Global Summit 2026: Innovations in Language Education

Core Synopsis

The operational truth that this presentation forces us to confront: AI now functions as a functional substitute for second language acquisition in the majority of practical communication contexts—real-time translation, multilingual interface access, cross-lingual document comprehension—making it necessary to ask, with intellectual honesty, whether SLA as a field is defending the right things for the right reasons. The presenter's diagnosis: SLA is a 60-year enterprise that has generated enormous empirical knowledge but has largely failed to consistently accelerate language learning outcomes at scale in institutional settings, and now faces a technological discontinuity (AI as real-time linguistic intermediary) that dissolves the rationale that motivated the field's founding. The AI discontinuity is not incremental improvement—it is a categorical shift: the friction of linguistic difference, which SLA sought to reduce through learning, is now addressable at machine speed. The stakes are high: if SLA does not pivot from "how to learn languages better" to "how societies configure language, identity, and human development in a linguistically flat world," it risks irrelevance to policy, resource allocation, and educational design. We are committing to naming this pivot explicitly, because ignoring it ensures the field is repositioned by others—technologists, economists, and policymakers who don't share its humanistic commitments.

SLA's Rationale Under AI Pressure

Core founding rationale: Acquiring another language expands communicative access, economic opportunity, cognitive development, and cultural understanding.

Now under pressure from: AI translation tools (DeepL, GPT-4+) that provide near-instant, high-quality cross-lingual communication without acquisition.

The question being forced: If the pragmatic access rationale is met by AI, what remains of the case for language learning?

The Core Diagnosis

Claim: SLA has been a 60-year field that failed its core promise—consistent, scalable improvement in language learning outcomes in institutional settings.

Evidence basis: Despite decades of research on input, interaction, form-focused instruction, and individual differences, average instructed language learning outcomes remain modest, slow, and variable.

Implication: The field cannot credibly defend its resource claims on the basis of delivering reliable language outcomes at scale.

The AI Discontinuity

Not incremental: Prior disruptions (computers, the internet) changed how SLA was practiced; AI changes whether it is necessary for many communicative goals.

Substitution domains: Real-time conversation (AI earbuds), document comprehension (AI translation), professional writing (AI drafting), software interfaces (AI localization).

Non-substituted domains: Deep cultural participation, identity construction through language, literary and aesthetic engagement, metalinguistic awareness as cognitive development, community belonging through shared linguistic practice.

The Stakes

For the field: SLA must decide whether it defends its traditional position (language learning as intrinsically valuable) or reconfigures around the non-substituted domains.

For learners: Policy and institutional support for language learning may decrease if the pragmatic rationale is perceived as met by AI.

For society: Linguistically flat worlds are not culturally flat worlds; loss of multilingual capacity has identity, political, and epistemic consequences not visible in efficiency metrics.

Counterarguments and Their Limits

Counterargument 1 (Humboldt/Bildung): Language learning shapes thought and personhood—AI translation does not replicate this formative function.

Limit: True, but this argument has not historically driven SLA research agendas or policy funding.

Counterargument 2 (Fluency and nuance): AI translation remains imperfect for nuanced, informal, or culturally embedded communication.

Limit: Rapidly improving; the argument becomes weaker each year.

Counterargument 3 (Identity and belonging): Authentic membership in a community requires its language.

Limit: Valid, but applies to motivated learners in immersive contexts—not to the institutional mass-instruction contexts where SLA's claims are under the most pressure.

Implications: The Research Agenda Pivot

From: How do learners acquire L2 grammar, vocabulary, and pragmatics more efficiently?

To: How do individuals, communities, and societies configure language, identity, and human development when AI removes communication friction?

New research questions include: What is lost when linguistic acquisition is bypassed? How does AI-mediated multilingualism change intergroup perception? What forms of language education retain value in a linguistically flat world? How do we design institutions for post-substitution SLA?

Action Items

• Organize a field-wide symposium on SLA's post-AI value proposition, bringing together researchers, policymakers, and AI developers to articulate the non-substituted functions of language learning —
• Commission a systematic review of AI substitution coverage across SLA's traditional communicative domains to map where the field's traditional rationale holds and where it has been superseded —
• Develop a public-facing position statement on what language education offers that AI cannot, grounded in identity, culture, cognition, and democratic participation rather than communication efficiency alone —
• Redesign SLA doctoral training to include AI literacy, philosophy of language education, and policy advocacy alongside traditional acquisition research methods —

Submitted Abstract

For over sixty years, SLA research has produced an extraordinary volume of theory, method, and published findings. Yet, by any hard external measure, the field has made no measurable, cumulative progress on its central promise of making adult language learning meaningfully faster, easier, or more successful. This talk argues that SLA's trajectory resembles alchemy more than chemistry. It is a genuinely sophisticated, well-funded discipline that has produced real secondary discoveries while never once achieving its central goals. Using the presenter's own research in CALL as a case study, the talk identifies a pattern common across the field: substituting theoretically-motivated proxies (negotiation of meaning, learner engagement, positive attitudes) for the one outcome that actually matters, measurable gains in language ability. This substitution has gone largely unchallenged for decades, protected by the fact that, until recently, there was no real alternative to learning a language yourself. That protection is now ending. Advanced AI models do not merely support second language acquisition the way earlier technologies did. Instead, they offer to bypass it entirely, providing instant, high-quality translation and conversation without the years of effort SLA has never learned to reduce. The talk closes with a proposal for what a defensible SLA research agenda might look like once the instrumental rationale for mass language education disintegrates.

Core Synopsis

The operational truth: ESP/ESAP students in managerial or professional-track programs can develop both English proficiency and professional identity when crisis communication is the content spine, but only if the course explicitly scaffolds the rhetorical moves that make crisis narrative credible—collective framing, transformative reframing, and strategic action—rather than defaulting to generic business English templates. The course at the center of this summary used a workshop model with B1–B2 students, deploying two crisis storytelling tasks and generating twenty reflective responses that reveal a consistent theme: students began the course as "English learners" and exited as "professionals who communicate in English"—a meaningful identity shift with pedagogical implications. The "AI-choice" component allowed students to select any AI tool as a drafting partner and then critically evaluate the output against the three-part crisis framing criteria, building AI literacy alongside language competence. The political layer—the instructor's explicit rejection of the "English teacher" identity in favor of "coach" or "communication strategist"—models the identity work that ESP/ESAP is actually trying to cultivate in students.

The Pedagogical Pivot

Standard ESP problem: Students learn business vocabulary and genre conventions but do not develop the rhetorical agency to adapt language strategically to high-stakes professional contexts.

This course's solution: Crisis communication as the content frame—inherently high-stakes, authentic, and motivating—that requires students to move beyond sentence-level accuracy toward discourse-level strategic choice.

Goffman's frame analysis: The course uses Goffman's framing concept to help students understand how the same situation can be narrated as catastrophe, challenge, or opportunity depending on framing choices.

Course Architecture

Level: B1–B2 students in a professional or academic track.

Format: Workshop-style; emphasis on production over passive reception.

Two anchor tasks:

• Task 1: Analyze a real crisis communication case (organizational, political, or personal) using the three-part framing model.
• Task 2: Draft and deliver a crisis narrative in English using AI tools, then critique the AI output against the framing criteria.

Reflection component: Twenty individual reflections submitted across the course; thematically analyzed for identity shifts and strategic awareness gains.

Crisis Framing Triad

• Collective Challenge: Narrating the crisis as a shared problem that mobilizes collective effort; removes blame, builds solidarity.
• Transformative Opportunity: Repositioning crisis as a catalyst for positive organizational or personal change; forward-looking framing.
• Strategic Action: Communicating a clear, confident, evidence-based response plan; establishes credibility and agency.

Application: Students evaluated AI-generated crisis narratives against these three criteria, identifying where AI defaulted to generic reassurance rather than strategic framing—a productive critical exercise.

Professional Discourse Socialization

Theoretical base: Professional discourse socialization (Duff, Belcher) — learners must be initiated into the rhetorical practices of their target discourse community, not just its vocabulary.

Instructor identity: The presenter explicitly positions herself as a "coach" or "communication strategist," not an "English teacher"—a stance she models for students as part of their own identity development.

Students as emerging professionals: By handling realistic, high-stakes scenarios in English, students begin to inhabit a professional identity rather than a learner identity—the affective and motivational shift reflected in the reflection data.

Reflection Themes

From 20 reflective responses, dominant themes:

• Shift from "I am learning English" to "I am communicating professionally in English."
• Recognition that English is a vehicle for professional authority, not just information transfer.
• Awareness of AI's tendency to produce safe, generic language that lacks strategic rhetorical force.
• Increased confidence in high-stakes speaking and writing scenarios.

