Closed-loop Experimental Psychology

[From Rick Marken (2010.07.02.1645)]

I've changed the subject line again; so many changes, so little
interest;-) Anyway, I think the new subject line better describes what
we're now talking about.

Bill Powers (2010.07.01.0615 MDT)--

RM: Not really. It's a closed-loop model, even if it's one shot. It's an
input that is produced in one shot, not an output.

BP: I think you're both ignoring the fact that this is by no means a
one-shot response when you see it from the subject's point of view.

I meant that that version of the model just went though one iteration.
I wanted the model's behavior to look like the subject's behavior
looks from the observer's point of view. My goal was just to show that
a psychophysical experiment can be modeled as a closed-loop task, even
when it _looks_ open loop to an observer.

Choosing to treat the process as a simple triggering of one event by another
one doesn't mean that is actually what is going on.

Of course. But experiments are designed to make it _look like_ that's
what's going on; and experimental psychologists are prepared to see it
that way (as a behavioral event triggered by the sensory consequences
of an environmental event) anyway. So what I'm trying to do now is
show that it's possible to see (and model) the behavior in these
experiments as closed-loop.

Last complaint. My experience with trying to model behavior has shown me
that most psychological experiments, even in psychophysics, are enormously
complex, though psychologists see them as simple measures of cause and
effect.

What's exciting to me about this discussion (or, to be less
politically correct, this argument) about how to model the behavior in
psychophysical experiments is that, after 30 years, I finally
understand how PCT relates to conventional psychology experiments. I
don't mean I know how to model these experiments in detail -- as you
say, they are just way too complex. I mean that I understand that the
behavior in these experiments is purposeful (like all behavior). In
every experiment the subject must be given a purpose -- through
instructions, if the subjects are humans, or though "establishing
operations" if the subjects are non-human. And I now see that this
purpose is essential or the subject will not behave as expected. So in
the reaction time experiment the subject is asked to adopt the purpose
of pressing a button only when a signal light comes on and not
otherwise. If the subject doesn't adopt this purpose the experiment
won't work at all. In the reaction time experiment, for example, this
is because the onset of a light doesn't cause people to press buttons
in the same way that the onset of a light causes an electrical
response in a photocell.

Once I realized that purpose is an essential component of an
experiment, I had a better feel for how PCT relates to conventional
experiments. Purpose implies controlled variables so the behavior in
an experiment must aimed at controlling some variable; that's what
carrying out a purpose is in PCT. Of course, we don't know what
variable(s) a subject is actually controlling but sometimes -- as in
the reaction time task -- we can make a pretty reasonable guess: the
relationship between light onset and button pressing and the speed
with which this relationship is brought to its reference state.

Since one of my main interests in PCT has always been its implications
for how we do psychological science, I now take my task to be trying
to show that the behavior in conventional psychological experiments
can be seen as closed loop. We'll see if I have any better luck with
this over the next 30 years than I had over the previous 30 trying to
show that the open-loop causal model of behavior doesn't work when the
system under study is closed loop.

Best regards

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

Rick,

This is well-done. I do believe it can show the fallacies in many classic psychology experiments and why their conclusions are neither scientific nor trustworthy.

Showing how PCT based experiments can provide accurate analyses of why people behave as they do and what you can and can’t do to change the behavior of others could well provide the practical value of learning PCT which I perceive has always been its Achilles Heel (or is that Hell?) which has constrained its wider acceptance.

My best wishes to you as a PCTer extradonair in your reorganized approach in teaching PCT and for a patriotic and happy Fourth of July!

Kenny

In a message dated 7/2/2010 7:50:00 P.M. Eastern Daylight Time, rsmarken@GMAIL.COM writes:

···

[From Rick Marken (2010.07.02.1645)]

I’ve changed the subject line again; so many changes, so little
interest;-) Anyway, I think the new subject line better describes what
we’re now talking about.

Bill Powers (2010.07.01.0615 MDT)–

RM: Not really. It’s a closed-loop model, even if it’s one shot. It’s an
input that is produced in one shot, not an output.

BP: I think you’re both ignoring the fact that this is by no means a
one-shot response when you see it from the subject’s point of view.

I meant that that version of the model just went though one iteration.
I wanted the model’s behavior to look like the subject’s behavior
looks from the observer’s point of view. My goal was just to show that
a psychophysical experiment can be modeled as a closed-loop task, even
when it looks open loop to an observer.

Choosing to treat the process as a simple triggering of one event by another
one doesn’t mean that is actually what is going on.

Of course. But experiments are designed to make it look like that’s
what’s going on; and experimental psychologists are prepared to see it
that way (as a behavioral event triggered by the sensory consequences
of an environmental event) anyway. So what I’m trying to do now is
show that it’s possible to see (and model) the behavior in these
experiments as closed-loop.

Last complaint. My experience with trying to model behavior has shown me
that most psychological experiments, even in psychophysics, are enormously
complex, though psychologists see them as simple measures of cause and
effect.

What’s exciting to me about this discussion (or, to be less
politically correct, this argument) about how to model the behavior in
psychophysical experiments is that, after 30 years, I finally
understand how PCT relates to conventional psychology experiments. I
don’t mean I know how to model these experiments in detail – as you
say, they are just way too complex. I mean that I understand that the
behavior in these experiments is purposeful (like all behavior). In
every experiment the subject must be given a purpose – through
instructions, if the subjects are humans, or though “establishing
operations” if the subjects are non-human. And I now see that this
purpose is essential or the subject will not behave as expected. So in
the reaction time experiment the subject is asked to adopt the purpose
of pressing a button only when a signal light comes on and not
otherwise. If the subject doesn’t adopt this purpose the experiment
won’t work at all. In the reaction time experiment, for example, this
is because the onset of a light doesn’t cause people to press buttons
in the same way that the onset of a light causes an electrical
response in a photocell.

