\J lilt‘dlUlHMWNWWWI

 

meets
‘ j r) LIBRA'R YVL‘

m”1111;11ng flfllfweiliflhfliflll‘ i Mich I
University ,

This is to certify that the

thesis entitled

&'u\zowr<emxc liOTEmAc-nop
AKD THE
CONCEPT OF THE NMBLE BIND

pre seentd bg

\XERBEH m. pomsfi

has been accepted towards fulfillment
of the requirements for

M degree in_'l_ 5:! C E] i QLOG‘f

flflffiifl

Majorroesspf

DaJ/lawfi/Bi (m

 

 

 

 

 

 

w".-
UM.

\Pf
v"'...g
Ifll‘.

Y

ABSTRACT

SCHIZOPHRENIC INTERACTION
AND THE
CONCEPT OF THE DOUBLE BIND

by Herbert M. Potash

This study sought to provide empirical validation
for the "double bind" hypothesis. This theoretical
viewpoint considers schizophrenia the end product of a
particular parent-child relationship in which the parent
.constantly communicates two contradictory messages to the
child, and the child cannot make any response which will
be positively reinforced. His ultimate solution to this
situation is thought to be his schizophrenic reaction.

The essential features of the double bind were simu-
lated in a two-person three-choice game. The optimal
game strategy, resulting in the highest monetary gain,
could only be achieved by choosing in such a manner as to
lose money every other trial (while reinforcing the other
player) and hoping that the other player would reciprocate
on the following trials. In following this strategy a
player could not know beforehand whether he was to be
(ultimately) rewarded or would lose money continually.

Thus the inability to discriminate between the alternative

Herbert M. Potash
meanings of another individual's behavior is common to
both situations. One of the possible game choices was a
situational withdrawal in which a player could neither
gain nor lose money. This was felt to be similar to the
schizophrenic's withdrawal which the theory states is
.prompted by frequent double bind exposure.

A control task was also utilized to insure that it
was specifically double bind frustration which the task
measured. The control task was a light guessing situ-
ation in which subjects were given money each time they
correctly predicted which of two lights would go on, and
lost an equivalent amount of money for each incorrect
guess they made. They were also permitted a withdrawal
reSponse, where they would neither gain nor lose money.

The control task was given to 14 normals and 14
undifferentiated schizophrenics. Fourteen pairs from
each of these groups were given the double bind task for
a total of 60 trials. The schizophrenic and normal
subjects were comparable in age, education and intel-
ligence.

The schizophrenics utilized withdrawal to a greater
extent than normals in both eXperimental and control

situations. While both groups withdrew more frequently

Herbert M. Potash
to the experimental task, attesting to the greater degree
of threat posed by a double bind situation, both groups
showed comparable increases in frequency of withdrawal
responses over the performance of their matched control
groups. Thus, the increase in withdrawal to a double
bind situation was neither exclusive nor differential in
the schizophrenic group. However, by choosing nonco-
operatively, the schizophrenics exposed themselves less
frequently to the double bind situation than did the
normals, though their actual frequency of exposure was
still quite high. The schizophrenics withdrew less with
time and tended to withdraw when they had the greatest
control over their earnings (hence, being less likely to
be placed in a double bind). In contrast, normals with-
drew more frequently when they had least control over
their earnings. Since the basic hypotheses of this
study were not confirmed, it is concluded that the double
bind poses no distinctive threat to schizophrenic indi-
viduals.

The individualistic, non-co-operative strategy of
the schizophrenic subjects is best explained by their
lack of interpersonal trust. None of the schizophrenic

subjects settled on a game strategy which involved a

Herbert M. Potash
permanent committment to the other player. The schizo-
phrenics were apparently motivated by a desire to avoid
such committments rather than to avoid the specific
threat of the double bind. The double bind is then
considered to be only one of many possible interpersonal

situations which may result in a lack of trust.

SCHIZOPHRENIC INTERACTION
AND THE

CONCEPT OF THE DOUBLE BIND

By

_ K"
Herbert M9 Potash

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Psychology

1964

ACKNOWLEDGMENTS

No research study is entirely an individual effort.'
This dissertation reflects the assistance of many people,
only some of whom I can directly acknowledge below. To
those not specifically mentioned, my sincere thanks for
all your help.

My greatest debt is to my wife, Jan, who has helped
in innumerable ways: from the many suggestions she
offered to help clarify the design, through the typing
of several revisions until the final copy, as well as the
extensive moral support she gave throughout the project.

I wish to express my deepest gratitude to Dr. Bill
Kell who was chairman for this dissertation. His many
suggestions were of immeasurable help, especially in
regard to data interpretation; and his comments have
greatly helped clarify the text of the paper.

I am very grateful to Dr. Stuart Armitage of the
Battle Creek Veterans' Administration Hospital for
extending me the use of the hospital facilities during
the research project. He was also of great assistance
in providing employees and patients to act as subjects
for the study.

I wish to thank my committee members, Drs. C. L.
Winder, M. Ray Denny, and Bertram Karon, for their

ii

comments concerning the dissertation. I am particularly
grateful to Dr. Karon for his suggestion of introducing
a control task in the procedure, which helped make the

results of the study more meaningful.

iii

TABLE OF CONTENTS

LIST OF FIGURES . . . . . . . . . . . . . . .
LIST OF TABLES . . . . . . . . . . . . .
Chapter

I. INTRODUCTION AND HISTORICAL BACKGROUND

General Introduction . . . . . . .
Hypotheses . . . . . . . . . .

II. METHODOLOGY . . . . . . . . . . . . .
Apparatus . . . . . . . . . . .

Subjects . . . . . . . . . . . . .

Procedure . . . . . . . . . . . . .

III 0 RESst 0 O O O O O O O O O O O O O 0
IV. DISCUSSION . . . . . . . . . . .

V. SUMMARY . . . . . . . . . . . . . .
BIBLIOGRAPHY . . . . . . . . . . . . . . . .

APPENDICES . . . . . . . . . . . . . . .

iv

'14

Page

vi

16

16
17
18

. 24

3O

. 4O

45
46

LIST OF TABLES
Table Page
1. Frequency of withdrawal responses (C) by
schizophrenic subjects over 60 trials
under control and experimental conditions . . . 24
2. Frequency of withdrawal responses (C) by
normal subjects over 60 trials under

experimental and control conditions . . . . . . 25

3. Frequency of withdrawal responses to the
control task . . . . . . . . . . . . . . . . . 25

4. Summary table for analysis of variance . . . . 27

LIST OF FIGURES
Figure Page
1. Game choices and returns . . . . . . . . . . . 12

2. Sample sheet . . . . . . . . . . . . . . . . . 20

vi

LIST OF APPENDICES
Table Page
5. Raw data for schizophrenic control subjects . . 46
6. Raw data for normal control subjects . . . . . 46

7. Background data on eXperimental schizo-
phrenic subjects . . . . . . . . . . . . . . . 47

8. Background data on exPerimental normal
subjects . . . . . . . . . . . . . . . . . . . 48

9. Game behavior of schizophrenic subjects
under simultaneous trials first condition . . . 49

10. Game behavior of schizophrenic subjects
under nonsimultaneous trials first condition . 50

11. Game behavior of normal subjects under
simultaneous trials first condition . . . . . . 51

12. Game behavior of normal subjects under
nonsimultaneous trials first condition . . . . 52

vii

1

Schizophrenic reactions are characterized by a
"disturbance in reality relationships and concept
formation, a retreat from reality, emotional disharmony
and unpredictable behavior" (American Psychiatric Asso-
ciation's Diagnostic Manual). However, the widespread
divergence in the application of this definition is
attested to by studies (7,10) demonstrating low relia-
bility in patient classification. Perhaps one of the
reasons for this disagreement lies in the inability to
qualify this definition in any way. That is, does
retreat from reality mean an unconditional retreat as
evidenced in a catatonic stupor, or a conditional retreat
dependent upon the immediate situation which the individ-
ual faces. The inability to answer questions such as
these seems in part responsible for the lack of consen-
sual agreement on classification. It appears that only
through an extensive study of the people labelled schizo-
phrenics can we arrive at an understanding of the common
behavior pathology which justifies the use of this diag-
nostic label.

