The Open World/Closed World Conundrum

2025-11-17 · Source: The Ontologist · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Data Science & Analytics, Software Development & Engineering · Depth: Advanced, long

Summary

The article explores the fundamental distinction between the Closed World Assumption (CWA) and the Open World Assumption (OWA) in knowledge representation, particularly within semantic web technologies. CWA, common in relational databases, assumes a knowledge base is complete, treating unproven statements as false. OWA, foundational to OWL and the Semantic Web, assumes incompleteness, treating unproven statements as unknown. This tension creates engineering friction in semantic technology stacks, as tools like RDF and OWL operate under OWA, while SPARQL queries often imply CWA, and SHACL enforces constraints under CWA. The author argues that the industry is inexorably moving towards closed-world systems due to performance, query complexity, security, and the demands of hypergraph architectures and AI integration.

Key takeaway

For AI Architects and Data Engineers designing knowledge graphs, recognize that while OWL embraces the Open World Assumption, practical systems and emerging AI integrations increasingly favor Closed World models. You should explicitly design for trivalent logic in SPARQL and SHACL to handle indeterminacy, preventing silent data loss or misinterpretation. Prioritize SHACL for constraint enforcement and consider SPARQL property paths for efficiency over extensive triple materialization.

Key insights

The semantic web is shifting from Open World to Closed World Assumptions, driven by practical system demands.

Principles

• Absence of evidence is evidence of absence (CWA).
• Absence of evidence is absence of evidence (OWA).
• Model domains with OWA, operations with CWA.

Method

SPARQL and SHACL require deliberate authoring, such as the BIND/IF/BOUND pattern in SPARQL or `sh:sparql` with `sh:message` in SHACL, to explicitly surface indeterminate states rather than silently collapsing them to false.

In practice

• Use `BIND/IF/BOUND` in SPARQL for trivalent logic.
• Employ `sh:sparql` with `sh:message` in SHACL for indeterminacy.
• Prioritize SPARQL property paths over materialization for `owl:sameAs`.

Topics

• Open World Assumption
• Closed World Assumption
• Semantic Web
• SHACL
• Knowledge Representation

Best for: AI Architect, AI Engineer, Data Engineer

Related on AIssential

• Off the Edge of the Map — Conflating absence of data with negative claims (CWA) in open-world systems leads to false conclusions.
• Converting from OWL to SHACL, Part 2 — Hybrid OWL and SHACL systems augment validation with open-world dynamics, rather than replacing OWL.
• Not an Ontology — Major data platforms market "ontologies" that lack true reasoning and inference, failing to meet core definitional requirements.
• Tractable Reasoning and Conjunctive Query Answering for Defeasible DL-Lite under Rational Closure — A plug-in architecture enables efficient defeasible reasoning and CQ answering for DL-Lite under Rational Closure.
• True Positive Weekly #166 — The issue highlights diverse AI advancements, from genomic challenges to novel LLM behaviors and practical tools.
• Towards Intrinsic Interpretability of Large Language Models:A Survey of Design Principles and Architectures — Intrinsic interpretability builds transparency directly into LLM architectures to enhance trustworthiness and safety.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Ontologist.