ONT Re: determining contingent synthetic truth
- To: Seth Russell <seth@robustai.net>
- Subject: ONT Re: determining contingent synthetic truth
- From: Jon Awbrey <jawbrey@oakland.edu>
- Date: Tue, 15 Jan 2002 10:56:16 -0500
- CC: Ontology <ontology@ieee.org>
- References: <3C1BF049.A59F9516@oakland.edu> <3C1CE3B5.44E0E460@oakland.edu> <3C1E43F9.B10FF51D@oakland.edu> <3C1FA753.7DDE2970@oakland.edu> <3C41BD11.FA7B5D7F@oakland.edu> <3C43055A.1F632E5@oakland.edu> <3C443482.6A9EB984@oakland.edu> <011801c19dd8$d4c101c0$657ba8c0@c1457248a.sttls1.wa.home.com>
- Sender: owner-ontology@majordomo.ieee.org
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤
seth, ont-list, as i already used my suo ticket for today.
but what you are talking about is what is commonly meant
by 'computing a predicate', say, a predicate q : X -> B.
that is, if somebody gives you a value x in X, you have
to return the value q(x) in B. 'recursive' just means
that you have 'recourse' to a finite base of resources,
like primitive functions that you are given for free.
it does not mean viciously circular, clever, loopy,
self-applied, or anything like that, except that
to 'make infinite use of finite means' you will
typically be forced by a brand of pragmatic
pigeonhole principle to return to the same
means over and over again on a 'recurring'
or an 'ultimately periodic' basis. when
you talk about programs in general, and
not just this or that finite mechanism,
then you are stuck with these issues.
jon awbrey
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤
JA = Jon Awbrey
SR = Seth Russell
SR: Consider a situation that tests real time reality or a situation where
we need to run a program or test a voltage against a threshold before we
can determine if an object has an attribute or is an instance of a class.
How do we express the relationship between the class (or attribute) and
a program which senses it? For example a program that asks a person
a question and returns a value {True, False, NoAnswer} according to
their answer to the question: "Are you married?" Where do we put
that *program* in our ontology? Why can't I find programs anywhere?
JA: seth, if you are really interested in predicates that are computable by programs,
and not just being coy, then you need to look into computability theory, of which
i don't think you'll find a bunch in sumo world. my first ventures into the area,
in many ways still my favorite books, were arbib, and minsky & papert. also good
for quick brushes with the topic are boolos & jeffrey, cutland, and davis. enjoy!
SR: No! I'm not talking about computing the predicates themselves, I'm talking
about a program that would ascertain the truth of a predicate in real time.
Nothing coy, or clever, or recursive needed here. I'm just talking about
the input/output ports of the computer. Suppose I wire my light switch to
a computer memory location, and contrive that if the light switch is on,
then address location 12345 contains hFF otherwise it contains h00.
Then I write a program, call it IfMyLight, which returns True when
address location 12345 contains hFF. My question is: in the sumo
ontology, how do I express the knowledge that the truth of the
predicate (attribute MyLight On) is contingent on the program
IfMyLight returning a value of True?
SR: Let's not make this more complicated than it needs to be!
¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤~~~~~~~~~¤