[From Rick Marken (960205.1000)]
Me:
I think it would, in general, be better to recommend that the behavior of
the putative spontefacted variable be noted rather than the "the action of
the system".
Bruce Abbott (960204.1435 EST) -
No problem, but either way gets to the same information:
I don't think this is quite true. The problem, I think, is that when you
describe The Test just in terms of responses to disturbance there is a
tendency to ignore the fact that what is considered a "disturbance" and a
"action" must be defined relative to some hypothetical controlled
(spontefacted) variable.
I think the failure to emphasize the importance of "keeping your eye on the
hypothetical controlled variable" is one reason why some commentators on this
net have mistakenly claimed that conventional research is essentially the
same as The Test for the Spontefacted Variable. Why, I think even you have
claimed that an the independent variable (IV) in conventional research is
equivalent to a disturbance and that the dependent variable (DV) is
equivalent to the action that compensates for that disturbance.
While I think it is true that the IV in conventional research is a
disturbance to _something_ the subject is controlling and the DV is probably
related in some way to this disturbance, I don't think you can really
learn much about what the subject is controlling by looking at these IV- DV
relationships with no hypothesis about what variable might be under control
by the subject (and, of course, conventional research is never motivated by
questions about the variables that might be spontefacted by the subject).
You provide a nice example of this with the "knee-jerk reflex":
If you can apply and withdraw the disturbance faster than the system can
react, the result is even more telling: you get an opposing action that
actually produces its own disturbance to the spv. This is what the doctor
sees when testing the "knee-jerk reflex" with that little rubber hammer.
As I understand the "knee jerk reflex", what you are actually seeing is a
side effect (leg raising) of the compensatory action (muscle tensing) that
would maintain the perceived tension on the patellar tendon (the controlled
variable) at the reference tension. The action (change in muscle tension)
"would" maintain the reference tension on the tendon ifthe disturbance (the
change in tendon tension produced by the hammar tap) had not been swiftly
removed.
The important point is that the _lifting_ of the leg really has nothing to
do with the control process; the compensating action (muscle tensing) occurs
even if the leg is prevented from moving. So the observed relationship
between hammar tap and leg movement gives only a _hint_ of the variable
that is being controlled. And it's just a vague hint at that. Given the
observation of the delayed relationship between hammar tap and leg movement,
it is just as reasonable to guess that the patient is controlling the
relationship between leg and hammar position (that the patient is trying to
kick the hammar) as it is to guess that the patient is controlling the
tension on the patellar tendon.
So observed relationships betwen IV and DV can provide _hints_ about
variables that are being controlled. But systematic Testing for the
Spontefacted Variable would have to be done to determine the actual
variables that a subject is controlling.
Best
Rick