Content uploaded by Judith L Herman
Author content
All content in this area was uploaded by Judith L Herman on Mar 19, 2015
Content may be subject to copyright.
Childhood Trauma in Borderline Personality Disorder
490 Am JPsychiatry 1 46:4, April 1989
Judith Lewis Herman, M.D., J. Christopher Perry, M.P.H., M.D.,
and Bessel A. van der Kolk, M.D.
Subjects with borderline personality disorder (N=
2 1) or borderline traits (N= 1 1) and nonborderline
subjects with closely related diagnoses (N=23) were
interviewed in depth regarding experiences of major
childhood trauma. Significantly more borderline sub-
jects (81 %) gave histories of such trauma, including
physical abuse (71 %), sexual abuse (68%), and wit-
nessing serious domestic violence (62 %); abuse histo-
ries were less common in those with borderline traits
and least common in the subjects with no borderline
diagnosis. These results demonstrate a strong associa-
tion between a diagnosis of borderline personality dis-
order and a history of abuse in childhood.
(Am JPsychiatry 1989; 146:490-495)
In the past two decades, borderline personality dis-
order has become the subject of intensive theoretical
and clinical investigation. Beginning with Stern (1),
successive investigators have refined their descriptive
formulations, culminating in the development of DSM-
III criteria for a reliably identifiable syndrome, stable
over time, with serious morbidity (2-9). It is generally
agreed that patients with borderline personality disor-
den are difficult to treat because of the intensity of their
engagement with caregivems, the sometimes overwhelm-
ing nature of their demands for care, and the strong
emotions and conflicts that they provoke in others
(10, 11).
Attempts to conceptualize the underlying pathology
of borderline personality disorder have generally in-
yoked either a biologic model of affective disorder
(12-15) or a psychodynamic model of developmental
arrest (10, 11, 16). In the developmental formulation,
disruptions in relations with primary caretakers are
thought to be an important factor in the genesis of the
disorder. Parental neglect and unprotectiveness are
cited by Walsh (17), Frank and Paris (18), Gunderson
(6), and Feldman and Guttman (19). Early, prolonged
separation from or permanent loss of primary caretak-
ens is described anecdotally by Adler (1 1 ) and demon-
strated in a significant proportion of patients in retro-
spective studies by Akiskal (13), Soloff and Millward
(15), and Bradley (20).
Although disruption of early attachments is fre-
quently cited, the role of childhood trauma, including
parental abuse, in the development of this disorder has
received less systematic attention. Data from three
small clinical studies offer suggestive evidence that his-
tories of childhood abuse may be especially common in
borderline patients. A study of 12 hospitalized border-
line patients reported by Stone (21) indicated that 75%
had a history of incest. In a chart review study of
psychiatric outpatients at an urban teaching hospital,
Herman (22) found that eight (67%) of 12 patients
diagnosed as borderline according to DSM-III criteria
had a history of abuse in childhood or adolescence;
such histories were found in only 22% of the entire
outpatient population. Bryer et at. (23), in an inter-
view study, found that 12 (86%) of 14 hospitalized
borderline patients diagnosed by DSM-III criteria had
a history of sexual abuse before age 16, whereas early
sexual abuse was reported by 21% of the entire in-
patient population. Although all of these studies in-
volve small numbers of patients, their findings are
consistent and provide sufficient evidence to warrant
further investigation.
The present study was undertaken to test the hy-
pothesis that a history of childhood trauma is partic-
ularly common among patients with borderline per-
sonality disorder. A fuller exposition of this hypothesis
has been published (24).
METHOD
Received March 28, 1988; revision received Sept. 22, 1988; ac-
cepted Oct. 18, 1988. From the Department of Psychiatry, Harvard
Medical School, The Cambridge Hospital, Cambridge, Mass. Ad-
dress reprint requests to Dr. Herman, 61 Roseland St, Somerville,
MA 02143.
Supported in part by NIMH grant MH-34123.
The authors thank Idell Goldenberg, M.A., Beth Hoke, M.A.,
Barbara Matthews, and Chris Pagano, who served as research
assistants.
Copyright © 1989 American Psychiatric Association.
