Re: clarifying clarifying ontologies (Pat Hayes)
Message-id: <>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Wed, 26 Jul 1995 14:42:35 -0500
To: Doug Skuce <>,,
From: (Pat Hayes)
Subject: Re: clarifying clarifying ontologies
Precedence: bulk
Well Doug, I guess I couldnt disagree more. I think your position is based
on a fundamental error, which is that the way to get formalizations is to
talk and talk until our thoughts are Clear, and then just jot them down in
a handy logic. The reason this doesnt work is one you may not have
discovered if you have never gotten as far as the last stage of
axiom-writing: the process of writing the axioms (or frames or whatever)
reveals distinctions, errors and inconsistencies which you simply hadnt
thought of before. It is impossible to get our intuitions clear just by
informal philosophizing, no matter how carefully done.

At the risk of sounding repetitive, I can cite temporal relationships,
which might seems about as clear as any intuitions could get, and which
have been really done to death by philosophers. Yet until you really get
those axioms written down and play with them a while and discover which
ones follows from wich others and what kinds of interpretation they can be
given, you wont even discover what some of the conceptual problems are, let
alone get them sorted out. The fact is, our intuitions arent as exact as we
think they are.

More specially, I think it is a real error to start at the top of the isa
heirarchies. You and John Sowa have decided that the Universe consists of
physical things and, er, what exactly? You call it (not very helpfully)
'nonphysical' (I can agree that the whole universe can be divided into
physical and nonphysical, but it can also be divided into boojum and
nonboojum) but John calls it 'information' which already makes my thumbs
prickle: I certainly dont think that the universe can be divided into
physical things and information. Here are a few problem cases: a law, a
conspiracy, the surface of the ocean, a long wait, a shed (the space in a
loom where the shuttle is thrown, but its a different shed depending on the
pattern of the warp threads), a bad mood, a photon, a direction.

But the main point is, what is the utility of doing this kind of high-level
classification? Its impossible to get agreement on it, and in any case we
dont need to get agreement in order to get useful reasoning done. If you
classify the surface of the ocean as a physical object and I classify it as
an abstraction, we can still agree on lots of its properties, and each can
make the other feel uncomfortable (what does it weigh? It has a location!).
Why not just put this kind of empty Aristotelian debate on one side and,
rather, look for ways in which we can get on with the reasoning *without*
having to have our is-a's lined up perfectly.

Pat Hayes

PS. The software case isnt analogous in an important way. With software,
the language we are using has its meaning fixed. But in our case, the
meaning of a relation or function symbol is *entirely* established by the
axioms you write which use it. Until axioms are provided, we are just
waving vague promises at ch other.

Until September: 
1916 Ivy Lane, Palo Alto, Ca. 94303               phone (415)855 9043