Re: This is not the roadmap
Le 31-juil.-06, à 23:32, John M a écrit : 1Z: I liked your examples, would have liked better if you do not base the entire list on matter to exist. It may not. I have a notion - cannot put my finger on an adequate formulation of it into words - that mathematics cannot be computed by mathamatics - I think Goedel would have some objections to that. Somebody tell me if this is a wrong idea. I will not fight it. (Not my table). It is ok. Godel would have approved: the whole of formal mathematics cannot be computed by any formal mathematics. It is a little vague but this convey the main godelian point. Concerning some of tyhe conversation between Brent, 1Z and Stathis, I would say that I don't see the relationship between computations and random string. Computations, or their description can be shown to be necessarily redundant, (and deep in Bennett' sense). For Tom and Georges: Take the Fi corresponding to 0-argument (fortran) programs. Any such program stops or does not stop. Consider the function which associates to n either 1 or 0 according to the fact that the nth program stop or does not stop. you get a deep complex and subtly redundant sequence of 0 and 1. If you decide to compress it maximally you will get Chaitin OMEGA number, which gives the probability that a Fi will stop or not, (but this cannot be done algorithmically). There is no reason to related consciousness to those random compression of computation. Look at nature from genome to the number PI: you will always see many redundancies. They are absent in the Putnam Chalmers rock. I don't think it makes sense to attribute computations in there (but then I don't care given that UDA makes us having to (re)define physics by winning (in some relative probabilistic sense) sheaf of relative computations existing in platonia. Bruno http://iridia.ulb.ac.be/~marchal/ --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
Brent Meeker wrote: And evolution constructs brains to be essentially deterministic for the same reason. So is it your theory that any deterministic sequence of states constitutes computation and the reason a rock doesn't instantiate computation is that, at the microscopic level its state changes are dominated by quantum randomness? My theory is that to implement an algorithm something needs to have the counteractuals that are part of the algorithm. A machine needs to have distinct states (unlike a rock) and to have them counterfactually/causally linked (unlike a cloud of gas), This thread started with a discussion of what computation could be counted as intelligent - or Stathis prefers conscious. Does your distinction entail that intelligence (or consciousness) is deterministic? I never said intelligence was computational in the first place ! Brent Meeker --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
Brent Meeker wrote: 1Z wrote: Brent Meeker wrote: 1Z wrote: Brent Meeker wrote: I'm considering rejecting the idea that a computation can be distinguished from noise by some internal characteristic of the computation. I don't think you can make the idea of information hidden in noise well defined. By Shannon's measure noise is information. You can easily distinguish computation from noise using counterfactuals Can you make that more concrete - an example perhaps? Counterfactuals come from the undertlying physics of the computation. Cups of coffee don't have any woth speaking about-- you can't force them into the same state twice. Sorry, but I still don't understand the counterfactual aspect. You have to be able to say what *would* have happened if the computation had gone down the other fork of an if-then. That requires some causal stability. http://en.wikipedia.org/wiki/Causality#Counterfactual_theories_of_causation Whether they are part of the internal characteristitcs of a computation depends, question-beggingly , ont what you mean by computation. I think I agree with that. I'm trying to come up with a non-question begging definition of computation and I think the idea that a rock implements all computations implies that computation can't be defined in terms of some chracteristic of its sequence of internal states. I think the idea that a rock implements all computations is the wrong place to start. If you think a computation is nothing but a string of 1's and 0's, counterfactuals will be very difficulty to find. So you're agreeing with me that it's impossible to distinguish noise and computation based their sequence of internal states (e.g. 1's and 0's)? No: I'm saying you do need to find counterfactuals, and since they aren't in bit-strings (movies or recordings), bit-strings aren't computations. Therefore, rocks don't compute merely by going through a succession of internal states. Brent Meeker --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
