On 10/1/2013 11:49 PM, Pierz wrote:
On Wednesday, October 2, 2013 3:15:01 PM UTC+10, Brent wrote:
On 10/1/2013 9:56 PM, Pierz wrote:
> Yes, I understand that to be Chalmer's main point. Although, if the
qualia can be
> different, it does present issues - how much and in what way can it vary?
Yes, that's a question that interests me because I want to be able to build
intelligent
machines and so I need to know what qualia they will have, if any. I think
it will
depend
on their sensors and on their values/goals. If I build a very intelligent
Mars Rover,
capable of learning and reasoning, with a goal of discovering whether there
was once
life
on Mars; then I expect it will experience pleasure in finding evidence
regarding this.
But no matter how smart I make it, it won't experience lust.
"Reasoning" being what exactly? The ability to circumnavigate an obstacle for instance?
There are no "rewards" in an algorithm. There are just paths which do or don't get
followed depending on inputs. Sure, the argument that there must be qualia in a
sufficiently sophisticated computer seems compelling. But the argument that there can't
be seems equally so. As a programmer I have zero expectation that the computer I am
programming will feel pleasure or suffering. It's just as happy to throw an exception as
it is to complete its assigned task. *I* am the one who experiences pain when it hits an
error! I just can't conceive of the magical point at which the computer goes from total
indifference to giving a damn. That's the point Craig keeps pushing and which I agree
with. Something is missing from our understanding.
What's missing is you're considering a computer, not a robot. As robot has to have values
and goals in order to act and react in the world. It has complex systems and subsystems
that may have conflicting subgoals, and in order to learn from experience it keeps a
narrative history about what it considers significant events. At that level it may have
the consciousness of a mouse. If it's a social robot, one that needs to cooperate and
compete in a society of other persons, then it will need a self-image and model of other
people. In that case it's quite reasonable to suppose it also has qualia.
> I'm curious what the literature has to say about that. And if
functionalism means
> reproducing more than the mere functional output of a system, if it
potentially means
> replication down to the elementary particles and possibly their quantum
entanglements,
> then duplication becomes impossible, not merely technically but in
principle. That
seems
> against the whole point of functionalism - as the idea of "function" is
reduced to
> something almost meaningless.
I think functionalism must be confined to the classical functions,
discounting the
quantum
level effects. But it must include some behavior that is almost entirely
internal -
e.g.
planning, imagining. Excluding quantum entanglements isn't arbitrary;
there cannot
have
been any evolution of goals and values based on quantum entanglement
(beyond the
statistical effects that produce decoherence and quasi-classical behavior).
But what do "planning" and "imagining" mean except their functional outputs? It
shouldn't matter to you how the planning occurs - it's an "implementation detail" in
development speak.
You can ask a person about plans and imaginings, and speech in response is an
action.
Your argument may be valid regarding quantum entanglement, but it is still an argument
based on what "seems to make sense" rather than on genuine understanding of the
relationship between functions and their putative qualia.
But I suspect that there is no understanding that would satisfy Craig as "genuine". Do we
have a "genuine" understanding of electrodynamics? of computation? What we have is the
ability to manipulate them for our purposes. So when we can make an intelligent robot
that interacts with people AS IF it experiences qualia and we can manipulate and
anticipate that behavior, then we'll have just as genuine an understanding of qualia as we
do of electrodynamics.
Brent
--
You received this message because you are subscribed to the Google Groups
"Everything List" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to everything-list+unsubscr...@googlegroups.com.
To post to this group, send email to everything-list@googlegroups.com.
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/groups/opt_out.