Re: Multiple Inheritance and "rigid" properties
Adam, thanks for the reply. I made the distinction so that it can be kept
in mind. As far as I can judge, terms of either sort may be useful.
Yours, Martin
e-business Services PHONE: +44-1707-363090 (Int 7-453090)
Rosanne House (RH2A) FAX: +44-1707-338732
Welwyn Garden City Internet: martin_king@uk.ibm.com
GB - AL8 6UB IBMMAIL: GBIBM3WS
apease <apease@teknowledge.com> on 22/06/2000 16:47:39
Please respond to apease <apease@teknowledge.com>
To: Martin King/UK/IBM@IBMGB, standard-upper-ontology@ieee.org
cc: Hugh Darwen/UK/IBM@IBMGB
Subject: Re: Multiple Inheritance and "rigid" properties
Martin,
I'm not sure I'm understanding you so let me try to rephrase. Are you
suggesting that only terms which have an extension should be included in
the SUO, or are you just mentioning the distinction between terms which are
primarily extensional or primarily intentional so that people will keep it
in mind as we construct the SUO?
Adam
At 09:15 AM 6/22/2000 +0100, martin_king@uk.ibm.com wrote:
>Adam, I would not object to any proposed concept on the grounds that it
>was defined in an axiomatic logic.
>
>I guess I left a hostage to fortune in the word "abstract". I guess I see
>a difference between Circle where I think of the word referring primarily
>to the abstraction and only secondarily to the necessarily approximately
>round shapes observed in the real world, and Person, where I see the word
>as primarily referring to the things in the real world, and only
>secondarily to the abstraction. Hence I would take the definition of
>Circle from Euclidean Geometry and consider anything to be a circle if and
>only if it conformed precisely to the theorems of that geometry.
>Conversely I cannot ever remember seeing a useful definition of Person
>which does not ultimately depend on someone declaring "Tom Jones is a
>Person". I think this corresponds to the distinction made between
>intentional and extensional definitions of members of a class. On this
>basis I would see the definitions of TangibleThing and Information as more
>likely to be like that of Person than that of Circle, and therefore your
>doubt about whether I would object ot them empty.
>
>If you can point to definitions of TangibleThing, Information, or even
>Person that challenge the distinction I am making,, I would be interested
>to see them.
>
>Yours, Martin
>
>e-business Services PHONE: +44-1707-363090 (Int 7-453090)
>Rosanne House (RH2A) FAX: +44-1707-338732
>Welwyn Garden City Internet: martin_king@uk.ibm.com
>GB - AL8 6UB IBMMAIL: GBIBM3WS
>
>
>apease <apease@teknowledge.com> on 21/06/2000 15:23:27
>
>Please respond to apease <apease@teknowledge.com>
>
>To: Martin King/UK/IBM@IBMGB, standard-upper-ontology@ieee.org
>cc: Hugh Darwen/UK/IBM@IBMGB
>Subject: Re: Multiple Inheritance and "rigid" properties
>
>
>
>
>At 09:49 AM 6/21/2000 +0100, martin_king@UK.IBM.COM wrote:
>
>
>
>[snip]
>
> >ple chosen to illustrate inheritance is not naturally
> >occurring objects, but the concepts of circle and ellipse which are
> >abstract concepts logically derived from a set of mathematical axioms.
> > From discussion with one of the authors, it is clear they did not
> >consciously choose these because of my argument above, but I believe he
> >would see it as much more complex, if not infeasible, to base their
> >discussion of inheritance on concepts classifying naturally occurring
> >objects. Finally I suggest the majority of the concepts we are likely
to
> >want to include in an SUO would be classifications of naturally
occurring
> >objects and not abstract objects defined in an axiomatic logic.
>
>Martin,
> Just to clarify, I trust you're not opposed to including concepts like
>Circle, or TangibleThing or Information in the SUO along with concepts
such
>as Person, Metal, and Country?
>
>Adam
>
>-----------------
>Adam Pease
>Teknowledge
>(650) 424-0500 x571
-----------------
Adam Pease
Teknowledge
(650) 424-0500 x571