Why PCT: The Rube Goldberg Justification

[rsmarken]

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/) and 2) discussing the process of mitosis with my a biologist sister-in-law.

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood. It was the part of the review that says “…PCT can predict some things but not much better
than competing theories”. This is true only when the behaviors being predicted are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).Â

RM: The discussion of mitosis – one type of cell division behavior– also shows the importance of understanding that PCT is an explanation of control. Mitosis involves a complex sequence of steps where the start of each step depends on the successful accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome. But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.Â

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in which a series of devices that perform simple tasks are linked together to produce a domino effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

 https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.Â

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.Â

BestÂ

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

[FredNickols]

[From Fred Nickols (2017.05.15.1625 ET)]

Hip, Hip, Hooray! Hip, Hip, Hooray! Hip, Hip, Hooray! Three cheers for Rick! Dude, I can use this! I can really use this! This is great! This is better than great! This is magnificent! It goes completely around the problem. Better yet, as some of you know, I write a monthly column for ISPI’s PerformanceXpress titled “Knowledge Worker� and I’ve been playing around with a column in which I aim to raise the notion that PCT or Perceptual Control Theory can also be viewed as “Performance Control Theory,� in other words, a theory of how the performer controls his or her performance (and not a theory of the performer’s behavior as commonly understood).

Many, many, many thanks for this, Rick. It has crystallized some of my own thinking. I owe you one, Dude!

Fred Nickols

···

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Monday, May 15, 2017 3:29 PM
To: csgnet@lists.illinois.edu
Cc: Richard Marken rsmarken@gmail.com
Subject: Why PCT: The Rube Goldberg Justification

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/) and 2) discussing the process of mitosis with my a biologist sister-in-law.

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood. It was the part of the review that says “…PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).

RM: The discussion of mitosis – one type of cell division behavior-- also shows the importance of understanding that PCT is an explanation of control. Mitosis involves a complex sequence of steps where the start of each step depends on the successful accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome. But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in which a series of devices that perform simple tasks are linked together to produce a domino effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.

Best

Rick

–

Richard S. Marken

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

[Kent]

This is a nice post, Rick, and it makes a really important point.

Kent

···

On May 15, 2017, at 2:28 PM, Richard Marken rsmarken@gmail.com wrote:

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear
why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was
not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The
Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/ )
and 2) discussing the process of mitosis with my a biologist sister-in-law.

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood.
It was the part of the review that says “… PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted
are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a
disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain
the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models
and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).

RM: The discussion of mitosis – one type of cell division behavior-- also shows
the importance of understanding that PCT is an explanation of control . Mitosis involves a complex sequence of steps where the start of each step depends on the successful
accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process
successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be
accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome.
But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this
is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence
of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in
which a series of devices that perform simple tasks are linked together to produce a domino
effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will
not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually
try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances
that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal
links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the
end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes
in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in
the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests
that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But
I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg
machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s
review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an
explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.

Best

Rick

–
Richard S. Marken

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

[rsmarken]

[From Rick Marken (2017.05.16.0955)]

···

On Mon, May 15, 2017 at 1:29 PM, McClelland, Kent MCCLEL@grinnell.edu wrote:

KM: This is a nice post, Rick, and it makes a really important point

Thanks Kent. If you (or anyone else listening) knows of any other critiques of PCT (besides the Alexander review of B:CP that I mentioned) I’d really appreciate getting the references. Again, my main goal is to show that critiques of PCT are based on ignorance of the phenomenon that PCT is designed to explain – the phenomenon of control!Â

BestÂ

Rick

Â

On May 15, 2017, at 2:28 PM, Richard Marken rsmarken@gmail.com wrote:

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear
why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was
not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The
Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/ )
and 2) discussing the process of mitosis with my a biologist sister-in-law.

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood.
It was the part of the review that says “… PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted
are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a
disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain
the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models
and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).Â

RM: The discussion of mitosis – one type of cell division behavior–Â also shows
the importance of understanding that PCT is an explanation of control . Mitosis involves a complex sequence of steps where the start of each step depends on the successful
accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process
successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be
accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome.Â
But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this
is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence
of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.Â

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in
which a series of devices that perform simple tasks are linked together to produce a domino
effect  in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will
not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually
try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances
that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

 https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal
links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the
end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.Â

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes
in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in
the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests
that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But
I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg
machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s
review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an
explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.Â

BestÂ

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

–

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

[FredNickols]

[From Fred Nickols (2017.05.16.1328 ET)]

Rick (and anyone else who cares to get involved):

Attached is a first draft of a Knowledge Worker column I intend publishing about Performance Control Theory.

