Answers to PCT Questions

CSGnet Archive CSG2003
1

[From Fred Nickols (2003.10.14.0915 EDT)] --

As some of you might recall, a while back I wrote a simplified overview of
PCT which I checked with the list for accuracy. I subsequently added to
that overview and created a paper titled "A Control Theory View of Human
Performance" which I put up on my web site for review and comment. A
long-time human performance technology specialist has read the paper and
asked me a few questions. His questions and my answers are reproduced
below. I'm interested in knowing if I've fully and correctly answered his
questions.

Thanks for the thoughtful reply, John. Responses follow...

I've looked at this article more than once and did so a bit again
tonight. It's a nice piece of work.

Thanks for the compliment.

Let me just share a few thoughts that it clicked for me.

1. I not read Powers but if the control a performer has is mostly of his
perceptions how does this eventually impinge on the external world.

As I understand Powers, his argument isn't that we don't impinge the
external world but, rather, that all we know of it we know through our
perceptions. Thus, although our behavior does indeed impinge the external
world, all we know of our behavior, the world about it and the effects our
behavior is having, we know through our perceptions. Consequently,
although we think we're controlling the external world (and well we may
be) we are in fact controlling our perceptions.

2. Is it possible that the theory is rather optimistic about a
performer's ability to obtain the information that a given desired
performance might require?

I don't think so. Perhaps I'm to blame here. Powers' theory focuses on
behavior, not performance per se. As far as the information (i.e.,
perceptual input) needed for behavior to control perception of things like
walking, talking, picking up a glass of water to take a drink, etc.,
there's probably no need to be concerned about adequate information. It's
when the condition the performer seeks to control (and relies on
perception to control) is removed in space and time that information flaws
come into play. This is also true when the condition one seeks to control
(i.e., the reference condition) pertains to matters that are not directly
observable (e.g., if I'm controlling for you liking me I might never
really know how you feel about me and so I make my judgment based on bits
and pieces over time; ditto for lots of results or accomplishments we seek
to attain). So, I agree that we need to be concerned about a performer's
ability to obtain the information that a given desired performance might
require, but I don't think Powers' theory is optimistic or pessimistic in
that regard.

3. I like your "prefigured" vs "configured" dichotomy but am not
convinced that much work in organizations below management is work in
which a "configured" approach is desired.

I probably think there's more of it than you do, John. A CSR in a call
center can required to hew closely to a script in some cases but exercise
great imagination and ingenuity in others. An electrician might well
follow some basic routines in installing ceiling fans (several of which
I've had installed lately) but have to configure that routine to meet the
circumstances at hand. Researchers, consultants, technical writers,
training developers and even assembly line workers can be spotted
configuring their response to the situation at hand. Sales people in
particular configure their work routines. I think all jobs, even those of
a manager or senior executive, have a mix of prefigured and configured
work. Execs and managers, for instance, don't have a lot more latitude
than their underlings when it comes to filling out certain required
forms. For me, the important issues are things like the following: Which
is the dominant form of work for this person? When and under what
conditions is a configured response called for? When should a prefigured
response be required/imposed?

4. I accept the distinction between "behavior" and "performance" and
agree that it is probably better to try to support performance
achievement that to attempt to control behavior. (Even Tom Gilbert, who
still uses the word "engineering" in his title, devoted a rather
passionate chapter critiquing an exclusive focus on behavior. (And
Henry Ford, the old bastard, famously said that he "never told anyone
how to so something, only what he wanted done." He said "he was
continually amazed at how ingenious people would be about accomplishing
a given objective. BUT there are still many, many situations in which
management is not only concerned about accomplishing give goals but
insist that they must be accomplished in given ways. Some of these
situation are rooted in legalities like "due process" in the claims
adjudication business, or in "disparate treatment" in the case of a too
creative approach to personnel decisions. But some of it is also
sourced in discovering that "instinctive" approaches to give things are
often mistaken. For example, one I have used before, many claims
examiners at first think they are being efficient by asking doctors for
ALL the medical evidence or COMPREHENSIVE medical reports. But it has
been found that it is far more efficacious to identify only the current
issues in a case and then ask precise questions of the doctor to develop
only the particular medical evidence in fact needed. The investigator
jobs I worked with were full of such things as well. This leads me to
the question of whether your treatment accentuates unduly the
desire/need of organizations for employee creativity and/or the latitude
they can afford to support with regard to mean needed to achieve given ends.