Political/Intent Layer

Instructor's explicit stance: She rejects the "English teacher" label as limiting and disempowering for both instructor and student; ESP must be taught by people who understand the professional discourse communities they are preparing students to enter.

Consequence: Course design, assessment, and feedback are all calibrated to professional standards, not language-learning milestones—a consequential design choice that students feel immediately.

Action Items

• Develop a crisis communication case library (5–10 cases across sectors: corporate, academic, public health, environmental) for use in ESP/ESAP courses —
• Create a three-part framing rubric (Collective Challenge / Transformative Opportunity / Strategic Action) as an assessment tool for crisis narrative tasks —
• Design a structured AI-critique protocol: students generate AI crisis narrative, score it against the rubric, identify rhetorical gaps, and produce a revised human version —
• Pilot the course with a pre/post professional identity measure to quantify the "learner → professional communicator" shift —
• Present the instructor identity framework (coach vs. English teacher) at ESP/ESAP professional forums to spark field-level conversation —

Submitted Abstract

Storytelling has been widely adopted in language teaching as a pedagogical tool for enhancing students' communicative competence, particularly speaking skills. However, within English for Specific Academic Purposes (ESAP), storytelling has received comparatively limited attention as a professional communicative competence in its own right. More importantly, its potential to socialize students into the communicative practices, identities, and discourse conventions of their future professions remains underexplored. This study addresses this gap by examining a workshop-based strategic storytelling approach as an innovative ESAP pedagogy in the Strategic Storytelling in Management Communication class. Adopting professional discourse socialization (Duff, 2010; Morita, 2004) as the central theoretical lens, this study examines how students construct professional identities and make sense of their experiences through workshop-based strategic storytelling. The study involves 20 undergraduate students with English proficiency ranging from B1 to B2 on the Common European Framework of Reference for Languages (CEFR) who were enrolled in the Strategic Storytelling in Management Communication course. Three datasets are analyzed: leadership storytelling, crisis storytelling, and end-of-course reflective journals. Students' leadership and crisis stories are analyzed using Goffman's (1974) Frame Analysis to examine how they frame managerial situations and construct professional selves through narrative. The reflective journals are analyzed using Reflexive Thematic Analysis (Braun & Clarke, 2021) to explore how students make sense of their professional discourse socialization and identity development throughout the storytelling workshops. The findings reveal that storytelling workshops socialize students into the discourse practices of management by enabling them to frame professional situations, enact managerial identities, and communicate strategically in English. The study contributes to innovative ESAP pedagogy by reconceptualizing language learning as the construction of professional selves through disciplinary communication.

Core Synopsis

The operational truth: in English-medium instruction (EMI) contexts with subject-matter content as the primary goal, language load—not concept difficulty—is the principal barrier to learner independence, and addressing it requires a principled scaffold that separates language from content, builds each systematically, and then integrates them through structured explanation. The presenter's AP Biology course at an English-medium international high school, serving primarily Japanese and Chinese learners, uses the Build→Apply→Explain sequence with a Five-R language scaffold (Replace, Reveal, Repeat, Reposition, Relearn) to move students from teacher-dependent to independently operational in English-medium academic discourse. The ~90% AP Biology pass rate over multiple years validates the instructional design as both linguistically and content-effective. The key insight that often gets buried in EMI discussions: confidence in the language is foundational to content engagement—without it, students opt out of class participation, group work, and assessment production even when they understand the concepts.

The Bottleneck: Language Load, Not Concept Deficit

Cummins' quadrants framework: Content that is cognitively demanding AND context-reduced (Quadrant 4) is the AP Biology default—high cognitive load, abstract concepts, no contextual scaffolding.

Observation: Students who struggle in EMI AP Biology are not failing biology; they are failing to decode, process, and produce English at the speed and register the course demands.

Consequence: Language load misidentified as conceptual difficulty leads to remediation in the wrong domain.

Instructional Architecture: Build→Apply→Explain

Three-phase sequence designed for Quadrant 4 independence:

• Phase 1 — Build: Explicit pre-teaching of the language structures and vocabulary needed to access the content before the content is introduced. Language first, biology second.
• Phase 2 — Apply: Students use the language scaffold to engage with the content—reading, data analysis, lab procedures—with language support available.
• Phase 3 — Explain: Students demonstrate content mastery through English production (oral explanation, written response) using the scaffold, then progressively without it.

Phase 1: Build

Five-R Language Scaffold:

• Replace: Substitute technical terms with accessible synonyms initially; then reintroduce the technical term once concept is secured.
• Reveal: Show the morphological structure of technical vocabulary (e.g., "photosynthesis" = photo- + synthesis); activates morphological awareness.
• Repeat: Multiple exposures across modalities (reading, listening, speaking, writing) to consolidate form-meaning mapping.
• Reposition: Move language from passive recognition to active production; from sentence completion to independent generation.
• Relearn: Targeted re-teaching of specific items that data shows remain weak after initial instruction.

Sentence Frame Examples: "The data suggest that… because…"; "One possible explanation is…"; "In contrast to…, this experiment demonstrates…"

Phase 2: Apply

Students work with AP Biology content (cell respiration, genetics, ecology) using pre-built sentence frames and vocabulary scaffolds.

Key move: Frames are not removed prematurely; they remain available as students develop fluency. Confidence is prioritized over independence in Phase 2.

Assessment in Phase 2: Formative; teacher monitors whether students are using the scaffolds productively or avoiding content production entirely.

Phase 3: Explain

Target: Students produce AP-level written and oral explanations without scaffold support.

Bridge activity: Peer explanation tasks ("teach it back") where students explain a concept to a partner—a high-production, low-stakes context that builds automaticity.

AP exam preparation: Practice prompts using AP free-response format; explicit instruction on the "claim–evidence–reasoning" structure that AP Biology scorers reward.

Non-Negotiables

Confidence as a prerequisite: Without explicit attention to linguistic confidence-building, content engagement collapses in EMI. This is not soft pedagogy—it is a design requirement.

Teacher role: Model the language you expect; use sentence frames yourself; slow your explanatory language and make its structure visible.

Assessment design: Never assess content with a language instrument that penalizes accurate content knowledge expressed imperfectly. Separate language and content assessment until students are at B2+ in the EMI domain.

Context and Results

School context: English-medium international high school; majority Japanese and Chinese learners; AP Biology as a university pathway course.

Outcome data: Approximately 90% AP Biology pass rate (score ≥3/5) over multiple cohorts—well above the global AP average (~66%).

Student feedback: Consistently positive ratings for language scaffold components; students report that explicit language teaching makes content more accessible, not less challenging.

Next Move

• Publish the Five-R scaffold as a usable template for other EMI subjects (AP Chemistry, AP Human Geography, AP Psychology) with subject-specific vocabulary inventories —
• Document the Build→Apply→Explain sequence as a replicable lesson planning framework and share with EMI teacher training programs —
• Develop a language confidence pre-assessment tool for EMI entry-level screening to identify students who need targeted language support before content instruction begins —
• Conduct a retrospective analysis of AP score data correlated with pre/post language scaffold performance to quantify the language-to-content relationship —

Submitted Abstract

What happens when students understand a scientific concept but struggle to explain their thinking in English? For the past decade, I have taught biology, chemistry, and AP Biology to Japanese students, many of whom are learning science through a second language. This experience has shown me that difficulty communicating an idea does not necessarily mean difficulty understanding it. In science classrooms, language can become an invisible barrier between what students know and what they are able to demonstrate. This presentation explores practical approaches for helping multilingual learners move from knowing scientific concepts to communicating their scientific thinking. Drawing from classroom experience, I will share strategies that integrate language development into science learning, including the intentional use of visuals and models, scientific vocabulary in context, structured discussion, sentence frames, collaborative reasoning, and writing-to-learn activities. Rather than simplifying scientific content or lowering cognitive expectations, these approaches provide students with the linguistic tools needed to engage with rigorous scientific thinking. Students are supported in explaining evidence, interpreting data, constructing arguments, asking questions, and communicating their ideas with increasing independence. Participants will consider how language demands influence students’ ability to demonstrate understanding and explore practical strategies they can adapt across subjects and grade levels. The goal is to help multilingual learners move beyond simply knowing the answer to confidently explaining **how they know and why they think it.**

Core Synopsis

The operational truth: language classrooms are among the highest-stakes identity spaces students navigate, and when instructors actively disrupt demographic assumptions—about race, name, authority, and representation—they create conditions for the kind of genuine intercultural learning that language education claims to value but rarely achieves structurally. Margaret Kim and Hosea Baker, an Asian American and African American instructor team, built their course around the Radical Hope Framework, which uses data-driven curriculum design and structured identity exploration to move students beyond stereotype and toward genuine cross-cultural curiosity. Their evidence base—144 student responses across 16 classes over three years—shows consistent gains in willingness to engage with cultural difference when psychological safety is explicitly constructed and instructor vulnerability is modeled. The dissonance is intentional: Hosea Baker's do-rag, worn deliberately in class, is not provocation—it is pedagogy; it forces students to confront their assumptions about who holds academic authority and what a teacher looks like.

