> >Obvious sources of nondetermination:
> > Quantum mechanics. [Granted, magnifying it to macroscopic level is
> >difficult. But not impossible--cf. Schrodinger's Cat.]
>
> Excellent point sir! I'm a real beginner in QM... almost totally hammered by
> even the simplest parts of the maths. I'd very much like to know what you
> think about the "universe cannot have perfect knowledge about how it works"
> proposal in my last posting. Because... well, QM is very very accurate, but
> not totally so... are there any physicists around who think that the reason
> you get such bizarre effects as measurement collapsing the Schroedinger
> equation is because... erm... we're at a level of detail where... erm...
> maths can't take us any further? Where, due to the limited ability of teh
> universe to compute its own nature, we're starting to get odd results? I
> know this sounds like a shitty cop-out, but I'm interested in the idea of
> humanity's search for truth reaching an asymptotic gap from the Truth...
It isn't. I'm currently studying both Quantum Mechanics and General
Relativity on the side, and computational difficulties abound.
On General Relativity's end, the equations are highly nonlinear--and
nonlinear spells COMPUTATIONAL DISASTER in the Math Department.
Quantum Mechanics is also relatively difficult computationally, even as a
free theory--one without interactions. [Sort of like dry water in fluid
dynamics.] Including interactions: Does anyone feel up to computing
probablistic trees whose branches are based on numerically unstable
matrix eigenvalues??? [I know that's technojargon. Key words are 'tree'
and 'numerically unstable'. The latter can be used to create
computer-based chaotic systems--see comment below.]
While Feynman diagrams give a linearized version of the above that is
reasonable, and gives correct results as far as can be measured, it is
known that the answers from the above diagrams are always inaccurate.
[This is called Haag's Theorem.]
And mixing the two? LAUGH! Instant PhD for anyone who solves *that*!
> > Badly posed Differential Equations. [Nonunique solutions to
> >problems can just ruin the determinists' day, even in a classical
> >universe.] "You need to learn Diff Eq so this doesn't happen: you
> >design the space probe for Venus, send it up only to see it crash into
> >the ocean, and *then* have some smart-mouth mathematician walk up and
> >say, 'Oh, those equations had five different solutions from those
> >initial conditions.'"
>
> But that IS the fault of your first go at the maths, isn't it?
NO. It can be the fault of the physical system you are modeling, that's
the problem. This is completely different from a chaotic system, which
is [in the abstract] completely deterministic, but still unpredictable at
long time scales.
> >I currently suspect that modern computers have souls, but not spirits.
> >Under this metaphor, a successful implementation of Strong AI would endow
> >the machine with a spirit.
> >
> >I am reluctant to describe a "soul" as "spiritual", obviously.
>
> I get a buzz from the idea of explaining meta-spiritual
> information-processing as being as deterministic and physics-bound as
> chemistry... the definitions are so lax here, though.
In this case, I don't view information-processing as spiritual. I view
programming the information processing as spiritual. [No, a genetic
algorithm hunting for an efficient program is still raw information
processing; it isn't getting to spiritual *yet*.]
//////////////////////////////////////////////////////////////////////////
/ Towards the conversion of data into information....
/
/ Kenneth Boyd
//////////////////////////////////////////////////////////////////////////