Science

What is Consciousness For?

Posted 27 Jul 2005 at 22:50 UTC by steve Share This

Scientists and philosophers have argued over what consciouness is and how to give it to machines for years. Lee Pierson and Monroe Trout have released a new paper that addresses a more fundamental question, What is Consciousness For? (PDF format). They also answer the question: "Our hypothesis is that the ultimate adaptive function of consciousness is to make volitional movement possible. All conscious processes exist to subserve that ultimate function. Thus, we believe that all conscious organisms possess at least some volitional capability. Consciousness makes volitional attention possible; volitional attention, in turn, makes volitional movement possible. There is, as far as we know, no valid theoretical argument that consciousness is needed for any function other than volitional movement and no convincing empirical evidence that consciousness performs any other ultimate function."


IAAP, posted 28 Jul 2005 at 15:13 UTC by dogsbody_d » (Master)

Consciousness needed for volition? Doesn't that strike you as a bit circular?

This is new..., posted 28 Jul 2005 at 17:25 UTC by Axiom » (Observer)

I haven't heard anyone talk about free will for a while now... I'd say this is a step in the right direction.
There are so many AI gurus out there who claim to have "solved AI" - the most popular approach being to claim that consciousness is superfluous, and that once computers become fast enough, qualia and human-like behaviour will just magically emerge. Our own AI4U seems to be under the impression that some cryptic Forth program running on his server contains all there is to human nature... The whole thing seems to me like one giant evasion on the part of the AI field i.e, "the problem of consciousness is too hard, so let's just ignore it and push ahead hoping things will work out somehow."
This paper doesn't have the answer of course, but the whole point of it is to stress the importance of the question. Searle et al have been saying as much for a while now, but let's home Mr. Pierson and Mr. Trout have better luck with it.

Freewill, posted 28 Jul 2005 at 17:40 UTC by steve » (Master)

On the topic of freewill, I highly recommend Daniel Dennett's book, Freedom Evolves. He's one of the few philosphers to address the topic in what seems to me a logical way. He points out the logical problems with the unfortunately common view that freewill is in some way incompatible with a deterministic universe. His more popular book, Consciouness Explained, is also quite fascinating. I find Dennett's ideas much more useful for AI and robotics application than Searle's, who mostly seems to argue that AI is impossible. Of course, Dennett also argues, and I agree, that the whole idea of "qualia" is bogus.

Anyway, I found this paper interesting in that the authors looked at the problem in a slightly different way than most of the papers on consciouness I've read in the last year or so.

A terrible admission, posted 29 Jul 2005 at 13:28 UTC by dogsbody_d » (Master)

Okay, so I didn't actually RTFA, my comments might only apply to the lovely summary presented here.

There's a good reason that AI researchers have been ignoring the problem of consciousness recently, namely the failure of Big AI in the Big Old Bad Old Days of Giant Room Filling Computers. Folks in those days used to think that intelligence amounted to one big algorithm, a bit like he-who-must-not-be-named. The people trying to find intelligence seemed focussed on the things that they thought were intelligent, like solving maths problems and playing chess. The irony is that if they'd been interested in playing football things would be different. If only folks had listened more to Grey-Walter, we might have got bit further.

There's nothing wrong with trying to find the low-level intelligence upon which our own must have involved. It's got us a lot of good results. If we are all trying to invent cars, that doesn't make us wrong for not inventing aeroplanes, we'll get to it.

Finbally, I was lucky enough to see Daniel Dennett lecture on "Freedom Evolves." I do indeed urge you to read his work,

Dennet, posted 29 Jul 2005 at 15:45 UTC by Axiom » (Observer)

Dennet is one of the people I was implicitly reffering to actually - his view is that free will is just a "user illusion", which as the paper in the article convincingly argues is not a tenable position. It's not tenable because 1. What's the point? i.e, in an evolutionary sense? 2. If it's just a user illusion, then the concept of truth versus falsehood has no meaning anymore. Regarding #2 specifically, think of it this way: the only thing Dennet can say is that "well, I may be programmed by genes/nature to believe in determinism, but I'm programmed correctly!"... but then he must admit "I'm programmed to believe that I'm programmed correctly" and then "I'm programmed to believe that I'm programmed to believe that I'm programmed correctly..." ad infinitum. There's no escaping the determinist paradox.

freewill an illusion?, posted 29 Jul 2005 at 19:12 UTC by steve » (Master)

Dennett definitely doesn't believe freewill is an illusion - quite the opposite. In fact, I think the recurring myth that he believes that is even addressed in Freedom Evolves somewhere. Perhaps the idea that he doesn't believe in freewill comes from cases in which he has pointed out logical fallacies in some of the common "thought exercises" used to debate the issue. As far as determinism in general, that's a seperate issue altogether. In other words, Dennett believes freewill can and does exist regardless of whether the Universe is deterministic or non-deterministic. Thought, if such a thing as non-deterministic freewill could exist, I'm not clear on how it could differ from randomness and if choice is constrained by the output of a random process it doesn't sound very free!

Like consciousness, I think a lot of the misunderstanding over things like determinism and freewill comes from no one defining the terms before they start debating their existence.

Dennett, posted 30 Jul 2005 at 00:59 UTC by Axiom » (Observer)

I'll admit that I haven't read Freedom Evolves, but I am reasonably sure that Dennett is indeed a determinist. He does not believe in free will in the classical sense - i.e, human choice as a first cause. Rather, he simply redefines it to something akin to soft determinism.

determinism, posted 30 Jul 2005 at 14:51 UTC by steve » (Master)

He's a determinist. I wouldn't say he redefines freewill so much as he defines it. The first part of the book looks at and dispells common myths about what freewill is, such as something supernatural. The second part presents Dennett's analysis of what it is and how it evolved. It's been a while since I read it so I couldn't give a verbatim definition. (somebody correct me if I'm too far off here...)

His general position is that determinism isn't the same as inevitability. Determinism means outcomes have causes, inevitiblility means outcomes can't be avoided. Intelligent agents can make real, moral choices that affect the outcomes of events. So the outcomes are "evitable" as he likes to say. The number and complexity of the choices has grown more complex along with the creatures making them.

I think he uses the example of a pitcher who throws a baseball directly at your head. The deterministic universe says the ball is going to reach a certain point at a certain time but you can choose to move your head out of the way or not. Did your choice have a "cause"? Sure, you probably didn't want to get hit in the head by the ball. So is a choice with a cause not freewill - does choosing to do something for a reason mean you aren't free to choose among multiple outcomes?

Anyway, read the book if you get the chance - you may still disagree but it's a fascinating read.

See more of the latest robot news!

Recent blogs

15 Nov 2014 mwaibel (Master)
14 Nov 2014 Sergey Popov (Apprentice)
14 Nov 2014 wedesoft (Master)
5 Nov 2014 shimniok (Journeyer)
5 Aug 2014 svo (Master)
20 Jul 2014 Flanneltron (Journeyer)
3 Jul 2014 jmhenry (Journeyer)
3 Jul 2014 steve (Master)
2 Jul 2014 Petar.Kormushev (Master)
10 Jun 2014 robotvibes (Master)
X
Share this page