A Minimal Deductive System for RDFS with Negative Statements

TL;DR

This paper extends RDFS to handle negative statements under the Open World Assumption by introducing a minimal, syntactic extension that preserves compatibility with existing reasoners and ensures satisfiability.

Contribution

It proposes a novel minimal extension of RDFS, called a ho df_ot^ egb, enabling negative statement reasoning without reification and maintaining compatibility with standard reasoners.

Findings

a ho df_ot^ egb remains syntactically a triple language.

Knowledge bases in a ho df_ot^ egb are always satisfiable.

Entailment decision complexity ranges from P to NP.

Abstract

The triple language RDFS is designed to represent and reason with \emph{positive} statements only (e.g."antipyretics are drugs"). In this paper we show how to extend RDFS to express and reason with various forms of negative statements under the Open World Assumption (OWA). To do so, we start from $\rho df$, a minimal, but significant RDFS fragment that covers all essential features of RDFS, and then extend it to $\rho df_\bot^\neg$, allowing express also statements such as "radio therapies are non drug treatments", "Ebola has no treatment", or "opioids and antipyretics are disjoint classes". The main and, to the best of our knowledge, unique features of our proposal are: (i) $\rho df_\bot^\neg$ remains syntactically a triple language by extending $\rho df$ with new symbols with specific semantics and there is no need to revert to the reification method to represent negative triples; (ii) the logic is defined in such a way that any RDFS reasoner/store may handle the new predicates as ordinary terms if it does not want to take account of the extra capabilities; (iii) despite negated statements, every $\rho df_\bot^\neg$ knowledge base is satisfiable; (iv) the $\rho df_\bot^\neg$ entailment decision procedure is obtained from $\rho df$ via additional inference rules favouring a potential implementation; and (v) deciding entailment in $\rho df_\bot^\neg$ ranges from P to NP.