the origin of life holodecker2: Life began with a simple molecule that had the strange property that it liked to make copies of itself. About 3.5 billion years ago, there was no life on earth. There was water, carbon, hydrogen, oxygen, all the elements necessary for life but there was no life. Wherever energy was available, sunlight, lightning, cracks in the ocean floor where magma and water collided, different molecules would combine, and breakdown randomly. The creative power of evolution was already at work in this lifeless world. Molecules that had properties that made them more “fit” to “survive” the harsh environment of early earth would tend to “survive” better than those less “fit” to do so. Eventually, and completely randomly a molecule with the ability to make copies of itself was produced and that molecule was the beginning of life. Although that molecule was not alive, its descendents would eventually evolve into something that would be considered alive by our definition, and would continue to evolve into the vast diversity of life we enjoy on earth today. The astonishing creative power of evolution, the simple idea that if a molecule can make copies of itself, and sometimes make a slight mistake and create a new molecule that is not an exact duplicate, thereby starting a new branch of self replicating molecules, led to more and more complex self replicating molecules. The creative power of evolution through natural selection was already at work on these lifeless, self replicating molecules. Eventually this process led to the creation of molecules we would consider viruses and eventually, to the most basic primitive forms of unicellular life. The astonishing creative power of evolution through mutation and natural selection is the method that eventually produced the first unicelled life forms on our planet. Yan26: But you didnt answer an important question. At what point in the journey from individual molecules of hydrogen , Carbon etc to humans did life originate? Was it when the first strand of self replicating DNA/RNA was created? or was it when this self replicating DNA/RNA built a cell around itself? In order to answer the questions you would first have to answer what is the criteria for considering something alive. So What do you consider life? Geoff: That's a scientific conundrum for which there is no international consensus. The distinct point between life and unlife is very hard to pin down, much as the dividing point between one species and the next (during the evolutionary process). Traditionally the definition of life was that something had to consume resources, excrete waste, reproduce, and die. CoIin: Another interesting question in a similar vein is if we consider a virus to be a lifeform, then it's not clear why a computer virus shouldn't be. Yan26: My understanding is that viruses are considered neither living nor non-living. I also think that one of the problem with being unable to define life has to do with the quickly changing scientific beliefs. Eg : The status of the virus has see-sawed between living and non-living over the last 100 years . Also stuff like emotions or the ability to reproduce may have been considered as part of being alive but today AI are able to reproduce human emotions to some extent and computer viruses can "reproduce" . The question holds a lot of importance as we have a lot more empathy for living beings (plants or animals) rather than for non-living. So in the future if a machine can perform all the functions Geoff mentioned plus feel emotions , imagine etc would we be willing to consider it alive although it may not contain any cells and tissues? Could we look at a machine made out of various metals and give it the same right to exist as a living organism? One Bar: What's the distinction point between intelligence and non-intelligence? Can a thermostat be considered a form of intelligence? Geoff: Yan and One Bar both raise interesting philosophical questions. As I said, the exact definition of life has been a matter of great debate. Both from a strictly scientific point of view, and from a philosophical stance. But, abstract concepts and technological devices aside, we are dealing with the simplest of life, so the traditional definition is, perhaps, sufficient. One Bar: What is the simplest known form (species?) of life that satisfies the traditional definition? CoIin: @ One Bar "What's the distinction point between intelligence and non-intelligence?" See the "Turing Test" See also John Searle's "Chinese Room" argument. He claims to debunk the Turing Test and I'm with him. P.S. But the Turing Test is used to distinguish human intelligence from artificial intelligence so I guess my comment is a bit irrelevant (Edited by CoIin) One Bar: I've heard of the Turing Test - something to do with answers being indistinguishable from a human's, if I remember right? Yes, I was referring to a species' intelligence. However, would an artificial intelligence be considered a species? An intelligent computer program might reproduce (copy) itself, say. Yan26: @One Bar - I am curious. What relationship do you draw between intelligence and life? Regarding Intelligence @ Colinian I read about the Chinese Room argument. It is interesting. I agree with some of what is said. But for me it raises an important question? Does an AI have to behave like a human to be considered intelligent. For eg Humans have the ability to empathise with others and do something for the greater good even at a loss to oneself. This by and large is an emotion and as emotions go it is not rational/logical. So while empathy may be a trait we would want in an AI ( infact empathy for humans may go hand in hand with the 1st law of robotics) can the presence of an irrartional/illogical response be considered a requirement for intelligence. In the case of language it is considered something similar to art, something tied to emotions(Complex language is required to convey complex emotions). It is therefore illogical/irrational. So how would you make an AI understand something illogical/irrational ? And how could you consider its ability to react irrationally/illogically as an act of intelligence? CoIin: Well, Searles point is that all computers do, and all they ever will do, is shuffle around syntactic symbols with no understanding of semantics. One Bar: @Yan26: I think I might consider a 'clever' AI to be alive. For example, a computer program that could indulge completely in general conversations and discussions, emotionally expressing itself just as an intelligent human might. If the program could manufacture copies of itself then I would certainly regard it as a life form (it would be satisfying the traditional definition mentioned above). Geoff: One bar - For the purposes of defining the moment of abiogenesis; the first discrete collection of molecules (since we are probably talking about quite simple strings of amino acids) that met the traditional criteria. And there are already computer programs that can reproduce. They could be said to consume the computer's resources. Because it could be deleted, it can be said that it can die. However, they do not excrete. Computer viruses are often mentioned when discussing those things on the boundary between life and unlife. //Edit - for typo. That'll teach me to post at 05:30 (Edited by Geoff) One Bar: Computers emit heat and light that the programs can't see - just like how we can't see heavenly souls. One Bar: ... Is it theoretically possible for a computer virus to be constructed such that it excretes? Excreting data I presume - I'm not sure how they work. Yan26: @colinian : This is not related to the topic but just a correction. Searle doesnt disprove the Turing test. However some scientists claim that a machine that passes the Turing test simulates a mind. Searle claims that such a machine actually only simulates the brain and that the brain and the mind are different. The brain is the anatomical thing inside your skull which gets squished if an elephant steps on it. The mind is the brain plus"actual physical-chemical properties of actual human brains"(Quote taken from Wikepedia). This is what Searle is trying to prove that the AI which passes the Turing test simply simulates the brain. As it does not "understand" the language it cannot be said to simulate the mind. However I believe that since the way language is used is not rational or logical it is impossible to create a logical algorithm for it. For an AI to be able to pass a Turing test and understand the language it would not be able to use a simple symbol manipulation algorithm and would infact need a different form of understanding one that would help it simulate a mind.(Speed and complexity: appeals to intuition section in Wikepedia) CoIin: @ Yan - Yeah, thanks, I don't think I explained myself very well above. My description of the Turing test isn't very accurate either (LOL . It was early ) Searle strongly rejects the Cartesian dualism which still permeates our society (even among those who have never heard of Descartes). His concept of mind is not that it belongs to a different ontological realm from matter, but that it is both caused by, and a feature of the brain, paradoxical though that may sound at first. However, he DOES claim to refute the Turing test (whether he succeeds or not is another question) on the grounds that it fails to distinguish true understanding from the mere appearance of understanding. If the guy in the Chinese room got good enough, he could conceivably pass the Turing Test and yet have absolutely no understanding of the symbols he was manipulating (just like a computer). Here are some notes from a lecture series by the man himself.:- Lecture 3 Strong Artificial Intelligence Many people who work in cognitive science and in the philosophy of mind think that the most exciting idea of the past generation, indeed of the past two thousand years, is that the mind is a computer program. Specifically, the idea is that the mind is to the brain as the computer program is to the computer hardware. This view I have baptized "Strong AI," but it is sometimes called "Computer Functionalism." In this lecture I explain the appeal of this view, but I also subject it to a decisive refutation, the "Chinese Room Argument." Required Reading Turing, Alan, "Computing Machinery and Intelligence" (Mind 1950, pp 433-60, reprinted in Anderson, ed) Searle, John, Minds. Brains. and Science, Chapter 2 I. The Theoretical Basis of Strong AI In order to explain the appeal of Strong AI, I have to introduce five somewhat technical notions. 1. Turing machines a. The idea of a Turing machine is an abstract, mathematical notion, but for practical purposes, ordinary computers, of the kind you buy in a store, are Turing machines. b. The remarkable feature of a Turing machine is that it performs only four operations: Print "0;" erase" I;" print" I ," erase "0;" move one square left; move one square right. Modern machines perform these operations at the rate of millions per second. 2. Algorithm a. An algorithm is a systematic procedure for solving a problem in a finite number of steps. Computer programs are algorithms. 3. Church's Thesis a. Church's Thesis states that any algorithm can be implemented on a Turing machine. For every computable function, there is a Turing machine that can compute that function. 