Please look it over and inform me of any egregious errors.

Thanks.

Fred Nickols

Performance Control Theory.docx (24 KB)

···

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Tuesday, May 16, 2017 12:55 PM
To: csgnet@lists.illinois.edu
Subject: Re: Why PCT: The Rube Goldberg Justification

[From Rick Marken (2017.05.16.0955)]

On Mon, May 15, 2017 at 1:29 PM, McClelland, Kent MCCLEL@grinnell.edu wrote:

KM: This is a nice post, Rick, and it makes a really important point

Thanks Kent. If you (or anyone else listening) knows of any other critiques of PCT (besides the Alexander review of B:CP that I mentioned) I’d really appreciate getting the references. Again, my main goal is to show that critiques of PCT are based on ignorance of the phenomenon that PCT is designed to explain – the phenomenon of control!

Best

Rick

On May 15, 2017, at 2:28 PM, Richard Marken rsmarken@gmail.com wrote:

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/) and 2) discussing the process of mitosis with my a biologist sister-in-law.

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood. It was the part of the review that says “…PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).

RM: The discussion of mitosis – one type of cell division behavior-- also shows the importance of understanding that PCT is an explanation of control. Mitosis involves a complex sequence of steps where the start of each step depends on the successful accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome. But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in which a series of devices that perform simple tasks are linked together to produce a domino effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.

Best

Rick

–

Richard S. Marken

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

–

Richard S. Marken

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

[rsmarken]

[From Rick Marken (2017.05.16.1040)]

···

Fred Nickols (2017.05.15.1625 ET)

Â

FN: Hip, Hip, Hooray! Hip, Hip, Hooray! Hip, Hip, Hooray! Three cheers for Rick! Dude, I can use this! I can really use this! This is great! This is better than great! This is magnificent!

RM: Gee, I feel like Sally Field getting the academy award. You like me, right now, you like me;-)

Â

FN: It goes completely around the problem. Better yet, as some of you know, I write a monthly column for ISPI’s PerformanceXpress titled “Knowledge Worker� and I’ve been playing around with a column in which I aim to raise the notion that PCT or Perceptual Control Theory can also be viewed as “Performance Control Theory,� in other words, a theory of how the performer controls his or her performance (and not a theory of the performer’s behavior as commonly understood).

RM: I think what better captures my point would be to say that PCT is a theory of performance as a process of control. Performers don’t really control performance (according to PCT they are controlling perceptions); I think it’s better to say that performance IS control. When we see a performance of any kind – a performance by a living system – we are seeing control. When we see a performance by a non-living system, like a Rube Goldberg machine, we are not seeing control; we are seeing what psychologists have mistakenly thought they are seeing when they see the performance of a living system: caused output.Â

RM: Psychological theories have been theories of behavior as caused output; PCT is a theory of behavior as control.Â

FN: Many, many, many thanks for this, Rick. It has crystallized some of my own thinking. I owe you one, Dude!

RM: Again, thanks so much. I look forward to seeing what you come up with for your monthly column! I hope you’ll keep us posted. And, again, if you know of any criticisms of PCT by people in your area I would appreciate it if you could send me a pointer to them.

BestÂ

Rick

Â

Fred Nickols

Â

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Monday, May 15, 2017 3:29 PM
To: csgnet@lists.illinois.edu
Cc: Richard Marken rsmarken@gmail.com
Subject: Why PCT: The Rube Goldberg Justification

Â

[From Rick Marken (2017.05.15.1230)]

Â

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/) and 2) discussing the process of mitosis with my a biologist sister-in-law.

Â

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood. It was the part of the review that says “…PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).Â

Â

RM: The discussion of mitosis – one type of cell division behavior– also shows the importance of understanding that PCT is an explanation of control. Mitosis involves a complex sequence of steps where the start of each step depends on the successful accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome. But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.Â

Â

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in which a series of devices that perform simple tasks are linked together to produce a domino effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

 https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

Â

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.Â

Â

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

Â

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.Â

Â

BestÂ

Â

Rick

Â

–

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

–
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 2017.05.16.13.23]

I'd like to add my compliments on this post, and make another

complementary (also complimentary) suggestion.

One of the core objectives of the "Rube Goldberg" apparatus of

“Predictive Control” is prediction of what actions are required in
order to produce certain perceptual effects. Complex algorithms are
needed to predict how to compensate for disturbances. Bill’s
hierarchic PCT takes care of this problem so simply that it may seem
unreal, by reorganizing layers of successively more complex but
consistent means of producing successively more complex effects on
successively more complex perceptions. If the external world is
real, there are consistencies in the way it works at these different
levels, and no complicated algorithms or explicit internal models
are required to find them. Building on already developed means of
controlling lower levels, the problem of reorganizing higher levels
ceases to be a real issue.