Once I realized that purpose is an essential component of an
experiment, I had a better feel for how PCT relates to conventional
experiments. Purpose implies controlled variables so the behavior in
an experiment must aimed at controlling some variable; that’s what
carrying out a purpose is in PCT. Of course, we don’t know what
variable(s) a subject is actually controlling but sometimes – as in
the reaction time task – we can make a pretty reasonable guess: the
relationship between light onset and button pressing and the speed
with which this relationship is brought to its reference state.

Since one of my main interests in PCT has always been its implications
for how we do psychological science, I now take my task to be trying
to show that the behavior in conventional psychological experiments
can be seen as closed loop. We’ll see if I have any better luck with
this over the next 30 years than I had over the previous 30 trying to
show that the open-loop causal model of behavior doesn’t work when the
system under study is closed loop.

Best regards

Rick

–
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bill Powers (29010.07.03.0708 MDT )]

Rick Marken (2010.07.02.1645)--

BP earlier: > Choosing to treat the process as a simple triggering of one event by another one doesn't mean that is actually what is going on.

RM: Of course. But experiments are designed to make it _look like_ that's
what's going on; and experimental psychologists are prepared to see it
that way (as a behavioral event triggered by the sensory consequences
of an environmental event) anyway. So what I'm trying to do now is
show that it's possible to see (and model) the behavior in these
experiments as closed-loop.

The theme I'm going to be presenting at the Manchester meetings is boiling down to about the same idea as yours. The actions that we call people's behavior are not what they are doing. The behaviors we observe are the means they use to control some aspect of their own experiences. If you want to understand their behavior, you have to understand what they are trying to accomplish and why that particular pattern of behavior is required by current circumstances if they are to accomplish it.

This is the message we should stay on until it finally penetrates. This will finally mark the end of both behaviorism and cognitive science as they are understood today.

I have in mind a demonstration that will blow all the fuses. The one I'm going to present in Manchester is simpler: the subject keeps a moving dot on a target, in two dimensions, by writing the word "hello" in script (or whatever pattern you want him or her to draw). It's working beautifully. The one that comes next, if I can manage to get it to work, is the big one: the subject makes a cursor trace out the word "hello" (already written on the screen in script) by writing the word "goodbye", or anything else, in script. If anyone wants to beat me to it, feel free.

Best,

Bill P.

1 Like

[From Rick Marken (2010.07.03.0930)]

Rick,

This is well-done.� I do believe it can show the fallacies in many classic
psychology experiments and why their conclusions are neither scientific
nor�trustworthy.

Gee, thanks Kenny.

Best

Rick

···

On Sat, Jul 3, 2010 at 6:09 AM, Kenneth Kitzke Value Creation Systems <KJKitzke@aol.com> wrote:

Showing how PCT�based experiments�can provide accurate analyses of
why�people behave as they do and what you can and can't do to change the
behavior of others could well provide the practical value of�learning PCT
which I�perceive has always�been�its Achilles Heel (or is that Hell?) which
has constrained its�wider acceptance.

My best wishes�to you as a PCTer extradonair�in your reorganized approach in
teaching PCT and for a patriotic and happy Fourth of July!

Kenny

In a message dated 7/2/2010 7:50:00 P.M. Eastern Daylight Time,
rsmarken@GMAIL.COM writes:

[From Rick Marken (2010.07.02.1645)]

I've changed the subject line again; so many changes, so little
interest;-) Anyway, I think the new subject line better describes what
we're now talking about.

Bill Powers (2010.07.01.0615 MDT)--

RM: Not really. It's a closed-loop model, even if it's one shot. It's an
input that is produced in one shot, not an output.

BP: I think you're both ignoring the fact that this is by no means a
one-shot response when you see it from the subject's point of view.

I meant that that version of the model just went though one iteration.
I wanted the model's behavior to look like the subject's behavior
looks from the observer's point of view. My goal was just to show that
a psychophysical experiment can be modeled as a closed-loop task, even
when it _looks_ open loop to an observer.

Choosing to treat the process as a simple triggering of one event by
another
one doesn't mean that is actually what is going on.

Of course. But experiments are designed to make it _look like_ that's
what's going on; and experimental psychologists are prepared to see it
that way (as a behavioral event triggered by the sensory consequences
of an environmental event) anyway. So what I'm trying to do now is
show that it's possible to see (and model) the behavior in these
experiments as closed-loop.

Last complaint. My experience with trying to model behavior has shown me
that most psychological experiments, even in psychophysics, are enormously
complex, though psychologists see them as simple measures of cause and
effect.

What's exciting to me about this discussion (or, to be less
politically correct, this argument) about how to model the behavior in
psychophysical experiments is that, after 30 years, I finally
understand how PCT relates to conventional psychology experiments. I
don't mean I know how to model these experiments in detail -- as you
say, they are just way too complex. I mean that I understand that the
behavior in these experiments is purposeful (like all behavior).� In
every experiment the subject must be given a purpose -- through
instructions, if the subjects are humans, or though "establishing
operations" if the subjects are non-human.� And I now see that this
purpose is essential or the subject will not behave as expected. So in
the reaction time experiment the subject is asked to adopt the purpose
of pressing a button only when a signal light comes on and not
otherwise. If the subject doesn't adopt this purpose the experiment
won't work at all. In the reaction time experiment, for example, this
is because the onset of a light doesn't cause people to press buttons
in the same way that the onset of a light causes an electrical
response in a photocell.

Once I realized that purpose is an essential component of an
experiment, I had a better feel for how PCT relates to conventional
experiments. Purpose implies controlled variables so the behavior in
an experiment must aimed at controlling some variable; that's what
carrying out a purpose is in PCT.� Of course, we don't know what
variable(s) a subject is actually controlling but sometimes -- as in
the reaction time task -- we can make a pretty reasonable guess: the
relationship between light onset and button pressing and the speed
with which this relationship is brought to its reference state.

