RE: SUO: RE: A proposed SUO content outline
Pat,
See my comments below.
-Ian
> -----Original Message-----
> From: pat hayes [mailto:phayes@ai.uwf.edu]
> Sent: Thursday, March 01, 2001 4:34 PM
> To: Ian Niles
> Cc: standard-upper-ontology@ieee.org
> Subject: RE: SUO: RE: A proposed SUO content outline
>
>
> >Pat,
> >
> > See my comments below.
> >
> >-Ian
> >
> > > -----Original Message-----
> > > From: pat hayes [mailto:phayes@ai.uwf.edu]
> > > Sent: Wednesday, February 28, 2001 6:49 PM
> > > To: standard-upper-ontology@ieee.org
> > > Subject: Re: SUO: RE: A proposed SUO content outline
> > >
> > >
> > >
> > > Ian Niles <iniles@teknowledge.com>, to Matthew West:
> > >
> > > >I understand your position. I'm still not convinced that a
> > > 4D orientation
> > > >cannot be incorporated into the merged ontology, and I think
> > > the replacement
> > > >of the 'Continuant'/'Occurrent' dichotomy with the
> > > >'ContinuantType'/'OccurrentType' division is a step in this
> > > direction,
> > > >because this division is consistent with the claim that
> > > things in themselves
> > > >are neither continuants nor occurrents (rather, type
> descriptions are
> > > >continuant-like or occurrent-like). Furthermore, nothing in
> > > the merged
> > > >ontology (to my knowledge) is inconsistent with the claim
> > > that everything as
> > > >a space-time worm. If you disagree, could you cite an axiom
> > > or subclass
> > > >relationship from the merged ontology that, by your lights,
> > > is inconsistent
> > > >with this claim.
> > >
> > > Oh, come on Ian! All the mereotopology, all the
> time-interval theory
> > > and all the 'holes'; the continuant/occurrent
> distinction; Virtually
> > > any concept that refers to 'ordinary things' (which is a
> loose phrase
> > > for 'continuant' here). Almost the entire upper
> classification will
> > > have to be re-done. Might be worth trying it, though. If
> I had any
> > > spare time I wouldnt mind having a go at it myself.
> >
> >I agree that the merged ontology makes liberal use of the
> ordinary notion of
> >a physical object,
>
> The very fact that you find this 'ordinary notion' so central and
> clear is in itself some of the "light" I referred to below. Is the
> coffee in a half-full mug an ordinary physical object (OPO, to put
> one of Awbrey's mannerisms to work for a moment)? Is the thread in a
> piece of fabric an OPO? Is the paint on my car an OPO? Is a bead of
> sweat on my brow an OPO? It is virtually impossible to walk through
> the world during daylight without seeing several hundred things every
> hour which would be problematic to classify firmly as either OPOs or
> non-OPOs.
Actually, to my mind, a piece of thread seems to be a pretty clear case of
an ordinary physical object (albeit a rather small one). As for paint and
sweat, I think they point out a need for subclassifying ordinary physical
objects into something like the stuff/object distinction that Cyc uses.
This doesn't really address your more general claim, though, that there are
far fewer ordinary physical objects in the world that we might be inclined
to think. I agree that there are many things in this world that aren't
ordinary physical objects and also that this notion is somewhat fuzzy, but I
do see an awful lot of paradigmatic objects as I look around my office. I
see a chair, a desk, a computer terminal, a keyboard, a cup, a bookshelf, a
whiteboard, a phone, a plant, a diskette, a computer print-out, a watch, a
mousepad, etc.
>
> >but I also think that this notion can be defined in terms
> >of the concepts that a 4D ontology would admit. In fact,
> you described one
> >possible way of doing this in your most recent exchange with
> Matthew West.
>
> No, that was one way that a *continuant* could be mapped into a 4-d
> ontology. Ordinary objects are already in the 4-d ontology (anything
> that occupies space and lasts for a time is in the 4-d ontology) but
> they aren't thought of as continuants there.
