Parsing Arabic sentences, specifically i c rāb , poses unique challenges due to the language’s intricate morphology, diverse syntactic structures, and rich contextual nuances. This study evaluates the performance of leading general-purpose Large Language Models (LLMs) in Arabic i c rāb parsing using a novel human-annotated dataset, systematically covering various grammatical phenomena. A tailored evaluation framework assesses performance across detailed syntactic and morphological features. Under matched multi-shot prompting (basis for cross-model comparisons), Claude-3.5-Sonnet achieved the highest overall F1 score (0.84), followed by GPT-4o (0.83) and Gemini-1.5-Pro (0.77). Conversely, less advanced models such as Claude2.1 and GPT-3.5-turbo struggled with complex constructions, highlighting persistent linguistic limitations. Multi-shot prompting substantially improved accuracy across proprietary models, yielding improvements of up to 18% in complex categories and underscoring the value of in-context learning. Additionally, evaluations of open-source models (DeepSeek-chat-v3-0324 and LLaMA-4-scout) established baseline performance levels confirming substantial gaps compared to proprietary models. The findings reveal ongoing challenges like diacritic sensitivity and semantic ambiguity while establishing a robust benchmark for Arabic grammatical parsing in general-purpose LLMs. All resources (dataset, codebase, and evaluation outputs) are available at https://github.com/alsabahi2030/Arabic-LLM-Parsing . • Human-annotated dataset of 1100 Arabic sentences across 11 categories and 34 subtypes • Tailored evaluation framework assessing Arabic syntactic and morphological parsing • Comparative analysis of Claude, GPT, Gemini, DeepSeek, and LLaMA-4 on Arabic • Multi-shot prompting boosts parsing accuracy considerably in complex categories

Assessment and monitoring of Huntington's disease (HD) symptoms remain limited to infrequent, clinic-based evaluations. We evaluated whether fully automated linguistic analysis of speech tasks recorded via smartphone can remotely capture core clinical markers of HD severity. In this cross-sectional multicenter study across Czech and German sites, 53 participants, including 30 HD (9 pre-symptomatic, 3 prodromal, and 18 manifest) and 23 healthy controls, completed a smartphone-based speech assessment, including spontaneous monologue and fairy tale retelling for 7 consecutive days. Recordings were automatically transcribed and analyzed using natural language processing to derive 3 lexical and 3 syntactic features. Predictive models for clinical outcomes based on the Unified Huntington's Disease Rating Scale and cognitive scales were built using multivariate linear regression with cross-validation. Linguistic features predicted HD severity across multiple domains with high predictive performance, explaining up to 57% of the variance in cognitive performance, 63% in motor impairment, and 59% in functional capacity. A median of 6 days for monologue and 3 for retelling was sufficient to reach 90% of maximal predictive performance. Compared to controls, HD participants showed reduced vocabulary range and increased phrase repetition in both tasks (p < 0.05), with additional monologue-specific deficits in sentence length (p = 0.018) and syntactic complexity (p = 0.004). Fully automated analysis of smartphone-based language assessment can remotely quantify cognitive, motor, and functional impairment in HD, offering a scalable, low-burden digital biomarker for clinical trials and decentralized monitoring.

This systematic review (SR) synthesizes 28 empirical studies conducted by the author between 2000 and 2025 on the teaching, learning, and assessment of English grammar among Saudi EFL college students. The corpus comprises seven experimental and quasi experimental studies evaluating the effectiveness of instructional approaches, online platforms, and learning environments; six model building studies proposing pedagogical frameworks for using podcasts, online tasks, and iRubrics; and fifteen analytical studies documenting learners’ grammatical weaknesses based on test responses and error corpora collected from assignments and projects. The studies were organized into four thematic clusters: teaching grammar with technology; grammar assessment and testing; grammar learning outcomes with three sub-clusters (morphological problems, grammatical difficulties in specialized contexts, and translation problems of grammatical structures); and factors influencing grammar learning outcomes. Across the four clusters, the findings reveal a coherent developmental trajectory in how Saudi EFL learners acquire, process, and apply grammatical knowledge under varying instructional, technological, and contextual conditions. Technology enhanced grammar instruction, whether through LMS platforms, online tasks, Elluminate sessions, or mind mapping, yielded significant gains when the tools were simple with an interactive learning community. Linguistically, students exhibited morphological weaknesses, particularly in plural formation and noun and adjective forming suffixes. Genre based studies showed difficulties in identifying syntactic features in advertisements, legal texts, and news discourse, while translation focused studies revealed challenges with collocations, pronouns, particles, SVO order, and agreement, indicating that grammatical difficulties extend beyond isolated forms to broader syntactic and discourse level processing. Findings from the fourth cluster demonstrate that pedagogical, technological, and socio cultural factors significantly shape grammar learning outcomes. Instructor qualifications, assessment practices, and instructional design influenced learners’ grammatical accuracy; simple LMS designs (as Nicenet) facilitated participation, whereas platforms with complex design, as Moodle and WebCT, hindered engagement; and gender segregated online collaboration introduced socio cultural constraints that affected interaction and learning. Moreover, the content and focus of grammar instruction should vary depending on whether it targets general academic purposes or professional/occupational needs. This SR fills a gap in the literature by offering an integrated thematic synthesis that brings together technology mediated instruction, assessment practices, morphological and syntactic difficulties, translation related errors, and contextual factors affecting grammar learning within a unified analytical framework.