Pedagogical Foundations and Icebreaker Strategies for Identity Exploration

Opening move: "Name game" icebreaker that invites students to share the story of their name—origin, meaning, family history.

Margaret Kim's name: Derived from actress Anne Margaret + Korean heritage; her name is itself a cross-cultural artifact.

Hosea Baker's name: His parents deliberately chose a name that would not be read as "black-sounding" (referencing Fryer & Levitt's research on name-based employment discrimination)—a story he shares openly with students to introduce systemic racism as a lived reality, not an abstraction.

Function: The name game establishes that identity is not a private matter to be bracketed from the classroom; it is the content.

The Radical Hope Framework and Data-Driven Curriculum Design

Framework origin: Radical Hope (after Jonathan Lear's philosophical framework applied to Native American cultural survival) repurposed for multicultural language education—the idea that genuine cross-cultural engagement requires hope in the face of difference and potential misunderstanding.

Curriculum data: 144 student responses across 16 classes over 3 years (longitudinal, multi-cohort design).

Key finding: Students in classes with explicit psychological safety construction and instructor vulnerability modeling showed significantly higher self-reported willingness to discuss cultural difference and personal identity.

Interactive Discourse Activities and Classroom Bias Disruption

Devil's advocacy technique: Instructor takes and defends an unpopular or uncomfortable position to model intellectual courage and to demonstrate that difficult conversations are survivable.

Cultural artifact show-and-tell: Students bring objects connected to their cultural background; language is the medium for explaining cultural significance.

Role reversal activities: Students are asked to argue a perspective opposite to their own—building perspective-taking capacity and exposing the rhetorical structure of identity claims.

Establishing Classroom Psychological Safety and Instructor Vulnerability

Psychological safety (Edmondson): The shared belief that interpersonal risk-taking will not result in punishment or humiliation—a prerequisite for genuine cultural disclosure and learning.

Instructor vulnerability as design element: Both Kim and Baker share personal stories of discrimination, identity challenge, and cultural navigation—not as therapy but as curricular content that gives students permission to share their own.

Onsen anecdote: Baker shared his experience of being refused entry to a Japanese hot spring (onsen) due to racial discrimination—a real-world incident that directly addresses the intersection of language learning, intercultural competence, and anti-black racism in Japan.

Challenging Demographic Archetypes and Faculty Representation

Baker's deliberate aesthetic: Wearing a do-rag in the academic classroom is a pedagogical choice that forces confrontation with assumptions about authority, professionalism, and racial representation in higher education.

Kim's experience: As an Asian American woman, she navigates student assumptions that she will be deferential, technically-oriented, and non-confrontational—and her course design explicitly contradicts these expectations.

Faculty representation argument: Students from non-majority backgrounds need to see faculty who look like them; students from majority backgrounds need to have their assumptions about who holds knowledge disrupted.

Audience Discussion on Navigating Extreme Ideologies

Discussion prompt: How do instructors handle students who arrive with hardened ideological positions (far right, far left, nationalist, etc.)?

Response framework: The course design emphasizes process over position—teaching students how to engage with difference rather than which positions are correct; this approach is more defensible and more durable.

Key insight: Psychological safety must extend to students with whom the instructor disagrees; modeling genuine engagement with uncomfortable ideas is more educational than protecting the classroom from them.

Submitted Abstract

“But you’re Japanese.” In many Japanese classrooms, that sentence carries an invisible assumption: if we share a nationality, we must also share a worldview. Students imagine that their peers think like they do, and that their teachers fit neatly into familiar boxes. Yet so much of our students’ identities are built on partial, unexamined understandings of the world around them—and even of the people they see every week in class. If what we assume is never challenged, how can our biases be exposed? How can we work toward a more equitable, inclusive, and just society? These assumptions do not stop at peers. They also shape how students see us, their teachers. When students assume that only “Japanese” teachers can truly understand them, or that non Japanese or racially marked faculty are temporary visitors rather than community members, they reproduce the same unexamined boundaries that keep inequality in place. This leads to a further question: how is inequality reduced when students begin to see their African American male professor and their non Japanese Asian American professor as legitimate members of their communities, rather than distant “others” who only visit for a 90 minute class? This discussion will focus on activities and methods designed to elicit real emotional vulnerability, measurable growth, and qualitative insight. Drawing on classroom data and student reflections, we will discuss how intentional language pedagogy can unsettle “But you’re Japanese” assumptions and open up space for more honest listening, more complex identities, and more just educational relationships.

Core Synopsis

The operational truth: Taiwan's 2030 bilingual policy creates a structural mandate for CEFR B2 English proficiency and EMI-capable graduates—a target current students cannot meet through conventional EFL instruction alone. The presenter's CLIL-driven SDG elective is a direct response: a course that uses the 17 UN Sustainable Development Goals as content spine, flipped classroom methodology as pedagogical architecture, and a deliberate handwriting requirement as an anti-AI measure to ensure authentic engagement. Evidence from student perception surveys shows that the most initially resistant pedagogical choice—the flipped model—becomes the highest-rated for perceived benefit by course end, suggesting that difficulty and value are not opposites but are sometimes the same thing. The strategic risk is that this course sits outside the mainstream of what Taiwanese universities reward (passive lectures, high enrollment, standardized testing), making it vulnerable to institutional marginalization even as it delivers exactly what the 2030 policy demands.

Policy Drivers and Mandates

Taiwan 2030 Bilingual Policy targets:

• 50% of university graduates at CEFR B2 in English by 2030.
• 25–50% of courses delivered via EMI (English-Medium Instruction) at participating universities.

Current gap: Most students enter university below B2; most courses are not EMI-capable; most faculty lack the pedagogical training to teach content in English effectively.

This course's position: A bridge elective—explicitly designed to build the linguistic and content competencies that would allow students to succeed in full EMI courses.

The Gap and the Chosen Mechanism

Identified gap: No existing course bridges EFL (language focus) and EMI (content focus) for Taiwanese university students at B2 entry.

Chosen mechanism: CLIL (Content and Language Integrated Learning) using SDGs as the content domain—globally relevant, intellectually substantive, and politically legitimate for a university elective.

Class size: Approximately 30 students — small enough for discussion-based, flipped instruction; large enough to generate diverse student perspectives.

Course Architecture and Pedagogy

Flipped classroom: Students complete assigned readings, videos, and structured note-taking before class; class time is used for discussion, debate, and student-led facilitation.

Student-led facilitation plans: Each student or pair prepares and delivers a 20-minute facilitation plan on one SDG topic — developing both English presentation skills and content expertise.

Handwriting requirement: All in-class notes and reflection tasks must be handwritten — explicitly stated as an anti-AI measure to ensure authentic thinking and language engagement.

Assessment structure: Participation (facilitation quality), written reflections (handwritten, in-class), a final English presentation on a chosen SDG topic.

Why SDGs Are the Content Spine

17 SDGs / 169 targets: Sufficient scope for a semester's worth of substantive content; each SDG connects to real-world data, policy debates, and cultural contexts students care about.

Motivation engine: Students are more motivated to engage in English when the content matters—climate change, inequality, food security, education access—than when the content is abstract or artificial.

Cross-cultural connection: SDGs provide a shared global framework that allows comparison of Taiwan's context with other countries—directly relevant to intercultural competence development.

EMI Competencies Targeted

Academic vocabulary: SDG-related discourse (sustainability, indicators, targets, stakeholders) with explicit instruction in academic vocabulary strategies.

Discussion and argumentation: Structured facilitation plans require students to generate discussion questions, manage turn-taking, and synthesize contributions—all core EMI skills.

Presentation English: Each student delivers at least one formal English presentation; feedback targets both content clarity and language accuracy.

Critical reading: Pre-class readings include UN reports, academic articles, and journalistic pieces on SDG topics—developing the reading stamina required for EMI coursework.

Evidence from Student Perception

Survey instrument: Post-course perception survey across multiple dimensions (most valuable activity, confidence gains, recommendation to others).

Key finding: Flipped homework initially rated least preferred by students; by course end, it ranks in the top 4 for perceived benefit — suggesting that the most cognitively demanding elements produce the most valued learning outcomes retrospectively.

