{3/ [5.511.231
. vii-12.10:)...I-lin2—‘y

\V

3' (r (t...
.‘rir :52355.

:1 I? It] 4.1121.

x. .linl... ?I?:ill.\.l!

ti. igiléllg

gitlilrto‘l‘l.

{3.9:}.
.Falv.)\vll...1a\x§ l‘gv:7.inf§ ul‘fl."l.lll¢l1ll.i

‘ {I it! . .. 3, :{rtvl'li v Isollu’toll.*xl3\ile)§llvnlllt:§.
.V ((5....Vv8I ..I!‘l Ill..1.\|(ftl.._l.:(

 

 

mam

AN STATEU

IIIIIIIIIIIIIIIIII lIIllIIIIlIlIIIIIIIlIIlIIlIIl

31293 00904 8202

 

 

 

 

 

 

 

 

 

This is to certify that the
thesis entitled

Psychodynamic Correlates of Depression:
Hostility and Orality

presented by

Eric J. Dammann

has been accepted towards fulfillment
of the requirements for

 

 

M.A. degreein Psychology
{m M

 

Major professor

Date May 18, 1993

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

LIBRARY

Michigan State ‘
University ‘

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.

DATE DUE DATE DUE DATE DUE

 

 

 

Il' .

1139-: «2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Affirmative Action/Equal Opportunity Institution
c:\drc\datedueorn3-o|

PSYCHODYNAMIC CORRELATES OF DEPRESSION:
HOSTILITY AND ORALI'IY

By

Eric J. Dammann

A THESIS

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

MASTER OF ARTS

Department of Psychology

1993

ABSTRACT

PSYCHODYNAMIC CORRELATES OF DEPRESSION:
HOSTILITY AND ORALITY

By

Eric J. Dammann

This study examined the relationship between hostility, orality, and depression.
Based on selected themes in early psychoanalytic writings on depression (e.g.
Freud, Abraham, Rado), it was expected that depressed patients would exhibit
more orality and hostility directed against the self than non-depressed patients.
Thirty (15 male, 15 female) psychotherapy clients seen at the Michigan State
University Psychological Clinic were assessed using a symptom checklist
(SCL—90-R). Transcripts taken from the first and third psychotherapy
sessions were scored for orality and hostility using scales developed by
Masling and Gottschalk-Gleser, respectively. Results demonstrated a
corrected correlation of .24 for depression scores and inward-hostility (-.06 5
p g .54). The hypothesis concerning depression and orality was not
supported (r = -.10), and regression analysis corroborated the importance of
inward hostility, but not orality, on depression scores. Examination of gender
differences revealed that women scored significantly higher on the oral
measure as well as on inward—hostility. Furthermore, the relationship between
depression and inward-hostility was much stronger for men than women, as
was the relationship between covert hostility and depression. Results are
interpreted as corroborating the importance of hostility (both inward and
outward) in depression, as well as the effect of gender on these relationships.
Discussion also addressed the high variability (over time) of the scales, as well
as their specific use in this context and with this sample.

ACKNOWLEDGMENTS

The help and support of many people made this work possible. First of
all, I would like to thank my committee members for their work and helpful
comments. Dr. Norman Abeles proved to be a patient and pragmatic
chairman, and his influence certainly helped in the completion of this project.
Dr. Joseph Reyher helped clarify both the theory and the methodology, and
our many discussions helped me to learn a great deal about both research and
psychoanalytic theory. Dr. Albert Rabin was able to bring his many years of
experience to this project, and Dr. Ralph Levine helped me to learn a new
way of thinking about statistics.

I would also like to acknowledge the help of two professors not
affiliated with Michigan State University, Drs. Louis Gottschalk and Joseph
Masling. They were both always willing to help, and their suggestions and
comments were appreciated.

Three undergraduate students helped in the long and rather tedious
task of transcribing and scoring psychotherapy sessions. Mary Beaudry,
Mark Donkers, and Lori Traicoff proved to be both patient and quite tolerant
during our many hours in the clinic research files.

Finally, I would like to thank my wife, Christine, for her support
throughout this project. Needless to say, being the spouse of a graduate
student involves a good deal of sacrifice, and I hope I can return the favor
some day.

TABLE OF CONTENTS

INTRODUCTION.

The Psychoanalytic Theory of Depression
Existing Research. .
Hostility and Depression .
Orality and Depression

METHOD

Subjects

Measures .

Design and Data Analysis
Hypotheses .

RESULTS

Reliability .
Descriptive Statistics
Hypotheses .
Gender Differences.
Males .
Females

DISCUSSION
Hostility Scales
Orality Scale.
Methodology
Results
Theory

BIBLIOGRAPHY .

APPENDICES

Appendix A-- Previous Research
Appendix B-- Research Consent Form.
Appendix C-- SCL-90-R Test

Page

l9
19
39

93

107

107
111
112

Appendix D-- Hopkins Psychiatric Rating Scale
Appendix E» Gottschalk-Gleser Hostility Scales
Appendix F-- Masling Orality Scale .
Appendix G—- Additional Tables .

118
119

125

Table

MAWN

10
11

12

13

15

LIST OF TABLES

Intercorrelations between verbal samples for inward hostility
and oral scores . . . . .

Intercorrelations between verbal samples for outward hostility
Descriptive statistics
Correlation matrix for all variables

Regression equation for depression scores on inward hostility
and orality . . . .

Regression equation for depression scores on orality and
overt, covert, and inward hostility

Correlations between depression scores and outward hostility
Correlations between gender and all variables.

Descriptive statistics for males

Correlation matrix for all variables for males .

Regression equation for depression scores on orality. and
inward hostility 1n males .

Regression equation for depression scores on orality and
covert, overt, and inward hostility in males

Correlations between depression scores and outward hostility
in males . . . . . .

Correlations between scales for males .

Descriptive statisu'cs for females .

I?
(D

125
59
61

62

63

125

65

67

67

125
126
69

Table

17

18

19

20
21

Correlation matrix for all variables for females

Regression equation for depression scores on orality and
inward hostility in females

Regression equation for depression scores on orality and
covert, overt, and inward hostility in females .

Correlations between depression scores and outward hostility
in females . . . . . .

Correlations between scales for females

Differences between correlations for males versus females .

71

126
126
72

LIST OF FIGURES

Figure

1 ‘ Formulation of confidence intervals

2 Formula for correcting for attenuation .

viii

esp:

INTRODUCTION

Depression is somewhat of a paradox in the history of psychiatry and
psychology. It has been recognized for over 2,000 years, but has to this day
left clinicians and researchers puzzled regarding its nature, classification,
etiology, and treatment. Furthermore, it remains one of the most powerful
experiences of the human condition. As Burton (1927) wrote, depression
constituted the:

...cream of human adversity, the quintessence, the upshot;
all other diseases are but flea-bitings.... They are in great
pain and horror of mind, distraction of soul, restless, full of
continual fears, cares, torments, anxieties, they can neither
eat, drink, nor sleep for them, take no rest, neither at bed
nor yet at board, will any rest despair afford (cited in
Anthony & Benedek, 1975, p. 545).

Possibly the first reference to depression occurs in the Old Testament,
in which King Saul (during 1033 BC.) described recuning symptoms of
depression including suicidal ideation. The systematic scientific inquiry into
depression begins with Hippocrates (around 400 BC), who believed that
melancholia was caused by an excess of black bile, often the result of long-
standing distressful situations. Since this early hypothesis theorists from
widely diverse backgrounds have attempted to understand the phenomenon
of depression, most with only limited success (see Georgotas & Cancro
[1988] for a discussion of the evolution of the scientific inquiry into
depression and mania).

Currently depression is the most common diagnosis in outpatients and
hospitalized psychiatric patients, and it also affects a large number of
untreated individuals (Lépez-Ibor Jr., 1991). Recent epidemiological studies
suggest a lifetime prevalence rate for major depression in the United States of
4.4 per 100 (Weissman, Leaf, Tischler, Blazer, Kamo, Bruce, & Florio, 1988).

The cost of this disorder in terms of mental anguish, medical expenses, and
even human life (depression is often a factor in suicide) behooves us to further
our understanding of this condition in order to improve treatment and
prevention.

Depression, due to its multifaceted nature, has been approached from a
number of theoretical standpoints. The present literature review will
concentrate primarily on psychoanalytic and psychodynarrric viewpoints;
research dealing with cognitive, interpersonal, and biological aspects of
depression will not be examined. Recently, some authors have made attempts
to pool data from these various areas in order to propose a truly
encompassing theory of depression (See Basch, 1975; Akiskal & McKinney,
1975, for examples of work towards this end).

Many of the early psychoanalytic contributions to the depression
literature continue to stand as models of clinical insight and description for
dynarrrically oriented theorists. This is in part due to the method of
investigation-- the case study-- which allows one to gather large amounts of
data on single cases. This method is obviously not without significant
drawbacks, however, and leads to problems due to observer bias, lack of
control group, etc. Thus there are few empirical studies that have attempted
to examine these early (and still prevalent) ideas about the dynamics
underlying depression.

The proposed study will attempt to isolate a few of the key variables
thought by many psychoanalytic theorists to be crucial in the etiology and
maintenance of depression. Drawing mostly from early psychoanalytic
theory (Abraham, Freud & Rado), the two variables involved in this study
will be inwardly-directed hostility and orality. Before further discussing the
proposed study, a brief exarrrination of the theory to be studied is in order.

The Psychoanalytic Theory of Depression

The title of this section is somewhat of a misnomer, as there really is no
one “psychoanalytic theory of depression.” There have been, however,
certain themes that have run through many of the major theories within the

psychoanalytic school, and this section will examine some of the main
contributors and theoretical positions.

The psychoanalytic literature on depression begins with a paper by
Karl Abraham published in 1911 (1927), in which he discusses six cases of
psychotic depression which he had treated. He begins this paper with an
explanation of the difference between depression and normal sadness:

One of the earliest results of Freud’s investigations of the
neuroses was the discovery that neurotic anxiety
originated from sexual repression; and this origin served
to differentiate it from ordinary fear. In the same way we
can distinguish between the affect of sadness or grief and
neurotic depression, the latter being unconsciously
motivated and a consequence of repression (p. 137).

Abraham believed that ambivalence was the main reason for this
repression, in that alongside the love for the lost object the depressive also felt
intense hatel (Abraham notes the similarity here with obsessive patients). The
patient cannot acknowledge this hostility (especially since they need the object
so greatly) and it is therefore subject to strict repression.

Abraham believed that aside from repressing this hatred, depressive
patients will project it onto others, therefore leading to the belief that others
hate them. Furthermore, he speculated that this repression often leads to
guilt, self-reproaches, etc. Finally, he also noted that their excessive self-
reproaches are of a narcissistic quality.

His second contribution emerged in 1916 (1927), and set out to
understand depression in light of Freud’s new work on the development of
the libido, found in the third edition of the IhLeg Egg g th_e Theog 9f
Sexuality (1949, first published in 1915). In this paper Abraham focuses on
the first pregenital stage of the libido, and gives some case examples to show
that the “instinctual life of the infant persists in some adults in a positive and
unmistakable fashion...” (p. 253). Throughout this paper he gives examples

 

“ It should be noted here that later theorists (see Freedman. 1986) believe that it is unconscious
ambivalence which is prevalent in depression, and that in fact the conscious experience of ambivalence is a
Slgn of clinical improvunent

from both psychopathology and everyday behavior of the lasting effect of
these earlier stages, especially the oral stage. Of interest here is his attention
to the vicissitudes of libido in regards to depression, in which he believed the
libido is regressed to the oral or cannibalistic stage of development:

In melancholic states of depression the libido seems to
regress to the earliest stage of development known to us.
That is to say, in his unconscious the melancholic
depressed person directs upon his sexual object the wish
to incorporate it. In the depth of his unconscious there is
a tendency to devour and demolish his object (p. 276).

He believes that this regression accounts for two similar symptoms in
many depressives-- the refusal to eat and the fear of dying of starvation. Both
of these are related to the unconscious wish to incorporate orally his/her
object. In effect, not eating is the only way to avoid carrying out their
cannibalistic urges.

One can see in these two early papers, therefore, a developing theory
which stresses the importance of hostility and orality in depression. However,
Abraham did not believe that these were necessarily to be considered
universal dynamics: “1 have attempted only to explain the wish-content of
certain depressive delusional ideas and the unconscious strivings that underlie
certain characteristics in the conduct of the melancholic and not the causes of
melancholic depression in gene ” (p. 278).

His third contribution to the literature on depression appeared in 1924
(1927). It should be noted that this paper was published after Freud’s own
work on the subject, soon to be discussed. In this extensive paper Abraham
set out to investigate the stages of libidinal development as they relate to
mental illness. Here again Abraham begins by noting the similarity between
the melancholic and the obsessive patient:

...what is especially interesting to the analyst is the fact
that in all cycloid illnesses the patient is found to have an
abnormal character-formation during his ‘free interval’;
and that this character-formation coincides in a quite

unmistakable way with that of the obsessional neurotic (p.
423).

Aside from the oral attributes, he noticed the anal character traits that
were associated with these two positions. In order to differentiate the two
disorders, however, Abraham proposed a subphase within the anal stage.
These two stages are labeled a “primitive stage” and a “later stage.” In the
primitive stage the goal is control of the object (as with the obsessive), and in
the later stage the goal is to expel it. This “anal expulsion” is what he often
found in the fantasies of depressives. Abraham further believed that once the
libido has crossed this “line” between the two subphases it will continue to
regress to earlier stages. This may be in part due to the fact that after the loss
of the internalized love object (due to expulsion) the depressive is left feeling
empty. As Freud (1917) suggested, the depressive therefore tries to gain
back the object through oral incorporation. Abraham cites the frequency of
the fantasy of eating one’s own feces as evidence of this wish to incorporate
the anally expelled object. Although he believes that introjection is also a part
of normal mourning, he points out the significant differences in depression:

Nevertheless, although introjection occurs in mourning in
the healthy person and in the neurotic no less than in the
melancholic, we must not overlook the important
differences between the process in the one and in the
other. In the normal person it is set in motion by real loss
(death); and its main purpose is to preserve the person’s
relations to the dead object, or- what comes to the same
thing- to compensate for this loss. Furthermore, his
conscious knowledge of his loss will never leave the
normal person, as it does the melancholiac. The process
of introjection in the melancholic, moreover, is based on a
radical disturbance of his libidinal relations to his object. It
rests on a severe conflict of ambivalent feelings, from
which he can only escape by turning against himself the
hostility he originally felt towards his object (p. 438)

(this idea of hostility turned against the self will be more fully discussed in

relation to Freud’s understanding of melancholia).

Just as Abraham differentiated two phases in the anal stage, he goes on
to do this with the oral stage, again noting the importance of the oral phase of
development in depression. The earlier phase is described as the oral-
incorporative phase, and corresponds with sucking in the infant. The later
phase is the oral-sadistic, and corresponds to biting. It is in this later stage
that ambivalence towards the object begins to become important.
Throughout this paper Abraham gives many clinical examples from his
patients of fantasies, symptoms, etc. which revolved around the mouth and
the theme of eating, all of which he took as evidence of the regression to the
oral sadistic phase. Abraham further summarizes this process by using an
interesting metaphor: “We may truly say that during the course of an attack
of melancholia the love-object goes through a process of psychological
metabolism within the patient” (p. 464). Interestingly, this metaphor is still
alive today-- when one is confronted with a difficult or painful situation one
will often say that they have “trouble digesting it.”

With this theoretical discussion of the developmental aspects of the
libido as a back-drop, Abraham posited several etiological factors in manic-
depression. These are: 1) A constitutional factor [possibly an over-
accentuation of oral eroticism]; 2) A fixation of the libido at the oral level;
3) Injury to infantile narcissism due to successive disappointments in love
[especially before the Oedipal wishes have been overcome]; 4) The
repetition of this primary disappointment in later life. As Robertson (1979a)
noted, the third factor listed here is what is primarily responsible for the
ambivalence which is felt about the object (and subsequent love objects).

In summary, Abraham noted, as did Freud, the importance of the
perceived loss of a loved object. He summarizes as follows:

When melancholic persons suffer an unbearable
disappointment from their love-object they tend to expel
that object as though it were feces and to destroy it. They
thereupon accomplish the act of introjecting and
devouring it- an act which is a specifically melancholic

form of narcissistic identification. Their sadistic thirst for
vengeance now finds its satisfaction in tormenting the
ego...(pp. 463-464).

Although Abraham was the first psychoanalytic theorist to write on
depression, it was his mentor, Sigmund Freud, who is still most influential in
this area. Freud’s “Morning and Melancholia,”2 published in 1917 (1957), is
still regarded as pivotal in the history of psychoanalysis. In part its
significance lies in its relation to Freud’s earlier writings and his overall
metapsychological theory, rather than its exploration of depression. For it
was in this paper that Freud first introduced the idea of a form of
psychopathology that was not fundamentally due to the vicissitudes of libido.
Furthermore, this paper emphasizes object-relations (Arieti & Bemporad,
1978). In this paper one can also catch early glimpses of the psychic agency
that would become the superego in The Egg an_d th_e I_d (1961, first published
in 1923). This agency was introduced in a paper written the year before
entitled “On Narcissism, an introduction” (1949, first published in 1914) as
the Ego Ideal.

Freud begins Morning and Melancholia3 with a warning to prospective
readers about the generalizations that can be drawn from this work. Freud
recognized that the definition of Melancholia fluctuates and takes various
forms which may not be a single identity. He also acknowledged that his
material was based on a small number of cases, and should therefore not
necessarily be taken as universal.

He begins his exposition into depression by noting the similarities
between melancholia and morning. He also, however, notes one importance
difference:

The distinguishing mental features of melancholia are a
profoundly painful dejection, cessation of interest in the

 

2 Freud did write on depression before this paper, but this work was never published and Freud later
abandoned many of these early ideas (Bemporad, 1988). Interestingly, Freud first believed that major
depression resulted from an abnormal sexual life, while periodic depression resulted from coitus interruptus
(see Deitz. 1989).

3 For further discussion of the precursors to this paper, see Strachey‘s comments in the Standard
Edition gt“ m Cowlete Psychological Works of Simund Freud, Vol. 14, pp. 239-242.

 

 

outside world, loss of capacity to love, inhibition of all
activity, and a lowering of the self-regarding feelings to a
degree that finds utterance in self-reproaches and self-
revilings, and culminates in a delusional expectation of
punishment. This picture becomes a little more intelligible
when we consider that, with one exception, the same traits
are met with in mourning. The disturbance of self-regard
is absent in mourning... (p. 244).

This difference was succinctly surrrrned by Freud when he stated that “in
mourning it is the world which has become poor and empty; in melancholia it
is the ego itself” (p. 246).

Freud further noted that both mourning and melancholia can result
from the loss of a loved object. However, here again there is a significant
difference in that in melancholia the object loss is unconscious, while in
mourning it is not (this idea is similar to Abraham, as noted above).
Furthermore, in melancholia the object has “not perhaps actually died, but
has been lost as an object of love” (p. 245).

Freud believed that during childhood the melancholic individual must
have had intense ties to an object which were undermined due to
disappointments in this relationship. This leads to a great deal of ambivalence
about the object. Once this relationship ended, the melancholic individual,
due to the narcissistic nature of the relationship, turned this libido onto their
own ego (instead of looking for new objects) in an attempt to hold on to it.
Freud stated that “...by taking flight into the ego love escapes extinction” (p.
257). Unfortunately, this then leads to the ambivalence (originally felt toward
the object) being directed against the person’s own ego.

Toward this end, Freud notes that in melancholia the patient has
delusions of moral inferiority, is often sleepless, will not eat, and is not
motivated by the “...instinct which compels every living thing to cling to life.”
(p. 246). He speculated that this self-reproach must originate in the ego, and
described in part what he would later call the superego:

We see how in him one part of the ego sets itself over

against the other, judges it critically, and, as it were, takes
it as its object. Our suspicion that the critical agency
which is here split off from the ego might also show its
independence in other circumstances will be confirmed by
further observation. We shall really find grounds for
distinguishing this agency from the rest of the ego. What
we are here becoming acquainted with it the agency
commonly called ‘conscience’; we shall count it, along
with the censorship of consciousness and reality-testing,
among the major institutions of the ego, and we shall
come upon evidence to show that it can become diseased
on its own account. In the clinical picture of melancholia,
dissatisfaction with the ego on moral grounds is the most
outstanding feature (pp. 247-248).

Freud then made a crucial observation about these self-accusations:

If one listens patiently to a melancholic’s many and
various self-accusations, one cannot in the end avoid the
impression that the most violent of them are hardly at all
applicable to the patient himself, but that with insignificant
modifications they do fit someone else, someone whom
the patient loves or has loved or should love... So we find
the key to the clinical picture; we perceive that the self—
reproaches are reproaches against a loved object which
have been shifted away from it on to the patient’s own
ego (p. 248).

What we see here is an early example of object-relationships being the
key feature of psychopathology along with the vicissitudes of libido. In
summary, Freud believed that this resulted from a shattering of a once
libidinized object-relationship. In melancholia, this libido is not displaced onto
a new object (as it usually is) but rather withdrawn into the ego, serving to
establish an identification with the lost object.“ In Freud’s now classic phrase:
“The shadow of the object fell upon the ego” (p. 249). Therefore the object-
loss becomes an ego-loss, and the feelings directed towards this object
become directed against the ego that has identified with this object.

 

‘ An interesting paper by Bak (1973) describes some similarities between “being in love" and many
of the dynamic factors important in depression (i .e. overvaluation of the object, previous object loss,
psychic regression).

10

Freud believed that identification is a more primitive form of object-
relationship than a mature genital relationship, and as already mentioned
Abraham suggested that this identification involves fantasies of oral
incorporation. Furthermore, like Abraham, Freud believed that this primitive
mode of relating was due to a regression to the narcissistic oral phase. This
phase is said to be narcissistic because all of the libido is invested in the
person’s own ego (during the oral phase the infant does not yet have the
mental ability to form object relationships, or differentiate objects from
him/herself. Therefore it is believed that at this point the child is narcissistic).

Another important aspect of this object relationship is ambivalence.
Freud believed that: “The melancholic’s erotic cathexis in regard to the
object has undergone a double vicissitude: part of it has regressed to
identification, but the other part, under the influence of the conflict due to
ambivalence, has been carried back to the stage of sadism...” (pp. 251—252).
This tendency towards sadism is what Freud believed to be at the root of the
suicidal preoccupation so often found in melancholia (Interestingly, Bellak
[1952] mentions that some authors feel that the wish to die is in fact an oral
wish to be able to relax, be passive, and to sleep, as if at the mother’s breast).

Although Freud never again discussed depression in similar detail, he
did return to the subject briefly in later writings, although they will not be
discussed here (See Robertson, 1979a, 1979b; Parkin, 1976, for a discussion
of these later ideas).s Although some of these later writings contradict his
earlier ideas (Bemporad, 1988; Pedder, 1982), they were never as fully
developed, or as influential, as those found in “Mourning and Melancholia.”

The final theorist who is considered a contributor to the “classical”
analytic views on depression is Sandor Rado. In 1928 (1968) he published a
paper which, although not ground-breaking in terms of theory, did offer a
concise view of the existing theory, especially in light of Freud’s new
structural theory which was published in m Egg and The Id (1923/1961).

Rado was able to incorporate this new theory into the existing theory of

 

5 Interestingly. independently of Freud an Englishman, Alfred Carver. came to similar conclusions as to
the dynamics involved in melancholia. although he was apparently unfamiliar with Freud’s paper (cited in
Mendelson, 1974).

11

depression, especially emphasizing the role of the superego and early infantile
situations which are the precursors of later depression.

Rado believed that the most striking feature of depression was the fall
in self-esteem and self-satisfaction (a notion that was to be especially
emphasized by later ego-psychologists). He states:

We find in them, above all, an intensely strong craving for
narcissistic gratification and a very considerable narcissistic
intolerance. We observe that even to trivial offenses and
disappointments they immediately react with a fall in self-
esteem. Their ego then experiences an urgent craving to
relieve in some way or other the resulting narcissistic
tension (p. 72).

Unfortunately, since their attitude towards others is basically narcissistic, they
need to gain self-respect from without, thereby leaving them forever
vulnerable to other’s actions. Their own accomplishments mean very little to
them if no one else recognizes and loves them. In this sense he believed that
depressed persons were very much like children.

The narcissistic object ties lead to a further pitfall:

But as soon as they are sure of the affection or devotion
of another person and have entered into a fairly secure
relation with him or her their behavior undergoes a
complete change. They accept the devoted love of the
beloved person with a sublime nonchalance, as a matter of
course, and become more and more domineering and
autocratic, displaying an increasingly unbridled egoism,
until their attitude becomes one of full—blown tyranny.
They cling to their objects like leeches (to use a phrase of
Abraharn’s) and feed upon them, as though it were their
intention to devour them altogether (p. 74)

Here we see that Rado, like Abraham and Freud, believed that this narcissistic
orientation had a strong oral-sadistic character to it.

Rado further examined to ego’s self-reproaches, and described this as
theego doing “penance,” while also exhibiting “a great despairing cry for

12

love.” However, at this point he points to an interesting paradox, in that the
theory posits that the depressed person has withdrawn their interest from the
object. Why then is the person striving for reconciliation and love from this

object? Rado states:

...but the melancholiac has transferred the scene of his
struggle for the love of his object to a different stage. He
has withdrawn in narcissistic fashion to the inner world of
his own mind and now, instead of procuring the pardon
and love of his object, he tries to secure those of his super-
ego. We know that his relation to the object was marked
by the predominance of the narcissistic desire to be loved,
and it is quite easy for this aspiration to be carried over to
his relationship with his own super-ego. It is as if the ego
of the melancholiac were to say to his super-ego: “I will
take all the guilt upon myself and submit myself to any
punishment; I will even, by ceasing to care for my bodily
welfare, offer myself as an expiatory sacrifice, if you will
only interest yourself in me and be kind to me”
(quotations in original, p. 75).

