RE: live meaning and dead languages

Richard Cyganiak wrote:

>On 8 Feb 2009, at 18:11, rick wrote:
>> As I have written before, the model theory on which the semantic web
>> is based is defined in Alfred Tarski's Semantic Conception of Truth.
>
>Rick, that's overstating the role of model theory on the Semantic Web.
>The formal semantics of RDF, as defined in [1], are based on model
>theory. But a lot of the deployed usage of RDF considers it simply as  
>a distributed graph data model, and ignores (or even violates) the  
>model theoretic semantics. Various non-RDF technologies, such as Topic  
>Maps or microformats are often lumped under the Semantic Web umbrella  
>as well.
>
>So, only a particular part of the Semantic Web technology portfolio is  
>based on model theory.

Certainly. 

It surprises me that some people seem to believe, just because the semantics of RDF (and OWL) is defined on model theoretic grounds, that the Semantic Web itself is somehow based on model theory, i.e. that model theory is sort of a philosophical foundation for the Semantic Web. 

My own understanding has always been that the single reason for choosing a model theoretic approach for the semantics of RDF and OWL is simplicity and clarity of specification. And, btw., as for most other technologies that are targeted to a mass market (computers, mobile communication, driver navigation, etc.), I believe that the actual technical specification behind the technology can easily be completely ignored by almost everyone in the user community without any harm, unless one actually wants to create systems based on the technology. 

Here are some points that IMO talk against the philosophical aspect:

* RDF(S) semantics could have been equally well defined directly by a set of triple rules (the one given in §7 of [1]), instead of first defining a model theory and then deriving the rule set from it. In this case, there would not be any model-theory around, and no one could say that the Semantic Web is based on such a model theory.
 
* If I would be asked to develop an ontology language comparable to that of OWL for a different purpose than the Semantic Web, for example for the restricted domain of health care and life science (HCLS) research, I would most probably also use a model theoretic approach, simply because it's the most straightforward approach I can think of for such a complex logic formalism. An additional advantage is that one can fairly well work with such a model theory, e.g. proving that the resulting ontology language actually conforms to all the original requirements. But eventually having such a model theoretically grounded ontology language wouldn't, of course, mean for HCLS that this field is suddenly based on a model theory.
 
In both cases, using a model theory is just a specific technical design decision by the group of people who create the specification, probably without having any (deep) philosophical considerations in mind.

>I agree, however, that it's the part that can  
>benefit most from armchair philosophizing.

Well, hardly. :-)

>Have fun,
>Richard

Cheers,
Michael

>[1] http://www.w3.org/TR/rdf-mt/

--
Dipl.-Inform. Michael Schneider
Research Scientist, Dept. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: schneid@fzi.de
WWW  : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555

==============================================================================

FZI Forschungszentrum Informatik an der Universität Karlsruhe
Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
Tel.: +49-721-9654-0, Fax: +49-721-9654-959
Stiftung des bürgerlichen Rechts
Stiftung Az: 14-0563.1 Regierungspräsidium Karlsruhe
Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael Flor,
Prof. Dr. rer. nat. Dr. h.c. Wolffried Stucky, Prof. Dr. rer. nat. Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus

==============================================================================