Subjects were drawn from an ongoing longitudinal
study of borderline personality disorder in comparison
to the closely related diagnoses of schizotypal person-
ality disorder, antisocial personality disorder, and bi-
polar II affective disorder. Subjects were originally me-
cruited from ambulatory mental health settings and
from advertisements for symptomatic volunteers. The
methods of subject selection have been previously de-
scnibed in detail (9, 25, 26). After full explanation of
HERMAN, PERRY, AND VAN DER KOLK
Am JPsychiatry 146:4, April 1989 491
the study, informed consent was obtained from all sub-
jects, who were paid for their time at each interview.
Most diagnostic interviews were conducted by the
principal investigator of the longitudinal study
(J.C.P.). Definite borderline personality disorder was
diagnosed if the subject met the cutoff of five or more
DSM-III criteria and had a score higher than 150 on
the Borderline Personality Scale, second version, a 52-
item precursor of the Borderline Personality Disorder
Scale, that rates major features of the disorder in nine
subcategories (25, 26). Borderline trait was diagnosed
if the subject met at least four DSM-III criteria and
had a score higher than 130 on the Borderline Person-
ality Scale. Antisocial and schizotypal personality dis-
orders were diagnosed according to DSM-III criteria.
Bipolar II disorder was diagnosed according to Re-
search Diagnostic Criteria (27).
Childhood histories were obtained by means of a
1 00-item semistructured interview, which generally
required 2 hours (Herman and van der Kolk, un-
published manuscript). The interview covered a de-
scniption of primary caretakers and other important
relationships in childhood and adolescence, major
separations, moves and losses, sibling and peer nela-
tionships, family discipline and conflict resolution,
family alcoholism, domestic violence, and physical and
sexual abuse. Subjects were encouraged to narrate
their experiences in detail, rather than simply provid-
ing yes or no answers to questions, so that the internal
consistency and credibility of the history could be eval-
uated. All interviews were conducted by one of the
authors (J.H. or B.vdK.). Interviewers were blind to
the subjects’ diagnoses and all other previously ob-
tamed information.
The interviews were scored for positive indexes of
trauma in three areas: physical abuse, sexual abuse,
and witnessing domestic violence. Instances of cultun-
ally-accepted corporal punishment and fighting or
consensual sexual exploration between peers were not
rated as abusive. Equivocal situations were scored as
negative.
Protocols were scored for occurrence of each type of
trauma at each of three developmental stages: early
childhood (0-6 years), latency (7-12 years), and ado-
lescence (13-18 years). Within each developmental
stage, no distinction in scoring was made between sin-
gle and repeated instances of abuse by the same per-
petrator; however, additional positive scores in each
category were given for abuse by different penpetra-
tors. A rough composite measure of trauma was con-
structed by adding the positive scores for each category
of trauma at each developmental stage. Thus a range
of scores was generated from 0 (no trauma at any
developmental stage) to 9 or higher (all three forms of
trauma at all three stages or multiple perpetrators at
one or more stages).
In addition to the structured interview, subjects
completed two brief self-report questionnaires: the Im-
pact of Event Scale (28), which was used as a measure
of current symptoms of posttraumatic stress disorder,
and the Dissociative Experiences Scale (29), which
probed familiarity with dissociative states.
Data analysis was conducted by means of cross-tab-
ulation and Kendall’s tau computation for ordinal by
categorical tables. General linear models procedure for
analysis of variance (ANOVA) with post hoc analysis
for comparison of means was used for continuous var-
iables. Speanman correlation coefficients were calcu-
lated for bivaniate relationships.
RESULTS
Of 75 subjects enrolled in the ongoing longitudinal
study, we were able to contact 58 (77%) during the
time period in which this investigation was conducted
(June 1986 to December 1987). Three subjects refused
to participate after being informed of the content of
the interview. Of the 55 subjects, 29 women and 26
men, who participated in the trauma interviews, 21
(17 women and four men) were diagnosed as having
definite borderline personality disorder, 1 1 (all men) as
having borderline traits, 1 1 (six women and five men)
as having bipolar II disorder, six (three women and
three men) as having antisocial personality disorder,
and six (three women and three men) as having schizo-
typal personality disorder.
The frequencies of abuse histories in each diagnostic
category are given in table 1. The great majority
(N= 1 7 or 8 1 % )of subjects with definite borderline
personality disorder gave histories of major childhood
trauma; 71% (N=1S) had been physically abused,
67% (N=14) had been sexually abused, and 62%
(N= 13) had witnessed domestic violence. Abuse his-
tories were less common in patients with borderline
trait and least common in the subjects with no border-
line diagnosis. Histories of trauma in early childhood
(0-6 years) were found almost exclusively in border-
line subjects, and over half of the borderline subjects
(N= 12 on 57%) reported such experiences in early
childhood. Borderline subjects also reported signifi-
cantly more abuse experiences in latency than other
subjects. The differences between the groups dimin-
ished with increasing age at onset of abuse, becoming
least significant in adolescence.
