Re: action of models - indirect effects

["Tim Gwinn" <***>]

Dan,

Thanks for those quotes. See interposed.

Regards,
Tim

> -----Original Message-----
> From: ROSEN Forum [mailto:*** Behalf Of Dan
> Fiscus
> Sent: Thursday, July 17, 2003 9:41 AM
> To: ***
> Subject: Re: action of models - indirect effects
>
>
> Tim,
>
> Re: your points...
>
> Tim Gwinn wrote:
>
> > The quotes from Patten are enlightening, especially the second
> one: it runs
> > contrary to a commonplace view that direct effects take precedence, and
> > indirect effects have diminishing importance. This would also mean that
> > stability, as well as alteration, of the biosphere would depend
> heavily on
> > such indirect effects. Did he comment any on if (or not) the
> net effect of
> > indirect influences tended generically toward or away from stability?
>
> He does say in a few places that he sees the general tendency of networks
> to be positive, which one could take to include stabilizing. He wrote
> (in that
> same chapter):
>
> "...the kind of biological interaction observed at a proximate level may
> become a qualitatively different type at a higher level of
> organization when
> all the consequences of context are taken into account. This principle of
> holism is another manifestation of hierarchical network organization...a
> tendency for locally negative interactions to become globally positive
> ones."
> p 298.
>
> "The present analysis supports the view of a nature dominated by "holistic
> positivism" (his in italics), and organized in consequence to provide the
> greatest good for the greatest number of individual lives playing
> out their
> collective existences together...To acknowledge the good fortune for life
> under network rule, I suggest that the still to be understood mechanism of
> this emergence might be referred to as a kind of 'hedonic
> property' (his in
> italics and single quotes) of ecological networks." p340.
>
> Here we see that despite the good ideas Patten still referes to
> mechanisms,
> even a mechanism of emergence - a new one for the oxymoron list!


Hmm, I doubt he meant 'mechanism' in the Rosennean sense. Even Rosen
sometimes used the word 'mechanism' in the more general sense of 'causal
process or agent'. But, I also wonder in what sense he means "emergence",
since surely these were computationally derived results, and therefore all
the result of algorithm. In that case, they might be the surprising results
of complication, but not complexity or emergence in the Rosennean sense.

The result of "greatest good for the greatest number of individual lives
playing out their collective existences together" is an encouraging one.
However, did his models take into account finite resources? I ask this
because I am not sure if "greatest number of individual lives" means 'no
upper limit' or 'within a certain maximum limit'.

This also interests me because I have been thinking alot about your earlier
comments, Dan, that a long-range goal for ecology is "open-ended evolution".
Of course, this can't mean maximizing all populations, so sheer fecundity
can't be a sub-goal. Instead, it seems that seeking "viability" is a more
apt criteria for optimal open-ended evolution. This may also tie into the
idea that Judith presented that perhaps all species are, in some sense,
parasitic on each other. If so, then maximizing fecundity is clearly
inappropriate, and optimizing mutual long-term viability could be thought of
as some kind of 'mutually inhibited parasitism'. But what does it mean to
talk of 'viability' instead of fecundity? At the least it seems to mean some
minimum population size; but perhaps it also means a maximum population
marked by where indirect effects in the ecological network become
destabilizing rather than stabilizing? I don't know...I am just speculating
out loud.

The remark that "interaction observed at a proximate level may become a
qualitatively different type at a higher level of organization" reminds me
very much of Rosen's activation-inhibition networks. [AS, EL ch. 22] Without
going into detail, they are informational networks of a system, comprised of
interlinked differential equations regarding the relationships of
rates-of-change between various observables or state variables that can have
multiple 'information levels' (e.g., rates-of-change of relationships of
rates-of-change, etc.). A profound thing is that, generically, these diffeqs
will be inexact. This means that, generically, they are not differentials of
some underlying single fixed function (and, therefore, the system is
complex). Instead, they may well indicate an underlying path-dependence
among the relationships between observables (or state variables) of the
system under study, which (I would think) opens up a big can of worms.


> > I also wonder about the role of indirect effects within an organism - it
> > seems likely the same kind of concerns would arise.
>
> Here I am reminded of a review article in Nature (vol 420, Nov. 14, 2002)
> called "The Community of the Self" in which Timothy Buchman says in the
> abstract "Physiology and computational biology now suggest that healthy
> dynamic stability arises through combination of specific feedback
> mechanisms and spontaneous properties of interconnected networks." He
> means in organisms and especially humans, and the community of self
> being the interaction networks of cells, organs, sub-systems, etc. He also
> wrote, "It is vital to create models that embed homeostatic
> mechanisms into
> larger networks that themselves confer robustness to perturbation and
> thereby protect the community of the self."


Very interesting! Of course, one can see in that last statement the
potential for infinite regress: what confers robustness on the "larger
network" into which the model is embedded? Another yet-larger-network? I
think this is where anticipatory systems alleviate that, by providing an
alternative to models with "homeostatic mechanisms", such that the "healthy
dynamic stability" is intrinsically determined.


>
> Some other cool stuff in this article too, like a quote from a D.
> Noble who
> speaks of the "ambiguity of causation in complex networks".
>
> Dan