What's the reinforcer?

[From Bruce Abbott (970910.1525 EST)]

Bill Powers (970910.0138 MDT) --

Bruce Abbott (970909.1315 EST)

In those cases, so was the appearance that behavior rate was related to
reinforcement rate an artifact.

The same experiment showed that the rate at which the rats responded on the
lever was low during initial training (before they had learned the
relationship between lever pressing and pellet delivery) and high
thereafter. Although we didn't do it, we easily could have shown that the
observed increase depended materially on the fact that pressing the lever
causes food pellet delivery. An increase in response rate that results from
making an event contingent on the response is called reinforcement; we
certainly observed that result.

Allow me to offer a different description.

What you offer is not a "different" description in the sense of
contradicting the one I offered, but a more detailed description, consistent
with mine but covering a longer span of time. Comments on specific parts
follow.

Before the rats were able efficiently to generate the action would produce
the food, they produced the relevant action only in passing.

"Only in passing" is an inference, not a description. What is observed is
that the _act_ that would produce the food (pressing the lever down to the
point of switch-closure) occurs very infrequently and by inconsistent means.

Then, as their
behavior patterns changed, their actions came to produce the food more and
more often, until the actions were producing food at a limiting rate.

Yes.

From
then on, the rate of food delivery was maintained by the maximum rate of
action that the rats could sustain.

Yes. I suggested that rate was limited in exactly this way a bit farther on
in my post. But not that this is theory again; you and I are _inferring_
that the rate of lever-pressing is limited by the rat's capabilities. We
did not observe that this is so.

Your description matches mine: behavior (lever-pressing) that was observed
infrequently at the start of the session is now occurring frequently. We
can demonstrate that the contingency between lever-pressing and food
delivery is crucial to this result, everything else equal. If we break the
contingency, the frequency with which the rat presses the lever soon
declines. If we then reestablish the contingency, lever-pressing returns to
its previously observed rate.

When much greater amounts of food are produced by a given behavior rate,
the rats do not change their behavior rate appreciably.

What do you mean, "greater amounts of food"? Don't you mean, "a greater
rate of food delivery"? A "greater amount" could mean a larger pellet.

They behave in a
stereotyped way, producing and eating food as quickly as possible until a
certain amount of food has been ingested, producing shorter and shorter
bursts of behavior as the normal meal size is approached and then doing
other things for a hour or more. Then another session of eating and
ingesting takes place (if possible) until a similar meal size is generated,
and so on during a normal day, maintaining the amount of daily intake that
the rats require, or an individual rat requires.

In our weight-control study, we discovered conditions under which the rat
could easily have consumed enough to maintain its body weight, but failed to
do so, until body weight fell to such an extent that I feared the rat would
starve to death, at which point I changed the conditions to those that would
allow the rat's weight to recover. Under the normal conditions the rat
encounters, however, it does behave as you describe, over the long run.
That is, the rat takes in sufficient food to maintain a steady or increasing
body weight.

What we observe is the behavior maintaining the food delivery rate to the
extent physically possible, in the dictionary sense of the transitive verb,
"to maintain." The subject is the rat, the object is the rate of food
delivery, and the means is the production of behavior in such a way as to
produce the food.

Yes. (And I have said nothing previously that would contradict this
description.)

The role of the experimenter is limited to setting up a means by which the
rat's behavior, should the particular act occur, can generate food which
the rat can eat. Since the experimenter is the more powerful of the two
participants in the situation, the experimenter can establish any relation
between any action and the appearance of food, and prevent any other action
from producing food.

This is true in the experiment (after all, an experiment is an arrangement
designed to exert strict control over the relevant variables for the purpose
of determining causation). It is not generally true in natural
circumstances. The fish changed their behavior, not because I am more
powerful than they, but because they learned that they could gain immediate
accesss to a new source of food by so doing.

The rat's requirements for food determine that the rat
will search for and find the key behavior or something sufficiently like it
to generate food, provided that the requirement is not beyond the rat's
ability to discover it, and that the action needed is not beyond the rat's
ability to sustain it.

To paraphrase: The rat's requirements for food determine . . . that it will
acquire behavior whose consequence is the generation of food. These are not
observations, they are either logical truisms (the requirements must fall
within the capabilities of the animal) or matters of theory (requirements
determine...).

If the experimenter changes the contingency, the expected effect with no
change in the rat's pattern of behavior would be an increase or a decrease
in the rate of food delivery.

Yes, depending on how the experimenter changed the contingency. For
example, increasing the ratio on a fixed ratio schedule would decrease the
rate of food delivery if response rate remained the same.