Although Rado does not mention the term introjection, one can see the
similarity between what he is describing and the introjective process as
described by Freud and Abraham.6 Furthermore, Rado’s explanation
attempts to shed light on the sometimes astounding way in which the
depressed person will let their life and even their own self go to pieces. Rado
also noted the similarity between the “desperate cry for love” that occurs
between the person and the object and the later attempt at forgiveness that
occurs between the ego and the superego. Rado notes the similarity of the
process to the normal formation of the supere go, but states that in the
depressed person this process goes too far and is especially problematic
because it is unconscious. Therefore, although Freud hinted at this concept in
his work, Rado was the first theorist to specifically posit an overly harsh
superego as one of the main problems in depression.

 

° It should be noted that in many of these writings the terms incorporation, introjection, and
identification are used interchangeably Robertson (1979b) gives a concise explanation of these terms and
their differences.

13

Rado uses infantile hunger as the prototype of the self-persecution and
feelings of depression:

I think that if we trace the chain of ideas, guilt-
atonement- forgiveness, back to the sequence of
experiences of early infancy: rage, hunger, drinking at
the mother’s breast, we have the explanation of the
problem why the hope of absolution and love is perhaps
the most powerful conception which we meet with in the
higher strata of the mental life of mankind (p. 80- italics in

original).

In Other words he believed that this sequence was the way in which the infant
learned to obtain forgiveness and love from the mother by appropriate
remorseful acts. It is this remorseful behavior in an attempt to win back love
that Rado sees as imperative in the understanding of melancholia. It is only
when this does not work with the object that the battle is internalized and
continued in the psychic plane.

In 1951 (1968) Rado again wrote on the subject of depression. He
continued the exploration of the role of repentance in depression, but
introduced some new factors which complicated the picture:

However, the patient’ s dominant motivation of repentance
is complicated by the simultaneous presence of a strong
resentment. As far as his guilty fear goes, he is humble
and yearns to repent; as far as his coercive rage goes, he is
resentful (p. 98).

It is easy to see from this formulation how a depressed person can feel so
helpless and stuck, and one can’t help but notice the similarity here with
Abraham’s first contribution, in which hate was seen as a paralyzing force.

Rado also explored this rage in relation to the depressed person’s guilt.
He believed that, due to their aggression towards the object, the person feels
that they are to blame for its loss (an idea that is not always far from the
truth).

14

In summary, Rado states:

Based on these findings, we view depression as a process
of miscarried repair. To a healthy person a serious loss is
a challenge. He meets the emergency by calming his
emotions, marshaling his remaining resources, and
increasing his adaptive efficiency. Depressive repair
miscarries because it results in the exact opposite.
Anachronistically, this repair presses the obsolete adaptive
pattern of alimentary maternal dependence into service
and by this regressive move it incapacitates the patient still
more (p. 101-).

It is this regression to old adaptational patterns which causes the ambivalence,
because the purpose of these patterns is: “...to destroy the frustrating aspect
of the beloved one (formed in the split-off image of the ‘frustrating mother’),
while retaining the gratifying aspect of the beloved one (formed in the split-
off image of the ‘gratifying mother’)” (p. 102, italics in original).

Although these three contributors are the main source of the classical
theory, two other authors have written important book chapters which
emphasize and elaborate on many of the theoretical points already described.
The first of these was Helene Deutsch, who in 1932 described an intense case
of depression. Although this case study adds little that is new to the existing
theory, she does offer corroboration of many of the earlier findings from her
extensive clinical rrraterial. For instance, she notes the presence of both
obsessional and melancholic symptoms in her patient. She also noted that:
“It not seldom happens that the outbreak of depressive states is brought
about by an apparently trivial loss, a change of abode or something similar.
These events are merely the immediate and welcome occasion for the break—
through of deeper, more significant, and hitherto suppressed reaction” (pp.
219-220). The following passage describes case material that is in line with
the material discussed so far:

By pursuing our patient’s psychical development we are
able to form a consecutive scheme of what went on within

15

her. First, hatred and aggression against her sister;
defense against these impulses through obsessional
neurotic mechanisms; afterwards successful over-
compensation for the hate through love and tenderness;
satisfaction for the narcissistic injuries through
identification with the sister; and finally, transformation of
the aggressions into a masochistically satisfying self—
sacrificing for her... After the disappointment at the hands
of her sister this psychical arrangement is not given up; it
is only added to by new quantities of aggressive impulses,
until the patient becomes seriously ill. The identification is
maintained, as well as the masochistic turning against the
ego. The punishment to which she had doomed the sister,
of being “thrown out into the street”, in order that she
should meet with a miserable end there, we hear the
patient demanding with monotonous regularity, no longer,
however, as a threat against her sister but against herself...
(pp. 220-221).

For the interested reader, Deutsch goes on to give many examples from this
case which support the theoretical ideas put forth by Abraham and Freud.
In Fenichel’s encyclopedic m Psychoanalytic Theog 9f Neurosis
(1945) he devotes a chapter to depression and mania. Fenichel begins this
chapter with an emphasis, following Rado, on self-esteem in depression:

A person who is fixated on the state where his self-esteem
is regulated by external supplies or a person whose guilt
feelings motivate him to regress to this state vitally needs
these supplies. He goes through this world in a condition
of perpetual greediness. If his narcissistic needs are not
satisfied, his self-esteem diminishes to a danger point (p.
387).

Furthermore, due to the narcissistic nature of their object relationships
depressives are almost never satisfied:

These persons, in their continuous need of supplies that
give sexual satisfaction and heighten self-esteem
simultaneously, are “love addicts,” unable to love actively;
they passively need to feel loved. Besides, they are

16

characterized by their dependence and their narcissistic
type of object choice. Their object relationships are mixed
up with features of identification and they tend to change
objects frequently because no object is able to provide the
necessary satisfaction (quotations in original- p. 387).

Fenichel goes on to discuss the oral and anal dynamics involved in
much the same way as Abraham. Furthermore, in discussing the introjection
(and the resulting inwardly directed hostility) that takes place, he states :

The outcome is that the struggle subject vs. introject
becomes complicated in two ways: in the foreground is
the struggle superego vs. ego + introject; but the ego, in
its ambivalence toward the superego, changes it also into a
struggle of ego vs. superego + introject (italics in original-
p. 393).

Fenichel, in a similar vein as Freud, also notes the ambivalence and
how it serves to differentiate mourning and depression:

Mourning becomes more complicated or even
pathological if the relationship of the moumer to the lost
object was an extremely ambivalent one. In this case the
introjection acquires a sadistic significance; the
incorporation then not only represents an attempt to
preserve the loved object but also an attempt to destroy
the hated object. If a hostile significance of this kind is in

the foreground, the introjection will create new guilt
feelings (pp. 394-395).

These guilt feelings are brought about by the superego which has enlisted the
sadism inherent in the introject, an idea which Freud was leading to in
“Mourning and Melancholia” (although the agency of the superego had not
yet been differentiated from a “part of the ego”). Fenichel states that in
melancholia it is as if the main emphasis of the personality is now the
superego, instead of the ego.

In summary, Fenichel gives a concise description of the etiology of

depression:

Now at last we are in a position to understand which
conditions actually make for the predisposition for
subsequent depressions. The decisive narcissistic injuries
must have taken the form of severe disappointments in
the parents, at a time when the child’s self-esteem was
regulated by “participation in the parents’ omnipotence”.
At this time, a dethroning of the parents necessarily means
a dethroning of the child’s own ego. Probably it is not
only so that after disappointments of this type the child
asks for subsequent compensating external narcissistic
supplies throughout his life, thus disturbing the
development of his superego; he also tries to compensate
for his parental insufficiencies by the development of a
specially “omrripotent”, that is, strict and rigid, superego...
(quotations in original- p. 405).

Therefore, in Fenichel’s view it is not necessarily the loss of an object that is
being mourned, but more the loss of the self-esteem that the object
symbolized. It is possible that this too leads to the depressive’s quest for
perfection and their feeling that they have to achieve for the whole family, a
dynamic that Cohen, Blake, Cohen, Fromm-Reichmann, and Weigert (1954)
found was quite prevalent in their intensive case studies of manic-depressives.
Although much of this theory was developed during the beginning of
this century, many of the hypotheses derived from it are still widely accepted
and used in the treatment of depressed individuals. For example, Stone
(1986), in a review of his work with 23 depressives, summarized that:

...the considerations of pathologic narcissism (including the
overexigent ego ideal), aggression toward the primary
object, oral ambivalent regression, and “identification”
with the disappointing object, are in varying degree, and
inner mutual relatedness, important in most if not all cases
of true depressive illness, and should be taken into account
in evaluating their pathology. While each of these factors
is a dynamic entity in its own right, they operate
synergistically in the complex of depressive illness
(quotations in original- p. 359).

18

Papers by Whitaker and Deikrnan (1980), Bellak (1981), Dalrl (1988),
Milrod (1988), and Teixeira (1990) also describe the therapy of depressed
individuals based on many of the theoretical propositions described herein.
Interestingly, even Sullivan (1956), although of a different theoretical bent
than those theorists discussed so far, noted the presence of hostility in

depression:

Although I have not been able to document my
observations to any extent, an equally important surmise I
would make is that other people must suffer from the
depression. Sometimes the performance of the depressive
is quite clearly punitive. And the troublesome
performances of a depressed patient occasionally
disappear when it becomes evident that there is nobody
who would suffer from them (italics in original, pp. 296-

297).

Furthermore, he notes the process of hostility turned against the self in

suicide:

And just as I have said at various times that
schizophrenics kill themselves by misadventure, I think
that depressives do too- the misadventure being that they
die in the process of making a supposedly unhealing
wound in an enemy... It left me with what I believe is a
supportable hypothesis- namely, that a particular person
who was a destructive influence in the patient ’s past is
the target of the patient’s self-destruction. The intention
is that that person shall sufler the rest of his life because
the patient has destroyed himself (italics in original, p.
298).

As is evident from this review, although important differences do exist,
all of these theorists (other than Sullivan) posit the importance of inwardly
directed hostility and oral fixation or regression in depression.7 Let us now

 

I It should be noted that there has been a good deal of more recent work on depression by
psychoanalytically oriented theorists. Overviews of work by such theorists as Gero, Klein. Lorand, Bibring,
Jacobson. Sandler, and Bowlby can be found in Mendelson (1974), and many of their seminal works have
been collected in Gaylin (1968).

l9

examine some of the empirical work which has attempted to elucidate these
hypothesized relationships.

Existing Research

As with other areas in psychoanalysis, although the analytic theory of
depression is quite prevalent in current thought on dynamics and treatment,
there is only a small amount of empirical investigation into the psychoanalytic
theory of depression. The following section will examine critically the
research that has attempted to assess the validity of certain aspects of the
theories previously discussed. In order to conserve space, only the major
conclusions will be discussed here, and readers interested in details such as
effect size are directed to Appendix A.

Hostility and Depression

In a paper examining epidemiological data about depression and
suicide, Kendell (1970) set out to check the hypothesis (based on the
psychoanalytic theory of aggression turned against the self) that depression
should be greater in situations where aggression is aroused but its expression
is prevented, than in situations where there is little frustration, or aggressive
outlets are available. He believes that this hypothesis is easier to test
empirically than the intrapsychic psychoanalytic hypotheses, even though
they are similar and would expect similar results. Using all major sociological
studies available he found several conclusions which corroborate this
hypothesis. For instance, he found a stable inverse relationship between
homicide and suicide in a number of geographical locations, as would be
predicted. Kendell states: “Taken together, these facts constitute an
impressive body of evidence, but the inverse relationship which they suggest
is not invariable” (p. 311). Other relationships he examined were cultural
variation, social class, gender, ethnicity and incidence of depression. In
conclusion Kendell states that:

All that can be said at present is that there is no flagrant

20

discrepancy between existing evidence and the
requirements of the hypothesis... Perhaps the strongest
argument in favor of the “inhibited aggression”
hypothesis is that it accounts for most of what we know
of the epidemiology of depression by means of a single
assumption (quotations in original- p. 316).

Kendell does note, however, that there are other possible explanations for
these findings and that this relationship is by no means completely
understood.

A group of studies have attempted to assess this relationship
empirically by using the Hostility and Direction of Hostility questionnaire.
(HDHQ- Caine, Foulds & Hope, 1967), a measure constructed from 52
MMPI items which is used to assess intropunitiveness (broken down into self-
criticism and guilt scales) and extrapunitiveness (broken down into acting-out
hostility, projected delusional hostility, and criticism of others). The first of
these was done by Foulds, Caine and Creasy (1960), who examined 40 (20
men and 20 women) psychiatric in-patients in each of the following diagnostic
categories: Hysterics with hysteroid personality, dysthyrrrics with hysteroid
personality, dysthyrrrics with obsessoid personality, melancholics, paranoid
states, and psychopaths (the way in which these diagnoses were reached was
not disclosed). A group of 40 hospital staff members (or members of their
families) made up the comparison group.

Only the results relevant to the current discussion will be addressed. It
was found that for males, the melancholics, along with paranoid patients and
psychopaths, differed significantly from all other groups on the projected
delusional hostility scale. The fact that the male melancholics displayed more
projected hostility is in contradiction to the psychoanalytic theory of
aggression turned against the self. No difference for female melancholics was
found on this measure, but it should be noted that the melancholic women
scored lower on this than any of the other groups of women patients. On the
self-criticism measure obsessoid dysthymic men scored significantly higher
than the rest. This also was not found for women. Both male and female

21

melancholics exhibited significantly more delusional guilt than all other groups
except psychopaths, as would be expected from psychoanalytic theory.
Finally, this study also points to the possibility that men and women
experience depression differently, a proposition that has not received adequate
empirical investigation.

It should be kept in mind when evaluating these results (and those of
the studies that immediately follow) that the measures being used are entirely
self—report, a method of research which has come under some scrutiny
(Kagan, 1988). This problem may be especially prevalent with certain
depressed individuals, who are hypothesized to avoid or deny their hostile
impulses.

The next study to use this questionnaire was done by Mayo (1967).

He examined 24 depressed inpatients (16 women, 8 men). Subjects were
tested twice, once upon admission, and once six weeks later. In the interim
subjects were treated with drugs or E.C.T. (the effects of which were not
addressed). At both pre— and post—test subjects were given the HDHQ, a
symptom-sign inventory (581- which discriminates between psychiatric
diagnoses), and a hysteroid-obsessoid questionnaire. Clinical improvement
was rated on a five point scale by a consultant psychiatrist.

The results of the SSI differentiated the subjects into 8 melancholics
and 14 neurotic depressives. It was found that none of the extrapunitive
scales changed significantly with clinical improvement. Both of the
intropunitive scales (self—criticism and guilt) showed significant decreases with
improvement. Differences were also found between the melancholic patients
and the neurotic depressives, such that the melancholics showed more overall
punitiveness, criticism of others, projected delusional hostility, and guilt.

Philip (1971) used this measure to compare a group of 18 depressed
women who improved after drug treatment to 18 women who did not (as
measured by the Beck Depression Inventory [BDI]; Beck, Ward, Mendelson,
Mock & Erbaugh, [1961]). Both groups were nearly identical in terms of
severity of depression upon admission.

It was found that the extrapunitive scales did not change over time, and

22

there was no difference between the groups. Intropunitiveness did differ
significantly both between testing occasions and between groups,
demonstrating that patients who did not show improvement were more
intropunitive at both pre- and post-test. Subjects who did improve
demonstrated significantly less intropunitiveness at post-test than at pre-test.
Also, compared to normative scores both groups of patients scored
significantly higher on the intropunitive scale.

The final study to use this measure was done by Blackburn (1974),
who attempted to assess the differences between six groups of patients in
terms of hostility. The patient groups were (as diagnosed by a consultant
psychiatrist): Bipolar- actively manic, Bipolar— recovered manic (no present
symptomatology, last episode was manic), bipolar- actively depressed, bipolar-
recovered depressed (no present symptomatology, last episode was
depression), unipolar— actively depressed, and unipolar- recovered depressive.
There were 18 patients in each group except the bipolar depressed group,
which had 16. The gender of the patients was approximately even in each
group.

Compared with the normative data (in Philip, 1971), both groups of
patients with active depression scored significantly higher on the
intropunitiveness scales. Both the manic and all of the recovered groups did
not differ significantly from the normals. On the extrapunitiveness scale, only
the active manic group was higher than the normals. In comparing the three
active groups, it was found that the manic group had significantly lower
intropunitiveness scores and higher extrapunitiveness than the depressives
(both uni- and bipolar). In comparing the active and recovered groups, it was
found that for both the bipolar and unipolar depressives the recovered group
were significantly lower than the active group on intropunitiveness. On
extrapunitiveness, the active manic group was significantly higher than the
recovered group; no differences were found for either depressed group.

Although this study still does not answer any questions regarding
etiology or cause and effect, it does demonstrate that internalized hostility is
only a factor in active depression, and that once recovery from depression has

23

taken place, inward hostility returns to “normal” levels.

. Taken as a whole, these four studies offer some corroboration for the
psychoanalytic theory of the co-existence of inwardly directed hostility and
depression, as well as demonstrating that upon recovery this hostility returns
to levels comparable with “normal” (non-depressed) subjects.

A study by Caine (1960) used a precursor of the HDHQ to explore the
expression of hostility and guilt in 26 hospitalized melancholic and paranoid
women. Other measures used were a sentence building task, the TAT (scored
for hostility and guilt), and a psychomotor tapping task. Caine found that on
the HDHQ scales the depressed subjects had significantly higher scores on the
self-criticisms/ guilt scale, but significantly lower scores on the projection of
hostility scale than the paranoid patients, as was expected from psychoanalytic
theory. The sentence building task demonstrated the same results, and on the
TAT the paranoid patients demonstrated more hostility than the
melancholiacs. The main problem with this study is that due to the lack of a
control group it is impossible to compare these results with “normals.”
Therefore even though the results were as expected in terms of the
psychoanalytic theory of depression, we do not know if the melancholics here
would have shown significant differences in regards to these scales when
compared to “normals.” However, based on the results of the previously
mentioned studies, this study adds further corroboration for the relationship
between intropunitiveness and depression.

Another self -report questionnaire that has been used in the study of
depression is the Buss-Durkee Inventory, 3 46 question inventory which
divides hostility into 7 subclasses (assault, indirect aggression, negativism,
verbal hostility, resentment, suspicion, and internalization of anger). Friedman
(1970) used the Buss-Durkee Inventory to assess aggression and hostility in
534 depressed inpatients (71% of whom were female). Friedman was
attempting to assess changes in hostility as related to clinical improvement. In
order to monitor change he had the patients rate how they were feeling each
day, as well as answer a 22-item inventory about behavior and feelings. At
baseline (when the patients were first admitted), the depressives reported

24

significantly less verbal hostility and more resentment than matched controls
from the community.

By examining the subtypes of depression from the Buss-Durkee
Inventory, some other interesting results were found. For instance, the
amount of extemalization of aggression and internalization were positively
correlated, suggesting that there is not a static quantity of aggression that is
either expressed inward or outward.

In terms of the relationship between hostility and patient-rated
improvement, it was found that by the seventh week of treatment, the greater
the improvement the less the hostile-aggression of all types. The most
significant effects were found for resentment and the amount of
internalization. He believes that the results are most parsimoniously explained
by the fact that as the patients improve, they feel better and have less need or
desire to be hostile, as they are feeling better. He also states that these results
are probably not due to the medication, as they were only taken for the first
five weeks.

Another finding was that depressed people expressed less verbal
hostility than controls while depressed, and that this difference increased even
further with clinical improvement. Friedman states that it is possible that in
depression-prone individuals the tendency (even when not depressed) is to
express very little verbal hostility. Even when not depressed, although they
express more hostility than is usual for them, it is still less than “normals.”
He states: “Their tendency to deny the‘bad’ in significant others, and to
perceive them selectively so that they do not consciously become angry and
depressed, may be greater during their ‘benign’ remitted, symptom-free
intervals; and may be one of their ways of attempting to ward off a disturbed
or depressive reaction” (p. 532- quotations in original).

One final finding of interest was that a significant positive correlation
was found between degree of depression and internalization, and the highest
correlation with the guilt-worthless scores from the Buss-Durkee Inventory
were also with internalization. Therefore, even though internalization did not

differentiate between depressed subjects and the controls, when considering

25

just the patient sample it did correlate with degree of depression and guilt
feelings (and there is a significant positive correlation between decrease of
internalization and improvement). He states that this finding lends some
corroboration to the classical view of depression.

This study, like those discussed previously, has the problem of using
only self-report measures of depression. Also, even though the patients only
took medication for the first 5 weeks, the effects of the medication could
conceivably still have been present at the seventh week. Another problem
with this study (and many others) is that, as Friedman rightly points out, the
definition of internalization used in the Buss-Durkee is not really in line with
psychoanalytic writings. Friedman describes this scale as being composed of
“...items...including such sorrratization items ‘When I am mad or angry, I
usually get a headache,’ and also tendencies to irritability.” (p. 526). As is
quite apparent, this is not the same conceptual framework as described by
Freud, Abraham, and Rado. Finally, a comparison between an outpatient and
a control group might have been better as it would have avoided any possible
effects of hospitalization.

Weissman, Klennan and Paykel (1971) used interview data to examine
hostility in 40 depressed women (outpatients) and 40 controls
(nonsymptomatic). At the initial interview a psychiatrist rated
“uncooperativeness” and hostility based on the interview. Also assessed by
the psychiatrist was the patient’s hostility towards others, their irritability
(both based on their self-report during the interview), and a research assistant
rated their friction with others based on the interview material.

It was found that depressed individuals demonstrated little hostility in
the actual interview, but hostility towards others was rated in the moderate
range, especially in regards to their immediate family. The authors state that
this is in agreement with the psychoanalytic theory, which posits ambivalence
towards love objects.

A study by Pilowsky and Spence (1975) attempted to examine the
relationship between hostility and “endogenous” and “nonendogenous”
depression. They cite past research which has been somewhat contradictory

26

in terms of the relationship between these variables, and also discuss past
factor analytic research which has demonstrated different patterns or
subgroups of depression, often involving an endogenous type and a second
that involved self-pity, hypochondriasis, complaining and demanding behavior,
irritability and hostility, and anxiety.

This study examined self-ratings of anger by 68 depressed inpatients
(53 were women) in relation to the “endogenicity” of their depression.
Classification of endogenous or nonendogenous was made by way of a self-
administered depression questionnaire (see Pilowsky & Spence, 1975, for an
example of this questionnaire). Anger and sadness were measured by way of
visual analogy scales in which the subject indicates on a line (the extremes of
which are labeled “1 do not feel at all angry” and “I feel as angry as I could
be”) how they are feeling. The authors believed that the nonendogenous
depressives would report more hostility than the endogenous group.

They found that the nonendogenous group did indeed obtain a mean
anger score that is “substantially higher” than the endogenous group,
although this difference failed to meet statistical significance. They state that
their results corroborate the use of hostility as a factor in determining
depressive typology, and state: “This finding is in keeping with the
importance that clinicians have ascribed to inhibited expression of hostility in
endogenous depression, and supports the View that the presence or absence
of anger constitutes an important variable in the genesis of depressive
syndromes” (pp. 1158-1159).

Certainly the main weakness of this study is the self-report nature of
the measure used to classify the patients and to assess their feeling states.
Also, the psychoanalytic theory does not state that depressed individuals will
necessarily report or even be conscious of hostility. In fact, according to
some theorists one would expect depressed individuals to report less overt
hostility, as this hostility is either being turned against the self or being
avoided so as not to threaten the loss of a love object (Rubinfme, 1968).

In Gottschalk and Gleser’s (1969) book examining the measurement of
psychological states through the content analysis of verbal passages, they

27

describe many validation studies of their hostility scales which speak to the
question of hostility and depression. Very briefly, they created three hostility
scales (hostility directed in, out, and ambivalently) for use in scoring five
minute samples of free associations (a more thorough discussion of this
method and scoring will be discussed in the method section of the present
work).

As part of their initial investigations into the efficacy of these scales,
they found that in a group of 19 psychiatric inpatients their Hostility Directed
Outward Scale (HDO) correlated negatively (-.28) with depressed state, and
their Hostility Directed Inward Scale (HDI) correlated .35 with depressed
state, as would be expected from psychoanalytic theory. They also described
a study comparing 12 depressed and 12 nondepressed patients and found the
HDO score to correlate -. 18 with the EDI, and the I-IDI Scale correlated .47
with the BDI. In a study of 50 outpatients the HDI scale correlated .34 with
the BDI and .52 with depression scores from an adjective checklist, while the
Ambivalent Hostility Scale correlated .37 with the BDI. In all, the
correlations from these validation studies are all in the expected direction
based on psychoanalytic theory (see Chapter 6 of Gottschalk and Gleser,
1969, for a more thorough discussion of these studies).

Since the introduction of these scales, a large body of research has
utilized them. Two studies by Klennan and colleagues (Klennan & Gershon,
1970‘; Gershon, Cromer, & Klerman, 1968) used these scales to examine
over time (11 weeks) a small number of patients who were taking
psychotropic medication. Because these studies involved only 3 and 6
patients, respectively, these results should be considered exploratory. In the
first study, no significant differences on any of the scales were found between
the periods before and during treatment (although there was a slight
reduction in hostility-in), even though the patients did show clinical
improvement. The second study utilized a modification of the Gottschalk-
Gleser scales, which involved separating the self-critical statements from
verbal expressions of depression, both of which are part of the hostility-in

 

' See this article for a discussion of other studies examining the effects of antidepressants on hostility.

