I descended into a rabbit hole of qualitative research methodology, spurred on by a reviewer’s to closely read Braun and Clarke, our matrons of thematic analysis (who would likely reject such titles). As I dove into Braun and Clarke’s twenty-year war against postpositivist readings of thematic analysis, one intriguing author that I encountered within the health and nursing literature is Margarete Sandelowski.
While applied linguistics conversations on qualitative methodologies and reflexivity often emphasize the importance of writing up clear, systematic methods, less attention gets paid to qualitative research as an aesthetic and literary exercise. Where is the line between creative, logical, and scholarly writing in qualitative inquiry? Qualitative research inspires me because it shows how researchers can think like fiction writers. Qualitative inquiry helps me appreciate research not only as a creative process, but a creative writing process.
I have loved creating things since I was young. In high school I helped my teammates assemble highlight reels to send to colleges. I wrote, recorded, and edited public service announcements for my high school’s closed-circuit television production club. I blogged about my experiences in South Korea. I therefore appreciated experience as a source of data before I thought of becoming a researcher. I love the creative latitude that one can traverse when representing experience.
Sandelowski is a researcher who makes clear how academic writing is creative writing. Even if such writing doesn’t take the oft-stereotyped shape of feminist arts-based research or other radically divergent representational formats; even when one writes within the legible bounds of traditional research reports, writing nonetheless remains a creative endeavor. This is because, as Sandelowski writes, reading qualitative research reports is more than just an ingestion of information; it is an affective experience.
When readers read qualitative research reports, they are located on what Rosenblatt (1978/1994) described as a continuum of reading anchored by ‘efferent’ and ‘aesthetic’ readings. Efferent reading refers to what readers take away from a text such as information and facts, or in the case of a research report, methodological procedure and the findings of a study. Aesthetic reading refers to the reader’s relationship and aesthetic experience with a particular text. Here the reading is less about taking away information as it is engaging with the experience of reading itself and such things as the rhythms of the language or the prose style of a text. (Sandelowski, 2014, p. 90).
Reading Sandelowski’s citation of Rosenblatt gives voice to something that has bubbled up in my past five years as a qualitative researcher: Why do I sometimes appreciate scholars as writers just as much as intellectual contributors? Can we appreciate scholarly writing independent of its logical arguments? According to Eric Hayot (2014), author of The Elements of Academic Style, of course we can. For example, I love how Joseph Sung-Yul Park can weave theory and story together in his anthropological approaches to English language ideologies and L1 Korean speakers; or how Suresh Canagarajah can employ illustrative vignettes to bring a plurilingual ethos to applied linguistic scholarship, focusing on negotiation and activity over rigid or fixed power dynamics. Even Michael Silverstein – a name that strikes fear into the hearts of many early career linguistic anthropologists — a scholar deemed so dense and inaccessible that it spawned an entire genre of conceptual papers focusing on metapragmatics, eucharistic dynamic figuration, or baptisms of indexicality. Yet I also appreciate Silverstein’s ability to move from the opaquest prose I ever read in my life to cracking a joke about a Chicago law student’s “metapragmatic boo-boo.”
This thought of aesthetic reading — of reading as an experience of art rather than just an ingestion of a bolus of facts and information — is an underdiscussed side of the AI debate raging among scholars these days. Dominant critiques against AI usage often involves hallucinations, false references, misinformation, or plagiarism – arguments anchored to flaws in logic or argumentation. Yet others bemoan the experience of reading what appears to be AI-generated text. Some talk of AI-generated text as simulating rhetorical sophistication; increasing lexical density while hollowing out writing’s aesthetic core (Jandric, 2026). But when AI can summarize or compare and contrast a batch of articles in seconds, I also wonder if reading has become an endangered practice. Or, will reading become a purely technological exercise rather than an affective experience? What do you think? Do you enjoy reading published articles? Or are AI summaries a welcome productivity hack?
There forms a cleavage between qualitative and quantitative scholars’ interpretation of this question. And within the qualitative side, another cleavage between what Kidder and Fine (1987) call “Big Q” and “small q” qualitative research. Small q qualitative research, with its commitment to systemized codebooks and inter-reliability statistics, lends itself to summarized reading – in other words, reading to extract themes and scrutinize methods. This compares to Big Q qualitative studies, which often adopt a more literary stance toward research report representation. This is particularly true with the increasing calls for reflexivity in Big Q research, where a researcher is not only beholden to presenting their findings, but to telling a story of their findings. Furthermore, reflexivity demands authors to reckon with how their narration is less than reliable – at least against an objectivist, realist reporting standard (Canagarajah, 1996). Put otherwise, one’s stance toward reading research reports likely correlates with one’s implicit epistemological and methodological commitments.
