Behavioral Illusion (was Re: p-hacking)

[rsmarken]

[From Rick Marken (2017.03.27.1130)]

Powers1978.pdf (1.46 MB)

···

Leeanne Wright (2017.03.27.1.39 AEST)

Hi Rick, Martin & Bruce,

LW:  I continue to be fascinated by this thread! The topic of significance-testing in psychology seems to be very salient at the present time. Therefore, good research fodder, perhaps! Maybe, the time has finally come…?Â

RM: At Martin&‘s behest I am continuing this discussion as a new thread with the subject head “Behavioral Illusion” even though I think that the behavioral illusion described in Powers’ 1978 Science article (which I have attached; the description of the behavioral illusion begins on p. 425) is directly relevant to significance testing that was being discussed in the “p-hacking” thread. The reason for the relevance is this: Significance testing is most often used in psychology to decide whether one can reject with a sufficiently small probability of being wrong (making a Type I error) the null hypothesis that the independent variable in an experiment did not have an effect on the dependent variable. When you reject the null hypothesis (with only a small probability of being wrong) you are deciding that the independent variable actually does have an effect on the dependent variable. The relevance of this to the behavioral illusion is that Powers shows that, when the systems under study are control systems (called N-systems in the paper), any apparent effect of an independent variable on a dependent variable is an illusion.Â

RM: The “p-hacking” thread noted some ways that significance testing can be abused. One way – the way that disturbed me most when I was a conventional psychologist – was to interpret the significance level as a measure of the probability of the null hypothesis being true. If the results are significant at the .001 level, that is taken to mean that null hypothesis more unlikely to be true than if the results are significant at the .01 level. Thus, the significance level is taken to be an inverse measure of how likely the alternative hypothesis is to be true. As Bruce A. so clearly pointed out, this is a very misleading way of reporting the results of a significance test; the significance level tells only the probability that your decision to reject the null hypothesis is a false alarm (a Type I error). Also, significance testing can be “hacked” in various ways, one being to just eliminate deviant data points and another being to keep increasing the sample size until your statistic is “significant” (with a large enough sample you can always get significance).Â

RM: But the behavioral illusions makes all the problems with significance testing moot. That’s because once you know that you are trying to understand the behavior of a living control system, the aim of your research is no longer to see if various independent variables (stimuli) have an effect on various dependent variables (responses). The aim of your research is to determine the variables that the system controls and how it controls them. The fundamental problem with significance testing, from a PCT perspective, is that is is based on a causal model of behavior; a model that says that the input to the system (the effect of the independent variable) is the ultimate cause of the output of that system (the dependent variable). That is, significance testing in psychology is based on the wrong model of the nature of living systems.

RM: So what I recommend for going forward on this “behavioral illusion” thread is that those of us interested in doing PCT research (which may be just you and me Leeanne;-) should read (or re-read) the attached 1978 Science paper by Powers and then discuss the implications of Powers’ analysis for research in psychology. Let’s see what’s going on in some experiments where the results were analyzed using significance tests and see how we might interpret those results in the context of the analysis in Powers 1978 Science paper. Then let’s propose some experiments consistent with the rather abstract proposals for research presented in that paper (as Experiments 5 and 6).Â

Best regards

Rick

–

Richard S. MarkenÂ

"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery

[MartinT]

[Martin Taylor 2017.03.27.15.18]

[From Rick Marken (2017.03.27.1130)]

Thanks for doing this, and for attaching the Powers 1978 Psych

Review paper. We can discuss the Behavioural Illusion without mixing
it up with the p-hacking problem. It’s not as simple as you make
out. I was in process of drafting something on the question to start
the thread when “the stimulus” of your post showed up, and am
posting this “response” now because I am not sure when I will be
ready to post it. I hope it will address your concerns, without
referring to them directly.

I think some of what follows belongs more in the p-hacking thread,

but I do acknowledge the connection to the Behavioural illusion. I’m
not going to respond to it in any detail here, for fear of leading
this thread away from the topic of the Behavioural Illusion, just as
introduction of the Behavioural Illusion redirected the thread on
p-hacking.

Yes, that's what's wrong with it.