28

scale in the original formulation. Gershon et al. (1968) believe that since they
were in part testing the formulation that hostility directed inwards may lead to
depression, the verbal expressions of depression should be scored in a
separate category, which they called “affect-in.” The remainder of the
hostility-in scale was named the “modified hostility-in scale.” The results
revealed no difference in any of the hostility scores based on whether the
patient was taking medication or not; therefore medication effects were
ignored (this result also speaks to some of the concerns raised previously in
this discussion). They found that depressive symptomatology was positively
correlated with hostility-in, as well as their two modified scales. No
correlation was found with hostility-out and depressive symptomatology.
However, they noticed that 2 patients who had hysterical traits appeared to
exhibit significantly higher hostility-out than the depressed patients who did
not have hysterical traits. They speculate that there might be two different
patterns in depression, a notion that has received support elsewhere (Hamilton
& White, 1959; Rosenthal & Klerman, 1966; Rosenthal & Gudeman, 1967).

Gottschalk, Hoigaard, Birch and Rickels (1979) used these scales to
examine the relationship between hostility and symptomatology in 35
outpatients who were receiving either psychotropic medication or placebo
(they did not mention the gender of the subjects). Subjects were assessed at
both pre- and post-dmg periods. Only the results pertaining to depression
measured at the pro-drug period will be addressed here; they found that
depression scores on a symptom checklist correlated .37 with inward hostility,
while the other hostility scales were uncorrelated with depression. In this
same year, Schofer, Koch, and Balck (1979) set out to undertake a normative
study of these scales in Germany, as an attempt to further validate their
usefulness. In a sample of 406 subjects they found a significant correlation
between depression, as measured by a mood adjective check list, and inward
hostility (r = .22). Unlike in previous studies, they also found a significant
correlation between depression and overt hostility-out (r = .16).

Lemaire & Clopton (1981) used the Gottschalk-Gleser scales to
examine 7 depressed (as assessed by the MMPI) and 11 control

29

undergraduate students (12 females, 6 males) over a six-week period. Each
subject was assessed once a week. They found that depressed subjects
expressed more inward, outward, and total hostility than non-depressed
subjects. For both groups there was an inverse relationship between inward
and outward hostility. They also found no significant changes in hostility over
time.

Another study which used the Gottschalk-Gleser scales was undertaken
by Rubin (1986), parts of which were later reported by Rubin, Abeles, and
Muller (1992). Rubin (1986) used the Gottschalk-Gleser scales to examine
hostility in 40 outpatient psychotherapy patients. Symptomatology was
assessed via the symptom checklist (SCL-90-R), which has a depression scale.
For the whole sample, he found a significant correlation (r = .34) between
inward hostility and depression; for females, this correlation was even higher
(r = .54).

Other methodologies have also been used to assess the relationship
between depression and hostility. Wessman, Ricks and Tyl (1960) used the
Rosenzweig Picture Frustration Test (a projective measure”) and a mood scale
to assess mood fluctuations in 25 college women. Contrary to prediction, the
authors found that the frequency of extrapunitive responses was significantly
higher, and intropunitive responses were significantly lower, in subjects
reporting depression. They suggest that the point at which hostility shifts
from extrapunitive to intropunitive may be when relatively minor depression
(note that the subjects were college students, and not patients) becomes a
severe affective disorder. Some corroboration for this notion is found in a
dissertation by Bulatao (1961, cited in Gershon, et al., 1968), which found
that in a sample of hospitalized depressed women there was a lower
frequency of extrapunitive responses and a higher frequency of intrapunitive
responses in depressed patients.

Cochrane ( 1975a) has examined the relationship between hostility and
depression using the Object-Relations Technique, a projective test for which

 

‘ This test assesses the direction of aggression as either “extrapunitve,” “intropunitive,” or
“impunitrve.”

30

he has devised an aggressive content scoring system. Simply put, the
projective material was scored based on whether the aggression was
controlled or uncontrolled (i.e. destructive), direct or symptomatic, and its
“containment” (identification of the victim and the agent, if applicable-- see
Cochrane [1975c] for a more systematic discussion of this technique and his
scoring scheme). He was specifically interested in comparing the classical
psychoanalytic (as described above), the Kleinian theory (which posits that the
aggression in depression should be uncontrolled and symptomatic), and the
“Inhibition of aggression” theory (as described above by Kendell [1970]-- see
Cochrane [19750] for further elaboration of these theories). Subjects were
200 psychiatric inpatients (124 female) of a variety of diagnoses, assessed for
depression on a battery of tests including the EDI and observer ratings.
Subjects were classified as being neurotically depressed, endogenously
depressed, or non-depressed. Subjects were withdrawn from medication 24
hours before testing.

Although the results are complex, as he examined each depression
measure and scoring category separately, many significant findings were
reported, demonstrating that there was a significant difference in many of the
aggression categories between depressed and non-depressed patients. In
terms of the classical theory, the results did not differentiate between the
endogenous and non—depressed subjects. However, they did differentiate
neurotic depressives from the non-depressed subjects. The Kleinian theory
successfully differentiated endogenous from non-depressed subjects, but failed
to differentiate neurotic depressives. The inhibition of aggression theory also
differentiated endogenous from non-depressed subjects, while only being
partially successful at differentiating neurotic from non-depressed subjects.
He. concludes that the Kleinian and inhibition of aggression theories are the
best predictors overall, while the classical theory was best for differentiating
nemotic depressives, although it was not very good. This finding is
somewhat surprising due to the fact that most of the classical theory is based
on material from rather severe cases, which one might assume to be
“endogenous” He concludes that:

31

Neither the classical psychoanalytic, the Kleinian, nor the
inhibition of aggression theories concerning the
relationship between aggression and depressive illness
seem satisfactory for either neurotic or non-specific
secondary depression. The classical theory is also seen to
be unsatisfactory for endogenous depression. Here,
however, both the Kleinian and inhibition theories appear
to be ‘good’ theories, with the latter perhaps being
favored. Yet some doubt is thrown on the claim of all the
theories that aggression is a causal factor in depression,
there being evidence to suggest that a drive frustration
theory of depressive illness could perhaps be tenable
(quotations in original- p. 126).

Certainly this study is not without problems. First of all, there is no
explanation of how the characterization of endogenous or neurotic depression
was made. Second, since the subjects had been on psychotropic medication
(the 24 hour period in which the drug was stopped would not be enough to
stop its effects) some of the results may be contaminated by the effects of the
medication. Also, the predictions he made based on the three theories appear
to bias in favor of the inhibition theory, which is in part similar to the classical
psychoanalytic theory. In other words, many of the predictions concerning
the inhibition of aggression hypothesis could also have been made for the
classical psychoanalytic theory. Unfortunately, because the results did not
examine each variable on its own, and simply compared the overall expected
differences between groups, it is impossible to examine more specific
hypotheses, such as the relationship between self-directed hostility and
depression. One of the strengths of this study is the use of a projective
measure of aggression instead of the typin self-report inventory.

Another interesting corpus of research attempts to assess the
relationship between hostility and depression using Silverman’s (1976)
subliminal psychodynanric activation technique. This method involves the
subliminal tachistoscopic presentation of words or phrases thought to be
either conflict-arousing or resolving, for example “mommy and I are one.”

32

Rutstein & Goldberger (1973) used this methodology to test Freud’s
(1917) theory that suicide (and depression in general) is the result of the
turning of hostile impulses inward (see Rutstein & Goldberger, 1973, for a
discussion of other studies examining suicide and hostility). In this study they
tested the hypothesis that suicidal subjects would show increases in inward
hostility and depression following subliminal presentation of aggressive stimuli
compared to themselves following presentation of a neutral stimuli, and
compared to control (non-suicidal) subjects. They also believed that these
effects would not be found following supraliminal presentation of these
stimuli, or following libidinal stimuli. Subjects were 64 female inpatients, 32
of whom had made serious suicidal attempts before or during hospitalization.
The other 32 had never made a suicidal attempt and were not diagnosed as
depressed.

The stimuli used contained both pictures and words. The aggressive
stimuli contained the words “Destroy Mother” along with the picture of a
young woman with a knife about to stab an older woman. The stimuli
intended to gratify the subject’s libidinal wishes was “Mommy Loves Me”
along with a picture of a little girl embracing a young woman. The control
(neutral) stimuli contained the words “People Are Walking” with a picture of
two men walking (4 other control stimuli were also used).

The Multiple Affect Adjective Check List and the Rorschach, which
was scored for aggression inward and outward, as well as libidinal content
(following Holt, 1968, cited in Rutstein & Goldberger, 1973) were the
measures used. To summarize, the methodology involves each subject
serving as her own control and being seen for four experimental sessions
(“subliminal aggressive,” “subliminal libidinal,” “subliminal control,” and
“supraliminal aggressive”). These sessions were counterbalanced. Each
session involved a baseline and a “critical” (after presentation of the given
stimuli) assessment of the subject’s functioning, using both the Rorschach
(different cards for baseline and critical) and the adjective checklist (see
Rutstein & Goldberger [1973] for a more thorough discussion of the

methodology and procedure).

33

They found that, as expected, the suicidal group showed a significant
increase in depression as measured by the adjective checklist following the
subliminal aggressive stimulus, but not following the control stimuli.
However, no change was found on the amount of inward-aggression on the
Rorschach, and when compared with the controls, no significant differences
were found. The control group demonstrated no significant changes
following the subliminal aggressive stimuli. When comparisons were made
between conditions for the suicidal subjects, it was found that there was more
outward aggression following the supralirninal aggressive stimuli compared
with the other conditions. Similarly, the controls scored higher on outward
aggression following supralirninal aggressive stimuli when compared with
subliminal aggressive stimuli. When compared with each other, the suicidal
group demonstrated more hostility (as measured by the checklist) than the
controls following the supralirrrinal aggressive stimuli.

In trying to explain the lack of a predicted difference between the
suicidal subjects and the controls in terms of inward hostility following the
subliminal aggressive stimuli, they reiterate that the control subjects were also
psychiatric patients, and therefore may have been affected by this stimuli.
However, the lack of any change in the measure within- group following this
stimuli brings into question this explanation. One other problem with this
study is that part of the directions involve the statement to subjects that: “I
am interested in slight changes in mood that people experience over brief
periods of time. Therefore, please indicate how much of these moods and
feelings you experience now” (p. 163). The fact that they ask for change
may have influenced subjects to respond differentially. Finally, we do not
know the effects of the pictures used, and it is conceivable that they are
conflict arousing in ways that are not expected (for example, two men
walking together could generate sexual orientation conflicts), thereby
confounding the results.

Slipp & Nissenfeld (1981) used this methodology with 48 depressed
females, as measured by the ED] (no cut-off point was given). This stimulus
used to arouse aggressive wishes was the same as that used in the previous

34

study. Another stimuli was “Mommy and I are one,” along with a picture to
two women merged like Siamese twins. This was believed to arouse fantasies
of symbiotic merging, and was believed to resolve conflicts around loss or
dependency. Finally, the authors wanted to explore the role of parental
demands for achievement in depression. Subjects also therefore received the
subliminal stimuli “Succeed for Mother” (or father, depending on which was
the dominant parent pressuring for success), accompanied by a picture of an
older woman (or man) standing behind and looking over the shoulder of a
young woman at a desk. A second success-related stimuli was used
(“Succeed for myself”) and accompanied by a picture of a woman standing
alone on a platform and speaking before an audience; this was assumed to be
adaptation-enhancing, and therefore would lead to a lessening of depression.

Changes in depressive symptomatology before and after the
presentation of the stimuli were assessed via an adjective checklist and the
TAT, scored using the scales developed by Welch, Schafer, & Dember (1961)
for depression and well-being. Also, post-test measures of self-esteem and
self -object differentiation were obtained from an adjective rating scale (see
Slipp & Nissenfeld [1981], for a more thorough discussion of the
methodology).

The results demonstrated that only the symbiotic stimuli (“Mommy
and I are one”) brought about significant changes. It produced an increase in
feelings of well-being on the TAT, and a trend toward heightened self-esteem
on the self -concept scale. When the measures of mood were combined into a
single measure of depression, it was found that the symbiotic stimuli
significantly decreased the overall depression score. The authors were unable
to explain their failure to replicate carlier findings from dissertations (see Slipp
& Nissenfeld, 1981) on the increase in depression following aggressive stimuli,
except that in previous studies the subjects were more disturbed. Therefore,
perhaps the level of intensity of the aggressive conflicts were not strong
enough to be affected by the subliminal stimuli. Also, as mentioned in regard
to the previous study, the so-called neutral stimuli could conceivably be far

from neutral. For example, the picture of a woman by herself talking to an

35

audience could possibly arouse dependency or loneliness conflicts, objective
self-awareness fears, etc.

These results are in contrast to those of Oliver & Burkham (1982),
who set out to extend the results of the dissertation by N issenfeld (1979-- the
results of which were later published in Slipp & Nissenfeld [1981], as just
discussed). In the Oliver et al. (1982) study, 30 heterogeneously depressed
female inpatients made up the subject pool. All of the patients scored at least
16 on the BDI. The stimuli used were: a baseline (neutral) stimuli (“Girls are
looking”) and three critical stimuli-- one control (“People talking”), and two
symbiotic (“Mommy and I are one,” and “Mommy loves me as I am”).
They found no significant effects on a variety of projective tests, adjective
checklists, and self -esteem measures for any of the independent variables.
The investigators were unable to explain this failure to replicate the earlier
findings of Nissenfeld (1979), although the small sample size created a
relatively low power test of the hypotheses.

The last study discussed here which used this methodology was done
by Newman and Hirt (1983). Their study attempts to extend the results of
three doctoral dissertations which found that subjects’ mood or self-esteem
was decreased following the presentation of a sublirninally aggressive stimuli.
This previous research also found no change in the conscious expression of
aggression. In the Newman and Hirt (1983) study subject’s level of field
articulation on coping with aggressive wishes was also examined. They
propose that field—dependent individuals would be more likely to use
introj ection, as their ego-boundaries are weaker. Therefore, depression
(thought to be in part related to introj ection and dependency) would be more
prominent in field-dependent persons. Also, field-dependent persons were
proposed to show greater physiological response to aggressive wishes, due to
their propensity for affective discharge.

Subjects were 60 undergraduates (30 male and 30 female), 32 of
whom were field—dependent (according to the rod-and-frame test), and 28 of
whom were field-independent. The measures used both before and after
tachistoscopic presentation included the Differential Emotion Scale for

36

Depression (a self-report inventory), a self-perception question to measure
self-esteem, and physiological measures (skin conductance and heart rate).
The procedure used was similar to the previously discussed studies except
that they used a verbal passage as a primer before the tachistoscopic
presentation. One was “neutral” and the second contained a passage about
two children’s reactions to their mother’s death. Subjects were asked to
recall both passages to insure attentiveness. Half of the subjects received a
“neutral” subliminal stimulus (the words “hike” or “swim”) while the other
half received aggressive stimuli (the words “attack” or “kill”).

They found that there was a main effect for field articulation, but not
for type of stimuli. Field dependent subjects revealed higher depression
scores regardless of the stimuli presented. Although this finding appears to be
in contrast with the psychoanalytic theory, subsequent analyses showed that
this effect was attributable to differences in mood as shown on the self-report
inventory such that field-dependent subjects demonstrated more distress and
inner-directed hostility. No effects were found on the self -esteem measure. In
terms of the physiological measures it was found that the response for field-
dependent subjects was consistently greater on skin-conductance
(demonstrating higher anxiety levels). A main effect for type of stimulus was
also found, the response being greater for the aggressive stimuli.

One final methodology used to examine depression is the content
analysis of dreams. Barrett and Loeffler (1992) scored the dreams of 20
depressed and 21 non—depressed female college students for anger,
apprehension, sadness, and confusion. They found that the depressed subjects
demonstrated significantly less anger, while results for sadness, apprehension
and confusion failed to reach statistical significance.

In trying to reach any conclusions based on this review of the research
examining the link between hostility and depression, one is immediately
struck by the lack of consensus among theorists about what is in fact being
studied, how to study it, etc. In an excellent review, Buss (1961) addresses
many of the problems that have plagued the study of hostility. He begins
with the basic and yet crucial problem of definition, and describes how the

37

terms aggression, hostility, and anger are often used interchangeably, even
though they are not the same thing.10 He describes aggression as: “...a
response that delivers noxious stimuli to another organism...” (italics in
original, p. 1); anger as a response with facial-skeletal and autonomous
components (other theorists, for example Tomkins [1963], consider anger to
be a primary affect); and hostility as: “...an attitudinal response that endures:
an implicit verbal response involving negative feelings (ill will ) and
negative evaluations of people and events” (italics in original, p. 12).
Obviously, the definition one chooses to use can greatly affect one’s results,
and yet most of the studies reviewed do not address this crucial point. This is
important especially considering the psychoanalytic theory, which is primarily
concerned with aggression turned against the self, and does not necessarily
suppose increased hostility or anger towards others (See Cochrane [1975b],
for a more thorough discussion of these issues).

Another of the major problems with many of the studies on hostility is
that they do not consider inhibitions to acting out or displaying one’s hostility
or aggressiveness. Again, based on psychoanalytic theory one might expect a
depressed individual to have excessive hostility or aggression but also to be
very inhibited from expressing this (or at least turn it against themselves
instead of directing it outward). Therefore, any instrument that does not take
this into account is likely to be useless. Finally, another important aspect of
hostility that is proposed in psychoanalytic theory, and has been ignored in
most of the studies mentioned, is the direction of the hostility (inward or
outward).

A review of the literature on hostility scales for TAT responses
demonstrates just how variable this concept is. Hostility or aggression has
been scored by: simple scales based on assigning “weights” for the hostile
content (Stone, 1956; Murstein & Wheeler, 1959; Hafner & Kaplan, 1960;
Beit-Hallahmi, 1971), scoring hostility as either overt or covert (Gluck, 1955),
scoring the type, intensity, and directness of the expression of hostility (James
& Mosher, 1967; Fiester & Siipola, 1972), and scoring both hostility and

 

V See Pedder (1992) for a review of psychoanalytic viewpoints on aggression.

38

punishment/ guilt or other inhibitions to the expression of hostility (Mussen &
Naylor, 1954; Lesser, 1958; Saltz & Epstein, 1963; Henry, 1981). It is clear
that by using these different scales one would expect to get different results
(see Buss, 1961, for a review of some of the earlier hostility scales).

Three scales which appear more adequate for testing psychoanalytic
hypotheses were created by Purcell (1956), Jensen (1957), and Murstein
(1968). Purcell (1956) scored TAT responses for aggressive content (both
quantity and quality), anticipated punishment from both external sources (i.e.
hero rejected) and internal sources (i.e. self-depreciation), as well as the
remoteness of the aggression from overt behavior (taking into account place,
level, object of aggression, time, and social context). Jensen (1957) created a
system which assigned weights for aggressive content (with the hero as
aggressor or victim), punishment, and defense against aggression (i.e. denial,
rejection of an act, etc). Finally, Murstein (1968) developed a scoring system
by combining aspects of earlier systems. He came up with a scale which
assigned a weight (1 to 5) for the hostile content (from Hafner et al., 1960),
internal punishment and external punishment (from Purcell, 1956), and
remoteness of hostile expression (similar to Purcell, 1956). Based on these
scores he then came up with ratios of internal punishment to hostility, external
punishment to hostility, internal punishment to external punishment, and
remoteness of hostility to hostility.

Based on the problems discussed it is not surprising that one is unable
to form any solid conclusions about hostility and depression based on this
review of the literature. All that can be said for certain is that under certain
conditions, the proposed relationships have been found, and sometimes quite
rigorously. The present study will attempt to improve on previous
methodology by using measures that are more explicit in what they attempt
to assess, as well as taking into account both the direction and inhibition of
the. impulse.

39

Orality and Depression

The link between orality and depression has received very little
empirical attention. Before examining what research has been done, a brief
theoretical discussion is warrant .“ The reason for this is that the term
orality, oral fixation or oral character has (like hostility) accrued a wide range
of definitions and characteristics over the years. This term was first used by
Freud in his Three Essays on th_e Theog 9f Sexuality (1949, first published in
1905) to describe a stage of libidinal development. During this stage the main
area for libidinal gratification was the mouth and oral cavity, and hence the
name “oral stage.” Since this initial theory, however, considerable changes
and additions have altered the way that this term is used, and what it denotes.
For example, Abraham (1924, as discussed above) split the oral stage into
two substages, the oral incorporative and the oral sadistic. Later theorists,
most notably Rado, began to move away from the libido or biologically
centered description of orality. For Rado the concept of orality included not
only pleasurable mouth sensations, but also other pleasurable feelings
associated with feeding, for example security, warmth, and nourishment.
Hirschfeld, Klennan, Chodoff, Korchin & Barrett (1976) take this idea a step
further by proposing that: “The pleasure experienced at the mother ’3 breast,
including the security, warmth, and nourishment, are the precursors of the
narcissistic gratification that is later experienced as self—satisfaction and self-
esteem.” (p. 377).12

Later theorists, such as Fenichel, added dependency to the attributes of
the oral character, and today the concepts of orality and dependence are often
used interchangeably (one sometimes hears the term “oral dependence”l3
used in this regard). Chodoff (1972) notes that this trend has continued in
psychoanalysis, such that: “...for many psychoanalysts, not only those within
the strict Freudian persuasion, the oral character has come to have lost its

 

" For a more extensive review of the psychoanalytic concept of orality. see Sandler and Dare (1970),
or Mendelson (1974).

'2 Perhaps, then, it is the lack of this narcissistic gratification that is experienced as depression.

" Masling (1986) has found in his extensive research that only the concept of oral dependence, and
not oral aggression, (as defined by Abraham [1927]) was able to predict overt behavior.

40

moorings in psychosexual and constitutional orality and to have become
synonymous with exaggerated affectional and supportive needs and with
traits expressing an excessive dependency” (p. 670).

Robertson (1979b) believes that this is how most analysts prefer to
view this concept now, and there has been some empirical evidence which
corroborates this. If one looks at the traits that have been found to hold
reliably together under the heading “oral,” one finds both actual mouth
activities and interpersonal processes thought to stem from this stage.
Gottheil (1965) and Gottheil and Stone (1968) found through factor analysis,
for instance, that the oral traits dependency, passivity, and demandingness
were correlated. Lazare, Klennan and Armor (1966) found passivity,
dependence, and self -doubt to load on an “oral” factor. For the interested
reader, Fisher and Greenberg (1985) have reviewed extensively the research
in this area, and also come to the conclusion that there is a good deal of
support for the notion of an “oral” cluster of character traits.

A good description of how the term is used most often today is that of
Shave (1974): “The oral dependency needs have been described as being the
most basic human needs, remaining so throughout life. The need to feel
important, loved, wanted, needed and cared for represent the oral needs” (p.
311). Shave continues:

This oral need of man is not left behind as a simple stage,
proposed in Freudian theory, but becomes diffused and
less obvious as emotional maturity is reached. Emotional
maturity is reached when the individual has diffused his
oral dependency needs in a great number of part objects
that ensures emotional comfortableness, while the
emotionally immature individual is dependent on too few
part objects (p. 312).

It is this second type of person who, due to their excessive dependency
needs on only one person, is prone to depression when confronted with an
object loss. This idea may help to shed light on other research which has

demonstrated the importance of social support in overcoming depression

41

(Oatley & Bolton, 1985; Harlow, Goldberg, & Comstock, 1991). Perhaps
someone with social support has the necessary part objects as described by
Shave.

I could find only five empirical studies which attempted to use content
analysis or projective measures to examine directly the relationship between
orality and depression.“ There has been a good deal of research relating
orality to a host of other disorders, and a summary of these studies can be
found in Fisher and Greenberg (1985) and Greenberg and Bomstein (1988).

The first empirical study that attempted to examine the relationship
between depression and orality was done by Wiener (1956). He used an oral
content scale developed for the Rorschach to compare depressed patients
with alcoholic patients (both of whom are assumed in psychoanalytic theory
to be orally fixated). He hypothesized that the neurotic depressives would
demonstrate more ambivalence and hostility than the alcoholics. The oral
content scale consisted of finding all references of food objects, anatomy used
in ingestion, eating or preparing food, use of the mouth, etc. These responses
were further scored as either positive, neutral (the mouth is used for non-
eating but non-hostile purposes), or hostile (i.e. devouring, biting).

Subjects were 27 (15 men and 12 women) patients diagnosed as
passive aggressive character disorder with alcohol addiction and 15 (10
women and 5 men) patients diagnosed as neurotic depressives. The study did
not explain how these diagnoses were obtained.

It was found that the depressive patients displayed more hostile and
neutral oral responses, and less positive responses than the alcoholic patients
(both groups produced a similar amount of oral responses overall). Within the
depressive group, a similar amount of positive and negative oral responses
were produced. This is what was expected based on the psychoanalytic
theory of ambivalence about dependency needs. Unfortunately, due to the
lack of a control group it is impossible to compare these results with non-
disturbed individuals. Also, no reliability or validity information was given

 

'3 There has been some rescarch which has used self -report measures to assess this relationship.
Examples of studies in this area include Lazare, Klerman & Armor (1966), Paykel, Klerman & Prusoff
(1976), and Alnaes & Torgersen (1991).

42

about the oral content scale.

An undergraduate honors thesis that attempted to examine this
relationship was undertaken in this same year (J osephthal, 1956). Josephthal
used the Blacky Test, a projective test designed to examine many aspects of
psychoanalytic theory, and tested 20 hospitalized female depressives
(determined by the diagnosis in their medical records), 20 hospitalized female
non-depressives, and 17 non-hospitalized “normal” women. The Blacky
protocols were scored according to Blum (1950), and the scoring was found
to be quite reliable.

A Chi-square was used to examine the differences between the three
groups in terms of the scoring categories of the Blacky Test. Based on an
“eclectic” psychoanalytic theory, Josephthal predicted that the depressives
would give more disturbed responses in the oral eroticism and oral sadism
categories than the other two groups. More disturbance on the guilt feelings
dimension for the depressives was also predicted.

