[From Bill Powers (930529.0800 MDT)]
Bob Clark (930528.1530) --
As I have increasing contact with "ordinary people" who are
interested in, and active with, the City Government, I find
more and more indications that they are, indeed, familiar with
the PCT view -- but tend to express it in rather different
language, intermingled with assorted other views.
I don't mean to minimize the understanding that people do have,
for I've found something similar. But people's mixing PCT
concepts with others often shows that they are unaware of the
conflicts involved.
From an earlier post:
"paying attention" to all the variables required to maintain
the bike upright, moving as desired, etc.
I remarked:
This is a non-explanation; it alludes to the existence of
explanatory terms without actually saying what they are.
And you objected,
Bill, OF COURSE this is a "non-explanation." It was NEVER
INTENDED to be any kind of "explanation."
My purpose was to DESCRIBE, in condensed summary, my concept of
what the rider is paying attention to as he observes his
changing surroundings.
I didn't express my comment well -- trying to be too brief. Does
the rider really pay attention to "the variables required to
maintain the bike upright," or is the perception under control,
and attended to, the uprightness itself? Output variables must be
adjusted to keep the bike upright, but the control system
concerned with uprightness does not perceive (or pay attention
to) them. My preference is to stress the way control systems
directly observe and control the outcome, not the means by which
the outcome is achieved. I don't like to characterize the
operation of a system like this in cognitive terms, because that
obscures the immediacy and automaticity of the operation of the
control loop.
This was really my objection in this exchange:
You:
The current perceptions are compared with the memories to
determine whether any form of "corrective action" is
appropriate.
I said:
This way of putting it seems, at least to me, to imply an
inappropriate amount of reasoning, and makes the control
process seem needlessly complex.
And you objected:
How you find an implication of reasoning here, puzzles me.
I would have said something like "The angle of the handlebars is
driven by the difference between the perceived vertical angle of
the bike and the specified [or desired] vertical angle." This
control system doesn't have any choice about whether to act. It
does not consider alternative forms of corrective action. The
appropriateness of any given action is neither considered nor
perceived. It makes no difference to the system whether the
reference signal comes from memories or directly from higher
systems. Neither do past values of the uprightness signal make
any difference.
Perhaps it's just my private word-associations at fault here.
When you say that current perceptions (plural) are compared with
memories (plural), I think of an assortment of perceptions and
remembered experiences, with the comparison involving pairing
them up and making multiple judgments about them. When you say
that the comparison is used to determine whether any form of
corrective action is appropriate, I take "determine" to imply
some sort of reasoning process, which could come out in several
ways. The reference to "any form of corrective action" seems to
imply the existence of several possible forms, and a
consideration of corrective versus non-corrective actions (as if
different possibilities for action exist, some corrective and
some not, so that one must choose among them on some unstated
basis). The term "appropriate" implies to me other considerations
beside the immediate error, with the possibility that sometimes a
corrective action might not be thought appropriate even if there
is an error. Your description seems to be better suited to a
cognitive situation, as in saying that you wanted to correct a
perceived error in the boss's grammar, but thought that doing so,
even if you could think of an effective way, might not be
appropriate.
My own preference is to fit the description of the control
process to the level of abstraction, and the degree of awareness,
that is probably involved.
You say
I am simply trying to describe, in summary, the sequence of
events as I think they occur.
Again, the language throws me. I see no "events" in keeping a
bicycle upright, and surely to analyze a control system as a
sequence of events is incorrect! The sequential analysis implies
that only one part of the system is active at a given moment, and
obscures the continuous and simultaneous character of the
process.
You say
Current perceptions, as I visualize the situation, bring
associated memories to mind. These memories are also current
perceptions, but not from current feedback sources. These
memories include those required for operation of the bike.
They are being selected and applied to the changing situation
as needed.
And while this is occuring, what is the error signal doing? What
output is occurring? What effect is the output having on the
current perceptions? The way you're describing keeping the
bicycle upright, you make it sound as if first the rider
perceives the current uprightness, which calls memories of former
situations to mind, which leads to selection of an action which
has been effective in the past for keeping the bicycle upright,
which finally leads to doing something to the bicycle. Then we
move forward in time another twentieth of a second, and the whole
cycle repeats.
This description really discards the simple mode of operation of
a continuous control system in favor of a far more complex
process than, in my opinion, is necessary for this situation.
Memory, instead of simply supplying an image of the desired state
of affairs, is used as a a source of information about what
outputs to produce, thus making the comparator and output
function superfluous. This one control system is now taking into
consideration not just the variable under control, the
uprightness of the bike, but the entire surrounding situation. I
just feel that you're trying to pack into a single control
process operations that are better modeled as many control
systems operating in parallel and many levels of control. Or at
least your description makes it seem that way to me.
Perhaps I'm the only one who sees any problems in this mode of
description. But I hope you'll agree that presenting a continuous
control process as a sequence of events is playing into the hands
of the opposition!
···
---------------------------------------
My initial suggestion for Sixth Order was (921205): Control of
Interpersonal Relationships. "Modes of Sixth Order" was
suggested for grouping topics being communicated. The Orders
of the Hierarchy provide a basis for such grouping.
I don't see control of interpersonal relationships as a basic
level of control. We control many kinds of relationships, some of
them being interpersonal. I think it's good to point out
relationships as a distinct class of perceptions, but why limit
this class to apply only to interpersonal situations? In driving
a car, isn't one controlling a relationship between car and road?
Isn't a conductor trying to control a relationship among sounds
made by different instruments? And is this fundamentally
different from a pianist controlling the relationship between
parallel melodic lines? Relationship control is a very broad
category, one that's easy to take for granted as a given of the
environment, so I definitely think it has to be recognized as
indicating a type of perception. But why limit the category to a
single application?
---------------------------------------
Finally:
It is my purpose, again, to describe the sequence of events as
it appears to the rider. Explanation? Perhaps, but not the
point at this time. Logic? Perhaps, in some sense, but not
the point at this time.
We have to explain how it is that the bicycle can be kept upright
without attention from the rider. I think we should eliminate all
unnecessary complexities from the explanation until something
absolutely forces us to include them. The inclusion of memory
association as an element in keeping a bicycle upright does not
seem forced on us, so I would eliminate it from a model for that
particular task.
And again, I hope we can avoid referring to a control process as
a sequence of events unless it is SPECIFICALLY a matter of
controlling sequence (ordering in time), and unless the elements
of the sequence are SPECIFICALLY events (brief packages of
transitions, configurations, etc., having a demarcated beginning,
middle, and end). Control systems themselves should never be
spoken of as if their elements took turns acting. That's one of
the models we're trying to replace.
This is enough for one post. I do approve strongly of your
objective of working from the existing vocabulary of an audience
toward a clearer understanding of PCT. I don't have the patience
to do it, but I can see that it is the best way for those who
have this enviable social skill.
-------------------------------------------------------------
Best,
Bill P.
-----------------------------------------------------------