- The effects of flying upon human performance - article ; n°1 ; vol.50, pg 629-638

Publié par

L'année psychologique - Année 1949 - Volume 50 - Numéro 1 - Pages 629-638
10 pages
Source : Persée ; Ministère de la jeunesse, de l’éducation nationale et de la recherche, Direction de l’enseignement supérieur, Sous-direction des bibliothèques et de la documentation.
Publié le : samedi 1 janvier 1949
Lecture(s) : 23
Nombre de pages : 11
Voir plus Voir moins

F. Bartlett
XI. - The effects of flying upon human performance
In: L'année psychologique. 1949 vol. 50. pp. 629-638.
Citer ce document / Cite this document :
Bartlett F. XI. - The effects of flying upon human performance. In: L'année psychologique. 1949 vol. 50. pp. 629-638.
doi : 10.3406/psy.1949.8479
by Sir Frederic Bartlett, F. R. S.
Cambridge Psychological Laboratory.
In considering what I could contribute to this Volume it
seemed to me that I should choose some topic having a definite
experimental basis, fundamental in its interests and yet at the
same time alive to current movements of thought and method.
For it is fair to say that these have been leading characteristics
of the long series of important contributions with which Pro
fessor Piéron has enriched the records of psychological science.
Especially during the late War, it became necessary for many
of the workers in this Laboratory to turn their attention to a
study of the changes in the performance of a variety of skills
when that performance must continue without interruption for
relatively long periods, often under stress. The types of activity
most involved rarely required excessive physical or muscular
effort at any stage. But they did require accuracy and usually
considerable speed in manipulation. They all, that is to say,
had their expression in co-ordinated bodily movements, and the
changes accompanying prolonged exercise could all be best stu
died as measurable changes in, or directly in relation to, these
1. I consider it an honour to have this opportunity to pay my tribute,
and that of many generations of students at the Cambridge Psychological
Laboratory, to the outstanding achievements of Professor Henri Piéron.
His work has been to us a persisting stimulus and a guide, for theincisiveness
and clarity of his ideas, for his strict adherence to the demands of scientific
method, and for the great breadth of his interests. We wish for him now a
long period of active leisure, during which his genius may move freely
wheresoever it desires. 630 PSYCHOLOGIE APPLIQUÉE
movements. At the same time the movements themselves were a
serial response to successive signals, or " displays", which made
up the working environment. The operations were in the form
of a continuous interplay between information and action, each
of which largely determined the next step in succession of the
other. The essential problems conformed in general to the
'familiar fatigue problem. But it very soon appeared from obser
vation that the results did not agree very well with any familiar
fatigue picture. Indeed it became speedily apparent that what
is often called «• fatigue » in the case of skill cannot be well
investigated in any experimental situation which calls for
the prolonged repetition of a relatively simple operation, and
attempts a measure in terms of a curve representing the course
throughout the whole active period of some simple character
of overall output. Skill " fatigue " — if we are to continue to
-use this word — has characteristics of its own. Its experimental
study demands special methods and the measures needed are
special measures.
One of the particular cases selected for special study was
the case of changes of human activity accompanying prolonged
air flights. Very broadly speaking there are two sets of problems :
fatigue in aircrew, and fatigue in the passenger. A little is
beginning to be known about the first, though far more remains
yet to be discovered. The second is still practically all a matter
for anecdote and speculation. I shall therefore limit this commun
ication to some remarks about fatigue in aircrew.
I. — Characteristics
There is so far no compelling evidence that any single process,
or any single group of processes, underlies all the many manif
estations of what is commonly called fatigue. For working
purposes fatigue must be held to include all those changes of
performance which appear to be due to the continued exercise
•of the activities involved; and in practice it is desirable to di
stinguish between continued exercise under what may be regarded
as the normal range of performance conditions, and continued
exercise when the normal conditions are exaggerated in some
special direction, or unaccustomed environmental circumstances
are present. Flying fatigue, for example, covers the possible F. BARTLETT, F. R. S. THE EFFECTS OF FLYING 631
effects of noise, of vibration, of a certain range of temperature
changes and of a number of other environmental circumstances
to the extent to which they are commonly present in flight,
and of course it covers also all possible direct effects of the
duration of the. activity up to the customary limits of continued
exercise. In addition there are special effects with marked oxygen
lack, or with abnormal acceleration, or at extreme altitude, or
with great excess of heat or cold, or with many other unusual
accompaniments. The present paper will be concerned only
with conditions that lie within the normal range.
