Varying Lever Pressure

[from Gary Cziko 2000.11.13 01:02 GMT]

Does anyone know of any research on how rats or other animals in an
operant conditioning situation vary the force of their lever presses
as necessary to overcome variable resistance in the lever?

I am writing a commentary on a Behavioral and Brain Science article
(see below) and I need to straighten the authors out about how
responses or characteristics of responses (such as response force)
cannot serve as units of selection in learning. They use the example
of selecting response force, so it would be nice to find research to
show that this varies as necessary to control the animal's intended
outcome.

--Gary

Hull, Langman & Glenn:
A GENERAL ACCOUNT OF SELECTION: BIOLOGY, IMMUNOLOGY AND BEHAVIOR
<http://www.cogsci.soton.ac.uk/bbs/Archive/bbs.hull.html>

[From Bruce Abbott (2000.11.12.2220 EST)]

Gary Cziko 2000.11.13 01:02 GMT --

Does anyone know of any research on how rats or other animals in an
operant conditioning situation vary the force of their lever presses
as necessary to overcome variable resistance in the lever?

I am writing a commentary on a Behavioral and Brain Science article
(see below) and I need to straighten the authors out about how
responses or characteristics of responses (such as response force)
cannot serve as units of selection in learning. They use the example
of selecting response force, so it would be nice to find research to
show that this varies as necessary to control the animal's intended
outcome.

Gary, maybe it's you who needs the straightening out, or maybe I just don't
understand your comment. Why _can't_ "responses or characteristics of
responses (such as response force) serve as units of selection in learning"?
I don't see any reason why they can't -- or don't you think that rats or
other animals can sense the characteristics of their own lever-presses, such
as how much force they exert on the lever?

Bruce A.

i.kurtzer (2000.11.13.1336)

[from Gary Cziko 2000.11.13 01:02 GMT]

Does anyone know of any research on how rats or other animals in an
operant conditioning situation vary the force of their lever presses
as necessary to overcome variable resistance in the lever?

See Slifken, Mitchell, and Brenner (1995) Variation of isometric response
force in the rat, Journal of Motor Behavior, 27, p 375-381.

[From Bill Powers (2000.11.13.0202 MST)]

Gary Cziko 2000.11.13 01:02 GMT--

I am writing a commentary on a Behavioral and Brain Science article
(see below) and I need to straighten the authors out about how
responses or characteristics of responses (such as response force)
cannot serve as units of selection in learning.

What does "unit of selection" mean? I've downloaded the paper but have only
skimmed the first part of it. Perhaps you could save me some time; in the
meantime, I'll guess on the basis of what you say about response force.

Perhaps what you're getting at is that it is not the action, but some
consequence of the action that is selected for. If so, then I can
anticipate your point. The action will vary immediately (in force, place,
direction etc.) with every disturbance of the consequence, so could hardly
be thought of as "evolving." It's _systematically_ related to disturbances
of the right variable.

It seems to me that we always come back to the same argument that you and I
have been having for some time. I view "selection" as an active process
that requires a system organized to select things, like a person selecting
which movie to see. Your view seems to be that selecting can be a passive
outcome of nonpurposive processes like sieving a collection of stones of
different sizes. It's true that we use the same word to describe both kinds
of processes, but their mechanisms are so different that they could hardly
be classified together -- _except_ on the basis of the similarity of the
words we use for them. This is similar to the way we use the word "learn"
to include the acquisition of a new behavioral organization and memorizing
a perceptual input: I learned to wiggle my ears; I learned your telephone
number. The two processes have nothing basic in common.

I might as well include Bruce Abbott's comment in the discussion. Bruce
points out that response force can itself be a controlled variable (the
means of controlling it would then be varying sensed muscle tension). I
think the answer to this is that nobody could train a rat to control
response force if response force were the _only_ manipulated consequence of
the action of tensing a muscle. There is no evolutionary advantage to a rat
in producing any particular response force: it must be some other
consequence of producing the muscle tension that is significant to the rat.
Otherwise rats would go around producing particular response forces, which
wouldn't be of much use. Unless responses forces are free to vary as
disturbances of their significant consequences come and go, the
consequences can't be controlled.

