Borderline Personality
and the Rorschach Test
Margaret Thaler Singer, PhD, Dale G. Larson, PhD
\s=b\ Rorschach responses of borderline persons, acute and
chronic schizophrenics, normals, and neurotics were compared
on summary, composite, and fabulized combination scores and
on a score reflecting decline in the quality of responses to
individual cards. The groups' summary scores were as ego
function theory would predict; normals had the highest scores,
followed by neurotics, borderline persons, acute schizophrenics,
and chronic schizophrenics. In a three-group comparison, dis-
criminant-function analysis correctly classified most of the bor-
derline and acute and chronic schizophrenic subjects. In a
two-group comparison, stepwise regression analysis correctly
classified most of the borderline and acute schizophrenic sub-
jects. The borderline persons tended to produce more fabulized
combination responses and show a greater decline in response
quality on each card. The associative drift and sporadic reason-
ing problems imputed to borderline persons clinically distion-
guished the borderline sample's Rorschach records.
(Arch Gen Psychiatry 1981; 38:693-698)
grows out of long-term efforts to
The present report
better
reviewing
conceptualize
the
psychological
borderline personalities. In
test literature on the border¬
line personality, Singer' and Gunderson and Singer2 noted
that although the reports were primarily case studies and
impressionistic descriptions, there was clear-cut agree¬
ment about the test patterns of borderline persons. These
See also 686.
persons show ordinary reasoning and communication in
highly structured test situations such as the Wechsler
Adult Intelligence Scale. However on projective techniques
such as the Rorschach test, where there is little structure,
for publication Jan 15, 1980.
Accepted
From the Department of Psychiatry, University of California, San
Francisco (Dr Singer); and the Department of Psychology, University of
California, Berkeley (Drs Singer and Larson).
Reprint requests to 17 El Camino Real, Berkeley, CA 94705 (Dr
Singer).
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 03/04/2016
they demonstrate flamboyantly deviant reasoning and
associative processes. In fact, as Singer1 noted, their
projective test responses are noted for being far more
openly filled with "primary process" associations and
"schizophrenic" thinking than are the Rorschach records of
most schizophrenics.
Both the test literature and clinical reports emphasize
that many borderline persons characteristically display
associative drift and sporadic reasoning problems in verbal
situations in which there is little structure. That is, in
unstructured interviews, in therapy sessions, and on pro¬
jective tests where the borderline person must provide his
own structure, his circumstantiality and drift are seen.
Although these drift and reasoning problems are beyond
those of normals and neurotics both in quality and quanti¬
ty, they are neither the gross thought-disordered conversa¬
tional slips of chronic schizophrenics nor the juxtapositions
of ideas of acute schizophrenics.
In the present research, the Rorschach was chosen as a
useful procedure to assess the presence and extent of
certain inferred thought and communication features
ascribed to borderline persons. A developmental-level scor¬
ing procedure was applied to the Rorschach responses of
borderline and certain matched comparison groups. This
was done to ascertain if the imputed associative drift and
sporadic reasoning problems that characterize borderline
persons clinically do indeed distinguish the Rorschach
records of borderline persons from those of other groups,
especially from those of acute schizophrenics.
METHOD
The research focused on formal, cognitive properties of Ror¬
schach responses and asked the following questions. (1) To what
extent do scores derived from a developmental-level scoring
system (Table 1, footnotes)35 differentiate a borderline group
from remitting and nonremitting schizophrenic groups and from
comparable normal and neurotic groups? (2) In addition, do
fabulized combination responses'6" distinguish the borderline
group from the others? In the fabulized combination response,
unrealistic relationships are posited between two or more percepts
solely on the basis of their spatial contiguity. Examples from our
borderline subjects' records are as follows: for card Í0, "A horse's
 Table 1—Mean Scores on Rorschach Variables4

