The Seven Core Dimensions of Conversational Context (CXC-7): A Framework Proposal for AI and Large Language Models

IMPORTANT NOTE: This paper presents a theoretical framework proposal without original empirical data. The CXC-7 framework is derived from systematic literature review and case analysis. This study addresses the lack of systematic contextual analysis frameworks in AI conversational interactions by proposing the Seven Core Dimensions of Conversational Context (CXC-7). Drawing on established psychological theories (Bowlby's attachment theory, Kahneman & Tversky's framing effects) and recent conversational AI safety taxonomies (Dinan et al., 2022), this framework encompasses seven core dimensions: Emotion & Attachment (E), Framing & Discourse Power (F), Ethical & Safety Boundaries (B), System & Surface Shifts (S), Prompt Ecology (P), Diffusion & Culture (D), and Transparency & Auditability (T). 中文摘要本稿提出 CXC-7:用以分析 AI 對話語境的七個核心維度,並主張「最低責任揭露(MSD)」作為治理與設計的共同基線。論文定位為理論提案,無原始實證數據;但提供可驗證、可設計、可治理的座標與圖組(互動矩陣、合成態、風險邊界)。並與 NIST AI RMF、ISO/IEC 42001、EU AI Act 之透明義務對齊。 版本 1.0 收錄圖組、MSD 檢核、跨語言擴散風險與範疇界定。 Version 1.0.1 Updates: Added Zenodo DOI to Citation sections in both PDFs Updated BibTeX references with DOI: 10.5281/zenodo.17247638 Complete citation formats (BibTeX and APA) included Contents: English Version ZIP: PDF, Typst source files, BibTeX, figures, compilation guide Chinese Version ZIP: PDF, Typst source files, BibTeX, figures, compilation guide