"situational awareness"

CSGnet Archive CSG1996
1

[Martin Taylor 960429 16:10]

Bruce Gregory (960429.1510 EDT)

> Figuring out reliable ways to divert
>attention from an ongoing control task should be fun.

How is this going to provide clues for understanding situational
awareness?

Only a hunch, but here goes. I suspect that "situational awareness"
is more akin to distractability than to attention. To be
situationally aware is to not be concentrating on controlling one
thing (the airspeed indicator) but to be open to "distractions" such
as the gear-up horn blowing in your ear.

"Situational Awareness" is one of those slippery terms like "consciousness"
that means different things to different people. There are research projects
on "situational awareness" and international committees on the topic. But
it means different things to different people. Bruce's "hunch" doesn't seem
to play well to any of the meanings with which I am familiar.

For example, a fighter pilot's "situational awareness" is usually taken to
mean the extent to which he is able to take control of differen potentially
important perceptions, which might include the relative locations of
friendly and enemy aircraft in the neighbourhood, the tactical interrelations
of those aircraft, the internal states of flight functions of his own
aircraft, the relation of his aircraft to the ground... At any time,
the pilot may not be controlling for any particular ones of these
perceptions, nor even be conscious of them, but if he is "situationally
aware" he can begin to control for them with no startup transient. (That's
looking at it from an HPCT, not a conventional, viewpoint).

"Distractability" doesn't seem to be at all the same thing. A situationally
aware person is not distracted in changing the set of controlled perceptions,
even to controlling perceptions that a short time previously were not
consciously being monitored.

My personal view on this question of conscious awareness has not changed
in the 3 or 4 years since I first promulgated the view that one is aware of
perceptions that are not being controlled but perhaps "should" be (e.g.
that are going out of some "safe" range of error), and of ones that are
being controlled but that might safely be left alone for a while, as
well as of ones for which control is being attempted but with poor
success (I think the latter just might be accounted for under the prior
headings, since the shifts of control of lower perceptions are presumably
an aspect of the unsuccessful control of a supported higher level perception).

One is not normally aware of perceptions that are being successfully
controlled and must continue to be controlled. Nor is one normally aware
of perceptions that will not be controlled. In either case, attention may
be directed to the perception in question, but I speculate that this only
means that the perception in question might begin or cease to be controlled,
in support of whatever higher-level perception influenced the directing
of attention to it.

But that's just a personal speculation. Kent McClelland (960428.1700 CDT)
has recently presented much the same speculation:

With all these processes going on in several perceptual modalities, what if
there were an "observer" system within the brain that got "reports" on the
currently stabilized perceptions (whatever their perceptual order) from
each of the perceptual modalities? Perhaps the function of that system
would be to resolve possible conflicts between the perceptions being
stabilized in different modalities by picking one modality to "pay
attention to" at any given moment, so that any conflicts which arose would
be resolved in its favor. What I'm groping to describe is an "attention
system" that moves the camera of attention across all the possible inputs
and fixes it from moment to moment on one perception or another.

I could spin this fantasy out a little further but don't have the time
right now. Let me know if anyone can see some plausibility to this
proposal.

Well, since it fits right in with my prior beliefs, I naturally find it
plausible. But is it right? I have no idea.

Martin

2

[From Bruce Gregory (960430.1020)]

(Martin Taylor 960429 16:10)

  For example, a fighter pilot's "situational awareness" is usually taken to
  mean the extent to which he is able to take control of different potentially
  important perceptions, which might include the relative locations of
  friendly and enemy aircraft in the neighbourhood, the tactical interrelations
  of those aircraft, the internal states of flight functions of his own
  aircraft, the relation of his aircraft to the ground... At any time,
  the pilot may not be controlling for any particular ones of these
  perceptions, nor even be conscious of them, but if he is "situationally
  aware" he can begin to control for them with no startup transient. (That's
  looking at it from an HPCT, not a conventional, viewpoint).

