Thread Links Date Links Thread Prev Thread Next Thread Index Date Prev Date Next Date Index Re: SUO: RE: Re: Comment #13 - 'NL' - To: "Horn, Graham" , standard-upper-ontology@ieee.org - Subject: Re: SUO: RE: Re: Comment #13 - 'NL' - From: apease - Date: Thu, 14 Sep 2000 15:46:02 -0700 - Cc: "Christensen, Joe" - In-Reply-To: - Sender: owner-standard-upper-ontology@ieee.org Graham, I think you are addressing in part a difficult issue that others have raised. I believe we need to be very clear about distinguishing an ontology from a natural language lexicon. Words in a lexicon can be ambiguous because humans use the same words in different ways, but terms in an ontology must have unique definitions. We may map terms in an ontology to a lexicon but the two artifacts are distinct. This also is related to the discussion that is occurring about dictionary definitions. The terms or labels we use in the SUO will just be markers to aid human understanding. A computational system will not care ultimately if we have a term 'Action' or a term 'Foo' if the axioms associated with the symbol are the same. Likewise, although evocative names for terms will be helpful, it is the logical definitions associated with the name that must be unambiguous. That said, I support the proposal to use long and descriptive names for concepts. Adam At 04:38 PM 9/13/2000 +1000, Horn, Graham wrote: >Jon Awbrey wrote: > >"How do you see the ontology of ambiguity working out? >So far I have, modulo a bit of guesswork on my part: > > : [Ambiguities] = [Context Specific Ambiguities (COSA)] + >[???]. > > : [COSA] = [Ambiguities Caused by Context?] > U [Ambiguities Solved by Context?] > U [???]." > >Jon, >. Actually, I think I can be more specific by saying that contexts >often result in particular: >* objects having particular potential uses, and >* processes having particular consequences. > >. For example, detonating a stick of dynamite in a mine can be used to >loosen rock and/or ore. The same action at sea may provide a haul of fish. >When devoid of context, the expression "detonating a stick of dynamite" >doesn't necessarily imply either. > >. Other types of ambiguities are caused by words with more than one >meaning, often called homonyms. Consider "sanctioning someone". This could >mean allowing them to do something, or penalising them for doing something. >The problem is that sanction derives from the concept of "setting apart", >and the purposes of the setting apart are implied, and are varied. I am >suggesting an approach that would avoid such implied extrapolations of >meaning, and replace the first meaning with a word like "allowed", and the >second with a word like "penalised". > >. This is what I mean by proposing that we are trying to come up with >a "simplified language". Basically it will ideally be more succinct and >unambiguous. > > > >Cheers Graham Horn > >Australian Institute of Health and Welfare >================================================ >Phone: 02.6244.1094 >Fax: 02.6244.1199 >E-mail: Graham.Horn@aihw.gov.au > > >-----Original Message----- >From: Jon Awbrey [mailto:jawbrey@oakland.edu] >Sent: Wednesday, September 13, 2000 2:01 PM >To: standard-upper-ontology@ieee.org >Subject: SUO: Re: Comment #13 - 'NL' > > >Horn, Graham wrote: > > > > While I acknowledge that some ambiguities are context specific, not all >are. > > While I'm happy with the proposed words, I don't feel they go far enough. > > > > I feel disambiguation should get a bit further attention. > > > > Cheers Graham Horn > > > > Australian Institute of Health and Welfare > > ================================================ > > Phone: 02.6244.1094 > > Fax: 02.6244.1199 > > E-mail: Graham.Horn@aihw.gov.au > > ><...> > >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > >Graham, > >How do you see the ontology of ambiguity working out? >So far I have, modulo a bit of guesswork on my part: > > : [Ambiguities] = [Context Specific Ambiguities (COSA)] + >[???]. > > : [COSA] = [Ambiguities Caused by Context?] > U [Ambiguities Solved by Context?] > U [???]. > >Please don't take this notation too seriously, as I am making it up as I go, >but that is just about as succinctly as I know how to summarize my array of >questions at this point. > >Even more of a wild guess here -- okay, it's more of a projection of my own >fixed idea -- do you recognize that the indexical context of interpretation >is only one of the many "pragmatic" factors that can affect the senses that >an interpretive agent or community will immediately seize on or ultimately >settle on, and that, as a consequence, the indicated task, in its broadest >terms, is to carry out an analysis that gets at the ontology of >interpreters, themselves? > >On a related note, I think that the relationships among Ambiguity, Entropy, >Uncertainty, and Information as the "average uncertainty reduction on >receiving each symbol" (AURORES?) ought to get some extra concerted >attention -- but maybe later. > >Time to do my dreaming off-line ... > >Au reverie, > >Jon Awbrey > >~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ ----------------- Adam Pease Teknowledge (650) 424-0500 x571 - References: - SUO: RE: Re: Comment #13 - 'NL' - From: "Horn, Graham" - Prev by Date: SUO: RE: Naming Conventions - Next by Date: Re: SUO: Comment #15 - ''English" - Prev by thread: SUO: RE: Re: Comment #13 - 'NL' - Next by thread: Re: SUO: RE: Re: Comment #13 - 'NL' - Index(es): - Date - Thread