My point is that the term "controlled variable" (cv) represents the
_observer's_ perception of the perceptual variable (p) that the actor
is controlling. Both the cv and p are perceptual variables.
Fine. Why are you belaboring the obvious? It's true; so what? Are you
arguing for solipsism?
My response is that information is not required as an explanatory principle
in control theory, it is not a quantum in modelling control, and it is not
required in order for control to happen. If you want to talk about
information and entropy you can say that the process of control creates
information. It does not use it.
Bruce Nevin (980308.1425 EST) --
Actually, the process of control destroys information. For example, if your
home HVAC system worked perfectly, the temperature of the air in your home
would offer no clue as to the outside temperature.
Actually, the process of control destroys information. For example, if your
home HVAC system worked perfectly, the temperature of the air in your home
would offer no clue as to the outside temperature.
Without control, the temperature inside the home would be the same as the
temperature outside. With control, they are different. The difference is
information created by control.
What offers a clue as to the outside temperature is not the air in my home
but a thermometer, or a person complaining about the temperature, and those
could be placed outside as easily as inside. Placing them, then, both
inside and out, with HVAC control we get twice as much information, a
different temperature reading (or judgement) in each place.
But the HVAC doesn't use information to bring this about. It has no sensors
for information, no functions for it, no place to store it. By the
recursive process of control it creates information, a difference from the
state the environment would have got to without it. Or so I understand from
those who control perceptions called information theory. I could be wrong.
Actually, the process of control destroys information. For example, if your
home HVAC system worked perfectly, the temperature of the air in your home
would offer no clue as to the outside temperature.
"Information" doesn't satisfy my criteria for correspondence with something
real. One of my reasons is that it depends on what is receiving it just as
much as on what it is about. We can't think of information as having some
independent existence in the environment, so that some of the original
information gets lost as sensory signals are generated or as outputs cancel
some of the effects of disturbing variables. If the receiving system
already knows about the state of the environment, it gets no information
from the environment even though the same sensory signals are generated and
even though control cancels the effects of disturbances. If the receiving
system is set up to recognize a different aspect of the environment than
what we assume, it gets no information at all of the kind we calculate. The
properties of information seem to be mainly artifacts of the way we
calculate it.