RE: ooRe: SUO: RE: Axiom and Intentionality vs Extentionality
Pierre,
> -----Original Message-----
> From: Pierre Grenon [mailto:pierre@cyc.com]
> Sent: Tuesday, November 06, 2001 8:21 AM
> To: Ian Niles
> Cc: standard-upper-ontology@ieee.org
> Subject: ooRe: SUO: RE: Axiom and Intentionality vs Extentionality
>
>
> Ian,
>
> Responses in line below.
>
>
> Ian Niles wrote:
> >
> > Pierre,
> >
> > > -----Original Message-----
> > > From: Pierre Grenon [mailto:pierre@cyc.com]
> > > Sent: Monday, November 05, 2001 2:30 PM
> > > To: Ian Niles
> > > Cc: standard-upper-ontology@ieee.org
> > > Subject: Re: SUO: RE: Axiom and Intentionality vs Extentionality
> > >
> > >
> > > Ian,
> > >
> > > I'm discovering too OrganisationalUnit, but pardon me,
> I'm still not
> > > following, I'd appreciate a few more inputs. In
> particular, it seems
> > > that OrganisationUnit should not subsume Organization.
> I.e., it seems
> > > that, instead of subClass, the relation should between
> > > OrganizationUnit
> > > and Orgnaization should be that instances of the former have
> > > collections
> > > associated with them at certain times. Well, you would say
> > > one and only
> > > one during any maximally continuous period of time.
> >
> > This strikes me as a good idea. We could remove the concept of
> > 'Organization' from the SUMO all together and then define a
> function, called
> > "OrganizationFn", say, that takes an 'OrganizationUnit' and
> a 'TimePosition'
> > as arguments and then returns a 'Collection' as its value, where the
> > 'Collection' returned would be the 'members', if any,
> associated with the
> > 'OrganizationUnit'. In cases where the 'OrganizationUnit' had no
> > corresponding 'members', the 'NullSet' would be returned.
> What do you
> > think? What do other people think of this idea?
>
> I can't tell for others :) But you realize that your talk of
> the result
> of this putative function plays on the ambiguity of the notion of
> &%Collection and &%members. The function would return a &%Collection.
> That has nothing to do with what the &%members of this
> &%Collection are
> (since the &%Collection's &%members can change.) In other word, the
> function will not return an enumeration of the members of the
> 'OrganizationUnit' during a given 'TimePosition', if it did that, most
> of the time it would simply fail to be a function.
I'm not sure I follow you here. Why would "OrganizationFn" fail to be a
function in this case. Because a 'Collection' can gain or lose 'members'
over a 'TimeInterval'? If so, then we can narrow the type restriction on
the second argument of the function to 'TimePoints'.
>
> Is what you are speaking about here is a relation holding between an
> &%OrganizationUnit, a &%TimePosition and an &%Organization, that would
> obtain whenever the members (kinda common sense) of the
> &%OrganizationUnit during the &%TimePosition are precisely
> the &%members
> (SUMO sense) of the &%Collection? This doesn't make any sense
> (maybe to
> me only) because there is nothing like 'the members of the collection
> during a given time interval'.
But do you think we have this problem if the second argument of
"0rganizationFn" is restricted to 'TimePoints'?
>
> I think it's pretty clear that this relation is not functional in its
> &%OrganizationUnit and &%TimePosition places since the same
> members can
> participate in the same OrgUnit at different times, at the
> same time in
> different OrgUnit.
I agree with you that "OrganizationFn" will be a many-to-one mapping, but
this doesn't prevent it from being a function. We would have a problem only
if you could produce a case where the putative function "OrganizationFn"
maps more than one 'Collection' to a particular 'OrganizationUnit' and
'TimePoint'. Can you think of such a case?
>
> My biase is that there is a more interesting relation. A relation that
> you could still construct by adding to the SUMO a category of
> 'aggreagate', say, and a function taking a SUMO-Collection and a
> &%TimePosition and returning one such aggregate.
Rather than adding a new category, perhaps we could just get rid of the
claim that adding/removing 'members' from a 'Collection' has no effect on
the identity of the 'Collection'. This would allow us to be agnostic about
identity conditions (my current preference) or to claim that some
collections have identity conditions that accord with the indiscernability
of identicals and others do not. How does that sound?