Since one of my main interests in PCT has always been its implications
for how we do psychological science, I now take my task to be trying
to show that the behavior in conventional psychological experiments
can be seen as closed loop. We'll see if I have any better luck with
this over the next 30 years than I had over the previous 30 trying to
show that the open-loop causal model of behavior doesn't work when the
system under study is closed loop.

Best regards

Rick

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Rick Marken (2010.07.03.1040)]

Bill Powers (29010.07.03.0708 MDT )--

Rick Marken (2010.07.02.1645)--

RM: So what I'm trying to do now is
show that it's possible to see (and model) the behavior in these
experiments as closed-loop.

The theme I'm going to be presenting at the Manchester meetings is boiling
down to about the same idea as yours...

I have in mind a demonstration that will blow all the fuses...the
subject makes a cursor trace out the word "hello" (already written on the
screen in script) by writing the word "goodbye", or anything else, in
script. If anyone wants to beat me to it, feel free.

This sounds great. But what I'm interested in doing now is trying to
develop demonstrations that will look familiar to experimental
psychologists. I think I have had difficulty convincing experimental
psychologists that PCT shows that their work is a meaningless waste of
time (other than because people don't like being told that;-) is
because the demonstrations I've developed to show this don't look like
the kinds of experiments that are done by experimental psychologists.
My little Mind Reading demo, which I think is such a clever way of
showing that you can't tell what a person is doing by looking at what
they are doing, is easily dismissed as being irrelevant to what
happens in a "real" psychological experiment. I think your "fuse
blowing" demo sounds great but I don't think it will blow any fuses in
the psychological research community. What we need, I believe, are
demonstrations using experimental situations that would be familiar to
experimental psychologists -- such as the perceptual and reaction time
experiments we've been talking about -- that show why such experiments
tell us very little about why the people behave as they do in these
situations.

Best regards

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[From Bill Powers (2010.07.03.1335 MDT)]

Rick Marken (2010.07.03.1040) --

RM: ...what I'm interested in doing now is trying to
develop demonstrations that will look familiar to experimental
psychologists.

BP: I understand the strategy, but the downside is that for all the
familiar demonstrations, they will already have familiar explanations
which they evidently think are adequate. Getting people to unlearn
what they think they already know is about as tough as problems get.
You end up pitting your explanation against theirs -- and most of the
time there isn't even a real phenomenon that you both understand the same way.

I think I'm going to focus on phenomena for which they have no ready
explanations. I said to Sara Tai (Manchester), "what I really want to
talk with you about is a series of demonstrations that show the utter
impossibility that any existing theory of behavior other than PCT
could be correct." That's the kind of challenge I want to issue.
Maybe someone can refute my claim. That will be useful in itself, and
will certainly be better than being ignored. But if nobody can refute
it even after trying, that will shut the book on the psychology of
the past, and we can start psychology on a second foundation. For
Asimov fans, that's spelled Second Foundation (see attached picture).

I think I have had difficulty convincing experimental
psychologists that PCT shows that their work is a meaningless waste of
time (other than because people don't like being told that;-) is
because the demonstrations I've developed to show this don't look like
the kinds of experiments that are done by experimental psychologists.

So you want to convince experimental psychologists that the
experiments which do look like theirs are a meaningless waste of
time? Won't this result in even greater obstinacy and opposition?
Does it help you do make your approach more effective if I tell you
it's a meaningless waste of time?

My little Mind Reading demo, which I think is such a clever way of
showing that you can't tell what a person is doing by looking at what
they are doing, is easily dismissed as being irrelevant to what
happens in a "real" psychological experiment. I think your "fuse
blowing" demo sounds great but I don't think it will blow any fuses in
the psychological research community.

The demo will cause anyone to blow a fuse who tries to find an
explanation for it in behaviorism or cognitive psychology. There is
no relationship between the stimulus and the response. There is no
way in which, even with infinite computing capacity, a brain could
analyze the environment, generate a plan of action, and formulate the
patterns of neural impulses needed to generate those actions to
produce the results we see in this demonstration. Since we can
program models that will closely approximate this kind of behavior,
we will be the last ones on the island.

What we need, I believe, are
demonstrations using experimental situations that would be familiar to
experimental psychologists -- such as the perceptual and reaction time
experiments we've been talking about -- that show why such experiments
tell us very little about why the people behave as they do in these
situations.

It's worth trying, despite my doubts. So have at it and good luck.

Best,

Bill P.

[From Rick Marken (2010.07.03.1750)]

Bill Powers (2010.07.03.1335 MDT)--

What a great picture! Thanks.

Rick Marken (2010.07.03.1040) --

BP: I understand the strategy, but the downside is that for all the familiar
demonstrations, they will already have familiar explanations which they
evidently think are adequate.

My experience is that the same thing happens with demos, like the
squared circle and my mind reading demo. As I said, what I usually
get when I present our demos is something like "sure, this is true of
closed loop situations like tracking but the behavior in experiments
is not closed loop". So my goal is to develop demonstrations that show
that the behavior in all psychological experiments is closed loop.

RM: I think I have had difficulty convincing experimental
psychologists that PCT shows that their work is a meaningless waste of
time (other than because people don't like being told that;-) is
because the demonstrations I've developed to show this don't look like
the kinds of experiments that are done by experimental psychologists.

So you want to convince experimental psychologists that the experiments
which do look like theirs are a meaningless waste of time?

I'm going to have to remember to stop telling psychologists that;-)

RM: �What we need, I believe, are
demonstrations using experimental situations that would be familiar to
experimental psychologists -- such as the perceptual and reaction time
experiments we've been talking about -- that show why such experiments
tell us very little about why the people behave as they do in these
situations.

It's worth trying, despite my doubts. So have at it and good luck.

Thanks. By the way, did I mention that you will be developing all the
programs for me;-)

Thanks, and have a happy 4th!

