From - Tue Nov 18 22:17:18 1997
Received: (from rhm@localhost) by pumpkin.cdepot.net (8.8.5/8.6.9) id WAA28037; Mon, 10 Nov 1997 22:53:28 -0800 (PST)
Date: Mon, 10 Nov 1997 22:53:27 -0800 (PST)
From: Richard H McCullough <rhm@pumpkin.cdepot.net>
To: "Patrick J. Hayes" <phayes@nuts.coginst.uwf.edu>,
        John McCarthy <jmc@cs.sanford.edu>
cc: David Kelley <ios@ios.org>, "David S. Burnett" <dsburnett@lucent.com>,
        "Richard H. McCullough" <rhm@cdepot.net>
Subject: Montague-Scott structure
Message-ID: <Pine.BSD/.3.91.971110215607.25667A-100000@pumpkin.cdepot.net>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
X-UIDL: 67a6dacfb54654aed377ebd44447e580
Status: O
X-Status: 
X-Mozilla-Status: 0001
Content-Length: 1752

I have done a lot of reading today in "Reasoning About Knowledge"
by Fagin, Halpern, Moses, Vardi.
The most pertinent issues are discussed in "Logical Omniscience"
(Chap. 9).  The Montague-Scott structure appears to be the closest
of their models to my KR.

My model is
in KR:
  knowledge := I do identify existent
in English:
  knowledge is the identification of a fact of reality

The axiomatic concept "identify" is the action of concept-formation.
It is not the same as "know".  "focus" (roughly "is aware of") is
one of its components.  I have worked at the decomposition of
"identify" off and on over the past year.  My latest decomposition is
  identify is "sense; perceive; focus; partition; measure; choose; define"
where the component actions are
  sense-data := I do sense entity      # transient
  percept := I do perceive sense-data  # permanent memory
  I do focus percept
  entity := I do partition percept
  attribute := I do measure percept
  name := I do choose word
  definition := I do define entity

In this decomposition the "focus; partition; measure"
sequence is the biggest problem area, in terms of my
being able to identify what is going on.  I expect to
get some new insights by reading "The Evidence of the
Senses" by David Kelley.  I have just started reading;
I think he is saying that a percept is actually a
proposition of the form
  this has that
where I have to discover what the entity "this" is,
and what its attribute "that" is.
If this is true, it is saying that I directly
perceive "low-level" attributes, but I consciously
measure "high-level" attributes.  For example,
I perceive the color of an object, but I measure
its length.

Dick McCullough
<rhm@cdepot.net>


