FW: What's perception got to do, got to do, with it?

[_Boris_Hartman]

Bruce,

it took me quite a long time to answer you. I admit I was deeply surprised about your answer.

I’m sorry Bruce, but it seems that we can’t come to some perfect agreement. Your opinion is quite clear. I think that we agree in general view how organism works and how it produces behavior, but there are some “detailsâ€? where we are practically on extreme, contrary positions.

1.  My opinion is that there is **no “subtractor�** (o-d) in environment in the “input quantity«. At least I never saw anywhere Bill using this combination. But “adder� (o+d) as it's seen from Bill's diagram as “**add effects**�, is something what is common in Bill’s theory, so effects of output and disturbances are “added� in environment. It's clear in his diagram.
   

It’s surprising that this is also Rick’s opinion if I may judge from his articles. One thing where we agree.

2.  I think that the only “controlled variable� in control loop is “perceptual signal� (hence Control of perception). Perceptual signal is to be “controlled� in comparator to some reference level, which is formed inside organism.
   

Bill P : So for “the organism to have control over any sensed condition of the environment, all that is necessary is for it to possess the means to cause that condition (as perceived) to vary in each of the possible ways it can varyâ€?….(B:CP).

&nbspp;

I see this as continuous “control”, and is diffferent to »sometimes controlled quantity" outside organism, which is by my opinion “imagined” construct, and “show” from time to time »counteracting« effects in environemnt. . Organism can’t control discretly from time to time. Their “Control of perception” is continuos.

I understand Bill as that organism »sense conditions« and control them. Not vica verse, that conditions are controlled and then organism sense »controlled conditions« through »controlled perceptual variable«. So I don’t think “perceptual signalâ€? can be something controlled from outside, to become “controlled perceptual signalâ€?, and in that sense “outside environmentâ€? could control organism. I think that “perceptual signal” is controlled inside organism to some genetically set references (directly or indirectly).

Bill P (LCS III) : Therefore changing the reference signal (with active cooperation from components in the control loop) very reliably changes the magnitude or state of the input quantity…

Comparator is by my oppinon the place for »control« and other components in control loop are helping control to be accomplished by »converting« effects through control loop.

I don’t see any »controlled sensed conditions« coming from environment.so that we could talk about »controlled perceptual variable«.

So I don’t think that »subtraction« in form of “Control of disturbances� (opposing) could possibly “control input�. I think that “perceptual signal� can not be “controlled� in advance from outer environment, but »perceptual signal« itself  is »controlled«. So I think that the only “controlled variable� in control loop is perceptual signal (hence »Control of perception«).

But people perceive »facts« obviously differently. The cause is probably variety of perceptions and variety of their control in comparator (nervous system).

3.  Behavior for me is not “control� because “output� is not controlled. So I think it's “control empty� and it can't control anything in outer  environment. Effects  of behavior to environment are in accordance with internal control of controlling system. So effects of internal control can by my opinion be limitly observed in outer environment as interpreted by observers own way.
   

As for equality of your and Bill’s opinion is concerned I think that you are not perfectly aligned as you pointed out. If you remember our conversation about “Kid who doesn’t talkâ€?, you accused Bill that he is seriously misleading CSGnet. You can check it. So I’m still wondering, how could author of theory and founder of CSGnet seriously mislead somebody about his theory ? So I have a reason to beleive that your and Bill’s opinion are not aligned. Speccially not about “adderâ€? and “subtractorâ€? in the form of “input quantityâ€?.

This is just my opinion. So members can decide whatever they wish.

As for cybernetics you talked about is concerned I made a separate thread.

Boris

···

From: Bruce Abbott [mailto:bbabbott@frontier.com]
Sent: Saturday, November 28, 2015 3:43 PM
To: csgnet@lists.illinois.edu
Subject: RE: FW: What’s perception got to do, got to do, with it?

[From Bruce Abbott (2015.11.28.0940 EST)]

From: Boris Hartman [mailto:boris.hartman@masicom.net]
Sent: Saturday, November 28, 2015 1:20 AM
To: csgnet@lists.illinois.edu
Subject: RE: FW: What’s perception got to do, got to do, with it?

Hi Boris,

In my replies to your questions, I will refer you to the screen shot of the LiveBlock program from LCS III shown below:

HB: I’d like to be sure if I understand clearly what you are saying. So I’ll try to sum only the part which I think is different by our oppinions.

BA : In the diagram below, the controlled variable is the small circle labeled as the »input quantity.« Its value is affected by the disturbance and by output quantity, operating on the input quantity via the feedback function. The feedback is negative, so adding the effects of feedback and disturbance accomplishes subtraction.

In the LiveBlock diagram above, the input quantity is represented by the small circle seen below the »System-Environment Boundary,« i.e., in the environment, and by the arrow leaving that circle and entering the Input Function. It is affected by disturbances (vertical arrow) and the feedback quantity (horizontal arrow coming from the Feedback Function). These two quantities add to the Input Quantity, as shown by the two plus signs. However, in normal operation the disturbance and feedback quantities change in opposite direction (because the feedback is negative), so the effect is one of subtraction, i.e., effect of feedback on the input quantity subtracts from the effect of the disturbance on the input quantity.

The circle represents the addition of these two influences to the input quantity. The arrow leaving the circle represents the resulting value of the input quantity. Both of these influences (disturbance and feedback), as well as the input quantity itself, lie in the environment of the control system.

If this were a home heating system, the input quantity would be the actual room temperature, the output quantity would be heat from the furnace, the feedback quantity would be the effect of that heat on the room’s temperature, and disturbances would be anything that disturbs actual room temperature, such as leakage of heat through the walls. Leakage of heat acts to reduce actual room temperature, whereas heat from the furnace acts to increase actual room temperature. Thus the action of the furnace (heat production) counteracts the effect of disturbances (heat loss).

HB : Do I understand right that »subtractor« is outside the »controlling system« ?

Yes, it is the little circle in the above diagram, found in the environment of the control system.

HB: And that »input function« is the place where »subtraction« is going on ?

No. The little circle is not the input function. The little circle (and the arrow leaving it) is the input quantity, qi. In my example, it’s the room temperature. The input function is the rectangle labeled as such in the LiveBlock diagram. It receives input from the input quantity and generates the perceptual signal as its output.

HB: And »input quantity« is fixed, really existing entity in environment ?

Yes, the input quantity is a really existing entity in the environment. (In the example it’s the room temperature.) It varies in response to the effects of the disturbance and feedback, so, no, it is not fixed.

HB: Did Bill mentioned somewhere specifically that this is happening in environment ?

Yes. Please note that the LiveBlock diagram is Bill’s diagram.

HB: Do I understand right that »output quantitty« is affecting controlled variable the »input quantity« ?

In the LiveBlock diagram, the output quantity is represented by the arrow leaving the output function and entering the feedback function. The output of the feedback function is the arrow entering the circle where the feedback and disturbances act on the input quantity.

In some versions of the diagram, the output quantity is shown directly affecting the input quantity. This is a convience done to simplify the diagram for purposes of exposition. When this is done, it is equivalent to assuming that the feedback function simply multiplies the output by a factor of 1.

HB: So can I understand that actions (output) are not controlled ? Just effects ?

In the diagram, actions (output) are not controlled. The output is simply a function of the error signal. The larger the error signal, the greater the magnitude of the output.

HB: And can I conclude that Warren’s and Tim’s conclussion in their article is not right :

Warren and Timothy : …«universal property of organisms : actions as the control of sensory input«.

Look at the LiveBlock diagram. Actions (output) are transformed into effects on the input quantity by the environmental feedback function. They oppose the effects of disturbances on the same input quantity. The input quantity affects the control system through its sensory inputs. (In the case of the home heating system the thermostat contains the sensor of room temperature.) The Input Function takes qi as input and converts this quantity to a perceptual signal that varies as a function of the input quantity. That signal is compared to the reference signal at the comparator (middle upper block in the diagram). The comparator produces the error signal, which determines the magnitude the the output quantity via the output function.

By opposing the effects of the disturbance on qi, the system acts to control its own sensory inputs, as the system perceives them.

HB : Or they are right ? If we go with backward analyses in your oppinion about »control« we can maybe say that »output quantity« is having some controlled effects on »input quantitty« ? Or not ? Or Rick is also not right when he is saying that »behavior is control« ?

That depends on what you mean by »controlled effects.« The output quantity, acting via the feedback function, affects the input quantity (e.g., heat from the furnace affects the room temperature). But this effect by itself is not control. Control emerges from the »behavior« of the system as a whole – from the fact that output is a function of error, that the feedback onto qi is negative, and that this feedback is scaled to an appropriate magnitude and applied without too much delay. If these factors are right, then the effect will be to oppose the effect of disturbances so as to bring the internal representation of qi (its perception) close to the reference value and hold it there – in other words, the system will control.

BA : The feedback compensates for the disturbance, so that its effect on the input quantity is minimized. That is, the value of the input quantity is controlled. The input quantity is thus the controlled variable in the environment.

HB : If I understand right »compensation« happens in outer environment, not in the whole control loop ? And the »input quantity« is the only »controlled variable« in the control loop ?

Yes, compensation happens in the outer environment, where feedback affects the input quantity. However, the input quantity determines the value of the perceptual signal, which is what the control system perceives. (It has no direct knowledge of the input quantity.) The control system controls its perception of the input quantity. To the extent that this perception corresponds to the value of the input quantity, then the input quantity is also controlled.

BA : The input function converts this input (sensed value of the controlled variable) to an internal representation of the controlled quantity called the perceptual signal. Because the perceptual signal tracks the value of the controlled variable in the environment, it is also controlled.

HB : Do I understand right that, the only »controlled variable« in environment »input quantitty« is converted to internal representation »perceptual signal«, which is thus »controlled perceptual signal« ?

Yes!

BA : The perceptual signal is the person’s inner representation of the controlled variable in the environment, so if the environmental variable is controlled, so is its perception. We can only directly know our perceptions, so from the organism’s point of view, what is controlled are its perceptions.

HB : Do I understand right that »controlled perceptions« which are the result of »control process« in environment are representing the organism’s point of view of what is being controlled ?

Yes!