Diagnostic Group
Acute Chronic
Normal Neurotic Borderline Schizophrenic Schizophrenic
1. W+ +% 3.00 1.96 2.44 1.46 0.45
2. D+ +'. 1.55 0.64 0.44 1.21 0.20
t3. W + % 4.55 7.20 2.84 4.21 1.00
t4. D + % 13.30 5.96 6.52 5.96 3.40
5. Wm% 14.70 12.60 10.52 10.50 10.20
6. Dm% 30.70 35.32 2828 28.29 25.50
7. Wv% 3.60 3.04 1.72 2.67 1.55
t8. Adx + Hdx% 0.10 0.36 0.08 0.21 1.65
9. Dd + % 3.35 4.76 3.60 2.68 1.15
10. Dv% 3.30 4.56 3.44 4.33 3.00
til- Da% 0.00 0.64 3.16 3.5B 4.65
12. D-% 11.45 13.60 15.24 12.33 15.50
13. DdD% 0.00 0.00 0.12 0.13 •0.10
14. Dd-% 2.20 2.32 4.40 4.08 2.85
t15. Wa% 1.10 0.28 1.08 3.29 5.80
16. W-% 5.55 3.84 4.08 6.63 7.30
M7. DW% 0.45 0.00 0.12 0.75 1.45
tie. FabC% 0.55 2.52 10.06 5.42 3.05
Í19. ConR% 0.00 0.28 2.00 3.45
|20. Per% 1.00 0.00 0.12 0.33 7.95
21. R 31.75 36.36 39.08 32.17 27.80
22. P% 28.15 24.60 22.68 26.33 20.20
t23. HCDL 39.95 38.52 34.76 35.42 32.50
124. R + % 78.60 78.80 67.00 66.38 52.80
¡"25. H% 24.45 20.96 22.44 22.42 13.15
t26. IPT 1.30 2.72 12.24 8.25 8.10
f27. Index of integration 23.40 17.68 15.66 13.96 6.20
t28. 1st RDL 3.85 3.76 3.54 3.25 2.78
t29. Becker 3.59 3.49 3.04 2.96 2.61
30. DDWC -0.25 -0.27 -0.50 -0.29 -0.17
*The first 20 items are described in detail by Becker* and Goldfried et al." The remaining ten items are as follows:
21. R, the number of responses in the record.
22. P%, the percentage of popular content (Beck's scoring).
23. HCDL, the developmental level of human-content responses. It is computed as the sum of the Becker weights for human percepts divided by the
number of human percepts.
24. R + %, form quality. It is computed as the percentage of responses scored F+ using Beck's system. All responses were scored.
25. H%, the percentage of total responses scored H, Hd, and (H).
26. IPT, the index of primitive thought. It is computed as ([FabC + ConR + DW + DdD]/R) 100.
27. Index of Integration is computed as follows: [(W ++) + (W + ) + (D++) + (D + )]/all W + D.
28. 1st RDL, the developmental level of first responses to cards. It Is computed as the mean of the Becker weights for first responses to cards.
29. Becker, the Becker developmental-level score. It is computed as the sum of the Becker weights for all percepts divided by R.
30. DDWC, developmental decline within cards. It is computed as follows: ([mean of Becker weights for all responses to a given card]—[Becker
weight for
first response to the card])/10. It is important to note that a high DDWC score (ie, greater decline) is represented by a large negative number.
tThe value is significant at < .05 by univariate analysis of variance.

head with two sea horses growing out of his ears"; for card 9, "Two used for each of the 20 Becker scoring categories is based on the
children's heads; they seem to be lying down next to each other. proportion of a scoring category to the total number of responses
The rest of the area, these clouds here, would be their breath"; and (eg, W+ +% [W+ +/total R] 100). This scoring system is
=

