Live Discourse Analysis

While not explicitly characterized as narrative analysis, thematic analysis (often framed as grounded theory [Corbin & Strauss, 1990]) uses participants small stories as a source for generating thematic codes – codes that authors often organize into a narrative when articulating overarching interpretations. As a result, Benson (2014, p. 156) uses narrative inquiry as “an umbrella term for a variety of approaches in which stories of the experience of language learning and teaching are in focus.” The literature is replete with examples of thematic analyses in qualitative applied linguistics research. It arguably marks the dominant paradigm when it comes to analysis of interview data (Benson, 2014).

Most such studies organize their findings along a handful of key themes shared among participants. Kang (2024), analyzing the beliefs of Korean learners of English while sojourning in the U.S., organizes themes kairotically (marking significant or serendipitous moments in time): (1) Learner beliefs about English language learning during study abroad, and (2) Shifts in language learning beliefs in relation to different aspects of study abroad. This is often the case in study abroad research, which often organizes participants’ attitudes into pre-study abroad and post-study abroad thematic bins. Howard (2019), on the other hand, in an analysis of professional identities among expatriate English teachers in South Korea, organized themes in relation to fluid, inhabited identities: (1) Distanced Other, (2) Commercial Asset, (3) Replaceable Part, (4) Cultural and Linguistic Ambassador, and (5) Professional and Respected Educator. In a grounded theory analysis, Ngyuen (2025), analyzing the language attitudes of ethnic minority college students in Vietnam, organizes findings in terms of behaviors: (1) Moving towards the mainstream norm, and (2) Moving away from the mainstream norm. These three truncated examples show how thematic analyses organize themes around many ontological axes including (but not limited to) (1) temporal dimensions, (2) identity dimensions, or (3) behavioral dimensions. While such studies do not often explicitly self-define as narrative inquiry, the excerpts in such studies often meet the criteria of small stories (e.g. Bamberg & Georgakopoulou, 2008; Marques & Koven, 2021, more on this below).

While thematic analysis or grounded theory ostensibly provides an empirical footing for organizing participants’ experience, often pushing against the literature by including heretofore unrepresented demographics or experiences, other scholars critique the practice of coding. They argue that coding interview data into themes (1) overgeneralizes participants’ experiences and (2) decontextualizes situated interactional events (Packer, 2017; Pavlenko, 2007). Coding, like other quantitative methodologies, also excludes outliers, anomalies, or contradictions in participants’ stories that do not fit neatly within bins of codes. As Packer (2017, p. 83) observes: “when we compare interviews and search for commonalities, we fail to take account ‘of the context and the diversity of the participants…[which also] tends to conceal the researcher’s reflexivity and underestimate [their] influence on the co-construction of the speech being analyzed” (cf. Barkhuizen & Cansoli, 2021; Prior, 2016). So contrary to the common assertion by grounded theorists that codes “emerged” from their data, “A theme never simply ‘emerges’; it is the product of interpretation” (Packer, 2017, p. 84; see also Pavlenko, 2007).

Finally, a review of the narrative inquiry literature would not be complete without acknowledging the concept of small stories (e.g. Bamberg & Georgakopoulou, 2008; Barkhuizen, 2016; Marques & Koven, 2017). Benson (2014) defines small stories as “narratives-in-interaction or more fragmentary accounts of experience in everyday talk.” In other words, small stories constitute segments of discourse that do not constitute self-contained structured narratives but do constitute accounts of lived experience and at times gesture toward narratives at larger scales, often called grand narratives (Lyotard, 1984). As Section 4 will argue, antenarrative (Boje, 2001a) approximates this definition of small stories, but also provides a powerful antidote to Barkhuizen and Consoli’s (2021) concerns over how narrative researchers define a story.

Ontogenetic Log

When I listen back, I recognize that I don’t like thematic analysis anymore. It’s not a methodology that appeals to me, but it is a methodology that once appealed to me. When I was an MA student at the University of Kentucky, I reported using thematic analysis in my MA thesis, where I used codes in alignment with Joseph Sung-Yul Park’s ideologies of necessitation, externalization, and self-deprecation within the South Korean Angloscene in order to argue for similarities and differences between the understandings of the ideological landscape between local Korean English teachers and foreign expatriate teachers of English. Thematic analysis appealed to me for several reasons: (1) it is replicable, (2) it is a great introductory method for green researchers, and (3) it performs a veneer of objectivity through saturation. Today, however, I think otherwise.

I view thematic analysis as, in some ways, the worst of both worlds when I consider narrative analysis. Narrative analysis serves a wonderful purpose within the field of qualitative methodologies in applied linguistics. Narrative analysis allows for more creativity in the analysis of data, and it affords the presentation of textured, layered, at times contradictory stories in order to remind ourselves in the field of language studies and language teaching that what we do is complex, situated, and ultimately sitting within never-ending intrapersonal, interpersonal, and political-economic tensions. To me, thematic analysis performs the objectivity and generalizability of the natural sciences while also refusing to claim the ground that it seeks to emulate. What I mean by that is researchers who use thematic analysis often also include performances of reflexivity, including positionality statements or other attempts to infuse reflexive undertones into their writing, and thereby accepting their interpretive report as one that is subjective, not generalizable. This is a bad deal for narrative inquiry.