>
> -An 'aggregate' is that sort of mereological sum that cease to exist
> whenever one of its part is removed, it is the only way in which it
> differs from a SUMO-collection. (It is much similar to a set in that
> respect. I'm not really sure of my claim here but I believe that Lowe
> defines that notion, in any event I didn't come up with
> this.) To me, it
> is more intuitive to think of a flock of sheep as an
> aggregate than as a
> SUMO-collection: if the pack of wolves kills a sheep, the flock is not
> the same, less heads, less wool, if it were a mother, less
> reproductive
> potential, less revenue, etc.
>
> -The function, call it here CollectionToAggreagateFn, takes a
> &%Collection and a &%TimePosition and returns the 'Aggregate' of
> &%members of a &%Collection at a given time (&%members could relate
> object to 'Aggregate').
>
> The relation I find interseting is one between an &%OrganizationUnit
> (vraisemblably a generalization of this), a &%TimePosition and an
> 'Aggregate', that hold (in this specific case of an &%Organisation)
> between an OU, a time position and the aggregate of members (kinda
> common sense) of the OU, so basically something to the effect of:
>
> (OrgOfOUDuring OU TP (CollectionToAggregateFn O TP))
>
> Indeed, I'd be more receptive to having a ternary relation relating
> something like an OrganizationalUnit (that persist despite being
> memberless), a time stamp (that I'd rather see being as much
> an interval
> as a point, in contrast to the SUMO that offers vocabulary
> biased toward
> situating &%Physical at intervals only)
This isn't true actually. Most temporal relations allow one to specify a
'TimePosition', and both 'TimeInterval' and 'TimePoint' are subclasses of
'TimePosition'.
>and an aggregate of
> the members
> of the first argument, i.e. a possibly transient object (that
> presumably
> should not be allowed to reduce to the null meaningfully,
> i.e., it would
> be obligatory for an 'aggregate' to have at least one members).
>
>
> My overall impression on the basis of my understanding of the SUMO is
> that your suggestions are appropriate and consistent with the current
> settings. However, my point is that I am uneasy with this setting.
>
> -A fundamental issue for me is the rigidity of the temporal vocabulary
> with respect to the emphasis put on 'interval'. IMHU, if you commit
> yourself to an ontology of time, the notion of moment /
> instant / point
> is decisive and needs to be better emphasized. I'm thinking of
> applications in situation / scenario analysis here.
>
> -I welcome things like &%OrganizationUnit, or as I said, even better a
> generalization of this: something that keeps its identity
> through time,
> despite period while it is memberless, etc. I believe there
> is more use
> for what I called 'aggregate' that &%Collection. It is not to say that
> &%Collection does not makes sense, it looks to me like some
> bastardized
> form of the generalization of &%OrganizationUnit kinda middle of the
> road closing toward 'aggregate'. I really don't quite see the argument
> for having this chimera of &%Collection that seems to be exactly like
> the putative generalization of &%OrganizationUnit because it
> can change
> its members but in the fact the &%Collection do not survive
> memberlessness. My (maybe naive or confused) view is that
> it's enough to
> have a sort of things that do not survive any change in
> membership and a
> sort of find that survives all.
>
> Maybe this additional blatter can make some sense,
> &%Collection appears
> to me closely indexed to the choice made in emphasizing time intervals
> over time points. Indeed, you can be in a setting such that
> over a time
> interval the members of an &%OrganizationUnit change, so you have more
> than one 'aggregate' associated, but this is cumbersome and the notion
> of &%Collection appears of much importance to save the appearance. But
> as a matter of fact, aggregate persist through time as well. It seems
> more truthful to consider a succession of aggregate.
>
> >
> > >I'd tend to prefer
> > > treating collections as mereological sums altered by
> > > membership chnages,
> > > i.e., a 'collection' is no longer _the same_ when it changes its
> > > members. Anyway, I understand that's not SUMO.
> >
> > That's not SUMO, and it also flies in the face of
> intuition. I don't think
> > most people would be happy with the idea that the very
> identity of the U.S.
> > Army changes every time a new recruit enlists or a veteran
> is discharged.
>
> I didn't mean that the identity of the US Army changes with the change
> in its contingent. Still you're probably right in that our intuitions
> diverge regarding the merit of the notion of SUMO-collection
> constrasted
> with those of 'Aggregate' as defined above.