I agree with what you say above, John. (By the way, I, too, have
considerable experience with the claims adjudication process and the work
of claims examiners. Perhaps we can swap war/horror stories some
day.) That said, I go back to my earlier comment: It's not a matter of
pushing for prefigured or configured routines, it's a matter of being able
to determine when and under what conditions each is appropriate and
preferred. I'm not advocating that we turn employees loose any more than
I'm advocating we rein them all in and put them in harness.

I think means are still very often part of what management is either
insisting on controlling or that they are convinced that they must
control. I agree that that is dysfunctional but it seems to me that the
alternative view does not recognize adequately the brackets the real
world still does or must place on employee creativity.

I agree. See my comments above. Looks like I need to spend some time
clarifying what I'm advocating regarding prefigured and configured routines.

Regards,

Fred Nickols
nickols@att.net
www.nickols.us

2

[From Bill Powers (2003.10.14.1207 MDT)]

Fred Nickols (2003.10.14.0915 EDT)-

A long-time human performance technology specialist has read the paper and
asked me a few questions. His questions and my answers are reproduced
below. I'm interested in knowing if I've fully and correctly answered his
questions.

By and large, yes. However, it could be made clearer that whether you're
concerned with controlling behavior or performance, and whether the
performance is prefigured or configured, it's still perception one must
control. If someone tells you to perform by moving your hands just so, you
have to remember how you are supposed to move them, perceive how you _are_
moving them, and adjust your outputs to keep the perception in a match with
the (remembered) reference condition. Of course this sort of instruction
leaves far less latitude for coping with disturbances than if the
instruction were simply a description of the end-result to be achieved,
leaving you free to move your hands (or your feet) in whatever way is
needed right now to achieve the result. But in all cases, the worker can
control only what is perceived, and never knows precisely what effects he
or she is having on the environment in the course of exerting this control.

Best,

Bill P.

3

[From Rick Marken (2003.10.15.0930)]

Fred Nickols (2003.10.14.0915 EDT)--

>> Henry Ford, the old bastard, famously said that he "never told anyone
>> how to so something, only what he wanted done."

>It's not a matter of pushing for prefigured or configured routines, it's a
matter of
>being able to determine when and under what conditions each is appropriate and
>preferred.

I think the hierarchical model of control suggests that whenever we say what is to
be done (what result is to be produced) we are, necessarily, saying how to do it
(what what actions to take). That's because "what is to be done" is "how to do
it" relative to a higher level specification for what os to be done. For example,
when Henry Ford says that what he wants done is "insertion of tab A into slot B"
he is not saying _how_ that insertion process should be done (in terms of detailed
actions used to produce that result) but he is saying how he wants a higher level
result produced. That higher level result may, for example, be building a
headlamp. With respect to producing that higher level result, Henry is saying that
it be done by, among other things, inserting tab A into B rather than by inserting
tab A into C, say.

So I think Henry Ford is not correctly describing the choice a manager has. It is
not that a manager can chose between telling an employee how to do something (the
actions that should be taken to produce a result) and telling the employee what to
get done (the perceptual result to be produced). The manager can't help but do
_both_ at the same time. The choice the manager has, I think, has to do with the
_level_ at which he or she specifies what is to get done. I think what the manager
should do is try to specify what is to get done at as high a level in the
perceptual hierarchy as is feasible. Since the hierarchical model is pretty
sketchy at the moment, doing this is currently far more art than science. But the
science suggests that specifying results at at as high a level as possible is good
because, the higher the level at which one specifies the goal result, the more
lower level degrees of freedom are available for producing the result in a
disturbance prone environment. Lower level degrees of freedom represent
"flexibility" ? the ability to act to protect lower level perceptions from
environmental disturbance on the many dimensions on which they can occur.