Josephthal found that the number of differences obtained was not
different from what would be expected by chance, so the results must be
interpreted with caution. Regarding the specific hypotheses based on
psychoanalytic theory, there was a decided lack of significant results, such that
none of the hypotheses were confirmed. Furthermore, the pattern of results
that were obtained appeared random and “meaningless.”

In attempting to explain the lack of any significant results, J osephthal
set out to examine the Blacky Test more thoroughly. He found, by
comparing different groups of subjects, that the test did not show any
differences in the amount of psychological disturbance between “normal”
and “abnormal” (hospitalized) subjects. In order to explain this result he
examined the hypothesis that the female subjects were not identifying with
Blacky (the dog and “hero” in the cards) while telling stories. This
hypothesis was based on the fact that the majority of subjects referred to
Blacky as male. Analysis of specific responses demonstrated that in fact the
subjects were not freely identifying themselves with Blacky, which led
J osephthal to conclude that the test fails as a projective device for women.

43

Neuringer (1968) has discussed other research which also questions the test’s
validity for women. Due to this problem, Josephthal concluded that the
original research question he set out to explore was not adequately addressed.

The third study was undertaken by Bomstein, Poynton and Masling
(1985). Subjects were 417 male undergraduates, split into two samples (N:
276 and 141) based on time of testing. Two measures were used, the
Depressive Experiences Questionnaire (DEQ), a 66 item self ~report measure,
and the Rorschach. Previous research using the DEQ has shown a three
factor structure in this measure; factor one is labeled dependency, factor two
is labeled self—criticism, and factor three is labeled efficacy (Blatt, D’Afflitti &
Quinlan, 1979). The scoring system used for the oral content of the
Rorschach was that of Masling (1986), originally developed in Masling, Rabie
and Blondheim (1967). This measure scores both oral percepts (foods, food
organs, food providers, activities of the mouth) and dependent percepts
(supplicants, nurturers, gifts, gift givers and good luck).

They found that factor 2 (self-criticism) correlated significantly with the
oral score for both samples, while factor one correlated significantly with the
oral score only in the second sample. The overall DEQ score correlated
significantly with orality in both samples. A separate analysis of each item of
the DEQ showed that 16 of the items correlated with orality in the first
sample, with 12 correlating in the second sample. When the samples were
combined 19 items showed positive correlations. The authors conclude:
“Although the correlations between individual items and orality, and between
total DEQ score and orality are small and account for relatively little of the
variance, the relationship is statistically quite dependable and stable, as shown
by the consistent findings across two samples. We have thus demonstrated
that orality and depression have common elements, as psychoanalytic theory
had predicted...” (p. 247). They state that one reason for the relatively small
correlations could be due to their use of a normal (not clinically depressed)
population.

A study by O’Neill and Bomstein (1990) used the same oral scale as in
the previous study to examine dependency and gender in 101 psychiatric

44

inpatients (39 men, 62 women). In addition to the Rorschach, all patients
completed the MMPI. A median-split was undertaken on the oral scores, and
the results demonstrated that orality was significantly related to depression for
females, but not for males. However, further examination of the sample
suggested that a greater proportion of the men versus women were alcoholic,
possible leading to distorted depression scores.

The final study was undertaken by O’Neill and Bomstein (1991), and
also used the Masling scale. They examined the relationship between orality
and depression, as assessed by the EDI and the MMPI, in 40 psychiatric
inpatients (20 rrrale, 20 female) with a variety of diagnoses. In contrast to the
previous study, they found that the predicted relationship between orality and
depression was significant for males, but not for females. Unfortunately, they
do not attempt to explain the discrepancy between this study and the
previously mentioned paper (which they also authored).

The literature review also revealed relatively few oral content scales,
which is not surprising considering the relatively small empirical literature in
this area. In a study designed to test the relationships between infant feeding
behavior and later adult personality traits, Thurston and Mussen (1951)
developed a TAT scoring scheme for oral personality traits as defined by
several theorists. These included needing to be ministered to, dependency,
talking, craving to receive, pessimism, aggressive demands, as well as oral
references (food, drink). No reliability information was given on this measure.

The previously mentioned study of Masling, et al. (1967) used a TAT
measure of orality along with the Rorschach measure. This scoring scheme
included oral dependent themes (passive dependence, asking for or receiving
help, presence of parental figures or nurturers, food sources, food organs,
food providers and objects, belief in good luck, magic, optimistic story
endings, helplessness, loneliness or depression, mouth behavior) and oral
sadistic themes (depriving others or being deprived, devouring figures and
aggression, overwhelming figures, burdens, oral assault). This measure was
reported to have adequate reliability. Furthermore, this measure has been
used extensively (see Masling, 1986, for a review of studies using this

45

measure).ls

Story (1968) developed a scoring scheme for free association material
which rated the intensity of oral drive-implicating content. The categories, in
order of descending weight, are as follows: mouth or throat activities; oral
objects, mouth parts; attributes of oral object parts or foods, hunger, thirst,
objects associated with receiving or giving; oral sadistic themes, protective
figures giving or receiving, places where food is prepared or consumed, food
containers, utensils, tobacco, abdomen and breasts; objects or activities
referring in their modal sense to passivity, incorporation, dependency or oral
sadism. The rational for the various weights was not given; however, the
reliability reported was quite high.

Fmally, orality has been measured in other studies using different
methods, but none of these appeared to be as adequate in terms of content
analysis for the present study as those already discussed (see Masling &
Schwartz [1979] for a summary of studies using both projective and
objective methods to assess orality).

In deciding on which scale to use for the present study,16 it was
believed that the scale of Masling was the best as it is the most theoretically
germane and has been used extensively (Masling, 1986). However, this scale
has never been used with psychotherapy transcripts. In fact, no studies which
examine orality in psychotherapy transcripts were found. Although the
present study used five minute verbal samples from psychotherapy sessions, it
was thought that this scale would be amenable to data gathered in this
fashion, and Masling (personal communication, 1991) also believed that the
scale was applicable in these circumstances.

 

1‘ A study by Whitson (19$, cited in Masling, 1986) found that scores derived from this measure
had essentially no correlation with oral scores derived from the Iarare-Klennan-Armor Trait Scales. a self
report measure. This highlights a problem often found in research, in which paper-and-pencil tests do not
seem to measure the same thing as projective tests. Unfortunately, this discrepancy has not been addressed
in terms of the oral scales.

“ Considering the problems addressed above with regards to the Blacky Pictures Test, it was not
considered for the present study even though it does have oral scales in its scoring scheme.

46

METHOD

The present study attempted to examine two of the variables which
have been linked to depression by psychoanalytic theorists. As noted in the
above review, the presence of inwardly-directed hostility is proposed by all of
the “classical” theorists, as well as many ego-psychologists (e. g., Jacobson,
1971 ). The other characteristic noted by the early theorists is the presence of
excessive oral or dependency demands. Although the theories of Abraham,
Freud, and Rado are certainly not identical, these two variables were believed
to be important by each of them. Therefore the hypothesis to be tested was
that persons exhibiting depressive symptomatology would demonstrate
greater inwardly-directed hostility and greater excessive dependency and
orality than persons who are not experiencing depression. Although both
hostility and orality have been studied in relation to depression, no
examination of these variables in combination has been undertaken.

Subjects

The initial subject pool consisted of 564 psychotherapy clients that had
been seen at the Michigan State University Psychological Clinic. The data
used was collected for research purposes as part of an ongoing data collection
effort. At the initial intake interview, all clients were offered the opportunity
to participate in the research project in exchange for a discount on their
psychotherapy fees. If they agreed, they signed a Research Consent Form
(Appendix B). From this initial pool of subjects, 30 were chosen for the
present study, as described below. Subject ages ranged from 19 to 60, with
the average age being 31. Males averaged 29, with a range of 22 to 46, while
females averaged 33, with a range of 19 to 60. Both SES and race were not
reported, but the clinic is a low cost community agency which serves
primarily low-income clients.

Measures
After agreeing to participate, all subjects filled out the Symptom

Checklist (SCL-90-R, Derogatis [1983} see Appendix C). This is a 90-item

47

symptom checklist developed for use with both clinical and “normal”
populations, in which subjects rate their degree of experience with a number
of problems (i.e “feelings of worthlessness”) on a scale from 0 (not at all) to 4
(extreme). Derogatis (1983) describes it as a measure of “...current, point-in
time, psychological symptom status...” (p. 4). The ninety items load on nine
symptom dimensions (including one which measures depression) and a global
severity index (GSI). This measure has been shown to have excellent
reliability and validity in several studies (see Derogatis, 1983). Although it
does not have formal validity scales, it does contain global measures which
can aid in detecting individuals who either “fake good” or “fake bad” while
responding.

The other measure used was the Hopkins Psychiatric Rating Scale
(HPRS- see Appendix D), a scale made up of the nine symptom dimensions
of the SCL-90-R. This was filled out by an advanced trainee during the
intake interview, who rated the client on each symptom dimension on a
seven-point scale (from 0 [none], to 6 [extreme]). The use of this along with
the client’s self-reported symptomatology avoids the problem of relying
entirely on self-report data to assess depression (Lewinsohn & Teri, 1982;
Kagan, 1988).

Participants’ scores on the SCL-90-R and the HPRS were used to
assess level of depression. Derogatis (1983) states that the SCL-90-R has
been used and researched as a screening measure for assessing psychiatric
disorders. This research has lead to the development of a formula for
assessing the “caseness” (the value or score on the measure that serves to
define a positive case) of a particular participant, such that “...if the
respondent has a GSI score (on Norm B, the non-patient norm) greater than
or equal to T-score 63, or any two primary dimension scores are greater than
or equal to T-score 63, then the individual shall be considered a positive
diagnosis or case” (p. 28). This formula has provid “...acceptable levels of
sensitivity and specificity across several populations. The definition presented
(above) has been developed via detailed comparisons of large samples of
psychiatric patients and non-patient cohorts...” (p. 28).

48

The Gottschalk-Gleser Hostility Scales (Gottschalk, Gleser & Springer,
1963; Gottschalk & Gleser, 1969-- Appendix E) and the orality scale of
Masling (Masling, 1986—- Appendix F) were used to assess levels of hostility
and orality (both of these measures will be discussed more fully below). The
data consisted of transcripts of three five-minute segments from the first and
third psychotherapy sessions of each subject. The first and third sessions
were chosen because I wanted to assess the subject’s psychological condition
at the beginning of therapy, before any possible treatment effects might have
occurred. Three different segments were used in an attempt to assess more
stable (trait) characteristics of the subjects. Although the Gottschalk-Gleser
scales were originally developed to tap transient feelings, Gottschalk (1986a)
states that “...affect scores derived from three or more five-minute verbal
samples... approximate trait measures in the sense of providing a measure of
the relatively unvaryin g central tendency of a psychological characteristic” (p.
44). Gottschalk et al. (1969) have offered some evidence for the validity of
this assumption. For the final scores, the average score of these three
segments was used.

The first five—minute sample was taken from the beginning of the first
session, at the point in which the client began describing the circumstances
which brought him/her into psychotherapy (for the most part this constituted
the very beginning of the session, but if the therapist began by explaining
ground rules, etc., the sample began when the client began to “tell their
story”). The second sample was taken from a period towards the end of the
first session. Specifically, it was taken when the tape counter read “600.”
The third segment was taken from the third session, at a point when the tape
counter read “400.” Although this procedure does not guarantee a similar
amOunt of data (words) for each subject (due to speed of speech, amount that
the therapist speaks, etc), the scales include a formula which takes into
account the number of words in the transcript when calculating the
magnitude of each affect.

Once these transcripts were obtained the verbal material was scored
using the Gottschalk-Gleser Scales (Gottschalk etal., 1969; Gottschalk,

49

Winget, & Gleser, 1969). After extensive review of the literature these scales

- were chosen as they are the most pertinent to the variables under question, as
well as being the only scales which are theoretically coherent with certain
aspects of psychoanalytic theory. In addition, they have been extensively
researched and demonstrate satisfactory reliability and validity, as well as
having published norms (Gottschalk et al., 1969; Gottschalk, 1979;
Gottschalk, 1986a). Furthermore, directions for using these scales have been
published in a thorough and understandable manual, which allows for their
use with (psychoanalytically) non-sophisticated raters (Gottschalk, nget, &
Gleser, 1969). Finally, a computer scoring system has been developed for
these scales and has been found to be as effective as hand scoring (Gottschalk,
Hausmann, & Brown, 1979; Gottschalk & Bechtel, 1982; Deffner, 1986;
Gottschalk & Bechtel, 1989).

The Gottschalk-Gleser Scales are verbal content analysis scales, and as
such share certain underlying assumptions and theoretical aspects with other
content analysis methods. A full discussion of the verbal content analysis
method is beyond the scope of the present work, but some of the underlying
assumptions are addressed in Lolas (1986) and Vmey (1986). The basic
reason for using verbal analysis is that language is believed to reflect certain
psychic structures and states. Lolas (1986) states that: “As the locus of
intersubjectivity, language behavior not only permits inferences about
individuals, it also shows the societal dimensions of the mind by relating inner
experience to overt behavior” (p. 18). Lolas summarizes the main
assumption underlying language behavior as such: “Our main contention is
more an axiom than a theorem: there is a link between communicative
structures and psychic structures” (p. 19).

The most concise description of the reasoning behind the use of verbal
content analysis comes from Gottschalk and Lolas (1986), in which they state
that:

the words people choose to express themselves reveal
significant information about how they are feeling and

50

thinking. That words may mask genuine feelings and
thoughts is not incompatible with the validity of content
analysis as a measurement method. Even when words are
used to cover up mental experiences or to deceive,
content analysis may help elucidate the genuine subjective
state of the individual. And whereas the reliability of
assessing mental processes through the analysis of verbal
communication reaches a convincing level of significance,
the assessment of mental processes based on non-verbal
communication is fraught with low reliability (p. 83).

The Gottschalk-Gleser scales were originally developed for use with
samples of speech gained through free-association, which is then tape
recorded and scored on three hostility scales. These scales are based on
“psychoanalytic, linguistic, and learning theory principles,” and are the most
sophisticated and comprehensive verbal content scales available. Due to their
psychoanalytic perspective, one of the advantages of these scales over those
previously discussed is that they attempt to take into account repressed affect
and defensive and adaptive mechanisms in the speech content. Gottschalk et
al. (1969) state that “We assume that the verbal content of spontaneous
speech, like dream content, contains the workings of primary and secondary
process thinking...” (p. 16). However, these scales are not tied to any
theoretical position, as Gershon, et al. (1968) state in their study which utilizes
these scales:

In this discussion, we have avoided the metapsychological
problems of the relationship of hostility as a feeling or
impulse to aggression as a primary instinctual drive. The
initial formulations concerning the role of hostility in
depression anteceded the development of Freud’s theory
of the dual instincts of libido and aggression. The way in
which hostility has been defined and measured in this
study is similar to the usage found in the hypotheses of
Freud and Abraham, which were phrased in terms of
hostility as a feeling-state (p. 234).

Another advantage of these scales is that they take into account the

51

direction of hostility, an important psychoanalytic variable (especially in the
study of depression). The three scales are labeled the Hostility Directed
Outward Scale (HDO), the Hostility Directed Inward Scale (HDI), and the
Ambivalent Hostility Scale (AH). The HDO Scale measures: “...the
assaultive and angry impulses and feelings that a person is aware of and can
describe in various self-report test procedures.” (Gottschalk et al., 1969, p.
158). This scale is further divided into covert and overt hostility. The AH
Scale includes aspects of both outward and inwardly-directed hostility;
however it has been shown in past research to be different enough from the
other two scales to wanant discrete classification. Basically it measures
statements by the speaker concerning hostility directed towards him/her from
the outside world. ‘7 The I-IDI Scale measures “...transient and immediate
thoughts, actions, and feelings that are self -critical, self-destructive, or self-
punishing” (Gottschalk, \Vrnget & Gleser, 1969, p. 93). A concise description
of these scales can be found in Gottschalk, et al., 1963.18

Although the scales were developed for use with five minute verbal
samples, a good deal of research has used these scales with psychotherapy
transcripts, and they have been found to be valid when used in this manner
(see Rubin [1986] for a discussion of studies using these scales with
psychotherapy transcripts). Furthermore, Gottschalk (1986b) recommended
the use of his scales with psychotherapy transcripts, and they have been
recommended for use in this way in an NIMH report (see Kiesler, 1975).

Once the transcripts were scored for hostility, they were scored by
myself and two upper-level undergraduate psychology students for oral
content using the Masling Scale (1986). The students underwent extensive
training in the use of this scale, and were blind as to the purposes of the study
and the status of the subjects (I was blind to the depression scores of the
participants).

The orality scale was first outlined in a paper by Masling et al. (1967).

 

'7 It should be pointed out that this definition of ambivalence is not in line with ambivalence as
described by Freud, Abraham or Rado.

“ Gottschalk & Hoigaard-Martin (1986) have recently used the same methodology and created a
depression scale for use with verbal samples.

52

The definition of oral dependence and oral sadism was derived from Schafer
(1954). Masling, following Abraham, divided oral responses into oral
dependent and oral sadistic categories. The oral dependent responses
included food sources, food providers, passive food receivers, food organs,
supplicants, nurturers, gifts, gift givers, and good luck symbols. The oral
sadistic responses included devouring, fighting or killing, overwhelming
figures, figures which deprive, deprivation, faulty oral capacity, oral assault,
and burdens. In this first study he used both the Rorschach and the TAT as
stimuli for the responses. However, only the Rorschach was used in the later
studies.

After the first series of studies were conducted using this scale, Masling
found that “...the oral aggressive phase was difficult to differentiate from the
second of Abraham’s anal stages-- the anal-sadistic. More to the point, it did
not predict anything. Only the oral-dependent concept seemed able to predict
overt behavior” (Masling, 1986, p. 74). The scale has therefore since been
altered and now only scores oral dependent responses, such that “...the
operational definition of the concept includes references both to activities and
the organs of the mouth as well as to dependent and nurturant behavior” (p.
74).

The final version of the scale contains 15 content categories, including
statements which deny oral material and statements which use baby-talk (The
scoring guide can be found in Appendix F). In Masling’s system, each oral
response is scored “1,” with a maximum score of 25. However, two changes
were necessary in using this scale for the present purposes. First, since the
amount of verbal material is likely to be highly variable (due to the speed with
which subject’s talk, etc.) it was proposed that a ratio of oral responses to
total word count would be the best method of scoring. For example, a
person who used 5 oral responses in a 100 word passage would get the same
oral score as a person who used 50 oral responses in a 1000 word passage.
Second, the upper limit of 25 was ignored

Interscorer reliability has been assessed many times and “...has always
been in the range of 89% to 95% agreement” (Masling, 1986). Furthermore,

53

the scale is relatively easy to master and can be used by inexperienced raters.
The reliability was continually checked in the present study to insure similar
reliability coefficients.

Finally, it should be noted that although the scoring scheme is simple
and straightforward and can be scored objectively, it does ignore some data
which might be quite relevant. As Masling (1986) stated: “The system is not
designed to elicit the greatest amount of information from single subjects but
to produce a meaningful score from a large number of subjects quickly and
reliably.” However, for the present purposes it appears to be the most
relevant, as well as the most researched, scale available.

Design and Data Analysis

Based on the methodology described above, a pool of 30 subjects with
a wide range of scores on the SCL-90-R was obtained. Because a relatively
small sample size was being used, it was believed that a random sample of 30
subjects would likely not vary a great deal in terms of depressive
symptomatology. Due to the lack of variance a true test of the hypotheses
would be difficult. Therefore, for each gender, an attempt was made to get 5
subjects with a depression T-score of 70 or above (depressed), 5 with a T-
score of 60-70, and 5 with a score between 50 and 60 (normal).‘9 Scores
were also checked on the global measures for evidence of “faking good” or
“faking bad.” Finally, in an effort to control for differences in the overall
amount of pathology or severity, an attempt was made to insure that all
subjects had a GSI T-score in the range of 50-70.

The HPRS was also be used so that measurement of depression was
not based solely on the clients’ self-report. An HPRS depression score of 4
or greater (along with the criterion stated above) was required for the
“Depressed” subjects (those with a T-score above 70), a score of 3 was
required for the “average” subjects, and an HPRS depression score of 2 or
below was required for the “Non-depressed” subjects (those with a T-score

 

‘° The middle group was included here in order to avoid inflated correlations due to the use of
extreme groups, a problem which can be avoided by including a moderate group (Shavelson, 1988).

54

under 60). Correlational analysis was used, as statistics such as ANOVA
would afford less power in this design (Humphreys, 1978).

Hypotheses

Before addressing the specific hypotheses of this study, an important
point with regards to the theory needs to be addressed. The classical
psychoanalytic theory of depression, as described herein, is a multi-faceted
and complex theoretical perspective on the etiology and maintenance of
depression. It is therefore difficult (if not impossible) to fully test this theory.
In order to adequately test this specific theory one would need to use subjects
who had experienced some type of object loss, or repeated disappointments at
the hand of a love object, during childhood. Furthermore, the study would
have to address variables such as identification, ambivalence, introj ection, and
a host of other processes all of which pose serious problems for the
researcher. Therefore, the present study was an attempt to examine some
important correlates of this theory, without actually being a direct test of the
theory.

Keeping this caveat in mind, the hypotheses were as follows:

1) There will be a positive correlation between inwardly-directed
hostility and depression scores.

2) There will be a positive correlation between orality and depression
scores.

3) Multiple regression analysis will establish that both inwardly-
directed hostility and orality contribute to depression scores.

Although not actually hypotheses, three other points need to be
addressed:

1) Although many of the theorists mentioned herein posit the presence

55

of “ambivalent hostility” in depression, the relationship between ambivalent

hostility and depression was not assessed in the present study since the
definition of ambivalence used in the Gottschalk-Gleser scales is different than

that used by Abraham, Freud, and Rado. Furthermore, the relationship
between depression and the other hostility measures was examined, although
no directional hypotheses were made.

2) Although somewhat different results might be expected for males
versus females, it is difficult to be specific in this regard due to past
inconsistencies in the research. For example, Gottschalk et al. (1969) report
more outward hostility for men than women, but no difference on inward
hostility or ambivalent hostility. Also, they found that the relationship
between inward hostility and depression was much stronger for women than
for. men. However, the study of Rubin et al., (1992) did not corroborate these
results. They found that men and women did not differ in their expression of
outward hostility, and that women were higher than men on inward hostility.
Finally, the paper of Masling et al. (1979) describes differing results for males
and females in many studies assessing orality. Therefore there were no
specific hypotheses regarding possible gender differences; nevertheless,
gender differences were examined using non-directional tests.

3) Aside from using traditional Fisherian tests of significance,
confidence intervals were given for each correlation and beta-weight. The
method used for this was derived from Hunter and Schmidt (1990), who
believe that confidence intervals give more information about the precision of
the sample values as point estimates than the dichotomous “yes-or-no”
results of simple tests of significance (see also Cronbach, 1975). The other
reason for utilizing confidence intervals was the low power of this study, due
to the sample size of 30. According to Cohen (1992), assuming a medium
effect size (a correlation of about .30), a study using an alpha of .05 would
need to have 85 subjects to have adequate power. The importance of power
is that without it one runs the risk of failing to reject the null hypothesis when

56

it is in fact false. Although this is not a new concept, Sedlmeier and
Gigerenzer (1989) state that most researchers still ignore the power of their
tests, which, paradoxically, leads to a lower chance of obtaining significant
results if there is a true effect.

The confidence intervals used in this study were a combination of the
best case and the worst case one-sided 95 percent intervals, making a 90
percent interval. The following figure will demonstrate this technique,
assuming a correlation (for example) of .24 between variable A and variable
B.

FIGURE 1
Formation of confidence intervals

 

Best Case interval (one-sided 95 percent)-- -.06 Z 1.00
Worst Case interval (one-sided 95 percent)-- -1.0 2 .54
-.06 2 .54

Combined interval (two-sided 90 percent)--

 

This figure reveals that, in this example, we can be 90 percent sure that the
true correlation between variables A and B lies between -.06 and .54. The
reason for using this type of interval was that it allows one to set a constant
error rate, instead of dealing with the often very large beta (type 11) error
which plagues much research (Sedlmeier et al., 1989). The tradeoff for this is
that, due to the large sampling variability, the intervals were quite large.
When no directional hypotheses were made, the 95 percent interval was used.
Finally, aside from the correlation and confidence interval, the inference
probability and odds ratio were reported to give further information about the
strength of the relationships. The inference probability is defined as the
probability that the correlation is greater than zero (or less than zero, if the

 

57

correlation is negative). The odds ratio is calculated by dividing the inference

probability by 1 minus the inference probability. It gives the odds that the
correlation is not zero.

58

RESULTS

Reliability

The alpha reliability of the SCL-90-R depression scale was assessed and
found to be quite adequate (alpha = .90). The hostility scales were all scored
using the computer scoring scheme of Gottschalk and his associates. This
scoring system has been found to have adequate coding reliability, and is
immune to the problems of rater inconsistencies, biases, etc. However, it is
difficult to interpret stability because affects, by definition, are expected to be
rapidly changing. Therefore, one might expect low test-retest reliabilities,
even though Gottschalk et al. (1969) believe that using three or more speech
samples will allow one to approximate a trait measure of hostility. Schofer et
al. (1979) assessed reliability using a split-half procedure, and found
correlations for all of the scales to be about .50. Table 1 gives the correlations
between different speech samples for inward hostility and orality.20

TAB LE 1
Intercorrelations between verbal samples for inward hostility and oral scores

 

 

Inwgd hostility HINl HIN 2 I-IIN3
HIN 1 1.000
I-IIN2 0061 1.000
HIN3 .171 -.008 1.000
Orality ORALl ORAL2 ORAL3
ORALI 1.000
ORAL2 .393 1.000
ORAL3 .352 .114 1.000

 

This table demonstrated that there was a great deal of variability in both

scores, suggesting that they are in fact measuring a more “state”
characteristic. In order to assure reliable measurement of orality, three raters

 

’5 Intercorrelations for the hostility-out scales can be found in Table 2 of Appendix G.

