On 13 Apr 2017 2:56 p.m., "Bruno Marchal" <[email protected]> wrote:
On 13 Apr 2017, at 13:35, David Nyman wrote: On 13 April 2017 at 10:09, Bruno Marchal <[email protected]> wrote: > Oops. > > Yet, you asked me about the modal logic, and I do think at some point I > have to say precisely what the G box [] is for. It is simple, but long and > tedious. I hope you have not too much difficulties with my previous post. > > To make this accessible, the universal dovetailer argument can be seen as > a layman version of the "machine interview (the G logics and its variants > imposed by itself). (even if it came befoe in my mind when observing the > amoebas). I replace the machine by a human willing to be that he is a > machine, and the thought experiments explains the reversal. The G logics > makes this precise, and more abstract/general, but it relies on the > extra-ordinary works on computability and logic of Gödel and others, which > are not that difficult, but is hard to sell without some precision. Don't > hesitate to encourage me to rewrite a passage in some more shorter way, in > case you lose the line. > Thanks for this Bruno. I suppose I ought to say I have some special interests here. In the first place, my interest is in understanding what ultimately could justify characterising any phenomenon as internal, intrinsic, or subjective; IOW what is called 1p in this forum. The reason is that I think this is an extremely under-examined, not to say almost terminally confused, notion in philosophy of mind and the one where things tend to go wrong at the start. For example, the idea that we can appeal to some "inner" nature of matter as in panpsychism, weaselling about "illusion" or "seeming" in what is variously called materialism, physicalism or naturalism, or the biological obscurantism resorted to by Searle in the face of the Chinese Room argument. This is also part of what I was getting at about unexamined extra presuppositions more or less tacitly assumed in informal reasoning - in this case the implicitly assumed "free lunch" of an inner or subjective point of view absent any explicit supporting rationale. So one of the most interesting things about comp for me has always been the possibility of explicating such a perspective explicitly in terms of a logic particular to the first-person. Always trying to go right at the heart of the problem. OK. The reason I see this as so central is to do with the gap between syntax and semantics, to use Searle's terminology. This gap has struck many as unbridgeable, whether frankly so as in mysterianism (e.g. McGinn) or implicitly in the belief that all further questioning will effectively be terminated by a completed neuroscience. So comp, here will go toward mysterianism. Comp gives a reason why there is a mystery from our limited terrestrial and locally mechanical points of view. So what I find of particular interest in comp is the possibility of bridging this gap in an intelligible manner. OK. But it is necessary a bit subtle. In a nutshell: G cannot bridge the gap. G* can bridge the gap (but G cannot believe in that bridge without becoming inconsistent) S4Grz (the first person) cannot see that there is a gap. They have to dialog. The choice will be, like Paul Valery said, between logic and war. I cited Wittgenstein's idea of the explanatory ladder that is discarded when the sought-after next level is reached. In this case the ladder is the temporary conceptual aid of an external interpretation (aka a view from nowhere) imposed on what is still fundamentally a third-person syntactical procedure. OK. This is, or should be, a provisional "transcendence" from syntax to semantics that is justifiable as an external prosthesis only so long as it leads to a replacement that can do away with the need for it. Hmm.... You seem to want to replace the Outer-God, by the Inner-God. That is a risky move toward solipsim. S4Grz does not see the gap, but it does not see the other minds either. O K , I've been re-reading the relevant section of T he Amoeba's Secret just to check my intuition about what I said above . There you remind us that in the final analysis the truth of p as distinct from its believability can't be proved in the language of the machine. Hence we postulate it as true by assertion. Let us accept this postulation for the moment as "provisionally" true. Then let us consider the case where a machine " believes " in the existence of a concrete perceptual reality, the 1-truth of which it also 3- asserts, but of course cannot communicate. Concealed i n the gap between the provability and communicability of the 3- belief and the unprovability and incommunicability of the 1- assertion is the elusive transition from syntax to semantics. The latter we will conceive as an additional " interpretation " of the former, but - by assumption of computationalism - any such interpretation *must be internally sustainable* . That is, it cannot depend, other than as a temporary pedagogical training tool, on the notion of any independently "external" mode of interpretation. Now we conceded a moment ago that the assertion of p was to be considered provisionally true. So let us consider the contra case that it is in fact false. IOW i n despite of the machine's explicit 3-belief in a concrete perceptual reality, we shall say that it is in not fact the case t hat any such 1-reality exists for the machine. The machine IOW is not truly a knower from its "own" point of view or - what is the same thing - it essentially possesses no such 1-point of view. But this is surely equivalent to saying that the attempt at a transition from the 3-syntax of the machine to its corresponding 1-semantics fails. Since this transition was to have be en effected by means of an interpretation internal to the machine that fails also. Furthermore since, again by assumption, it is we who find ourselves in the selfsame position of the machine, it must also follow that we (i.e. our 1-selves) likewise " fail " in precisely the same sense . IOW we too lack a 1-point of view. But this of course is patently false. To accept it would amount to the total annihilation of any concrete perceptual reality and equally of any intelligible terms in which it could possibly be discussed. So w e can see that under these circumstances our Wittgenstein ladder - the ultimately replaceable pedagogical assumption of a putative