Stathis Papaioannou wrote: John M writes: Peter Jones writes: Hmm. Including limitations in time? Yes, if an infinite number of finite computations are run simultaneously on a system with a finite number of physical states. Stathis Papaioannou - So if I have a system with finite number of physical states, it will take a matching finite number of (base)-computations leaving an infinite number untreated. Out of them I can take a deduction for muiltiplying the finite number of physical states by the finite number of the base-states to get to the total number of computability on that system in parallel - still a finite number. I still have an infinite number of unbtreated cases left. Damn that infinite! Cantor's curse. John M Suppose there is a very simple physical system that goes through two states, on and off. You wish to map these states onto a binary sequence which at first glance seems too long: 10110100... You write down the following: on the first run, on-1 and off-0; on the second run, on-1 and off-1; on the third run, on-0 and off-1; and so on, for as long as you like. It is not common practice to change the code from run to run when designing a computer, but that is just a matter of convenience. If you specify exactly how the code changes the meaning is unambiguous, and in principle the two physical states can encode any number of binary states, or even more complex computations. A computation is not a series of states. A computation is an implementation of an algorithm, and algorithms include conditional statements which must be modelled by something with counterfactual behaviour -- by something which *could have* execute the other branch. The above probably seems silly to most people reading this, because the burden of the computation falls on the specification of the code, the physical processes being essentially irrelevant. Nevertheless, we may have the situation where the code specification is documented in a big book while the computer (such as it is) carries out the physical processes which, if we to refer to the book, performs perfectly legitimate computations. We could even design a driver for a monitor to display the computations, again using the book. Now, suppose the last copy of the book is destroyed. The computer would still do its business, but it may as well be a random number generator for all the good it does us without the code specification. But what if, by the book, the computer is actually carrying out *conscious* computations? Would it suddenly cease being conscious as the book is burned in a fire, or gradually lose consciousness as the book's pages are ripped out one by one? No amount or arbitrary mapping can transofrm a situation without counterfactuals into one with them --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
Peter: As I recall all I wrote (and the post marked it as was: So if I have a system with finite number of physical states, it will take a matching finite number of (base)-computations leaving an infinite number untreated. Out of them I can take a deduction for muiltiplying the finite number of physical states by the finite number of the base-states to get to the total number of computability on that system in parallel - still a finite number. I still have an infinite number of unbtreated cases left. Damn that infinite! Cantor's curse. John M * I wanted to point to the 'flipside of it' which was not addressed in your reply: mixing finite and infinite. Those marks drive me crazy. too. John - Original Message - From: 1Z [EMAIL PROTECTED] To: Everything List everything-list@googlegroups.com Sent: Tuesday, August 01, 2006 9:17 AM Subject: Re: Bruno's argument Stathis Papaioannou wrote: John M writes: Peter Jones writes: Hmm. Including limitations in time? Yes, if an infinite number of finite computations are run simultaneously on a system with a finite number of physical states. Stathis Papaioannou - So if I have a system with finite number of physical states, it will take a matching finite number of (base)-computations leaving an infinite number untreated. Out of them I can take a deduction for muiltiplying the finite number of physical states by the finite number of the base-states to get to the total number of computability on that system in parallel - still a finite number. I still have an infinite number of unbtreated cases left. Damn that infinite! Cantor's curse. John M Suppose there is a very simple physical system that goes through two states, on and off. You wish to map these states onto a binary sequence which at first glance seems too long: 10110100... You write down the following: on the first run, on-1 and off-0; on the second run, on-1 and off-1; on the third run, on-0 and off-1; and so on, for as long as you like. It is not common practice to change the code from run to run when designing a computer, but that is just a matter of convenience. If you specify exactly how the code changes the meaning is unambiguous, and in principle the two physical states can encode any number of binary states, or even more complex computations. A computation is not a series of states. A computation is an implementation of an algorithm, and algorithms include conditional statements which must be modelled by something with counterfactual behaviour -- by something which *could have* execute the other branch. The above probably seems silly to most people reading this, because the burden of the computation falls on the specification of the code, the physical