4. Turing's Theorem a. Turing's theorem states that there is a Universal Turing machine which can simulate the behavior of any other Turing machine. 5. The Turing Test a. The Turing test states that if an expert cannot distinguish the behavior of a machine from that of a human, then the machine has the same cognitive abilities as the human. II. Strong AI If we put all these together, we get the idea that the brain is a Universal Turing machine and that human cognitive abilities are computer programs. We test this with the Turing test and come to the conclusion that artificial intelligence is, in principle, capable of creating minds. III. Refutation of Strong AI Strong AI and, with it, the Turing test, are subject to a decisive refutation, the Chinese Room Argument. 1. The Chinese Room Argument a. The Chinese Room Argument claims that a monolingual English speaker locked in a room with a set of computer rules for answering questions in Chinese, would in principle be able to pass the Turing Test, but would not thereby understand a word of Chinese. If the man doesn't understand Chinese, neither does any digital computer. (Edited by CoIin) CoIin: Here's more:- (hope this isn't hogging the thread) Lecture 5 Can a Machine Think? The Chinese Room Argument was originally intended as a specific refutation of a very specific thesis in the philosophy of mind, Strong AI. However, it raises a very large number of other issues, and in this lecture I try to go through these in a systematic fashion. Could a machine think? Could an artifact think? Could we build an artificial brain as we have built artificial hearts? What is the significance of the Chinese Room argument? Does it really show that "computers can't think?" What exactly of a general philosophical nature is established by the Chinese Room Argument? I begin with the possibility that "connectionist" (parallel distributed processing, neuronal net) architectures might escape the Chinese Room. I. How do Connectionist machines differ from von Neumann machines? A. They operate by interrelated parallel processing. 1. As shown on the diagram. B. Does this evade the Chinese Room objection? 1. No. By Church's thesis, any computation they can perform can be performed on a Turing machine. Connectionism is faster, but has no additional computational powers. 2. If we are talking about the physics of specific architectures then we are doing speculative neurobiology, not computation. Output Hidden Layer Input II. Could a machine think? A. If by machine we mean a physical system capable of performing functions, then the brain is such a machine, so, of course, machines can think. But that is not what is meant. B. Could an artifact think? If you could duplicate the causal powers of the brain, then you would produce a thinking machine. It is, in principle, no more impossible to have an artificial brain than to have an artificial heart. C. But then, why couldn't a computer think? Why not an artificial brain made of silicon chips? Answer: Of course, the Chinese Room does not show that Something couldn't both be a computer and be thinking. Indeed, that is what our brains are. We know that our brains can compute, and we can think, it follows that we are thinking computers. D. But then, what does the Chinese Room show? Does it show that "computers cannot think?" No, what it shows is that implementing the program by itself is not sufficient for thinking. It might both implement and think, but the program is not constitutive of thinking. III. What would a machine have to have in order to think? A. To the three premises of the Chinese Room, let's add a fourth: Brains do it, brains cause mental phenomena. 1. From this it follows that any other system would have to duplicate the causal powers of the brain. B. Yes, but what are those causal powers? 1. We don't know specifically. We believe that it involves neurons and synapses, but we do not know the details. But we do know two facts: a. Just implementing a program would not be sufficient by the Chinese Room argument. b. Any other system that could think would have to have causal powers equal to the brain, and a program by itself would not be sufficient to do that. IV. Silicon brains? Imagine your brain slowly replaced by silicon chips. There are at least three possibilities: 1. Behavior and mental states remain intact 2. Behavior remains intact, but mental states shrink to zero 3. Mental states remain intact, but behavior shrinks to zero All three possibilities illustrate the bankruptcy of the Turing test. V. Biological Naturalism as a theory of the mind A. Descartes' questions can be answered by taking the mind seriously as a biological phenomenon. Mental states are caused by and realized in the brain, in the same sense in which digestion is caused by and realized in the stomach and the digestive tract. VI. The distinction between observer independent and observer relative features of the world. A more important distinction than the invalid distinctions between mind and matter or machines and nature. One Bar: Has anyone here read Shadows of the Mind by Roger Penrose? I tried a few years back but found it too heavy. It's still on my shelf, so I might give it another go. One Bar: It's his sequel to The Emperor's New Mind. Both are great books to 'casually' leave on the coffee table if you wish to impress a lady. Geoff: Actually, I find - "how to make love to a lady when you have a huge penis, but don't want to hurt her" is the ideal book to leave on coffee tables to impress ladies. | Science Chat Room 1 Person Chatting Similar Conversations |