Autonomy; Words and perceptions

[From Bill Powers (960420.0300 MDT)]

Fred Nickols [960419.1700] -- (replying to Rick Marken)

Rick described four theories of "non-autonomous" behavior:

1. Control by consequences (reinforcement)
2. Control by input (stimulus-response)
3. Control of output (cognitive psychology, psychodynamics, eg. Freud)
4. Control by equilibration processes (chaos, non-linear dynamics)

     I think I have a working grasp of the first three in Rick's list
     and I confess to knowing absolutely nothing about the fourth. That
     said, I don't see how the first three (even the S-R model) are
     inconsistent with a view of people as "autonomous."

You're going to have to define what you mean by "autonomous." To me, the
term means _not_ controlled by anything in the environment. Actually, I
could class 3 as a theory of autonomous behavior, in that it proposes
that behavior is generated by internal processes.

     (As an aside, I don't think the pure S-R model, that is, the notion
     of a stimulus eliciting a completely unmediated response, has any
     currency anywhere in today's world.)

From my point of view, S-R theories and S-O-R theories are basically the

same: they propose that inputs produce outputs. It took me a long time
to realize that "pure" S-R theories were actually proposing that stimuli
cause responses WITHOUT mediation -- to me, a mind-boggling notion. But
it's just as mind-boggling to me to think that excitation of sensory
endings would be sufficient to create the complex behaviors that we see.

     I don't see how all that stuff is automatically mutually exclusive.
     Could it not be the case, for instance, that the consequences of
     our behavior (which we know chiefly through our perceptions of them
     and our perceptions of their linkages to our behavior) might have a
     hand in shaping the referents we seek to control?

The key question is "HOW?" I don't see how physical happenings in the
environment can do all the things to the nervous system of an organism
that would be required to produce organized actions, or to change it
from producing one organized action to producing another. With regard to
reinforcers, what is there in a pellet of food that could alter the
nervous system in such a way as to make a rat lift a paw and press down
on a lever 150 times in a row to get the next pellet?

It has always seemed to me that people who propose such effects are
ignoring the questions of mechanisms and means. I have never seen a
plausible explanation for _how_ a complex stimulus could produce a
complex response, or _how_ a reinforcer could alter the probability that
a particular response will be emitted again. All of the (rare) attempts
to model such processes that I have seen end up proposing models of the
organism, not of its environment. When you start trying to explain HOW
these apparent effects are created, you end up talking about
capabilities of organisms, not effects of their environments.

To put it simply, if organisms didn't want food, and didn't contain
mechanisms for altering their own organization until food is obtained,
there would be no such thing as a reinforcer.

     Maybe the problem is that I'm just old.

Unless you, will be more than 70 at the end of this summer, I say
"Nonsense, sonny." Otherwise, I will say "Yeah, that must be the


Avery Andrews 960420 --

I've read your WWW posting and found it fascinating.

     `Inferred' vs. `Direct' perception is a hot potato in psychology
     (or used to be at least), and I also think it is content-free, so
     I'm trying to describe it in observationally meaningful terms like
     `intermittent', etc.

You probably should make it clear that you're not tackling an
epistemological problem, but an engineering problem. The engineering
approach is necessarily naive-realist; you assume that the environment
is the same for everyone, and that you can know both what is inside and
outside the brain.

     ... certainly the catogories associated with words, such as `tall',
     underspecify the corresponding perceptions.

Words in general underspecify perceptual experience, not so? I'm still
considering a demo in which one person verbally guides another to create
some sort of result on a computer screen; the guide can't see the
screen. I think this demo could show very dramatically what it is about
experience that words always leave out.
Best to all,

Bill P.