Re: Roles and email@example.com (Pat Hayes)
Content-Type: text/plain; charset="us-ascii"
Date: Mon, 2 Oct 1995 12:03:11 -0600
To: firstname.lastname@example.org (John F. Sowa), email@example.com, firstname.lastname@example.org
From: email@example.com (Pat Hayes)
Subject: Re: Roles and dependence
Cc: firstname.lastname@example.org, email@example.com
At 10:10 PM 9/30/95 +0500, John F. Sowa wrote:
>From Pat Hayes:
>>An example, from CYC as of a few years ago. I have always had a very clear
>>mental distinction between an individual thing and a mere patch of
>>space-time. CYC also had this, but the writers discovered after a lot of
>>work, and to their surprise, that the distinction seemed to be
>>unnecessary.(Doug, please comment on this in case I have it wrong.) This
>>was a real empirical result, on a par with CERN discovering only one
>>particle where theory has predicted two.
>If they seriously believe that, then they must be more muddled than
>I thought they were. And if you seriously believed that, you would be
>able to reproduce their proof from memory.
Your response nicely illustrates my point, and your failure to understand
it. There is no such 'proof': it was an observation. They DISCOVERED that a
distinction (which they themselves had drawn clearly when designing CYC and
which was embodied right at the top of their categorical heirarchy) was of
no practical use, and in fact served only to enforce categorical
differences which were obtrusive and misleading. (Their interim solution
was to merge the heirarchies, but I believe they have a better fix now.)
>What I can believe is that for their particular system, they were not
>able to find any case where an individual thing and a patch of space-time
>had any distinguishable consequences within the system. That may well be true,
>but it would by no means be a fact of science or a fact of logic, but a
>mere artifact of their particular axiomatization.
They tried hard to avoid it, using several different axiomatisations, but
found the conclusion - which they themselves found unpalatable - forced
On your view, would it be POSSIBLE to discover something about ontological
structure, or would any deviation from a high-level structure extracted
from the writings of Peirce or Aristotle always have to be relegated to the
role of an 'artifact' ? (And if it could be discovered, how?)
>>The top levels of our classification networks need to be be DISCOVERED, not
>>imposed from above by 'theoretical work'. None of this ontological
>>semiphilosophising is remotely precise enough to count as 'theory' relative
>>to any empirical discovery, in any case. As soon as one comes up with
>>slightly peculiar examples, this classification gets just as muddled as any
>>other a-priori top-down classification.
>Or for that matter, just as muddled as any other a-posteriori bottom-up
>empirical classification. As I said in my paper for the ontology workshop,
>Aristotle established a paradigm for doing ontology that has never yet
>been improved upon: a synthesis of abstract metaphysics, linguistic
>analysis, and detailed empirical studies. You need all three, and any
>attempt to work purely top-down or purely bottom-up is doomed to failure.
I agree, of course. But (a) we now have 2000 years of more or less
free-rein top-down thinking, so maybe we could focus on the bottom-up for a
while, and (b) in order to work with all three, there must be some way of
connecting them together and allowing a discovery or insight in one area
constrain or direct the thinking in another.
>>My suggestion is that we simply put 'thing' at the top, and immediately
>>split it into the highest-level categories that make actual clear sense in
>>the domains we are describing....
>Some very good people have been doing that for a very long time. They
>are called lexicographers, and their results are known as dictionaries.
No, that is NOT what I mean, as I am sure you know. A dictionary is not a
KB; dictionaries don't come with inference rules.
>> ... We are in a
>>better position than most of the philosphers ever were to discover what the
>>REAL categories of human thought actually are.
>We are only in a better position if we take into account the best work
>that has been done by all of our predecessors and contemporaries and
>do the very hard work of synthesizing it with the latest results that are
>available from linguistics, physics, and all the other empirical sciences.
>But I have seen no evidence whatever that anyone has done or is doing such
>a grand synthesis.
I think our position is rather similar to that of science at the time of
the founding of the Royal Society. We have lots and lots of thinking, much
of it very fine thinking by great thinkers, but we can't trust any of it.
Our job is to discover things and try to make sense of what we discover. If
it turns out to be inconsistent with Aristotle, so much the worse for
Aristotle. Its much too soon to be talking of a 'grand synthesis'. We have
years and years of scutwork to do before anyone should have the hubris to
propose any kind of 'grand' anything.
For example, we have a number of attempts at what might be called a locally
top-down analysis of categories of events and processes to explain
linguistic regularities such as verb parsings, etc.. Ive only seen these
done for indo-european languages, however. Do they work for Japanese, or
African languages, etc..? If so, that is very encouraging; if not, it
should oblige our theoreticians (among who I number yourself) to come up
with a better account.
Beckman Institute (217)244 1616 office
405 North Mathews Avenue (415)855 9043 or (217)328 3947 home
Urbana, Il. 61801 (217)244 8371 fax