Hi, it seems you start with the assumptions that an AI can't do science as humans... to conclude just that.
Regards, Quentin 2009/8/6 Colin Hales <c.ha...@pgrad.unimelb.edu.au>: > Man this is a tin of worms! I have just done a 30 page detailed refutation > of computationalism. > It's going through peer review at the moment. > > The basic problem that most people fall foul of is the conflation of > 'physics-as-computation' with the type of computation that is being carried > out in a Turing machine (a standard computer). In the paper I drew an > artificial distinction between them. I called the former NATURAL COMPUTATION > (NC) and the latter ARTIFICIAL COMPUTATION (AC). The idea is that if COMP is > true then there is no distinction between AC and NC. The distinction should > fail. > > I found one an one only situation/place where AC and NC part company. Call > this situation X. > > If COMP is false in this one place X it is false as a general claim. I also > found 2 downstream (consequential) failures that ultimately get their > truth-basis from X, so they are a little weaker as formal arguments against > COMP. > > FACT: Humans make propositions that are fundamentally of an informal nature. > That is, the utterances of a human can be inconsistent and form an > fundamentally incomplete set (we don't 'know everything'). The > quintessential definition of a scientist is a 'correctable liar'. When a > hypothesis is uttered it has the status indistinguishable of a lie. Humans > can participate in the universe in ways which can (apparently) violate any > law of nature. Humans must be able to 'violate' laws of nature in the > process of accessing new/novel formal systems to describe the unknown > natural world. Look at the world. It is not hard to see how humans exemplify > an informal system. All over the world are quite normal (non-pathologically > affected) humans with the same sensory systems and mental capacities. Yet > all manner of ignorance and fervently held contradictory belief systems are > ‘rationally’ adopted. > =================== > COMP fails when: > a) You assume COMP is true and build an artificial (AC/computer) scientist > <Sa> and expect <Sa> to be able to carry out authentic original science on > the a-priori unknown....identically to humans. To do this you use a > human-originated formal model (law of nature) ts to do this.... your > computer 'computes ts, you EMBODY the computer in a suitable robotic form > and then expect it to do science like humans. If COMP is true then the human > scientist and the robot scientist should be indistinguishable. > > b) You then discover that it is a fundamental impossibility that <Sa> be > able to debate/propose that COMP is a law of nature. > > c) Humans can debate/propose that COMP is a law of nature. > > BECAUSE: (b) <> (c) they are distinguishable. NC and AC are different > THEREFORE: ts cannot be the 'law of nature' for a scientist. > THEREFORE: COMP is false in the special case of (b) > THEREFORE: COMP is false as a general claim. > > (b) is not a claim of truth or falsehood. It is a claim that the very idea > of <Sa> ever proposing COMP (= doubting that COMP is true) is impossible. > This is because it is a formal system trying, with a fixed, formal set of > rules (even self modifying according to yet more rules) to construct > statements that are the product of an informal system (a human scientist). > The very idea of this is a contradiction in terms. The formal system is 100% > deterministic, unable to violate rules. When it encounters a liar it will be > unable to resolve what falsehood is being presented. It requires all > falsehoods to be a-priori known. Impossible. How can a formal system > encounter a world in which COMP is actually false? If it could, COMP would > be FALSE! If COMP is true then it can't. Humans are informal....ergo we have > some part of the natural world capable of behaving informally....=> GOTCHA! > > This argument is has very 'Godellian' structure. That was accidental. > > When you say 'physics is fundamental'. I don't actually known what that > means. > > What I can tell you is that to construct an authentic ARTIFICIAL SCIENTIST > (not a simulation, but an 'inorganic' scientist), you have to replicate the > real physics of cognition, not 'compute a model' of the cognition or a > 'compute a model of the physics underlying cognition'. Then an artificial > scientist is a scioentist in the same sense that artificial light is light. > > R.I.P. COMP > > => Strong AI (a computer can be a mind) is false. > => Weak AI (A computer model of cognition can never be actual cognition) is > true. > > It's nice to finally have at least one tiny little place (X) where the seeds > of clarity can be found. > > Cheers > colin hales > > > > > 1Z wrote: > > On 31 July, 22:39, David Nyman <david.ny...@gmail.com> wrote: > > > I note that the recent posts by Peter Jones - aka the mysterious 1Z, > and the originator of the curiously useful 'real in the sense I am > real' or RITSIAR - occurred shortly after my taking his name in vain. > Hmm....... > > Anyway, this signalled the resumption of a long-running debate about > the validity of causal accounts of the first person based on a > functional or computational rationale. I'm going to make an attempt > to annihilate this intuition in this thread, and hope to encourage > feedback specifically on this issue. You will recall that this is at > the heart of Bruno's requirement to base COMP - i.e. the explicitly > computational account of mind - on the the number realm, with physics > derived as an emergent from this. Step 8 of the UDA addresses these > issues in a very particular way. > > However, I've always felt that there's a more intuitively obvious and > just as devastating blow that can be dealt to functional or > computational notions based on physical entities and relations > conceived as ontologically foundational and singular (i.e. no dualism > please). So as not to be misunderstood (too quickly!) let me make it > clear at the outset that I'm addressing this to first person conscious > experience, not to third person descriptions of 'mentality' - so > eliminativists can stop reading at this point as there is nothing > further that requires explanation in their view (as odd as I trust > this sounds to you