Cosmology, Quantum Mechanics & Consciousness Message Board › What IS "consciousness"?
Hi Peter, Andrew, Camilla (and others .. ?)
Firstly Peter (in white; Camilla in olive-green):
Hope you – and readers in general – won’t object to a running resume of outstanding points:
“For what it's worth, since you introduced the topic earlier as an objection to something I put forward, I believe I am an adherent of "Strong AI", and that I have clearly proved that the difference between Weak and Strong AI is null. Note: the definition I am using here is that used by John Searle as part of his Chinese Room argument. "Weak AI only claims that machines can act intelligently. Strong AI claims that a machine that acts intelligently also has mind and understands in the same sense people do."
“Laplace is long dead, and so (I trust) are his daemons. He was indeed a great mathematician: the key error he made was caused by not knowing about Chaos Theory which would have required time travel. However, if we are going to refer to such concepts as his " infinitely precise calculator who comprehends every single..." we need to understand the flaw. It's the same flaw I was describing in my previous post. Such a calulator cannot, in principle and in theory, exist. He is envisioning a causal network which (as per my original definition) is apart from, a separate entity to, and capable of being executed independently of, the original system, but which can still predict the original system. As I showed earlier such a causal network cannot exist for any non-analytic system.”
“This somehow loses the FAR MORE important point, at least as far as I'm concerned, that chaotic systems are sensitive to the resolution of the computation of the functions. Only infinite resolution (i.e. zero time interval and zero chunk size) is sufficient. Any finite resolution yields the possibility of arbitrarily large errors”
“The issue there is for physics, that predictability ought not to rest solely on those within the set of the universe as it unfolds, despite this being Popper's fervent wish.”
“I am unable to determine what this means. The Universe itself cannot be a set, of course”
Yes I admit to being a little unclear about what Camilla was precisely trying to say.
“However, moving on... I am quite certain that this:
once we're satisfied that we've harvested the important data abstracted from some system of interest then we should be able to account for the entirety of the system's behaviour,
“I disagree profoundly; but that is really quite irrelevant, because what matters is that this explains your viewpoint; and, right or wrong, it is tenable and self consistent.”
[Hrrrmmmppphhh!! .. As opposed to .. ? ]
“What comes though loud and clear is that you really do not like emergence or emergent phenomena. If you view all mental activity at the level of neurons firing and interacting (or analogous activity in a non-human AI) then all that you say is trivially true: i.e. conscious decisions would not exist
. we would indeed simply sense everything mental”
Ah, but there’s the rub you see. As far as l’m concerned my own admittedly trimmed-down version of David Chalmers’ famously-named Hard Problem is precisely that we do not have the faintest inkling, scientifically speaking, of what “sensation” means, and therefore are not even in prospect of building any machine – whether made of semiconductor-based chips, living neurons or whatever – which could be conscious.
So such an understanding is certainly not “trivial”. It is l am claiming at the nub of the issue as to what consciousness means! (l admit that Camilla remains to some extent fixated by a residual fascination with the “freewill issue”.)
Note that l am certainly not claiming that such an accomplishment would be impossible in principle, because as said we are of course conscious!
l am a straightforward materialist. That’s not philo-techno jargon. It’s colloquial English and communicates pretty exactly with any intelligent English speaker.
.. Unless l need to add – for the benefit of some perhaps less intelligent English speakers (they shouldn’t really be reading this board!) – that in this context “materialist” does not mean “one obsessed with acquiring consumer goodies, services, and miscellaneous financial services such as exotic holidays, and so on! (l’ve actually met people – too many! – who as soon as the word “materialist” is mentioned jump to that very conclusion, which is why in accordance with the prevailing philosophical parlance of the past 50 years or so l have dedicatedly stuck to the alternative and more precise nomenclature “physicalist” (Otto Neurath, logical positivist, 1935. Scarcely “venerable”, but, rather, coeval with the dads or at most granddads of each of our correspondents! Neurath chose the nomenclature in view of the impingement of QM on physicists’ thinking, and the dawning of the realisation that matter is certainly not after all the be-all-and-end-all, ontologically speaking!)
