[From Rick Marken (940404.1545)]
Ok. I couldn't resist. I was overcome by a brief wave of civility
so I posted the following reply to the sci.cognitive post from
James Speer:
···
-----
In article <1994Mar28.164926.1@ccsvax.sfasu.edu> f_speerjr@ccsvax.sfasu.edu
(James Speer) writes:
In article <EXS.94Mar24234617@emotsun.bham.ac.uk>, exs@emotsun.bham.ac.uk
(Edmund Shing) writes:
Could someone please give me a brief summary of William Powers' Perceptual
Control Theory - I am about to get hold of his 1973 book "Behavior: The
Control of Perception", but wondered if I could get hold of a concise
summary first to enable me to see if it is relevant to my research interests
I did not see Mr. Shing's question about Powers' Perceptual Control Theory
(PCT) except here, in Mr. Speer's reply. But I will take the liberty of
answering Mr. Shing's request for a brief summary of PCT.
PCT is a model of purposeful behavior which, in living organisms, consists
of intentionally produced results of efferent neural activity. An intentional
result is a controlled result -- one which is maintained, against disturbance,
in a fixed or varying reference state, by the actions of the organism. PCT
is a model of how organisms produce controlled (intentional) results; they
do it by keeping perceptual representations of the results to be controlled
matching internal specifications (references) for those perceptions. Organisms
are always controlling many hierarchically related perceptual variables
simulataneously; in many cases they control perceptions (such as the perception
of a a typed word) by specifying lower order perceptions (such as the per-
ceptios of the letters that make up the word). At all levels of behavioral
organization, the PCT model produces behavior by specifying what is to be
perceived, not what is to be done. In general, this is how purposeful behavior
is generated by the PCT model -- behavior is the control of perception (which
is a coin of the title of Powers' first book on PCT).
The PCT model performs purposeful mental behavior (imagining, remembering,
planning, etc) in the same way that it performs "regular" purposeful behavior
except that, in mental behavior, the connection between the organism's
efferent activity and the perceptual consequences of that activity is
"short-circuited"; efferent activity that usually influences perception
via the environment now influences the same perceptions by direct neural
connection. This removes the actual contraints of physical law so we are able
to produce perceptual results in imagination (such as lifting an elephant
with one hand) that could not be produced in fact.
PCT redefines the goals of a science of living systems; rather than looking
for the variables that influence behavior (as in traditinal behavioral psych-
ology) or for the mental processes that underly the production of behavior (as
in conventional cognitive psychology), PCT suggests that the basic goal of a
science of living systems is to determine the perceptual variables that
organisms control, how they control them and why.
From what I have been told, his theory seems to bear relations to Gibson's
(1979) concept of "affordances", Ballard's (1991) animate vision paradigm
and Heidegger's concept of "Dasein" (cf. Chapman, 1990).
If you're looking for CLOSE relations to the foregoing theories / ideas, I
think you're in for a disappointment. Powers' view, I have always thought, is
insufficiently constrained to count as a theory. A "meta-theory" or research
program, perhaps.
I think Mr. Speer is quite right on the first point; Powers' model may _seem_
similar to many other theories -- but it is really essentially different.
It is differnet in terms of the phenomenon it explains (purposeful behavior,
also known as "control"), in the way it works (by controlling input, not
by causing, planning or "constraining" output) and in how it is tested (a
method called "the test for controlled variables" instead of standard
independent/dependent variable based statistical techniques).
PCT is not a meta-theory; it is working model of purposeful behavior. There
are many interesting computer implementations of the model and there is
a substantial set of research results which support the basic predictions
of the model to within 1% error.
Because PCT is so differnet from conventional approaches to understanding
living systems, there are only a few people who are actually doing research
and modelling based on PCT. There is a PCT discussion group on InterNet and
William T. Powers himself is a member; so there's no need to poke though
those dusty archives; if you're interested in the relevance of PCT to
what you are studying, just ask Bill.
I believe that you can subscribe to the PCT discussion group (called
csg-l for "control systems group - list) by sending your name
and address after the command "subscribe" to
listserv@vmd.cso.uiuc.edu
or send a message to the list god (Gary Cziko) at g-cziko@uiuc.edu
and he'll tell you what to do.
Best
Rick Marken