 

59

were used, all of whom had extensive training in the use of this scale. When
90% inter-rater reliability was achieved, all transcripts were scored by all
three raters. For the final score, only those words agreed upon by all three
raters were used, to ensure even greater reliability (inter-rater reliability was
therefore not calculated). Therefore, the high variability demonstrated was
somewhat disappointing, in that the aforementioned problems with the
hostility scales might be expected to be absent with the oral scale, which is
measuring orality, a supposedly more stable personality characteristic. 'In this
sample, then, the oral scale appears to be measuring a more state based
characteristic which is quite variable. In order to approximate the more stable
underlying personality characteristic, the average of the three scores was used
(Spielberger, 1966).

 

Descriptive statistics:

Table 3 contains the descriptive statistics for all of the major variables,
including DEPRESS (depression score), GSI (Global Severity Index),
HSCORE (Hopkins rating), HOTOT (total hostility-out), HOOVE21 (overt
hOStility-out), HOCOV (covert hostility-out), HIN (hostility-in), and ORAL

 

(oral score).
TABLE 3
Descriptive statistics

DEPRESS GSI HSCORE HOTOT HOOVE
MEAN22 19.400 0.981 2.900 1.416 0.559
STANDARD DEV 11.581 0.447 1.626 0.193 0.122
MEDIAN 19.000 1.100 3.000 1.403 0.533

Hooov HIN ORAL
MEAN 1.297 0.548 0.017
STANDARD DEV 0.202 0.150 0.009
MEDIAN 1.282 0562 0.016

 

 

2' As a quick check of the convergent validity of this scale, hostility scores were correlated with the
hostility scale from the SCL-90-R. This scale correlated .42 with HOOVE, and only .02 with HOCOV, as
would be expected.

’2 Gottschalk et al. (1969) provide tables to compare means with their original normative sample.
Using this table, the means of the present study put these subjects in the 60-70th percentile for hostility-in,
the 80-90th percentile for total hostility-out, the 50th percentile for overt hostility-out, and the 90th
perwntile for covert hostility-out.

As a function of the sampling technique, in which an attempt was made
to get subjects with a wide range of depressive symptomatology, depression
scores were bi-modally distributed, and therefore a median split was
undertaken in order to dichotomize depression scores. Point-bi serial
correlations were then used to analyze the relationship between depressive
symptomatology (as measured by the SCL—90-R), inward and outward
hostility (as measured by the Gottschalk-Gleser Scales), and orality (as
measured by the Masling scale). Due to the artificial median split that was
undertaken, each correlation was divided by .8 in order to correct by 20%
(Hunter et al., 1990).

An attempt was also made to correct all correlations for attenuation
due to measurement error, according to the formula of Hunter et al., (1990--
see Figure 2).

FIGURE 2
Formula for correcting for attenuation

 

 

rxy corrected = rxy observed / ’ rxx\’ l'yy \

rm = reliability of variable x
ryy = reliability of variable y

 

Unfortunately, the variability of most of the scales was such that correcting
for attenuation was not applicable. Therefore, the only corrections which
were undertaken were for correlations with DEPRESS; all other correlations
were uncorrected.

Table 4 gives the corrected correlations below the diagonal and the
original correlations above.

61

TABLE 4
Correlation matrix for all variables

 

DEPRESS GSI HOTOT HOOVE HOCOV HIN ORAL
DEPRESS 1.000 0.888 0.073 0.196 0.031 0.183 -0.074
GSI 1.000 1.000 0.047 0.067 0.030 0.059 —0.029
HOTOT 0.096 0.047 1.000 0.161 0.969 0.473 0.054
HOOVE 0.258 0.067 0.161 1.000 -0.035 0.234 0. 197
HOCOV 0.041 0.030 0.969 -0.035 1.000 0.420 0.035
HIN 0.237 0.059 0.473 0.234 0.420 1.000 0.056
ORAL -0.097 —0.029 0.054 0. 197 0.035 0.056 1.000

DEPRESS=depression score, GSI=Global Severity Index, HOTOT=total hostility-out,
HOOVE=overt hostility-out, HOCOV=covert hostility-out, HIN=hostility-in, ORAL:

oral score.

 

 

As can be seen from this table, there was a very large correlation between

depression scores and GSI. Although an attempt was made to use subjects
who had variable depression scores but similar GSI scores, this was not
possible. As might be expected, those subjects with significant levels of
depression also tended to have high levels of other symptomatology, as
reflected by this correlation.

Hypotheses
In order for the most information to be given regarding the

hypotheses, 90 percent confidence intervals were reported for each of the
hypothesized relationships, as well as the inference probability and the odds

ratio, as described above. Significant correlations are those in which zero is
not in the interval.

Hypgthesis #1-- There will be a positive correlation between inwardly directed
hostility and depression scores.

 

I W Inference Probame Odds Ratio
.2423 -.06 g p g .54 .91 10.1/1

 

’5 This is the corrected correlation, as are all following correlations and beta weights which involve
depression.

62

Although this correlation failed to reach traditional significance, the inference
probability demonstrated that in all likelihood the true correlation was not

zero, and more than likely it was positive.

Hypothesis #2" There will be a positive correlation between oraligv and

depression scores.

1 Confidence Interval Inference Probability Odds Ratio
-.10 -.39 g p _<_ .20 .71 2.4/1

This hypothesis was not corroborated, as shown by the extremely low
correlation and odds ratio.

Hypothesis #3
The third hypothesis predicted that using multiple regression analysis,

depression scores would be significantly related to orality and inwardly-
directed hostility. These results are found in Table 5, including 90 percent
confidence intervals and inference probabilities for the beta weights and the
multiple correlation.

TABLE 5
Regression equation for depression scores on inward hostility and orality

 

_B_etp Confidence Interval Standard Error Inf. Probability
HIN .243 -.046 5 B S .532 .176 .92
ORAL -.111 -.413 _<_ B _<_.191 .184 .73
R: .26 -.02 sR_<_.54 .17 .94

 

This hypothesis received marginal support, although the multiple correlation
just failed to reach traditional significance levels. The inference probability

suggests that the true multiple correlation was not zero.

63

Although no direct hypothesis was made, multiple regression was
performed using each of the independent variables in order to assess their
influence on depression scores. Because HOTOT is made up of HOCOV and

HOOVE, HOTOT was not included in this equation to avoid problems of
multicollincarity. These results are found in Table 6, using 95 percent

 

 

confidence intervals.
TABLE 6
Regression equation for depression scores on orality and overt, covert, and
inward hostility
Beta Confidence I___nterval _S__tandard _E_rr__or Inf. Probabilig

HOOVE .238 - 06.7 _<_ B _<_ .543 .186 .90
HOCOV -.033 - .374 < B < 3.08 .208 .56
HIN .202 -. 133 _<_ B _<_ .537 .204 .84
ORAL -.154 -.462 5 B 5 . 154 .188 .79
R: .35 .07 5R5.63 .17 .98

 

This table demonstrated that this multiple correlation was significant,
suggesting that this model fits the data well. In this model overt and inward
hostility accounted for most of the variance.

The relationship between depression scores and the other hostility
scales was undertaken using 95 percent two-sided intervals, as no directional
hypotheses were made. These results are shown in Table 7, found in
Appendix G.

 

Insert Table 7 about here

 

Only the correlation between depression and overt hostility approached

64

significance, as demonstrated by the inference probability of .92.

Further exarrrination of Table 4 revealed three other significant
correlations. The first was between total outward hostility and covert hostility
(r = .97). This is to be expected, as total outward hostility is made up of the
sum of covert and overt hostility. Since overt hostility did not correlate
significantly with total outward hostility, in this sample HOTOT was more
related to HOCOV than to HOOVE. The second significant positive
correlation was between total outward hostility and inward hostility (r = .47),
and the third was between covert hostility and inward hostility (r = .42). This
demonstrated that in this sample outward hostility (especially covert) was
positively related to inward hostility.

 

Gender Differences

In order to assess the importance of gender, correlations were
undertaken between each of the variables and gender. The results were all
displayed using two-tailed 95th percentile confidence intervals, as shown in

 

Table 8.
TABLE 8
Correlations between gender and all variables
[ Confidence Interval I_n_f_._ Probability Odds Ratio
Gender and ORAL .49 .22 5 p 5 .77 1.00 NC
Gender and HIN .43 .14 5 p 5.73 1.00 NC
Gender and HOTOT .19 -. 16 5 p 5 .54 .85 6/1
Gender and HOOVE .05 -.32 5 p 5 .42 .60 1.5/1
Gender and HOCOV .16 -.19 5 p 5 .52 .82 4.5/1
Gender and DEPRESS .26 -.07 5 p 5 .60 .94 15.7/1
Gender and GSI .26 -.07 5 p 5.60 .94 15.7/ 1

DEPRESS=depression score, GSI=Global Severity Index, HOTOT=total hostility-out,
HOOVE=overt hostility-out, HOCOV=covert hostility-out, H[N=hostility-in, ORAL=oral score

 

65

As can been seen from these intervals, significant relationships were found
between gender and orality and gender and inwardly-directed hostility. Also,
the relationships between gender and total outward hostility, covert hostility,
depression scores, and global severity approached significance, as
demonstrated by the inference probabilities. In all of these cases women
scored higher that men.

. Because previous research has suggested that the relationship between
hostility and depression might be different for men versus women (see
above), each hypothesis was examined separately for each gender.

Males
The descriptive statistics for males are found in Table 9.

TABLE 9
Descriptive statistics for males

 

DEPRESS GSI HOTOT HOOVE HOCOV HIN ORAL

IVflEAN 16.467 0.870 1.380 0.553 1.264 0.484 0.013
STANDARD DEV 10.710 0.452 0.127 0.131 0.126 0.128 0.005
MEDIAN 1 1 0.960 1.367 0.533 1.280 0.460 0.013

DEPRESS=depression score, GSI=Global Severity Index, HOTOT=total hostility-out,
HOOVE=overt hostility-out, HOCOV=covert hostility-out, HIN=hostility-in, ORAL-_-
oral score

 

Table 10 gives the correlation matrix for each of the variables for the
males, using the same corrections as described above (corrected correlations
are below the diagonal).

66

 

TABLE 10
Correlatron matrrx for all vanables for males

DEPRESS GSI HOTOT HOOVE HOCOV HIN ORAL
DEPRESS 1.000 0.848 0.558 0.226 0.548 0.203 -0.080
GSI 1.000 1.000 0392 0.070 0.421 0.155 0.034
HOTOT 0.734 0392 1.000 0.176 0.924 0.229 -O308
HOOVE 0.300 0.070 0.176 1.000 0.171 0.031 0.030
Hocov 0.720 0.421 0.924 0.171 1.000 0.211 -O36O
HIN 0.227 0.155 0.229 0.031 0.211 1.000 0.363
ORAL 0.110 0.034 0.308 0.030 0.360 -0363 1.000

DEPRESS=depression score, GSI=Global Severity Index, HOTOT=total hostility-out,
HOOVE=overt hostility-out, HOCOV=covert hostility-out, HIN=hostility-in, ORAI;
oral score

 

_ Each hypothesis was exarrrined for males, using 90 percent confidence
intervals.

HvDothesis #1-- There will be a positive correlation between inwardly-diiected

hostjlity a_nd depression scores.

1 Confidence Interval
.27 -.07 5 p 5 .61

Inference Probability Odds Ratio
.90 9/ 1

Although this correlation failed to reach traditional significance levels, the
confidence interval demonstrates that the true correlation was probably not

zero, and more than likely it was positive.

Hypothesis #2-- There will be a positive correlation between oralig and

 

 

 

gepression scores.
_r_ Confidence Intervpl Inference Probability Odds Ratio
-.11 -.52 5 p 5 .30 .67 2/1

This hypothesis was not corroborated, as shown by the extremely low
correlation and odds ratio.

Hypothesis #3
The third hypothesis predicted that using multiple regression analysis,

67

depression scores would be related to orality and inward hostility. As might
be expected due to the previous findings, this hypothesis was also not
corroborated, as demonstrated by Table 11. However, the inference
probability of the multiple correlation was at the 90th percentile, suggesting
that the true multiple correlation was not zero.

 

 

TABLE 11
Regression equation for depression scores on orality and inward hostility in
males
Beta Confidence Interval Standard Error [i Probability
HIN .311 -.1195 B 5 .741 .262 .88
ORAL .003 -.464 5 B 5 .470 .285 .05
R = .31 -.08 5R 5 .70 .24 .90

 

Although no directional hypothesis was made, multiple regression was
performed using each of the independent variables in order to assess their
influence on depression scores. These results, using 95 percent confidence

intervals, are found in Table 12.

 

TABLE 12
Regression equation for depression scores on orality and covert, overt, and
inward hostility in males
_Be_ta Confidence Interval Standard Error Inference Probabilig
HOOVE .430 .00 5 B 5 .87 .268 .95
HOCOV .841 .493 5 B 5 1.0 .212 1.0
HIN .213 -.228 5 B 5 .654 .269 .79
ORAL .257 -.258 5 B 5 .772 .314 .79

R=.88 .63 5R51.0 .15 1.0

 

68

This table demonstrated that the regression model fit the data well, with both
overt and covert hostility accounting for most of the variance.
The relationship between depression scores and the other hostility

scales in males was undertaken using 95 percent intervals, as shown in Table
13 of Appendix G.

 

Insert Table 13 about here

 

 

As demonstrated by these intervals, for men there was a significant
relationship between depression scores and outward hostility, especially
covert. The relationship between the different scales was also examined,
using 95 percentile confidence intervals. These results can be found in Table
14 of Appendix G.

 

Insert Table 14 about here

 

These intervals demonstrated that for men there was a significant negative
relationship between covert hostility and orality, and the other correlations,
although not statistically significant, are more than likely not zero.

Females
The same analyses were undertaken for females as for males. The

descriptive statistics are found in Table 15.

69

 

TABLE 15
Descriptive statistics for females
DEPRESS GSI HOTOT HOOVE HOCOV HIN ORAL
MEAN 22.33 1.093 1.451 0.565 1.329 0.612 0.021
S'I‘ANDARDDEV 12.028 0.397 0.242 0.117 0.257 0.147 0.010
MEDIAN 25.00 1.300 1.497 0.560 1377 0.643 0.020

DEPRESS=depression score, GSI=Global Severity Index, HOTOT=total hostility-out.
HOOVEzovert hostility-out, HOCOV=covert hostility-out, HIN=hostility-in, ORAL-_-
oral score

 

 

In order to examine each of the main hypotheses, point-biserial
correlations were used, and the matrix is found in Table 16, with corrected

correlations below the diagonal.

 

TABLE 16
Correlatron matrrx for all vanables for females

DEPRESS GSI HOTOT HOOVE HOCOV HIN ORAL
DEPRESS 1.000 0.928 0.234 0.150 0.271 0.026 0.278
GSI 1.000 1.000 0.239 0.039 0.258 0.029 0.296
HOTOT 0310 0.239 1.000 0.168 0.980 0.558 0.021
HOOVE 0.200 0.039 0.168 1.000 0.031 0.457 0314
HOCOV -0360 0.258 0.980 0.031 1.000 0.490 0.019
HIN 0.034 0.029 0.558 0.457 0.490 1.000 0.143
ORAL 0370 0.296 0.021 0.314 0.019 0.143 1.000

DEPRESS=depression score, GSI=Global Severity Index, HOTOT=total hostility-out,
HOOVEzovert hostility-out, HOCOV=covert hostility-out, HIN=hostility-in, ORAL:
oral score

 

Hypgthesis #1-- There will be a p_ositive correlation between inwardly-directed
hostilig and dyression scores.

1; Confidence Interval

.03 -.395 p 5 .46

Inference Probame
.55

Odds Ratio
1211

As can be seen from the confidence interval, this correlation was essentially

70

zero, and did not corroborate the hypothesis.

Hypgthesis -- There will be a positive correlation between oralig and
d_epression scores.

; Confidence Interval Inference Probame Odds Ratio
-.37 -.68 5 p 5 -.06 .97 32.3/1

This hypothesis was not corroborated, and was in fact refuted by the
significant negative correlation.

Hymthesis #3

The third hypothesis predicted that using multiple regression analysis,
depression scores would be significantly related to orality and inward hostility.
The results of this analysis are found in Table 17.

 

TABLE 17
Regression equation for depression scores on orality and inward hostility in
' females
m Confidence Interval Standard Error fl Probability
HIN -.019 -.462 5 B 5 .424 .270 .53
ORAL -.373 -.756 5 B 5 .011 .234 .94
R = .37 -.01 5 R 5 .75 .23 .94

 

Although the multiple correlation failed to reach traditional significance levels,
it was quite large and more than likely not zero.

Although no directional hypothesis was made, multiple regression was
performed using each of the independent variables in order to assess their
influence on depression scores. These results are found in Table 18.

71

 

TABLE 18
Regression equation for depression scores on orality and covert, overt, and
inward hostility in females
Beta Confidence Interval Standard Error I_nf_. Probability
HOOVE .386 -. 198 5 B 5 .97 .356 .86
HOCOV -.340 -.8355 B 5 .155 .302 .87
HIN -.046 -.7045 B 5 .612 .401 .55
ORAL -.491 -.96 5 B 5 .-.02 .286 .96
R::.62 .28 5R5.96 .21 1.0

 

 

This table demonstrated that this regression model fit the data well, with
orality accounting for most of the variance.

The relationship between depression scores and the other hostility
scales for women was undertaken using 95 percent intervals, as no directional
hypotheses were made. These results are found in Table 19 of Appendix G.

 

Insert Table 19 about here

 

These intervals demonstrated that for women there was a significant negative
relationship between depression scores and covert hostility.

The relationship between the hostility and oral scales was also analyzed,
and the results are found in Table 20 of Appendix G.

 

Insert Table 20 about here

 

This table demonstrated significant relationships between outward hostility

72

(both overt and covert) and inward hostility, as well as a substantial (although
not significant) relationship between orality and overt hostility.

In order to further examine these gender effects, correlations for
women versus men were examined to see if they differed significantly. Table
21 gives the difference between the correlations for males versus females on

each variable, as well as the 95 percent confidence interval.

TABLE 21
Differences between correlations for males versus females

 

 

difference Confidence Intefll Inf. Probpbilitv om; Ratio

 

DEPRESS and HIN .24 -.41 5p 5.88 .76 3.2/1
DEPRESS and ORAL .26 -.37 5p 5.89 .79 3.8/1
DEPRESS and HOTOT 1.00 .53 5p 51.00 1.00 NC
DEPRESS and HOOVE .10 -.51 5p 5 .71 .63 2.7/1
DEPRESS and HOCOV 1.00 .55 5p 51.00 1.00 NC
HIN and HOTOT .33 -. 14 5p 5 .80 .92 12.5/1
HIN and HOCOV .38 -.11 5p 5.87 .94 15.7/1

DEPRESS=depression score, GSIzGlobal Severity Index, HOTOT=total hostility-out,
HOOVE=overt hostility-out, HOCOV=covert hostility-out, HIN=hostility-in, ORAL=oral score

 

This table demonstrated that the relationship between outward hostility
and depression scores, especially covert, was much stronger for males versus
females. Furthermore, the relationship between inward hostility and outward
hostility (especially covert) is stronger for females than males, although this
difference just failed to reach traditional significance levels.

73

DISCUSSION

This study attempted to examine the relationship between depressive
symptomatology and orality and hostility. It was hypothesized, based on
selected themes found in psychoanalytic writings on depression, that subjects
with higher levels of depression would exhibit greater hostility directed
against the self and greater orality. The following discussion will first examine
issues of scale variability, and problems with the methodology will also be
addressed. Next, the results will be exarrrined, and put in the context of other
research in this area. Finally, the relevance of these results for theory and
future research will be evaluated.

In order to gain a full understanding of the results of this study, a
number of issues regarding the scales and the methodology need to be
addressed. Perhaps the biggest problem encountered in this study was with
the high variability of the scores over different times. Each scale will be
addressed separately below.

Hostility Scales

Perhaps the major assumption in using the Gottschalk-Gleser scales in
the present context was that the use of three five minute samples would give
an adequate estimate of the “traits” in question. This question has been
addressed before, and Gottschalk (1986a) states that three samples is adequate
for all of their scales except perhaps hostility-out (especially for males), as the
generalizability is lower. However, it appears from these results that this
assumption might be questioned, as all of the scales demonstrated high
variability over time (low generalizability was also found by Gershon et al.,
1968; SchOfer, et al., 1979). Because the final score used in this study was an
average of the scores for the three segments, it was obviously affected by
high variability. These results highlight an important issue regarding
reliability, and one that is often overlooked in current research. Most
researchers are well-aware of inter-rater reliability, and are usually diligent in

 

74

reporting this in their studies. However, test-retest reliabilities are often
ignored. For example, the previously mentioned study of Lemaire et al.
(1981), assessed hostility repeatedly over a six week period, but did not
address test-retest reliability. Because these scales were not necessarily
designed to have high test-retest reliabilities, this issue is often ignored, but it
seems to be an important area that will need to be examined in the future so
that a better understanding of what these scales are actually measuring can be
determined. Based on this study, the use of these scales as “trait” measures
appears to be questionable.

This problem also highlights the state-trait issue regarding affects such
as hostility. 2“ Researchers in the area of affect usually accept their transient
nature, but often describe these transient states as being affected by some
underlying “trait” for that certain affect. For example, Spielberger (1966)
hypothesized that these transient states would be influenced by “differences
between individuals in the probability” that the states would be manifested
under varying circumstances. While this may certainly be true, it seems from
the present study that perhaps more samples are needed to in fact begin to
tap into this “probability.”

Before dismissing the use of three segments to assess more enduring
personality characteristics, the specific use of these scales in the present study
needs to be addressed. Although these scales have been recommended for
use with psychotherapy transcripts (Gottschalk, 1986b; Kiesler, 1975), this is
not the way they were originally designed to be used. The actual
methodology entails a subject being told to “talk about any interesting or
personal life experiences you have had.” One fact which becomes
immediately obvious is that this is a much more standardized situation in
contrast to a psychotherapy session. Furthermore, in the original
methodology experimenters are instructed to say very little to the subject, and
simply let them talk. This too is quite different than many psychotherapy
sessions, in which the therapist (especially depending on their theoretical bent)

 

’1 In fact, the whole issue of “traits” has become somewhat controversial lately, although this will not
be addressed here (discussions can be found in Mischel, 1969; Ebel, 1974; Cronbach, 1975).

75

might speak and lead the client a great deal. In fact, examining the
relationship between what the therapist says and the client's responses (and
hostility scores) would seem to be an interesting road for future research.
One final point about these scales involves the issue of fully repressed
hostility. As Gottschalk et al. (1969) mention, these scales are not necessarily
able to measure fully repressed hostility, although they state that 'some of the
defensive and adaptive mechanisms in language signal the presence of
suppressed and repressed hostility. According to some theorists, however, it
is this type of hostility which might lead to depression. As Klennan et al.
(1970) point out, it remains an open question whether or not these scales are
able to tap preconscious or unconscious hostility or aggression. They do
appear to be, however, the best hostility scales available today, and they
continue to be widely used. The question of repressed hostility in these scales,
then, would seem to be an important question for future research. This
problem applies to all hostility measures which I found, including those that
have been developed for projective tests. It seems that the creation of a
measure which might allow for a better test of the “repressed hostility”
hypothesis is needed. And finally, as addressed earlier in regard to the
ambivalent hostility scale, there is some question as to how similar the
conceptualization of hostility is between these scales and Freud’s and other’s
writings. This seems to be a problem in a good deal of research, in which the
measures used are from a different theoretical background than what is being
tested. In effect, this makes a true test of hypotheses very difficult (Meehl,

1978; Dar, 1987).

Orality Scale

As with the hostility scales, the Oral scores proved to be quite variable.
This was especially noteworthy because orality, unlike hostility, might be
assumed to be more stable trait than hostility (an affect), and therefore not
vary all that much.” To my knowledge, the test-retest reliability of this scale

 

’3 Some variability based on the current state of the subject (for example, if they are hrmgry) might be
expected to alter oral responses, however.

 

76

has never been assessed, because it has never been used repeatedly on the
same subject. This would seem to be an important avenue for future work.

. It should be pointed out, however, that the Oral scale (like the hostility
scale) was used differently in this study than the manner for which it was
developed. The standard procedure for using this scale is that subjects
respond to a set of Rorschach cards, a situation that is much less variable than
psychotherapy sessions. In this respect the same problems mentioned above
in using the hostility scales for psychotherapy sessions hold true for the Oral
scale. As this was the first study to use this scale with psychotherapy
transcripts, more work is needed in this manner before any decision about
their usefulness in this area can be established.

Another aspect of the Oral scale which seemed problematic is that it
includes scoring of both the vicissitudes of the drives (an id- or libido-centered
description), and object relations (a more ego-oriented approach). As
mentioned earlier, most analysts today use the term “oral” to mean both
psychosexual fixation and needs such as dependency, self-esteem, etc. (see
Chodoff, 1972; Shave, 1974; Robertson, 1979b). Although factor analytic
studies have corroborated this notion (Gottheil, 1965; Gottheil and Stone,
1968; Lazare, Klennan and Armor, 1966), this is not entirely in line with
orality as described by Freud and Abraham. It also seems possible that this
description is too broad to be meaningful in a study such as this. In this
regard, Maslin g (1986) separated the total oral score into dependent responses
(passive, nurturant, and supplicant images, as well as responses mentioning
pregnancy and reproductive organs), and oral responses (food, food
providers, food related content, oral activity, and digestive organs) for one
study. The correlation between these two subscores was -.06, demonstrating
that this scale is not homogeneous, and in fact seems to be measuring too
different things.