No. It's the mechanism of the effect that is an issue, not the

existence of the effect under the conditions of the experiment. It’s
the following interpretation that is the “illusion”.

All true, and should be copied to the p-hacking thread. As I said

earlier, my rejoinder to the editor who rejected my paper because it
didn’t include a significnce test was “With enough data you can find
the effect of the orientation of the rings of Saturn on the curl in
a puppy-dog’s tail.”

OK if you replace the initial "The" with "One possible".

Basically true, but often taken to an unwarranted extreme.

The basic problem with that is one that has often been discussed on

CSG-L and CSG-net: the relevant data for a PCT analysis are
unavailable.

Your own demos are good examples, but I think you have to go both

simpler and more carefully than Powers did, if you want to make a
research program out of it. To begin with, you have to understand
how to reconcile the Behavioural illusion with the apparent success
of modelling and parameter optimization. I go into that in what I
was drafting.

Martin

···

            Leeanne Wright

(2017.03.27.1.39 AEST)

Hi Rick, Martin & Bruce,

              LW:  I continue to be fascinated by this thread! 

The topic of significance-testing in psychology seems
to be very salient at the present time. Therefore,
good research fodder, perhaps! Maybe, the time has
finally come…?Â

          RM: At Martin's behest I am continuing this discussion

as a new thread with the subject head “Behavioral
Illusion”

          even though I think that the behavioral illusion

described in Powers’ 1978 Science article (which I
have attached; the description of the behavioral illusion
begins on p. 425) is directly relevant to significance
testing that was being discussed in the “p-hacking”
thread. The reason for the relevance is this: Significance
testing is most often used in psychology to decide whether
one can reject with a sufficiently small
probability of being wrong (making a Type I error) the
null hypothesis that the independent variable in an
experiment did not have an effect on the
dependent variable.

          When you reject the null hypothesis (with only a small

probability of being wrong) you are deciding that the
independent variable actually does have an effect on the
dependent variable. The relevance of this to the
behavioral illusion is that Powers shows that, when the
systems under study are control systems (called N-systems
in the paper), any apparent effect of an
independent variable on a dependent variable is an
illusion.

          RM: The "p-hacking" thread noted some ways that

significance testing can be abused. One way – the way
that disturbed me most when I was a conventional
psychologist – was to interpret the significance level as
a measure of the probability of the null hypothesis being
true. If the results are significant at the .001 level,
that is taken to mean that null hypothesis more unlikely
to be true than if the results are significant at the .01
level. Thus, the significance level is taken to be an
inverse measure of how likely the alternative hypothesis
is to be true. As Bruce A. so clearly pointed out, this is
a very misleading way of reporting the results of a
significance test; the significance level tells only the
probability that your decision to reject the null
hypothesis is a false alarm (a Type I error). Also,
significance testing can be “hacked” in various ways, one
being to just eliminate deviant data points and another
being to keep increasing the sample size until your
statistic is “significant” (with a large enough sample you
can always get significance).

          RM: But the behavioral illusions makes all the problems

with significance testing moot. That’s because once you
know that you are trying to understand the behavior of a
living control system, the aim of your research is no
longer to see if various independent variables (stimuli)
have an effect on various dependent variables (responses).
The aim of your research is to determine the variables
that the system controls and how it controls them.

          The fundamental problem with significance testing,

from a PCT perspective, is that is is based on a causal
model of behavior; a model that says that the input to the
system (the effect of the independent variable) is the
ultimate cause of the output of that system (the dependent
variable). That is, significance testing in psychology is
based on the wrong model of the nature of living systems.

          RM: So what I recommend for going forward on this

“behavioral illusion” thread is that those of us
interested in doing PCT research (which may be just you
and me Leeanne;-) should read (or re-read) the attached
1978 Science paper by Powers and then discuss the
implications of Powers’ analysis for research in
psychology. Let’s see what’s going on in some experiments
where the results were analyzed using significance tests
and see how we might interpret those results in the
context of the analysis in Powers 1978 Science paper.

          Then let's propose some experiments consistent with the

rather abstract proposals for research presented in that
paper (as Experiments 5 and 6).Â

[rsmarken]

[From Rick Marken (2017.03.28.1055)]

···

Martin Taylor (2017.03.27.15.18)–

MT: No. It's the mechanism of the effect that is an issue, not the

existence of the effect under the conditions of the experiment.

