[Paper Review] The observational roots of reference of the semantic web

Simon Scheider, Krzysztof Janowicz+1 more|arXiv (Cornell University)|Jun 27, 2012

Semantic Web and Ontologies18 references3 citations

TL;DR

This paper proposes a human-centric, operational approach to reference in the Semantic Web, grounded in perceptual, cognitive, and technical human competences. By building reference systems based on reproducible observation procedures—such as geodetic systems, measurement scales, and address systems—it enables bottom-up semantic integration, overcoming limitations of top-down ontologies and formal semantics.

ABSTRACT

Shared reference is an essential aspect of meaning. It is also indispensable for the semantic web, since it enables to weave the global graph, i.e., it allows different users to contribute to an identical referent. For example, an essential kind of referent is a geographic place, to which users may contribute observations. We argue for a human-centric, operational approach towards reference, based on respective human competences. These competences encompass perceptual, cognitive as well as technical ones, and together they allow humans to inter-subjectively refer to a phenomenon in their environment. The technology stack of the semantic web should be extended by such operations. This would allow establishing new kinds of observation-based reference systems that help constrain and integrate the semantic web bottom-up.

Motivation & Objective

• Address the persistent problem of reference in the Semantic Web, where shared referents are essential for linking data across users.
• Challenge the dominance of top-down, syntax-driven approaches in semantic web research by emphasizing human cognitive and perceptual operations.
• Overcome limitations of formal semantics and natural language descriptions, which are context-dependent and fail to ensure inter-subjective reference.
• Introduce reference systems as a new semantic integration layer that constrains meaning through reproducible human observations and conventions.
• Extend the Semantic Web technology stack to incorporate human-centric, operational foundations for grounding meaning in real-world phenomena.

Proposed method

• Propose an operational view of reference as a human speech act involving attention to reproducible phenomena, grounded in perception and joint attention.
• Reconstruct semantic concepts (e.g., geographic places, water bodies) not top-down via definitions, but bottom-up via perceptual operations and affordances.
• Use reference systems—such as geodetic systems, measurement scales, and address systems—as models for new kinds of semantic integration layers.
• Anchor abstract or vague concepts in conventionally established, reproducible observation procedures to ensure inter-subjectivity.
• Draw on cognitive frameworks like conceptual spaces (Gärdenfors) and observation sentences (Quine) to formalize how human cognition grounds reference.
• Integrate these systems into the Semantic Web stack to enable bottom-up constraint of meaning, complementing top-down ontology reasoning.

Experimental results

Research questions

• RQ1 How can shared reference be established in the Semantic Web without relying solely on formal semantics or URIs?
• RQ2 What role do human perceptual and cognitive operations play in grounding semantic reference?
• RQ3 How can reference systems based on observation procedures serve as a foundation for semantic integration across heterogeneous data?
• RQ4 In what ways do existing systems like geodetic or measurement systems exemplify operational reference, and how can they be generalized?
• RQ5 How can the Semantic Web technology stack be extended to incorporate human-centric, bottom-up reference mechanisms?

Key findings

• Shared reference is essential for the Semantic Web’s network effect, enabling multiple users to contribute to the same referent, such as a geographic place.
• Formal semantics and natural language descriptions alone cannot resolve the problem of reference due to context dependence and polysemy.
• Reference is not an abstract mapping but a human action involving perception, joint attention, and convention, rooted in observable, reproducible phenomena.
• Existing reference systems—like geodetic systems and calendars—are successful because they are grounded in human-observable procedures and conventions.
• Bottom-up reference systems based on human operations can constrain meaning more effectively than top-down ontologies, especially in heterogeneous domains.
• New reference systems can enable semantic integration across diverse conceptualizations by anchoring concepts in perceptual operations and convex regions in conceptual spaces.