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[Martin Taylor 2010.07.05.00.05]

[From Rick Marken (2010.07.04.0840)]

  Martin Taylor (2010.07.04.10.15)_-
As a passing comment, very little of psychophysics is concerned with
magnitude estimation. Most psychophysics has to do with the capabilities of
the processing channels, such as timing effects, noise limits,
cross-interference, cross-support, and the like.

Yes, but since all psychophysical research is based on a an open-loop
causal model of the systems under study, those who study psychophysics
(which used to include me;-) can't really be learning much about the
"capabilities of the processing channels" of these systems using
conventional methods. Agreed?

As I said earlier [Martin Taylor 2010.07.04.17.02], definitely not agreed.

I've been thinking about this more and more over the course of this thread, and I'm beginning to see that a good part of what is taught as good experimental practice in psychophysics actually is performing the test for the controlled variable, whether the experimenters know it or not (and most don't). "Good practice" is intended to ensure that the subject is trying to do what the experimenter wants, and that the subject's output mechanism doesn't interfere with measuring the properties under study (unless, of course, the output pathways themselves are of interest, as sometimes is the case; in that case, the perceptual input is usually made very simple).

Of course, most practitioners will not realize that they are doing "The Test", any more than most people realize they are always controlling their inputs, but it seems to me that good practice nevertheless usually, if not always, results in the experimenter performing the Test for the controlled variable. Experimental technique has developed by normal evolutionary methods -- keep what works, change what doesn't -- and in my opinion that evolutionary development has approached what would have been prescribed from first principles by PCT analysis.

In any experiment, both the experimenter and the subject are controlling many variables, but from the experimenter's point of view only two of the subject's controlled variables matter very much, namely (1) that the subject is controlling for perceiving that the experimenter perceives that the subject is performing as the experimenter wishes, and (2) that the subject is in fact acting to control against disturbances in the way the experimenter intended. The controlled variable and reference value in (2) are part -- but only a part -- of the output structure of (1), but could take the experimenter's desired values in the absence of (1). The experimenter's perception of both of these variables is controlled, meaning that if the experimenter does not perceive their state to be as wished, the experimenter acts to bring their state toward their reference values. (The experimenter also has quite a few other controlled perceptions, mostly relating to the proper functioning of the equipment, but we don't need to be concerned much with those).

Even if (1) is satisfied, (2) may not be, and good experimental technique ordinarily implements "the test for the controlled variable" to determine whether the subject is actually doing what the experimenter hoped for (controlling the intended variable at the intended reference value). Poor experimental technique does not. Poor experimental setup may make it impossible (e.g. a setup in which the subject finds it hard to provide an environmental output precisely related to the output of the control unit that controls the experimental variable).

I'll give an example of a real case in which (1) was satisfied, but (2) was not, because the subject misinterpreted what variable the experimenter wanted the subject to control. A colleague of mine, when he was an undergraduate, was a subject in an experiment that was intended to assess tracking ability, as measured by the proportion of time the subject could keep a stylus on a metal spot on a rotating disk (maybe there was mirror viewing or some such). My colleague interpreted the instruction "Keep the stylus on the small metal spot as long as you can" as an intelligence test, and instead of tracking the spot while the disk rotated, he dismounted the disk and laid the stylus on the metal spot. He was controlling a variable he thought the experimenter had intended, but it was not the actual variable the experimenter had intended. The experimenter could test for this by observing first the record of the tracking error, and when that was obviously not a track any but a superhuman could attain, he then directly observed the placement of the disk and stylus. The experimenter might not have been able to know what variable my colleague was controlling for, but "The Test" showed clearly that it was not the variable the experimenter had intended. (I would say that this experimenter's technique was not very good, but at least it was adequate to ensure that my colleague's data were not included in the experimenter's interpretation of the results).

"The Test for the controlled variable" has several components, which can be stated in the form of questions.
1. Is there a possibility for a control system to perceive the effect of a disturbance to the putative environmental correspondent of the controlled variable? (Let's call that the ECV -- "environmental controlled variable" -- for short, even though the environmental variable is NOT the controlled variable).?
2. Is there a possibility for the output of a control system to influence the ECV?
3. When the disturbance is applied, is its effect on the ECV less than would be expected?
4. Can the lessening of the effect of the disturbance on the ECV be attributed to the output of the putative control system? In other words, can the output of the control system be correlated with the disturbance?

[Parenthetical note: In the past, we have used "CEV" or "Complex Environmental Variable" to represent the function of environmental variables that truly corresponds to the perceptual input function of a control system. I use "ECV" here to distinguish a guess as to the CEV from the actual CEV.]

In that form, "The Test" only determines that the perception corresponding to the ECV is correlated with the controlled perception. It does not determine that the corresponding perception IS the controlled perception (as Rick showed in his comparisons of different possible controlled variables in when the length and width of a rectangle were independently disturbed and the subject was trying to control (I think) "size" -- it's not on Mindreadings.com).

One of the well known issues of experimental technique in psychophysics is to determine whether the subject is actually doing the task the experimenter intends -- controlling the "right" variable. The "clever Hans" phenomenon illustrates an early example of where no such test was done until fairly late in the game, when the horse Hans was shown to be controlling for a relationship with the handler's movements rather than to the intended "stimulus". In psychophysical detection studies, experimenters go to considerable lengths to ensure that nothing correlated with the "stimulus" is perceptible to the subject -- no clicks when a signal goes on, no visible signs of experimenter movement (and the experimenter should not know what stimulus is presented on a particular trial, anyway). Good experimental technique is, in part, aimed at ensuring that the disturbance provided by the stimulus in only to the variable intended by the experimenter.