“We then sneak an idea from special relativity - namely the equivalence of all inertial observers - and then pervert it beyond all measure to apply to philosophical viewpoints, and use that idea to assert that you are completely entitled to view all mental activity at the level of neurons interacting in the brain, or any other level you choose, secure in the knowledge that Einstein agrees your viewpoint is at least as valid as any other.”
Yes. You see, Peter, the – or at least my – aim and object in corresponding is to understand how consciousness works ..
.. just like any biologist surveying and analysing some biological sub-system or other. My first academic calling at Hons Degree level is biology. Should have mentioned that earlier. Apologies for the oversight! This lifelong search has carried me in as conscientious and in-depth a manner as necessary over the territories of physics, philosophy and cybernetics as well (quite naturally) and l quite frankly don’t see the relevance of any other specialised areas of expertise, and since no-one except research zooneurologist Nick Humphrey at the LSE has apparently had the remotest inkling of how to do it, l have been obliged to put together some sort of admittedly ramshackle garden-shed synthesis of my own. So yes, you’re absolutely right:
“It seems the viewpoint you have selected is not necessarily that of neurons firing and interacting, but still (what I would call) "fairly low level", and rooted in the physical reality of the brain - an actual biological organ.”
Yes, just one step upward actually – neuronal circuits rather than individual neurons. Why? Because that’s the level at which one expects to find the answer to the mystery! Not in how many different flavours of quarks and leptons there are, not in the chemistry of neurotransmitters, not in the composition of nodes within some information-processing network (as earlier mentioned l'm not a type materialist) and certainly not in programmable Turing machines ..
[ .. to be continued in about 1 hour! .. ]
I'm happy to accept that whatever process or procedure you are putting forward here does not result in intelligence or consciousness emerging, as I don't understand what your procedures are, but they clearly bear no relation to ideas I am proposing to explain these phenomena.
On a matter of principle, if something is unclear it cannot be employed in the business of clarifying relations to other ideas, or not.
Now returning to procedures:
Function is a mathematical term, meaning some formula or computational process defined on a collection of input parameters and yielding some result, often some new collection of values for those same parameters. You spoke originally of "the function(s) describing the system", and I assumed you were referring to the function that constituted the causal network of the system (note italics). Why? Because those are the only Functions in view. The system itself is NOT described by ANY functions - it's non-analytic. Its causal network comprises the system itself and its environment. It is part of reality and acts (deterministically of course by virtue of the definition you have supplied) in accordance with the laws of physics. These behave smoothly, of course, but they are not "Functions" however hard you stretch that term.
So the Functions must be those that describe the causal network of the system and the only thing they can diverge with respect to is the behaviour of the system itself. One fundamental property of non-analytic systems is that the predictions made by its causal network diverge from its actual behaviour in sudden an unpredictable ways. Indeed if you refer back to my definition of predictable you will see a definition that captures the idea of "smoothness" in a rigorous way (I hope). Non-analytic systems are not predictable: the Functions that describe the behaviour of their causal networks (and which might be used to try to predict their behaviour) diverge from reality in ways that are NOT smooth.
Point one. There are smooth interpretations.
Two, the real world system behaviour may be unique but it is generally describable by some (fn) (fx) e.g. to take but one, the Taylor series. For a treatment of a n/a (fn) I cite:
Ariel Scolnicov. "infinitely-differentiable function that is not analytic" (version 2). PlanetMath.org. Freely available at http://planetmath.org...
Therefore, there is some (fn)(fx) that attempts to capture whatever non analytic system we may discuss.
Pattern recogniton is mentioned further on as not syntactic and I might as well nail that one here - computation does pattern recognition ultimately in binary as machine code; written back and forth to stacks thence modifying databases. The machine code is syntax.
If they are not procedural or syntactic, go on surprise me, what are the base units for the mathematics of your non analytic (I am not allowed by you to call them "functions" here, you somehow regarded that as woolly, so systems will have to do)? Please straighten me out if there is some difference, as to in what measurement type non analytic systems' unpredictability or non predictability in principle is cashed out? If it turns out to be mathematical then my claim is that is syntactic and that the actions/sequences so related in any real word n/a system, whilst of an apples and pears type difference to mathematics in themselves, are still referred to strictly by the same base units of the descriptive n/a system's mathematical numerical-function. That surely is what maths does? Attempts to encode the procedures? This surely has comprehensively covered this problem? So now you could just stop questioning what the word function in n/a context refers to, it means mathematical function (fn).
