[PLUG-TALK] Re: Re: Omniscience vs. Freewill

[Keith Lofstrom]

The capital-free but heavily-initialled glen e. p. ropella writes:

> I have to disagree with your general message, Keith.  Science doesn't
> work the way you're suggesting.  Quantum Mechanics is just a theory.
> It's a _model_ of how the universe works at very small scales.
> (Note that there are _two_ prongs to that point:  1) QM is _just_
> a model... the map is NOT the territory and 2) it only works at 
> very small scales.)

Keith responds:

Point 1) Of course it is a model.  We are discussing models, and models
only (reality is hard to render in ASCII).  Determinism is a model. 
Omniscience is a feature of a model.  What we can do is compare models,
and show that two models that cannot be mapped to each other are
contradictory.  A theory is a way of relating models - the theory of
gravity relates a mathematical model to an observational model of
moving bodies, for example.  Relativity relates more sophisticated
mathematics to more sophisticated observations.  Glen, as you noted
to Russ, the weak description of two models (robot world, and
omniscient God) do not automatically result in a consistent theory
relating them.  

Point 2) The QM model can be applied to many macroscopic observations.
My background is semiconductors and superconductors, and many of the
effects we see are rooted in behavior most concisely described within
the QM model.  That said, I am specifically thinking in this context
of "integration over paths", a common (but very difficult) operation
used in particle physics to accurately measure the behavior of systems.
The photon goes through slit A, AND through slit B, and around the
edge of the experiment, and three loops around Alpha Centauri, and
you can compute a probability for all of the maybes (very small for
Alpha Centauri) and sum them together and see what they add up to. 
For larger systems, the calculations usually average out to a single high
probability state, so we can usually dispense with the complicated
math.  But researchers are now constructing macroscopic systems where
this is not the case, and multiple moderate probability outcomes may
occur.  Quantum computing is one such attempt at such a system.  

There are some conjectures about the brain that suggest that the
brain itself may be such a QM amplifier (I think Penrose says that),
but lacking both detailed mathematical/logical models of the brain,
or precise observations, such conjectures cannot be elevated to
a hypothesis, much less a theory.  Neither can such conjectures be
ruled out (for now) by the inadequate models or measurements we have.

The chaos mathematics comes into my thesis to demonstrate nonlinear
amplification of very small initial perturbations.  Quantum sized
or smaller events matter given these processes.  Maybe the
butterfly/hurricane analogy doesn't twirl your propeller, but one
can connect the emergence of life to geochemistry to planetary 
physics (a fascinating set of calculations I can share with you
someday).  When you look closely, contingency on small events is
everywhere.  But convergence is also everywhere (there is usually
a maximum probability path), so the model stays managable (and
survivable, though some models will kill you).  

With QM and nonlinear amplification (if you prefer that label to
"chaos"), we can show that a complete observational model of the
universe in one state - any slice through N dimensional space-time
- cannot completely predict any other slice.  With a deterministic
model, you could precisely connect the states.  Still, the paths
usually converge.  So while some amplified quantum effect may make
the next word flabutanically unpredictable, we converge to the same
high-probability period at the end of the sentence.  

Or not,

All the paths are there in some sense, there are no "other universes". 
We inhabit a multiverse that branches and converges, creating new
information (path history) and destroying old information.  Every
path is marked by a probability (the flapping butterfly can "cause"
any number of low probability hurricanes) and without the
probability math, 1/0 determinative models are highly misleading.

As a side note, and relating this roughly to computers, we are well
on the way towards computers where indeterminism will be an important
constraint in their design and operation.  In other words, on the
Pentium 14 processor of the year 2020, you can repeat same 1000
instruction sequence and end up in two different states.  Making
such computers converge on a usable result, cost effectively, will
be a challenge for computer science - and put all the determinists
out of work.  This stuff does have real consequences.


> All the garbage about how QM relates to philosophies like Determinism
> is just a bunch of meta-physical hooey.  You might as well be arguing
> about how many angels can dance on the head of a pin.

I can't refute a sneer.  As you (correctly) demanded from Russ, could
you put some traction on those rather unbecoming statements?


> Granted, QM is the single most successful scientific theory we 
> monkeys have developed to date.  But, that doesn't mean it's true.
> As you so vividly point out, we're wrong about alot of things alot
> of the time.  I would bet my soul that we're wrong about QM, too.

Wrong in what sense?  Of course we will find more sophisticated
mathematics, and make more sophisticated observations.  But what
we have works for most of the observations we can make, even for
cosmological questions that I would have never thought related.
There are still numerous cosmological questions that we cannot
measure or map onto mathematics, so it is premature to claim that
QM does or doesn't apply there.  And QM is certainly a real bastard
to calculate with - Occam is dying of a million cuts here - so a
more calculable model is desperately wanted.

However, I have a sick feeling that more accurate, inclusive models
will be beyond the capabilities of unaugmented "monkey" brains.  The
really sick feeling is that even with augmentation we won't get there,
and HAL9000 will just say "you fleshboys will *never* understand it"
and leave us all in the dust.  I'm hoping for 'augment'.

> Now, you glibly drop a few references to G?del, chaos theory, and some
> unnamed set of other things that supposedly demonstrate that
> Determinism (another bit of metaphysics) is false.  And, this is
> hooey, too, not for the least reason that one cannot _disprove_
> metaphysical philosophies.  (enough negatives for you? ;-)

You are very determined about indeterminate determinism.  Ow, my
head hurts just writing that.  I cannot prove that determinism
is FALSE.  That takes something outside the system.  As you just
said, and go on to say in a different way.  All I can do is argue
observations and probabilities, and then engineer from it. 
Determinism may be TRUE somewhere out there in lala land, but
it makes for lousy predictivity and results in lousy design.