Even greater disagreement is encountered when the
question of the etiology of schizophrenia is considered.
A myriad of theories abound in the literature, all

attempting to explain the factors responsible for

2

psychoses; these range from a strict hereditary position
(arguing that psychoses are genetically induced)(3) to
strict environmental explanations (conceiving of partic-
ular interactions as causative agents, responsible for
the lack of interpersonal trust and retreat from reality)
(14). A new kind of approach to the area is offered by
Ruesch (11) who considers mental illness a disturbance
in communication. "The condition which the psychiatrist
labels 'psychosis' is essentially the result of the
patient's misinterpretation of messages received" (11,
p. 88). The misinterpretations are thought due to the
particular experiences the individual has undergone. The
recent work by Jackson, Bateson, Haley, Weakland and Lidz
involving intensive study of the families of schizo-
phrenics, including long range behavioral observations,
family therapy, group therapy and individual therapy has
resulted in the conception that the familial interaction
is the precipitative agent in schizophrenic reactions.
Since this theoretical position has recently been
expounded, it still does not have any experimental veri-
fication. It is the aim of this study to put some of
these authors' conceptualizations to an indirect test.
While the aforementioned writers are using somewhat

different approaches, there is a great deal of consensus

3

in their resultant theorizing. Basically, they all feel
that schizophrenia is a product of a disturbed familial
interaction. Lidz (6) observes two basic kinds of rela-
tionships between the parents of schizophrenics which he
labels the schism and the skew. The schism represents a
direct and open split in which the parents are constantly
at odds with each other. In the skewed relationships one
parent, generally the mother, assumes an overadequate
role without the encouragement of her mate, whereby she
exercises the power in the family even though incapable
of wielding it effectively. None of the authors have
reported cases in which one or both of the parents are
considered emotionally healthy individuals. Jackson and
’Weakland state "the (schizophrenic) patient's symptom of
withdrawal is largely shared by the parents but ordinar-
ily they have better conventional covers for this" (5,

p. 620).

Prior to the birth of a child the parents of schizo-
phrenics are thought by Bowen (3) to engage in a very
cyclical relationship where they move from extreme
closeness, which creates discomfort or anxiety within
them, to aloofness, which they also view as unfavorable.
When the child is born, the father becomes psycholog-

ically distant and isolated from both mother and child;

4

and the mother-child relationship becomes very emotion-
ally intense. Bowen's description would seem to
correspond to Lidz' conception of the skewed parental
relationship, but he does not address himself to families
characterized by a schism.

Jackson and Weakland outline the common core of the
schizophrenic family as containing"(1) a child whose
mother becomes anxious and withdraws if the child
responds to her as a loving mother... (2) a mother, to
whom feelings of anxiety and hostility towards the child
are not acceptable and whose way of denying them is to
express overt loving behavior to persuade the child to
respond to her as a loving mother and to withdraw from
him if he does not... (3) the absence of anyone in the
family such as a strong and insightful father who can
intervene in the relationship between the mother and
child and support the child in the face of the contra-
dictions involved" (5, p. 236). Thus, the child is
constantly exposed to two levels of messages from his
mother, one of which covertly contradicts the other.
Moreover, the child cannot comment on the messages to
correct his discrimination and determine which level of
message he should respond to. He is therefore placed in

what these authors call the "double bind." An example of

5

the double bind which the authors cite is the case in
which a mother visits her schizophrenic son at the
hospital and he impulsively puts his arm around her. She
stiffens and he withdraws his arm leading the mother to
comment, "Why did you take your arm away; don't you love
me?" Frequent double bind exposure is what Jackson and
Weakland consider the causative agent in schizophrenic
behavior.

When faced with the double bind situation the child
cannot leave the field because of his dependency on his
mother. "This dependency is fostered by other messages
of double bind communication to a degree far beyond the
physical or emotional 'realities' of the person's current
life situation" (16, p. 376). While it is necessary to
respond adequately to the communication, "an adequate
response is difficult to achieve because of the
concealment, denial and inhibition inherent in or added
to the basic contradictory pair of messages" (14, p. 377%
Of the three kinds of responses the child can make (a)
labelling the incongruity verbally (b) giving a double
bind message in return or (c) giving a humorous response,
the schizophrenic child is most likely to choose the
second probably because it is the only one reinforced.

Bateson feels that in order to maintain the homeostasis

6

in the family, one member, the schiZOphrenic individual,
must sacrifice his normality and remain confused in this
interaction. "The boundary of sanity is, however,
crossed when the subject uses these tricks of communi-
cation in situations which the common man -- one
hesitates to say the normal -- would not perceive as the
schizophrenic seems to perceive them" (1, p. 134-135).
While weakland and Jackson consider repeated
exposure to a double bind situation the determinant of
schizophrenic behavior, they do not quantify this or
tell us how frequent this exposure must be to produce
a schizophrenic reaction. Perhaps Bowen's theorizing
(3) will be of help. He feels that the period of ado-
lescence threatens the mother-dependent child rela-
tionship for the child then attempts to achieve a balance
between his close relationship with his mother and the
demands of the adult world. The inability of the mother
to prepare the child for adult living through her
constant use of the double bind makes it impossible for
him to deal effectively with reality and prompts the
schizophrenic withdrawal. Perhaps in the case of child-
hood schizophrenia, the exposure to a double bind has
been severe enough to prompt the schizophrenic withdrawal

at that point.

7

The concept of the double bind can also be stated in
the language of social psychology. It can be charac-
terized as a power relationship between the mother and
child in which the mother, by virtue of her ability to
manipulate the rewards desired by the child, exercises a
great deal of authority in this relationship. Her
contradictory attitude toward the child makes all of his
behavior inappropriate, in that it fails to lead to
reward. Such actions generate a feeling of futility in
the child, and this generalizes so that he is unable to
gain any need satisfaction outside the home. He becomes
schizophrenic either as a means of gaining the necessary
need satisfactions (Bateson) or in order to break from
his dependency upon his mother (Haley). Perhaps both
are involved, but the overt symptomatology of the schizo-
phrenic which we encounter involves a withdrawal from
interpersonal communication.

Mills' study of "Power Relations in Three-Person
Groups" (8) offers some experimental data which may be
extrapolated and applied to family relationships. He
noted the tendency in an ad hoc three person group to
split into a pair and an isolate. He characterizes the
relationship between members of a pair as solidary (where

they are above the median in support of each other),

8

contending (where the less active member is below and the
more active one above the median in support), conflicting
(where they are below the median in mutual support), and
dominant (where the more active one is below and the less
active one above the median in mutual support). A
further exPeriment by Mills, once again with ad hoc
social groups, led to the formulation that "one important
personality condition affecting the persistence of the
coalition structure is the level of anxiety that iso-
lation generates; the higher it is, the more persistent
the structure; the lower, the more likely is the coa-
lition structure to dissolve" (11, p. 666).

