A Triad Architecture for Knowledge Organization Ecosystems: Designing Explainable and Sustainable Semantic Infrastructures

Abstract

Knowledge Organization Ecosystems (KOEs) represent the next evolutionary stage in the management and curation of digital knowledge, demanding conceptual models that transcend static classification and embrace dynamic, multi-actor semantic processes. This paper introduces and elaborates a comprehensive triad-based conceptual architecture, in which fundamental knowledge units are modeled as structured entities composed of three interdependent and co-evolving layers: referential grounding (Base), representational encoding (Representation), and procedural-relational linkage (Linkage). This framework is philosophically grounded in the triplet model of scientific concepts, which overcomes the limitations of reductive singlet and duplet models by capturing the asynchronous evolution of a concept’s ontic commitments, symbolic forms, and the practices that bind them. Moving beyond traditional Knowledge Organization Systems (KOSs), which prioritize stable, hierarchical structures, the proposed architecture conceptualizes knowledge objects as dynamic, operationally embedded semantic entities. These entities evolve within complex digital ecosystems shaped by algorithmic mediation, collaborative curation, and institutional governance. The triad architecture directly supports semantic interoperability across distributed systems, version-managed knowledge resources that track conceptual drift, and explainable Human–Artificial Intelligence (human–AI) knowledge workflows that make semantic assumptions transparent. By aligning conceptual modeling with ontology engineering practices, knowledge graph infrastructures, and the principles of responsible AI, this paper provides a scalable, epistemically robust foundation for building transparent, accountable, and sustainable knowledge organization infrastructures. The framework is demonstrated through a detailed application scenario tracing the evolution of the ‘Open Science’ concept within the European Open Science Cloud (EOSC), illustrating the model’s utility for analyzing and designing real-world semantic infrastructures. It contributes a significant theoretical advance to the study of KOEs and offers a concrete, transferable design framework for next-generation semantic infrastructures in data-intensive, AI-mediated research and cultural heritage environments.