The fierce and terrible Marken gives feedback

[From Rick Marken (960724.1100)]

Bill Powers (960724.0915 MDT) --

Feedback is the effect of a variable on ITSELF.

Rick Marken (960724.0930) --

"Feedback" that is "given" (whether it is wanted or not) is just a
_disturbance_ from a PCT percspective.

Bruce Gregory (960724.1310 EDT) --

Seen with the eye (I) of faith, I _know_ that Marken and Powers
are saying the same thing. It only _seems_ they are differing.

Yes. I was referring to "feedback" as that term is often used in informal
conversation; as something that is "given" (like advice). In PCT, we use the
word "feedback" to refer to only one thing; the effect of a variable on
itself.

The intent of my post was simply to say that _unsolicited_ "feedback" (as
that term is conventionally used) is what we call a "disturbance" in PCT.

It _seems_ to me that feedback of the form "you're getting warmer" or
"you're getting colder" enhances, rather than detracts from our ability to
control.

If you are controlling for hearing "you're getting warmer" then these phrases
are states of a controlled perception; they are not "feedback" (in the PCT
sense) and they are not disturbances (if they are not independent of you
actions -- and they aren't independent of your actions if someone has agreed
to say "warmer" or "colder" based on your actions). These phrases don't
really enhance control; they are what IS controlled. You are trying to move
so that you keep hearing "you're getting warmer" until finally you hear "you
got it!"

My point in the previous post was that _unsolicited_ feedback (as in "let me
give you a little feedback about your presentation") is likely to be a
disturbance to a controlled variable (like your self esteem); it is a
disturbance even if the person says they LOVED your presentation; it is a
disturbance because it was unsolicited; it was independent of your actions.
If I had wanted to know how you felt about my presentation, then I would ask.
Then the "feedback" you give me about my presentation (GOOD OR BAD) is NOT a
disturbance (independent of my actions) becuase it is not unsolicited;
instead, it is an aspect of a perception I am trying to control -- my
perception of being informed about how you feel about my presentation.

Best

Rick

[From Bruce Gregory (960724.1445 EDT)]

(F.& T. Marken 960724.1100)

My point in the previous post was that _unsolicited_ feedback (as in "let me
give you a little feedback about your presentation") is likely to be a
disturbance to a controlled variable (like your self esteem); it is a
disturbance even if the person says they LOVED your presentation; it is a
disturbance because it was unsolicited; it was independent of your actions.
If I had wanted to know how you felt about my presentation, then I would ask.
Then the "feedback" you give me about my presentation (GOOD OR BAD) is NOT a
disturbance (independent of my actions) because it is not unsolicited;
instead, it is an aspect of a perception I am trying to control -- my
perception of being informed about how you feel about my presentation.

O.K. Very clear. I realize that the most useful "feedback"
anyone can give me about a presentation is not how good (or
bad) it was, but what they learned, or think they learned, from
it. You can tell what _I'm_ controlling for...

Regards,

Bruce

[from Jeff Vancouver 960725.13:00]

Rick Marken (960724.1100)]

&

Bill Powers (960724.0915 MDT) --

&

Rick Marken (960724.0930) --

>"Feedback" that is "given" (whether it is wanted or not) is just a
>_disturbance_ from a PCT percspective.

Although I would agree that much confusion can be attributed to the
multiple meanings of feedback. This last statement from Rick seems to me
not totally correct. I think the problem can be found in Bill P.'s
response. He suggests that solicited "feedback" (I will use quotes for
the non-technical definition of feedback), when received, corrects the
error in the not-knowing unit. In this sense, is it a disturbance or, in
fact, feedback for the not-knowing system, whose output was to ask for
"feedback"?

Moreover, the not-knowing unit was a subunit of the unit controlling for
the perception of the content of the feedback (e.g., how you liked my
paper). The information in the "feedback" will be used to form a
perception of liking, which the controller is presumably interested in.
In this loop, the "feedback" is, in fact, feedback (also perhaps with some
disturbance added in). That is the information in the "feedback"
indicates the current state of the variable. That current state is likely
a function of the actions of the focal person and disturbances. (It is
not necessarily an accurate indication - the "feedback" giver could be
lying).

Thus, I do not see how "feedback" is just _disturbance_ except in special
cases? For example, if Dr. Nichols was attempting to achieve antagonism
in his friend, but the day he got the "feedback" indicating antagonism was
the same day that his friend had a nail in his shoe. The information in
the "feedback" might have everything to do with the nail and nothing to do
with Dr. Nichols' previous actions. But is seems that the nail should be
called the distrubance, not the "feedback."

As for unsolicited "feedback," if the information in it (which may
include just the fack that it is given) is not used to create perceptions
of controlled variables, it will have no impact. But the degree to
which that information is a function of past actions (and therefore
feedback) or disturbances is independent of it being "feedback."

I think "feedback" is very important because without it it is often very
difficult to control many kinds of variables I suspect we attempt to
control. That is, many of the variables we attempt to control involve
how others feel, think, or perceive. Imagine how difficult it would be
for the husband who just wants his wife to be happy, when he gets no
information about her happiness. On the other hand a pot thrown threw
the air if at least "feedback" ("feedback" need not be verbal).

Bottom line, I think some control theorists need to study both kinds of
feedback. I am.

Later

Jeff

<[Bill Leach (960728.2244 EDT)]

[from Jeff Vancouver 960725.13:00]

Maybe this is "just me" (but I don't think so)...

