Defining Borderline Patients: An Overview

In 1953 Knight noted that the term “borderline” was applied to patients who could not be classified in other ways, i.e., as psychotic or neurotic, hence making it something of a wastebasket diagnosis (1). No doubt much of the dissatisfaction with recognizing such a category, whether it is termed borderline or any of its alternative labels, stems from the wish to keep schizophrenia as a clearly distinct disorder. Yet despite this objection, its use has steadily expanded.

Stern (2) was the first to use the term borderline, but the real parentage for this unwanted category is traceable to the “as-if” personality described by Deutsch (3), the ambulatory schizophrenia of Zilboorg (4), and the latent schizophrenia as introduced and developed by Rorschach (5), Bleuler (6), and Federn (7). Latent schizophrenia was sanctioned by Bleuler in 1911 to classify persons whose conventional social behavior he felt concealed underlying schizophrenia. Ambulatory schizophrenia was subsequently offered by Zilboorg in 1941 to combat the therapeutic nihilism that clinicians felt the latent schizophrenia label implied. Deutsch’s article on the “as-if” personality described persons whose superficial social appropriateness masqueraded highly disturbed personal relationships.

Before and even after Knight’s paper popularized the term borderline, many other names were suggested and then silently retired in favor of the ever-widening use of this term. Among these are preschizophrenia (8), schizophrenic character (9), abortive schizophrenia (10), pseudopsychopathic schizophrenia (11), psychotic character (12), subclinical schizophrenia (13), borderland (14, 15), and occult schizophrenia (16).

The most serious competition in the nomenclature has come from the term pseudoneurotic schizophrenia. This term was particularly popular in New York because of the local influence of its originators, Hoch and Polatin and associates (17, 18). Hoch urged replacing the term borderline with pseudoneurotic schizophrenia, which defined a specific psychopathological condition characterized by the combination of panneurosis, pananxiety, and pansexuality together with symptoms of schizophrenia. The broadening of the concept of schizophrenia which followed in New York may be responsible for a discrepancy in diagnostic habits between New Yorkers and other American psychiatrists as well as the British (19). Yet this term, too, has given way in the most recent APA diagnostic manual (20).

Although the use of the term borderline has become more common, disagreement over its definition has not subsided but has merely been displaced and camouflaged. Many who accept this term now disagree about whether it refers to borderline patient (21), state (22), personality organization (23, 24), character (25), pattern (26), schizophrenia (27, 28), condition (29), or syndrome (30). The increasing frequency of “borderline” patients (31–36), the greatly expanding literature, and the existing confusion about diagnosis and treatment make it imperative that some method for defining the patient group be devised that is replicable and from which research can proceed and conclusions can be generalized. In this paper we will attempt to survey the major relevant descriptive accounts of the borderline and to chronicle the common and discriminating features of these accounts.

Before we expand on these descriptive accounts, certain features about the literature itself should be noted. Since Knight’s pivotal article appeared, there has been a large and still expanding descriptive literature on the borderline patient. There were approximately 25 articles on borderline patients up to 1955, and that number has doubled in the past 10 years. Nevertheless, there remains a confusing overlap and discrepancy among authors in their descriptive attempts to define borderline patients. While most authors pay lip service to the previous literature, they proceed to describe borderline patients anew without noting how their descriptions add to or simply repeat earlier contributions. It is not clear whether this provincialism stems from an unfamiliarity or an unspoken dissatisfaction with the existing literature.

In this article we will discuss the major descriptive accounts of the borderline syndrome. Clearly, not all of the writings are of equal importance. The work of some authors is extremely well known and widely quoted, while others’ work is obscure and/or limited to a single publication. Certain writers attribute every imaginable trait to borderline patients, while others are quite selective in their descriptions. Furthermore, some authors expand or contract their definitions of borderline patients in later publications. In addition, some descriptive accounts are secondary or preliminary to a discussion of other issues, e.g., psychodynamics (37, 38), theory (27, 38), treatment (21, 31–33, 36, 37, 39–42), testing technique (43, 44), and schizophrenia (45). Yet in many the descriptive accounts are the primary goal of the paper (12, 17, 23, 29, 30, 46–49). Among the latter, only Grinker and associates (30) have undertaken a prospective and systematic collection of observations and data analysis.