Syntax of simulations versus semantics in our heads of real in-the-world analytic systems, is all that Searle was illustrating in the Chinese Room.
You dislike comparison of your shifting views but still it is necessary to move the argument along.
At various stages you have claimed the following and you cannot adhere to all of them at once, that is the logical issue:
1) Understanding = procedural/syntactic mathematics as that particular (fn) progresses.
2) Understanding ¬= procedural/syntactic mathematics as that particular (fn) progresses.
I've already observed that procedure does not, in general, lead to understanding.
Procedure does not lead to understanding, at lease not in general, and not in my experience.
Maybe, sometimes... of course.
Please explain that last sentence.
3) Understanding = learning.
4) Understanding = learning = consciousness.
A Turing machine (or any conventional computer) can be programmed to simulate the sort of architectural structure that does resemble real biological brains.
The whole point, even then, was that those programs were simply learning systems - programs with some in-built desires/motivation and an ability for self modification. I don't think you can fairly characterize such programs by the phrase "Whenever software is written, we export into coded format expertise at specific types of problem-solving" even if some expertise in the nature of the learning process may be coded".
To the extent that a brain is (in it's simplest form) just a network with desires and the ability to modify itself, then those programs do simulate brains as well as minds.
However, the big difference between us seem to be:
"... But from the mere fact that the machine is talking (or printing, or whatever) and convincingly, in the sense that it fulfils the terms of Turing's challenge -- it does not follow that the machine is conscious. It could be simply following a highly "intelligent", adaptive script."
As far as I'm concerned. that's what, that's all consciousness IS. If the script is good enough, the script is conscious.
5) Consciousness ¬= learning but requires a bit of it.
I certainly do not hold that Consciousness = Intelligence;
but I do think they are related, in that consciousness without some intelligence is a bit hard to swallow, and also because I believe that given enough intelligence (plus some other necessary pre-conditions) consciousness is likely to arise.
6) Understanding = learning ¬= consciousness because they are "apples and pears"
How do you get from that fact to the idea that I think (according to you - you have a shortcut into my head?) that Consciousness is Strong AI. I have never held such a view, nor said anything to suggest such a view.
See 4) above.
How could they possible be equal? One is a state exhibited by a mind, the other an abstract academic concept. Not so much "you can't compare oranges with lemons" as "oranges with adverbs".
So, your inconsistent thesis appears now to be that intelligence both is and isn't consciousness, that one is a process and the other a concept and that they can both arise, repeatably, from non analytic systems?
[ .. carrying on directly from above .. ]
.. and certainly not “spiritual” (whatever that’s supposed to mean) or at the “personal” level either. lf l were interested in the musings of Hemingway, l’d be corresponding on some literary website instead of this one. A good procedural recommendation lMV to describe the system of interest in terms of variables “at the next level down”. Thus for instance civil engineers do not need to bother themselves with the niceties of general relativity, special relativity, or quantum mechanics even though these theories are as far as we know “truer” than the Newtonian mechanics which entirely suffices not only for the building of suspension bridges but even to land people on the Moon! Obviously, “deeper” physical theories are in fact irrelevantly precise for the civil engineer’s purposes, and specifying the behaviours of e.g. all the quarks within the metal of the suspension bridge would be to substitute every pine-cone on every single tree for the simply and parsimoniously comprehensible wood which contains them! (l sincerely apologise for the lack of precision in the above-given, clearly hand-waving “procedural specification”, Peter!)
But as already stressed several times, my (and Camilla’s) approach is strictly reductionist. (Boo, hiss!) That is, only “weak” emergence is acceptable because “strong emergence” would violate any of the known laws of physics or would produce effects thus far unseen in what we would normally regard as fairly parochial circumstances. For instance, if QM is responsible for contributing to the possibility of either some biological or technologically speaking topologically and causally isomorphic systems becoming conscious, then we would want (as said earlier) to know why no other alternatively configured system and/or system with radically different causal properties isn’t also conscious! (for instance, Mount Everest, or the Amazon jungle, or fork handles. Whatever!)
(This is the path chosen by philosopher David Chalmers. It’s not QM-centred. Rather, it’s “panprotopsychist” in that every electron and quark allegedly contains “a grain of consciousness”. I just don’t buy that. OK, call me conservative!)