The same applies to "Deep Learning" approaches to developing complex

perceptions. Instead of feedback from hundreds of thousands of
pictures of different things causing restructuring throughout a
multilevel network, HPCT feedback from the success of otherwise of
control reorganizes the perceptions so that complex levels build on
simple ones. The difficulty of knowing what is the significance of
points within the Deep Learning network is always an issue. That
problem goes away when control is happening at all the levels
simultaneously and individually.

Martin

···

[From Rick Marken (2017.05.15.1230)]

      RM: Two recent events have led me to realize how important

it is, when “promulgating” PCT, to make it clear why
PCT was developed: * PCT was developed to explain the
phenomenon of control as it is seen in the behavior of
living systems*. It was not developed to explain
the behavior of living systems, where “behavior” refers to
what a system is seen to be “doing” – its observable outputs.
The two events that led me to this realization were 1)
re-reading the review of “Behavior: The Control of Perception”
by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/ )
and 2) discussing the process of mitosis with my a biologist
sister-in-law.

      RM: Scott Alexander's review of B:CP was actually

reasonably positive. But one part of that review, in
particular, led me to realize the importance of making it
clear that PCT is an explanation of control and not “behavior”
as conventionally understood. It was the part of the review
that says “… PCT can predict some things but
not much better
than competing theories”. This is true only when the
behaviors being predicted are produced in a disturbance-free
environment. That is, PCT is “just another theory of
behavior” when one ignores the fact that the behavior of
living organisms is normally produced in a disturbance-prone
environment; behaviors that appear to be emitted output
are actually controlled consequences of output – they are
controlled inputs. When this fact is taken into account
then PCT turns out to be the only theory that can explain
the behavior (controlling) of living systems (as
demonstrated in Bourbon and Powers (1999) Models and their Worlds, * International
Journal of Human-Computer Studies*, 50, 445-461).Â

          RM: The discussion of mitosis --

one type of cell division behavior–Â also shows
the importance of understanding that PCT is an explanation
of control . Mitosis
involves a complex sequence of steps where the start of each
step depends on the successful accomplishment of the
previous step. My biologist sister-in-law was convinced
that the process of mitosis was completely explained by the
laws of physics and chemistry; no control involved at all.
But it seemed to me that consistently carrying out this
process successfully implied that some kind of control
involved since the process is being carried out in a
disturbance-prone environment. One step, in particular,
caught my attention; the process of “centering the
centrosome”. Apparently the centrosome has to be accurately
maintained in the center of the cell before the next step in
the process can proceed. The biological explanation of this
centering is a causal model that involves building filaments
that exert equal force at different points around the
centrosome. But I realized that this process was carried
out in a cell that was likely to be varying in shape
somewhat over time. So the centering of the centrosome had
to be a control process. This idea was firmly rejected by my
sister-in-law. And I realized that this is because she (and,
apparently, most biologists) saw the process of producing
cell division behavior (mitosis) in the same way
psychologists see the process of producing overt behaviors
(like walking) – as a cause-effect process where a complex
sequence of effects (results) are produced with each result,
once produced, becoming the cause of the process that
produces the next result. I will call this the Rube Goldberg
view of behavior.Â

        RM: The Rube

Goldberg view of behavior sees a complex behavior like cell
division as the activity of a Rube Goldberg machine. A Rube
Goldberg machine is a “deliberately complex contraption in
which a series of devices that perform simple tasks are
linked together to produce a domino
effect  in which activating one device triggers the
next device in the sequence”. The end result of this complex
process is some ridiculously simple achievement. But for
present purposes what’s interesting about these devices is
that the end result will not be achieved unless each of the
intermediate steps in the process is carried out
successfully. Rube Goldberg was a cartoonist and his amusing
devices always worked on paper (just like the causal models
of mitosis always work on paper). But when you actually try
to build a Rube Goldberg machine that works in the real
work, you discover that getting it to produce the end result
reliably is a virtual impossibility. This is because, in the
real world, there is many a slip 'twixt cup and lip because
there are disturbances that can disrupt the process at each
step. Here is a video of a real-world Rube Goldberg machine
in action:

      Â [https://www.youtube.com/watch?v=YWk9N92-wvg#t=27](https://www.youtube.com/watch?v=YWk9N92-wvg#t=27)