Nevertheless, each of the authors cited has attempted to articulate his conceptualization of borderline patients and in so doing has implied either his agreement or disagreement with others. Taken in their entirety, these various views reflect the present clinical opinion about these patients. Any attempt by researchers to reliably define borderline patients should encompass these major clinical impressions. In this review, the most common and distinguishing characteristics will be identified, and a rational guide for standardizing clinical criteria for diagnosing borderline patients will be offered.

METHODOLOGICAL PROBLEMS

There are essentially three types of descriptive accounts: symptomatic and behavioral observations, psychodynamic formulations, and psychological test findings. Although these sources of descriptive data are parallel, they seldom seem touched or influenced by each other’s proximity. This independence, or ignoring of other sources, arises in part from the different contexts in which the observations are made and reported. Symptom and behavior observations tend to be published in the psychiatric literature and are the purview of clinical researchers and of those involved in residential treatment. Psychodynamic formulations frequent the psychoanalytic literature and usually are developed within iñdividual psychotherapeutic office practice. Finally, the psychological literature contains the accounts made of the borderline patient by psychologists who have administered controlled clinical testing procedures. In addition, the independence of these three groups no doubt grows out of a traditional suspicion each group holds for each other’s methodologies. In any event, the psychologist tends to focus upon intrapsychic structure, the psychoanalyst upon theory and therapy, and the general psychiatrist upon diagnosis, prognosis, and outcome.

Clearly, the amount of structure provided by a setting in which the borderline phenomenology is observed will influence how these patients are described. For example, clinicians using psychoanalytic techniques and psychologists using the Rorschach test agree in emphasizing the major ego defects and primitive intrapsychic mechanisms and thinking found in these patients. Yet clinicians observing these patients in structured hospital settings or evaluating them with structured interview techniques emphasize their stable personality features and interpersonal patterns. Certainly the broad agreement among authors from all persuasions (4, 21, 22, 27, 29, 31–33, 35, 37, 39–41, 43, 50, 51) about the borderline patient’s proclivity for regression in unstructured settings draws attention to the critical need to define the context in which the observations and descriptions of the borderline patient are being made. This propensity to regress when structure is low becomes an important and perhaps pathognomonic criterion for defining any sample of borderline persons.

The circumstances that lead the borderline person and his evaluator to meet are also important. For example, there appears to be a remarkable contrast between borderline outpatients voluntarily seeking treatment and borderline inpatients, who may be referred by others for treatment. A comparison of the observations made by Hoch and Cattell (17) with those made by Grinker and associates (30), the only authors whose descriptive data were collected in a systematic manner, sheds light on the influence of sample selection upon the conclusions reached about borderline patients.

Both research groups viewed schizophrenia as a distinct pathological condition and selected borderline patients who were free of overt schizophrenic symptomatology, such as clear-cut delusions and hallucinations, on the basis of their history and mental status examination. Hoch and Cattell further limited their sample to patients who presented mainly severe psychoneurotic symptoms, but who, on closer evaluation in psychoanalytic therapy and eventually in hospitals, revealed primary signs of schizophrenia in their thinking, feelings, and physiological functioning. Grinker and associates, on the other hand, selected patients on the basis of good functioning between hospitalizations and the presence of an ego-alien quality to any psychotic behavior. The diagnosis of borderline was made on outpatients, who were then hospitalized for participation in a prospective study. Thus Grinker and associates chose as their sample subjects with good premorbid features, and hospitalization was not a clinical necessity but rather for research purposes. It is not surprising, then, that their sample showed rare psychotic phenomena and developed virtually no schizophrenia during a five-year follow-up (52). In contrast, slightly more than one-fourth of Hoch and Cattell’s hospitalized population later developed manifest schizophrenia (53).

It is clear that the initial selection of samples influenced the conclusion that Grinker and associates reached that borderlines and schizophrenics have separate and distinct disorders while Hoch and Cattell concluded that their patients were a subgoup of schizophrenics. It is somewhat like packing a suitcase and then being surprised later to find what is in it when it is opened.¹ Most authors choose to regard borderline patients as a group somewhere between, and perhaps including, both of these extremes. Thus the need for hospitalization is a critical variable in comparing samples. Borderline patients who are referred for hospitalization because of severe symptoms would be expected to be more disordered than those functioning as outpatients.

