[From Bill Powers (920713.1200)]
Leaving for Denver/Boulder tomorrow (Tuesday 14th) morning -- back
Saturday.
Bruce Nevin (920713.0803) --
RE: individual and social hierarchies
In some forms of hierarchy theory, a hierarchy consists of a group of
elements seen individually, in small groups, in groups of groups, and so
on. The "levels" in such a hierarchy don't introduce any new individuals;
they are more like the result of a single observer taking successively more
abstract points of view toward the one set of individuals, in effect re-
perceiving the original individuals in different ways.
In the PCT hierarchy as I have conceived it, new levels consist of new
control systems; they aren't simply the same bottom-level systems looked at
from farther and farther away. If a level of control is added, it is added
EXPLICITLY, as input, comparison, and output functions physically distinct
from the components of already-existing systems at lower levels.
This principle of explicitness distinguishes, I think, purely conceptual
models from models intended to represent a physical system. In a conceptual
model, you could say that a collection of intensity signals is conceptually
equivalent to a sensation perception, that the set of conceptually-defined
sensations is equivalent to a conceptually defined configuration level, and
so on. After all, the sensation is implicit in the collection of
intensities; the configuration is implicit in the collection of sensations,
and so on for as many levels as you like. However many levels you add in
this conceptual way, however, in the physical system there is still only
the original set of elements; nothing has in fact been added.
In a model intended to be physical in nature, however, nothing that is
merely implicit can have any effect. A collection of control systems
controlling individual intensity signals will behave exactly the same way
whether an observer attends to the individual systems or to the
conceptually implicit control of sensations. In a physically-oriented
model, there will be no control of sensations until some neural function
receives a set of intensity signals and creates a new signal explicitly
dependent on that set, according to some functional form. There will be no
control until that explicit perceptual signal is compared with a reference
signal, and the difference is routed to the reference inputs of some of the
sensation-controlling systems. All those functions and signals must
physically exist, distinct from the systems of the lowest level, before
sensation control can become explicit, and not just represent a viewpoint
of the observer.
Frank Rosenblatt elicidated this principle a long time ago, when he
insisted that for any variable pertinent to behavior to have effects in a
real system, it must be embodied as an explicit signal. This is my basis
for saying that anything experienced has to exist as a neural signal, a
perceptual signal. It's not enough that something COULD be perceived in a
collection of elements; it must BE perceived, exist as an explicit signal,
before the rest of the brain can do anything with it. That's why my system
uses explicit signal paths even to represent imagined information. Without
this principle of explicit representation, there would be no hope of
committing any model to hardware.
So much for the preamble; now to your post.
... 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).
If these cells constitute higher-order systems ONLY IMPLICITLY, by virtue
of the way we conceptualize their interactions, then there is no physical
higher-order system. The higher-order system exists only as a conceptual
level in the mind of the observer, and is not actually part of the physical
cells or cell assemblies. It is not part of the organization of the real
system.
In order for an actual higher-order system to exist, some cells must take
on roles that the others do not; they must become concerned with an
environment consisting of the other cells, and act on that same
environment, explicitly sensing something about the other cells, explicitly
acting on something that alters what is sensed about the other cells. In
the brain, the configuration level of control consists of neurons in the
midbrain physically distinct from those in the brainstem and spinal systems
that carry out lower levels of control. The configuration system senses
signals that are the perceptual signals of sensation systems; its outputs
go not to the spinal systems, but to the sensation-control systems, as
reference signals. This is an explicit new level of control that is
physically distinct from and does something different from the systems of
lower order.
So the mere existence of a collection of control systems of a given level,
and the mere fact that they interact with each other in the course of their
control actions, can't by itself result in a new level of control. There
can be no higher levels of variables until some set of cells explicitly
computes the higher variables; there can be no higher level of control
action until some set of cells generates an explicit error signal that
reaches specific members of the lower-order group of systems. New cells
with new specializations must appear in the proper relationship to the old
population of cells before control at a new level can explicitly exist.
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.
But I claim that we cannot yet see how levels of dependency come into
being. A special kind of specialization is needed. Valentin Turchin
characterizes the required kind of development as a "metasystem
transition," which is very different from a mere proliferation of systems
at an existing level. This required specialization removes some cells from
the population that previously existed, and puts them into a superordinate
position, so now the old population is in their environment; they no longer
share the same environment with the older systems of cells. A metasystem
transition produces a physically new population of cells with different
functions in the whole system: the functions we associate with higher
levels of control.
We can see, of course, what is gained from such a metasystem transition,
but we haven't a clue as to what is different about that kind of
specialization.
So you can see that I must reject your extension of this analysis to social
systems:
...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 analogy to a metasystem transition within the cells of an organism
would be the appearance of a new kind of organism which senses the
condition of other organisms, compares that with the condition it wishes to
perceive, and acts by setting the highest reference signals of other
organisms. I know of no such superordinate organisms. Some people claim
that they do, and call them gods, or God: creatures so superior to human
individuals that they control for variables of kinds inconceivable to
mortals, and who act by injecting reference signals into the highest levels
of human consciousness, for purposes beyond human ken.
Perhaps such individuals exist. I wouldn't know, and neither would anyone
else who isn't one of them. Whatever the case, it's certain that there are
no social control systems consituted by the mere control behaviors and
interactions of ORDINARY people -- and until dramatic evidence to the
contrary appears, I will assume that all people now on earth are quite
ordinary human beings. You can make up any stories to the contrary that you
like, but they won't be germane to the point we have been disputing.
To sum up: your argument seems to depend on the emergence of higher levels
from populations of systems of an existing level, without the addition of
any new kinds of physical systems. I would claim that you are relying on
IMPLICIT organization to create new levels. I, on the other hand, argue
that the new levels must be EXPLICIT.
ยทยทยท
--------------------------------------------------------------------
Best,
Bill P.