An Ontology-Driven Graph RAG for Legal Norms: A Hierarchical, Temporal, and Deterministic Approach

Standard Retrieval-Augmented Generation (RAG) systems fail in the legal domain, as they are blind to the hierarchical, diachronic, and causal structure of law, leading to anachronistic and unreliable answers. This paper introduces the Structure-Aware Temporal Graph RAG (SAT-Graph RAG), an ontology-driven framework designed to overcome these limitations by explicitly modeling the formal structure and diachronic nature of legal norms. We ground our knowledge graph in a formal, LRMoo-inspired model that distinguishes abstract legal Works from their versioned Expressions. To validate this architecture, we present a qualitative, trace-based evaluation demonstrating how an agentic framework can leverage our model’s deterministic primitives to resolve complex temporal and provenance queries that are intractable for standard RAG. The result is a practical and auditable framework for building trustworthy and explainable legal AI systems.