All members of air crew have special duties to perform and
it is reasonable to seek a measure of fatigue in terms of the
recordable changes in the ways in which they carry out their
particular technical tasks. Although these operations differ from
one member of aircrew to another, they all have important
broad likenesses. None normally involves long continued heavy
muscular effort. All are of the nature of skilled adjustment to
environmental signals which to an increasing extent are given
by a variety of instrument readings, the instruments being a
part of the normal aircraft equipment. The operations are the
refore correctly to be regarded as requiring a continued process
of interplay between receptor interpretations of environmental
signals and groups of signals, and timed and co-ordinated effector
response. When an instrument record is perceived to have
changed by a certain amount, it initiates an effector response,
and this itself brings about a further change of instrumental
record which in due course leads to a further bodily adjustment.
An important and never-to-be-forgotten part of the significance
of this is that, within limits, whenever any trained and efficient
aircrew operator does something, he is able to some degree to
anticipate and prepare for the probable next adjustment in the
series. This is one of several reasons which make it impossible
ever to express change of activity as a measure of efficiency
in the execution of any single item of the performance alone,
or as a measure of the overall final product of the total number
what' is regarded as the finished task, or in any of items in
single cycle of operations of that task.
In a general sense the extraordinary feature of any series of
receptor-effector interchange of this type, if both the display
— the signals for action — and the control — the bodily move
ments - — are well designed, with a due consideration of what
the human body and mind are adapted to do, is the very long PSYCHOLOGIE APPLIQUEE 632
period of continuous exercise that is possible for any fit man
before notable changes in performance become manifest, either
during or after exercise. It is safe to say that if significant dete
rioration can be shown in any bodily fit and mentally interested
member of aircrew within the limits of a flight of -ordinary dura
tion, or of any continuations of such flights from day-to-day,
either the operation has been badly designed, or factors are at
work which lie outside those involved in the continued exercise
Changes of activity leading to actual or potential deterioration
of efficiency do however commonly occur, even, say, in the
course of an ordinary trans-Atlantic flight. They indicate that
there is still very much to be learned about the correct design
and matching of display and control in the case of aircrew
operations. This is of course well known and widely recognised,
and it dictates what is at present an over-riding group of pro
blems in research into the psychological conditions affecting
human activity in flying operations. What is the optimum design
and positioning for instrument dials and scales and their recorder
devices? How may direction of control most naturally fit corre
sponding directional change of signals? What are the normal
human tolerance limits for " speed " and " load " when linked
interpretation and action have to be prolonged? Perhaps most
important and least known of all, how can the tempo and other
features of serially arranged signals and controls be so planned
as best to fit the inevitable anticipatory pattern of human
The answers to these and other questions of the same basic
rationale which are now being actively investigated all over the
world, vary in detail according to the particular function concer
ned, and also probably, as Russell Davis (1948) has shown,
according to whether the operator belongs to an over- or under-
active type. The details cannot be properly considered in a
short communication such as this, but there are at least four
very closely related principles which have to be kept in mind
all the time, and these can be stated.
a) Threshold of indifference.
A man is interpreting a change of signal in his environment.
When this varies a certain amount from its starting state he F. BARTLETT, F. R. S. THE EFFECTS OF FLYING 633
must do something about it. There are two thresholds involved.
There is a minimal detectable change, and there is a minimal
effective change. In the single test, when the operator, fresh
or tired, is, so to speak on his mettle and specially alert, these
two thresholds are almost, if not exactly, identical. But in
continued exercise they are apt to drift apart, and the body
remains indifferent to changes which the mind, or the senses,
can still continue to detect. Action is then delayed, the timing
of responses in a series becomes ragged and the performance,
though for a long time yet its overall measure may show little
variation, is on the way to dissolution. It is not the ordinary
threshold of acuity that has suffered, but the threshold of indi
fference that has extended.
There need be nothing conscious about this. The widening of
the limits of stimulus change about which the muscles remain
indifferent may be a purely neurological affair, the mechanisms
of which are at present most imperfectly understood.
