controlling for

[from Wayne Hershberger]

Bill (WTP), I've been out of town for 2 weeks and
missed your post regarding a joint effort (paper).
Could you please post it to me. Thanks.

Bruce Nevin (930427.1232) --
          A while back (it has been a while) as I was
     going over something that I was preparing to send
     to CSG-L, I made the following correction:
     controlling for the perception of X ==>
     controlling the perception of X. Since then, I
     have noticed this locution a number of times, even
     in something that Bill wrote. The reason I
     thought of it as an error and corrected it is
     because it seems to say "controlling actions for
     the perception of X," when what we want to say is
     that perception and not action is what is
     controlled.

Bill Powers (930428.0815)
     This was something that my students in a 1972-1973
     seminar started saying, and I picked it up.

     Bruce, I share your concern with the expression
"controlling for," essentially for the reason you
mention. I am afraid that the students in Bill's 1972-
73 seminar did not understand HPCT as well as he would
like to believe. Clearly, they came to believe him,
but not necessarily to understand him--at least, not as
well as you do. To me, your correction of their
mislocution is a clear indication that you understand
HPCT far better than they did.
     During the final session of our 8th annual meeting
last summer in Durango, I initiated an hour-long
colloquy among those present by asking what the
expression "controlling for" means. My purpose was not
to answer a rhetorical question, but to do some
consciousness raising. I wish you had been there to
participate because your point is very well taken.
     The symbol X in the expression "controlling for X"
appears to be something that is wanted, either (a) as a
reference signal or (b) as a wish. An example of (a)
is a cruise-control system "controlling for" a road
speed of 60 mph (i.e., it wants to keep the sensed road
speed at that reference value). An example of (b) is a
man buying a lotto ticket, wanting to win the prize
money; that is, he controls his eligibility for the
prize which he wishes (i.e., only hopes) to win.
     Since, a want is a wish (i.e., only a hope)
precisely to the degree that it is NOT a reference
signal of a control system, "controlling for the prize
money" is simply an awkward oxymoron. And, more often
than not, the expression "controlling for" is used
precisely in this way; that is, as a terribly
inappropriate substitute for the expression "wishing
for." In my view, it is a bad habit.
     This is not to say that wishes are improper
conative concepts, but only that they do not imply
control, but, rather, the lack of control, or the need
to develop control. I think that this is what Bill was
talking about in the remainder of his reply cited
above. But the precise role that wishing may play in
reorganization has yet to be systematically examined.

Warm regards, Wayne

Wayne A. Hershberger Work: (815) 753-7097
Professor of Psychology
Department of Psychology Home: (815) 758-3747
Northern Illinois University
DeKalb IL 60115 Bitnet: tj0wah1@niu

[From Bill Powers (980809.0805 MDT)]

Bruce Nevin (980809.0420 EDT)--

So "controlling for x" just means "trying to control x."

Not quite. When I say I am controlling for 72 degrees F (as I adjust the
thermostat), I am saying that I am controlling a temperature (x) and that
the set-point is 72 degrees (x-ref). I am also saying that even the
temperature isn't very close to 72 degrees, that is still what I am aiming
to obtain (as I pound on the controller trying to get it unstuck).

Rick's notion of control as "organized as a control system" sounds like
"controlling for":

To me, "organized as control system" means "possessing all the components
of a control system in working order, hooked up together in the proper way,
and in an environment to which that control system is adapted."

The attempt to move the house is the behavior of the mover. The attempt to
resist overwhelming force is the behavior of the victim of coercion. The
coercer does not control this behavior of the victim. The coercer only
forces it to be "controlling for" rather than "controlling."

Yes. However, I think you're overlooking the case where the coercer's goal
is specifically to produce a particular action by the coercee. The coercer
is pre-empting the output quantity of the coercee.

But all this is unnecessary. If we just draw the model of the two systems
with the interaction point properly designated, we can see what the
combination of systems will do under various assumptions about gains,
reference levels, etc.. If anyone still wants to.

Best,

Bill P.

[From Bruce Nevin (980809.0420 EDT)]

Bill Powers (980807.1501 MDT)--

So "controlling for x" just means "trying to control x."

Rick's notion of control as "organized as a control system" sounds like
"controlling for":

Rick Marken (980806.1600)--

your house mover (the person pushing
on the house) is controlling because he is organized as a control
system with respect to his perception of house position.
[...] the mover is controlling
(unsuccessfully, if the feedback connection is zero) because he
is organized as a control _system_ with respect to perception
of house position.

The attempt to move the house is the behavior of the mover. The attempt to
resist overwhelming force is the behavior of the victim of coercion. The
coercer does not control this behavior of the victim. The coercer only
forces it to be "controlling for" rather than "controlling."

  Bruce Nevin