Short version of this post: social engineering is impossible. We cannot draw those graphs with lines and numbered axises because we do not know how to reliably measure either factor or have equations to draw the lines.
The concepts should be useful antidotes to social engineers’ hubris. But they are not.
I would say difficult rather than impossible; you might not be able to exactly quantify the curves, but you can usually guess that making the prisoners less angry and improving communication won't cause a riot
Back in the before times when I dabbled in dynamic simulations that were linear transformations I came across non linear transformations. One thing stuck in my peabrain; in a path dependent situation you jump from one state to the other but you don't get the jump when you reverse the path. The possibilty of this behavior in our warming planet indicates we can't just stop adding carbon and return to a cooler world.
As they say, "It's hard to stop a train." Can you undrink the beer you just scarfed? No one knows how to stop an earthquake, at least once it gets started. Maybe it could be prevented by relieving stresses?
Can chain reactions get stopped as easily as they get started? Anybody here do nuclear reactors? They have rods they push in.
When they send the rocket to the moon it must exceed the "escape velocity," then gravity can no longer pull it back.
Not sure how to apply this to prison populations. "Warden, all you have to do is..."
"Oh...but, I cannot do that."
"I cannot release the prisoners."
We may be mixing our meta-phors...hope they don't blow up.
I'm intrigued by this definition of riot as derivative of disorder. What's going on in the maximum tension, minimum alienation scenario, that's nice and smooth to reach but just as disorderly as a full-blown riot?
I would guess that's a situation you would never actually allow the prison to reach unless you jumped there from one that was basically tolerable to the authorities
I'm not a huge fan of game theory, but I think it provides a better approach to understanding multiple equilibria. That's in part because we can describe them in a static setting. Dynamics would be more appealing if we had any idea how to model them in relation to social phenomena but we mostly don't and probably never will have.
>I made a joke about Stafford Beer’s diagrams in the book, but I actually love hand-drawn graphs in articles and hope we never lose the art.
Maths departments in older universities will often have a collection of exotic surfaces with various curves traced on them rendered in plaster. It's particularly common in Germany, which was the real centre of geometry until WWII but I've seen them all over the place. In Strasbourg (Thom's university) they had a collection of model catastrophes. They're just curios now you can rotate surfaces to your heart's content in mathematica, but I find them charming too.
Ivar Ekeland has a nice little book called Mathematics and the Unexpected where he does a good job of connecting the maths and the metaphor without losing his head.
Oh man, that brings back some memories. I dabbled in catastrophe theory a long time ago, we had a Big Idea that the Thom classification theorem could be adapted to phase transitions in condensed-matter physics, but I was not able to get my head around the theorem.
My recollection of the vocabulary is different -- there's a "control space" (the 2D space in the picture), and a "response space" (the vertical axis), and the big deal is, as you say, that small changes in the control space can have discontinuous and large changes in the response space. I don't remember "control surface" being a thing. I have Arnol'd's book to hand, but it doesn't have an index.
Anyways, thanks for the nostalgia trip, is what I'm really saying.
I'll confess that I'd never heard of catastrophe theory until today but the Wikipedia article seems pretty good and suggests that a "control surface" is a set in the control parameter space, which was my intuition.
yep, that's the one. I am playing around with whether "control surface" is a suffiicently intuitive idea that I can use it in an expanded and metaphorical sense to describe the way that managers and policymakers interact with the world.
People probably try to use direct control methods first, then later come to understand that only indirect methods fit the situation.
Sometimes you need a hammer, sometimes a kiss.
We also tend to prefer hard and simple and understandable methods to fuzzy or soft or non-graspable ones.
The first stealth aircraft like the F-117 had very angular shapes, because the computers they designed them with could only do so many calculations. With newer computers, they could calculate more angles, and were able to design shapes that were more fluid, and stealthy, like the B-2 bomber. Speaking of the B-2, it is so inherently unstable that a human can't control it, only a computer can respond fast enough. The pilot just tells it which way to go. It is like working with AI to solve a problem that you can't on your own.
So maybe having the right amount of data and the right amount of power to use it may be necessary conditions to solve certain problems. This may exceed what is available. Then, the problem doesn't get solved. Plus, even if you figure out a "solution," it may be too late (or too soon!) to use it. The world keeps changing around you. You can't step in the same river twice.
But you can step in the same s--t twice!
Helps to be a hedgehog AND a fox. Didn't a wise person say: get a practical person to solve a difficult problem, and an impractical person to solve an impossible problem?
