From newshub.ccs.yorku.ca!torn!cs.utexas.edu!usc!sdd.hp.com!uakari.primate.wisc.edu!ames!olivea!bu.edu!buphy0.bu.edu!leao Mon Aug 24 15:41:36 EDT 1992
Article 6675 of comp.ai.philosophy:
Path: newshub.ccs.yorku.ca!torn!cs.utexas.edu!usc!sdd.hp.com!uakari.primate.wisc.edu!ames!olivea!bu.edu!buphy0.bu.edu!leao
>From: leao@buphy0.bu.edu (Joao Leao)
Newsgroups: comp.ai.philosophy
Subject: Re: Freewill, chaos and digital systems
Message-ID: <94238@bu.edu>
Date: 21 Aug 92 14:35:55 GMT
References: <Bt4xt1.MA0.1@cs.cmu.edu> <1992Aug19.210204.29868@mp.cs.niu.edu> <Bt9Kq2.CLy.1@cs.cmu.edu>
Sender: news@bu.edu
Lines: 104

In article <Bt9Kq2.CLy.1@cs.cmu.edu>, rudis+@cs.cmu.edu (Rujith S DeSilva) writes:
|> In article <1992Aug19.210204.29868@mp.cs.niu.edu> rickert@mp.cs.niu.edu (Neil
|> Rickert) writes:
|> >In article <Bt4xt1.MA0.1@cs.cmu.edu> rudis+@cs.cmu.edu (Rujith S DeSilva)
|> (that's me!) writes: 
|> >>(1) Does freewill arise solely through the mechanism of chaos?
|> >
|> >Usually we think of free will as the ability to make a decision, and
|> >have that decision affect what we do in the future.  That is, we make a
|> >choice, and can stick to that choice.
|> 
|> I prefer a more operational definition.  Let me suggest the following
|> different flavours of freewill.  I assume that the brain/mind/soul obeys the
|> `usual' physical laws.
|> 
|> (a) All your `decisions' can be predicted in advance, given sufficiently
|> advanced equipment.  In this scenario, I would say that one does not have
|> freewill.  Note that being able to predict some or most of your decisions is
|> insufficient.
|> 

I think you mean initial information, not equipment. In any case this is also
arguable because the amount of information would be relevant for the prediction
in case such prediction could be made (i.e. we had a description of the dynamical law relevant for the prediction to be made). In physical terms
both of these (initial conditions and dynamical evolution laws ) are 
important for determinism. Even if the two can be specified completely
you may run into serious problems if you assume, as you should that the
nervous system is non-linear because of sensitivity to initial conditions.

|> (b) All your decisions can be predicted, given sufficient time on the advanced
|> equipment.  All of them can no longer be predicted in advance.  I'm not sure
|> whether this is significantly different from (a).
|>

Unclear again. If it takes more time to procure the initial conditions and
compute the results than the time it takes for the deterministic evolution
to take place it becomes unusable as a prediction because you cannot verify
it. 
 
|> (c) Your decisions are pre-determined, but they are not predictable.  I think
|> chaos has to be invoked here.
|> 

This again may mimic free-will but I guess most people would not accept
it as such. Incidentally exertion of free will has little similarity with
the kind of phase maps you see associated with chaotic systems. It looks
more like bifurcation behaviour...No?

|> (d) Your decisions are not pre-determined.  I think current quantum theory
|> (specifically, the absence of `hidden variables') allows this possibility, but
|> would like an expert opinion.
|> 

As far as QM is concerned there is a residual form of indeterminism 
associated with quantum correlations as evidenced by the photon pair
correlation experiments which verified the violation of Bell's
inequalities. In principle these results just invalidate the type
of determinism called "local". As some people (Carl Brans and John
Bell himself) have pointed out this would still allow a "demiurge"
type of determinism where everything could still be set globally
at some initial time so that all observed correlated events live
within one mega-light-cone (the demiurge). This is hardly a decent
framework for determinism ("fatalism" is a better word in this case)
for it would leave little rationale for the occurence of local
natural laws, for which we have pretty good evidence. 

Bells's inequalities apply only to statistical hidden variable models
(not the deterministic variety), by the way.
The type of "no-go argument" for hidden variables presented reccently
by Greeenberger, Horn and Zeilinger and further developed by Mermin
is more damaging for the cause of determinism because it shows that
Quantum Mechanics can give predictions with probability 1 that are in
conflict with deterministic hidden variable model predictions.

To my knowledge no one has pondered the effects of this new argument to 
the "demiurge hypotheses" but I don't see much hope for such a scenario
anyway (unless we want to go back to medieval arguments...)

|> Have I left out any possibilities?  What do you believe is the true situation?
|> Personally, I don't have a clue, but I would not like it to be (a).  Actually,
|> (d) is not too bad.  It could be that most of our decisions are based on our
|> nature/nurture, with only the borderline cases being resolved randomly.
|> 
|> >Well, yes, your decision is determined in advance, but no, free will is not a
|> >sham, and life is not a charade.
|> 
|> In this case, as a previous poster pointed out, are we morally responsible for
|> our actions?  I think the concept of rehabilitation makes sense, but
|> punishment does not, except as a deterrent.  Actually, none of the four
|> flavours above supports the idea of punishment.
|> 
|> Rujith de Silva.
|> Carnegie Mellon.

I strongly doubt one could find any reasonable exemption from moral
responsability on ones' own actions from either Physics or Computer 
Science, for that matter. Don't you?

-- 
Joao Pedro Leao (Artificial Iconoclast and Director of Computer Resources
Artificial Physics Lab * Boston University - Physics Dept. Boston MA 02215)
 	leao@buphy.bu.edu | leao@buphyc.bitnet | BUPHYC::LEAO 
"Well I am sitting here in Tahiti/ I am laying in the sun and sipping a...
...chartreuse tropical drink!/ and I say: I know those Bermuda shorts!..."


