Re: November questions

Jay Lemke (jllbc who-is-at cunyvm.cuny.edu)
Tue, 09 Nov 1999 22:47:21 -0500

It's good to poke around at a theory to see what it will and won't do for
you, what you can and can't figure out how to use it for, as Phil is doing.

The model I presented does NOT tell you what to focus on on scale N. You
have to put that in. It then tells you something about the nature of the
complexity and possible logics of relations across timescales of processes
(N-1, N+1; and beyond that heterochrony, cascading bifurcations, etc.) that
you should pay attention to and look for. It also tells you that if you
have identified two adjacent relatively weakly interacting level about two
orders of magnitude different in timescales, that there are conditions
under which a new intermediate level of organization will emerge in between
them. But it can't tell you what the properties of that level N will be;
they are emergent; that's the point.

I have just sent a posting about patchiness in time and space for
ecosystems and ecosocial systems. That picture in some ways says how I
might address issues like the Great Schism or the reunification of Berlin.
Note that there are two different meanings of 'bifurcation' floating around
here. One is the technical meaning in nonlinear dynamics: a system that
under one range of conditions has only one possible behavior enters a zone
of conditions under which it has two possible modes of behavior, or
develops itself into a new kind of system that has this larger repertory of
options. The other is a more bio-geographical meaning: one population finds
itself distributed over a range of different conditions and develops into
two distinct subpopulations, moreso as the coupling between the subgroups
is weakened. I think Phil's examples are more of the second sort, though
there could be a place for genuine bifurcation if one had a developmental,
rather than an evolutionary model of the origins of cultural or language
diversity. For the relation between developmental and evolutionary modes of
discourse, see Stan Salthe's books (easier than, and maybe more broadly
applicable than comparable accounts in Kauffman). The basic difference is
whether we think in terms of an individual (development, one system
changes) or a population (evolution, distributions change). Bringing in
relations across scales makes for a kind of dialectic between these two
perspectives (see Salthe).

Phil also raises some intriguing, perhaps rhetorical, questions, about the
smallest nail without which kingdoms can be lost. This is the cascade
model, small events trigger bigger ones, that trigger bigger ones, etc. It
does not happen very often, in the sense that it's unlikely that all the
dominoes across scales will be neatly lined up to permit the cascade to
continue; normally adaptive systems are well buffered against most events
at lower scales, neutralizing their differences so they don't make a
difference up the line (equifinality). Where they do happen is in systems
that have no intermediate levels of organization to do the buffering
(supersaturated solutions in crystalization, physical phase changes,
renormalization cascades in quantum systems, etc.) They also sometimes
happen when a lower scale process can cross a scale boundary in magnitude:
cancer -- cell scale to organ-scale; HIV or prions -- macromolecular to
cellular to systemic. Maybe something like this happens with successful
political revolutions, but I doubt it. More likely it's akin to the
supersaturated solution: conditions were ready on the larger scale, they
just needed an initiating event. The dominoes were all lined up. The
initiating event was not the _cause_; it was just the trigger. There is no
simple linear causality across scales in complex self-organizing systems.
What happens happens because of the state of the whole, not because of the
state of any one part.

A good analysis of the kind Phil is looking for, independently convergent
with my own (though less focussed on time and dynamics) is Latour's, say in
_Aramis_, where he looks at the sociotechnical network that did not achieve
closure (no self-sustaining complex of interdependent elements and
processes; no auto-catalysis at N+1) in the case of a proposed new
transport system in Paris. He basically shows how not just the absence of
some bits, but the absence of a complete network of connections among the
bits, doomed the project. Mature sociotechncal networks are also
well-buffered; if one element is withdrawn, the rest of the system will
adapt, find a substitute, etc.

The issue of mass is of course important. I am not saying that only
timescales matter; just that they are a more reliable starting point in
general than spatial scale or mass-energy scale or even information-scale.
There are good general reasons in physics, and in the history of the
lineages of complex systems, for why small is fast and big is slower. I
would be willing to venture as a first approximation (though not the whole
story) that basically a system can normally only have a major effect on
another system at its own scale. Individuals do not make or destroy
organizations or nations. If you want to influence an organization, you
need another organization to do so; preferably somewhat smaller and faster.
Ideally a subnetwork within the organization. This is not really news,
except for another nail in the Great Man myth's coffin. N-1 assaults are
too small and don't last coherently over long enough timescales. Of course
there are apparent exceptions, and we hear entirely too much propaganda
about these anomalous cases; but they are all like the dominoes -- level N
was ready to change, the N-1 event or individual was just a trigger. In
some cases of real bifurcations, yes, one event or person can make the
difference between the N-level system going one way or a different way.
That's history. But we do not make it just as we please. We cannot as
individuals (or small groups) set up the conditions that allow such
'butterfly' effects to happen. Nor can we predict the longterm consequences
of the N-1 scale event. Our actions have consequences, but they do not have
control over the future on scale N+1 (much less N+k).

As a general suggestion to would-be revolutionaries: breaking taboos and
making connections (couplings) where they have not been made before and are
not supposed to be made is a really promising strategy. Nothing makes a
larger scale system change its behavior radically more than altering its
coupling scheme, how its parts interact with one another. Maybe just a few
strategic re-wirings here and there could lead to major change -- but still
this would be unpredictable change. What is needed beyond a critical mass
and a good (lucky?) re-wiring strategy is longterm persistence and just the
right degree of weak coupling to the rest of the system. All those paranoid
fantasies about Hidden Masters (Eco's Templars) who guide human history by
maintaing their organizational coherence across generations (even with
evolving and emergent changes in their goals) speak a genuine insight, I
think. The fantasies do fall prey to Elite Group versions of the Great Man
fallacy; you cannot make up in time for lack of mass. You need a lot more
participants than the fantasies imagine. Time will not compensate because
unpredictability, including within the organization (its schisms for
example, leading to cross-purposes that neutralize each other), accumulates
too fast.

There is of course also another kind of answer to Phil' s question about
minimal-mass methods for social change. Take away all the paper (prior to
electronic memory) and the scale of organization collapses radically. Take
away the semiotic artifacts that mediate heterochrony. But while each one
may be small, in aggregate there's quite a bit of mass (though still small
compared to the systems they enable to persist). For other recipes for the
collapse of social systems, see various accounts of what happened to the
Incas, Aztecs, West African kingdoms. (A file-overwriting computer virus
and an antibiotic-resistant bacterium that eats paper, quickly, would do
the trick today.)

All power to the Griots!

JAY.

---------------------------
JAY L. LEMKE
PROFESSOR OF EDUCATION
CITY UNIVERSITY OF NEW YORK
JLLBC who-is-at CUNYVM.CUNY.EDU
<http://academic.brooklyn.cuny.edu/education/jlemke/index.htm>
---------------------------