Going back to the subject of selection, it seems to me that the authors are
trying to find a classification scheme rather than talking about underlying
mechanisms. It is, of course, the underlying mechanisms that make things
work as they do, not the way we classify phenomena. We may decide that two
phenomena ought to be classified together, but unless we can show that they
share a common mechanism there is really nothing more to it than that -- a
subjective impression of similarity. Are the authors talking about anything
more than that?

You will find, as I have done, that Isaac is a very useful resource for
finding information in the literature.

Best,

Bill P.

[From Bruce Abbott (2000.11.13.0735 EST)]

Bill Powers (2000.11.13.0202 MST) --

Going back to the subject of selection, it seems to me that the authors are
trying to find a classification scheme rather than talking about underlying
mechanisms. It is, of course, the underlying mechanisms that make things
work as they do, not the way we classify phenomena. We may decide that two
phenomena ought to be classified together, but unless we can show that they
share a common mechanism there is really nothing more to it than that -- a
subjective impression of similarity. Are the authors talking about anything
more than that?

I've only skimmed the paper myself, and I'm not really sure what the authors
are trying to do. However, I'd like to point out that things can be
classified according to mechanism -- finding a classification scheme and
talking about underlying mechanisms are not necessarily the mutually
exclusive activities you imply they are. I _think_ what the authors are
trying to do is find the elements in each case that correspond to the three
components of Darwinnian evolution: variation, selection, and replication.
This, it seems to me, is no different than, for example, trying to identify
what physical elements correspond to the comparator, error signal, and so on
in a specific biological control system.

With respect to the "evolution" of response force, one might note that in
the set of lever-presses recorded during an experimental session, there is
variation in response force. If the apparatus now is made to require that
the force fall within a particular "window" before it will deliver a food
pellet to the hungry rat, this constitutes a selection criterion for
response force. If, as a result of this selection, the population of
response forces shifts until most are within the window, this corresponds to
replication, and the situation being examined qualifies as an example of
Darwinnian evolution, broadly conceived.

Bruce A.

[From Bill Powers (2000.11.13.0718 MST)]

...things can be
classified according to mechanism -- finding a classification scheme and
talking about underlying mechanisms are not necessarily the mutually
exclusive activities you imply they are.

I agree, but this implies that you know the mechanisms before you do the
classification. Then, of course, you can group phenomenona according to the
_known_ mechanism that produces them.

I _think_ what the authors are
trying to do is find the elements in each case that correspond to the three
components of Darwinnian evolution: variation, selection, and replication.

Each of these phenomena, however, can be produced by different mechanisms,
so not everything that is classed as "variation" (for example) has
something in common with all other phenomena put in the same category.
These are not the names of mechanisms, but of phenomena that are generated
by mechanisms. And the phenomena are not even the same across all instances.

This, it seems to me, is no different than, for example, trying to identify
what physical elements correspond to the comparator, error signal, and so on
in a specific biological control system.

I think it's quite different, even though something of the same problem
still exists. A comparator creates one clearly-defineable relationship
between two inputs and one output. It's true that this relationship can be
created by several different mechanisms (for example neurons, vacuum tubes,
and transistors, each operating according to its own laws), but the
underlying quantitative relationship can be defined in a way that is the
same for each known mechanism: e = r - p. There is no problem parallel to
that of defining "learning," where one term includes clearly different
phenomena.

With respect to the "evolution" of response force, one might note that in
the set of lever-presses recorded during an experimental session, there is
variation in response force. If the apparatus now is made to require that
the force fall within a particular "window" before it will deliver a food
pellet to the hungry rat, this constitutes a selection criterion for
response force. If, as a result of this selection, the population of
response forces shifts until most are within the window, this corresponds to
replication, and the situation being examined qualifies as an example of
Darwinnian evolution, broadly conceived.

This applies only in a disturbance-free environment, which gives a
distorted picture of all relationships. In a realistic environment, one
response force will not result in one unique environmental consequence. In
that kind of environment, it would become clear that it is control of the
consequence of a particular response force that is learned: a control
process is acquired, not a particular act. However, when the environment is
arranged so the same response force always has the same effect on other
variables, we have a degenerate case in which it is not possible to see
what is actually learned. This is true of all experiments in which a
behavioral action always has the same consequence. The only way to see what
is actually going on is to introduce disturbances and see what, if
anything, is maintained the same or nearly so.

Best,

Bill P.

[From Bruce Gregory (2000.1113.1000)]

Bill Powers (2000.11.13.0718 MST)

There is no problem parallel to
that of defining "learning," where one term includes clearly different
phenomena.

If we define "learning" as a process that allows an organism to accomplish
something it was unable to accomplish before, it is not clear to me that we
must be dealing with "clearly different phenomena."

BG

[From Rick Marken (2000.11.13.1010)]

Gary Cziko (2000.11.13 01:02 GMT) --

Does anyone know of any research on how rats or other animals
in an operant conditioning situation vary the force of their
lever presses as necessary to overcome variable resistance in
the lever?

i.kurtzer (2000.11.13.1336)

See Slifken, Mitchell, and Brenner (1995) Variation of isometric
response force in the rat, Journal of Motor Behavior, 27,
p 375-381.

This sounds like it could be a _very_ interesting paper. Does it
really deal with how the rats vary force to compensate for varying
force disturbances to a controlled variable (which I think is what
Gary was asking about). For the sake of those (like myself) who
don't have quick access to the _Journal of Motor Behavior_
could you give a quick summary of what Slifken et al did and what
they found.

Thanks

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
Life Learning Associates e-mail: marken@mindreadings.com
mindreadings.com

[from Gary Cziko 2000.11.13 21:44 GMT]

[I thought I had sent a version of this message earlier, but can now
find no trace of it. So I apologize if anyone gets two versions of
this.]

I asked:

[from Gary Cziko 2000.11.13 01:02 GMT]

Does anyone know of any research on how rats or other animals in an
operant conditioning situation vary the force of their lever presses
as necessary to overcome variable resistance in the lever?

i.kurtzer (2000.11.13.1336) responded:

See Slifken, Mitchell, and Brenner (1995) Variation of isometric response
force in the rat, Journal of Motor Behavior, 27, p 375-381.

Thanks Isaac. However, this study actually turns out to demonstrate
what Bruce Abbott was talking about--how (perceived) response force
can be selected (learned) if the force required for reward is kept
constant.

In this study the required lever force was kept constant for six
days, with no feedback during the response as to whether the force
being applied was sufficient (click and food were presented only
after a response with the required force was terminated). Then the
force required was increased slightly and kept constant for another
six days (sorry, Rick!).

When I encounter a wet, heavy shovelful of snow, I automatically
produce more force to lift and throw it. When I pump up my bicycle
tires, I produce whatever force is necessary as the tire fills and
there is increasing resistance to movements of the pump handle. Can
there really be no experimental studies of such phenomena for
animals and/or humans?

--Gary

[From Bill Powers (2000.11.13.1849 MST)]

Bruce Gregory (2000.1113.1000)--

There is no problem parallel to
that of defining "learning," where one term includes clearly different
phenomena.

If we define "learning" as a process that allows an organism to accomplish
something it was unable to accomplish before, it is not clear to me that we
must be dealing with "clearly different phenomena."

The question is, what is that process? You can claim that there is some one
process that it accounts for all ways in which an organism becomes able to
accomplish something it could not accomplish before, but I'll bet you can't
say what it is or how it works. I, of course, would claim that there is no
one process that can account for all the things we mean when we say
"learning." You can create a category at will, but whether it actually
contains any items is a different question.

