How Is Meaning Organized in Judicial Decisions? Evidence from Semantic Polarity, Entropy, and Network Structure

Institutional discourse is shaped by normative and communicative constraints that influence how meaning is organized in text. This study examines the semantic structure of judicial decisions across two types of cases using a computational linguistic approach. Semantic polarity and lexical entropy are measured in shared lexical items, while semantic organization is explored through dimensionality reduction using t-distributed Stochastic Neighbor Embedding (t-SNE) and shared-word semantic networks. Comparative statistical analyses reveal systematic differences in semantic configuration between the two domains. Decisions related to economic disputes exhibit more formal and procedural semantic patterning, whereas divorce decisions show greater heterogeneity and contextual engagement. These findings demonstrate that semantic organization in institutional texts adapts to differing normative contexts, contributing to quantitative approaches to language-in-use.