IntroductionEarly detection of Alzheimer’s disease (AD) is critical for timely intervention, particularly during the mild cognitive impairment (MCI) stage. This study aimed to develop and evaluate a multidomain speech analysis framework to support cognitive screening, biomarker prediction within the amyloid, tau and neurodegeneration (ATN) framework, and estimation of cognitive function across the AD continuum.MethodsThis study analyzed speech from 2,320 individuals spanning the cognitive spectrum-including those with subjective cognitive decline (SCD), MCI, and Alzheimer’s disease dementia (ADD)-using three spoken tasks (∼3 min) and extracted multidomain features including acoustic, lexical, syntactic, and semantic features. Machine learning models were trained to classify cognitive status, predict amyloid, tau and neurodegeneration (ATN) biomarker positivity, and estimate scores across six neuropsychological domains.ResultsMultidomain speech models achieved high performance in differentiating cognitive stages, with AUC values of up to 0.94 for SCD vs. ADD and 0.82 for SCD vs. MCI classifications. In biomarker prediction, the models yielded AUCs of 0.71, 0.74, and 0.73 for ATN classification, respectively. Speech-based models also showed strong correlations (up to 0.83) with cognitive function scores. Feature importance analysis revealed that verbal fluency measures were the most predictive. Explainability analyses indicated minimal dependency on age, sex, or education, supporting model fairness.DiscussionThese findings show that multidomain speech features capture clinically and biologically relevant information across the AD continuum, enabling cognitive classification, biomarker prediction, and cognitive estimation. These results underscore the potential of speech analysis as a non-invasive, accessible tool for scalable cognitive screening and early detection of AD. These results underscore the potential of speech analysis as a non-invasive, accessible tool for scalable cognitive screening and early detection of AD.

A significant increase in the number of patent publications in recent years has created difficulties in conducting classical manual analysis and searching for patent analogues. Automation of the search for patent analogues is a key tool for reducing time and financial costs at the stages of patent application formation and patent examination. The use of Big Data technologies and distributed systems makes it possible to build an effective architecture of a system of patent analogues and improve the quality of patent search results. The theoretical significance of the work lies in the development of the architecture and concept of a full-text patent search system based on a comparative analysis of the effectiveness of various distributed systems for searching and processing textual Russian-language information, taking into account its morphological and syntactic features. The practical significance of the work lies in the implemented software, which includes tools for parsing patent documents into a distributed file system, searching taking into account the features of the natural Russian language, as well as a web interface for visualizing search results. Modern frameworks and technologies are used in the process of work: Apache Hadoop, Spark, Hive, Elasticsearch, PostgreSQL, ClickHouse. Elasticsearch showed the best results in both response time and search quality (accuracy – 0.87, completeness – 0.82, F-measure – 0.84) for complex queries reflecting the specifics of the Russian language.