Victoria Clarke posted an interesting LinkedIn post asking about the relationship between rigor and trustworthiness. At that time, Clarke was referring back to Guba and Lincoln and their framework of trustworthiness as a methodological benchmark in qualitative research. Guba and Lincoln’s framework contains four components:

1. Credibility: this maps on to notions of accuracy. In other words, does the research report accurately reflect the reality of participants?
2. Transferability: do the findings apply to other contexts? This parallels generalizability or external validity.
3. Dependability: can another researcher follow the same trail of observations in order to reach similar conclusions? In other words, is the thinking replicable?
4. Confirmability: can the research report demonstrate that findings emerge from data rather than the researcher’s own personal bias?

To be fair to Guba and Lincoln, they published this in 1985. I think about qualitative research quite differently in 2026.

For credibility, I substitute in critical presence. Qualitative data, particularly interview data, is never an accurate representation of a participant’s reality. It is not only constructed through language; it is constructed through two or more people’s languaging practices. Narratives are co-constructed, and importantly, they are storied. I substitute credibility for critical presence because I think that the goal is not to show the findings as accurate to reality, but to show a mix of centering participants in an ethical manner while also not allowing the researcher to fade into the background as a neutral view from nowhere. This requires more than a sequestered positionality statement. This requires a pendular movement from participant to researcher and back again, a recursive loop of storytelling that can present the most present form of analysis, what I would call live discourse analysis.

Instead of transferability, I aim for resonance. When one opens a piece of fiction, they do so knowing that (1) the story is not true, and (2) the story might involve characters’ experiences that are completely different from their own. Does this prevent fiction from moving us to thinking and seeing the world differently? No. Does it prevent us from changing the course of action having read a piece of fiction? No. This is because we don’t always look for comparable experiences. What we look for are resonant experiences. Resonant experiences do not linearly map on a one-to-one basis the connection between the reader’s life, the writer’s life, and the participant’s life. What we create are resonances, affective movements that cannot often be explained through a logical trail of propositions, but rather through a connection that often exceeds language. Because this connection often exceeds language, I have little more to say about it at this time. But in the same ways that the resonances of two different vibrating bodies can still harmonize together, the diverse experiences of participants, researchers, and readers can still harmonize together if a research report allows or provides the conditions for that harmony to take place. Scarily, this involves writing beyond logic and proposition and cognitive bases in favor of more poetic concerns (Kramsch, 2026).

Instead of dependability, what I would argue for is integrity. This integrity comes across through reflexivity. Yet, I do have a problem with the ways that people often discuss dependability in contemporary applied linguistics scholarship. This is particularly through what I call an ideology of transparency. I read, particularly from scholars like Sal Consoli or Matthew Prior and other leading narrative researchers in the field, that it’s important for a researcher to make their positionality, their thinking, and their methods of analysis transparent. I would argue, however, that such presentations and articulations of transparency within a research report are ultimately also narrativized. One cannot lay bare one’s entire methodological process. This is not only because we have limited word counts when preparing a manuscript. This is also because transparency presupposes that the researcher knows themself. And while it is a low bar for a researcher to know themself in terms of intersectional identity categories and motivations that often constitute positionality, I would also argue that the thinking of a researcher constitutes more of a black box than we are comfortable to admit. The directions that our thinking takes are oftentimes nonlinear and opaque to what we can consciously articulate in language. What I try to argue for in live discourse analysis is one of integrity. This involves incorporating the pendular movements of thinking with participants and thinking with oneself by also narrating it in the present tense, being able to articulate a storied stream of consciousness that not only makes an argument but can also articulate feelings, emotions, vulnerabilities, doubts, things that oftentimes are stereotyped within research reports as elements of weakness. I think we should reposition these as elements of strength.

And finally, in lieu of confirmability, I’m more interested in the practice of intra-action. I borrow this famous concept from Karen Barad, who discusses the concept of space-time mattering and the fact that the tool used to measure a phenomenon—and this includes the researcher as a thinking-embodied tool—has a direct effect on the measurement that comes out. This is not a calibration issue. Rather, this is a quantum entanglement issue. A phenomenon is only made visible through the limiting lens of the way that it is measured. Therefore, to separate the bias of the researcher from the credibility of the findings is a fool’s errand in the quantum age. What we have instead is an honest—not transparent, but an honest—accounting of the interactions that occur. The simplest way that this is invoked concretely is by including the interviewer’s voice within interview data. It is also involving the researcher better storying their methodographic trajectory. As well as treating the traditional practice of member checking not as an accounting of truth, but as a newly emerging layer of re-storying that does not allow for a confirmation, but rather allows for a more complex and layered reading of both participants’ and researchers’ biases, thoughts, and experiences.

In short, I don’t knock the use of thematic analysis. Yes, I do find the research articles that invoke thematic analysis to be rather boring and predictable, akin to a movie that explains all of its themes and plot points before starting. I prefer methods of analysis that allow a reporting of surprise. In fact, this is my biggest gripe with qualitative research and applied linguistics today. We feel such a need to compress our findings within the context of the introduction for the sake of those who do not wish to read, that we do a disservice to those who actually take the time to read articles in full. That’s not to say that everyone must read articles in full, but it can be worth asking whether or not we ought to reward those who put in effort, rather than feel like we want to give equal accessibility to those who do not give an article equal access as a reader.