RM: Right. Conventional research assumes that the mechanism of the effect is mediated by the organism; the behavioral illusion shows that the mechanism of the effect is mediated by the feedback function. That’s why I call it an “apparent” effect. What is apparent (and illusory) is that the observed effect of the independent on the dependent variable reveals something about the the organism under study when, in fact, it mainly reveals something about the environment in which the behavior is taking place.Â

RM: Powers puts it this way in the “behavioral illusion” section of the 1978 paper: “This comparison [between the closed loop and open loop equations relating independent (disturbance) to a dependent (output) variable – RM] reveals a behavioral illusion of such significance that one hesitates to believe it could exist. If one varies a distal stimulus q.d and observes that a measure of behavior q.o shows a strong, regular dependence of q.d there is certainly a temptation to assume that the form of the dependence reveals something about the organism. Yet the comparison that we have just seen indicates that the form of the dependence may reflects only properties of the local environment”.Â

Â

MT: Basically true, but often taken to an unwarranted extreme.

 RM: Perhaps this thread will reveal what that unwarranted extreme is.Â

MT: The basic problem with that is one that has often been discussed on

CSG-L and CSG-net: the relevant data for a PCT analysis are
unavailable.

 RM: I think it’s worth it to try again. Bill’s rather devastating analysis in the 1978 Psych Review paper (sorry I mistakenly said Science paper earlier) was rather abstract. I think it might help to get his message across (after over 30 years of it being ignored by scientific psychology) if we came up with a paper explaining the implications of Bill’s analysis in the context of real psychology experiments, particularly well-known classics.

MT: Your own demos are good examples, but I think you have to go both

simpler and more carefully than Powers did, if you want to make a
research program out of it.

RM: Actually, my goal is just to show in a more concrete way than Bill did in the 1978 paper why the conventional approach to research tells you very little about the mechanisms of behavior.Â

Â

MT: To begin with, you have to understand

how to reconcile the Behavioural illusion with the apparent success
of modelling and parameter optimization. I go into that in what I
was drafting.

RM: I agree and I look forward to seeing your paper.Â

Best regards

Rick

–
Richard S. MarkenÂ

"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery

          RM: When you reject the null hypothesis (with only a small

probability of being wrong) you are deciding that the
independent variable actually does have an effect on the
dependent variable. The relevance of this to the
behavioral illusion is that Powers shows that, when the
systems under study are control systems (called N-systems
in the paper), any apparent effect of an
independent variable on a dependent variable is an
illusion.

          RM: The fundamental problem with significance testing,

from a PCT perspective, is that is is based on a causal
model of behavior; a model that says that the input to the
system (the effect of the independent variable) is the
ultimate cause of the output of that system (the dependent
variable). That is, significance testing in psychology is
based on the wrong model of the nature of living systems.

          RM: So what I recommend for going forward on this

“behavioral illusion” thread is that those of us
interested in doing PCT research (which may be just you
and me Leeanne;-) should read (or re-read) the attached
1978 Science paper by Powers and then discuss the
implications of Powers’ analysis for research in
psychology. Let’s see what’s going on in some experiments
where the results were analyzed using significance tests
and see how we might interpret those results in the
context of the analysis in Powers 1978 Science paper.

          RM: Then let's propose some experiments consistent with the

rather abstract proposals for research presented in that
paper (as Experiments 5 and 6).Â

[John_and_Leeanne_Wri]

Hi Rick and Martin.,

I am still following with enthusiasm, although I will probably be offline for a few weeks with school holidays, my two children’s birthdays (and parties), parents visiting, uni assignments, not to mention a week long Intensive School that I need to travel interstate for. If you had asked me 3 years ago would I be interested in this, I would have said an emphatic “No�. However, now the answer to that question is an emphatic “YES!�.

[From Rick Marken (2017.03.28.1055)]

LW: I think I understand this but could you perhaps explain in a bit more detail…

LW: That would certainly catch people’s attention, especially as it is something they are familiar with. Although, in this case I think it is the familiarity that would be important….

LW: Me too. Really helpful explanations for me in this thread. Thanks!

