[From Bruce Abbott (2015.11.13.1830 EST)]
Fred Nickols (2015.11.12.1532) –
BA: I agree with you here:
When we say “Behavior is the control of perception” different meanings can attach to that statement.
The statement could be taken to mean that someone who is behaving is controlling one or more perceptions. I think that’s accurate and I don’t have any problem with it.
Similarly, the statement could be taken to mean that behavior controls perception. This is the equivalent of saying that behavior serves to control perception or that behavior is the means by which perceptions are controlled. I think all of these are accurate and I don’t have any problem with any of them.
And, the statement could be taken to mean that the term “behavior” refers to the entire negative-feedback, control loop that undergirds PCT. I struggle with that.
As I’ve thought and believed for years, one use of the term “behavior” refers to the activity of an organism, most often to its observable actions but not limited to those. It is my view that the actions of an organism are the means by which it controls its perceptions. I see no conflict between this view of behavior and the earlier statements, except for the one that equates behavior with the control loop that enables the control of perception. The loop includes behavior and it includes a lot more (e.g., reference signals and perceptual signals to name just two additional elements).
So, for me, behavior is part of the loop and the entire loop is required in order for behavior to control perception, but there’s more to the loop than just behavior.
I can describe the loop’s functioning in action terms as follows:
-
perceiving existing conditions,
-
formulating desired conditions,
-
comparing perceived conditions with desired conditions and, when perceived and desired conditions are not a match,
-
acting to align perceived conditions with desired conditions.
BA: But I have problems with this part:
That list clearly indicates behavior permeates the entire loop and, conversely, that the entire loop is involved in behavior.
What just dawned on me is that my notion of “the activity of the organism” includes all the items listed immediately above. The activity of the organism permeates the entire loop and is not limited to the output quantity. The activity of the organism controls perception. The activity of the organism is the control of perception. The activity of the organism is behavior.
Behavior IS the control of perception.
BA: In science any variable can be said to exhibit “behavior,” when it varies. Scientific laws specify how a variable’s behavior is determined by the influences of other variables. For example, the gravitational attraction between two bodies varies inversely with the square of the distance between them. We can use this law, together with an initial position, velocity, and acceleration of the objects with respect to each other, to predict their behaviors over time (relative positions, velocities, accelerations), such as the two objects orbiting a common center of mass. But this behavior is not control.
BA: Your visual system creates visual perceptions, usually dependent on the images being projected on the retinas of your eyes. The changes in the outputs of the photoreceptors might be considered “behaviors” of the photoreceptors in response to changes in the image, but they are not control. The resulting “behaviors” of the neural signals arising from the associated neural networks are not control. The perceptions arising from the complex bottom-up and top-down neural influences might be considered behaviors of the perceptual signals but they are not control. The perception of the position of a nail-head with respect to a board surface is likewise a state arising from the behavior of the perceptual system but it is not control.
BA: Control is an emergent property of a particular organization of parts, it is not a property of each individual component of the control loop.
BA: There are many things you perceive that you do not control. How those things vary might be described as their “behavior,” but such behaviors are not control.
BA: Behavior, defined as the actions (output) of a control system, is not control. Such behavior may occur and yet fail to establish control over a perceptual variable. Behavior is the means of control; it is not itself control. Behavior, defined as the way a variable changes over time, is not control. It is just variation.
BA: Behavior as Doing
BA: In ordinary language, when we speak of what someone is “doing,” we refer to the effect of the person’s actions as we perceive them. For example, we may perceive that the person is using a hammer to drive a nail into a board. We may even infer that the intended result is to have the head of the nail flush with the board. Such results depend on control – randomly flailing the hammer about is unlikely to produce the observed movements of the hammer and nail. But in all cases, what is being controlled (when control is successful) are the perceptions of the person who is doing the doing. What the observer perceives are the observable effects of that control.
BA: Consider a person whose index finger-tip is observed to be tracing an approximate circle in the air. To accomplish this, the person must vary a set of muscle tensions so as to move the finger-tip in the required arc while opposing the effects of gravity and perhaps other disturbances. As a result, the behavior of the finger-tip, as observed by the person and by the observer, is to trace an approximate circle in the air. But this is “behavior” in the sense of observed variations in position of the finger-tip. What is actually being controlled is the perception of the person doing the movements, of the finger-tip moving in a roughly circular pattern. And the behaviors (actions) being employed to accomplish this control consist of a complex pattern of muscle contractions.
BA: Those muscle contractions are themselves the products of control systems acting to contract the muscles in a specific pattern according to reference values being varied by the higher-level system that is attempting to perceive a circular movement. So actions may be both the means of control of one set of variables and simultaneously the products of perceptual control at the next higher level.
What we say the person is doing is “drawing a circle in the air,” and if you believe that there is a real fingertip out there, it may indeed be moving in a roughly circular pattern (a “behavior”). In that sense the “behavior” is being controlled (or more accurately the perception of that behavior). But the actions that generate this perceived behavior are the means of control, not control itself. Similarly, the controlled behaviors (perceptions) are not control, they are the products of control.
Bruce Abbott