Applying these concepts to family interaction we
can conceive of the parents in a "normal" family as a
pair in a solidary relationship which persists over time.
Pairs of mother-child and father-child also exist, but
the child becomes an isolate in those instances where he
deviates from the parental demands. In the schizophrenic
family, however, the isolate is the father rather than
the child. Moreover, the child is placed in an extremely
dependent position which increases his anxiety and thus
strengthens the existing mother-child relationship. The
mother increases her power in this 'conflicting rela-

tionship' due to her use of the double bind. The child's

9

only possible resolution of the frustration inherent in
his position rests in developing a schizophrenic reaction.

All hypotheses regarding the etiology of schizo-
phrenia remain difficult, if not impossible, to put to
experimental test. Studies of contemporary interaction
of families can only furnish partial verification of the
double bind theory since we cannot be sure of what the
familial interaction was like when the patient was a
child. Individual therapy can be an added measure of
verification, but successful therapy with a schizo-
phrenic is at best a time consuming operation covering
several years; and even here we cannot be sure of the
accuracy with which the patient perceived and/or reported
his parents' behavior. Therapy with the family may be
the best technique, but again inaccurate recall and
misperceptions remain possible confounding variables.

It seems that by conceding the inability to put
these hypotheses to a direct test we nonetheless might
reap some benefits by attacking the problem indirectly.
If the "double bind" is responsible for schizophrenic
behavior, re-exposure to such a situation should bring
about a situational withdrawal of greater intensity in
schizophrenics than in "normals," since the former group

have allegedly structured their lives to avoid such

10

o?

situations. To simulate this paradigm we would thus have
to create a situation in which the subjects are partially
dependent upon others for rewards. They would be given
messages from these reinforcing agents which imply that
they are being rewarded as well as being punished, and
the subjects will not be given any further clarification
about the nature of the reinforcement until they have
made a response. Among the possible alternative actions
they should be permitted some sort of behavior involving
withdrawal from the situation.

If these conditions can be set up, we would be simu—
lating the conflict which Bateson and others consider
the common dynamic underlying schiZOphrenic withdrawal.
If exposure to a double bind situation prompts a greater
withdrawal by schizophrenic than by nonschizophrenic
subjects, then the reenacting of the paradigm furnishes
indirect proof of the hypothesis. Further verification
can be offered by demonstrating that schizophrenics
withdraw to a greater degree in double bind situations
than in other frustrating interpersonal situations. If
nonschizophrenics withdraw from the double bind situation
with the same frequency as the schizophrenics do, then
the "double bind" hypothesis will be cast in serious

question.

11

The medium which seems quite suitable in fulfilling
the necessary criteria is that of a game, specifically
a non-zero-sum game where the gains and losses of the
players do not cancel each other out. This media seems
especially appropriate in that it would minimize face-
to-face verbal interaction which is felt to be
threatening to the schizophrenic. It also would permit
us to design alternative choices which would be suffi-
ciently ambiguous so as to carry the different conno-
tations of reward and punishment. Further, game choices
lend themselves more easily to analysis and offer less
opportunity for disagreement in interpretation than does
verbal behavior.

A two person three choice game was designed in which
the combinations involved loss as well as gain of money
to the participants. The three choices open to each
player were labelled A, B, and C. There were thus nine
possible combinations of choices: AA, AB, AC, BA, BB, BC,
CA, CB, and CC (with the first letter representing player
number one's choices, and the second letter representing
player number two's choices). The game is outlined in
Figure'l.

In this game a player receives the greatest gain

(10¢) by selecting B when the other person chooses A.

12

However, the person choosing A would be unhappy with this
combination for it results in a loss of 2¢ a trial. It
is therefore expected that he would abandon this line of

play with repeated trials.

FIGURE 1. Game choices and returns.
Player number two's choices

A B C

 

 

A 29, 2¢ -2¢, 10¢ 2¢, 0¢
Player number «
one's choices B 10¢, -2¢ -1¢, -1¢ 1¢, 0¢

 

 

 

 

 

C 0¢, 2¢ 0¢, 1¢ 0¢, 0¢

 

(Player number one's returns are listed
first in each cell.)

The greatest gain over time for both players would
be achieved by alternating A and B play; that is, if on
trial 1 player one plays A and player two plays B, on the
next trial player one would play B and player two, A.
With this strategy both players will average a gain of 4¢
a trial. Such play demands that one relinquish the
attempt to perpetually receive 10¢ a trial and also
requires mutual trust; for it results in an immediate
loss of money, hopefully to be followed by future gain.
However, one can never be sure that the other person will

reciprocate in this strategy and allow the original A

13

player to receive 10¢ on the succeeding trial. Thus, by
playing A when the other individual plays B, one never
knows if he will be rewarded or punished, as is the case
with the double bind where the schizophrenic does not
know if his behavior is being rewarded or not.

If both subjects strive for the greatest immediate
gain and play B continually, they would lose 1¢ on each
trial and find this line of play unsatisfactory. Only
through cooperation can they make money consistently;
for with both the A and B selections the subjects can
either gain or lose money, depending on the other
person's choices.

The last alternative, C, amounts to a decision not
to play the game. The C choice brings no financial gain
or loss and amounts to a situational withdrawal. When
faced with a C strategy the selection of A brings a 2¢
gain and B brings a return of 1¢ a trial.

It is recognized that any analogue may be questioned
as an adequate replication of a hypothesized situation.
With regard to this study it is felt that the game
measures the frustration resulting from exposure to a
double bind. In order to insure that double-bind
reactions are being measured as contrasted with a gener-

alized reaction to frustrating situations, a control task

14

was introduced in the procedure. This was similar to the
original task except that the situational outcome was not
in the hands of another person.

The control task was a light guessing situation in
which subjects were to predict which of two lights would
go on. They were given 5¢ for each correct prediction
and lost 5¢ for each incorrect guess. They were also
afforded a third choice -- making no prediction and
consequently neither gaining nor losing money. The
sequence with which the lights came on was in a random
50% - 50% schedule insuring that the subjects would have
frequent losses as well as gains in the experiment. The
subjects made their predictions for a total of 60
trials.

Hypotheses

 

The following hypotheses were therefore devised:
1. There will be a greater number of withdrawal responses
made by schizophrenics in the double bind game situation
than in the control task.
2. The schizophrenics will be unwilling to settle on a
'hutually co-operative strategy and will withdraw rather
than let the other person control the rewards and find
themselves in a double bind.

a. This will happen either at the outset of the game

15

and persist or
b. Withdrawal responses will increase in frequency

as the game proceeds.
3. The normal subjects not having experienced excessive
frustration from double bind situations in the past will
demonstrate a lesser tendency to withdraw than will the
schizophrenic subjects.
4. Schizophrenics will show a greater tendency to with-
draw when they start out with nonsimultaneous game trials
(in which one subject's choices are known before the
second subject makes his selections) than when they begin
with simultaneous trials. This should hold true for

normals as well.

Methodology

 

Apparatus:

Subjects for the experiment were seated side by side
at the narrow end of a 36" X.70" table. A black wooden
partition measuring 24" X 25%" was placed on the table
between them, extending in such a manner that the
subjects could not see one another. Directly in front of
each subject was a 10" X 5" switch box; each box
contained three switches marked A, B, and C. At the far
end of the table was a 37" X 14%" board containing six
7 watt light bulbs arranged in pairs. Above the pairs of
lights were the one inch metal letters, A, B, and C,
reading from left to right, respectively. Above the left
light in each pair was the number, one; and above the
right light was the number, two. In back of the board
on the left side was another switch which the experi-
menter used to prevent the lights from going on until
both subjects made their choices.

In the control condition where subjects were tested
individually, the partition was removed and the fifth
light (C #1) was taken off the board. The first switch
box was placed in back of the light board and was manu-

ally controlled by the experimenter.

16

17
Subjects:

The schizophrenic sample was recruited from the
Battle Creek Veterans' Administration Hospital. They had
all been given the diagnosis of undifferentiated schizo-
phrenia by the hospital diagnostic staff, and were either
open ward patients or patients with some grounds privi-
leges. (An undifferentiated group was selected because
they comprised the bulk of the schizophrenic population,
permitting the selection of an adequate sample.) The
mean age of the schizophrenic group was 31 years, and
ages ranged from 22 to 42 years. Mean length of hospi-
talization was 23 months and ranged from two months to
five years. Their average amount of schooling was 10.6
years, and the range was from seven to 16 years. The
subjects were also given the Thorndike-Lorge Multiple
Choice Vocabulary Test as a measure of intelligence.
Their mean vocabulary score was 9.2 with a range of one
to 19 (highest possible score is 20).

The normal group consisted of employees at the
hospital (male aides and housekeepers). Their mean age
was 30, and ages ranged from 19 to 60. The average
amount of schooling was 11.4 years with a range of five
to 15 years. Their mean score on the Thorndike-Lorge

Test was 10, and scores ranged from four to 16. Normals

 

 

18
and schizophrenics did not differ significantly in age,
schooling or vocabulary score (rank sums and t test).

A total of 42 subjects were used from each group.
Twenty-eight of the subjects in each group were run in
pairs, and 14 were tested individually (control subjects).
The assignment to conditions was on a random basis.

(Half of the pairs in each group were tested with 30
simultaneous trials first, followed by 30 nonsimultaneous
trials; and the other group were first given nonsimulta-
neous trials followed by simultaneous trials.) The pairs
of aides all worked on the same ward; and the pairs of
patients were also recruited from the same wards, making
the subjects somewhat equivalent in terms of prior
knowledge about the other player.

Procedure:

The subjects tested under the experimental
conditions were seated by the table and given the
following instructions:

we are interested in finding out how people make
decisions and have asked you both here to help us in this
study. To make it more interesting for you, we have
introduced money into the situation; and as you will soon
see, different choices will bring you gains and losses of
money. You will receive your earnings after the exPer-
iment is over.

In front of you are three switches marked A, B, and

C. Each switch represents a choice so that both of you
can choose either A, B, or C. There are nine combi-

19

nations that can occur. When you choose A, the other
person can choose A, B, or C; when you choose B, he may
choose either A, B, or C; and when you choose C, he may
choose A, B, or C. Do you understand the different kinds
of choices that can occur?

You make your choices by pushing down the appro-
priate switch. If you want to pick A, you will push the
switch marked A; if you want to choose B, you push that
switch down; and to choose C, you push down the switch
marked C.

In front of you is a board containing the letters
A, B, and C. Underneath each letter are two lights, and
as you notice both of the lights have a number. The
number 1 stands for Mr. '3 (player 1) choices; and the
number 2, for Mr. '3 (_Iayer 2) choices. Thus if Mr.
(1) chooses A, after he presses the switch the light
marked #1 under A will go on. If he chooses B, the light
marked #1 under B will go on; and if he chooses C, the
light marked #1 under C will go on. Likewise, if Mr. 52;
chooses A, the light marked #2 under A will go on; if he
chooses B, the light marked #2 under B will go on; and if
he chooses C, the light marked #2 under C will go on. Do
you understand how the lights work? Just for practice I
want each of you to push a switch now. Mr. (1), what did
Mr. (2) choose? And Mr. (2), what did Mr. (1) choose?
That's right. Now would you each push the switch back
and make another choice. Mr. (2), what did Mr. (1)
choose, and Mr. (1), what did Mr. (2) choose? (If either
subject had difficulty in identifying the choices made,
more practice trials were given until they had learned
how to identify the choices.)

 

Would you now turn to the sheet you have been given.
The first column lists your possible choices; the second
column lists the other person's choices; the third column
is your returns; and the last column is the other
person's returns.

When you choose A, the first three rows, you can
either receive 2¢ (if the other person chooses A or C) or
lose 2¢ (if he chooses B). When you choose B, rows 4-6,
you can receive 10¢ if he chooses A (row 4), one cent if
he chooses C (row 6), and lose one cent if he chooses B
(row 5). When you choose C (rows 7-9) you never gain or
lose money.

20
FIGURE 2. Sample sheet.

YOUR CHOICES HIS CHOICES YOUR RETURNS HIS RETURNS

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A A 2¢ 2¢
A B -2¢ 10¢
A c 2¢ o¢
B A 10¢ -2¢
B B - 1¢ - 1¢
B c 1r: o¢
c A o¢ 2::

L. c B o¢ 1¢

F c 0 oc o¢

 

Let's look at each combination individually. When
you both choose A, you each receive 2¢; when you choose
A and the other person chooses B, you lose 2¢ and he
receives 10¢. When you select A and he chooses C, you
receive 2¢ and he gets nothing. When you choose B and he
selects A, you receive 10¢ and he loses 2¢. If you both
choose B, you each lose 1¢; and if you choose B while he
chooses C, you get 1¢ and he receives nothing. If you
choose C you will never gain or lose money, but the other
person gets 2¢ by choosing A when you choose C and 1¢ by
choosing B when you choose C. 'Do you understand the
amount of money you gain or lose with each choice?

You will notice then that the most amount of money
you can make on any one trial is 10¢, by choosing B while
the other person picks A. However, since you will be
making your choices for 60 trials, it is unlikely that
you will keep getting 10¢ a trial because each time you
do, the other person loses 2C, and he will not want to
continue losing money. If he did though, you could
receive a total of $6.00. If you both try to receive
10¢ a trial, by choosing B, you will each lose 1¢. You

21

would be best off by alternating A and B so that one of
you gets 10¢ on one trial and the other gets 10¢ on the
next trial. This way you will average 4¢ a trial
(winning 10¢ on one trial and losing 2¢ on the next
trial), receiving a total of $2.40 apiece. However, you
can never be sure if the other person will let you get
10¢ after you've let him do so. The only other combi-
nation which gives both of you money is AA, but you will
only get 2¢ a trial this way. Also you cannot be sure
that the other person will play A when you do. You will
notice that whenever you play B, the other person cannot'
make any money, and as was mentioned before, your choice
of C never gives you any money.

Do you understand how the choices work now? For
practice I want each of you to push a switch now. Mr.
ill, will you push A, and Mr. 5&2, B. Mr. $11, will you
tell me what you would receive for this choice and what
Mr. £32 would receive. Now would you both push B. Mr.
£223—what would you have gotten for this combination and
what would Mr. £12 receive? Now would you both push A.
Mr. fill, what would you receive and what would Mr. £22
receive for this combination? Mr. ill, will you push B
and Mr. (2), A. Mr. £22, what would you receive and what
would MrTfZLl receive for this combination? (If there
was any difficulty in the subjects understanding the
outcome of the different combinations, they were given
additional practice until they had no trouble with the
choices.)

Remember that after you have made your choice and it
has shown on the board, you must push the switch back
before we can have another trial. While we are doing
this study, we do not want you to talk to each other at
any time. If you try to talk or communicate in any way,
we will have to stop and neither of you will receive any
money. ‘Do you have any questions?

 

NONSIMULTANEOUS TRIALS lst Condition

One last thing, for the first 15 trials Mr. $11 will
see Mr. (2)'s choices before he makes his, and for trials
16-30 MrTfZg; will see Mr. Lil's choices before he makes
his own selections. 0n the last 30 trials neither of you
will know the other person's choices until you have made
your own. Any questions about that?

22

SIMMLTANEOUS TRIALS lst Condition

One last thing, for the first 30 trials you will not
know the other person's choices until you have made your
own. On trials 31-45, Mr. £1) will see the choices made
by Mr. £2) before he makes his own selections; and on
trials 46-60 Mr. $22 will see Mr. gll's choices before he
makes his own choices. Any questions about that?

 

Remember to push the switch off after a trial is
over and never to push more than one switch on any trial.
Do not forget that you are not to talk to each other at
any time during the experiment. we will begin now.

After each trial the eXperimenter turned the master
switch and waited until both subjects had made another
choice before he turned the current on, illuminating
their choices. The subjects were told when they were to
change conditions (either from simultaneous to nonsimul-
taneous, or from nonsimultaneous to simultaneous). At
the conclusion of 60 trials the subjects were asked what
strategy they used and why they chose it. They were then
given the Thorndike-Lorge Test and paid the total amount
they had earned.

Subjects in the control condition were seated and
given the following instructions:

we are interested in finding out how people make
decisions and have asked you here to help us in this
study. Before you are a group of five lights, lettered
A, B, and C and numbered 1 and 2. One of the two lights
marked #1 will go on and we want you to predict which one
it is. That is, light A numbered one or light B numbered
one will go on. You make your predictions by using the

levers in the box by your hand. If you think A will go
on you push the lever forward marked A, and if you think

23

B will go on you push the lever marked B. After you have
made your choice the lights will go on and you will see
which #1 light went on, and the light you predicted
(light #2) will also go on. If you predicted correctly
then either both.A lights or both B lights will be on;
but if your guess was wrong then one A light and one B
light will go on. Each time you predict correctly you
will receive 5¢ and each time you guess wrong you will
lose 5¢. You will be making predictions for a total of
60 trials. After you have seen whether you are right or
not on each trial, you should push the lever back and
make another prediction. You are to make only one guess
on a trial. If you do not want to make a guess on any
trial push lever C and you will neither gain nor lose
money. When the study is completed you will be given the
total amount you have earned. ‘Do you have any questions?
Just for practice, would you make a prediction. 'What did
you predict? Were you right? WOuld you gain or lose 5¢?
(The eXperimenter then lit the other combinations, asking
the subject if he would be right or not in each case.
This was repeated until it became evident that the
subject understood the outcome of each combination. Then
he was told:) Now we can begin.

In the control condition the #1 switch box was
placed in back of the board. The eXperimenter depressed
the A.and B levers according to a predetermined random
50% - 50% strategy. When the subjects had made their
predictions on each trial, the experimenter turned the
master switch on, illuminating the choices. At the
conclusion of 60 trials the subjects then took the
Thorndike-Lorge Test and were paid the total amount they

had earned.

Results

The frequency with which schizophrenics chose the
withdrawal response in the experimental and control tasks
is depicted in Table 1. A chi-square performed on this
datum yields a value of 7.79 which is significant at the
.01 level. Thus the schiz0phrenics withdrew signifi-
cantly more often in the experimental task than in the
control situation.

TABLE 1. Frequency of withdrawal re5ponses (C) by

schizophrenic subjects over 60 trials

under control and exPerimental conditions.

Withdrawal responses

 

 

0 or 1 2 or more
Control ‘
task 7 7
Experimental
task 2 26

 

 

 

 

chi-square = 7.79, significant at .01

The frequency of withdrawal responses by the normal
subjects under experimental and control conditions is
listed in Table 2. The resultant chi-square value of
11.46 is significant at the .001 level, leading us to the
conclusion that the normals withdrew significantly more
often in the experimental task than in the control situ-

ation.

24

TABLE 2.

25

Frequency of withdrawal responses (C) by
normal subjects over 60 trials under
experimental and control conditions.

”Withdrawal responses

 

 

0 or 1 2 or more
Control
task 13 1
Experimental
task ~ 9 19

 

 

 

 

chi-square = 11.46, significant at .001

Schizophrenics and normals were compared in terms of

the frequency with which they chose the withdrawal

response in the control condition. Table 3 summarizes

this datum and the resultant chi-square of 4.37 is signif-

icant at the .05 level. The schizophrenics, therefore,

withdrew more frequently than normals in the control task.

TABLE 3.

Frequency of withdrawal responses to the
control task.

Withdrawal responses

 

 

0 or 1 2 or more
Normals
13 1
Schizo-
phrenics ' 7 7

 

 

 

 

chi-square = 4.37, significant at .05

In order to compare the behavior of schizophrenics

and normals in the eXperimental task as well as to gauge

the effects of initial testing condition (simultaneous

26

vs. nonsimultaneous trials) upon the frequency of with-
drawal responses, an analysis of variance was performed.
In order to satisfy the assumption of homogeneity of
variance inherent in this analysis, the F max test was
calculated. The obtained value of 1.96 permits the
retention of the hypothesis that the variances are homo-
geneous.

Table 4 summarizes the analysis of variance. The
schizophrenics chose the withdrawal response signifi-
cantly more often than the normals did (means of 12.9 and
7.4, respectively), which is significant at the .01 level.
There were no significant differences between the effects
of initial experimental condition and no significant
subject-condition interaction.

Since the schizophrenics withdrew to»a greater
degree than the normals in both experimental and control
conditions, it was necessary to discover if there was a
differential withdrawal to eXperimental conditions. That
is, if the double bind poses a greater threat to schizo-
phrenics than to normals, the schizophrenics should
withdraw to a greater degree under experimental
conditions as compared to control conditions, than the
normals do. A.t-test was therefore computed comparing

the difference between schizophrenics under control and

27

TABLE 4. Summary table for analysis of variance.

Source degrees of sum of mean F
freedom squares square

 

Order (simultaneous
vs. nonsimultaneous) 1 16.08 16.08 1

 

Subjects (schizo-

 

 

 

 

 

 

 

phrenics vs. normals) 1 424.22 424.22 9.8?*
Subjects X order 1 9.32 9.32 1
Within cells 52 2234.60 42.97 |
Total 55 2684.22

 

 

 

 

 

 

 

* significant at .01 level

experimental conditions with the difference between
normals under these two conditions. The obtained value
of .44 does not permit us to reject the hypothesis that
there is no significant difference between these two
differences. In other words, there was no differential
withdrawal by either group to the experimental task.
While a joint alternation strategy will yield the

greatest earnings in the game and necessitate double bind
exposure, it is possible for subjects to make their
choices in such a manner as to avoid the double bind (not

play A) as well as avoid withdrawal (not play C). (see

28

Figure 1.) That is, continual selection of B will avoid
double bind exposure. Since the only way that subjects
can be placed into a double bind is with their choice of
A, a t-test was performed comparing normals and schizo-
phrenics in terms of the frequency with which they
risked double bind exposure (chosehA). The test for
homogeneity of variance yielded an F of 1.51 permitting
retention of the hypothesis that both groups have compa-
rable variances. The mean frequency of A selections was
26.0 for the normals and 20.14 for the schizophrenics.
The resultant t yields a value of 5.32 which is signif-
icant at the .005 level, leading to the conclusion that
normals risked double bind exposure to a greater degree
than the schizophrenics did.