Implication: Student resistance to demanding pedagogies is not evidence that those pedagogies are wrong; it is evidence that they require more scaffolding and transparent rationale communication.

Strategic Read: Leverage vs. Resistance

Leverage points:

• Direct alignment with national 2030 policy mandates makes the course administratively defensible.
• SDG content has cross-disciplinary appeal that can attract students from science, social science, and humanities tracks.
• Small class size enables the discussion intensity that builds genuine EMI readiness.

Resistance sources:

• Institutional reward structures favor passive lectures, large enrollment, and standardized assessment.
• Faculty workload: Flipped instruction requires substantially more preparation time than lecture delivery.
• Student habits: Taiwanese university culture has historically rewarded passive reception; active facilitation is disorienting for students initially.

Action Items

• Scale the course to two or three sections and compare pre/post B2 proficiency gains using a standardized measure (IELTS Academic, TOEFL iBT) —
• Develop an open-access SDG facilitation plan template library that other instructors can adapt for their own CLIL contexts —
• Document the anti-AI handwriting rationale explicitly in the syllabus and share as a model policy statement for courses concerned about AI-mediated inauthenticity —
• Build a bridge from this course to existing EMI courses at the same institution, creating an articulated pathway for students pursuing EMI readiness —
• Present at EMI/CLIL conferences in Taiwan and internationally to position the course as a replicable 2030 policy implementation model —

Submitted Abstract

This practice-oriented presentation examines the design and implementation of a sustainability-focused CLIL course for B2-level undergraduates at a technological university in Taiwan. The course was intended to facilitate their transition from conventional EFL study to EMI by developing academic English, content-learning strategies, and learner autonomy. Rather than treating English as an object of study, the course aimed to help students use English to access, interpret, discuss, and communicate disciplinary knowledge. Authentic multimodal materials from United Nations reports, international media, and academic videos exposed students to formal registers, specialized terminology, and the information density typical of EMI courses. The course integrated sustainability content, academic English, learner autonomy, and four 21st-century competencies: communication, collaboration, creativity, and critical thinking. Students were expected to comprehend complex texts and lectures, develop advanced and topic-specific vocabulary, evaluate claims and evidence, and express reasoned positions in academic speaking and writing. Equally important, they practiced EMI-related learning behaviors: preparing independently, contributing to discussions, collaborating with peers, and taking responsibility for meaning-making. The learning sequence combined scaffolded worksheets with case studies, student-led discussions, Quizlet vocabulary practice, peer evaluation, and reflective journals. These activities supported a progression from guided comprehension to independent inquiry and knowledge production. Pre- and post-course self-assessments and classroom outputs indicated gains in students’ sustainability knowledge and their ability to understand and communicate content-specific ideas in English. Student-led activities also appeared to strengthen engagement, responsibility, and confidence, although many students continued to value explicit teacher support when encountering unfamiliar concepts and terminology. The presentation argues that autonomy and EMI readiness should not be equated with the withdrawal of teacher support. A carefully scaffolded CLIL course can provide a transitional environment in which students develop academic English, content knowledge, learning strategies, and the mindset needed to move from learning English to learning through English.

Core Synopsis

The operational truth: AI tools now generate grammatically fluent, structurally coherent English that allows students to perform comprehension and production without actually developing either — creating a new pedagogical pathology the presenter names "facade fluency." The course design response is structural: jigsaw peer teaching forces students to become the expert and the teacher in their cohort, eliminating the passive reception strategies that AI makes attractive. Google AI Essentials (replacing ESL textbooks) gives the course contemporary relevance while embedding AI literacy directly into the language learning experience. The two-week expert→teach cycle, with Bloom's taxonomy as the design backbone, creates a learning architecture where passive compliance is not viable: if a student cannot explain the material to peers, the gap is immediately visible — to themselves, their group, and the instructor.

The Problem Mechanism (Diagnosis)

AI facade fluency: Students use AI-generated text to meet surface-level production requirements without developing underlying language competence; the output looks fluent; the learner is not.

Surface compliance vs. deep learning: Traditional ESL tasks (gap-fills, multiple choice, reading comprehension) have low resistance to AI substitution; peer teaching tasks have high resistance.

Passive bypass: Students who are not required to teach or explain can coast on AI support indefinitely without detection under conventional assessment methods.

The Structural Countermeasure (Solution Architecture)

Core logic: Teaching forces encoding — explaining a concept to another person requires deeper processing than receiving an explanation; jigsaw creates a context where every student must teach.

Jigsaw model: Class divided into "expert groups" (each assigned different content) and "teaching groups" (each expert teaches their piece to a mixed cohort). No one can freeload; every student is accountable for both learning and teaching.

Google AI Essentials as curriculum spine: Replaces traditional ESL textbook; students learn about AI tools, applications, and limitations — directly relevant to their lives and embedding AI literacy in language instruction.

Two-Week Cycle: Week 1 (Expert Knowledge Formation)

Activities:

• Script shadowing: Students shadow recorded AI experts (Google AI Essentials videos) to develop oral fluency with technical vocabulary.
• Reader's theater: Students perform scripted dialogues about AI concepts — pronunciation, prosody, and comprehension simultaneously.
• Expert group work: Students become authorities on their assigned AI topic; prepare to teach it to non-experts.

Anti-passivity controls:

• Keyword checkpoints: Before teaching phase, each student must correctly identify and explain 5 key terms from their topic.
• Padlet contributions: Each student posts at least one artifact (definition, example, question) to the shared class Padlet before the teaching session.

Two-Week Cycle: Week 2 (Synthesis and Teaching)

Teaching session: Each expert teaches their AI topic to a mixed group using self-prepared materials (slides, handouts, or Padlet boards); instructor observes but does not intervene.

Pronunciation integration: Cool English Whisper used for pronunciation feedback; students must achieve 80% accuracy threshold before teaching session to ensure intelligibility.

AI Literacy Tech Log: Each student maintains a running log of new AI tools encountered, evaluating each against a structured rubric (accuracy, usefulness, potential bias).

Anti-Passivity Controls

Design principle: Every structural element of the course must create a moment where passive reception is not possible.

Specific controls:

• Jigsaw accountability: Cannot hide behind group work; must teach individually.
• Keyword checkpoints: Cannot progress to teaching without demonstrating expert knowledge.
• Padlet artifacts: Cannot contribute without producing something original (not AI-generated).
• Cool English Whisper threshold: Cannot teach until pronunciation is intelligible.

Outcomes and Validation

Self-reported gains: Students report higher confidence in English speaking and higher engagement with AI tools than in previous ESL courses.

Peer assessment data: Teaching quality consistently improves from Week 2 to Week 4 as students internalize the expert-teach cycle.

AI Literacy development: Tech Log analysis shows increasing sophistication in evaluating AI tools over the course; students move from "cool tool" to "critical evaluator" orientation.

Challenges and Iterations

Challenge 1: Students initially resistant to replacing textbooks with Google AI Essentials — unfamiliar format, no clear "right answers."

Response: Explicit framing of AI literacy as a career skill; student buy-in increased once they saw the content's relevance.

Challenge 2: Pronunciation threshold (80% on Cool English Whisper) was initially stressful for lower-proficiency students.

Response: Multiple attempts allowed; threshold reframed as a preparation check, not a gatekeeping measure.

Challenge 3: Jigsaw teaching quality is uneven — high-proficiency students inadvertently over-teach, reducing peer learning.

Response: Structured teaching protocols introduced (one concept per 5 minutes; pause for questions; never paraphrase the textbook) to level the facilitation quality.

Strategic Principle

Structural resistance to AI-assisted passive compliance is not about banning AI — it is about designing tasks where AI assistance does not substitute for learning. Jigsaw peer teaching is one such design: it creates an authentic communicative demand (teaching a naive audience) that AI cannot fulfill on the student's behalf without exposing the student's own competence gap.

Next Move

• Develop a validated "facade fluency detection" rubric that assesses whether students can explain concepts in their own words without AI support —
• Expand the Google AI Essentials curriculum into a full semester sequence with progressive topic complexity —
• Publish the two-week expert-teach cycle design as a replicable model for ESL/EAP courses at B1–B2 level —
• Pilot with a control group (traditional textbook instruction) to generate comparative data on engagement, AI literacy, and speaking confidence —

Submitted Abstract

In an English for Specific Academic Purposes (ESAP) course on artificial intelligence (AI), students are encouraged to use AI tools as they build language proficiency and disciplinary understanding. A recurring concern is that AI use may allow students to offload the cognitively demanding work central to learning. The jigsaw method can help balance AI-supported learning with sustained cognitive engagement, encouraging students to draw on AI while remaining active participants in processing course content. Each topic followed a two-week instructional cycle. Students first encountered key content and terminology through guided materials and shadowing activities. In expert groups, they studied one key concept of the topic, discussing its meaning and applications, evaluating strengths and weaknesses, drafting quiz questions, and preparing to explain it clearly. In home groups, each student taught their area of expertise, answered peer questions, and helped integrate components into a coherent understanding of the topic. Student feedback supported these aims. Preparing content in advance reduced speaking anxiety, while repeated oral explanation and peer teaching encouraged active participation. Teaching home-group members also helped students understand and remember key terms and concepts. However, some initially found the jigsaw process confusing, and heavy reliance on AI sometimes led students to focus only on their assigned subtasks, limiting their grasp of the broader topic. Several students also felt the cycle moved quickly, leaving little time to absorb new terminology. Building on this, I plan to tighten the instructional cycle to allow more guided, hands-on time with AI tools for instant feedback on pronunciation and writing, while requiring students to attempt writing tasks independently first. This will include techniques such as story-relay, where students take turns extending a shared piece of writing, so they engage with the content before turning to AI, preserving early-stage idea building while still leveraging AI's feedback strengths.

Core Synopsis

The operational truth: learner autonomy in EAP/ESL contexts is not a single variable but a spectrum — and high-autonomy learners, while showing the strongest skill and experience gains on average, are also the most at risk of misaligning their choices with course learning objectives. The quantitative study (N=284 post-course survey; one-way ANOVA across low/moderate/high autonomy cohorts) confirms that autonomy correlates positively with academic skills and lived experience outcomes; but the qualitative subtext is that programs offering autonomy without alignment structures inadvertently create conditions where motivated students work hard on the wrong things. The two structural responses — Quality Talk (discussion rigor with topic choice) and the SPSE Poster Presentation (situation–problem–solution–evaluation) — are the course's attempt to offer genuine choice within a disciplined framework that ensures learning objectives are actually met, regardless of which topic the student has chosen to pursue.

Empirical Signal: Autonomy → Skills + Experience

Study design: N=284 EAP/ESL students; post-course survey assessing self-reported gains in academic skills and lived experience; one-way ANOVA comparing low/moderate/high autonomy cohorts.

Key finding: High-autonomy learners report significantly greater gains in both academic skills and lived English experience than low-autonomy counterparts.

Effect sizes: Varied across skill domains; strongest for discussion confidence and academic writing; weakest for pronunciation and grammatical accuracy.

Interpretation: Autonomy drives engagement, and engagement drives skill development — but only when the autonomous activity is linguistically productive.

Political/Subtext Reality

The tension: High-autonomy learners are self-directed but not always curriculum-aligned. They pursue topics they find interesting but may inadvertently avoid the linguistic challenges the course is designed to develop.

Example: A student who chooses to research and present only on topics they already know well is maximally comfortable but minimally challenged.

Course design obligation: Autonomy structures must create real choice while ensuring that every student, regardless of topic, is practicing the target academic language skills.

Design Response: Structured Autonomy Channels

Quality Talk

What it is: Structured academic discussion protocol where students choose their own topic but must follow a rigorous discussion format (evidence-based claims, uptake of peers' ideas, explicit reasoning).

What it preserves: Topic choice — students pursue content they find compelling.

What it enforces: Discussion quality — the protocol ensures that linguistic rigor is non-negotiable regardless of topic.

SPSE Poster Presentation

What it is: A four-part academic presentation format — Situation (context-setting), Problem (identifying a gap or challenge), Solution (proposing and defending a response), Evaluation (assessing the solution's effectiveness and limitations).

What it preserves: Topic autonomy — students choose their SPSE topic from their own disciplinary or personal interests.

What it enforces: Rhetorical structure — the four-part format develops the argumentation pattern central to academic discourse.

Presentation format: Poster + brief oral explanation; encourages synthesis and distillation rather than comprehensive coverage; post-presentation discussion with audience.

Boundaries

Risk of high autonomy: Alignment failure — students working autonomously but not toward course learning objectives.

Detection method: Formative checkpoints (Quality Talk monitoring, SPSE draft review) to identify alignment failure before summative assessment.

Response protocol: Individual conferences with misaligned high-autonomy students; redirect without removing autonomy; frame the alignment requirement as a professional skill (all autonomous academic work must be accountable to an audience).

Next Move

• Develop an autonomy-alignment checklist for high-autonomy learners to self-assess whether their chosen topic and activities are producing the target linguistic outcomes —
• Expand the SPSE Poster format to include a digital presentation option (recorded video or slide deck) to increase accessibility for reticent presenters —
• Conduct a longitudinal follow-up with high-autonomy learners to assess whether their skill gains are durable and transfer to academic contexts beyond the course —
• Publish the Quality Talk protocol and SPSE rubric as open-access assessment tools for EAP/ESL programs —

Submitted Abstract

As part of Taiwan’s Bilingual 2030 Policy, English for Specific Academic Purposes (ESAP) has become a key component of higher education, preparing students for English-medium instruction (EMI) through disciplinary knowledge and academic English. Because ESAP requires students to integrate disciplinary knowledge with academic English, learner autonomy is essential for effective learning. Although learner autonomy has been widely studied in general English contexts, little research has examined its role in interdisciplinary ESAP settings. This study examined the role of learner autonomy in students’ learning skills and perceptions of learning tasks, instructional materials, and the learning environment in ESAP courses. The study was conducted at a university of technology in Taiwan, where interdisciplinary ESAP courses were offered in disciplines including computer science, management, and engineering. A school-wide questionnaire was administered to students enrolled in these courses. The questionnaire measured learner autonomy, learning skills, and students’ perceptions of learning tasks, instructional materials, and the learning environment. A total of 284 valid responses were collected. Participants were classified into three groups according to their level of learner autonomy (high, moderate-to-high, and low-to-moderate), and one-way ANOVA was conducted to compare the three groups. The findings revealed significant differences across all measured variables. Students with higher levels of learner autonomy demonstrated significantly stronger academic learning skills, such as identifying the main ideas of discipline-related academic texts, than those with lower levels of autonomy. They also reported significantly more positive perceptions of learning tasks, instructional materials, and the learning environment. These findings suggest that learner autonomy is an important factor associated with students’ learning experiences and academic development in interdisciplinary ESAP courses. The study contributes to the emerging literature on ESAP by highlighting the value of fostering learner autonomy and provides pedagogical implications for designing ESAP curricula that better support students’ academic learning and preparedness for EMI.

Core Synopsis

The operational truth: targeted, contrastive pronunciation instruction paired with human-mediated feedback significantly increases learners' self-rated accuracy and confidence, but gains stall at real-life transfer because off-campus practice remains scarce. The course at NTUST centered on three pillars—Mandarin–English contrastive drills, peer/instructor feedback loops, and weekly reflective journals—moved students from a mean of 1.98 to 3.57 across 25 self-report items over nine semesters (n≈213 pairs), with the strongest improvements in pronunciation knowledge and in-class practice opportunities and the weakest in daily-life English use. The instructional frame explicitly aligns with the intelligibility principle (clarity over native-like accent), backed by prior meta-analyses showing strong effects for structured pronunciation training. The subtext: the instructor intentionally de-emphasized automated AI correction to avoid the accuracy–confidence trade-off where ASR feedback boosts error awareness but depresses willingness to speak; human feedback preserved encouragement and participation. The stakes are clear: classroom gains in prosody, segmentals, and confidence will plateau unless practice bridges to authentic contexts; we must redesign the course-to-life pipeline to convert controlled gains into durable communicative competence beyond the classroom.

From "Correction" to "Confidence": The Operating Logic and Gaps

The Pivot to Intelligibility

Thesis: Pronunciation teaching should target intelligibility (clarity/effectiveness), not native-like accent.

Evidence Base: References to Jenkins (intelligibility principle), Saito's Japanese L2 studies, and Lee/Chan/Plonsky meta-analytic findings showing consistent, large effects from systematic pronunciation instruction.

The Instructional Architecture (Three Pillars)

Contrastive Instruction (Mandarin–English):

• Focus on voiced consonants, final consonant deletion, tense–lax vowel pairs, and prosody.
• Rationale: L1-aware contrasts accelerate perception–production mapping for Taiwanese learners.

Peer Feedback and Correction Cycles:

• Recording, peer listening, and iterative correction to expand practice volume and engagement.

Reflective Pronunciation Journals:

• Weekly logging of errors to build metacognitive awareness of rhythm, intonation, and segmentals.

Measurement Framework and Outcomes

Design: 25-item pre/post survey across four domains: self-assessment (overall ability, accuracy, clarity), daily-life communication, classroom communication, and practice opportunities.

Sample: Pre n=238, post n=233, matched pairs n=213 over nine semesters (~32 instructional hours/semester).

Results:

• Aggregate gain: mean 1.98 → 3.57 (significant).
• Effect sizes: varied (.09 to 2.9), with strongest gains in course understanding/practice; weakest in daily-life communication.
• Interpretation: Instruction and in-class feedback effectively grow knowledge and confidence; transfer to off-campus use is limited.

The Confidence–Accuracy Dialectic

Observation: Automated speech feedback can raise accuracy by highlighting errors but may reduce confidence and willingness to communicate due to heightened error salience and "auto-correction masking" that doesn't generalize to human listeners.

Countermeasure: Human-mediated feedback (peer/instructor) used to preserve interpersonal encouragement and mitigate anxiety, aligning with MacIntyre & Gardner and related findings on confidence and WTC.

The Transfer Failure

Phenomenon: Classroom gains do not fully generalize to life; students default to Chinese off campus.

Mechanism: Lack of authentic, structured speaking opportunities beyond class; no enforced or incentivized practice environments in daily contexts.

Risk: Without transfer, improvements remain context-bound, limiting real communicative payoff.

Strategic Implications

• Keep the contrastive phonetics + prosody core; it is the engine of accuracy gains.
• Redesign for transfer: embed authentic tasks, off-campus practice scaffolds, and controlled exposure to non-automated, human interaction to convert classroom gains into real-world competence.
• Calibrate AI use: leverage ASR for drills where confidence risk is low; keep formative evaluation human-led to maintain WTC.

Next Move

• Design and pilot two off-campus "authentic speaking" tasks per semester (e.g., service encounter role-play with community partners; English-only peer meetups) with mandatory participation and rubric-based assessment —
• Integrate a prosody-focused micro-syllabus (stress, rhythm, intonation) with week-by-week drills and performance checks to institutionalize the strongest gain driver —
• Implement a structured peer feedback protocol (checklists for segmentals and prosody, positive-first, two concrete suggestions) to standardize human-mediated encouragement —
• Add an "AI-light" practice track: limited ASR drills for segmental accuracy, followed by human debrief to align error awareness with confidence preservation —
• Establish a transfer metric: track weekly minutes of off-campus English use via logs verified by peer signatures; set a minimum threshold and tie to course credit —

Submitted Abstract

This study examines the effectiveness of an English pronunciation course designed around contrastive phonetic instruction (Mandarin–English vowels and consonants), prosody training, peer feedback, and real-time error correction. Data were collected across multiple course offerings from 2022 to 2026. Each offering was delivered over 16 weeks (32 hours) and integrated three stages of instruction—phonetic knowledge building, guided pronunciation/speaking practice, and error correction—supported by regular classroom recordings, pronunciation journals, and one-on-one instructor consultations. To evaluate learning outcomes, pre- and post-course surveys measured students' self-reported performance across four dimensions: overall self-assessed English/pronunciation ability, communication in daily-life contexts, communication in classroom contexts, and understanding of pronunciation knowledge/practice opportunities. Using student ID matching across administrations, 213 students with valid paired pre- and post-test responses (collected between 2022 and 2026) were identified, and paired-samples t-tests were conducted on each dimension. Results showed statistically significant improvement across all four dimensions (p < .001), with effect sizes (Cohen's d) predominantly in the large range (d = 1.0–2.1). The largest gains were observed in students' understanding of pronunciation knowledge and practice opportunities (pre M = 2.1, post M = 4.2, d ≈ 1.9), indicating that intensive in-class practice and correction substantially strengthened phonological knowledge. Self-assessed pronunciation accuracy and fluency also improved markedly (d ≈ 1.5), reflecting increased learner confidence. Gains in daily-life communication, while still significant and large in magnitude (d ≈ 1.0–1.3), were somewhat smaller than classroom-context gains, suggesting room to strengthen transfer of classroom learning to authentic communicative settings. These findings, drawn from four years of consistent implementation, support the integration of explicit phonetic instruction, intensive practice, peer feedback, and immediate correction in pronunciation pedagogy, and point to increased one-on-one consultation and better-calibrated practice materials as directions for further improving learning transfer.

Core Synopsis

The operational truth: William Sadowski's English fiction and film course at NTUST Language Center faces a foundational problem he names the "sea of nothingness"—a classroom reticence so pervasive that students will not speak, respond, or engage unless a very specific sequence of scaffolding moves creates the psychological safety and cognitive access that makes participation feel possible. His response is a multimodal, drama-infused, project-based curriculum that integrates PBL, drama-based instruction, ZPD scaffolding, dual coding, and digital channels (Spotify podcasts, Instagram) to meet students where they are and pull them toward where they need to be. The empirical result: a pre/post test with approximately 40 students across one semester showed a 19-point score improvement (p=0.001)—statistically significant, practically meaningful. The social media observation—Facebook abandoned as "for old people," Instagram retained as the only platform students actually use—is a reminder that the technology choices in course design are not neutral; they signal whether the instructor is designing for students or for an imagined classroom that no longer exists.

Contextual Background and Instructional Challenges

Teaching context: NTUST (National Taiwan University of Science and Technology) Language Center; English elective for engineering and science undergraduates.

Core challenge: "Sea of nothingness"—students arrive silent, unexpressive, and resistant to participation; not from lack of English but from classroom culture, anxiety, and habits of passive reception.

Student population: Primarily engineering/STEM track; highly technically competent; limited experience with creative, performative, or interpretive English tasks.

Course goal: Develop English fluency, cultural literacy, and communicative confidence through engagement with fiction and film as authentic language and culture artifacts.

Theoretical Foundations and Scaffolding Paradigms

Vygotsky's ZPD (Zone of Proximal Development): Content and task complexity calibrated to what students can almost do—not too easy (no challenge), not too hard (no purchase); with instructor/peer mediation bridging the gap.

Dual Coding Theory (Paivio): Simultaneous verbal and visual processing deepens memory encoding; film and fiction naturally generate dual-coded representations.

Multimodality (Kress): Language learning through multiple semiotic modes (image, sound, movement, gesture) increases both engagement and comprehension.

Drama-Based Instruction (DBI): Embodied, performative engagement with fictional scenarios reduces linguistic anxiety by shifting focus from accuracy to meaning-making and character.

Pedagogical Methodologies and Curriculum Design

PBL (Project-Based Learning → Project-Based): Semester culminates in student-produced projects (not just assignments); each project has an audience, a purpose, and a performance dimension.

Drama-Based Instruction: Students perform scenes, voice characters, role-play scenarios; instructor provides feedback on performance quality and language simultaneously.

Greek mythology role-play: Students adopt mythological personas (Hercules, Athena, Medusa) and interact in English from within the character—a low-stakes, high-engagement route into creative language production.

Talk-show format: Cross-cultural intercultural presentations structured as mock talk-show interviews; students are "guests" discussing their cultural background and perspectives in English.

Film library: 200+ films curated for language and cultural learning value; used for scene analysis, vocabulary extraction, and cultural comparison tasks.

Multimodal Activities and Digital Engagement Channels

Spotify podcasts: Instructor-designed English listening activities tied to published podcast episodes; authentic listening with discussion tasks.

Instagram: The only social media platform that student data confirmed students actually use; used for visual-English content sharing, vocabulary challenges, and short video production tasks. Facebook was discontinued — students described it as "for old people."

Video production: Students create short English videos (1–3 minutes) on cultural or literary topics; combines speaking, scripting, editing, and public presentation skills.

Empirical Evaluation, Student Feedback, and Operational Challenges

Study design: Pre/post test with ~40 students; one semester; test covering reading comprehension, vocabulary, and cultural literacy.

Result: 19-point improvement (p=0.001); statistically significant at conventional threshold.

Student feedback themes: High engagement ratings for drama activities and film analysis; lower engagement for reading-heavy units; strong positive response to social media integration.

Operational challenges:

• "Sea of nothingness" requires sustained patience; early lessons are often met with silence even with scaffolding in place.
• High content complexity (fiction analysis, cultural history) demands significant preparation for both instructor and students.
• Technology platform decisions require annual review; student social media habits shift faster than curricula can adapt.

Submitted Abstract

This study examines the implementation of an interdisciplinary Fiction and Film curriculum for undergraduate and graduate students (CEFR B2-level learners) at National Taiwan University of Science and Technology. Although integrating literary texts and film adaptations into university EFL instruction offers valuable opportunities for language development and critical engagement, many non-English majors struggle to participate in English discussions and to analyze texts beyond surface-level comprehension. To address these challenges, I designed a multimodal instructional model that combines Project-Based Learning (PjBL) and Drama-Based Instruction (DBI) through both classroom-based and digital learning activities. The 16-week course adopted a blended learning approach that combined line-by-line novel readings, film analyses, Live Action Role-Playing activities, student and instructor podcasts, and intercultural language exchanges and interviews. To provide additional opportunities for English interaction outside class, students participated in synchronous online movie club meetings on Google Meet, asynchronous discussions, and peer-shared media recommendations through Facebook, Padlet, and Instagram. To investigate students’ learning experiences and perceived development, I collected student-produced language samples, including short stories, film reviews, and podcast scripts, alongside pre- and post-course surveys. These surveys examined changes in comprehension, writing, and speaking confidence, as well as perceptions of digital and intercultural learning activities. The survey data indicated improvements in reading comprehension, critical engagement with literary texts, and improved writing confidence. Qualitative analysis of student work highlighted increased vocabulary use, more coherent narratives, and collaborative learning. Furthermore, over 93% of participants expressed satisfaction with the intercultural language exchanges and reported reduced anxiety when speaking in public. These findings suggest that combining literary analysis with digital interaction and intercultural communication activities can offer new ways to increase student participation in university EFL literature courses and create learner-centered environments that foster students’ critical literacy and ability to communicate ideas in English. Keywords: Multimodal Pedagogy, Drama-Based Instruction, Blended Learning

Core Synopsis

This session covered two distinct presentations delivered in a shared space — notably a kitchen area of the venue, booked three months in advance rather than the standard six, reflecting the conference's rapid growth. The heat was a recurring challenge, with the main lights kept off to reduce temperature.

The first presentation, by a presenter drawing on Rod Ellis's work and Marcos Benavides (JF Oberlin), focused on Task-Based Language Teaching (TBLT). The core argument: meaning must precede form in principled task design — students should engage authentically before grammar becomes the focus. The presentation outlined four task types applied in a Japanese university context: one-minute impromptu speeches, split reading tasks (each student holds half the information), AI image generation prompt-writing (students write English descriptions for Midjourney/DALL-E to generate images, then evaluate the output), and ranking tasks (students rank items and defend their choices in discussion). The audience engaged with how ranking tasks in particular can generate sustained negotiated interaction.

The second presentation described a COIL (Collaborative Online International Learning) project connecting Nagoya University of Foreign Studies with a partner institution in San José, Costa Rica. The 8-week virtual exchange ran weekly Friday sessions (Japan time) / Wednesday sessions (Costa Rica time). Tools used: Padlet for asynchronous sharing and Google Forms for structured reflection. Three student trajectory cases were presented in detail — Chihiro, Utah, and Rena — illustrating a Dunning-Kruger pattern across the exchange: initial overconfidence in intercultural competence, followed by a mid-exchange crisis of uncertainty, followed by genuine growth in intercultural awareness. Kevin Hodson was mentioned as a key collaborator. A "3D virtual environment" called "Exhaling Cloud" was briefly mentioned as a potential future platform for the exchange.

TBLT: Meaning Before Form

Core principle (Ellis): Tasks should position students as language users before language learners — meaning-first, then formal attention as needed.

Four task types in use:

• One-minute speeches: Students speak for 60 seconds on an assigned topic without preparation — focus on fluency and communicative effectiveness.
• Split reading: Each student reads a different text containing complementary information; must communicate in English to complete the information gap.
• AI image generation prompts: Students write English descriptions for image-generation AI; the visual output provides immediate feedback on the precision and vocabulary of their English.
• Ranking tasks: Students rank items (e.g., survival priorities, qualities of a good teacher) and defend rankings through negotiated discussion — generates sustained, authentic language production.

COIL: Japan–Costa Rica Virtual Exchange

Structure: 8-week exchange; weekly synchronous sessions; asynchronous Padlet contributions and Google Form reflections.

Institutions: Nagoya University of Foreign Studies (Japan) and partner in San José, Costa Rica.

Student cases:

• Chihiro: Initially highly confident about intercultural communication; mid-exchange struggled with actual cross-cultural misunderstandings; ended with more nuanced and humble self-assessment.
• Utah: Showed consistent growth throughout; particularly strong in Padlet contributions and willingness to revise initial assumptions.
• Rena: Most dramatic Dunning-Kruger trajectory — early overconfidence followed by a significant mid-point crisis, then substantial growth in the final weeks.

Key insight: The Dunning-Kruger pattern is pedagogically productive — the crisis of mid-exchange disorientation is not a failure but the moment of genuine intercultural learning.

Submitted Abstract

In task-based language teaching (TBLT), tasks are generally expected to fulfil four key principles. However, the extent to which these principles are realised can vary, suggesting that some classroom activities are more "task-like" than others. This presentation explores how existing tasks were adapted to more closely reflect these four principles. While all four principles will be discussed, particular attention will be given to the importance of ensuring a meaningful non-linguistic outcome, a feature that is often misunderstood or overlooked. Drawing on a range of classroom examples, the presentation will illustrate how carefully designed outcomes can enhance learners' motivation, engagement, and enjoyment while supporting successful language learning. Through these examples, the presenter will demonstrate how relatively small adaptations can strengthen task design, with practical implications for materials development.

Core Synopsis

This session captured the behind-the-scenes reality of running a multi-presenter, multi-room conference when AV infrastructure, scheduling documents, and recording consent frameworks are misaligned — and the human cost of those gaps falls disproportionately on the presenters who are most prepared and most professional. Two presenters, identified as "Justin" and "Philo," navigated a session that involved dongle swapping, incompatibility between projected systems, discrepancies between the Google Form scheduling document and the printed conference booklet, and unclear recording consent protocols. The session is documented here not as a failure but as a learning record: a frank account of what goes wrong in conference logistics when setup time is insufficient, backup equipment is absent, and consent procedures are not communicated in advance.

AV and Technical Challenges

Dongle conflicts: Multiple presenters with different computer systems (Mac/PC, USB-C/HDMI) competing for a single projector connection without a universal adaptor available at the venue.

Projector compatibility: The room's primary projection system could not read one presenter's file format without software that was not installed on the venue computer.

Resolution: Presenters improvised — one used a personal hotspot and screenshared via Zoom to display slides; the other restructured their presentation to be delivered without slides.

Lesson: Every conference room should have a universal adaptor kit (USB-C to HDMI, HDMI to VGA, USB-A hub) and a venue laptop with common software (PowerPoint, Keynote, Google Slides, PDF reader) pre-loaded and pre-tested.

Scheduling Document vs. Printed Booklet Discrepancies

Google Form schedule: The digital scheduling system showed one time slot assignment for Justin and Philo's session.

Printed booklet: The conference booklet, finalized and distributed before late registrant scheduling changes, showed a different room and time.

Impact: Attendees arrived at the wrong room; the presenters' session audience was split; both presenters and attendees lost time to navigation and confusion.

Lesson: Print materials must be locked before scheduling changes are finalized, or a real-time digital correction mechanism (QR code linking to current schedule) must replace or supplement printed materials.

Recording Consent Friction

Protocol gap: The conference had no standardized, pre-distributed recording consent form. Consent was sought ad hoc, at the room level, immediately before sessions began.

Impact: Presenters who objected to being recorded — for various legitimate reasons — had to raise the issue in front of the audience, creating awkwardness. One presenter asked the recorder to stop; the subsequent negotiation consumed several minutes of session time.

Lesson: Recording consent must be collected in advance (at registration or via pre-conference email) and must specify: who is recording, how the recording will be used, and how long it will be retained. Room-level consent is too late and too coercive.

Key Operational Learnings

For future conference organizers:

• AV kit standard: Universal adaptor set, venue laptop with common presentation software, HDMI extension cable, spare clicker — every room, every session.
• Schedule freeze: Print materials locked 72 hours before conference; any late changes communicated only digitally, with a prominent QR code on printed schedules linking to the live version.
• Consent protocol: Pre-conference recording consent form distributed at registration; session-level consent collection only for unexpected situations.
• Setup time: Minimum 20 minutes between sessions for full AV and presenter setup; 10 minutes is insufficient when system incompatibilities arise.
• Point of contact: Each room should have a designated volunteer or staff member whose sole responsibility is AV and schedule support during the conference day.