In summary, we have cited four major variables to be considered in any descriptive account of borderline patients: who is describing them, the methods used to collect descriptive data, the context in which the patients were observed, and how the sample was selected. Each has an important impact on the resulting description of borderline patients.

What follows is a selective review of the literature covering three major descriptive conceptualizations of the borderline: first, the literature on symptoms and behavior; second, the psychological test literature; and third, the analytic literature as it views ego functioning.

SYMPTOMS AND BEHAVIOR

In the descriptive behavioral and symptom literature for borderlines, a number of characteristics are repeatedly mentioned that can be grouped under the general headings of affect, behavior, and psychosis. Among the studies that have considered the behavior and symptoms of borderlines, the study by Grinker and associates (30) deserves special citation as the only prospective, systematic one. Despite the slanted selection of patients in that sample, the findings must be considered as the marker against which all other reports should be measured. It is thus of special value to compare the findings of Grinker and associates with those other descriptions based on purely clinical impressions.

Affect

Of the four prevailing characteristics that Grinker and associates found in their borderline patients, two were qualities of affect.

1. Anger. “Expressed more or less directly to a variety of targets, anger seems to constitute the main or only affect that the borderline patient experiences” (30, p. 90). The expression of this anger—or the defenses against it—are a major discriminating feature used to identify four separate subgroups of borderline patients.

2. Depression. “Not the typical guilt-laden, self-accusatory, remorseful ‘end-of-the-rope’ type, but more a loneliness as the subjects realize their predicament of being unable to commit themselves in a world of transacting individuals” (30, p. 91). Grinker and associates pointed out that this depression was not present in their healthiest borderline group, which suffered instead from a clinging, childlike, anaclitic depression.

These conclusions by Grinker and associates are given substantial albeit inconsistent support by others. Some authors have noted the prevalence of anger but do not mention depression (21, 33), and vice versa (17, 36, 54). It seems likely that the confusion over the qualities of affect is traceable to at least two major sources. First, there is confusion about whether one is describing affects the patient presents with, affects that are covert, or affects that emerge in treatment. Second, as some authors have noted, the borderline patient’s anger and depression have peculiar qualities.

How these two sources of confusion influence the descriptions of borderline affects becomes apparent when one examines some of the statements about depression in the literature. For example, there are frequent qualifying phrases used by those clinicians who note a predominance of depression. Cary (29) has noted that the borderline patient is characterized by a “sense of futility and pervasive feelings of loneliness and isolation” that he says do not constitute a “true” depression. Hoch and Cattell (17) stated that they found frequent secondary depressions due to the persistent illness in their pseudoneurotic patients, but that primary depressions were infrequent. Kernberg (23) noted the prevalence of depressive-masochistic character traits in some of his borderline patients but differentiated these from depressive symptoms. Further, he advised that depression “as a symptom should not be used directly as an indicator of borderline personality organization” but suggested that only severe depression approaching psychosis in the form of “ego depersonalization” should be a presumptive indicator for a diagnosis of borderline personality. Gruenewald (55) commented that there is often a “covert” depression that emerges later. Chessick (56) described a chronic “existential despair” in borderline patients.

Anger is less controversial than depression. Many authors have noted a prevailing anger in borderline patients. This one feature seems to have been used progressively to discriminate borderlines from the original description of the “as if” personality (3, 30) and from “schizoid” personalities (29), where withdrawal from frustration is considered more characteristic. One author (57) felt that the borderline patient’s anger is so prominent that he suggested changing the name to “choleric.” Despite the apparent agreement about the prevalence of anger, such a broad range of behaviors is cited as being “angry” that a high degree of inference may be required. Some examples are hostility (50), rage reactions (31, 49), acting out (3), self-destructiveness (23, 31, 33, 58), detachment (59), mutism (33, 57), and demandingness (29, 57). Several authors have said that anger is not a presenting theme but one that emerges in the course of treatment (42, 50, 59). Kernberg (23) and Meza (57) speculated about excessive aggressive drive, while Modell (38) saw the anger as “mostly defensive.”