      RM: This video shows the Rube Goldberg machine achieving

its end result: rolling out the new Honda. It succeeds
because, as you can see, each of the complex steps in the
process is carried out successfully. So it looks like a
complex set of causal links can produce a particular end
result. But it turns out that this machine made it through all
steps successfully only after 606 failures! On 606 prior
“takes” something went wrong at some point in the process, so
the machine didn’t manage to produce the end result until take
number 607, when, apparently, by luck disturbances failed to
interfere with the production of the correct results at each
step in the process.Â

      RM: Clearly, this Rube Goldberg process -- a series of

complex causal steps – cannot be what is going on in mitosis
(or other biological of behavioral processes). The end result
of mitosis – cell division with the correct compliment of
chromosomes in both daughter cells – is achieved nearly 100%
of the time; the end result of the Rube Goldberg machine in
the video is produced, at best, .16% of the time. This seems
like pretty strong evidence that mitosis is not a Rube
Goldberg process; the steps in the process of cell division
may look like the steps in a Rube Goldberg (causal) process
but the fact that the end result (two perfect daughter cells)
is produced with nearly perfect reliability in what is known
to be a disturbance-prone environment suggests that what we
are looking at when we look at mitosis is control. A model
that correctly accounts for the behavior of cell division
(like a model that correctly accounts for any consistently
produced behavior of organisms) will be a model of control –
PCT.

        RM:

The idea that PCT is a model of control and not of behavior
as conventionally understood – the behavior of a Rube
Goldberg machine – is one that is very hard to get across.
But I think it is at the heart of the problems PCT has had
with becoming accepted in the behavioral and other life
sciences. But I want to try to get this idea across to
behavioral and life scientists and I think maybe I can do it
using the idea of a Rube Goldberg machine as an example of a
device that produces the kind of behavior that does not
involve control – and, therefore, can’t work reliably in
the disturbance prone real world. I’d like to do this in the
context of criticisms of PCT, like those in Scott Anderson’s
review of B:CP. So I would appreciate it if anyone could
point me to other articles where PCT has been criticized.
Again, I think all the criticisms of PCT that I have seen
(and I’m doing this from memory) are based on the mistaken
assumption that PCT is an explanation of behavior as caused
output (as is that of the Rube Goldberg machine) rather than
what it actually is: an explanation of behavior as control.Â

BestÂ

Rick

–
Richard S. MarkenÂ

                                "Perfection

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

[DagArchive]

[From Dag Forssell (2017.05.16.11:55 PST)]

Fred, Nice, and I think Rick's comment is significant.

Why two different addresses for Benchmark? Alice moved from Connecticut to New Jersey some years ago.

Best, Dag

···

At 10:30 AM 5/16/2017, you wrote:

[From Fred Nickols (2017.05.16.1328 ET)]

[FredNickols]

I think I used the address in the books. I will check.

Fred Nickols, CPT

Writer & Consultant

DISTANCE CONSULTING LLC

“Assistance at a Distance”

View My Books on Amazon

···

Sent from my iPad

On May 16, 2017, at 2:54 PM, Dag Forssell csgarchive@pctresources.com wrote:

[From Dag Forssell (2017.05.16.11:55 PST)]

Fred, Nice, and I th
ink Rick’s comment is significant.

Why two different addresses for Benchmark? Alice moved from Connecticut to New Jersey some years ago.

Best, Dag

At 10:30 AM 5/16/2017, you wrote:

[From Fred Nickols (2017.05.16.1328 ET)]

[DagArchive]

I think you are right about that. But why give the impression that there
are two different publishers? The current address is NJ.

Dag

···

At 12:06 PM 5/16/2017, you wrote:

I think I used the address in
the books. I will check.
Fred Nickols, CPT
Writer & Consultant
DISTANCE CONSULTING LLC"Assistance at a Distance"View My Books on
Amazon

Sent from my iPad

On May 16, 2017, at 2:54 PM, Dag Forssell > < > csgarchive@pctresources.com> wrote:

[From Dag Forssell
(2017.05.16.11:55 PST)]

Fred, Nice, and I th ink Rick’s comment is significant.

Why two different addresses for Benchmark? Alice moved from Connecticut
to New Jersey some years ago.

Best, Dag

At 10:30 AM 5/16/2017, you wrote:

[From Fred Nickols
(2017.05.16.1328 ET)]

[FredNickols]

I think it’s an academic requirement pertaining to the accuracy of citations.

Fred

···

From: Dag Forssell [mailto:csgarchive@pctresources.com]
Sent: Tuesday, May 16, 2017 3:39 PM
To: csgnet@lists.illinois.edu
Subject: Re: Why PCT: The Rube Goldberg Justification

I think you are right about that. But why give the impression that there are two different publishers? The current address is NJ.

