- From: Antoine Zimmermann <antoine.zimmermann@emse.fr>
- Date: Tue, 21 Jan 2020 09:48:21 +0100
- To: Patrick J Hayes <phayes@ihmc.us>
- Cc: Henry Story <henry.story@bblfish.net>, Jos De Roo <jos.deroo@agfa.com>, semantic-web <semantic-web@w3.org>
Le 20/01/2020 à 21:14, Patrick J Hayes a écrit : > > [skip] >> >> Simple entailment can be translated to FOL using either: relations of >> arity 2 (where the predicate IRI is used as a binary FOL predicate); >> or using a single relation of arity 3 (where <s> <p> <o> becomes >> Triple(s,p,o)). >> >> If you assume the set D to be empty, then RDF-entailment and >> RDFS-entailment can be translated to FOL as well. But as soon as you >> add datatypes and literals to the picture, it's a whole different >> story. RDF 1.1 Semantics imposes that RDF(S)-entailment MUST recognise >> xsd:string and rdf:langString. This totally cripples the RDF-to-FOL >> translation. > > Well, as with RDFS, it maps into a FO axiomatic theory (in this case, > countably infinite) but the actual reasoning is still simple FO > reasoning. It does not require a different /logic/. Right, but the RDF-to-FOL translation of D-entailment, even with simple datatypes like xsd:string, requires some tricks (see below). > > Pat > >> >> For instance, in RDFS-entailment recognising >> {xsd:string,rdf:langString}, is the following graph consistent >> (written in Turtle, assuming the obvious prefixes)? >> >> rdfs:Resource rdfs:subClassOf xsd:string . >> >> (left as an exercise to the reader :) > > :-) I am not sure myself. If a language-tagged string a kind of string? In D-entailment, the IRIs in D have to be interpreted as the datatype they are associated to. For instance, in RDFS semantics, xsd:string has to be interpreted as the datatype (LS,VS,L2V) where LS -- the lexical space of xsd:string -- is the set of valid XML schema strings, VS -- the value space of string -- is the same as LS, and L2V -- the lexical-to-value mapping -- is iedentity. So, the universe of all RDFS interpretations has to contain the triple (LS,VS,L2V). This triple is clearly not a sequence of UNICODE character itself. But the statement: rdfs:Resource rdfs:subClassOf xsd:string . in RDFS means that all resources (all things in the universe) must belong to the extension of the class denoted by xsd:string. In turn, D-entailment says that all datatypes in D are classes whose extension is exactly their value space (crucially, this last constraint is removed by ter Horst in his D*-entailment [1]). So, the triple above constrains everything in the universe to be character strings, whereas xsd:string is required to denote a triple of set-theoretic structures. Obviously, it cannot be the case, so there are no RDFS-models of such triple. Thereby, it is RDFS-inconsistent. D-entailment can be horribly complicated. Consider this example: rdf:Property rdfs:subClassOf xsd:boolean . Is this RDFS-recognising {xsd:boolean}-consistent? --AZ >> >> Datatype semantics in RDF is horribly complicated to properly handle >> and I doubt there is (and even will be) any reasoner that correctly >> and completely implements datatype entailment as defined in the >> standard. Notably, in his excellent paper from 2005, Herman J. ter >> Horst provides a sound and complete algorithm for RDFS with datatype >> entailment, but he dismisses the official D-entailment semantics to >> propose a more computable one, that he calls D*-entailment [1]. >> D*-entailment is what practical reasoners usually implement, if they >> support datatypes at all. >> >> However, De Bruijn and Heymans show that, if we are cautious in >> selecting supported datatypes, it is still theoretically possible to >> reason efficiently with standard D-entailment [2]. But first, they do >> not provide an effective algorithm,* and second, it forbids quite a >> lot of datatype combinations. >> >> >> * Only indirectly via translation to F-Logic, and it may not be >> efficient at all. >> >> >> --AZ >> >> >> >> [1] Herman J. ter Horst. Completeness, decidability and complexity of >> entailment for RDF Schema and a semantic extension involving the OWL >> vocabulary. In Journal of Web Semantics, Volume 3, Issues 2–3, October >> 2005, pages 79-115. >> [2] Jos De Bruijn and Stijn Heymans. Logical Foundations of RDF(S) >> with Datatypes. In Journal of Artificial Intelligence Research, Volume >> 38, August 2010, pages 535-568. >> > > [skip] >-- Antoine Zimmermann Institut Henri Fayol École des Mines de Saint-Étienne 158 cours Fauriel CS 62362 42023 Saint-Étienne Cedex 2 France Tél:+33(0)4 77 42 66 03 Fax:+33(0)4 77 42 66 66 http://www.emse.fr/~zimmermann/ Member of team Connected Intelligence, Laboratoire Hubert Curien
Received on Tuesday, 21 January 2020 08:48:30 UTC