Maybe kayaking is a good metaphor for running the rapids we are in right now? Just don't tell McKinsey about this, or they'll start peddling rafting tours:
Kayaking yadda yadda yadda Kayaking yadda yadda yadda Kayaking yadda yadda yadda
Thom's "Structural Stability and Morphogenesis" is a bit like Topology written by Derrida, but it was unfairly criticized when it came out and many of its insights are yet to be incorporated into biology and social science. Unfortunately, I think the differential topology that Thom was pioneering wasn't the ideal mathematical formalism for the conceptual ideas he was trying to share.
Interesting. Does the diagram mean that when, after the riot, the administrators get tension back down to a certain level, the situation suddenly resolves and disorder suddenly drops just as it suddenly rose in the riot? (And is that accurate? I know nothing about the practicalities of prison administration but would have guessed that reducing tension post-riot would be difficult, maybe reflected by a different slope / curve, but not necessarily discontinuous in the way that the riot itself was.)
The jump downward is marked on a different version of the same diagram that Zeeman used in a talk as "truce", which seems fair enough. (The full model has a homeostatic parameter which basically says that people get tired of high disorder and bored with low disorder)
In the pendulum wave, there is order in disorder and vice versa, like yin containing yang. One property seems to predominate, then the other. It is also quite beautiful and moving:
After an avalanche or rockslide you cannot put the hill back together the same way.
All it takes is one skier to trigger an avalanche. The ski areas with big, steep slopes try to preemptively trigger avalanches with explosives, before they can kill anybody. Out in the back country, you are on your own.
There are certain things, that once seen, you cannot unsee. The investigators who were sent to document the massacre at Sandy Hook Elementary, including taking pictures of "everything," were severely traumatized by the experience. The town had to tear down the school and build a new one. To continue, they had to forget, as well as remember.
You cannot wind back time and make things un-happen, at least in this universe, at human scale.
I remember "catastrophe theory" when it was "hot" and was tried to fit a number of discontinuous phenomena. Even biologists were trying to use it. But, like other theories, it proved very difficult to apply, and more importantly, other non-linear models were just as applicable. AFAIK, it fell out of favor as a model. Fractals and Chaos Theory became the next "hot idea".
When I saw the words "control surface" I immediately thought of the control surfaces on aircraft, such as the small "wing" called the aileron that makes the big wing rise or fall, thereby causing the plane to turn by directing lift to one side, or the little "trim tabs" that make the control surfaces such as the elevator or rudder move, which then make the whole plane move on its lateral or vertical axes.
Buckminster Fuller famously used the trim tab as a metaphor for how little things have the power to move big things. He regarded himself as a "trim tab" in human form.
The other aviation metaphor that immediately comes to mind is how you make a plane "stall," or suddenly lose lift, by trying to create too much lift, usually by pulling the nose too high, thereby exceeding a critical "angle of attack," the angle at which the wing moves through the air. The plane starts going down, and may even go into a spin if it is not flying straight into the airflow at the time. The has real consequences. I know three people who were killed this way.
Things happen gradually...
...until they happen SUDDENLY!
You go were going up. Then you are going DOWN. Even though the nose may be up! It is not intuitive.
The interesting thing is that nature tends to repeat, and patterns found in one part of it also show up in another part.
Then there's fractals...
Our culture has a bias then leans us towards mechanistic thinking, where we think about doing simple things, like pushing a button, that have predicable consequences. Sometimes this is true.
But the things that vex us the most usually can't be "solved" by pushing a button. You can't "solve" weather. By its nature, it must constantly be changing.
What engineers sometimes forget is that not all things can be "solved." We must learn nature's ways, and live in harmony with them. Feynman spelled it out: "Nature cannot be fooled."
Sometimes a 3-map can show more than a 2-D one can. What if it were animated? Or colors or sounds were added?
Was that a model of a prison riot, or an earthquake?
Short version of this post: social engineering is impossible. We cannot draw those graphs with lines and numbered axises because we do not know how to reliably measure either factor or have equations to draw the lines.
The concepts should be useful antidotes to social engineers’ hubris. But they are not.
I would say difficult rather than impossible; you might not be able to exactly quantify the curves, but you can usually guess that making the prisoners less angry and improving communication won't cause a riot
1/ Guessing about directionality of the curves is not remotely like drawing the curves.
2/ Try telling that “insight” to a prison warden. My guess is that he or she will not be impressed. It’s too vague to be useful.