BA : I hope I’ve made it clear by now that Bill and I are saying the same thing; I am not proposing any changes. If I had that wrong, surely Bill would have pointed that out to me.

HB : I think that equating your and Bill’s knowledge can wait a little, because whatever you presented till now is not what is »equating« your and Bill oppinion. I’ll explain later as we »define« control.

Well, that’s disappointing, because as I see it, Bill’s opinion and mine are the same.

BA : Boris, how do you define »control«? As I define it, control involves acting on a variable in such as way (via negative feedback) as to cancel the effects of disturbances to that variable.

HB : This is maybe general definition of the »control« in machines. But not in organism. Maybe here is misunderstanding. In organism control is defined as :

Bill P. and others (50th Anniversary) :

1.  **Perceptual Control Theory (PCT) provides a general theory of functioning for organisms.**
   

2.  **At the conceptual core of the theory is the observation that living things control the perceived environment by means of behavior**
   

Also Ashby used defitnition of »control« in organisms in very similar way when he defined homeostasis :

Ashby (1960) : I propose defitnition that a form aof behaviour is adaptive if it maintains the essential variables within physiological limits.

As I see it, control in PCT is going on in organism continuosly (with milions of cooperating control units) and through actions on environment, which also cancel the effects of disturbances.

Behavior is supporting to control in organism. So it’s not necesary that it appears in every moment. The output »o« can be zero, and also »d« can be zero. And sometimes both can limit zero. Whatever imagined construct »input quanttity« represent, it’s not a general concept which could »cover« all control situations, as there are actually cases when there is no »control« in »input quantity« by your definition. But in organism control runs all the time, it can’t be that control in organisms don’t exist or stop. Organism dies.

If you are lying on the sun (sunbathing) there is no action »o«, but disturbances »d« are acting on organism, including sensors.

The sun is heating my body and I am just lying there. But absence of action does not imply absence of control. Absence of action in this case means absence of error. My skin temperature as I perceive it is close enough to my internal reference that there is no error and therefore I take no action. But if my skin temperature rises enough due to the disturbing effect of the sun on my skin’s temperature, error will develop and I will move to the shade or cover up.

HB: So there is no »control« in outer environment and still organism is controlling. Orgnisms control inside not outside. I think that whatever you tried to represent as »control« is not »control in organisms«, which has to be in accordance with physiological »facts«, but control of »machines«, what by my oppinion is not the same.

While I am lying there in the sun, my skin (or body) temperature may not be hot enough to generate sufficient error that I will take action via my musculature. However, internal control systems are still monitoring skin and body temperature and comparing these as sensed to reference values. If there is error these physiological, homeostatic mechanisms will initiate actions of their own, such as sweating, that act to oppose the sun’s disturbance. If ultraviolet rays are damaging the skin, other control systems that sense this damage will take action to oppose this effect of the sun, such as generating melonin in the skin, which absorbs the damaging rays.

All of these systems – those that require the action of the muscles and those that act inside the body in other ways (via the autonomic nervous system, glands, etc. – act all the time, and all operate via the physiological systems of the body. There simply isn’t any difference. They are all control systems. The muscles act on the environment outside the skin, but those other control systems, though they act internally with respect to the body surface, still exert actions the affect variables in the environments of those control systems. It just happens that those invironments are inside the skin.

HB: You also exposed similar view as Bill and Ashby did in your Synopsis :

BA : At the heart of perceptual control theory is the idea that human beings are essentially intricate control mechanisms that function to keep certain intrinsic (or essential, see Ashby, 1952) variables within survivable limits. Intrinsic variables (variables intrinsic to the organism) include basic physiological variables such as blood glucose levels or body temperature, as well certain high-level variables whose maintenance in certain states are crucial to the individual’s well-being; the references of these intrinsic variables are genetically specified. With respect to physiological quantities, the body is known to house numerous control mechanisms that help to maintain them within the narrow limits required for efficient operation and survival. These mechanisms are capable of sensing the current levels of these controlled quantities and automatically initiating physiological changes as necessary to correct deviations of these levels from reference values, a process that the early 20th century physiologist Walter Cannon (1932) termed homeostasis.

BA: Although at any given moment a tremendous number of physiological quantities is being automatically regulated through nonbehavioral (purely physiological) means, the regulatory mechanisms by themselves are not capable of countering all the sources of potential disturbance to the intrinsic variables. To take one example, because humans are not rooted in the soil like plants, we must seek out and consume food and water. Automatic physiological mechanisms do act against disturbances to internal levels of water and nutrition, but these only can take the form of actions to reduce the rate of depletion of these quantities. To replenish them, we must behave. That is, we must move our muscles in a way that ultimately leads to locating, obtaining, and consuming food and water. Behavior, then, is a means by which humans (and other animals) defend their intrinsic variables against disturbance.

HB : Your view upon PCT in Synopsis as I said many times before, influenced my view. In Synopsis you are clearly saying that organisms »control« inside »at any given moment«, but with behavior organisms have to »replenish essential quantities« from time to time. Behavior is not continuous. So effects of behavior on environment are occasional, when organisms needs it. And that can’t be control, because »control in organism« is continuous inside and behavior is just means of control (inside).

Where you are going wrong, in my opinion, is assuming that lack of action implies lack of control. Of course, if we are not controlling a given variable, we take no action when the variable changes its value. But what of the variables that we do control? Lack of action means lack of sufficient error to support any action.

For many variables that we control, there is a »deadband« or range of values across which no error will be generated. The current value is »close enough« to reference. We don’t eat continuously because there is a range of nutrient levels in the digestive system and elsewhere in the body over which you will not feel hungry. Only when these levels fall below some threshold will there be an error that will impell you to action.

Another misconception you may be having is believing that the term »environment« applies only to variables outside the body. That is the environment that your muscles act on in order to control certain environmental variables, but for other control systems the environment in which they act is internal. In HPCT, for example, the »envronment« of a Level 2 control system is the Level 1 control system through which it acts and from which it receives its perceptual signals.

On further point: Imagine that I’m lying in the sun and because of some drug effect I am no longer conscious of my skin temperature. My skin temperature is rising well above the level at which I would normally take action by moving into the shade. But my perception of skin temperature is close to my internal reference for it, so I do nothing. I’m still controlling perceived skin temperature, but no longer controlling actual skin temperature.

Now imagine the opposite scenario: the drug is affecting my perception of skin temperature so that it does not relate to actual skin temperature. Suddenly my skin feels uncomfortably hot and I immediately move into the shade, even though my actual skin temperature is within the range of values over which I normally would not take action. I’m still contolling percieved skin temperature, but now this control is allowing actual skin temperature to fluctuate wildly. The latter is no longer controlled, even though perceived skin temperature still is.

The two scenarios above demonstrate why Bill placed such an emphasis on control of perception. Much of the time, by controlling a perception, we also control the environmental analog of that perception, the input quantity. In such cases the input quantity is properly referred to as the controlled quantity. (Bill used the term »quantity« to refer to physical variables, as distinguished from the term »signal,« which conveys information internally.) Sometimes the perceptual signal does not correlate well with any environmental variable, yet the control system may still be able to control its perception, as in the cases described above. Control is always control of perception, although environmental correlates of those perceptions may (or may not) be controlled as well.

HB: Your oppinion in Synopsis is by my oppinion in accordance with Ashby and Bill and I of course support it.

HB: I hope it’s O.K. that before answering on your whole post I checked whether I understand clearly what are you saying, because I have impression that something disturbed your »whole« judgement about how organisms control. Well if I’m sincere I blame Rick for this confussion. I just wanted to avoid misunderstandings with you. And if possible I’d like to come with you in agreement. As I said I highly value your oppinion.

That’s certainly O.K. with me, Boris. I hope my replies have helped to bridge the differences between us.

Best wishes,

Bruce A.

[_Boris_Hartman]

Bruce,

Your theoretical explanation of Wieners »control theory« is bellow so I’ll not »copy-paste« it here. I’ll just try to look for differences in Bill’s and Wieners theories. If it is »the same«, why did Bill wrote his theory ?

BA : Boris, you can’t be a proponent of PCT and at the same time hold that we shouldn’t be equating machines and organisms. I’ll try to make clear why this is so.

BA : Now let’s examine what you take Bill to be saying about control systems.

HB : I think that Bill criticized Wiener and “engineering control theoryâ€?. So I think there are more problems in just “equatingâ€? theories …

Bill P : For now the relevant aspect of Wiener’s diagram is the fact that it is drawn so that one may consider the whole unit in terms of stimulus and respons, just as behaviorists have always done.

Bill P : Unspoken in Wiener’s diagram, by the way, are the reorganizing system’s intrinsic reference levels….

<

Bill P :

Regarding the underpinnings of PCT: There was no one in cybernetics/systems theory after Ashby’s book in 1953 (Design for a Brain) from whom I learned anything about control theory and its role in behavior. Wiener’s book of 1948, which I read in 1952 thanks to Kirk Sattley, got me started: the concept of feedback control, and the particular relations to behavior that he laid out, clicked in my mind as the obvious successor to all the psychological models I had ever heard of, including the one in which I then believed. Ashby’s book gave me an organized view of how one would start applying these principles on a grander scale—itt was as much his organization as his ideas that turned me on.

Bill P : Wiener and Ashby inspired me to go back to the sources of the ideas that they had adopted. When I did, I gradually came to realize that neither of them had learned very much about control systems.

HB : I’m sure you can find more Bill’s critics on Wieners model. what by my opinion tell us that thinking about “Control theory� wasn’t simply transferred to PCT.

But I think that most comprehesible critics of engineering “control theory� was given by Henry Ying. His article is worth of reading as I think he is the only one whom I saw upgrading PCT till now. I think that he could be very helpful in conflicting understanding of PCT.

Henry Yin :

4.2. Control of Input. A control system always controls its input, not output [7]. Only perceivable consequences of behavior can be controlled.

According to mainstream engineering control theory, a control system controls its outputs, not its input. This is perhaps the most common fallacy today, both in engineering and in the life sciences [49, 55, 56]. This fallacy, an unfortunate legacy of cybernetics, is the result of imposing the perspective of the observer rather than using the perspective of the organism or controller. The mistake is to assume that what the engineer perceives and records, the “objective� effect of the system, is the output of the system.