In contrast, several authors have regarded anxiety as the typical affect shown by their patients. Hoch and Cattell (17) gave this anxiety the status of a “defining secondary diagnostic symptom.” Although others (23, 32, 41, 47, 49) have also commented on anxiety in borderline patients, it is difficult in most instances to know if they are describing a symptomatic problem among borderlines or are making an inference based on a theoretical role given anxiety in personality theories.

Finally, another term frequently applied to borderline patients is anhedonia (32). In fact, there is considerable agreement that they lack a capacity for pleasure and rarely experience truly satisfied feelings. Their anhedonia has been discussed in terms of borderline dysphoria, unhappiness (47, 60), anguish (56), and lack of tenderness (3, 38).

In summary, the affective state of borderline patients is characterized by the prominence of anger and depression plus varying degrees of anxiety and anhedonia. If a generalization can be made, it is that these patients are not flat in their affect tone; they tend, in fact, to experience intense and variable affects, although this does not seem to include the experience of pleasure.

Behavior

Much of the literature on the treatment of borderline patients describes behavior during therapy, especially during hospitalization. Here we are interested in those behaviors which characterize borderline patients when they come for evaluation and that would therefore be used as criteria for making the diagnosis of borderline and in planning treatment. This is an important issue, since there may be a typical and highly distressing behavioral regression following hospitalization whose active prevention may be required from the start (21, 29, 31, 33, 35, 50, 61). Within the repertoire of hospital behaviors, such acts as window breaking, wrist slashing (31, 33, 58), and repeated overdosing (50, 62) emerge as quite specific to this kind of patient.

One historical factor, which led psychoanalysts to the conception of and interest in borderline patients as a distinct entity, was the discovery that many patients who by their histories and demeanors seemed relatively healthy underwent regressions on the couch. This disparity between good social behavior and poor intrapsychic structure has been mentioned repeatedly by both analysts (18, 48, 63, 64) and psychologists (28, 60, 65). What is meant by good social behavior seems to be good appearance and manners combined with superficial interpersonal relationships, and—more surprisingly—good functioning at work (12, 36, 42, 48, 60, 66, 67). The latter is surprising because this impression is noted almost in passing by many writers despite its apparent conflict with the behavioral record the patients have elsewhere in their lives. Schmideberg (47) took exception to this view. She described her borderline patients as marginal and transient in their work histories and cited their “sense of entitlement” as a source of their work problems. Grinker and associates summarized their follow-up data by saying, “Although gainfully employed and largely self-sufficient economically, the facts suggest that the group was occupationally and academically static at a fairly low level” (30, p. 132). Frosch (12, 66) noted that a borderline person may have a surprisingly stable work record when he is employed in a highly structured environment.

The characteristic most frequently and consistently ascribed to the behavior of borderline patients is that of impulsive and self-destructive acts. “Self-destructive” is used here to indicate a broad range of behaviors whose result is self-destructive although their intent or purpose is not. Examples include sexual promiscuity and perversions in search of affection (56, 68), self-mutilation with the goal of object manipulation (58, 61) or establishing self-identity (66), and addiction in search of escape (15, 32). Generally, borderline patients do not regard these behaviors as self-destructive, self-degradative, or guilt provoking. Although relatively little of the literature on borderline patients has dealt with actual suicide, repeated suicidal gestures and threats have been noted (50, 54, 57) and specific manipulative behavior attributed to such patients (69, 70).

Diverse sexual problems are attributed to borderline persons, but there is little agreement on their prevalence. Certain authors have noted a preoccupation with sex (17), and others have described polymorphous perverse sexuality (17, 23, 35, 46). Some authors (3, 12) have even included within the borderline category most persons with specific sexual deviances. Greenson (35) noted a “prominence of organ pleasures at the expense of object relations” among this group. Several authors (17, 41, 42) believe that these behaviors reflect a basic confusion in the sexual identity of the individuals.

