SUO: RE: RE: More KIF-ified Ontology Content
Hi Ian,
I appreciate that many emails have been sent since this before on this
topic, but it may be worthwhile making myself clear.
You wrote:
If I understand you right, you're saying that each chunk of
ontological content (whether subsumption relations or axioms) presupposes a
high-level theory and, unless this theory is made explicit, it will be very
difficult to merge the chunks into a single, comprehensive, and consistent
ontology.
I do not think I am claiming this. There may well be some low level
'theories' that do not presuppose any higher theory - I suspect that most
people's everyday 'theories' are like this.
I am saying something more along the lines that if we want to get an
integrated ontology, then any (every) low level theory will need to be
integrated (somehow) into this theory - and this will involve, typically,
integrating with a high level framework. I do not see how you can avoid this
if you want to build a SUO.
If you are saying that we start with low level theories as part of the
process of building up to the high level - this seems sensible. However, my
original point re: holes was that the specific hooks to the high level were
ignored - which seems odd if the goal is to regiment them.
Chris
-----Original Message-----
From: owner-standard-upper-ontology@ieee.org
[mailto:owner-standard-upper-ontology@ieee.org]On Behalf Of Ian Niles
Sent: 05 December 2000 18:52
To: 'Chris Partridge'; Standard-Upper-Ontology (E-mail)
Subject: SUO: RE: More KIF-ified Ontology Content
Hi Chris,
Thanks for your comments. I'll try to reply to your general point,
and I'll leave to Chris Menzel (the author of the KIF formalization) your
question about the concepts of 'material' and 'immaterial'.
If I understand you right, you're saying that each chunk of
ontological content (whether subsumption relations or axioms) presupposes a
high-level theory and, unless this theory is made explicit, it will be very
difficult to merge the chunks into a single, comprehensive, and consistent
ontology. I agree that, to the extent possible, definitions of underlying
concepts should be made explicit. However, this is not always possible.
Scientific theories often depend upon unexplicated concepts like gravity,
force, behavior, etc. In addition, many problems of high-level conceptual
analysis seem to be philosophically complete, in the sense that agreement
about all foundational issues would be necessary before the analysis could
be expected to withstand any argument. At some point theorizing has to come
to an end or progress will never be made.
I think what we should strive to do in our ontological construction
is to strike a balance between high-level conceptual analysis and
creating/formalizing axiomatic chunks like C & V's theory of holes.
Creating self-consistent chunks of ontological content in a common language
sheds some light on higher-level concepts, while explicating the latter
concepts partially determines the nature of subsumed concepts. Eventually
(if we allow ourselves to be optimistic) we will reach a stage where there
is balance between the high-level concepts and the lower-level chunks. At
this stage, a merge of both sorts of content should be possible. This merge
will undoubtedly not contain all of the original content in the axiomatic
chunks or the definitions of high-level concepts, but it should be
comprehensive in the sense that it provides a hook on which to hang any
domain-specific ontology.
-Ian
> -----Original Message-----
> From: Chris Partridge [mailto:chris_partridge@csi.com]
> Sent: Tuesday, December 05, 2000 10:19 AM
> To: 'Ian Niles'; Standard-Upper-Ontology (E-mail)
> Subject: RE: More KIF-ified Ontology Content
>
>
> Ian, et al.
>
> It seems to me that the formalisation skips over some key ontological
> points (but maybe I have missed this) - which I believe at
> least some of
> the members of this list regard as important.
>
> Excerpt:
> "(documentation "The main thesis is that a hole is an immaterial body
> located at the surface (or at some surface) of a material
> object. Since the
> notion of a surface is essentially a topological one, and since the
> property of being immaterial is reflected in the
> morphological property of
> being fillable, the ontological basis is concerned first and
> foremost with
> the general dependence of a hole on its host.")"
>
> This excerpt (and Achille and Roberto's book) seems to
> presuppose two basic
> (and distinct) ontological kinds; material and immaterial.
> However, these
> do not seem to get mentioned in the KIF, only indirectly on
> their impact on
> the notion of hole. I would have thought that basic ontological
> distinctions such as these should be stated up front. Maybe
> they are som
> ewhere else in? But would it then not be sensible to refer to them?
>
> This is closely related to another point. This is not the only way of
> thinking about holes - as shown by comments on the book
> (some, if I recall
> correctly, by David Lewis). If you have a different high
> level ontology,
> then you will end up with different axioms. This is a good
> illustration of
> the importance of having a clear high level ontology.
> Obviously if another 'theory' takes a different stance which leads to
> different ontological structures, and neither this nor the
> other theory,
> makes clear its high level ontological commitment - then
> amalgamating the
> theories will be more difficult.
>
> This is a general point that I think and others have raised
> before. Unless
> we get agreement on the high level structure (the ontological
> paradigm) and
> this will involve hard choices then I do not see how we can expect to
> amalgamate the lower level theories.
>
> Regards
> Chris
>
>
> -----Original Message-----
> From: Ian Niles [SMTP:iniles@teknowledge.com]
> Sent: Tuesday, December 05, 2000 5:14 PM
> To: Standard-Upper-Ontology (E-mail)
> Subject: SUO: More KIF-ified Ontology Content
>
> << File: Holes.kif >>
> Hi All,
>
> Attached is a SUO-KIF formalization (done by Chris
> Menzel) of Casati
> and Varzi's formal theory of holes, outlined in HOLES AND OTHER
> SUPERFICIALITIES, MIT Press, 1995. The idea is that this
> formalization,
> along with the other KIF-ified ontology sources, will
> eventually be folded
> into the SUO. Any questions, criticisms or comments regarding this
> formalization or the original content would be greatly appreciated.
>
> -Ian Niles
>
>