non-eliminativists out there). > > The argument runs as follows. To take what physics describes with > maximal seriousness - as standing for ontological reality - is just to > take its entities and causal relationships seriously to the same > extent. God knows, physicists have gone to enough trouble to define > these entities and relationships with the most precisely articulated > set of nomological-causal principles we possess. Consequently, taking > these with maximal seriousness entails abjuring other causal > principles as independently efficacious: i.e. showing how - or at > least being committed to the belief that - all higher order causal > principles somehow supervene on these fundamentals. Any other > position would be either obscurantist or incoherent for a physical > realist. > > Now I should say at this point that I'm not criticising this position, > I'm merely articulating it. It follows from the foregoing that > although we may speak in chemical, biological, physiological or > historical narratives, we believe that in principle at least these are > reducible to their physical bases. We also know that although we may > speak of cabbages and kings, weather, oceans, processes, computations > and untold myriads of equally 'emergent' phenomena, we still must > retain our commitment to their reducibility to their physical bases. > So of course, we can - and do - legitimately speak, in this way, of > physical computers as 'performing computations', but following the > foregoing principle we can see that actually this is just a convenient > shorthand for what is occurring in the physical substrates upon which > the notion of computation must - and of course does - rely for its > realisation in the world. > > To be more explicit: The notion of a 'program' or 'computation' - when > we place it under analysis - is a convenient shorthand for an ordered > set of first person concepts > > > In what sense "first person"? Surely not in the sense that qualia are > supposed to be mysteriously and incommunicably first-person. > > Presumably in the sense that something is only a computer > when regarded as such, (like certain pieces of paper being money). > But that is quite contentious. It is not enough to say "under > analysis", > one must actually analyse > > > > which finds its way into the physical > account in the form of various matter-energy dispositions. The > macroscopic media for these are variously paper and ink, actions of > computer keyboards, patterns of voltages in computer circuitry, > illumination of pixels on screens, etc. All of these, of course, can > - and must - reduce to fundamental relations amongst physical > 'ultimates'. At some point after entering the physical causal nexus, > this chain of dispositions may re-enter the first person account > (don't ask me how - it's inessential to the argument) at which point > they may again be construed *by someone* in computational terms in a > first person context. But at no point is the 'computation' - qua > concept - in any way material (pun intended) to the physical account; > a fortiori, in no way can it - or need it - be ascribed causal > significance in terms of the physical account. After all, what could > this possibly mean? Are these spooky 'computational' relationships > 'reaching across' the energy-transfers of the computer circuitry and > changing their outcomes? Of course not. How could they? And why > would they need to? Everything's going along just fine by itself by > purely physical means. > > I hope the foregoing makes it clear that computer programs and their > computations - at the point of physical instantiation - literally > don't exist in the world. > > > It doens't remotely. Just because something (eg a horse) > isn't a *fundamental* constituent of the world doesn't make it > non-existent in the sense that unicorn is. > > > > They're semantic formulations - ways of > speaking - that have applicability only in the first-person context, > and we can see that this is true any time we like by performing the > kind of 'eliminativist' demonstration performed above: i.e. we can > eliminate the concept without affecting the action on the ground one > whit. Of course, this is the insight that makes the strictly physical > account of mind - as presently understood - problematic if one wishes > to take the first person seriously, because it shows the notion of > 'emergence' to be redundant at the level of causation. It's just > another way of speaking, however much insight it carries - for us. > However, it isn't my wish to make that point again here. Rather my > intention has been to show that whatever options are left in strict > physicalism to address the first person issues seriously - without > eliminating them - emergence is emphatically not one of them. > > I hope this makes the argument clear, and also illustrates the point > of Bruno's reversal of numbers and physics to save the computational > account of mind (and body, as it happens). To be absolutely explicit: > if functional-computational relations are to be taken to be > fundamentally causally efficacious, they must be held to be real and > foundational in exactly the sense (RITSIAR) ascribed to those in the > physical account. But for that to be the case, all other causal > relations must supervene on them - again just as in the physical > account. But now, of course, this must include physics itself. > > > What must include physics itself? > > > > Now, you don't of course have to accept COMP. But if you want to be a > physical realist, it means you can only hang on to the computational > explanation of mind by eliminating the mind itself from reality. > > > I don't see how that follows at all. > > > > Personally, not being committed to such an explanation, this doesn't > in itself constitute my problem with current physical accounts. The > alternative is rather that physics as an account of mind must be > incomplete, or else it is wrong. But that's another story. > > David > > > > > > > -- All those moments will be lost in time, like tears in rain. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "Everything List" group. To post to this group, send email to email@example.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 -~----------~----~----~----~------~----~------~--~---