>
> >Furthermore, a 4D ontology that could not formalize our talk
> about ordinary
> >physical objects would, it seems to me, be of very limited utility.
>
> Of course; but it can. (Well, Im not sure what you mean by "talk". In
> one sense I don't think we should be trying to formalize talk.)
>
> >Most of
> >our daily interactions with the world and our speech about these
> >interactions crucially involve this notion.
> >
> >There are really two distinct issues here that I think
> you're conflating.
> >One issue is about the concepts that we admit into our
> ontology, and the
> >other is how we cash out these concepts and/or map them to
> other conceptual
> >schemes. The merged ontology locates physical objects
> towards the top of
> >the hierarchy, but, as much as possible, it leaves open the
> question of how
> >this notion is to be philosophically construed.
>
> I am talking about the concepts used by the axioms which
> comprise the ontology.
>
> These distinctions are not mere philosophical decorations; they arise
> sharply as soon as you try to actually write axioms. The axioms of a
> 4-d ontology will look very, very different from those of an
> endurantist ontology. So as soon as someone wants to start on the
> business of creating an actual ontology, they will need to take a
> stance on issues like this.
>
> There really is not much point in having a standard upper ontology
> which just tells you that you need to talk about everyday physical
> objects. We know that before we start. The relevant ontological issue
> is, HOW are you going to do that talking? What ontological concepts
> are going to be used to refer to the everyday categories of cabbages
> and kings? It is those concepts - which might be the apparatus of an
> endurantist approach, or of a 4-d approach, or of some other approach
> - which are the 'concepts that we will admit to our ontology', not
> everyday things.
>
> Let me see if I can make the point with an analogy. Matthew and Chris
> P. and myself are like carpenters arguing about whether to use oak,
> ash or elm for the legs of chairs, and you are telling us that what
> we need are legs. Its not that you are wrong, exactly; but why are
> you in the carpenter's shop?
I think you have a point here, but I think you could make it much more
effectively if you provided some examples. Could you cite two axioms having
roughly the same meaning but framed, respectively, in Endurantist and 4D
terms. I'm not doubting that you can do this; it would just help me (and
maybe some others as well) to understand what's at stake here.
>
> > > > >
> > > > > MW: The inferences (in both directions) constitute the
> > > > > mapping. A mapping
> > > > > specification may be developed for members of
> particular classes.
> > > > >
> > > > > MW: The advantage of this approach is that different world
> > > > > viewpoints are
> > > > > respected and kept separate, whilst we learn explicitly of
> > > > > the relationship
> > > > > between them, and we have a basis for challenging the
> utility and
> > > > > completeness of particular world viewpoints.
> > > >
> > > >Well, in compiling the first few versions of the merged
> > > ontology, I've done
> > > >my best to avoid adopting a "world viewpoint".
> > >
> > > Ian, it is IMPOSSIBLE to avoid adopting a world view when
> making an
> > > ontology. Your view shines through almost every comment
> you make like
> > > a bright light. It is clear that you think like what
> might be called
> > > an Aristotelian endurantist: you see "ordinary" physical things as
> > > real and lasting through time, you find solid physical
> things as more
> > > real than things like spaces and surfaces; you think that
> things of
> > > zero thickness can't be spatiotemporally located; you think that
> > > universals exist.
> >
> >An "Aristotelian endurantist", can I put that on my resume?
> Honestly, I
> >don't know why you think I believe all of these
> philosophical claims, and
> >they certainly have not been incorporated into the merged ontology.
>
> Oh, they indeed have. The mereotopological axioms for example do not
> admit surfaces into the domain of quantification.
Well, this is really more a product of sloppiness on my part in merging the
various ontologies, rather than a philosophical claim that informs the
merged ontology. I think we can resolve this matter simply by adding a
clause excluding surfaces to the antecedent of each mereotopological rule.
>
> >I'm not
> >a realist about physical objects or universals, and when did
> I ever say that
> >spaces and surfaces are less real than objects or "things of
> zero thickness
> >can't be spatiotemporally located"?