In order to further clarify the degree of threat
which the "double bind" task imposed, the choices on
nonsimultaneous trials were analyzed. When a subject
chooses first and his choice is displayed to the other
player, he has the least control over his returns. When
he is in the position of choosing last, knowing the other
player's choices, he has greatest control over his
returns. The frequency of withdrawal responses made by
the schizophrenics under these two conditions was

compared by means of a matched t-test. The obtained

29
means of 2.57 (choosing first) and 4.28 (choosing last)

over 15 trials, yields a t of 3.05, which is significant
at the .02 level. Thus schizophrenics withdrew to a
greater degree when they chose last and had greatest
control over the situation, than when they chose first
and had least control over their returns.

The frequency of withdrawal responses by normals
during nonsimultaneous trials was analyzed in a similar
manner. Their mean frequency of withdrawal responses was
2.71 when choosing first, and 1.85 when choosing last
(over 15 trials). The obtained t of 2.263 is significant
at the .05 level. Thus, normals withdrew significantly
more often when choosing first than when choosing last,
on nonsimultaneous trials.

A matched t-test was performed in order to compare
the frequency with which schizophrenics chose withdrawal
responses over time in the experimental task. Behavior
on the first 30 trials was compared with behavior on the
last 30 trials (obtained means of 7.25 and 5.68, respec-
tively). The resultant t of 2.70 is a difference signif-
icant at the .05 level. The schiZOphrenics, therefore,

withdrew less frequently with repeated trials.

Discussion

 

The finding that both normals and schizophrenics
withdrew more frequently in the double bind situation
than in the control condition confirms the fact that the
double bind is a threatening situation. While the degree
of threat varies among individuals, it does not result in
a generalized withdrawal. The two subjects with the
highest frequency of withdrawal responses were in the
normal group, and they withdrew on only 40% of their
choices (23 of a possible 60). Furthermore, none of the
schizophrenics selected the withdrawal choice as their
most preferred response.

Although the total frequency of withdrawal responses
was rather minimal, it was demonstrated that withdrawal
is a more favored re5ponse in the schizophrenic's
repertoire of behavior than in the normal's. That is,
schizophrenics withdrew significantly more often in both
the experimental and control tasks than the normals did.
However, the normals'and schizophrenics showed comparable
increases in frequency of withdrawal reSponses to the
double bind situation, over their initial frequency of
withdrawal in the control task (mean gains of 7.57 for
the schiZOphrenics and 6.28 for the normals, which is a
nonsignificant difference). If the double bind carries

30

31

a special threat to schizophrenics, it would seem likely
that they would show a disproportionate increase in
withdrawal responses to the eXperimental task in contrast
to the increase of the normal group. Their failure to
behave in this manner casts some serious doubt upon the
validity of the double bind hypothesis as it has been
outlined.

There are two other factors which place the double
bind theory in serious question. The first is the
finding that schizophrenics withdrew less as the game
progressed. Since they failed to achieve the highest
paying 'mutual co-operative' strategy, it would seem that
in accordance with the theory schizophrenics would try to
escape from the double bind, just as their schizophrenia
represents an attempt at such a solution. Instead they
chose to remain in the situation more frequently than
they had initially.

The other finding in contradiction to the hypotheses
is the fact that schizophrenics withdrew more frequently
when they had greatest control over their returns than
when they had least control over their earnings. In
other words, they withdrew when the possibility of
entering a double bind was minimal rather than maximal.

Thus, it does not appear that their primary motivation

32

was to avoid being placed in the dependent position in a
double bind. If being controlled was the critical issue
then too, order (simultaneous vs. nonsimultaneous first
testing conditions) should have had an effect upon
withdrawal, which it did not.

There seem to be several alternative eXplanations
for the lack of support of the double bind theory.
Perhaps the game procedure is an inadequate simulation
of the situation: It is true that it involved peers
rather than a parent-child relationship, and the moti-
vation here was overtly economic rather than emotional
(though both emotional rewards in the game and economic
motivation in double bind situations cannot be ruled
out). Another factor is that the patients used were all
from open wards or with some grounds privileges, and were
less confused and perhaps in a greater state of recovery
from their schizophrenic symptomatology than would be
closed ward patients. However, the dependency and uncer-
tainty of reinforcement which appear crucial to the
double bind are reenacted in the game. This is confirmed
by the fact that both groups had a greater frequency of
withdrawal responses to the eXperimental situation than
to the control task. While the eXperiment may not

simulate the same degree of personal investment as the

33

original double bind, some generalization would be
expected if the theory is correct. Thus the lack of a
differential increase in withdrawal responses by the
schizophrenic subjects points to the erroneous nature of
the theory.

It is possible to reason that the hospital employees
are an inappropriate normal sample and perhaps not the
best adjusted group of subjects. However, their behavior
in the game was similar to that of college students who
also failed to achieve consistent co-operative strategies
when playing similar games. (Wilson, et. al.)

When pilot work was performed to test this procedure
and instructions, seven pairs of psychology graduate
students acted as subjects in the game. All six pairs
who had no trouble understanding the payoffs settled on
a co-operative strategy which persisted from the first
trial to the last. This stands in marked contrast to
the overt performance of the aides. However, the
graduate students were playing for points rather than
money.

In evaluating the behavior of normal subjects, how-
ever, the issue does not appear to be whether "normal"
subjects actually succeed in achieving a co-operative

strategy. Both the graduate students and the aides

34
attempted to achieve this goal, and both groups used the

co-operative choice of A more frequently than the schizo-
phrenics did. Both normal groups also used withdrawal
less frequently than did the schizophrenics. Thus the
same general results would be expected with either aides
or psychology graduate students serving as the normal
group. The difference seems to be only one of degree.
Schizophrenics and aides had a mean frequency of 12.9
and 7.4 withdrawal responses in the experimental task,
respectively, which are significantly different. If the
psychology graduate students' lack of withdrawal
responses was statistically compared with the schizo-
phrenics', this too would be a significant difference.
Since other research using similar games with under-
graduate students demonstrates that a coaoperative
strategy is rarely attained, it would seem that the
graduate students are a unique group rather than repre-
sentative of a broader spectrum of the normal population.
Since the double bind was not a distinctive threat
to the schizophrenics, their behavior was evidently a
result of other factors. Information about the nature of
these factors was offered by the subjects' verbal
statements of the strategies they employed in the eXperi-

mental procedure. Only three of 20 schizophrenics stated

35

that they tried to achieve a co-operative strategy, 14
said they were trying to win or make the most money for
themselves (individualistic strategy), and three
mentioned irrelevant strategies (tried to get all combi-
nations, too lazy to think). In the normal group 14 of
21 stated that they were trying to achieve a co-operative
strategy, four said they were trying to gain the most for
themselves, and three mentioned irrelevant strategies
(two were trying to get the different combinations, and

a third said he was trying to give the other player money
and did not care about his own returns). -If these verbal
strategies concerning game behavior were implemented,
then the schizophrenics and normals had different motives
in the game.