Dag

At 12:06 PM 5/16/2017, you wrote:

I think I used the address in the books. I will check.
Fred Nickols, CPT
Writer & Consultant
DISTANCE CONSULTING LLC"Assistance at a Distance"View My Books on Amazon

Sent from my iPad

On May 16, 2017, at 2:54 PM, Dag Forssell < csgarchive@pctresources.com> wrote:

[From Dag Forssell (2017.05.16.11:55 PST)]

Fred, Nice, and I th ink Rick’s comment is significant.

Why two different addresses for Benchmark? Alice moved from Connecticut to New Jersey some years ago.

Best, Dag

At 10:30 AM 5/16/2017, you wrote:

[From Fred Nickols (2017.05.16.1328 ET)]

[_Boris_Hartman]

Down…

···

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Monday, May 15, 2017 9:29 PM
To: csgnet@lists.illinois.edu
Cc: Richard Marken
Subject: Why PCT: The Rube Goldberg Justification

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems.

Bill P. at all (50th Anniversary, 2011) :

Perceptual Control Theory (PCT) provides a general theory of functioning for organisms. At the core concept of the theory is the observation that living things control perceived environment by means of their behavior. Consequently, the phenomen of control takes center stage in PCT, with observable behavior playing an important but supporting role.

HB : To be excact it is not only the phenomenon of control that is seen in »Behavior of living systems« but how living things control perceived environment and of course how they function. Control theory is trying to explain why and how orgsnisms behave but it’s priority is to explain how they function. But as I wrote many times before PCT is not finnished yet. It’s a lot of work to be done, so that it will become recognized as leading theory on the field of psychology and medicine.

RM : It was not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs.

HB : You are all the time promoting what a system is seen to be “doing” – its observable outputs, which by your oppinion.are control what is missing the point as the behavior on stimulu is missing the point.Although you are not pointed out this here.Your writings are all the time circling arround behavior, obervation of behavior and so on…¦.

PCT is trying to understand how organisms function as control systems and how they produce behavior. It’s not that »obervable outpus are controlled«or produced by stimuli as the central problem. Important are internal processes of living systems which are control.

The behavior is not the main point in PCT. It’s important. It’s the part that is opposing classical psychology and this is one of the places where Scott Alexander was wrong. PCT is much stronger theory if you try to understand how living beings function than just to understand how you drive a car r how you observe behavior as control.  And as I could see Scott is not some expert in physiology like you are not Rick. »Blind« will criticize »blind«. What a combination ?

RM : The two events that led me to this realization were 1) re-reading the review of “Behavior: The Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/) and 2) discussing the process of mitosis with my a biologist sister-in-law.

HB : Well I answered you on that too. See bellow. You will see that you are inventing »hot-water«.Bill from the beggining realized that »control« is in essence of organisms functioning. And the organism consists of 100 bilion cells. So the essense of cell fucntioning is the essence of organisms functioning

Bill P: (B:CP)Â Â

PCT…«can exxplain a fundamental aspect of how every living thing works, form the tiniest amoeba to the being who is reading these words.«Â

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood.

BH : Behavior has something to do with control but not as much as you think. See above the Anniversary.

RM : It was the part of the review that says “…PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs.

HB : Behavior is consequence of organisms control. Behavior do not »carry« any control. You must be sure what you are talking about. There isn’t any controlled consequences of output. Behavior is consequence of control. See Bills’ physiological evidences in B:CP and diagram in LCS III…

RM : When this fact is taken into account then PCT turns out to be the only theory that can explain the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).

HB : As you pointed out Tom Bourbon you should listen to him and you wouldn’t invent »hot water«. Â

Tom B.

We think that the self-replicating molecules, like those in DNA, are control systems, complete with their equivalents of reference signals that specify which ‘perceptions’ of molecular shape, or of chemical states, they will sense. On this construal, gense are not “commands” for what we will become, but they are control systems that control for certain molecular states; of the rest that happens is in a way one big side-effect of control at the biochemical level. If that is so, then it must be the case that, more often than not, creatures like us, with perceptual hierarchies like ours, end up being good environments for DNA to achieve its own control.

HB : Listen to him Rick, he seems to be a wise man. PCT behavior (controlling) of living systems is :

Bill P :

It means that we produce actions that alter the world of perception, and that we do that specifically to make the state of that world conform to the reference conditions we ourselves have choosen (to the extent we can change the perceptions by our actions).

Bill P : There is one explanation for the existence of reference states that has been proposed over and over the centuries : they are determined by the intentions of the behaving organism. The driver has, inside him, the intention that the door be open. He acts to achieve this purpose, doing whatever is required (if possible) to achieve it…

HB : Control is happening in organism and the behaviors are jsut means to support control in organism.

