. perhaps the closer analogy (see below) is Aristotle .

*****

Peter Tillers            <http://tillers.net/> http://tillers.net
Professor of Law
Cardozo School of Law, Yeshiva University
55 Fifth Avenue, New York, NY 10003

(212) 790-0334; FAX (212) 790-0205


[EMAIL PROTECTED]

-----Original Message-----
From: Peter Tillers [mailto:[EMAIL PROTECTED]
Sent: Saturday, February 07, 2004 2:26 PM
To: '[EMAIL PROTECTED]'; '[EMAIL PROTECTED]'
Subject: RE: [UAI] functional_vs_causal models

Dear All,

I know -- or believe -- that Zadeh's theory is not Platonic. But there
is a loose isomorphism between the Platonic theory of being, Platonic
ontology, and Prof. Zadeh's notion of partial truth. In Plato's
scheme, things or entities have varying degrees of being. In Zadeh's
scheme matters have varying shares or amounts of truth. In Zadeh's
partial truth scheme attributes are distributed unequally among things
or entities -- or things and entities have varying amounts of an
attribute (which is a very Platonic way of thinking about the
relationship between concepts and instances or elements) -- and fuzzy
logic is a grammar that allows a description of that sort of varying
sharing of an attribute.

Is Zadeh's way of thinking about the relationship between classes or
sets and elements useless? I wonder. In some contexts -- e.g., in the
context of reasoning about and interpretation of legal norms and
principles -- Zadeh's way of picturing reasoning seems helpful.

Is it necessary for Zadeh or proponents of fuzzy sets to show that no
other way of reasoning about the relationship between concepts and
elements is helpful? I wouldn't think so.

Is it necessary for a proponent of fuzzy sets to show that the theory
of partial truth teases out implications that other forms of reasoning
do not?  I would think so.

* Professor (emeritus but fully-active) Lothar Philipps of the
University of Munich has done a bit of work along this line,
suggesting that fuzzy logic can explain some legal results. (To get
details, you need to turn to him, and not to me.)

But here matters get difficult in contexts such as law -- because of
general issues about the appropriateness of using a formal logic or
procedure to reason in real-world legal settings about matters such as
interpretation of legal rules or, even, about seemingly non-normative
factual issues.

Sincerely,

  Peter T

*****

Peter Tillers           http://tillers.net
Professor of Law
Cardozo School of Law, Yeshiva University
55 Fifth Avenue, New York, NY 10003

(212) 790-0334; FAX (212) 790-0205

[EMAIL PROTECTED]

- -----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of
[EMAIL PROTECTED]
Sent: Saturday, February 07, 2004 12:10 AM
To: [EMAIL PROTECTED]
Subject: RE: [UAI] functional_vs_causal models



Greetings :-



    Thank you, Professor Zadeh, for your reply to my inquiry. Your
remarks reinforce my impression that these exchanges would benefit from
an authentic definition of partial truth.

    I confess anew that I am unable to guess the meaning from the
examples. "Check-out time is 1 pm," for instance,  is laconic but clear.

    After 1 pm, one has something to negotiate with the
inn-keeper. The negotiation may be trivial, and the sign does not say
otherwise. The sign simply tells when there need be no negotiation at
all.

    As for the count of tall men in a room with no definition of
"tall," then I imagine a situation similar to being asked for the
count of men who stand 190 or more cm. with no tape measure.

    Both problems dissolve if I am supplied with a standard for
classification which can actually be applied. Failing that, you will
in each task get an estimate, my estimate, rather than the count.

    If you would prefer to call your estimate in either situation "the
count," then go ahead, so long as we both know that you give an
unusual meaning to a common word; two common words, in fact, since
"the" rarely means "my" in careful speech.

    I thank Professor Zadeh for moving a bit beyond examples in his
discussion suggesting that one might understand "partial truth" by
analogy to other partial things.  _Partial_ is indeed a fine concept
with many varied applications.

    I do not know what it means in connection with the truthfulness of
a proposition.  I also do not know what it means in connection with
other categorical estates, such as pregnancy, humanness, or
left-handedness.

    That prevents my understanding of partial truth by analogy to
other uses of _partial_. I am not denying that there could be a
meaning for _partial_ adjoined to any of those things, only saying
that the meaning would not be obvious, and so asking Professor Zadeh
what his meaning is in the case of partial truth.

    Please define partial truth, Professor Zadeh.

Reply via email to