Ovi -
we cover every issue
Philosophy Books  
Ovi Bookshop - Free Ebook
Stop human trafficking
Ovi Language
Ovi on Facebook
Stop violence against women
Tony Zuvela - Cartoons, Illustrations
International Red Cross and Red Crescent Movement
BBC News :   - 
iBite :   - 
A Reflection on Artificial Intelligence and Human Consciousness
by Dr. Emanuel Paparella
2009-02-27 10:56:43
Print - Comment - Send to a Friend - More from this Author
DeliciousRedditFacebookDigg! StumbleUpon

To continue with book presentations, some twenty years ago a book came out which became an immediate best seller and irritated many experts on artificial intelligence, at least those who advocated artificial intelligence as the salvation of a flawed human consciousness and personality leading to a trans-human ideal world devoid of flaws and imperfections; a wholly rationalistic and determined world.

book2_400 The book contradicted their certitudes, for it attacked the very foundations of strong artificial intelligence. It is still being discussed and it is still worth reading. It is titled The Emperor’s New Mind. Its author is a top scientist and physicist Roger Penrose who in the sixties worked on the astrophysics of black holes with Stephen Hawking. Nobody can accuse him of having some kind of secret spiritual religious agenda pleading against the materialism of modern science, for the man is a non-believer and an avowed materialist.

It is a crucial book for anyone interested in the history of thinking about AI (artificial intelligence) and consciousness and whether or not this consciousness is determined or is the foundation of our freedom and our very humanity. Part survey of modern physics, part exploration of the philosophy of mind, the book is not for casual readers--though it's not overly technical, it rarely pauses to let the reader catch a breath. The overview of relativity and quantum theory, written by a master, is priceless. 

Penrose claims that there is an intimate, perhaps unknowable relation between quantum effects and our thinking, and ultimately derives his anti-AI stance from his proposition that some, if not all, of our thinking is non-algorithmic. Of course, these days we believe that there are other avenues to AI than traditional algorithmic programming; while he has been accused of setting up straw robots to knock down, this accusation is unfair. Little was then known about the power of neural networks and behavior-based robotics to simulate (and, some would say, produce) intelligent problem-solving behavior. The Emperor's New Mind offers powerful arguments useful to believer and nonbeliever alike. 

Penrose is the kind of physicist who believes that some aspects of the human mind will never be duplicated by artificial intelligence and supports his view with material drawn from quantum mechanics, brain structure and other theories. His book is not "creation science," it doesn't address evolution, or the existence of God, or the existence of the human soul. What it is rather, is a solid study of cognition, theories of artificial intelligence, and the enduring problem of the nature of human consciousness.

What Penrose suggests in such a book (a theory that he expands on in his subsequent Shadows of the Mind), is that science, and specifically physics, is inadequate now, and more importantly will always be inadequate, to describe the nature of human intelligence, cognition, and consciousness--a thesis similar to the showing of Godel's 1931 Theorem that certain fundamental axioms of mathematics were incapable of proof within any mathematical system.

In other words, Penrose suggests that there are elemental restrictions within science itself limiting our understanding of our own mental processes, which concomitantly limit the possibilities for development of artificial intelligence. And that obviously doesn't sit well with those for whom naturalistic science is itself a kind of "religion."  It's a challenging read, not for philosophical featherweights, but it will richly reward those with enough curiosity and patience to make it through.

When I first read the book it occurred to me that Giambattista Vico, all the way back in the 18th century (1725) in his New Science had already intuited Penrose’s discovery, that the world of nature created by God will never be exhaustively known by a creature such as man. The more realistic, albeit more humble, attitude is that of focusing on what man himself has made (the institutions, history, language, artifacts) and which he can in principle know exhaustively.

Print - Comment - Send to a Friend - More from this Author

Get it off your chest
 (comments policy)

Alexander Mikhaylov2009-02-28 05:14:17
But we have already developed artificial intelligence!Just look around you!

Alexander Mikhaylov2009-02-28 05:18:05
'In other words, Penrose suggests that there are elemental restrictions within science itself limiting our understanding of our own mental processes, which concomitantly limit the possibilities for development of artificial intelligence.'