I think the idea of "specifying results at as high a level as possible" is what
lies behind Henry Ford's idea that it is better to tell an employee what you want
done rather than how you want it done. I think how high the level is at which you
can specify results depends, to a large extent, on how much the employee can be
assumed to know about _how_ to produce the higher level result. I think the
problem in factories has been that workers have been assumed to know very little
about how to produce higher level results so they have been asked to produce low
level results, which is like specifying "how" to get things done. This is surely
bad for morale when workers are perfectly capable of seeing how their actions
contribute to the production of a higher level result. I think the Japanese have
handled this better by actually listening to their employees and realizing that
the employees often know a lot more about how to produce higher level results than
had been assumed.

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

4

[From Fred Nickols (2003.10.15.1230 EDT)] --

Bill Powers (2003.10.14.1207 MDT)]

Fred Nickols (2003.10.14.0915 EDT)-

> A long-time human performance technology specialist has read the paper and
>asked me a few questions. His questions and my answers are reproduced
>below. I'm interested in knowing if I've fully and correctly answered his
>questions.

By and large, yes. However, it could be made clearer that whether you're
concerned with controlling behavior or performance, and whether the
performance is prefigured or configured, it's still perception one must
control. If someone tells you to perform by moving your hands just so, you
have to remember how you are supposed to move them, perceive how you _are_
moving them, and adjust your outputs to keep the perception in a match with
the (remembered) reference condition. Of course this sort of instruction
leaves far less latitude for coping with disturbances than if the
instruction were simply a description of the end-result to be achieved,
leaving you free to move your hands (or your feet) in whatever way is
needed right now to achieve the result. But in all cases, the worker can
control only what is perceived, and never knows precisely what effects he
or she is having on the environment in the course of exerting this control.

Thanks for taking the time to respond, Bill.

Your closing remark (But in all cases....) prompts a question.

If I can perceive how I'm moving my hands and comparing that with the "just
so" picture in memory, why can't I also perceive the effects that my
movements are having on my environment? This seems especially crucial if
the performance in question is defined in terms of effects upon the
environment. Then again, maybe I'm misreading your remark.

Regards,

Fred Nickols
nickols@att.net
www.nickols.us

5

[From Bill Powers (2003.10.15.1101 MDT)]

Rick Marken (2003.10.15.0930)--

I think the idea of "specifying results at as high a level as possible" is
what
lies behind Henry Ford's idea that it is better to tell an employee what
you want
done rather than how you want it done.

Great reply, Rick. Better thought out than mine.

Best,

Bill P.

6

[From Bill Powers (2003.10.15.1112 MDT)]

Fred Nickols (2003.10.15.1230 EDT) --

If I can perceive how I'm moving my hands and comparing that with the "just
so" picture in memory, why can't I also perceive the effects that my
movements are having on my environment? This seems especially crucial if
the performance in question is defined in terms of effects upon the
environment.

I used perceiving how you move your hands as an example of telling someone
what to do, rather than what to accomplish. You are perfectly right. See
Rick's comment -- much more useful than mine.

Best,

Bill P.

7

[Bruce Nevin (2003.10.15 20:50 EDT)]

Fred Nickols (2003.10.14.0915 EDT)--

···

At 09:14 AM 10/14/2003 -0400, Fred Nickols wrote:

1. I not read Powers but if the control a performer has is mostly of his
perceptions how does this eventually impinge on the external world.

The 'impingement' of actions on the environment is your means of affecting your perceptual input. If the loop is not closed through the environment, then either control fails, or you are controlling imagined input.