If it is then assumed that the schizophrenics were
using an individualistic strategy, their behavior on
nonsimultaneous trials can be logically explained. When
the game is played with nonsimultaneous trials, if the
first player chooses B, the second player can either
withdraw or lose money. He cannot earn any money in that
situation.. Since the second players in both the normal
and schiz0phrenic groups had to select against comparable

totals of B choices (179 and 196, respectively), the

normals by not withdrawing more frequently chose a

36

response involving loss of money. This occurred because
the normals were trying to influence as well as punish
the other player in order to ultimately achieve a co-oper-
ative strategy. A withdrawal response would have
produced little change in the other player's strategy.
However, if one did not consider the other player's moti-
vation, when faced with a B choice which allows either
withdrawal or loss of money, one would be best off eco-
nomically by withdrawing, which is exactly what the
schizophrenics did. 'When choosing last on nonsimul-
taneous trials, schizophrenics chose C 70 times against
a B strategy, 34 times against an A strategy, and 12
times against a C strategy.

Along with the possibility that schizophrenics were
choosing in order to maximize their short term gain, it
is also possible that they withdrew rather than assume
the responsibility for the amount of mutual returns.
Perhaps they were trying to avoid self-assertion in an
interpersonal situation. Both of these factors may have
been operating, but in view of the fact that the
frequency of withdrawal responses on nonsimultaneous
trials was less than five (out of 15 trials), neither of
these possibilities could have been the preponderant

motivation. It remains for further research to

37

demonstrate the factors involved in the low but signif-
icant use of the withdrawal response by schizophrenics.
The question to be answered is why the schizo-

phrenics chose an individualistic rather than a co-oper-
ative strategy in the game procedure. It is possible
that the double bind posed a great deal of threat, and an r-
individualistic strategy was one means of avoidance. Yet
if this was the case, they should not have used the A ;

alternative at all, since playing A exposed one to the

 

Jainc- ' .4; v.

double bind. Instead, they chose A on an average of 20
times (33% of their selections), a frequency we would not
expect if the double bind was such a threat. It is
possible to go a step further and argue that choice of A
represents an attempt to master the double bind situation,
but this seems to be an effort to stretch the data to fit
the theory. Furthermore, no reference is made in double
bind theory concerning repeated attempts to master the
situation.

A more plausible explanation is that past experience
has taught the schizophrenic that he cannot trust other
people. When faced with the game situation, he chooses a
strategy that will exclude the necessity of mutual trust.
The schizophrenic's past experience negates the exam-

iner's explanation that co-operative strategy would be

38
the most profitable.

Thus the explanation for the game behavior of the
schizophrenics would appear to rest primarily on their
lack of trust of other people. If the underlying cause
for their behavior were specifically the double bind, we
would have seen a pervasive use of withdrawal; for if
withdrawal is a symptom of schizophrenia, it would be
expected to occur most frequently in situations similar

to the ones which caused the schiz0phrenia.

 

With lack of trust as the key to the schizophrenic's
game behavior, the double bind functions solely as
another situation in which the schizophrenic must charac-
teristically keep on guard. He seeks the greatest mone-
tary gain which does not demand an interpersonal
commfltment.

Thus, the experimental procedure offers no support
for the double bind as the primary causative agent in
schizophrenic reactions. Double bind situations can be
seen as one kind of interaction which has taught the
schizophrenic individual that he cannot trust other
peOple. However, these situations represent only a small
proportion of the possible interactions leading to this
lack of trust. Double bind situations may be the primary

cause for some schizophrenic reactions, but they do not

39

seem particularly important for the subjects in this

study.

 

Summary
This study was designed to investigate the

importance of the double bind in the etiology of schizo-
phrenic reactions. ‘Double bind theory considers the
parent-child relationship the determinant of schizo-
phrenic reactions; specifically, it states that a parent
can continually present contradictory messages to the
child so that none of the child's reSponses will be
rewarded. The child's solution to this situation is
thought to be his schizophrenic reaction.

The double bind situation was simulated in a two-
person three-choice game. In this game one could achieve
the greatest monetary returns by losing money every other
trial (thus rewarding the other player) and having the
other reciprocate on the following trial. Since verbal
communication was not allowed in the game, subjects could
not know beforehand whether their solutions would ulti-
mately be rewarded or punished. If a player lost money
on a trial, the other player's strategy was open to
contradictory interpretations (as leading to an ultimate
reward or representing immediate punishment). In a
similar manner the behavior of the schizophrenic's
parents represents frequent contradiction. A game alter-
native of situational withdrawal was permitted which

40

41

parallels the withdrawal inherent in a schizophrenic
reaction.

A control task was introduced into the procedure to
demonstrate that the frustration being measured was
specifically double bind frustration. This control task
was a light guessing situation, where subjects had to
predict which of two light would go on. The monetary
reward for correct guesses was equivalent to the monetary
loss for those incorrect. Since the lights went on
randomly, subjects were rewarded approximately as often
as they were punished. In the control task subjects were
also permitted a withdrawal response which prevented them
from either gaining or losing money.

Fourteen pairs of undifferentiated schizophrenics
and 14 pairs of normal subjects played the game for a
total of 60 trials. Another group of 14 subjects from
each category were given the 60 trials of the control
task. The schizophrenic and normal groups were comparable
in age, education and intelligence.

Both groups of subjects withdrew more frequently in
the experimental task than in the control task, attesting
to the greater threat posed by the double bind situation.
While schizophrenics utilized withdrawal more frequently

than did normals in both eXperimental and control situ-

42

ations, they did not show a differential increase in
withdrawal responses to the eXperimental task. However,
in utilizing a nonco-operative strategy the schizo-
phrenics avoided double bind eXposure more frequently
than did the normal subjects; although the schizophrenics'
exposure to the double bind was still quite high. ‘
Normals tended to withdraw when they had least control
over their earnings. In contrast, schizophrenics with-
drew when they had greatest control over their earnings
and hence were less likely to be placed in the double
bind. The schizophrenics also withdrew less over time.
Since the basic hypotheses of the study were not
confirmed, it was concluded that the double bind poses
no distinct threat to the schizophrenics and is, there-
fore, not the primary factor in the etiology of schizo-
phrenic reactions.

A lack of interpersonal trust seems to best explain
the individualistic, nonco-Operative game strategy used
by the schizophrenics. None of the schizophrenic subjects
settled on a strategy which involved a permanent inter-
personal committment. It was apparently the desire to
avoid committments rather than a desire to avoid the
double bind, which motivated the schizophrenics' game

behavior. The double bind is then considered only one of

43

many possible interactions which could result in the

development of a lack of trust.

10.

11.

BIBLIOGRAPHY

Bateson, G., Cultural problems posed by a study of
schizophrenic process in Schizophrenia, an Inte-
grated Approach. Auerbach, A., ed., 125-141,
Ronald Press, New York, 1959.

 

 

Bateson, G., Jackson, D.D., Haley, J., and Weakland,
J.H., Towards a theory of schizophrenia. Beh. Sci.
1956, 1, 251-264.

Bowen, M., A family concept of schizophrenia in The
Etiology of Schizophrenia. Jackson, D.D., ed., 346-
372. Basic Books Inc., New York, 1960.

 

Jackson,'D.D., A critique of the literature on the
genetics of schizophrenia in The Etiology of Schizo-
phrenia. Jackson, D.D., ed., 37-91. Basic Books
Inc., New York, 1960.

 

Jackson, D.D., and weakland, J.H., SchiZOphrenic
symptoms and family interaction. A.M.A. Arch. Genl.
Psychiat. 1959, 1, 618-621.