Voice cloning enables realistic fake speech in which a speaker’s identity is preserved while the spoken message is semantically altered. This paper asks whether such meaning-level manipulation leaves detectable traces in transcripts alone. To study this problem, we introduce FakeSpeech+, a paired real–fake dataset built from authentic speech clips and their matched semantically altered counterparts, re-embedded into cloned voices while preserving speaker identity. Using this dataset, we conduct a transcript-first analysis based on interpretable text-only features from two groups: (i) linguistic content organization and discourse dynamics, and (ii) compact production-related proxy cues, including hesitation and disfluency markers. We evaluate these cues under transcript-length control through residualization and compare authentic and manipulated transcripts using statistical and experimental analyses. The results show that only a limited subset of features retains strong separation after length control, with coordination-related structure and emotion anchoring emerging as the clearest cues, while several production-related and discourse-variability features show weaker but still informative differences. In contrast, a number of syntactic, lexical-diversity, and other discourse-level features show substantial overlap after residualization. These findings indicate that transcript-level structure and selected production-related cues remain informative under realistic content-manipulation threats, supporting the value of transcript-based analysis for identity-preserving fake speech.

Proverbs are paremiological units that express a people’s worldview, life experience, and moral values in a concise yet profound manner. Among these, proverbs devoted to the concept of “mind” occupy a special place. These proverbs reflect generalized conclusions about human cognitive activity, thinking patterns, behavioral models, and social positions. Proverbs representing the notion of “mind” hold a central position in the paremiological systems of many languages. By expressing collective wisdom briefly and compactly, they encode cultural attitudes toward intellect, reasoning, consciousness, memory, and judgment. This article examines proverbs on the theme of “mind” in terms of their semantic, lexical, syntactic, stylistic, pragmatic, and cognitive aspects. The study demonstrates that proverbs on the “mind” are not only moral and educational tools but also linguistically rich structures that reveal how societies conceptualize human cognition. The article analyzes the lexical, semantic, syntactic, and stylistic features of proverbs about “mind,” exploring their expressive potential and communicative functions.

Language is a primary means of expressing a society’s cultural, historical, and social life. Literary language and dialects constitute two main manifestations of this broader linguistic system. Literary language represents the normative and codified form of language, employed in education, official documents, literature, and media. Dialects, in contrast, are linguistic varieties associated with specific regions and social groups, distinguished by phonetic, lexical, syntactic, and stylistic features. A complex and dynamic relationship exists between literary language and dialects. Historically, dialects have played a crucial role in the formation of literary language, providing lexical material, phraseology, idioms, and folklore elements that enrich its expressive capacity. Conversely, literary language exerts a normative influence over dialects, standardizing certain features and shaping them into socially accepted forms. From a sociolinguistic perspective, the relationship between dialects and literary language also involves considerations of prestige, value, and social acceptance. The integration of dialectal elements into literary language and their use in literary style contributes to the emergence of new standard forms, thereby ensuring the vitality of the language. The interactions between literary language and dialects are also significant for preserving regional linguistic diversity and cultural heritage. Dialects serve not only as forms of spoken communication but also as carriers of cultural and historical identity. Therefore, dialectal elements incorporated into literary language should be evaluated not only normatively but also functionally. In conclusion, the reciprocal influence between literary language and dialects ensures a balance between standard and regionally specific styles, preserving the richness of the language, enabling the expression of social and cultural phenomena in literary texts, and aligning with societal language policies and normative requirements.

Background: Effective screening and cohort enrichment remain major challenges in clinical trials for Alzheimer’s disease (AD), where traditional diagnostic pathways rely on costly, invasive, and time-consuming procedures. Speech and language analysis has emerged as a scalable, low-burden approach for detecting subtle cognitive-linguistic and motor-speech changes that may appear early in the disease course. Summary: This review synthesizes current evidence on acoustic, prosodic, lexical, semantic, and syntactic speech features associated with AD and mild cognitive impairment (MCI) and evaluates their reported utility across a range of elicitation tasks including picture description, verbal fluency, narrative recall, spontaneous speech, and reading. Across studies, machine-learning models trained on speech and language features have reported consistent performance, although results vary substantially depending on task design, feature sets, and cohort characteristics. Task-dependent variability is evident, with picture description and verbal fluency tasks capturing lexical-semantic and timing markers, while narrative and spontaneous speech tasks capture impairments in coherence, information content, and prosody. Hybrid approaches integrating hand-crafted and machine-extracted features have also been explored to improve interpretability and model performance. Speech and language analytics may support digital pre-screening, cohort enrichment, and quality-assurance monitoring within clinical trials; however, their application depends on methodological considerations and validation across diverse settings. Key Messages: Despite encouraging findings, several methodological challenges persist, including inter-individual variability, limited dataset sizes, differences in recording conditions, and limitations in automatic speech recognition performance in cognitively impaired populations. Continued development of standardized protocols, disorder-adapted speech models, and multimodal analytic pipelines is needed to support clinical translation. Collectively, current evidence suggests that speech and language features represent candidate digital markers that may improve screening efficiency and support clinical trial enrichment in AD, although further validation is required to establish their reliability and generalizability.