mode of interpretation external to all machines - would fail to arrive at the sought-after, self-sustainable terminus . So on the foregoing argument we seem to have achieved a reductio of any putative element of falsity on the part of our "provisional" acceptance - assuming comp - of the axiomatic truth of p by assertion, as applied in the philosophy of mind. T he intelligibility of the entire comp theory can now be seen quite simply to depend on it. IOW, it is now abundantly clear that comp can admit of no "illusions" or "seemings" on pain of explanatory collapse. ISTM that something not unlike this way of thinking about the matter may be at least a component of a possible intermediate level of "pedagogical" intuition between those of the UDA and the AUDA. At least it seems to play that role for me, in priming the pump of intuition with respect to what comp can distinctively tell us concerning the nature of the oh-so-elusive internal-external dichotomy in the philosophy of mind, and in so doing at least indicate how that selfsame gap may be populated with an intelligible semantics. Or more simply it seems on this basis not quite so impossible that I myself could some day be willing to accept the possibility of my machine-nature without any concomitant grief for the irretrievable loss of my humanity. David The needed replacement is an intelligibly internalised semantics, or self-interpretation, which ISTM must be the very nub of a first-personal logic. OK. And as you know, that will be given, at the meta-level, by S4Grz. But the machine cannot involves its own S4Grz box. From its machine's view, it has simply no name. It is not genuinely Turing emulable. The machine can discover it by *postulating* computationalism. If this can be explicated in terms of a logic that is itself emulable in computation, then - on the comp assumption that we occupy the selfsame first-personal situation - the perceptual facts of an inter-subjective concrete reality can then retroactively illuminate the theoretical framework that has explicated them. Through the bet in a sharable reality, but at the price of having to justify its physical aspect without invoking a physical reality. Equally, denying such facts becomes an effective nullification of the sense of any framework paradoxically adduced in their support (as in any Churchland or Dennett-style weaselling about "illusions"). We agree that such weaseling in the nth attempt by materialist to put the first person under the rug. I think these aspects may be so important to communicating these ideas to a wider appreciation, as opposed to a technical/professional audience, that if at all possible some Hilbert-style man-in-the-street way of encapsulating their most essential features, without involving a detailed point for point grasp of the bulk of the technicalities, would be valuable indeed. Most people use (often unconsciously) authoritative argument. So, they will listen to the layman account when they are reassured by the authority. They are unaware (unlike you) of the existence of the problem to begin with. They confuse G and S4grz, the 3-p and the 1-p. By the most essential features I mean for example the distinctions between belief (or judgement) and perception, communicability and incommunicability, formal and informal. No doubt you might point to others. The UDA, while valuable in itself, assumes the first-person position (as indeed it is entitled to do on the comp assumption) but without providing any explicit rationale for it. Of course, that is the intended function of AUDA and the machine interviews. But then that takes us immediately into the complexities of the formidable technical batteries of mathematical logic. Is it not possible that there might be some intermediate pedagogical step? To explain the formal, a minimum account of the formal is necessary. I think. You will not escape the (very easy) exercise I gave in my yesterday post :) Please don't panic. If you can't do it, which is quite possible for diverse (psychological) reasons, you will still understand, when I will provide the solution, that it was extremely easy. And you will learn something important, I think. Perhaps crucial. And also you will see that we are very near the definition of the G box, which I need to explain the Solovay's theorem, which is the main reason why the machine point of view explodes in the 8 views. Without Gödel's incompleteness theorem, everything I say would collapse in triviality. The KEY is that a consistent machine cannot proves its consistency, but still can prove exactly: that if she is consistent she cannot prove its consistency, a bit like a conscious being can grasp that it cannot prove its consciousness to another. Raymond Smullyan has made an account of G by replacing the "diagonal lemma of Gödel" by his "island of Knight and Knaves" (in his book "Forever Undecided"). It does not really work. The puzzle remains technically difficult for most, and, worst, most people believe even less in G, as it seems to require a fairy tale construct. Also, most of his philosophical remark shows that he confused, like everybody, G and S4Grz, once he refers to himself or to the mathematicians. I tried to suggest to some people to read The Theaetetus of Plato as an intermediate between UDA and AUDA. But here they tend to agree with Socrates' refutation of Theaetetus, and again, I am forced to (try to) explain to them that Gödel's incompleteness put some doubt on the validity of Socrates definition (in the comp frame). I will think about searching that intermediate account, between the machinery and its intuitive meaning, meanwhile you might to try to progress in my yesterday post, taking all your time, and asking any question if something looks unclear or need complimentary information. Maybe read only the (long) last paragraph. Possibly *you* will succeed in finding that intermediate account! That would be wonderfully helpful, for me and the others. Anyway, thanks for the interest. It is only by open dialogs that we can grasp the difficulties and ease the search toward solutions. Bruno David -- 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 https://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 https://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 https://groups.google.com/group/everything-list. For more options, visit https://groups.google.com/d/optout.