>
> I think that what most people are interesting in is the
> number of active
> soldiers in the US Army, that would be either considered by
> people as a
> 'set' or as an 'aggregate' (you cannot shoot a set, you can fire at an
> aggegate.) The fact that there is a US Army that can survive
> annihilation (maybe as long as the USA exists) is interesting
> indeed, US
> Army is an &%OrganizationalUnit. If the army is annihilated
> and then the
> US governement starts recruiting again, there's no
> US-Army-collection-before and US-Army-collection-After, there
> is one US
> Army and a shitload of 'aggregate', indeed one each time there's a
> casualty, a defection (mereological difference) or enrolling
> (mereological sum).
>
> Was I stubborned enought to give you the impression that this makes
> sense? :)
>
> yours,
> Pierre
>
>
> > >
> > > Help me with this example. What about Bruce Springsteen's E
> > > Street Band?
> > > Is it an OrganizationalUnit? An organization? How does it
> > > relate to its
> > > members, say how does it relate to Bruce Springsteen? It
> seems that
> > > Bruce Springsteen is a member of the E Street Band before
> dissolution.
> > > The band had no member for a few years, then reassembled.
> So it seems
> > > that E Street Band qua OrganisationUnit survived the
> dissolution of E
> > > Street Band qua Collection, then 'another' E Street Band qua
> > > Collection
> > > was formed (that it had the same members is inessential).
> So it looks
> > > like an OrganisationUnit is related to the members of
> possibly several
> > > Collections, is that right?
> >
> > That's right, so it may be better to recast 'Organization'
> as a function, as
> > suggested above. On the other hand, I think we need to be
> wary of the
> > methodology of analytic philosophy here. Philosophers are
> wont to concoct
> > cases that disprove the validity of a general claim, but
> that only arise so
> > infrequently that they are, for most purposes, irrelevant.
> This sort of
> > methodology is acceptable when one's aim is to construct a
> definition that
> > holds in every possible world. But if one is happy with
> some version of
> > satisfying the overwhelming majority of the people the
> overwhelming majority
> > of the time, then I think we can relax our standard a bit.
> This isn't to
> > deny the usefulness of your example, but it is to caution you about
> > overplaying the significance of such examples in the
> present context.
> >
> > >
> > > A music band which has a continuous lifespan of three years, is an
> > > OrganizationalUnit identical with a collection, i.e., it's an
> > > Organisation. But if that band reassemble ten years later
> is no longer
> > > an Organization?
> >
> > It's still an 'Organization', but it's not the same
> 'Organization' that it
> > was before. I agree that this is counterintuitive, so we
> may want to make
> > 'Organization' a function, as suggested above.
> >
> > >
> > > I guess my puzzlement comes from this: if during a decade or so, E
> > > Street Band was an Organization, it's an
> OrganizationUnit. However, it
> > > then ceased to exists to be another Organization which
> > > actually was the
> > > very same OrganizationUnit as the one before. How come can
> > > this band be
> > > two different Organizations (because 2 diffreent
> collection) but only
> > > one OrganizationUnit?
> >
> > Right, that's the problem. Let me know if you think the suggested
> > redefinition solves the problem.
> >
> > >
> > > All right, I should stop there, confused and drivelling.
> > >
> > > yours,
> > > Pierre
> > >
> > > Ian Niles wrote:
> > > >
> > > > Mike,
> > > >
> > > > I agree with you that there is a sense of
> > > "organization", according
> > > > to which the "organization" may have no members. In fact,
> > > as several SUO
> > > > subscribers have pointed out, there are many examples of
> > > corporations and
> > > > churches that have assets, are liable for certain claims
> > > etc, even though
> > > > they have no members. This was the motivation for
> distinguishing
> > > > 'OrganizationUnit' from 'Organization' and making the
> > > latter a subclass of
> > > > the former in the SUMO. The latter is required to have
> > > members, because it
> > > > is a subclass of 'Collection'. However,
> 'OrganizationUnit' is not a
> > > > subclass of 'Collection', because it is meant to cover both
> > > organizations
> > > > and positions, so there is no requirement that an
> > > 'OrganizationUnit' have
> > > > any members. Accordingly, we do have a notion of
> > > organization that covers
> > > > "empty" organizations.
