[From Bob Clark (930411.0345 EDT)]
RE: IT/PCT Discussion: Observer Viewpoint
The following is from the Viewpoint of one Observing the Discussion.
Is this IT/PCT Discussion important? Certainly, from at least three
standpoints: 1) If we can't agree on our language, our communications
become unreliable, at best; 2) The importance of recognizing how differing
"Viewpoints" affect perceptions; 3) The interactions provide an interesting
example of complex feedback systems in the process of constructing
Problem-solving Programs, without any apparent Intrinsic Error Signals
showing up anywhere!
1) Language.
To an Observer it was soon apparent that a single word was being used with
(at least) two significantly different meanings. Great effort was expended
in the attempt to bring these meanings into some kind of agreement. This
effort eventually took the form of a vote on a secondary question --
summarized as: "Is there any "information in" the Disturbance of an
Environmental Variable as Perceived by a Control System?"
After many challenges, logical analyses, and other interactions, Powers
[Bill Powers (930408.0700)] looked up the Information Theory entry in his
Britannica (I had looked in mine also, but my version was too old (1949)
and there was no entry!) He found that a suitable way to apply IT would be
to regard the Disturbance as the "message," with incidental unrelated items
as the "noise."
Since IT says nothing about the content of the message, it "contains" no
information relevant to the operation of the Control System in question.
Thus it seems to me that "Information" in IT pertains only to the
characteristics of the transmission line. If some special "Mystery Box" is
added (by an "Engineer") to provide some form of decoding, this ceases to
be a single elementary feedback control system and becomes a multi-level
system.
On the other hand, for the Control System, a signal does not become
"Information" until it is related to other Perceptions available to the
System. This seems more apparent if a higher order system is considered,
where its Feedback Signal consists of a combination of lower order Signals.
Thus, a disturbance in one of the lower order inputs can affect the higher
order system in a unique way -- possibly even forming a feedback signal for
a different System! To become recognized as "Information," comparison with
related, remembered events would be necessary. In some situations, one or
other of existing Problem-Solving Programs might also be activated. If
this happens, "Information" may become perceived as "Knowledge."
2) Differing Viewpoints.
Most of the discussion seems to be based on looking at the "innards" --
details of a single system and its reporting of the condition of a single
environmental variable. This is essentially the Viewpoint of the Engineer
(the User and the External Observer do not have access to the internal
workings of a single system.) The Engineer is, properly, concerned with the
reliability and consistency of the transmission line -- with no interest at
all in the content of the message being carried. From this standpoint, the
Feedback Signal "contains" all the "information" he (the Engineer) cares
about. On the other hand, the User is concerned primarily with the message
(the Perceptual Signal), other matters are regarded as "Noise." Ordinarily
this Noise can be ignored, but if it is excessive, the User brings into
play some of his other Systems to improve the accuracy and reliability of
the communication systems.
3) Problem-solving Programs.
We have here a demonstration of several people (arrays of complex
hierarchical control systems) finding their communications are not received
as intended. Thus, whatever higher order systems are setting the reference
levels resulting in the rather complex procedures needed to produce these
postings, the perceptions do not match the expectations -- error signals
result, producing further and more extensive efforts to obtain some form of
acceptance.
As these efforts are continued unsuccessfully, they become more complex and
cover a larger range of activities.
During these activities, there seems to be neither a threat of Intrinsic
Error, nor any actual Intrinsic Error developing. That is, the efforts
seem well organized, pertinent and pretty much to the point. There seems
to be little truly random activity -- as expected if Intrinsic Error is
involved.
Those Problem-solving Programs brought into use are clearly unsuccessful,
although separately well organized. They simply seem to be insufficiently
related to the problem at hand. Indeed, the real problem has rarely been
stated at all, namely: "What, exactly, do we mean by "Information?" The
wrong problem has been attacked!
NOTE: A few participants did point out a language problem, and there was
some discussion of the meaning of the word, but it did not become the
primary focus until late in the discussion.
We see here a rather detailed and extensive display of a group of
individuals engaged in a non-threatening decision making activity. There
were multiple explorations of memories (including reference documents) and
experimental efforts to resolve the situation. From these activities,
additional memories of various kinds at various levels were acquired. In
addition, some memories of procedures were formed, available for later use
in the form of new Problem-solving Programs. For convenience, one such
Program could be given a name: "Look for a trouble-making word!" Or pick
your own label.
I hope you find these comments interesting and "useful" (whatever that
means!) -- Regards, Bob Clark