The Braitenberg's paradox: or whether control is everywhere

[Alex_Gomez-Marin]

If a Braitenberg vehicle can be rightfully called a “control system”, does it qualify immediately as a “perceptual control system”?

If it does so, then let’s take the Earth orbiting around the Sun: one could say that the forces at work (Gravitational, in this case) generate behavioural “outputs” that are sensed (via Gravitational forces, again) so as to keep the perception of “conservation of energy” constant. Actually, this is one of the great achievements of celestial mechanics: starting with patterns in the sky (geometry and kinematics), then casting laws (actually, scaling relations! liked Kepler’s 2/3power law!), then discovering the forces (a kind of “underlying mechanism”), and finally formulating it in terms of energies and conservation principles.

Then, by the same token, would the Earth be a “perceptual control system”?

If so, then anything in the entire Cosmos is. And so, if all is, then nothing is or it does not matter. What is the point, then? Is PCT just a kind of emphasis or loud speaker onto what all things do, within a gradient?

[EetuP]

[Eetu Pikkarainen 2018-02-15_12:16:13 UTC]

Good. As Aristotle said: ”A stone is a system which controls it’s perception of standing on the ground.” Newton corrected that it is a system which
controls its perception of staying in rest – or in stable movement.

I understand that a stone could perceive hits, pressure, warmth etc. but how can it perceive its position or state of movement?

I think I have learned that a PCT system has at least these kind of requirements:

···

From: Alex Gomez-Marin [mailto:agomezmarin@gmail.com]
Sent: 15. helmikuuta 2018 13:38
To: csgnet csgnet@lists.illinois.edu; AGM agomezmarin@gmail.com
Subject: The Braitenberg’s paradox: or whether control is everywhere

If a Braitenberg vehicle can be rightfully called a “control system”, does it qualify immediately as a “perceptual control system”?

If it does so, then let’s take the Earth orbiting around the Sun: one could say that the forces at work (Gravitational, in this case) generate behavioural “outputs”
that are sensed (via Gravitational forces, again) so as to keep the perception of “conservation of energy” constant. Actually, this is one of the great achievements of celestial mechanics: starting with patterns in the sky (geometry and kinematics), then casting
laws (actually, scaling relations! liked Kepler’s 2/3power law!), then discovering the forces (a kind of “underlying mechanism”), and finally formulating it in terms of energies and conservation principles.

Then, by the same token, would the Earth be a “perceptual control system”?

If so, then anything in the entire Cosmos is. And so, if all is, then nothing is or it does not matter. What is the point, then? Is PCT just a kind of emphasis or loud
speaker onto what all things do, within a gradient?

[Abbott]

[From Bruce Abbott (2018.02.15.1410 EST)]

[Eetu Pikkarainen 2018-02-15_12:16:13 UTC]

Good. As Aristotle said: ”A stone is a system which controls it’s perception of standing on the ground.” Newton corrected that it is a system which controls its perception of staying in rest – or in stable movement.

I understand that a stone could perceive hits, pressure, warmth etc. but how can it perceive its position or state of movement?

Aristotle said that? (Are you referring to Aristotle’s claim that a stone falls to Earth because the Earth is its natural place?) And I don’t recall Newton saying anything about a system “controlling its perception of staying at rest” – he only said that an object in motion stays in motion, and an object at rest stays at rest, unless acted upon by an outside force (the principle of inertia).

To suggest that a stone could perceive hits, pressure, warmth, etc. stretches the meaning of the word “perceive” beyond its usefulness. Aspects of its structure are affected by these variables, but we usually reserve the word “perceive” for cases in which there is a specific sensor that actively transduces such variables onto an internal signal, as opposed to a passive accommodation to changes in physical variables acting on the object. The perceptual signal acts on what we might term a “perceiver” – a mechanism that does something, based on the perception, as when a thermostat turns on the furnace when its perceived temperature falls below the set point. The only objects of which I am aware that meet this criterion are living systems and inanimate systems, like cruise control, that have been designed by human beings.

I think I have learned that a PCT system has at least these kind of requirements:

•      differentiation of the system and its environment (like subject and object)
  

•      two-way interaction between the system and the environment (both affect each other)
  

•      differentiation between input and output in that interaction (this is not possible in Newton’s law of force and counterforce)
  

•      asymmetry between input and output (input weak and output strong)
  

•      dependence of the output not only from input but from the error between internal fixed or changing reference value
  

•      negative feedback of output to input which tends to stabilize the input to a value which depends on the reference value
  

I think Bruce A [From Bruce Abbott (2018.02.14.1810 EST)] already explained how a Braitenberg vehicle is (at least implicitly) a PCT system. The simple Vehicle 1 (https://en.wikipedia.org/wiki/Braitenberg_vehicle) is not because does not stabilize. Instead the vehicles with two light sensors tend to stabilize their position in relation to light source depending on built in and in principle settable reference.