> > > >
> > > > -Ian
> > > >
> > > > > -----Original Message-----
> > > > > From: Mike Pool [mailto:mpool@iet.com]
> > > > > Sent: Monday, November 05, 2001 11:29 AM
> > > > > To: Ian Niles; standard-upper-ontology@ieee.org
> > > > > Subject: RE: SUO: RE: Axiom and Intentionality vs
> Extentionality
> > > > >
> > > > >
> > > > > Ian,
> > > > >
> > > > > response below.
> > > > >
> > > > > Mike
> > > > >
> > > > > At 10:01 AM 05/11/01 -0800, Ian Niles wrote:
> > > > > >
> > > > > > Mike,
> > > > > >
> > > > > > See my comments below.
> > > > > >
> > > > > > -Ian
> > > > > >>
> > > > > >> -----Original Message-----
> > > > > >> From: Mike Pool [mailto:mpool@iet.com]
> > > > > >> Sent: Saturday, November 03, 2001 3:01 AM
> > > > > >> To: Jean-Marc Orliaguet; Pierre Grenon
> > > > > >> Cc: David Whitten; iniles@teknowledge.com;
> > > > > standard-upper-ontology@ieee.org
> > > > > >> Subject: Re: SUO: RE: Axiom and Intentionality vs
> > > Extentionality
> > > > > >>
> > > > > >> At 12:40 PM 03/11/01 +0100, Jean-Marc Orliaguet wrote:
> > > > > >>>
> > > > > >>> Pierre Grenon wrote:
> > > > > >>>>
> > > > > >>>> Dear David, dear Ian,
> > > > > >>>>
> > > > > >>>> Pardon my confusion, maybe I'm too unfamiliar with the
> > > > > SUMO, but I find
> > > > > >>>> puzzling your interpretations of this axiom.
> > > > > >>>>
> > > > > >>>> To be consistent with the claim that "collections have
> > > > > position in time"
> > > > > >>>> (assuming I'm not mistaken in understanding that it
> > > > > means that at
> > > > > >>>> different time, members may not be the same) shouldn't
> > > > > an axiom stating
> > > > > >>>> the necessary non emptiness of a collection be alike the
> > > > > following:
> > > > > >>>>
> > > > > >>>> (=>
> > > > > >>>> (and
> > > > > >>>> (instance ?COLL Collection)
> > > > > >>>> (instance ?TIME TimePosition))
> > > > > >>>> (exists
> > > > > >>>> (?OBJ)
> > > > > >>>> (holdsDuring ?TIME (member ?OBJ ?COLL))))?
> > > > > >>>>
> > > > > >>>> That is, at any time a collection has at least
> one member?
> > > > > >>>>
> > > > > >>>> But, one way or another, I confess that it seems odd to
> > > > > me to require of
> > > > > >>>> a collection that it be non empty. If a collection
> > > > > really preserves its
> > > > > >>>> identity under alteration of membership, i.e., "members
> > > > > can be added and
> > > > > >>>> subtracted without thereby changing the identity of the
> > > > > &%Collection".
> > > > > >>>> Does that mean that collections are intermittent beings?
> > > > > >>>
> > > > > >>>
> > > > > >>>
> > > > > >>> A collection is not an "intermittent being", it is an
> > > > > abstraction, and a
> > > > > >>> collection that has no member is called "nothing"
> > > > > >>>
> > > > > >>> Peirce in 4.650 "[...] two apparent and highly
> > > > > interesting ornithological
> > > > > >>> collections, the one of whatever phoenixes there ever
> > > > > were or will be, the
> > > > > >>> other of whatever cockatrixes there are at this moment,
> > > > > are one and the
> > > > > >>> same collection, having one and the same essential
> > > > > character. It is that
> > > > > >>> quite unique collection that goes by the name of Nothing.
> > > > > >>> Some writers whose logical conceptions would seem to be
> > > > > in a state of
> > > > > >>> disintegration have supposed the collection whose sole
> > > > > member is Gaius
> > > > > >>> Julius Caesar to be identical with Gaius Julius Caesar
> > > > > himself -- a
> > > > > strange
> > > > > >>> confusion considering that the latter was a man of
> > > > > immense force of
> > > > > >>> intellect who was brought into the world by a grossly
> > > > > unskillful operation
> > > > > >>> of surgery, while the other is nothing but an ens
> > > > > rationis brought into
> > > > > >>> being by the idea of that man being chosen without any
> > > > > surgery at all and
> > > > > >>> utterly deprived of any force of intellect or life."