Best,

Bill P.

[From Bill Powers (2000.11.14.0406 MST)]

Gary Cziko 2000.11.13 21:44 GMT--

Thanks Isaac. However, this study actually turns out to demonstrate
what Bruce Abbott was talking about--how (perceived) response force
can be selected (learned) if the force required for reward is kept
constant.

In this study the required lever force was kept constant for six
days, with no feedback during the response as to whether the force
being applied was sufficient (click and food were presented only
after a response with the required force was terminated). Then the
force required was increased slightly and kept constant for another
six days (sorry, Rick!).

How you see this experiment depends on the bias you bring to it. If you
believe that the environment can, by providing food, select one behavior of
an organism out of a range of possible behaviors, then you will interpret
what goes on according to reinforcement theory. The action of pressing with
a certain force produces a food consequence, which increases the
probability that the rat will press with the same force the next time. Thus
it is the contingency set up in the environment together with the
reinforcer that selects the behavior of pressing with a certain force. If
you want to see the environment as being in control of behavior, this way
of thinking is how to achieve that desire.

On the other hand, if you believe that environments merely convert actions
into physical consequences, then it will be the rat that varies the force
with which it acts until it succeeds in creating the consequence of getting
food. In the experiment cited by Isaac, as you note, the food does not
appear until the end of the response, so the animal can only vary the
reference force over many trials until it finds the reference force that
produces food. Thus it is the rat that create variations (in its own
actions), experiences the consequences (appearance or nonappearance of
food), and selects the action that will produce the desired or intended
consequence (by ceasing to try different actions). The environment is a
passive component in this process and has no special effects on behavior.
If you want to see organisms as autonomous agents that control perceived
aspects of the nonliving and nonpurposive world, this way of thinking will
help to achieve that desire.

Of course neither kind of organization, as fans of PCT know, requires any
miracles to work. The same kind of purposive behavior that traditional
science rejected as depending on magic we know now can be accepted as
perfectly real and nonmagical, so the Darwinian concept is no longer the
only alternative to divine guidance. The deliberate purposive selection by
an organism of consequences of deliberately varied actions is a perfectly
scientific and feasible concept of how animals learn, one that does not
depend on miracles any more than natural selection does.

As long as we let the desire to see the situation in one or the other way
determine what we believe, we will be stuck between these choices. On the
other hand, if anyone wonders whether one interpretation might be more
defensible than the other, there might be experiments that a clever person
could devise that would tip the balance.

Does anyone care to find out the truth? Or is the contention that we
already know the truth?

Best,

Bill P.

···

When I encounter a wet, heavy shovelful of snow, I automatically
produce more force to lift and throw it. When I pump up my bicycle
tires, I produce whatever force is necessary as the tire fills and
there is increasing resistance to movements of the pump handle. Can
there really be no experimental studies of such phenomena for
animals and/or humans?

--Gary

[From Bruce Gregory (2000.1114.0646)]

Bill Powers (2000.11.13.1849 MST)

Bruce Gregory (2000.1113.1000)--
>
>If we define "learning" as a process that allows an organism to accomplish
>something it was unable to accomplish before, it is not clear to me that we
>must be dealing with "clearly different phenomena."

The question is, what is that process? You can claim that there is some one
process that it accounts for all ways in which an organism becomes able to
accomplish something it could not accomplish before, but I'll bet you can't
say what it is or how it works. I, of course, would claim that there is no
one process that can account for all the things we mean when we say
"learning." You can create a category at will, but whether it actually
contains any items is a different question.

You can make claims at will, but whether they are supported by evidence is
a different question.

BG

[From Bruce Gregory (2000.1114.1252)]

Unless one holds a dualistic view of life, all learning requires a
modification of the existing neural structure. It's true that I can modify
your neural structure by hitting you over the head with a hammer, so we
need to say something about how the modification occurs in order to qualify
as learning. The simplest assumption that I can think of is that learning
takes place whenever the organization of the neural system is altered as a
result of its own operation. There are many neural network models of such
processes. I am reluctant to say that is obvious that they are all worthless.