The presence of obvious behavioral disturbances in a variety of spheres including drug use and sexual deviance often causes the borderline diagnosis to overlap with various character problem diagnoses in which chronic acting-out patterns such as antisocial, addictive, alcoholic, and homosexual behavior are seen. Because of this, Kernberg (71) has argued for a new classification of character types based upon what he believes are more fundamental personality features than behavior. He and many others have included a number of specific character disorders within the broader diagnostic category of borderline syndrome (21, 38, 41, 46, 47, 54, 68). Jan Frank (35) has suggested that various acting-out behaviors provide outlets for many persons now diagnosed as borderline who previously would have become overtly schizophrenic. He and others (71, 72) believe that inadequate impulse control is the dominant ego defect in these persons.

One concludes that in considering the behavioral evidence for the diagnosis of borderline, a clinician should not be deterred by the presence of a stable work history or superficial social adequacy but should examine other areas, where he may often find evidence of impulsive sexual, drug-taking, or other activities whose results are self-destructive even though the patient’s intent or purpose is not.

Psychosis

While there has been general agreement that borderline syndrome is a stable personality disorder (12, 23, 30, 33, 38, 39, 47), there is also widespread recognition that a number of these patients may develop psychotic symptoms (1, 17, 36, 39, 45, 46, 49, 51, 66, 73). Indeed, as indicated earlier, the borderline person’s capacity to develop regressive psychotic symptoms may be a pathognomonic feature. Weiner would seem to concur with this conclusion in his review of the literature (28). However, there is a consensus that when psychoses do occur, they have the following differential features: l) stress related (21, 23, 36, 43, 49, 68), 2) reversible (21, 30, 41, 49), 3) transient (12, 15, 30, 49), 4) ego-alien (1, 12, 30), and 5) unsystematized (23, 29, 30, 43, 45, 60, 74). Numerous authors have used some or all of these features to differentiate borderline psychoses from the psychoses of schizophrenia and other disorders (21, 23, 29, 30, 38, 45, 54, 66). Thus there is general agreement as to absence of stable or clear delusions or hallucinations, with only a few dissenting opinions (17, 33). Some authors have postulated that the borderline’s psychoses occur in response to intolerable rage (29, 30, 41, 73).

Interestingly, despite the consensus about the vulnerability of certain borderline persons to psychotic-like episodes and regressions, only a few authors have viewed this as an essential feature of the borderline syndrome (1, 12, 45, 73, 75). Most authors have hastened to note that the occurrence of psychoses is the exception, not the rule (23, 29–31, 39, 46, 76). A few have taken a determined position that psychoses do not occur at all in borderline persons (33, 50, 63). At the opposite extreme, Hoch and Cattell (53) found that the psychoses in their sample of pseudoneurotic schizophrenic patients were not necessarily transient and reversible. A more recent report (77) suggested that the psychoses of this group cannot be differentiated from those of schizophrenics, and another (46) reported that they are brief and rare. This last report would thereby place pseudoneurotic schizophrenia within the mainstream of thought about psychoses among borderline persons.

There have also been widespread references to the similarity between the psychotic thought processes of borderline and schizophrenic persons. Some borderline persons demonstrate fears of being controlled (17), ideas of reference (38), externalization (1, 56), and other paranoid tendencies (12, 23, 56). Some writers have noted that they have vividly loose associations and other symptoms of formal thought disorder. For example, Knight (1) stated that loose associations can be detected by use of the Rorschach test and free-associative interviews. On the other hand, Grinker and associates (30) emphasized that they found no looseness of associations. (However, they did not use projective tests or free-associative interviews in their assessment of their patients.) Hoch and Cattell (17) took an intermediate position, stating that “approximate” or “parallel” associations are frequent. Thus there is little agreement among clinicians about the presence of thought disturbances in borderline persons. Some say there are none (23, 30), while others say there are many (1, 47, 50, 65, 73). It is clear that these differences result from problems in defining and assessing thought disorder as well as from the methodological problems cited earlier.

Some authors have underscored the frequent occurrence of disturbed states of consciousness. These peculiar ego states, which were first described by Deutsch (3), have been variously categorized as depersonalization, dissociation, and derealization (23, 29, 36, 45, 61, 66). They have been called borderline “states,” to be differentiated from borderline “personalities.” These states are seen as responses to anxiety (66), depression (23), and rage (29) and as prepsychotic experiences (3).