I recently engaged in an interesting comment dialogue with an L2 political science researcher about this issue of AI and style on LinkedIn. Many researchers that favor AI usage point to AI leveling the linguistic playing field within the discriminatory apparatus of academic publishing. It is a fair consideration, as this apparatus long tilted in favor of L1 English users and against L2 English users. This is because first, English marks the dominant scholarly language. It is an indispensable lingua franca for scholars interested in establishing a global research footprint. Second, L2 English scholars long been punished by the publishing apparatus — not for the quality of their ideas but for how they represent those ideas in their own idiosyncratic form of English. It is almost cliche to talk of editors’ requests for non-native English speakers to enlist the expensive support of editing services, to find native-speaker co-authors, or to carefully proofread sentence by sentence in order to ensure that the manuscript is free of grammatical errors — notwithstanding that grammatical errors are not neutral but conditioned by inner-circle Englishes, their norms, and their conventions. I would say it was cliché if it wasn’t equally consequential and unjust.
So, while I, Mr. L1 English speaker, fear the loss of stylistic autonomy as AI adopts a prominent role in the drafting and review of academic manuscripts[1], I can also appreciate others’ enthusiasm for a tool that can save countless hours and dollars – folks like the political science scholar I engaged with on LinkedIn. I can see at first glance how AI can be a great equalizer that relieves L2 English speakers of the long oppressive burden of writing, editing, and enlisting the services of native speakers as they craft scholarship in an alien tongue. And I cannot deny how efficiently AI can proofread prose to avoid the ire and the censure of prescriptivist editors or reviewers.
On a side note, this emerging genre of “AI-written” text may connect to native speakerist ideologies that link such texts to an author’s inability to produce error-free English writing on their own. In other words, will “this was written by AI” become the new discriminatory backdrop for arguments previously framed as “this person needs to learn English?” Will the claim of AI-generated writing become the new field upon which native speakers assert privilege and domination within an applied linguistics scholarshipscape striving to embrace global Englishes? Or is AI-generated text itself a new flavor of global Englishes?
After all, that political science professor made a persuasive point: stylistic novelty is a burden he would rather not assume. He instead thinks of AI as a powerful tool that democratizes the production of scholarship by removing the stylistic barriers that have so long barred many L2 English speakers. Yet I also wonder whether the democratization of academic language through AI use, in turn, produces the homogenization of academic language (Bal et al., 2026) – a technologization of language that Martin Heidegger (1998) predicted over 60 years ago. As Thomas Manabu (2026) argues, AI can render the writing process both lifeless and frictionless. By reducing friction, we devalue the blood, sweat, and tears that have historically gone into writing scholarship.
In short, that political science professor made me realize that many scholars see writing as more burden than blessing – more of a task to accomplish than a craft to embrace. And reflexivity compels me to recognize my native speaker privilege in arguing that we, as scholars, ought not surrender the transformative practices of reading and writing to AI machines. There is a privilege in arguing that the potential benefits of removing barriers for L2 English or non-enthusiastic writers can also yield downstream effects that render scholarship even more opaque than it has already become. Studies often flood my LinkedIn that argue how unedited AI-assisted prose is more dense, complex, and inaccessible than the average human writer. Yet this is hardly a new development. Anyone who has gone through graduate school or read academic papers can sense the exclusionary jargon – shibboleths that separate the disciplinary insiders from the outsiders. Manabu (2026) argued that this disjunct in clarity comes from different audiences. Humans write for other humans. AI writes for computers to read. Graduate school socializes us to think, read, and write like computers, not to appreciate the aesthetic experience of reading. Academic language, in my experience, aligns with AI to prioritize comprehensiveness over clarity. This comprehensiveness appears opaque and unsettling to human readers. Which makes me suspect that academia socializes us away from our humanity. Or, academia is one of many arenas in which a post-AI world socializes our language to support information and reject poetics (Kramsch, 2026).
As this opacity of style colonizes academic writing, it will also devalue the importance we as scholars put onto the transparency of research methods and their representation. As Sandelowski writes,
Transparency is itself not an attribute of any inquiry procedure but rather a judgement made about its clarity and whether clarity ought to be judged at all (Sandelowski, 2014, p. 91).
This year, as I prepare to gather data for my ethnographic investigation of a science policy graduate student organization, I have grown obsessed with applied linguistics qualitative researchers’ obsession with the notion of transparency. The term gravitates to articles that discuss reflexivity, rigor, or argumentative strength within qualitative research in applied linguistics. I previously wrote how I often read transparency in these articles in two different ways; how transparency can become a Trojan horse that smuggles in realist and postpositivist assumptions within a postmodern or poststructural performance.
However, today I pivot to another discussion of transparency – discussions with my applied linguistics colleagues here at Penn State. People often advocate for more transparency in annual reports. People want greater transparency for the requirements of comprehensive examinations, dissertation proposals, or the criteria through which faculty allocate departmental graduate student awards. More transparency, more transparency!
As Sandelowski writes, transparency means clarity. I imagine transparency as a window – a window whose clarity helps us notice and realize predictable processes for predictable outcomes. If you have visited a Krispy Kreme Donut shop, you may have seen a window transparently displaying a herd of donuts bubbling in a river of oil passing under a waterfall of glaze. You can appreciate your donut because you have a transparent look into its production. Such a view of transparency, it seems, can allow graduate students, or reviewers, or editors, or readers to have a clear view of internal mechanisms so they can actively participate in decision-making processes. In other words, there seems to be an idealized endpoint of transparency where, if one presents their positioned reflexivity, their sampling and data gathering, and analytical procedures in rigorous and systematic detail, then there can be but only one achievable conclusion: an analysis that will speak for itself; an award everyone can predict. In this late-stage transparency, we can relieve ourselves of the burden of opacity and ambiguity.