Borderline subjects not only suffered from abusive
experiences more commonly than others but also ne-
ported more types of trauma, beginning earlier in
childhood and repeated over longer time periods, re-
sulting in higher total trauma scores. The distribution
of childhood trauma scores is given in figure 1. Scores
ranged from zero (18 subjects) to 10 (one subject). An
ANOVA of mean trauma score by borderline diagno-
sis was significant (F=7.82, df=2, 54, pO.OO1), and
a post hoc analysis (alpha=0.OS) indicated that the
mean trauma score for the group with definite border-
line personality disorder (4.29±2.89) was significantly
higher than the means for the group with borderline
traits (1.73± 1.95) and the means for those with any
nonbordenline closely related diagnosis (1.63 ±2.05).
FIGURE 1. Distribution of Childhood Trauma Scores Among 55
Subjects in a Longitudinal Study of Borderline Personality Disorder
EJ No Borderline Personality Disorder (N =23)
J Borderline Trait (N=11)
-Borderline Personality Disorder (N =21)
70
60
50
40
30
20
10
0II
-‘---A
0
CUMULATIVE TRAUMA SCORE
CHILDHOOD TRAUMA IN BORDERLINE SUBJECTS
492 Am JPsychiatry 146:4, April 1989
TABLE 1. Traumatic Childhood Experiences in 55 Subjects in Longitudinal Study of Borderline Personality Disorder
No
Borderline Borderline
Personality Borderline Personality
Disorder Trait Disorder
Age at Onset and Type
of Trauma
(N=21 )
N % N
(N= 1 1)
%
(N=23)
N %
Analysis
Kendall’s tau Z p
Early childhood (0-6 years)
Physical abuse 7 33 0 0 1 4 0.34 3.14 <0.005
Sexual abuse 4 19 1 9 0 0 0.28 3.39 <0.001
Witness to domestic violence 7 33 0 0 2 9 0.28 2.20 <0.05
Any trauma 12 57 1 9 3 13 0.40 3.48 <0.001
Latency (7-12 years)
Physical abuse 15 71 2 18 S22 0.42 3.63 <0.001
Sexual abuse 7 33 2 18 2 9 0.26 2.22 <0.05
Witness to domestic violence 10 48 4 37 S22 0.23 1.87 <0.10
Anytrauma 17 81 6 55 8 35 0.39 3.49 <0.001
Adolescence (13-18 years)
Physical abuse 13 62 3 27 8 35 0.23 1.79 <0.10
Sexual abuse 11 52 2 18 6 26 0.23 1.80 <0.10
Witness to domestic violence 10 48 2 18 522 0.23 1.83 <0.10
Any trauma 17 81 S 45 12 52 0.25 2.07 <0.05
All ages (0-18 years)
Physical abuse 15 71 4 36 9 39 0.27 2.20 <0.05
Sexual abuse 14 67 3 27 6 26 0.34 2.87 <0.005
Witness to domestic violence 13 62 4 36 7 30 0.27 2.16 <0.05
Anytrauma 17 81 8 73 12 52 0.31 2.67 <0.01
3-4 5-6 7-8 9-10
Rating total childhood trauma as a continuous van-
able allowed for correlation with degree of personality
pathology, as measured by the Borderline Personality
Disorder Scale, antisocial personality disorder lifetime
symptom count, and schizotypal personality disorder
lifetime symptom count (table 2). Degree of borderline
psychopathology was positively correlated with all
three forms of childhood trauma. No such correlation
TABLE 2. Correlation
and Childhood Traum
Study of Borderline Pc
s Between Personality Disorder Measures
a Scores of 55 Subjects in a Longitudinal
rsonality Disorder
.
Type of Childhood
Trauma
Spearman Correlation (r)
Schizotypal Antisocial Borderline
Physical abuse
Sexual abuse 0.03 0.15 #{216}47a
0.11 0.12 0#{149}40b
Witness to domestic
violence -0.07 0.12 #{216}#{149}4#{216}i
Total 0.03 0.22 0.SY
ap<o.ool.
b<001
was found for antisocial or schizotypal personality pa-
thology, although a trend relationship was found be-
tween antisocial symptom count and total trauma
scone (p<O.lO).