Regards

Leeanne

···

Martin Taylor (2017.03.27.15.18)–

MT: No. It's the mechanism of the effect that is an issue, not the

existence of the effect under the conditions of the experiment.

RM: Right. Conventional research assumes that the mechanism of the effect is mediated by the organism; the behavioral illusion shows that the mechanism of the effect is mediated by the feedback function. That’s why I call it an “apparent” effect. What is apparent (and illusory) is that the observed effect of the independent on the dependent variable reveals something about the the organism under study when, in fact, it mainly reveals something about the environment in which the behavior is taking place.

RM: Powers puts it this way in the “behavioral illusion” section of the 1978 paper: “This comparison [between the closed loop and open loop equations relating independent (disturbance) to a dependent (output) variable – RM] reveals a behavioral illusion of such significance that one hesitates to believe it could exist. If one varies a distal stimulus q.d and observes that a measure of behavior q.o shows a strong, regular dependence of q.d there is certainly a temptation to assume that the form of the dependence reveals something about the organism. Yet the comparison that we have just seen indicates that the form of the dependence may reflects only properties of the local environment”.

MT: Basically true, but often taken to an unwarranted extreme.

          RM: When you reject the null hypothesis (with only a small

probability of being wrong) you are deciding that the
independent variable actually does have an effect on the
dependent variable. The relevance of this to the
behavioral illusion is that Powers shows that, when the
systems under study are control systems (called N-systems
in the paper), any apparent effect of an
independent variable on a dependent variable is an
illusion.

          RM: The fundamental problem with significance testing,

from a PCT perspective, is that is is based on a causal
model of behavior; a model that says that the input to the
system (the effect of the independent variable) is the
ultimate cause of the output of that system (the dependent
variable). That is, significance testing in psychology is
based on the wrong model of the nature of living systems.

RM: Perhaps this thread will reveal what that unwarranted extreme is.

MT: The basic problem with that is one that has often been discussed on

CSG-L and CSG-net: the relevant data for a PCT analysis are
unavailable.

RM: I think it’s worth it to try again. Bill’s rather devastating analysis in the 1978 Psych Review paper (sorry I mistakenly said Science paper earlier) was rather abstract. I think it might help to get his message across (after over 30 years of it being ignored by scientific psychology) if we came up with a paper explaining the implications of Bill’s analysis in the context of real psychology experiments, particularly well-known classics.

          RM: So what I recommend for going forward on this

“behavioral illusion” thread is that those of us
interested in doing PCT research (which may be just you
and me Leeanne;-) should read (or re-read) the attached
1978 Science paper by Powers and then discuss the
implications of Powers’ analysis for research in
psychology. Let’s see what’s going on in some experiments
where the results were analyzed using significance tests
and see how we might interpret those results in the
context of the analysis in Powers 1978 Science paper.

MT: Your own demos are good examples, but I think you have to go both

simpler and more carefully than Powers did, if you want to make a
research program out of it.

RM: Actually, my goal is just to show in a more concrete way than Bill did in the 1978 paper why the conventional approach to research tells you very little about the mechanisms of behavior.

MT: To begin with, you have to understand

how to reconcile the Behavioural illusion with the apparent success
of modelling and parameter optimization. I go into that in what I
was drafting.

          RM: Then let's propose some experiments consistent with the

rather abstract proposals for research presented in that
paper (as Experiments 5 and 6).

RM: I agree and I look forward to seeing your paper.

[MartinT]

[Martin Taylor 2017.03.28.23.36]

This is not a response to anything said in this thread, so much as

an explanation of the attachment. When I started to write what I had
intended as the initiating contribution to this thread (before Rick
beat me to it), I found that it was turning into a rewrite of the
“Behavioural Illusion” section of the book I am writing on “the
barely travelled paths” of PCT ( It is Chapter 2 Section 5 at the
moment). Since even as a first draft I like it better than the
original, I have left the in-book material such as cross references
in what I am sending. I have, however, added at the top Figure 1.2,
which occurs some 20 pages earlier in the book, because it is
referenced in the text in a way that probably won’t make sense
without the actual figure.

