Re: PHIL/AI : Humongous Lookup Table

From: Hal Finney (hal@rain.org)
Date: Thu Feb 13 1997 - 12:19:28 MST


From: John K Clark <johnkc@well.com>
> I don't see how the subjective could ever be made objective. Suppose a mind
> reading machine told me that when you see the color red you have same
> subjective experience that I do, red looks the same to you as it does for me.
> How could I ever know if the machine was giving me correct information?
> The operating principles of the machine would have to be based on some theory
> of consciousness, but I'd have no way of knowing if that theory was correct.

Ok, suppose it is done without the aid of a machine, but directly. It turns
out that ESP and telepathy is possible with the right training. In time
you learn to directly perceive the innards of another person's mind. You
see what he sees, follow his thoughts, feel what it is like to be him.

Now, if you're going to start suggesting that you don't know if this
is real, that you can't be sure that you are really feeling what he
feels, then this starts to sound like solipsism. You would be directly
perceiving his mind, and to deny these perceptions would be no more
justified than to deny the other perceptions you have. How do you
know that the world is really there, that your senses are giving you
the right information? I think the answer is that we accept it because
our senses give us a consistent picture. Other people report the same
things, our sense cross check each other consistently, and it all adds up.

If the same thing were true of these mental perceptions, then I'd say we
would have equal grounds to accept them as real. If the person whose mind
I am reading agrees with my statements about what he seems to be feeling,
if other people reading his mind perceive the same things I do, then I have
grounds for believing them to be real.

There could also be an underlying theory which puts our observations into
a logical framework. If I perceive your reds to be greenish, then it
might be logical that you would perceive my reds to have a purple shade.
With a body of consistent observations like this we have further grounds
for accepting these perceptions.

Granted, the existence of this kind of telepathy may be hard to accept,
but I don't think we could say that it is impossible. And even if it
turns out that mechanical intervention is necessary, it would still seem
that the reasoning above would apply.

>There is an even more fundamental problem. My brain is different than yours,
>so If I want to know what it's like to be you I'd need something that could
>translate the states of your brain into states that my brain could understand.
>If I did that then I'd know what it's like for John Clark to be Hal Finney.
>Unfortunately that's not what I want, and the only way for me to know what
>it's like for Hal Finney to be Hal Finney is to turn my brain into an exact
>copy of yours. Even then "I" wouldn't know what it's like to be you, because
>"I" wouldn't be John Clark anymore, I'd be Hal Finney.

This is a little too philosophical for me. It's not clear what it means
for John Clark to be Hal Finney, and how that differs from Hal Finney being
Hal Finney. Suppose that while John Clark is being Hal Finney, he feels
exactly the same as Hal Finney being Hal Finney. Then, when the mind reading
is over, he goes back to feeling like John Clark, but he still has his memories
of what it felt like to be Hal Finney (or to be Hal Finney feeling like he
was Hal Finney? the mind reels...).

In any case, as I said I think this overanalyzes the situation, like the
guy who said he wasn't seeing the world, he was seeing his retina. Given
a world in which mind reading was possible with consistent, logical results
as I described, I think we would naturally accept these perceptions on a
par with those of our other senses. There might be some people who refuse
to, just as there are no doubt people today who refuse to accept the
evidence of their senses. But this could put them at a disadvantage, and
I suspect they would be relatively few.

Hal



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:44:10 MST