Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: New paradigm again

Author: karen Dall Lynn

Date: 08:37:03 02/19/02

Go up one level in this thread


On February 19, 2002 at 10:44:31, José Carlos wrote:

>On February 19, 2002 at 10:07:12, karen Dall Lynn wrote:
>
>>On February 19, 2002 at 09:24:07, José Carlos wrote:
>>
>>>  As it has been brought up again and I didn't give my opinion in the past, I'd
>>>like to say what I think about all of that.
>>
>>
>>  [DIET QUOTE PLEASE REFER TO THE PARENT POST ABOVE]
>
>  Sorry, I don't understand what you mean :(
>
>>  I think it is good to distinguish between fast-dumb and slow-smart, and that
>>>they can be cosidered two paradigms in computer chess programming, at least, two
>>>schools (I don't know if this direct translation is correct in english). But
>>>magic doesn't exist. It's all about 1's and 0's...
>>>
>>>  José C.
>>
>>
>>I agree with your detachment of two main "rationales" for chess programming.  I
>>also agree that so far chess programs may have all their routines reducible to
>>even less than 010100101001101010 tapes, indeed to any set of tokens that
>>restricts the whole processes to syntactical processing.
>>
>>But the philosophical point here is not that the chess programs, as similar
>>behaviorial programs, lack psychological or intentional characteristics in their
>>bottom; the interesting point is that in a blind experiment we humans can
>>**remain in doubt** if our opponent is a program or not -- this doubt being
>>imprinted on us for no more thant the syntactical competence of the program.
>>
>>In this sense, chess programs *do have* the psychological accent int heir
>>rockbotom they lack. For all practical proposes, when some cyborg wins an online
>>game while cheating by the use of a program, and the other complains: "you're a
>>program" there remains the benefit of doubt when the cheater replies: "I am
>>not", provided he/she is really cheating.
>>
>>Allan Turing was honored when contemporary cognitive philosophers gave his name
>>to this test: Turing Test. A program is said to have passed in the turing test
>>when, in the best of their effort, humans cannot be sure (again in a blind
>>exeperiment) if they are playing chess against a program or against a human.
>>
>>Now, if some future chess programming will be smart in a different way - for
>>instance, semantically (not only syntactically) smart - and if some
>>unconvencional programming will come to use (for instance, fuzzy programming,
>>paraconsistent programming etc) -- this would break your dichotomy between
>>slow-smart and fast-dumb into a new qualitative category. But that's for the
>>future, imho.
>>
>>Karen
>
>  I agree. That category might exist one day. But my point is that even when
>that happens, it still will be an illusion in the sense that behind all of that
>abstraction we still will have the 1001001110...
>
>  José C.

Tks for replying.

We cannot possibly know in advance that all future programming will be
discretely syntactical, involving a huge generalization of a Turing Machine (a
function-box that receives a continuous tape with one of two symbols of an
alphabet; in each read, it either changes the symbol or keeps it unaltered).

There are concepts that cannot be formalized that way - for instance,
continuously graded concepts like the concept of "baldness" (What would be the
0/1 blend you'd attribute for a man - if 1 stands for "hair" and 0 for "no hair"
in each of his head points - in order to start to call him a bald man?). Or -
how could you formalize, in purely syntactical terms, the funny side of the
following question:

"Laden, where have you bin?"

in the face of which most of the English native speakers would perhaps laugh?

Some has said that the intelligent act must involve fuzzy concepts; in this
case, future intelligent programs may not be able to restrict themselves to
synthactical procedures and syntactical programming.

Karen


P.S.: [DIET QUOTE...] means I made your posting "thinner", supressing part of it
so that my posting would not be too long. [...PLEASE REFER TO THE  PARENT POST
ABOVE] is a warn to the reader that he/she has to read your complete posting in
the message to which mine was a follow-up  (the "parent" post in the tree) in
order to know the complete statement of your position.

Karen



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.