On Wed, Oct 14, 2015 at 10:31 AM, Bruno Marchal <[email protected]> wrote:
> > On 14 Oct 2015, at 07:45, Jason Resch wrote: > > > > On Tue, Oct 13, 2015 at 11:59 PM, Brent Meeker <[email protected]> > wrote: > >> >> >> On 10/13/2015 8:50 PM, Jason Resch wrote: >> >> >> >> On Tue, Oct 13, 2015 at 10:21 PM, Brent Meeker <[email protected]> >> wrote: >> >>> >>> >>> On 10/13/2015 7:22 PM, Jason Resch wrote: >>> >>> >>> >>> On Tue, Oct 13, 2015 at 9:04 PM, Brent Meeker < <[email protected]> >>> [email protected]> wrote: >>> >>>> >>>> >>>> On 10/13/2015 6:36 PM, Jason Resch wrote: >>>> >>>> >>>> >>>> On Tue, Oct 13, 2015 at 6:16 PM, Brent Meeker < <[email protected]> >>>> [email protected]> wrote: >>>> >>>>> >>>>> >>>>> On 10/13/2015 3:53 PM, Stathis Papaioannou wrote: >>>>> >>>>> >>>>> >>>>> On 14 October 2015 at 09:46, Brent Meeker < <[email protected]> >>>>> [email protected]> wrote: >>>>> >>>>>> >>>>>> >>>>>> On 10/13/2015 3:36 PM, Stathis Papaioannou wrote: >>>>>> >>>>>> >>>>>> ... >>>>>> Standard computationalism does not say anything about whether matter >>>>>> is primary or not. It says that you can make a computer (or a robot) that >>>>>> thinks and acts like a human. Those opposed to computationalism disagree. >>>>>> They believe either that is not possible to make a computer that behaves >>>>>> like a human because there is non-computable physics in the brain (eg. >>>>>> Roger Penrose), or that it is possible to make a computer that behaves >>>>>> like >>>>>> a human but not one that thinks like a human (eg. John Searle). >>>>>> >>>>>> >>>>>> But the problem with what you say is that on this list >>>>>> "computationalism" tends to mean much more than, "you can make a computer >>>>>> (or a robot) that thinks and acts like a human". Bruno claims to have >>>>>> proven that your simple statement logically entails that all of physics >>>>>> and >>>>>> consciousness. But that is not so generally accepted and so when someone >>>>>> "reject computationalism" here, it may be they are just rejecting the >>>>>> inferences Bruno claims it entails. >>>>>> >>>>> >>>>> I take Bruno's term "comp" to mean computationalism plus the >>>>> conclusions he draws from it. >>>>> >>>>> >>>>> But with that extended meaning, the following two sentences are not >>>>> true: "It says that you can make a computer (or a robot) that thinks and >>>>> acts like a human. Those opposed to computationalism disagree." Those who >>>>> reject (extended) computationalism, may very well agree. >>>>> >>>> >>>> If someone believes computationalism does not lead to the extended >>>> conclusions Bruno drew from it, it is on them to show where Bruno's >>>> argument is wrong. >>>> >>>> >>>> Or why it's conclusion doesn't follow necessarily - which I and others >>>> have done. >>>> >>>> >>> I missed this. Could you point to the posts where you showed this? >>> >>> >>> No, I don't keep posts that long, but I can summarize some of the >>> arguments. >>> >>> 1. Bruno's argument that recapitulating a recording doesn't instantiate >>> consciousness is a mere argument from incredulity and doesn't logic entail >>> its conclusion. In fact the whole MGA is an argument from incredulity. >>> >> >> The computations involved in playing back a recording of a given X by Y >> pixel image are constant regardless of what the recording shows. Therefore >> I don't see how one could argue (assuming computationalism) that a >> recording invokes the computation one presumes is necessary to instantiate >> a conscious state. >> >> >> "Presume" is the operative word there. >> >> > Okay, I think we agree then. You can have a theory that says a recording > is consciousness, but then that isn't computationalism. And if it isn't > computationalism then you can't use that as an argument against the logic > used in the UDA which starts from the assumption of computationalism. > > >> >> >>> Mallah, who has his own version of the argument, takes the reductio to >>> prove computationalism is false. >>> >> >> So then what is his theory of mind? >> >> >> Dunno, probably physicalism. But he's not obliged to have an alternative >> in order to reject one premise of a reductio. >> >> > I'm nor familiar enough with his argument to comment. > > >> >> >>> >>> 2. Bruno's argument depends on the MG being conscious while isolated. >>> He tries to make this plausible by supposing that his scenario takes place >>> as a dream; but this is not convincing. Even a dream needs prior >>> experience of an outside world. >>> >> >> Do you deny the possibility of mind uploading on the account that it >> could be isolated from the outside world? >> >> >> Yes. A brain in a vat with no connections would not be able to sustain >> consciousness. >> > > Can you dream? > > >> >> >> >> >>> So on reflection it is not plausible that consciousness can exist >>> independent of an external world. It may be that physics can be derived >>> from arithmetic, but it is not enough to say that self-evaluation exists in >>> arithmetic. For the theory