None of this of course denies the very real possibility of QM "doing some work behind the scenes" within biological systems. For instance it has been known since the '90s that in the later-stage, electron-transfer stage of photosynthesis, the electron liberated as a result of being chemically suppied by energy captured as a result of the absorption of photons does not travel "classically" from the site of liberation to the site at which it is required to reduce water, i.e. it does not travel diffusively, as would molecules, because that would entail energy loss and "thermalisation" of the electron, instead, it travels non-locally, exactly in the manner of a single photon "through 2 slits simultaneously, as in a Young's Interferometer! This definitely took me by surprise so Andrew, there's hope yet! However, these little very low level -- in Peter's definitional sense -- activities aren't essential to understanding the principle of how to transduce the energy of light into electrical energy. It's just "how nature happened to do it". (First, as usual!) It's obviously not obligatory to carry the process out in that specifically quantum mechanical way, and thus, l stress, would be also too low-level and of no explanatory avail in understanding the essential principle which lies behind the possibility of conscious systems even being possible in the first place. (Whatever principle that happens to be. As you know l have advocated one, but l think that we need to do a lot more contextual groundwork before anything that l might have to say in that regard stands the remotest possibility of being taken seriously!
“That being so, consciousness does exert a causal influence, because l'm assuming as does just about everyone else that it is perfectly legitimate to invoke a causal theory of reference.
The problem is... I looked up "causal theory of reference" on Google, and came up with a plethora of different meanings. However, NONE of them made any sense in the context of what you said.
So I have to guess what you meant by looking for meanings that make it true. I come up with this:
If you refer to anything you must be aware of it. If you are aware of it, it must have some effect on you , since even awareness is an effect. So "consciousness" must have some causal influence on the mind that somehow exhibits consciousness.
A bit of a leap in the dark there. Have I landed in the mud and covered myself in foolishness? or ... got it more or less right? lol
If you meant something like that, I agree entirely.”
“In fact, I suspect you mean that the sort of brain you have in mind can be simulated by software and that, once we have written it correctly, the simulation will show behavior and characteristics similar to the original.”
[ .. continued .. ]
[ .. continued .. ]
“Function is a mathematical term, meaning some formula or computational process defined on a collection of input parameters and yielding some result, often some new collection of values for those same parameters. You spoke originally of "the function(s) describing the system", and I assumed you were referring to the function that constituted the causal network of the system (note italics). Why? Because those are the only Functions in view. The system itself is NOT described by ANY functions - it's non-analytic. Its causal network comprises the system itself and its environment. It is part of reality and acts (deterministically of course by virtue of the definition you have supplied) in accordance with the laws of physics. These behave smoothly, of course, but they are not "Functions" however hard you stretch that term.”
Each “rerun” of observationally indistinguishable reconfigurations of the initial conditions of some suitably specified chaotic system will pursue a trajectory in dynamical phase space which diverges from each of the others. (Smoothly, as it happens, although clearly not "second-order smoothly". That is, the rates of curvature along each line which mark (or generate) such mutual divergence would in general not be uniform, nor indeed would they onset at identical times from one trajectory to the next!)
Each of these trajectories could be functionally specified.
“Non-analytic systems are not predictable:”
“the Functions that describe the behaviour of their causal networks (and which might be used to try to predict their behaviour) diverge from reality in ways that are NOT smooth.”
“What, please, is your problem with strong emergence? Weak Emergence is defined as Emergence where NO unpredictable behaviour becomes apparent, and Strong Emergence as Emergence where unpredictable behaviour DOES in fact appear. We have established that deterministic tells us nothing about predictability and that non-analytic systems are NOT predictable. It is therefore not much of a shock to learn that even in a deterministic Universe non-analytic systems are nevertheless able to exhibit Strong Emergence.”
My objection is – and this is the standard definition within physics, don’t forget: Strong emergence would mandate the operation of new and hitherto unknown laws of physics !!! ..
.. which is pretty drastic according to my standards. (Don’t know about anyone else!)
Thanks for the honest reply Andrew. Please sit and watch developments then unless and until you feel an uncontrollable urge to dive in. Camilla and Peter, l’ll need to add my comments to your ongoing tug-of-war tomorrow. Alas, l’ll definitely be timed out today.
(Phew! Finished at last! No doubt l'll be greeted by a Hiroshima's-worth of further clear-up-necessitating work tomorrow.)
Peter, you say:
So, I can agree with you entirely on this: "I dispute that Strong Emergence conflates entirely new strange & rare properties with mathematical unpredictability.".
Strong Emergence is not in fact some sort of metaphysical food blender: "conflate" means to combine, meld or blend. Strong Emergence provides for the sort of commonday, easily understood phenomena that populate out daily lives, and appear robust, reliable and even predictable; but which in fact are NOT predictable in the formal sense or "foreseeable from the standard reductive level physics" (to use your terminology).
I disagree that once found, they are not reducable to standard physics.
I said:Switching from a deterministic system to an unpredictable one simply means that for that system whilst still causally dependent, logical indeterminism now rules.
Google has never heard of "logical indeterminism", so it would help if you defined your terms.
Agreed, I now think this is wrong. I was seeking to tease apart the notion of causal necessity (since you insisted n/a systems are in practise unpredictable) from their in principle predictability as it gets confusing. But I put it the wrong way around, it should have been both causal and logical are deterministic but that would not be to say anything useful really! It's not key to any development I then had in mind.
But then you go on in response to my unrepeatablility assertion about properties you want to instantiate more than once, in systems that will work like our brains: I said:
"This gains you no advantage in the consciousness forming stakes as each run of such a non analytic system will produce its own unique 'outcome', if outcome is the right term, since the function continues to run on chaotically." By function I meant mathematical (fn) and perhaps I should have said "unstably".
If you have some means to re-start and re-run a non-analytic system it will develop in the same way each time (ignoring QM uncertainty). It remains determinisitic as per your definition: its causal network works the same each time.
By your own reliance in all your arguments until now that the important feature is the in practise indeterminacy, n/a systems would not run to the same guaranteed property again and again as the environmental conditions change so much over time.
You're now suddenly hopping around from foot to foot relying on the opposite premise.
Why have you invoked this?
It does look a bit like just to disagree for the sake of it, you know, whereas my remit is that I want to move the debate onwards a bit each time.
You then end with this also puzzlingly unargued for position:
Let me propose the opposite view, because it is something I believe deeply but cannot (yet) see how to prove.
Let us postulate a universe with no QM, no uncertainty. Let us further postulate that all systems within it are analytic. Hence we have a Universe that is not only deterministic but also predictable.
Then we ask:
In such a Universe can intelligence or consciousness develop? Can free-will exist?
If you are right, and this is true:
And such internal properties do not rely on outright analytical unpredictability in principle.
then the Answer is:
YES, In such a Universe intelligence or consciousness can develop Free-will can exist (depending on what internal properties you had in mind, of course).
HOWEVER, I think:
NO, In such a Universe intelligence or consciousness cannot develop Free-will cannot exist.
Why not? As I pointed out, you NEED non randomness for causality, hence the possibility of repeatablilty of any phenomena. And obviously that's the basis of any lawlikeness and decision making.
Earlier on, we had used the term instability for n/a functions.
I prefer it because it allows for the suggestion that the environment i.e. the universe is affecting the real time function through feedback. On this view, if we could go back in time without changing anything to monitor the function causally progressing we would see the same process. So it is a knowledge argument not an ontological one and no appeal to the statement "the universe is unpredictable, so there!" will do. I don't even agree that unpredictability is logically proven, whilst it just happens that unpredictability in practise for n/as by definition does hold, as simply a working pragmatic assumption.
Why are you suggesting your predictable vs unpredictable outcome theory based on a feeling?
There is no actual argument given, you are simply relying on mathematical unpredictability once again. Which we have examined into irrelevance in getting your agreement in the first paragraph far above, repeated here below.
So, I can agree with you entirely on this: "I dispute that Strong Emergence conflates entirely new strange & rare properties with mathematical unpredictability.". Strong Emergence is not in fact some sort of metaphysical food blender: "conflate" means to combine, meld or blend. Strong Emergence provides for the sort of commonday, easily understood phenomena that populate out daily lives, and appear robust, reliable and even predictable;
I dispute that unpredictability comes into it, in fact I want mathematical predictability to be able to get a stable dynamic set of neural net systems off the ground.
