Download as pdf or txt
Download as pdf or txt
You are on page 1of 24

Three Preliminary Studies of a

Psychoanalytic Theory of
Alcohol Addiction 1
Robert J. Gibbins 2 and Richard H. Walters 3

J\N EXPLANATION of alcohol addiction which claims a great


f l many adherents, having its origin in the writings of Freud,
postulates an affinity between addictive drinking and re-
pressed homosexuality. Although there are rather wide differences
in the etiological views of psychoanalysts, a large number-includ-
ing Abraham ( 1), Fenichel ( 2), Lolli ( 3), Schilder ( 4) and Weijl
( 5 )-award the homosexual factor a prominent position in the etio-
logical picture, while still others-Juliusberger ( 6), Read ( 7) and
Tabori ( 8 )-assign to it the decisive role. The early ontogenetic de-
velopment of each individual is believed to contain a homosexual
component which is, under favorable circumstances, sublimated. In
persons who become addicted to alcohol this sublimation has failed
to occur and latent homosexual trends remain in the personality.
If there is any substance to the belief that alcoholic males are
latent homosexuals, one might expect that they and nonalcoholic
manifest homosexuals would perceive certain stimuli in their en-
vironment in a somewhat similar manner, i.e., that in some respects
alcoholics would differ from normal subjects and resemble homo-
sexuals. Consequently, methods used in "perception and personality"
studies provide a possible approach to exploring further the plausi-
bility of the psychoanalytic theory of alcohol addiction. Two meth-
ods of this kind were used in the studies reported herein. The first
1
From the Ontario Alcoholism Research Foundation, Toronto, Canada.
Received for publication: 5 October 1959.
2
Lecturer, Department of Psychology, University of Toronto; Research Associate,
Alcoholism Research Foundation.
3
Assistant Professor, Department of Psychology, University of Toronto.
AcKNOWLEDGMENTS.-The authors wish to express thanks to Dr. D. Paitich, Mr.
W. R. Hooper and Miss Phyllis Young for their assistance in the collection of data.
We are especially indebted to Messrs. J. R. Seeley and R. E. Popham for their
valuable technical assistance.

618
STUDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 619

involved the determination of recognition thresholds for words with


sexual connotation; the second involved determination of a prefer-
ence between two classes of sex symbols in the form of art drawings.
The perceptual defense hypothesis assumes that recognition of
anxiety-arousing stimuli tends to be distorted or repressed in some
way. Assuming that a latent homosexual is in some respects similar
to an overt homosexual (e.g., in his lack of an assured masculine
identification), both homosexuals and alcoholics might be ex-
pected to use defensive mechanisms of a similar kind. For example,
it might be predicted that alcoholics and homosexuals would find
difficulty in recognizing tachistoscopically presented sexual words.
On the other hand, both groups might be sensitized to certain sexual
stimuli and thus perceive them more rapidly than normal subjects.
In an experiment by Daston ( 9), which gives some indirect sup-
port to the latter supposition, paranoid patients, also latent homo-
sexuals according to psychoanalytic theory, were shown to have
lower recognition thresholds for words with homosexual connota-
tion than did nonparanoid controls.
By another line of reasoning, however, alcoholics and overt
homosexuals might be expected to differ from nonnals, but in
opposite directions, especially if the homosexuals are relatively
accepting of their way of life. The latent homosexual is presumably
not overt because his homosexual impulses are repressed; the overt
homosexual, on the other hand, is necessarily conscious of his devi-
ant sexual pattern. As soon as the term "latent" is introduced into
the description of a group of subjects, it is difficult to predict their
behavior in an experimental situation. However, since psycho-
analytic theory stresses certain similarities between the "character
structure" of overt and latent homosexuals, and since some of the
"perception and personality" theorists have assumed that perceptual
distortions and idiosyncrasies are an index of the underlying needs,
wishes or motives of their subjects, a demonstration that alcohol
addicts and homosexuals respond similarly under conditions that
facilitate misperception, perceptual distortion, or sensitization to
perceptual stimuli, would tend to support the psychoanalytic theory
of alcohol addiction.
Perceptual defense phenomena may, however, be simply a func-
tion of prior reinforcement in certain social contexts ( 10, 11). On
the basis of this theory, whether or not alcoholics and homosexuals
will respond in a manner similar to one another and different from
620 ROBERT J. GIBBINS AND RICHARD H. WALTERS