As anticipated, gender differences were also signifi-
cant. The mean total trauma score was 3.64±2.97
for women and 1.58±1.81 for men (alpha=0.OS).
Women reported more physical and more sexual abuse
in childhood, whereas witnessing domestic violence
was reported equally by men and women. After con-
trolling for diagnosis, the gender difference disap-
peared with respect to reports of physical abuse
(F=0.28) but remained significant with respect to me-
ports of sexual abuse (F=8.74, df=4, SO, p=O.OOS).
The positive association between a borderline diagno-
sis and total childhood trauma score remained signif-
icant when the effects of gender differences were con-
trolled. An ANOVA showed main effects for gender
(F=10.46, df=4, SO, p=O.0O2) and diagnosis (F=
4.S1, df=4, SO, p=O.Ol6) and no significant interac-
tion effect between gender and diagnosis (F=0.02).
HERMAN, PERRY, AND VAN DER KOLK
Am JPsychiatry 146:4, April 1989 493
After controlling for diagnosis, the gender effect dimin-
ished to a significant trend (F=3.28, df=4, SO, p
0.08). After gender was controlled for, the effect of
diagnosis remained significant (F3.71, df4, SO,
p=O.O3).
The subjects with personality disorders generally re-
ported high levels of dissociative symptoms. Bivaniate
analyses with scones on the Dissociative Experiences
Scale demonstrated a significant correlation with
scores on the Borderline Personality Disorder Scale
(r=O.29, N=SS, p0.O3) and a trend correlation with
lifetime schizotypal symptoms (r=O.26, N=SS, p=
0.06) but no correlation with lifetime antisocial symp-
toms (r5=O.19). Heirarchical regressions were con-
ducted to predict the Dissociative Experiences Scale
score on the basis of the Borderline Personality Disor-
der Scale and the total childhood trauma scores. The
scores on the Borderline Personality Disorder Scale
were significant when entered first (F=4.83, df2, 52,
p=O.03) but not significant when the childhood
trauma scones were entered first (F=.34); however, the
total childhood trauma score was significant even
when entered second (F=S.42, df2, 52, pO.02). No
differences were found between borderline, borderline
trait, and nonborderline subjects with respect to post-
traumatic symptoms as measured by the Impact of
Event Scale (F0.27).
DISCUSSION
These results demonstrate a strong association be-
tween borderline personality disorder and a reported
history of childhood abuse. The great majority of bor-
denline subjects reported such a history. Although
abuse experiences were also reported by some subjects
with closely related diagnoses, they were less common
and cumulatively less severe. Early childhood histories
of abuse and multiple childhood abuse experiences as
reflected by very high trauma scores were found almost
exclusively among borderline subjects.
The importance of our findings is enhanced by the
conservative definitions that were used in scoring
trauma histories as positive. We found no evidence to
suggest that such histories were exaggerated or fabni-
cated. The following case examples, disguised for pro-
tection of subjects’ identities, illustrate the contrasting
types of histories reported by our subjects and the
judgments that were made in assigning trauma scores.
Case 1 .Ms. A, a 35-year-old woman with borderline per-
sonality disorder, was the fourth of five siblings born to two
alcoholic parents. After her father deserted the family when
she was 3 years old, the household became increasingly cha-
otic; her mother enforced unclear and inconsistent rules by
screaming, hair pulling, hitting on the head and face, and
kicking in the knees and genital area. In Ms. A’s words,
“You never knew when to expect it. You could do something
really wrong and she wouldn’t notice, and then you could
knock over your milk and she would fly off the handle.” Her
mother remarried when Ms. A was 9 years old, and the level
of violence in the home diminished. Shortly after entering the
home, however, the stepfather began sexually molesting Ms.
A and her three sisters. The incestuous relationship, which
proceeded to oral sex and intercourse, continued until she
ran away from home at age 15. While on the road, Ms. A
frequented bars where she would pick up older men, offering
sex in exchange for shelter. At age 17 she was brutally raped
and beaten in one such encounter, requiring hospitalization
for her injuries. Ms. A received a trauma score of S (physical
abuse in early childhood and latency, sexual abuse in latency
and adolescence, and rape in adolescence).