Initially this separation seemed possible to do in this study. However, a
quick subjective examination of the protocols revealed that certain subjects,
who had an excessive amount of oral responses, may have been alcohol
dependent because much of the sessions were spent discussing drinking.

77

Weiner (1956) demonstrated that alcoholics exhibit different levels and types
of orality compared to depressives. Furthermore, the use of alcohol may
serve to lower depression scores, thereby distorting the relationship between
orality and depression (O’Neill & Bomstein, 1990). It is therefore possible
that the negative correlation found between orality and depression scores for
women in this study might have been in part due to the presence of alcohol
dependent subjects, who might have lower depression scores combined with
high oral scores. Future work examining the link between depression and
orality will therefore have to account for the effect of alcohol or other
addictive disorders.

Methodology
Certain methodological issues must also be taken into account in

interpreting the results from this study. The first important issue relates to the
sample. Much of the theoretical background for this study was from the
psychoanalytic literature, which in many cases is based on what might be
assumed to be a more disturbed population than in the present work.

Mendel son (1974) points out that most of the early theory is predicated on
what might be called “psychotic” rather than “neurotic” depressions. The
sample in this study consisted entirely of outpatients, and might therefore be
less disturbed than the population that much of this theory is based on. In
terms of the issues involved, both Stone (1986) and Parkin (1976) believe that
mild depression can be quite different psychologically than more severe
depression in terms of the unconscious dynamics that accompany it (see also
Bibring, 1953). There has also been some research which has suggested that
the relationship between hostility and depression is different for more severe
depression compared to mild depression (see below).

Another problem with the sample was that they were assessed only for
depression, and therefore were probably quite heterogeneous on other
personality characteristics which might interact with the variables in question
(Cronbach, 1975). There is a growing body of evidence that there might in
fact be many types of depression, each with its own symptom structure and

78

dynamics (Riley, Mabe & Davis, 1991; Blazer, Woodbury, Hughes, George,
Manton, Bachar & Fowler, 1989; Gershon et al., 1968; Rosenthal &
Gudeman, 1967; Rosenthal & Klerman, 1966; Hamilton et al., 1959). Some
authors have begun to try to examine the personality structures of these
different types of depression, as well as the effect of underlying personality
disorders on depression (Alnaes et al., 1991), although many conflicting
results have been reported (Overholser, Kabakoff, & Norman, 1989;
Matussek & Feil, 1983; Paykel, Klerman & Prusoff, 1976; Lazare &
Klennan, 1968). From a theoretical standpoint, this issue has been addressed
by Freedman (1986), Arieti & Bemporad (1980), Basch (1975), and Blatt
(1974). Blatt and colleagues have also been able to demonstrate some
empirical evidence to back up their theory of two types of depression. In one
study, they demonstrated that “depression that focuses on dependency may
be characterized by different types of defenses [than depression based on self-
critical attitudes]-- particularly denial and displacement-- rather than turning
against the self in marked negative attitudes including self—criticism.” (Blatt,
Quinlan, & Chevron, 1990, p. 112). Factor analytic studies by Grinker, Miller,
Sabshin, N unn, & Nunnally (1961) demonstrated several independent factors
in depression, one of which described a depression primarily concerned with
experiences of deprivation and manipulation to obtain oral supplies, and
another that focused on feelings of guilt and restitution. Finally, Westen and
colleagues (Benjamin, Silk, Lohr, & Westen, 1989; Nigg, Westen, Lohr, Gold,
& Silk, 1992) have explored the effect of underlying borderline personality
disorder on depression. For example, Benjamin et al. (1989) found that
depressed borderline patients exhibited more hostility, interpersonal sensitivity,
and paranoia on the SCL-90-R than did pure depressed patients.

Although the effect of personality on depression (and vice versa) is still
not well understood, its influence is unquestionable, leading to a great deal of
heterogeneity under the heading of depression. As far back as 1917 Freud
recognized this dilemma: “Melancholia, whose definition fluctuates even in
descriptive psychiatry, takes on various clinical forms the grouping together
of which into a single unity does not seem to be established with certainty”

79

(p. 243). This notion was addressed by Milrod (1988), who outlined three
major types of depression. The first is based on a narcissistic injury which is
the result of the “...self representation falling too far short of the goals and
standards in the individual’s wished-for self image” (p. 87). Milrod believes
that this depression involves an intrasystemic process. A second type of
depression is caused by the self-representation falling too far short of the
“moral and ethical values built into the ego ideal,” leading to guilt and
superego punishment. He describes this type of depression as involving an
intérsystemic process. Finally, the third type “...occurs when an ambivalent
love object continues to be the essential source of libidinal supplies for the self
representation” (p. 86). Milrod believes that it is this third type that Freud
was addressing in “Mourning and Melancholia,” and he goes on to warn that:
“Many authors have made the mistaken assumption that the same structure
and the same dynamics apply to all depression. Some have even confused
the dynamics of mourning with those of depression” (p. 87).

In this study other personality characteristics of the sample were not
examined for two reasons.26 First, the small sample size would have made
any statistical tests of these subgroups impossible. Second, and perhaps more
importantly, Chodoff ( 1972) has warned against assessing enduring
personality characteristics during the period of illness, as this may affect what
is uncovered.

. Another methodological issue which may have affected the results of
this study involved the use of the first few psychotherapy sessions. The initial
sessions were used in order to minimize any treatment effects, but it is
possible that this strategy might have obfuscated any possible results due to
the “face-saving techniques” (Goffman, 1967) which are assumed to occur in
the early stages of any relationship. In other words, it is possible that during
the early stages of psychotherapy, clients are still concerned with how their
therapists view them, so they might be less likely to reveal their dependent or

 

1' Another area which was not able to be assessed was the presence of any manic symptoms, which
could have affected the results. For example, Blackburn (1974) formd more outward hostility for manics
than depressives. Hirschfeld et al. (1979) found that manic patients exhibited less orality than either
“normals” or depressed patients, while Alnaes et al. (1991) found that cyclothymic patients had a much
greater level of orality than bipolar patients.

80

hostile impulses. This notion received some corroboration in the study of
Weissman et al. (1971), which demonstrated that the depressed patient’s
behavior during initial interviews with a psychiatrist was quite different, and
much less hostile, than her behavior outside the interview (it should be
pointed out, however, that this is referring to behavior, and not verbal hostility
scales). Other research (Leff, Roatch, & Bunney, 1970) has demonstrated
that initial psychotherapy interviews did not always uncover the core conflicts
which were behind the depressive episode. Finally, it is possible that at this
early stage of therapy the client is feeling nurtured and taken care of, so they
are not yet experiencing oral frustration.

Finally, this study did not address what the therapists said, which could
certainly have had a major effect on what the client said. In this regard,
gender of therapist/patient interactions were also not examined, although
Gottschalk et al. (1969) have found them to be important.

 

Results27

Before addressing the specific hypotheses, the correlation between
depression scores and global severity needs to be addressed. In designing this
study, the aim was to get a sample with varying degrees of depression, but
with similar global severity levels. This was to ensure that the results could be
reasonably interpreted as being due to differing levels of depression, and not
simply the subject’s level of “sickness.” As pointed out earlier, however, this
proved to be impossible, as those subjects with higher levels of depression
almost invariably had higher GSI scores. An examining of Table 4 reveals,
however, that there were no significant correlations between GSI and any of
the. independent measures, and that each of these correlations was smaller
than for the same measure and depression scores. Therefore, it seems
reasonable to assume that any findings obtained for depression are not simply
due to the severity of overall pathology.

The first hypothesis assumed a positive relationship between depression
scores and inwardly-directed hostility. Although this correlation did not reach

 

’7 The reader is again referred to Appendix A for quick reference to previous research.

81

traditional levels of significance, it was in the expected direction and at the
9lst probability level. This result is not unlike those found in work by
Gottschalk et al. (1969) and associates. In their original validity studies for
these scales they found the correlation between inward hostility and the BDI
to be between .34 (outpatients) and .47 (inpatients). Gottschalk et al. (1979)
found this correlation to be .37, and Schofer et al. (1979), using a random
non-patient sample in Germany, found a small but significant correlation
between depression, as measured by the mood adjective check list, and
inward hostility (.22). Lenmire et al. (1981) found that mildly depressed
students demonstrated more inwardly-directed hostility than controls, and
other studies, using different measures and methodology, have also found a
significant relationship between depression and inwardly-directed hostility (cf.
Newman et al., 1983; Bulatao 1961, cited in Gershon, et al., 1968), although
some conflicting results have also been reported (Foulds et al., 1965; Slipp et
al., 1973).

° Before any final conclusion can be reached regarding this relationship,
however, one point regarding the hostility-in scale needs to be addressed.
Because the Gottschalk-Gleser scales are based in part on psychoanalytic
theory, some of the scoring category for hostility-in entails statements about
feeling depressed. Therefore, based on these statements alone one might
expect at least a slight positive correlation between depression scores and
inward hostility. This issue was addressed by Gershon, et al. (1968), who
modified this scale by separating the self -critical statements from verbal
expressions of depression. Verbal expressions of depression were scored in a
separate category, which they called “affect-in.” The remainder of the
hostility-in scale was named the “modified hostility-in scale.” Even with this
modification, however, they found a significant positive correlation between
depression scores and inward hostility, suggesting that this relationship cannot
be accounted for simply due to expressions of depression.

The second hypothesis, which assumed a positive relationship between
depression scores and orality, was not corroborated. Very little empirical
study of this relationship has taken place, but studies by Paykel et al. (1976)

82

and Hirschfeld and Klerman (1979), also found little relationship between
orality (assessed via the Lazare-Klerman-Armor scale) and depressive
symptoms. Hirschfeld et al. (1979) did, however, find significantly lower
levels of orality in manic patients, and Alnaes et al. (1991) found that both
depressed and cyclothymic patients had higher oral scores than patients with
other psychiatric disorders. Finally, Matussek et al. (1983) found that
“neurotic depressives” demonstrated little orality but some hostility. Results
from the current study and some of those just discussed, then, appear to
discount the relationship between orality and depression (other studies used
either only males or only females, and will therefore be discussed below).
Before dismissing this hypothesis, however, problems with the orality scale, as
described above, as well as the sample, should be taken into account.

Multiple regression analysis demonstrated a marginally significant
prediction of depression scores based on orality and hostility scores.
Examination of the beta weights revealed, however, that inward hostility was
the primary source of this prediction, as would be expected based on the
results discussed thusfar. In order to examine more fully the relationship
between hostility and depression scores, a further regression analysis was
undertaken using all of the independent variables. This equation proved
significant, with overt hostility and inward hostility accounting for most of the
variance. With this sample, then, hostility directed both towards the self and
others did have a significant relationship with depression scores, albeit a slight
one.

The only other relationship which approached significance for the
whole sample was a positive correlation (at the 92nd probability level)
between depression scores and overt hostility. This is in contrast to the
studies of Gottschalk et al. (1969), which found that outward hostility was
negatively correlated with depression scores. Results more in line with the
present study were reported by Schofer et al. (1979), who obtained a
significant correlation between depression, as measured by the mood adjective
check list, and overt hostility (r = .16). Lemaire et al. (1981) also found that
depressed subjects expressed more outward and total hostility than non-

83

depressed subjects. These results have also been corroborated in research
using other methods, such as that of Paykel (1971), who found a subgroup of
depressed patients which he called “hostile depressives.” In this regard,
Mayo (1967) compared psychotic and neurotically depressed patients, and
found that psychotic patients were higher on criticism and projected hostility,
as well as on guilt. Therefore, it appears that, at least for a portion of
depressed patients, depression is associated with higher levels of overt
hostility.

Because previous research has suggested that the relationship between
depression and variables such as hostility and orality might be different for
males versus females, each analysis was undertaken separately by gender. If it
is true that men and women are different in their handling of hostility and
orality, then results which are based on mixed samples may tend to obscure

 

relationships that are more obvious when studying only one gender.

An examination of gender differences on each of the variables revealed
that women scored significantly higher on the orality and inward hostility
scales than men. Previous research using the orality scale with Rorschach
protocols did not reveal any significant gender differences (Masling, 1986;
however, see O’Neill & Bomstein, 1991), so this finding is somewhat
unexpected. Furthermore, Wiener (1956) found no sex differences on his
oral measure with neurotic depressives. The higher inward hostility for
women is also surprising based on the normative data of Gottschalk et al.
(1969), which showed no significant gender difference on hostility-in. The
previously mentioned work of Rubin et al. (1992), however, did find higher
levels of inward hostility for women (this research was carried out at the same
clinic as the present study).

Women also scored higher on the depression measure, as might be
expected based on previous research (for analysis of higher rates of
depression in women, see Nolen-Hoeksema, 1987 ; Herman, 1983). In fact,
women were higher than men on all variables, although not always
significantly. In the normative data of Gottschalk et al. (1969), the only
significant gender difference was for Hostility-out, with men scoring higher;

84

however, this difference was not significant in their patient sample.

When examining males only, ’8 the relationship between depression
scores and inward hostility was nearly significant, and not unlike the results of
earlier studies. The relationship between depression scores and orality,
however, was insignificant, corroborating the earlier work of O’Neill et al.
(1990). This is in contrast to the studies of Bomstein et al. (1985) and O’Neill
et al., (1991), which found a significant correlation between these variables in
men. These results are confusing, because these studies all used the same oral
scale. Although all of these studies (except Bomstein et al., 1985) used
psychiatric patients, a more detailed analysis of differences in the samples used
might prove interesting, and help shed light on the conflicting results. At this
point, then, the relationship between orality and depression in men remains
unclear.

 

Other significant relationships for men in the present study were found
between depression scores and outward hostility, especially covert. In
Gottschalk's scoring system covert hostility-out refers to statements of people
other than the participant involved in fighting or other hostile acts.
Gottschalk et al. (1969) assume that this may be a defensive reaction, and
therefore represent unconscious or repressed hostility. This finding could be
interpreted as lending corroboration to the inhibition of aggression theory of
Kendell (1970). This result also corroborates the results of Foulds et al.
(1960), who found that for melancholic men there was a high level of
projected hostility, while for women this was much lower. The previously
mentioned work of Paykel (1971) and Mayo (1967) also seems to fit with
these results.

In the present study multiple regression analysis, using only inward
hostility and orality, revealed a marginally significant equation, with inward
hostility accounting for most of the variance. When all of the independent
variables were included, however, a highly significant equation was found,

 

" Because the results by gender were based on samples of 15 subjects, the standard errors were quite
large. These small sample sizes are especially problematic m multiple regression (Tabachnick & Fidell,
1983) and, therefore, the following findings should be considered starting points for future research using

larger samples.

85

which accounted for 77% of the variance in depression scores. In this
eqr'ration, covert hostility was a major contributor, while the effect of overt
hostility was also significant. The beta weights for inward hostility and orality
were not significant. Furthermore, the fact that the beta weight for orality
was positive, even though the correlation between orality and depression
scores was negative, suggests that in this equation orality was acting as a
suppressor variable. In this regard it should be noted that marginally
significant negative correlations were found between orality and inward
hostility and covert hostility (both correlations = -.36). In conclusion, for men
there appears to be a significant relationship between depression scores and
both hostility directed against the self and others, while the relationship
between depression scores and orality appears insignificant.

For women, significant positive relationships were found between the
different hostility measures; if a subject displayed more inward hostility she
tended to display more outward hostility also. This was in contrast to some
previous work which has suggested a negative relationship between inward
and outward hostility (Lemaire et al., 1981). The hypothesized relationship
between depression scores and inward hostility was not supported. This is in
contrast to the work of Gottschalk et al. (1969), which found this relationship
to be quite strong for women-- often stronger than for men. This finding also
contradicts that of Rubin (1986), who found a correlation of .54 between
depression and inward hostility for women. As mentioned earlier, one
possible reason for these conflicting results might be the presence of alcohol
dependent subjects in the present study, although the actual effect of alcohol
dependence on depression scores and hostility in this study is unknown.

Studies using other methodologies and measures have also
demonstrated mixed results regarding the relationship between depression
and inwardly—directed hostility in women. The earlier cited work of Rutstein
et al. (1973) found corroboration for this relationship using Silverman’s
subliminal psychodynamic activation technique, while Slipp et al. (1981) did
' not; they mention that their study used less disturbed people than the Rutstein
et al. (1973), so perhaps this relationship only holds for more depressed

86

patients, a notion that has received some corroboration in dissertations cited
in Slipp et al., (1981).

Other studies have also demonstrated that this relationship only holds
for more severely depressed people. Friedman (1970) found that this
relationship was not significant for the whole sample, which included controls,
but it did hold within the depressed patient sample. In order to examine this
possibility in the current study, correlations were undertaken between
depression scores and the independent measures for just the 15 depressed
subjects. The correlation between depression scores and inward hostility was
insignificant, but the correlation between depression scores and covert
hostility was significantly negative (-.71, before correction), suggesting that
within this depressed population the greater the covert hostility, the lower the
depression. Perhaps, then, if depressed persons are able to express some of
their hostility outward (especially in disguised covert fashion), they can direct
the hostility away from themselves and in turn deter more severe depression.
This dynamic has also been suggested by Lemaire et al. (1981), and Wessman
et al. (1960).

With regard to the second hypothesis, the relationship between
depression scores and orality was significant, although in the opposite
direction to what was predicted. Similar results were reported by Josephthal
(1956), who found higher levels of orality (as measured by the Blacky test)
for normals and non-depressed women patients than for those who were
depressed. The studies by O’Neill and colleagues (O’Neill et al., 1990;
O’Neill et al., 1991) found mixed results for this relationship; the first found a
significant positive relationship, the second found no relationship. In this
sample, then, women who demonstrate more orality admit to fewer
depressive symptoms, although the previously mentioned effects of possible
alcohol dependence should be considered.

Although not hypothesized, a significant negative relationship between
depression scores and covert hostility was found, showing that the more
depressed the women were, the less covertly hostile they were. This is in
contrast to Lazare et al. (1968), who found that in a subgroup of depressed

87

women (those who also had hysterical traits), hostility was quite high during
depression. Perhaps these contradictory results can be explained in part by
referring to the study by Gershon et al. (1968), which found two patterns for
hostility in depressed women such that it was higher with patients who had
hysterical personality features compared to those who did not. The negative
correlation between depression scores and covert hostility was also
contradicted by Wessman et al. (1960), who found, using the Rosenzweig
Picture Frustration Test, that women were more extrapunitive and less
intrapunitive when they were feeling more depressed (it should be noted,
however, that this study used “normal” college students). Furthermore, there
is some difference between covert hostility and extrapunitiveness. Wessman
et al. (1960) suggest that at the point when this outward hostility begins to
turn inward, more severe depression begins. This notion was corroborated in
a study by Zuckerman, Persky, Eckman, & Hopkins (1967), which, in
contrast to Wessman et al. (1960) study, used all male subjects. Zuckerman et
al. (1967) found that hostility scores derived from the Buss-Durkee inventory
correlated positively with depression in normals, but not in depressed
outpatients.

In order to examine the effect of gender on hostility and orality in
depression, the differences between these correlations for men versus women
were examined. The relationship between depression scores and both total
outward hostility, and, more specifically, covert hostility, was much greater for
men than women. Men with high levels of covert hostility tended to have
higher depression scores, while women with high levels of covert hostility
tended to have lower levels of depression. Furthermore, the relationship
between inward hostility and covert hostility was much greater for women
than men, demonstrating that for women, but not for men, the amount of
hostility directed against the self is related to the amount of covert hostility.
These results strengthen the conclusion that hostility and its relation to
depression is quite different for men versus women, and suggest that
analyzing men and women together may serve to obfuscate important
relationships.

88

In summary, then, this study adds to the body of evidence which
demonstrates the important relationship between hostility and depression.
Although many of the correlations did not account for a great deal of
variance, significant effects were found, often corroborating previous
research. The fact that the correlations were small should not be surprising,
however, due to the subject matter and methodology. Masling et al., (1979)
state that: “... it is as naive to expect to find impressive correlations between
experiment and a psychoanalytic description of personality as it would be to
expect to find a real person who is identical with a character in a novel” (p.
265).

The relationship between orality and depression, however, seems much
more tentative. For women, a negative relationship between orality and
depression scores was found, but for reasons stated above, this finding should
not be considered conclusive. For men, no relationship was found. Aside
from the issues already discussed, the effect of the sample should also be
noted. As with the relationship between hostility and depression, it is likely
that this sample was less severely depressed than those on whom the original
theories were based. This was addressed by Milrod (1988):

Depressed patients have lost their sense of worth and are
engaged in efforts to regain it. If easily regained, the
depression is short-lived and will hardly become a clinical
problem. When restorative efforts are unsuccessful or
blocked, regression to early oral modes of gaining self-
esteem are brought into play... Although orality plays a
universal role in the depressive’s efforts to restore self-
directed libidinal supplies, it is more obvious in the more
severe depressions (pp. 94-95).

Because there has been so little research examining this relationship, it
remains a question in need of further empirical analysis. One possible road
for future research would be to divide the Masling Orality scale into two
scales as he did (see Masling, 1986) and examine the relationship of each

subscale to depression. This would allow for the testing of the “dependency”

89

hypothesis put forth by Hirschfeld, et al. (1976) and Birtchnell (1984).

In conclusion, this study demonstrated that there are significant gender
differences found on the orality, hostility, and depression measures, and the
relationships between them. Future research in depression will need to be
aware of this, so that studies can be designed explicitly to examine these
differences and the theoretical reasons for them.

Although there have been some findings in the area of hostility and its
relationship to depression, one can not help but be struck by the large amount
of conflicting results. Certainly part of this is due to the wide range of
samples, measures, and methodologies used to assess this relationship.
Masling et al. (1979) point out how much of the research in psychoanalytic
theory is plagued by this problem:

The field is dominated by single, non-replicated studies.
With few exceptions, authors have published only one
study on a particular topic, thus denying the field the
opportunity of profiting from sustained, continuing
programmatic research efforts. The next investigator of
the topic has then adopted his own measures and his own
operational definitions, and published his one study, which
invariable contradicted the results of other studies (p. 295).

A further problem is that often studies (this one included) attempt to
isolate a few variables, without accounting for the importance of a host of
variables on the depressive experience (see Cronbach, 1975). Psychoanalytic
writings on depression have emphasized the importance of hostility, orality, a
severe superego, a lofty ego-ideal, loss of self-esteem, and narcissistic
character structure. As Bellak (1981) pointed out, many of these ideas are
testable, but perhaps future studies should attempt to assess a wider range of
phenomenon so that the possible confounding effects of unexamined variables
can be minimized.

Theory

Finally, I will address how the results described herein fit with existing
theories of depression. Although, for the many reasons described above, it is
difficult to say that this study was a true test of the early psychoanalytic
theory of depression, the results do coincide with certain theoretical
vieWpoints better than others.

The hypothesized relationship between depression and inwardly-
directed hostility, as described by Freud, Abraham, and other early
psychoanalysts, does seem to hold for a certain group of depressed patients.
Other theorists, however, posit a relationship between outwardly-directed
hostility and depression, an idea which also found some corroboration. The
presence of outwardly-directed hostility is in line with what Rado called
“coercive rage,” in which the patient would attempt to “terrorize the beloved
one.” Another theoretical viewpoint which receives some corroboration from
these results is that of Bowlby (1963), who believes that “... the problem in
understanding pathological mourning, it seems, is that of understanding not
the simple presence of hostility directed against the lost object but its
repression and/or displacement towards other objects including the self” (p.
524). This notion receives some corroboration from the significant
correlation between depression scores and covert hostility. It should be
remembered that covert hostility consists of statements about someone else
killing or hurting something, which Gottschalk et al. (1969) believe implies
some form of defensive adaptation. Shave ( 1974) supposes that in depression
anger may be repressed, but it is not necessarily turned against the self.
Instead, it finds its outlet in unconscious guilt, and it is this guilt, and not
aggression turned against the self, which leads to death wishes, etc.

Other theorists, although acknowledging that orality and hostility
sometimes play a role, see depression more as an ego phenomenon, especially
exemplified by the feeling of helplessness. The idea is most closely linked
with Bibring (1953), although other theorists (Rubinfme, 1968) have made
this connection. Bibring (1953) differentiates between the “ego killing itself”
and the “ego letting itself die”; he beheves that aggression is only important

91

in the first case. In effect, he highlights the important notion that there are
perhaps more than one type of depression, as did Milrod (1988-- see above).
Therefore, in studies which mix together different types of depression, small
correlations might be expected, as more powerful ones are perhaps being
hidden.

As a quick examination of the “helplessness” hypothesis, all references
to helpless or passive states were scored, and then correlated with depression
scores. As would be expected based on Bibring’s theory, this correlation was
significant (corrected r = .47), lending some corroboration to this idea

(interestingly, this relationship was much stronger for men than for women).
This relationship is certainly an area that is ripe for further study.

In the final analysis, what can be said about psychoanalytic theories of
depression? Certainly some of their tenets have proven vigorous under

experimental scrutiny, even with the many flaws in design and methodology
discussed herein. In the future, however, it seems that research must move to

a more biopsychosocial view, one that takes into account the underlying
personality, the social context, the interpersonal interactions, other
psychopathology, and biology. Goldberg (1975) concludes that:

...psychoanalysis has made unique and fundamental
contributions to our understanding of depression, but it
has not explained it; yet without the concepts of
psychoanalysis there will be no explanation of depression.
No one discipline today seems capable of carrying the
burden alone... fractionation of the observing field may be
another problem; we notice the child’s aggression and
depression while we are unaware of the mother’s feelings
and reactions and totally oblivious to the social factors that
may be operating. This is not a call for a holistic approach
but rather a reminder that observation per se is an artifact
in scientific investigations (pp. 137-138).

