[From Rick Marken (950221.1500)]
Lars --
At long last, a short comment about my conflict demo. The conflict demo is
just another version of a tracking task (naturally). You (the subject) keep a
dot on a target by moving a mouse. The dot is disturbed in the x and y
dimensions and you can compensate for these disturbances by moving the mouse
in the x and y dimensions. The catch is that mouse movement in the x
dimension is a disturbance to the y position of the dot and mouse movement
in the y dimension is a distrubance to the x position of the dot. That is,
x = mx+my+dx
y = mx+my+dy
where x and y are the horizontal and vertical positions of the dot, mx and my
are the horizontal and vertical positions of the mouse and dx and dy are
disturbances to the x and y positions of the dot, respectively.
In this task you are controlling two variables, x and y. But the environment
is set up so that that there is really only one degree of freedom available
to control both variables. So it is impossible (unless the disturbances are
very cooperative) to maintain both x and y in their goal states at the same
time.
If you try to keep the spot on target in this situation (maintain both the x
and y reference states of the dot) you will be in conflict with yourself; the
harder you try to correct the error in the x dimension, the more you increase
the error in the y dimension; and the harrder you try to decrease the error
in the y dimension, the more you increase the error in the x dimension. In
this demo, you are WIRED for positive feedback as long as you try to maintain
the goal of keep the dot on target; and positive feedback is VERY negative;
it eliminates control -- FAST.
You "go up a level" when you can dissociate yourself from the task and see
that 1) you don't have to be doing this task (when a person realizes this
they just stop) or 2) that there are other goals that might be achievable
(controllable) such as keeping the dot located at the same x position while
it wanders up and down, uncontrolled, in the y dimension; or keeping the dot
located at the same y position while is wanders back and forth in the x
dimension. Both of these solutions to the conflict require making the two-
dimensional control task itself an OBJECT of awareness; that is, they require
that you consiously get OUT OF the task -- dissociate from it -- to see it as
something you are doing. You go up a level.
Bill Leach (950217.18:19 EST) --
While I absolutely agree with you, Bill, Tom, etc. concerning the
usefulness of the concept that there is information in the perceptual
signal about the disturbance, Martin IS NOT absolutely wrong in his
assertions.
In any REAL control system, gain is less than one and therefore any
disturbance MUST result in a NON-ZERO net error output change (even if
response speed is still instantaneous).
But we don't know what component of this error is due to the disturbance.
Error is r-p or (r-o+d); even if this error is diminishingly small, we still
don't know what component of the error is due to variations in r, o or d.
So even when control is nearly perfect, the residual error contains no
information about whatever disturbances might be contributing to the
difference between perception and reference.
I think it is important to understand that, in PCT, perception is just THERE.
It is only good or bad, right or wrong with respect to reference signals.
This is an important distinction between PCT and other approaches to
psychology (and philosophy); the latter tend to see perception as having a
lot of intrinsic properties, some theorists seeing perceptions as "rewarding"
and "punishing", others seeing them as "affording", still others seeing them
as "informing", or "stimulating", etc. In PCT perceptions are just there;
they differ from one another in kind but they have no intrinsic properties.
Perceptions get their "punch" only in relationship to the reference signals
in one's own head. In PCT, perception by itself is just perception; it is r-p
that matters in behavior.
Best
Rick