I agree. I was simply trying to say that an accident seems often to
be the result of a loss of situational awareness triggered by the
effort to control one anomalous variable. This variable is not a
"distraction" in any conventional sense. It is very relevant (such
as a misleading airspeed caused by an iced pitot tube), but in trying
to control it, the pilot "forgets" to monitor other variables, some
of which would have cast doubt on the variable he is trying to
control.

Lets look at the airspeed indicator problem. I recall an accident
report in the 1960's in which the crew ferrying a 707 suffered pitot
icing and resulting ever-increasing airspeed readings. In the effort
to control airspeed, the pilots reduced the throttles to idle and
pulled back on the stick. As the plane approached a stalling angle
of attack, the stick shaker was activated. Since the airspeed
indicator was still showing a very high reading, the crew interpreted
this shaking as Mach buffet, and increased the angle of attack
further. The aircraft stalled, spun, and augered in from 20,000 ft.
Had the crew check the airspeed reading against the artificial
horizon or the altimeter, they would have known something was wrong.
Their failure to cross-check suggests that they were concentrating
exclusively on controlling the airspeed. That is what I was trying
to say, perhaps in too few words.

Bruce G.

3

[From Richard Thurman (960430.0930)]

(Martin Taylor 960429 16:10)

For example, a fighter pilot's "situational awareness" is usually taken

to

mean the extent to which he is able to take control of differen

potentially

important perceptions, which might include the relative locations of
friendly and enemy aircraft in the neighbourhood, the tactical

interrelations

of those aircraft, the internal states of flight functions of his own
aircraft, the relation of his aircraft to the ground... At any time,
the pilot may not be controlling for any particular ones of these
perceptions, nor even be conscious of them, but if he is "situationally
aware" he can begin to control for them with no startup transient.

(That's

looking at it from an HPCT, not a conventional, viewpoint).

Could this "no startup transient" be an illusion based upon the assumption
that "the pilot may not be controlling for any particular perception?"
What if the pilot is indeed controlling those perceptions in question --
they just happen to have very small errors? When error becomes large (via
environmental disturbance or change in reference level) it seems as if
there is a very small startup transient, but in reality the variable was
already under control.

It may be the case that the pilot is not consciously controlling the
perception and therefore may not be consciously aware, but if there is no
startup transient, I would hazard a guess that the perception in question
is being unconsciously controlled and therefore the pilot is
"unconsciously aware."

"Distractability" doesn't seem to be at all the same thing. A

situationally

aware person is not distracted in changing the set of controlled

perceptions,

even to controlling perceptions that a short time previously were not
consciously being monitored.

I think that you and Bruce are saying the same thing. Both of you agree
that when people are distracted they have lost situation awareness. Or
are you saying something different here?

My personal view on this question of conscious awareness has not changed
in the 3 or 4 years since I first promulgated the view that one is aware

of

perceptions that are not being controlled but perhaps "should" be (e.g.
that are going out of some "safe" range of error),

I cannot follow this. If one is aware that a perception is going out of
some safe range of error then it is already under control. It may not be
tight control, but it is control nonetheless.

and of ones that are
being controlled but that might safely be left alone for a while,

I cannot see the difference between this kind of awareness and the first
kind mentioned above.

as well as of ones for which control is being attempted but with poor
success (I think the latter just might be accounted for under the prior
headings, since the shifts of control of lower perceptions are presumably
an aspect of the unsuccessful control of a supported higher level
perception).

As far as I can tell, all three kinds of conscious awareness you mentioned
above are simply control. Their qualitative differences come from
increased gain, or varying reference signal values.

This is a very fascinating discussion.
Thanks to all for such willingness to share ideas and concepts!

Rich

···

--------------------------------------------------
Richard Thurman
Air Force Armstrong Lab
6001 S. Power Rd. BLDG. 558
Mesa AZ. 85206-0904
(602)988-6561
Thurman@hrlban1.aircrew.asu.edu
---------------------------------------------------

4

[Martin Taylor 960430 11:30]

Bruce Gregory (960430.1020)

I was simply trying to say that an accident seems often to
be the result of a loss of situational awareness triggered by the
effort to control one anomalous variable. ... It is very relevant (such
as a misleading airspeed caused by an iced pitot tube), but in trying
to control it, the pilot "forgets" to monitor other variables, some
of which would have cast doubt on the variable he is trying to
control.

