RE: SUO: RE: Re: More KIF-ified Ontology Content
Pat,
. Thanks for this further elaboration, with which I agree.
. The problem to me is that it seems what you have described
can be applied to objects as well as relations.
. Basically it seems to me you are discussing histories, or
memories of time-histories. I am not seeing at this stage why such an
approach does not apply equally to both objects and relations.
. I can easily think about my great-grandparents, none of whom
I ever met, and their doings. The same applies to my grandparents, all now
deceased, but some of whom I did met. Of course the same applies to
"inanimate" objects such as cars, etc.
Cheers Graham Horn
National Data Standards Unit
Australian Institute of Health and Welfare
================================================
Phone: 02.6244.1094
Fax: 02.6244.1199
Email: Graham.Horn@aihw.gov.au <mailto:graham.horn@aihw.gov.au>
-----Original Message-----
From: pat hayes [mailto:phayes@ai.uwf.edu]
Sent: Friday, January 05, 2001 5:29 AM
To: Horn, Graham; Matthew.R.West@IS.shell.com
Cc: standard-upper-ontology@majordomo.ieee.org
Subject: RE: SUO: RE: Re: More KIF-ified Ontology Content
>Dear Matthew,
> . Thanks for this explanation. I certainly agree about
>the restrictions with 3D models
>
> . However, I still don't understand why relations do not also
>have a time dimension. For example, what about two beings which have
>friendly relations at some periods and hostile ones at other times?
In the 3d (actually 3.5d would be a better way to put it) ontology
one might write
(Friendly A B t1) and (not (Friendly A B t2))
or maybe
(Holds (Friendly A B) t1) and (not (Holds (Friendly A B) t2)). In the 4d
ontology one distinguishes the 'parts' of each person at the relevant times
- thinking of people, like everything else in this ontology, as extended
through time and so having 'temporal slices' - and then would write
something like:
(Friendly A@t1 <mailto:A@t1> B@t1 <mailto:B@t1> ) and (not (Friendly A@t2
<mailto:A@t2> B@t2 <mailto:B@t2> ) where X@Y <mailto:X@Y> means the
temporal Y-slice of the thing X. The relations are now timeless (they don't
have a temporal parameter and they don't hold at one time but not at
another), but the things themselves are temporally indexed. For more
details of the various possibilities, take a look a the discussion in
http://www.coginst.uwf.edu/users/phayes/TimeCatalog1.ps
<http://www.coginst.uwf.edu/users/phayes/TimeCatalog1.ps> The advantages of
this framework can be best appreciated by tryi ng to use it to axiomatize
things like machinery and liquid flow. For example, it gives one the ability
to talk about the temporal 'shapes' of various actions, and distinguish
velocities from accelerations, etc.. The disadvantages are more subtle, and
include the fact that it is just as easy to write things like (Friendly A@t1
<mailto:A@t1> B@t2 <mailto:B@t2> ) which are at best odd and at worst
meaningless. CYC uses a basic temporal framework of this sort, and it
enriches it with a lot of relational categories which are designed to rule
out oddities like this, such as 'isotemporal' relations which are required
to only hold between things which exist during the same time interval. There
are other complications as well concerned for example with concepts which
intrinsically involve certain kinds of change, such as divorce. (If marriage
is isotemporal, then the death of a spouse ends the marriage, but it doesnt
make the surviving spouse a divorcee.)
Pat Hayes
---------------------------------------------------------------------
IHMC (850)434 8903 home
40 South Alcaniz St. (850)202 4416 office
Pensacola, FL 32501 (850)202 4440 fax
phayes@ai.uwf.edu <mailto:phayes@ai.uwf.edu>
http://www.coginst.uwf.edu/~phayes <http://www.coginst.uwf.edu/~phayes>