*really* good questions

[From Rick Marken (920922.1000)]

eric harnden (920922.0940) asks some VERY GOOD questions:

1) mr marken, in his letter to estes, reasserts the essential argument
that study should be directed not toward behavioral outputs but toward
controlled inputs.

We're all friends here on the net -- you can call me Dr. Marken (just
kidding -- please call me Rick).

now, i was
under the impression that behavioral studies often involved the presentation
of stimulus to an organism, and an observation of response.


i was further
under the impression that these stimuli were presented under a not entirely
unfounded set of assumptions of what, in the local terminology, the organism
is controlling for.

Not formally. This is certainly true in operant conditioning studies where
the researchers "know" that the organism is controlling for some aspect of
reinforcement. Of course they wouldn't say it that way. But they sometimes
break down and say things that imply that they know that their exper-
imental manipulations will work only if the animal WANTS the food, water,
or whatever is the "reinforcer". The problem is that they don't really
know what "wanting" is (it is a reference for a particular level of a
perceptual variable) so they don't know how to study it. Instead, they
are content to study the noticable side effects of achieving wanted
perceptual states -- the disturbance resistance process. They never think
of trying to really nail down WHAT perceptual variable(s) is (are) being
controlled. For a good example of the process of testing for controlled
perceptual variables in an operant conditioning situation, see Powers
article in Behavioral Science (I think in 1971). You will see in that
article that it is possible to be far more precise about a controlled
variable than, say, "wants to avoid pain". Bill shows quantitatively how it
is possible to tell that the rat in a "shock avoidance" experiment
is controlling (as best it can) the interval between shocks (as I recall).
He also showed that this variable (interval between shocks) is a
better representation of the controlled variable than at least one
reasonable alternative -- the probability of a shock. So "the test"
is a process of honing in on the best representation of a controlled
variable -- where "best" means a representation that allows you to predict
EXACTLY how the organism will react to disturbances of that variable.

please explain to me why it is invalid for a mainstream
behaviorist to say 'yes, well of course the organism is going to avoid pain,
and prefer food. in this supposedly new jargon, i am being told what i
already know: that it is controlling for lack of pain and hunger.

Because the behavioral psychologist does NOT already know what is being
controlled without doing the test. PCT is not just a change of jargon;
it is a working model of behavior. It says that behavior is organized
around the control of PERCEPTUAL VARIABLES. An observer cannot see what
perceptual variables are being controlled just by looking. Well, perhaps
the observer can get some COARSE idea -- like "avoid pain". But that's too
vague for PCT modelling. As Bill's Behavioral Science paper (also
reported in the last chapter of his book, BCP) shows, it is possible to
get quite precise about exactly what perceptual variables are being controlled.
The "coin game" described in BCP is also a good illustration of the how to
of get a good idea of what perceptual variable(s) another organism (in
this case, a person) is controlling -- and how difficult this can be. I also
report a study (in my Mind Reading collection) where I show that it
is possible to tell that a person is controlling the area rather than the
perimeter of a quadrangular shape. The determination of controlled
perceptual variables is not necessarily easy to do -- there is a lot of
skill and creativity involved. And it has not been done AT ALL in
conventional psychological research (actually, I think Piaget got close
to doing a version of the test).

Once you know what is being controlled, you can understand how the organism
wiii act in MANY different situation -- because it MUST act certain ways
if it is to keep the perceptual variable under control. But the fundemental
thing to understand about organisms, according to PCT, is what they are
controlling -- what kinds of perceptual variables they control. Such a
catalogue of controlled variables is not yet available in the literature
of conventional psychology; this is what we need in PCT. Developing
such a catalogue is one piece of work that we PCT researchers can
"get on with" -- once we stop kvetching about conventional psychology.

interests me is the characterization of its outputs, since it is the nature
of these that ultimately affects itself, its environment,and other organisms.

Yes, this is the interest of conventional psychologists. But such studies
reveal nothing much about the nature of the organism that produces such
outputs. The relationship between inputs (usually disturbances) and outputs
will always be puzzling, different and usually statistical because the
outputs that are observed depend on 1) other influences that are also
acting on the controlled variable 2) secular variations within and
between organisms in the level at which they intend to keep a controlled
perception and 3) differences across organisms in the perecptual variable that
is actually controlled. These three factors make it CERTAIN that the
results of the conventional methodology (which looks at arbitrary measures
of action -- arbitrary with respect to the perceptual variables that are
actually controlled) will reveal nothing about what the organism is actually
doing (controlling certain perceptual variables). The outputs may be fun to
watch but they are truly superficial -- and will always be puzzlingly
different across organisms, situations and times.

Best regards




Richard S. Marken USMail: 10459 Holman Ave
The Aerospace Corporation Los Angeles, CA 90024
E-mail: marken@aero.org
(310) 336-6214 (day)
(310) 474-0313 (evening)