Hi Rick and Martin.,

    I am still following with enthusiasm, although I

will probably be offline for a few weeks with school holidays,
my two children’s birthdays (and parties), parents visiting, uni
assignments, not to mention a week long Intensive School that I
need to travel interstate for. Â If you had asked me 3 years ago
would I be interested in this, I would have said an emphatic
“No�.  However, now the answer to that question is an emphatic
“YES!�.

That's nice to hear. I hope you continue to feel that way for a long

time.

            [From Rick Marken

(2017.03.28.1055)]

      LW: Â I think I understand this but could you perhaps explain

in a bit more detail…Â

If I understand correctly, Rick's point, with which I basically

agree, is that there’s no value in measuring the wrong things, no
matter how sophisticated the analysis. My “unwarranted extreme”
refers to the fact that even in PCT simple cause-effect
relationships do sometimes occur between observable variables. There
are a couple of examples in the attachment. It’s also true that even
when the basic behaviour depends on interacting control loops as
part of a pattern of social interactions, there are occasions when,
say, the strength of these interactions (effect) can be related to a
cause (introduction of the internet, for example). Even in these
cause-effect cases, significance testing can be quite misleading.

To be strict, Bill's analysis depends on control being perfect,

which it never is. I go into the ramifications of this and its
relation to the ability of PCT simulation models to say anything
about the internal processes.

      LW:  That would certainly catch people’s attention,

especially as it is something they are familiar with.
 Although, in this case I think it is the familiarity that
would be important….

So its not a research program you are contemplating, but another

demo or two?

      LW: Â Me too. Â Really helpful explanations for me in this

thread. Â Thanks!

Well, here it is, for better or for worse. I hope for some

thoughtful criticism of the attachment from any reader, not just
Rick and Leanne! I am quite willing to admit that there’s a lot that
could be improved.

Martin

BehaviouralIllusion.pdf (289 KB)

···

Martin Taylor (2017.03.27.15.18)–

                    MT: No. It's the mechanism of the effect

that is an issue, not the existence of the
effect under the conditions of the experiment.

…
Â

                    MT: Basically true, but often taken to an

unwarranted extreme.

                                RM: When you reject the

null hypothesis (with only a small
probability of being wrong) you are
deciding that the independent
variable actually does have an
effect on the dependent variable.
The relevance of this to the
behavioral illusion is that Powers
shows that, when the systems under
study are control systems (called
N-systems in the paper), any
apparent effect of
an independent variable on a
dependent variable is an illusion.

                                RM: The fundamental

problem with significance testing,
from a PCT perspective, is that is
is based on a causal model of
behavior; a model that says that the
input to the system (the effect of
the independent variable) is the
ultimate cause of the output of that
system (the dependent variable).
That is, significance testing in
psychology is based on the wrong
model of the nature of living
systems.

                  Â RM: Perhaps this thread will reveal

what that unwarranted extreme is.Â

                    MT: The basic problem with that is one

that has often been discussed on CSG-L and
CSG-net: the relevant data for a PCT analysis
are unavailable.

                  Â RM: I think it's worth it to try

again. Bill’s rather devastating analysis in the
1978 Psych Review paper (sorry I
mistakenly said Science paper
earlier) was rather abstract. I think it might
help to get his message across (after over 30
years of it being ignored by scientific
psychology) if we came up with a paper explaining
the implications of Bill’s analysis in the context
of real psychology experiments, particularly
well-known classics.

                                RM: So what I recommend

for going forward on this
“behavioral illusion” thread is that
those of us interested in doing PCT
research (which may be just you and
me Leeanne;-) should read (or
re-read) the attached 1978 Science paper by
Powers and then discuss the
implications of Powers’ analysis for
research in psychology. Let’s see
what’s going on in some experiments
where the results were analyzed
using significance tests and see how
we might interpret those results in
the context of the analysis in
Powers 1978 Science paper.

                    MT: Your own demos are good examples, but

I think you have to go both simpler and more
carefully than Powers did, if you want to make a
research program out of it.

                  RM: Actually, my goal is just to show

in a more concrete way than Bill did in the 1978
paper why the conventional approach to research
tells you very little about the mechanisms of
behavior.

