Re: Granny Objections to Computers' Having Minds

From: HARNAD Stevan (harnad@cogsci.soton.ac.uk)
Date: Wed Apr 24 1996 - 18:53:01 BST


> From: "Bollons Nicholas" <NSB195@psy.soton.ac.uk>
> Date: Mon, 22 Apr 1996 17:07:19 GMT
>
> Q If we are so similar to computers in respect to the fact that we both
> incorporate some sort of programming (computers in computer language
> and ours in our D.N.A)? Then we created computers, did some one then
> create us?

What has one thing to do with the other? If we can build a synthetic
liver, does that mean "someone" built ours?

All evidence is that we were "built" by the Blind Watchmaker, which is
to say, by no one at all, but merely by chance variations and their
consequences for survival and reproduction. (The successful ones were
passed on -- that's what success means -- the unsuccessful ones weren't.
That was the "building" process.)

> And if 'they' did, did they produce the same outcome that
> we have done with computers? That is : computers can do some things
> allot more competently than us humans. Are we better than 'him
> upstairs' ?

Having established that a "him" has nothing to do with it, hence neither
does "competence," we CAN ask whether evolution, blind and impersonal as
it is, has not somehow outdone itself in coming up with us: It looks as
if we, unlike other creations of evolution, can understand -- explain,
predict, control -- the evolutionary process itself. We cure diseases
that would otherwise have wiped some of us out; we selectively breed
species -- like SIGHTED watchmakers -- and now we are even able to alter
the genetic code itself, directly.

But are we, in this respect, any different from any other outcome shaped
by the blind process of evolution? Is our capacity to alter our own DNA
or that of other species really any different from our capacity to alter
our blood sugar (by eating) or to alter the blood sugar of another
species (by eating it)?

Besides, it ain't over till it's over: We have yet to see what all
our special "competence" leads to in the long run...

> Q If we are products of the 'blind watchman' of evolution. Are
> computers going to also become touched by this 'mutant manipulator' ?

What could this mean? There is no "blind watchmaker." That's just a
metaphor for the blind, impersonal, purposeless, mechanical PROCESS by
which mutation and its chance effects on survival and reproduction can
blindly "shape" structure. What has that got to do with the things we
build with our hands? They are not genetically coded, hence they do not
work according to the biological survival/reproduction measure of
success.

Of course, we can create artificial measures of success: In "Artificial
Life," computational "forms" compete to "survive" and "reproduce" in a
computational "game of life." But clearly all we're doing there is
modeling or mimcking life: The same computer could just as well be
running a payroll programme as an artificial life programme.

One thing is sure: Darwinian survival/reproduction will not touch or be
touched by robots as long as they are not part of the biosphere (as long
as they cannot mate and reproduce with any of us): Till then they are
out of the genetic lottery that evolution is all about.

> Are computers going to start adapting themselves to the
> environment ?

What do you mean by "adapt"? The usual meaning is the survival and
reproduction of gene-based systems. If computers don't have genes and
don't reproduce, why would you talk about them as "adapting" any more
than you would talk about a glacier or a river adapting?

> Do they even have an environment other than the one to
> which we have created them for?

As far as I know, there is just one environment, and we're all in it,
robots and all.

> In that case, are we the 'blind
> watchmaker' ? And if we are, hadn't we better watch out what we produce.

I think you have not quite grasped the "Blind Watchmaker" metaphor (it
IS a metaphor, not a person, or anything like a person: on the contrary,
it is a blind, mechanical process whose products merely LOOK as if they
must have been fashioned by a person).

And sure we need to watch what we produce -- but that's true entirely
independently of Granny issues and computation: Air pollution is already
evidence enough.

> Q The creativity of Einstein is predominately logical and
> computational. This can of course be compared to computers and their
> computational abilities. And no, there have not been that many
> generally computationally 'creative' people in human history.

Could I ask on what evidence you base your conclusion that (1)
Einstein's creativity was computational and that (2) there have not
been many computationally creative people in human history? There have
certainly been few people who made contributions anywhere as big as
Einstein's, but how do we know that his creativity was computational?
Maybe computation captures ordinary, average cognition, but the rare
Einsteins have something else: Let's get back to this when we get to
the Creativity lectures.

> But what of art ? There is an awful lot of artistically creative people
> in history - Michealangelo, Monet, Shakespeare - the list is endless.
> Computers can no way do anything that these chaps could do, for they
> created whole new individual artistic trends/pieces, unseen before..

The Einsteins of art are as rare as the Einsteins of science. As far as
the rest of us are concerned, the punchline is the same: It is NOT an
objection against computers that they cannot do what these Einsteins and
Michelangelos and Shakespeares can do, because 99.999% of the rest of us
can't so it either. So you can't fault computation with being unable to
do what we're unable to do either.

Besides, although I chided you for being so sure that Einstein's creativity
WAS computational before (how can you know, until a computer CAN do
what Einstein did?), I have to remind you that by the same token you have
no way of knowing that a computer WON'T be able to do what Monet or
Shakespeare did. Give 'em time: So far they can't even do most of what
WE (ordinary mortals) can do...

> Can computers do this artistic creativity, can they produce new unseen
> individual artistic pieces ?

New is no problem. "Artistic" is the problem.

We'll get to creativity later. Suffice it to say that besides requiring
something to be new in order to qualify to be called creative, we also
require that it somehow be "unexpected," which is to say, not the result
of the mere mechanical application of something that is already
familiar. But computers work on the basis of formulas (rules,
algorithms), that can be applied mechanically ("like a computer").
So there may be a contradiction between wanting something to be
creative, hence not governed by a mechanical rule, and at the same time a
computer, hence governed by a mechanical rule.

But lest you rush to say: "Aha, that's what Granny meant all along, with
her "creativity" objection!", let me point out that an artistic work
merely needs to APPEAR unexpected to our senses: It has to look (or
sound) as if it was not generated by the mechanical application
of a rule. But what if a rule could have that effect? A complicated,
subtle rule perhaps, one with which we are totally unfamiliar.

> (No use of Fractal generated imagery examples here, because they use
> input to output system, and i think that there is no way that the whole
> of the Sisteen (spelling ?)
> Chapel could have been 'outputted' by Michealangelo due to the 'input'
> he received from the 14th century).

Ah, but doesn't that just mean that, if there was some rule derived from
what came before, you can't see it? It seems as if it's pure
inspiration. But remember that in all things cognitive, great and small,
we tend to be unaware of the underlying rule: We don't know HOW we do
it.



This archive was generated by hypermail 2b30 : Tue Feb 13 2001 - 16:23:39 GMT