processes being essentially irrelevant. Nevertheless, we may have the situation where the code specification is documented in a big book while the computer (such as it is) carries out the physical processes which, if we to refer to the book, performs perfectly legitimate computations. We could even design a driver for a monitor to display the computations, again using the book. Now, suppose the last copy of the book is destroyed. The computer would still do its business, but it may as well be a random number generator for all the good it does us without the code specification. But what if, by the book, the computer is actually carrying out *conscious* computations? Would it suddenly cease being conscious as the book is burned in a fire, or gradually lose consciousness as the book's pages are ripped out one by one? No amount or arbitrary mapping can transofrm a situation without counterfactuals into one with them --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: This is not the roadmap
Thanks, Bruno, for your 1st par below. My idea was based on (my) common sense using that tiny little I read (and heard) about Gödel. To the 2nd par: I disagree with any 'random' in the 'existence' (nature etc.) - except for the mathematical use like: take ANY number. However: a 'random' string (unfettered by 'order') IMO cannot provide reasonable computational results as seeable e.g. from a 'function' with unidentified and unlimited variables. It may lead to anything at all. (says the layman - after a friend who teaches math at a NY univ.). Your 3rd par, however, (For Tom and Georges:) sounds to me like musical noise and I prefer Beethoven. I needed some 20-30 years of intensive study to get it right. Thanks anyway. John - Original Message - From: Bruno Marchal [EMAIL PROTECTED] To: everything-list@googlegroups.com Sent: Tuesday, August 01, 2006 6:38 AM Subject: Re: This is not the roadmap Le 31-juil.-06, à 23:32, John M a écrit : 1Z: I liked your examples, would have liked better if you do not base the entire list on matter to exist. It may not. I have a notion - cannot put my finger on an adequate formulation of it into words - that mathematics cannot be computed by mathamatics - I think Goedel would have some objections to that. Somebody tell me if this is a wrong idea. I will not fight it. (Not my table). It is ok. Godel would have approved: the whole of formal mathematics cannot be computed by any formal mathematics. It is a little vague but this convey the main godelian point. Concerning some of tyhe conversation between Brent, 1Z and Stathis, I would say that I don't see the relationship between computations and random string. Computations, or their description can be shown to be necessarily redundant, (and deep in Bennett' sense). For Tom and Georges: Take the Fi corresponding to 0-argument (fortran) programs. Any such program stops or does not stop. Consider the function which associates to n either 1 or 0 according to the fact that the nth program stop or does not stop. you get a deep complex and subtly redundant sequence of 0 and 1. If you decide to compress it maximally you will get Chaitin OMEGA number, which gives the probability that a Fi will stop or not, (but this cannot be done algorithmically). There is no reason to related consciousness to those random compression of computation. Look at nature from genome to the number PI: you will always see many redundancies. They are absent in the Putnam Chalmers rock. I don't think it makes sense to attribute computations in there (but then I don't care given that UDA makes us having to (re)define physics by winning (in some relative probabilistic sense) sheaf of relative computations existing in platonia. Bruno http://iridia.ulb.ac.be/~marchal/ -- No virus found in this incoming message. Checked by AVG Free Edition. Version: 7.1.394 / Virus Database: 268.10.5/404 - Release Date: 7/31/2006 --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
If the characters bother you
John, you can download a freelittle program at http://www.papercut.biz/emailStripper.htm that strips all those things from any file you feed it. If the "" characters bother you, give it a try. Norman - Original Message - From: "John M" [EMAIL PROTECTED] To: everything-list@googlegroups.com Sent: Tuesday, August 01, 2006 1:43 PM Subject: Re: Bruno's argument . . . Those marks drive me crazy. too. John --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
