http://richarddawkins.net/article,2580,Is-Science-Killing-the-Soul,Richard-Dawkins-Steven-Pinker-Edge
It raises two important points, which I will address in separate posts.
First, I'd like to address the issue of qualia. Here is Pinker's response to one of the questions:
The part that remains a mystery [about consciousness] is why the purely subjective aspect of experience should exist at all. Some philosophers, such as Dan Dennett, argue that that isn't a scientific problem and may not even be a coherent question -- since, by definition, pure subjective experience has no observable consequences, we're wasting our time talking about it. I think that goes too far, but it is possible that the existence of subjective first-person experience is not explainable by science. When cognitive neuroscience completes the story of how the brain works and predicts every last itch, every last nuance of color and sound in terms of the activity of the brain, one can still wonder why it feels like something to see and touch and taste. My own hunch is that this unsatisfied curiosity may itself be an artifact of how our brains work.
Qualia refers to this subjective experience--the feeling of pain, or sensation of the color red. So the question is why do I feel these things subjectively, rather than just behave like a robotic automoton. And to me the answer is: the only way to get a complicated thing to reliably and successfully navigate a complicated environment is to make it have subjective feelings. An organism with more computational power in its brain needs to have deeper, more visceral feelings in order to adapt to more varied environments and solve more complicated problems over a longer period of time. By contrast, an organism with very little computational power can successfully replicate under a very simple set of rules, e.g., grow roots downward, grow branches upward, turn leaves toward sunlight, etc. Let's say you build an organism as smart as a dog. Its brain is capable of navigating the animal through varied terrain, using its jaws to attack other animals, use its paws to dig holes, etc. How do you ensure that this rather smart brain does the right things in order to spread its genes? There is no simple rule-book that you can write down for it, like "take ten paces forward, bend head downward, extend tongue to drink". Life is too complicated--this organism must figure things out on the fly. But in order to get the organism to behave in evolutionarily successful ways, it has to enjoy fresh water, and dislike salt water, and enjoy sex, and dislike physical wounds.
Let me borrow on Dennett's idea of the zombie, which is a person who does all of the things that a normal person would do but has no subjective experience. If I remember correctly, I think that Dennett's argument is that even if we were all zombies, we should still treat each other the same way. I guess I would say that zombies simply don't exist. That is, there is no way to make an evolutionarily successful organism as complex and intelligent as a person without giving that organism subjective feelings. Otherwise, they would forget to eat. They would forget to have sex. They would forget to take their hand out of the fire. Their genes would die off rather quickly.
I think this is an important lesson for artificial intelligence. To get an individual to adapt to complex surroundings, it needs to have desires. The way to create those desires is with neural nets.
No comments:
Post a Comment