Edited by Camilla Martin on Apr 3, 2012 9:39 PM
|A former member||
Responding only to points where significant disagreement remains
I don't think you responded to my assertion that your statement is untrue. I depends on what you mean by "account" of course, but assuming "account" means something rigorous, then my view is:
I do not believe we can necessarily/always fully account for, or predict, or deduce from lower level physics, all the behaviour of a system.
This point clearly needs discussion. I trust we can agree that it boils down to the issue of Emergent Phenomena in non-analytic systems, and in particular whether those Emergent Phenomena exhibit Strong or Weak Emergence.
I don't want to discuss this now, because it comes up later. Neither do I seek to prejudice the discussion in any way. I wish simply (at this point) to establish that we agree on the terminology and that we agree where we differ.
“What comes though loud and clear is that you really do not like emergence or emergent phenomena. If you view all mental activity at the level of neurons firing and interacting (or analogous activity in a non-human AI) then all that you say is trivially true: i.e.
I think I see another clear difference here - maybe two differences, but still predicated on the disagreement as to whether Strong Emergence exists.
My view is that "sensation" (and all the things you want to associate with it) are simply Emergent Phenomena. At one level, that is a sufficient explanation in itself, and if you believe (as I do) that these are Strong Emergent phenomena, then no further explanation is possible, or necessary. In my view, if we want to create a conscious being, we must create the correct conditions for a consciousness toi emmerge and allow it to happen.
You differ. I accept that. I am, yet again, trying only to establish, clearly, what our respective views are: not to argue for either camp. You believe in Weak Emergence, so you think that these emergent phenomena are capable of an explanation in terms of lower level physics and indeed you believe we need to have those explanations available to us in order to create a conscious being.
Have I correctly understood your position?
Yes. You see, Peter, the – or at least my – aim and object in corresponding is to understand how consciousness works ..
And we return again to this same point, because my views lead inexorably to the conclusion that:
No explanation of how consciousness works can exist at the level at which you seek it.
Edited by User 26,619,642 on Apr 4, 2012 10:49 AM
|A former member||
Again, addressing only points of significant disagreement.
Two points here:
1. Repeatabe Re-Runs
The definition you have supplied for deterministic (and which I have accepted) ignores QM uncertainty, and provides that the entire Universe is deterministic. Specifically, everything that can have any effect on anything is contained within the Universe, the initial conditions are known with infinite precision, the causality is subject to no uncertainty.
Under these (slightly hopeful) assumptions I am quite certain that each re-run of the universe would produce identical outcomes. For if not, some difference in the initial state, or in the causal laws, or some uncertainty must have caused this difference. But none such can exist, by hypothesis....
However, I'm not convinced this is a significant disagreement: i.e. I'm not sure it matters either way.
2. Smooth Divergence of Prediction from Reality
Again, this looks trivial but I think it may be covering something important. I asserted:
the Functions that describe the behaviour of their [ed: i.e. some non-analytic system under discussion] causal networks (and which might be used to try to predict their behaviour) diverge from reality in ways that are NOT smooth.”
and you disagreed - saying (I presume) that they diverge smoothly. Of course it rather depends on what we mean by "smooth". I concede that the divergence would be smooth MOST OF THE TIME by any meaning of "smooth".
However, there exist mathematical definition of "smooth", and all of them require "smooth" behaviour ALL OF THE TIME. I am confident that:
. if we had such mathematically smooth divergence between "the systems themselves and the Functions that describe the behaviour of their causal networks"
. then we could exploit that smoothness and establish that those Functions converged on some set of limit Functions (as time interval and chunk size tended to zero), thus converting a non-Analytic system into an Analytic system. That would be a neat trick indeed!
A mere technicality?
No, I fear.
First: I used the word "Functions", and maybe I should not have done so. I would like to apologize, especially if this has caused any mis-understanding.
The issue lies in what, exactly, these Functions are: and what people might take them to imply. Sorry, but we need to go back a bit.
I trust you you agree we are discussing non-analytic systems and that you are happy with my earlier definions under which:
All systems have a causal network that is fundamentally a part of that system, or an intrinsic description of it, which operates under the laws of physics and determines the behaviour of that system.
All systems also have a causal network (note italics) which is a separate entity, an attempt to capture the causal network in question, and re-cast it as a collection of mathematical, logical or computational units which we might use to TRY to predict the future (or past) behaviour of the system. I was using the word Functions to refer to this "collection of mathematical, logical or computational units".
Now, for an analytic system the causal network really does represent the the system's causal network exactly, and this equivalence could justify writing down the Functions and ascribing them to the causal network rather than to the associated causal network. It think this would be a mistake, because:
1. The causal network is in fact part of reality, not a piece of mathematics, so it isn't a collection of Functions.
2. But speaking in this way permits us to imagine that the causal network will always have an associated set of Functions which somehow embody the laws governing the (determinsitic) behaviour of the system, making it not just determinsitic but also predictable.
Which is not the case in general, it is true ONLY for analytic systems, as I will now prove.
It may seem a natural idea to assume that if something is deterministic there must exist some collection of rules (i.e. which can be abstracted into a collection of Functions) that capture the behaviour of the system exactly. Let me be clear: in a deterministic Universe a collection of equations that define the behaviour of any system must always exist, but those equations do not necessarily have any analytic solution. Indeed, their (frequently provable) non-existence is the definition of a non-analytic system.
So, in fact, we can see that, in general, it is NOT true that:
. if something is deterministic there must exist some collection of rules (i.e. which can be abstracted into a collection of Functions) that capture the behaviour of the system exactly.
Because, if that were true, all non-analytic systems would immediately turn into analytic systems[ since the Functions would be an exact (and hence analytic) solution to the physical equations that define the behaviour of the system in question.
But we know that non-analytic systems exist... QED
That's not all... we'll get back to the Functions and even to "smooth" eventually. Next post (silly software, silly limit)
Edited by User 26,619,642 on Apr 4, 2012 4:30 PM
|A former member||
I can contain myself no longer and must speak out!
Ian, according to my reading you are wrong about strong emergence and Peter is nearer the mark. I have read definitions including by Chalmers which do not support you. I would add that in my view the definition (as given by Chalmers) is highly subjective, totally unsatisfactory for use in any rigorous scientific work, and so from my point of view this debate about whether consciousness is a weak or strong emergent feature is just an example of what keeps philosophers talking and writing. I say that this whole aspect of the discussion is a total waste of time. My opening remarks in this thread described consciousness as evidently an emergent feature, that much is obviously right and science will clarify and explain to you in due course, if you will be patient.
I have other things to say but the purpose of holding back is to let you three get somewhere. But for a lighter comment may I direct you all to this video of a recent presentation of half an hour or so by two scientists and one philosopher, who were introducing the subject to the general public. I have to say that I found myself nodding a lot at the clarity and good sense of all three speakers. A model for this forum!
|A former member||
carrying on. Funnctions ans Smoothness
So, what exactly are these Functions?
Well... these functions describe the causal network of the system in question. For a non-analytic system no analytic solution can exist, so the Functions must be an approximation of some sort. In many cases the Functions represent an attempt to predict (not predict - that's impossible by definition) future (or past) states of the system via a simulation using finite time intervals and non-zero chunk sizes. I suspect the Functions that describe the causal network of a non-analytic system must always take this form, but I don't want to preclude some other form simply because I can't figure out what it might be.
So what is the problem?
The problem is that the Functions themselves are mathematically precise. They contain no error. It is terrifyingly easy to forget that these Functions claim only to approximate the actual behaviour of the system and transfer their precision to the "causal network" itself thus imbuing it with accuracy and properties such as the ability to predict the behaviour of the system. These are properties that the causal network of a non-analytic system cannot, in principle, posses.
Indeed, once we start talking about the causal network of a non-analytic system and the Functions that describe it in such a quantitative way we have to be very careful. We are in danger of asserting that the causal network exists (seems reasonable enough) that it has some Functions that describe it (also seems reasonable) and that therefore, if you select the right functions and/or the right inputs (like time_interval = 0 and chunk_size = 0 !) you will obtain the Function(s) that exactly describe the causal network and hence can exactly describe, and predict the behaviour of the system itself.
Nobody would (I hope) present such an argument explicitly, but within the context of a broader discussion it is all to easy to present it implicitly.
Even the idea that the causal network of a non-analytic system really exists at all can lead to falling into this fallacy. I think a better way to view the causal network of a non-analytic system as a collection of (need a new term, grrrrrrr)causitive networkseach of which has its own Fuction(s) that describe it exactly, and each of which offers an approximation to the actual behaviour of the system itself. These causitive networks may differ from eachother only in terms of time interval and chunk size, but I am prepared in principle to consider other differences also as I said earlier.
OK, that is (sadly) rather more complex, but it does eliminate entirely the dangerous idea of a causal network described by precise Functions for non-analytic systems.
Now... we all know what's going to happen next. Some bright spark with a little knowledge of A-level maths is going to take that collection of causitive networks and their associated Functions and try to put them into some suitable sequence, and get them to converge on some limit, and then claim the limit of this sequence of causitive networks is the causal network that I hold does not really exist, and further claim that the limit of the associated sequence of Functions is the Function that describes this causal network - a Function which I'm absolutely certain cannot exist (because it would transform the non-analytic system into an analytic system - see above).
Nice try. No.
The definition of predictability that I originally supplied puts requirements on the error (i.e. divergence from reality) between the Functions output and reality. Specifically, for any given required accuracy, you must be able to choose a suitable causitive network and its associated Function(s) that guarantee the error cannot exceed the required limit.
Which is PRECISELY what you CANNOT do.
Ho Hum.. Actually:
If Ian is right, and:
. the Function(s) that describe the behaviour of the causal networks of non-analytic systems diverge from reality "smoothly"
. then you can construct such a sequence and it will converge, so the limit will exist, so all systems will have genuine exact causal networks with exact Functions describing them, making the entire Universe predictable and ensuring that no non-analytic systems can ever exist.
By contrast, I hold that the divergence is NOT smooth, thus allowing for a deterministic[i/] but unpredictable universe, in which non-analytic systems can exist.
I hope it clear the difference is far from trivial
And this is why I think the issue of "smoothness" is so important.
Edited by User 26,619,642 on Apr 4, 2012 5:38 PM
|A former member||
But isn't that EXACTLY what we see all around us.
Apart from the "new and hitherto unknown" bit, of course.
So it comes down to this:
Strong emergence would mandate the operation of new and hitherto unknown laws of physics
I think the problem is that these "new and hitherto unknown laws of physics" are in fact "old and utterly commonplace laws of physics" that we are so comfortable and familiar with that we scarcely notice them at all, and certainly NOT as "Strong Emergent Phenomena".
So let's consider Boyle's law. It talks about the volume, pressure and temperature of a gas. Hmm... Pressure and Temperature don't really exist. They are statistical phenomena. In order top get from physics at the level of molecules to Boyle's law you have to do some statistics (i.e. take an average or two - pretty trivial stuff I grant you). Nevertheless you have taken a massively complex description (one involving the mass and velocity of every molecule of gas in some container, not to mention the molecular description of the container walls itself) allowed a small amount of uncertainty (by doing the necessary statistics) and ended up with a massively simpler, massively more useful, description of the system, which involves a new law of physics - Boyle's law.
I trust we can agree that this is an Emergent Phenomenon?
The question is, of course:
Is it a Strong or Weak Emergent Phenomenon?
Now, I have deliberately chosen this example because it is so simple. The gas in question shows no tendency towards self-organization, the statistics required are trivial, and the level of uncertainty introduced tiny. Whether or not this is Strong Emergence depends entirely on how much statistics you are prepared to allow into physics.
What this simple example shows is that whether or a particular Emergent Phenomena is Strong, depends on how you define "predictability".
So long as you deny the existence of Strong Emergence, I can go on pushing more and more extreme examples at you. Eventually, you will either have to admit to the existence of Strong Emergence or claim that you can deduce all science from first principles starting at (say) the elements and single atoms.
Camilla clearly believes you can deduce all science in this way:
Peter: Strong Emergence provides for the sort of commonday, easily understood phenomena that populate out daily lives, and appear robust, reliable and even predictable; but which in fact are NOT predictable in the formal sense or "foreseeable from the standard reductive level physics" (to use your terminology).
Do you, Ian? does anybody else here, believe this?
I do not. Clearly.
This is without doubt another cusp - another point where our views diverge and where this fundamental difference is responsible for all the minor issues that follow from it.
So yet again, my first objective is to ensure we agree on terminology and that we correctly understand each other's pposition.