Logical conclusionbs

[From Bill Powers (950716.1805 MDT)]

Bruce Abbott (950716.1635 EST) --

     Skinner, Watson, and other early behaviorists believed that such
     questions would ultimately lead one back out into the environment,
     and in a sense they were right. You wouldn't do anything,
     according to PCT, if all your references were satisifed.

This is a common misconception (Locke, Bandura, et. al. are always
harping on this point). It's simply wrong.

Suppose your reference level is to maintain a comfortable temperature on
your skin. The means available for doing so is a crank that turns a fan.
The rate of turning the crank will increase until the felt temperature
matches the reference temperature. But you can't then stop cranking,
because if you do the fan will stop and the error will go back to where
it was.

You can have a hierarchy in which all the errors are very close to zero,
with complex and even strenuous behavior continually going on to
maintain that state. In systems with true integral output functions, the
error can actually become zero; the integrator will hold the output
signal. But there are no perfect integrators, so the best we can say is
that errors will remain small -- but will never be exactly,
mathematically, zero. "Zero error" is a _quantitative state_, one that
almost never occurs.

···

----------------------------------
     Whether a food pellet acts as a reinforcer would depend on a
     variety of variables, including whether food was already available
     in the situation, the animal's state of deprivation, the taste of
     the pellet, and so on. The ability to reinforce a response would
     not be considered an absolute property of the food pellet but a
     relationship between the properties of the food pellet (taste,
     size, nutritional content, etc.) and the animal's biology. To
     claim that food pellets have some absolute ability to reinforce
     responding would be analogous to claiming that water has some
     absolute ability to power a steam locomotive. Water certainly can
     power the locomotive, but first you have to put it in the boiler,
     heat it, and allow its pressure to build to a certain level. Then
     you have to allow it to do work against a piston.

An animal will try to get a piece of food only if it doesn't have it and
if it wants it. Why make things any more complicated than that? The food
pellets have no power to make the organism do anything. The ability to
reinforce a response is imaginary; it is not a property of the food
pellet. Food pellets contain organic materials, minerals, filler, and so
forth -- and that is all. The only effects they can have on an organism
are mechanical and biochemical. A complete analysis of a food pellet
will reveal nothing reinforcing about it. Food has no "abilities."

If you keep this up, I'm going to start calling you an animist again.
------------------------------------------
     I didn't say they were zero-error conditions. I said that the
     _nutrient_ control system rarely experiences error. Animals such
     as cows usually maintain a steady supply of food moving through the
     gut--the nutrient level experienced in the bloodstream and tissues
     of these animals rarely varies.

That is only because it's under active control. Active control is error-
driven. Always.

     So, there must be other controlled variables at work to keep the
     animal eating more-or-less all day. Controlling those variables
     normally prevents any error from appearing in the nutrient-control
     system.

No. It is the error in the nutrient-control systems that keeps the cow
eating and looking for more grass. If the cow were ever satiated, it
would stop eating. As metabolism and activity drain the stores of
nutrients, the small errors that are always present drive the behavior
that keeps the nutrients resupplied, from internal storage or from
active behavior, or both. All of this, at every level, is error-driven.

What you have to understand is that _small_ errors can lead to _large_
behaviors. That is what keeps the errors small. There is no such thing
as zero error unless external disturbances momentarily supply all of
what is needed.
---------------------------------
     Well, if you want to provide a general explanation but have no data
     to guide that explanation in a specific case, the best you can do
     is to speculate.

In that case it would be better just to say "I can't explain it."

     But real data can be offered to support the general explanation,
     that certain histories of reinforcement or alternative sources of
     reinforcement can have effects such as those to which I refer.

But you don't know the histories. You know only a few scattered
incidents which strike you as important, and then only for specific
organisms. In specific cases you can say, perhaps, that there was a
specific source of reinforcement that was at work -- but you can't then
say that in all similar cases, the same source of reinforcement must
have been responsible, or even that in the present it is the only
reinforcement that matters.

The problem with explaining behavior in terms of specific past
influences is that most of the time you will have to guess about what
past influences were present and you can never know what they all were.
In PCT this isn't necessary. What matters in PCT is the _present_
organization, which doesn't depend on how it got that way.

     If a friend of mine tried to keep a cursor on a target in this
     video game we were playing, but he couldn't do it, what would you
     say? (I give you no other information.)

I would say "What is it he's having trouble with? Is he trying to
control the right variable? Is he able to move the control handle
freely? Is the handle working? Is he aware of the relationship between
the handle and the cursor? Does he want to do this task? How is his
eyesight?" I would watch how he reacts to disturbances, and try to
estimate the speed and amplitude of his actions in relationship to
disturbances. Given time and cooperation, I would set up special
experiments to test various hypotheses. In the end, I would have
estimates of his control parameters, and from them I could draw some
conclusions. If the loop gain turned out to be very low, I might guess
that he isn't really trying -- errors don't matter enough to him. If he
was oscillating, I would suspect some kind of organic problem. Whatever
the reason for the poor performance, it is a _present-time_ reason.

     [Would you guess that] ... control theory must be inadequate to
     explain tracking behavior? I think not.

That ALWAYS remains a possibility. If you ever start discounting that
possibility, you will start doing bad science. All the tests I would
apply amount to Tests of the controlled variable. If the Test is always
failed, I can only conclude that this person isn't controlling any
variable, and that therefore control theory can't be applied to this
behavior.

... how many of these authors have speculated
that perhaps reinforcement does NOT maintain behavior ...

     That's an interesting question, but I dare say that the problem is
     not unique to EAB. Science is practiced by human beings, who have
     learned to perceive their experiments and the results they generate
     from a particular perspective. How many scientists questioned
     Newton's work? It took better than 300 years for Einstein to come
     along and show that Newton's conception of the universe was
     fundamentally wrong.

I'm not interested in bad science. Bad science comes from adopting a
point of view and then defending it no matter what. People who want to
be good scientists more than they want to be right do not do that sort
of thing.

I preach a different kind of discipline to PCTers: never defend control
theory. Attack it. Every experiment must be designed so that if there is
no control going on, the Test will be failed and the experiment won't
work. An experiment without a disturbance is no experiment. An
experiment in which the states of the output, the disturbance, and the
controlled variable are not all recorded is no experiment. If someone
throws doubt on an experiment or a conclusion, lay out the conditions,
provide the means for doing the experiment, describe the reasoning
behind the conclusions, and say "Show me where I went wrong." There's no
glory to be had in defending a wrong theory. A right theory doesn't need
to be defended because it will work every time. So there is no point in
defensiveness. If there's something wrong with PCT, I want to be the
first to know, not the last.

     It isn't necessary to know the complete "history of reinforcement,"
     not if you can show that certain elements of that history are of
     overriding importance.

But if you don't know the complete history, how can you know that a
certain element is more important than something you haven't
investigated? If you know that one element was important to one organism
in one experiment, how do you know it will continue to be important in
the future, to that organism or others? Everything that's important to
your explanation must be tested in _every_ experiment with _every_
individual. You can think that one thing is important when it's really
two other things that matter, with the supposedly important thing being
only a side-effect or a coincidence.

If you can analyze the _present_ organization of a system, there's no
need to guess at historical influences. They simply don't matter.

     For example, you aren't going to get steady responding on a VI 2-hr
     schedule of reinforcement if that's the schedule you start with.
     However, you may be able to do so if you begin with VI 30-s and
     work your way up. To explain stable responding on the VI 2-hr
     schedule, you must appeal to the effects of those previous
     exposures to the higher-rate schedules.

But if you could come up with a model that would behave that way, you
wouldn't need to memorize that rule. And the fact that an animal does
respond on a VI 2-hr schedule doesn't prove that it worked up from a VI
30-sec response. All you've done is describe one history that could lead
to the final result. There are probably many others that would lead to
the same result. That proves that something beside the history is at
work. It is also possible that some animals would never get to the Vi 2-
hr level no matter what path you tried to lead them through. There is
always _some_ history that leads to the present state of affairs. But
there is no way to prove that it is the ONLY history that the present
state of affairs could have had, or that repeating the history would
lead to the same result.

If you found the ball in a pinball machine in the 5000 slot, and kept a
video of the whole play, you could trace the history of the ball and
show how at every moment its path was determined by its interactions
with the bumpers and flippers and the tilts created by the player. So
you could point to the exact historical events that led the ball to be
where it is. Do you think that this historical record would do you any
good as an explanation the next time you found the ball in the 5000
slot?
----------------------------------------------------------------------
Best,

Bill P.