for card 2, "A butterfly flying in front of a spaceship, and maybe a
rabbit or a bird got in the picture here." (3) Does a greater decline
in developmental-level scores within the series of responses given
to each Rorschach card characterize the borderline group in
contrast to the other groups? To answer this question, we created
the developmental decline within cards (DDWC) score, which is
defined in Table 1.
A developmental-level scoring system developed by Friedman,"
with modifications by Becker' and the present authors, was used
to quantify certain Rorschach response properties. Becker
derived from Werner's"·"' theory of perceptual and cognitive
development.
In this procedure, a numerical score is assigned to each response
reflecting the articulation (area of blot), specificity (specified or
vague content), accuracy (form quality), and any "cognitive slip¬
page" problems (eg, contaminations, confabulations, fabulized
combinations, and perseverations) that may be present in that
response.'1"" Each Rorschach response is assigned a score from
1 to 6. A low score reflects poor articulation and conceptualization;
a high score reflects a well-articulated, integrated, and realistic

assigned weights to each scoring category and added the following
three scores to Friedman's system: oligophrenic detail
(Adx + Hdx), minus unusual detail (Dd-); and plus unusual detail
(Dd + ). Friedman's summary scores were based on the proportion
of a scoring category to the type of location choice (eg,
[W++% W+ +/total W] 100). Here the summing method
=
response.
The scores for a person's entire record are summed, and the
average is obtained. This summary score indicates the average
articulation and synthesis qualities of the person's Rorschach
responses. Analyses of these summary scores and of various
subscores within the record were made in the present work.

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 03/04/2016

 Table 2—Demographic Features of Sample
Sex of Index
Severity of Offspring Age of Education of
Disorder of No. of Index, yr Index, yr
Groups Index Offspring* Families (Mean ± SEM) (Mean ± SEM)
Normal 1-2 20 11 19.0 ± 0.7 12.5 ± 0.5
Neurotic 16 17.4 ± 0.4 11.0 ± 0.4
Borderline 25 17 16.2 ± 0.5 11.9 ± 0.4
Acute schizophrenic 24 10 14 22.4 ± 1.3 13.1 ± 0.4
Chronic schizophrenic 6-7 20 10 10 23.0 ± 1.0 11.5 ± 0.5
Total 114 62 52 19.9 ± 4.6 12.0 ± 2.1

SA global, seven-point rating scale for severity of diagnosis was used as described in previous work.' These diagnostic ratings were made ¡ndepen-
dently of the psychological testing.