This was echoed by Basch (1975), who stated that:

A satisfactory causal explanatory theory of depression
must consider the entire gamut of the depressive

92

syndrome, must be independent of any given school of
psychological thought or method of therapy, must go
beyond clinical appearances, and should not contradict
what has been properly established to be the case in other
sciences (p. 530).

Recently, a few theorists have moved in this direction, although the
majority of work in this area continues to be narrow and one-dimensional.
Two examples of more encompassing theories include the work of Basch
(1975) and Freedman (1986), who have brought new and interesting ideas
into this area. Until we are able to devise more precise experiments which
are not too far removed from theory, however, we will remain unable to tease

apart fact from artifact.

BIBLIOGRAPHY

93

BIBLIOGRAPHY

Abraham, K. (1927). Notes on the psychoanalytic investigation and
treatment of manic-depressive insanity and allied conditions (1911). In
Selected papgrs pp psychoanalysis. London: Hogarth Press.

Abraham, K. (1927) The first pre-genital stage of the libido (1916). In
Selected pamrs pp psychoanalysis. London: Hogarth Press.

Abraham, K. (1927). A short study of the development of the libido, viewed
in the light of mental disorders (1924). In Selected papers 9_n_
psychoanalysis. London: Hogarth Press.

Akiskal, H. S., & McKinney, W. T. (1975). Overview of recent research in
depression. Archives gt General Psychiag, §_2_, 285-305.

Alnaes, R., & Torgersen, S. (1991). Personality and personality disorders
among patients with various affective disorders. Journal o_f Personalig
Disorders _5, 107-121.

 

Anthony, E. J ., & Benedek, T. (1975). Depression _a_Ld human existence.
Boston: Little, Brown and Company.

Arieti, S., & Bemporad, J. R. (1978). Severe and mild depression. New
York: Basic Books.

Arieti, S., & Bemporad, J. R. (1980). The psychological organization of
' depression. Americap Journal at Psychiatry, 137, 1360-1365,

Bak, R. C. (1973). Being in love and object loss. International Journal pf
Psychoanalysis, _S_4, 1-8.

Barrett, D., & Loeffler, M. (1992). Comparison of dream content of
depressed vs nondepressed dreamers. Psychological Remus, 10, 403-
406.

Basch, M. F. (1975). Toward a theory that encompasses depression: A
revision of existing causal hypotheses in psychoanalysis. In E. J.
Anthony & T. Benedek, (Eds), Depression and human existence.
Boston: Little, Brown and Company.

 

94

Beck, A. T., Ward, C. H., Mendelson, M., Mock, J ., & Erbaugh, J. (1961).
An inventory for measuring depression. Archives pf General
Psychiatg, 4, 561-571.

Beit-Hallahmi, B. (1971). Sexual and aggressive fantasies in violent and non-
violent prison inmates. Journal 9f Personalig Assessment. 3_5_, 326-
329.

Bellak, L. (1952). mfic-dggressive psychosis and allied conditions. New
York: Grune & Stratton.

Bellak, L. (1981). Brief psychoanalytic psychotherapy of nonpsychotic
depression. Americap Journal of Psychotherapy, 3_5_, 160-172.

Bemporad, J. R. (1988). Psychodynamic models of depression and mania. In
A. Georgotas & R. Cancro, (Eds), Depression an_d mania. New York:
Elsevier.

Benjamin, J ., Silk, K. R., Lohr, N. E., & Westen, D. (1989). The relationship
between borderline personality disorder and anxiety disorders.
American Journal pf Orthopsychiatry, 52, 461-467.

Bibring, E. (1953). The mechanism of depression. In W. Gaylin, (Ed) m
meaning pf despair. New York: Science House, Inc.

Birtchnell, J. (1984). Dependence and its relationship to depression. British
Journal 9f Medical Psychology, fl, 215—225.

Blackburn, 1. M. (1974). The pattern of hostility in affective illness. British
Journal 9_f Psychiag, Q, 141-145.

Blatt, S. J. (1974). Levels of object representation in anaclitic and introjective
introjective depression. T_he ppychoanalfl'c study of phg child, _2_9, 107-
157. Connecticut: Yale University Press.

Blatt, S. J ., D’Afflitti, J. P., & Quinlan, D. M. (1979). Depressive exppriences
questionnaire. Unpublished manual, Yale University, New Haven.

Blatt, S. J., Quinlan, D. M., & Chevron, E. (1990). Empirical investigations of
a psychoanalytic theory of depression. In J. Masling (Ed.), Empirical
studies 9f psychoanalfl'c theories. New Jersey: The Analytic Press.

 

 

95

Blazer, D., Woodbury, M., Hughes, D. C., George, L. K., Manton, K. G.,
Bachar, J. R., & Fowler, N. (1989). A statistical analysis of the
classification of depression in a mixed community and clinical sample.
Journal qf Affective Disorders, Q, 1120.

Blum, G. S. (1950). The Blacky Pictures: A technique fg me exploration pf
rsonality dynamics, Manual and instructions. New York: The
Psychological Corporation.

 

Bomstein, R. F., Poynton, F. G., & Masling, J. (1985). Orality and depression:
An empirical study. Psychoanalytic Psychology, 2, 241 -249.

Bowlby, J. (1963). Pathological mourning and childhood mourning. Journal
qf L113 American Psychoanalytic Association, 11, 500-541.

Buss, A. H. (1961). m psychology pf aggression. New York: John Wiley
and Sons, Inc.

Caine, T. M. (1960). The expression of hostility and guilt in melancholic and
paranoid women. Journal pf Consulting Echology, 24, 18—22.

Caine, T. M., Foulds, G. A., & Hope, K. (1967). Manual Q” th_e hostiliy an_d
direction pf hostilig questionnaire. London: University of London
Press.

Chodoff, P. (1972). The depressive personality. Archives g” General
Psychiatg, 21, 666-673.

Cochrane, N. (1975a). The role of aggression in the psychogenesis of
depressive illness. British Journal 91” Medical Psychology, fl, 113-130.

Cochrane, N. (1975b). Assessing the aggressive component of personality— 1.
Problems. British Joumfl qf Medical Psychology, g, 9-14.

Cochrane, N. (19750). Assessing the aggressive component of personality- II.
A new approach. British Journal _o_f Medical Psychology, g, 1530.

Cohen, J. (1992). A power primer. Psychological Bulletin, 1L2, 155- 159.

Cohen, M. B., Blake, G., Cohen, R. A., Fromm-Reichmann, F., & Weigert, E.
V. (1954). An intensive study of twelve cases of manic-depressive
psychosis. Psychiag, 1_7_, 103-138.

96

Cronbach, L. J. (1975). Beyond the two disciplines of scientific psychology.
Americap Psychologist, fl, 116-127.

Dahl, A. A. (1988). Aspects of the analysis of a patient with severe
depression. Scandinaviap Psychoafllytic Review, Q, 3-23.

Dar, R. (1987). Another look at Meehl, Lakatos, and the scientific practices
of psychologists. Americap Psychologist, 42, 145-151.

Deffner, G. (1986). Microcomputers as aids in Gottschalk-Gleser rating.
Psychiatgy Research, _1_8, 151-159.

Deitz, J. (1989). The evolution of the self-psychological approach to
depression. American Journal qf Psychotherapy, E, 494-505.

Derogatis, L. R. (1983). SCL-90-R: Administration scoring é procedures
manual-II aid other instruments _o_f tlrg psychopathology rating scale
series. Towson, MD: Clinical Psychometric Research.

 

Deutsch, H. (1932). Psychoanalysis qf th_e neuroses. London: Hogarth
Press.

Ebel, R. L. (1974). And still the dryads linger. Americg @chologig, 2_9_,
485-492.

Fenichel, O. (1945). The psychoanalfl'c theog pf neurosis. New York:
Norton and Company, Inc.

Fiester, S., & Siipola, E. (1972). Effects of time pressure on the management
of aggression in TAT stories. Journal qf Personalig Assessment, 36,
230-240.

Fisher, 8., & Greenberg, R. P. (1985). ILe scientific credibility pf Freud’s
theories app therapy. New York: Columbia University Press.

Foulds, G. A., Caine, T. M., & Creasy, M. A. (1960). Aspects of extra- and
intro-punitive expression in mental illness. Journal qf Mental Science,
1_0.§,599-610.

Freedman, N. (1986). On depression: The paralysis, annihilation, and
reconstruction of meaning. In J. Masling, (Ed), Empirical studies qf
psychoanalfl'c theories. New Jersey: The Analytic Press.

 

97

Freud, S. (1949). On narcissism: An introduction (1914). The collected
pagrs qf Sigmund Freud, fl 4, 30-59. London: Hogarth Press.

Freud, S. (1949). Three essays pp th_e theog _o_f sexualig (1915). London:
Imago Publishing Company, Limited.

Freud, S. (1957). Mourning and melancholia (1917). m standard edition o_f
the complete psychological works 91' Sigmund Freud, SQL. 14, 243-258.
London: Hogarth Press.

Freud, S. (1961). The ego and the id (1923). 115 standard edition qf gig
complete psychological works qf figmund Freud, Vol. 194 12-66.
London: Hogarth Press.

Friedman, A. S. (1970). Hostility factors and clinical improvement in
depressed patients. Archives o_f General Psychiam, _23, 524—537.

Gaylin, W. (Ed). (1968). m meaning qt" despair. New York: Science
House, Inc.

Georgotas, A., & Cancro, R. (1988). Depression an_d mania New York:
Elsevier.

Gershon, E. S., Cromer, M., & Klerman, G. L. (1968). Hostility and
depression. Psychiatry, 3_1_, 224-235.

Gluck, M. R. (1955). The relationship between hostility in the TAT and
behavioral hostility. Journal qf Projective Techpiques, _12, 21-26.

Goffman, E. (1967). The interactio_n ritual: Essays o_n face; _tp face begvior.
New York: Anchor Books.

Goldberg, A. I. (1975). The evolution of psychoanalytic concepts of
depression. In E. J. Anthony & T. Benedek (Eds), Depression ad
human existence. Boston: Little, Brown and Company.

Gottheil, E. (1965). An empirical analysis of orality and anality. m Journal
of Nervous an_d Mental Disease _14_1, 308-315.

 

Gottheil, E., & Stone, G. C. (1968). Factor analytic study of orality and
anality. T_hg Journal pf Nervous a_ng Mental Disease, £6, 1-17.

Gottschalk, L. A. (1979) (Ed). m content analysis o_f verbal behavior-
Further studies. New York: SP Medical and Scientific Books.

 

 

98

Gottschalk, L. A. (1986a). Research using the Gottschalk-Gleser content
analysis scales in English since 1969. In L. A. Gottschalk, F. Lolas, &
L. Vmey, (Eds), Content analysis o_f verbal behavior: Sign_1f' rcance ip
clinical medicine an_d psychiafl. Berlin: Springer-Verlag.

Gottschalk, L. A. (1986b). Content category analysis-- the measurement of
the magnitude of psychological dimensions in psychotherapy. In R. L.
Russell, (Ed), Language in psychotherapy: Strategi;e_s 9_f discoveg.
New York: Irvington Press.

Gottschalk, L. A., & Bechtel, R. J. (1982). The measurement of anxiety
through the computer analysis of verbal samples. Comprehensive
Psychiatry, _2_3, 364-369.

Gottschalk, L. A., & Bechtel, R. J. (1989). Computerized content analysis of
natural language. Artificial Intelligence in Medicine, 1, 131-137.

Gottschalk, L. A., Gleser, G. C., & Springer, K. J. (1963). Three hostility
scales applicable to verbal samples. Archives pf General Psychiatry, 2,
254-279.

Gottschalk, L. A., & Gleser, G. C. (1969). m measurement o_f psychological
states through the content analysis qf verbal behavior. Berkeley:
University of California Press.

Gottschalk, L. A., Hausmann, C., & Brown, J. S. (1979). A computerized
scoring system for use with content analysis scales. In L. A.
Gottschalk, (Ed) The content analysis pf verbal beh_avior- Further
studies. New York: SP Medical and Scientific Books.

Gottschalk, L. A., Hoigaard, J. C., Birch, H., & Rickels, K. (1979). The
measurement of psychological states: Relationships between
Gottschalk-Gleser content analysis scores and Hamilton Anxiety Rating
Scale Scores, Physician Questionnaire Rating Scale Scores and
Hopkins Symptom Checklist Scores. In L. A. Gottschalk (Ed), m
content analysis pf verbal behavior- Further studies. New York: SP
Medical and Scientific Books.

Gottschalk, L. A., & Hoigaard-Martin, J. (1986). A depression scale
applicable to verbal samples. Psychiagg Research, _1_‘Z, 213-227.

99

Gottschalk, L. A., & Lolas, F. (1986). The Gottschalk-Gleser content analysis
method of measuring the magnitude of psychological dimensions: Its
application in transcultural research. Transcultural Psychiatric Research
Review, E, 83-111.

Gottschalk, L. A., Mnget, C. N., & Gleser, G. C. (1969). Manual qf
instructions f9; using the Gottschalk-Gleser content analysis scales.
Berkeley: University of California Press.

Greenberg, R. P,. & Bomstein, R. F. (1988). The dependent personality: II.
Risk for psychological disorders. Journal qf Personalig Disorders, 2,
136- 143.

Grinker, R. R., Miller, J., Sabshin, M, Nunn, R., & Nunnally, J. (1961). m
phenomena qf depressions. New York: Hoeber.

Hafner, A. J ., & Kaplan, A. M. (1960). Hostility content analysis of the
Rorschach and TAT. Journal qf Projective Techniques, E, 137 - 143.

Hamilton, M., & White, J. (1959). Clinical syndromes in depressive states.
Joru'nal pf Mental Science. 105, 485-498.

Harlow, S. D. Goldberg, E. L., & Comstock, G. W. (1991). A longitudinal
study of risk factors for depressive symptomatology rn elderly
widowed and married women. American J___oumal o_f Epidemiology,
1_34, 526-538.

Henry, R. M. (1981). Validation of a projective measure of aggression-anxiety
for five-year-old boys. Journal qf Personalig Assessment 4_5, 359-369.

Herman, M. F. (1983). Depression and women: Theories and research.
Journal qf gm American Academy 91‘ Psychoanalysis, ii, 493-512.

Hirschfeld, R. M. A., & Klennan, G. L. (1979). Personality attributes and
affective disorders. American Journal pf Psychiatry, 136, 67-70.

Hirschfeld, R. M. A., Klerman, G. L., Chodoff, P., Korchin, S., & Barrett, J.
(1976). Dependency- self-esteem— depression. Journal _o_f th_e American
Academy qf Psychoanalysis, 4, 373-388.

Hurnphreys, L. G. (1978). Doing research the hard way: Substituting
analysis of variance for a problem in correlational analysis. Journal qf
Educational Psychology, _7_Q, 873-876.

100

Hunter, J. E., & Schmidt, F. L. (1990). Methods of meta-analysis. London:
Sage Publications.

Jacobson, E. (197 1). Depression. New York: International Universities
Press.

James, P. B., & Mosher, D. L. (1967). Thematic aggression, hostility-guilt and
aggressive behavior. Journal o_f Projective Techniqtgs fl
Personalig Assessment. 3.1, 61-67.

Jensen, A. R. (1957). Aggression in fantasy and overt behavior.
Psychological Monogpaphs: General m Applied, 21, 1-13.

J osephthal D. H. (1956). Ane e_pirical investiggrtion_ of the Lychoanalytic
theo eo_1_'y of depression bl_ use o_f the Blacky Pictures Test. Unpublished
undergraduate honors thesis, Wesleyan University, Middletown,
Connecticut.

Kagan, J. (1988). The meanings of personality predicates. American
Psychologist, Q, 614-620.

Kendell, R. E. (1970). Relationship between aggression and depression.
Archives pf General Psycpiatry, 22, 308-318.

Kiesler, D. J. (1975). Patient measures based on interviews and other speech
samples. In I. E. Waskow & M. B. Parloff, (Eds), Psychthherapy
chan g e measures. Rockville, Maryland: National Institute of Mental
Health.

Klennan, G. L., & Gershon, E. S. ( 1970) Irnipramine effects upon hostility
in depression. J__oumal of Nervous and Mental Disease 1_5_0_, 127- 132.

 

Lazare, A., Klerman, G. L., & Armor, D. J. (1966). Oral, obsessive, and
hysterical personality patterns. Archives pf General Psychiag, i4,
624-630.

Lazare, A., & Klerman, G. (1968). Hysteria and depression: The frequency
and significance of hysterical personality fcatures in hospitalized
depressed women. American Journal o_f Psychiapy, &, 48-56.

Leff, M. L., Roatch, J. F., & Bunney, W. E. ( 1970). Environmental factors
preceding the onset of severe depression. Psychiag, fl, 293-311.

101

Lemaire, T. E., & Clopton, J. R. (1981). Expressions of hostility in mild
depression. Psychological Repprts, 48, 259-262.

Lesser, G. S. (1958). Conflict analysis of fantasy aggression. _J___oumal of
Personaligy, 2_6, 29-41.

Lewinsohn, P. M., & Teri, L. (1982). Selection of depressed and
nondepressed subjects on the basis of self-report data. Journal pf
Consulting an;d Clinical Psychology, g, 590-591.

Lolas, F. (1986). Behavioral text and psychological context: On pragmatic
verbal behavior analysis. In L. A. Gottschalk, F. Lolas, & L. L. Vmey,
(Eds), Content analysis o_f verbal behavior: Significance 1p clinical
med_igine and psychiatry. Berlin: Springer-Verlag.

Lopez-Ibor Jr., J. J. (1991). The masking and unmasking of depression. In J.
P. Feighner & W. F. Boyer, (Eds), m diagposis qf depression
Chichester: John Wiley & Sons.

Masling, J. (1986). Orality, pathology, and interpersonal behavior. In J.
Masling, (Ed), Empirical studies 91“ psychoanalytic theories. New
Jersey: The Analytic Press.

Masling, J., Rabie, L, & Blondheim, S. H. (1967). Obesity, level of aspiration,
and Rorschach and TAT measures of oral dependence. J oumal o_f

Consulting Psychology, 3__], 233-239.

Masling, J ., & Schwartz, M. (1979). A critique of research in psychoanalytic
theory. G__enetic Psychology Monogrpphs, M, 257-307.

Matussek, P., & Feil, W. B. (1983). Personality attributes of depressive
patients. Archives pf General Psychiam, 4_(_), 783-790.

Mayo, P. R. (1967). Some psychological changes associated with
improvement in depression. B__1itish Journal o_f Social _aQ— Clinical
Psychology, 6, 63-68.

Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Sir
Ronald, and the slow progress of soft psychology. Journal qf
Consulting app Clinical Psychology, fl, 806-834.

Mendelson, M. (1974). Psychoanalytic concepts o_f depression (2nd Ed).
New York: Spectrum Publications, Inc.

102

Milrod, D. (1988). A current view of the psychoanalytic theory of
depression. Psychoanalytic Study qf £13 Child, Q, 83-99.

Mischel, W. (1969). Continuity and change in personality. American
Psychologist, E, 1012-1018.

Murstein, B. 1. (1968). Effect of stimulus, background, personality, and
scoring system on the manifestation of hostility on the TAT. Journal qf
Consulting an_d Clinical Psychology, 22, 355-365.

Murstein, B. I., & Wheeler, J. I. (1959). The projection of hostility on the
Rorschach and thematic stories test. Journal qf Clinical Psychology,
Q, 316-319.

Mussen, P. H., & Naylor, H. K. (1954). The relationships between overt and
fantasy aggression. Joumal 9f Abnormal Social Psychology, 42, 235-
240.

Neuringer, (1968). A variety of thematic techniques. In A. I. Rabin, (Ed),
Projective techniques '21 personality assessment. New York: Springer
Publishing Company.

Newman, R. S., & Hirt, M. (1983). The psychoanalytic theory of depression:
Symptoms as a function of aggressive wishes and level of field
articulation. Journal qf Abnormal Psychology, 22, 42-48.

Nigg, J. T., Westen, D., Lohr, N. E., Gold L. J., & Silk, K. R. (1992).
Malevolent object representations in borderline personality disorder and
major depression. Journal pf Abnormal Psychology, &, 61-67.

Nissenfeld, S. (1979). m effects _o_f fo_ur m 9! subliminal stimuli o_n female
depressives. Unpublished doctoral dissertation, Yeshiva University.

Nolen-Hoeksema, S. (1987). Sex differences in unipolar depression:
Evidence and theory. Psychological Bullet1_n,' _191, 259-282.

Oatley, K., & Bolton, W. (1985). A social-cognitive theory of depression in
reaction to life events. Psychological Review, 22, 372-388.

Oliver, J. M., & Burkham, R. (1982). Subliminal psychodynamic activation in
depression: A failure to replicate. Journal pf Abnormal Psychology,
22, 337-342.

103

O’Neill, R. M., & Bomstein, R. F. (1990). Oral- dependence and gender:
Factors rn help-seeking response set and self-reported psychopathology
in psychiatric inpatients. Journal _o_f Personalig Assessmeni,__, 55 28-40.

O’Neill, R. M., & Bomstein, R. F. (1991). Orality and depression in
psychiatric inpatients. Journal qf Personality Disorders, 5, 1-7.

Overholser, J. C., Kabakoff, R., & Norman, W. H. (1989). The assessment of
personality characteristics in depressed and dependent psychiatric
inpatients. Journal o_f Personalig Assessment, _52, 40-50.

Parkin, A. (1976). Melancholia: A reconsideration. Journal qf the American
Psychoanalytic Association, 24, 123-139.

Paykel, E. S. (1971). Classification of depressed patients: A cluster analysis
derived grouping. British Journal qf Psychiatpy, 118, 275-288.

Paykel, E. S., Klerman, G. L., & Prusoff, B. A. (1976). Personality and
symptom pattern in depression. British Journal qf Psychiatry, 122,
327-334.

Pedder, J. R. (1982). Failure to mourn, and melancholia. British Journal pf
Psychiatry, 141, 329-337.

Pedder, J. (1992). Psychoanalytic views of aggression: Some theoretical
problems. British Journal qf Medical Psychology, _6_5_, 95—106.

Philip, A. E. (1971). Psychometric changes associated with response to drug
treatment. British Journal o_f Social and Cliniggi Psychology, m, 138-
143.

Pilowsky, I., & Spence, N. D. (1975). Hostility and depressive illness.
Archives qf General Psychiag, 22, 1154—1159.

Purcell, K. (1956). The TAT and antisocial behavior. Journal pf Consulting
Psychology, _2_0, 449456.

Rado, S. (1968). The problem of melancholia (1928). In W. Gaylin, (Ed)
m meaning qf despair. New York: Science House, Inc.

Rado, S. (1968). Psychodynamics of depression from the etiologic point of
view (1951). In W. Gaylin, (Ed), m meaning 9_f despair. New York:
Science House, Inc.

104

Riley, W. T., Mabe, P. A., & Davis, H. C. (1991). Derivation and implications
of MMPI cluster groups in clinically depressed inpatient females.
Journal o_f Psychology, ,1_2_5_, 723-733.

Robertson, B. M. (1979a). The psychoanalytic theory of depression. Part 1-
The major contributors. Canadiap Journal pf Psychiatry. E, 341-352.

Robertson, B. M. (1979b). The psychoanalytic theory of depression. Part H-
The major themes. Canadian Journal pf Psychiatry, 2_4, 557-573.

Rosenthal, S. H., & Gudeman, J. E. (1967). The self-pitying constellation in
depression. British Journil _o_f Psychiatry, 113. 485-489.

Rosenthal, S. H., & Klerman, G. L. (1966). Content and consistency in the
endogenous depressive pattern. British Journal _cg‘ Psychiafl, Q,
471-484.

Rubin, D. M. (1986). Client expression pf hostiliy in psychotherapy.
Doctoral dissertation, Michigan State University, East Lansing, MI.

Rubin, D., Abeles, N., & Muller, R T. (1992). Client expression of hostility in
psychotherapy. Psychotherapy, E, 393400.

Rubinfine, D. L. (1968). Notes on a theory of depression. Psychoanalytic
Quarterly, _31, 400-417.

Rutstein, E. H., & Goldberger, L. (1973). The effects of aggressive
stimulation on suicidal patients. Psychoanalysis Q4 Contempo_r_a_ry
Science 2, 157-174.

 

Saltz, G., & Epstein, S. (1963). Thematic hostility and guilt responses as
related to self-reported hostility, guilt, and conflict. Journal pf
Abnormal Social Psychology, fl, 469-479.

Sandler, J ., & Dare, C. (1970). The psychoanalytic concept of orality. Journal
9_f Psychosomatic Rescarcl_1, _14, 211-222.

Schafer, R. (1954). Psychoanalm'c intemretation i_r_r Rorschach testing. New
York: Grune & Stratton.

Schofer, G. , Koch, U., & Balck, F. (1979). The Gottschalk-Gleser content
analysis of speech: A normative study. In L. A. Gottschalk (Ed), T_he
content analysis 52' verbal behavior- Further studies. New York: SP
Medical and Scientific Books.

105

Sedlmeier, P., & Gigerenzer, G. (1989). Do studies of statistical power have
an effect on the power of studies? Psychological Bulletin, 1_05, 309-
316.

Shave, D. W. (1974). Depression as a manifestation of unconscious guilt.
Journal pf the Americap Acaderriy _o_f Psychoanalysis, 2, 309-327.

Shavelson, R. J. (1988). Statistical reasoning fg the behavioral sciences (2nd
Ed). Boston: Allyn and Bacon, Inc.

Silverman, L. H. (1976). Psychoanalytic theory: “The reports of my death
are greatly exaggerated.” Americap Psychologi_s§, 2;, 621-637.

Slipp, S., & Nissenfeld, S. (1981). An experimental study of psychoanalytic
theories of depression. Journal qf the American Academy 91'
Psychoanalysis, 2, 583-600.