In psychophysical experiments, the controlled variable is usually quite simply described, and it is usually a relationship variable, such as to match a selection of a response with a particular value of the stimulus presentation -- e.g. "did the left or the right light come on?". Since the question of interest is some property of the perceptual channel that leads to one input of the relationship control perceptual input function, good technique requires that the pathway from the output function of the relationship control system to the outer world be as simple as possible. In particular, since most perceptual pathways have much higher bandwidths than the muscular part of the output pathway, if the perceptual bandwidths are under study, it is often impractical to have feedback through the external environment (which means that tracking methods cannot be used). The feedback pathway must, in those cases, be internal (in imagination), and not visible to the experimenter.

So now we have to ask: If the Test requires that the output of the control system be observed and checked for its relationship with the disturbance, but the experimental question cannot allow for externally observable tracking, how can the Test be performed in practice? The answer must be that the externally observed output has to be very nearly perfectly correlated with some phase (ordinarily the final phase) of the output of the tested control system. If that can be guaranteed -- and good experimental technique demands that this be tested initially by making the disturbance-creating "stimuli" such as to make the control very easy -- then the output visible to the experiment can reasonably be taken as a surrogate for the output of the control system whose input pathway is the object of study.

As an example, imagine a test of the ability to detect a 500 Hz tone burst in the presence of some masking noise. Good technique suggests that the experimenter should first present to the subject similar tone bursts that are loud enough for the subject to hear them clearly, and to ask "Do you hear the tone bursts?". The subject may answer that she is not be able to hear them, or that she does not know what it is that should be heard. That would generate error in a control system in the experimenter, to be corrected, perhaps, by checking the wiring or by instructing the subject more carefully, until the subject says: "Yes, I hear them clearly now".

Next, the experimenter checks whether the subject knows what is to be controlled, and how to report the output of the control system: "You will hear four noise bursts, and in one of them you will hear the kind of tone I just showed you. Your job is to determine which of the four bursts had the tone in it, and push the button with the corresponding number." So then the experiment tries giving four bursts, with a clearly audible tone in one of them. If the subject presses the correct button, the experimenter's relevant control system experiences no error, but otherwise, the experimenter does whatever is needed to get the subject to press the correct button and have the result recorded. Usually the experimenter would provide several stimuli to be sure that the output pathway is accurately representing what the subject should have matched to the interval containing the perceived tone.

From the subject's viewpoint, the loop is closed by the experimenter's telling him that the response was correct or that it was wrong. When the subject has learned to press correctly and consistently the button that the experimenter wants to have associated with each presentation interval, neither the subject nor the experimenter will have error in the relevant control systems.

Next, the experimenter will decrease the loudness of the tone until it is quite difficult to hear, and make sure that the subject presses a button on every trial, by seeing whether this happens and letting the subject know when a trial occurs but no button response was observed. Maybe the subject says: "But I did push the button every time" or "I didn't push a button because I didn't hear a tone". In that case, perhaps the subject needs instruction about pushing every time, or about pushing hard enough, or maybe there is a faulty connection in the wiring. Either way, there's an error that needs to be corrected in one of the experimenter's control systems.

Finally, when the tight relationship between observable button pushes and the output of the subject's relationship control system has been established, the experiment proper can begin, and the experimenter can start investigating whatever property of the input perceptual pathway is of interest.

I don't think there would be any difference to this technique if the experimenter realized that what he was doing was making sure of the controlled variable, and of the relationship between the output of that control system and the visible environmental output by the subject. It's possible, however, that in some cases a knowledge of PCT might alter the experimenter's interpretation of the results.

···

=========

The next question is whether this kind of experiment can provide any information about the perceptual pathway between the physical sensor and the corresponding input to the control system for the variable being controlled. For some properties, it can. For example, in the kind of experiment used as an example above, if the noise burst has a spectral gap around the 500Hz tone, changing the upper and lower bounds of the gap will influence the correlation between the stimulus and the response for some intensities of the tone. If the gap is narrow, to be heard the 500Hz signal tone will need to be louder than if the gap is wide. Varying the bounds of the gap tells the experimenter something about the acceptance bandwidth of whatever internal mechanism separates the tone from other acoustic input.

As another example in the same vein, if the noise spectrum is flat, the actual level at which the tone is correctly reported x% of the time allows a determination of the efficiency of the internal signal pathway as compared to the mathematical ideal observer. As a third example, the inverse of the first, if a tone at 500+-n Hz is present in all four intervals, how does the value of n affect the intensity of the tone at 500Hz required to allow a correct report x% of the time? This latter would seem to be studying the same property as the first example -- the acceptance bandwidth of the signal detection mechanism, but in fact it gives a different estimate for that bandwidth; this difference tells something about how the detection mechanism works.

Yes, a lot can be determined about perceptual pathways from conventional psychophysical experiments done with good technique.

Martin

From Bill Powers (2010.07.08.1547 MDT)]

Martin Taylor 2010.07.05.00.05 --

MMT: I've been thinking about this more and more over the course of this thread, and I'm beginning to see that a good part of what is taught as good experimental practice in psychophysics actually is performing the test for the controlled variable, whether the experimenters know it or not (and most don't).

... Of course, most practitioners will not realize that they are doing "The Test", any more than most people realize they are always controlling their inputs, but it seems to me that good practice nevertheless usually, if not always, results in the experimenter performing the Test for the controlled variable.

Experimental technique has developed by normal evolutionary methods -- keep what works, change what doesn't -- and in my opinion that evolutionary development has approached what would have been prescribed from first principles by PCT analysis.

Very nice discussion, Martin. Have you considered possibly writing an article introducing PCT to psychophyics along these lines? From inside our field we can see good reasons for each of the tests, precautions, etc. that you describe here, but I doubt that even a good experimentalist knows exactly why each of them evolved. PCT offers a framework in which each of them has a specific function based on a model of how behavior works. It might turn out that having a systematic theory behind the practices of psychophysics could help to make sure that the most important conditions are always met, instead of the proper procedures being treated as a somewhat arbitrary ritual.

Do you have access to any good sources of experimental data relating to the question of log or power-law perceptions? Did Stevens ever publish any of the raw data?