 

Lidz, T., and Fleck, S., Schizophrenia, human inte-
gration and the role of the family in The Etiology
of Schizophrenia. Jackson, D.D., ed., 323-345,
Basic Books Inc., New York, 1960.

 

 

Mehlman, B., The reliability of psychiatric diag—
nosis. J. Abn. Soc. Psych. 1952, 41, 577-579.

 

Mills, T.M., The coalition pattern in three person
groups. Amer. Soc. Rev. 1954, 19, 657-667.

 

Mills, T.M., Power relations in three person groups
in Group Dynamics Research and Theory, 2nd ed.,
Cartwright, D., and Zander, A., eds., Row Peterson
& Co., Evanston, Ill., 1960.

Passarmanick, B., Dinitz, 5., and Lefton, M.,
Psychiatric orientation and its relation to diagnosis
and treatment in a mental hospital. Amer. J.
Psychiat., 1959, 116, 127-132.

Ruesch, J., and Bateson, G., Communication the
Social Matrix of Psychiatry. W.W. Norton & Co.,

 

44

 

 

11.,

45

cont. Inc., New York, 1951.

12.

13.

14.

15.

16.

Scodel, A., Minas, J.S., Ratoosh, P., & Lipetz, N.,
Some descriptive aspects of two-person non-zero-
sum games. J. Conflict Res. 1953, Q, 114-119.

 

von Neumann, J., and Morganstern, 0., Theory of
Games and Economic Behavior. Princeton Univ.
Press, Princeton, N.J., 1947, 2nd ed.

 

Weakland, J.H., The double-bind hypothesis of
schizophrenia and three party interaction in The
Etiology of Schizophrenia. Jackson, D;D., ed.,

 

373-388. Basic Books Inc., New York, 1960.

Wilson, K.V., and Bixenstine, V.E. Personal
communication.

Winder, C.L., Some psychological studies of
schizophrenia in The Etiology of Schizophrenia.
Jackson, D.D., ed., 191-247. Basic Books Inc.,
New York, 1960.

 

APPENDICES
TABLE 5. Raw data for schiZOphrenic control subjects.

Subject Vocabulary Birth Years of Frequency of

number score year education withdrawal

responses
1. 12 1936 12 0
2. 5 1934 10 0
3. 12 1929 12 13
4. 6 1927 10 5
5. 10 1930 10 2
6. 9 1923 16 3
7. 9 1926 13 0
8. 10 1934 12 0
9. 1 1923 7 13
10. 13 1934 12 0
11. 13 1940 12 0
12. 9 1933 10 19
13. 6 1927 12 20
14. 5 1930 9 0

TABLE 6. Raw data for normal control subjects.

Subject Vocabulary Birth Years of Frequency of

number score year education withdrawal

responses
15. 4 1906 8 0
16. 12 1941 11 O
17. 16 1921 13 0
18. 8 1934 12 15
19. 11 1942 12 0
20. 8 1916 10 0
21. 10 1938 12 0
22. 13 1942 12 0
23. 10 1936 15 1
24. 8 1984 9 0
25. 11 1934 13 0
26. 10 1915 10 0
27. 12 1941 12 0
28. 8 1937 12 0

46

47

TABLE 7. Background data on eXperimental schizo-
phrenic subjects.

Subject Vocabulary Birth Years of

number score year education
29. 7 1930 9
30. 8 1922 7
31. 8 1930 11
32. 12 1936 9
33. 9 1939 10
34. 15 1929 9
35. 7 1931 10
36. 7 1931 12
37. 11 1936 8
38. 15 1924 11
39. 8 1939 10
40. 11 1923 12
41. 9 1934 12
42. 7 1933 11
43. 8 1933 12
44. 19 1931 15
45. 10 1925 15
46. 7 1932 12
47. 10 1933 13
48. 15 1933 10
49. 9 1937 12
50. 9 1941 12
51. 5 1925 12
52. 12 1924 13
53. 7 1923 7
54. 8 1921 8
55. 6 1921 8
56. 9 1927 8

48

TABLE 8. Background data on experimental normal
subjects.

Subject Vocabulary Birth Years of

number score year education
57. 10 1938 12
58. 10 1942 12
59. 7 1944 11
60. 7 1944 10
61. 16 1907 11
62. 13 1932 12
63. 11 1940 12
64. 9 1937 12
65. 9 1941 12
66. 11 1941 14
67. 9 1917 12
68. 9 1935 12
69. 10 1937 10
70. 11 1937 12
71. 8 1903 5
72. 10 1941 12
73. 9 1941 12
74. 8 1944 12
75. 10 1928 10
76. 11 1937 12
77. 10 1926 7
78. 9 1929 12
79. 11 1934 14
80. 8 1942 13
81. 6 1932 9
82. 12 1927 12
83. 9 1940 12

84. 2 1927 8

49

TABLE 9. Game behavior of schizophrenic subjects under
simultaneous trials first condition.

 

Nonsimultaneous
trials
Subject Total C Total C on Total C Total C Total A
number choices simultan- when when choices
eous trials first last
29. 15 7 0 8 17
30. 10 9 0 1 26
31. 7 6 1 0 22
32. 16 12 2 2 14
33. 19 10 4 5 19
34. 13 6 3 4 25
35. 18 8 4 6 26
36. 0 0 0 0 29
37. 14 3 6 5 10
38. 11 9 2 0 22
39. 17 8 4 5 10
40. 15 0 0 5 22
41. 16 9 4 3 20
42. 18 8 5 5 19

50

TABLE 10. Game behavior of schizophrenic subjects
under nonsimultaneous trials first

 

condition.
Nonsimultaneous
trials
Subject Total C Total C on Total C Total C Total A
number choices simultan- when when choices
eous trials first last
43. 9 1 0 8 2
44. 11 0 0 11 17
45. 16 8 4 5 26
46. 17 6 3 8 30
47. 20 11 1 8 18
48. 5 0 0 5 28
49. 21 9 4 8 12
50. 12 5 1 6 21
51. 20 10 5 5 22
52. 0 0 O 0 26
53. 20 10 5 5 22
54. 15 7 6 2 19
55. 7 5 2 0 23
56. 10 3 4 3 27

51

TABLE 11. Game behavior of normal subjects under
simultaneous trials first condition.

 

Nonsimultaneous
trials
Subject Total C Total C on Total C Total C Total A
number choices simultan- when when choices
eous trials first last
57. 8 0 3 5 35
58. 5 1 1 3 19
59. 0 0 0 0 32
60. “0 0 0 0 31
61. 1 0 1 0 24
62. 4 0 4 0 28
63. 18 4 2 2 19
64. 23 10 5 8 16
65. 0 0 0 0 32
66. 0 0 0 0 32
67. 22 6 12 4 11
68. 2 0 0 2 35
69. 6 0 l 5 23
70. 11 6 4 1 29

52

TABLE 12. Game behavior of normal subjects under
nonsimultaneous trials first condition.

 

Nonsimultaneous
trials
Subject Total C Total C on Total C Total C Total A
number choices simultan- when when choices
eous trials first last
71. 14 6 4 4 24
72. 8 2 5 1 24
73. 0 0 0 0 30
74. 0 0 0 0 30
75. 7 6 0 1 26
76. 1 1 0 0 37
77. 18 11 5 2 19
78. 15 9 3 3 21
79. 0 0 0 0 25
80. 11 6 4 1 23
81. 11 5 3 3 22
82. 15 6 5 4 22
83. 8 2 5 1 29
84. 9 3 4 2 26

"71211111111711.11111111““