normal subjects will depend largely on the conditions under which


testing is carried out. This theory is not the one under test in the
present series of studies; it nevertheless provides an alternative
approach to the interpretation of any results that may be obtained.
According to F enichel ( 2, p. 331 ) , one of the crucial factors in
the development of homosexuality in the male is the substitution
of identification for object relationships:
"Following the loss of an object or disappointment in an object, every-
one tends to regress from the level of object love to that of identification;
he becomes the object which he cannot possess. Thus, the homosexual
individual identifies himself with the object (his mother), after having
been disappointed by its genitals; what determines whether he will be-
come homosexual is how and in what respect this identification takes
place. The homosexual man identifies himself with his frustrating
mother in one particular respect: like her, he loves men."
The result of this crucial identification is that male homosexuals,
like normal women, will be more attracted to males than to females
as sexual partners. Although the specific type of homosexuality that
emerges depends upon the stage of libidinal development at which
fixation occurs, these later variations do not affect the male homo-
sexual's fundamental preference for male sexual partners.
In his writing on the interpretation of dreams, Freud ( 12) sug-
gested that sexual objects may be expressed in disguised, symbolic
form. Certain symbols usually represent the male, and others the
female. Male or phallic symbols are characteristically linear, upright
objects, whereas female symbols typically have a curvilinear, hollow
and enclosed form. 4 This, Freud asserted, is true for the normal as
well as for the neurotic individual.
From the foregoing considerations, it may be inferred that an in-
dividual's sexual identification determines whether his choice of
sexual object will be male or female; that the object choice of
homosexual males is similar to that of heterosexual females; and
that the nature of the cathected object may be revealed symbolically.
In an investigation concerning the relationship between attitudes
toward sex and p ersonality structure, Franck ( 13) devised and
made use of a test based on the Freudian theory of symbols. This
test consisted of nine pairs of pictures each showing one male and
one female symbol. She stated ( 13, p. 81):

• While interpretation is now no longer so ri gid, by and large, for quantitative


purposes, the distinction will suffice.
STUDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 621

"Considering the pictures as environmental stimuli, it was hoped the


subjects would react to male symbols as they tend to react toward men
and to female symbols as they tend to behave toward women. Therefore
they were asked, for each pair, to note which picture attracted them
more .. . . It was assumed that in this way those women with more female
identifications would allow themselves to be attracted by the male
symbols and that the opposite would be true for men."
The results of this investigation seemed sufficiently promising to
justify further work with the symbol test, particularly since it sug-
gested a way of examining further the psychoanalytic theory of
alcohol addiction. For example, a group of manifest homosexual
males should, under the hypothesis, prefer more of the male symbols
than a group of heterosexual males; and, granted certain assump-
tions, male alcoholics-if they are latently homosexual-shoul d be
more like homosexuals than heterosexuals in their choice.
The following two hypotheses were thus under test:
1. In a perceptual defense experiment, the responses of alco-
holics will differ from those of normal subjects ( Ss) in the same
direction, though not to the same extent, as those of homosexuals.
2. (a) When presented with sexual symbols, a group of homo-
sexual males should choose a greater number of male symbols than a
group of normal males. ( b) There should be little or no difference
between an alcoholic and a homosexual group with regard to the
number of male symbols chosen.

EXPERIMENT 1
Method
Subjects.-The Ss were 40 men, all between the ages of 20 and 50, of
whom 15 were alcoholic inpatients at the Brookside Clinic, Toronto,
15 were volunteer control Ss from Workers' Educational Association and
YMCA adult evening classes, and 10 were self-confessed manifest homo-
sexuals. The possibility of including a disproportionate number of homo-
erotic males in the control group was reduced by including only men
who attended YMCA educational and social functions and not those
who were residents of hostels. The relation of the experiment to homo-
sexuality was not indicated to the control or alcoholic subjects. The
homosexuals were volunteers, none of whom was currently under medical
or legal supervision.
Apparatus.-A tachistoscope, set at a constant speed of 20 millisec.,
was used to project words on a screen. A rheostat, graduated from 1 to
100, allowed control of light intensity in terms of the maximum 300 w.
illumination. The words were photographed on 24 mm. slides from
622 ROBERT J· GIBBINS AND RICHARD H. WALTERS

