[From Rick Marken (2017.02.25.1210)]
···
Bruce Abbott (2017.02.25.1250 EST)-- Â Â Â Â Â Â
Â
BA: Note that, in Rick’s version, both p and Qi are equal to the product of v1 and v2. Therefore Qi = p, save for a conversion factor to rescale p from the units in which the environmental variables v1 and v2 are expressed to the units in which p is expressed (e.g., neural impulses/second). Rick, if I understand him correctly, says that Qi is the observer’s perception of the controlled variable, and therefore does not exist unless there is an observer.Â
RM: Correct!
Â
BA: But this leads to the problem that Qi is no longer an input quantity (located in the environment) but rather another p, located in the observer.
RM: It’s not a problem because Qi need exist as an input quantity only when there is an experimenter present who is interested in discovering what the controller is controlling. That is, when an experimenter is present doingthe TCV.
Â
BA: So would it be better to identify the little circle in Rick’s system with another p? And how should it be computed? To compute this p, wouldn’t we need to know the observer’s input function, rather than the control system’s?
RM: Qi is itself the output of a perceptual function in the observer. If the observer is a human it is unnecessary to know how it is computed; the observer’s nervous system does it for him. This is what happens in the coin game where E can perceive the “arrangement of the coins” as the variable being controlled (Qi) and “zig zag” as the reference state of that variable. In more precise studies of controlled variables, the computer computes functions of environmental variables as candidates for the perceptual function that produces Qi. This is what happens in the “What is Size” demo, for example, where the computer computes two possible functions of height and width – h + w  and h * w – as candidates for the variable, Qi, that is actually being controlled.Â
Â
BA: This whole problem can be avoided by assuming that the quantities entering the input function are real environmental variables like light intensities or water temperatures rather than an observer’s perceptions of them.Â
RM: But that’s what PCT assumes to be entering the input function of both the observer and the controller. The problem that you imagine to exist simply doesn’t exist. I have done quite a bit of research using the TCV and the problem you imagine to exist simply doesn’t.Â
Â
BA: We eliminate the observer from our description of how control systems work (they appear to work just as well when not being observed) and we don’t have to deal with silly philosophical questions similar to the one asking, if a tree falls in the woods and there is nobody there to hear it, is there a sound? (The answer depends on how one defines “sound.�)
RM: The observer is not part of the description of how a control system works. The observer is the person who observers the phenomenon that is to be explained – the fact that various aspects of the environment are controlled. The aspects of the environment that are observed to be controlled – things like one’s balance while walking, taking sips of tea or defending a principle – are Qi; controlled variables from the point of view of the observer. Control theory is the explanation of these observations.Â
Â
 BA: Put THAT in your pipe and smoke it!
RM: I prefer a cigar, thanks. But I’ll put that in my pipe and smoke it if you will stop putting the cart before the horse. I think you and some others would do well to heed Bill’s (and my) advice and deal with phenomena first, theory second.Â
BestÂ
Rick
–
Richard S. MarkenÂ
"Perfection is achieved not when you have nothing more to add, but when you
have nothing left to take away.�
                --Antoine de Saint-Exupery