PICTORIAL PERCEPTION AND CULTURE
Do people of one culture perceive a picture differently froln people
of another? Experilnents in Africa sho,v that such differences exist,
and that the perception of pictures calls for some fonn of learning
�Picture is a pattem of lines and
shaded areas on a flat surface that
depicts some aspect of the real
world. The ability to recognize objects in
pictures is so common in most cultures
that it is often taken for granted that
such recognition is universal in man. Although
children do not leam to read
until they are about six years old, they
are able to recognize objects in pictures
long before that; indeed, it has been
shown that a 19-month-old child is capable
of such recognition. If pictorial
recognition is universal, do pictures offer
us a lingua franca for intercultural communication?
There is evidence that they
do not: cross-cultural studies have shown
that there are persistent differences in
the way pictorial information is interpreted
by people of various cultures.
These differences merit investigation not
only because improvement in communication
may be achieved by a fuller understanding
of them but also because
they may provide us with a better insight
into the nature of human perceptual
mechanisms.
Reports of difficulty in pictorial perception
by members of remote, illiterate
tribes have periodically been made by
missionaries, explorers and anthropologists.
Robert Laws, a Scottish missionary
active in Nyasaland (now Malawi) at
the end of the 19th century, reported:
"Take a picture in black and white and
the natives cannot see it. You may tell
the natives, 'This is a picture of an ox
and a dog,' and the people will look at
it and look at you and that look says that
they consider you a liar. Perhaps you
say again, 'Yes, this is a picture of an ox
and a dog.' Well, perhaps they will tell
you what they think this time. If there
are a few boys about, you say: 'This is
really a picture of an ox and a dog. Look
at the hom of the ox, and there is his
82
by Jan B. Deregowski
tail!' And the boy will say: 'Oh! yes and
there is the dog's nose and eyes and
ears!' Then the old people will look again
and clap their hands and say, 'Oh! yes,
it is a dog.' When a man has seen a picture
for the Hrst time, his book education
has begun."
Mrs. Donald Fraser, who taught
health care to Africans in the 1920's, had
similar experiences. This is her description
of an African woman slowly discovering
that a picture she was looking at
portrayed a human head in proHle: "She
discovered in tum the nose, the mouth,
the eye, but where was the other eye? I
tried by tuming my proRle to explain
why she could only see one eye but she
hopped round to my other side to point
out that I possessed a second eye which
the other lacked."
T
here were also, however, reports of
vivid and instant responses to pictures:
"When all the people were quickly
seated, the Hrst picture flashed on the
sheet was that of an elephant. The wildest
excitement immediately prevailed,
many of the people jumping up and
shouting, fearing the beast must be
alive, while those nearest to the sheet
sprang up and fled. The chief himself
crept stealthily forward and peeped behind
the sheet to see if the animal had
a body, and when he discovered that the
animal's body was only the thickness of
the sheet, a great roar broke the stillness
of the night."
Thus the evidence gleaned from the
insightful but unsystematic observations
quoted is ambiguous. The laborious way
some of these Africans pieced together
a picture suggests that some form of
leaming is required to recognize pictures.
Inability to perceive that a pattern
of lines and shaded areas on a flat
surface represents a real object would
render all pictorial material incomprehensible.
All drawings would be perceived
as being meaningless, abstract
patterns until the viewer had learned
to interpret and organize the symbolic
elements. On the other hand, one could
also argue that pictorial recognition is
largely independent of learning, and
that even people from cultures where
pictorial materials are uncommon will
recognize items in pictures, provided
that the pictures show familiar objects.
It has been shown that an unsophisticated
adult African from a remote village
is unlikely to choose the wrong toy
animal when asked to match the toy to a
picture of, say, a lion. Given a photograph
of a kangaroo, however, he is likely
to choose at random from the array
of toys. Yet one can argue that this sample
was not as culturally remote as those
described above. It is therefore probably
safer to assume that utter incomprehension
of pictorial material may be observed
only in extremely isolated human
populations.