BG

[From Rick Marken (2000.11.14.1030)]

Bill Powers (2000.11.14.0406 MST)--

How you see this experiment depends on the bias you bring to it. If you
believe that the environment can, by providing food, select one behavior of
an organism out of a range of possible behaviors, then you will interpret
what goes on according to reinforcement theory...

On the other hand, if you believe that environments merely convert actions
into physical consequences, then it will be the rat that varies the force
with which it acts until it succeeds in creating the consequence of getting
food...

As long as we let the desire to see the situation in one or the other way
determine what we believe, we will be stuck between these choices. On the
other hand, if anyone wonders whether one interpretation might be more
defensible than the other, there might be experiments that a clever person
could devise that would tip the balance.

I think Gary Cziko and Bruce Abbott are clever enough to devise such
an experiment. Perhaps they could take the lead in designing such an
experiment on the net. Then someone could run it and we would be able
to see which way the balance tips in terms of what is doing the selecting:
the environment or the organism. And the nice thing about a well designed
experiment is that the result is never a draw.

Best

Rick

···

---
Richard S. Marken Phone or Fax: 310 474-0313
MindReadings.com mailto: marken@mindreadings.com
www.mindreadings.com

[From Bruce Abbott (2000.11.14.1330 EST)]

Bill Powers (2000.11.14.0406 MST) --

How you see this experiment depends on the bias you bring to it. If you
believe that the environment can, by providing food, select one behavior of
an organism out of a range of possible behaviors, then you will interpret
what goes on according to reinforcement theory. The action of pressing with
a certain force produces a food consequence, which increases the
probability that the rat will press with the same force the next time. Thus
it is the contingency set up in the environment together with the
reinforcer that selects the behavior of pressing with a certain force. If
you want to see the environment as being in control of behavior, this way
of thinking is how to achieve that desire.

Those who subscribe to this view believe that the selecting is being done by
mechanisms within the organism; the notion that the environment is doing the
selecting is simply a shorthand way of noting that the environmental
consequences of behavior, as perceived by the organism, result in a change
in what the organism is likely to do in the future under similar conditions.
Certain consequences have this effect because of the way the organism is
organized.

Note that this account is strictly descriptive -- it does not propose any
specific mechanism to accomplish these changes, any more than Darwin's
theory proposes any specific mechanism by which characteristics of organisms
vary across organisms and are transmitted to their offspring. It says that
these things happen, but it does not explain how.

On the other hand, if you believe that environments merely convert actions
into physical consequences, then it will be the rat that varies the force
with which it acts until it succeeds in creating the consequence of getting
food. In the experiment cited by Isaac, as you note, the food does not
appear until the end of the response, so the animal can only vary the
reference force over many trials until it finds the reference force that
produces food. Thus it is the rat that create variations (in its own
actions), experiences the consequences (appearance or nonappearance of
food), and selects the action that will produce the desired or intended
consequence (by ceasing to try different actions). The environment is a
passive component in this process and has no special effects on behavior.
If you want to see organisms as autonomous agents that control perceived
aspects of the nonliving and nonpurposive world, this way of thinking will
help to achieve that desire.

I hold that these accounts are not necessarily opposing views. I don't know
of anyone in EAB who would disagree with your assertion that "environments
merely convert actions into physical consequences." The notion that
environments select is meant in the same sense as this phrase is used when
describing Darwin's theory -- biologists do not mean, when they say that the
environment selects for certain traits, that the environment is some
purposive being that picks out those traits; rather, they mean that
characteristics of the environment determine which characteristics of the
organism will tend to increase or decrease in frequency within the
population over generations. Darwin proposed that such natural selection
occurs, but he could not specify the underlying mechanisms by which the
traits are made to vary, or by which they get passed on to the offspring.
Today we know about the genetic machinery and the sources of variation.
However, we do not claim that genetics offers a clear _alternative_ to
Darwin's theory of evolution. We do not say the people with one sort of
bias believe in Darwin's theory, whereas others with another sort of bias
believe in genetic theory. Genetics is not in conflict with evolution;
instead, genetics _provides the mechanism_ by which Darwinnian evolution occurs.