RM: The discussion of mitosis – one type of cell division behavior-- also shows the importance of understanding that PCT is an explanation of control.

HB : Vauuu. What a discovery. In his diagram p. 191 (B:CP) Bill is clearly determining the control of Living beings with »genetic control system« that is the driving power of organisms control. Although he didn’t exactly talk about »genetic control system (Tom Bourbon did), it is obvious that he meant just that.

I’m sure that Bill was acquainted with some medicine knowledge (he seems to be employed in hospital and he has a friend physiologist). So I don’t doubt that he knew that in physiological books you could find »genetic control« in cells at least from arround 1956.

So you are out of time Rick. Bill already defined that organisms function on the basis of »genetic control«. We can conclude that from diagram on p. 191. It’s clear that references come from genetic source. And as PCT is about control units in organism it’s not difficult to conclude that he is talking about »genetic control unit«.

Genes (DNA) controls the forming of organism including cell divission »mitosis« (Eucarionts) … Genes control also other type of cell divission (for example bacteria), but it’s different from »mitosis«. »Binary fission« can be explained only with »genetic control«.

So the point is that we have to finnish diagram on p. 191 (B:CP) what I’m proposing for a long time, so that you Rick will not invent »hot-water« all the time with your theory RCT.

RM : Mitosis involves a complex sequence of steps where the start of each step depends on the successful accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment.

HB : Well unwillingly I have to agree with you. But if you would be acquanted with physiology you wouldn’t blindly guess, because you would know that in the basiscs of cell regulation is for along time know »genetic control«. O.K. I’ll repeat again that Bill showed this »fact« on p. 191 (B:CP) and probably in his other literature.

RM : One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome. But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.

HB : You could conclude that from diagram on p. 191 (B:CP) if you would understand it of course. It’s true that diagram needs upgrade, but the main point of »genetic control« in organisms is done. I even think that with improved version of diagram on p. 191 (B:CP) we could make simulation of cell division from the view of »genetic control«. I think that’s what PCT is about. To show how cotnrol works in organism from genes to 11.level of hierarchy.

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in which a series of devices that perform simple tasks are linked together to produce a domino effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

HB : Well at least I know this from year 2007 when I think I read »Behavior : Control of Perception«. But it’s good to discover it once in time than never.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across.

HB : So far so good…J. But it’s not that hard to get across of control in organism if you understand what you are reading in B:CP. Then it’s very easy.

RM : But I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.

HB : Well you were doing so good and on the end you »bullshitted« everything. It’s not behavior that is controlled, it’s perception. So if you’ll continue criticizing non-PCT articles, don’t do it with non-PCT theory like RCT is. If you’ll do it, I’ll have to cirticize you.

Bill P :

The flaw of this reasoning is hard to understand if one does not know (as the founders of scientific psychology did not know) of organizations capable of complex internal activities that are essentially independent of current external events. By ruling out the possibility of significant causes of behavior inside the organism, where they could not be observed, early behavioral scientist in effect commited themselves to a whole chain of deductions following from the assumption that everything of significance with regard to behavior could be observed from outside of organism. They were betting everything on the assumption that such internal causes would never be found to exist.

HB : I’ll just repeat again that PCT organism on p. 191 is showing quite well how internal functioning of organisms start in »genetic control« which is known in physiology for a long time. I’m not acquainted with the state in biology, but it seems to be behind PCT and physiology for at least 60 years.

So I’ll again offer the possibility that Powers ladies could organize proffessional project on improving diagram on p. 191 (B:CP) so that members will not have problems understading that PCT is how control function in organisms from the core of its existance. .

                                    Â

HB : So we see that Bill beleived that PCT can explain also how cells function as control system. And the diagram on p. 191 is showing also that.

Best,

Boris

Best

Rick

–

Richard S. Marken

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

[_Boris_Hartman]

Down…

···

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Tuesday, May 16, 2017 7:37 PM
To: csgnet@lists.illinois.edu
Subject: Re: Why PCT: The Rube Goldberg Justification

[From Rick Marken (2017.05.16.1040)]

Fred Nickols (2017.05.15.1625 ET)

FN: Hip, Hip, Hooray! Hip, Hip, Hooray! Hip, Hip, Hooray! Three cheers for Rick! Dude, I can use this! I can really use this! This is great! This is better than great! This is magnificent!

RM: Gee, I feel like Sally Field getting the academy award. You like me, right now, you like me;-)

FN: It goes completely around the problem. Better yet, as some of you know, I write a monthly column for ISPI’s PerformanceXpress titled “Knowledge Workerâ€? and I’ve been playing around with a column in which I aim to raise the notion that PCT or Perceptual Control Theory can also be viewed as “Performance Control Theory,â€? in other words, a theory of how the performer controls his or her performance (and not a theory of the performer’s behavior as commonly understood).