Case 2. Mr. B, a 28-year-old man with borderline traits,
described an intact family with a capricious and domineering
father and a compliant, submissive mother. He described
very restrictive family rules, “like boot camp,” and frequent
corporal punishment (hitting with a belt) and stated that “If
my father had been left to his own devices, I would have been
a battered child, but my mother protected me.” He described
one incident at age 1 1 in which his father became enraged,
chased him into his room screaming “I’ll kill you,” cornered
him, and began to strangle him. His mother attempted to
intercede, at which point his father attacked her, struck her
in the face with a closed fist, and knocked her to the ground.
After this attack, his father was remorseful, and no similar
incidents occurred. Mr. B received a trauma score of 2
(one incident each of physical abuse and witnessing domestic
violence).
Case 3. Ms. C, a 38-year-old woman with bipolar II dis-
order, was the youngest daughter among eight children. Her
father was a severe alcoholic, and she described her mother
as raising the family single handedly. She described her
mother’s discipline as very strict: “She was very old country;
she was trying to cope too.” Although discipline was carried
out by means of frequent hitting with a cane or a “bony Irish
hand,” Ms. C stated that “It was not traumatic, she did it
with everybody, it seemed all right; the nuns at school did it
too.” When Ms. C was 10 years old, her 14-year-old brother
began to involve her in sexual games, including showing and
fondling of genitals, kissing, and imitation of activities
shown in pornographic magazines. She idealized her brother,
was grateful for the attention, and in spite of the age differ-
ence did not perceive the sexual relationship as exploitative
until she was 12 years old, at which time her brother at-
tempted to bribe her to perform the same activities with his
friends. She felt deeply betrayed and angrily refused. Ms. C
received a trauma score of 0 (harsh but nonabusive corporal
punishment, equivocal sexual abuse).
Although no definitive conclusions regarding the eti-
ology of borderline personality disorder can be drawn
from correlations based on retrospective data, the hy-
pothesis that childhood abuse has a major formative
role in the development of the disorder is strongly sup-
ported by our findings. The strength of the association
between childhood trauma and borderline personality
disorder suggests that it is an important factor but not
alone sufficient to account for borderline psychopa-
thology. It is possible that trauma is most pathogenic
for children with vulnerable temperaments or for those
most lacking protective factors, such as positive rela-
tionships with other caretakers or siblings.
Despite severe abuse histories, the borderline sub-
CHILDHOOD TRAUMA IN BORDERLINE SUBJECTS
494 Am JPsychiatry 1 46:4, April 1989
jects did not report current symptoms of posttnaumatic
stress disorder, at least as measured by the Impact of
Event Scale. It appeared that memories of the abuse
had become integrated into the total personality orga-
nization and had become essentially ego syntonic. The
subjects generally did not perceive a direct connection
between their current symptoms and abusive experi-
ences in childhood. This finding is compatible with
observations from follow-up studies of trauma victims
(30, 31) which indicate that fragments of the trauma
may be transformed oven time and relived in a variety
of disguised forms, e.g., as somatic sensations, affect
states, visual images, behavioral reenactments, or even
dissociated personality fragments. Our finding that
dissociative symptoms were more strongly correlated
with childhood trauma than with borderline psycho-
pathology per se is consistent with the recent finding of
Spiegel et al. (32) that dissociation and trauma are
highly correlated.
Childhood trauma has been implicated as an etio-
logical factor in such diverse psychiatric conditions as
somatoform disorder (33), panic disorder (34), and
multiple personality disorder (35-38). Thus, it might
be possible to conceptualize a range of adaptations to
childhood trauma, or trauma spectrum disorders, with
multiple personality disorder representing an extreme
adaptation to severe chronic abuse, borderline person-
ality disorder representing an intermediate form of
adaptation to chronic abuse, and some forms of somat-
oform, panic, and anxiety disorders representing dis-
sociated somatic reexpeniencing of more circumscribed
traumatic events (30).
Childhood abuse as an important antecedent to the
development of borderline personality disorder could
explain in part the higher prevalence of borderline per-
sonality disorder in women. Epidemiologic data on
child abuse (39) indicate that although boys and girls
are at approximately equal risk for physical abuse,
girls are at two to three times greater risk for sexual
victimization. Moreover, sexual abuse is apparently
more prevalent, and often more prolonged, than phys-
ical abuse (40). Thus, girls may be more frequently
exposed to conditions favoring the development of
borderline personality disorder.