Stathis Papaioannou wrote: Brent Meeker writes: Would you allow that one machine or computation may be emulated by another following some sort of mapping rule, and that consciousness may be preserved in this process? This would seem to be an assumption at the basis of functionalism and computationalism. But what if the mapping rule were the equivalent of what in cryptography is called a one-time pad, determined by some stochastic process such as radioactive decay? The states of the emulated machine would then seem to vary randomly, but if you had access to the mapping rule you would be able to read it (and perhaps interact with it) just as if it followed some simpler code, like shifting each letter of the alphabet by one. Are you prepared to argue that the emulated machine is only conscious if an external observer has the relevant mapping rule at hand and/or is actually reading it or interacting with it using this information? Stathis Papaioannou Yes, that's roughly my idea. Of course you can't insist that a computation interact continuously to count as computation, only that it does occasionally or potentially. In your example I would say that you can only know that there is computation, as distinct from noise, going on if the computer, via the emulation code, can still interact with its environment (i.e. you). I don't believe the simplicity or complexity of the internal operations is relevant. For example, if you could see the movements of electrons in my computer, you couldn't tell whether it was displaying this email or just doing something random - but if you look at the dispaly screen you can. On the other hand, to the alien from alpha centauri, the screen might also look random. Brent Meeker That's fine in the case of an email, but consider a computer which is conscious and spends its time musing or dreaming. Would you say that this computer's consciousness is contingent on the existence of external observers who might be able to figure out what it's up to? Stathis Papaioannou Consider a computer which is doing something (whether it is dreaming or musing or just running is the point in question). If there is no interaction between what it's running and the rest of the world I'd say it's not conscious. It doesn't necessarily need an external observer though. To invoke an external observer would require that we already knew how to distinguish an observer from a non-observer. This just pushes the problem away a step. One could as well claim that the walls of the room which are struck by the photons from the screen constitute an observer - under a suitable mapping of wall states. The computer could, like a Mars rover, act directly on the rest of the world. Brent Meeker --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
Stathis Papaioannou wrote: John M writes: Peter Jones writes: Hmm. Including limitations in time? Yes, if an infinite number of finite computations are run simultaneously on a system with a finite number of physical states. Stathis Papaioannou - So if I have a system with finite number of physical states, it will take a matching finite number of (base)-computations leaving an infinite number untreated. Out of them I can take a deduction for muiltiplying the finite number of physical states by the finite number of the base-states to get to the total number of computability on that system in parallel - still a finite number. I still have an infinite number of unbtreated cases left. Damn that infinite! Cantor's curse. John M Suppose there is a very simple physical system that goes through two states, on and off. You wish to map these states onto a binary sequence which at first glance seems too long: 10110100... You write down the following: on the first run, on-1 and off-0; on the second run, on-1 and off-1; That one's not gonna work :-) on the third run, on-0 and off-1; and so on, for as long as you like. It is not common practice to change the code from run to run when designing a computer, but that is just a matter of convenience. If you specify exactly how the code changes the meaning is unambiguous, and in principle the two physical states can encode any number of binary states, or even more complex computations. The above probably seems silly to most people reading this, because the burden of the computation falls on the specification of the code, the physical processes being essentially irrelevant. Nevertheless, we may have the situation where the code specification is documented in a big book while the computer (such as it is) carries out the physical processes which, if we to refer to the book, performs perfectly legitimate computations. We could even design a driver for a monitor to display the computations, again using the book. Now, suppose the last copy of the book is destroyed. The computer would still do its business, but it may as well be a random number generator for all the good it does us without the code specification. But what if, by the book, the computer is actually carrying out *conscious* computations? Would it suddenly cease being conscious as the book is burned in a fire, or gradually lose consciousness as the book's pages are ripped out one by one? The implication is that the computer was conscious before the book was burned - but I would ask, What was it's interaction with the world? If the answer is that the person with the book interpreted the output and was informed by that or acted on that, then I'd say the book+computer was conscious - but not the computer alone. Brent Meeker --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
