Leo, (01)
I was a bit hasty last night when I said "I agree" to the
following point: (02)
> I think the value is not in the particular upper ontology
> one chooses, but in the ontological analysis (even metaphysical
> analysis) that leads you to propose a given upper ontology. (03)
The simplest revision I would make is "There is no such thing
as an ideal upper level, but there are some very bad ones." (04)
I should also have said that I agree with the general principle
that Lenat expressed about 20 years ago: The middle and lower
levels of the ontology are far more important for detailed
reasoning than the upper level. (05)
Given those caveats, there are many qualifications: (06)
1. A very detailed ontology, which is intended for *integrating*
all the components of a system design, would need some very
detailed constraints. But such an ontology can be included
as a "subontology" or "microtheory" underneath a less detailed
upper level that is intended for a broad range of applications. (07)
2. Highly underspecified lexical ontologies such as WordNet have
proved to be very useful for a wide range of applications.
But it is important to recognize that their upper levels
should be considered a highly underspecified default for
mapping NLs to and from an ontology. (08)
3. When two detailed ontologies A and B have been "aligned"
to WordNet or any other lexical ontology, that does *not*
imply that A and B have been aligned to one another. (09)
4. Fundamental principle: you can't make mutually inconsistent
ontologies consistent just by "aligning" their categories. (010)
5. But for many applications, an alignment to WordNet works
almost by accident. The reason why it works is that the
the inconsistencies were not relevant to the applications
on which they were tested. (011)
6. But it is better to design ontologies in a way that does
not depend on serendipity in order to be successful. (012)
Short summary: If an upper level ontology has proved to be
useful for certain applications, that does not mean it is good.
It only means that the application developers managed to ignore
the implications they did not like or understand. (013)
John (014)
_________________________________________________________________
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J (015)
|