12-pt. IBM type. The illumination on the screen was kept at approx. 1.5
ft-candles. The tachistoscope was located 8 ft. from the screen, with Ss
seated in front and slightly to the left of the projector.
Ten words of ambiguous sexual meaning were selected, after refer-
ence to Henry (14), in such a way that 2 fell into each of 5 categories:
homosexual oral, homosexual anal, homosexual derogatory (slang terms
for homosexuals), male anatomy, and neutral sexual (not specifically
homosexual). Each of the experimental words was matched with a non-
sexual word taken from the Thorndike-Lorge word count ( 15). The
words in each pair were matched for length and general word form as
well as for frequency of occurrence. The 20 words, given in Chart I,
were then arranged in 15 different random orders. Each order was used
once in testing the alcoholics and normals. The first 10 random orders
were used in testing the homosexuals.
Procedure.-A recognition threshold for each word was established
for each S. Starting at an illumination level well below S's recognition
threshold, the experimenter ( E) exposed the first word on the list. The
illumination level was then successively increased by one graduation un-
til S correctly reported the word. E continued in this manner until S
had correctly reported all 20 words.
S was instructed to report each word as soon as he thought he recog-
nized it, even if this involved guessing.

Results
Ten difference scores were obtained for each S by subtracting the
illumination level at which he recognized each control word from
the level at which he recognized the corresponding experimen-
tal word. Difference scores based on words in the same category
(e.g., oral) were then combined, and an analysis of variance was
carried out.
As Table 1 indicates, there were no significant differences be-

CHART !.-Paired Words Used in Experiment 1


Category Experimental Words Control Words
Homosexual oral Blow Slow
Suck Sock
Homosexual anal Brown Train
Rim Rip
Homosexual derogatory Fruit Truth
Queer Quiet
Male anatomy Prick Track
Balls Bills
Neutral sexual Screw Shrew
Lay Law
STIJDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 623

TABLE !.-Analysis of Variance of Difference Scores for


Recognition Thresholds of Alcoholic, Homosexual and
Control Groups (Experiment 1)
Source Sums of Squares df Mean Square F p
Between groups 41.027 2 20.514 1.50 >.05
Between individuals
within groups 504.793 37 13.643
Between categories 334.670 4 83.668 4.52 < .01
Groups X Categories 44.223 8 5.529 <l
Residual 2741.107 148 18.521

tween the groups of Ss. Nevertheless, the mean difference score of


the alcoholic group was intermediate between the scores of homo-
sexuals and controls (Table 2). This was true not only for scores
based on all 10 pairs of words but also for scores on 3 of the 5 word
categories. A significant difference between categories is perhaps
of little importance; this may in part reflect inadequate matching
of experimental and control words in some categories.

EXPERIMENT 2
Failure to find significant differences between groups in Experi-
ment 1 may have been due to weaknesses in the experimental de-
sign. Words with both homosexual and nonhomosexual connotations
were deliberately chosen for use in Experiment 1 because common
derogatory terms for homosexuals are all ambiguous, and it was
felt that all or none of the experimental words should be ambiguous.
The equating of matched words for word form may have compli-
cated the results. Ss sometimes misperceived the word of a pair that
was shown second, thinking that the word shown first had again
been projected. Moreover, evidence that responses to a sexual word

TABLE 2.-Means and SD's of Difference Scores for Recognition Thresholds


of Homosexuals, Alcoholics and Controls for Five Categories
of Words (Experiment 1 )
HOMOSEXUALS ALCOHOLICS CONTROLS
(N = 10) (N = 15) (N = 15)
Category Mean SD Mean SD Mean SD Fo
Oral -0.80 2.29 -0.23 2.70 -1.50 3.22 1.62
Anal -0.90 2.84 -0.47 2.80 0.00 2.14 0.72
Derogatory -0.30 2.72 1.20 3.48 1.83 3.20 3.05
Anatomy 0.45 3.26 -0.70 2.66 0.03 3.42 0.84
Sexual -0.40 3.28 0.37 3.05 0.57 2.72 0.69
Total -0.39 2.72 0.03 3.17 0.19 2.95 1.50
0
No F-ratio is signilicant at the .05 level of confidence.
624 ROBERT J. GIBBINS AND RICHARD H. WALTERS

may influence responses to an adjacent nonsexual word has already


been provided ( 16, 11), consequently control words should per-
haps not have followed experimental words immediately.