Back in the before times when I dabbled in dynamic simulations that were linear transformations I came across non linear transformations. One thing stuck in my peabrain; in a path dependent situation you jump from one state to the other but you don't get the jump when you reverse the path. The possibilty of this behavior in our warming planet indicates we can't just stop adding carbon and return to a cooler world.
Hmmm.
As they say, "It's hard to stop a train." Can you undrink the beer you just scarfed? No one knows how to stop an earthquake, at least once it gets started. Maybe it could be prevented by relieving stresses?
Can chain reactions get stopped as easily as they get started? Anybody here do nuclear reactors? They have rods they push in.
When they send the rocket to the moon it must exceed the "escape velocity," then gravity can no longer pull it back.
Not sure how to apply this to prison populations. "Warden, all you have to do is..."
"Oh...but, I cannot do that."
"I cannot release the prisoners."
We may be mixing our meta-phors...hope they don't blow up.
Sandpile metaphors are good too, although the diagrams aren't as evocative.
I'm intrigued by this definition of riot as derivative of disorder. What's going on in the maximum tension, minimum alienation scenario, that's nice and smooth to reach but just as disorderly as a full-blown riot?
I would guess that's a situation you would never actually allow the prison to reach unless you jumped there from one that was basically tolerable to the authorities
I'm not a huge fan of game theory, but I think it provides a better approach to understanding multiple equilibria. That's in part because we can describe them in a static setting. Dynamics would be more appealing if we had any idea how to model them in relation to social phenomena but we mostly don't and probably never will have.
This is basically what killed catastrophe theory - all the concepts are useful but there is almost always a better way to model any one of them
>I made a joke about Stafford Beer’s diagrams in the book, but I actually love hand-drawn graphs in articles and hope we never lose the art.
Maths departments in older universities will often have a collection of exotic surfaces with various curves traced on them rendered in plaster. It's particularly common in Germany, which was the real centre of geometry until WWII but I've seen them all over the place. In Strasbourg (Thom's university) they had a collection of model catastrophes. They're just curios now you can rotate surfaces to your heart's content in mathematica, but I find them charming too.
Ivar Ekeland has a nice little book called Mathematics and the Unexpected where he does a good job of connecting the maths and the metaphor without losing his head.
Oh man, that brings back some memories. I dabbled in catastrophe theory a long time ago, we had a Big Idea that the Thom classification theorem could be adapted to phase transitions in condensed-matter physics, but I was not able to get my head around the theorem.
My recollection of the vocabulary is different -- there's a "control space" (the 2D space in the picture), and a "response space" (the vertical axis), and the big deal is, as you say, that small changes in the control space can have discontinuous and large changes in the response space. I don't remember "control surface" being a thing. I have Arnol'd's book to hand, but it doesn't have an index.
Anyways, thanks for the nostalgia trip, is what I'm really saying.
I'll confess that I'd never heard of catastrophe theory until today but the Wikipedia article seems pretty good and suggests that a "control surface" is a set in the control parameter space, which was my intuition.
https://en.wikipedia.org/wiki/Catastrophe_theory
yep, that's the one. I am playing around with whether "control surface" is a suffiicently intuitive idea that I can use it in an expanded and metaphorical sense to describe the way that managers and policymakers interact with the world.
Maybe sometimes what we need to do is indirect, the way curling sweeping is:
https://www.greatcurling.com/post/mastering-simple-curling-sweeping-techniques-a-step-by-step-guide
People probably try to use direct control methods first, then later come to understand that only indirect methods fit the situation.
Sometimes you need a hammer, sometimes a kiss.
We also tend to prefer hard and simple and understandable methods to fuzzy or soft or non-graspable ones.
The first stealth aircraft like the F-117 had very angular shapes, because the computers they designed them with could only do so many calculations. With newer computers, they could calculate more angles, and were able to design shapes that were more fluid, and stealthy, like the B-2 bomber. Speaking of the B-2, it is so inherently unstable that a human can't control it, only a computer can respond fast enough. The pilot just tells it which way to go. It is like working with AI to solve a problem that you can't on your own.
So maybe having the right amount of data and the right amount of power to use it may be necessary conditions to solve certain problems. This may exceed what is available. Then, the problem doesn't get solved. Plus, even if you figure out a "solution," it may be too late (or too soon!) to use it. The world keeps changing around you. You can't step in the same river twice.
But you can step in the same s--t twice!
Helps to be a hedgehog AND a fox. Didn't a wise person say: get a practical person to solve a difficult problem, and an impractical person to solve an impossible problem?