In one version of my infrared beam following program I arranged to make the forward speed of the vehicle inversely proportional to distance from the beam transmitter. In this it resembles a Braitenberg vehicle 1. The vehicle moves quickly when at some distance from the transmitter but slows as it approaches, and finally stops a short distance from it when the power to the motors is too low to turn them. If you then back the transmitter away from the vehicle, the vehicle resumes its forward motion until it again closes the distance.

So, all the earmarks of control are present – a controlled perception (sensed proximity to transmitter), negative feedback (output opposes an increase in distance), resistance to disturbance, and counteracting action powered by an external energy source (the vehicle’s battery) rather than by the disturbance itself. The Test for the controlled variable would even reveal the value of the (virtual) proximity reference. However, there is no internal reference signal and no error signal. It functions as a control system, but because the apparent reference emerges from a balance of forces (the vehicle stops when the motor power reduces to the point that friction prevents rotation), we might classify it as a somewhat “degenerate” type. However, control is control, even if it is achieved in a way that does not strictly follow the architecture of the prototypical PCT control-system diagram.

Bruce

[MartinT]

[From
Bruce Abbott (2018.02.15.1410 EST)]

        [Eetu Pikkarainen 2018-02-15_12:16:13 UTC]



        Good.

As Aristotle said: ”A stone is a system which controls it’s
perception of standing on the ground.” Newton corrected that
it is a system which controls its perception of staying in
rest – or in stable movement.

        I

understand that a stone could perceive hits, pressure,
warmth etc. but how can it perceive its position or state
of movement?

…

        To

suggest that a stone could perceive hits, pressure, warmth,
etc. stretches the meaning of the word “perceive” beyond its
usefulness.

[rsmarken]

[Rick Marken 2018-02-15_13:13:58]

···

On Thu, Feb 15, 2018 at 3:38 AM, Alex Gomez-Marin agomezmarin@gmail.com wrote:

AGM: If a Braitenberg vehicle can be rightfully called a “control system”, does it qualify immediately as a “perceptual control system”?Â

RM: A Braitenberg vehicle is called a control system because it is a closed-loop negative feedback system; it is what Bill called an N-system in the 1978 Psych Review paper. It’s an N system because it’s outputs have a negative effect on the inputs that drive those outputs; outputs drive the inputs toward zero. As Bill notes in that 1978 paper, all N-systems – all control systems – control their perceptual inputs. So all control systems are perceptual control systems: Braitenberg vehicles are perceptual control systems; water clocks are perceptual control systems; thermostats are perceptual control systems; and living organisms are perceptual control systems.Â

RM: Before about 1987 or so all of these N-systems were just called control systems. And that included living organisms. And Bill referred to his theory of the behavior of living organisms as “control theory” because that’s what the theory was. Bill didn’t invent control theory; he just applied it to understanding the behavior of organisms. But, as noted in his 1978 paper, control theory was already being applied to understanding the behavior of organisms when Bill published (in 1960) his first paper on what has come to be called perceptual control theory. Indeed, a whole field of psychology dedicated to understanding behavior in terms of control theory started years before Bill developed his own application of control they to behavior. What Powers did was see the correct way to apply control theory to behavior. This is all explained in the 1978 paper. So perceptual control theory is the theory formerly known as control theory.

RM: What Bill realized is that, when control theory is correctly applied to the behavior of living systems, it is important to understand that what control systems control is their perceptual input. Previous applications of control theory to behavior – influenced by the causal concept of behavior and the illusion that behavior is caused by perceptual input (the behavioral illusion) – saw perceptual input as the cause of behavioral output. So
we now refer to Powers application of control theory to behavior as perceptual control theory rather than just control theory to
 emphasize the fact that what is controlled when people control is perceptual input; the phrase distinguishes Powers’ correct application of control theory to behavior from the incorrect ones that were and are still prevalent.
But Powers’ theory is still just control theory – the theory of systems that control. It’s just applied correctly to behavior, which makes all the difference! By the way, the incorrect applications of control theory see the “error” in the control loop as being outside the system; error becomes a perceptual input that causes the output that corrects the error. This misapplication of control theory can still be used to evaluate the quality of control carried out by organisms; so these applications have produced a lot of research on how the spectral characteristics of “driving functions” (what we see as disturbances) affect the quality of control. But, as Bill mentions in that 1978 paper, that miss the whole point of control, which is to maintain perceptual input variables in autonomously determined reference states; that is, they miss the point that behavior is organized aroud the control of perception.Â

AGM: If it does so, then let’s take the Earth orbiting around the Sun: one could say that the forces at work (Gravitational, in this case) generate behavioural “outputs” that are sensed (via Gravitational forces, again) so as to keep the perception of “conservation of energy” constant. Actually, this is one of the great achievements of celestial mechanics: starting with patterns in the sky (geometry and kinematics), then casting laws (actually, scaling relations! liked Kepler’s 2/3power law!), then discovering the forces (a kind of “underlying mechanism”), and finally formulating it in terms of energies and conservation principles.