If physically possible, the rat's pattern of
behavior always changes so as to restore the rate of food delivery to its
original level, or as close to it as possible.

Now we again have entered the realm of theory, control theory to be
specific. You expect that increasing the ratio requirement will lead to an
increased rate of responding, so as to compensate (at least partly) for the
decline in rate of food delivery that would occur if response rate did not
change. This of course is not what we observed.

The only exception is when
the rat is already producing as much behavior of the right kind as it can,
and the change of contingency reduces the rate of food delivery. Then we
observe no counteracting change in the rat's behavior, because no more
behavior can be generated.

And here is the theoretical explanation. If reinforcement theory had
required the invocation of a limit on response rate to explain the
observations, and PCT did not, I am quite sure you would now be talking
about this rate-limit idea being just what is needed to save the theory.

What we have here is an animal presented with a problem which it proceeds
to solve to the best of its ability, unaided except for any hints provided
by the experimenter by way of shaping. The experimenter sets the problem by
creating a contingency. The rat solves it by varying its patterns of action
until it is producing as near to the amount of food it needs as it can
under the given circumstances --neither more nor less.

I agree.

This, I submit, is a fairly complete and accurate account of what is
observed, at least as complete and accurate as any behaviorist account of
similar length and detail. All references to agency and direction of
causation are expressed correctly in terms of conventional interpretations
and direct observations; the behavior is produced by the rat, the
requirements for food are the rat's, and the rate of food delivery is
completely dependent on the rat's actions. The experimenter's action
consists entirely of setting up the passive environmental link between some
action and food delivery rate, and this link runs in one direction only,
from action to food delivery.

I submit that it is part observation and part theory. If your goal was to
provide a theory-free description of the facts, you failed.

I suggest that the reason why the rate of pellet delivery turned out not
vary across the different values of the ratio schedule we tested is that the
rats were not controlling for rate of pellet delivery, but rather for access
to food pellets in the cup as soon as the rat desired another pellet for
consumption. Because completing the ratio imposed a delay (owing to the
fact that lever presses require time to execute), the best that the rat
could do was respond as rapidly as possible at all ratio values.

This implies that under other conditions, the rats would press at lower
than maximum rates. I say this is false: the rats are incapable of
producing a systematically variable rhythm of lever pressing. All they can
do is produce a rapid repetitive action on the lever, or cease pressing
altogether. They are either producing this stereotyped action, or they are
doing something else.

You are being mislead by the fact that all our observations were conducted
using ratio schedules. On these schedules, behavior tends to alternate
between responding at a high, relatively constant rate (while completing the
ratio requirement) and pausing (after delivery of the food pellet). On
other schedules the rat (or pigeon) is definitely able to vary its rate of
responding in a fairly smooth way. I have observed rats on a Sidman
avoidance schedule, for example, sitting at the lever and pressing it rather
steadily at a rate of about one press per second -- about 1/5th the rate we
observed on ratio schedules of pellet delivery.

The only reason that "rate of pressing" caught on as
a measure of behavior was that the measure was obtained by dividing total
presses by session duration. As a result, the rat's changing the _kind_ or
_location_ of behavior was indistinguishable from its changing the _rate_
of behavior of a single kind at a single location. In fact, the latter,
which is assumed to occur, does not occur. So it does not need an explanation.

Both sorts of change have been observed, reported and, incidentally,
explained (descriptively) in terms of reinforcement theory. Both require an
explanation.

Reinforcement theory would account for the same constancy in rate by noting
that, on ratio schedules, higher rates of responding yield shorter times to
reinforcement, which in turn would be expected to generate yet higher rates
of responding. This is a positive feedback loop that pushes response rate
up to the maximum, or at least to an equilibrium value established by the
conflicting effects of reinforcement and response effort. (Effort increases
with rate; a response rate is reached where the decrease in delay to
reinforcement is balanced by the increase in effort.

There is no need for this explanation, because rates of responding do not
actually change. If they did change, the explanation might be appropriate,
but as they do not, the explanation is empty of meaning. It explains
something that does not happen.

You lost me there. I explain why response rates are constant across ratios,
and you tell me that there is no need for this explanation, because response
rates are constant across ratios.

I should also point out that whatever we
observe about behavior, it is not "responses." To observe a response one
must also observe the stimulus; otherwise all that one is observing is an
action.

The term "responses" is a holdover from the old S-R days in psychology. As
used in operant studies, it does not imply a prior stimulus. What we
observe are actions, but the one of interest in the analysis are those that
have a common consequence in the environment; they are what Skinner called
"operants."

