Re: Converses (Re: Availability of the ANSI standard proposal?)
phayes@uiuc.edu (Pat Hayes)
X-Sender: phayes@casper.beckman.uiuc.edu
Message-id: <v02140b04ad7e816e7a9a@[128.174.210.75]>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Date: Wed, 27 Mar 1996 00:47:13 -0600
To: cg@cs.umn.edu, brayman@zuben.ca.boeing.com (Bill Brayman)
From: phayes@uiuc.edu (Pat Hayes)
Subject: Re: Converses (Re: Availability of the ANSI standard proposal?)
Cc: goolsbey@cyc.com, interlingua@isi.edu, murray@cyc.com, srkb@cs.umbc.edu
Sender: owner-srkb@cs.umbc.edu
Precedence: bulk
>>I agree with Fritz that at a fundamental level there is only one relation,
>>i.e., that converse(s) should not be represented. However, at a more
>>"user-friendly" level, they seem to be needed, e.g., it is convenient
>>to represent both of the notions "parents-of" and "children-of".
I bet this only happens when English has the alternative words handy. Take
the relation between three points of A being between B and C. Ive seen this
written as between(A, B, C) and also as between(B, A, C) , and both
seem reasonable.
This raises another issue, however. Its all very well to say that only one
ordering needs to be kept and not the converses, but how is it specified
what the canonical ordering IS? If I write between(d,e,f), what order are
those points supposed to be in on a line? Where is that information
represented?
..
>What Fritz only hinted at is the ontological work that still has to be
>done to >represent views. It vaguely reminds me of Robin Lakoff's (or
>someone's) >observation that kids learn about dogs and cats first and only
>much later about >taxonomies and felines and other more canonical forms.
I think what Bill is referring to here is that kids learn the concepts in
the 'middle' of the heirarchies first, dogs and cats before mammals or
before pekinese. This actually is suggestive that these middle classifiers
are the most informative: they cut up the set of examples more efficiently
than the finer ones because they are more evenly balanced. But the more
things you know about, the higher the 'most useful' splits get to be, which
is why we tend to make abstractions as we get more knowledgeable.
>So, what cyc and other systems have to do is be able to make statements
>about >dogs and cats even tho they may be built upon notions of taxonomies
>and felines >and canines.
>
Yes, and what this suggests is that the classifier has always got to be
ready to make 'higher' divisions. Maybe you never came across a fruit: when
you do, you have to somehow contrast 'fruit' with 'nonfruit'. The wierder
the new thing you find, the higher in the classification tree you have to
be ready to make the split.
Question: do isa-trees always have to have a top node?
Pat Hayes
----------------------------------------------------------------------
Beckman Institute (217)244 1616 office
University of Illinois (217)328 3947 or (415)855 9043 home
405 North Mathews Avenue (217)244 8371 fax
Urbana, IL. 61801 Phayes@uiuc.edu
----------------------------------------------------------------------