Separate from the purely clinical literature already summarized, a literature grew in the 1960s that recognized a vulnerability among borderline persons to psychoses when exposed to pharmacological stress, namely, that produced by marijuana (78), LSD (68, 79), and mescaline (68). These reports suggest that the borderline person is unique in his sensitivity to pharmacological stress. This special sensitivity or vulnerability seems to support Schmideberg’s often quoted characterization of the borderline personality as “stably unstable” (47).

From the many foregoing clinical reports, which vary in their positions on whether psychoses occur among the borderline group, a series of clinical questions arise: Do all borderline persons have a vulnerability to psychosis even if they are not psychotic when assessed? Are most borderline persons free from psychosis throughout their life despite such a vulnerability? Clinicians are uncertain and divided over these issues. Grinker and associates (30) contended that psychoses occur in only one subgroup of borderline persons. Equally important, they are the only authors who have attempted to identify subgroups of borderlines who they contend will not develop psychoses. Other authors have implied or hinted that many or even most borderline persons could develop psychoses, given properly stressful circumstances. Parallel issues concerning the extent and type of their reality testing and the nature and type of their cognitive style will be considered in a later section of this paper.

To conclude, reports in the clinical literature generally agree that an undetermined number of borderline persons do develop psychoses in stressful situations. Moreover, when such psychoses occur, they are characterized by their limited symptoms and limited duration. There are, however, few actuarial data on the prevalence of vulnerability to psychosis among borderline persons. Some authors have suggested that dissociative states may be quite marked in these patients.

INTRAPSYCHIC PHENOMENA

The psychological test literature devoted to characterizing borderline patients has generally been in agreement that they demonstrate ordinary reasoning and responses to structured tests such as the Wechsler Adult lntelligence Scale (WAIS), but that less structured tests such as the Rorschach reveal deviant thought and communication processes (27, 28, 43, 60, 74, 80–83). As in the clinical literature, most articles on which this conclusion is based are impressionistic. Their many methodological issues require attention before this broadly held viewpoint is accepted (6). The seminal contributions were made by Rorschach, Rapaport and associates, and Schafer, with subsequent authors generally being content to add confirmatory evidence.

Rorschach (5) and later Rapaport and associates (8) used the terms “fabulizing, combinatory, and confabulated” thinking to describe the propensities of borderline persons to overspecify secondary elaborations of their associations and to combine and reason oddly. They are prone to reason circumstantially rather than logically. Their separate perceptions tend to become intermingled and related simply because they occur close together in time or space. Borderline persons read more affective elaboration into their perceptions than others can validate, i.e., they tend to add too much and too specific affective material to simple perceptions. Other persons then have trouble accepting this affective meaning, although they might accept the same basic percept that the borderline person sees.

Rorschach, in 1921 (5), was the first to call attention to some seemingly adequately functioning persons whose responses to inkblot tests resembled those given by schizophrenic patients. He applied the term latent schizophrenic to those persons who had average surface behavior but Rorschach test features in common with schizophrenics, such as self-references, belief in the reality of the cards, scattered attention, variability in quality of ideas, and absurd and abstract associations.

It was Rapaport and associates (8) who first described the borderline person’s intact performance on the WAIS and a pervasively odd Rorschach record; this has subsequently become an almost axiomatic diagnostic rule for later writers presenting single case studies. Although these authors defined two groups of what they called preschizophrenics, namely, the coarctated group and the overideational group, it has been largely the overideational borderline patient who is referred for psychological testing and upon whom subsequent literature has concentrated. Stone and Dellis (74) reiterated Rorschach’s observation about the disparity between social functioning and thinking. They went on to confirm in a prospective study (74) Rapaport and associates’ finding of a discrepancy between the Rorschach test and the WAIS in their evidence for pathology in borderline patients.

Schafer (82) introduced a third distinguishing feature about the disturbed thinking of borderline patients when he suggested that they are more comfortable about their bizarre and distorted thinking than are schizophrenics. Although Schafer’s finding seems to differ from the general impression in the clinical literature that psychoses are ego-dystonic, it was later repeated in a report by Fisher (60).

