NetNews Usenet Archive 1993 #1/NN_1993_1.iso/spool/bit/listserv/csgl/2092

Text File | 1993-01-08 | 8.0 KB | 159 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!usc!howland.reston.ans.net!paladin.american.edu!auvm!MCIMAIL.COM!0004912499
Return-Path: <@VMD.CSO.UIUC.EDU,@pucc.Princeton.EDU:[e-mail]>
Message-ID: <12930108165021/[e-mail]>
Newsgroups: bit.listserv.csg-l
Date:         Fri, 8 Jan 1993 16:50:00 GMT
Sender:       "Control Systems Group Network (CSGnet)" <[e-mail]>
Comments:     <Parser> W: TO field duplicated. Last occurrence was retained.
From:         "Robert K. Clark" <[e-mail]>
Subject:      RESPONSES RE:DME FROM RKC
Lines: 146

January 8, 1993

TO: Greg Williams, Allan Randall, Rick Marken

FROM: Bob Clark

Thanks for the "WELCOME," Greg.

Just after sending my post of Dec 5, I became involved in working out a
problem between an insurance client and the Company.  This involved a
combination of three different rates of interest, the effects of
compounding, and interpretations by the people involved.

Then came the Holidays.  Had a good visit with my son and family in
Dearborn where he is an engineer for Ford.

When I looked at my INBOX, there were 228 messages!  After a superficial
review, only about 20 remained for discussion.  I will respond to them more
or less in sequence, but it will take time.  And they are continuing to
come in!

PROBLEM: In the process of reviewing and down-loading these posts, I find I
have somehow mislaid one from Powers.  It was posted between Dec 30 and Jan
4, I think, and included remarks about some of his activities before
retirement.  This post also included other items of interest.  Is there
some way I can get that Post?

QUESTION: I see that posts commonly list about 4 addresses in addition to
CSG-L.  Is this necessary?  This seems to increase the probability of
errors in addressing.  No doubt this could be automated, but I don't have
that worked out just yet.

Greg, as usual, it is much easier to raise questions than to answer them.
This is to be expected, since the questioner regards the world from his own
viewpoint combined with his available store of ideas.

You ask (Williams 92107): "Why might the DME direct attention to certain
memories, rather than others, at some particular time?  Do you have a
theory of attention "selection" other than the broad viewpoint that the DME
tries to (as you say later) "improve his well-being"?  Is there some
calculus for tradeoffs among various possible way to "improve" (more or
less)?

You suggest: "running "imagination connection" trials ... to "examine" and
"select" some of them for actual performance."

Allan Randall (921207) comments on Bill's Chapter 15 model of memory.

Also Rick Marken (921208) prefers "to look at decisions as the conscious
result of conflict -- reorganization" ...  "and tolerate the error
resulting from not doing the other" ...  he notes "A better way to solve
such conflicts is to "go up a level""  Rick further seems to accept the view
that decision making "is an inherently statistical phenomenon."

In response to these questions, comments, and remarks, let me point out
that I am concerned with the process of decision making.  Statistically,
the only common element among people is that they all make decisions.
Methods, reasoning, procedures, etc can differ drastically from one person
to another.

To me, decision making is a peculiarly individual matter.  For this phrase
to have meaning, there must be at least two alternatives available.  This
implies at least a minimal conflict in that they cannot both be selected.
The alternatives need not be particularly important (although they could
be).  There must be some way in which they can be examined.  There must be
some basis for selection.  And there must be some entity capable of putting
all this together.

We already have all these elements except the Decision Making Entity --
which is implied in Powers Chapter 15 since something somewhere must
operate the diagrammed switches.  In fact, my view of the relation among
the DME, current perceptions and memory(ies) might be regarded as an
extension of the concept Bill illustrated with two single pole double-throw
switches.  Using this diagram, these switches are controlled by the DME --
one of its major functions.

In addition, the Recording Function (would this term be better than
"Memory?"), can be considered a multi-dimensional recorder, including not
only all perceptual signals, but also all consciously imagined combinations
and projected conclusions, and decisions.  When examined (imagination), the
memories are much like multi-dimensional video tapes.  These memories are
not necessarily logically related, nor otherwise coherent.  They may arise
simply through some accidental event that provides some connection (a
"reminder") to the specific remembered event.  It could be an odor, a face,
a sound, an idea, a word, etc.  Or it could be a problem ("conflict?") with
recognizable aspects bringing related memories to mind.

The Recording Function is mostly undirected, but the DME can make it more
easily available by consciously assigning some kind of labels to suitable
recordings.  How do you learn the name of someone you have just met?
However many "labels" are acquired more or less accidentally.  Thus the
word "chocolate" easily brings an image (images?) to mind.  But there are
many forms of labels: the appearance of a house, a date, a period of time,
and so on.  There are many ways to locate specific memories.

Most memories are inactive most of the time.  (What a confusion otherwise!)
They appear to be "forgotten" until some form of "reminder" occurs.  (I
have been intrigued by the Questions that "pop" into mind in response to
JEOPARDY's Answers.)

A great many of the decisions needed are simple, requiring very little
attention or analysis.  A very rapid (perhaps on the order of a few
milliseconds) switching may occur between alternative memories when little
analysis is needed and anticipated conclusions are quickly formed.  These
alternatives are selected for their relevance, often simply by being
"reminded" of similar situations.  But the DME may find more thorough
investigation necessary in seeking a satisfactory selection.  This tends to
be related to the Level in the Hierarchy involved -- as implied by Rick
Marken's remark about "going up a level." Thus, in ordinary situations, the
DME can make its selection quickly.

In my remarks above, I noted that the Recording Function records
conclusions and decisions.  These are more likely to pertain to higher
level situations, where a previously formed conclusion/decision can provide
a quicker response.  "I'll push the button the instant I perceive a flash
of light," rather than, "There's a flash of light, what do I do now?" Or,
"The moment the light turns green, I'll hit the throttle," results in a
much faster response.

The DME's attention is directed by the need to select among alternatives.
The characteristics of the situation serve to remind the Memory which
recording to present.

It is interesting to observe that the DME cannot direct its attention to
its own acts as it is performing them.  Its only information about its own
activity (self-knowledge?) is through examining the relevant memories.
These are not necessarily readily available.

By and large, these are mostly ordinary and familiar observations, but they
seem to have been left out of much behavioral discussion.

Greg, you note that, "It appears to me that the DME is basically directed
(not completely random) reorganization." In its origins, the Reorganizing
Function was proposed to explain the observation that individuals change
their behavior when faced with a conflict -- especially if it is hard to
resolve.  However, that is an "outsider's" viewpoint.  To the DME involved,
it is not intrinsically different from any decision-making situation.
Available alternatives (including, perhaps, violent movements, or whatever)
are reviewed and the behavior offering the most promising anticipated
results is put into operation.  To the outsider, who may not even suspect
the alternatives available, this will tend to appear "random," that is,
"unpredictable."

Regards, and more later as I try to get caught up.

Bob