Beyond the Axiom: The Scaffolding of Confidence
When we teach, we are also learning. This axiom is almost a cliché in higher education, yet it becomes newly real when one is removed from familiar routines and asked to teach in a different linguistic, cultural, and pedagogical environment. My Fulbright experience in Taiwan did exactly that. It brought me to National Cheng Kung University (NCKU) in Tainan—one of Taiwan’s premier research universities, world-renowned for its engineering and semiconductor programs—where I taught two undergraduate English-medium instruction (EMI) courses on Latin American literature and culture. What I had imagined as primarily an intercultural teaching project quickly became, for me, a lesson in voice, anxiety, and the pedagogical design required for academic agency in a foreign language.
My Fulbright project was designed to strengthen students’ academic English literacy and intercultural awareness, using Latin American literary and cultural materials as the vehicle for rigorous disciplinary engagement. I taught two courses: Mosaics of Latin America: Culture, Arts and Sports (拉丁美洲文化拼圖：文化、藝術與運動), a General Education course, and Latin American Voices: The Short Story in Translation (來自拉丁美洲的話語：其短篇故事與翻譯), an upper-level elective. Both courses asked students to read, discuss, write, and present in English while engaging with intellectually demanding content: indigenous civilizations, colonial histories, visual culture, sports, identity, and major Latin American writers in translation.
Teaching humanities in a university so defined by STEM excellence provided a unique vantage point. I arrived well prepared in many respects, with years of experience teaching Spanish as a second language and Latin American literature in English. I had carefully designed my syllabi to support both disciplinary English development and substantive humanistic inquiry. What I was not fully prepared for was a classroom dynamic quite different from my own, where silence and hesitation often signaled not disengagement, but a careful negotiation of language, confidence, and risk. I soon realized that the true task was deeper: helping them find their own voices within an advanced language level they often felt hesitant to inhabit.
I was also acutely aware of my own limitations. I do not speak Mandarin, Taiwanese, or any of the Indigenous languages spoken by some of my students. The classroom was only possible because of the bridge they had already built through years of studying English. My presence depended on their linguistic labor. I wanted to honor that reality—not by lowering academic standards, but by creating conditions in which they could use that English with confidence, without fear of imperfection.
On the first days of class, students were warm, generous, and clearly curious. But they were also hesitant. When I asked questions, many first wrote down their answers, checked their phones for grammatical accuracy, and then read their responses aloud. I quickly realized that the central obstacle was not lack of knowledge or preparation, but anxiety: the fear of speaking imperfect English in public in a high-stakes academic setting. In other words, the challenge was affective (fear of use) rather than cognitive (lack of knowledge). In EMI contexts, especially in the humanities, silence is rarely empty. It can signal caution, fear of negative evaluation and public shame, linguistic insecurity, or the difficulty of thinking across languages while under pressure.
That challenge was intensified by another force hovering in the background: digital support tools, including generative AI. In the end-of-semester anonymous survey, 86% of students reported using digital tools to support their English, especially for understanding readings, translation, and grammar. Rather than centering these tools in the classroom, I chose to design activities that asked students to work without them at key moments. Recent work by Lin and Chen (2025) calls for ethical, pedagogically grounded engagement with generative AI rather than panic or simplification. I agree. But my Fulbright experience convinced me that before students can negotiate AI well, they need to experience themselves as capable without it.
Looking back, I now see that my most important work was affective before it was technological. By designing classroom structures that lowered anxiety while maintaining high expectations, I was trying to help students move from fear to agency. In the end-of-semester survey, a clear majority of students reported higher confidence speaking English in class, with no students reporting lower confidence than at the beginning of the semester. A substantial number of students also reported lower anxiety when speaking English. Those results matter to me not because they prove success in a narrow quantitative sense, but because they confirm what I felt in the room: students were beginning to trust their own voices.
What follows are four scenes from that process.
Scene 1: Accent as Permission to Speak
One of the earliest turning points came in a workshop I called “Speak Up with Confidence in English.” I began with a slide on accent and identity. I told students that everyone has an accent, and that an accent is not a flaw to erase but part of a person’s history. The goal, I explained, was not perfect pronunciation or sounding “native.” The goal was clear communication and confident connection.
I used myself as an example. Spanish is my first language, and my English carries traces of that history. For years, I too had felt apprehension about speaking in English, especially about my accent, but I eventually came to understand that accent tells a story: where we come from, what languages have shaped us, and how we have moved through the world. Students were visibly surprised by this framing; several immediately pulled out their phones to take pictures of the slide. I also shared public-speaking ideas that reframed nervousness as something manageable, including the notion that speakers often look less nervous than they feel (Wrench et al.) and the memorable reminder that “it’s okay to have butterflies; just get them to fly in formation” (Bryant).
This moment mattered because it reframed English not as a test of proximity to a native speaker norm, but as a medium for intellectual participation. It was an exercise in lowering the “affective cost” of speaking—an essential step in building a student’s willingness to communicate in a foreign language. In a Freirean sense, it also challenged a subtle hierarchy that often governs language learning: the idea that legitimacy belongs to those who sound effortless or polished in a specific way (Freire). Instead, I wanted accented English—including my own—to be understood as a legitimate vehicle for thought.
Students later confirmed how important this was. One wrote that what helped most was “not stressing out about accents and utilizing the anxiety to perform better.” Another reflected, “Teacher keeps emphasizing that accent is not really important, and that everyone has their own accent as an identity made me feel less anxious.” A third wrote simply, “Embracing my imperfections.”
These responses showed me that the accent conversation was not a decorative preface. It was foundational. Before students could speak more, they had to feel that their way of sounding in English was not a disqualification from speaking at all.
Scene 2: Timed Writing and AI-Free Authorship
If Scene 1 addressed the emotional threshold of speaking, Scene 2 addressed a related issue: students’ dependence on digital mediation when producing academic English.
In the short story course, I assigned two in-class reading reports. These were closed-book, timed writing exercises in which students produced analytical essays by hand in response to literary prompts on Horacio Quiroga, Jorge Luis Borges, Julio Cortázar, and Augusto Monterroso. They were demanding assignments, especially in an EMI setting. Precisely for that reason, I wanted students to discover that they could think and write in English without relying on Grammarly, translation software, or text-generating AI.
My syllabus took a firm stance against AI-generated work submitted as original writing. But the deeper issue was not prohibition alone. I did not want students merely to avoid AI because it was forbidden; I wanted them to experience themselves as capable without it. If students believe that only digitally polished English counts as “real,” then AI becomes psychologically necessary. A more durable pedagogy must restore confidence before it debates tool use.
This is where Vygotsky helps me understand what was happening. Students did not arrive ready to perform these tasks automatically; they needed scaffolding and a classroom climate that made risk tolerable. Learning, in this sense, was socially mediated before it became individually demonstrated (Vygotsky). Once the conditions were in place, many students were surprised by what they could do.
Post-exam conversations revealed students’ pride at having produced coherent literary analysis in real time and without devices. One survey response captured the cumulative effect: “It wasn’t really one thing specific, but the repeated practice in class was really helpful to me in improving my confidence.” Another wrote that “multiple practices made me feel less anxious about using English.” These reports were acts of confidence-building, showing students that their English did not need to sound machine-polished to be intellectually valid.
Scene 3: Paraphrasing and Owning Academic English
A third key scene emerged through paraphrasing workshops in Mosaics of Latin America. I designed a workshop around sentence starters, modeling, and peer-checking, focusing on difficult conceptual passages. We worked through complex arguments from Bradford Burns’s The Poverty of Progress, a foundational text on 19th century Latin America. The task was not just to summarize, but to restate those dense historiographical ideas clearly and accurately in their own words.