Conceptualizing borderline personality disorder as a
complicated posttraumatic syndrome has direct impli-
cations for the treatment of patients. Clinical literature
on the treatment of posttraumatic syndromes (41-48)
has shown the importance of recovery and integration
of traumatic memories with their associated affects
and the necessity for validation of the patient’s trau-
matic experiences. The integration of the trauma is a
precondition for development of improved affect tol-
erance, impulse control, and defensive organization;
the validation of the trauma is a precondition for nes-
toration of an integrated self-identity and the capacity
for appropriate relationships with others. Posttraumat-
ic states are often undiagnosed in cases in which se-
crecy or stigma prevents recognition of the traumatic
origins of the disorder; such patients may show me-
mankable improvement when the connection between
symptom and trauma is recognized. Whether some of
the negative therapeutic reactions so frequently ob-
served in borderline patients might be avoided by early
and appropriate recognition of the relationship be-
tween the patient’s current symptoms and traumatic
experiences in childhood remains to be determined.
REFERENCES
1. Stern A: Psychoanalytic investigation of and therapy in the bor-
derline neuroses. Psychoanal Q1938; 7:467-489
2. Perry JC, Klerman GL: Clinical features of the borderline per-
sonality disorder. Am JPsychiatry 1980; 137:165-173
3. Andrulonis P, Glueck B, Stroebell C, et al: Borderline person-
ality subcategories. JNerv Ment Dis 1982; 170:670-679
4. Pope H, Jonas J, Hudson J, et al: The validity of DSM-III bor-
derline personality disorder. Arch Gen Psychiatry 1983; 40:23-
30
S. McGlashan T: The Chestnut Lodge follow-up study, II: long-
term outcome of borderline personalities. Arch Gen Psychiatry
1984; 41:586-601
6. Gunderson JG: Borderline Personality Disorder. Washington,
DC, American Psychiatric Press, 1984
7. Frances A, Clarkin JF, Gilmore M, et al: Reliability of criteria
for borderline personality disorder: a comparison of DSM-III
and the Diagnostic Interview for Borderline Patients. Am JPsy-
chiatry 1984; 141:1080-1084
8. Barasch A, Frances A, Hurt S, et al: Stability and distinctness of
borderline personality disorder. Am JPsychiatry 1985; 142:
1484-1486
9. Perry JC, Cooper SH: Psychodynamics, symptoms and outcome
in borderline and antisocial personality disorders and bipolar
type II affective disorder, in The Borderline: Current Empirical
Research. Edited by McGlashan T. Washington, DC, American
Psychiatric Press, 1985
10. Kernberg 0: Borderline Conditions and Pathological Narcis-
sism. New York, Jason Aronson, 1975
1 1 .Adler G: Borderline Psychopathology and Its Treatment. New
York, Jason Aronson, 1985
12. Klein D: Psychopharmacology and the borderline patient, in
Borderline States in Psychiatry. Edited by Mack J. New York,
Grune & Stratton, 1975
13. Akiskal HS: Subaffective disorders, dysthymic, cyclothymic and
bipolar II disorders in the borderline realm. Psychiatr Clin
North Am 1981; 4:25-46
14. Akiskal HS, Chen SE, Davis GC, et al: Borderline: an adjective
in search of a noun. JClin Psychiatry 1985; 46:41-47
15. Soloff P, Millward J: Psychiatric disorders in the families of
borderline patients. Arch Gen Psychiatry 1983; 40:37-44
16. Mahler M, Kaplan L: Developmental aspects in the assessment
of narcissistic and so-called borderline personalities, in Border-
line Personality Disorders: The Concept, the Syndrome, the Pa-
tient. Edited by Harticollis P. New York, International Univer-
sities Press, 1977
17. Walsh F: The family of the borderline patient, in The Borderline
Patient. Edited by Grinker R, Werble B. New York,Jason Aron-
son, 1977
1 8. Frank H, Paris J: Recollections of family experience in border-
line patients. Arch Gen Psychiatry 1981; 38:1031-1034
19. Feldman RB, Guttman HA: Families of borderline patients: lit-
eral-minded parents, borderline parents, and parental protec-
tiveness. Am JPsychiatry 1984; 141:1392-1396
20. Bradley SJ: The relationship of early maternal separation to
borderline personality in children and adolescents: a pilot study.
