Vicarious conditioning a la eyeblink "reflex"

[From Dick Robertson,2008.02.18.1920CST]

[From Bill Powers (2008.02.17.0730 MST)]

Bruce Nevin (1:21 PM 2/16/2008 -0500):


I can’t figure out from this what (among other terms)
“conditioning” is, what is meant by a
“stimulus”, what “elicit” means, or what model
is implied by “reflex.” One is just supposed to know, I guess.
Or maybe nobody knows, but no one wants to be the first one to ask.

In PCT, we offer the theory that such actions are part of a control
process organized to maintain some variable in a reference state, with
the independently-manipulated variable acting as a disturbance of the
controlled variable. The controlled variable is affected in one way by
the independently-varied environmental variable, and in the opposite
direction by the action. We would say that the pin’s entry created a
perceptual signal in some control system having a reference level of zero
for that signal. The resulting error signal was routed and amplified to
operate the muscles that moved the hand in the way that reduced the
perceptul signal to zero.

Dick Robertson, some time ago, had some useful comments to make about the
eye-blink response to a puff of air. Perhaps he would repeat them for us

Glad to. I’ll paste in from Intro to Modern Psych, pp. 116-117

116 Introduction to ModernPsychology

Some investigators have objectedto generalizing about learning from the work on the “withdrawal reflex,”claiming that it only demonstrates the conditioning of a reflex, and is not anexample of forming an association. We have tried to show in the abovediscussion that this is a misperception resulting from the way events werelabelled in the first place.

If you have an “association”between what you call a perception and what you call a reflex, that is whatPavlov called “conditioning.” If you have an “association” between two“perceptions” and don’t consider what variable is being controlled at the time,you could call it “association of ideas.” The boundaries of what you see as thephenomenon are dictated by the way you label it. Thus, early psychologistsnoted associations between ideas, or between perceptual categories, such asbetween the name and color “orange.” Since, following Descartes, they thoughtof the perception as coming into the organism, rather than as controlled by theorganism, they had no reason to think that there was any “doing” going on inthe organism, as there is when muscles are contracting and the body moves.True, the “doing” in the case of theassociation between a sensation of color and an act of naming involvesrelationship control, which occurs at the level of computing in the brain anddoes not necessarily result in physical activity. This distinction wasperceived previously as fundamental in distinguishing reflexes fromassociations. Reflexes were thought of as mechanically invariant and alwaysinvolving some sort of action, such as the knee jerk or secretion of saliva, tomention two of the classic cases upon which traditional theory was built. Oncethat way of looking at events had been constructed, scholars simply didn’t seefacts which didn’t fit into that frame, as shown next by a discussion of areflex which failed to occur regularly, as required by traditional theory.

8.5 An Experiment with theControl of Eyeblinking

Some early students ofcontrol-theory psychology became interested in repeating a classical laboratoryexperiment to determine whether they could identify the controlled variable ina simple control system. The classic “eyeblink reflex” was chosen as a goodcandidate. There is a legend among psychologists that it provides a niceillustration of the conditioning of a reflex. The apparatus is simple. Youconstruct a chin rest in which a subject places his or her head, and next to ityou place a pipette which is connected by a tube to a rubber ball or syringewhich can be used to send a blast of air onto the side of the subject’seyeball. The “conditioning” involves making a noise, like a click, just beforethe air blast, and the subject is supposed eventually to blink at the tonewithout the air blast.

My students set up thisexperiment with one added feature. They asked the subject to work arithmeticproblems which were on the blackboard in the room, telling him or her that theywere studying whether the process is affected by the experimental conditions,which they did not clarify. This complication was imposed in hopes of keepingthe subject from attending to the air blast as something to be concerned about.

The first subjectperformed as expected. After a time, he blinked even when the air blast did notcome. However, with the second subject, something different happened. He beganto blink more frequently, but randomly as related to the time of the air blast.

The explanations forthese differences draw upon higher levels of control, in which the eyeblink isonly the final output. The subject who showed conditioning turned out uponfollowup questioning to have accepted the task as it was construed for him bythe experimenters. He tried to go ahead with working the problems, paying aslittle attention to outside distractions as possible. The other subject said hequickly noticed the distraction, figured that it played a part in theexperiment, and assumed that he was to get the distraction “under control.” Hedid this by resetting his reference signal for more frequent blinking. In otherwords, the two subjects defined their overall purposes in different ways,resulting in different kinds of overt activity.

The perception of the eyeblink as the main subject of interestwas from the experimenters’ point of view, and it did not take into account thefact that each of these subjects constructed a different control system for the(different) variable defined by his purpose. To reach that conclusion, it is necessaryto start from a paradigm in which the nervous system is considered to be aninterlocked pyramid of nested control systems, rather than a collection ofisolated systems which can or cannot become linked by some hypothetical processcalled “association.”

One further observationresulting from the analysis of these experimental results might haveconsiderable consequence for the labelling of such phenomena as various“reflexes.” It seemed to us, after going down through the different controlledvariables which our subjects were controlling, that it would be appropriate torename the “eyeblink reflex” a “corneal-lubrication control system,” for thatis basically what it is.

8.6 Are There Different Kinds ofLearning?

There is currently considerabledebate about whether there is more than one “kind” of learning: whether “skill”learning is different from “perceptual” or ° “complex” learning, etc. (See, forexample, Fox, 1983.) This confusion resulted in part from the associationisttendency to conceive of every behavior as a thing in a itself. Later, when itbecame apparent that some learned behaviors are parts of other learnedbehaviors, this tendency already had become entrenched, and it interfered withthe idea that all behavior could be interlocked in an overall hierarchy.

Thus, we find Gagne (1970)describing seven categories (or kinds) of learning. They are arranged In termsof increasing complexity—but it is “complexity” as defined by subjectiveimpressions, not in terms of identifying the number of orders of controlsystems involved.

Sorry for the irregular text. I can’t figure out why I can’t edit the pasted material properly, but it’s all there anyway. I can add one remark that didn’t go into the text. I asked the “experimental” man in my department about the eyeblink reflex and he told me, “Oh, it’s a hard one to duplicate. You don’t always get the result.” That didn’t seem to bother him at all.


Dick R