In exactly parallel fashion, reinforcement theory merely describes how
certain environmental consequences of actions lead to certain changes in the
future liklihood of those actions. If it is true that those physical
consequences of actions have the effect of increasing the liklihood that
such actions will occur under similar circumstances in the future, then
there must be some mechanism within the organism that mediates this change.
What the underlying mechanisms are that accomplish the observed changes in
behavior is an important question, and PCT offers one possible mechanism.
Such a mechanism would explain why certain consequences of actions bring
about those changes in the future liklihood of those actions; it would not
disprove that such changes do occur.

Bruce A.

[From Bruce Gregory (2000.1014.1351)]

Bruce Abbott (2000.11.14.1330 EST)

In exactly parallel fashion, reinforcement theory merely describes how
certain environmental consequences of actions lead to certain changes in the
future liklihood of those actions. If it is true that those physical
consequences of actions have the effect of increasing the liklihood that
such actions will occur under similar circumstances in the future, then
there must be some mechanism within the organism that mediates this change.
What the underlying mechanisms are that accomplish the observed changes in
behavior is an important question, and PCT offers one possible mechanism.
Such a mechanism would explain why certain consequences of actions bring
about those changes in the future liklihood of those actions; it would not
disprove that such changes do occur.

My view is that of Divine Selection. God causes us to act in ways that
accomplish His goals. I do not propose a specific mechanism, but simply
note that Divine Selection is compatible with what we observe. PCT offers
one possible mechanism by which the Divine Will is exercised.

BG

[From Bruce Gregory (2000.1114.1600)]

Your idea of Divine Selection asserts a causative mechanism. Bruce Abbott
said that EAB does not assert a causative mechanism, but is frankly limited
to external-observer description of a correlation. By your facetious
parallel are you saying that this is not so, and that EAB in fact does
assert a causative mechanism comparable to Divine Selection?

        Bruce Nevin

My point was simply that in the absence of a model one explanation is hard
to distinguish from another. EAB appears to be light on models and
therefore encourages hand waving. Not that I have anything against hand
waving, as long as you realize that is what you are doing.

BG

[From Bruce Abbott (2000.11.14.1835 EST)]

Bruce Gregory (2000.1114.1600) --

My point was simply that in the absence of a model one explanation is hard
to distinguish from another. EAB appears to be light on models and
therefore encourages hand waving. Not that I have anything against hand
waving, as long as you realize that is what you are doing.

One could say the same about Darwin's theory of evolution. However, that
doesn't seem to have prevented it from being seen by biologists as the most
significant organizing principle in biology.

Bruce A.

Your idea of Divine Selection asserts a causative mechanism. Bruce Abbott said that EAB does not assert a causative mechanism, but is frankly limited to external-observer description of a correlation. By your facetious parallel are you saying that this is not so, and that EAB in fact does assert a causative mechanism comparable to Divine Selection?

         Bruce Nevin

···

At 01:52 PM 11/14/2000 -0500, Bruce Gregory wrote:

[From Bruce Gregory (2000.1014.1351)]

My view is that of Divine Selection. God causes us to act in ways that
accomplish His goals.

[From Bruce Gregory (2000.1115.0932)]

Bruce Abbott (2000.11.14.1835 EST)

>Bruce Gregory (2000.1114.1600) --

>My point was simply that in the absence of a model one explanation is hard
>to distinguish from another. EAB appears to be light on models and
>therefore encourages hand waving. Not that I have anything against hand
>waving, as long as you realize that is what you are doing.

One could say the same about Darwin's theory of evolution.

One could, but one would be wrong.

BG