block diagrams

CSGnet Archive CSG2001
21

[From Bruce Gregory (2000.1012.0950)]

[From Bruce Gregory (2000.1011.2003)]

Rick Marken (2000.10.11.1440)

Bruce Gregory (2000.1011.1712)--

> O.K. Suppose you put a dish of Cherry Garcia in front of me,
> and I eat it. Does this tell me that there is no higher-order
> control system lowering my reference for eating Cherry Garcia
> while writing e-mail messages?

I don't think so.

Oh?

I guess I don't understand the Test, then. How do you test for the
existence of a higher level controlled perception that sets the reference
level for eating Cherry Garcia ice cream at zero and for writing e-mail to
a high level?

BG

22

[From Rick Marken (2000.10.12.0800)]

Bruce Gregory (2000.1012.0950)--

I guess I don't understand the Test, then. How do you
test for the existence of a higher level controlled
perception that sets the reference level for eating
Cherry Garcia ice cream at zero and for writing e-mail
to a high level?

You can't just Test for the "existence of a higher level
controlled perception". The Test always starts with a
_hypothesis_ about a controlled variable -- a hypothesis
that is continuously revised as various disturbances do
or don't have the predicted effect on the hypothesized
variables.

If you suspect that not eating a certain kind of ice cream and
writing e-mails at a certain clip are the particular values
of some of the actions involved in the control some higher
level perception, then you have to guess (develop a hypothesis
about) what that perception might be before you can _start_
Testing for its existence. And the Testing itself will be
an iterative process that converges on the best definition
of the controlled perception.

Off hand, I can think of a couple hypotheses about the higher
level perceptions that might be kept under control by stopping
ice cream eating and increasing e-mail writing: rate of
writing and accuracy of writing come to mind. Some ingenuity
would then be required to Test these hypothesis.

Best

Rick

···

--
Richard S. Marken Phone or Fax: 310 474-0313
MindReadings.com mailto: marken@mindreadings.com
www.mindreadings.com

23

[From Bruce Nevin (2000.1012.1236 EDT)]

Bruce Gregory (2000.1012.0950)--

Bruce Gregory (2000.1011.1712)--

> O.K. Suppose you put a dish of Cherry Garcia in front of me,
> and I eat it. Does this tell me that there is no higher-order
> control system lowering my reference for eating Cherry Garcia
> while writing e-mail messages?

[...]

How do you test for the
existence of a higher level controlled perception that sets the reference
level for eating Cherry Garcia ice cream at zero and for writing e-mail to
a high level?

I'm afraid this example leads to that can of worms labelled "resolving conflicts" and "making decisions". If you have to go away from your PC in order to get some Cherry Garcia to eat then control of eating Cherry Garcia conflicts with control of writing email. If someone puts the ice cream next to your keyboard, there is no (or little) conflict. You can write about your ice cream and eat it too.

It may be that some higher-level control system has the effect of resolving the conflict when it resets the references (I don't want Cherry Garcia after all), or orders the two in a sequence (ice cream will have to wait, or I need an ice cream break before I write that), or the like, or that some gain-adjusting mechanism has the effect of deferring the conflict when it resets the gain ("deferred gratification"?). But testing for any of these possibilities is complicated by the existence of the conflict.

Methodology for applying the Test when a controlled variable is under conflict is so far as I know unspecified. And in general, when we test for controlled variables, how do we know that conflict is not involved? How can we distinguish the following two cases:

1. Subject is controlling X at value m.
2. Subject is controlling X at value a, and is also controlling X at value z, and the outcome is that the subject appears to be controlling X at the value m.

There may not be a better example for your purpose:

Bruce Gregory (2000.1011.1402)--

> The hierarchical control model, as far as I can tell, does not
> tell why I am controlling one perception rather than another.

Looks like the above can of worms to me. Or ice cream on your keyboard.

         Bruce Nevin

···

At 09:50 AM 10/12/2000 -0400, Bruce Gregory wrote:

24

[From Bill Powers (2000.10.12.1434 MDT)]

Found 39 messages waiting when I got on line at my daughter's house here in
Boulder Co. I found a very easy way to reply to most of them: as soon as I
came across anything sarcastic or a complaint about somebody or a defense,
or another random tennis serve, I dumped the post into the "printout"
mailbox for the archives, and went on to the next one. That brought the
number down to 7 which look interesting or need some sort of reply from me
(or both).

I'll get to them in due time. This is just to let those whose posts fall
into one or more of the above categories not to wait around for my reply or
comment. This is the only one there will be.

Best,

Bill P.

25

[From Bruce Gregory (2000.1012.1815)]

Bill Powers (2000.10.12.1434 MDT)

Found 39 messages waiting when I got on line at my daughter's house here in
Boulder Co. I found a very easy way to reply to most of them: as soon as I
came across anything sarcastic or a complaint about somebody or a defense,
or another random tennis serve, I dumped the post into the "printout"
mailbox for the archives, and went on to the next one. That brought the
number down to 7 which look interesting or need some sort of reply from me
(or both).

I'll get to them in due time. This is just to let those whose posts fall
into one or more of the above categories not to wait around for my reply or
comment. This is the only one there will be.

Is this post supposed to be informative? Are we supposed to put ourselves
into the appropriate category? I don't think I posted anything to you, but
if I did, please ignore it.

BG

26

[From Bruce Gregory (2000.1013.1423)]

Bruce Nevin (2000.1012.1236 EDT)

Bruce Gregory (2000.1011.1402)--

> The hierarchical control model, as far as I can tell, does not
> tell why I am controlling one perception rather than another.

Looks like the above can of worms to me. Or ice cream on your keyboard.

Indeed. In my understanding you can always construct a tentative HPCT model
to explain why and how you are controlling a particular perception. That
model is not always easy to test, however. Am I really typing this message
in order to convince you that I understand HPCT? That certainly is one
model. You could try posting me with, "Gee, Bruce you _really_ understand
HPCT," and see if go to the refrigerator for a bowl of Cherry Garcia.

BG