Conventions for depicting the spatial
arrangement of three-dimensional obPICTORIAL
DEPTH PERCEPTION is tested by showing subjects a picture such as the top
illustration on the opposite page_ A correct interpretation is that the hunter is trying to
spear the antelope, which is nearer to him than the elephant. An incorrect interpretation is
that the elephant is nearer and is about to be speared. The picture contains two depth cues:
overlapping objects and known size of objects. The bottom illustration depicts the man, el·
ephant and antelope in true size ratios when all are the same distance from the observer.
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
83
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
ELECTROLUMINESCENT PANEL�
c--
__
_
__
jects in a Bat picture can also give rise to
difficulties in perception. These conventions
give the observer depth cues that
tell him the objects are not all the same
POLAROID SHEET
SPOT OF LIGHT APPEARS
TO VARY IN DEPTH
HALF-SILVERED
MIRROR
�
J
J
J
J
"
' \
, \
, \
, \
, \
�
J \ t1
\ t1
\ t1
\ 11
\t1
SHEET
POLAROID'b \�
CARD WITH
distance from him. Inability to interpret
FIGURE such cues is bound to lead to misunderstanding
of the meaning of the picture
as a whole. William Hudson, who was
then working at the National Institute
for Personnel Research in Johannesburg,
stumbled on such a difficulty in testing
South African Bantu workers. His discovery
led him to construct a pictorial
perception test and to carry out much of
the pioneering work in cross-cultural
studies of perception.
SPOT OF LIGHT
DSCALE
Hudson's test consists of a series of
pictures in which there are various combinations
of three pictorial depth cues.
The first cue is familiar size, which calls
for the larger of two known objects to be
drawn considerably smaller to indicate
that it is farther away. The second cue
is overlap, in which portions of nearer
objects overlap and obscure portions
of objects that are farther away; a hill
is pmtly obscured by another hill that is
closer to the viewer. The third cue is
perspective, the convergence of lines
known to be parallel to suggest distance;
lines representing the edges of a road
converge in the distance. In all but one
of his tests Hudson omitted an entire
group of powerful depth cues: density
gradients. Density gradients are provided
by any elements of uniform size:
bricks in a wall or pebbles on a beach.
The elements are drawn larger or smaller
depending on whether they are nearer
to the viewer or farther away from
him.
H
udson's test has been applied in
APPARATUS FOR STUDYING PERCEIVED DEPTH enables the subject to adjust a spot
of light so that it appears to lie at the same depth as an object in the picture. The light is
seen stereoscopically with both eyes but the picture is seen with only one eye. Africans un·
familiar with pictorial depth cues set the light at the same depth on all parts of the picture.
many parts of Africa with subjects
drawn from a variety of tribal and linguistic
groups. The subjects were shown
one picture at a time and asked to name
all the objects in the picture in order to
determine whether or not the elements
were correctly recognized. Then they
were asked about the relation between
the objects. (What is the man doing?
What is closer to the man?) If the subject
takes note of the depth cues and
makes the "correct" interpretations, he
is classified as having three-dimensional
perception. If the depth cues are not
taken into account by the subject, he is
said to have two-dimensional perception
[see illustration on preceding pageJ.
The results from African tribal subjects
were unequivocal: both children and
adults found it difficult to perceive depth
in the pictorial material. The difficulty
varied in extent but appeared to persist
84
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
through most educational and social
levels.
Further experimentation revealed that
the phenomenon was not simply the result
of the pictorial material used in the
test. Subjects were shown a drawing of
two squares, one behind the other and
connected by a single rod [see top illustration
at right]. They were also given
sticks and modeling clay and asked to
build a model of what they saw. If Hudson's
test is valid, people designated
as two-dimensional perceivers should
build flat models when they are shown
the drawing, whereas those designated
as three-dimensional perceivers should
build a cubelike object. When primaryschool
boys and unskilled workers in
Zambia were given Hudson's test and
then asked to build models, a few of
the subjects who had been classified
as three-dimensional responders by the
test made flat models. A substantial
number of the subjects classified as twodimensional
perceivers built three-dimensional
models. Thus Hudson's test,
although it is more severe than the construction
task, appears to measure the
same variable.
The finding was checked in another
experiment. A group of Zambian primary-school
children were classified into
three-dimensional and two-dimensional
perceivers on the basis of the modelbuilding
test. They were then asked to
copy a "two-pronged trident," a tantalizing
drawing that confuses many people.