AGM: Then, by the same token, would the Earth be a “perceptual control system”?Â

AGM: If so, then anything in the entire Cosmos is. And so, if all is, then nothing is or it does not matter. What is the point, then? Is PCT just a kind of emphasis or loud speaker onto what all things do, within a gradient?

RM: Again, control theory (or perceptual control theory, same thing) is a theory of control. If you can demonstrate that control is involved in the behavior of any of these systems then control theory applies. My understanding is that the behavior of systems like the solar system is explained pretty well by causal models, suggesting that control is not involved and, thus, a control model is not needed.Â

Best

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

[Alex_Gomez-Marin]

thanks, Rick. that is helpful.Â

···

On Thu, Feb 15, 2018 at 10:14 PM, Richard Marken rsmarken@gmail.com wrote:

[Rick Marken 2018-02-15_13:13:58]

On Thu, Feb 15, 2018 at 3:38 AM, Alex Gomez-Marin agomezmarin@gmail.com wrote:

AGM: If a Braitenberg vehicle can be rightfully called a “control system”, does it qualify immediately as a “perceptual control system”?Â

RM: A Braitenberg vehicle is called a control system because it is a closed-loop negative feedback system; it is what Bill called an N-system in the 1978 Psych Review paper. It’s an N system because it’s outputs have a negative effect on the inputs that drive those outputs; outputs drive the inputs toward zero. As Bill notes in that 1978 paper, all N-systems – all control systems – control their perceptual inputs. So all control systems are perceptual control systems: Braitenberg vehicles are perceptual control systems; water clocks are perceptual control systems; thermostats are perceptual control systems; and living organisms are perceptual control systems.Â

RM: Before about 1987 or so all of these N-systems were just called control systems. And that included living organisms. And Bill referred to his theory of the behavior of living organisms as “control theory” because that’s what the theory was. Bill didn’t invent control theory; he just applied it to understanding the behavior of organisms. But, as noted in his 1978 paper, control theory was already being applied to understanding the behavior of organisms when Bill published (in 1960) his first paper on what has come to be called perceptual control theory. Indeed, a whole field of psychology dedicated to understanding behavior in terms of control theory started years before Bill developed his own application of control they to behavior. What Powers did was see the correct way to apply control theory to behavior. This is all explained in the 1978 paper. So perceptual control theory is the theory formerly known as control theory.

RM: What Bill realized is that, when control theory is correctly applied to the behavior of living systems, it is important to understand that what control systems control is their perceptual input. Previous applications of control theory to behavior – influenced by the causal concept of behavior and the illusion that behavior is caused by perceptual input (the behavioral illusion) – saw perceptual input as the cause of behavioral output. So
we now refer to Powers application of control theory to behavior as perceptual control theory rather than just control theory to
 emphasize the fact that what is controlled when people control is perceptual input; the phrase distinguishes Powers’ correct application of control theory to behavior from the incorrect ones that were and are still prevalent.
But Powers’ theory is still just control theory – the theory of systems that control. It’s just applied correctly to behavior, which makes all the difference! By the way, the incorrect applications of control theory see the “error” in the control loop as being outside the system; error becomes a perceptual input that causes the output that corrects the error. This misapplication of control theory can still be used to evaluate the quality of control carried out by organisms; so these applications have produced a lot of research on how the spectral characteristics of “driving functions” (what we see as disturbances) affect the quality of control. But, as Bill mentions in that 1978 paper, that miss the whole point of control, which is to maintain perceptual input variables in autonomously determined reference states; that is, they miss the point that behavior is organized aroud the control of perception.Â

AGM: If it does so, then let’s take the Earth orbiting around the Sun: one could say that the forces at work (Gravitational, in this case) generate behavioural “outputs” that are sensed (via Gravitational forces, again) so as to keep the perception of “conservation of energy” constant. Actually, this is one of the great achievements of celestial mechanics: starting with patterns in the sky (geometry and kinematics), then casting laws (actually, scaling relations! liked Kepler’s 2/3power law!), then discovering the forces (a kind of “underlying mechanism”), and finally formulating it in terms of energies and conservation principles.

AGM: Then, by the same token, would the Earth be a “perceptual control system”?Â

AGM: If so, then anything in the entire Cosmos is. And so, if all is, then nothing is or it does not matter. What is the point, then? Is PCT just a kind of emphasis or loud speaker onto what all things do, within a gradient?

RM: Again, control theory (or perceptual control theory, same thing) is a theory of control. If you can demonstrate that control is involved in the behavior of any of these systems then control theory applies. My understanding is that the behavior of systems like the solar system is explained pretty well by causal models, suggesting that control is not involved and, thus, a control model is not needed.Â

Best

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

Rick

[EetuP]

[Eetu Pikkarainen 2018-02-15_19:23:30 UTC]

Bruce,

[From Bruce Abbott (2018.02.15.1410 EST)]

[Eetu Pikkarainen 2018-02-15_12:16:13 UTC]

Good. As Aristotle said: ”A stone is a system which controls it’s perception of standing on the ground.” Newton corrected
that it is a system which controls its perception of staying in rest – or in stable movement.

