[Avery Andrews 930125.1610]
The bad guys say:
Perception guides action
What Rick Marken says may (I hope) be paraphrased as:
Perception of the net effects of disturbances on a controlled
variable guides action
Which is clearly more specific, & therefore more useful.
Avery.Andrews@anu.edu.au
[From Rick Marken (930125.0800)]
Avery Andrews (930125.1610) --
The bad guys say:
Perception guides action
What Rick Marken says may (I hope) be paraphrased as:
Perception of the net effects of disturbances on a controlled
variable guides action
I hate to be a pain in the ass about this, but that's not quite
what I meant either. I meant:
Perception (of ANYTHING) does not guide action: action
guides (controls) perception.
All this is just based on the two basic laws of control system
operation:
1) p = r
2) o = -kd
The equals signs are actually approximations (which I can't
easily notate on the net) but VERY VERY CLOSE approx-
imations when loop gain is HIGH.
Law 1 says that perception depends on the setting of the
reference signal: perception is a dependent variable, the
reference signal is the independent variable. This law (fact)
of control is the basis of the title of Bill's book Behavior:
The control of perception. This is because we usually think
of behavior as a dependent variable that the actor is in
some sense "responsible" for. In a control loop, the only
dependent variable that fits this description is perception.
Law 2 says that output (or action) is negatively related to
the disturbance. The disturbance is the net influence of the
environment on the controlled perception, p; the action is
the net influence of the actor on the controlled perception.
The important (and AMAZING) thing to notice about
equation 2 is that p does NOT appear in it (on the left
OR right). The output (in a high gain, negative feedback
loop) depends ENTIRELY on the disturbance, which in
many cases (such as compensatory tracking) CAN NOT
EVEN BE PERECIVED. But whether d can be perceived
or not is irrelevant; the fact of the matter is that action
(output) depends ONLY on the disturbance AS AN UN-
SENSED ASPECT OF THE ENVIRONMENT.
From a conventional (S-R) point of view, law 2 is
just flat out magic or mysticism. As I said, it is just
something that you have to accept (like mass attract-
ion) even though it seems to completely contradict
everyday experience.
This is why modelling (even in very simple situations)
is so important for understanding and accepting PCT;
it helps if you can see with your own eyes that Law 2
really does hold for a high gain closed negative feedback
loop.
Best
Rick
[Martin Taylor 930125 12:15]
(Rick Marken 930125.0800) to Avery Andrews
Perception of the net effects of disturbances on a controlled
variable guides actionI hate to be a pain in the ass about this, but that's not quite
what I meant either. I meant:Perception (of ANYTHING) does not guide action: action
guides (controls) perception.All this is just based on the two basic laws of control system
operation:1) p = r
2) o = -kd
The equals signs are actually approximations (which I can't
easily notate on the net) but VERY VERY CLOSE approx-
imations when loop gain is HIGH.
A lot of the confusion might perhaps be removed if these laws were always
accompanied by "in the limit of infinite time." In practice, at the levels
of our experiments, infinite time is reasonably well approximated in seconds
or parts of seconds. But the dynamic behaviour of the control loop is
important. And dynamically, some part of the disturbance DOES appear in p.
That part is what leads people to think in S-R terms.
You can look at a closed loop starting anywhere, provided you get back to
where you started before you quit. Rick showed how one can get various
flavours of psychology by looking only at parts of the loop, in his
Blind Men and Elephant paper.
The fact that you can start anywhere carries with it the implication
that effects inserted anywhere have implications all round the loop.
We prefer to see that changing the reference results in changing the
percept (which is the point where the loop closes, starting at the
reference injection at the comparator). One could equally see the
effects as starting at the disturbance, which is compensated by the
results of the output on the world. And that's what an outsider
(experimenter) can see. The basis of The Test is the fact that the
disturbance has no long-term result in the world. But it does have
a dynamic effect that depends on the delays, gains, and impedances
around the loop. That dynamic effect is easily, but misleadingly,
seen as the effect of stimulus on response.
From a conventional (S-R) point of view, law 2 is
just flat out magic or mysticism. As I said, it is just
something that you have to accept (like mass attract-
ion) even though it seems to completely contradict
everyday experience.
It would be mystical, if there were no way for perception to be
controlled (if changes in perception were not reflected as error
that affected output that affected perception). It's the dynamics
that removes the mystical quality.
Martin