[Top] [All Lists]

Re: [ontolog-forum] Using controlled natural languages for ontology

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Tue, 15 Mar 2011 10:19:29 -0500
Message-id: <4D7F8381.2080401@xxxxxxxxxxx>
Rich and Doug,    (01)

>> ... events and situations, as the full expression of those symbols in
>> language, are not SEMANTIC - they are syntactic just as surely as verbs
>> and nouns are syntactic.    (02)

> I hold that the concepts of Event and Situation are semantic concepts
> since relations can be defined for them that apply to their subtypes.    (03)

I agree with Doug, but the critical issue is to distinguish the object
language from the metalanguage.  To be specific, let's assume that
English is the object language and Common Logic is the metalanguage.    (04)

In CL, we could write rules that specify the grammar of English.
For that, we would use relations like (Verb x) or (Noun x).    (05)

But in CL, we could also write rules that specify the semantics
of English.  For that, we would use relations like (PhysObject x),
(Event x), (Situation x), etc.    (06)

CL, as a system of logic, has both a syntax and a semantics.
Those expressions, written in CLIF, use one kind of syntax.
Their translation to CGIF, would have a slightly different
syntax:  (Verb ?x) (Noun ?x) (PhysObject ?x) (Event ?x)...    (07)

But both the CLIF versions and the CGIF versions would have
the same semantics:  A domain that contains three subdomains:    (08)

  1. A subdomain that consists of instances of English grammatical
     constructions represented by n-tuples of syntactic relations,
     such a Verb and Noun.    (09)

  2. A subdomain that consists of instances of semantic constructions
     represented by n-tuples of semantic relations.    (010)

  3. A subdomain that consists of instances of relations that relate
     English syntax to English semantics.  Those n-tuples would relate
     relations about words and sentences to relations about objects,
     events, situations, and other kinds of things in the world.    (011)

These are technical issues about logic.  But I agree with Rich that
in order to make these logical symbols relate to the actual world,
there must be some theory about the way humans and/or robots
relate their mechanisms of perception and action to their
use of language and their behavior in the world.  But that is
a separate issue about symbol grounding.  Following are some
slides that I presented about such issues:    (012)

    http://www.jfsowa.com/talks/situ.pdf    (013)

Given all that, I would also like to comment about primitives.
In CL, the word 'primitive' is not a technical term.  That is
a metalevel term about theories of semantics that choose to
anoint certain relations with the honorific status of Primitive.    (014)

That is a totally different issue.  We've discussed it many times
on this email list, and it always gets us into a swamp of confusions.
I don't have any use for the term, and I would prefer to avoid it.    (015)

If would be happy to say that Event and Situation are important
relations that would be fairly high in the ontology.  But I would
prefer not use the word 'primitive' to talk about them,    (016)

John    (017)

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    (018)

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