What I am saying here, Bruce, is that the behaviorist account, which you
present as a simple factual description of observations, is nothing of the
sort. It is a biased account slanted toward encouraging the listener to
conclude that the environment is controlling behavior -- that behavior is
controlled by its consequences, which is exactly the opposite of the truth.

If that is what you are saying, you have failed to make your case. In fact,
your own description failed as a simple factual account of the observations.
Also, I have said nothing about behavior being controlled by its
consequences. I have simply noted that when certain events are made
contingent on certain behaviors, those behaviors increase in frequency.
When the contingency is broken, those behaviors decrease in frequency,
returning more or less to their original levels. When asked "why did the
behavior increase in frequency?", I might answer, "because it was followed
by the contingent event." That is a purely descriptive explanation, which
makes no appeal to mechanism of any sort. Nowhere does this explanation
assert that the environment is controlling behavior -- not in the sense in
which the term "control" is used in PCT. What it asserts is that behavior
changes as a function of establishing (or breaking) certain contingencies
between that behavior and its consequences, and that much is a simple
description of the observations.

All the strange, contorted, backward reasoning in behaviorist accounts, all
the special terminology, all the special definitions of auxiliary terms
that ignore the primary usages of the words, all the insistence on a
particular set of terms in which to express observations -- all this points
in only one direction. It points to a concerted attempt to put the
environment into the causal role and remove purposiveness from organisms.

You haven't demonstrated much of this, Bill. You haven't demonstrated that
behaviorist accounts use strange, contorted, or backward reasoning. You
haven't demonstrated that the terms used ignore the primary usages of the
words. You haven't demonstrated that anything points to a concerted attempt
to do anything.

I don't see how you can deny this: this has been the avowed purpose of
behaviorists like Skinner and his followers from the very beginning.
Skinner said you must always chooose the way of speaking that attributes
the initiation of behavior to the environment; that science itself demands
this overt bias; that nobody who speaks otherwise can really be a
scientist. And his words have been echoed again and again; they are part of
the behaviorist credo. The term "radical behaviorism" is well chosen; this
movement is a extremist one based not on science but on ideology.

We seem to have departed entirely from the issues under consideration,
lapsing instead into an attack on B. F. Skinner in particular and
behaviorism in general. And by the way, all proponents of a particular
outlook or theory are ideologs, you included. Your dichotomy between
science and ideology is a false one: it is quite possible to have both
simultaneously.

One last point. Even though I am a control theorist and think that PCT is
basically correct, and wish to persuade others to my point of view, I can
still describe the facts on which PCT is based without using a single
special term. I can do it without mentioning control, controlled variables,
reference levels or signals, errors or error signals, disturbances (in any
special sense), input functions, comparator functions, or output functions.
I can describe these phenomena in such a way that nobody who doesn't know
me (or control theory) would ever guess what explanation I might offer.

Perhaps so, but in this post you failed to demonstrate that you can stick
purely to observations. Theory and inference kept creeping in.

Can you do the same for EAB? I very much doubt it. The language of EAB uses
special terms and auxiliary words each one of which is carefully defined
(mostly in unusual ways) to support the theoretical position that is
asserted with every breath. Theory and observation are so intertwined that
there is no way to separate them; take away the special terms, and the
observations can't even be described. If you're not allowed to use the word
"reinforcement," what do you say instead? Any other terms you might use
would lay out for all to see what the theoretical bias is. Try it and see.

I have shown that the terms employed in EAB are not used in unusual ways,
but you have chosen to try to make that case anyway, because, I suspect, it
raises a red herring that you can employ to divert the discussion from the
real issues. The problem is not language. By claiming that my use of EAB
language makes communication impossible (or whatever it is you are trying to
have us believe), you simply stop the discussion cold without having to come
to grips with the substantive issues I have raised.

Can I provide a theory-neutral description of the observations? Yes I can.
It was never my intention to do so, because I was attempting to convey the
relationships that exist between certain EAB terms and control system
operation. I'll take up your challenge in my next post. (I have to get
prepared for a lecture at the moment.) Perhaps _then_ we can return to such
substantive questions as what, in control theory, constitutes the event
referred to as the "reinforcer" in EAB. That question, which is the
supposed subject of this series of posts, seems to have been entirely displaced.

Regards,

Bruce

[From Bruce Gregory (70910.1655 EDT)]

Bruce Abbott (970910.1525 EST)]

Perhaps _then_ we can return to such
substantive questions as what, in control theory, constitutes the event
referred to as the "reinforcer" in EAB. That question, which is the
supposed subject of this series of posts, seems to have been entirely displaced.