This is the principal reason why remarkable results can be
secured by any method of regular intermittent feeding back to
the central nervous mechanism of the operator at the level of
his current performance. It then appears that the operator,
somewhat after the style of a modern servo-mechanism (see
e. g. Hick and Bates 1950) is able to effect a practical compar
ison of what he is doing now with what he was doing a little
while ago, or with w^hat he has set out to do, and in the result
the " threshold of indifference " is kept within its initial limits.
The striking effects that can be produced in this way were
beautifully demonstrated by L. Carmichael and W. S. Dearborn
in their well known experiments on Reading and Visual Fatigue
and our own work has yielded many other illustrations.
Moreover this widening of indifference threshold is one of
several reasons why nobody ever has discovered, and probably
nobody ever will discover, a single quick test of skill fatigue.
The principle diagnostic sign of such fatigue is a change in the
rhythm of sequence of receptor-effector responses which will not
be recorded at all outside of the rhythm.
b) The awareness of discomfort.
When the rhythm of skill becomes involuntarily somewhat
ragged and the timing of the constituent items suffers, genuinely 634 PSYCHOLOGIE APPLIQUEE
psychological results normally follow. As action becomes jerky
and there are snatched responses here and halts in response
there, bit by bit the operator may become more aware of how
his body is working, and less of what it is that he is trying to do.
An obscure mass of proprioceptive impulses tend to push thems
elves into the forefront of his consciousness, until they have
almost obsessive force. For these the operator very often blames
anything but his own behaviour and he may assert that he is
doing just as well as he ever did, even just the same as he
ever did, but he complains of a variety of more or less localised
discomforts. This is the second broad character of the onward
march of fatigue, but though it usually marks a fairly advan
ced stage, it cannot be taken to show any completely constant
relation to overall efficiency of performance, and there is no
known way by which it can be used to constitute a reliable
c) Anticipation span.
Much less is known in detail about the third broad character
istic of advancing fatigue that I wish to mention. I have said
already that the expert member of aircrew, engaged upon an
operation which is laid out in a sequence of closely related steps,
normally uses anticipation to facilitate his action. Anticipation
as used in human performance can experimentally be shown to
have different forms, each with its own type of fundamental
psycho-physical mechanism. For instance there is a receptor-
effector form in which signals for action are accepted before the
moment of action has arrived, and prepare for such action. There
is also a perceptual form in which the signal and movement
patterns of the immediate past and of the present are used to
swing action, so to speak, towards the future. The working
assumption of the perceptual form is that the signal and move
ment patterns which have recently occurred or are now occur
ring will retain their relational characters as work continues.
It can be experimentally shown that both of these forms of
anticipation operate by producing a smoother type of control
performance, and in this and no doubt in other ways as well may have much to do with the length of continued
flight that can be sustained without change or deterioration.
Experimental evidence is now available (see e. g. Pouiton, 1950) 1". BARTLETT, F. R. S. THE EFFECTS OF FLYING 635
•which tends to show that, perhaps as a result of increasing speed
and load stress, perhaps following some unexpected break in a
regular sequence of performance, perhaps directly in conse
quence of long continued exercise, and perhaps with increasing
age, the range or span of certain forms of anticipation becomes
retracted. The smoothness of performance then suffers. Snatchy
and hurried forced reactions occur, and the rhythm which is
the best indicator of effective skill suffers. All this however is as
yet a topic in need of much further controlled research.
d) Speed, Load, Anxiety.
The three characteristics so far considered are best regarded
as the most important broad diagnostic signs of fatigue due to
difficult or long continued flying within the range of normal con
ditions. There is firm ground for saying that speed, load and
anxiety are the three most important conditions which are
likely to give rise to these signs. That these last few years have
seen an enormous potential speeding up of many special mental
and bodily activities on the part of air crew whether for civil or
military flying, is known everywhere. Increase of working speed
in the air, as in almost all other circumstances, tends to be accom
panied by increase of load. Partly as a direct result of increase
of speed and load up to near human tolerance limit, and partly
on account of a special feature of speed-load increase in the air
which I will consider briefly later, there is a considerable ten
dency for piled up, though often unacknowledged, anxieties to
accumulate for all members of air crew. Unresolved anxiety is
in fact by far the most potent conditioning agency in air fatigue.
Mackworth (1950) and Conrad (1951) have shown experi
mentally that speed and load are related but independent con
ditions in human performance. Both have tolerance limits which
can be demonstrated for a variety of working conditions, and
which do not vary much from one fit person to another within a
fairly wide range of conditions. The amount of skill deterioration
normally set up by increase of speed and load can be measured
by methods that are now available. It is also now known that
if either of these two conditions is increased or decreased by an
amount sufficient to produce decrease or increase in the efficiency
of performance, an equal and opposite variation of the others
is not normally able to maintain the status quo. 63G PSYCHOLOGIE APPLIQUÉE
Now increased speed in the air is due not only to the very rapid
and radical improvements which have been effected in aero
engines and the design of aircraft, but also to a correspondingly
rapid development of varieties of self -regulating mechanisms
largely electronic in character. Much that a little while ago
would have demanded unremitting attention from a member of
air crew is now done by recording and controlling instruments
« Load » in reference to skilled performance of any kind refers,
not to the amount of mental or muscular effort required for any
item of performance, but to the number of items which have to
be dealt with simultaneously or near simultaneously. The exten
sive use of self-regulating instruments in flying therefore can be
held to reduce the load, and so to enhance the possibilities of
increased speed. But the self-regulating instruments concerned
are often electronically complex and they may at any time, and
particularly under critical conditions, develop faults. Conse
quently a manually operated control has also, in many cases, to
be provided. The manifest load has decreased but the latent
load has not; it may even be increased. In several ways, which
every psychologist will be able to identify, this has the effect of
greatly increasing the lurking anxiety which can never be wholly
dispelled in flying. The result is to produce a number of problems
of severe fatigue in air crew, and particularly in the air-pilot,
which are to all intents and purposes, new problems, specialised
to the operation of the new types of aircraft. To deal with these
far more research, with fresh methods and more adequate mea
surements are urgently needed.
II. Diagnostic examination for fatigue.
Although it is now almost certain that nobody will ever di
scover a simple, uniform, fatigue test, which can be applied to
identify and measure either skill fatigue in general or fatigue set
up by special circumstances such as flying, this certainly does
not mean that controlled diagnostic examination for fatigue is
impossible. Recent experiments carried out on groups of civi
lian air crew before and after flight, appear to show that, pro
vided certain principles of method are observed, such examinat
ions are both possible and useful (see Welford, Brown and
Gabb, 1950). UARTLETT, F. lt. S. THE EFFECTS OF FLYING 1)37 F.
The principles are these :
(a) The operator must be given a task set in a pattern of serial
performance involving a sequence of interrelated moves both
mental and bodily;
(b) The performance must involve continued exercise, the
limits of such continuation being determined empirically;
(c) In the most significant cases it appears that the starting
signals of the performance have some element of ambiguity
which becomes progressively less and less as performance con
tinues. That is to say the operator can begin in more than one
of a known number of ways, but as he moves to the completing
stage of any cycle of manipulation the " open " possibilities
become fewer and eventually only one-way manipulation is
(d) Scoring must yield some measure of overall performance
but also some measures of the incidence and especially the,
timing of contributory items of performance. The criteria will
have to do with the relation of overall measures to the number
and timing of internal items or moves of manipulation. Whate
ver the form such measures may take in particular cases, fatigue
will always be indicated by some reduction in the internal ec
onomy of performance relative to the final level of achievement.
So far the particular settings for diagnostic examination
which have proved most significant are : for radio operators
simple practical electrical problems which must be solved in a
series of related steps; and for both radio operators and stewards
a practical grid matching task requiring transition from one set
of space co-ordinates to another related set.
The work is as yet in its very early stages. It raises a number
of unsolved problems both of experimental method and of
fundamental interpretation. These diagnostic examinations are
built directly upon controlled observation of changes of perfo
rmance accompanying continued exercise in the experimental
cockpit and in the air. They open up a new and extensive field
for research; but it is reasonable to say that they indicate that
the problem of flying fatigue in aircrew may be amenable to
experimental study to a degree that seemed quite impossible a
few years ago.
(1) Davis (Russell). — Pilot Error. — Air Force Publication. II. M.
tionery Office, London, 1948.
(2) Hick (W. É.) and Bates (J. A. V.). — The Human Operator of Control

Soyez le premier à déposer un commentaire !

17/1000 caractères maximum.