                                RM: Then let's propose

some experiments consistent with the
rather abstract proposals for
research presented in that paper (as
Experiments 5 and 6).Â

Â

                    MT:

To begin with, you have to understand how to
reconcile the Behavioural illusion with the
apparent success of modelling and parameter
optimization. I go into that in what I was
drafting.

                  RM: I agree and I look forward to

seeing your paper.Â

[FredNickols]

[From Fred Nickols (2017.03.29.1155 ET)]

For what it’s worth, Richard, I don’t see the two statements as mutually exclusive. To me, to say that a stimulus (an input) causes a response (an output), ignores the issue of what is between the input and the output. So, to say that the response (the output) is determined by the processing (of the input) inside the organism does not conflict with nor negate the first statement. However, I am far from an authority on this stuff so I’ll wait to see what the experts have to say.

Fred NickolsÂ

···

From: richardpfau4153@aol.com [mailto:richardpfau4153@aol.com]
Sent: Wednesday, March 29, 2017 10:57 AM
To: csgnet@lists.illinois.edu
Subject: Re: Behavioral Illusion (was Re: p-hacking)

[From: Richard Pfau (2017.03.29 11:00 DST)]
Re: [Martin Taylor 2017.03.28.23.36]

I’m confused! I thought that the “Behavioral Illusion” was the illusion that “a stimulus causes a response” – not, as stated in your attachment, that “It is the illusion that the ‘response’ that follows a ‘stimulus’ is determined by the processing that occurs inside the organism.”

-----Original Message-----
From: Martin Taylor mmt-csg@mmtaylor.net
To: csgnet csgnet@lists.illinois.edu
Sent: Wed, Mar 29, 2017 12:08 am
Subject: Re: Behavioral Illusion (was Re: p-hacking)

[Martin Taylor 2017.03.28.23.36]

This is not a response to anything said in this thread, so much as an explanation of the attachment. When I started to write what I had intended as the initiating contribution to this thread (before Rick beat me to it), I found that it was turning into a rewrite of the “Behavioural Illusion” section of the book I am writing on “the barely travelled paths” of PCT ( It is Chapter 2 Section 5 at the moment). Since even as a first draft I like it better than the original, I have left the in-book material such as cross references in what I am sending. I have, however, added at the top Figure 1.2, which occurs some 20 pages earlier in the book, because it is referenced in the text in a way that probably won’t make sense without the actual figure.

Hi Rick and Martin.,

I am still following with enthusiasm, although I will probably be offline for a few weeks with school holidays, my two children’s birthdays (and parties), parents visiting, uni assignments, not to mention a week long Intensive School that I need to travel interstate for. If you had asked me 3 years ago would I be interested in this, I would have said an emphatic “No�. However, now the answer to that question is an emphatic “YES!�.

That’s nice to hear. I hope you continue to feel that way for a long time.

[From Rick Marken (2017.03.28.1055)]

Martin Taylor (2017.03.27.15.18)–

RM: When you reject the null hypothesis (with only a small probability of being wrong) you are deciding that the independent variable actually does have an effect on the dependent variable. The relevance of this to the behavioral illusion is that Powers shows that, when the systems under study are control systems (called N-systems in the paper), any apparent effect of an independent variable on a dependent variable is an illusion.

MT: No. It’s the mechanism of the effect that is an issue, not the existence of the effect under the conditions of the experiment.

…

RM: The fundamental problem with significance testing, from a PCT perspective, is that is is based on a causal model of behavior; a model that says that the input to the system (the effect of the independent variable) is the ultimate cause of the output of that system (the dependent variable). That is, significance testing in psychology is based on the wrong model of the nature of living systems.

MT: Basically true, but often taken to an unwarranted extreme.

LW: I think I understand this but could you perhaps explain in a bit more detail…

If I understand correctly, Rick’s point, with which I basically agree, is that there’s no value in measuring the wrong things, no matter how sophisticated the analysis. My “unwarranted extreme” refers to the fact that even in PCT simple cause-effect relationships do sometimes occur between observable variables. There are a couple of examples in the attachment. It’s also true that even when the basic behaviour depends on interacting control loops as part of a pattern of social interactions, there are occasions when, say, the strength of these interactions (effect) can be related to a cause (introduction of the internet, for example). Even in these cause-effect cases, significance testing can be quite misleading.