         /Bruce Nevin

8

[From Rick Marken (2003.10.16.1030)]

Fred Nickols (2003.10.16.1215 EDT) --

Question: I've long known that ends and means are labels that have meaning
only in relation to one another (i.e., an end in relation to one means can
also be a means to some other end). It seems you are saying the same is
true of "what" and "how." Is that the case? If so, I agree with you.

Yes. If "what" refers to a reference signal (what perception you want) and "how"
refers to how you go about getting the perception into the state you want, then
what PCT says is that what you want at one level is part of how you get what you
want at a higher level.

To clarify and restate my concern above: I agree that specifying a result
necessarily restrains and constrains the person who is to accomplish it
(i.e, rules in and rules out some alternatives and options). The lower
(i.e., more tightly) the result is specified, the tighter the restraints
and constraints on the performer. But, specified at a sufficiently high
level, I don't see how the specification of a result also specifies "how"
that result is to be produced.

It constrains (to some extent) _how_ a _high level_ result can be produced.
Suppose I ask you to control for a specific value of a high level perception by
asking you to be "fair" to your employees. This leaves a lot of leeway in terms
of the low level means you can use to control this perception. But it still
constrains how you can control for still higher level perceptions. For example,
you can't control for higher level perceptions -- like being a "team player" -- by
treating your employees unfairly. Of course, this constraint on how you can
control for the higher level perception only exists as long as you agree to
continue controlling for being fair to your employees. My point, however, was
only that asking a person to control for a higher level perception leaves far more
(lower level) degrees of freedom (reference signals) free to vary than it takes
away from control of still higher level perceptions.

Best regards

Rick

···

--
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

9

[From Rick Marken (2003.10.16.1115)]

Let me try this one more time:

Rick Marken (2003.10.16.1030)--

> Fred Nickols (2003.10.16.1215 EDT) --
> ...

> But, specified at a sufficiently high
> level, I don't see how the specification of a result also specifies "how"
> that result is to be produced.

It constrains (to some extent) _how_ a _high level_ result can be produced.

I think it would have been better if I had written:

It constrains (to some extent) _how_ an even _higher level_ result can be produced.

The specification of a result specifies (to some extent) "how", not _that result_ ,
but an even higher level result can be achieved.

I hope that's clearer.

Best

Rick

···

---
Richard S. Marken, Ph.D.
Senior Behavioral Scientist
The RAND Corporation
PO Box 2138
1700 Main Street
Santa Monica, CA 90407-2138
Tel: 310-393-0411 x7971
Fax: 310-451-7018
E-mail: rmarken@rand.org

10

[From Fred Nickols (2003.10.16.1215 EDT)] --

Rick Marken (2003.10.15.0930)]

I think the hierarchical model of control suggests that whenever we say
what is to
be done (what result is to be produced) we are, necessarily, saying how to
do it
(what what actions to take). That's because "what is to be done" is "how
to do
it" relative to a higher level specification for what os to be done.

Question: I've long known that ends and means are labels that have meaning
only in relation to one another (i.e., an end in relation to one means can
also be a means to some other end). It seems you are saying the same is
true of "what" and "how." Is that the case? If so, I agree with you.

For example,
when Henry Ford says that what he wants done is "insertion of tab A into
slot B"
he is not saying _how_ that insertion process should be done (in terms of
detailed
actions used to produce that result) but he is saying how he wants a
higher level
result produced. That higher level result may, for example, be building a
headlamp. With respect to producing that higher level result, Henry is
saying that
it be done by, among other things, inserting tab A into B rather than by
inserting
tab A into C, say.

I get what you're saying above (and below) but it leaves me wondering where
alternatives and options come into play. Clearly, the higher up the
hierarchy I specify the result I'm after, the more latitude the performer
has in accomplishing it in a disturbance prone environment. So, if I
specify the result at a level high enough that it leaves the performer a
lot of latitude in accomplishing it, how have I also dictated
"how"? Granted, by specifying any result at all I both restrain and
constrain the performer in terms of relevant actions. The lower I go in
the hierarchy with my performance specification the more tightly I restrain
and constrain the performer. Not good news in a "disturbance prone"
environment but, in a tightly controlled (i.e., "engineered") environment,
that kind of precise (i.e., lower level) specification is
commonplace. It's the essence of industrial engineering. I'm not say
that's good or bad; just that it's commonplace. But, in disturbance prone
(i.e., turbulent, messy, chaotic what-have-you) environments, low level
specifications are asking for trouble.

