On 11 Nov 2009, at 19:52, Brent Meeker wrote:

> > Bruno Marchal wrote: >> >> On 10 Nov 2009, at 19:29, Brent Meeker wrote: >> >> >>>> >>> But this seems like creating a problem where none existed. The >>> factorial is a certain function, the brain performs a certain >>> function. >>> Now you say we will formalize the concept of function in order to >>> study >>> what the brain does and perhaps understand what is consciousness. >>> But >>> there is nothing that requires that you start over with all possible >>> computations. That is like explaining the factorial function by >>> considering all possible computations that produce it (like the >>> above). >>> It's not wrong, but it doesn't follow from saying that the >>> factorial is >>> a function. That's why I say I take it as an ansatz - "Let's >>> consider >>> all possible computations and see if we can pick out physics and the >>> brain and consciousness from them." >> >> >> Hmm... The seventh step comes after six other steps. I think you >> confuse UDA and Tegmark or Egan speculation on the mathematical >> nature >> of physics. But when we assume comp, > > But it seems there is a shift the meaning of "assume comp" here. We > start with comp = "Consciousness is something a brain does. A brain > does a lot of things (metabolizes, takes up space,...) but the thing > it > does that produces consciousness is a kind of computation, i.e. > information processing. " Almost all scientists and philosophers > think > that is good hypothesis and one they would assume. But then it seems > you use comp2 = "We - our stream of consciousness - IS a computation > that exists in the Platonic sense." This seems slightly different. When some materialists, like some neuroscientists, assume the brain does computations at some high level, like with neuronal processing. I am not sure all are aware of the digitality assumption needed for making sense to the word computation. But if we assume the brain obeys to the physical laws, then we assume comp already, given that the physical laws are, as far as we know, Turing emulable. Comp is NOT the hypothesis that "consciousness is something brain does". This is not even a precise definition of comp, given that strictly speaking we don't know what a brain is, nor what consciousness. But that's ok, because comp is very close to that, perhaps equivalent deending of what you mean by "consciousness is something brain does". Comp is two things: "yes doctor", which means: there is a level such that I survive (or feel nothing) when may brain is substitute for digital part made at that level". This does not use any neither a definition of brain, except it is something which can be described by some machine (but not necessarily the one in the skull), nor any idea of consciousness, above what we can ask to a doctor (will I make it Doc?). And then I believe you did understand the seventh step. In a "concrete universe with a UD running in it (even materially), well, if my comp state is S, my future subjective experience is given by a sum on all computations going through S. So if the laws of physics still applies, they have to be recover from the mathematics of computations, which is a branch of math. In physics the notion does not admit definition which is different from the one by mathematicians. And quantum computation, actually a mathematical notion, is already close to the idea of sum on infinities of computations. After the seven step, you can still invoke, to save both comp and materialist, that we live in a universe which is too little for a universal dovetailing to be executed integrally or on some large portion. That is why there is an eight step, which shows that the universal machine (that we all provably are, even if comp is false) cannot, once comp is true, make the difference, for a short time on which bear the probabilities, distinguish between any implementations made below its substitution level, notably between "physically implemented" or "arithmetically implemented". Step 8 shows that consciousness is not produced by the brain. Consciousness is "produced" by the arithmetical relation (or XYZ-ical relation with XXX being a first order specification of any universal system. In that case the move to a little universe can not work, and reality, or more aptly the realities (corresponding to each points of view) have to emerge from the XYZ-relations. I don't need any more "platonic reality" than I need to explain what is Church thesis, what is a universal machine or number, etc. No book of physics assume less. > >> the physical appearance cannot be describe by *any* computation a >> priori: > But the main evidence for the comp hypothesis is that physics is so > successfully described by computations. Not necessarily. If you take Old QM (SWE + Collapse), the collapse of the wave, and its result, IS NOT the output of a computation. If you take the "new" QM, which assumes comp, the physical appearance is already described by infinities of computations. The only problem is that by assuming comp and taking it seriously into account (meaning taking the first and third person distinction into account) the UDA- MGA argument shows that the SWE has to be derived from comp, that is from number theory (or the XYZ-theory). I take XYZ = Arithmetic, just because Arithmetic is already universal, and is already taught in high school. As I said, it would help if you could say where in SANE04 you have a problem. I will send to the list a new version of MGA quite close to the one already sent. (MGA is just sketched in SANE04). > > >> it *has* to be retrieve from all computation. Roughly speaking, if we >> are universal machine, > > But assuming "we are a universal machine" is assuming more than "our > brains do computations and that produces consciousness." No, it is less. I can prove to you that you are at least a universal machine. Comp is that you are Turing emulable, and this means that you are not MORE than a universal machine. Universal here is used in respect of computability, or emulability, not in the sense of provability or anything else. > >> we do belong to an infinity of computation, and matter, or anything >> below our substitution level, has to be described by "all >> computations". It is an open problem if that sum converge toward >> something we could describe by "one" computation. > If it is proven that it doesn't, would that refute comp2? Why? Not