At first, I worried that this workshop would feel mechanical. Yet students realized that paraphrasing was not merely a technical safeguard against plagiarism; it is a mode of comprehension and intellectual ownership. To paraphrase well, they had to author their own understanding. In an age where digital tools generate plausible reformulations instantly, I wanted to cultivate human paraphrasing as an exercise in disciplined attention—a way to reclaim the space between thought and expression.
Freire’s critique of the “banking model” of education is useful here. In such a model, students receive finished language and deposit it elsewhere; AI can intensify that logic by presenting a finished product before the student has had to work through the idea on their own. My goal was the opposite: to ensure the classroom remained a space where students could “name the world” in their own words, however imperfectly (Freire).
The survey data suggests that beneath the surface of these tasks, students were drawing on a complex reservoir of linguistic and conceptual resources. My students reported, at varying frequencies, that they first planned ideas in Mandarin or another familiar language before expressing them in English, often drawing on bilingual strategies to navigate course content. Heugh understands “transknowledging” as the two-way translation, exchange, production, and transfer of knowledge across languages and knowledge systems in multilingual education. In our classroom, this made possible a quiet but powerful encounter: Latin American history and literature were not simply rendered into English for comprehension but reinterpreted through students’ local linguistic repertoires and the situated perspectives of Taiwanese culture (Heugh; Gu et al.).
Scene 4: Poker Chips and the Redistribution of Risk
The most visibly innovative strategy I used was also one of the simplest: poker chips.
Because participation counted significantly, I wanted a system that was fair, transparent, and less face-threatening than cold-calling. Students worked in groups, and each group received poker chips corresponding to participation points. A chip was surrendered each time someone spoke, and speakers were required to rotate. This shifted responsibility toward the students while preventing a few highly confident speakers from dominating the room.
The strength of the system was affective. It redistributed the emotional risk of speaking. Participation became a shared responsibility. Students held the chips themselves and decided when and how to use them, returning them to me as they spoke. Students had time to confer, decide who would speak, and build confidence together. Because the chips had to be used, the system created a gentle but persistent pressure toward speech while avoiding the face threat of direct cold-calls.
Students embraced this structure. One wrote: “I think the tokens really help us speak more… I used to be afraid to raise up my hand… but in this class, I speak much more than before.” Another described it as building relationships: “It gives me an opportunity to work with my team… we seem to build good relationships afterwards.” Another simply called it “very creative.” These comments confirm that participation is about more than language proficiency; it is about the social distribution of risk. In a large EMI class, poker chips became a way of organizing courage.
Putting the Butterflies in Formation
These four scenes taught me that effective EMI humanities teaching depends on careful attention to affect and classroom risk. My students did not need me to lower standards. They needed structures—explicit guidance, repeated practice, and group support—that made those standards reachable.
That, in the end, is also how I now understand the AI question. Generative AI is part of students’ academic reality, and ethical conversations about its use are necessary (Lin and Chen). But the more fundamental pedagogical question comes earlier: do students experience themselves as authors of meaning, or only as managers of linguistic deficiency? If the latter, AI will feel indispensable. If the former, AI becomes one tool among many, rather than a substitute for voice.
After the impromptu speaking exercise, a noticeably quiet student waited for me after class. In a very Taiwanese way, he said: “If I may, I believe that I did quite well. I cannot believe it.” I agreed—his group had moved beyond simply describing the object and used it as a launchpad to discuss a more meaningful topic. Then he added, with a slight smile, “I believe I managed to put my butterflies in formation,” echoing Bryant’s idea that nervousness is inevitable, but can be channeled rather than eliminated. That moment stayed with me.
That phrase stayed with me because it also captured my own Fulbright experience. I had arrived in Taiwan with my own butterflies: uncertainty about classroom culture, student expectations, and how my teaching would translate into an unfamiliar environment. What I learned there was that effective EMI humanities teaching is not a matter of bringing a set of techniques —or a ‘bag of tricks’— into a new classroom. It requires attentiveness, humility, and careful scaffolding. It also requires confidence in students—not confidence that they will perform perfectly, but confidence that they can grow when given the right structures.