The common assumption that output is controlled ignores the perspective of the organism that is doing the controlling. By imposing his own desire and perspective, the engineer ignores the autonomy of the negative feedback controller, for he is always trying to make the machine do what he wants. He can only accomplish this by adjusting the reference signal, as the user operates a thermostat by

adjusting the temperature setting. Since this is the signal generated by the user, it is usually labeled as the input to the system. In a biological organism, however, the reference signal is always internal to the organism.

The real input is the perceptual variable that can be affected by feedback [49]. In a temperature controller using negative feedback, the perceptual input is from the temperature sensors. Of course, in a man-made thermostat, the user can adjust the “set point,� but that is a unique feature of these systems, because that is the only way to use a negative feedback control system. The man-made controller, at least so far, is not designed to adjust its own references. Rather it is designed as a “servo,� to serve the needs of the user. A biological organism, in contrast, has reference signals of its own, not accessible to any user. It is autonomous, because it does not serve the needs of another, but those of itself only and always.

From the perspective of the engineer, negative feedback control is about injecting an error signal to get the desired output. In traditional cybernetic applications of control theory to the study of behavior, the comparison between error and reference is placed outside of the organism, where the engineer designing the system also performs the comparison function. Thus for decades such control systems have been treated as stimulus-response or input-output devices: error in, behavior out.The tendency to resort to linear causation is so strong that even closed loop controllers have been treated as devices that receive error signals and generate behaviors.

In the end, the appropriate output signals must be computed somehow. The question is how. The negative feedback organization simply eliminates the effects of disturbance by subtracting them from the internal reference. The effect of its own output is monitored with its own sensors and actively controlled. This elegant solution to the calculation problem avoids calculations on the disturbances in advance.

The misidentification of the inputs and outputs of a control system resulted in persistent mistakes in the application of control theory even when the correct mathematical equations were used. What is worse is that it has made it impossible to perform the appropriate experiments to measure the actual properties of the living control systems.

As a result of these conceptual confusions, in traditional models negative feedback is always misunderstood. Placing the comparator outside the organism has the unintended effect of inverting the inside and outside of the system (Figure 5).What should be part of the organism is considered To be a part of the environment, and what should be part of the environment, namely, the feedback function, is considered a part of the organism. Consequently, the equations that describe how forces act on loads and accelerations and decelerations of the loads are assumed to be computed by the nervous system [50]. These conceptual confusions have largely prevented any progress in the study of behavior for many decades.

Best,

Boris

···

From: Bruce Abbott [mailto:bbabbott@frontier.com]
Sent: Tuesday, November 24, 2015 4:06 PM
To: csgnet@lists.illinois.edu
Subject: RE: FW: What’s perception got to do, got to do, with it?

[From Bruce Abbott (2015.11.24.10:05 EST)]

From: Boris Hartman [mailto:boris.hartman@masicom.net]
Sent: Tuesday, November 24, 2015 4:58 AM
To: csgnet@lists.illinois.edu
Subject: RE: FW: What’s perception got to do, got to do, with it?

Hi, Bruce,

HB: sorry not to answer your part about thermostat, as I think Henry Yin did good explanation where could be problems with equating »machines« and »organiams«. So I’ll just continue where you make some conclussions about »control of external environment«.

Boris, you can’t be a proponent of PCT and at the same time hold that we shouldn’t be equating machines and organisms. I’ll try to make clear why this is so.

Norbert Wiener’s seminal book, the one that Bill Powers read that put him on the path to PCT, was entitled Cybernetics: Or Control and Communication in the Animal and the Machine. That title announces Wiener’s central insight, which is that the same principles of control and communication can be applied toward analyizng and understanding both animals and machines.

What Wiener understood was that there can be a »science of machines,« abstracted from any particular physical realization of the machine. A given machine in this more abstract sense can be characterized as a set of functions (transformations) that are linked together (communicate) in a particular way. He understood that the machine will behave in the same general way regardless of the material of which it is constructed. Thus, for example, one can build an amplifier out of electronic components connected together in a particular way, or out of hydraulic components connected together in the same way. Schematically, they are the same system if the components perform the same transformations and are connected in the same way.

During World War II there had been a rapid development of control systems for doing such things as stabilizing guns on a battleship against the rocking the the ship, accurately aiming anti-aircraft guns to track and shoot down aircraft, and automatically flying aircraft (autopilots). Wiener realized that one could understand the behaviors of animals, just as one could understand the behavior of an autopilot, by identifying how the relevant components are organized into a system. In particular, the system Wiener had in mind is the control system, which is why he coined the term »cybernetics« to label this new field. It comes from the Greek, meaning »steersman.«

All control systems whose components carryout the same general functions and are connected in the same way will behave in identical fashion, whether those systems are implemented by the electro-mechanical components of a car’s cruise control or the driver’s iono-mechanical sensory receptors, neurons, and muscles, which when organized to form a speed control system, can similarly keep the car’s speed at the driver’s reference level (by means of adjusting foot pressure on the accelerator pedal).

PCT is in essence the application of these machine-derived principles of control to living creatures. The problem isn’t with equating machines and organisms, because at the systems-analysis level it doesn’t matter how the components and signals are formed so long as they perform analogous functions and are connected so as to form the same system of components.

The real problem with equating machines and organisms is that organisms are extremely complex. Even worse, it is notoriously difficult to observe the workings of their inner parts so that one can construct an accurate functional »wiring« diagram. Often the best we can do (given current technology) is theorize a particular organization of parts, create a simulation based on that proposal, and observe whether the simulation reproduces the organism’s observed behavior. The many PCT simulations of tracking behavior are examples of this strategy, where the control exerted by the simulation is compared to that produced by the individual in the tracking task.

Now let’s examine what you take Bill to be saying about control systems.

BA : Of course, we humans can know only what our sensors tell us. We have no direct access to the outside reality, whatever that may be. One can even adopt the position that such a reality does not exist. Bill Powers recognized this fact, but like most people, he preferred to believe that there is a reality beyond our perceptions, and that by controlling our perceptions, we also control numerous variables belonging to the reality beyond our perceptions.

HB : Bruce, I highly value your oppinion, and I highly value you as a person, but Bill was explicit in his »definitions« (B:CP) and his diagram (LCS III). It’s his theory. We don’t control external environment, we affect it. Whatever Bill beleived, his statements in FUNCTION and COMPARATOR and his DIAGRAM are very clear. Just effects to external environment what is plausible with physiological »facts«. So I’d like to know what is wrong with his statements.

PCT is Bill’s theory, yes, but it’s an application of control theory, originally developed to understand the behavior of machines that control.

Bill P (LCS III):

FEED-BACK FUNCTION : The box represents the set of physical laws, properties, arrangements, linkages, by which the acrtion of this system feeds-back to affect its own input, the controlled variable. That’s what feed-back means : it’s an effect of a system’s output on it’s own input.

In the diagram below, the controlled variable is the small circle labeled as the »input quantity.« Its value is affected by the disturbance and by output quantity, operating on the input quantity via the feedback function. The feedback is negative, so adding the effects of feedback and disturbance accomplishes subtraction. The feedback compensates for the disturbance, so that its effect on the input quantity is minimized. That is, the value of the input quantity is controlled. The input quantity is thus the controlled variable in the environment.

The input function converts this input (sensed value of the controlled variable) to an internal representation of the controlled quantity called the perceptual signal. Because the perceptual signal tracks the value of the controlled variable in the environment, it is also controlled. The perceptual signal is the person’s inner representation of the controlled variable in the environment, so if the environmental variable is controlled, so is its perception. We can only directly know our perceptions, so from the organism’s point of view, what is controlled are its perceptions.

HB: So you can beleive Bill or not. I do.

HB: But nobody is saying that you are not free to propose changes to generic diagram and put »controlled effects« into »feed-back function« or »CV« (controlled variables) into immediate environment of control system, and consequently change all the loop, as Rick is proposing last years.

I hope I’ve made it clear by now that Bill and I are saying the same thing; I am not proposing any changes. If I had that wrong, surely Bill would have pointed that out to me.

BA : Indeed, if one assumes that such a reality exists, it follows that evolution would produce living control systems whose control over their various perceptions thereby controls those aspects of reality that must be controlled if they are to survive.

HB : All organisms »control« for constant internal environment – »homeostasis«, including plants. That’s what BBill theory is about :

Bill P :

Perceptual Control Theory (PCT) provides a general theory of functioning for organisms.

HB : And that’s what control is about :

Bill P :

Boris, how do you define »control«? As I define it, control involves acting on a variable in such as way (via negative feedback) as to cancel the effects of disturbances to that variable. Isn’t that what happens to the cursor in a tracking experiment as disturbances act on it but are opposed by the participant’s actions? Isn’t the cursor out there in the environment (on the computer screen)? Isn’t its movement being opposed by the participant’s arm movements, which move the mouse, which acts on the position of the cursor?

BA : Thus, when I move my perceived arm to a peceived position so as to reach a perceived coffee cup, I believe that I am moving a real arm toward a real coffee cup. By controlling my sensed perceptions of my perceived arm’s position, I am thereby also controlling the actual position of my real arm.

HB : You are free to think so. It’s your perception or what you are experiencing… I suppose that you probably with your control reeasoning perceive any behavior as »control«. But how many people on the world percive behavior in the same way as you do ? And I think that many self-regulation theorist do perceive it in the way you do. Maybe most of the people.

No, I don’t perceive just any behavior as control. Only behavior that acts to oppose changes in some observed variable. That’s the definition of control.

HB: But my oppinon is, that Bill was right when he put just effects to outer environment »even to a cup of coffee« to control perception. If you don’t think so, please show where you think he was wrong. I’m speccially interested in how would you change Bill general diagram about control loop to suit your oppinion, How you imagine that »control« comes to outer environment ?

BA : I don’t think I can explain it any better than that.

HB : You don’t need to Bruce. Thank you for your oppinion.

Well, you’re welcome, but I hope you now understand that the view I expressed was Bill’s opinion, too.

Bruce A.

[_Boris_Hartman]

BA : In the diagram below, the controlled variable is the small circle labeled as the »input quantity.« Its value is affected by the disturbance and by output quantity, operating on the input quantity via the feedback function. The feedback is negative, so adding the effects of feedback and disturbance accomplishes subtraction.

