Re: Good and Bad IS-A email@example.com (Pat Hayes)
X-Sender: firstname.lastname@example.org (Unverified)
Content-Type: text/plain; charset="us-ascii"
Date: Mon, 24 Jul 1995 16:48:40 +0000
To: email@example.com (Eduard Hovy)
From: firstname.lastname@example.org (Pat Hayes)
Subject: Re: Good and Bad IS-A hierarchies
Cc: email@example.com, firstname.lastname@example.org, email@example.com
Eh Hovy wrote:
>So I disagree with Peter Clark. I *would* like to see a lot of concept
>organizations (isa hierarchies at the very simplest), presented in a way
>that is easily browsable so that I can pick and choose bits and pieces
>from here and there and weld them together to suit my needs of the moment.
But its not clear what this "welding" process is supposed to be. The point
is that just having a concept listed in an organization tells you almost
nothing about it. The content of the concept is entirely held in the
axioms, if there are any, which relate the concept to other concepts. If
this welding is going to actually capture any meaning, it will have to
involve some kind of gathering together of these axioms. But if you want
the concept of 'horn', and there is an axiom telling you that horns are
found on rhinos, you will have to accept the concept of 'rhino'. There is
no way to compartmentalize the meanings of concepts into little packets
which can be taken or not to suit your needs of the moment.
I discussed this issue long ago in the naive physics 'manifesto'. Concepts
dont occur in neatly seperable packages, but they do seem to occur in
clusters. If you imagine a (hyper)graph with nodes as concepts and
(hyper)edges as axioms connecting them, then the entire graph is connected.
However, some regions seem to be more tightly connected than others. The
real issue is to try to identify these tightly connected 'clusters' and get
them axiomatized. These will provide the organization which we need to make
this off-the-shelf construction of useful concept bundles possible. The
is-a heirarchies are a side issue.
>(Or is anyone going to say they know the true, correct, way to model the
This is a side issue. Your point here - and I agree its a strong one - is
that there will have to be alternative ways of describing things. But this
holds before we even get to the off-the-shelf construction problem. Take
time for example. Sometimes its best to talk about timepoints, sometimes
timeintervals are the best basic idea. Or you can have both, and then there
are several ways to have them related to one another (eg points can be the
meetingplaces of intervals, or intervals can be approximations to points,
or points can be infinitely short intervals). None of this is usefully
reflected in any kind of isa heirarchy: its not a choice between different
concepts, but different ways to axiomatise the same collection of concepts.
Moreover, there is no way to organize the concepts, or even the axioms,
into neat little packets so that the various alternatives can be assembled
by choosing some and ignoring others.
There just are genuine alternatives, and one has to make committments in
selecting a temporal theory to work with.
I seem to detect, in the object-oriented flavor which informs so much
current work in ontologies, a residue of the old bias against the use of
axioms. But theres no way around it: if you want to answer questions, you
have to be able to draw conclusions, and conclusions involve making
connections between things. The basic unit of meaning is not a concept but
a theory, ie a collection of axioms.
1916 Ivy Lane, Palo Alto, Ca. 94303 phone (415)855 9043