This scoring method differs markedly from traditional Ror¬

Table 3.—Contrast Matrix of t Statistics schach scoring procedures. The 30 subscores derived for each
for Comparisons of Three Groups person's Rorschach record quantify features that can be termed
"aspects of ego functioning." These scores reflect reality aware¬
Borderline Borderline Acute ness and logical reasoning as these features are displayed by one
vs Acute vsChronic vs Chronic
Variable*
person communicating with another while viewing the Rorschach
Schizophrenic Schizophrenic Schizophrenic cards.
1. W+ +% 1.09 2.10 1.06
2. D+ +' 0.36 1.50 Sample
3. W + % 1.15 1.47 2.55t The 114 persons whose test records were studied were a young
4. D + % 0.33 2.01 1.63 adult sample with a mean age of 19.9 years (Table 2). A global,
5. Wm% 0.19 0.17
0.01 seven-point rating scale for severity of diagnosis was used to
6. Dm% 0.002 0.63 0.62 classify each person studied. Levels 1 and 2 included persons who
7. Wv% 0.84 0.14 0.90 were within ordinary limits of psychiatric normality. All those
8. Adx + Hdx% 0.29 3.41t 3.12t with a score of 3 or greater were inpatients. Level 3 included
9. Dd + % 0.71 2.29 1.60 persons with symptomatic (primarily severe obsessional or depres¬
10. Dv% 0.64 0.30 0.90 sive states) neurotic or personality disorders but without current
or past evidence of psychotic features. Level 4, the borderline
11. Da% 0.28 0.96 0.68
12. D-% 1.14 0.10
sample, was chosen prior to the appearance of both the Gunderson
and Singer2 lists and the Spitzer et al12 lists. Our raters used
13. DdD% 0.06 0.12 0.15
criteria offered by Kringlen12 and Wynne et al," but they did not
14. Dd- 0.21 0.95 0.75 subdivide the group as did Grinker et al." The Kringlen and the
15. Wa% 1.35 2.74t 1.45 Wynne et al descriptions emphasized deviations in communica¬
16. W-% 1.30 1.57 0.32 tion, referential or paranoid ideation, odd perceptual experiences,
17. DW% 1.24 2.46f 1.27 and social isolation and anxiety; these features are central to the
18. FabC% 2.08 2.99Î 1.00 Spitzer et al, as well as the DSM-III, cluster of schizotypal
19. ConR% 0.10 1.29 1.18 features.'2'" Thus, the majority of our sample would have met the
20. Per% 0.09 3.16t 3.04t schizotypal criteria; the remainder would have fit the DSM-III
"borderline" criteria.
21. R 1.32 2.05 0.79
Patients at level 5 had a common, good premorbid history, an
22. P% 1.05 0.69 1.68 acute onset, and a definite remission1"1" and typically showed
23. HCDL 0.02 0.09 0.11 what Wynne and Singer1" have described as "fragmentation" of
24. R + % 3.09t 2.94f the thinking process. The 20 patients at levels 6 and 7 had never
25. H% 0.01 3.32t 3.30t shown a clinically important remission from a schizophrenic
26. IPT 1.34 1.35 psychosis and would be called poor-premorbid or "process" schiz¬
27. Index of ophrenics using the prognostic criteria of the Phillips Scale.2"
integration 0.57 3.04t 2.48t RESULTS
28. 1st RDL 1.56 0.38 0.24
29. Becker 0.58 2.95t 2.39 To determine whether sex should be included in all analyses as
30. DDWC 1.66 2.47t 0.38 an independent variable, a multivariate analysis of variance was
"The first 20 variables are described in detail by Becker5 and Goldfried et
performed using Finn's21 multivariance. Since neither the sex
al." The remaining variables are described in Table 1.
effect nor the sex by group interactions were significant, sex was
tThe value is significant at < .05. eliminated as a variable.
General Group Differences

The mean Becker developmental-level scores (item 29 in Table 1)

For each person, each subscore in the Becker system was
computed as a proportion of the total number of responses within
that person's record. The 20 subscores from the Becker method and
the ten other subscores selected by the authors were included in
the statistical analysis (Table 1).
The developmental-level scoring and our extension of it is a
reliable procedure. Both authors scored all protocols independently
and were blind to diagnosis; 90% agreement was obtained for all
scorings. Discrepant scores were inspected and an agreed-on score
was derived and used in the data analysis.
for the five groups, as well as their mean first-response scores
(item 28), fell into the sequence that theory predicts. A univariate
analysis of variance distinguished the groups at < .001. That is,
the scores in descending order were as one predicts from ego
function theory; the highest scores were those of normals, followed
by neurotics, then borderline persons, then acute schizophrenics,
and, lastly, chronic schizophrenics (Table 1). Furthermore,
Scheffe22 contrasts were calculated for the scores of each possible
pairing of groups (normal vs neurotic, etc). Each pairing revealed a
significant difference between both the normal and neurotic