But since when has research ever been so linear? So predictable? So devoid of unexpected surprises? Part of the joy of qualitative research, paralleling the joys of creative expression, is its nonlinearity; its unpredictability; its ability to surprise us. If I opened a novel, and the narrator explained the ending and the step-by-step approach by which the ending was achieved, I would put it down in disgust. To appreciate a novel, we need surprise. We need the author to withhold information until the perfect moment – that moment when revelation both surprises us and makes us recontextualize all we have read before. I crave a suspension of disbelief. This does not mean that we do not attend to rigor when we read a novel or watch a film. The most common critique of poor novels or films is that the plot is (1) too predictable or (2) too unbelievable. This extends to characters that likewise can be too predictible or too unbelievable.
As McKinley and Zhou (2026) wrote in a recent editorial for TESOL Quarterly, relevant studies know the stories circulating in literature. A publishable study both knows and subverts those stories. The best qualitative papers leave space for surprise. They present readers with new interpretations of old assumptions. They grab us and force us to reckon with the simplicity of stories told before. They make us realize life is complicated. This is not achievable when one overcommits to transparency. Instead, it requires a coherence of opacities. AI produces text opaque to humans and transparent to machines. Effective human writers produce stories that oscillate between opacity and transparency to give readers more than just information, but a reading experience where tension builds and releases in seasonal rhythms.
Scholars often associate transparency with systematicity and systematicity with predictability. But what if a Big Q qualitative researcher explained their uncertainty about the provenance of their findings? Or their selection of excerpts? Sometimes I select an excerpt not because it illustrates a theme, but because a vibration in my gut tells me that there is an important story to tell here — even if I cannot yet articulate how that story will unfold. Sometimes, as qualitative researchers, we do not make inductive or deductive moves of logic, but abductive leaps. We drift over our skis only to recover, as we backfill context and analysis.
Pratt (2008) talks about the value of checklists over procedures. Checklists provide a list of elements for compelling qualitative study, but do not prescribe a necessary order. Nor do they advocate for a rigidity of procedure that hardens into doxa — a series of arbitrary conventions so deeply overwritten by habit and tradition that they present themselves as objective and self-evident.
I always return to something Suresh Canagarajah told me: the best writing is risky writing. To follow the formula without deviation is to render your writing invisible. Furthermore, a formula is a construct hungry for critique – a structure that demands genealogical questioning; a regime of conventions that writers ought not take as a self-evident given. I sometimes fear that scholarship will be dead in the next ten years; that an over-reliance on AI will create a death spiral of structural repetition — an inescable doxa of acceptable style; that AI will become author, editor, and reviewer; that AI will rob us of our creativity. But if qualitative researchers continue to appreciate the practice of writing and reading, then all hope is not lost. If writers want to lean into AI for its democratizing potentiality, let them. I will create otherwise.
References:
Bal, P. M., Orhan, M. A., Van Rossenberg, Y., & Alhnaity, R. (2026). Reproduction of form in management and organization research: Why and how compulsively repetitive publishing kills academia. Organization. Advance online publication.
Canagarajah, A. S. (1996). From critical research practice to critical research reporting. TESOL Quarterly, 30(2), 321–331.
Hayot, E. (2014). The elements of academic style: Writing for the humanities. Columbia University Press.
Heidegger, M. (1998). Traditional language and technological language (W. T. Gregory, Trans.). Journal of Philosophical Research, 23, 129–145.
Jandrić, P. (2026). GenAI and academic writing. Postdigital Science and Education, 8(2), 372–386.
Kidder, L. H., & Fine, M. (1987). Qualitative and quantitative methods: When stories converge. In M. M. Mark & L. Shotland (Eds.), New directions for program evaluation (No. 35, pp. 57–75). Jossey-Bass.
Kramsch, C. (2026). A critical reflection on where applied linguistics is taking us. Language Teaching, 1-8.
Manabu, T. (2026). The machine consumes itself: Artificial intelligence and the end of publish-or-perish. Postdigital Science and Education, 1–24.
McKinley, J., & Zhou, S. (2026). From relevant to publishable in TESOL research: Contribution, access, and accountability: An editorial statement from the incoming editors of TESOL Quarterly. TESOL Quarterly.
Pratt, M. G. (2008). Fitting oval pegs into round holes: Tensions in evaluating and publishing qualitative research in top-tier North American journals. Organizational Research Methods, 11(3), 481–509.
Rosenblatt, L. M. (1978/1994). The reader, the text, the poem: The transactional theory of the literary work. Southern Illinois University Press.
Sandelowski, M. (2015). A matter of taste: Evaluating the quality of qualitative research. Nursing Inquiry, 22(2), 86–94. (Advance online publication 2014.)
[1] I know I’m not alone in this feeling whenever I visut my LinkedIn profile and read scholars’ stories about receiving AI-generated manuscripts as editors or receiving AI-generated reviews as authors.
Leave a Reply