SUO: RE: Metaphysical choices - position. mereology and constitution
Chris,
See my comments below.
-Ian
> -----Original Message-----
> From: Chris Partridge [mailto:chris_partridge@csi.com]
> Sent: Tuesday, June 11, 2002 9:04 AM
> To: standard-upper-ontology@ieee.org
> Cc: apease@ks.teknowledge.com
> Subject: SUO: Metaphysical choices - position. mereology and
> constitution
>
>
>
> Adam,
>
> I have changed the name of the thread - as it seems to have
> moved on and
> this forms a natural break.
>
> I originally made the comment to illustrate some points I
> have been making
> for some time. In no particular order:
> 1) There are some metaphysical questions that can usefully
> inform the design
> of an ontology (this is in response to Ian's angels on a pin
> head comment).
> That is not to say that some metaphysical questions may be as
> irrelevant as
> angels to building SUO type ontologies.
> 2) The fact that useful discussion can take place without it
> being framed in
> terms of axioms - though it will typically need to be cashed
> out in the end
> in some kind of formalisation (such as axioms).
> 3) That there is some useful top level regimentation that
> still needs to be
> done to the SUMO.
>
> It seems to me that unless I explain the underlying issue
> here, pointing at
> inconsistencies or multiple interpretations of particular
> axioms will appear
> a bit ad hoc. So I will start by trying to sketch quickly the
> issue - which
> will be familiar to some people.
>
> Bill Burkett made a point to you some time ago that it was
> difficult to
> ensure against misinterpretation of an ontology. And I think
> it was agreed
> that reducing the risk of misinterpretation was 'a good thing'. I am
> guessing (and I will rely on your and Ian's knowledge of it)
> but I believe
> that the SUMO, as it stands, allows for some situations to be
> ontologised in
> a number of ways, that all conform with the SUMO. And that
> regimentation of
> the top level can eliminate this 'ambiguity'.
If what you mean by "regimentation" is that one ought to make explicit,
whenever possible, the principles and distinctions that are guiding the
development of the ontology, then I agree with you. Of course, you're not
going to find too many people who will challenge the claim that clarity
tends to eliminate ambiguity.
>
> To see the issue that arises as a result of the original
> change made to
> position, consider the following topical, but trivial,
> example. The Queen
> (of England) at her Jubilee Party eats a cake. We are happy
> to say that it
> is true that there is a cake eating event and that Queen
> Elizabeth eats a
> cake and that Elizabeth Windsor eats a cake. Can we focus on
> this 'cake
> eating event'
>
> Under the original SUMO scheme - let's call it the O-C scheme - Queen
> Elizabeth is a position that is a role of Elizabeth Windsor.
> So there is a
> single 'cake eating' event.
>
> Under the new SUMO scheme Queen Elizabeth is a Position that is a
> CognitiveAgent and Elizabeth Windsor is a Human that is a
> CognitiveAgent. I
> presume also that SUMO somewhere has an 'occupies' relation
> that relates the
> two cognitive agents (is the occupiesPosition?).
>
> My question is whether it is possible within this second
> scheme to determine
> whether there is one or two cake eating events - and if so
> how? And whether
> this is what we want to do a conscious policy?
Actually, even if there are two cognitive agents here, viz. the Queen and
Elizabeth Windsor, it does not necessarily follow that there are two
distinct events. There are many events that have more than one agent, and
it may be that the particular event of cake-eating that you cite here is one
of these. However, I doubt that this will appease you. You'll still be
worried about how many agents there are in this example. My response is
that I just don't know how many agents there are here. In fact, this
situation seems perfectly parallel to the classic problem about identity,
viz. morning star and evening star. Just as it had to be established by
empirical investigation that both of these "stars" are in fact the planet
Venus, we might have to do some work to determine whether the Queen of
England and Elizabeth Windsor are the same person. In any case, I don't
think this is a matter to be settled a priori by one's preferred ontology.