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 03/04/2016

 Statistics
univariate statistics from this three-group analysis were exam¬
Table 4.—Discriminant Analysis ined. Those 11 variables that had significant (P < .05) univariate t
Standardized Discriminant- statistics were then used as independent variables in a discrimi¬
Function Coefficients i nant-function analysis with group membership as a dependent
variable (Tables 4 and 5).
Variable* 1st Function 2nd Function The classification of the subjects into the three groups using
W+% 0.092 0.625 these 11 variables is given in Table 5. Classification was correct for
Adx + Hdx% 0.770 -0.233 80% of the borderline subjects, 63% of acute schizophrenic subjects,
Wa% -0.002 0.434 and 70% of the chronic schizophrenic subjects. Overall, 71% of 69
DW% 0.146 0.409 subjects were correctly assigned.
FabC% -0.422 -0.580 Borderline vs Acute Schizophrenic Groups
Per% 0.146 -0.669
R+% -0.543 0.551 Finally, since the discrimination between the borderline and
H% -0.332 0.578 acute schizophrenic groups is of most importance diagnostically,
Index of integration —0.069 —0.190 data from subjects in those two groups were used in a stepwise
Becker -0.132 -1.034 regression analysis, with group membership as the dependent
variable and the same 11 scores as independent variables. This
DDWC 0.182 0.330
analysis indicated that eight of the variables were the "best"
•The first seven variables are described in detail by Becker' and discriminators between the borderline and acute schizophrenic
Goldfried et al." The remaining variables are described in Table 1. groups in that stepwise regression analysis. Using this group of
tFor the first function, Wilk's lambda 0.400; 2 56.74, df 22,
= = =
eight variables produced the smallest mean square error figure. A
= .000. For the second function, lambda 0.B05; ' =
13.46, df 10, = =
cutoff score that minimized total selection error was calculated for
=
.199. this composite score, and the percentage of correct classifications
was found (Table 6). Using this procedure, classification was
correct for 96% of borderline subjects and 67% of acute schizo¬
Table 5.—Classification Into Diagnostic Groups phrenic subjects.
by Discriminant-Function Analysis* Overall, 82% of 49 subjects were correctly classified. The regres¬
sion analysis was performed on the scores from the borderline and
Predicted Group Membership, acute schizophrenic groups because differentiating between these
No. of Subjects two groups is of the most clinical interest; furthermore, this
Actual analysis provided a conceptual basis for determining the best
Diagnostic Acute Chronic
Group Borderline Schizophrenic Schizophrenic
equation based on a criterion other than significance (by specify¬
Borderline 25 20
ing the criterion equation beforehand). The equation as a whole
Acute
significantly discriminated between the two groups, and the
24 6 15
FabC%, DDWC%, and Becker variables were significant univar¬
schizophrenic iate discriminators (P < .05). Table 7 gives comparisons of the
Chronic scores of the acute schizophrenic and borderline groups on the
schizophrenic 20 3 3 14
eight variables selected as the best discriminators between
*The following variables were used in the analysis: W + %, Adx + Hdx%, them.
Wa%, DW%, FabC%, Per%, R + %, H%, Index of Integration, Becker, and No simple interpretation can be given for the findings in Table
DDWC. See Table 1 tor definitions. 7. The acute schizophrenic group tended to give more human (H%)
content than the borderline group. Five of the items (FabC%,
Wa%, Adx + Hd%, W + %, and DW%) illustrate how the groups
differed in their approach to synthesizing percepts. The borderline
Table 6—Results of Classification Analysis* group's higher Becker score combined with their propensity to
Diagnostic Classification
produce fabulized combinations indicates that they maintained
Based on Cutoff Score slightly better quality control over their percepts per se, but they
Actual reasoned that percepts were related merely because they occurred
Diagnostic Acute in close proximity on the cards. The acute schizophrenic group's
Group Borderline Schizophrenic high Wa%, Adx + Hdx%, W + %, and DW% scores indicate that
Borderline 25 24 attentional instability may have been a prominent underlying
Acute problem at the time of testing; this feature was combined with
24 8 16 more variability in where they looked for meaning (from the
schizophrenic
amorphous figures to tiny detail areas). The borderline group
'The following variables were used in the analysis: W + %, Adx+ Hdx%, sought meaning in ordinary parts of the cards and assigned fairly
Wa%, DW%, FabC%, H%, Becker, and DDWC. See Table 1 for defini¬
realistic content, but they demonstrated reasoning errors by
tions.
permitting contiguity to indicate to them relatedness among the
adjoining percepts. In contrast, the acute schizophrenic group
groups' scores and those of the other groups. attended to or focused on fragments (Adx + Hdx% and DW%) or
In addition, inspection of Table 1 reveals that the other major only vaguely articulated a percept (Wa%). Further research could
summary variables (form quality [R + %], index of integration, and pursue these leads about differences between the two groups in
index of primitive thought [IPT]) also clearly separated the attentional controls and in forms of reasoning slippages.
normal and neurotic groups from the remaining groups. Although For example, our inspection of the DDWC scores for each group
comparison of the normal and neurotic groups with these other revealed that the normals and neurotics tended to begin each card
groups is of theoreticalinterest, we shall focus here on the with adequately realistic responses and terminated their associa¬
differences between the borderline, acute schizophrenic, and tions still on that level. The chronic schizophrenics began with
chronic schizophrenic subjects. Having distinguished the normal poorly fitting responses, tended to have their better responses in
and neurotic groups from the other groups, we went on to attempt the middle of their sequence of associations, and ended close to the
to make these more interesting clinical distinctions. low level of quality of their opening responses. The acute schizo¬
Group differences among borderline, acute schizophrenic, and phrenics tended to begin and end their series of associations on
chronic schizophrenic subjects were examined through planned adequate levels, inserting their ill-fitting and idiosyncratic
pairwise contrasts for all comparisons among the groups (Table 3). responses in between. However, the borderline persons tended to
To develop a score (or a set of scores) that could be used as a begin with an adequate response for each card but to end their
diagnostic tool in separating these three diagnostic groups, the series of responses with a very ill-fitting one. (Having to start

Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 03/04/2016

 Table 7.—Directional Comparison of Scores on
_W+%_Adx + Hdx%_Wa%_DW%
Acute schizophrenic (n =
24) High High High
Borderline (n =
25) Low Low Low
•See Table 1 for definitions of variables.
fresh on each card as it was presented seemed to refocus and
restructure the borderline person's efforts, and his best responses
came forth, only to gradually decline.)
Characteristics of Borderline Group
The borderline group's tendency to produce fabulized combina¬
tion responses and to show a decline in response quality on each
card are features consistent with the repeated clinical observation
that at least one large segment of the borderline group is
characterized by the presence of low-level thought disorder and
associative drift in their remarks.
As given in Table 8, the presence of two or more fabulized
combination responses characterized the records of nearly two
thirds of our borderline group. This agrees with the test literature
cited by Singer,1 Gunderson and Singer,2 Rapaport et al," and
Larson,* which contains many references to this type of response
as one that is characteristic of borderline persons. It might be
speculated that the third of our subjects whose records did not
contain such responses, yet who had been classed by clinicians as
borderline personalities on the basis of clinical materials but
without test data, resemble those borderline persons whom Rapa¬
port et al" termed "coarctated"; they comprise the more verbally
constrained but somewhat withdrawn subgroup or are the "stably
unstable"12'23 borderlines whose clinical manifestations center
more around instability.
COMMENT
The following three research questions were posed: (1) if
scores derived from the developmental-level scoring sys¬
tem could distinguish among Rorschach records of five
groups, (2) if fabulized combination responses distin¬
guished the borderline group from the other groups, and (3)
if a decline in developmental-level scores within the series
of responses given to each Rorschach card characterized
the borderline group in contrast to the other groups? For
the first question, the answer was that the major summary
scores clearly distinguished the normal and neurotic
groups from the borderline, acute schizophrenic, and
chronic schizophrenic groups. At a second stage in the
analysis, we compared the borderline, acute schizophrenic,
and chronic schizophrenic groups exclusively. A discrimi¬
nant analysis was performed using the 11 variables having
significant univariate t statistics. By this method, 80% of
the borderline, 63% of the acute schizophrenic, and 70% of
the chronic schizophrenic subjects were correctly classified.
A final analysis then compared only the borderline and
acute schizophrenic subjects. Eight of the 11 variables of
the preceding analysis were selected as the best discrimi¬
nators among these two groups.
Composite scores based
on these eight variables were
calculated. A cutoff that minimized total selection
score
error permitted correct classification of 96% of the border¬
line and 67% of the acute schizophrenic subjects. Three of
these eight variables (the DDWC, FabC%, and Becker
summary scores) were also significant univariate discrimi¬
nators. Thus, the acute schizophrenic and borderline groups
were successfully discriminated. It is of interest to the
clinician that the FabC%, DDWC, and Becker summary
scores, which can be easily calculated in daily practice, can
aid in the diagnosis of borderline persons. Also, these
findings suggest that further study of subscores in the