to work it must produce a world to be conscious >>> of, and so far it doesn't do that. Bruno just writes things like, IF comp >>> is true then physics must follow. But that's like saying IF Christianity >>> is true Jesus will return. >>> >> >> There is a lot of independent evidence for computationalism. >> >> >> Name some that's not also compatible with physicalism. >> > > Cochlear implants and artificial retinas give evidence toward multiple > realizability,and therefore, against mind-brain identity theory. > > > I think I guess your intuition here. But it is hard to convince Peter > Jones like argument with the mutiple realizabity, as they use their > ontological commitment in primary matter, and would say that whatever > realis-zability is used, it has to be built on the primary matter to > "create" consciousness. That is why I make the MGA, to show that this does > not follow, without making that primary matter having a non turing emulable > role in that generation of consciousness, which is a way to say that comp > is false, or true by magic. > > > > > They show that it is functional equivalence, rather than > material/compositional equivalence that matters. Since computers can > realize any finite function, then assuming there are no necessary > infinities within the brain, computers can realize any functional state the > brain is able to realize. For physicalism to be correct, you have to > believe either that functional states are irrelevant to consciousness, or > that physics can instantiate functional states which Turing machines cannot. > > > > OK. > > > > >> >> >> The Church-Turing Thesis means a computer can perfectly replicate all >> human behaviors. >> >> >> First, it's a "thesis", not a fact. >> > > It's as well accepted as any theory in physics. > > >> Second, it doesn't mean that an abstract computation can replicate human >> behavior. >> > > It does. But you would say that it would result in a zombie. > > >> If human cogitation is Turing emulable, it may still have to be >> physically realized, >> > > The Church-Turing thesis says nothing about computers having to be built > out of electrons and quarks. > > >> which means its finite, which means the infinities of arithmetic are not >> necessary to intelligence or consciousness. >> > > I don't think infinities are necessary to consciousness. It just so > happens that in arithmetical realism, there are infinite numbers of > platonically existing computations. > > >> >> >> A rejection of zombies, or a rejection of the idea that we can have no >> reliable knowledge of our own conscious states + Church-Turing Thesis gives >> you computationalism. >> >> >> But what do you mean by "computationalism"? >> > > That the proper computation results in consciousness. > > >> Just that consciousness can be instantiated by an artifact?...by a >> digital computer? Or does "computationalism" imply all the inferences >> Bruno argues for, but which are not commonly accepted. >> > > My above insertion was that the "No zombie principal" + "Church-Turing > thesis = true" -> computationalism. > > > > Interesting. I will think about that. It is of the type of "non solipsism > + comp" implies first person plural (which means that if I am duplicated in > the normal computations, you too". > > The problem is that comp, strictly speaking allow quasi zombie. Entities > which looks like being conscious for a short time, and "short" is hard to > define. > > > Apparent zombies can exist, but a true zombie (one that is functionally identical to a conscious being but not conscious) would violate the "no zombie principle". Since Turing machines can implement any function to any arbitrary precision, it follows that a turing machine could be functionally identical to a conscious being. But if it weren't conscious it would violate the no zombie principle. Therefore the turing machine must be conscious. > > > It is the UDA which says computationalism -> the reversal. If the UDA is > correct, then transitively "No zombie principal" + "Church-Turing thesis = > true" -> the reversal. > > >> >> >> >> >>> >>> 3. Peter Jones wrote several critiques pointing out that there is no >>> reason to suppose a UDA exists, it's merely a hypothetical abstraction. A >>> related criticism is that Bruno assumes arithmetic is infinite in order to >>> use Godel's theorems about what a system cannot prove about itself. But >>> physics doesn't need infinities, they are just calculational conveniences. >>> >> >> Ultrafinitism is a fringe theory which leads to a break down of >> mathematics as we know it. I think it is an extreme length to go to reject >> the UDA, to say there is a biggest number to which 1 cannot be added to. >> >> >> That's just your prejudice. Try reading Feng Ye and Jan Mycielski. I >> think it's telling that you look at the mere existence of alternative >> number theory as destroying "mathematics as we know it". Mathematics is >> just a bunch of axiom/theorem systems. There's no one really real >> mathematics any more than there's one real language. >> > > Well then, if there are many mathematics, and they are all equally valid > (as you say), then we inhabit the ones that contain Universal Dovetailers. > > > All mathematical theories agrees on the truth of the sigma_1 sentences of > arithmetic. If Turing complete, they prove the same behavior for all > machines. The sigma_1 truth is accepted by all mathematician (except the > ultrafinistic). Philosophical nuances begins on the negation of the > sigma_1. All universal dovetailing do the same in all mathematical > realities. > > Computationalism has an ontology on which everyone agree. Those who claim > to disagree usually add philosophical commitment which is not used in the > reasoning. > > That is a subtle point rarely explained, but the book by Torkel Franzen on > "Inexhaustibility" makes a good try. > Thanks I will look into this book. Jason > > > > > > >> >> >> >> >>> >>> 4. Bruno leans heavily on saying his theory explains QM, but it's not >>> clear to me that it's even consistent with QM. For example how is the >>> operation of Shor's algorithm consistent with the multiple threads of the >>> UDA? >>> >>> I think Bruce Kellet has also made some critiques of Bruno's argument. >>> >> >> Bruce's argument is that computationalism is false, and arithmetical >> realism is false. If you reject these, it is no conflict with the UDA, >> whose logic depends on those assumptions. >> >> >> Again it's not clear what you mean by computationalism. Bruce can speak >> for himself, but I think he agrees that strong AI is possible. >> >> > Strong AI implies consciousness is substrate independent. But I thought > Bruce argued against consciousness being derivable from mathematical > computations, which would mean consciousness is substrate dependent: that > it depends on physically implemented Turing machines. > > >> >> Your argument in #1 and #2, also rests implicity on a rejection of >> computationslim. #1 implies the computations don't matter, and #2 implies >> the right computations don't matter if they are isolated. >> >> >>> It is a red herring to ask "where is the error" because I don't think >>> his argument is a fallacy; I think it is less than logic entailment. >>> >> >> You can dispute the assumptions (computationalism, infinity, arithmetical >> realism, etc.) but if you reject infinity or arithmetical realism, you must >> also reject Church-Turing's thesis, >> >> >> Why? Arithmetic is system of propositions. Whether it is real or not >> has no effect on Church-Turing. >> > > Then you should be equally happy to have your brain implemented by "unreal > computations" as "real computations" :-) > > >> >> >> and it's difficult to make sense of computationalism if you can no longer >> define computation or computability. So if you accept computationalism, you >> are implicitly accepting infinity and arithmetical realism. Given this, the >> rest of Bruno's result is a logical proof, which is either correct or has >> an error. >> >> >> No, you missed the point that it is not a logical proof. It's an >> argument from incredulity. >> > > "A Turing Machine either reaches a halting state or runs forever" > "A Turing Machine can emulate any other Turing machine" > "A Turing Machine has an unlimited tape" > > None of the above items are true for physical approximations of Turing > machines. If these basic principals of computer science are not true for > physical Turing machines, then what kind of Turing machines are they true > for? Is Computer Science founded on lies, or does it concern itself with a > recently-discovered mathematical object for which these statements are true? > > > > OK. Good question. Sometimes people refutes science with "bad philosophy". > Now, I agree that the FPI need the observers in believing in (analytical) > limits, but then a lot of physics assumes this too and not always in the > machine epistemology, like we can do with c. > That difficulty is unavoidable. By incompleteness no machine or theory can > define in its own language the semantic of that theory or of the machine's > beliefs. <>t is already religious, and indeed belongs to G* \ G. > > Bruno > > > Jason > > > -- > You received this message because you are subscribed to the Google Groups > "Everything List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/everything-list. > For more options, visit https://groups.google.com/d/optout. > > > http://iridia.ulb.ac.be/~marchal/ > > > > -- > You received this message because you are subscribed to the Google Groups > "Everything List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To post to this group, send email to [email protected]. > Visit this group at http://groups.google.com/group/everything-list. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Everything List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. Visit this group at http://groups.google.com/group/everything-list. For more options, visit https://groups.google.com/d/optout.