>
> Could you tell us what is your (and maybe Ian's) intuition is
> (maybe based
> upon SUMO)? And how is this backed up in SUMO?
Note that we've been over this ground before. I tried to give a
comprehensive summary of the revised SUMO view on these matters in an old
message to the SUO list: http://suo.ieee.org/email/msg06357.html.
>
> There are quite a lot of intertwined issues. But keeping
> things simple. Some
> people will say that there are two completely distinct
> CognitiveAgents and
> that the occupies relation that links them is a kind of
> constitution. Under
> this strategy one typically ends up with two events. Others
> will say that
> the occupies relation is one of sharing temporal parts and so
> is similar to
> a temporalised (improper) part of relation. Under this strategy one
> typically ends up with one event. It seems to me that would
> be useful to
> formalise the preferred strategy to ensure that conforming
> ontologies end up
> with the same number of objects.
You seem to be assuming here that an ontology should legislate the precise
number of objects in the world. Otherwise, how would "conforming"
ontologies always agree about the number of objects in the world? If you
are making this assumption, it strikes me as utterly wrong. There are many
things that we in our current state of ignorance don't know about the world,
and the ontology, I think, should leave room for settling these matters as
the facts stack up. To go back to the morning star/evening star example, do
you suppose an ontology constructed before these two things were known to be
the same thing, should have taken a stand about whether they were the same
thing? If it didn't take such a stand, how could it satisfy your criterion
that "conforming ontologies end up with the same number of objects".
>(NB I could not find
> 'constitution' or
> 'part of' using the browser - probably my fault).
I'm not sure what you mean by "constitution". As for 'part of', look under
'part'.
>
> The metaphysical choice I referred to in my original mail
> turns on whether
> one uses part of (mereology) or constitution to link
> co-located objects such
> as the position and its occupier (or our old friends the
> statue and the
> clay).
>
> I hope you can now see the underlying issue. As a matter of
> process, it
> seems to me proper to clarify what the issues are, in rough
> way, and the
> strategies for resolving them - refining our intuitions
> before diving into
> formalising them.
Well, my worry is that we'll spend an enormous amount of intellectual work
getting clear about a small range of cases, which may feature prominently in
one's favorite metaphysics textbook but which don't have many real world
ramifications. I think one has only to consider the 2,500 year history of
philosophy to give this worry some substance.
>
> Regards,
> Chris
>
> -----Original Message-----
> From: apease@ks.teknowledge.com [mailto:apease@ks.teknowledge.com]
> Sent: 11 June 2002 00:43
> To: Chris Partridge; standard-upper-ontology@ieee.org
> Subject: RE: SUO: RE: Copyright issues
>
>
> Chris,
>
> At 03:56 PM 6/6/2002 +0200, Chris Partridge wrote:
>
>
> [snip]
>
> > >I don't know where Cyc stands on this issue, but in the
> SUMO 'Position'
> is
> >a
> > >subclass of 'CognitiveAgent', so that the identity of a
> position can be
> > >"tracked".
> > >CP2> But did you also amend the relevant bits of the
> mereology axioms to
> > >cater for what you have introduced? And so on....
> >
> >AP>What are the axioms in the mereology that would you like amended?
> >
> >The change that was made was from 'Position' being a role of
> a person, to
> >being CognitiveAgent. It seems to me that making this shift
> raises obvious
> >mereological questions - and also questions about
> metaphysical choices
> >(hence me offering it as an example to Ian). I presume from
> your question,
> >that it does not do this for you. Am I right?
>
> Could you explain further what you see as the mereological
> choices, and how
> they might be made with respect to the existing content of
> SUMO? We're not
> aware of any impacts from the change that currently need to
> be addressed,
> but would welcome detail to the contrary.
>
> Adam
>
> >Chris
>
> Adam Pease
> Teknowledge
> (650) 424-0500 x571
>
>
>
>