RM: I think what better captures my point would be to say that PCT is a theory of performance as a process of control. Performers don’t really control performance (according to PCT they are controlling perceptions); I think it’s better to say that performance IS control.

HB : It’s not we. You see »behavior is control« and some other members.Cahnging the term »behavior« into »performance« does not change the essence of your RCT.

You see behavior as control. That’s your mistake in thinking about PCT. You are thinking as ordinary psychologist. Like for example, Carver. Behavior (performance) is control. You control »cup of coffee« bringing it with your hands from table into mouth. It is not PCT. In PCT you act on »cup of coffee« to change perception from on table to the mouth. Do you understand the difference between RCT and PCT ???

Here is real PCT :

Bill P : Our only view of the real world is our view of the neural signals that represent it inside our own brains. When we act to make a perception change to our more desireble state – when we make the perception off the glass change from »on the table« to »near the mouth« - we have no direct knowledge of what we are doing to the reality that is the origin of our neural signal; we know only the final result, how the result looks, feels, smells, sounds, tastes, and so forth…It means that we produce actions that alter the world of perception

RM : When we see a performance of any kind – a performance by a living system – we are seeing control.

HB : You see control as »performance» of LCS and your RCT. You see »behavior as control« what is not. And the essence of PCT is not only control. It’s also how living beings function. Organism is not functioning entirely on the basis of Control mechanisms. You aagreed in 2011 :

Bill P. at all (50th Anniversary, 2011) :

Perceptual Control Theory (PCT) provides a general theory of functioning for organisms. At the core concept of the theory is the observation that living things control perceived environment by means of their behavior. Consequently, the phenomen of control takes center stage in PCT, with observable behavior playing an important but supporting role.

RM : When we see a performance by a non-living system, like a Rube Goldberg machine, we are not seeing control; we are seeing what psychologists have mistakenly thought they are seeing when they see the performance of a living system: caused output.

RM: Psychological theories have been theories of behavior as caused output; PCT is a theory of behavior as control.

HB : How many times do I have to repeat that PCT is not theory of »Behavior as control«. It’s theory about »Control of Perception«. If behavior in Psychological theories are »controlled« by environmental »stimuli«, your »Control of Behavior« is based on »internal processes« of organism. But organisms can’t control their Behavior. But they can »Control perception«.Â

FN: Many, many, many thanks for this, Rick. It has crystallized some of my own thinking. I owe you one, Dude!

RM: Again, thanks so much. I look forward to seeing what you come up with for your monthly column! I hope you’ll keep us posted. And, again, if you know of any criticisms of PCT by people in your area I would appreciate it if you could send me a pointer to them.

HB : If you’ll be criticizing other non-PCT articles with your RCT, it will be the same as »non-PCT« theories are cirticized by »non-PCT« theory.

All in all it’s not bad idea to criticize Non-PCT theories. But you have to do it with PCT not with »Behavior (performance) is Control« means. I’ll have to criticise you.

Best,

Boris

Best

Rick

Fred Nickols

From: Richard Marken [mailto:rsmarken@gmail.com]
Sent: Monday, May 15, 2017 3:29 PM
To: csgnet@lists.illinois.edu
Cc: Richard Marken rsmarken@gmail.com
Subject: Why PCT: The Rube Goldberg Justification

[From Rick Marken (2017.05.15.1230)]

RM: Two recent events have led me to realize how important it is, when “promulgating” PCT, to make it clear why PCT was developed: PCT was developed to explain the phenomenon of control as it is seen in the behavior of living systems. It was not developed to explain the behavior of living systems, where “behavior” refers to what a system is seen to be “doing” – its observable outputs. The two events that led me to this realization were 1) re-reading the review of “Behavior: The Control of Perception” by Scott Alexander at his blog (https://slatestarcodex.com/2017/03/06/book-review-behavior-the-control-of-perception/) and 2) discussing the process of mitosis with my a biologist sister-in-law.

RM: Scott Alexander’s review of B:CP was actually reasonably positive. But one part of that review, in particular, led me to realize the importance of making it clear that PCT is an explanation of control and not “behavior” as conventionally understood. It was the part of the review that says “…PCT can predict some things but not much better than competing theories”. This is true only when the behaviors being predicted are produced in a disturbance-free environment. That is, PCT is “just another theory of behavior” when one ignores the fact that the behavior of living organisms is normally produced in a disturbance-prone environment; behaviors that appear to be emitted output are actually controlled consequences of output – they are controlled inputs. When this fact is taken into account then PCT turns out to be the only theory that can explain the behavior (controlling) of living systems (as demonstrated in Bourbon and Powers (1999) Models and their Worlds, International Journal of Human-Computer Studies, 50, 445-461).

