Re: A dilemma

/snip/
Sea Wasp (Ryk E. Spoor) wrote:
Note: I PERSONALLY would love to believe that there's something Special About Me that means that no machine will ever match what I am.

But I don't believe it, and see no reason to, any more than I see a reason to believe in God.

Well, isn't that just as religious a statement as Tina's?

Not at all. We have "Universe exists with natural law" and "universe exists with natural law, plus God".

As God is at least as complex as the entirety of the rest of the universe, Ockham goes utterly berserk with his razor.

Similarly, we're physical beings

I meant 'religious' in the sense that you asserted a belief without empirical evidence in support. I'm not arguing for God. I'm arguing for evidence, and there doesn't appear to be enough on either side of the argument yet to justify belief, only open-minded, scientific enquiry.


Consciousness is one of the least well understood phenomena (in spite of being highly researched - we know a huge amount *about* it). It's not even known whether it's even in the domain of concepts that are human-understandable - it may not be.

And it may not be necessary to understand it in order to duplicate it. We didn't understand fire, but we could reproduce it quite well.

Well, sure, we already can and do all the time. It's called having kids. But that doesn't imply that you can build one out of Lego (i.e. a Universal Turing Machine). I'm not saying you *can't*. I'm saying that there is not yet a convincing argument that it's possible, and may prove to be impossible. All that's known is that we don't Know.

/snip/

It may well be possible for a
digital computer to simulate a human mind; equally it may be just as impossible for a discrete system to simulate something complex and continuous as it is for the alphabetical letters in this sentence to represent my thoughts. They do a reasonable job of representing me, but no-one would argue the letters forming this sentence that you are reading right now are equivalent to the thoughts I'm thinking, the beliefs I hold, and the sum of my awareness.

Not all things can be digitized and retain their essence.

That's an assertion without evidence.

Not sure which statement you mean there. There's a vast amount of evidence to support "Not all things can be digitized and retain their essence." Not knowing your math or computing background I'm not sure what to point in your direction though.

Words don't equate to your thoughts because they're the equivalent of a Reader's Digest of your thoughts. In digital terms, that's like doing a four-bit sampling of Beethoven's Ninth and saying that you can't digitize music because the resultant recording sucks.

That's true, but digitizing something to a finer and finer resolution won't allow a Universal Turing Machine to breach its computational limits. A hypercomputer working with real numbers (as opposed to digital approximations) could theoretically calculate things (in a single step) that a UTM could only ever approximate whilst taking all the time and space in the universe to do so. Seigelmann proved that analog recurrent neural networks can perform computations provably uncomputable by a UTM. If the brain is an ARNN, and consciousness is an epiphenomenon of that hardware, then you'll never build one out of Lego, no matter how much you have. Sorry.

There are some processes that cannot be PERFECTLY represented by digital ones, yes. But the question is whether it's reasonable to think anything of that nature is involved in the operation of our brains, which derive from progressively less complex structures until you hit single-celled organisms.

Consciousness may well be -- probably is -- an emergent phenomenon, one that doesn't show up in individual cells as part of their function. The cells operate on relatively gross (compared to quantum/subatomic) levels.

Quite a lot of neural structure function depends quantum mechanics actually. No matter how powerful the UTM it ran on, nor how fine the approximation, it's unlikely that a digital simulation of a brain would behave at all like the original it was approximating. It's be a very interesting experiment if we could try it. Greg Egan wrote some nice stories along those lines.

Unless we postulate something mystical, something nonphysical,
there's no reason at all to think that you can't duplicate the functionality of a cell, or a billion cells, of any given type.

That's not true either. Computers (assuming they're UTMs, i.e. of the type we use every day) need only classical physics to operate which is a subset of the physics the rest of the universe uses, including our brains. It's not unreasonable to suppose that the minds they support might be a result out of that larger set.

They're chemical engines, and we can simulate chemistry quite nicely.

Last time I looked the schrödinger equation only yielded analytical solutions for two bodies (i.e. one electron and one proton: the Hydrogen atom), add just one more electron and the perturbative algorithms are needed that can only approximate answers. Algorithms that solve N-body (non-relativistic I might add) Schrödinger equations become extremely hard to calculate with any accuracy as the number of charged particles is increased. This is why is takes the world's most powerful supercomputers to calculate the folding properties of _single_ protein molecules (and often the results are best guesses). I believe the current limit with _tolerable_ error levels in computational chemistry is a system with about 40 electrons. (And this is using equations that don't model the physics properly. And even if they did, using physical models that we don't know for certain are actually correct.)

> Do this
with a few billion brain cells, and you have the potential for consciousness there

Well, let's see, a brain has 10^11 neurons, and about 10^27 atoms with atomic numbers between one and (say for simplicity) eight (for oxygen), i.e 16 charged particles each so let's say we have to simulate the chemistry of a system with 10^28 charged particles... and our simulations barf after... oh. Forty.

-- unless you contend that there is some nonphysical element involved in consciousness.

There is no requirement to invoke non-physical elements to argue against a UTM consciousness.

Gruff
.

Relevant Pages

  • Re: Symbol Grounding Problem (attn: Vend)
    ... For example, Physical CT covers operations on real-valued quantities, continuous dynamical processes operating on strings (as in certain kinds of connectionist computing and molecular computing), and quantum processes. ... brain over biological time. ... not make the physical system a Turing Machine. ... "e have good reason to believe that the laws of physics are computable, so that we at least ought to be able to simulate human behavior computationally." ...
    (comp.ai.philosophy)
  • Re: Trying ti install FreeBSD 5.1 on Sun Ultra 10
    ... In the computing world, nobody cares, even if they do. ... > And it seems that physics in the high-schools in the US has pretty ... what can you expect from brain damaged software on brain ... and hardware builders that never seem to learn from history, ...
    (comp.unix.bsd.freebsd.misc)
  • Re: consciousness
    ... The mind is the brain. ... because the other person doesn't have the same pattern detectors in their ... because of the normal laws of physics. ... and memory chips that store the picture of the car. ...
    (comp.ai.philosophy)
  • Re: Penrose vs the Robot
    ... then is it true that it is STRONGER than a Turing machine? ... by a universal model computing machine operating by finite means." ... "Quantum mechanical measurements on a physical system are ... "Penrose argues that the nondeterminism of physics ...
    (sci.logic)
  • Re: Problem for physicalist evolutionists
    ... A lag of 500 milliseconds is plenty of time for the brain to stitch ... If the lag were too long then the robot might seem a bit slow on the ... visual fixation our brain builds a fleeting representation of the ... laws of physics, or how the robot is built, there could be no way to ...
    (talk.origins)