What's positive?

Tom Bourbon [950208.1615]

Several people replied to, or alluded to, my remarks directed at a post
by Lars-Christian Smith, about purported instances of positive feedback.



[From: Bruce Nevin (Tue 93027 16:41:29 EST)]
responsibility and intrinsic wants

(Tom Bourbon [950206.1435]) --

What I *think* I see here is an example in which manufacturers altered
their output (the nature of their products) until the products matched
the intended perceptions of their customers, a state of affairs that
probably allowed the manufacturers' perceptions of their profit margins
come closer to the margins they intended to perceive.

Bruce N:
Tom, the following is not intended to contradict anything essential about
what you are saying. [TB: material deleted]

Tom, now:

There was no contradiction, Bruce. I was talking about feedback control
systems. You wrote about different topics: the history of competition
between manufacturers of Betamax and VHS systems, and some interesting
ideas about the history of clocks. I am still trying to understand why
several people think those are examples of "runaway positive feedback." So
far, I haven't a clue what all of you are talking about.

[Martin Taylor 950207 18:10]

Tom Bourbon [950207.1147]

Re: Intro & Positive FeedBack ???


(I agree with nearly all
of what Bill said, and little of what Martin said, but that's no surprise,
either :wink:

I haven't seen anything of what Bill said, but what is it that you disgree
with about what I said? It seemed uncontroversial, from a PCT point of
view, to me.

Tom, now:
Martin, the important part of my statement was the following: :wink:
But now that you mention it:

Are you saying that that it is NOT possible for two (or
more) interacting negative control systems to generate a positive feedback

Tom, now:
Not at all. Often deliberately, but more often stupidly, I have created
positive feedback interactions between pairs of PCT models. However, in the
examples presented so far on the net, I have seen nothing that looks, to me,
like the kind of explosive runaway process I always see with my models.
Were there some examples on the net that were clear to everyone else and I
just didn't comprehend them? That certainly is possible.

[Lars Christian Smith 950208 12:30 CET]
Re: Intro & Positive FeedBack ???
To: Tom Bourbon


The example with driving on the right side of the road is an example of
spontaneous order. Such order may be the outcome of a positive feedback

Fine, but I don't know what you mean when you say this *is* an example of
"spontaneous order" and *may be* the "outcome of a positive feedback
process." How is this an example of spontaneous order -- what is one
of those, in general, and why do you think this is an example of one, in
particular? I don't know what you mean --- a failure on my part. Once you
have explained those two ideas, then I would like to see why you think a
positive feedback process might have created the order. In my experiences
with positive feedback, whether within a single control loop or in an
interaction between two control systems (I work on a very small scale), the
controlled variable(s) *rapidly* went out of bounds -- in essence, the
system(s) "blew up." I am not denying the veracity of your examples, but I
would like to understand how positive feedback can lead to results so
radically different from the ones I have always seen.


It is, of course, true that underlying this process there is a
negative feedback process, people are avoiding accidents.


Let us assume
there was no rule, and driving was chaotic. Individuals driving left,
center or right on the road will avoid hitting other vehicles, which
means driving on the side of the road where most of the others are
driving in the same direction.

On the assumption that all of the drivers are perceptual controllers with
references to get on down the road and to avoid accidents while on the way,
I can imagine such a scenario. In what ways is this an example either of
"a positive feedback process," or of "spontaneous order?" It all looks like
negative-feedback perceptual control, to me.

More people driving on one side of the
road would lead to more people driving on the same side. A rule would
quickly become established.

Well . . .. OK. But see my questions and comments immediately above.

Do you have a disagreement with the characterization of this process as a
flip-flop positive feedback process?

I am in no position to know if I agree or disagree. I don't know what there
is about the example that you think can be characterized as, "a flip-flop
positive feedback process." Whatever you think it is, and I would like to
know, all *I* see is actions in a common space (the road) by a large number
of independent negative-feedback perceptual controllers. All of the rest
looks, to me, like side effects of their individual control.

About your examples of the Betamax-VHS Wars, and the Windows-Apple Wars,
see my remarks to Bruce Nevin, above.

From Bruce Buchanan (950208.1300 EST)

Re: Positive Feedback

Bruce Buchanan:

I would like to
provide a description of the roles of negative and positive feedback from
the book Thinking by Machine: A Study of Cybernetics, by Pierre de Latil,
foreward by Isaac Asimov (1957), which has influenced my thinking greatly
for nearly 40 years. The following is all direct quotation.

re "Constancy Effectors and Tendency Effectors: (p. 89:) ...

[TB: quoted material omitted here]

Bruce, this must not be my day. As was the case in my replies above, I just
don't get it. The first kind of device you mentioned certainly looks like a
negative-feedback controller; I am not certain I would characterize the
second kind as a controller at all. A lamp? How is a lamp an example of a
controller run amock -- a positove-feedback controller? I can think of
certain lamps as *neg-fb* controllers, if they are built like the
ingeniously contrived devices from long ago -- the "lamps of God" that never
ran out of oil and always had a fresh, new wick.

Can you describe what it is about the behavior of a less-ingenious modern
lamp that might lead you to call it a positive feedback controller? Once
again, my understanding must be clouded by apparently limited experiences
with positive feedback in models of perceptual control systems.

What does Marc Abrams think about all of this? Marc, didn't you say a few
days ago that you had seen no discussion on csg-l about positive feedback?
Do the events which followed that remark qualify as an example of a positive
feedback process that might lead to spontaneous order? :wink:



[From: Bruce Nevin (Thu 93029 12:47:20 EST)]

( Tom Bourbon [950208.1615] ) --

I am still trying to understand why
several people think those are examples of "runaway positive feedback." So
far, I haven't a clue what all of you are talking about.

Tom, I think the claims rest on the notion that what we model as the
feedback loop through the environment may itself contain (participate in)
feedback loops within the environment, and that some of the loops in the
environment may, while they last, involve positive feedback. Gregory
Bateson made a case for what he called cybernetic structure in the
environment of organisms, notably in _Mind in Nature: A necessary unity_.