Submitted Abstract

Virtual reality (VR) has been increasingly adopted to enhance vocabulary acquisition, a foundational and critical component of foreign language learning. However, little research has examined how VR supports the learning and retention of vocabulary across different parts of speech. Addressing this gap, the present project investigated the effects of spherical video-based virtual reality (SVVR) on English as a Foreign Language (EFL) learners’ vocabulary learning and retention of nouns, verbs, and adjectives, as well as their cognitive load. Forty college students were randomly assigned to either a control group receiving traditional paper-based instruction or an experimental group engaging in SVVR-based learning. Four immersive VR scenes were developed, integrating target vocabulary items with their parts of speech, definitions, example sentences, and virtual avatars. Students in the experimental group interacted with the SVVR environments using mobile-rendered head-mounted displays (HMDs), whereas those in the control group learned the same content through printed handouts. Data were collected through vocabulary pretests and posttests, delayed retention tests, and semi-structured interviews. Although no significant difference was found between the two groups on immediate posttest performance, the SVVR group demonstrated significantly higher vocabulary retention, indicating a positive effect of SVVR on long-term memorization. Among the different parts of speech, noun retention was the highest, followed by adjectives and verbs. Interview findings further revealed that the highly immersive and interactive learning environment facilitated the retention of nouns and adjectives, while the incorporation of virtual avatars particularly supported verb learning. Additionally, the experimental group reported significantly lower cognitive load than the control group, suggesting that SVVR may reduce learners’ cognitive demands when processing concurrent visual and textual information. Based on these findings, pedagogical implications for VR-assisted vocabulary instruction are discussed.

Core Synopsis

The operational truth: 360° scene-based virtual reality (SBVR) — using accessible, non-specialist tools like DeliteX Edu (formerly CoSpaces) and Google Street View — produces a measurable, statistically significant advantage in delayed vocabulary recall compared to conventional instruction, without requiring the cost or equipment of full immersive VR. The study (N=40 TOEIC-matched elementary-level Taiwanese EFL undergraduates, between-groups design) found no significant immediate post-test advantage for SBVR, but a significant delayed-test advantage and significantly lower cognitive load during learning in the SBVR condition. The mechanism: 360° scenes (Taipei MRT, museum, aquarium, airport) provide spatial anchors that reduce the cognitive overhead of vocabulary learning, allowing more processing capacity for encoding. Adjective anchoring ("automatic, prehistoric, spacious") and verb enactment via avatar movement are the two instructional mechanisms most likely responsible for the retention advantage. The accessibility argument is the course's strongest long-term claim: SBVR is actionable for practitioners without VR labs, specialized hardware, or large budgets.

Pedagogical Gap VR Targets

Vocabulary retention problem: EFL learners encounter vocabulary in decontextualized lists or minimally contextualized sentences; retention is shallow and decay is rapid.

Standard instruction limitation: Even communicative tasks often present vocabulary in low-sensory, low-spatial environments that provide few memorable contextual anchors.

VR's theoretical promise: Immersive, spatial environments should provide richer contextual encoding; multiple sensory channels (visual, proprioceptive, spatial) should support deeper memory traces than text alone.

The Accessibility Pivot: SBVR Instead of Full VR

Full VR constraint: Requires headsets, dedicated hardware, lab space, and specialist technical support — prohibitive for most EFL programs.

SBVR solution: 360° photographic scenes displayed on standard computers or tablets; students navigate with mouse/trackpad; no headset required.

Tools used:

• DeliteX Edu (formerly CoSpaces): Platform for creating interactive 360° educational environments with embedded vocabulary tasks, avatar controls, and teacher dashboards.
• Google Street View: Real-world 360° photography used for authentic scene exploration; students "enter" real spaces (Taipei MRT, international museum) and complete vocabulary tasks within them.

The Experimental Spine

Design: Between-groups, TOEIC-matched; N=40 elementary-level EFL Taiwanese undergraduates.

• Group A (SBVR): Vocabulary instruction embedded in four 360° scenes (Taipei MRT station, natural history museum, aquarium, international airport).
• Group B (Control): Same vocabulary items taught via conventional instruction (definitions, example sentences, matching tasks).

Assessment points: Immediate post-test (within 24 hours); delayed post-test (one week later); cognitive load self-report scale.

Vocabulary target: 20 items across four semantic domains aligned with the four scenes; mix of adjectives (automatic, prehistoric, spacious, translucent) and verbs (navigate, exhibit, preserve, depart).

Results: What Moved, What Didn't

Immediate post-test: No significant difference between SBVR and control groups — both acquired vocabulary at similar rates initially.

Delayed post-test: SBVR group significantly outperformed control group on retention of target items (p < .05); effect size moderate to large.

Cognitive load: SBVR group reported significantly lower cognitive load during the learning phase despite engaging with a more complex environment — counterintuitive but consistent with the cognitive load theory prediction that meaningful context reduces extraneous load.

Diagnosis: Why Retention Wins

Spatial encoding advantage: 360° scenes provide location-specific memory anchors; vocabulary items are mentally "placed" in the MRT station or museum, providing retrieval cues unavailable in list-based learning.

Adjective anchoring: Descriptive adjectives were linked to perceptually salient scene properties (spacious airport hall; prehistoric dinosaur exhibit); the perceptual match strengthens the semantic encoding.

Verb enactment: Avatar movement in the scene allowed students to "perform" action verbs (navigate through the MRT, exhibit an artifact) — enactment effects on memory are well-documented in embodied cognition research.

Reduced extraneous cognitive load: Authentic spatial context removes the need for abstract imagination that decontextualized vocabulary learning requires; the scene does the contextualizing work, freeing cognitive resources for encoding.

Limitations

Small N (40): Between-groups design with only 40 participants; effect sizes may not replicate at larger scale.

Single institution: NTUST context may not generalize to different learner populations or L1 backgrounds.

Self-report cognitive load: Single-item self-report is a rough proxy; more granular cognitive load measures (pauses, error rates, secondary task performance) would strengthen the causal claim.

No productive use test: Post-tests measured receptive recognition; whether SBVR advantage extends to productive use (writing, speaking) was not tested.

Strategic Leverage for Practitioners

Immediate applications:

• Use Google Street View to create vocabulary contextualization activities for any topic with a physical location dimension (airport English, hospital English, restaurant English).
• Use DeliteX Edu/CoSpaces to embed target vocabulary in interactive scenes with clickable definitions and pronunciation models.
• Focus SBVR design on adjectives (perceptual anchoring) and action verbs (enactment) for maximum retention advantage.

Cost argument: Google Street View is free; DeliteX Edu has a low-cost educator license; no VR hardware required; implementation is within reach of any instructor with a computer lab.

Action Items

• Replicate the study with a larger sample (N=80+) across multiple institutions to validate the delayed-test advantage and cognitive load findings —
• Extend the vocabulary target to include collocations and phrasal verbs, not just single items, to test whether SBVR's contextual advantage extends to multi-word units —
• Design a productive-use post-test (speaking or writing) to determine whether SBVR retention advantage transfers from recognition to production —
• Develop an open-access SBVR lesson template library (5 scenes × 20 words each) for common EFL vocabulary domains —
• Present findings at CALL and vocabulary acquisition conferences to build awareness of SBVR as an accessible alternative to full-immersion VR —

Submitted Abstract

Virtual exchanges like COIL offer equitable and affordable opportunities to internationalize the curriculum while providing sustained intercultural interaction across borders without the need for physical mobility (Blake, 2013; Ceo-DiFrancesco, 2015; O’Dowd, 2023; Rubin, 2023). Despite the growing literature on COIL, assessment often focuses on overall outcomes, potentially obscuring individual differences in what students report learning from their intercultural encounters. This study follows Japanese university students participating in an eight-week COIL virtual exchange with university students in Costa Rica. Data were collected through weekly reflections, online discussion-board posts, and post-project and post-then-pre retrospective questionnaires. Drawing on dynamic approaches to interculturality (Dervin, 2016) and the notion of developing an “intercultural stance” through telecollaboration (Ware & Kramsch, 2005), the analysis follows individual students longitudinally rather than treating intercultural development as a single outcome. It examines how students’ understandings, communication, and responses to cultural difference evolved across weekly Zoom encounters, allowing diverse learning trajectories to emerge that may be missed in group-level results. Initial analysis suggests different but overlapping pathways of learning. For some students, interactions prompted reconsideration of cultural assumptions and greater attention to the contexts behind cultural differences. For others, communication difficulties led to greater awareness of interactional strategies such as active listening, follow-up questions, and persistence. For still others, developing interpersonal relationships with Costa Rican peers reduced perceived distance and contributed to more complex understandings of Costa Rica and its people. These pathways were not necessarily linear, as broad generalizations remained alongside more nuanced understandings. Findings from this small-scale study suggest that the value of COIL may lie less in producing a uniform increase in intercultural competence than in creating repeated opportunities for intercultural, communicative, interpersonal, and reflexive learning. Implications for designing COIL interactions and reflection activities to support these different pathways will be discussed.