Yes, and I think that this ties in both with the "attention and awareness"
thread and with the hemispheric thread that somewhat diverged from the
consciousness thread.

One of the functions of attention seems to be to restrict what would otherwise
be multiple parallel streams of processing down to one (or a small number).
There are hints of this in a wide variety of situations, none definitive.
Here are a couple of examples.

1. Hypnosis.
Several years ago, I helped with the analysis of an experiment done during
her grad school year by a summer student of mine. All subjects had tested as
being highly skilled in being hypnotized (yes, it's an ability of the
hypnotized person, not of the hypnotist). I forget now the precise details,
but it's something like this: The subjects were played two streams of
words, one into each earphone, and had to detect target words, pressing
one button for a target in one ear and another for a target in the other
ear. Under normal conditions, detection performance deteriorated when
both streams were playing, in the sense that detection in each stream was
worse than when that stream was played by itself. But under hypnosis there
was no difference between performance when the stream was played alone or
when the other stream was also being played. The hypnotic state seemed
to eliminate some inhibition of one on the other, and I speculate that
this relates to a lack of conscious awareness of performing the task.

I make this speculation based partly on personal experience in a different
situation--auditory signal detection, listening for tone bursts in noise.
I've done this a lot, and on quite a few occasions I have finished a sequence
of 100 without having been aware that I was pushing the response buttons
at all. I have had to ask the experimenter whether I had actually been
responding. Invariably, the results in these conditions have been more
consistent and at a slightly higher performance level than when I have been
aware of, and thinking of, the signal and which button to press.

2. Masked priming in reading
Tony Marcel (Attention and Performance VIII) did a study that various
people have claimed to have been refuted, but I cannot see that any of
the refutations are valid. What he did was as follows (it uses statistics,
which may be sufficient refutation for some on CSGnet):
Subjects were asked to determine whether a letter string flashed on a
screen was a word or not. The experimenter measured how long it took them
to push the button corresponding to their decision. It's a standard task
in the reading literature, though what it has to do with everyday reading
I have always questioned. Be that as it may, subjects are assumed to be
controlling for pleasing the experimenter by pushing the button as fast
and accurately as they can, and experiments usually get quite consistent
results.
One of the consistent results is that if a word is preceded by a strongly
associated word (e.g. doctor->nurse, or table->chair), the "yes" button
push is faster than if it preceded by a neutral word (e.g. "blank"->nurse).
Some words have multiple meanings. "Palm" can refer to a part of the hand
or to a kind of tree, for example. To cut a long story short, if three
words are shown in sequence--Hand->Palm->Wrist (related meanings),
Tree->Palm->Wrist (unrelated meanings) or Speed->Palm->Wrist (neutral),
the results differ depending on the relationships among the three words
(timings are always measured for the last word only). Related is fastest,
then neutral, then unrelated.
  What Marcel did was to mask the central word in a sequence of three, in two
ways, using a pattern of letter parts or using a noise field. These two masks
have different kinds of effects, as many experiments have shown. With the
pattern mask, a subject will often say that no word was presented, and yet
under appropriate test conditions it can be shown that some aspects of the
word (specifically its semantic meaning, in many cases, but not in all cases)
affected the subject, and in a forced choice experiment the subject will
usually choose correctly. With the noise-field mask, usually when the subject
says there was no word presented, the word has no effects.
  Marcel's critical result was this. When the first word was associated with
with the other two (related meanings), button-push speed was essentially
the same whether the central word was unmasked or pattern masked, both being
faster than when the central word was noise-masked. That's to be expected.
But when the first word was associated with the _wrong_ meaning of the
central word (e.g. tree->palm->wrist), the unmasked condition was the
slowest in the whole experiment, the noise-masked condition was like the
other control conditions, but the pattern-masked condition was almost as
fast as when the first word had the related meaning (e.g. hand->palm->wrist),
and much faster than the other control conditions (though the neutral
pattern-masked condition was almost as fast).