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 03/04/2016
Best Variables* for Discriminating Two Groups
FabC% H% Becker_DDWC
High Low High Low Little decline
Low High Low High Much decline
Table 8.—Fabulized Combination Responses
FabC% Responses*
No.(%) of Subjects
Group Total >1 >2
Borderline_25 93 22(83) 16(64)
Acute
schizophrenic 24 38 18(75) 8(33)
Chronic
schizophrenic 20 21 7(35) 4(20)
*FabC% is defined in detail by Becker5 and Goldfried et al."
developmental-level scoring system might be profitable.
The findings support the often-reported features that
clinicians have noted about borderline persons, namely,
that the role of structure in transactions appears to be a
central issue for the borderline person. Here, two of the
features, the presence of fabulized combination responses
and a decline in developmental-level scores within the
series of responses given to each card, characterized the
borderline group in contrast to the comparison groups.
These features reflect an inability to maintain a focus, to
reason realistically, and to perform consistently when
structure is not provided by the task itself or by another
person.
The chronic schizophrenics, when compared with the
borderline persons and acute schizophrenics, showed much
less adequate and integrated perceptual-cognitive func¬
tioning on most of the variables in the present study.
Becker,' Fine and Zimet,'-'4 and Zimet and Fine,'8 used
Rorschach developmental-level scores to compare chronic
and acute schizophrenic samples, but, to our knowledge, the
attempt to distinguish borderline persons from these
groups has not been presented heretofore. The earlier work
found "reactive" (acute) and "process" (chronic) schizo¬
phrenics to differ in their developmental-level scores. In
the present study, perhaps the most striking finding is the
similarity of the borderline and acute schizophrenic groups.
Both showed certain forms of "cognitive slippage" in the
Rorschach test situation. However, certain qualitative
differences are worth noting, especially the tendency for
the borderline person's associations to drift to a lesser
quality as he or she proceeded, whereas the acute schizo¬
phrenics tended to give a better-quality response subse¬
quent to giving a poor (ie, ill-fitting and unrealistic)
response.
The present report adds confirmation to the test litera¬
ture on borderline conditions, which contends that a major¬
ity of those classed by clinicians as borderline do show
certain cognitive slippage problems. It is consonant with
the work of Spitzer et al,12 thus indicating that schizotypal
thought processes are found within a segment of the
borderline group. Future research on borderline personali¬
ties should follow the lines suggested by Spitzer et al'2 and
the American Psychiatric Association Task Force for
DSM-IIP" using such criteria to assess the unstable and the
schizotypal features of these persons. In addition, using
 the Rorschach and the Wechsler procedures should permit
quantification of certain aspects of the schizotypal dimen¬
sion in borderline personalities.
This research was supported in part by Research Scientist Award 5K05
00175 (Dr Singer).
This work was done in collaboration with Lyman Wynne, PhD, and
Margaret Toohey, PhD, formerly of the National Institute of Mental
Health, Bethesda, Md, and now of the University of Rochester (NY) School
of Medicine and Dentistry.
References
1. Singer MT: The borderline diagnosis and psychological tests: Review
and research, in Hartocollis P (ed): Borderline Personality Disorders: The