I thought a pellet was a reinforcer. A pellet, however, is not
an event. I also thought you had answered this question. A
reinforcer is a part of the control loop external to the
controlling system. You said that the heat generated by a
furnace is a reinforcer as far as a thermostat is concerned.
What am I missing?

Bruce

[From Bruce Abbott (970910.2045 EST)]

Bruce Gregory (70910.1655 EDT) --

Bruce Abbott (970910.1525 EST)

Perhaps _then_ we can return to such
substantive questions as what, in control theory, constitutes the event
referred to as the "reinforcer" in EAB. That question, which is the
supposed subject of this series of posts, seems to have been entirely >>

displaced.

I thought a pellet was a reinforcer. A pellet, however, is not
an event. I also thought you had answered this question. A
reinforcer is a part of the control loop external to the
controlling system. You said that the heat generated by a
furnace is a reinforcer as far as a thermostat is concerned.
What am I missing?

We can call the pellet itself a reinforcer because, when produced as a
consequence of a response, its delivery has the particular effect of
reducing the error between a given CV and its reference. But to have this
effect it must be _delivered_ as a consequence of the operant. If the
system were controlling for absence of the pellet, then delivery of the
pellet would _create_ an error, and removal of the pellet would then
function as a reinforcer. (It would then be termed a negative reinforcer.)
So it is important when talking about the reinforcer to know whether its
delivery or removal reduces the error. Delivery and removal of the
reinforcer are events.

I had answered the question, "what's the reinforcer?", but as far as I know,
there has not yet been any agreement from the other side that this answer is
acceptable. Instead of discussing the merits of this proposal, Bill Powers
wants to obfuscate the issue by arguing instead about language and who can
provide the most theory-free description of the rat's behavior in the
operant chamber.

I have explained in what sense the "reinforcer" can be said to strengthen an
operant on which it is contingent: When the experimenter arranges a
contingency between the operant and delivery of the food pellet, the rate at
which the operant is observed to occur rises above its former level. If an
operant can be said to be "stronger" when it occurs with greater frequency,
then what has been observed is a strengthening of the operant that is
(indirectly) due to the fact that the operant now produces the pellet. A
synonym for "strengthen" is "reinforce" -- in the ordinary meaning of the
term "reinforce."

For reasons I am unable to fathom, Bill Powers refuses to accept this simple
and logical explanation, asserting instead that this use twists the ordinary
dictionary meaning of "reinforce." I suppose he must try to make this case
in order to give an appearance of support for his current drumbeat, to wit:

  Not only are the basic terms
  composites of observation and theory, but the auxiliary terms are also
  given special meanings, so we would have to go through a large part of the
  English language, redefining common terms like transitive verbs so they
  could be used without implying any subject -- so that, for example, "to be
  maintained" means the same thing as "continues."

That reminds me: I did not claim that "to be maintained" means the same
thing as "continues." Referring to the dictionary, I said:

  maintain: 1. to keep in existence or continuance; preserve; retain.

This implies that something is doing the maintaining. In reinforcement
theory it is the contingent delivery of the reinforcer, but only in the
sense that this is a necessary condition: all else being equal, we observe
that the response rate remains elevated so long as the response produces the
pellet.

I'll try to get my promised "theory-free description" done next (if that is
still wanted), but probably not tonight.

Regards,

Bruce

[From Bill Powers (970910.1548 MDT)]

Bruce Abbott (970910.1525 EST)--

I submit that it is part observation and part theory. If your goal was to
provide a theory-free description of the facts, you failed.

Good, let's fix it. In some cases the problem is the difference between
intended and understood word-meanings, so we can change some words to void
that.

Before the rats were able efficiently to generate the action would produce
the food, they produced the relevant action only in passing.

"Only in passing" is an inference, not a description. What is observed is
that the _act_ that would produce the food (pressing the lever down to the
point of switch-closure) occurs very infrequently and by inconsistent means.

Change "only in passing" (meant to suggest "only as the rat passes by the
lever") to "infrequently." Actually, the _motor action_ that presses down
on the lever is occurring very frequently, but not at the position of the
lever.

From then on, the rate of food delivery was maintained by the maximum

rate >>of action that the rats could sustain.

Yes. I suggested that rate was limited in exactly this way a bit farther on
in my post. But not that this is theory again; you and I are _inferring_
that the rate of lever-pressing is limited by the rat's capabilities.

By "limiting" I meant "asymptotic." Change to read

was maintained by the action at the maximum rate the rats were observed to
generate.

Your description matches mine: behavior (lever-pressing) that was observed
infrequently at the start of the session is now occurring frequently. We
can demonstrate that the contingency between lever-pressing and food
delivery is crucial to this result, everything else equal. If we break the
contingency, the frequency with which the rat presses the lever soon
declines. If we then reestablish the contingency, lever-pressing returns to
its previously observed rate.