RM: The discussion of mitosis – one type of cell division behavior-- also shows the importance of understanding that PCT is an explanation of control. Mitosis involves a complex sequence of steps where the start of each step depends on the successful accomplishment of the previous step. My biologist sister-in-law was convinced that the process of mitosis was completely explained by the laws of physics and chemistry; no control involved at all. But it seemed to me that consistently carrying out this process successfully implied that some kind of control involved since the process is being carried out in a disturbance-prone environment. One step, in particular, caught my attention; the process of “centering the centrosome”. Apparently the centrosome has to be accurately maintained in the center of the cell before the next step in the process can proceed. The biological explanation of this centering is a causal model that involves building filaments that exert equal force at different points around the centrosome. But I realized that this process was carried out in a cell that was likely to be varying in shape somewhat over time. So the centering of the centrosome had to be a control process. This idea was firmly rejected by my sister-in-law. And I realized that this is because she (and, apparently, most biologists) saw the process of producing cell division behavior (mitosis) in the same way psychologists see the process of producing overt behaviors (like walking) – as a cause-effect process where a complex sequence of effects (results) are produced with each result, once produced, becoming the cause of the process that produces the next result. I will call this the Rube Goldberg view of behavior.

RM: The Rube Goldberg view of behavior sees a complex behavior like cell division as the activity of a Rube Goldberg machine. A Rube Goldberg machine is a “deliberately complex contraption in which a series of devices that perform simple tasks are linked together to produce a domino effect in which activating one device triggers the next device in the sequence”. The end result of this complex process is some ridiculously simple achievement. But for present purposes what’s interesting about these devices is that the end result will not be achieved unless each of the intermediate steps in the process is carried out successfully. Rube Goldberg was a cartoonist and his amusing devices always worked on paper (just like the causal models of mitosis always work on paper). But when you actually try to build a Rube Goldberg machine that works in the real work, you discover that getting it to produce the end result reliably is a virtual impossibility. This is because, in the real world, there is many a slip 'twixt cup and lip because there are disturbances that can disrupt the process at each step. Here is a video of a real-world Rube Goldberg machine in action:

https://www.youtube.com/watch?v=YWk9N92-wvg#t=27

RM: This video shows the Rube Goldberg machine achieving its end result: rolling out the new Honda. It succeeds because, as you can see, each of the complex steps in the process is carried out successfully. So it looks like a complex set of causal links can produce a particular end result. But it turns out that this machine made it through all steps successfully only after 606 failures! On 606 prior “takes” something went wrong at some point in the process, so the machine didn’t manage to produce the end result until take number 607, when, apparently, by luck disturbances failed to interfere with the production of the correct results at each step in the process.

RM: Clearly, this Rube Goldberg process – a series of complex causal steps – cannot be what is going on in mitosis (or other biological of behavioral processes). The end result of mitosis – cell division with the correct compliment of chromosomes in both daughter cells – is achieved nearly 100% of the time; the end result of the Rube Goldberg machine in the video is produced, at best, .16% of the time. This seems like pretty strong evidence that mitosis is not a Rube Goldberg process; the steps in the process of cell division may look like the steps in a Rube Goldberg (causal) process but the fact that the end result (two perfect daughter cells) is produced with nearly perfect reliability in what is known to be a disturbance-prone environment suggests that what we are looking at when we look at mitosis is control. A model that correctly accounts for the behavior of cell division (like a model that correctly accounts for any consistently produced behavior of organisms) will be a model of control – PCT.

RM: The idea that PCT is a model of control and not of behavior as conventionally understood – the behavior of a Rube Goldberg machine – is one that is very hard to get across. But I think it is at the heart of the problems PCT has had with becoming accepted in the behavioral and other life sciences. But I want to try to get this idea across to behavioral and life scientists and I think maybe I can do it using the idea of a Rube Goldberg machine as an example of a device that produces the kind of behavior that does not involve control – and, therefore, can’t work reliably in the disturbance prone real world. I’d like to do this in the context of criticisms of PCT, like those in Scott Anderson’s review of B:CP. So I would appreciate it if anyone could point me to other articles where PCT has been criticized. Again, I think all the criticisms of PCT that I have seen (and I’m doing this from memory) are based on the mistaken assumption that PCT is an explanation of behavior as caused output (as is that of the Rube Goldberg machine) rather than what it actually is: an explanation of behavior as control.

Best

Rick

–

Richard S. Marken

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

–

Richard S. Marken

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