So I think Henry Ford is not correctly describing the choice a manager
has. It is
not that a manager can chose between telling an employee how to do
something (the
actions that should be taken to produce a result) and telling the employee
what to
get done (the perceptual result to be produced). The manager can't help
but do
_both_ at the same time. The choice the manager has, I think, has to do
with the
_level_ at which he or she specifies what is to get done. I think what the
manager
should do is try to specify what is to get done at as high a level in the
perceptual hierarchy as is feasible. Since the hierarchical model is pretty
sketchy at the moment, doing this is currently far more art than science.
But the
science suggests that specifying results at at as high a level as possible
is good
because, the higher the level at which one specifies the goal result, the more
lower level degrees of freedom are available for producing the result in a
disturbance prone environment. Lower level degrees of freedom represent
"flexibility" ? the ability to act to protect lower level perceptions from
environmental disturbance on the many dimensions on which they can occur.

I think the idea of "specifying results at as high a level as possible" is
what
lies behind Henry Ford's idea that it is better to tell an employee what
you want
done rather than how you want it done. I think how high the level is at
which you
can specify results depends, to a large extent, on how much the employee
can be
assumed to know about _how_ to produce the higher level result.

I agree.

I think the
problem in factories has been that workers have been assumed to know very
little
about how to produce higher level results so they have been asked to
produce low
level results, which is like specifying "how" to get things done. This is
surely
bad for morale when workers are perfectly capable of seeing how their actions
contribute to the production of a higher level result. I think the
Japanese have
handled this better by actually listening to their employees and realizing
that
the employees often know a lot more about how to produce higher level
results than
had been assumed.

I agree again.

To clarify and restate my concern above: I agree that specifying a result
necessarily restrains and constrains the person who is to accomplish it
(i.e, rules in and rules out some alternatives and options). The lower
(i.e., more tightly) the result is specified, the tighter the restraints
and constraints on the performer. But, specified at a sufficiently high
level, I don't see how the specification of a result also specifies "how"
that result is to be produced.

By the way, the notion of relating specification of results to the
hierarchy is very useful. Thanks.

Regards,

Fred Nickols
nickols@att.net
www.nickols.us

11

[From Fred Nickols (2003.10.16.1220 EDT)] --

Bruce Nevin (2003.10.15 20:50 EDT)]

Fred Nickols (2003.10.14.0915 EDT)--

>>>1. I not read Powers but if the control a performer has is mostly of his
>>>perceptions how does this eventually impinge on the external world.

The 'impingement' of actions on the environment is your means of affecting
your perceptual input. If the loop is not closed through the environment,
then either control fails, or you are controlling imagined input.

Thanks, Bruce. That's consistent with what I thought. Your comment about
"imagined input" prompts some further speculation.

When the results people seek are removed in space and time, perceptual
input is still limited to "here and now" sensory input. We don't and can't
have any perceptions "over here" or "right now" of what is going to happen
"over there later" (except for imagined input). So, I've thought that the
perceptions we try to control in such situations are whatever our "mental
models" tell us about the workings of such situations. In other words, I
behave so that my perceptions of what I'm doing are consistent with a
mental model (i.e., a reference condition) for what is required to bring
about the results I'm after. That's an imagined reference condition, not
an imagined perceptual input, but I think it's related to what you say above.

Is what I'm describing the same as what you meant by "imagined input"?

Regards,

Fred Nickols
nickols@att.net
www.nickols.us

···

At 09:14 AM 10/14/2003 -0400, Fred Nickols wrote: