[From: Bruce Nevin (Mon 920713 08:03:56)]
(Bill Powers (920710.1330) ) --
The hypothesis: differences in the environment of an organism that make
a difference within the organism (error, but especially intrinsic error
and conflict resulting in chronic error) must also make differences that
make a difference in the environments of the organism's constituent
cells (error in the intra-cellular control system). The actions of
cells to reduce intra-cellular error must (in part?) amount to the
supra-cellular changes that we call reorganization, and perhaps also
some forms of learning. As you put it:
The individual (neural) cells that constitute an ECS are themselves
independent living entities. . . . they know nothing of the larger
system of which they are the components. The variables for which they
control are only those that they can sense. The actions they employ for
control are those that affect the same variables. Disturbances that alter
the controlled variables are opposed by the actions of the [cell's] system.
Nonetheless, as a byproduct of their autonomous self-control in an
environment comprising other cells and their byproducts, the cells
together do in fact constitute higher-order control systems of which
they can have no ken (because they lack the perceptual means).
It occurs to me that this
approach is an attempt to deduce the existence of a higher order of control
systems -- the neural heirarachy -- by referring only to the reference
signals and control systems inside the cellular components of the larger
system. Can we get there from here? I'm thinking of neurotaxis, which seems
to be a phenomenon of a level higher than the cell. Can we express
neurotaxis in terms of reference signals and controlled variables inside
the cell itself? Could a cell that needs negative feedback from somewhere
emit a chemical signal from the spot where it's needed? Could a cell with
surplus neurotransmitter to unload grow itself toward a spot that wants and
can accept that kind of transmitter?
I am not trying to *deduce* the existence of a higher order of control
systems in the case of cells and neural control systems. They are
observational givens (within the theory). (Perhaps you are here looking
ahead to deducing suprapersonal control systems, by analogy? We'll get
to that below.) My question at this point is how supracellular control
systems can come to be, using only the means that cells have at their
disposal.
Taxis in general is construed as the movement of an organism toward a
stimulus. We reject the explanatory framework presupposed in the word
"stimulus," of course. There is some disturbance to a controlled
variable within the cell, such that the observationally perceived
behavioral output called taxis counteracts the internal effect of that
disturbance.
Could a cell that needs negative feedback from somewhere
emit a chemical signal from the spot where it's needed? Could a cell with
surplus neurotransmitter to unload grow itself toward a spot that wants and
can accept that kind of transmitter?
A "need for negative feedback" can be observed only from the point of
view of the supracellular control system (actually, from a point "above"
that even). How might such a need manifest as a disturbance to one or
more cells in a control system? And which cell or cells?
We can see how complex interdependencies of cells can come to exist over
evolutionary time. By establishing symbiotic relationships cells
mutually experience the advantage of each providing a more stable and
more easily controlled environment for the others. Observationally,
"above" the cell's point of view, we see specialization in terms of the
morphology and behavioral outputs of different kinds of cells.
As you suggest, a waste product of a cell's error-correcting behavioral
outputs may serve as what we observationally perceive as a signal or as
a neurotransmitter. Possible analogy: the particular smell of scat may
signal to a predator that one of the deer in a herd is sick and ripe for
culling, or the smell of testosterone-laced urine warns of territorial
limits.
I think there's a limit to how far we can go with this kind of emergence --
maybe.
Absent an antecedently given teleology, this has to be the route for
explanation of ontogeny. For phylogeny, symbiotic interdependencies
(giving the benefit of a more stable and predictable environment for each
cell) provide supracellular scaffolding that was not present for
evolutionary forebears, but explanation still must take the point of
view of the cell, not of the control system that it participates in
constituting. (Social institutions, customs, traditions, etc. provide
suprapersonal scaffolding that was not present for evolutionary
forebears, cf. Bruner's Language Acquisition Support System. But I'm
jumping ahead.)
The behavioral hierarchy gets most of its negative feedback through
the external world, where physical phenomena foreign to the body get into
the loop. And the effects of controlling for different external (that is,
sensory) variables in different ways are important to the body in places
remote from the controlling systems: in the stomach, the bloodstream, the
gonads, and so on. Something has to link these remote effects back to the
very organization of the behavioral systems (that's what my reorganizing
system is supposed to do). Could these remote effects get into the loop in
any meaningful way at the level of a single cell trying to maintain itself?
Somehow I think not: the effects of a single neural signal on the external
world, outside the body, would be lost in the general effects from all the
nerve-cells that participate in behavior. We're talking, I think, about a
much smaller-scale environment, including only a small volume around the
nerve-cell.
You have identified the problem. I think you have also identified the
solution to it, long since: the control hierarchy. Variables in the
environment of the organism are far beyond the immediate environment of
a cell in the neural control system. But the neural control system is
so structured as to bring news of a difference in the organism's
environment (a difference that might make a difference to the organism)
from cell to cell, each making a difference that makes a difference to
its neighbor (each time just in immediate cellular environment of its
neighbor), until there is a difference in the environment of the
particular cell that we have singled out for observation. Setting aside
for the moment the question of how this marvellously articulated
ramification of neural bucket brigades came into being (ontogeny and
phylogeny), its existence explains how a difference in the environment
of an organism can be transformed (through the control hierarchy) into a
difference in the environment of any given cell in the hierarchy.
ยทยทยท
==========================
You have understood and articulated the social analogy very clearly, and
you have also stated a reason (I suspect) you have been reluctant to
broach it in other discussions:
I feel as if the bottom has dropped out and the
ceiling has been removed and I don't know whether to fall or fly.
There is no need to do either. Nothing has changed, this is just where
you always were. More to the point, the same choice to limit your focus
to the proper purview of control theory is still available. The only
difference is your acknowledgement that it is a choice and not a
preconditional Reality.
Unsolicited homily #37: Limitation is the first step of any creative
process. You define the scope of the work. You draw your magic circle,
and you ignore everything outside it. But as any magician worth his
salt knows, you don't just forget about what lies outside the ordered
realm of the work, and you don't live there inside the circle. From
time to time you dissolve the circle, take a lunch break, whatever.
When you re-form the circle, you may well bring something in that wasn't
there before, and throw out some baggage that turned out just to be in
the way. All familiar process, by small trial-and-error steps. I'm not
advocating that you expand your magic circle too far, beyond your (our)
means to control. I'm just suggesting that you take your lunch breaks
in interesting places. And that you not be so troubled by others
drawing intersecting circles. In general you don't. There are just
some strong commitments respecting social control that sometimes
blindside you.
What this does first is to give license to say "I don't know" about
whether there are suprapersonal control systems or not. And to be
comfortable with that, since (on the proposed analogy) it is only by
virtue of individual human "cells" controlling for just what matters
most to each of them that such higher-order systems can be constituted.
The intuitive grasp which I am seeking to articulate is that the pursuit
and realization of one's (evolving) heart's desire turns out to be one's
way of helping to constitute a healthy, well-controlling higher-order
control system (though one inherently cannot control for constituting
such a system per se--as with the cell constituting the ECS, it is
beyond one's perceptual means), and that participation in one's
particular capacity in such a control system turns out to be the most
personally fulfilling thing one can find to do (all by trial and error,
of course). This is related to Ruth Benedict's ideas on synergy (the
post I sent to CSG-l shortly after coming on board a little more than a
year ago).
Of course it is possible for people to control for reference perceptions
other than those that (by trial and error) they find most fulfilling.
They can define personal achievement in terms of ability to deny
fulfilment to others "under" them, for example. There's a lot of that
going around. I understand that the precursors of symbiosis are
competitive relations destructive to the rivals for the same niche and
parasitic relations destructive to the host. How might such changes
come about, from the cell's point of view? Our shared animosity for
abuses and abusers of social relations and institutions must be
separated from our advocacy of control theory as a science. Control
theory (or an understanding and acceptance of it) does not preclude such
abuse, alas. Indeed, a conviction that it does could, with appropriate
missionary mind-set, support rationalization and outright ignoring of
one's participation in such abuses, as human beings have demonstrated
again and again, with remarkable creativity and imagination. (Al
Capone, Dale Carnegie tells us, thought of himself as a benefactor to
humanity. Dale Carnegie did too.)
The purview of control theory is limited to the same purview as one has
for conflict resolution within the control hierarchy. What you can see
is what you can "get," so to speak. Beyond that we can make suggestive
analogies upward and downward. How does it all work from the point of
view of a cell? Of a molecule? What does the process of cells evolving
to a control system potentially tell us about our ongoing social
evolution?
Unlike the cell, and to a much greater extent than other animals (so far
aw we know), we have limited capacity to extrapolate beyond our
immediate perceptual means. From what we can determine from observation
of cells and molecules (of their behavioral outputs constituting
structures and systems), and of control systems on the scale of animals
and humans, what analogies are there to groups, cultures, ideologies,
etc., and are those analogies useful?
Between levels of control there is a control relationship. Between
orders of control (e.g. cellular vs ECS) there is a constitutive
relationship. The limits of our direct perception are related to the
constitutive bounds of our perceptual control hierarchy. By using
imagination and analogy to interpret "meter readings" of various kinds,
we can extend our understanding lower in the constitutive hierarchy--and
perhaps higher as well. (Science is imagination and analogy
systematized.) It seems to me implausible that the constitutive
hierarchy of control systems has its upper bound the order of control
evolved by multicellular organisms such as humans. I am prepared to
entertain the possibility of persons (or person-like organisms) whose
"cellular" control participates in constituting a higher order of
"supra-cellular" control. We can't test that notion, but it may be
important context for devising and interpreting tests that are within
our grasp.
Things to think about when control theorists are out to lunch.
Bruce
bn@bbn.com