 |
 |
 |
 |
 |
 |
 |
 |
 |
|
|||||||||
Helping you to sip from the information firehose
Well, actually I'm not. By the time this edition of web.scan is published, I'll be on vacation. In fact, I might even be dead, yet still mysteriously publishing new work... rather like L. Ron Hubbard. Leaving such morbid and unpleasant thoughts aside, though...
Many religions put a lot of effort into answering the question: "Why am I here?" Amidst all the discussion of the meaning of life, it's easy to overlook a question which I think is much more interesting: "How and why am I aware that I'm here?"
To me, the existence of complex lifeforms isn't a big mystery; I've studied enough biology, and seen enough demonstrations of the power of natural selection. No, the questions which still stump me are the ones that deal with consciousness and self-awareness.
One controversial question is whether it's possible to produce a computer which has consciousness, or even mere intelligence. Read anything on the subject, and pretty soon you'll hear about the Chinese Room Argument. It goes something like this:
Imagine a man sitting in a sealed room. In the room are many books filled with symbols, and instructions on how to manipulate and re-write those symbols.
People outside the room feed pieces of paper into the room via a slot. The man looks at the symbols on the paper, refers to the instructions printed in the books, writes the appropriate results on another piece of paper, and passes that paper back out through the slot.
Now, suppose that the pieces of paper passed into the room are questions in Chinese; suppose also that the books contain enough complicated rules that to the people outside, it looks as if the replies are answers to those questions.
The man in the room does not understand Chinese. The books cannot be said to understand anything. Nor does the room. Therefore there is no understanding of Chinese -- just symbol-processing.
So, the argument goes, a computer which could converse intelligently would not have any understanding -- it would just be processing symbols.
Now, I personally think that the Chinese Room argument is very similar to a classic fallacy of composition -- no individual bit of the system has an understanding of Chinese, therefore there is no understanding of Chinese. It doesn't convince me for a moment.
Many people are convinced, though, and it can lead to strife. I once overheard an anguished computer scientist mentioning that his girlfriend thought the Chinese Room argument was convincing, and that they'd had some terrible rows about it.
For an excellent overview of some of the controversy surrounding the Chinese Room Argument, check out the Chinese Room Home Page. It's part of the PT Project pages at Illinois State University. There are many personal web pages attacking the Chinese room argument, sometimes in strong terms. Jay Scott bluntly states "I personally think that this argument is stupid", and goes on to outline why. Andy Ylikoski contributes a more detailed refutation, of a kind which I hadn't seen before. David Thornley describes Searle's argument as "hogwash" and offers another detailed refutation; and Larry Hauser's refutation is even more formal and academic in style.
But as Jason Corley puts it, "there is something unshakeably appealing about the Chinese Room", in spite of its flaws. The argument certainly seemed to appeal to mathematician Roger Penrose; his book The Emperor's New Mind includes a spirited defense of Searle's point of view. But I leave it to Professor Daniel C. Dennett to point out perhaps the biggest flaw with the Chinese Room Argument -- that it's not actually an argument at all.
Bruschi Mario produces a reductio ad absurdum argument, based on the structure of Searle's argument: Just as the Chinese Room does not understand Chinese, because no single part of it has understanding -- so the human brain does not understand anything, because no single part of it has understanding.
Which brings up an interesting question -- what do Searle and Penrose think gives the human brain intelligence and self-awareness? The answer can be found on Stuart Hameroff's web pages, where he describes the "Penrose-Hameroff Model", which theorizes that the brain actually performs quantum computation, and that this large-scale coherent quantum state is what is necessary for intelligent conscious thought. Our consciousness might be quite literally tied to the universe.
It's an interesting idea, and it's easy to see why it might appeal to a physicist: for starters, it explains why conscious observers have such a strange role in quantum mechanics. It also seems as though it might explain why we perceive time in a fundamentally different way to any other dimension of spacetime.
Ultimately, though, I'm still a little more convinced by the Emergent Behavior argument. Kevin Kelly's book Out Of Control provides an enjoyable introduction to the subject; don't be put off by the fact that he's executive editor of fashion magazine WIRED...
The Emergent Behavior argument is that as systems become more and more complex, their behavior doesn't simply become more complex; at some point, the system begins to behave in totally new ways -- so-called 'emergent behavior'. Intelligence and self-awareness, the theory goes, are forms of emergent behavior. Give a neural network the right kind of complexity and the right kind of training, and it just might be possible to get it to behave intelligently.
Then we can teach it to watch TV, and bring it back down to our level.
mathew
|
[ e-mail the URL of this page ]
[top of page]
|
|||||||||
|
|||||||||
|
|
|||||||||
|
Copyright© Internet Infidels® 1995-Present. All rights reserved.
|
Last updated: Wednesday, 30-Nov-2005 17:06:10 CST