A visit later in the semester to the National Taiwan Museum in Taipei, where I explored an exhibit on the natural history of butterflies, gave that phrase a different resonance. The exhibit describes Taiwan as a “Kingdom of Insects,” where a vast mosaic of micro-climates allows hundreds of species—including its renowned butterflies—to thrive. I was struck by the parallel with Taiwan’s linguistic landscape. While public debates often center on English and Mandarin, the lived reality is a far more complex ecosystem that includes Taiwanese, Hakka, and Indigenous languages. In my classroom, validating accented English was a conscious refusal to flatten this linguistic richness into a single ‘native’ standard. It was an exercise in English as a Lingua Franca—creating a space where students could move from the silence of hesitation toward the agency of expression without having to erase their own histories.
Perhaps the most unexpected feedback came from several students who told me that the confidence they found in our classroom had traveled with them into their Mandarin-language STEM courses. They reported participating more actively in their other classes, and some even shared that their professors had noticed the change. This suggests that the transformation was not just about English; it was about the transferability of agency across contexts. When a student learns to trust their own voice in a challenging EMI humanities course, that newfound confidence becomes a tool they can carry across disciplines and languages. Consequently, these practices are not limited to one classroom; they can inform broader EMI program design and student success initiatives, offering a scalable model for faculty development across disciplines. This reinforces a claim I have explored in other contexts: that voice in language learning is not simply a matter of proficiency, but of legitimacy and self-recognition (Martínez).
In a broader sense, helping students find their voices in English was about much more than speaking up in class. It was about legitimacy, agency, and presence. It meant recognizing that a student’s voice includes accent, hesitation, perspective, and lived experience, and refusing the idea that meaningful academic participation belongs only to those who sound effortless. And, for me, it became something shared: not just their voices, but our voices.
Fulbright challenged me to teach in a new context, but it also asked me to listen. In helping students find their voices, I was reminded that voice is larger than language. It is how we present our values and histories to others. If I learned anything in Taiwan, it is that voice does not begin with perfection. It begins with permission, practice, and trust. By helping them find their voices, I was able to rediscover my own.
And that may be one of the deepest lessons of my Fulbright journey.
References
- Bryant, Adam. “How to Speak in Public.” The New York Times, Nov. 1, 2018.
- Freire, Paulo. Pedagogy of the Oppressed. Continuum, 1970.
- Gu, Michelle Mingyue, Amy Wanyu Ou, and Chi-kin John Lee. “Translanguaging and Transknowledging Practices Among STEM Teachers in EMI Higher Education.” Applied Linguistics Review, 17, 2025. pp. 467-489.
- Heugh, Kathleen. “Southern Multilingualism, Translanguaging and Transknowledging in Inclusive and Sustainable Education.” Language and the Sustainable Development Goals, edited by Philip Harding-Esch and Hywel Coleman, 2021, pp. 37–47.
- Lin, Angel M. Y., and Qinghua Chen. “Towards Ethical and Responsible Engagement of Generative AI in Education: The PAA Model and 4T Lenses in Action.” Designing Learning with Multimodality in English Medium of Education (EME) Classrooms Across Asia, edited by Fei Victor Lim and Jack Pun, Bloomsbury Academic, 2025, pp. 235–258.
- Martínez, Luciano. “Who We Really Are? Disciplinary Struggles and the Role of Literature in Language Departments.” ADFL Bulletin, vol. 47, no. 1, 2022, pp. 89–95, https://doi.org/10.1632/vjyp5250.
- Vygotsky, Lev S.. Mind in Society: The Development of Higher Psychological Processes. Harvard University Press, 1978.
- Wrench, Jason S., et al. Stand up, Speak out. The Practice and Ethics of Public Speaking. FlatWorld 2011.