Best,

Bill P/

[Martin Taylor 2010.07.08.23.29]

From Bill Powers (2010.07.08.1547 MDT)]

Martin Taylor 2010.07.05.00.05 --

MMT: I've been thinking about this more and more over the course of this thread, and I'm beginning to see that a good part of what is taught as good experimental practice in psychophysics actually is performing the test for the controlled variable, whether the experimenters know it or not (and most don't).

... Of course, most practitioners will not realize that they are doing "The Test", any more than most people realize they are always controlling their inputs, but it seems to me that good practice nevertheless usually, if not always, results in the experimenter performing the Test for the controlled variable.

Experimental technique has developed by normal evolutionary methods -- keep what works, change what doesn't -- and in my opinion that evolutionary development has approached what would have been prescribed from first principles by PCT analysis.

Very nice discussion, Martin. Have you considered possibly writing an article introducing PCT to psychophyics along these lines? From inside our field we can see good reasons for each of the tests, precautions, etc. that you describe here, but I doubt that even a good experimentalist knows exactly why each of them evolved. PCT offers a framework in which each of them has a specific function based on a model of how behavior works. It might turn out that having a systematic theory behind the practices of psychophysics could help to make sure that the most important conditions are always met, instead of the proper procedures being treated as a somewhat arbitrary ritual.

Interesting idea. I've never considered such a thing. If I did that, it probably wouldn't look very like the message on which you are commenting, since that was addressed to PCT experts, and a discussion for psychophysicists would have to explain the perceptual control basis.

Do you have access to any good sources of experimental data relating to the question of log or power-law perceptions? Did Stevens ever publish any of the raw data?

I never was much interested in Stevens when he was around (despite having had that tea with him), so I never investigated his work. In my youth I had the opinion that among his hundreds of publications, he had written three papers, many times each. One was the power-law, which I never took seriously, and I've forgotten what the other two were.

As for experimental data on log or power-law perception, I don't know of any specifically, but I have a vague memory from decades ago that there was something showing that in some facet of the neural system the intensity of light input was accurately logarithmic over 7 log units. It might have been related to an article in (I think possibly) Kybernetik that showed (I think individual) neurons to be capable of calculating correlation very accurately. But this was so long ago, and at the time it wasn't central to my interests, that my memory is extremely hazy and quite probably wrong.

I imagine that the current state of neurophysiological knowledge would have much better answers than anything from decades ago. But even with that knowledge, it wouldn't tell you about the subjective perception -- only about the levels of the neural signals, which are the signal values in the PCT models, but not necessarily the values in consciousness.

Martin

[From Rick Marken (2010.07.08.2210)]

�Martin Taylor (2010.07.05.00.05)--

Rick Marken (2010.07.04.0840)--

Yes, but since all psychophysical research is based on a an open-loop
causal model of the systems under study, those who study psychophysics
(which used to include me;-) can't really be learning much about the
"capabilities of the processing channels" of these systems using
conventional methods. Agreed?

As I said earlier [Martin Taylor 2010.07.04.17.02], definitely not agreed.

I've been thinking about this more and more over the course of this thread,
and I'm beginning to see that a good part of what is taught as good
experimental practice in psychophysics actually is performing the test for
the controlled variable, whether the experimenters know it or not (and most
don't).

Yes, I agree. Instructions tell the subject what to control and the
experimenter must make sure that the subject is indeed following the
instructions.

In any experiment, both the experimenter and the subject are controlling
many variables, but from the experimenter's point of view only two of the
subject's controlled variables matter very much, namely (1) that the subject
is controlling for perceiving that the experimenter perceives that the
subject is performing as the experimenter wishes, and (2) that the subject
is in fact acting to control against disturbances in the way the
experimenter intended.

I can't think of a situation where I have ever controlled for (1). I
don't even know how one would control for it. It seems to me like all
the experimenter ever really controls for is (2). As an experimenter
I want to be fairly confident that the subject is controlling what is
to be controlled (such as the relationship between their button press
and the onset of a light in a reaction time experiment). As the
experimenter I can see that the subject is controlling this variable
if, in pilot runs, the subject is properly compensating for
disturbances, such as the light going on and off, by making the
appropriate responses, such as pressing or not pressing the button.

I'll give an example of a real case in which (1) was satisfied, but (2) was
not

The experimenter could have told that (2) was not satisfied by just
watching the subject. I can't see how the experimenter could have
known that (1) was satisfied. In fact, your friend was indeed trying
to do as instructed (as he understood it) but what the experiment saw
could have easily been interpreted as the subject acting to violate
the experimenter's wishes. But, again, I can't imagine an experimenter
caring whether the subject is trying to cooperate or not. The
experimenter just wants the subject to control what the experimenter
wants the subject to control.

In psychophysical experiments, the controlled variable is usually quite
simply described, and it is usually a relationship variable

Yes! I'd say the controlled variable in psychophysical experiments is
always a relationship between a stimulus and response.

Since the
question of interest is some property of the perceptual channel that leads
to one input of the relationship control perceptual input function, good
technique requires that the pathway from the output function of the
relationship control system to the outer world be as simple as possible.

Yes, a button press is a pretty simple path from output (muscle force)
to outside world (button depression).

In particular, since most perceptual pathways have much higher bandwidths than
the muscular part of the output pathway, if the perceptual bandwidths are
under study, it is often impractical to have feedback through the external
environment (which means that tracking methods cannot be used). The feedback
pathway must, in those cases, be internal (in imagination), and not visible
to the experimenter.

But if the feedback path exists only in imagination how can the
subject control the variable that the subject has been instructed to
control? In fact, in all experiments, there is always an overt
response (like a button press) that affects the state of the
controlled variable. The output may occur after a lot of imagination,
but it occurs and that's how the experimenter (and the subject!) can
see that the variable the subject was instructed to control is,
indeed, under control.