What this says to me is that the pattern-masked condition allows a person
to "read" the word without being conscious of it, in the sense that all
the possible meanings of the word are evoked (much as JW could get the
meaning of a word presented to his right hemisphere without being able to
identify the word). The business of the "conscious" process is to inhibit
the "wrong" possibilities in favour of the one that fits the local context.

There is, to my mind if not to that of other reading researchers, a lot
of evidence suggesting that there are two "tracks" of reading processes,
one that rapidly provides all possible readings, and another than more
slowly determines the unique "correct" reading, inhibiting all other meanings.
The former set of processes occurs in both hemispheres, the latter only
in the left, and is associated with concepts such as sequence and syntax.

Extrapolating wildly, one might suggest that _conscious_ control involves
the inhibition of multiple parallel control processes at the _same_
level of the perceptual hierarchy, in favour of the one on which attention
is concentrated (I think "concentration" is a most apt metaphor here, both
in its everyday sense--orange juice--and in its Latin roots).

Now if we do a wild extrapolation to the aircraft accident situation, we
can speculate that the _conscious_ attention to the faulty airspeed reading
and attempts consciously to control that may actually inhibit the perception
and control of "wrong" variables in the aircraft situation. Under normal
flight conditions, multiple parallel control is easy and not (very?)
conscious.

I realize this is very vague, and brings together stuff that might not
belong together, and is wildly speculative...but what the heck. It's worth
considering. And it does tie in with the flip-flop/associative-group
mechanism I have mentioned in the past. But that's a story for another day.

Martin

5

[Martin Taylor 960430 13:30]

Richard Thurman (960430.0930)

(Martin Taylor 960429 16:10)
At any time,
the pilot may not be controlling for any particular ones of these
perceptions, nor even be conscious of them, but if he is "situationally
aware" he can begin to control for them with no startup transient.

Could this "no startup transient" be an illusion based upon the assumption
that "the pilot may not be controlling for any particular perception?"
What if the pilot is indeed controlling those perceptions in question --
they just happen to have very small errors?
...
It may be the case that the pilot is not consciously controlling the
perception and therefore may not be consciously aware, but if there is no
startup transient, I would hazard a guess that the perception in question
is being unconsciously controlled and therefore the pilot is
"unconsciously aware."

I'm looking at the opposite situation, in which the pilot is aware of the
state of a perception (perhaps vaguely "consciously"), but is currently
performing no action to influence it. It maintains a low error value
because external disturbances have in the recent past been weak or
predictable, as would be the case in a flight through calm air, for example.
A perceptual variable (aircraft attitude) is monitored, but the pilot is,
figuratively speaking, just sitting back and watching until the error
gets large enough to require active control. The pilot is monitoring
potentially many more perceptual variables than can possibly be controlled
by simultaneous action, so the output is time-multiplexed. At any moment
only a subset of the perceptions involved in his "situational awareness"
are actively under control, but at any moment the membership of that subset
might switch, and it would do so with very little startup transient.

This is in contrast to the startup transient that might occur if the pilot
had fallen asleep, to be awakened by a stentorian "PULL UP...PULL UP," or
if he suddenly became aware that the dot on the horizon was an aircraft
on a collision course with his.

I think that you and Bruce are saying the same thing. Both of you agree
that when people are distracted they have lost situation awareness.

I think you misread Bruce Gregory. As I understood him, it is an _inability_
to be distracted that signals a loss of situational awareness. Attention
is too solidly focused on a perception that is the subject of unsuccessful
attempts at control. Various lower-level perceptual controls are being
switched around (rather like reorganization) so that the problematic one
can be brought successfully under control. A person who is so absorbed in
one control process at a given level is inhibiting the control of other
perceptions at the same level--is not "distractable." At least that's both
how I read him and how I understand the situation.

