Disertasi

Incorporating lexically-based language teaching and google gemini in paragraph writing class exploring students’ collocation and writing style / Pandu Prasodjo

Abstrak
In Indonesia English language teaching programs (ELT) face a fascinating paradox. ELT bachelor graduates are expected to be highly skilled in both linguistic skills and pedagogy. However the reality paints a different picture despite their theoretical grounding many graduates still struggle with expanding their own vocabulary size and developing a robust mental lexicon. Indonesian ELT undergraduate students often navigate a complex world of vocabulary acquisition. While many are aware of their vocabulary limitations they might fall into the trap of relying heavily on a group of familiar words known as lexical teddy bears (sometimes called phraseological teddy bears). Lexical teddy bears offer a sense of security and comfort due to their frequent use. However this overreliance significantly hinders natural language use for them. The skill to overcoming the lexical teddy bear trap lies in expanding the vocabulary bank especially fixed and strong collocations through lexically-based language teaching. The lexically-based language teaching emphasizes the importance of lexico-grammatical chunks such as fixed and strong collocations to develop natural language use. However traditional teaching often fall short of providing the personalized and context-specific guidance which needed to internalize these chunks appropriately. This gap could be addressed by incorporating Google Gemini an AI-driven tool into the teaching approach. Google Gemini provided AI-assisted feedback to enhance lexical precision in developing ELT students rsquo paragraph writing particularly in learning fixed and strong collocations instead of isolated words. Gemini served as a tool that bridged the gap in lexically-based language teaching by providing real-time and immediate feedback correcting inaccurate collocations and introducing alternatives which enhanced students rsquo vocabulary bank. Additionally by consistently exposing them to accurate collocations Gemini helped to gradually diminish the reliance on ldquo lexical teddy bears rdquo . In terms of writing style Gemini rsquo s feedback played a noticeable role in fostering contextually appropriate accurate and concise writing by identifying inconsistencies awkward phrasing or language tone that may detract from the quality of paragraph writing. Gemini provided suggestions that encouraged students to use varied sentence structures transitions and advanced vocabulary leading to a more refined writing style. This continuous exposure to both corrective feedback and guidance inspired students to become more aware and reflective writers capable of writing paragraphs that were not only grammatically sound but also nuanced. Therefore Gemini did not merely assist students in identifying errors it enhanced their ability to craft expressive and accurate paragraphs ultimately improving both their lexical proficiency and overall writing competence. This dissertation was conducted qualitatively with an exploratory research design to incorporate lexically-based language teaching as a teaching approach and the potential of an AI-driven technology Google Gemini to expose collocations and writing style development among 14 ELT undergraduate students programming Paragraph Writing class in a private university in Batam. The study focused on how Gemini could strengthen lexically-based language teaching to (a) expose students with a vast corpus of text through Google Gemini rsquo s big data introducing accurate collocations (fixed and strong) (b) assist students AI-driven feedback on contextually appropriate and suggestions for accurate and concise writing (c) personalize opportunities to individual student needs and learning styles and promoting vocabulary learning engagement. The data were collected from three paragraph writing assignments students rsquo reflections and Gemini rsquo s feedback logs as the research instruments. The first writing assignment was considered the students rsquo writing performance baseline. There were 2 revision sessions and 2 writing assignments respectively. The data analysis began with compiling and classifying students writing assignments compiling and classifying the collocation usage quantifying fixed and strong collocation for each student compiling Google Gemini s feedback from all revision sessions categorizing feedback thematically comparing pre- and post- feedback performance summarize findings and highlighting gaps and outliers. This study claimed that incorporating lexically-based language teaching and Google Gemini concluded compatibility. Lexically-based language teaching was strengthened by Google Gemini which provided immediate and contextually relevant feedback thus encouraging students to acquire vocabulary through meaningful contexts rather than rote memorization assisted students in refining their writing appropriateness accuracy and conciseness and involved repetitive exposure to collocations and conscious attention using authentic materials within paragraph writing class.

Informasi Detail

DDC

Rd 421 PRA i

Prodi

Universitas Negeri Malang. Program Studi Pendidikan Bahasa Inggris, 2025.

Deskripsi Fisik

xix, 268 lembar. : ilus. ; 30 cm.

Bahasa

No Reg

00015/RD/25

Edisi

Disertasi (Pascasarjana) - Universitas Negeri Malang. 2025

Subjek

1. BAHASA INGGRIS - MENULIS - STUDI DAN PENGAJARAN
2. ENGLISH - WRITING - STUDY AND TEACHING

Pembimbing

1. Prof. Dr. Siusana Kweldju, M.pd;2. Prof. Dr. Hj. Nur Mukminatien, M.pd;3. Francisca Maria Ivone, S.pd, M.a., Ph.d

Lampiran Berkas