So now we have to ask: If the Test requires that the output of the control
system be observed and checked for its relationship with the disturbance,
but the experimental question cannot allow for externally observable
tracking, how can the Test be performed in practice? The answer must be that
the externally observed output has to be very nearly perfectly correlated
with some phase (ordinarily the final phase) of the output of the tested
control system. If that can be guaranteed -- and good experimental technique
demands that this be tested initially by making the disturbance-creating
"stimuli" such as to make the control very easy -- then the output visible
to the experiment can reasonably be taken as a surrogate for the output of
the control system whose input pathway is the object of study.

I have no idea what this means.

As an example, imagine a test of the ability to detect a 500 Hz tone burst
in the presence of some masking noise. Good technique suggests that the
experimenter should first present to the subject similar tone bursts that
are loud enough for the subject to hear them clearly...

Next, the experimenter checks whether the subject knows what is to be
controlled, and how to report the output of the control system: "You will
hear four noise bursts, and in one of them you will hear the kind of tone I
just showed you. Your job is to determine which of the four bursts had the
tone in it, and push the button with the corresponding number."

Right. The controlled variable is the relationship between the
sequential position of the tone and the sequential position of the
button: the reference state is a match between the two, i.e., tone in
position 2, press button 2.

From the subject's viewpoint, the loop is closed by the experimenter's
telling him that the response was correct or that it was wrong.

Doesn't the subject have to know what they did also? I would think the
subject would have to know that button 2 was pushed when the tone was
perceived to be in position 2. The loop has to be closed with respect
to the variable the subject is controlling or the subject can't
control that variable. If the subject is not trying to control the
relationship between tone and button position, then the experimenter's
feedback is meaningless.

When the
subject has learned to press correctly and consistently the button that the
experimenter wants to have associated with each presentation interval,
neither the subject nor the experimenter will have error in the relevant
control systems.

Yes, agreed.

Finally, when the tight relationship between observable button pushes and
the output of the subject's relationship control system has been
established, the experiment proper can begin, and the experimenter can start
investigating whatever property of the input perceptual pathway is of
interest.

He will think he is investigating a property of the perceptual
pathway. But since he is dealing with a closed loop system that is
controlling the relationship between S and R, the observed
relationship between S and R will depend on the nature of the feedback
connection from R to to the controlled variable.

I don't think there would be any difference to this technique if the
experimenter realized that what he was doing was making sure of the
controlled variable, and of the relationship between the output of that
control system and the visible environmental output by the subject. It's
possible, however, that in some cases a knowledge of PCT might alter the
experimenter's interpretation of the results.

Yes, it certainly would.

The next question is whether this kind of experiment can provide any
information about the perceptual pathway between the physical sensor and the
corresponding input to the control system for the variable being controlled.
For some properties, it can. For example, in the kind of experiment used as
an example above, if the noise burst has a spectral gap around the 500Hz
tone, changing the upper and lower bounds of the gap will influence the
correlation between the stimulus and the response for some intensities of
the tone. If the gap is narrow, to be heard the 500Hz signal tone will need
to be louder than if the gap is wide. Varying the bounds of the gap tells
the experimenter something about the acceptance bandwidth of whatever
internal mechanism separates the tone from other acoustic input.

Yes, there will be an observed relationship between disturbance (in
this case, signal spectrum) and output (button pressed). But the
_nature_ of this relationship will depend on characteristics of the
feedback path from output to controlled input, not on characteristics
of the perceptual system.

Yes, a lot can be determined about perceptual pathways from conventional
psychophysical experiments done with good technique.

I am not convinced. I think psychophysical experiments, like all
experiments in psychology, tell you very little about the nature of
the system under study if you just take observed relationships between
S and R as reflecting an open-loop causal path through the system.
Since you yourself recognize that subjects are controlling perceptions
(such as the relationship between S and R) in psychophysical
experiments then you must realize that the nature of these
relationships depend on the nature of the feedback path from R to
controlled variable, not on the nature of the perceptual functions
between controlled variable and R.

This fact is not called a behavioral illusion because the relationship
between S and R looks illusory. It's called an illusion because the
relationship between S and R looks compellingly like S is causing R
via the organism. You can only tell that this apparent causal
connection is an illusion by analyzing the situation using control
theory. Scientific psychologists (psychophysicists included) would not
have been laboring under this illusion for over 100 years if it were
obviously an illusion. What Bill discovered is extremely hard for
scientific psychologists to believe. That's why they haven't bought
into PCT. They just can't believe (to say nothing of not _wanting_ to
believe) that they have been studying an illusion for the last
century.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com

[Martin Taylor 2010.07.09.17.31]

[From Rick Marken (2010.07.08.2210)]
Martin Taylor (2010.07.05.00.05)--
Rick Marken (2010.07.04.0840)--

In any experiment, both the experimenter and the subject are controlling
many variables, but from the experimenter's point of view only two of the
subject's controlled variables matter very much, namely (1) that the subject
is controlling for perceiving that the experimenter perceives that the
subject is performing as the experimenter wishes, and (2) that the subject
is in fact acting to control against disturbances in the way the
experimenter intended.
I can't think of a situation where I have ever controlled for (1). I
don't even know how one would control for it.
You mean you trust the data you get from subjects who don't try to

do it right, or who intend to cross you up by inserting random wrong
answers to screw up your data? I wouldn’t. I’d like to see that the
subjects at least seem as though they are honestly trying to do
their best to follow instructions. Mind you, (1) could be satisfied
by a deliberately deceptive subject who is controlling for
perceiving that the experimenter perceives the subject to be
performing as the experimenter wishes, while not actually doing so.
In that case, the subject would have to also be deceptive about (2),
and generate data that would be plausible enough to satisfy the
experimenter. I think that in most cases, such deception would be
hard for a subject to carry off, so in the normal course of events,
(1) is pretty easy to control for, if only by rejecting a subject
that seems to be careless or resentful.

