Path: utzoo!attcan!uunet!lll-winken!ames!xanth!nic.MR.NET!shamash!tank!staff_bob@gsbacd.uchicago.edu
From: staff_bob@gsbacd.uchicago.edu
Newsgroups: comp.ai
Subject: Re: Thought/Emotion/Feeling
Message-ID: <1380@tank.uchicago.edu>
Date: 11 Jan 89 22:23:16 GMT
Sender: news@tank.uchicago.edu
Organization: University of Chicago Graduate School of Business
Lines: 89

>In article <1867@cadre.dsl.PITTSBURGH.EDU> geb@cadre.dsl.pittsburgh.edu (Gordon E. Banks) writes:
>>Now I know where that idea about language and consciousness
>>evolving in the last 10000 years came from!  Jaynes' thesis is
>>considered quack by every anthropologist I have talked to.
> 
>I'm not sure why you think anthropologists are the right people to ask.
> 
>Disagreements on this subject are often due to different ideas about
>what "consciousness" means.  Many people think animals are conscious, which
>seems rather unlikely given how much of our internal experience involves
>language.  I'm not sure what it means to say a rat is conscious.  But
>some people think consciousness is something rats might have, and they
>would certainly disagree with Jaynes.  But then I'd say they were talking
>about different things.
> 
>And, in any case, Jaynes says some interesting things about consciousness
>that seem to be independent of his "bichameral mind" ideas.

Have you ever noticed that psychologist, including Jaynes, tend to equate
consciousness with the ability to speak? That is to say, if a subject has
perceived a stimulus, or reacted to it, without being able to say that
he perceived it or reacted to it, he is thought not to be conscious of it.
This is manifest in Jaynes, who is at best a very sloppy academic, but
who nonetheless raises many points which should not be overlooked. that
fact is that by his own definition, consciousness cannot have existed
before language, if only because it could not have been measured!

One often hears the assertion that no matter what, we will never create
a machine that is truly conscious, as we are. Certainly, we will never
have to admit that one is. To date, there is no general concensus that
any animal is consciously aware, and this is because that we cannot
observe consciousness, except directly. DesCartes once 'proved' that
animals were not conscious because they could not speak. What would
he say now, that we have taught chimps to use sign language? I have
observed my dogs making logical deductions, and yet I could not prove
this to anyone, let alone to a skeptical philosoper. We shall never
be able to demonstrate consciousness in a machine, the best we will
be able to do is the best we've ever done. We can ask it, 'Are you
conscious?'

As I have said, I don't think much of Jayne's method, but the fact that
establishment anthropologist consider him a quack speaks in his favor,
as far as I'm concerned. It seems quite probable to me that in the
last three thousand years mankind has been transformed from a right-brain
dominated culture to a left-brain dominated culture. I'm not prepared
to accept the bicameral hyposthesis, but I don't think that it should
be rejected out of hand. In any case, it's clear that 3000 or so years ago,
men started to write things down. They began to study mathematics.
The march of reason through history has not been steady and consistent,
but it is nonetheless clear that the rationality of mankind has progressed
during that interval, at the expense of what may be characterized at
right-brained ideals.

So why all this in an AI discussion? I think that if we are going to
understand intelligence, we need to delve more deeply into the nature
and character of the irrational, illogical, non-symbolic, intuitive,
superstitious mind, which is to say, the right half of the brain. It
seems emminently likely that without reproducing this functionality,
we could never pass the Turing test. Moreover, it seems clear from what
little I have read of these discussions that there is little interest in
these issues on the part of the AI community. Most AI researchers seem
to believe that we can duplicate that functions of the mind through symbol
processing, and that intelligence is in fact only that which is performed
by the left hemisphere of the brain! The point about reductionism is
well taken in the sense that most AI seems focused upon reproducing 
left-hemispheric processing, thereby ignoring at least half of what
constitutes human intelligence. Logic, analysis, speech, and language
are only part of the problem. I suppose that pattern matching and image
identification are most properly right-brained functions, but it seems to
me that in the main we *have* reduced the problem of intelligence too 
far.

In this light, I would like to also comment on some of the proposed
'definitions' of intelligence. A definition is a rational, symbolic
construct. It is a product of speech and the left hemisphere of the
brain. As such, it may satisfy the needs of that hemisphere for a 
verbal catagorization of phenomena, but it can never capture the 
'thing in itself' or essense of that which it describes, except in the
case of so-called _a priori_ reason, since _a priori_ reason is itself
a construct of the left hemisphere. In particular, defintions are ill
suited to capturing the essense of the right brain and conveying that to
the left brain. Since we have no reason not to expect that something of
the right hemisphere should be included in what we think of as intelligence
(certainly it must be so by the Turing test defintion), it follows that
no definition will ever suffice. If this bothers you, recall that many
very important constructs, such as 'truth', 'beauty' and 'justice' have
defied definition for thousands of years. The search for a pleasant
definition of intelligence is a vain effort to reduce a process to
a concept.