Behavior, Equilibrium and Control

[From Rick Marken (950712.073y0)]

Bruce Abbott (950712.0845 EST)

Perhaps you could summarize for us just what my "approach to PCT" is--in
your view. Maybe then I can see what it is you don't like about it.

I think your approach to PCT is based on the notion that PCT is an alternative
model of "behavior". In conventional psychology, "behavior" is any measurable
result of an organism's actions: behavior is "output". This approach to
behavior does not distinguish between intended (controlled) and unintended
(uncontrolled) results of action; results is results. This view of behavior
is completely superficial; psychologists who are willing to call certain
behaviors "intended" or "goal-oriented" do this on the basis of the
appearance of the behavior -- not based on an understanding of control. This
is why psychologists don't know the difference between the movement of a
point toward an attractor basin and the movement of the thumb to cover the
image of the earth; in both cases there is a particular result (point at
basin, thumb over image of earth); it looks (superficially) like these are
the same phenomenon. So we get mathematically inclined psychologists
applying chaotic attractor models to limb positioning tasks -- when, in fact,
limb positioning and movement to an attractor basin are two different

An understanding of PCT must (I think) begin with a clear understanding
of the nature of the phenomenon that PCT explains. PCT does NOT explain
"behavior" as conventionally defined (and studied in conventional
psychological research). PCT explains "behavior" as "controlled
results of action". This is what my "mind reading" demo is about. It shows
that the conventional definition of behavior tells you nothing about what
an organism is "doing" (controlling) by just looking at the visible results
of the organism's actions. Conventional psychologists have been in the
position of the person watching the five numbers move around the screen;
their only basis for choosing which number is actually being moved
intentionally is by looking at the movements and seeing which number APPEARS
to be the one moved intentionally; in other words, they are just guessing.
And even when they guess right, they have no idea how the movement of
the intentionally moved number differs from the movement of any of the
other numbers (it differs, of course, in the sense that the actor is
systematically resisting the invisible disturbances and keeping the
number in the intended state-- the one that is seen on the screen).

PCT begins with a re-evaluation of the nature of behavior itself. It shows
that behavior -- intentionally produced results of action -- are controlled
results of action. Conventional psychologists are looking at behavior from
the wrong perspective; they are studying the wrong phenomenon. This is
why they don't know the difference between equilibrium and control and it is
why many think control theory is just an alternative model of "behavior".
Control theory is about a phenomenon that has not yet been studied in
conventional psychology: it is not about "behavior" as conventionally
defined; it is about CONTROL.



[From Oded Maler (950712)]

Re: Rick Marken (950712.073y0)

I probably asked this one once, but I don't remember the answer:

Why *in principle* are Equilibrium and Control different? Consider a high
dimensional universe consisting of many physical variables interconnected
via complex dynamical laws (you agree that something like this is underlying
everything, don't you?) Now you pick two such physical variables that
happen to correspond to some neural signals and you call them P (perception)
and R (reference). The set of points of this huge state-space that satisfy
P=R are indeed the bassins of attraction of the system, to which the system
will flow from other regions in the state-space.

The only point is that although the projection of the trajectory on the
P-R plane might be simple (we can assume R to be constant for simplicity)
these two variables are not observable at all (and will probably never be),
and the projection of the trajectory on the *rest* of the world (the other
variables) might look very complex and not reveal what is going behind
the scene.

In some sense, the soap bubbles have a very limited interface with the
observable (by us) world and hence they do not confuse us as much as
humans do - they do not cause so many side-effects in their way to
equilibrium as does humans.

Well, it started as a question and ended as an answer, but anyway it's
a side-effect of something else..



<[Bill Leach 950712.19:40 U.S. Eastern Time Zone]

[From Oded Maler (950712)]

Why *in principle* are Equilibrium and Control different? ...

In equilibrium phenomenom the forces opposing a disturbance are precisely
and always equal to the value of the disturbing force.

In control the same is only generally true (though for steady state
references and zero transport lag it is also exactly true).

In equilibrium the effect(s) (ie: deformation of a soap bubble) caused
by the disturbing force(s) is/are precisely and always equal to that
value predicted by application of physics laws.

In control the effect(s) (ie: deformation of a soap bubble) caused
by the disturbing force(s) is/are NOT equal to that value predicted by
application of physics laws. This is indeed exactly how one can
determine that a control phenomenon exists! When the disturbing forces
should cause a certain amount of change in a CCEV but do not (and there
are no unknown physical phenomenon involved) then a control system is
also operative upon that same CCEV.

The so called phenomenon of "equilibrium" is really only a physcial
situation where non-linear forces are in balance.

In a control system "balance" or "equilibrium" situation typically a
potentially overwhelming force "is balanced" against disturbing forces by
applying only as much of the overwhelming force capability as is needed
to "neutralize" the disturbing force.

As has been demonstrated both physically and theoritically, this can not
be done in the general case by calculation of output values. Only a
control system can match this overwhelming output capability against
random vectors of disturbance to a CCEV.

Even in that case the "match" is only to the portion of the disturbance
vector that would affect a controlled aspect of the CCEV. All other
force effects of the disturbing force remain unopposed by the control