Event Meaning Construction and Knowledge Representation in Multilingual and Multimodal Environments: A Computational Linguistics Framework for Large Models in Southeast Asian Event Analysis

Abstract

With the rapid development of large language models and multimodal artificial intelligence, event analysis has gradually evolved from traditional information extraction toward event intelligence, which integrates semantic understanding, knowledge representation, and intelligent reasoning. In multilingual and multimodal environments characterized by heterogeneous information sources, existing studies mainly focus on event extraction, modality fusion, and model performance optimization, while paying insufficient attention to the mechanisms of event meaning construction and unified knowledge representation. Taking Southeast Asian multilingual and multimodal event analysis as an application context, this paper adopts a computational linguistics perspective and draws upon Systemic Functional Linguistics, Frame Semantics, and Multimodal Discourse Analysis to propose the concept of Event Meaning Construction. It further explores the mechanisms underlying meaning construction in multilingual and multimodal environments and introduces a Multilingual Multimodal Event Knowledge Representation Framework (MME-KR Framework). Building upon this framework, the paper discusses the evolutionary path of event intelligence and its potential implications for knowledge-enhanced large models in complex regional contexts. It is argued that event analysis should move beyond structural information extraction toward meaning construction, knowledge organization, and intelligent reasoning, thereby providing a unified theoretical foundation for knowledge-enhanced large models, complex event understanding, and regional intelligent analysis.