Maybe kayaking is a good metaphor for running the rapids we are in right now? Just don't tell McKinsey about this, or they'll start peddling rafting tours:
Kayaking yadda yadda yadda Kayaking yadda yadda yadda Kayaking yadda yadda yadda
;-)
I’m indebted as this article has led me to a delightful archive of Zeeman’s work https://www.lms.ac.uk/archive/zeeman_archive/listing?field_archive_category_value=All&published_date%5Bmin%5D=&published_date%5Bmax%5D=&title=&page=3
Thom's "Structural Stability and Morphogenesis" is a bit like Topology written by Derrida, but it was unfairly criticized when it came out and many of its insights are yet to be incorporated into biology and social science. Unfortunately, I think the differential topology that Thom was pioneering wasn't the ideal mathematical formalism for the conceptual ideas he was trying to share.
Interesting. Does the diagram mean that when, after the riot, the administrators get tension back down to a certain level, the situation suddenly resolves and disorder suddenly drops just as it suddenly rose in the riot? (And is that accurate? I know nothing about the practicalities of prison administration but would have guessed that reducing tension post-riot would be difficult, maybe reflected by a different slope / curve, but not necessarily discontinuous in the way that the riot itself was.)
The jump downward is marked on a different version of the same diagram that Zeeman used in a talk as "truce", which seems fair enough. (The full model has a homeostatic parameter which basically says that people get tired of high disorder and bored with low disorder)
In the pendulum wave, there is order in disorder and vice versa, like yin containing yang. One property seems to predominate, then the other. It is also quite beautiful and moving:
https://www.youtube.com/watch?v=b8O1jlHyKpU
https://www.youtube.com/watch?v=LLwClbpnIaM
Humpty Dumpty fell off the wall.
After an avalanche or rockslide you cannot put the hill back together the same way.
All it takes is one skier to trigger an avalanche. The ski areas with big, steep slopes try to preemptively trigger avalanches with explosives, before they can kill anybody. Out in the back country, you are on your own.
There are certain things, that once seen, you cannot unsee. The investigators who were sent to document the massacre at Sandy Hook Elementary, including taking pictures of "everything," were severely traumatized by the experience. The town had to tear down the school and build a new one. To continue, they had to forget, as well as remember.
You cannot wind back time and make things un-happen, at least in this universe, at human scale.
You might find this interesting or at least entertaining: a series of “mutinies” not riots resolved
https://www.canada.ca/en/navy/services/history/dissension-in-the-ranks.html
I remember "catastrophe theory" when it was "hot" and was tried to fit a number of discontinuous phenomena. Even biologists were trying to use it. But, like other theories, it proved very difficult to apply, and more importantly, other non-linear models were just as applicable. AFAIK, it fell out of favor as a model. Fractals and Chaos Theory became the next "hot idea".
When I saw the words "control surface" I immediately thought of the control surfaces on aircraft, such as the small "wing" called the aileron that makes the big wing rise or fall, thereby causing the plane to turn by directing lift to one side, or the little "trim tabs" that make the control surfaces such as the elevator or rudder move, which then make the whole plane move on its lateral or vertical axes.
https://en.wikipedia.org/wiki/Trim_tab
https://www.linkedin.com/pulse/power-trim-tab-effective-change-management-ian-moorhouse-zimnf
Buckminster Fuller famously used the trim tab as a metaphor for how little things have the power to move big things. He regarded himself as a "trim tab" in human form.
The other aviation metaphor that immediately comes to mind is how you make a plane "stall," or suddenly lose lift, by trying to create too much lift, usually by pulling the nose too high, thereby exceeding a critical "angle of attack," the angle at which the wing moves through the air. The plane starts going down, and may even go into a spin if it is not flying straight into the airflow at the time. The has real consequences. I know three people who were killed this way.
Things happen gradually...
...until they happen SUDDENLY!
You go were going up. Then you are going DOWN. Even though the nose may be up! It is not intuitive.
The interesting thing is that nature tends to repeat, and patterns found in one part of it also show up in another part.
Then there's fractals...
Our culture has a bias then leans us towards mechanistic thinking, where we think about doing simple things, like pushing a button, that have predicable consequences. Sometimes this is true.
But the things that vex us the most usually can't be "solved" by pushing a button. You can't "solve" weather. By its nature, it must constantly be changing.
What engineers sometimes forget is that not all things can be "solved." We must learn nature's ways, and live in harmony with them. Feynman spelled it out: "Nature cannot be fooled."
Sometimes a 3-map can show more than a 2-D one can. What if it were animated? Or colors or sounds were added?
Was that a model of a prison riot, or an earthquake?