necessarily. Unless you mean that all physical result is given by that "one" computation, which is already not the case with "new QM". > Would we be > left with no explanation of the perceived unity of individual > consciousness? Consciousness is when (universal) machine bets (instinctively at first) on "transcending realities", like their parents, their local neighborhood (to begin with), up to things like "physical universe", "Gods" or the Number Realm. It is always hypothetical. It is not always consciously "hypothetical". > > >> It is the whole point of the reasoning. It is "theorem" in the comp >> theory that matter emerges from all computations. From this you can >> prove that comp implies the non-cloning of any piece of matter, like >> it proves the existence of a strong form of indeterminacy, etc. > > What's the non-cloning proof? UDA1--7 is enough. To predict the exact state of any object, you have to run the whole dovetailer. This is not possible, so the exact state of any piece of matter, although it may be simulated quickly with a big accuracy, would need the whole running of the UD to be emulated. > >> >> >>>>>> >>>>> Could you remind me what the phi_i are? The enumerated partial >>>>> functions? >>>> >>>> >>>> The enumerated so called "partial /computable/ functions". >>>> >>>> To get them, take your favorite universal system (fortran, lisp, c >>>> ++, >>>> java, whatever), write down the grammatically correct description >>>> of >>>> function (of one argument, say, that is, from N to N). Put those >>>> codes >>>> in lexicographical order, and you get the corresponding phi_i: >>>> phi_1, >>>> phi_2, ..., and their domain W_1, W_2, W_3, ... >>>> >>>> With Church thesis, all the computable functions (having the >>>> domain N) >>>> will belong to that list, but there will be no algorithm capable of >>>> telling in advance for any phi_i if it compute partial computable >>>> function or a computable functions. >>>> >>>> Given that this is a key point for everything which will follow, I >>>> have to be sure that most people understand exactly why this has to >>>> be so. >>> >>> Ok, I think I understand. It's probably not relevant here, but >>> physicist usually think of functions which can be computed >>> approximately >>> by a uniformly convergent algorithm as computable (e.g. compute >>> pi) but >>> I think in the above you mean computable in the Turing sense that >>> the >>> computation stops with the answer (e.g. compute pi to 100 decimal >>> places). Right? >> >> >> Right. There is a vocabulary problem about what is a function, and >> unfortunately english speaker and french speaker have different >> conventions, and sometimes I slip from one to other, and this does >> not >> help. Usually a function from N to N is supposed to be defined on all >> element of N. And thus a computable function will have an algorithm >> which does stop on all of its input. >> But the Kleene diagonalization shows that there is no computable list >> of all computable functions, so if a language is universal, it means >> that the computable functions can only belong to a list of something >> else. That something else are called partial computable function: >> they >> are allowed to be not necessarily define on some natural number. So >> to >> get ALL functions, in some computable way, we have to take something >> larger: all partial functions, and to get all execution of all >> algorithm, we will have to dovetail, > Thanks, I did understand, but sometimes I need reassurance that I've > grasped it. Tanks for telling. I think many does not grasp it, or does grasp it since a longer time. I will sum up, but people would help by telling if they have grasped Cantor diagonalization proof of the non enumerability of N^N. Then I can explain with more assurance Kleene very similar, yet completely different result and their consequence. People should reread perhaps the "coffee bar machine" thread, or the "combinator" thread, or learn some programming language to familiarize themselves with example of universal system. But all what I need does not depend on any particular system. I refer to the numbers because everyone know them, but once a universal system is very simple, it becomes rather long and tedious to show that is universal. There is an hard pedagogical dilemma here. > >> and from the first person point of view, there is an emerging >> continuum of computations, and it plays the role of the background >> roots of the physical laws, below our substitution level. > > But how is the "first person point of view" defined? Can this theory > tell me how many persons exist at a given time? In the UDA, the first person view is defined very simply by the personal memory, or the personal diary of the teleported candidate. The key is that such a memory is annihilated and reconstituted when the experiencer is annihilated or reconstituted. (Which explain what the first person will not "feel the split" (to talk like Everett). The third person description is given by what is not annihilated, like a local observer, or eventually the number theoretical relations themselves. IN QM, roughly speaking the third person view can be described by Everett universal wave, or Deutsch (Heisenbergian) universal matrix. And the first person views are given by what the machine observer memories in each branch of that differentiating waves. But with comp, eventually the third person description is given by the relation between number (and definable with + and *), and the "Everett wave" is retrievable as a first person (plural) point of view. Thanks to Everett QM, we have very strong evidence, assuming comp, that we are multiple (infinities) but share deep and long computations. QM saves comp from solipsism, but matter, well Matter, does not go through. That's a good thing because no one really can defined it, nor provide evidence that it exists, beyond a naïve extrapolation of the boolean mind from its local neighborhood. And once you assume Matter, the mind body problem become non resoluble, you can even get divorce between human and exact science, leading to people having no idea what is going on, yet having all means of destruction/creation at their disposal ... Brrr... 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 everything-list+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/everything-list?hl=en -~----------~----~----~----~------~----~------~--~---