The descriptions differ in that mine says the behavior maintains the food
delivery rate, while yours says the behavior "is maintained" (by the
continued presence of the contingency, or by the occurrance of
reinforcers). My description is based on the observed properties of the
apparatus. Yours assumes a causal connection that is not observed. We will
get back to this later in the post.

If I were to expand the description, I would note that it is not "behavior"
that appears when the contingency is established, but _contact closures_.
Behavior is continuous, before and after the contingency is established.
The location in which behavior takes place is continually changing, until
the contingency is enabled. The appearance of foot pellets after enablement
does not increase the behavior; it stops the switching of the location of
behavior from one place to another. The continuing behavior then continues
to produce food pellets because it is no longer switched away from the lever.

When the contingency is broken, sooner or later behavior is switched away
from the lever to other locations. It does not diminish. Only the contact
closures diminish. While behavior is still directed to the lever, the
initial effect of breaking the contingency, if any, is an increase in
behavior rate.

When much greater amounts of food are produced by a given behavior rate,
the rats do not change their behavior rate appreciably.

What do you mean, "greater amounts of food"? Don't you mean, "a greater
rate of food delivery"? A "greater amount" could mean a larger pellet.

That's what I meant: more food produced by the same efforts, so the rat can
actually cease pressing for a while when it is full. In our experiment, it
could achieve this condition.

They behave in a
stereotyped way, producing and eating food as quickly as possible until a
certain amount of food has been ingested ...

In our weight-control study, we discovered conditions under which the rat
could easily have consumed enough to maintain its body weight, but failed to
do so, until body weight fell to such an extent that I feared the rat would
starve to death, at which point I changed the conditions to those that would
allow the rat's weight to recover.

Evidently, this situation was not "easy" for the rat. Remember, this was
perhaps the 15th change in conditions to which this rat had been subjected
during its short life. I would hestitate to draw any conclusions from the
rat's reaction to this manipulation.

Under the normal conditions the rat
encounters, however, it does behave as you describe, over the long run.
That is, the rat takes in sufficient food to maintain a steady or increasing
body weight.

Yes, that is what I had in mind.

What we observe is the behavior maintaining the food delivery rate to the
extent physically possible, in the dictionary sense of the transitive verb,
"to maintain." The subject is the rat, the object is the rate of food
delivery, and the means is the production of behavior in such a way as to
produce the food.

Yes. (And I have said nothing previously that would contradict this
description.)

Except that you keep saying it is the _behavior_ that is maintained by the
contingency or the reinforcements, instead of the rat's maintaining the
behavior which maintains the reinforcement rate via the contingency. What
you're saying is, in fact, the exact opposite of what I said.

The role of the experimenter is limited to setting up a means by which the
rat's behavior, should the particular act occur, can generate food which
the rat can eat. Since the experimenter is the more powerful of the two
participants in the situation, the experimenter can establish any relation
between any action and the appearance of food, and prevent any other action
from producing food.

This is true in the experiment (after all, an experiment is an arrangement
designed to exert strict control over the relevant variables for the purpose
of determining causation). It is not generally true in natural
circumstances. The fish changed their behavior, not because I am more
powerful than they, but because they learned that they could gain immediate
accesss to a new source of food by so doing.

I conjecture that the fish would not have approached you if they already
had a readily-available supply of equally tasty or nutritious food. I
presume the fish were in no position to wrest the food from your hands. If
you had been standing in the water and the fish were sharks, the situation
would have been different.

Throwing a bread crumb into the water and seeing the fish approach and eat
it does not illustrate any change in the fishes' behavior: this is how they
behaved already toward any objects that smelled of food. The only change in
behavior was to approach _before_ any food was forthcoming, as if they
believed that doing so would make the food appear (as indeed, it did). In
theoretical terms, they were controlling one perception (the relative
position of the blob) as a means of controlling another (appearance of the
food). Controlling appearance of the food, of course, was only a means of
controlling nutrient level (among other things).

What you did was to play the part of an apparatus, by behaving according to
the rule that if the fish approached you you would give them bread. Before
you threw the first bit of food into the water, the fish were already
searching for food. If they had been replete, they would not have directed
their positions toward the food. The fact that they did so as soon as the
bread appeared (after the initial alarm due to the splash) shows that they
were already prepared to approach and eat any food that showed up within
the range of their senses.

The rat's requirements for food determine that the rat
will search for and find the key behavior or something sufficiently like it
to generate food, provided that the requirement is not beyond the rat's
ability to discover it, and that the action needed is not beyond the rat's
ability to sustain it.

To paraphrase: The rat's requirements for food determine . . . that it will
acquire behavior whose consequence is the generation of food. These are not
observations, they are either logical truisms (the requirements must fall
within the capabilities of the animal) or matters of theory (requirements
determine...).

We determined the rats' requirement for food by allowing them to feed
freely. This is an observation, not theory. We tried to determine the limit
on the rat's ability to produce behavior; the fact that this enquiry
fizzled out does not mean that it's impossible to determine the maximum
rate of pressing of which the rat is capable (or rather, of food production
-- rate of pressing is not the only critical part of the pattern of
behavior. The pattern of behavior includes where the behavior is occurring).

If the experimenter changes the contingency, the expected effect with no
change in the rat's pattern of behavior would be an increase or a decrease
in the rate of food delivery.

If physically possible, the rat's pattern of
behavior always changes so as to restore the rate of food delivery to its
original level, or as close to it as possible.

Now we again have entered the realm of theory, control theory to be
specific. You expect that increasing the ratio requirement will lead to an
increased rate of responding, so as to compensate (at least partly) for the
decline in rate of food delivery that would occur if response rate did not
change. This of course is not what we observed.

It's not what _we_ observed, but many others have observed the constancy of
food (and water) delivery and consumption over wide ranges of ratio
schedules -- a range of 5000 to 1, in one report I remember, and even
higher in others.
This is an observation, not theory -- although theory is required to
explain why we observed only a moderate degree of this effect.

The only exception is when
the rat is already producing as much behavior of the right kind as it can,
and the change of contingency reduces the rate of food delivery. Then we
observe no counteracting change in the rat's behavior, because no more
behavior can be generated.

And here is the theoretical explanation. If reinforcement theory had
required the invocation of a limit on response rate to explain the
observations, and PCT did not, I am quite sure you would now be talking
about this rate-limit idea being just what is needed to save the theory.

The asymptotic limit on behavior rate is not theory. We have observed it
and so have many others.
----------------------------------->

You are being mislead by the fact that all our observations were conducted
using ratio schedules. On these schedules, behavior tends to alternate
between responding at a high, relatively constant rate (while completing the
ratio requirement) and pausing (after delivery of the food pellet). On
other schedules the rat (or pigeon) is definitely able to vary its rate of
responding in a fairly smooth way. I have observed rats on a Sidman
avoidance schedule, for example, sitting at the lever and pressing it rather
steadily at a rate of about one press per second -- about 1/5th the rate we
observed on ratio schedules of pellet delivery.

I said nothing about pressing at lower rates. However, your observation
hints that it is still possible that we might have seen a rat pressing at
less than the asympotic rate we observed, so my assumption that rats can't
press at controlled rhythmic rates may be wrong. In that case, the
remaining explanation is that the reference level for food intake was
simply higher than the rat could achieve at its maximum rate of pressing.
This is equivalent to your more complex proposal that the rat wanted the
pellet "right now."

a measure of behavior was that the measure was obtained by dividing total
presses by session duration. As a result, the rat's changing the _kind_ or
_location_ of behavior was indistinguishable from its changing the _rate_
of behavior of a single kind at a single location. In fact, the latter,
which is assumed to occur, does not occur. So it does not need an
explanation.

Both sorts of change have been observed, reported and, incidentally,
explained (descriptively) in terms of reinforcement theory. Both require an
explanation.

How have they been observed? Recording contact closures does not reveal
what the rat is doing between closures. In choice experiments, you get a
record of two different contact closures, but there is still no record of
behavior when neither contact is closed. I read of one experiment in which
the position of pressing along an elongated bar could be detected, and the
rats did indeed vary the position of pressing according to which position
yielded food -- but still no record existed of what they did between
presses. Recording contact closures is still ambiguous as a measure of
behavior, the most ambiguous when there is only a single operant.

From our videos, we can see something that the apparatus never could have

picked up. Many of the rats exhibited repeated pawings of the lever at a
high rate, when they were first learning and early in each session, but
without sufficient force to close the contact. So we recorded much too low
a rate of behavior. The lever was much too massive and insensitive to
forces to pick up what the rats were actually doing.

As to these behaviors being explained by reinforcement theory, since you
can select any aspect of the situation you like as "the reinforcer," it is
always possible to show that behavior correlates with the occurance of
_some_ state of _some_ variable. Lacking any definition of a reinforcer but
its assumed effect, and having no equivalent of the Test, it is not likely
that you will fail to find an explanation. It is not possible for the
proposal that behavior is maintained by reinforcement to be disproven.