Method
The method used in Experiment 2 was identical with that employed
in Experiment 1, except in the following respects:
Subjects.-Although the same number of Ss were employed in each
group as in Experiment 1, no S served in both experiments. Controls
were this time drawn from University evening extension classes and
from YMCA adult classes.
Apparatus.-Thirty- two words were used (Chart II). There were 10
"sexual" words, 2 in each of the 5 categories under study; 10 control
words, each matched with a sexual word for length and word frequency,
but not for word form; 10 buffer words; and 2 practice words. The
sexual and control words were again set in 15 random orders. The buffer
words, all relatively common 2-syllable words, were added to the
series, one after each sexual word. The order in which the buffer words
were introduced was systematically varied from one random order to the
other, since it was not certain that these words would be equally
"neutral" for the three groups of Ss. The practice words were placed
at the beginning of each series. Responses to the practice and buffer
words were ignored in the analysis of data.

Results
The analysis of data was carried out in exactly the same way as for
Experiment 1. This time, significant differences were found between
groups (Table 3). Again, the mean score of alcoholics in most cases
CHART IL-Paired, Buffer and Practice Words Used in Experiment 2
Category Experimental Words Control Words
Homosexual oral breast pillar
suckle oblong
Homosexual anal bowel polar
buttocks helmsman
Homosexual derogatory womanish renegade
effeminate impoverish
Male anatomy penis newsy
testicles ironstone
Neutral sexual sexual recede
rape musk
BUFFER WORDS: doctor, system, nature, moment, river, person, member,
office, island, army
PRACTICE WORDS: paper, story
STUDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 625

T w enty Pairs of Symbolic Pictures (Cards 1-20 ) Used


in Experiment 3. ( Reduced about 55 per cent.) Cards 2,
3, 4, 5, 7, 9, 14, 17 and 20 are those us ed by Franck ( 13 ).
Th e remaining 11 were d esign ed to correspond. Repro-
du ced by permission of the Journal Press, Provincetown,
Mass.

CARD l

a b

CARD 2

a b
626 ROBERT J. GIBBIN S AND RICHARD H . WALTERS

CARD 3

a b

CARD 4

a b

CARD 5

a b
STUDIES OF PSYCHOANALYTIC THEOHY OF ALCOHOLIS:tvl 627

CARD 6

a b

CARD 8

a b
628 ROBERT J. GIBBINS AN' D RICHARD H. WALTERS

CARD 9

a b

CARD 10

a b
STUDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISlVI 629

CARD 11
0 '
a b

i't
\
CARD 12

a b
630 ROBERT J. GIBBINS AND RICHARD H . WALTERS

CARD 13

a b

CARD 14

a b

CARD 15

a b
STUDIES OF PSYCHOANAL YTIC T HEORY OF ALCOHOLI SIII 631

CARD 16

a b

CARD 17

a b

CARD 18

a b
632 ROBERT J. GIBBINS AND RICHARD H . WALTERS

CARD 19

a b

CARD 20

a b
STUDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 633

TABLE 3.-Analysis of Variance of Difference Scores for Recognition


Thresholds of Alcoholic, Homosexual and Control
Groups (Experiment 2)
Source Sums of Squares df Mean Square F p
Between groups 196.388 2 98.194 4.77 <.05
Between individuals
within groups 761.407 37 20.578
Between categories 213.370 4 53.342 2.82 <.05
Groups X categories 90.670 8 11.342 <l
Residual 2797.160 148 18.900

fell between those of homosexuals and controls (Table 4). However,


if the categories are taken singly, significant differences among the
groups occurred only in one category, and this category did not con-
tain words with predominantly homosexual implications.

EXPERIMENT 3
Method
Subjects
(a) Homosexual Group: This group (N=l6) was an unbiased sam-
ple of the manifest homosexual males (excluding pedophiles) referred
to a Toronto forensic clinic for psychiatric examination during an 8-
month period. Apart from homosexuality, these subjects did not exhibit
any gross psychiatric disabilities. They ranged in age from 15 to 47 years.
( b) Alcoholic Group: This group comprised 13 consecutive admis-
sions to a public alcoholism clinic. These Ss were not manifest homo-
sexuals, nor did they exhibit any other obtrusive psychiatric symptoms.
Their age range was from 26 to 51.
( c) Controls: Twenty men enrolled in an extension course in psy-
chology served as nonpsychiatric controls. Nearly all of these subjects

TABLE 4.-Means and SD's of Difference Scores for Recognition


Thresholds of Homosexuals, Alcoholics and Controls for
Five Categories of Words (Experiment 2)
HOMOSEXUALS ALCOHOUCS CONTROLS
(N = 10) (N = 15) (N = 15)
Category Mean SD Mean SD Mean SD F
Oral -0.10 1.87 0.80 1.99 1.37 3.38 2.09
Anal -1.50 2.48 -1.13 4.23 0.o7 2.19 1.81
Derogatory -0.75 1.92 0.13 2.64 -0.27 3.68 0.48
Anatomy -0.55 3.42 -0.13 3.62 0.50 3.97 0.52
Sexual -0.25 1.64 -0.53 2.72 1.33 3.64 3.56°
Total -0.63 2.36 -0.17 2.93 0.60 3.56 4.77°
0
Significant at the .05 level.
634 ROBERT J. GIBBINS AND RICHARD H. WALTERS

were public-school teachers. They ranged in age from 20 to 50 years.


A group of 235 second-, third- and fourth-year students enrolled in
psychology courses at the University of Toronto contributed to the
study in a way that will be fully described below.
Materials
Twenty pairs of pictures, each pair consisting of one male and one
female symbol, were used (Cards 1-20, pp. 625-632). Nine pairs (Cards
2, 3, 4, 5, 7, 9, 14, 17 and 20) were reproductions of those used by
Franck ( 13). The remaining pairs were designed to correspond as
closely as possible with these in style, abstractness, size and aesthetic
quality. A preliminary study involving 32 Ss showed that responses to
the Franck series were positively correlated ( r=.57, p< .Ol) with re-
sponses to the series used in this study.
Procedure
Following the procedure outlined by Franck ( 13), each of the 20
pairs of pictures was presented to the 235 judges who were asked to
identify each picture as either masculine or feminine. They were in-
structed to consider each picture independently of its partner, so that
both pictures of a given pair could be judged as masculine or as
feminine. The distribution of judgments obtained in this way is shown
in Table 5. It can be seen (columns 2 and 4) that the male symbol of
each pair received the highest proportion of judgments of masculine.
Since the pairs were held constant and presented one at a time, the
results should not be affected by the fact that Ss awarded the female
symbols of some pairs more judgments of masculine than the male
symbols in other pairs.
To obtain a measure of the degree to which a male symbol was con-
sidered to be more masculine than its female partner, "separation" values
were calculated for the individual pairs. The per cent judgments of "mas-
culine" were expressed as deviates under the normal curve in sigma units
(Table 5, columns 5 and 6) . By subtracting the value thus obtained for
the female symbol in each pair from that obtained for the male symbol,
the "separation" value for each pair was found (Table 5, column 7).
Separation values for male symbols were designated positive ( +) and
those for female symbols negative ( - ) .
In order to estimate how "masculine" each male symbol was, com-
pared with the other male symbols, or how "feminine" each female
symbol was, compared with the other female symbols, each pair of
pictures was assigned a "scale" value. The midpoint between the two
"separation" values obtained for each pair (Figure 1), expressed in
sigma units, is the scale value given in Table 5, column 8.
Finally each symbol was assigned a value (position value) which ex-
pressed its absolute scale position and separation from its partner. To
obtain this quantity each separation value was multiplied by its corre-
sponding scale value. To reduce the size of the position values and to
TABLE 5.-Per Cent Judgments of Masculine for Each Picture by 235 Judges, and Derivation of Separation,
Scale and Position Values of Pictures
INTENDED INTENDED SEPARATION SCALE
MALE CARDS FEMALE CARDS SIGMA VALUE OF PAIR VALUE 0 VALUEt POSITION V ALUEt
Judgments Judgments Original Original Final
Picture Masculine Picture Feminine Female Male Female Male Male "'
( 1)
lB
2B
(2)
57
83
(3)
lA
2A
(4)
32
25
(5)
-0.88
-1.22
(6)
0.25
1.47
(7)
11
27
-3
1
(B) (9)
33
-27
(10)
-33
27
( 11)
-21
-15
i
0
"1
3B 84 3A 53 0.06 1.49 14 B -112 112 -6 ;ii
4A 64 4B 34 -0.79 0.58 14 -1 14 -14 -19 ><
5A 76 5B 52 0.01 1.15 11 5 -55 55 -12 @
0
6B 51 6A 28 -1.09 -0.04
z>
11 -6 66 -66 -24
7B 75 7A 23 -1.32 1.09 24 -1 24 -24 -20 >
BB BA
~
86 58 0.30 1.58 13 9 -117 117 -6
9A 29 9B 22 -1.37 -1.06 3 -12 36 -36 -21 0
!OB 55 lOA 19 -1.50 0.17 17 -7 119 -119 -29
llB
12A
82
45
llA
12B
63
19
0.52
-1.49
1.39
-0.29
9
12
10
-9
-90
108
90
-108
-9
-28
~
0
13A 91 13B 59 0.37 1.83 15 11 -165 165 -1 ~
14A 52 14B 18 -1.54 0.04 16 -8 128 -12B -30 0
15A 50 15B 44 -0.35 "1
-0.06 3 -2 6 -6 -lB
16A
17A
63
58
16B
17B
29
28
-1.06
-1.09
0.54
0.28
16
14
-3
-4
48
56
-48
-56
-22
-23
§
~.....
lBA 83 18B 62 0.50 1.45 10 10 -100 100 -B
19B 70 19A 47 -0.19 0.84 10 3 -30 30 -15
20A 70 20B 23 -1.31 O.B5 22 -2 44 -44 -22 "'~
0 Separation
values are d erived by adding the sigma values of each pair and rounding; the male pictures are then given a positive sign (+) and the
female pictures a negative sign ( - ).
t For the d erivation of the scale values see Figure 1.
:j: Original position values are derived by multiplying the separation values by the scale values. The final position values are derived by subtracting the ~
constant 175 from the male position values and by adding the same constant to the female position values. The final position values of only the male w
pictures are shown, the values of the female pictures being the same except with a positive (+) sign. °'
1 e • -v.-J
I

a
2
J .lf. I ~
0:.

J S.."' -• s.v. II

5 ~If-~ ~v. S
6 ~
~
s.w. _,
7
S -V. 'I
8 ':-'
<l s.11.-11
~

I
9 0
< ~.V. -7
"- l
s.v. 10 a
"'IX l l 0 I

~
::i
~V. -'I
f-< 12
u
s.v. "

~
;;: 13
~-11. -8
l
15 Q .H.-l.. 0

16 .J.I/. "3 ~

I
} .V.-f
17
1.v. 10
18
~- "- 3
19
i.. v . -l.
2

-1.6 -1.4 -1.2 -1~0 - -:s -.6 -.4 -.2 0 .2 ,4 l~O l. 1.4 1.
s.o. s.o.
FIGU RE 1.-( 1) Location of Each Symbol Pair as Derived by Converting Proportion of Judgments of Masculine into
Sigma Units under the Normal Curve. (2) Derivation of Scale Values for Each Pair of Pictures. O=female sym-
bol; D =male symbol; s.v.=scale value.
STUDffiS OF PSYCHOANAl..YTIC TIIBORY OF Al..COHOLISM 637
make male values positive and female values negative, a constant of 175
was added to all female position values, and subtracted from all male
position values. The results of this operation were divided by 10 and
rounded to serve as final score units. These ranged in size from 1 to 30.
(Table 5, column 11, shows the final values of the male pictures.)
The score assigned to each S was the arithmetic sum of the final
position values of all the symbols he had chosen. In this way, a negative
total score for an individual would indicate that predominantly male
symbols were chosen; a positive total score would reveal that the sub-
ject had selected a greater number of female symbols.
The pictures were then presented to the Ss in the three experimental
groups. Each S was told that he would be shown a series of paired pic-
tures and that he was to indicate the one picture of each pair that
attracted him the most. Each pair was exposed for approximately 10 sec.

Results
The scores indicating final position values of the combined male
and female symbols chosen by each S are shown in Table 6.

TABLE 6.-Scores and Ranks Obtained by Each Subject in Choice


of Male and Female Symbols"'
NORMALS A.LeoHoucs HOMOSEXUALS
(N=20) (N=l3) (N=l6)
Score Rank Score Rank Score Rank
-141 9 -157 7 -233 1
-121 13 -127 11 -2.31 2
-109 15 -95 20 -225 3
-103 17 -83 22.5 -207 4
-101 18.5 -71 24 -167 5
-89 21 -63 25 -163 6
-83 22.5 -27 30 -149 8
-45 27.5 -25 31.5 -135 10
-43 29 +57 37.5 -125 12
-25 31.5 +71 40 -119 14
-5 33 +71 42.5 -107 16
+5 34 +151 42.5 -101 18.5
+23 36 +154 48 -49 26
+49 39 -45 27.5
+59 41 +21 35
+87 44 +41 37.5
+97 45
+129 46
+133 47
+155 49
0 Negative scores indicate predominantly male symbols chosen; positive scores, predominantly

female symbols.
638 ROBERT J. GIBBINS ANl> lUGHARi> H. WALTERS

The Kruskal-Wallis Test ( 17) was used to determine whether or


not significant differences existed among the three groups. This test
yielded an H of 14, which, with 2 degrees of freedom, is significant
at less than the .005 level.
Mann-Whitney U Tests (17) were used for a more detailed
analysis of the results. These tests show that ( 1 ) the homosexuals
chose more male symbols than the non psychiatric controls ( Z=2.2,
p < .05); ( 2) the alcoholics chose more male symbols than the non-
psychiatric controls (Z=5.9, p<.01); (3) the alcoholic and the
homosexual groups were not significantly different with respect to
preference for male symbols ( Z=0.6, p=.55).

DISCUSSION

Although the results of the first two experiments suggest that the
responses of alcoholics to certain perceptual tasks fall between those
of normal Ss and homosexuals, they do not by themselves provide
strong support for the psychoanalytic theory of alcoholism.
The category that most clearly differentiated between the three
groups of Ss in the perceptual defense experiments was that con-
taining words without specifically homosexual connotations. Conse-
quently, one possible explanation of the findings is that alcoholics,
like homosexuals, have difficulties in sexual adjustment, but that
these difficulties are not predominantly homosexual in nature. Per-
haps, as Keller has suggested, 5 alcoholics unlike homosexuals have
not identified with the wrong sex but have failed to identify firmly
with either sex. The results may, however, reflect the past experi-
ences of Ss in quite a different way. The experiment was carried
out in a clinic to which the homosexuals came as volunteer Ss. It
must have been clear to them, even though they were not given
any relevant information, that they were selected on account of
their preferred sexual pattern. It was quite evident that many of
them were willing to talk about their sexual behavior, although
they were not asked to do so. They were thus in a social situation
in which their status was accepted, a situation probably similar in
this respect to ones in which their use of sexual words had been
reinforced. The alcoholic Ss while in the clinic setting are questioned
about their sexual behavior and are even encouraged to give details.

• KELLER, M. [Personal communication, 6 October 1959.]


STUDiEs OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 639

For them, too, there had probably been some reinforcment of sexual
responses. The case would be quite different for the control Ss, who
-at least before the experiment-had no idea that the study had
any reference to sexuality. The findings may thus simply reflect the
influence of prior reinforcement in the particular social setting cre-
ated by the experiment ( 10).
When the results of the third experiment are also taken into con-
sideration, however, the support provided for the psychoanalytic
theory becomes somewhat stronger. A cautious interpretation of
results is nevertheless necessary because of certain limitations of
this study. One serious limitation is the absence of comparison
groups composed of psychiatric patients who are neither homo-
sexual nor alcoholic. Until such groups are studied, we cannot be
sure that responses of the type made by the alcoholic and homo-
sexual groups are not characteristic of psychiatric patients gener-
ally. Moreover at least two of the three groups used in the present
study were in certain respects rather select. The homosexual subjects
had all been arrested for homosexual offenses. At present, it is
impossible to be sure just what this means in relation to our findings.
It does, however, indicate the need for further study of homosexuals
who are not under medical or legal supervision. The nonpsychiatric
controls were nearly all male school teachers enrolled in a university
extension course in psychology. This means that they were more
homogeneous with respect to education than the other two groups
and probably more sophisticated about psychological tests generally.
Nevertheless, there was no indication that this group was less
naive about the purpose of the symbol test than the other two. All
Ss were requested to state what they thought the test was designed
to do and only two (both homosexuals) were sure that it had
something to do with sex.
The symbol test appears to be a potentially useful research tool.
A great deal more work is, neve1theless, required with the test.
Pictures that are judged male nearly as often as they are judged
female should be replaced, and other methods of presentation re-
quire investigation, e.g., Ss might rate each picture on a rating
scale or rank the pictures in order of preference. In the absence of
adequate research tools for exploring sexual deviation, further in-
vestigation of the strengths and weaknesses of this test should pro-
vide valuable information.
640 llOBERT J. GIBBINS AND lliCHARD H. WALTEllS

SUMMARY

Three preliminary investigations of a psychoanalytic theory of


alcohol addiction, which links this disorder with repressed homo-
sexuality, were carried out with alcoholic patients at a clinic.
In the first two experiments the visual recognition thresholds of
a group of 15 clinic patients were compared with those of 15 overt
homosexuals and 10 normal control subjects. The first of these ex-
periments failed to provide significant findings, though the pattern
of results was somewhat as expected. The second experiment, in
which certain defects of the first were remedied, yielded significant
differences, with alcoholic Ss falling between homosexuals and
controls in the speed at which they recognized sexual words. These
experiments do not, however, afford unequivocal support for the
psychoanalytic hypothesis, since the groups of Ss differed in their
responses to sexual words without homosexual connotations in the
same manner, as in their responses to homosexual words, and to a
more marked degree. Alternative explanations of the findings were
offered.
In the third study the responses of 13 alcoholic and 16 homo-
sexual Ss to pictures symbolizing male and female genitalia were
compared with those of 20 normal Ss. Again, the responses of alco-
holics were intermediate between those of homosexuals and normal
Ss. In conjunction with the first two experiments, this study pro-
vides some support for the psychoanalytic theory. However, the
limited nature of the study necessitates a cautious interpretation of
results.
While the experiments as a whole do not provide strong evidence
for the psychoanalytic theory, the results suggest that this theory
should not be lightly discarded.

REFERENCES
l. ABRAHAM, K. Psychological relations between sexuality and alcoholism. Int. J.
Psycho-Anal. 7: 2-10, 1926.
2. FENICHEL, 0. The Psychoanalytic Theory of Neurosis. New York; Norton; 1945.
3. LOLLI, G. The addictive drinker. Quart. J. Stud. Ale. 10: 404-414, 1949.
4. SCHILDER, P. The psychogenesis of alcoholism. Quart J. Stud. Ale. 2: 277-291,
1941.
5. WEIJL, S. Theoretic and practical aspects of psychoanalytic therapy of problem
drinkers. Quart. J. Stud. Ale. 5: 200-211, 1944.
6. JuLIUSBERGER, 0. Alkoholismus und Psychosexualitat. Z. Sexualwiss. 2: 357-366,
1916.
STUDIES OF PSYCHOANALYTIC THEORY OF ALCOHOLISM 641

7. READ, C. S. The psychopathology of alcoholism and some so-called alcoholic


psychoses. J. ment. Sci. 66: 233--244, 1920.
8. TABOR!, J. Ueber die seelischen Hintergrund des Alkoholismus. Psychoanal.
Praxis 3: 10-19, 1933.
9. DASTON, P. G. Perception of homosexual words in paranoid schizophrenia.
Percept. Mot. Skills 6: 45-57, 1956.
10. BANKS, R. K. and WALTERS, R. H. Prior reinforcement as a determinant of
recognition thresholds. Percep. Mot. Skills 9: 51-54, 1959.
11. WALTERS, R. H., BANKS, R. K. and RYDER, R. R. A test of the perceptual
defense hypothesis. J. Pers. 27: 47-55, 1959.
12. FREUD, S. The Interpretation of Dreams. New York; Basic Books; 1956.
13. FRANCK, K. Preference for sex symbols and their personality correlates. Genet.
Psychol. Monogr. 33: 73-123, 1946.
14. HENRY, G. W. Sex Variants. New York; Roeber; 1944.
15. THORNDIKE, E. D. and LORGE, I. The Teacher's Word Book of 30,000 Words.
New York; Bureau of Publications, Teacher's College, Columbia University;
1944.
16. McGINNIS, E. and SHERMAN, H. Generalization of perceptual defense. J. abnorm.
soc. Psychol. 47: 81-85, 1952.
17. SIEGEL, S. Non-Parametric Statistics for the Behavioral Sciences. New York;
McGraw-Hill; 1956.

You might also like