This article addresses the significance of Gang of Four (GoF) design patterns as formal architectural solutions in objectoriented programming and emphasizes the importance of their automated detection in modern software systems. This study examines the challenges of identifying architectural solutions in extensive software systems and the constraints of conventional analytical approaches. The scientific novelty of the suggested method is in the utilization of contemporary transformer-based language models trained on source code integrated with conventional machine learning techniques for identifying structural patterns. The proposed approach employs the DeepSeek-Coder-V2 framework to generate multidimensional vector representations (embeddings) of code segments. We employ Principal Component Analysis to reduce dimensionality. The resultant embeddings serve as features for training and testing various classifiers, encompassing both linear and nonlinear models. The objective is to autonomously identify design trends. We developed a bespoke annotated dataset of 23 GoF patterns and additional architectural patterns derived from actual open-source projects. Experiments demonstrate that transformer-based code embeddings significantly outperform conventional feature extraction techniques, achieving a macro-averaged F1-score of up to 0.82. The test demonstrates that the embeddings accurately represent both the syntactic and semantic characteristics of the source code. The proposed strategy is more versatile and capable of managing a broader array of scenarios compared to manual or heuristic-based solutions. It functions effectively for pattern recognition tasks and can be utilized to analyze extensive codebases. Potential applications encompass rectifying, sustaining, and enhancing the quality and comprehension of software architecture. This approach establishes a unified framework for subsequent research and advancement in software engineering.

This study examines the effectiveness of the URUBAH game as an Android-based Arabic learning application developed using the ADDIE instructional model and Unity 3D software, with the aim of improving second-language (L2) learners’ Arabic writing proficiency. The game was designed to support Indonesian learners of Arabic who experience difficulties in spelling, grammar, and sentence structure. It addresses these challenges by focusing on the morphological and syntactic features of Arabic while considering structural differences between Arabic and English writing systems. The objectives of this research are twofold: first, to develop an interactive digital learning tool that enhances Arabic writing skills through gamification; and second, to evaluate its effectiveness in comparison with traditional teaching methods. A mixed-methods approach was employed, integrating quantitative data derived from pre- and post-tests with qualitative insights gathered from surveys and learner feedback. The findings indicate that the game positively improves writing outcomes, including learners’ accuracy, fluency, and overall engagement. In addition, learners’ motivation increased, accompanied by reduced anxiety when practicing Arabic writing through the interactive platform. In conclusion, the URUBAH game presents a promising instructional approach with practical applications for teaching Arabic as a second language. It also demonstrates the potential of integrating advanced digital technology into language classrooms to enhance learning experiences

The subject of the research is the capabilities and limitations of modern generative language models ChatGPT and DeepSeek in conducting a comprehensive analysis of media texts. The work examines their ability to process qualitative and quantitative characteristics of news materials, including lexical-statistical parameters (word frequency, unique vocabulary), syntactic features (sentence complexity, use of grammatical constructions), and stylistic markers (emotional coloring, clich&amp;#233;d expressions). Special attention is paid to the models' ability to identify manipulative techniques and potentially misleading information in media discourse. The main goal of the research is to conduct a comparative assessment of the effectiveness of generative models in analyzing media content of various themes and volumes. The study aims to determine the degree of reliability of automated analysis, the models' ability to identify key topics and their sentiment, as well as their potential to prevent the spread of fake information. The research utilized a sample of 10 texts from RIA Novosti, differing in theme and volume. The analysis was conducted in three directions: lexical-statistical, syntactic, and stylistic. Manual analysis was used for result verification. The methodology included a standardized prompt for both models, which ensured uniform evaluation criteria. The main findings of the conducted research are as follows. Firstly, the fundamental possibility of using generative language models for comprehensive analysis of media texts has been proven, including the identification of stylistic features, key topics, and sentiment of the materials. Secondly, it has been established that modern models demonstrate high effectiveness in qualitative analysis, comparable to expert evaluation, but have significant limitations in the accuracy of quantitative measurements of textual parameters. Thirdly, the models show the ability to recognize manipulative language techniques and potentially unreliable information. A significant contribution to the research topic is the development of a standardized methodology for applying generative models for media content analysis, which includes a comprehensive three-component approach (lexical-statistical, syntactic, and stylistic analysis) and criteria for assessing the reliability and objectivity of media texts. The results obtained are of great importance for the development of digital media linguistics and the improvement of editorial processes. Promising directions for further research include enhancing the accuracy of quantitative analysis, adapting the methodology for multilingual texts, and integrating with other AI fact-checking technologies.

Dataset for drone problem identification and severity estimation.

Complex syntax in psychological test items may hinder some respondents’ comprehension, yet the prevalence of specific syntactic complexity features and their implications for item responding and validity are not well understood. This two-part study provides new empirical insight into these relationships. Study 1 examines whether different types of dependent clauses are associated with item responses and whether these associations vary according to respondents’ vocabulary familiarity. Study 2 uses a corpus-based register analysis to determine how frequently dependent clause types occur in psychological test items, providing context for the practical significance of the Study 1 findings. Of the five clause types examined in Study 1, only nonfinite complement clauses showed a significant association with item responses, and this association was moderated by vocabulary familiarity. Study 2 further showed that nonfinite complement clauses are the most prevalent dependent clause type in a corpus of psychological test items. Together, these findings suggest that dependent clause types should not be treated as uniformly difficult to process. This study demonstrates how integrating psychometric modeling with corpus-based linguistic analysis can build validity evidence and inform fairer assessment practices for linguistically diverse populations.

Understanding the distinction between signs in sign language and gestures has been a topic of interest among researchers over the past decade. Recent arguments emphasize that sign languages exhibit syntactic, semantic, morphological and phonological features, whereas gestures are considered external to language. This theoretical distinction is importantly reflected at the brain level, with distinct brain regions being involved in sign language comprehension and gesture processing. Using a fast periodic visual stimulation (FPVS) EEG design, the present study investigated whether this distinction is modulated by sign language knowledge in hearing adults. We compared neural responses of late hearing signers and hearing non-signers to finger-number configurations presented either as number signs (finger-montring) or number gestures (finger-counting). Results revealed that hearing signers showed significantly stronger neural discrimination responses to number signs than hearing non-signers, whereas both groups exhibited similar responses to number gestures. These findings demonstrate that number signs and number gestures elicit distinct neural responses as a function of linguistic experience. Crucially, they show that even late acquisition of sign language is sufficient to modulate neural processing of manual number representations, supporting the view that sign language recruits language-specific neural mechanisms. More broadly, these results provide neurophysiological evidence that manual configurations can be processed either as linguistic signs or as gestures depending on the observer's language knowledge, thereby underscoring the linguistic status of sign languages.

The subject of the research is the phenomenon of the "artificial correlate" of folk tales – a text generated by a large language model that imitates the stylistic and plot framework of the folkloric original. The authors thoroughly examine the lexical, morphological, and syntactic features of authentic folk tales and their generated counterparts. Special attention is given to the analysis of compositional, ritual-mythological, and semiotic characteristics of the folkloric fairy tale text and the artificially generated tale. The aim of the work is to construct a theoretical model that allows for the parameterization of the "naturalness" of fairy tale discourse and to identify the ontological gaps between authentic folklore and its algorithmic imitation. The research material comprises authentic folk tales in Russian (from D.K. Zelinin's collection), as well as fairy tale texts generated by the following LLMs: GigaChat and AliceAI. The research methodology includes corpus analysis, conceptual modeling, methods of computational linguistics, as well as elements of quantitative and statistical analysis using the Python programming language. The scientific novelty of the research lies in the development of a multi-level model for assessing the naturalness of fairy tale narratives, which, unlike existing technical approaches, takes into account the mythological, ritual, and ethical constants of the genre. For the first time, a comprehensive analysis of the deficiencies of the generated text is conducted, not as technical errors, but as symptoms of the model's failure to comprehend the culturally significant code. The main conclusion is the demonstration that the artificial correlate largely successfully reproduces the superficial attributes of the genre ("the texture" of the tale), but exhibits significant deficiencies at the level of motif structure and ethical causality. It has been established that a key difference between a natural fairy tale and its digital counterpart is the presence of a rigid ritual-mythological foundation that ensures the teleology of the plot. The generated correlate, on the contrary, demonstrates a "fragmentary" nature: mechanical combinations of folkloric clich&amp;#233;s without maintaining the internal logic of the fairy tale world, as well as euphemization of archaic motifs. The developed model of substantive parameters allows not only to diagnose the nature of the text but also raises questions about the limits of artificial intelligence's ability to reproduce culturally significant codes.