The only reason that PCTers discuss "both forms" of feedback it to point out
that the "common" use is absurd. In electronics, physics, and most
especially Control Theory the term feedback is a term with precise meaning.
The common use and indeed the so called "technical" use in the field of
psychology came about as the result of an almost complete failure to
understand the meaning of the term. The common "meaning" that everyone
"knows and understands" is devoid of meaning in any useful "qualitative"
sense much less any quantitative sense.

The scientific _meaning_ of the term feedback is value neutral (positive and
negative referring only to phase relationship between the control system
output as seen at the output and the effect of control system output as seen
at the the system input (net negative or the system is unstable if open loop
gain greater than unity). A positive feedback path transfer function will
have the same polarity sign as the input/output transfer function and a
negative feedback path transfer function has the opposite sign of the
input/output transfer function.

"Giving another feedback" means expressing your opinion of their performance
(honest or otherwise), is a value judgement and NEVER anything more.

In a sense, when another askes you for "feedback" your response NO MATTER
WHAT THAT RESPONSE MIGHT BE is indeed feedback in the true scientific sense
to the loop containing the reference for perceiving a response to the question.

However other preceptions related to the interaction are probably AWAYS more
"important" than just the fact that you responded to the question (much less
for the case where you provide "feedback" without the request).

I suggest that use of the term in the "common" sense is likely to be very
confusing to all parties concerned on the net and definately adds nothing to
understanding.

Thus, I do not see how "feedback" is just _disturbance_ except in special
cases? For example, if Dr. Nichols was attempting to achieve antagonism
in his friend, but the day he got the "feedback" indicating antagonism was
the same day that his friend had a nail in his shoe. The information in
the "feedback" might have everything to do with the nail and nothing to do
with Dr. Nichols' previous actions. But is seems that the nail should be
called the distrubance, not the "feedback."

The preceeding is, I think, a good example of what I am trying to say... you
need to know so much about what perceptions are under control by each of the
systems for a "feedback" discussion to mean anything that it seems a bit
pointless to discuss "feedback". If you know all of the perceptions why
bother talking about feedback and if you don't then you also don't know if
what you are seeing is feedback at all.

As for unsolicited "feedback," if the information in it (which may
include just the fack that it is given) is not used to create perceptions
of controlled variables, it will have no impact. But the degree to
which that information is a function of past actions (and therefore
feedback) or disturbances is independent of it being "feedback."

That is a "pretty big" IF in that first sentence even though the statement
is essentially correct (the "information" does not "create" _controlled_
perceptions)! The second part of that is a little vague to me.

"the degree to which that information is a function of past actions" is far
too vague for me to assign meaning to what you are saying with any degree of
certainty. The unsolicited feedback might disturb or otherwise alter
controlled perceptions. It might also just be perceived (no control
involved) and it might possibly not even be perceived.

Part of the problem with trying to talk about this "feedback" in terms of
feedback is that neither the transfer function nor the sign of the function
can be related to the original systems' output in any rigid or rigorous
sense. We CAN do that with our models but only because of our knowledge in
the absolute sense concerning the perceptions that are being controlled by
the models and their transfer functions (thus one model's behaviour can be
described rigorously in terms of an environmental feedback function for the
interacting model).

I think "feedback" is very important because without it it is often very
difficult to control many kinds of variables I suspect we attempt to
control.

Of course the "wise guy" remark is that control in the sense that we deal
with it here is not possible without feedback (negative of course) for the
control of anything period.

That is, many of the variables we attempt to control involve
how others feel, think, or perceive.

This is, of course true but ask why?

Imagine how difficult it would be for the husband who just wants his wife
to be happy, when he gets no information about her happiness. On the other
hand a pot thrown threw the air if at least "feedback" ("feedback" need not
be verbal).

Bottom line, I think some control theorists need to study both kinds of
feedback. I am.

Calling these things (how other feel, think, perceive or act) feedback begs
the important questions. Deciding that you wife's throwing a pot through the
air is "feedback" is a silly exercise.

All people do things (including what they say) for the purpose of
controlling one or more perceptions. What is truely important in the human
interactions (if you have a reference for a high level of "caring" for this
other human) are the references for these controlled perceptions (as well as
possibly "higher level" references) and not the behaviour itself (or more
exactingly... your perception of their behaviour).

Many practictioners have an "intuitive" feel for many of the implications of
PCT, the unfortunately do not have the theoritical base to apply either to
"check" intuition or to use when intuition seems to fail.

As has been said here many times:
         PCT does not answer the "BIG" questions of life.

However what PCT tells us in its' answers to the "little" questions is more
profound than all of the rest of the "knowledge" in the field of behavioural
science combined. PCT explains what control _is_ with respect to living beings.
Even in its' present infant state PCT provides profound insight into how
even a high level reference converts to output and the nature of what must
happen if that output is to result in control of perception (or attainment
of goals in "common speak"). PCT provids significant insights into the
nature of conflict both internal and between individual control systems.

If one "wants to study PCT" with respect to "real situations" involving
"real people" then what one really needs to do is to learn the nature of the
questions that are important to the understanding human action from the
perspective of PCT. Trying to "apply" PCT to the sort of questions that come
from other behavioural "science" is a futile waste of time except to the
extent that one is sufficienty honest to allow "PCT thinking" to show why
the question is irrelevent (unless the question does not really come out of
the "classic" theory in which case there is a remote chance that the
question is valid)!

bill leach
b.leach@worldnet.att.net
ars KB7LX