[Paper Review] Disentangling Domain Ontologies

Mayukh Bagchi, Subhashis Das|arXiv (Cornell University)|Mar 21, 2023

Semantic Web and Ontologies4 citations

TL;DR

This paper introduces Conceptual Entanglement—a five-level phenomenon arising during incremental domain ontology development involving perception, labelling, semantic alignment, hierarchical modelling, and intensional definition. It proposes Conceptual Disentanglement, a multi-level strategy using normative principles to enforce semantic bijections across levels, enabling the creation of conceptually disentangled, reusable, and harmonized ontologies for knowledge integration and digital libraries.

ABSTRACT

In this paper, we introduce and illustrate the novel phenomenon of Conceptual Entanglement which emerges due to the representational manifoldness immanent while incrementally modelling domain ontologies step-by-step across the following five levels: perception, labelling, semantic alignment, hierarchical modelling and intensional definition. In turn, we propose Conceptual Disentanglement, a multi-level conceptual modelling strategy which enforces and explicates, via guiding principles, semantic bijections with respect to each level of conceptual entanglement (across all the above five levels) paving the way for engineering conceptually disentangled domain ontologies. We also briefly argue why state-of-the-art ontology development methodologies and approaches are insufficient with respect to our characterization.

Motivation & Objective

• To identify and formalize the phenomenon of Conceptual Entanglement as an inherent, ordered manifoldness in incremental domain ontology development.
• To address the limitations of existing ontology development methodologies that fail to account for layered semantic heterogeneity across conceptual representation levels.
• To propose a principled, multi-level strategy—Conceptual Disentanglement—that enables the engineering of conceptually disentangled, reusable domain ontologies.
• To establish a foundation for harmonizing diverse conceptual representations (e.g., Burj Khalifa as skyscraper, hotel, corporate complex) into a unified ontological schema.
• To support knowledge organization in digital libraries and repositories through classificatory ontologies that are purpose-driven, exhaustive, and logically consistent.

Proposed method

• Identifies five distinct levels of conceptual entanglement: perception, labelling, semantic alignment, hierarchical modelling, and intensional definition.
• Proposes a normative framework for disentanglement at each level using guiding principles: canon of purpose, relevant succession, exhaustiveness, and modulation.
• Applies the canon of purpose to ensure that successive classification characteristics are selected based solely on the modeling purpose.
• Employs the canon of relevant succession to enforce a purpose-driven, non-arbitrary sequence of differentiating characteristics in taxonomic hierarchies.
• Uses the canon of exhaustiveness to ensure all concepts at a given depth are fully and mutually exclusive at the next depth in the hierarchy.
• Applies the canon of modulation to eliminate missing links in taxonomic paths, ensuring full connectivity and preventing many-to-many ambiguities in relations.

Experimental results

Research questions

• RQ1 How does conceptual entanglement emerge as a layered, ordered phenomenon across the stages of domain ontology development?
• RQ2 Why are current ontology development methodologies insufficient for handling the full spectrum of semantic heterogeneity arising from conceptual representation levels?
• RQ3 What normative principles can be systematically applied to disentangle conceptual representations across perception, labelling, alignment, hierarchy, and intensional definition?
• RQ4 How can a multi-level disentanglement strategy ensure semantic bijections and logical consistency in domain ontologies?
• RQ5 What is the role of purpose-driven design in eliminating ambiguity and ensuring completeness in classificatory ontologies?

Key findings

• Conceptual Entanglement is a five-level phenomenon rooted in the cognitive and representational processes of human conceptualization, from perception to logical formalization.
• The proposed Conceptual Disentanglement strategy enforces semantic bijections across all five levels through a set of normative principles, enabling unambiguous and reusable ontologies.
• The canons of purpose, relevant succession, exhaustiveness, and modulation collectively eliminate ambiguity, incompleteness, and missing links in taxonomic hierarchies.
• The approach enables a novel correspondence between the language-agnostic mental model, the formal ontology (e.g., OWL), and its underlying first-order logic axiomatization.
• State-of-the-art ontology methodologies fail to recognize or address the full spectrum of layered conceptual entanglement, limiting their utility for classificatory and knowledge-organization tasks.
• The framework is particularly effective for engineering classificatory ontologies in domains like healthcare and digital libraries, where consistency and reusability are critical.