Semantic Cluster Theory – Part I: Abstract Concepts as Networked Semantic Clusters

This paper develops Semantic Cluster Theory, arguing that abstract concepts—such as love, justice, freedom, or trust—are not fixed semantic units but networked structures defined by their surrounding vocabulary (“semantic clusters”). A concept’s meaning arises from: the configuration of its peripheral lexical nodes, the directional vector connecting these nodes, and the strength of their internal edges. Because multiple clusters may coexist within an individual mind, Selective Normalization emerges: people adopt one cluster as the “true meaning” while reclassifying alternative clusters under different labels. This framework explains why misunderstanding, value conflict, and conceptual divergence occur even when speakers use the same word. The paper formalizes cluster-based meaning, analyzes vector alignment as the basis of interpersonal understanding, and introduces the phenomenon of expansion events, in which a concept’s network grows when new lexical nodes are integrated without contradiction. Part I establishes the theoretical foundation for Parts II and III, which address specialized concepts and AI–human cognitive compatibility.