W3C home > Mailing lists > Public > semantic-web@w3.org > May 2017

Re: I am.

From: Pat Hayes <phayes@ihmc.us>
Date: Fri, 19 May 2017 11:32:51 -0700
Message-Id: <0551A483-9BDD-4210-A6FE-34C08FE4ADC8@ihmc.us>
Cc: Martin McEvoy <martin@weborganics.co.uk>, Sarven Capadisli <info@csarven.ca>, Semantic Web <semantic-web@w3.org>
To: Simon Spero <sesuncedu@gmail.com>

> On May 19, 2017, at 9:58 AM, Simon Spero <sesuncedu@gmail.com> wrote:
> 
> On Fri, May 19, 2017, 1:58 AM Pat Hayes <phayes@ihmc.us <mailto:phayes@ihmc.us>> wrote:
> Indeed, and also the simple ‘am’ is an assertion of existence, which cannot be said (or, perhaps, is trivial) in RDF because RDF (like most formal logics) is based on a semantics in which anything that is named, exists. 
> 
> Adding RDFS does allow for partial use Quine's hack,  if proper names are converted to "classes" (I think he'd accept the notational shorthand for a universally quantified unary predicate, but not the intensional gubbins). 
> 
> PatHayes a Class; subClassOf Person . 
> _:x a PatHayes . 

Hmm, never thought of myself as a class before. But that does allow for non-existence, ie the empty class. Neat hack, but it only works if everyone agrees to use it. Unlikely. 
> 
> NonCurmudgeonlyPatHayes subClassOf PatHayes, NonCurmudgeonlyThing
>  #     ;sameAs Unicorn #  wvoq@harvard.edu <mailto:wvoq@harvard.edu>
> . 
>> 
>> Would it be cheating to represent things as an RDF node of type Cycl / IKL / CLIF  assertions, with the actual representation in an attached literal? 
> 
> Well, it’s legal, but it rather takes away the point of using RDF in the first place. Any RDF entailments will not be Cyc/IKL/CLIF entailments and vice versa. 
> 
> But cf the OWL 2 Direct Semantics,  with the  RDF -> Abstract Syntax mapping. 

Yes, but you still won’t actually get the Cycl/whatever inferences happenig by doing OWL inference on text literals. The RDF/OWL is just carrying the other syntax aound like paintings on pieces of china. Its not doing anything with them.

Pat
Received on Friday, 19 May 2017 18:33:29 UTC

This archive was generated by hypermail 2.3.1 : Friday, 19 May 2017 18:33:30 UTC