1Z wrote: Stathis Papaioannou wrote: John M writes: Peter Jones writes: Hmm. Including limitations in time? Yes, if an infinite number of finite computations are run simultaneously on a system with a finite number of physical states. Stathis Papaioannou - So if I have a system with finite number of physical states, it will take a matching finite number of (base)-computations leaving an infinite number untreated. Out of them I can take a deduction for muiltiplying the finite number of physical states by the finite number of the base-states to get to the total number of computability on that system in parallel - still a finite number. I still have an infinite number of unbtreated cases left. Damn that infinite! Cantor's curse. John M Suppose there is a very simple physical system that goes through two states, on and off. You wish to map these states onto a binary sequence which at first glance seems too long: 10110100... You write down the following: on the first run, on-1 and off-0; on the second run, on-1 and off-1; on the third run, on-0 and off-1; and so on, for as long as you like. It is not common practice to change the code from run to run when designing a computer, but that is just a matter of convenience. If you specify exactly how the code changes the meaning is unambiguous, and in principle the two physical states can encode any number of binary states, or even more complex computations. A computation is not a series of states. A computation is an implementation of an algorithm, and algorithms include conditional statements which must be modelled by something with counterfactual behaviour -- by something which *could have* execute the other branch. I think this something is an interaction with something outside the computer, i.e. a different input or a real-time sensor input. I could also be a random variable generated internally - but I'm not clear on whether that satisfies lz's idea - it doesn't satisfy mine. Brent Meeker --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
RE: Bruno's argument
Peter Jones writes: A computation is not a series of states. A computation is an implementation of an algorithm, and algorithms include conditional statements which must be modelled by something with counterfactual behaviour -- by something which *could have* execute the other branch. Whatever else a computation is, it is a series of states. My computer is going through a series of physical states, with the earlier states determining the later states. If the earlier states were different, then the later states would also be different, hence the computer handles counterfactuals. However, this is so with any physical system: it goes through a series of states, the earlier states determine the later states following the laws of physics, and had the earlier states been different, so would the later states. Now, I suppose you would say that the states in a rock are random, while those in a computer are not. But what is to stop someone from designing a computer so that there is no pattern to its internal states unless you have the key? Suppose you find two inputless electronic devices, powered up, with complex and at first glance random currents circulating in their internal components. One of these devices is in fact implementing a computation, deliberately scrambled to keep it secret from prying eyes, while the other is just a decoy with random electrical activity. Without access to the key, would you be able to tell which is which? Another question: I can see why a computer should be able to handle counterfactuals if it is to be of practical use, but what is wrong with saying that a recording implements a computation, whether that is adding two numbers or having a conscious experience? Stathis Papaioannou _ Be one of the first to try Windows Live Mail. http://ideas.live.com/programpage.aspx?versionId=5d21c51a-b161-4314-9b0e-4911fb2b2e6d --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---
Re: Bruno's argument
On Wed, Aug 02, 2006 at 10:05:37AM +1000, Stathis Papaioannou wrote: Another question: I can see why a computer should be able to handle counterfactuals if it is to be of practical use, but what is wrong with saying that a recording implements a computation, whether that is adding two numbers or having a conscious experience? Stathis Papaioannou In the Multiverse, there is a huge difference between a recording and the actual computation. Only in one single universe (or history) of the ensemble do the two coincide. The recording is a computation issue is only a problem for single universe theory IMHO. Cheers -- *PS: A number of people ask me about the attachment to my email, which is of type application/pgp-signature. Don't worry, it is not a virus. It is an electronic signature, that may be used to verify this email came from me if you have PGP or GPG installed. Otherwise, you may safely ignore this attachment. A/Prof Russell Standish Phone 8308 3119 (mobile) Mathematics0425 253119 () UNSW SYDNEY 2052 [EMAIL PROTECTED] Australiahttp://parallel.hpc.unsw.edu.au/rks International prefix +612, Interstate prefix 02 --~--~-~--~~~---~--~~ You received this message because you are subscribed to the Google Groups Everything List group. To post to this group, send email to everything-list@googlegroups.com To unsubscribe from this group, send email to [EMAIL PROTECTED] For more options, visit this group at http://groups.google.com/group/everything-list -~--~~~~--~~--~--~---