L2 writing

This review paper explores the potential benefits and limitations of Grammarly, a widely used Automated Corrective Feedback (ACF) tool, in the context of Second Language (L2) writing. Specifically, while Grammarly offers advantages such as improved grammatical accuracy, reduced writing anxiety, and support for self-directed learning, it also presents several concerns, including unclear feedback, cognitive overload, and a limited focus on content and organisation. Drawing on empirical research studies published primarily between 2021 and 2024, this review suggests that Grammarly can serve as a valuable supplementary tool when used strategically. Its effective integration into the classroom requires active collaboration between teachers and students, supported by explicit instruction and guidance. More importantly, students are encouraged to engage critically with the Grammarly feedback to develop their writing skills. The review concludes with recommendations for future research in this area.

This study is the first to explore the neural correlates of cognitive processes proposed in models of writing, combining fMRI scanning with keystroke logging to bring together insights from neuroscience and behavioural research. We focused on a key writing mechanism, pausing; specifically, we examined how pause location (within words, between words/sentences) may relate to neural processes. The participants were 21 intermediate-level Japanese university students of L2 English, who carried out four decision-making writing tasks in English. While the participants were writing, we recorded their brain activity through fMRI scanning and their writing behaviours with keystroke logging. As predicted, pauses at lower textual units (within and between words) were associated with increased activation in language-related brain regions, whereas pauses at higher textual units (between sentences) were related to greater brain activity in planning-related brain areas. In addition, the study revealed heightened brain activity in sensory and motor regions for within-word pauses, relative to pauses between words and sentences. Given the convergence between our behavioural and neural data, the results suggest that the cognitive processes posited in models of writing correspond to identifiable neural correlates.

While existing research has largely focused on students' perceptions of GenAI tools, there remains limited empirical evidence of how EFL university students engage with these tools during real-time writing processes, particularly within Asian higher education contexts. Addressing this gap, the present study investigates how Vietnamese EFL university students use GenAI tools when completing academic essay tasks and examines the functions these tools serve across different stages of writing. Drawing on 39 screenrecorded writing sessions from students enrolled in academic English courses, the study adopts a qualitative, process-oriented approach to analyse GenAI use during composition. The analysis showed that students used GenAI across three interrelated domains of writing: content construction, where GenAI supported idea generation, summarisation, and content development; language mediation, where it assisted with vocabulary, translation, paraphrasing, and proofreading; and process regulation, where it provided feedback, explanations, and guidance during drafting and revision. The findings further revealed several pedagogical and ethical issues, including uneven evaluative engagement with GenAI output, blurred authorship boundaries, and ethical ambiguity when GenAI was used to generate or shape writing content and structure. While some students engaged with GenAI reflectively, others incorporated generated, translated, or paraphrased content with limited visible transformation, raising questions about authorial control, acceptable assistance, and students' engagement with core writing processes.

The present study was an attempt to investigate the impact of collaborative writing of Iranian learners of English as a Foreign Language (EFL) on improving their argumentative writing skills. Having administered a writing test taken from TOEFL as a pre-test, 69 EFL students from intact writing classes at the University level were assigned into three groups: collegial collaborative, sequential collaborative, and no collaboration groups. At the end of the treatment, which took 12 sessions each for 80 minutes, all groups received the post-test. Findings made it clear that learners benefit equally well from either collegial or sequential cooperation. However, no collaboration led to poor argumentation skills in the writings of the participants. In other words, the collaboration of the types analyzed can potentially develop the reasoning and argumentation skills among EFL students. The findings of the present study can provide EFL teachers with an understanding of the application of different types of collaborative writing to create a positive classroom environment.

This paper tries to investigate different types of hedges and the functions of hedges used in short
story, Friend (1986), written by Tariq Rahman. The present research is qualitative in nature and is based on an
interpretivist paradigm. The short story, Friend (1986), by Tariq Rahman is taken as the source of data. The
theories of hedges and its functions, proposed by Coates (1996), and the types of hedges proposed by Salager
Meyer (1997) are used as a theoretical framework in the present study. The researchers have selected five types
of hedges from Salager-Meyer (1997) classification.  This paper concludes that the hedges which occur most
frequently in the short story is modal auxiliary verb which occurred with the frequency of 9(36%), then the ‘if’
clauses which occurred with the frequency 6(24%), other types such as modal lexical verbs and introductory
phrases occurred with same frequency 4(16%), then compound hedges with frequency 2(8%). There are three
functions of hedges out of four found out by the researchers, these functions are; expression of doubt and
confidence, searching for the right word, and sensitivity to others’ feelings.

Automated Writing Evaluation (AWE) tools have emerged as potential pedagogical aids, leveraging technology to support students' writing. Despite their growing presence, research on students' perceptions of AWE in higher education remains limited. This study examines the autonomous use of AWE tools by Moroccan English as a Foreign Language (EFL) students. A closed-ended questionnaire was administered to 61 EFL students to collect quantitative data on their independent use of AWE tools, motivations for adoption or avoidance, usage patterns, and satisfaction. Statistical analysis revealed that most students view AWE feedback favorably, recognizing its benefits in vocabulary, spelling, and punctuation. These findings align with previous studies highlighting the effectiveness of AWE tools in enhancing writing skills. Insights from non-users highlight barriers to adoption. The findings suggest that integrating AWE tools into writing instruction improves students' writing proficiency while fostering critical thinking and self-editing skills.

Automated Writing Evaluation (AWE) tools have emerged as potential pedagogical aids, leveraging technology to support students' writing. Despite their growing presence, research on students' perceptions of AWE in higher education remains limited. This study examines the autonomous use of AWE tools by Moroccan English as a Foreign Language (EFL) students. A closed-ended questionnaire was administered to 61 EFL students to collect quantitative data on their independent use of AWE tools, motivations for adoption or avoidance, usage patterns, and satisfaction. Statistical analysis revealed that most students view AWE feedback favorably, recognizing its benefits in vocabulary, spelling, and punctuation. These findings align with previous studies highlighting the effectiveness of AWE tools in enhancing writing skills. Insights from non-users highlight barriers to adoption. The findings suggest that integrating AWE tools into writing instruction improves students' writing proficiency while fostering critical thinking and self-editing skills.

Writer-reader interaction is considered a crucial component of a well-written academic text, largely achieved through the use of metadiscourse resources. This paper investigates the frequency and range of 'attitude markers', i.e. one of the five sub-categories of interactional metadiscourse resources (Hyland and Tse 2004) in two learner corpora composed of Czech (L2) and native speaker (L1) university students' academic texts. The survey has shown that research-oriented attitude markers are preferred over topic-oriented ones in both corpora, with the former being used mainly to evaluate the writer's own research and its findings. Attitudinal adjectives were identified as the most frequent and most varied category in both corpora, followed by attitudinal nouns, adverbs, and verbs in their respective order. The results also indicate some interesting differences, for example, that Czech students, unlike L1 students, tend to overuse attitudinal adverbs in the sentence initial position. By highlighting the similarities and differences in the use of attitude markers in L2 and L1 student writing, the study may be of value to both students and their tutors and help to raise students' awareness of the dialogic and interactional character of academic texts.

Foreign language instructors inconsistently implement instructional practices to support Chinese high school students’ learning of oral English as a second language (ESL). The purpose of this basic qualitative study was to explore the perceptions of foreign language teachers working in the Hangzhou region of China, regarding their use of instructional strategies to support Chinese high school students’ learning of oral English language. Sato and McNamara’s general-purpose second-language oral competency theory, which suggests that instructional practices can be evaluated in terms of formal and interaction components, served as the basis for the conceptual framework to examine foreign teachers’ perspectives regarding instructional practices related to lesson plans, classroom management, and extrinsic pressures created by tests and assessments. Research questions were designed to obtain foreign language teachers' perceptions concerning implementation of instructional practices. Data were obtained from interviews with eight purposively sampled foreign ESL instructors teaching at a public school and learning center. Inductive, deductive, and NVivo coding were used to analyze data. Themes included preparation methods for ESL instruction, classroom management strategies for courses and individual learners, and ways institutional-related educators prepare students for external demands for success. Instructors perceive that their classroom management methods are strong; however, practices to improve student engagement and anticipation of classroom challenges could enhance outcomes.
This study could affect positive social change through creation of best practices to improve ESL education in public schools and learning centers in China as well as promote proactive foreign ESL teacher strategies for classroom management.

This paper examines the gaps between the objectives of the curriculum and actual practice in the English Language Teaching (ELT) in Bangladesh. The English curriculum used in the country is based on Communicative Language Teaching (CLT), but in most cases, the lessons are conventional and test-oriented. This restricts learners’ ability to achieve communicative competence. Surveying 40 secondary and higher secondary English teachers and 107 pupils through structured questionnaires, and semi-structured interviews with 10 teachers, constituted a mixed-methods approach to the survey.

The quantitative data reveal that very often communicative activities are not used at the classroom. The qualitative data, on the other hand, reveal numerous interconnected problems, including pressure to perform well on tests, a high number of classes, teacher training, and low student competency. These variables make it harder to put CLT into practice and keep the gap between the planned curriculum and what really happens in the classroom.

The research shows that the relationship between the intended curriculum and students’ classroom activities remains tenuous and systemic. According to it, the only solution to this issue is significant and includes incorporating communicative skills into the assessment system, providing intensive, practical teacher training, and aligning Communicative Language Teaching (CLT) with current school practices in the region.

 

Keywords: Bangladesh, classroom practices, Communicative Language Teaching, English Language Teaching, and implementing curriculum into practice.

The growing integration of Generative AI (GenAI) tools in language education presents new opportunities for enhancing learner engagement. However, empirical evidence on their effectiveness remains limited, particularly in project-based and collaborative contexts. This study examined the impact of integrating GenAI tools into collaborative Project-Based Language Learning (PBLL) on EFL undergraduates' engagement and agency. Using a quasi-experimental pre-/post-test control group design and a sequential mixed-methods approach, the study combined quantitative questionnaire data with qualitative reflections. Results showed significant improvements across emotional, behavioural, cognitive, and agentic engagement dimensions, with the most substantial gains in emotional and agentic domains. Student reflections supported these findings, revealing increased motivation, confidence, participation, and a stronger sense of autonomy. Emotional responses evolved from initial uncertainty to enthusiasm, while students demonstrated leadership and self-directed learning. Participants valued the combination of GenAI with peer collaboration and teacher support, highlighting the importance of balancing technology with human interaction. The study underscores the potential of GenAI-driven PBLL to create engaging, learner-centred environments and offers practical insights for educators seeking to integrate AI meaningfully. It calls for inclusive, autonomy-supportive learning conditions that scaffold the use of GenAI while fostering reflective, active, and personalized language learning experiences.

This study investigates the progression of syntactic complexity in children's story texts from the BOOKR Class digital reading application across nine graded levels. Using a quantitative approach, syntactic complexity was measured through sixteen indicators grouped into four dimensions: Subordination, Length of Production, Coordination, and Nominalization. Descriptive statistics, one-way ANOVA, and Games-Howell post hoc tests were employed to examine differences across levels. The results reveal statistically significant variation in all indicators, indicating a systematic and cumulative increase in syntactic complexity as reading levels advance. While adjacent levels often show non-significant differences, substantial contrasts emerge between distant levels, reflecting a gradual and developmentally appropriate progression. Minor non-linear fluctuations further suggest dynamic redistribution of syntactic features rather than regression. Overall, the findings provide empirical evidence that BOOKR Class implements a linguistically principled leveling system and demonstrate the usefulness of syntactic complexity measures for evaluating graded digital reading materials designed for young language learners.

As one of the pioneering applications of machine learning for quantitative data analysis in applied linguistics, this study investigates the predictive power of linguistic features on English as a Foreign Language (EFL) writing quality. To overcome the limitations of traditional linear regression, specifically multicollinearity and non-normal data distributions, we employed a Multilayer Perceptron (MLP) neural network to analyse 96 essays from Chinese secondary school students. The model evaluated the relative predictive importance of various lexical, syntactic, fluency, and accuracy features. Demonstrating exceptional fit, the neural network accounted for approximately 80% of the variance in overall writing scores with no evidence of overfitting. Sensitivity analyses revealed that lexical diversity was the most robust predictor of writing quality (100% normalised importance), followed by grammatical accuracy (60.4%). Conversely, syntactic complexity, text length, and lexical sophistication exhibited comparatively minimal influence. These findings underscore the critical role of vocabulary diversity in evaluating adolescent EFL writing, while successfully establishing neural networks as a powerful, innovative methodological tool for future applied linguistic research.

Artificial intelligence (AI) is reshaping academic writing instruction. However, there is limited evidence on how students actually interact with AI and how pedagogy can guide ethical and reflective engagement. Piloted in a first-year undergraduate English language and literacy unit at an Australian university, this paper presents the design and trial of curriculum-aligned, task-based writing materials comprising an AI-assisted guide with structured prompts to support essay preparation with AI. Analysis draws on AI-interaction logs from a case study, complemented by a survey of 207 students. Thematic coding and interactional move analysis of AI-interaction logs observed four recurrent moves-clarification, uptake, adaptation, and rejection-in which the learner negotiated AI feedback. Survey data indicated that while students viewed AI as useful for efficiency and confidence, concerns about plagiarism and over-reliance persisted. This study offers early insights into how structured scaffolds can support ethical, reflective engagement with AI and inform future research.

As artificial intelligence (AI) use in English academic writing instruction has increased, this scoping review reviews empirical studies in higher education. However, the studies differ in theory use and teaching practice. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) protocol and the Population-Concept-Context (PCC) framework, a systematic Web of Science (WoS) Core Collection search yielded 20 eligible peer-reviewed studies published post-2020. Most studies were carried out in English as a Foreign Language (EFL) contexts, aiming to improve writing performance and learner motivation via AI-assisted tools. Common platforms encompass automated writing assessment systems. Theoretical frameworks were frequently limited to self-efficacy and scaffolding, with scant consideration for cognitive load, sociocultural adaptation, or ethical governance. Methodologically, most studies relied on short-term experimental designs, lacking longitudinal or classroom-based rigor. Future research should employ mixed-methods and longitudinal designs, include diverse learner demographics, and prioritize teacher preparation and ethical AI literacy to guarantee sustainable and equitable implementation.

The rapid emergence of generative artificial intelligence (GAI) tools has disrupted student writing practices (Barrot, 2023), highlighting the need to understand how English language student writing and GAI writing differ. With the inefficacy of GAI-generated text identification tools well-established (Grassini, 2023), it is necessary to examine linguistic differences between English student-written and AI-generated writing samples. This research supports two complementary objectives: facilitating the identification of student-generated versus GAI-generated writing and aiding the development of Auto-Peer (McCarthy et al., 2021), an established automated writing evaluation tool developed in the UAE
and currently being adapted for the Japanese EFL context as part of ongoing government-funded research (Talandis Jr. et al., 2025).
We compare first drafts of 28 argumentative (“for and against”) essays produced by university-level English learners at the A2 CEFR proficiency level, specifically second-year students in Japan and the UAE. This human-authored data was created through a flipped interaction (Schmidt et al., 2023) activity, where students respond to GAI-generated questions, resulting in initial drafts compiled without textual corrections. As a comparative baseline, fully AI-generated texts were also produced from similar prompts using three prominent tools: Gemini 2.0 Flash, ChatGPT-4-turbo, and DeepSeek-V3, expanding on Mizumoto et al. (2024). Texts were computationally analyzed using Coh-Metrix (McNamara et al., 2014)
and Gramulator (McCarthy et al., 2012) to identify differences in vocabulary, sentence complexity, text cohesion, and overall argumentative structure. This comparison helps identify culturally specific patterns among L2 writers while distinguishing these from purely AI-generated content. Results will guide the linguistic and pedagogical refinement of Auto-Peer, addressing the specific needs and challenges faced by foundational EFL and ESL students, thereby contributing to a broader
understanding of how GAI may be leveraged to ethically and effectively support authentic L2 academic writing. Our findings also have practical applications in distinguishing student-generated from GAI-generated
text.

(References available)

This poster presents an ongoing government-funded research project (JSPS Kakenhi Grant #JP24K04103) that is adapting Auto-Peer (McCarthy, et al., 2021), an automated writing evaluation and feedback tool, for Japanese EFL classrooms. The study explores how student writing data can guide Auto-Peer’s development as a pedagogical aid for building feedback literacy and ethical AI use. Specifically, this poster reports preliminary findings showing how linguistic insights from Japanese learner writing and generative AI output reveal a path toward integrating GAI within Auto-Peer’s workflow. By leveraging systematic differences in how leading LLMs produce prose, Auto-Peer aims to generate linguistic resources that support foundational learners—recognizing GAI fluency not as an endpoint, but as a developmental milestone that can accelerate authentic writing growth and free teachers to focus more on building higher-order skills.

We introduce a project to tailor the automated writing evaluation (AWE) tool Auto-Peer for English academic writing classrooms in Japan that includes a prompt feature to guide productive and ethical large language model-based generative artificial intelligence tool use. The project is titled Feedback Literacy and AI Ethics: Leveraging Auto-Peer for Productive Interaction with Generative AI Tools in L2 Writing Education in Japan (JSPS KAKEN grant JP24K04103). Auto-Peer currently offers immediate, personalized automated feedback, fostering student feedback literacy, but is optimized for higher-level ESL learners (Thomas & McCarthy, 2023). In Japan, where many English learners are at foundational writing levels (Mizusawa, 2020), it is necessary to update its algorithms to effectively serve students' needs (McCarthy et al., 2021). Here we outline our plans to develop Auto-Peer by collecting quantitative and qualitative data through a mixed-methods quasiexperimental design to assess its effectiveness in improving English writing skills and by developing the prompt library feature for generative AI tools like ChatGPT, Claude, and Gemini. By documenting the project's initial objectives, desired outcomes, and expected challenges, we establish the foundation for the research and development of Auto-Peer and its role visa -vis the ethical and effective use of generative AI tools in L2 writing education. This manuscript outlines a project aimed at adapting the automated writing evaluation (AWE) tool, Auto-Peer, for L2 English academic writing classrooms in Japan and integrating generative AI prompt engineering into the software 1. The project, titled Feedback Literacy and AI Ethics: Leveraging Auto-Peer for Productive Interaction with Generative AI Tools in L2 Writing Education in Japan (JSPS KAKEN grant JP24K04103), details the research plan, foreseeable challenges, and limitations, establishing a public record of researcher perspectives at the project's outset. As the project evolves, our perspectives will serve as benchmarks for tracking inevitable and unexpected challenges, consistent with Open Science Framework principles of transparency, collaboration, and reproducibility (Foster & Deardorff, 2017). We begin by discussing the necessity of AWE tools and how this research contributes to their development. We conclude with reflections on the project's goals and its role in maintaining academic standards while integrating technology in L2 writing education.

There is very little research on writing teachers' perceptions of multilingual writers' needs and the teachers' preparation in working with L2 writers in first year composition courses. This present study was conducted to identify the extent to which first year writing teachers recognize the presence and needs of multilingual writers and the preparation those teachers need to have to better work with multilingual writers. A total of 22 writing teachers completed a 21-item survey. The 22 participants were teaching English writing courses at the university. The participants include graduate teaching assistants with different educational backgrounds, currently pursuing their master degree in English and full-time and part-time adjunct teachers with master's or Ph.D. degrees. The participants completed a survey to identify: 1) to what extent do teachers recognize the needs of multilingual writers and 2) what preparation (professional development needs) that those teachers need to better work with multilingual writers. The research reveals that teachers recognize multilingual writers' presence and needs, and suggest external professional preparation opportunities to address the unique needs of L2 writers. The findings of the study serve as an informative source for college writing programs to identify the existing shortcomings and devise administrative and training plans to address them.

The American University of Sharjah (AUS) Writing Center, one of the first writing centers in the Gulf region, supports a multilingual student body in the transnational context of the United Arab Emirates. The profile gives an account of the Center's history, peer-tutoring program, tutor-training course, and Writing Fellows initiative, with particular attention to how the Center adapts Western writing center models to local linguistic and cultural realities through these practices. It also describes strategies that draw on tutors' multilingual and intercultural competence, such as code-switching, politeness strategies, and attention to affect. Current challenges include navigating generative AI, countering the perception of the Center as remedial, and supporting tutor well-being. The profile concludes by reflecting on how the AUS Writing Center contributes to transnational writing center scholarship and offers insights for writing centers working with culturally and linguistically diverse student populations.

This research explores the guided experimental implementation of ChatGPT as a tool for developing teachers' skills in teaching English. The intervention involved 24 in-service English as a Foreign Language (EFL) teachers who engaged in crafting activities and assessments using researcher-designed prompts. Utilizing a mixed-methods approach, the researchers assessed the participants' ChatGPT integration proficiency through a custom-designed assessment tool aligned with the technological pedagogical content knowledge framework. The eight-week intervention introduced educators to various applications of ChatGPT in EFL teaching, including lesson planning. A quantitative analysis revealed statistically significant improvements in the teachers' ChatGPT integration proficiency across all measured dimensions. The qualitative findings highlight the perceived benefits, challenges, and future prospects of ChatGPT in EFL education. While the practical significance of the improvement was modest, the results suggest that the guided integration of generative chatbots can bolster teachers' ability to leverage this technology appropriately. This study contributes to the limited body of empirical research on integrating large language models into teaching and offers insights into the practical applications and challenges of using ChatGPT in EFL contexts.

Cohesion is important in writing because of its role in writing quality development (Abdi Tabari & Johnson, 2023) and in making texts comprehensible (Ghasemi, 2013). It can differentiate between proficient and deficient writing (Crossley et al., 2016b, 2019), demonstrating its use as an index of language proficiency. Despite being an index of proficiency, little is known about cohesion use in second-language contexts (Abdi Tabari & Johnson, 2023; Crossley, 2020; Crossley et al., 2016a). To address this, this study investigated the use of connectives, givenness, and lexical overlap in essays from the Philippines, Singapore, and China available in a corpus. A freely available text analysis tool was employed to compute the cohesion values of the essays based on the three indices. Although the relationships found were generally weak, the quantitative analysis revealed that connective use was in line with the literature results, while data on givenness and lexical overlap contradicted previous findings. Chinese students also used more connectives, givenness, and lexical overlap than Filipinos and Singaporeans. These findings can be traced to Holliday's (1999) large and small cultures. Pedagogical implications and future research directions are also discussed.

Secondary school-leaving English examination results are often regarded as indicators of students' competence to study in English-medium universities, which is usually demonstrated through source-based academic writing. In English as a second or foreign language (ESL/EFL) contexts, many English-medium universities require local undergraduates to enrol in an academic writing course, unless they received outstanding English results in secondary school-leaving examinations. This study investigates the relationship between ESL undergraduates' secondary school-leaving English examination results and their academic writing performance through latent profile analysis. Results show that students can be grouped into four classes of academic writing performer (AWP), namely Complex-AWP, High-AWP, Medium-AWP and Low-AWP. Surprisingly, the Complex-AWP group had the highest means in structure, argument and language, but the lowest in citation. Secondary school-leaving English examination results can generally predict students' class membership in language, argumentation and, to some extent, structure, but not citation. An important implication is that students with high English proficiency do not necessarily do well in all aspects of academic writing. This study can inform university senior management on how to set policies about who needs an academic writing course and provide appropriate training in various aspects of academic writing for university students with diverse English proficiency.

Although AI integration in education is accelerating, professional development for Chinese EFL teachers remains predominantly techno-centric and lacks disciplinary specificity, resulting in persistently lower levels of AI literacy. To address this, this study used an eight-week AI-Integrated Genre-Based Pedagogy (AI-GBP) module grounded in Situated Learning Theory and TPACK. A quasi-experimental, explanatory, sequential, and mixedmethods design was employed to compare the AI-GBP module with a generic AI-literacy program among 64 Chinese college EFL teachers (n = 32 per group). Quantitative data from pre-and post-intervention surveys were analyzed using ANCOVA and MANCOVA to determine differences, while qualitative data from semi-structured interviews underwent thematic analysis. Quantitative findings indicated statistically significant gains in overall AI literacy as well as in its subcomponents, particularly regarding practical application and critical appraisal, among teachers in the AI-GBP group compared with the control group. Qualitative findings revealed certain developmental shift from tool-oriented AI use toward more critical and pedagogically relevant engagement, including teachers' increased ability to identify model bias, contextual mismatch, and 'agreeableness' AI possesses. These findings suggest that embedding AI learning within more pedagogically and contextually appropriate situations (e.g., genre-based pedagogy) can support EFL teachers' sustainable and critical integration and understanding of AI in their professional practice.

The transnational flow of resources, information, capital, and commodities in the current era of globally networked economy has made communications increasingly dependent on internet-based multimedia technologies. Accordingly, communications for everyday life and for business are becoming more multimodal. English learners (ELs) are engaged with multimodal language and literacy practices in and out of school for various purposes. Studies of ELs' literacy practices out of school have shown that they are engaged in multimodal communications for entertainment, friendship, and family bonds in transnational settings and affinity groups online (Gee & Hayes, 2011; Kim, 2015; Lam & Warriner, 2012; Yi, 2010). Their communications often are translanguaging processes that mix and remix different languages, genres, modalities, modes, and styles. These literacy practices often lead to learners' development of semiotic competences and more opportunities to express their identities with expanded meaning-making resources, not to mention the expanded agency to represent themselves while learning English (Cimasko & Shin, 2017; Nelson, 2006). The affordances of multimodal literacy have been gradually extended to K-12 contexts of TESOL in supporting the literacy development of ELs. Although traditional alphabet-based literacy practices still dominate language and literacy education, a growing number of teachers have started to employ multimodal literacy in their curriculum and instruction, drawing on various mediums (e.g., digital storytelling, graphic novels, multimedia posters, PowerPoint, social media, performance). Research has examined multimodal composing and literacy practices in K-12 settings, drawing on a range of different theories including social semiotics, sociocultural theories, multiliteracies, and systemic functional linguistics. This chapter presents theoretical and methodological approaches to examining multimodality in K-12, followed by the main affordances and constraints of applying multimodal instruction from the existing literature. It concludes with applications and directions for TESOL in K-12 education.

This study investigates a sixth-grade L2 writer's composition of digital multimodal texts, and his development of the metalanguage of modal and intermodal resources of language and image. Based on a systemic functional approach to multimodal discourse analysis (SF-MDA) and principles of sociosemiotic ethnography, it examines both the student's composing processes and his composed texts. In the writing center of an inclusive sixth-grade classroom in a U.S. elementary school, the student created multimodal PowerPoint slides explaining the greenhouse effect, and Glogster-based multimedia texts arguing for banning guns. The collected data include the student's texts, interview responses, and observation notes. Grounded in SF-MDA, the analytical framework draws on codes such as employed modes, modal resources, intermodal relations, and constructed meanings (i.e., ideational, interpersonal, and textual). The findings suggest that a) the student employed linguistic and visual modes in the mediums drawing on concurrence and complementarity intermodal relations; b) the student used non-linguistic resources for the ideational meanings of texts as a primary semiotic mode; and c) the student developed awareness of intermodal relations and metafunctions of sign systems in a non-linear way. The findings contribute valuable insights into understanding L2 learners' development of metalanguage for multimodal composing.

This study investigates how Large Language Models (LLMs) encode second language (L2) writing proficiency distinctions compared to human learners, focusing on the structural alignment between synthetic outputs and human developmental patterns. We analyzed CEFR-graded Write & Improve 2024 learner essays and matched LLM generations (A2-C1) using statistical models controlling for prompt and text length. With a CALF feature set (Complexity, Accuracy, Lexical Complexity, Fluency), level explained 2.9 % of variance in human texts (R 2 = 0.029) but 10.6 % in pooled LLM outputs (R 2 = 0.106), indicating clearer A2-C1 features in LLM-generated texts overall. LLM level effects were robust (R 2 range = 0.038-0.454), with sharper differences between neighboring CEFR bands than in learner essays (e.g., A2-B1 distance, D 2 = 4.81 versus 1.07). Prompt-wise analyses found effects for all 10 prompts in the LLM set but only 2/10 in the learner set, with much greater separation than in the learner cohort. These results support CALF and CEFR construct validity: with prompt and length controlled, A2-C1 texts show ordered CALF gradients, especially for LLM texts across prompts. The sharper LLM separation likely reflects idealized, low-variance production; therefore, CALF captures core structural proficiency (syntax, lexis, accuracy, fluency) but not discourse-pragmatic qualities or human variability. The analyses provide level-calibrated evidence that LLM texts show clearer distinctions than learners', under identical prompts, whose development is gradual and overlapping. This positions LLMs as both tool (exemplars, rubric calibration) and challenge (assessment validity and fairness), while offering SLA researchers insight into how proficiency constructs are encoded in human versus model-based writing.