The confusion is a direct result of attempting
to interpret the drawing as a
three-dimensional object [see top illustmtion
on next page]. One would expect
that those who are confused by the
trident would find it difficult to recall
and draw. The students actually made
copies of two tridents: the ambiguous
one and a control figure that had three
simple prongs. To view the figure the
student had to lift a flap, which actuated
a timer that measured how long the flap
was held up. The student could view
the figure for as long as he wanted to,
but he could not copy it while the flap
was open. After the flap was closed the
student had to wait 10 seconds before
he began to draw. The delay was introduced
to increase the difficulty of copying
the figure. The results confirmed that
the students who were three-dimensional
perceivers spent more time looking at
the ambiguous trident than at the control
trident, whereas the two-dimensional
perceivers did not diller significantly
in the time spent viewing each of the
two tridents.
Do people who perceive pictorial
a b
CONSTRUCTION·TASK FIGURES consist of two squares connected by a single.rod. Most
subjects from Western cultures see tbe figure a as a three·dimensional object, but wben
the figure is rotated 45 degrees (right), they see it as being flat. Africans from a variety of
tribes almost always see both figures as being flat, with the two squares in the same plane.
STICK·AND·CLAY MODELS of the figure a in the top illustration were made by test sub·
jects. Almost all the three.dimensional perceivers built a three·dimensional object (le/t).
Subjects who did not readily perceive depth in pictures tended to build a flat model (right).
"SPLIT" DRAWING was preferred by two·dimensional perceivers when shown a model
like figure c and given a choice between the split drawing and figure a in top illustration.
85
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
AMBIGUOUS TRIDENT is confusing to observers who attempt to see it as a three-dimensional
object. Two-dimensional perceivers see the pattern as being flat and are not confused_
depth really see depth in the picture or
are they merely interpreting symbolic
depth cues in the same way that we
learn to interpret the set of symbols in
"horse" to mean a certain quadruped?
An ingenious apparatus for studying
perceived depth helped us to obtain an
answer. This is how the apparatus is described
by its designer, Richard L. Gregory
of the University of Bristol:
"The figure is presented back-illuminated,
to avoid texture, and it is viewed
through a sheet of Polaroid. A second
sheet of Polaroid is placed over one eye
crossed with the first so that no light
from the figure reaches this eye. BeSPLIT-ELEPHANT
DRAWING (left) was generally preferred by African children and
adults to the top-view perspective drawing (right)_ One person, however, did not like the
split drawing because he thought the elephant was jumping around in a dangerous manner.
86
tween the eyes and the figure is a halfsilvered
mirror through which the figure
is seen but which also reflects one
or more small light sources mounted on
an optical bench. These appear to lie
in the figure; indeed, optically they do
lie in the figure provided the path
length of the lights to the eyes is the
same as that of the figure to the eyes.
But the small light sources are seen with
both eyes while the figure is seen with
only one eye because of the crossed Polaroids.
By moving the lights along their
optical bench, they may be placed so
as to lie at the same distance as any selected
part of the figure."
A Hudson-test picture that embodied
both familiar-size and overlap depth
cues was presented in the apparatus to
a group of unskilled African workers,
who for the most part do not show perception
of pictorial depth in the Hudson
test and in the construction test [see illustration
on page 84]. The test picture
showed a hunter and an antelope in the
foreground and an elephant in the distance.
The subjects set the movable light
at the same apparent depth regardless
of whether they were asked to place it
above the hunter, the antelope or the
elephant. In contrast, when three-dimensional
perceivers were tested, they
set the light farther away from themselves
when placing it on the elephant
than when setting it on the figures in the
foreground. The result shows that they
were not simply interpreting symbolic
depth cues but were actually seeing
depth in the picture.
When only familiar size was used as
the depth cue, neither group of subjects
placed the movable light farther back
for the elephant. The result should not
be surprising, since other studies have
shown that familiar-size cues alone do
not enable people even in Western cultures
to see actual depth in a picture,
even though they may interpret the picture
three-dimensionally.
The fact that depth was seen in the
picture only in the presence of overlap
cues is of theoretical interest because it
had been postulated that a perceptual
mechanism for seeing depth cues where
none are intended is responsible for certain
geometric illusions, for example
overestimating the length of the vertical
limb of the letter L. If the mechanism is
the same as the one for the perception
of pictorial depth in Hudson's tests, then
one would expect a decrease in the perception
of geometric illusions in people
who have low three-dimensional
scores.
Do people who find pictures of the
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
perspective type difficult to interpret
tend to prefer pictures that depict the
essential characteristics of an object
even if all those characteristics cannot
be seen from a single viewpoint? Here
again the first systematic cross-cultural
observations were carried out by Hudson.
He showed African children and
adults pictures of an elephant. One view
was like a photograph of an elephant
seen from above; the other was a top
view of an elephant with its legs unnaturally
split to the sides. With only one
exception all the subjects preferred the
drawing of the split elephant [see bottom
illustration on opposite page]. The
one person who did not prefer the drawing
said that it was because the elephant
was jumping about dangerously.
A GOOD
SLIDE PROJECTOR
SHOULD BE SEEN BUT
NOT HEARD.
The GAF 2690 slide projector
has a unique
noiseless operation,
a utoma tic focusing,
pop-up editor, remote
control,500-watt
brilli ance, even a
spiII -proof 100slide
tray.
In fact, it gives
you everything
b ut a lot of
noise.
Other studies have shown that preference
for drawings of the split type is
not confined to meaningful pictures but
also applies to geometric representations.
Unskilled Zambian workers were
shown a wire model and were asked to
make a drawing of it. Only an insignificant
proportion of them drew a figure
that had pictorial depth; most drew a
fiat figure of the split type [see bottom
illustration on page 85]. They also preferred
the split drawing when they were � ®
.
shown the model and were asked to _ We put 130 years ofphotographic experience into everything we make.
choose between it and a perspective t------------------,,-----------------­
drawing. Then the process was reversed, .
and the subjects were asked to choose
the appropriate wire model after looking
at a drawing. Only a few chose the
three-dimensional model after looking
at the split drawing; instead they chose
a fiat wire model that resembled the
drawing. Paradoxically the split drawing
had proved to be less efficient than
the less preferred perspective drawing
when an actual object had to be iden­
tified.
�though preference for drawings of
the split type has only recently been
studied systematically, indications of
such a preference have long been apparent
in the artistic styles of certain cultures,
for example the Indians of the
northwestern coast of North America.
Other instances of the split style in art
are rock paintings in the caves of the
Sahara and primitive art found in Siberia
and New Zealand. What art historians
often fail to note is that the style
is universal. It can be found in the drawings
of children in all cultures, even in
those cultures where the style is considered
manifestly wrong by adults.
Perspective drawings and drawings of
the split type are not equally easy to interpret.
Even industrial draftsmen with
a great deal of experience in interpreting
� ..
ecutlve
�=tial
*1500 to*8500•
A nation wide service used by
over 20,000 executive and professional
men as a quick and
convenient source of credit,
either for immediate needs or
a credit line for future use.No
collateral ...no embarrassing
investigation. All details handled
by personal mail in the
privacy of your office.
References:
The First National Bank of Minneapolis
Continental Illinois National Bank
of Chicago
•
Phone, toll free. 800-328-7328 or write
Mr. T. K. Lindblad
Executive Loan Director
fDIndustrial CreditPlan, Inc.
271 Hamm Building
St. Paul, Minnesota 55102
Our 20th
year of service to the World's
finest craftsmen and technicians.
� Nationa' Camera
.,..". I 2000 West Union A.,e. Dept. JAC
....,.. En9'ewoocl, Colorado,80rrO
.. --------- ..
I
Senda FREE copy ofthe nc Flasher
I
I�-----------------I
name
I��--------------I
I address
I
I state zip
� Nationa' Camera
""'"
I
2000 West Union A"e. Dept. JAC
� En9'ewood, Colorado,80nO
87
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
engineering drawings, which are essentially
of the split type, find it more difficult
to assemble simple models from
engineering drawings than from perspective
drawings.
One theory of the origin of the split
style was put forward by the anthropologist
Franz Boas. His hypothesis postulated
the following sequence of events.
Solid sculpture was gradually adapted
to the ornamentation of objects such as
boxes or bracelets. In order to make a
box or a bracelet the artist had to reduce
the sculpture to a surface pattern and
include an opening in the solid form, so
that when the sculptured object was
flattened out, it became a picture of the
split type. It is possible that this development
led to the beginnings of split
drawings and that the natural preference
of the style ensured its acceptance.
There is no historical evidence that this
evolution actually took place, however,
and it does seem that the hypothesis is
unnecessarily complicated.
The anthropologist Claude LeviStrauss
has proposed a theory in which
the split style has social origins. According
to him, split representation can be
explored as a function of a sociological
theory of split personality. This trait is
common in "mask cultures," where privileges,
emblems and degrees of prestige
are displayed by means of elaborate
masks. The use of these mask symbols
apparently generates a great deal of perSTYLIZED
BEAR rendered by the Tsimshian Indians on the Pacific coast of British Co·
lumhia is an example of split drawing developed to a high artistic level. According to an·
thropologist Franz Boas, the drawings are ornamental and not intended to convey what an
object looks like. The symbolic elements represent specific characteristics of the object.
88
sonality stress. Personalities are tom
asunder, and this finds its reflection in
split-style art.
Both Boas' and Levi-Strauss's hypotheses
ignore the universality of the phenomenon.
If one accepts the existence of
a fundamental identity of perceptual
processes in all human beings and extrapolates
from the data I have described,
one is led to postulate the follOWing.
In all societies children have an
aesthetic preference for drawings of the
split type. In most societies this preference
is suppressed because the drawings
do not convey information about
the depicted objects as accurately as
perspective drawings do. Therefore aesthetic
preference is sacrificed on the altar
of efficiency in communication.
Some societies, however, have developed
the split drawing to a high artistic
level. This development occurs if the
drawings are not regarded as a means
of communication about objects or if the
drawings incorporate cues that compensate
for the loss of communication value
due to the adoption of the split style.
Both of these provisions are found in the
art of the Indians of the Pacific Northwest.
These pictures were intended to
serve primarily as ornaments. They also
incorporate symbolic elements that enable
the viewer to interpret the artist's
intention. Every such code, however,
carries the penalty that communication
is confined to people familiar with the
code. Highly stylized art is not likely
to be easily understood outside of its
specific culture. Thus whereas the same
psychological processes under the influence
of different cultural forces may lead
to widely different artistic styles, the
styles arrived at are not equally efficient
in conveying the correct description of
objects and evoking the perception of
pictorial depth.
What are the forces responsible for
the lack of perception of pictorial
depth in pictures drawn in accordance
with the efficacious conventions of the
West? At present we can only speculate.
Perhaps the basic difficulty lies in the
observers' inability to integrate the pictorial
elements. They see individual symbols
and cues but are incapable of linking
all the elements into a consolidated
whole. To the purely pragmatic question
"Do drawings offer us a universal lingua
franca?" a more precise answer is available.
The answer is no. There are significant
differences in the way pictures
can be interpreted. The task of mapping
out these differences in various cultures
is only beginning.
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC
This memol'Y is
I'ight in the gl'oove.
Here's a small part of
our new 1024 bit random
access memory. The veeshaped
grooves are
what's new. They provide
air isolation between
active components on
the chip.
This exclusive V-ATE
process reduces individual
circuit cells to a tenth
the size of those in conventional
bipolar devices.
The result is a low-cost
memory with an unmatched
combination of
high speed, low power
dissipation, high bit density
and small chip size.
For data processing
designers this means high
performance systems at
lower cost.
We bring this kind of
creative engineering to
all aspects of our microcircuit
capability - from
high volume commercial
devices to.modularized
military subsystems.
This is just one of many
technologies where we're
finding lower cost solutions
for today's system
requirements. Raytheon,
the Systems Managers.
Raytheon Company,
Lexington, Mass. 02173.
t[AYTHEO:J
© 1972 SCIENTIFIC AMERICAN, INC© 1972 SCIENTIFIC AMERICAN, INC