[From Bill Powers (920522.1100)]
Greg Williams (920521) --
My point in bringing up the history-determination of current
behavior/actions of PCT circuits was to counter claims that PCT >supports a
notion of strict individual autonomy.
Let's go into the meanings we might give to "autonomy." I use the term in a
way that has a rather complicated meaning with a basis farther back than
current behavioral interactions with the current environment. I'll slip in
my hypotheses about the role of control without marking them; I'm sure
others will be able to tell what is hypothesis from what is generally-
accepted "fact" in this story.
Start with DNA. While the surface appearance is that genetic
characteristics are transmitted via the DNA molecule, in fact a lot more
passes from generation to generation than just DNA. Much cellular material,
as in mitochondria, is passed along with the DNA through the mother's egg;
in the lowest orders, the cellular material simply divvies up during the
reproductive divisions. Immediately after a new individual is launched, the
DNA is in an environment that is continuous with the previous environment,
at least locally.
So the biochemical control systems whose reference signals are carried in
DNA can operate right across the boundary between generations. These
control systems, finding themselves isolated, begin again building the
control systems that build the control systems that build the control
systems that constitute the adult organism. The entire milieu interieur
(sp?) is regenerated, with whatever changes occurred during the division of
genetic material. The continuity proceeds, as people have long suspected,
through the mother.
One of the final products of this process is the a set of intrinsic
reference signals. These reference signals are the basis of reorganization
or learning through which the new organism establishes control in the
environment it first and subsequently encounters. The intrinsic reference
signals represent the target states of some as yet poorly defined set of
variables critical to the survival of the individual. There is no reason to
think that the reference signals are identically set from one person to the
next, or even that they are all of the same kind. Each individual differs
in details of organization at all levels from DNA through cellular through
organ-structures through gross bodily structure through neural circuitry.
And the mix of intrinsic reference signals will differ from individual to
individual.
Intrinsic reference signals are part of a system, probably distributed
rather than lumped in one place, that controls for zero intrinsic error.
The means of control is blind variation of the organization of the nervous
system and the biochemical control systems. Reorganization is driven by
intrinsic error, and ceases when intrinsic error drops below some
threshold.
As a result, the organism acquires control systems that can maintain
perceived aspects of the external world at learned reference levels by
means of motor behavior (and at the biochemical level, changes in such
things as strength, speed, organ size and activity, and so on). The
criterion for acquiring any behavioral control system, and for setting its
reference signal to any specific value, is that intrinsic error be
maintained at the lowest possible level.
Thus the overriding concern of the reorganizing system, and the purpose for
which it causes any behavioral organization to appear, is to control its
own basic physical state; to maintain its component variables at
endogenously-determined reference levels. It neither knows about nor cares
about nor CAN care about any processes external to the body. Everything it
causes to be done by way of interacting with an external world is done for
the purpose of controlling an internal state. It is therefore completely
and absolutely autonomous in its purposes.
It does, to be sure, have a history. But this is not so much a history of
antecedent events as it is a history of gradually changing organization.
The reorganizing system of one generation is continuous with the
reorganizing systems of previous generations: it is the same system,
evolving. At the center of this system are reference signals that have not
changed in billions of years, having survived even speciation.
Reorganizations that preserve these basic reference signals have led to the
development of instrumental reference signals and associated control
systems, and those have led to still more elaborate control mechanisms, and
so on to the various physical forms that life has ultimately adopted -- as
a means of preserving the fundamental function, which is to control. And to
control is the ultimate meaning of being autonomous.
If the criterion for stopping reorganization is bringing intrinsic
variables to their respective reference levels, it follows that only those
behavioral control systems will survive reorganization that do entail
actual effects of the right kinds on the intrinsic variables.
The effect of any given behavioral act is not determined by the organism:
it is determined by the nature of the surrounding world (including the
behavioral organization of other organisms in that world). So
reorganization can't cease until the actual effects on intrinsic states,
via that external world, are correct for maintaining zero intrinsic error.
