It is not clear (from this example) that the jumping out *has* to be
done by non-computable means. This is a presumption.
> Seen in this light, computers are always followers,
> never leaders... computers mimic something that
> was already there...never originate.
I think you are confusing computers and designed computer programs
here. Designed computer programs (i.e. ones that formalise a
specific intension) do always follow (at least on one level). This
does not mean that any program would, or that it might "follow" on
one level but not on another (the program might be designed to
produce other non-designed programs). To say that they never
originate, is not true, for example, proof search programs have come
up with what would be normally called highly original proofs.
> One could argue "what about genetic algorithms?"
> These algorithms originate on compuers by random
> alteration and a seletive pressure. But in the
> case of genetic algorithms, there is a single
> meta system in place that provides a context for
> the whole process to take place. There really is
> no jumping out of a system.
In a sense, this is true. There is always a sufficiently "high"
meta-level context, from which a system is viewed, such that (in that
description) it can't jump out of that level. The same is true of us
humans. This does not mean that there are not sufficiently powerful
meta-levels, such that the jumping-out done by the lower level, can
not be significant. After all, just because we can't jump out of a
matter/energy context does not stop *us* jumping contexts in card
games.
If the meta-level is lower than this (as in the above Genetic
algorithm example above), then jumping *is* possible. The genetic
algorithm could be evolving machine-code programs and it could be
programmed such that if a "child" program were successful enough, it
would pass control on to it to "develop" more "children" (possibly
in a different way), etc...
> Also consider that the two people playing cards
> are able to reach a consensus on the trick
> without utilizing randome choices or selection
> pressures on future generations...
Yes, but how do they do this? And why couldn't a similar mechanism
be used by a computer program (designed or otherwise)?
----------------------------------------------------------
Bruce Edmonds
Centre for Policy Modelling,
Manchester Metropolitan University, Aytoun Building,
Aytoun Street, Manchester, M1 3GH. UK.
Tel: +44 161 247 6479 Fax: +44 161 247 6802
http://bruce.edmonds.name/bme_home.html