Concept, the Syndrome, the Patient. New York, International Universities
Press Inc, 1977, pp 193-212.
2. Gunderson JG, Singer MT: Defining borderline patients: An overview.
Am J Psychiatry 132:1-9, 1975.
3. Friedman H: Perceptual regression in schizophrenia: An hypothesis
suggested by use of the Rorschach test. J Genet Psychol 81:63-98, 1952.
4. Friedman H: Perceptual regression in schizophrenia: An hypothesis
suggested by the use of the Rorschach test. J Project Techn 17:171-185,
1953.
5. Becker WC: A genetic approach to the interpretation and evaluation of
the process-reactive distinction in schizophrenia. J Abnorm Soc Psychol
53:229-236, 1956.
6. Rapaport D, Gill MM, Schafer R: Diagnostic Psychological Testing.
Chicago, Year Book Medical Publishers, 1946, vol 2, pp 329-366, 473-491.
7. Watkins JG, Stauffacher JC: An index of pathological thinking in the
Rorschach. J Project Techn 16:276-286, 1952.
8. Larson DG: The Borderline Syndrome in the Rorschach: A Comparison
With Acute and Chronic Schizophrenics, thesis. University of California,
Psychology Department, Berkeley, Calif, 1974.
9. Werner H: Comparative Psychology of Mental Development, rev ed.
Chicago, Follett Publishing Co, 1948, pp 213-336.
10. Werner H: The concept of development from a comparative and
organismic point of view, in Harris DB (ed): The Concept of Development.
Downloaded From: http://archpsyc.jamanetwork.com/ by a University of California - San Diego User on 03/04/2016
Minneapolis, University of Minnesota Press, 1957, pp 125-148.
11. Goldfried MF, Stricker G, Weiner IB: Rorschach Handbook of Clinical
and Research Applications. Englewood Cliffs, NJ, Prentice-Hall Inc, 1971,
pp 19-55, 288-305.
12. Spitzer RL, Endicott J, Gibbon M: Crossing the border into borderline
personality and borderline schizophrenia: The development of criteria. Arch
Gen Psychiatry 36:17-24, 1979.
13. Kringlen E: Heredity and Environment in the Functional Psychoses:
An Epidemiological-Clinical Twin Study. Oslo, Oslo University Press,
1967.
14. Wynne LC, Singer MT, Bartko JJ, et al: Schizophrenics and their
families: Research on parental communication, in Tanner JM (ed): Develop-
ments in Psychiatric Research. London, Hodder & Stroughton Ltd, 1977, pp
254-286.
15. Grinker R, Werble B, Drye RD: The Borderline Syndrome. New York,
Basic Books Inc, 1968, pp 73-181.
16. American Psychiatric Association: Diagnostic and Statistical Manual
of Mental Disorders, ed 3. Washington, DC, American Psychiatric Associa-
tion, 1980.
17. Vaillant GE: An historical review of the remitting schizophrenias. J
Nerv Ment Dis 138:48-56, 1964.
18. Vaillant GE: Prospective prediction of schizophrenic remission. Arch
Gen Psychiatry 11:509-518, 1964.
19. Wynne LC, Singer MT: Thought disorder and family relations of
schizophrenics: II. A classification of forms of thinking. Arch Gen Psychia-
try 9:199-206, 1963.
20. Phillips D: Case history data and prognosis in schizophrenia. J Nerv
Ment Dis 117:515-525, 1953.
21. Finn J: Multivariance: Univariate and Multivariate Analysis of
Variance, Covariance, and Regression, version 5. Ann Arbor, National
Educational Resources, 1972.
22. Scheff\l=e'\H: The Analysis of Variance. New York, John Wiley & Sons
Inc, 1959, pp 67-69.
23. Schmideberg M: The borderline patient, in Arieti S (ed): American
Handbook of Psychiatry. New York, Basic Books Inc, 1959, vol 1, pp
398-416.
24. Fine HJ, Zimet CN: Process-reactive schizophrenia and genetic levels
of perception. J Abnorm Soc Psychol 59:83-86, 1959.
25. Zimet CN, Fine HJ: Perceptual differentiation and two dimensions of
schizophrenia. J Nerv Ment Dis 129:435-441, 1959.