hiebeler@turing.cs.rpi.edu (Dave Hiebeler) (10/08/90)
I apologize if this has been mentioned already; I missed the very
beginning of this discussion on emergent phenomena.
I've been working with the Center for Nonlinear Studies, and the
Complex Systems Group in the Theoretical Division at Los Alamos Lab
for the past year and a half or so. There is certainly some
interesting work going on there. Among other things, they had a
conference on "Emergent Computation" last summer.
Stephanie Forrest (of CNLS) organized it, and wrote up an
introduction for the proceedings. Although the conference was about
emergent *computation* rather than emergent phenomena in general, the
people following this discussion might want to check out the
proceedings, which appeared as Physica D, volume 42 (1990), Nos. 1-3.
It will also appear as a book ("Emergent Computation", edited by S.
Forrest, MIT Press).
Anyways, in her Intro, Steph lists the following as a definition of
an emergent computation (there is no claim that this the best
definition, of course):
(i) A collection of agents, each following explicit instructions.
(ii) Interactions among the agents (according to the instructions),
which form implicit global patterns at the macroscopic level,
i.e. epiphenomena.
(iii) A natural interpretation of the epiphenomena as computations.
Later on, she states "Thus, the concept of emergent computation cannot
contribute magical computational properties. Rather, we are
advocating a way of thinking about the design of computational systems
that could potentially lead to radically different architectures which
are more robust and efficient than current designs."
I interpret this as a somewhat Sophistic view implying that it's not
particularly important that we have a strict definition of emergence,
since the concept of emergent phenomena is simply a convenient way to
describe certain systems or behaviors; i.e. each of us has our own
"truth" about what is emergent. I agree with it to some extent, since
I feel that any definition we come up with will be inadequate in some
way.
There are several very interesting articles in the conference
proceedings; I recommend checking out a copy if possible.
--
Dave Hiebeler | Internet: hiebeler@turing.cs.rpi.edu
Computer Science Dept., Amos Eaton Bldg.| hiebeler@heretic.lanl.gov
RPI | Bitnet: userF3JL@rpitsmts
Troy, NY 12180-3590 |