I understand that a stone could perceive hits, pressure, warmth etc. but how can it perceive its position or state of
movement?

Aristotle said that? (Are you referring to Aristotle’s claim that a stone falls to Earth because the Earth is its natural place?)
And I don’t recall Newton saying anything about a system “controlling its perception of staying at rest” – he only said that an object in motion stays in motion, and an object at rest stays at rest, unless acted upon by an outside force (the principle of inertia).

I am sorry that, inspired by the question of Alex whether Earth controls, I took some (joker’s) freedoms and put words to others’ mouths. Aristotle did not use the word control but he
was a great forerunner of our control theory. He thought that every being has a purpose and we know that purpose came to world with control systems. So, if a stone had a purpose to stay on the ground then it should be a control system. Newton showed that this
is not the case but he (or his followers) went too far and taught that no physical being can have any purpose. Applying Alex’s saying that “if everything is then nothing is” I thought that if nothing controls the we can say that everything controls. But for
Newton the physical being is full passive staying in rest or stable movement if no external force affects: the control becomes a passive resistance.

To suggest that a stone could perceive hits, pressure, warmth, etc. stretches the meaning of the word “perceive” beyond its usefulness.
Aspects of its structure are affected by these variables, but we usually reserve the word “perceive” for cases in which there is a specific sensor that actively transduces such variables onto an internal signal, as opposed to a passive accommodation to changes
in physical variables acting on the object. The perceptual signal acts on what we might term a “perceiver” – a mechanism that does something, based on the perception, as when a thermostat turns on the furnace when its perceived temperature falls below the
set point. The only objects of which I am aware that meet this criterion are living systems and inanimate systems, like cruise control, that have been designed by human beings.

I admit that it sounds strange in this context. I meant that perceiving is based on being affected by external effects. Any system or sensor can perceive only something which affects it.

I think I have learned that a PCT system has at least these kind of requirements:

[Abbott]

[From Bruce Abbott (2018.02.16.1015 EST)]

[Eetu Pikkarainen 2018-02-15_19:23:30 UTC]

Bruce,

[From Bruce Abbott (2018.02.15.1410 EST)]

[Eetu Pikkarainen 2018-02-15_12:16:13 UTC]

Good. As Aristotle said: ”A stone is a system which controls it’s perception of standing on the ground.” Newton corrected that it is a system which controls its perception of staying in rest – or in stable movement.

I understand that a stone could perceive hits, pressure, warmth etc. but how can it perceive its position or state of movement?

Aristotle said that? (Are you referring to Aristotle’s claim that a stone falls to Earth because the Earth is its natural place?) And I don’t recall Newton saying anything about a system “controlling its perception of staying at rest” – he only said that an object in motion stays in motion, and an object at rest stays at rest, unless acted upon by an outside force (the principle of inertia).

I am sorry that, inspired by the question of Alex whether Earth controls, I took some (joker’s) freedoms and put words to others’ mouths. Aristotle did not use the word control but he was a great forerunner of our control theory. He thought that every being has a purpose and we know that purpose came to world with control systems. So, if a stone had a purpose to stay on the ground then it should be a control system. Newton showed that this is not the case but he (or his followers) went too far and taught that no physical being can have any purpose. Applying Alex’s saying that “if everything is then nothing is” I thought that if nothing controls the we can say that everything controls. But for Newton the physical being is full passive staying in rest or stable movement if no external force affects: the control becomes a passive resistance.

Ah, no need to apologize. I just didn’t catch the twinkle in your eye!

By the way, based on some very limited reading of Aristotle some years ago, it seemed to me that his conception of purpose (teleology) had been misunderstood. Aristotle was perhaps most interested understanding the biological world, and the characteristic of living things that stands out to distinguish them from nonliving is that their various parts apparently are as they are to serve particular purposes – e.g., teeth to cut or grind, wings to fly (form follows function). Aristotle’s “final cause” was interpreted by later writers to assert tha,t somehow, the need for some function in the future causes it to develop now. This idea was firmly rejected because (in this interpretation) it requires time reversal – the future affecting the present, and this was seen as violating the scientific principle that causes must precede their effects. But my view was that Aristotle was merely noting that one can understand why a given structure is as it is by discovering its function – e.g., the heart functions as a pump, as such it must have structures that exert pressure on the blood in certain ways (properly arranged muscles and properly timed contractions of them) and one-way valves. This is far from the notion that the need to pump blood caused a heart to develop. As to how such functional structures emerge, in one passage Aristotle comes very close to stating the theory of evolution, and I suspect would have taken the final step had he known that evolution occurs and requires explanation!

