meekerdb wrote:
On 5/16/2015 4:44 PM, Bruce Kellett wrote:
Telmo Menezes wrote:
On Sat, May 16, 2015 at 2:48 PM, Bruce Kellett
    So you think that Darwinian evolution produced intelligent zombies,
    and then computationalism infused consciousness?

No. What I am saying is that consciousness is not a plausible target for gradual evolution for the following reasons:

1) There is no evolutionary advantage to it, intelligent zombies could do equally well. Every single behaviour that each one of us has, as seen for the outside, could be performed by intelligent zombies;

Do you find it an advantage to be conscious in your everyday life? Do you really think that your partner and/or children are zombies?

I think the obvious inference is that one way to gain intelligence is to become conscious. But being conscious isn't just one all-or-nothing ability. There's being aware of the environment...being aware of oneself...being able to model oneself in an environment, including a social environment in which you model other beings as aware and similar to yourself. I think there can be different kinds of awareness just as there are different modes of perception. Whether intelligence at the human level could be realized without the kind of self-modeling implicit in consciousness seems doubtful, but it might be possible in some radically different kind of AI. That wouldn't necessarily mean it wasn't conscious, but only that it might be conscious=self-aware in such a different way we'd have a hard time recognizing it. For example, most of our memories are reconstructions as opposed to faithful recordings. If there were an AI that had enormously greater memory capacity, it might make a lot more intelligent decisions just by repeating some remembered action (Big Blue played chess sort of like this as compared to a human player).

Yes. If you take an evolutionary perspective there are clearly degrees of consciosuness and degrees of intelligence. I think you are right that full human intelligence without consciousness, the inner narrative, and subconscious activity, is essentially impossible. I don't find the idea of philosophical zombies very likely. Also, things that make for 'intelligent' behaviour in humans are often no more than heuristics, rough rules of thumb. But if one is to reflect on these heuristics, and improve them according to experience, one must be conscious, and have that inner narrative, emotions, values, goals, etc. Zombies wouldn't have any of this, so would not appear adaptable or intelligent to us.


2) There is no known mechanism of conscious generation that can be climbed.

If consciousness is just a matter of self-modeling, then it is an advantage that would show up in intelligence of decisions and natural selection could act on it.

For example, we understand how neurons are computational units, how connecting neurons creates a computer, how more neurons and more connections create a more powerful computer and so on. Evolution can climb this stuff. There is no equivalent known mechanism for consciousness.

This is the tired old creationist crap saying that the eye is too complicated to be explained by evolution; the bacterium's flagellum is too complicated...; the ...... is too complicated .....

At bottom, it is just an argument from ignorance. You do not happen to know a mechanism whereby consciousness could develop from simpler forms. But that does not in any way mean that such is not possible. Creationist anti-intellectualism yet again....


I don't know if intelligent zombies are possible. Maybe consciousness necessarily supervenes on the stuff necessary for that level of intelligence. But who knows where consciousness stops supervening? Maybe stuff that is not biologically evolved is already conscious. Maybe stars are conscious. Who knows? How could we know?

I see it as an engineering question. When we can reliably create AIs that act intelligently and appear as conscious as other human beings and we can create AIs that have more or less humor, guilt, ego,... Then we will have understood it.

I agree. The only way we will ever finally understand intelligence and consciousness is to build it -- the engineering problem.

Bruce

--
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 [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/everything-list.
For more options, visit https://groups.google.com/d/optout.

Reply via email to