It seems to me like all
the experimenter ever really controls for is (2). As an experimenter
I want to be fairly confident that the subject is controlling what is
to be controlled (such as the relationship between their button press
and the onset of a light in a reaction time experiment). As the
experimenter I can see that the subject is controlling this variable
if, in pilot runs, the subject is properly compensating for
disturbances, such as the light going on and off, by making the
appropriate responses, such as pressing or not pressing the button.
That's more or less true, but a clever deceptive subject may be able

to fake it, which is why experimental protocols are usually designed
so that most fakery attempts will show up as clearly anomalous.
Controlling for perceiving the subject to be trying to do the right
thing is a pretty good start in ensuring that (2) will be satisfied,
but as my colleague’s experience as a subject illustrates, it isn’t
sufficient by itself.


I'll give an example of a real case in which (1) was satisfied, but (2) was
not
The experimenter could have told that (2) was not satisfied by just
watching the subject.
Most psychophysical experiments are set up so that subject and

experimenter can’t see each other. At any rate, this one was.

I can't see how the experimenter could have
known that (1) was satisfied. ... again, I can't imagine an experimenter
caring whether the subject is trying to cooperate or not. The
experimenter just wants the subject to control what the experimenter
wants the subject to control.
Well, I most certainly would care whether the subject was intending

to cooperate! As for perceiving whether they were, that’s normal
social interaction. There are lots of indications as to whether
someone you are talking to is sincere and cooperative, or evasive
and deceptive, unless the other is a well practiced actor and/or
deceiver.


In psychophysical experiments, the controlled variable is usually quite
simply described, and it is usually a relationship variable
Yes! I'd say the controlled variable in psychophysical experiments is
always a relationship between a stimulus and response.
"Always"? You are prepared to narrow down the possibilities to that

extent? Why?


But if the feedback path exists only in imagination how can the
subject control the variable that the subject has been instructed to
control?
I don't understand the question. Wherin is there a problem? The

(let’s say) relationship control system’s output function generates
different imagined values until it finds one that satisfies the
relationship reference value. That’s control, isn’t it?

I have the impression that Bill and I are on the same page in this.

Here are a couple of quotes from earlier:

···
[From Bill Powers (12010.06.17.1016 MDT)]The subject presumably

imagines different numbers until their perceived magnitudes on the
logarithmic scale match the log of stimulus magnitude at the inputs
of
the relationship detector.

Of course the experimenter knows only of the final result and

doesn’t see
any preliminary activity that happens in imagination. So it looks
like a
simple stimulus-response sequence and it can then be analyzed as if
it
really were one.

[From Bill Powers (2010.06.21.0550 MDT)]...it's not clear where (at

what level) the
imagination loop is closed. I think people will often think to
themselves
exactly what you said: two, no it’s more like three, well say 2 and
a
half. Then they say the final number aloud. Sometimes they even
mutter
those thoughts to themselves, not yet intending to deliver the
result to
someone else – “thinking out loud.”

-----------------
In fact, in all experiments, there is always an overt
response (like a button press) that affects the state of the
controlled variable.
It's very doubtful that the button press "interval 2" affects the

subject’s perception that the tone had been in interval 2. The
button press “interval 2” is perceptible by the experimenter, and by
the subject, but I would be extremely surprised if hitting the
button caused the subject to change from thinking the tone had been
in interval 3 to thinking it had been in interval 2.

The output may occur after a lot of imagination,
but it occurs and that's how the experimenter (and the subject!) can
see that the variable the subject was instructed to control is,
indeed, under control.
That much, I can agree with.

Martin

[From Bill Powers (2010.07.09.1720 MDT)]

Martin Taylor 2010.07.09.17.31 –

Marken: In fact, in all
experiments, there is always an overt
response (like a button press) that affects the state of the
controlled
variable.

MMT: It’s very doubtful that the button press “interval 2”
affects the subject’s perception that the tone had been in interval 2.

BP: That isn’t the controlled variable that Rick is referring to. In the
model he and I have been thinking of, the controlled variable is the
state of the relationship between the button press and the tone, which is
affected by both the press and the tone. The relationship is disturbed by
occurrance of the tone and the button press restores the relationship to
its reference condition.
The perceived relationship might be something like (A XOR B) with a
reference level of false.

not[(A and not-B) or (not-A and B)]

which is equivalent to requiring that both A and B be true (or
false) at the same time – or something similar such as A times B. In the
latter case, A*B would be positive if A and B are both positive or both
negative, and otherwise negative.

Best,

Bill P.

[Fom Rick Marken (2010.07.09.1650)]

Martin Taylor (2010.07.09.17.31)--

Rick Marken (2010.07.08.2210)]

RM: But if the feedback path exists only in imagination how can the
subject control the variable that the subject has been instructed to
control?

I don't understand the question. Wherin is there a problem? The (let's say)
relationship control system's output function generates different imagined
values until it finds one that satisfies the relationship reference value.
That's control, isn't it?

I'd call it imagination; but imaginary control would be ok too. Real
control -- the kind that we have to do when we get out of our
armchairs -- is a process of maintaining perceived aspects of the
external environment in pre-specified states. In an experiment, the
perceived aspect of the environment that is to be controlled is the
relationship between a stimulus and a response. Some of the process of
control might take place in imagination but at some point you have to
do something that either brings the controlled variable to the
reference state or doesn't. That's how both you and the experimenter
know whether or not you are keeping the variable under control.

Controlling in imagination is not really controlling because there are
no constraints; you're guaranteed to succeed (bring the controlled
variable to the reference state). So I can imagine that I made the
response that keeps the relationship between stimulus and response
under control, but that's quite a different than actually doing
something in the world where the best laid plans... well, you know.

Best

Rick

···

--
Richard S. Marken PhD
rsmarken@gmail.com
www.mindreadings.com