NetNews Usenet Archive 1992 #27/NN_1992_27.iso/spool/alt/usage/english/8763

Internet Message Format | 1992-11-22 | 4.3 KB

Path: sparky!uunet!decwrl!waikato.ac.nz!maj
From: [e-mail]
Newsgroups: alt.usage.english
Subject: Re: Uniquely fuzzy
Message-ID: <[e-mail]>
Date: 23 Nov 92 18:01:56 +1300
References: <[e-mail]> <[e-mail]> <[e-mail]> <[e-mail]>
Organization: University of Waikato, Hamilton, New Zealand
Lines: 79

In article <[e-mail]>, [e-mail] (Mandar M. Mirashi) writes:
> In article <[e-mail]> [e-mail] writes:
>>In article <[e-mail]>, [e-mail] (Detlef Lannert) writes:
>>> In article <[e-mail]> [e-mail] (Roger Lustig) writes:
>>>
>>>  [...]
>>>>Moreover, a thing can be "quite unique" wrt *one* basis for comparison,
>>>>by not even coming close to anything else.  After all, most of the comparisons
>>>>we do are not simply same/different; they take into account degrees of
>>>>difference.
>>>  [...]
>>>
>>> The times of the clear "unique"/"not unique" distinction are over.
>>> If there is fuzzy logic in every washing machine you can buy, why
>>> shouldn't it be in the language as well? English couldn't survive
>>> on the world market if it ignored the modern concepts.
>>>
>>> And speakers will always (have to) adapt to their language. So it's
>>> time for everyone to get fuzzy. Or should I say "quite fuzzy"?
>>>
>>
>>Both languages and washing machines would be far better without
>>fuzzy logic. It's the wrong way to handle uncertainty about real
>>events. Because it's purely formal and without any clear semantics
>>I think it's probably the wrong way to handle conceptual vagueness
>>as well.
>>--
>
>       That's the funniest thing I've heard today! :)  Even computers
> have difficulty in representing events precisely. For example, take
> the case of numbers. When you say "He's 5'10" ", when has the person
> referred to ever been exactly 5'10" feet tall? He could be 5'9.99"
> or 5'10.01" for that matter! Ambiguity always occurs in human speech
> and this because each of us has a different notion about certain
> concepts. Yes, there may be certain events which always result in a
> simple "Yes" or "No" decision (binary for all you computer lingo
> aficionados), but many events are conceived differently by different
> persons. Let me relate the case of our fuzzy logic class. The professor
> asked us to write down at what age a person was considered to be "young".
> When the responses were tallied, they varied in the range from 5-30 years!
>
>         Human beings will never be able to converse in precise terms,
> however many milleniums may pass.
>
>                                                             Mandar.
>
>
>
> --
> --------------------------------------------------------------------------------
> "Space the final frontier" - Star Trek.
> "Try to see it my way, only time will tell if I am right or I am wrong"
>                                                          - The Beatles
I am not saying that some concepts are not vague or fuzzy,
I'm just saying that 'fuzzy logic' or 'fuzzy set theory'
is the wrong way to handle them. The calculus of fuzzy logic
assigns numerical truth values between zero and one to
statements like 'Fred is old'. Deductions are made purely
by calculating truth values for conclusions of logical
rules purely as functions of the truth values of the
premisses.

This kind of theory cannot do justice to all the complex
linkages that can exist between statements. To take
your example if A and B are 'young' to the degree of
0.6 what is the truth value of the statement 'Both
A and B are young'?  But what if I tell you that A and B
are twins?

The proper way to handle uncertainty is through the theory
of probability, which treats the important notions of
independence of events and conditioning on new information.

--
Murray A. Jorgensen [ [e-mail] ]    University of Waikato
Department of Mathematics and Statistics     Hamilton, New Zealand
__________________________________________________________________
'Tis the song of the Jubjub! the proof is complete,
if only I've stated it thrice.'