RM: Perhaps this thread will reveal what that unwarranted extreme is.

RM: So what I recommend for going forward on this “behavioral illusion” thread is that those of us interested in doing PCT research (which may be just you and me Leeanne;-) should read (or re-read) the attached 1978 Science paper by Powers and then discuss the implications of Powers’ analysis for research in psychology. Let’s see what’s going on in some experiments where the results were analyzed using significance tests and see how we might interpret those results in the context of the analysis in Powers 1978 Science paper.

MT: The basic problem with that is one that has often been discussed on CSG-L and CSG-net: the relevant data for a PCT analysis are unavailable.

RM: I think it’s worth it to try again. Bill’s rather devastating analysis in the 1978 Psych Review paper (sorry I mistakenly said Science paper earlier) was rather abstract. I think it might help to get his message across (after over 30 years of it being ignored by scientific psychology) if we came up with a paper explaining the implications of Bill’s analysis in the context of real psychology experiments, particularly well-known classics.

To be strict, Bill’s analysis depends on control being perfect, which it never is. I go into the ramifications of this and its relation to the ability of PCT simulation models to say anything about the internal processes.

LW: That would certainly catch people’s attention, especially as it is something they are familiar with. Although, in this case I think it is the familiarity that would be important….

RM: Then let’s propose some experiments consistent with the rather abstract proposals for research presented in that paper (as Experiments 5 and 6).

MT: Your own demos are good examples, but I think you have to go both simpler and more carefully than Powers did, if you want to make a research program out of it.

RM: Actually, my goal is just to show in a more concrete way than Bill did in the 1978 paper why the conventional approach to research tells you very little about the mechanisms of behavior.

So its not a research program you are contemplating, but another demo or two?

MT: To begin with, you have to understand how to reconcile the Behavioural illusion with the apparent success of modelling and parameter optimization. I go into that in what I was drafting.

RM: I agree and I look forward to seeing your paper.

LW: Me too. Really helpful explanations for me in this thread. Thanks!

Well, here it is, for better or for worse. I hope for some thoughtful criticism of the attachment from any reader, not just Rick and Leanne! I am quite willing to admit that there’s a lot that could be improved.

Martin

[MartinT]

[Martin Taylor 2017.03.29.11.11]

      [From: Richard Pfau (2017.03.29 11:00 DST)]

      Re: [Martin Taylor 2017.03.28.23.36]



      I'm confused!  I thought that the "Behavioral Illusion" was

the illusion that “a stimulus causes a response” – not, as
stated in your attachment, that “It is the illusion that the
‘response’ that follows a ‘stimulus’ is determined by the
processing that occurs inside the organism.”

No. There's no illusion about the "stimulus" causing the "response".

If the organism controls well and the reference value stays
constant, it does (insofar as we can isolate a cause for any
observation).

The illusion is that the form of the response is determined by what

happens inside the organism. To the casual observer it looks as
though this response is caused by that stimulus because that’s
the way the organism is made. But if that stimulus is actually a
disturbance to a controlled variable, it doesn’t matter how the
organism is made, the “response” must be whatever is required in
order to counter the disturbance by acting on the environment,
typically done by changing lots of muscular tensions.

All of that doesn't matter to the values in the simple control loop.

If the disturbance value is “d”, the countering effect must
approximate -d. The output is “o”, and the environmental feedback
path to the point where the disturbance has its effect can be
written as a function f( ), so the actual effect of the output is
f(o). So f(o) = -d for perfect control.

Apart from the possibility that the organism might choose different

means to achieve -d (walking, cycling, driving to change one’s
perception of one’s location) and these choices imply different
functions, the output “o” is f^-1 (-d) for perfect control.
That’s independent of how the organism processes the “stimulus” to
produce this precise value. All it has to do is control.

I should have pointed out in the text that this argument applies to

the quasi-static analysis of the control system of the kind used by
Powers in the Psych Review paper. It does not apply to the dynamics
of how the final response is reached after an abrupt change to the
disturbance. The entire loop has its own dynamic characteristics
that cannot be separated into environmental and internal components
by observing what happens between two distinct places in the loop.

Martin