X-Message-Number: 11607
From: Thomas Donaldson <>
Subject: The utter failure of Turing machines
Date: Sat, 24 Apr 1999 23:33:12 +1000 (EST)

To Mike Perry:

A real person must not only be able to think but also act, with all the
extra stuff required to do that. That "extra stuff" actually involves
lots more in both programming and physical devices; I will note the
attempts to devise a computing machine able to act in the world, beginning
with a very stripped-down world --- but such attempts have consistently
failed to produce a machine able to act in the REAL world, with all its
confusion and refusal to fit simple ideas. The classic case of this 
would be black swans: prior to the discovery of black swans in Australia,
swans were DEFINED as white. But for some reason those who saw the black
swans realized at once what they were. Ultimately our language and
thinking comes down to the level of pointing at things, and our thinking
does not depend on definitions but on that pointing. This process cannot
be done solely with a computer program, no matter how complex. It takes
a good deal of extra (HIGHLY PARALLEL) machinery.

Turing machines can only think and report what they are thinking (and
that alone requires a quite advanced program!). A really bright Turing
machine with the right peripherals might even be able to receive language
input from someone, but that too fails to be sufficient: our language is
only our (faulty) representation of the world, not the world itself.

So Turing machines can (alone, with no other faculties) never come close
to a real person. Their original purpose, to represent a computer, 
remains very useful, but even now they have become very poor
representations of ANY computer currently in use.

As someone who has been very involved with parallel computing ever since
I "left" mathematics, I will also add that no device could really deal
with the real world unless it were parallel. Too many things are going on
at the same time, and all of them must be dealt with now or soon. (Think
about driving a car on a current road --- adding lots of devices to the
road to help the computer isn't allowed for this example. For that
matter, think about what's involved in walking along a path). Just by 
being sequential, Turing machines could not cope with the real world.

Finally, I too want to see what comes of attempts to amalgamate both
quantum theory and general relativity. And I'm glad that such attempts
are proceeding. But I do want a valid amalgamation, proven by experiment,
not just a theoretical one.

			Best and long long life,

				Thomas Donaldson

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=11607