Awareness (was Choice)

[From Dan Palmer (2000.01.31.1229)]

I enjoyed Bill Power's (2000.01.30.0630 MST) description of "awareness" in PCT
terms. I'm wondering if the following would fit in anywhere. Somewhere I
picked up the idea that conscious awareness has to do with keeping a verbal
tab on what we are presently doing. This makes awareness consist of a verbal
description of present action. So, for example, awareness would consist not
in seeing something, but in realizing that we are seeing something in the
sense of verbally framing that act of seeing. A lecturer of mine used to give
the example of driving through a small town on the way to your destination,
and then, on being asked whether you had yet passed through that town, not
being sure, because you weren't keeping a verbal commentary of what you were
doing as you drove though, even though you obviously did see the town.

I think this might be compatible with PCT in that, as Bill says, the hierarchy
can smoothly proceed at its various levels without feeding into awareness, yet
awareness is able to observe levels of the hierarchy when the need arises
(when the book hits your toe, for example). So would this (rather loose) idea
of a kind of "verbal spotlight" be any help?

It is an odd notion that in trying to explain this ability to observe
different parts of the hierarchy seems to require that we "add" something
other than more levels, signals and functions. It's as if awareness requires
that we jump out of our own heads to look at our own hierarchy from outside
itself. Interesting stuff.

Cheers,
Dan Palmer

[From Bill Powers (2000.01.30.11476 MST)]

Dan Palmer (2000.01.31.1229)--

[Hello, Dan!]

I enjoyed Bill Power's (2000.01.30.0630 MST) description of "awareness" in

PCT

terms. I'm wondering if the following would fit in anywhere. Somewhere I
picked up the idea that conscious awareness has to do with keeping a verbal
tab on what we are presently doing.

I'm sure that we can keep verbal tabs on what we are presently doing, and
that we can be aware of doing so. But I don't think that keeping verbal
tabs _is_ awareness. I admit that it's hard to think of attaching verbal
labels to things as possible to do _unconsciously_, but perhaps we call it
something else when we're not conscious of doing it -- stereotyping, for
example, or forming associations.

Rather than argue about that, it's probably better to try to think of
experiments that involve the direction and misdirection of attention as a
way of finding out how awareness affects control processes (if it does).

The problem with your proposal is that verbalization, overt or covert,
might well have something to do with the retrievability of memories, so it
could affect the brain's ability to retrieve stored information whether or
not awareness is required to do it. And even if awareness _were_ always
involved when we create verbal descriptions, that would not mean that the
two processes are the same thing. My hands are always involved when I eat
with a fork, but that doesn't mean that forks are hands.

Best,

Bill P.

  This makes awareness consist of a verbal

description of present action. So, for example, awareness would consist not
in seeing something, but in realizing that we are seeing something in the
sense of verbally framing that act of seeing. A lecturer of mine used to

give

the example of driving through a small town on the way to your destination,
and then, on being asked whether you had yet passed through that town, not
being sure, because you weren't keeping a verbal commentary of what you were
doing as you drove though, even though you obviously did see the town.

I think this might be compatible with PCT in that, as Bill says, the

hierarchy

can smoothly proceed at its various levels without feeding into awareness,

yet

awareness is able to observe levels of the hierarchy when the need arises
(when the book hits your toe, for example). So would this (rather loose)

idea

ยทยทยท

of a kind of "verbal spotlight" be any help?

It is an odd notion that in trying to explain this ability to observe
different parts of the hierarchy seems to require that we "add" something
other than more levels, signals and functions. It's as if awareness requires
that we jump out of our own heads to look at our own hierarchy from outside
itself. Interesting stuff.

Cheers,
Dan Palmer

[Martin Taylor 20000131 9:56]

[From Bill Powers (2000.01.30.11476 MST)]

Rather than argue about that, it's probably better to try to think of
experiments that involve the direction and misdirection of attention as a
way of finding out how awareness affects control processes (if it does).

There's an experiment related to this that we hope will soon be under
way. It isn't aimed specifically at "awareness" (which has a lot of
different meanings), but it is aimed at directing attention.

In 1994, Bill P provided a set of tracking tasks that were used in a
study on the effects of drugs supposed to help against sleep
deprivation. We are now intending to use some of those tasks again in
a study of proposed aircraft audio wornings. The idea is that the
subject will be controlling some perception against an easy or a
difficult disturbance, and at some moment while this is going on, an
audio warning will occur. The subject will have to report which
warning it was. In the analysis, we will (try to) estimate the
control model parameters in the seconds before, during, and after the
warning, to see what effect (if any) the diversion of attention (if
it occurs) had on the control system.

We might also be able to determine the effect of tracking difficulty,
and perhaps of the hierarchic level of the controlled perception, on
the accuracy of the subject's perception of the warning signal. But
this last is rather a long shot.

"Awareness" might apply to the warning signal, the tracking task, or
both, I guess. But so far as I can see, there would be no
experimental way to determine whether the subject was "aware" of
either. All we will know are the tracking data and the reports on the
warnings.

Martin

[From Bill Powers (2000.01.31/1131 MDT)]

Martin Taylor 20000131 9:56--

We are now intending to use some of those tasks again in
a study of proposed aircraft audio wornings. The idea is that the
subject will be controlling some perception against an easy or a
difficult disturbance, and at some moment while this is going on, an
audio warning will occur. The subject will have to report which
warning it was. In the analysis, we will (try to) estimate the
control model parameters in the seconds before, during, and after the
warning, to see what effect (if any) the diversion of attention (if
it occurs) had on the control system.

I'm looking forward to seeing the results. I hope that this time the
physical conditions will be precisely standardized, with a stable arm
position and only one degree of freedom of arm movement (a sideways swing
with the elbow in a fixed position, if a mouse is used), with assurance of
a good surface for the mouse to prevent slippage, a grip on the mouse that
does not let any part of the hand drag against the supporting surface, with
careful attention to establishing baseline measurements, and with a
standardized screen-viewing distance and position. Also, I hope there will
be sufficient practice so that each individual in the study has reached
asymptotic performance and the parameters of control are stable before any
variations or manipulations of conditions are introduced. If all these
conditions are met (and any others you can think of along the same lines),
unnecessary noise will be removed from the data and the measurements should
be far more sensitive than they were in the first uses of these tasks.
Considering the amount of labor that will be put into running the
experiments and collecting and analyzing the results, a reasonable
investment of time and attention to detail at the beginning will greatly
improve the return on investment.

I have been looking (on the Web) into the subject variously called
"optimization" and "minimization," trying to find out how to do the Powell
method of model-fitting that is used in Vensim. Using the Vensim package, I
was able to get far better fits of the model to data than my simple-minded
method could achieve -- even using integration factor, leakage factor,
delay, and reference level as four parameters to be adjusted. I think you
would be well-advised to explore this approach -- and of course when you've
figured it all out, I would much appreciate being taught how to do it!

Best of luck,

Bill P.