Martin

6

[From Bruce Gregory (960430.1455 EDT)]

(Martin Taylor 960430 13:30)

  I'm looking at the opposite situation, in which the pilot is aware of the
  state of a perception (perhaps vaguely "consciously"), but is currently
  performing no action to influence it. It maintains a low error value
  because external disturbances have in the recent past been weak or
  predictable, as would be the case in a flight through calm air, for example.
  A perceptual variable (aircraft attitude) is monitored, but the pilot is,
  figuratively speaking, just sitting back and watching until the error
  gets large enough to require active control. The pilot is monitoring
  potentially many more perceptual variables than can possibly be controlled
  by simultaneous action, so the output is time-multiplexed. At any moment
  only a subset of the perceptions involved in his "situational awareness"
  are actively under control, but at any moment the membership of that subset
  might switch, and it would do so with very little startup transient.

That's the way I think of the situation, too.

  This is in contrast to the startup transient that might occur if the pilot
  had fallen asleep, to be awakened by a stentorian "PULL UP...PULL UP," or
  if he suddenly became aware that the dot on the horizon was an aircraft
  on a collision course with his.

Agreed.

  I think you misread Bruce Gregory. As I understood him, it is an _inability_
  to be distracted that signals a loss of situational awareness. Attention
  is too solidly focused on a perception that is the subject of unsuccessful
  attempts at control. Various lower-level perceptual controls are being
  switched around (rather like reorganization) so that the problematic one
  can be brought successfully under control. A person who is so absorbed in
  one control process at a given level is inhibiting the control of other
  perceptions at the same level--is not "distractable." At least that's both
  how I read him and how I understand the situation.

That's my understanding too.

Bruce G.

7

[From Richard Thurman (960501.0845)]

Bruce Gregory (960430.1455 EDT)

(Martin Taylor 960430 13:30)

I think you misread Bruce Gregory. As I understood him, it is an

_inability_ >>to be distracted that signals a loss of situational
awareness. Attention

is too solidly focused on a perception that is the subject of

unsuccessful

attempts at control. Various lower-level perceptual controls are being
switched around (rather like reorganization) so that the problematic one
can be brought successfully under control. A person who is so absorbed

in

one control process at a given level is inhibiting the control of other
perceptions at the same level--is not "distractable." At least that's

both

how I read him and how I understand the situation.

That's my understanding too.

Bruce, are you agreeing with Martin's assessment of your proposal. Or are
you agreeing that he has interpreted your proposal correctly.

He is saying that you are saying that it is an inability to be distracted
that causes a loss in situational awareness (or that it is the ability to
be distracted that leads to SA).

If you are saying that, then I should be able to predict that the
'attention deficit disorder' kids I mentioned last week should make the
best pilots. These kids, at least on the surface, seem not to be able to
focus on any one thing for very long. They attend to everything, but in
so doing can't concentrate on anything.

Yet I also can agree with parts of what you seem to be implying. When
individuals are so focused on one single thing they can't possibly be
keeping track of any other (possibly) important perceptual variables.

Rich

···

--------------------------------------------------
Richard Thurman
Armstrong Lab
6001 S. Power Rd. BLDG. 558
Mesa AZ. 85206-0904
(602)988-6561
Thurman@hrlban1.aircrew.asu.edu
---------------------------------------------------

8

[From Richard Thurman (960501.0810)]

[Martin Taylor 960430 13:30]

Richard Thurman (960430.0930)

It may be the case that the pilot is not consciously controlling the
perception and therefore may not be consciously aware, but if there is

no

startup transient, I would hazard a guess that the perception in

question

is being unconsciously controlled and therefore the pilot is
"unconsciously aware."

I'm looking at the opposite situation, in which the pilot is aware of the
state of a perception (perhaps vaguely "consciously"), but is currently
performing no action to influence it. It maintains a low error value
because external disturbances have in the recent past been weak or
predictable, as would be the case in a flight through calm air, for

example.

If a pilot is aware of the state of a perception and performing no action
to influence it (due to its "low error") then he is controlling.