Reinforcement theory would account for the same constancy in rate by noting
that, on ratio schedules, higher rates of responding yield shorter times to
reinforcement, which in turn would be expected to generate yet higher rates
of responding. This is a positive feedback loop that pushes response rate
up to the maximum, or at least to an equilibrium value established by the
conflicting effects of reinforcement and response effort. (Effort

increases

with rate; a response rate is reached where the decrease in delay to
reinforcement is balanced by the increase in effort.

There is no need for this explanation, because rates of responding do not
actually change. If they did change, the explanation might be appropriate,
but as they do not, the explanation is empty of meaning. It explains
something that does not happen.

You lost me there. I explain why response rates are constant across ratios,
and you tell me that there is no need for this explanation, because response
rates are constant across ratios.

If you don't observe the same rat under conditions where the response rate
is less than the observed maximum (as we didn't), there are no
less-than-maximum response rates to explain. Your explanation applies to
observations in which the response rate _is_ sometimes less than the
maximum. You're trying to explain imaginary data.

I should also point out that whatever we
observe about behavior, it is not "responses." To observe a response one
must also observe the stimulus; otherwise all that one is observing is an
action.

The term "responses" is a holdover from the old S-R days in psychology. As
used in operant studies, it does not imply a prior stimulus.

Then why do they talk about "stimulus control of behavior?"

What we observe are actions, but the one of interest in the analysis are

those >that have a common consequence in the environment; they are what
Skinner >called "operants."

I am trying to point out that an "operant" is an ambiguous measure with
respect to motor behavior. Your concept of the "observation of interest" is
equivalent to saying that I will point my telescope at one spot in the sky,
and measure the orbit of a planet by counting the number of times it
crosses the field of view. As we know from the videos taken of our
experiments, the behavior of an animal is continuous and varied. The
recording apparatus, however, only sees the behavior when it happens to
occur at the position of the lever, and is of sufficient force to close the
switch operated by the lever. Therefore the apparatus can't tell the
difference between an animal sitting at the lever and pressing it once
every minute, and animal roaming around the cage and chancing to step on
the lever once a minute as it passes by. Furthermore, if the only record
kept is the total number of presses divided by session length, the data do
not distinguish between periods of concentrated bar-pressing and periods
of, for example, sleeping.

What I am saying here, Bruce, is that the behaviorist account, which you
present as a simple factual description of observations, is nothing of the
sort. It is a biased account slanted toward encouraging the listener to
conclude that the environment is controlling behavior -- that behavior is
controlled by its consequences, which is exactly the opposite of the truth.

If that is what you are saying, you have failed to make your case. In fact,
your own description failed as a simple factual account of the observations.

Where it did, that defect can be repaired.

Also, I have said nothing about behavior being controlled by its
consequences. I have simply noted that when certain events are made
contingent on certain behaviors, those behaviors increase in frequency.
When the contingency is broken, those behaviors decrease in frequency,
returning more or less to their original levels. When asked "why did the
behavior increase in frequency?", I might answer, "because it was followed
by the contingent event." That is a purely descriptive explanation, which
makes no appeal to mechanism of any sort.

By this way of describing events, you can answer the question "Why did the
contacts in the thermostat control unit close?" by saying "because the
furnace turned on." The assumed direction of causality makes a great deal
of difference. In your "description", you could also ask "why did the
contingent event occur?" and answer it with "because the behavior occurred
and the contingency was enabled."

When you enable the contingency, sooner or later a food pellet appears. Why
did it appear? Because the animal moved the location of its behavior to the
lever with the result of pressing it enough to close the switch. After
that, we observe that food pellets continue to appear, and at an increasing
rate. Why do they show this behavior? Because the animal stays near the
lever and presses it with increasing success at closing the contact. And
when the contingency is disabled, why do food pellets stop appearing?
Because the contingency is disabled. And why do the lever presses
eventually stop? Because the animal moves the location of its behavior
elsewhere.

These, too, are descriptive explanations. The difference from yours is that
they treat the behavior of the animal as the independent variable, the
cause of the changes in pellet delivery rate, rather than treating them as
the effect. The other main difference is that the assumed direction of
causality in my decription coincides with the observable direction of
causal connections that we can see in the environment, while yours assumes
the opposite direction of causality, corresponding to no causal connections
that we can see.

Nowhere does this explanation
assert that the environment is controlling behavior -- not in the sense in
which the term "control" is used in PCT. What it asserts is that behavior
changes as a function of establishing (or breaking) certain contingencies
between that behavior and its consequences, and that much is a simple
description of the observations.
We seem to have departed entirely from the issues under consideration,
lapsing instead into an attack on B. F. Skinner in particular and
behaviorism in general. And by the way, all proponents of a particular
outlook or theory are ideologs, you included. Your dichotomy between
science and ideology is a false one: it is quite possible to have both
simultaneously.

When I am giving PCT-based explanations, I am an ideolog. However, I am
quite capable of describing what happens without explanation, and without
any slant toward PCT.

One last point. Even though I am a control theorist and think that PCT is
basically correct, and wish to persuade others to my point of view, I can
still describe the facts on which PCT is based without using a single
special term. ...
Can you do the same for EAB? I very much doubt it.

I have shown that the terms employed in EAB are not used in unusual ways,
but you have chosen to try to make that case anyway, because, I suspect, it
raises a red herring that you can employ to divert the discussion from the
real issues.

Nonsense. When you use transitive verbs, which take or imply a subject, as
if they were equivalent to things just happening, you are using language in
a special way. When you say "unusual ways" you mean only "unusual in EAB
communications." They are certainly unusual in terms of ordinary usage.

Can I provide a theory-neutral description of the observations? Yes I can.
It was never my intention to do so, because I was attempting to convey the
relationships that exist between certain EAB terms and control system
operation. I'll take up your challenge in my next post.

I'm looking forward to it. It's interesting that you seem to be admitting
now that your descriptions so far have not actually been theory-neutral.

Best,

Bill P.

[Hans Blom, 970911b]

(Bruce Abbott (970910.2045 EST))

That reminds me: I did not claim that "to be maintained" means the
same thing as "continues." Referring to the dictionary, I said:

maintain: 1. to keep in existence or continuance; preserve; retain

This implies that something is doing the maintaining. In
reinforcement theory it is the contingent delivery of the
reinforcer ...

In my humble opinion, it is not a _thing_ that reinforces, but the
_perception_ (even though it may still be partial or hazy) _of a
relationship_: it is the perceived _contingency_ of the delivery that
reinforces (changes behavior). The rat somehow (mechanism to be
elucidated...) "discovers" that pushing a lever is related to
("causes") delivery of a food pellet. It is this "knowledge" of a
relationship that is "strengthened" (becomes more certain) when
examples of it come about repeatedly and consistently. Concurrent
with this "knowledge discovery" is _using_ that knowledge in
accomplishing the rat's current goals, whether that be eating or
getting away from the stuff.

I'll try to get my promised "theory-free description" done next (if
that is still wanted), but probably not tonight.

Don't try, Bruce, you won't succeed. Notions like reinforcement (or
control, for that matter) are several levels removed from raw
observations and thus _not_ theory-free. At best we will be able to
agree that our raw observations are more or less identical.

On second thought: the description need not be theory-free. It need
only be acceptable to two theories, yours and Bill's.

Greetings,

Hans

[From Bruce Gregory (970911.1025 EDT0]

Bruce Abbott (970910.2045 EST)]

I have explained in what sense the "reinforcer" can be said to strengthen an
operant on which it is contingent: When the experimenter arranges a
contingency between the operant and delivery of the food pellet, the rate at
which the operant is observed to occur rises above its former level. If an
operant can be said to be "stronger" when it occurs with greater frequency,
then what has been observed is a strengthening of the operant that is
(indirectly) due to the fact that the operant now produces the pellet. A
synonym for "strengthen" is "reinforce" -- in the ordinary meaning of the
term "reinforce."

I suspect that the problem arises because of the words "due to"
which imply that a component of the control loop causes a
behavior. Strictly speaking, what is observed is that the
behavior is strengthened, i.e., occurs more frequently. We
observe no "due to". Connecting a furnace to a thermostat makes
it possible for the thermostat to control its perception of the
temperature. The establishment of the connection, however, does
not make the furnace "cause" the thermostat to turn the
switch on and off.

Bruce

[From Bruce Gregory (970911.1045 EDT)]

Bill Powers (970910.1548 MDT)

As to these behaviors being explained by reinforcement theory, since you
can select any aspect of the situation you like as "the reinforcer," it is
always possible to show that behavior correlates with the occurance of
_some_ state of _some_ variable. Lacking any definition of a reinforcer but
its assumed effect, and having no equivalent of the Test, it is not likely
that you will fail to find an explanation. It is not possible for the
proposal that behavior is maintained by reinforcement to be disproven.

This seems to me to be fundamental. I can't tell which component
of the control loop is to be labeled "the reinforcer", since the
entire loop plays a role in reducing the error. Singling out a
single component seems arbitary and potentially misleading. Nor
does it seem to contribute to our understanding of the mechanism
involved.

Bruce