Dear Matthew, Gian Piero, and Doug, (01)
MW
> Why is it important that a record is say a temporal relationship rather
> than a qua-individual? Would you have the same objection if it were a state?
>
> I disagree that there is a proliferation of records, for accidental
> properties you have to have some record that represents the period of time
> the object has the accidental property. The rest is just about what you want
> to call it. (02)
I agree. The fundamental unit of ontology or knowledge representation
of any kind is the relation (which includes propositions as 0-adic
relations, types as monadic relations, and classes as monadic relations
that determine a set in each model). (03)
You can classify those relations in many different ways. Some of them
are true of some x for as long as x exists. Others are true of x
during some interval. Both are important. (04)
DF
> One solution would be to attach the statements to temporal contexts,
> so that instead of storing: (Joe instanceOf Student) -- which can
> not be timelessly true, it stores
> ((that (Joe instanceOf Student) holdsDuring (GregorianYearFn 2011))
> Another solution is to define Student as a 'role' instead of as a
> subclass of 'Person' and make 'role' disjoint with 'class'. Then
> a statement such as
> (Joe hasRoleDuring Student (GregorianYearFn 2011))
> could be stored. [Make it more complicated if you like triples.] (05)
I would dispense with the notion of class, since it just invites
confusion. Treat monadic relations as fundamental and classes
as a notion of relation+set for use in special purpose languages. (06)
GPZ
> On the other hand, I am not interested in Facebook or in extracting knowledge
> from natural language but, mainly, in high-level inferencing starting from
> complex conceptual representations. (07)
Every version of logic is an abstraction from some aspect of natural
language semantics. The data used in that inferencing comes from some
source that is defined or explained in NLs. The results derived from
that inferencing must translatable and expainable in NLs. (08)
Any ontology that cannot be translated to and from NLs is meaningless. (09)
GPZ
> I am surprised you have never been faced, e.g., with an apparently
> stupid but concrete problem like that of giving meaningful "names"
> to the instances created at run time - (010)
The basic solution in both KR and DB systems is to use internal
"surrogates" for the named entities -- e.g. GENSYM in LISP. (011)
The meaningful names come from the NL "named entities" and phrases.
By appropriate inferencing, you identify which internal surrogates
are distinct or coreferent, and you map them to the NL. (012)
John (013)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (014)
|