> > > > > >>
> > > > > >> Is this the same sense of 'Collection' as that being
> > > > > defined in the SUMO,
> > > > > >> i.e., in the definition of &%Collection? &%Collection
> > > seems to be
> > > > > something
> > > > > >> very pedestrian, I'm understanding it to mean something
> > > > > like "group of
> > > > > >> physical objects" or maybe "identifiable group of physical
> > > > > objects." (Are
> > > > > >> all groups of physical objects instances of &%Collection,
> > > > > e.g., is any
> > > > > >> subgroup of a given flock of sheep an instance of
> > > > > &%Collection?) A group
> > > > > >> of physical objects doesn't seem to be "brought into being
> > > > > by the idea" of
> > > > > >> its constituent member(s). Also, it doesn't seem that a
> > > > > &%Collection is
> > > > > >> an "abstraction"; a flock of sheep (example from the
> > > definition of
> > > > > >> &%Collection) isn't an abstraction is it?
> > > > > >>
> > > > > >> But even with this earthy defintion of &%Collection, I
> > > > > share Pierre's
> > > > > >> concerns here. Given that a &%Collection can survive
> > > > > membership changes,
> > > > > >> it seems rather arbitrary to declare that a &%Collection
> > > > > can't survive a
> > > > > >> time period in which it has no members.
> > > > > >>
> > > > > >> Hmm. This doesn't seem arbitrary to me. In fact, it is
> > > > > precisely the
> > > > > >> situation that we have with physical objects that are not
> > > > > aggregates.
> > > > > Parts
> > > > > >> of the chair I am sitting in right now can be replaced
> > > > > with other parts
> > > > > >> without affecting the identity of the chair. However, if
> > > > > the chair were
> > > > > >> completely destroyed, it would not be possible to refer to
> > > > > this same chair
> > > > > >> in any future time interval.
> > > > > >>
> > > > > >> Consider an example given in the definition, a football
> > > > > team. If all the
> > > > > >> members of a football team die in a bus crash and the team
> > > > > is replenished
> > > > > >> the next day in an emergency draft, would we declare that
> > > > > the team had to
> > > > > >> start with a new 0-0 record and that the members couldn't
> > > > > pursue old club
> > > > > >> records? Or would we see this simply as a radical
> > > > > personnel change that
> > > > > the
> > > > > >> organization survived? If we maintain that
> > > > > &%Organization is a subclass
> > > > > >> of &%Collection, then it is not clear that no &%Collection
> > > > > can survive a
> > > > > >> loss of all of its members.
> > > > > >>
> > > > > >> Well, this is just recasting the "Ship of Theseus" paradox
> > > > > in terms of
> > > > > >> collections. Given that this paradox has not been
> > > > > resolved in over 2,000
> > > > > >> years, my suggestion would be that we focus on problems
> > > > > that are likely to
> > > > > >> be more tractable.
> > > > > >
> > > > >
> > > > > I guess I wasn't being very clear. I was simply noting that
> > > > > it seems that
> > > > > some
> > > > > organizations may, in fact, survive being memberless.
> > > > > Hence, I wonder if it
> > > > > makes sense to axiomatize '&%Organization' as if they do not.
> > > > >
> > > > >
> > > > >
> > > > >
> > > > >
> > > > > ________________________________
> > > > > Mike Pool
> > > > > Information Extraction & Transport, Inc.
> > > > > (703) 841-3500 x632
> > > > > (703) 841-3501 Fax
> > > > >
> > > > >
> > > > >
> > > > >
> > >
> > > --
> > > Pierre Grenon, Cycorp, Inc, Austin, Texas.
> > > Tel: 512 514 2982 - Fax: 512 342 4040
> > > mailto:pierre@cyc.com http://www.cyc.com
> > >
>
> --
> Pierre Grenon, Cycorp, Inc, Austin, Texas.
> Tel: 512 514 2982 - Fax: 512 342 4040
> mailto:pierre@cyc.com http://www.cyc.com
>