[Top] [All Lists]

Re: [ontolog-forum] Ontologies for hybrid connectionist-semantic systems

To: ontolog-forum@xxxxxxxxxxxxxxxx
Cc: Uta Priss <u.priss@xxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Fri, 03 Dec 2010 23:23:54 -0500
Message-id: <4CF9C25A.8030403@xxxxxxxxxxx>
Leo,    (01)

I have no quarrel with anyone who chooses to use DLs or frame-based
languages when they really understand the options.  There are often
good reasons for using DLs, and I would have no objection whatever.    (02)

> OWL specifically is both frame-based and DL in nature, by design.
> You may not like it, but that is a different story.    (03)

I am not expressing my preferences.  I am saying that the hype
behind OWL has led many people to adopt it when they should be
using something much simpler and more efficient.    (04)

> Because indeed description logics are primarily useful for
> classification reasoning. And often one wants more.    (05)

Yes indeed.  Your examples (which I did not copy) prove my claim:
you need a *hybrid* such as a classification system combined with
a more expressive language, such as a rule-based system.    (06)

And for many (perhaps a majority) of applications for which people
have been told to use OWL, there are vastly better classification
systems.  A good example is Formal Concept Analysis (FCA).  See the
FCA home page, which is maintained by Uta Priss (on cc list above):    (07)

    http://www.upriss.org.uk/fca/    (08)

BFO is a prime example of an ontology for which OWL is *worse*
than useless.  See the BFO ontology at the following web site:    (09)

    http://www.ifomis.org/bfo    (010)

Following is the BFO definition for 'Continuant' expressed in OWL:    (011)

  <owl:Class rdf:about="&snap;Continuant">
     <rdfs:subClassOf rdf:resource="&bfo;Entity"/>
         <owl:unionOf rdf:parseType="Collection">
           <owl:Class rdf:about="&snap;DependentContinuant"/>
           <owl:Class rdf:about="&snap;IndependentContinuant"/>
           <owl:Class rdf:about="&snap;SpatialRegion"/>
     <owl:disjointWith rdf:resource="&span;Occurrent"/>
     <rdfs:label rdf:datatype="&xsd;string">continuant</rdfs:label>
     <rdfs:comment rdf:datatype="&xsd;string">Definition: An entity 
[bfo:Entity] that exists in full at any time in which it exists at all, 
persists through time while maintaining its identity and has no temporal 
     <rdfs:comment rdf:datatype="&xsd;string">Examples: a heart,
a person, the color of a tomato, the mass of a cloud, a symphony 
orchestra, the disposition of blood to coagulate, the lawn and 
atmosphere in front of our building</rdfs:comment>
     <rdfs:comment rdf:datatype="&xsd;string">Synonyms: 
   </owl:Class>    (012)

Note that the complicated statements in the last few lines of this
definition are enclosed in comment brackets.  They sound meaningful,
but none of that meaning is recognized or processed by OWL tools.
The OWL statements in the first 13 lines of that definition can
be translated to the following sentences in Controlled English:    (013)

   Every continuant is an entity.
   Every continuant is either
     a collection,
     a dependent continuant,
     an independent continuant,
     or a spatial region.
   No continuant is an occurrent.    (014)

These 7 lines can be read and understood by any physician who might
use the BFO ontology.  Furthermore, they (or their restatement in
a table) can be processed very simply and efficiently by FCA tools.    (015)

If you look through the entire BFO ontology at the above web site,
you'll notice that not a single statement goes beyond FCA or Aristotle.    (016)

Furthermore, FCA automatically generates a minimal lattice that is
guaranteed to be consistent.  Many OWL users actually apply the FCA
tools to test OWL statements for consistency.  But if you have a
hybrid with rule-based system, anything that goes beyond FCA can be
handled by the rule system.  (As I said, OWL is either overkill
or underkill.)    (017)

> Of course they [PowerSet] don't use RDF or OWL (at least to my limited
> knowledge), since the latter are subsequent to LFG and HPSG.    (018)

I have talked with them, and they had access to the Cyc ontology,
but they did not use any of the Cyc axioms.  The only thing they
used was the type hierarchy.  If they didn't have Cyc, they could
have used FCA tools.  OWL would add complexity, but no added value.    (019)

John    (020)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (021)

<Prev in Thread] Current Thread [Next in Thread>