The Generative Mechanism of Reporting Potential in Academic Discourse

转述结构作为学术话语多声性的核心载体，其生成机制和功能意义尚未在不同学科中得到合理解释。本文基于系统功能语言学与构式语法，构建跨学科学术语篇语料库，提出“学术增强型语义–语用词嵌入模型(Academic Enhanced Semantic-Pragmatic Embedding)”，量化分析转述动词的学科特异性语义分布规律；并提出指数衰减模型(Exponential Decay Model)，揭示了不同转述动词的语义相似度衰减速率。研究发现：1) 学科认识论驱动转述结构的句法分化，人文社会科学倾向使用“主体–行为–内容”构式，使用逻辑连接词以强化立场协商，自然科学倾向使用“现象–结论”构式，依赖名词化与简化主谓宾结构；2) 转述潜势(Reporting Potential)是不同学科中句法规则与语法资源的概率化生成系统，它通过及物性过程反映知识表征功能。本研究为学术写作规范与跨学科话语分析提供了理论框架与实证依据。 As a core carrier of heteroglossia in academic discourse, the generative mechanisms and functional significance of reporting structures remain underexplored across disciplinary contexts. This study develops a cross-disciplinary academic discourse corpus through the integration of Systemic Functional Linguistics and Construction Grammar. We propose an Academic-Enhanced Semantic-Pragmatic Embedding model to quantify disciplinary-specific semantic distribution patterns of reporting verbs, complemented by an Exponential Decay Model that reveals semantic similarity attenuation rates across verb categories. Key findings demonstrate: 1) Disciplinary epistemologies drive syntactic differentiation of reporting structures, humanities and social sciences favor agent-process-content constructions with logical connectives for stance negotiation, while natural sciences prefer phenomenon-conclusion patterns through nominalization and simplified SVO structures; 2) Reporting potential constitutes a probabilistic generative system of syntactic conventions and grammatical resources, operationalized through transitivity processes to fulfill knowledge representation functions. This research establishes a theoretical framework and empirical foundation for academic writing standardization and interdisciplinary discourse analysis.