|
Authored by: Anonymous on Friday, July 20 2012 @ 05:42 PM EDT |
I don't actually care whether neuroprocessing is classicaly
chemoelectric or whether quantum effects play some role.
(I haven't seen the TV show you refer to, but it sounds like
a large leap from "microtubules are small" to "neurons must
be massively parallel quantum computers.")
The point is that the brain operates entirely on physical
processes.
I can't disprove PolR's idea that brains might operate in
some non-algorithmic fashion; I don't even have a clear
conception of what non-algorithmic operation would look
like.
But I suspect that there's nothing magic about the way
humans do semantics. All we do is associate (brains, being
neural networks, are rather like massive associative
memories, but that's still algorithmic if I'm not mistaken)
some input with some behavior.
There's no way to test whether we "understand" semantics
other than to examine our behavior. Do I know what a "frog"
is? Sure, that's an English word that I've been trained on,
probably hundreds of thousands of times over my lifetime so
far. I can produce behavior that will convince you that my
understanding of "frog" is as good as yours - possibly
superior, as I've dabbled in biology and taxonomy, and also
history and etymology. There is no doubt in my mind that a
properly programmed computer, given similar training, could
do exactly the same.
Where I'm coming from is that this whole argument reminds me
of the idea that humans have a "soul". It's invisible,
nobody can prove it's there. "Understanding semantics" is
like that. Fine, assume it's there. But be honest and
don't assume that when a human behaves a certain way, it's
because of the invisible thing called
soul/consciousness/semantics, and when a computer or non-
human animal behaves the same way, they're just faking it. [ Reply to This | Parent | # ]
|
|
|
|
|