=46rancis
--------------------
Date: Fri, 27 Dec 1996 14:03:05 -0500
X-Sender: jearls@macareo.pucp.edu.pe
To: Multiple recipients of list PRNCYB-L <PRNCYB-L@BINGVMB.BITNET>,
Francis Heylighen <fheyligh@VNET3.VUB.AC.BE>
=46rom: jearls@pucp.edu.pe (John Earls)
Status: RO
[...]
At the moment I am very interested in applying the theory of neural
networks to social systems, in particular to the Andean socio-cultural
systems. I have mentioned before that these systems are particularly
interesting, in part because they incorporate information "gathering"
(perceiving), "processing" (filtering), transducing (into digital code),
synoptic re-presentation to all of the 500 -2,000 members of a community,
and from all this construct consensual decisions. All within a few days or
less. In thinking about the PCP nettalk on superbrain and the MST, + the
fragments of documents on memes and related topics, I am evermore convinced
that these systems (at least for the last 1,500 years, from archaeological
data) are operating as massively parallel metasystems, even without the Si-C
interface to emerge in the coming MST. My point is that we just can't afford
to sit back and work on (or wait for) the Si-C synthesis to resolve many of
the basic social problems that are prerequisites for "least hassle"
integration at a higher level. Many basic problems, such as that described,
can in principle be resolved at a purely social level; the Andean blend of
hierarchical and heterarchical organization (at least at the demographic
levels indicated) in the world's highest variety (/ unit area) physical
environment, demonstrates that there are social technologies now existing
which are quite compatible with the global metasystem's organisational
requirements, and should be taken seriously. What I want to do is to find an
application of the Hebbs rule (i.e. the necessary algorythms -- a Hopefield
connection?) which I can start trying out to find "fits" (without being
forced) with the body of ethnological, ethnohistorical and archaeological
information now exixtent. I have the Hertz, Krogh, & Palmer "Introduction to
the theory of neural computation" but find the level of the book too high
for me for comfortable practical use. I am hoping that there exist PCP nodes
which contain more elementary descriptions of connectionist models, and
something on how to use them. ~ 20-25 years ago I worked at trying to apply
the McCulloch-Pitts model and von Foerster's reworkings of it to the Andean
organisation, but I was then focusing on the more conventional
anthropological lineage and alliance structures, and they didn't fit to
well. But now with much rethinking of the data, and new data, I think I can
do it. Or at least, to formally state the problem for people with more
pertinent scientific backgrounds who may be interested.
A final point: I have talked informally with Andean Indian peasants on the
idea of the Superbrain and Gaia (the Quechua word is "Pachamama"), and the
reaction is usually of great eagerness to get into it (i.e. the C-Si
interface), especially as they quickly understand how it will greatly
improve their own organisation. They know what "organisation" is are always
attent to forms of improving what they have -- Western peck-order like
managerial hierarchies are usually now laughed at, since whenever they have
been implemented have produced factionalism and the breakdown of the
recursive coordinations of coordinations which maintain the systems'
identity, and so provoke a rapid return to the Andean (cybernetic)
organising principles.
Well here's to a better 1997
John Earls
jearls@pucp.edu.pe
Pontificia Universidad Cat=F3lica del Per=FA
________________________________________________________________________
Dr. Francis Heylighen, Systems Researcher fheyligh@vnet3.vub.ac.be
PESP, Free University of Brussels, Pleinlaan 2, B-1050 Brussels, Belgium
Tel +32-2-6292525; Fax +32-2-6292489; http://pespmc1.vub.ac.be/HEYL.html