Thus the organism learns first what variables are critical to perceive in
that external world, and second what specific states of those variables are
critical to maintain. This process of learning has been going on through
geological time, with the appearance of control structures of greater and
greater generality, and what we recognize as higher and higher levels of
control. As each new level of control appeared, new and more important
aspects of the environment became perceivable and came under control by the
organism. The actions of the organism adapted themselves to the environment
in more and more subtle ways.
The means of action did not change nearly as much as the neural control
systems that use actions to control ever more complex variables. A human
being and a monkey share nearly identical means of motor action. Both have
hands at the ends of jointed limbs; but the human being can accomplish
things with its hands that a monkey cannot. This is not because of having
an opposable thumb, but because of having higher levels of control. Human
beings can do more even with their thumbs cut off than a monkey can do with
ten digits.
So we arrive finally at the question of autonomy in the individual human
being. Autonomy is clearly not freedom from physical constraints (which
include, in the final analysis, social constraints). The environment, not
the organism, dictates the effects of any given action. But the environment
does not dictate the desired consequences of any action. It is the organism
that chooses those consequences, and learns how it must act in order to
produce them.
In a hierarchical control system, built, I presume, level by level over the
eons and recapitulated in the individual, the lower systems give up their
autonomy to the higher systems that manipulate their reference signals. At
whatever level is currently the highest, the reference signals are set from
within the organism by the process of reorganization; the purpose of
choosing a particular setting is to maintain intrinsic error as close to
zero as possible -- as the purpose has always been. In order to bring the
highest level of perception into a match with this autonomously-set
reference signal, the highest control systems must, as usual, be altered to
produce actions which are among those that will have the required effects.
Now those actions are determined by properties of the existing lower levels
as well as by the characteristics of the world external to the organism.
The organism can't choose what properties the external world will have, no
matter what the level of perception. Once its lower levels have been built
and brought into mutual harmony, the organism has less than a completely
free choice even as to the kinds of actions it can produce (without
starting again from scratch, which is probably no longer possible in the
adult organism, in the time remaining to it). So the particular behavioral
organizations that appear in the adult are shaped by the properties of the
world around it and by the properties of its own already-acquired lower
levels of control.
However, the highest levels of reference signal remain autonomous and are
changed only in service of maintaining the individual organism's mix of
intrinsic variables at their unique mix of reference settings. The external
world has no influence over that basic requirement. Intrinsic error remains
the organism's sole criterion for judging the value of any aspect of its
experiences. This is true of all organisms from the amoeba to the human
being.
If the highest levels of reference signals are autonomously determined,
then the next-to-highest levels of reference signals are varied so as to
prevent the environment (as perceived through all the lower levels) from
making the highest perceptions depart materially from their reference
settings. This means that the next-to-highest levels of perception will
also be shaped to meet the requirements of the highest reference signals.
But the next-to-highest reference signals will be determined by what the
environment requires, for the highest perceptual signals in general contain
effects of uncontrolled elements. To make the net result match what the
highest system requires, the reference signals for the controllable parts
of the next-to-highest world must be varied, and those variations much be
matched to the properties of the lower systems and the external world. The
organism can't choose the settings freely, because only certain settings
will result in the required perceptions. There may be many alternative
settings that will produce the required perception, but there is freedom to
choose only among those alternatives, given that the highest reference
signal is to be satisfied. All other alternatives are ruled out by
properties of the external world.
The general picture is that the environment determines behavior, while the
autonomous organism determines consequences of behavior. Given the intended
consequences, the environment sets the limits as to what lower-level
actions can in fact bring those consequences about.
So we can see where autonomy begins and ends. It is the organism that
selects consequences that keep its intrinsic errors as close to zero as
possible. It is the environment -- and other organisms in it -- that
determines what actions must be produced in order that those consequences
be brought about and maintained. The external world sets the stage on which
existence is played out. But the reorganizing system writes the play.
And even the reorganizing system is just the product of a deeper control
process, at the core of which lies a tiny and unimaginably ancient spark
of purpose that makes life different from everything else.
···
---------------------------------------------
Best,
Bill P.
