X-Message-Number: 21661
Date: Fri, 25 Apr 2003 00:21:30 -0700
From: Mike Perry <>
Subject: Computation and Language

A Reply to Thomas Donaldson, #21647:

>1. Quantum computers don't just compute faster. They compute
>    instantaneously (for some problems).

Some questions have been raised about such issues as error tradeoffs. How 
hard it may be, in terms of total energy input, to really get useful 
results by quantum computing has yet to be established. And setting up a 
problem to be solved must consume nonzero time, effectively limiting the 
speed of computation at least a little. I've also heard claims that quantum 
computers can solve the halting problem, unsolvable by Turing machines, but 
there are some difficulties with that position too, such as what is really 
meant by a "computation." A working halting-problem solver meanwhile has 
yet to be produced, along with other practical devices based on quantum 
computing. That, of course, does not mean it can't be done, but clearly 
there are major problems to be solved, if they can be.

>Nor is a qubit the same
>    as a bit. Regardless of what Deutsch or any other philosopher
>    may say, a quantum computer doesn't look to me at all like
>    the device moving slowly over a tape marked with single
>    bits which was the original Turing machine ... and out of
>    which the definition of a Turing machine came.

It still appears to me that a Turing machine could simulate a quantum 
computer, albeit inefficiently. This would mean that a classical computer 
could also simulate a quantum computer, though again inefficiently. In 
practical terms it could mean that a classical computer will not be a 
feasible simulator of natural nervous systems. That much I'll concede, 
though clearly it remains to be seen what the powers of such devices will 
be, as well as what will be desired of them.


>2. You claim, despite my examples, that a language could be
>    developed that would be clear to any creature capable of
>    symbolism. Mathematics itself did not come purely from our
>    imagination but from our attempt to deal with the world.

Which could make it easier not harder to decipher by aliens who would also 
know about the world.

>    As mathematicians discovered in the 19th Century, systems
>    of axioms for geometry could have several interpretations.

In that case the ambiguity is inherent in the system, something that could 
be recognized for what it is.

>    Yes, math does have one advantage, in that we could begin
>    our discussion of it with numbers ie. * 1  ** 2  *** 3,
>    etc. But in doing so we are attaching symbols to things
>    in the world, once more.
Well, numbers aren't exactly "things in the world" but are already 
abstracted. It seems reasonable to me, though, that the whole of 
mathematics could be expressed in decipherable form starting with simple 
patterns to represent the whole numbers and other elementary concepts. You 
could also have digitized visual images, in a format not hard to decipher, 
as part of the language. Whole movies running to considerable length could 
thus be presented (with added sound and color, let's say), to aid in 
decipherment and conveying information.

Mike Perry

Rate This Message: http://www.cryonet.org/cgi-bin/rate.cgi?msg=21661