Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: SUO: RE: CYC event vs. SUMO Process -- really different?



At 08:59 AM 6/13/2003 -0700, Richard Cooper wrote:

Patrick Cassidy wrote:

[snip]
> > However, in OpenCyc, while I can say
> >
> >         (#$objectFoundInLocation #$MySwissArmyKnife #$MyLivingRoom)
> >
> > I cannot say
> >
> >         (#$objectFoundInLocation
> &$MikeReprogramsHisVCROnJune102003 #$MyLivingRoom)
> >
> > or even:
> >
> >         (#$inRegion &$MikeReprogramsHisVCROnJune102003
> #$MyLivingRoom)
> >
> > because #$MikeReprogramsHisVCROnJune102003 is not an
> instance of #$SpatialThing or #$SpatialThing-Localized.
> >
> > In OpenCyc one uses a different predicate, #$eventOccursAt,
> to relate events to the places at
>  > which they occur, but it is neither a generalization nor a
> specialization of the predicates
>  > used to specify relative spatial positions of physical objects,
> i.e., #$inRegion and its
>  > specializations.  Events or processes are not located in the same
> way that physical objects are
>  > in OpenCyc; they are in SUMO.
>

Not having learned the complexities of OpenCyc or SUMO, it
seems bad design to either

1. Require events to occur at locations (in regions) because
   there are events that are not spatial at all, e.g.,
   "what if Bob thought about Mary's birthday party?"
   is a hypothetical, with no spatial relationships at all.
   It shouldn't be necessary to represent this in any
   spatial way. 


But I don't think that either &%Process or #$Event are *required* to occur at a location.

Note, also, that the hypothetical you state is not (at least not obviously to me) itself an event or process, it may be a state of affairs or some such.  Also, thinking about it is an #$Event and a %&Process (presumably) and were it to actually occur it would be an #$Event or %&Process.  Note that in either case many would argue that such events do indeed have locations, i.e., in some human's cognitive gunk.  (Having said that, I don't disagree with the more general point, i.e., that we may want to allow for non-localized events but I'm not sure that either ontology rules them out.) 

or

2. Make the mathematics of temporal logic unavailable to
   any class of events, whether spatial or nonspatial in
   description. 
I don't think that either ontologies have done so.


Something seems wrong in this example.

Could you say a bit more about what's wrong?


HTH,
Rich