I think the New Testament already pointed out this pecularity of human outlook (don't remember how it goes precicely but anyway...'When a blind person is trying to lead other blind persons, all of them will end up lying in the ditch'...)

Emanuel Paparella2009-02-28 09:48:33
Indeed we have artificial intelligence in abundance and computers can beat us at chess most of the times making their cognitive rational operations seem superior to that of flesh and blood human beings, but the puzzle Penrose is exploring is whether rationality by itself can explain consciousness exhaustively and indeed can replace it. Apparently it does not and in his opinion it will never be able to do so to the consternation of those who think that they will be able to leave the essence of their lives in a computer and be able to go on living: the famous brain in a vat conundrum. As Einstein put it, one cannot get out of the box of rationality by using rationality, and if that is so, Penrose's book is still relevant.

Eero Nevalainen2009-03-01 16:18:00
This is a very interesting subject, something that was part of my studies back in the day. Penrose's book is fascinating for a couple of reasons -- for example his exposition of the Cantor diagonal argument and basic computability theory is the best condensation/popularisation on the subject I have ever read -- it's better (in the pedagogical sense) than a perhaps more rigorous and formal textbook treatment.

Penrose's quantum-mechanical argument regarding cognition is, as far as I know, considered as crackpottery by neuroscientists and cognitive scientists, but his exposition of the relationships between algorithmics, the Church-Turing thesis (idea that any function that can be physically computed can be computed by the Turing Machine), and the relationship of that to materialism is worth the book alone, the "conclusion" nonwithstanding.

There is one seeming misconception in this article that I feel pressed to point out -- neural networks and emergent robotics etc are not in any way computationally "stronger" than classical algorithmics. We are still dealing with classical Turing-complete stuff anyway, so although for example a neural network is a very good statistical black-box model fitting tool, it is not such a "breakthrough" as to make a difference to the underlying things Penrose discusses. It still runs on a very classical computer after all.

Of course, NNs being a model lifted from biology and showing that mimicing the workings of biological neurons we can achieve interesting "soft pattern matching" effects that approach what we see in the brain proves that there is some relationship to cognition and computability theory. Now, the big question of strong AI is whether cognition really IS a materialistic computational process, and whether, by Church-Turing thesis (if it is true), as all those processes, cognition included, can be "run on the computer".

Now, here we of course run into Gödel and his incompleteness theorems. Somehow, in a given formal symbolic system we always end with unprovable sentences that we can still "see" as being true (and not only that, there is a way to construct an infinite number of those sentences). What is it in the human mind that lets us "see" the correctness of Gödel's proof, but that makes that sort of stuff intractable from within the system itself?

Conversely, if the human mind is a materialistic computational device, is it subject itself to Gödel sentences that are not provable within cognition? Can we make the mind "crash"? :)

Quantum mechanics alone is not going to let us out of the trap; quantum computers are actually not THAT magical as far as computability theory goes. In particular, the common misconception that so-called NP-complete problems could be solved in polynomial time (a kind of holy grail in complexity theory) by quantum computers is believed to be false. Knowing the exact additional strengths quantum computation brings to the table would probably let us have a better idea of the merits of Penrose's ideas regarding the AI issue.

Emanuel Paparella2009-03-01 19:33:24
Mr. Nevalainen, thanks for the complementary information and the correction. I am no physicist or neuro scientist myself but I do have a hunch as a mere layman on the subject that Penrose remains topical for no other reason that it adds a missing dimension to the current debate on the desirablity of a "trans-human" world, which in some way is the world that Nietzsche in a positive and speculative mode, and the Nazis in a negative more practival mode were also exploring. What does it mean to be human remains as topical today as when Socrates and the ancients put that question. It is the ancient question of self-knowledge and it is perhaps more important than the answer to it, is the ability to still pose the question.

© Copyright CHAMELEON PROJECT Tmi 2005-2008  -  Sitemap  -  Add to favourites  -  Link to Ovi
Privacy Policy  -  Contact  -  RSS Feeds  -  Search  -  Submissions  -  Subscribe  -  About Ovi