To suggest that a stone could perceive hits, pressure, warmth, etc. stretches the meaning of the word “perceive” beyond its usefulness. Aspects of its structure are affected by these variables, but we usually reserve the word “perceive” for cases in which there is a specific sensor that actively transduces such variables onto an internal signal, as opposed to a passive accommodation to changes in physical variables acting on the object. The perceptual signal acts on what we might term a “perceiver” – a mechanism that does something, based on the perception, as when a thermostat turns on the furnace when its perceived temperature falls below the set point. The only objects of which I am aware that meet this criterion are living systems and inanimate systems, like cruise control, that have been designed by human beings.

I admit that it sounds strange in this context. I meant that perceiving is based on being affected by external effects. Any system or sensor can perceive only something which affects it.

O.K., makes sense.

I think I have learned that a PCT system has at least these kind of requirements:

•      differentiation of the system and its environment (like subject and object)
  

•      two-way interaction between the system and the environment (both affect each other)
  

•      differentiation between input and output in that interaction (this is not possible in Newton’s law of force and counterforce)
  

•      asymmetry between input and output (input weak and output strong)
  

•      dependence of the output not only from input but from the error between internal fixed or changing reference value
  

•      negative feedback of output to input which tends to stabilize the input to a value which depends on the reference value
  

I think Bruce A [From Bruce Abbott (2018.02.14.1810 EST)] already explained how a Braitenberg vehicle is (at least implicitly) a PCT system. The simple Vehicle 1 (https://en.wikipedia.org/wiki/Braitenberg_vehicle) is not because does not stabilize. Instead the vehicles with two light sensors tend to stabilize their position in relation to light source depending on built in and in principle settable reference.

In one version of my infrared beam following program I arranged to make the forward speed of the vehicle inversely proportional to distance from the beam transmitter. In this it resembles a Braitenberg vehicle 1. The vehicle moves quickly when at some distance from the transmitter but slows as it approaches, and finally stops a short distance from it when the power to the motors is too low to turn them. If you then back the transmitter away from the vehicle, the vehicle resumes its forward motion until it again closes the distance.

I could not find a better description of the Braitenberg’s vehicle 1 than the short and indeterminate one in Wikipedia. That your vehicle clearly controls the strength of received beam but that Braitenberg’s vehicle 1 sounds to act so that it does not clearly control anything, but moves in surprising ways. But anyway it must be controlling the warmth somehow.

The description of Braitenberg’s vehicle 1 given in Wikipedia does describe a somewhat different arrangement than the one I described for my EV3 vehicle. In the former, forward speed is directly proportional to temperature. If the vehicle is passing through a temperature gradient in which the temperature first rises and then falls, the vehicle will speed up and then slow down. If the temperature falls sufficiently it will eventually stop. One can think of this as temperature control system with a virtual reference located at the temperature at which the vehicle stops.

This reminds me of the behavior of pill bugs (armadillididium), whose locomotion also depends on light intensity. They move until the intensity reaches a low value and, consequently, they generally end up under rocks, leaves, etc. where the moisture levels are higher. Pill bugs are not insects and their exoskeletons are not as good at retaining moisture, so they are in danger of drying out if they cannot find a relatively moist environment. As such environments also tend to be shielded from the light, their “aversion” to light helps to keep them moist.

So, all the earmarks of control are present – a controlled perception (sensed proximity to transmitter), negative feedback (output opposes an increase in distance), resistance to disturbance, and counteracting action powered by an external energy source (the vehicle’s battery) rather than by the disturbance itself. The Test for the controlled variable would even reveal the value of the (virtual) proximity reference. However, there is no internal reference signal and no error signal. It functions as a control system, but because the apparent reference emerges from a balance of forces (the vehicle stops when the motor power reduces to the point that friction prevents rotation), we might classify it as a somewhat “degenerate” type. However, control is control, even if it is achieved in a way that does not strictly follow the architecture of the prototypical PCT control-system diagram.

Yes, that is interesting. thank you

Bruce

[Alex_Gomez-Marin]

Let me add: “respect for Aristotle!”. So many philosophically illiterate lazy scientists (nowadays, most of them — a century ago probably none of them) make fun of hhim, and ridiculize his views —I have seen it done so many times in physics and in biology. He was one of the greatest minds of all times, but we think that with our internet, some cybernetic demos, and starbucks coffee, we can render their insights not only outdated, but irrelevant.

Read the first two paragraphs of the section “Explaining explanation” here.Â

And look at figure 4a here.

It is a must that we respect philosophy, and stop saying “this is just philosophy” to diminish the importance of something.

By the way, I was recently intrigued, surprised (and disappointed) by Wiener’s chapter in Cybernetics about Newtonian versus Bergsonian time. Another example of how a clever scientific mind can be pretty stupid philosophically.

Cheers,

Alex

···

On Fri, Feb 16, 2018 at 4:16 PM, Bruce Abbott bbabbott@frontier.com wrote:

[From Bruce Abbott (2018.02.16.1015 EST)]

Â

[Eetu Pikkarainen 2018-02-15_19:23:30 UTC]

Â

Bruce,

[From Bruce Abbott (2018.02.15.1410 EST)]

Â

[Eetu Pikkarainen 2018-02-15_12:16:13 UTC]

Â

Good. As Aristotle said: â€?A stone is a system which controls it’s perception of standing on the ground.â€? Newton corrected that it is a system which controls its perception of staying in rest – orr in stable movement.

I understand that a stone could perceive hits, pressure, Â warmth etc. but how can it perceive its position or state of movement?

Â

Aristotle said that? (Are you referring to Aristotle’s claim that a stone falls to Earth because the Earth is its natural place?) And I don’t recall Newton saying anything about a system “controlling its perception of staying at restâ€? – he only said that an object in motion stays in motion, and an object at rest stays at rest, unless acted upon by an outside force (the principle of inertia).

Â

I am sorry that, inspired by the question of Alex whether Earth controls, I took some (joker’s) freedoms and put words to others’ mouths. Aristotle did not use the word control but he was a great forerunner of our control theory. He thought that every being has a purpose and we know that purpose came to world with control systems. So, if a stone had a purpose to stay on the ground then it should be a control system. Newton showed that this is not the case but he (or his followers) went too far and taught that no physical being can have any purpose. Applying Alex’s saying that “if everything is then nothing isâ€? I thought that if nothing controls the we can say that everything controls. But for Newton the physical being is full passive staying in rest or stable movement if no external force affects: the control becomes a passive resistance.

Â

Ah, no need to apologize. I just didn’t catch the twinkle in your eye!

Â

By the way, based on some very limited reading of Aristotle some years ago, it seemed to me that his conception of purpose (teleology) had been misunderstood. Aristotle was perhaps most interested understanding the biological world, and the characteristic of living things that stands out to distinguish them from nonliving is that their various parts apparently are as they are to serve particular purposes – e.g., teeth to cut or grind, wings to flly (form follows function). Aristotle’s “final causeâ€? was interpreted by later writers to assert tha,t somehow, the need for some function in the future causes it to develop now. This idea was firmly rejected because (in this interpretation) it requires time reversal – the future affecting the present, and this was seen as violaating the scientific principle that causes must precede their effects. But my view was that Aristotle was merely noting that one can understand why a given structure is as it is by discovering its function – e.g…, the heart functions as a pump, as such it must have structures that exert pressure on the blood in certain ways (properly arranged muscles and properly timed contractions of them) and one-way valves. This is far from the notion that the need to pump blood caused a heart to develop. As to how such functional structures emerge, in one passage Aristotle comes very close to stating the theory of evolution, and I suspect would have taken the final step had he known that evolution occurs and requires explanation!

Â

To suggest that a stone could perceive hits, pressure, warmth, etc. stretches the meaning of the word “perceiveâ€? beyond its usefulness. Aspects of its structure are affected by these variables, but we usually reserve the word “perceiveâ€? for cases in which there is a specific sensor that actively transduces such variables onto an internal signal, as opposed to a passive accommodation to changes in physical variables acting on the object. The perceptual signal acts on what we might term a “perceiverâ€? – a mechanism that does sommething, based on the perception, as when a thermostat turns on the furnace when its perceived temperature falls below the set point. The only objects of which I am aware that meet this criterion are living systems and inanimate systems, like cruise control, that have been designed by human beings.

Â

I admit that it sounds strange in this context. I meant that perceiving is based on being affected by external effects. Any system or sensor can perceive only something which affects it.

Â

O.K., makes sense.

Â

I think I have learned that a PCT system has at least these kind of requirements:

Â

-Â Â Â Â Â Â Â Â Â differentiation of the system and its environment (like subject and object)

-Â Â Â Â Â Â Â Â Â two-way interaction between the system and the environment (both affect each other)

-         differentiation between input and output in that interaction (this is not possible in Newton’s law of force and counterforce)

-Â Â Â Â Â Â Â Â Â asymmetry between input and output (input weak and output strong)

-Â Â Â Â Â Â Â Â Â dependence of the output not only from input but from the error between internal fixed or changing reference value

-Â Â Â Â Â Â Â Â Â negative feedback of output to input which tends to stabilize the input to a value which depends on the reference value

Â

I think Bruce A [From Bruce Abbott (2018.02.14.1810 EST)] already explained how a Braitenberg vehicle is (at least implicitly) a PCT system. The simple Vehicle 1 (https://en.wikipedia.org/wiki/Braitenberg_vehicle) is not because does not stabilize. Instead the vehicles with two light sensors tend to stabilize their position in relation to light source depending on built in and in principle settable reference.

Â

In one version of my infrared beam following program I arranged to make the forward speed of the vehicle inversely proportional to distance from the beam transmitter. In this it resembles a Braitenberg vehicle 1. The vehicle moves quickly when at some distance from the transmitter but slows as it approaches, and finally stops a short distance from it when the power to the motors is too low to turn them. If you then back the transmitter away from the vehicle, the vehicle resumes its forward motion until it again closes the distance.Â

Â

I could not find a better description of the Braitenberg’s vehicle 1 than the short and indeterminate one in Wikipedia. That your vehicle clearly controls the strength of received beam but that Braitenberg’s vehicle 1 sounds to act so that it does not clearly control anything, but moves in surprising ways. But anyway it must be controlling the warmth somehow.

Â

The description of Braitenberg’s vehicle 1 given in Wikipedia does describe a somewhat different arrangement than the one I described for my EV3 vehicle. In the former, forward speed is directly proportional to temperature. If the vehicle is passing through a temperature gradient in which the temperature first rises and then falls, the vehicle will speed up and then slow down. If the temperature falls sufficiently it will eventually stop. One can think of this as temperature control system with a virtual reference located at the temperature at which the vehicle stops.Â

Â

This reminds me of the behavior of pill bugs (armadillididium), whose locomotion also depends on light intensity. They move until the intensity reaches a low value and, consequently, they generally end up under rocks, leaves, etc. where the moisture levels are higher. Pill bugs are not insects and their exoskeletons are not as good at retaining moisture, so they are in danger of drying out if they cannot find a relatively moist environment. As such environments also tend to be shielded from the light, their “aversionâ€? to light helps to keep them moist.

Â

So, all the earmarks of control are present – a controlled perception (sensed proximitty to transmitter), negative feedback (output opposes an increase in distance), resistance to disturbance, and counteracting action powered by an external energy source (the vehicle’s battery) rather than by the disturbance itself. The Test for the controlled variable would even reveal the value of the  (virtual) proximity reference. However, there is no internal reference signal and no error signal. It functions as a control system, but because the apparent reference emerges from a balance of forces (the vehicle stops when the motor power reduces to the point that friction prevents rotation), we might classify it as a somewhat “degenerateâ€? type. However, control is control, even if it is achieved in a way that does not strictly follow the architecture of the prototypical PCT control-system diagram.

Â

Yes, that is interesting. thank you

Â

Bruce

[FredNickols]

I read the first two paragraphs, Alex, and it seems to me you’ve got something backward. Last time I looked, it’s the mouse (prey) that tries to escape the cat (predator). You’ve got the cat trying to escape the mouse.

Fred Nickols

···

From: Alex Gomez-Marin [mailto:agomezmarin@gmail.com]
Sent: Friday, February 16, 2018 10:24 AM
To: csgnet csgnet@lists.illinois.edu
Subject: Re: The Braitenberg’s paradox: or whether control is everywhere

Let me add: “respect for Aristotle!”. So many philosophically illiterate lazy scientists (nowadays, most of them — a century ago probably none of them) make fun of him, and ridiculize his views —II have seen it done so many times in physics and in biology. He was one of the greatest minds of all times, but we think that with our internet, some cybernetic demos, and starbucks coffee, we can render their insights not only outdated, but irrelevant.

Read the first two paragraphs of the section “Explaining explanation” here.

And look at figure 4a here.

It is a must that we respect philosophy, and stop saying “this is just philosophy” to diminish the importance of something.

By the way, I was recently intrigued, surprised (and disappointed) by Wiener’s chapter in Cybernetics about Newtonian versus Bergsonian time. Another example of how a clever scientific mind can be pretty stupid philosophically.

Cheers,

Alex

On Fri, Feb 16, 2018 at 4:16 PM, Bruce Abbott bbabbott@frontier.com wrote:

[From Bruce Abbott (2018.02.16.1015 EST)]

[Eetu Pikkarainen 2018-02-15_19:23:30 UTC]

Bruce,

[From Bruce Abbott (2018.02.15.1410 EST)]

[Eetu Pikkarainen 2018-02-15_12:16:13 UTC]

Good. As Aristotle said: â€?A stone is a system which controls it’s perception of standing on the ground.â€? Newton corrected that it is a system which controls its perception of staying in rest – or in stable movement.

I understand that a stone could perceive hits, pressure, warmth etc. but how can it perceive its position or state of movement?

Aristotle said that? (Are you referring to Aristotle’s claim that a stone falls to Earth because the Earth is its natural place?) And I don’t recall Newton saying anything about a system “controlling its perception of staying at restâ€? – he only said that an object iin motion stays in motion, and an object at rest stays at rest, unless acted upon by an outside force (the principle of inertia).

I am sorry that, inspired by the question of Alex whether Earth controls, I took some (joker’s) freedoms and put words to others’ mouths. Aristotle did not use the word control but he was a great forerunner of our control theory. He thought that every being has a purpose and we know that purpose came to world with control systems. So, if a stone had a purpose to stay on the ground then it should be a control system. Newton showed that this is not the case but he (or his followers) went too far and taught that no physical being can have any purpose. Applying Alex’s saying that “if everything is then nothing is� I thought that if nothing controls the we can say that everything controls. But for Newton the physical being is full passive staying in rest or stable movement if no external force affects: the control becomes a passive resistance.

Ah, no need to apologize. I just didn’t catch the twinkle in your eye!

By the way, based on some very limited reading of Aristotle some years ago, it seemed to me that his conception of purpose (teleology) had been misunderstood. Aristotle was perhaps most interested understanding the biological world, and the characteristic of living things that stands out to distinguish them from nonliving is that their various parts apparently are as they are to serve particular purposes – e.g., teethh to cut or grind, wings to fly (form follows function). Aristotle’s “final causeâ€? was interpreted by later writers to assert tha,t somehow, the need for some function in the future causes it to develop now. This idea was firmly rejected because (in this interpretation) it requires time reversal – the future affecting the presentt, and this was seen as violating the scientific principle that causes must precede their effects. But my view was that Aristotle was merely noting that one can understand why a given structure is as it is by discovering its function – e.g., the heart functions as a pump, as such it musst have structures that exert pressure on the blood in certain ways (properly arranged muscles and properly timed contractions of them) and one-way valves. This is far from the notion that the need to pump blood caused a heart to develop. As to how such functional structures emerge, in one passage Aristotle comes very close to stating the theory of evolution, and I suspect would have taken the final step had he known that evolution occurs and requires explanation!

To suggest that a stone could perceive hits, pressure, warmth, etc. stretches the meaning of the word “perceiveâ€? beyond its usefulness. Aspects of its structure are affected by these variables, but we usually reserve the word “perceiveâ€? for cases in which there is a specific sensor that actively transduces such variables onto an internal signal, as opposed to a passive accommodation to changes in physical variables acting on the object. The perceptual signal acts on what we might term a “perceiverâ€? – a mechanism that does something, based on the perception, as when a thermostat turns on the furnace when its perceived temperature falls below the set point. The only objects of which I am aware that meet this criterion are living systems and inanimate systems, like cruise control, that have been designed by human beings.

I admit that it sounds strange in this context. I meant that perceiving is based on being affected by external effects. Any system or sensor can perceive only something which affects it.

O.K., makes sense.

I think I have learned that a PCT system has at least these kind of requirements:

•      differentiation of the system and its environment (like subject and object)
  

•      two-way interaction between the system and the environment (both affect each other)
  

•      differentiation between input and output in that interaction (this is not possible in Newton’s law of force and counterforce)
  

•      asymmetry between input and output (input weak and output strong)
  

•      dependence of the output not only from input but from the error between internal fixed or changing reference value
  

•      negative feedback of output to input which tends to stabilize the input to a value which depends on the reference value
  

I think Bruce A [From Bruce Abbott (2018.02.14.1810 EST)] already explained how a Braitenberg vehicle is (at least implicitly) a PCT system. The simple Vehicle 1 (https://en.wikipedia.org/wiki/Braitenberg_vehicle) is not because does not stabilize. Instead the vehicles with two light sensors tend to stabilize their position in relation to light source depending on built in and in principle settable reference.

In one version of my infrared beam following program I arranged to make the forward speed of the vehicle inversely proportional to distance from the beam transmitter. In this it resembles a Braitenberg vehicle 1. The vehicle moves quickly when at some distance from the transmitter but slows as it approaches, and finally stops a short distance from it when the power to the motors is too low to turn them. If you then back the transmitter away from the vehicle, the vehicle resumes its forward motion until it again closes the distance.

I could not find a better description of the Braitenberg’s vehicle 1 than the short and indeterminate one in Wikipedia. That your vehicle clearly controls the strength of received beam but that Braitenberg’s vehicle 1 sounds to act so that it does not clearly control anything, but moves in surprising ways. But anyway it must be controlling the warmth somehow.

The description of Braitenberg’s vehicle 1 given in Wikipedia does describe a somewhat different arrangement than the one I described for my EV3 vehicle. In the former, forward speed is directly proportional to temperature. If the vehicle is passing through a temperature gradient in which the temperature first rises and then falls, the vehicle will speed up and then slow down. If the temperature falls sufficiently it will eventually stop. One can think of this as temperature control system with a virtual reference located at the temperature at which the vehicle stops.

This reminds me of the behavior of pill bugs (armadillididium), whose locomotion also depends on light intensity. They move until the intensity reaches a low value and, consequently, they generally end up under rocks, leaves, etc. where the moisture levels are higher. Pill bugs are not insects and their exoskeletons are not as good at retaining moisture, so they are in danger of drying out if they cannot find a relatively moist environment. As such environments also tend to be shielded from the light, their “aversion� to light helps to keep them moist.

So, all the earmarks of control are present – a controlled perception (sensed proximiity to transmitter), negative feedback (output opposes an increase in distance), resistance to disturbance, and counteracting action powered by an external energy source (the vehicle’s battery) rather than by the disturbance itself. The Test for the controlled variable would even reveal the value of the (virtual) proximity reference. However, there is no internal reference signal and no error signal. It functions as a control system, but because the apparent reference emerges from a balance of forces (the vehicle stops when the motor power reduces to the point that friction prevents rotation), we might classify it as a somewhat “degenerateâ€? type. However, control is control, even if it is achieved in a way that does not strictly follow the architecture of the prototypical PCT control-system diagram.

Yes, that is interesting. thank you

Bruce