A perceptual variable (aircraft attitude) is monitored, but the pilot is,
figuratively speaking, just sitting back and watching until the error
gets large enough to require active control.

In this example, aircraft attitude is definitely being controlled. Just
sitting back and watching until the error gets large enough is control.
Sometimes sitting back is exactly what is necessary to maintain a
perceptual variable at a reference level.

Take the rubber band demo for example. If I am the experimenter and you
are the subject, I can let the knot sit on top of the dot and you will do
nothing -- yet you are still controlling for the knot to be on top of the
dot.

The pilot is monitoring
potentially many more perceptual variables than can possibly be

controlled

by simultaneous action, so the output is time-multiplexed. At any moment
only a subset of the perceptions involved in his "situational awareness"
are actively under control, but at any moment the membership of that

subset

might switch, and it would do so with very little startup transient.

Sounds like control at the program level to "multiplex".

I would guess that if any one of the variables under control were to
deviate rapidly from its predicted value (model based control?) then one
would see startup transients (and the pilot would be startled).
Otherwise, very little startup time would be needed.

This is in contrast to the startup transient that might occur if the

pilot

had fallen asleep, to be awakened by a stentorian "PULL UP...PULL UP," or
if he suddenly became aware that the dot on the horizon was an aircraft
on a collision course with his.

When control is lost (however loose it may be) - control is lost.

···

-------------------------------------------------------------------------

I think that you and Bruce are saying the same thing. Both of you agree
that when people are distracted they have lost situation awareness.

I think you misread Bruce Gregory. As I understood him, it is an

_inability_

to be distracted that signals a loss of situational awareness. Attention
is too solidly focused on a perception that is the subject of

unsuccessful

attempts at control. Various lower-level perceptual controls are being
switched around (rather like reorganization) so that the problematic one
can be brought successfully under control. A person who is so absorbed in
one control process at a given level is inhibiting the control of other
perceptions at the same level--is not "distractable." At least that's

both

how I read him and how I understand the situation.

You are completely right. I read him as saying the opposite. This is
kind of like talking to behavioral researchers about "the control of
perception" but all they hear is "control by perception." I guess I was
filtering out what Bruce was saying so that it would sound like what I
wanted to hear.

In light of this I agree with your assessment (assuming I understand it
correctly). It does not seem correct to say that situational awareness is
a result of the ability to be distracted.

Rich

--------------------------------------------------
Richard Thurman
Armstrong Lab
6001 S. Power Rd. BLDG. 558
Mesa AZ. 85206-0904
(602)988-6561
Thurman@hrlban1.aircrew.asu.edu
---------------------------------------------------

9

[From Bruce Gregory (960501.1245 EDT)]

(Richard Thurman 960501.0845)

  Bruce, are you agreeing with Martin's assessment of your proposal. Or are
  you agreeing that he has interpreted your proposal correctly.

  He is saying that you are saying that it is an inability to be distracted
  that causes a loss in situational awareness (or that it is the ability to
  be distracted that leads to SA).

Distracted is a bit of a loaded word. If you are focusing on the
artificial horizon, is an indication of electrical failure a
"distraction"? Maybe. Then again, probably not.

  If you are saying that, then I should be able to predict that the
'attention deficit disorder' kids I mentioned last week should make the
  best pilots. These kids, at least on the surface, seem not to be able to
  focus on any one thing for very long. They attend to everything, but in
  so doing can't concentrate on anything.

Interesting point. I guess one question would be if they are able to
control all the variables a pilot needs to control. At the other end
(maybe) of the spectrum, we had a group of autistic young people
working in the library for about a year. They were very effective,
because they didn't seem to get bored doing very routine things. In
fact, they seemed to flourish -- they reshelved books much more
accurately than the scientists do. If we could teach them the
instrument scan, they might making outstanding pilots. At least
until they encountered a truely novel problem.

  Yet I also can agree with parts of what you seem to be implying. When
  individuals are so focused on one single thing they can't possibly be
  keeping track of any other (possibly) important perceptual variables.

That's all I was really trying to say.

Bruce G.