then there are other things happening in the moment of truth-telling. It is this “other” that we believe qualitative researchers ought to trace—not to account for but to show its constitution and its effects, effects that are immanent to, and a product of, relations; effects that are constrained and enabled by regulating rules and beliefs that arrange the possibilities of truth-telling; effects that position and elaborate the “I” through the telling and that becomes a failure in itself because of the traces of the unaccountable in “giving an account of oneself” (Butler, 2005). Voice-as-effect side-steps the simplicity of representing a multiplicity of voices and instead hones in on analyzing dimensions and textures of voice, especially the fragility and failure of voice to provide coherence, comfort, and presence.

We borrow here from Deleuze and Guattari’s concept of the “body without organs.” When they conceived the body without organs, they did so in an attempt to enact thinking without a subject and to confront our reliance on essential objects or material representations to understand and explain. A voice without organs then is our move toward a voice that is not linked directly to a speaking subject but rather one that is constituted in the intersection of images, intertitles, out-of-field noises, and other elements in the continuum of sound.

We think it is possible then to think a voice without organs, a voice too easily taken as a “thing,” or a voice without a subject in order to think the very limit of how voice is constituted, constrained, manipulated, and held constant in our efforts as researchers to craft a narrative, tell a story, give voice, and make meaning. Such an unthinking of voice, given a new vocabulary, calls us to question our motives as researchers as we continue to rely on and interrogate “a” subject to tell “a” story that is uncomplicated and that in fact relies on an unreliable subject that doesn’t exist. Or worse perhaps, to use ourselves as reliable narrators to convey or to “give voice.” If we are to use Deleuze to help us think voice beyond its already constituted forms, or to think voice that is not contained by a speaking subject, a “voice without organs” if you will, that is not bounded by the binaries between the discursive and the material, then what might be possible for us as qualitative researchers is a further extension of the elements of an image-less voice of which Deleuze writes (Deleuze, 1985/1989). One that complicates voice in a refusal to let participants speak for themselves.