Posted by
glen e. p. ropella-2 on
Oct 07, 2009; 7:00pm
URL: http://friam.383.s1.nabble.com/EMERGENCE-SEMINAR-V-Dennett-et-al-WAS-emergence-seminar-what-s-next-tp3771294p3783816.html
Thus spake Nicholas Thompson circa 09-10-06 11:39 PM:
> In fact, it is not clear to me that Rosen's Life Itself was
> not an attempt to create that very formalization. Have you ever looked at
> Rosen????
I know you were talking to Owen (am I hijacking the thread, here?); but
I'd like to say that I _do_ think Rosen's work on _complexity_ is a
start towards the ability to create complex (computational) formalisms
-- where Rosen's claim is that all current (computational) formalisms
are _simple_ because of the way we define and implement them.
The trouble with Rosen's work and its extensions is that, in order to
construct such formalisms, we _must_ include construction loops. And
when we include construction loops in computational systems, we get
ambiguity (multivalence... multiple, equally correct, answers to the
same question). In the most strict situations, the ambiguity is
realized as things like "deadlocks" (where multiple blocking processes
are waiting for the same resource) and (I speculate) race conditions
(where multiple concurrent processes race to see which will get its way
in the end).
So, while we can build these formalisms, they are unsatisfying to the
little engineering homunculus in our heads because they violate a sacred
requirement: they don't reduce to a single outcome. No SANE computer
scientist would want to build an ambiguous computing device. Right?!?!
;-) Or perhaps I should say no sane computer _engineer_ would want
to... By all rights, a computer SCIENTIST would love to create such
things and study them.
I qualify "formalism" with "computational" because we do have
non-algorithm, mathematical, philosophical, and logical formalisms that
express complexity in this sense. But they require us to toss out the
axiom of regularity (which says that sets can't be members of
themselves). This makes any computation we formulate in such a wacko
formalism open to running forever (infinite regress, race conditions,
deadlocks, etc.) or coming up with multiple different results. Also
note that all the standard computer programming languages are "turing
complete", which, according to Rosen's work, means any program written
in them will be a _simple_... not complex ... system.
In any case, sorry for the distraction. It's not at all clear how these
formalisms relate to "emergence" UNLESS we define emergence as a measure
of complexity.
--
glen e. p. ropella, 971-222-9095,
http://agent-based-modeling.com============================================================
FRIAM Applied Complexity Group listserv
Meets Fridays 9a-11:30 at cafe at St. John's College
lectures, archives, unsubscribe, maps at
http://www.friam.org