Bill & Gary on Autonomy

From Greg Williams (920524)

Bill Powers (920523.0800)

It doesn't matter how the universe got into its present condition
-- by what paths the current variables came to be in their present states.
Only the states of the variables NOW, and the laws connecting them, matter
in determining their next states. And when the next states occur, there is
a new NOW. All real phenomena and interactions occur NOW. That is what
differential equations are all about.

It matters to the folks who have pitched the idea of what I've been referring
to as "traditional" free will (which is where this thread started). They claim
that an individual has a "transcendental self" (their term) which can make
"free deliberate choices," where "free" means completely independent of
anything besides the "self" and "deliberate" means "not-random." So models of
how behavior/action happens which don't include such things -- such as what
I've termed "unaugmented" PCT models -- provide no basis for a belief in such
things. For the "traditional" free will believers, it is important that at
least some of an individual's current behavior NOT be subject to influence of
EITHER past OR present events outside the "self," otherwise the "self" would
not able to be absolutely autonomous. In PCT models, rules (in some cases,
perhaps with "random" components) are postulated connecting the present state
with the previous state (differential equations, for time-continuous models),
and there are also rules connecting the present form of the governing
differential equations with previous forms (with reorganization, the
differential equations are non-time-stationary). Further, all of those rules,
in general, are non-autonomous; that is, they have forcing functions (think of
these as boundary conditions which change with time) set by the niche, not by
the "self" (even if it is claimed that a whole PCT model is a model of a
"self"). The trajectory of states describing the evolution of behavior/action
through time will be different, in general, if the forcing functions are

PCT models imply that an individual's current control structure is the result
of past interactions between the (changing-over-time) control structure and
its niche, and NOT the result SOLELY of an autonomous dynamics of the control
structure (which would mean that when reorganization stopped, in any
particular case, would be independent of the niche, denying the very reason
for reorganization). And PCT models imply that the current behavior/action is
the "automatic" result of the current control structure and the current
forcing functions. Again, I agree with you: PCT models say that at any moment,
the environment (via forcing functions) influences action, but cannot
influence behavior (that is, which goals are set); PCT models say that at any
moment, the existing control structure sets the goals and thereby influences
action; and PCT models say that the existing control structure is the result
of the past interactions between the evolving control structure and the niche.

For me, the question isn't whether miracles could occur, by anyone's
definition, but what is the best explanation of what we observe? As long as
you can't put on a convincing demonstration of your explanation, you
haven't ruled out anything, even magic. There's no need to settle once and
for all on any one mode of explanation, although with experience you will
probably try certain modes of explanation before others.

Right on. My only complaint is with people who claim that PCT supports
"traditional" free will -- absolute autonomy. They can believe what they want,
but I, for one, won't let them get away with claiming to support that belief
with a model which doesn't support it.

It's perfectly possible that there are entities which exist in space and
time, but are purely causal in nature: that is, they are generators of
phenomena, but are not generated by phenomena external to themselves.

Sure it's possible. But no such entities are to be found in PCT models.
(Although PCT models COULD be augmented, as I have said before, to include
such entities.)

On the other hand, such entities may not exist at all. Basically, who
cares? We're a long way from needing that answer now. Looking for short-
cuts to ultimate answers has never done science much good.

Again, right on. In all this, I've been attempting to head-off such shortcuts.
It is tempting, but premature, to claim support for religion in scientific

Gary Cziko

I realize now that the environment (and the other organisms in it)
DO determine certain perceptions, but only the LOWER ones in the hierarchy.
If the living control system is successful, the environment will not
determine what the PERCEPTIONS THAT COUNT will be (those higher in the

I don't understand this claim. As I understand the HPCT model, at any given
moment the "relevant" or "interactive" environment (what I have been calling
the "niche") cannot SOLELY determine ANY perceptions in the hierarchy --
rather, the perceptions at ALL levels are determined (possibly with a random
component -- sorry Bill; maybe "are set" sounds better?) by the niche AND the
current structure of the hierarchy. How could it be any different at the top
of the hierarchy? Surely you aren't saying that the top level is a solipsist
and compares its reference signals (about system concepts?) with COMPLETELY
MADE-UP perceptual signals. Bill mentioned physiological homeostasis variables
("intrinsic reference signals"); are you alluding to these as somehow "not
determined" by the environment? If so, you should be claiming that the
REFERENCE SIGNALS, not the PERCEPTIONS, are "not determined." BTW, intrinsic
reference signals ARE determined (sorry again, Bill; set!) evolutionarily
by organism-environment interactions (not, of course by the environment
ALONE). And intrinsic perceptions (i.e., of core body temperature?) are set
from moment to moment by the current hierarchy and the niche, acting