Am JPsychiatry 1979; 136:424-426
21 .Stone MH: Borderline syndromes: a consideration of subtypes
and an overview, directions for research. Psychiatr Clin North
Am 1981; 4:3-13
22. Herman JL: Histories of violence in an outpatient population.
Am JOrthopsychiatry 1986; 56:137-141
HERMAN, PERRY, AND VAN DER KOLK
Am JPsychiatry 146:4, April 1989 495
23, Bryer JB, Nelson BA, Miller JB, et al: Childhood sexual and
physical abuse as factors in adult psychiatric illness. Am JPsy-
chiatry 1987; 144:1426-1430
24. Herman JL, van der Kolk B: Traumatic antecedents of border-
line personality disorder, in Psychological Trauma. Edited by
van der Kolk B. Washington, DC, American Psychiatric Press,
1987
25. Perry JC: Depression in borderline personality disorder: lifetime
prevalence at interview and longitudinal course of symptoms.
AmJ Psychiatry 1985; 142:15-21
26. Perry JC, Cooper SH: A preliminary report and defenses and
conflicts in borderline personality disorder. JAm Psychoanal
Assoc 1986; 34:863-893
27. Spitzer RI, Endicott J, Robins E: Research Diagnostic Criteria:
rationale and reliability. Arch Gen Psychiatry 1978; 35:773-
782
28. Horowitz M, Wilner N, Alvarez W: Impact of Event Scale: a
measure of subjective stress. Psychosom Med 1979; 41:209-
218
29. Bernstein EM, Putnam FW: Development, reliability and valid-
ity of a dissociation scale. JNerv Ment Dis 1986; 174:727-735
30. van der Kolk B: The trauma spectrum: the interaction of bio-
logical and social events in the genesis of the trauma response.
JTraumatic Stress 1988; 1:273-290
3 1 .Brett EA, Ostroff R: Imagery and posttraumatic stress disorder:
an overview. Am J Psychiatry 1985; 142:417-424
32. Spiegel D, Hunt T, Dondershine HE: Dissociation and hypno-
tizability in posttraumatic stress disorder. Am JPsychiatry
1988; 145:301-305
33. Krystal H: Trauma and affects. Psychoanal Study Child 1978;
33:81-116
34. Faravelli C, Webb T, Ambonetti A: Prevalence of traumatic
early life events in 3 1 agoraphobic patients with panic attacks.
Am JPsychiatry 1985; 142:1493-1494
35. Bliss EL: Multiple personalities: a report of 14 cases with im-
plications for schizophrenia and hysteria. Arch Gen Psychiatry
1980; 37:1388-1397
36. Horevitz RP, Braun BG: Are multiple personalities borderline?
Psychiatr Clin North Am 1984; 7:69-87
37. Putnam FW, Post RM, Guroff JJ, et al: One hundred cases of
multiple personality disorder. JClin Psychiatry 1986; 47:285-
293
38. Sachs R, Goodwin J, Braun B: The role of childhood abuse in
the development of multiple personality, in Multiple Personality
and Dissociation. Edited by Braun B, Kluft R. New York, Guil-
ford Press, 1986
39. Finkelhor D: Sexually Victimized Children. New York, Free
Press, 1979
40. Pagelow M: Family Violence. New York, Praeger, 1984
41 .Kardiner A: The Traumatic Neuroses of War. New York,
Hoeber, 1941
42. Niederland WG: The role of the ego in the recovery of early
memories. Psychoanal Q1965; 34:564-571
43. Horowitz MJ: Stress Response Syndromes. New York, Jason
Aronson, 1976
44. Herman JL: Father-Daughter Incest. Cambridge, Harvard Uni-
versity Press, 1981
45. Danieli Y: The treatment and prevention of long-term effects
and intergenerational transmission of victimization: a lesson
from Holocaust survivors and their children, in Trauma and Its
Wake: The Study and Treatment of Post-Traumatic Stress Dis-
order. Edited by Figley CR. New York, Brunner/Mazel, 1985
46. Rieker PP, Carmen E: The victim-to-patient process: the dis-
confirmation and transformation of abuse. Am JOrthopsy-
chiatry 1986; 56:360-370
47. Herman JL, Schatzow E: Recovery and verification of memories
of childhood sexual trauma. Psychoanalytic Psychol 1987; 4:1-
14
48. van der Kolk B: Psychological Trauma. Washington, DC,
American Psychiatric Press, 1987