HB : Bruce, where did you see that Bill ever talked about »input quantitty« in this way ??? I can’t think of anything else but of your pure imagination that »subtraction« is done in outer environmnet. Where did Bill ever metnioned »subtraction« in outer environment ???

Subtraction is clearly done in »comparator« inside organism, where real control is going on. Only »nervous system controls«- Not environment.

Bill P (LCS III diagram) : The effects of the output quantity adds to the effects of the disturbances on the input quanttiy.

HB : It’s only »adding«, no »subtracting«.

Please read Henry Yin about »illusions« of »subtractor« in outer environment. You are definitelly changing PCT with putting »subtractor« into outer environment.

In diagram (LCS III) is clearly seen that effects are »adding«. I can’t conclude anything else but what is obviously. It seems to be your illusion that behavior is control, that subtaction is going on in outer environment, that perceptual signal is controlled from outside….Â

into »perceptual controlled variable«. I never saw Bill used any of this terms in such a way.

So I think that in diagram (LCS III) what is happening is :

Bill P : So for “the organism to have control over any sensed condition of the environment, all that is necessary is for it to possess the means to cause that condition (as perceived) to vary in each of the possible ways it can varyâ€?….(B:CP)

&nbspp;

Mary P :

The only known organization that can maintain itself in a variable world is a control system. A control system receives input—perceptions—from its en environment. This input is a combined function of environmental effects plus the perception of its own actions. The input is compared to a reference state, and the difference drives the output, which is immediately and continuously perceived, along with its effect or lack of effect on the environment. The output varies to reduce the difference between input and reference states.

HB : I also included Mary P. oppinion about »control in environment«, because it seems that is clear making a point and make it more understandable that the input is compared (subtracted) and output is contntinuosly perceived whether it affects environment or not. Output always (generally speaking) varies to affect perception. Try it by observing yourself.             Â

It’s clearly seen that effects are added, not subtracted. Even if effects of distubances would be subtracted from »output« it wouldn’t be the general case as Bill’s diagram is tending to present.

There is no »subtractor« in outer environment, it’s inside control system. So we are talking about »added« effects of »disturbances and ouptut« on sensors (input function) and empahasis is on effects, not on »controlled« effects which would turn perceptual signal into »perceptual controlled variable«.

I’ll stick to Bill’s explanation of »added« effects. It’s obvious for me that he is right and I also think that diagram (LCS III) is supported with physiological evidences when »addition« is accomplished. But that’s not generally the case. I think this is crucial for understanding why only »effects of control" are seen in environment, not »controlled effects«.

I also think that it would be helpfull to finish »whole picture of organisms control« where there are still problems :

1.   with input »references on 11.level« and
   

2.   how process of »reorganization« function in the organisms control when in the same time »essential variables« are affected by »genetical control system«.
   

It’s hard task. My oppinion is that without help of Henry Yin there will be no progress, and »the whole picture« of »PCT organism« will not be finished so soon. And thus place for speculations, phylosophy and manipulations with terms will be always present on CSGnet. If phylosophy will prevail I doubt that »picture« on p. 1919 (B:CP) wil make it clear how organisms and speccially nervous system function. This is the final goal of PCT if I understoood Bill well. To explain how organisms function.

Best,

Boris

Best wishes,

Bruce A.

···

Subject: RE: FW: What’s perception got to do, got to do, with it?

[Abbott]

[From Bruce Abbott (2015.12.20.1030 EST)]

···

From: Boris Hartman [mailto:boris.hartman@masicom.net]
Sent: Sunday, December 20, 2015 1:21 AM
To: csgnet@lists.illinois.edu
Subject: RE: FW: What’s perception got to do, got to do, with it?

Bruce,

it took me quite a long time to answer you. I admit I was deeply surprised about your answer.

I’m sorry Bruce, but it seems that we can’t come to some perfect agreement. Your opinion is quite clear. I think that we agree in general view how organism works and how it produces behavior, but there are some “detailsâ€? where we are practically on extreme, contrary positions.

My opinion is that there is no “subtractorâ€? (o-d) in environment in the “input quantity«. At least I never saw anywhere Bill using this combination. But “adderâ€? (o+d) as it’s seen from Bill’s diagram as “add effectsâ€?, is something what is common in Bill’s theory, so effects of output and disturbances are “addedâ€? in environment. It’s clear in his diagram.

Boris, I agree with you: There is no »subtractor« in the environment. But I think you are getting confused between adding variables and adding the values of variables.

The two plus signs entering qo in Bill’s diagram represent the addition of two variables. In a negative feedback system, the values of those two variables will have opposite signs, in the steady state. The variables are added, but because the values have opposite signs, adding them results in a value for qi equal to the difference between d and qo!

As proof, examine the figure below, which is taken from a screen shot of LiveBlock:

I moved the Disturbance slider control to the extreme right to produce a steady disturbance value of 15.00. The LiveBlock control system responded by adjusting the Output Quantity to minus 14.85. The result of adding these two quantities to qi is shown in the box labeled »Input Quantity«: 0.15.

qi = 15.00 + (-14.85) = 15.00 – 14.85 = +0.15

Bruce

[Abbott]

[From Bruce Abbott (2015.12.20.12:00 EST)]

···

From: Boris Hartman [mailto:boris.hartman@masicom.net]
Sent: Sunday, December 20, 2015 1:25 AM
To: csgnet@lists.illinois.edu
Subject: RE: FW: What’s perception got to do, got to do, with it?

Bruce,

Your theoretical explanation of Wieners »control theory« is bellow so I’ll not »copy-paste« it here. I’ll just try to look for differences in Bill’s and Wieners theories. If it is »the same«, why did Bill wrote his theory ?

BA : Boris, you can’t be a proponent of PCT and at the same time hold that we shouldn’t be equating machines and organisms. I’ll try to make clear why this is so.

BA : Now let’s examine what you take Bill to be saying about control systems.

HB : I think that Bill criticized Wiener and “engineering control theoryâ€?. So I think there are more problems in just “equatingâ€? theories …

Bill P : For now the relevant aspect of Wiener’s diagram is the fact that it is drawn so that one may consider the whole unit in terms of stimulus and respons, just as behaviorists have always done.

Bill P : Unspoken in Wiener’s diagram, by the way, are the reorganizing system’s intrinsic reference levels….

Where did you find these quoted sentences? Winer offered several diagrams; I have no idea which one Bill refers to here.Â

Bill P :

Regarding the underpinnings of PCT: There was no one in cybernetics/systems theory after Ashby’s book in 1953 (Design for a Brain) from whom I learned anything about control theory and its role in behavior. Wiener’s book of 1948, which I read in 1952 thanks to Kirk Sattley, got me started: the concept of feedback control, and the particular relations to behavior that he laid out, clicked in my mind as the obvious successor to all the psychological models I had ever heard of, including the one in which I then believed. Ashby’s book gave me an organized view of how one would start applying these principles on a grander scale—it was as much his organization as his ideas that turned me on./span>

Bill P : Wiener and Ashby inspired me to go back to the sources of the ideas that they had adopted. When I did, I gradually came to realize that neither of them had learned very much about control systems.

Bill did not invent control theory; he took an existing body of knowledge from control system engineering and from it borrowed the negative feedback control system as the basic control model for PCT. As Bill noted above, his initial inspiration for doing this was provided by reading Winer’s book. Ashby provided the idea for the reorganizing system, including its »intrinsic variables« (which Ashby called »essential« variables).

HB : I’m sure you can find more Bill’s critics on Wieners model. what by my opinion tell us that thinking about “Control theory� wasn’t simply transferred to PCT.

HB: But I think that most comprehesible critics of engineering “control theory� was given by Henry Ying. His article is worth of reading as I think he is the only one whom I saw upgrading PCT till now. I think that he could be very helpful in conflicting understanding of PCT.

Henry Yin :

4.2. Control of Input. A control system always controls its input, not output [7]. Only perceivable consequences of behavior can be controlled.

According to mainstream engineering control theory, a control system controls its outputs, not its input. This is perhaps the most common fallacy today, both in engineering and in the life sciences [49, 55, 56]. This fallacy, an unfortunate legacy of cybernetics, is the result of imposing the perspective of the observer rather than using the perspective of the organism or controller. The mistake is to assume that what the engineer perceives and records, the “objective� effect of the system, is the output of the system.

The problem here is that »mainstream engineering control theory« uses the terms »input« and »output« differently than we do in PCT. When the control-system engineer speaks of »output«, he or she is referring to the variable that the system is controlling, whereas we use »output« to refer to the actions by means of which the control system acts to exert that control. When the control-system engineer speaks of »input« he or she is referring to the reference. Consequently, when Henry says that, »according to mainstream engineering control theory, a control system controls its outputs,not its input,« he is correct – but given the engiineer’s definitions of those terms, the statement is not a fallacy, it’s a correct description!

As Henry notes, the engineer’s problem is to design a system to be used by human beings. If you want a room to be a certain temperature, you change the reference setting on the thermostat (you input that setting) and if the system works properly, the result is that the room temperature is brought to the set temperature (the system produces or »outputs« the user’s desired room temperature). So using »input« to refer to the setting of the reference and »output« to refer to the system’s sensed value of the variable under control makes good sense.

The problem is not that mainstream engineering control theory has it wrong, the problem is that some folks who have endeavored to apply control theory to human and animal behavior have incorrectly understood the engineer’s term »output« to mean »behavior« or »actions« and »input« to mean »stimulus«.

The common assumption that output is controlled ignores the perspective of the organism that is doing the controlling. By imposing his own desire and perspective, the engineer ignores the autonomy of the negative feedback controller, for he is always trying to make the machine do what he wants. He can only accomplish this by adjusting the reference signal, as the user operates a thermostat by adjusting the temperature setting. Since this is the signal generated by the user, it is usually labeled as the input to the system. In a biological organism, however, the reference signal is always internal to the organism.

Again, it’s not the control engineer who is doing this. Those who attempt to apply control theory to living organisms while misunderstanding the engineer’s meaning of those terms are doing this.

Please don’t confuse »traditional engineering control theory« with »modern engineering control theory.«  Bill relied on traditional engineering control theory and its analytic techniques. What he objected to was modern control theory, with its Kalman filters, feed-forward mechanisms, and a set of mathematical techniques that yield »optimal control« while making it just about impossible to understand how the resulting system accomplishes what it does.

Bruce A.

[rsmarken]

[From Rick Marken (2015.12.20.1830)]

TCVModelwColor.xlsx (37.3 KB)

···

Martin Taylor (2015.12.19.12.39)–

MT: I look for functions in the environment that result in variables

that appear to be controlled, and hypothesize that for each there is
a perception in the organism generated by a similar function, and
that this perception is the only variable that is controlled.

RM: This is very close to a description of what I do too, which leads me to believe that our apparent difference may be only semantic. In the hopes of reaching agreement I have created a spreadsheet that does what I believe both of us are saying. The spreadsheet is attached.

RM: The spreadsheet is a model of a control system controlling in an environment made up of only two physical variables, x.1 and x.2. The control system controls a linear combination of these variables. The coefficients of this linear combination, shown in the upper left corner of the spreadsheet, defines the aspect of the environment, q.i, that the controller controls. I call this the “Controller’s q.i”. The spreadsheet arrives with q.i = -.2x.1+.9x.2. [Note that there is nothing in the environment that corresponds to q.i. The only things that exist in the environment of this control system are the variables x.1 and x.2. There are an infinite number of aspects of this environment that could be controlled, corresponding to the infinite number of possible values of the coefficients of the linear equation that defines q.i]

RM: The variables x.1 and x.2 can be thought of as the intensity of two different wavelengths of light (ignore the fact that they can go negative for now; I’m going to try to develop a version of this sheet that only allows positive physical variable values but I haven’t figured out how to do it yet). So q.i can be thought of as a sensation, like color, that results from a combination of input wavelength intensities.

RM: The variable x.1 is a disturbance variable because it varies sinusoidally independently of the controller; x.2 is an output variable because it is varied by the controller. The situation is equivalent to having two sources of light shown on a screen, the intensity of one source varying independently of the controller and the intensity of the other being varied by the controller. The controller’s job is to maintain a perception of the combination of lights, q.i, in some reference state. Of course, q.i exists only as a perception in the controller; the perception controlled by the controller is assumed to be simply proportional to q.i: p = k*q.i. The spreadsheet allows you to adjust k (the “input gain”) as well as other parameters of the control system. These parameters can be set in the cells labeled “Control System Parameters”. Leave them as they are for now because they are currently set in a way that allows the control system to control.

RM: The main reason for building the spreadsheet was to illustrate what I believe is going on when an experimenter (Observer) does the TCV. What the Observer is trying to do is come up with a description of the aspect of the environment (the controlled quantity, q.i) that is equivalent to that which the Controller is actually controlling. You do this in the spreadsheet by trying to come up with coefficients for x.1 and x.2 that are equivalent to those that define q.i for the Controller. Those coefficients can be set in the gray cells labeled “Observer’s q.i (q.i’)”. In practice, setting these coefficients would be equivalent to the observer looking at the combination of light intensities, x.1 and x.2, that are passed through two filters differing in terms of amount of attenuation.

RM: The correctness of the Observer’s guess, q.i’, at the definition of the aspect of the environment the Controller is controlling, q.i, is evaluated using a measure of how well the variable is being controlled. The measure of control used here is Stability, defined as 1 - sqrt(Vobs/Vexp), where Vobs is the observed variance of the controlled quantity and Vexp is the variance in q.i that is expected if q.i were not controlled.Perfect control would yield a stability measure of 1.0 and no control would yield a stability measure of 0. The stability measure in the cell labeled “q.i Stability” is the stability of the actual controlled quantity; at .97, it’s nearly perfect. The stability measure in the cell labeled “q.i’ Stability” is the stability of the hypothesized controlled quantity, q.i’, the one defined by the coefficients for the Observer’s definition of the aspect of the environment that the Controller is controlling, q.i; at .41 it is clear that this variable is not controlled nearly as well as the variable the Controller is actually controlling.

RM: So that’s it; the spreadsheet shows how the TCV can be used to reveal the controlled quantity, q.i, that a Controller is actually controlling. Since q.i is the definition of the perception, p, the Controller is controlling (p = kq.i = k (-.2x.1+.9*x.2) in this case) once you find the coefficients for q.i’ that are equivalent to those for q.i (the coefficients that result in a stability value for q.i’ that is close to 1.0) you have discovered the perception that the Controller is controlling.

RM: The other things on the spreadsheet are designed to give you a better feel for what this means. For example, the column labeled “Controller’s q.i” is all a medium shade of blue; this is the reference state of the color perception, p, that results when p = .1* (-.2x.1+.9x.2) and the reference signal is set to .3. The column labeled "Observer’s q.i’ " is the color perception that would result if the Controller were controlling p = .1* (-.5x.1+.5x.2). The main thing to note there is that the color perception varies considerably as you scroll down that column while it stays the same as you scroll down the column for the actual value of q.i.

RM: The graph shows temporal variations in q.i, q.i’ and p; q.i and p remain completely stable (the difference in levels of q.i and p on the graphs is just a result of the perceptual scaling factor, k; q.i’ varies sinusoidally because the disturbance, x.1, varies sinusoidally and the disturbance to that definition of the controlled quantity, q.i’, is not completely resisted.

MT: The

appearance of a controlled variable in the environment is always
what you have called a “spandrel”, as the taste of lemonade example
suggests.

RM: Hopefully, this demonstration will show that you and I agree that there is, indeed, no controlled variable (or controlled quantity) in the environment. As I said above, the only things the environment in this demonstration are the physical variables, x.1 and x.2. The controlled variable (or, more appropriately, the controlled quantity, q.i) is an aspect (or function) of the variables in the environment that is computed by the perceptual input functions in the Controller; it is computed either by the perceptual functions of the Observer – if it can be – or by instruments, like the filters I assume here, that produces the observer’s “view” of q.i, which I call q.i’.

RM: This is a lot of stuff so feel free to ask questions if you have any. But I would really appreciate it if you (and anyone else interested in this) could let me know what you think of the demonstration; whether it makes sense to you and what you might suggest to improve it. What I would like to do is develop a demonstration of how to do research based on PCT – research aimed at determining the kinds of perceptions people (or other organisms) control. So any comments/ criticisms would be very helpful.

Best regards

Rick

–
Richard S. Marken

Author, with Timothy A. Carey, of Controlling People: The Paradoxical Nature of Being Human.

[bnhpct]

[Bruce Nevin (2015.12.21.08:40 ET)]:

Hi, Martin.

Martin Taylor 2015.12.19.12.39–

MMT: Rick usually asserts that the controlled quantity, qi, is objectively in the environment and can be directly influenced by an external observer. […]

MMT: My position is quite the opposite. I say that we ONLY can control our perceptions. A perception is produced by some function whose input arguments may be composed entirely from sensory variables that are influenced by a supposed-to-exist environment […].

MMT: When we talk, we omit the “supposed-to-exist” caveat … and simply assert that there exists a real world through which our PCT model suggests our perceptions (real to us) are controlled.

BN: Perhaps you can allow Rick (and me) the grace occasionally to omit the “supposed-to-exist” caveat as well?

BN: There are two questions in play. The first question is the ever-recurrent epistemological one: How real is our control? That devolves to: How real are our perceptions? I think successful control of a perception is our only warrent for attributing reality to it, that collective control strengthens that warrent by intersubjective agreement, that successful collective control implicitly includes and is based upon corroborating (testing for) controlled variables, and that we have the strongest warrent for attributing reality to q.i as determined by the TCV (best when replicated). Doing science is our best intellectual way of determining what is real.

BN: The variable q.i is an excellent instance of a perceptual variable to which we are strongly warrent to attribute reality because we determine it by doing science.

BN: The second question is: Are q.i and the CEV equivalent. q.i is a single variable perceived and measured by the observer; the CEV is by definition a complex of more than one variable.

BN: Rick (2015.12.20.1830) proposes that you and he are in violent agreement, with a spreadsheet as means of verification. Below, I offer verbal evidence in support of this. You may not have seen this, which was posted during the time in which you weren’t reading CSG-net.

Rick Marken (2015.12.12.1245)

RM: I would say that the PCT stance on perception (or, epistemology) is that the perceptions that we control are constructions based on sensory data that is caused by an external (environmental) reality that is approximated by the models of physics and chemistry, and these constructions are not arbitrary but have functional significance in the sense that they allow us to control the aspects of physical reality that we have to be able to control in order to be able to survive. In other words, I would say that the PCT stance is constructivist – realist – functionalist.

BN: So Rick is a realist in regard to the models of physics and chemistry. These models of course are perceptions. Realism here is a belief that doing science in the fields of physics and chemistry constructs perceptions that accord so well with reality (whatever that may be) that we are warrented to omit the “supposed-to-exist” caveat. That warrent is functional and provisional.

BN: On this basis, my point was about the character of the “supposed-to-exist” CEV. My understanding is that intensity receptors perceive variables identified by physics and chemistry (in the example, sugar and acid) and generate perceptions. Input functions at the sensation level combine these intensity perceptions. In the example, they construct the unitary perception of the taste of lemon. The term CEV suggests that, to the same extent that we are warrented in supposing that the sugars and acids of lemon juice are really in the environment (warrented as variables identified by physics and chemistry), we are likewise warrented in supposing that lemon flavor is really in the environment.

BN: I may have a slightly different approach to your disagreement with each other.

Rick Marken (2015.12.06.1625)

RM: What you call the CEV is q.i, the controlled quantity. I prefer to stick with Powers’ terminology, not only because it’s Powers’ terminology but because CEV is kind of misleading; it implies that there is actually some “complex” variable in the environment that is controlled. The controlled quantity is actually an aspect or function of environmental variables, like the function of acid and sugar that is the perceived as the taste of lemonade, that doesn’t necessarily exist as a variable in the environment.

BN: I understand, yes, that you reply “No, no, the CEV is not controlled, only the perception of it is controlled.” This quarrel hinges on answers to question 1 above.

BN: My quibble here is with this definition of q.i as a (singular) function of (plural) environmental variables. The making of them into a complex unity happens in the perceptual hierarchy. Fructose is perceived as sweet; citric acid is perceived as sour. Fructose and citric acid are commingled in a lemon. Intensity perceptions of sweet and sour are combined to construct the sensation of the taste of lemon juice. But at the interfaces between the taste sensors in the organism’s mouth and the physical environment outside those sensors there is a q.i for sugar and a q.i for acid, aggregated for populations of sensors. Each q.i is transformed by sensors into an intensity perception. At any higher level of the hierarchy we are no longer speaking of q.i, we are speaking of signals p that are constructed at the level below. The above formulation suggests that there is a single, unitary q.i = “a ratio of x sucrose to y citric acid” corresponding to the taste of lemon. From your term CEV a reader is liable to infer that there is, and likewise the same inference might be made from Rick’s definition of q.i as a (singular) function of (plural) environmental variables.

BN: The variable q.i is an instrumental measurement that can be observed by more than one person and replicated by anyone. Such instrumental measurements have a certain claim to be objective, the only claim of which I am aware, namely, intersubjective agreement.

BN: Of course the measurement, the process of measuring, the instruments of measurement themselves, cannot be demonstrated to be anything more than controlled perceptions. When we crank the epistemological implications of PCT up to the paradoxical sticking-place, intersubjective agreement appears to me to be our only route out of solipsism. Equally of course, as each of us in turn have acknowledged at one time or another, we assume that our perceptions are veridical, at least in some functional sense. Seriously not to do so would surely be paralyzing. I doubt that anyone has carried that line of epistemological doubt any farther than mind games.

/Bruce Nevin

[JohnK]

[John Kirkland 2015.12.22]

Reading about Rick’s description from fiddling about with colour mixing reminded me of a series of publications generated from an approach we designed (and others, internationally, continue to deploy) for probing into colour blending perceptions.

By way of background we started this line of enquiry many years ago when playing around with Benham discs. See http://www.michaelbach.de/ot/col-Benham/index.html My colleague (David Bimler, who’s become something of an expert in colour perception) wrote a little programme for creating these discs, that were then printed on white paper and stuck onto 45rpm vinyl records. We used a $5 stackable turntable so when the change lever was flicked a new one dropped onto the spinning turntable for viewing. Briefly, the black/white flicker of a Benham disc tends to be reported as what’s known as “subjective colour” usually a pastel green, blue, red, brown, etc… OK, that’s by way of background to our question: What would colour deficient (aka colour-blind) people report?

First off we needed to design a simple procedure for assaying anybody’s colour vision, to ensure there was no hint of CVD. We based this work on the D15, a standard colour-perception tool. We’d been doing a lot of work on multi-dimensional scaling and knew how to go about this task. The upshot from many dead-ends was a set of 15 representational swatches, each about the size of your little finger nail. The Munsell color-perception distorted sphere shape notorises three dimensions: value (light - darkness), hue (the ROYGBIV colours), and chroma (saturation). See: http://munsell.com/about-munsell-color/how-color-notation-works/ We selected evenly spaced hue/saturation and two values of value. Our initial studies were to do with validation.

Secondly, we assembled trilemmas (sets of three swatches) and arranged these as corners of a small implicit triangle, printed onto credit-card sized card.

Thirdly, the task itself. A deck consists of 75 cards (we have two parallel decks). Working under controlled light conditions participants shuffled the deck and deal from the top. Their job was to report which triad is different from the remaining ones being careful not to focus on value but stick to differences in hue (with examples). We’ve found these results are extremely sensitive to for identifying the sorts of perceptual errors people make, even those of us with “normal” trichromat vision. Of the many studies we’ve reported using this technique: identical twins make the same sorts of errors, smokers’ eyes are screwed, and across three generations (grandfather, mother, son) if the CVD is particularly “strong” we are able to distinguish female “carriers”.

Fourthly, and finally, what’s this to do with PCT? After all these years, and triggered by Rick’s recent report of his method, I wonder if we’d been unknowingly carrying out PCT research all along.

Years ago, courtesy of our explorations into Personal Construct Theory (the other PCT) we realised three is always a crowd. There’s so much lost with paired comparisons, when only two items are presented. With sets of three there is always a reference, a point of difference. Hence trilemmas (rather than dilemmas).

Each of our trilemmas is a disturbance. The resolution is to “blend” two and report about the one most different from the other two.

OK Rick (anybody), here’s where I need some help. What we’ve got here is “visual lemonade”. But because the referents (swatches) are standardised it is possible to make some inferences about the underlying built-in mechanisms. That’s not the focus of my current PCT-related question. Instead, it is, “What is a PCT-related explanation of what we’ve been doing with this approach?” Diagrams welcomed too.

As an aside #1. We created lozengers too, and asked people to report what they saw in screen-presented patterns (like many 3-faced cubes). When colours blend it is reported as if a line appears, immediately highlighting the odd-one-out “different” trilemma member.

Aside #2. David has been carrying out some reanalyses of others’ data and that team reported: a. pre-clinical indicators of Type II diabetes, and b. pre-clinical signs of heavy-metal poisoning (mercury in Brazil gold-fields).

Not bad from messing around with Benham discs and asking a naive question.

BTW, we never got around to conducting research aimed at answering the original question, what CVD people report when watching spinning black-and-white patterns.

Season’s greetings to one and all, from summer.

JohnK

.

···

On Tue, Dec 22, 2015 at 4:43 AM, Bruce Nevin bnhpct@gmail.com wrote:

[Bruce Nevin (2015.12.21.08:40 ET)]:

Hi, Martin.

Martin Taylor 2015.12.19.12.39–

MMT: Rick usually asserts that the controlled quantity, qi, is objectively in the environment and can be directly influenced by an external observer. […]

MMT: My position is quite the opposite. I say that we ONLY can control our perceptions. A perception is produced by some function whose input arguments may be composed entirely from sensory variables that are influenced by a supposed-to-exist environment […].

MMT: When we talk, we omit the “supposed-to-exist” caveat … and simply assert that there exists a real world through which our PCT model suggests our perceptions (real to us) are controlled.

BN: Perhaps you can allow Rick (and me) the grace occasionally to omit the “supposed-to-exist” caveat as well?

BN: There are two questions in play. The first question is the ever-recurrent epistemological one: How real is our control? That devolves to: How real are our perceptions? I think successful control of a perception is our only warrent for attributing reality to it, that collective control strengthens that warrent by intersubjective agreement, that successful collective control implicitly includes and is based upon corroborating (testing for) controlled variables, and that we have the strongest warrent for attributing reality to q.i as determined by the TCV (best when replicated). Doing science is our best intellectual way of determining what is real.

BN: The variable q.i is an excellent instance of a perceptual variable to which we are strongly warrent to attribute reality because we determine it by doing science.

BN: The second question is: Are q.i and the CEV equivalent. q.i is a single variable perceived and measured by the observer; the CEV is by definition a complex of more than one variable.

BN: Rick (2015.12.20.1830) proposes that you and he are in violent agreement, with a spreadsheet as means of verification. Below, I offer verbal evidence in support of this. You may not have seen this, which was posted during the time in which you weren’t reading CSG-net.

Rick Marken (2015.12.12.1245)

RM: I would say that the PCT stance on perception (or, epistemology) is that the perceptions that we control are constructions based on sensory data that is caused by an external (environmental) reality that is approximated by the models of physics and chemistry, and these constructions are not arbitrary but have functional significance in the sense that they allow us to control the aspects of physical reality that we have to be able to control in order to be able to survive. In other words, I would say that the PCT stance is constructivist – realist – functionalist.

BN: So Rick is a realist in regard to the models of physics and chemistry. These models of course are perceptions. Realism here is a belief that doing science in the fields of physics and chemistry constructs perceptions that accord so well with reality (whatever that may be) that we are warrented to omit the “supposed-to-exist” caveat. That warrent is functional and provisional.

BN: On this basis, my point was about the character of the “supposed-to-exist” CEV. My understanding is that intensity receptors perceive variables identified by physics and chemistry (in the example, sugar and acid) and generate perceptions. Input functions at the sensation level combine these intensity perceptions. In the example, they construct the unitary perception of the taste of lemon. The term CEV suggests that, to the same extent that we are warrented in supposing that the sugars and acids of lemon juice are really in the environment (warrented as variables identified by physics and chemistry), we are likewise warrented in supposing that lemon flavor is really in the environment.

BN: I may have a slightly different approach to your disagreement with each other.

Rick Marken (2015.12.06.1625)

RM: What you call the CEV is q.i, the controlled quantity. I prefer to stick with Powers’ terminology, not only because it’s Powers’ terminology but because CEV is kind of misleading; it implies that there is actually some “complex” variable in the environment that is controlled. The controlled quantity is actually an aspect or function of environmental variables, like the function of acid and sugar that is the perceived as the taste of lemonade, that doesn’t necessarily exist as a variable in the environment.

BN: I understand, yes, that you reply “No, no, the CEV is not controlled, only the perception of it is controlled.” This quarrel hinges on answers to question 1 above.

BN: My quibble here is with this definition of q.i as a (singular) function of (plural) environmental variables. The making of them into a complex unity happens in the perceptual hierarchy. Fructose is perceived as sweet; citric acid is perceived as sour. Fructose and citric acid are commingled in a lemon. Intensity perceptions of sweet and sour are combined to construct the sensation of the taste of lemon juice. But at the interfaces between the taste sensors in the organism’s mouth and the physical environment outside those sensors there is a q.i for sugar and a q.i for acid, aggregated for populations of sensors. Each q.i is transformed by sensors into an intensity perception. At any higher level of the hierarchy we are no longer speaking of q.i, we are speaking of signals p that are constructed at the level below. The above formulation suggests that there is a single, unitary q.i = “a ratio of x sucrose to y citric acid” corresponding to the taste of lemon. From your term CEV a reader is liable to infer that there is, and likewise the same inference might be made from Rick’s definition of q.i as a (singular) function of (plural) environmental variables.

BN: The variable q.i is an instrumental measurement that can be observed by more than one person and replicated by anyone. Such instrumental measurements have a certain claim to be objective, the only claim of which I am aware, namely, intersubjective agreement.

BN: Of course the measurement, the process of measuring, the instruments of measurement themselves, cannot be demonstrated to be anything more than controlled perceptions. When we crank the epistemological implications of PCT up to the paradoxical sticking-place, intersubjective agreement appears to me to be our only route out of solipsism. Equally of course, as each of us in turn have acknowledged at one time or another, we assume that our perceptions are veridical, at least in some functional sense. Seriously not to do so would surely be paralyzing. I doubt that anyone has carried that line of epistemological doubt any farther than mind games.

/Bruce Nevin

[bnhpct]

[Bruce Nevin (2015.12.22.18:30 ET)]

(Martin Taylor 2015.12.21.12.39) –

Mirage and black ice: the higher in the hierarchy, the farther from q.i, the more ambiguity, and the greater the need for additional perceptual input. I don’t see that these bear on the question of justifying a distinction between q.i and CEV.

As we both have pointed out, a strict understanding of q.i would apply only to a single sensor. Or maybe to a set of sensors of the same type (e.g. retinal cells sensitive to green). Not awfully useful for PCT experiments.

A more useful alternative would be a complex input quantity. We might expect a complex q.i to be a measurement of what the observer perceives as a complex environment variable. But it can be a simple measurement of what the observer perceives to be a simple environment variable, because the sensors do their own parsing. For example, the wavelength and intensity of a light source can be measured, so I suppose that yields two quanta, q.i and q.i. Considering only q.i, I see that the three types of cone cells in the retina are most sensitive at wavelengths near 564–580 nm, 534–545 nm, and 420–440 nm, rem, respectively. (Ignoring those people with four types of rod cells, giving them tetrachromatic vision.) Unless it is right at one of these wavelength peaks, a color measured as q.i is transformed to a neural signal most strongly by one type of cell, less strongly by one or perhaps even both of the others. So that measured value q.i maps immediately to level 2 of the hierarchy.

So q.i is not a simple variable, but these problems don’t justify a distinction between q.i and CEV. What is relevant to a quantitative model is q.i. What is the relevance of a CEV?

···

On Mon, Dec 21, 2015 at 2:02 PM, Martin Taylor mmt-csg@mmtaylor.net wrote:

[Martin Taylor 2015.12.21.12.39]

  In response to [From Rick Marken

(2015.12.20.1830)] and

[Bruce Nevin (2015.12.21.08:40 ET)]:

        BN: There are two questions in play. The first question

is the ever-recurrent epistemological one: How real is our
control? That devolves to: How real are our perceptions? I
think successful control of a perception is our only warrent
for attributing reality to it, that collective control
strengthens that warrent by intersubjective agreement, that
successful collective control implicitly includes and is
based upon corroborating (testing for) controlled variables,
and that we have the strongest warrent for attributing
reality to q.i as determined by the TCV (best when
replicated). Doing science is our best intellectual way of
determining what is real.

This is essentially the answer I was intending to write to Rick.

Anyone can imagine and simulate arbitrary perceptual functions, as
in Rick’s spreadsheet, but unless the results of applying those
functions is effective in controlling some perception,
reorganization will soon modify or eliminate them. The perceptual
functions that persist are the ones that prove useful. It’s not so
much that the perception produced by the function is controllable,
though that is often true, it is that controlling that function is
useful, eventually in sustaining the intrinsic variables in states
that allow life to continue, but more immediately in forming parts
of other controllable perceptions. [Parenthetically, this was J. G.
Taylor’s main insight, as expressed in “The Behavioral Basis of
Perception”, Yale UP, 1962].

Consider the mirage, as depicted in many a cartoon of a person

crawling over a desert. The person perceives a lake. We, the
omniscient outside observers, know there is no lake. But why does
the person have a perceptual function that would allow him to see –
not a shimmering patch AS a lake – but a lake?

The reason is, I think, that lakes may have been useful perceptions

in control of other perceptions such as slaking thirst, rowing a
boat, appreciating beauty, and so forth. A perceived lake is a
perceived lake, not a perceived combination of coherently transient
patches of light and shade over a vertically limited extent, etc.
etc. It’s just a lake. Is the lake a real lake? Who knows? But it’s
a good perceptual function to have for the purpose of controlling
one’s perceptions if every time one has that catenation of
properties, one can slake thirst, row a boat, and appreciate the
beauty.

Someone flying over a "real" lake would see that lake just as would

the person on the ground. The catenation of properties at the
sensory level would not be the same, but the higher level properties
of “boundedness” (is that in the environment?), wateriness, etc.
would be the same. But not so for the lake that is a mirage. It is
still a perceived lake, but that lake doesn’t stay put when the
crawling man moves over the desert. He never gets to drink from it.
We omniscient observers say it isn’t a real lake. That’s not a
question of interpersonal agreement, but a statement that the
perceptions that seem real come from functions that usually produce
single valued (or vector) results that are useful in control or that
are usefully controlled, even if sometimes those functions produce
perceptions that don’t allow for such uses or that cannot themselves
be controlled.

        BN: The variable q.i is an excellent instance of a

perceptual variable to which we are strongly warrent to
attribute reality because we determine it by doing science.

        BN: The second question is: Are  q.i and the CEV

equivalent. q.i is a single variable perceived and measured
by the observer; the CEV is by definition a complex of more
than one variable.

The CEV and qi have the same status in this respect. Both, at the

sensory level, are complexes of more than one variable, unless qi is
specifically the physical input to a sensory unit such as a visual
cone or an inner hair cell. Both are single valued results of
applying the same function, but in computing qi, the inputs come
only from the environment, whereas the CEV has no such limitation.
It might also incorporate inputs from memory and imagination, or be
produced entirely from imagination, as are probably most perceptions
in dreams. Even then, the CEV is perceived AS being in the
environment.

        BN: Rick (2015.12.20.1830) proposes that you and he are

in violent agreement, with a spreadsheet as means of
verification. Below, I offer verbal evidence in support of
this. You may not have seen this, which was posted during
the time in which you weren’t reading CSG-net. Â

Rick Marken (2015.12.12.1245)

          RM: I would say that the PCT stance on perception (or,

epistemology) is that the perceptions that we control are
constructions based on sensory data that is caused by an
external (environmental) reality that is approximated by
the models of physics and chemistry, and these
constructions are not arbitrary but have functional
significance in the sense that they allow us to control
the aspects of physical reality that we have to be able to
control in order to be able to survive. In other words, I
would say that the PCT stance is constructivist – realist
– functionalist.Â

        BN: So Rick is a realist in regard to the models of

physics and chemistry. These models of course are
perceptions. Realism here is a belief that doing science in
the fields of physics and chemistry constructs perceptions
that accord so well with reality (whatever that may be) that
we are warrented to omit the “supposed-to-exist” caveat.
That warrent is functional and provisional.

        BN: On this basis, my point was about the character of

the “supposed-to-exist” CEV. My understanding is that
intensity receptors perceive variables identified by physics
and chemistry (in the example, sugar and acid) and generate
perceptions. Input functions at the sensation level combine
these intensity perceptions. In the example, they construct
the unitary perception of the taste of lemon.

I agree that the combination is likely to produce a taste. I cannot

say that this taste would be “lemon” for someone who has never seen
a lemon or been told that this taste is “lemon”. However, for
someone who does associate this taste with lemon, the perception
of “lemon taste” would be as real as any other, and would be
perceived as a property of a substance, even if there happened to be
no substance, but only electrical stimulation of some neurons.
Nobody else would be able to perceive that particular CEV, which, if
there was a real substance, would be controllable by, say, adding
sugar or tomato juice to the mix.

        The term CEV suggests that, to the same extent that we

are warrented in supposing that the sugars and acids of
lemon juice are really in the environment (warrented as
variables identified by physics and chemistry), we are
likewise warrented in supposing that lemon flavor is really
in the environment.

That is indeed as it is perceived to be. Is it a mirage? I think

not, because it can be used in controlling other perceptions, and
moreover, to someone else who has a perceptual function that
produces a perception of how “lemony” a taset might be, my addition
of sugar or tomato juice will influence that person’s “lemoniness”
perception. That kind of cross-personal influence is an argument for
considering the taste to be in an environment accessible to both
people. The taste of that drink could be a Collective Complex
Environmental Variable (CCEV), controllable by a virtual controller
composed of all the people who control their own perceptions of the
taste by contributing ingredients to the mix. They won’t perceive
the same taste, but they would all be controlling it as though there
was a virtual controller with a controlled taste perception.

        BN: I may have a slightly different approach to your

disagreement with each other.

Rick Marken (2015.12.06.1625)

          RM: What you call the CEV is q.i, the controlled

quantity. I prefer to stick with Powers’ terminology, not
only because it’s Powers’ terminology but because CEV is
kind of misleading; it implies that there is actually some
“complex” variable in the environment that is controlled.
The controlled quantity is actually an aspect or function
of environmental variables, like the function of acid and
sugar that is the perceived as the taste of lemonade, that
doesn’t necessarily exist as a variable in the
environment.Â

        BN: I understand, yes, that you reply "No, no, the CEV is

not controlled, only the perception of it is controlled."
This quarrel hinges on answers to question 1 above.

That's not the essence of my reply. My essential point is that a

controlled variable is one that is compared with a reference value.
The difference between those two values determines the influence
that alters the controlled variable’s value so that it approaches
the reference value. In the environment where qi is supposed to
exist, no such reference value is supposed to exist. Therefore qi
cannot be a controlled variable.

 Â

        BN: My quibble here is with this definition of q.i as a

(singular) function of (plural) environmental variables. The
making of them into a complex unity happens in the
perceptual hierarchy. Fructose is perceived as sweet; citric
acid is perceived as sour. Fructose and citric acid are
commingled in a lemon. Intensity perceptions of sweet and
sour are combined to construct the sensation of the taste of
lemon juice. But at the interfaces between the taste sensors
in the organism’s mouth and the physical environment outside
those sensors there is a q.i for sugar and a q.i for acid,
aggregated for populations of sensors. Each q.i is
transformed by sensors into an intensity perception. At any
higher level of the hierarchy we are no longer speaking of
q.i, we are speaking of signals p that are constructed at
the level below. The above formulation suggests that there
is a single, unitary q.i = “a ratio of x sucrose to y citric
acid” corresponding to the taste of lemon. From your term
CEV a reader is liable to infer that there is, and likewise
the same inference might be made from Rick’s definition of
q.i as a (singular) function of (plural) environmental
variables.

I think this quibble is mistaken, both as regards qi and as regards

the CEV. Imagine any function of N variables, but we take N=2 for
easy visualisation, and call those variables x and y. For every
possible x-y pair, the function produces a value z, which can be
visualized as a height above the x-y plane. The function defines a
surface of hills and valleys over the plane. The volume in which
this surface exists is 3-dimensional, but the surface is only
two-dimensional. Some other function would produce a different
surface of hills and valleys in the same 3-dimensional volume. What
the functions do is reduce by one the dimensionality of possible
variation. Every point on the 2-D surface produces the same output
from the function. When N is greater than 2, the function that
produces a single value defines an analogous set of “hills and
valleys” that form a hypersurface of N-1 dimensions.

There's no "complex unity". There is only a complex set of

possibilities that all produce the same result, the same pattern of
hills and valleys. The job of the Test for the Controlled Variable
is to try to find, from the infinite set of possible functions (an
infinity greater than the infinity of real or complex numbers), some
function that produces pretty much the same surface as the function
that produces the controlled variable – the perception.

        BN: The variable q.i is an instrumental measurement that

can be observed by more than one person and replicated by
anyone. Such instrumental measurements have a certain claim
to be objective, the only claim of which I am aware, namely,
intersubjective agreement.Â

Here I contradict you, to argue that intersubjective agreement is a

perception that has no different status in “reality” from any other
perception. To me, the test of “objectivity”, if that word is not
intrinsically paradoxical, is the stability of the perceptual
function as used in control, either of itself or as a component of
other perceptions. Intersubjective agreement is indeed useful in
controlling many perceptions, but can be tested in most cases of
apparent agreement on a particular item by attempting to use that
item in controlling some different perception. That two men crawling
across the desert both agree that they see the same lake makes the
lake no more real than if there were only one person crawling. Of
course, there might “really” be an oasis in sight, and this reality
could be tested when the two men found they could drink from it.

Again "of course", there might be no desert, and both might be

hypnotized to perceive themselves in that situation. I don’t think
that matters, because the perception each would have is that the
lake was no mirage, but was a real lake. It would matter, however,
if the hypnosis persisted to the extent that the subject died of
thirst thinking that the water was real.

        BN: Of course the measurement, the process of measuring,

the instruments of measurement themselves, cannot be
demonstrated to be anything more than controlled
perceptions. When we crank the epistemological implications
of PCT up to the paradoxical sticking-place, intersubjective
agreement appears to me to be our only route out of
solipsism. Equally of course, as each of us in turn have
acknowledged at one time or another, we assume that our
perceptions are veridical, at least in some functional
sense. Seriously not to do so would surely be paralyzing.

Do you really believe that? Do you always assume that your

perception of a dark patch on a cold sidewalk is wetness? That could
be very dangerous in a Toronto winter. Or do you just assume that
you veridically see a dark patch and not use that perception in
controlling any other perception? (For the benefit of those living
in more tropical climes, a dark patch on the sidewalk might, but
probably does not, mean that the sidewalk is covered by ice, and a
careless step might mean a dangerous, possibly fatal, fall). I don’t
find it paralyzing to step carefully onto such a dark patch to see
whether my foot slips before I put my full weight on it.

        I doubt that anyone has carried that line of

epistemological doubt any farther than mind games.

/Bruce Nevin

It's not mind games to try controlling something that depends on a

perception that may not be veridical, as the desert crawlers
discover when they die of thirst, or the walker on a possibly icy
sidewalk finds when the tentative step determines that the wet patch
is or is not ice.

Martin

[bnhpct]

[Bruce Nevin (2015.12.23 ET)]

O tell me where is q.i bred,

Or in the world or in the head?

What we perceive we cannot show,

For to perceive is all we know.

The canonical PCT diagram shows q.i in the environment, along with d and q.o.Â

Reflection suggests that what is labeled “Environment” could more properly be labeled “Observer’s perceptions”. Of course the observer assumes the reality of those perceptions.They are the observer’s perceptual inputs, as distinct from items on the “Organism” side of the dotted line, which refer to imagined perceptions inferred from the PCT model.Â

Also shown in the environment is that which is influenced by d and q.o, and measured as q.i. Years ago we called this the controlled variable (CV), then we called it the environmental controlled variable (ECV) because p is the controlled variable, and now the complex environmental variable CEV, and perhaps other things in the interim. This thing-which-cannot-be-named is not equivalent to the environmental feedback function, but it is co-located with it in the environment, or perhaps within it.

That which is influenced by d and q.o and measured as q.i, whatever we call it, has in itself no quantitative value or computational function in the model. It is completely neutral, a grey blank, an empty box with an unstable label. It is a mere placeholder for our conviction of the reality of our perceptions. It resists naming because that reality remains unknowable other than as the perceptions which are otherwise labeled in the diagram. The observer perceives it; the subject perceives it, and when during the TCV the observer disturbs its state and the subject resists that disturbance, the observer and the subject are in conflict over its state. Interpersonal conflict is a form of collective control. Collective control is our principal means of discerning what is real from what is illusory by establishing intersubjective agreement. Obviously, this means of discernment is not perfect—illusion and uncertainty persistâ—but with the refinements of scientific method collective control is the best we’ve got.Â

And so the canonical PCT diagram includes a placeholder for that which is controlled in the environment, although all we can know of it is perceptions, and the relevant perceptions all have their own labels in the diagram, and that is why we don’t know what to call it. Our perceptions tell us that its state is an effect of actions q.o plus disturbances d, though only the perception of its state is controlled. But it’s still there in the diagram.

And if it weren’t, we’d have a devil of a time explaining PCT to anyone.

···

On Tue, Dec 22, 2015 at 9:29 PM, Richard Marken rsmarken@gmail.com wrote:

[From Rick Marken (2015.12.22.1830)]

Martin Taylor (2015.12.21.12.39)–

  In response to [From Rick Marken

(2015.12.20.1830)] and

[Bruce Nevin (2015.12.21.08:40 ET)]:

        BN: ... I

think successful control of a perception is our only warrent
for attributing reality to it, that collective control
strengthens that warrent by intersubjective agreement, that
successful collective control implicitly includes and is
based upon corroborating (testing for) controlled variables,
and that we have the strongest warrent for attributing
reality to q.i as determined by the TCV (best when
replicated)…

MT: This is essentially the answer I was intending to write to Rick.

Anyone can imagine and simulate arbitrary perceptual functions, as
in Rick’s spreadsheet, but unless the results of applying those
functions is effective in controlling some perception,
reorganization will soon modify or eliminate them.

RM: Â Well, this isn’t quite what I was after as a comment on the spreadsheet. The perceptual functions in that sheet are, indeed, “arbitrary”, but only so as to make the point that what is called the controlled quantity and notated q.i is the observer’s view of the perception, p, that a controller is controlling.Â

RM: The spreadsheet is meant to demonstrate that the controlled quantity is not an entity in the environment; it is a variable aspect of the environment that is defined by the perceptual function. The observer doing the Test for the Controlled Variable (the Test) is trying to determine what this perceptual function is. The observer does this by testing hypotheses about the perceptual variable that the controller is controlling. These hypotheses are actually guesses about the nature of the perceptual function that produces the controlled perceptual variable. These hypotheses can be implemented using mathematical functions (as they are in the spreadsheet) or using one’s own perceptual systems (as they are in the example of “The Coin Game” on pp. 236-238 of B:CP, 2nd edition).Â

RM: Regardless of how hypotheses about the perceptual function (and, hence, about the perception that is under control) are implemented, the resulting perception is the controlled quantity, q.i, seen from the perspective of the observer. If the Test shows q.i to be under control (it is protected from disturbances that should cause to vary) then we conclude that our perception of the controlled quantity, q.i, corresponds to the perceptual variable (the controlled variable) that the controller is actually controlling. (Emphasis mine;-)

RM: So when our perception of q.i passes the Test – it is shown to be controlled-- then we can say that q.i (our perception of the controlled variable) is equivalent to the perception,p  the controller is controlling. This  has led me to make some revisions of my spreadsheet demonstrating the TCV. These are mainly changes in notation and the data available to the user when doing the Test. The new spreadsheet is attached.Â

RM: Now all you see in the upper left is gray cells that can be used to change your hypothesis about the perceptual function that is producing the perception that the controller is controlling. That perceptual function is hidden so that you are now in the same position as an observer doing the Test for real. You (the observer) have reason to believe that the controller is controlling a perceptual variable that is a linear function of two environmental variables, x.1 and x.2. So you adjust the coefficients of x.1 and x.2 (under the constraint that the coefficient of x.2 must be >0, otherwise the feedback effect of the output on the controlled quantity will be zero or positive).Â

RM: Each change of the coefficients changes q.i, your hypothesis about the perception that is under control. You test the accuracy of each hypothesis by seeing whether the controller protects the hypothetical controlled variable, q.i, from disturbance. The extent to which this happens is measured by the Stability value in the green cell. When the stability of q.i is up above .9 then q.i, as defined by the coefficients of the perceptual function, is being protected from disturbance and it can be concluded that q.i is very close to the perception, p, that is actually under control.Â

RM: Once you find a definition of q.i that has a high Stability value – i.e. your perception of the controlled variable, q.i, seems to correspond to the perception, p, that the controller is controlling – you can see how well you did (something you can’t do in the real Test) by hitting Ctrl u which unhides the actual perceptual function that defines p.Â

RM: So that’s it for now. Again, I’d really appreciate some comments (from anyone) on this spreadsheet. I know that methodology is not everyone’s cup of tea but I do think that understanding the Test can give you a better understanding of PCT even if you are not into PCT as a researcher.

BestÂ

Rick

–
Richard S. MarkenÂ

Author, with Timothy A. Carey, of  Controlling People: The Paradoxical Nature of Being Human.Â

[rsmarken]

Content-Type: application/vnd.ms-excel.sheet.macroEnabled.12;
  name="TCVModelwColor2.xlsm"
Content-Disposition: attachment; filename="TCVModelwColor2.xlsm"
X-Attachment-Id: f_iinfr3ku0

TCVModelwColor2.xlsm (38.1 KB)