There has been a continuous rise in the emergence of media, including print, audio, and visual forms such as newspapers, television, and radio. One of the forms of media is newspapers, which is used in an extensive way as a tool for damaging the reputation of others. This research conducts a critical discourse analysis of defamation in selected English newspapers. The researcher examines the syntactic and lexical features used in newspapers in conducting the act of defamation, as well as the ideologies embedded in these linguistic features. The researcher aims to discover the hidden ideologies that are embedded in language words and structure by conducting a critical discourse study. To conduct such a study, the researcher adopts an eclectic model consisting of Fairclough's three-dimensional model and van Dijk's ideological discursive strategies. The data are two cases taken from two English newspapers. The researcher concludes that defamatory articles contain syntactic features such as modality, word order and nominalization. Each of these strategies contains a specific ideology, such as criminalization and discrimination.

This paper explores methodological approaches to forensic stylistics, focusing on the role of style analysis in legal contexts. Forensic stylistics examines how lexical, syntactic, pragmatic, and discourse-level features of language provide evidence in authorship attribution, text authenticity verification, and the interpretation of disputed documents. The study highlights both qualitative and quantitative methods, including corpus analysis, comparative stylistics, and computational techniques, which enable experts to detect distinctive authorial patterns. By combining these approaches, forensic stylistics ensures a more balanced and reliable evaluation of linguistic evidence. The paper also addresses challenges such as deliberate disguise, stylistic variability across genres, and limited text samples, emphasizing the need for interdisciplinary collaboration between linguists and legal practitioners. Ultimately, methodological rigor in stylistic analysis strengthens the objectivity and transparency of forensic linguistic expertise, ensuring that linguistic evidence serves as a credible and scientifically grounded tool within judicial processes.

To investigate the diachronic changes in syntactic complexity in argumentative writing by Chinese English-major students, this study takes the English-major writing subcorpus of the Chinese Learner English Corpus (CLEC) and a self-built contemporary corpus of argumentative writings by English-major students as research objects. Based on the 14 syntactic complexity measures proposed by Lu, a diachronic comparison is conducted on the syntactic complexity features in argumentative writing by junior English-major students two decades ago and at present. The results reveal that no significant differences exist between learners of the two periods in such measures as unit length, sentence complexity and the use of subordinate structures, and the complexity of verb phrases remains relatively stable. In contrast, the complexity of coordinate structures and noun phrase structures is significantly higher in contemporary texts than in texts from 20 years ago. The study indicates that changes in the teaching environment and language input conditions for writing in recent years have to some extent facilitated learners’ expansion of syntactic expressions, yet the development of subordinate structures remains relatively slow. The findings can provide references for optimizing the focus of syntactic complexity training in English-major writing instruction.

This study investigates the grammatical structure of Bisaya as spoken in Santa Josefa, a linguistic variety that serves as a marker of community identity and reflects the speakers’ deep appreciation for their language. Anchored in a pedagogic grammar framework and employing a qualitative descriptive research design, the study utilized discourse analysis of data drawn from Pear Film narratives, procedural (“how-to”) interviews, and a 200-word translation task. These data were supplemented by acoustic vowel analysis using PRAAT to capture phonetic and phonological features. The findings reveal salient morphological, phonological, morphosyntactic, and syntactic characteristics of the Bisaya variety in Santa Josefa. Its phonemic inventory comprises sixteen consonants and three vowels, with notable phonological features including minimal pairs, diphthongs, and distinct phonotactic patterns. The language demonstrates productive morphological processes and exhibits an ergative–absolutive alignment alongside a predominantly verb-initial sentence structure. The study concludes that the Bisaya spoken in Santa Josefa possesses a systematic and culturally grounded grammatical system that reinforces linguistic identity. The grammar notes generated from this research contribute to the documentation of this local variety and offer a valuable resource for the development of instructional materials in Mother Tongue-Based education, thereby supporting language preservation and pedagogical effectiveness. Keywords: Grammar Notes, Bisaya, Santa Josefa, Philippines gideon.tiongson@deped,gov.ph Pag-asa Elementary School, Philippines