4/21/2004 9:18
AM
Key questions on Common Upper Ontology
Leo, Irene, Paul, et al.,
You know the controversy is un-resolvable by any objective criteria when people start writing things like that:
Ontologies are the basic
infrastructure for the Semantic Web. Everybody agrees on this...
In other words, the author(s) couldn't find any convincing arguments, so they just declared that everybody agrees. Along the same lines, I'd like to cite a passage from my KR book (copied below).
I'd also like to support the anonymous note by the AI scholar:
I have never felt quite as
strong as Herb, but have taken a more agnostic approach: I don't really care if
machines think. If they are doing tasks
that relieve people of some of their cognitive tasks, analogous to the ways
that physical machines have relieved people of physical tasks that humans
previously had to do throughout history, I will be quite happy.
This is a sensible position that engineers would be happy to support. Furthermore, neuro-physiologists are the first to agree that nobody in the world really knows how people think and that anybody who claims to know is either a liar or a charlatan.
In response to my earlier note, Irene asked
Does this definition mean that
any ontology designed for re-use is an upper ontology? I thought that an "upper" ontology
is the one that you use by subclassing the lowest classes.
This leads us to the lattice of all possible theories. Every ontology (or program specification), when formally represented in some version of logic, is a theory. Every specialization of a theory by subclassing (i.e., adding more constraints or conditions, a.k.a. axioms) is lower in the hierarchy -- it is a specialization of every theory that can be derived from it by erasing axioms or by replacing them with less constrained axioms.
The more specialized theories specify (parts of) a smaller number of programs, and the more generalized theories specify (parts of, but smaller parts of) a larger number of programs.
At any point in time, the number of theories that have been written is finite, but the number of all possible theories is infinite. Where you make the cut between upper and middle or between middle and lower is a matter of personal taste, preference, and perseverance.
John Sowa
From _Knowledge Representation_ by John F. Sowa, Chapter 2, "Ontology":
Kant considered this table [his 12 upper-level categories] a principled
framework for organizing the categories, not a rejection of all the work that
had been done within Aristotle's framework:
"If
one has the original and primitive concepts, it is easy to add the derivative
and subsidiary, and thus give a complete picture of the family tree of the pure
understanding. Since at present, I am concerned not with the completeness of
the system, but only with the principles to be followed, I leave this
supplementary work for another occasion. It can easily be carried out with the
aid of the ontological manuals, for instance, by placing under the category of
causality the predicables of force, activity, passivity; under the category of
community the predicables of presence, resistance; among the categories of
modality the predicables of origin, extinction, change, etc." (A:82,
B:108)
In continuing this discussion, Kant seriously underestimated the amount
of effort required to complete his tree of concepts:
"From the little I have said, it will be obvious that a dictionary
of pure concepts with all the requisite explanations, is not only possible, but
easy to complete." (A:83, B:109)
Whenever a philosopher or mathematician uses words like "easy" and "obvious," that is a sure sign of difficulty. Kant used those words several times in the course of a page or two; after two hundred years, his easy task is still unfinished.