Spielberger, C. D. (1966). Anxieg an_d behavior. New York: Academic
Press.

Stone, H. (1956). The TAT aggressive content scale. Journal pf Projective
Techniggs an_d Personalig Assessment, E, 445-452.

Stone, L. (1986). Psychoanalytic observations on the pathology of depressive
illness: Selected spheres of ambiguity or disagreement. Journal o_f th_e
American Psychoanalytic Association, 3_4_, 329-362.

Story, R. I. (1968). Effects on thinking of relationships between conflict
arousal and oral fixation. Journal qf Abnormal Psychology, 13, 440—
448.

Sullivan, H. S. (1956). Clinical studies ip psychiag. New York: W. W.
Norton and Co.

Tabachrrick, B. G., & Fidell, L. S. (1983). Using multivariate statistics. New
York: Harper & Row Publishers.

Teixeira, M. A. (1990). Psychoanalm'c theog ad therapy i_n_ t_l_1§ treatment qf
manic-depressive disorders. Paper presented at the 98th Annual
convention of the American Psychological Association, Boston,

Massachusetts.

Thurston, J. R., & Mussen, P. H. (1951). Infant feeding gratification and adult
personality. Journal qf Personality, _12, 449-458.

106

Tomkins, S. S. (1963). Affect, imageg, consciousness (Vol. II). New York:
Springer.

Vmey, L. L. (1986). Assessment of psychological states through content
analysis of verbal communication. In L. A. Gottschalk, F. Lolas, & L.
L. Vmey, (Eds), Content analysis pf verbal behavior: Sigm'ircance jp
clinical medicine and psychiag. Berlin: Springer-Verlag.

Weissman, M. M., Klerman, G. L., & Paykel, E. S. (1971). Clinical evaluation
of hostility in depression. American Journal pf Psychiatry, 1%, 41-46.

Weissman, M. M., Leaf, P. J., Tischler, G. L., Blazer, D. G., Kamo, M., Bruce,
M. L., & Florio, L. P. (1988). Affective disorders in five United
States communities. Psychological Medicine, 1_8_, 141-153.

Welch, B., Schafer, R. & Dember, C. F. (1961). TAT stories of hypomanic and
depressed patients. Journal o_f Projective Techniques, 25, 221-232.

Wessman, A. E., Ricks, D. F., & Tyl, M. M. (1960). Characteristics and
concorrritants of mood fluctuation in college women. Journal _o_f
Abnormal and Social Psychology, 60, 117-126.

Whitaker, L., & Deikman, A. (1980). Psychotherapy of severe depression.
Psychotherapy: Theog, Research an_d Practice, 1_7_, 85-93.

Wiener, G. (1956). Neurotic depressives’ and alcoholics’ oral Rorschach
percepts. Journal qf Projective Techniques, 22, 453-455.

Zuckerman, M., Persky, H., Eckrnan, K. M., & Hopkins, T. R. (1967). A
multitrait multimethod measurement approach to the traits (or states)
of anxiety, depression, and hostility. Journal qf Projective Techniques
aqd Personalig Assessment. 3_1, 39-48.

APPENDICES

107

Appendix A29
Previous Research

Abbreviations--
F 2 female
M = male
NC = unable to calculate
RPFT = Rosenzweig Picture Frustration Test
HDHQ = Hostility and Direction of Hostility Questionnaire
G-GHS = Gottschalk-Gleser Hostility Scales
BDI = Beck Depression Inventory
Buss-Durkee = Buss-Durkee Hostility Inventory
ORT = Object Relations Technique
ACL = Adjective Checklist
DEQ = Depressive Experiences Questionnaire
MAS = Masling Orality Scale

Hostility

Caine, 1960
Subjects-- 26F psychiatric in-patients, Measures-- precursor to HDHQ, Findings--
depressed patients higher on self -criticism and guilt, lower on projection of hostility
than paranoid patients, Effect Size-- r = .70, .54, respectively.

Foulds et al., 1960
Subjects-20M, 20F psychiatric in-patients, Measures-- HDHQ, Findings-- More
outward hostility in depressed patients (only for Males), Effect Size-- NC.

Wessman et al., 1960
Subjects-- 25F students, Measures-- RPFT, Findings» depressed subjects
demonstrated more extrapunitive, and fewer intropunitive, responsespositive
correlation between both covert hostility-out and hostility-in and depression, Effect
Size-- NC.

Mayo, 1967
' Subjects-- 8M, 16F psychiatric in-patients, Measures-- HDHQ, Findings--
melancholics more outward hostility (criticism of others, projected delusional
hostility) than neurotic depressives, both groups less inward hostility with clinical
improvement, Effect Size-- r = .62, .65, respectively

Gershon et al., 1968
Subjects-- 6F psychiatric in-patients, Measures-- G-GHS, Findings-- positive
correlation between hostility-in and depression, Effect Size-- r = .45.

 

j The formulas for calculating effect sizes, reported as “r”, are taken from Hunter and Schmidt (1990).

 

108

Gottschalk et al., 1969
Subjects-- 4M, 15F psychiatric in-patients, Measures-- G-GHS, Findings--
negative correlation between hostility-out and depression, positive correlation
between hostility-in and depression, Effect Size-- r = -.28, .35, respectively.

Gottschalk et al., 1969
Subjects-- 12M, 12F psychiatric in-patients, Measures-- G-GHS, BDI, Findings--
positive correlation between hostility-in and depression, Effect Size-- r = .48.

Gottschalk et al., 1969
Subjects-- 20M, 30F psychiatric out-patients, Measures-- G-GHS, BDl, Findings-
positive correlation between hostility-in and depression, Effect Size-- r = .34.

Friedman, 1970
Subjects-- 155M, 379E psychiatric in-patients, Measures» Buss-Durkee,
Finding -- Si gnificant correlation between internalization of anger and depression,
Effect Size-- r = .30.

 

Klennan et al., 1970
Subjects-- 3F psychiatric in-patients, Measures-- G-GHS, Findings-- No
significant results, Effect Size-- NC.

Philip, 1971
Subjects-- 36F psychiatric in-patients, Measures-- HDHQ, Findings-- more
intropunitiveness for patients who did not improve, Effect Size-- r = .34

Rutstein et al., 1973
Subjects-- 64F psychiatric in-patients, Measures-- ACL, Rorschach, Findings--
increased depression following subliminal aggressive stimulus for suicidal patients,
but no effect on inward hostility, Effect Size-- r = .27.

Blackburn, 1974
Subjects-- 51M, 55F psychiatric in-patients, Measures-- HDHQ, Findings--
depressed patients demonstrated more intropunitiveness than compared to normative
data, this decreased with clinical improvement, Effect Size-- NC, r = ..37 for second
finding.

Cochrane, 1975
Subjects-- 76M, 124E psychiatric in-patients, Measures- ORI‘, Findings--
difficult to interpret regarding individual variables, Effect Size-NC.

Pilowsky et al., 1975
Subjects-- 15M, 53F psychiatric in-patients, Measures-- Depression questionnaire,
anger and sadness levels, Findings-- nonendogenous depressives higher anger
ratings than endogenous, Effect Size-- r = .28.

Gottschalk et al., 1979
Subjects-- 35 psychiatric out-patients (gender not given), Measures-- G-GHS,
Finding -- positive correlation between hostility-in and depression, Effect Size-- r =
.37.

109

Schiifer et al., 1979
Subjects-- 203M, 203F random normative sample, Measures-- G-GHS, Findings--
positive correlation between both covert hostility-out and hostility-in and depression,
Effect Size-- r = .16, .22, respectively.

Lemaire et al., 1981
Subjects-- 6M, 12F students, Measures-- G-GHS, Finding -- depressed subjects
demonstrated significantly more inward and outward hostility than controls, Effect
Sign r = .81, .75, respectively.

Slipp et al., 1981
Subjects-- 48F psychiatric in-patients, Measures-— ACL, TAT, Findings-- No
change in depression following subliminal aggressive stimulus, Effect Size-— NC.

Oliver et al., 1982
Subjects-- 30F psychiatric in-patients, Measures-- ACL, TAT, Findings-- No
change in depression following subliminal symbiotic stimuli, Effect Size-- NC.

Newman et al., 1983
Subjects-- 30M, 30F students, Measures-- self -esteem scale, physiological
measures, Finding - No main effect on depression following subliminal aggressive
stimulus, Effect Size-- NC.

Rubin, 1986
Subjects--15M, 25F psychiatric out-patients, Measures-- G-GHS, Findings-- a
positive correlation between depression and hostility-in was found for females, but
not for males, Effect Size-- r = .54.

Barrett et al., 1992
Subjects-- 41F students, Measures-- content analysis of dreams, Findings--
depressed subjects dreams contained significantly less anger than nondepressed
subjects, Effect Size- r = .42.

Oraligy

Josephthal, 1956
Subjects-- 57F psychiatric in-patients, Measures-- Blacky Pictures Test, Findings--
no significant relationship between orality and depression, Effect Siz -- NC.

Wiener, 1956 .
Subjects-- 15M, 12F psychiatric in-patients, Measures-- Rorschach, Findings--
depressives had more hostile and neutral oral responses than alcoholics, Effect size-
r = .50, .47, respectively.

Bomstein et al., 1985
Subjects-- 417M students, Measures-- DEQ, MAS, Findings-- overall depression
score, dependency score, and self -criticism score all significantly correlated with
orality, Effect Size-- r = .20, .11, .15, respectively.

110

O’Neill et al., 1990
Subjects-- 39M, 62F psychiatric in-patients, Measures» MMPI, MAS, Findings--
depression significantly related to orality for females but not for males, Effect Size-
(for gender/orality interaction) r = .27.

O’Neill et al., 1991
Subject -- 20M, 20F psychiatric in-patients, Measures-- MMPI, BDI, MAS,
Finding -- depression significantly related to orality for males but not for females,
Effect Size-- (males) r = .44, (females) r = .14.

 

111

Appendix B
Research Consent Form

Dear Client:

The clinic is conducting an evaluation to assess the helpfulness of the
services offered here in meeting the needs of our clients. We expect that
through this evaluation we will be able to find better ways to serve you.

In order to carry out this evaluation, we request your assistance. We
will ask you to fill out one or two questionnaires during your initial intake
interview, after your last therapy session and sometime after your therapy has
ended. In addition, we would like to tape record occasional therapy sessions.
These questionnaires and tapes will help us understand your reasons for
coming to the clinic and how useful therapy has been for you. All
questionnaires and tapes will be held in strict confidence and you will remain
completely anonymous. Your right to therapy will not be affected by your
decision on whether or not you participate in the evaluation. You also have
the right to drop out of the evaluation at any time.

If you are willing to participate in this research, please sign the
statement below.

Sincerely,

The Staff of the Psychological Clinic

I hereby agree to take part in this evaluation research and grant
permission for some of my/my child's therapy sessions to be tape recorded. I
grant this permission with the understanding that names, questionnaires and
recorded materials will be held in strict confidence.

 

Name

 

 

112

Appendix C
SCL-90-R Test

Copyright © Leonard R. Derogatis, Ph.D., 1975

INSTRUCTIONS: Below is a list of problems and complaints that people
sometimes have. Please read each one carefully. After you have done so,
please circle one of the numbers to the right that best describes how much
that problem has bothered or distressed you during the past couple of weeks
including today. Circle only one number for each problem and do not skip
any items. Please read the example before beginning.

CATEGORIES: 0 - Not at all
1 - A little bit
2 - Moderately
3 - Quite a bit
4 - Extremely

EXAMPLE: How much were you bothered by: 1. Backaches
By circling #1, this person answered that he/she was a
little bit bothered by backaches.

1. Headaches 0 l 2 3 4

2. Nervousness or shakiness inside 0 1 2 3 4

3. Unwanted thoughts, words, or
ideas that won't leave your

mind 0 1 2 3 4
4. Faintness or dizziness 0 1 2 3 4
5. Loss of sexual interest or

pleasure 0 1 2 3 4
6. Feeling critical of others 0 1 2 3 4

7. The idea that someone else can
control your thoughts 0 1 2 3 4

 

113

8. Feeling others are to blame
for most of your troubles 0 1 2 3 4

9. Trouble remembering things 0 1 2 3 4

10. Wonied about sloppiness or
carelessness 0 1 2 3 4

11. Feeling easily annoyed or

 

imitated 0 1 2 3 4
12. Pains in heart or chest 0 1 2 3 4
13. Feeling afraid in open

spaces or on the streets 0 1 2 3 4
14. Feeling low in energy or

slowed down 0 1 2 3 4
15. Thoughts of ending your life 0 1 2 3 4
16. Hearing voices that other

people do not hear 0 1 2 3 4
17. Trembling 0 1 2 3 4
18. Feeling that most people

cannot be trusted. 0 1 2 3 4
19. Poor appetite 0 1 2 3 4
20. Crying easily 0 1 2 3 4

21. Feeling shy or uneasy with
the opposite sex 0 1 2 3 4

22. Feeling of being trapped
or caught 0 1 2 3 4

23. Suddenly scared for no
reason 0 1 2 3 4

24. Temper outbursts that you
could not control 0 1 2 3 4

25.

26.
27.

29.
30.
31 .
32.
33.
34.

35.,

36.

37.

38.

39.

41.
42.

114

Feeling afraid to go out of
your house alone.

Blaming yourself for things

Pains in lower back

Feeling blocked in getting
things done

Feeling lonely

Feeling blue

Worrying too much
Feeling no interest in things

Feeling fearful

Your feelings being easily
hurt

Other people being aware of
your private thoughts

Feeling others do not understand
you or are unsympathetic

Feeling that people are
unfriendly or dislike you

Having to do things very
slowly to insure correctness

Heart pounding or racing
Nausea or upset stomach
Feeling inferior to others

Soreness of your muscles

O

OOOOOO

OOOOO

N

NNNNNN

NNNNN

DJ

WWWUJUJUJ

030300000.)

#hhh-b-b

h-b-h-h-fi-

 

43.

45.

47.

49.

51.
52.

53.

55.

57.

59.

1 15
Feeling that you are watched
or talked about by others
Trouble falling asleep

Having to check and double-
check what you do

Difficulty making decisions

Feeling afraid to travel on
buses, subways, or trains

Trouble getting your breath
Hot or cold spells

Having to avoid certain things,
places, or activities because
they frighten you

Your mind going blank

Numbness or tingling in parts
of your body

A lump in your throat

. Feeling hopeless about the

future
Trouble concentrating

Feeling weak in parts of your
body

Feeling tense or keyed up

Heavy feelings in your arms
or legs

Thoughts of death or dying

Overeating

 

61.

62.

63.

65.

67.

69.
70.

71.
72..
73.

74.
75.

76.

116

Feeling uneasy when people are
watching or talking about you

Having thoughts that are not
your own

Having urges to beat, injure
or harm someone

Awakening in the early
morning

Having ideas or beliefs that

f others do not share

Sleep that is restless or
disturbed

Having urges to break or
smash things

Having ideas or beliefs that
others do not share

Feeling very self-conscious
with others

Feeling uneasy in crowds such
as shopping or at a movie

Feeling everything is an effort
Spells of terror or panic

Feeling uncomfortable about
eating or drinking in public

Getting into frequent arguments

Feeling nervous when you are
left alone

Others not giving you proper
credit for your achievements

 

77.

78.

79.

81.

83.

85.

87.

58

117

Feeling lonely even when you
are with people

Feeling so restless you
couldn't sit still

Feelings of worthlessness

Feeling that familiar things

are strange

Shouting or throwing things

Feeling afraid you will
faint in public

Feeling that people will take advantage
of you if you let them

Having thoughts about sex
that bother you a lot

The idea that you should be
punished for your sins

Feeling pushed to get things done

The idea that something serious
is wrong with your body

Never feeling close to another person
Feelings of guilt

The idea that something is
wrong with your mind

 

118

Appendix D
Hopkins Psychiatric Rating Scale

Clinician-
Please rate the client on each of these variables, with 1 being absent
and 5 being extremely characteristic.

Intake Ratings

Hopkins Psychiatric Ratings
1. Somatization 1 2 3 4 5
2. Obsessive-Compulsive 1 2 3 4 5
3. Interpersonal Sensitivity 1 2 3 4 5
4. Depression 1 2 3 4 5
5. Anxiety 1 2 3 4 5
6. Hostility 1 2 3 4 5
7. Phobic Anxiety 1 2 3 4 5
8. Paranoid Ideation 1 2 3 4 5
9. Psychoticism 1 2 3 4 5
10. Global Pathology Index 1 2 3 4 5

Feel I had adequate information to rate (1 = Yes, 2 = No)

119

Appendix E
Gottschalk-Gleser Hostility Scales

Hostility DimctedOutwald Sca1_e: Destructive, Injurious, Critical Thoughts
and Actions Directed to Others

I. Hostility Outward-Overt
Thematic Categories

a3* Self killing, fighting, injuring other individuals or threatening threatening
to do so.

 

b3 Self robbing or abandoning other individuals, causing suffering or anguish
to others, or threatening to do so.

c3 Self adversely criticizing, depreciating, blaming, expressing anger, dislike
of other human beings.

a2 Self killing, injuring or destroying domestic animals, pets, or threatening
to do so.

b2 Self abandoning, robbing, domestic animals, pets, or threatening to do so.
02 Self criticizing or depreciating others in a vague or mild manner.
d2 Self depriving or disappointing other human beings.

al Self killing, injuring, destroying, robbing wild life, flora, inanimate objects
or threatening to do so.

bl Self adversely criticizing, depreciating, blaming, expressing anger or dislike
of subhumans, inanimate objects, places, situations.

cl Self using hostile words, cursing, mention of anger or rage without
referent.

II. Hostility Outward-Covert
Thematic Categories

a3 Others (human) killing, fighting, injuring other individuals or threatening
to do so.

120

b3 Others (human) robbing, abandoning, causing suffering or anguish to
other individuals, or threatening to do so.

c3 Others adversely criticizing, depreciating, blaming, expressing anger,
dislike of other human beings.

a2 Others (human) killing, injuring, or destroying domestic animals, pets, or
threatening to do so.

b2 Others (human) abandoning, robbing, domestic animals, pets, or
threatening to do so.

 

c2 Others (human) criticizing or depreciating other individuals in a vague or
mild manner.

d2 Others (human) depriving or disappointing other human beings.

e2 Others (human or domestic animals) dying or killed violently in death-
dealing situation or threatened with such.

f2 Bodies (human or domestic animals) mutilated, depreciated, defiled.

a] erd life, flora, inanimate objects, injured, broken, robbed, destroyed or
threatened with such (with or without mention of agent).

bl Others (human) adversely criticizing, depreciating, expressing anger or
dislike of subhurrrans, inanimate objects, places, situations.

cl Others angry, cursing without reference to cause or direction of anger.
Also instruments of destruction not used threateningly.

(11 Others (human, domestic animals) injured, robbed, dead, abandoned or
threatened with such from any source including subhurrran and
inanimate objects, situations (storms, floods, etc).

e1 Subhurnans killing, fighting, injuring, robbing, destroying each other or
threatening to do so.

fl Denial of anger, dislike, hatred, cruelty, and intent to harm.

* The number serves to give the weight as well as to identify the category.

121

The letter also helps identify the category.

Hostility Directed Inward Scalp: Self-Destructive, Self-Critical Thoughts and
Actions

1. Hostility Inward
Thematic Categories

a4* References to self (speaker) attempting or threatening to kill self, with or
without conscious intent.

b4 References to self wanting to die, needing or deserving to die.

a3A References to self injuring, mutilating, disfiguring self or threats to do so,
with or without conscious intent.

b3 Self blaming, expressing anger or hatred to self, considering self worthless
' or of no value, causing oneself grief or trouble, or threatening to do so.

c3 References to feelings of discouragement, giving up hope, despairing,
feeling grieved or depressed, having no purpose in life.

a2 References to self needing or deserving punishment, paying for one's sins,
needing to atone or do penance.

b2 Self adversely criticizing, depreciating self; references to regretting, being

sorry or ashamed for what one says or does; references to self
mistaken or 1n error.

02 References to feeling of deprivation, disappointment, lonesomeness.

al References to feeling disappointed in self; unable to meet expectations of
self or others.

bl Denial of anger, dislike, hatred, blame, destructive impulses from self to
self.

cl References to feeling painfully driven or obliged to meet one's own
expectations and standards.

* The number serves to give the weight as well as to identify the category.

122

The letter also helps identify the category.

A This code is reduced to a weight of 2 if the injury is slight. It is then
written Ia3.

123

Appendix F
Masling Orality Scale

1. Foods and drinks

a. Anything that can be eaten or drunk in its present slate is
scored—meat, bread, milk, whiskey, carrot, mushroom
(“mushroom-shaped clou ” is not scored).

b. An animal is scored only if it is invariably associated with being
edible (e.g., chicken). Duck or turkey is not scored unless
phrases descriptive of food are used (e.g., “roast duck,” “turkey
leg”). Descriptions of animal life in the sea (e. g., crabs, lobsters,
fish on Card 10) are not scored unless they are qualified as edible
(“boiled lobster”).

2. Food sources
a. Obvious sources of food—restaurant, bar, saloon, picnic,
birthday party.
b. Inferred sources—breast, bra, well-endowed woman, buxom
wormn.

3. Food objects—kettle, decanter, silverware, drinking glass. Score “pot”
and “cauldron” on Card 3 only if the act of cooking is implied.

4. Food provider— waiter, cook, bartender.
5. Passive food receiver—bird in nest, piggy bank. References to either “fat”

body shapes or “thin, slender” shapes (e.g., fat man, thin man, skinny
legs, big stomach, flat stomach) are scored.

6. Beggars, those praying for help—person saying prayers, two dogs
begging, person asking forgiveness.

7. Food organs— mouth, stomach, lips, jaw, liver, intestines, teeth.

8. Oral instruments—lipstick, clarinet, tuba, cigarette, whistle, dentures,
toothbrush, telephone, megaphone.

9. Nurturers— parent, mother, father, doctor, nurse, God, Jesus, genie, angel,
.life preserver, Good Fairy.

10. Gifts and gift givers—Xmas tree, Santa Claus, cornucopia.

11.

12.

13.

14.

15.

16.

124

Good luck— wishbone, rabbit's foot, or horseshoe if these are in the
context of a good-luck talisman; four-leaf clover.

Oral activity—eating, talking, gossiping, arguing, singing, shouting,
smoking, kissing, dogs howling or barking, lipstick stains.

Passivity and helplessness
a. Explicit statements of helpless or passive condition— ”he looks
confused,” “unable to protect himself,” “he is lost.”
b. Embryo is scored. Baby is not scored unless there is some
suggestion of passiveness, frailness.

Pregnancy and reproductive organs—placenta, womb, ovaries,
pregnancy. Vagina, penis, pelvis, and sex organs are not scored.

Baby talk in the subject's responses—teeny-weeny person, bunny rabbit,
playing patty-cake, pussycat.

Negations of oral percepts are scored— ”there is no mouth here,” “she is
not pregnant,” “empty cupboar ,” “woman with no breasts,”
“definitely not a Christ figure.”

125

Appendix G
Additional Tables

TABLE 2
Intercorrelations between verbal samples for outward hostility

 

 

 

 

Outward hostiliiy HOTOTI HOTOT2 HOTOT3

HOTOT 1 1.000

HOTOTZ -.126 1.000

HOTOT3 .263 .053 1.000
Overt hostility HOOVEI HOOVE2 HOOVB

HOOVEl 1.000

HOOVE2 -.063 1.000

HOOVE3 .043 .084 1.000
Covert hostility HOCOV] HOCOV2 HOCOV3

HOCOV 1 1.000

HOCOV2 -.053 1.000

HOCOV3 .347 -.015 1.000

TABLE 7

Correlations between depression scores and outward hostility

 

; Confidence Interval Inf. Probabilig Odds Ratio
70

 

DEPRESS and HOTOT .10 -.26 5p 5 .45 . 2.3/1

DEPRESS and HOOVE .26 -. 10 5 p 5.61 .92 11.5/1

DEPRESS and HOCOV .04 -.33 5p 5.41 .58 1.4/1
TABLE 13

Correlations between depression scores and outward hostility in males

 

' ; Confidence Interval inf, Probabiligy Odds Ratio
DEPRESS and HOT OT .73 .49 5 p 5 .97 1 00 NC

DEPRESS and HOOVE .30 -.11 5p 5.71 .952 11.5/1
DEPRESS and HOCOV .72 .48 5p _<_.96 1.00 NC

 

126

TABLE 14

Correlations between scales for males

 

I.

Confidence Interval Inf. Probability Odds Ratio

 

HOTOT AND HIN .23 -. 18 5p 5.64 .86 6.1/1

HOTOT AND ORAL -.31 -.68 5p 5.06 .94 15.7/1

HOCOV AND HIN .21 -.20 5p 5.62 .84 5.3/1

Hocov AND ORAL -.36 -.69 5 p 5.00 .98 49/1
TABLE 19

Correlations between depressi

on scores and outward hostility in females

 

 

[ Confidence Interval Inf. Probabiliy Odds Ratio
DEPRESS and HOTOT -.31 -.72 5 p g . 10 .93 13.2/1
DEPRESS and HOOVE .20 -.25 _<_ p _<_ .65 .81 4.2/1
DEPRESS and HOCOV -.36 -.75 5 p g -.03 .96 24/1
TABLE 20

Correlations between scales for females

 

; Confidence Interval

HIN and HOTOT .56 .32 5 p _<_ .79 1.00

HIN and HOOVE .46 .165p5.75 1
HIN and HOCOV .49 .22 5p 5.76 1.

ORAL and HOOVE .31 -.04_<_ p 5 .67 .94

Inference Probabilng Odds Ratio

NC
NC
NC

15.7/1

 

IIISIIII1I|I2III||IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIZIIIIIIZIIIII