AN INVESTSGA‘TION OF TH£
ENFLUENCE OF A FIGUREflROUND
UPON THE PERCEPTION OF IT'S

CONSTITUENT ELEMENTS

‘f’huis fer i‘ho D‘gru of M. A.
MICHQGAN STATE COLLEGE
S’swnay ‘W; Weaiharhoad
W48

 

Thisistocertifgthatthe

thesis entitled

"An investigation of the influence of a
figure-ground upon the perception of its
constituent elements"

presented by

Sidney W. Weatherhead

has been accepted towards fulfillment
of the requirements for

Ma s ter mafia m—Bsychology

flame/€37

Major professor

Date / /75/_

 

‘4';

5N INVESTIGATION OF
THE INFLUENCE OF‘A FIGUREFGROUND
UPON THE PERCEPTION
0F IT'S CONSTITUENT
. EEEMENTS

BY
SIDNEE’W} WEATHERHERD

A.THESIS

submitted to the School of Graduate Studies of Michigan
State College of Agriculture and applied Science
in partial fulfillment of the requirements
for the degree of
IISTEB OF.ARTS
Department of Psycholog

1948

Smears

 

 

ECKNGLEDGEENI

The author wishes to acknowledge the very kind and patient
guidance of Dr. Maurice R. Denny under whose supervision

the aperimental work to be discussed was carried out.

2881 9 O

meme 9; Contents

SubJect

The Problem and; It's Setting
lpparatus

Procedure

l)iscus si on of Results

Conclusion

References

12
19
20

‘ - 93 protrem an; 1333 sectgg

' rhe present study is an attempt to test the hypothesis
tint the perception and learning of a constituent part at a
total field till be facilitated when the field and the part
have identical characteristics and inhibited when these
characteristics are not identical.

This hypothesis was formulated as a restatment of a
hypothesis derived from the emergent behavior theory of social
science developed by ltr'. Ergo Engelmann of Michigan State
College. The following statements are taken from an un-
published paper by m. mam; ‘

'Ihe construction of a unified theoretical social
science seems to demand the acceptance of the theory
of configurational emergentism as a description of
behavioral dynamics. This theory asserts that every
behavior pattern must be considered at the same

time as an emergent at the intersection of two or
more, less complex behavior patterns, and as a dif-
ferentiation from a more complex field. Every be-
havior pattern must also be considered a new differ-
entiation and a new emergent at the time of its oc-
currence, regardless of how many times an identical
behavior attern has been performed before by the
same sin e biological organism. If this biological
organism has performed this behavior pattern before
we may speak of a re etitive pattern, otherwise of
a nan-repetitive pat rn.‘

'Ihe hypothesis selected for experimental investigation
can be stated as follows. The newly emerging non-
geggtitive behavior pattern re-emer s as a rcpetitive
e vior pattern, because it consti utes a behavior
pattern of such a nature that if it intersects with
a different or identical behavior pattern, a new dif-
ferentiating field emerges in Which the original nar-
repetitive behavior pattern can again emerge as a
ripetigiye behavior pattern at a higher level of can-
P ex 0

The task of the author was shaply one of unbiased exper-
imentation to lay bare the facts as to what visual pattern would
emerge most readily and to discover the possible reasons for
its emergence. (Visual patterns were used to represent the
intersecting behavior patterns discussed by mr. Englemann).
The goal was either to substantiate the theory or to supply
data for its possible revision.

Antecedent to this investigation are many experiments on
form perception. However, past experimentation is not directly
related to the problem under investigation although it does
furnish a background from which the results of this eXperiment
might be predicted.

Most of the experiments in the past have been performed
with the purpose of substantiating or demolishing some tenet
of Gestalt rheory. whose experiments have in general led
to positive evidence. for the following laws of field organ-
ization ( e ).

L. Law of Pregnanzx Psychological organization moves
toward good gestalt, e.g., stability, simplicity,
and regularity.

2. Law of Proximity: Old impressions are less well
recognized and recalled than new ones because the
recent trace is nearer in time to the present active
process.

3. Law of Closure: Closed areas are more stable than

unclosed ones, and therefore more readily form

figures in perception.

4. Law of Similarity: Similar "items (form, color,

etc.) or similar transitions (alike in steps
separating items) tend to form groups in perception.

5. Law of good Continuation: Organization in.perception

tends to occur in such a.manner that a straight
line will continue as a straight line, a circle
as a circle, etc.

In general, Gestalt Theory has been based on the assump-
tion that experience is made up of definite configurations
in which a figure protrudes from a more or less undifferentiated
background. It is further postulated that “A structure once
present in experience creates favorable conditions for its
own reappearance or that of a similar structure (5’)." ,In
traditional Gestalt Theory figure and ground present two as-
pects of a total field one or which predominates while the
other sets it off.

In the configurational emergentist theory here postulated,
the concern is not with a clearly perceived form and an un-
differentiated background but rather the emphasis is upon
the relationships existing between two or more patterns of
behavior experienced simultaneously one of Which.may emerge
from a more complex field.

The experhmental set up used in this inveStigation.pro-
vides for partawhole relationships in that there is a homo-

geneous ground for the more complex configuration out of

which the small figure differentiates. As an example the
triangle would perhaps serve best. The intersecting phenomena
are the three lines of the triangle, the differentiating field
is the background field. In a triangle made up of triangles
the intersecting phenomena are triangles which intersect

in such a way that in the differentiating field the Spatial
arrangement of a triangle appears. As far as the little tri-
angle in this signation is concerned it differentiates from

a field which in turn differentiates from a still larger

field attaches“ the same structural properties as the emergent
phenomenon (the little triangle) which emerges at the inter-
sections of other phenomena, 1.6., lines. The intersections
are not thought of as taking place between the large and the
small figure. The large figure constitutes a field and not

an intersecting phenomenon. The difference with traditional
gestalt experiements lies in the fact that the ground in our
experiment for the small figure is a highly organized and form.-
like field.

Our problem is to determine whether or not the parts
will be recognized more effectively when emerging from a
figure-ground of identical form characteristics than when
emerging from a figure-ground of different form character-
istics. ‘

Apparatfis

Thirty-six slides each consisting of a different com-
bination of figure-ground were used. Emery large figure was
made up of every small figure and each small figure had it's
counterpart in each large figure - see Figure 1. Six figures
were used and each figure (lagge) was made up of every other
figure (small); The unidimensional ratio of large figure to
small figure as drawn is 8% inches to 1 inch. Therefore, the
ratio of large figure area to small figure area is (8%)g/l.
The figures when projected upon the screen were about the same
size as before photographic reduction to a size suitable for
the making of the slides.

. The slides were made by placing the positive photographs
on stiff white cardboard which furnished a homogeneous back-
ground on Which the field and it's elements appear.

A Balopticon Opaque projector was used to project the
designs upon a white canvass. The projector was placed 133
inches from the screen. The aperture was opened to maximal
'width. The lens was adjusted to give such a focus that the
perception of the small figure was liminal. After prOper ad-
justment was obtained the lens was kept at this point through-
out the eXperiment. A thumb-tack was placed on the canvass

as a fixation point. ‘mxposure time was controlled by means e
of a Keystone tachistosc0pic fixture called a Flashmeter.

Each.design was eXposed for exactly.l/5th of a second.

‘000000
0

.00000
Q

mmmmmflddfinbm

 

10

Toocooc_s

: <<l
qqqqqq,
<1<<14<1<T
<I<l .
QQ
QC]
g QC] .
CIGCICICIC] r-
CIQQQQC}.
‘ QC]
h QC]

‘ <4
;. <1<1
:<<<<<<
f<<<<<<

teenage
manage

3%
i an
smegma
immense

gadQGQQ~
gdddflflfl
1 an;
i dd
.anmdd
idddfldd’

_._ ‘—-—-——-r— » —
,,i__————-‘———'

Ioooaoo‘
lfifiGflQfi

on;
e co»
menace;

{£22229
000000
000000,
+ 00

00
000000'

900000

 

18

17

16

15

Figure I. The Desigrs in Numerical Order

 

 

— - e
{QQQQ a
Q

id <1
U dd
dflflflg
G .
a 3'
QQGQE'

4 Eli];
'EEEE E,
* E
E E]
5,3333%-
D

DD -
[IUD a:
[3 EJ'
DDDDDB
['60
30000 O.
O

O 0
OO '
OO O~

L ___ _99
#0
$Q%#.¢,
1'43” "
a 9f
#fififig.

 

23

21

20

19

DD
D
D
Di

_~___.

D
Q

n
nnngnn*

l
l

1

fl

o
~cocoon

0
Gfl
@ Q

@
fi

\
l
I

A
{AAAAAA

A
AA
A A

A

. A
A

D+'

E1 1 A
D 1
[:1 .

D
D
D
D
D
LDDDDDD‘

 

000000.

30

29

28

26

25

nmmemmmmd

ggfigofifzgg

 

an an NM!
QQQG J EDD—U 00000
G D D 0
G D flu £0
GGQQ DDDD W 000
mm

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Procedure

Forty-one college students were used as subjects. Most
of these subjects were undergraduate psychology majors and a
few were graduate psychology majors. The subjects were seated
approximately from eight to sixteen feet away from the screen.
They were tested in groups of from three to fifteen members.
A record blank - see Figure 2 - was handed to each subject.
The following directions were then read:

LfListen carefully. You will be shown a series of
designs. Each design will consist of a large figure
made up of numerous small figures. mch design will
be shown for only a fraction of a second. You are to
try and observe both the large figure and the small
figures in each design.

Ch the record sheet that has been given to you
place a check under the figures you observe during
each trial. You are to check the large figure ob-
served under the correct figure in the column headed
LARGE FIGURE You are to check the small figure ob- -
served under the correct figure in the column headed

A SMALL FIGURE Thus for each eXposure you will have
two checks to record, one for the large figure ob-
served and one for the small figure observed.

If you are in doubt as to what the large figure
is or what the small figure is, then guess. But, do
your best to record two checks for each exposure.

Each eXposure was preceded by the commands: ”Ready; "
'Watch;" and then the design was flashed. Just before the
experimnt was started for each group, the illumination of
the room was dimmed to twilight vision conditionS. ,

The slides were numbered as shown in Figure 1. These
numbers were then picked at random in the order presented

in Table '1’... This random order made up the first series.

 

 

 

 

 

 

'L- no LanEE FIGURE L- ,SWLLL FIB!
‘ 1 H1 IHL L_L',+ E] O A BELOW: U707"

I
w ‘

 

 

 

 

O-.. p «7.4.; v-Cq

 

 

 

 

 

 

 

 

I
- ._iJ.--__L
-1- .L

 

“F”

 

 

 

 

—o

 

 

 

L

 

ELL“) cowoe UN: (.0 Pk)“—
-J
_._L -1

 

 

 

 

.A. f\.)
L— —-—4—-
“1

 

F

C

 

 

 

 

 

{A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. Figural.
Form Used By Subjects to Record the Figures Seen

Table 1'...

Order of Presentation of Stimulus
Cards by Groups and Individual

 

 

 

 

Designed
fife—r16 8___Gmrm_u 8 in W

I. I II. III. IV. V. VI.

1' 10,26. ‘11-’28, 2.14. 32.23. 19.22. ‘.33.
309 3. [31.35. 29.17, 56.16. 27920. 8.21.

250 5.H18.34' 12079 24.1, 13s 9s 15sec

I13 Iv. III. II. V. VI. I.
III. v0 III. II. No 10 VI.

 

 

 

 

 

 

 

 

 

a Note: Numbers refer to numbered
designs in 1"igure 1.

This order was divided into six groups of six designs each. The
groups were then picked in two different random orders making
up the arrangement for trials two and three. The order within
the separate groups remained constant but the groups themselves
were always arranged in random order.

It should be noted that the slides made up of a large
circle and a small cross and a large circle and a small U were,
by error, made up into slides containing figures too large for
use 'in the experiment. — However, the slides were left in the
experiment to serve as interesting relief exposures to keep
the interest and morale of the subjects at maximum. But, in
analysing the data these two slides were eliminated. Because
of this certain ”comparisons could not-be made.

Discussion gf'nesultsh

The data, consisting of 8,85; reaponses,has been analyzed
several different ways and the results obtained are discussed
below. An analysis of the data reveals the following information.

In Table II. we see that there is no evidence in support
of the hypothesis, although, as will be indicated later, a
more detailed analysis of the data shows that the evidence
against the hypothesis is not particularly conclusive.

The small figure was seen correctly in combination with
an identical field 54% of the time. The small figure was seen
correctly in combination with a field different in character
dog: of the time. '

This is a difference of a; against the mothesis in a situation
where the perceptibility is roughly 503% accurate or liminal,
that is, in a situation presumably optimal for the disclosure

of a variable relevant to the perception of the small figure.

- The 6% difference against the hypothesis is significant beyond
the 1% level. of confidence, as based upon the standard error

of the difference between twO uncorrelated percentages. The
It‘s used were 738 and 3,444.

- It should also be noted that the high percentage of acc-
urate perceptions of the larger figure - see Table II - means
that the variable (large figure) was potentially operative.

An analysis. of variance gives an F Ratios. of 2.34 which is
only significant between the 5% and 110% levels of confidence.

Table I 1.

Percent With Which All «Figures are Perceived
Correctly for All 41 Subjects

Difference
A i -Grnd Series Series 1;; 1530mm! 3’3,qu gni'iggflge
Small Fig 30
Identical 48% 51% 3% 54% 9[|.%

 

shall Fi
Different 49‘ 51% 13% - 60% i 92% L

 

 

 

 

 

 

 

' There is considerable variability in the perception
of the small figures. Their perception ranged from 4
correct to 74% correct as may be seen by reference to
Table 111 . In order of least difficult to perceive to
most difficult to perceive, the small figures rank as
follows: triangle, square, trapezpid, "U“. circle, and
cross - see TableIV . If we compare the ease with which
the small figure is perceived with the accuracy with which
it is perceived, as may be done by examination of the last
two rows in Table III., we see that, in general, the relative
order is about the same.

The accuracy with which the small figure is perceived
is about the same for all figures except for the cross which
was well below threshold in all reapects. In other words,
guessing was not an inportant factor in determining the re-
sults as, in general, the most perceived small figures were
also seen most accurately. Also the low percentage accuracy
of the cross further indicates that the data on the cross
cannot be validly compared with the data on the perception
of the other small figures.

The background figures are about equal in their general
influence on the perception of the small figure as may be
seen by inapection of‘ithe last column on the right in Table
Neglecting the data for thecircle background, since the data
for the most difficult small figures to see are eliminated,

Table III._

each Small figure Is Seen Correctly With
Bch Large Figure Seen Correctly «-

Frequency

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

u+noAAm°%
. , ct ° .33.—
L__.| 51 6 101 16 96 62 352 738 45%
I ' 66 5 92 57 91 81 5‘70 738 50%
D 64 6 97 lo 88 53 318 758 43%
‘ Not Hot .
0 sad Use 76 23 95 77 271 492 55%
A 68 3 93 51 108 or see 738 53%
A 76 ,3 77 19 66 78 319 .768 45%
Total
Correct 325 21 541 156 544 412
Total
Exposed 615 615 738 .738 758 758
g? , ..
Correct 53‘ 4‘ {7d 21% 74% 56%
93 Accura 1
as 59% 13% 55% 59% 65% 61¢
a» Note: Each cell represents the small figure
frequency for any particular figure-
ground for 41 subjects for all three
. series. .
*4:- Note: %Accuracy was found by; Total Borrect/Toral Reaponse:

Table IV

Rank of Perceg tibility Compred With
Position of I entical Element - Snell

Figure
Bank of ‘ Position of
Large Pe ,cepti- Order of Perceptibility Identical
biiLity'S" From High to Low Element

 

 

 

Inland. + ‘ 1

 

2 IDALLA + 1

 

 

InDiLLA —|— 1

 

4 IDA ALJ.+ 4

 

 

 

 

5 {AA LC) 4-6

 

+OEDD>E

 

 

 

6 IDAALJO-l-

 

 

we see that the backgrounds seem to influence the perception
of the small figures in some general way in descending order
as follows: triangle, cross, "U" trapezoid, and square.
However, the difference in their effect is only slight .-
range from 45% to 55%.

Mamination of Table III. shows that except for the circle,
each set of small figures was. seen about equally often in each
background. A more Specific analysis, on the other hand, in-
dicates the possibility of some significant relationships.
the such relationship is that‘figural contrast seems most
effective when the small figures are either below or at the
visual threshold. For examme, the small circle which is con-
siderably under visual threshold‘is seen best when the triangle
- a sharply contrasting figure - serves as the background,

(see Table, III). The circle and the triangle appear to be
highly contrasting in structure. The circle is a continuous
unbroken closed figure within which every point is of equal
accentuation. On the other hand, the triangle is a figure
cmnposed of three straight lines joined sharply at three dis...
tinct points all equidistant from each other. There is far

more accentuation at the exact points than between the points.
Furthermore , the three ankles ofthe triangle contrast with
the continuity of the circle.

when the small figure is well above threshold, there
is evidence that identity of small figure and figure-ground
is a factor'in the perceptibility of the small figure - see-
bottom of Table III. The small triangle is seen best with
the large triangle and the small square and trapezoid are
the most often perceived small figures with the large square
and trapezoid bacgrounds reSpectively. I ‘

Similarity - not identity - where it can be most dlearly
drawn, as between the triangle and the trapezoid seems to
show inhibition of perception. Disregarding the cross and
the circle the trapezoid is the least often figure perceived
in the triangle background and the triangle is the least often
seen in the trapezoid background.

(From Table I,'we see that, although there is not much
learning from session.to session, there is 9% more learning
when the figure-ground is different than when identical.

Advocates of Gestalt Theory often cite the circle as a
model example of good gestalt configuration. Hewever, the
circle in this experiment does not stand out over the other
figures in relative perceptibility even thOugh it is seen
about as accurately as the other figures. Actually, when .
the cross is eliminated, the circle is seen fewer times than any
other figure. This may be explained in part by the fact that

the circle was quite often mistaken as a diamond.

However, in the framework ofiMr.‘Englemann's theory it might
be interpreted as follows: The variant behavior of the circle
might be eXplained in terms of the greater complexity of the
intersecting phenomena. With a triangle you have lines inter-
secting in a field Which is a plane. If you take the circle
you have gg§,line intersecting with a planecin a field which
is a.more complex plane. If we consider the plane as a phe-
nomenom.more complex than a line, the difference in.recognition
can probably be explained.

‘COnclfiSidfi

In.a situation presumably Optimal for the disclosure of
a variable relevant to the perceptibility of the small figures,
the results with reference to perceptibility and learning do
not support the hypothesis that the small figure will be per-
ceived more readily when emerging from a field of identical
form.

By reference to the bottom of TableZII it seemSchat
contrast is most operative when the small figures are at or
below threshold and idientity to be more operative when the
figures are above threshold.‘ Therefore, further research to
retest the hypothesis with all figures higher above the limen

seems warranted.

1.

2..

5.

9.

10.

11.

- 20 -

References

 

Bartlett, F. C. An eXperimental study of some problems
of perceiving and imaging. Brit. J. Psychol., 1916,
8 ’ 222-226 0

 

Brigden, R. L. A tachistoscopic study of the differ-
entiation of perception. Psychol. Monog., 1955, Mb,
155-166.

 

Fehrer, E. V. An investigation of the learning of
visually perceived forms. Amer. J. Psyphol., 37,
187 -221 o

 

Gibson, J. J. The reproduction of visually perceived
forms. J. Exper. Psychol., 1929, 12, 1-59.
I

Helson, H. Psvchology of gestalt. Amer. J. P3 chol.,
56. 1925. aha-57o, Len-526; 57. 1M . 5-0 . - 25.

Hilgard, E. R. Theories of Learning. Appleton-Century-
Crofts Inc., I9h8, New‘YOrk.

 

Koffka, K. Principles of Gestalt Psychology. 1955.
Harcourt and Brace and Co., New York.

Kohler, W. Principles of Gestalt Psychology. 1929.
Liveright Publishing Corp., New York.

‘--.

 

Perkins, F. T. Symmetr in visual recall. Amer. J.
Ps c1101.. 1932. at. 754190. "“""‘—

Vever, E. G. Figure and ground in the visual perception
of form. Amer. J. Psychol., 1927, 38,119h-22 6.

 

Woodworth, R. S. Egperimental Psychology. 1958. Henry
Holt and Co., NewfiYork. 1“

 

fit-4’51? ailing! E.wal.nm§ll. 35.VVJ.$.1.F .E. -.

I.

.. [x'tllla-Bw‘ ,3 .

 

"'TIWITHILHIL 11;" MN 11 injlijfy‘flafit'mmfimi'“