> If you're going to quote someone in meta-mathematics that did damage
> to our logic and the philosophy that results from it, then cite
> Tarski.  He did far more work to address truth and the concepts of
> truth than G?del ever did (or wanted to do).  There's a tidy little
> article called "The Concept of Truth in Formalized Languages" that I'd
> highly recommend.  In it he independently proves G?del's theorem and
> carries the ideas much further.

The 100 page paper ( http://en.wikipedia.org/wiki/Alfred_Tarski ) is
"tidy little?"  Perhaps you are referring to a summary.  If you are
referring to the full paper, it appeared in the collection "Logic,
Semantics, and Meta-Mathematics" which I can spend $24 on if you think
it is important.  Otherwise, point me at the summary and we can argue
from equipotential ignorance.

Note, however, that I am an engineer.  I use theories to get things
done, to map models onto "stuff" and vice-versa.  So far, I have
found that THE TRUTH is a useless hypothesis for this task.  Still,
it is an important hypothesis to some folks, so I will check it out.


> As for chaos theory, it, like G?del's theorem, is an artifact of our
> _mathematics_.  Mathematics is thought... monkey thought, to be
> specific.  [grin] Chaos theory demonstrates certain consequences that
> ensue from our mathematical infrastructure.  That's all it is.  It
> says nothing about truth, God, or the philosophy of Determinism.

No, "chaos theory" relates observations of systems to the mathematics of
nonlinear amplification.  I used to design and build Analog to Digital
Converters, a type of circuit prone to a phenomena called metastability. 
That is, if the voltage into the converter is at 11.5000, it can take a
very long time to resolve to 11 or to 12.  Yes, thermal noise can change
the resolution path, as can quantum effects, but the result is still that
you better not do anything irrevocable (like converting the result to
1011 or 1100) with the information before it has had sufficent time to
resolve.  If you try to prematurely resolve the result, you may end up
with 1000 or 1111 or 1001 or 1110 out of the converter (a "sparkle code"
if you are converting video).  I used chaos theory to prove to colleagues
that this or that design of theirs did not bypass the problem.  (BTW,
the trick is to measure with something called "gray code" and convert
to binary after pipelining it through multiple stages of non-destructive
XOR logic - which Tektronix turned into US Patent 5,459,466 six years
after I left [and couldn't prevent] ).  Sure, "chaos theory" can be
used for metaphysics, but I use it in engineering.  Paint me wierd.

> But, if you _insist_ on haggling on this point, the mathematics of
> chaos theory IS deterministic.  Chaos theory supports a deterministic
> (but unpredictable... with the calculus, anyway) view of dynamical
> systems.  So, chaos theory argues that even if the universe were a
> huge system of partial differential equations, for example, it might
> be unpredictable but it is definitely deterministic.

I guess we differ on the definition of deterministic, then.  I connect
it to the verb "determine", which implies an actor and two subjects
(a predictive mental construct, and observations to describe).  You 
seem to be implying a weaker definition like "mathematically describable"
or "partitionable result set".  The problem is that the human actor turns
predictions into other outcomes (that rock is falling, I'd better move!)
with cascading effects that are very difficult to model.  People
"determine", then they act, then their actions have consequences,
usually mingling with the outcome predicted.  The result is a furball of
events that render many human predictive models useless.  Maybe God's
models are rendered useless, too (I'm cheating, today is an "atheist" day).

Freedom increases the unpredictability - more actors can make more
models and affect more stuff.  That is why determinism survives as
a concept - it permits the insecure to shackle others without feeling
too bad about it.  Observe the purveyors of of determinism, and the
things they demand of others, and you will find this is a pretty good
theory (not perfectly determinative, of course).  The determinists
want to increase predictability because they are not very adaptable,
and want to increase their relative advantage at the expense of others. 
Straightforward monkey politics.  Can't prove it, but it works for me.

I am not much of a metaphysician (take two hypotheticals and call me
in the morning).  I observe things IN the universe, and make models
from them - in that same universe.  A metaphysician leaves out the
observations, while the mystic leaves out the model.  Some folks
leave out both - we call them politicians.

One thing I can say for sure is that the appearance of the human mind
made the world  become quite unpredictable - we can make better theories
about the last billion years than we can about the next hundred.  That
observation alone should render highly suspect any mechanistic models
of human mentation.

> [grin] Shame on you for wielding these weapons so irresponsibly!  I
> charge you to slay a calf at the alter of G?del (make sure you burn
> it... you know how paranoid he is) and iterate 12 fractals in the
> temple of Lorenz!

Howzabout I sleigh the calf at the altar?  I don't kill calves, but
I will use them to get around!  And I am more attracted to Lorentz
than Lorenz, though you have to c 1 to know 1.  

glen, it is always a pleasure.  Between the two of us, we make a
whole wit ...

Keith

PS:  Tomorrow is a theist day.  I can talk more about God then...

-- 
Keith Lofstrom          keithl at keithl.com         Voice (503)-520-1993
KLIC --- Keith Lofstrom Integrated Circuits --- "Your Ideas in Silicon"
Design Contracting in Bipolar and CMOS - Analog, Digital, and Scan ICs