>
> I inferred it from your reactions to some of Matthew's earlier
> messages; but the inference was not deductively valid, I confess.
>
> > The merged ontology contains just one
> >existence claim, viz. that there is at least one instance of
> the root node
> >'Entity'.
> >
> >As I see it, what we're trying to do in constructing an
> upper-level ontology
> >is to provide a precisely specified vocabulary that supports
> anything that
> >we would every want to say in a formal context. As
> paradoxical as it might
> >sound, I think we should strive for an ontology that makes as few
> >ontological commitments as possible.
>
> It is paradoxical, yes, as you will find out if you try to write any
> reasonably dense collection of axioms about almost anything.
>
> >The more philosophical assumptions we
> >build into the ontology, the more philosophers are going to
> be alienated by
> >the ontology, and the more engineers are going to shake
> their heads in
> >stupefaction.
>
> It's the engineers who know most acutely why the conceptual
> vocabulary is central. Why do you think that the people you are
> arguing with here are working for Shell, developing business
> ontologies, and in my case worrying about physical reasoning? You are
> arguing with the practical guys here, trying to make us swallow a
> wierd cocktail distilled from Peirce, Whitehead, Lesniewski and God
> knows who else, and you have the brass balls to tell us that
> engineers might be too confused by hyperspace??
When did I ever claim that engineers might be too confused by hyperspace?
In any case, let me see if I can cut through the hyperbole here and
reconstruct your argument. I've compiled an ontology from various sources,
Matthew West and I have been disagreeing about whether a 4D orientation is
preferable to a 3D one, you and I have been arguing about the extent to
which philosphical concerns are relevant in this context, ergo... Oops, I
guess I can't do it; I can't find an argument here.
>
> > >
> > > >I'm sure that many
> > > >assumptions about the real nature of the world have crept
> > > in, but the aim
> > > >has always been to create a high-level language that
> > > supports anything that
> > > >we would ever want to say in a formal context.
> > >
> > > Anything, ever? That is a very large claim.
> > >
> >
> >That's the aim. I think that the more we approach it, the
> more successful
> >we'll be.
>
> I'm tempted to invoke monkeys, trees and the moon, but I will refrain.
I think we saw enough of monkeys, trees, and the moon in your previous
comment.
>
> > > <snip>
> > >
> > > > >
> > > > > MW: Continuant does not compute in "my" ontology.
> > > >
> > > >If this implies that you are unable to ascribe
> > > continuant-like properties to
> > > >(what we ordinarily regard as) objects, then that sounds
> > > like a limitation
> > > >of your ontology.
> > >
> > > Ian, "continuant" is not a natural kind, a term of nature; it is a
> > > term which is only meaningful within one philosophical approach to
> > > ontology. There are other approaches, and I agree with
> Matthew that
> > > the 4-d one is of considerably greater utility. (It is (still) the
> > > only ontological framework within which one can even begin to
> > > describe fluids adequately, for example, and it is the basic
> > > framework used throughout qualitative physical reasoning, for good
> > > reasons: nothing else cuts the mustard.) You may not agree, of
> > > course, many people don't. But to say that a rejection of the
> > > concept of continuant is a "limitation", is to be blinkered by one
> > > world-view. Of course everyone agrees that what we
> ordinarily regard
> > > as objects must be describable in any ontology. But it
> doenst follow
> > > that they must be described as continuants.
> >
> >I really don't follow you here. To my mind, what we
> ordinary regard as
> >objects are regarded in this way precisely because we ascribe
> >continuant-like properties to them.
>
> Well, yes, you do seem to be an endurantist, as Ive said, so one
> would expect that your mind would work that way. Nothing personal,
> just an observation.
>
> Pat
>
> ---------------------------------------------------------------------
> IHMC (850)434 8903 home
> 40 South Alcaniz St. (850)202 4416 office
> Pensacola, FL 32501 (850)202 4440 fax
> phayes@ai.uwf.edu
> http://www.coginst.uwf.edu/~phayes
>
>