In an interesting report DeSlullitel and Sorribas (84) compared the Rorschach test results of normal subjects, borderline persons, and creative artists. They found that the negative unpleasant content within “fabulized combination” responses by borderline persons distinguished them from the creative artists, who presented positive content within similar types of responses.

Gruenewald (55) reported on psychological test batteries given to 10 of Grinker and associates’ original 51 borderlines five or more years after their discharge. She noted that based solely on test results, 2 would have been diagnosed as schizophrenic. However, when these data were combined with other information, the results were consistent with the borderline diagnosis and fit within the subgroups Grinker and associates had derived. She noted that “maladaptive primary process manifestations” were sometimes present in thought content and organization. Unfortunately, she made no mention of any discrepant functioning on individual tests.

In summary, borderline persons are believed to connect unrelated percepts illogically, overelaborate on the affective meaning of percepts. and give circumstantial and unpleasant associations to the Rorschach inkblots. This disturbed thinking may be more flamboyant and more ego-syntonic than that found among schizophrenic persons. Yet such borderline persons are reported to function adequately on the WAIS, showing few or none of the ideational deviances.

EGO FUNCTIONS

There are various ways to assess and classify ego functions. Thus it is difficult to select from the literature comparable descriptions of specific ego-function characteristics of borderline persons. However, two functions do emerge as particularly relevant to this group, namely, reality testing and interpersonal relationships. While the latter remains almost solely within the purview of clinical impressions, several approaches to evaluating reality testing have been made.

Reality Testing

Any discussion of psychosis is based upon the concept of reality testing. Frosch (12, 40) has stated that an intactness in reality testing differentiates borderline persons from schizophrenics. Distinguishing among reality testing, sense of reality, and relationship to reality, he concluded that borderline and psychotic persons share a poor sense of reality and relationship to reality, but that the borderline person can test out his experiences whereas the psychotic cannot. In a panel report on the “as-if” personality, a similar conclusion was reached about this subgroup of borderline persons (85). This position is akin to the frequent comments of several authors that borderline persons, when compared to schizophrenics, have more distance from their psychotic experiences and regard the episodes as ego-alien or ego-dystonic. Zetzel (21) added a twist to Frosch’s viewpoint by stating that it is the capacity to reverse impaired reality testing, given a good situation (in treatment), that distinguishes borderline persons from schizophrenics.

A number of authors seem to agree with Frosch that reality testing is generally maintained in borderline persons (23, 30, 85). Wolberg (85) has stated that the reality distortions that do occur are defensive in nature and that the borderline person’s actual perception of reality is always extant. However, Kernberg (51) has noted that “under special circumstances—severe stress, regression induced by alcohol or drugs, or transference psychosis”–they may lose this capacity. Kernberg and others have noted that these patients are prone to develop psychotic transference reactions. Of course, unless this feature is cited in the patient’s history, it would not be of use in an initial diagnostic evaluation. Authors using psychological tests have assigned relatively greater importance to the vulnerability of the borderline person’s reality testing to stress (31, 43, 60). Knight (1) noted the borderline person’s inability to distinguish between dreaming experiences and reality. Hoch and Cattell (53) later drew a similar conclusion about the impairment in reality testing in their group of pseudoneurotic schizophrenic patients.

A number of authors have emphasized that reality testing should not be viewed as a phenomenon that one has or does not have but, rather, that there is a spectrum of reality testing (1, 16, 38, 50). Modell (38) and others (21, 23) have discussed the borderline person’s reality testing problems in the context of self-object differentiation. Modell noted,

Such a graded view of reality testing helps to reconcile certain discrepancies noted earlier among authors who have presented contradictory views of reality testing in borderline persons.

As Hurvich (86) has pointed out, there is a need for quantifiable measurements of this ego function, which could be used to evaluate whether and to what extent impairment exists. Although Grinker and associates (30) included “relation to reality” as one of the seven ego functions they intended to evaluate in their borderline patients, they did not include any instruments to measure this directly. Their evaluation was based on behavior that was generally labeled as “positive” or “negative” and upon ratings of perception devised from global judgments of the patient’s awareness of self, others, time, events, and things. These assessments did not give any consideration to latent vulnerability to disruptions in reality testing. Further, their validity as reflections of reality testing is not always obvious. Grinker and associates seemed aware of these problems and did not draw any definite conclusions about the relationship to reality and the capacity for reality testing in their sample. Until instruments or methods of measuring reality testing are developed, the literature on borderline persons will continue to reflect the ambiguity of the concept, and clinical impressions will be subject to the unstated and therefore unknown biases of the researchers.

At this time the consensus is that there is a definite impairment of reality testing in borderline persons, but it is not as severe as that in psychotic persons and, under most circumstances, is not apparent.

Interpersonal Relationships

A number of authors have pointed to borderline patients’ style of relatedness as the most distinguishing diagnostic feature of this group. Zetzel (21) said that for the borderline patient in particular, “the kind of doctor-patient relationship that is established may prove to be a crucial factor in reaching a definitive diagnostic evaluation.” This relationship is best illustrated by the following observations.

A frequently cited feature of the borderline patient’s object relationships is a predictable superficiality and transiency (30, 32, 60, 87). Fisher (60) suggested that “superficiality in relationships” distinguishes borderline patients as a group from neurotic patients. Knight alluded to this somewhat differently by noting, “Other ego functions, such as conventional (but superficial) adaptation to the environment and superficial maintenance of object relationships, may exhibit varying degrees of intactness” (1, p. 6). This echoes the observation made earlier about the borderline person’s surprising capacity for adequate social functioning but adds that this apparent behavioral normality may depend on superficiality. Dyrud (32) commented that case material cited by Grinker and others demonstrates more than anything else the remarkably short sequences of interactive behavior that these patients are capable of maintaining.

This quality of superficiality and transiency is supported and explained in some measure by the original formulations of the “as-if” personality by Deutsch (3). Deutsch described the borderline person’s interpersonal relationships as “plastic” and “mimicry.” She stated that the “essential characteristic is that outwardly he conducts his life as if he possessed a complete and sensitive emotional capacity.” Eventually, the absence of real emotional responsiveness leads to repeated dissolution of relationships. This disparity between adequate superficial relatedness and inadequate internal relatedness has been used to characterize borderline patients more generally (29, 30, 48–50, 60, 80). It may be attributable to a lack of coherent self-identity, which Grinker and associates (30) found in their sample and which other authors have also noted (1, 21, 23, 45, 49, 56, 66).

In contrast to the recurrent theme that borderline persons’ interpersonal relationships are superficial and transient is the claim that they are prone to form intense, clinging relationships. Adler (31) stated that their “readiness to form rapid, intense, engulfing relationships” is what differentiates them from schizoid and schizophrenic persons. Similarly, Modell (38) stated, “My principal reason for considering this group homogeneous is that they develop a consistent and primitive form of object relationships in the transference.” Grinker and associates (30) referred to this when they cited as one of the four identifying characteristics of borderlines “a defect in affectional relationships. These are anaclitic, dependent, or complementary, but rarely reciprocal.” These accounts of borderline persons’ intense relationships are underscored by the consensus that their initial relationships with therapists are dependent and demanding (12, 21, 23, 29, 31, 38, 50, 76). In addition, many authors have noted that the intense therapeutic relationship is characteristically devaluative and manipulative (21, 31, 38, 50, 57). The emergence of such angry behaviors may in turn lead to the repeated disruption of such close relationships (49, 50, 57). These qualities of the borderline person’s close relationships may not be immediately discernible. Houck (50) pointed out that this initial deception can lead to later problems for unsuspecting therapists.

Thus in their everyday relationships borderline persons relate in a fairly normal but superficial and transient manner, while in their close relationships they become intense, dependent, and manipulative. In any event, these individuals are actively involved with other people and are not particularly socially withdrawn.

SUMMARY

In this paper we have surveyed the large literature of descriptive accounts of borderline patients. Within the major variations in these accounts we have attempted to identify certain themes and prevailing clinical impressions. We have discussed the four methodological issues that significantly influence the resulting descriptive accounts. Taking these methodological issues into account, we have identified a number of features that most of the authors believe seem to characterize most borderline persons. These features are as follows:

1. The presence of intense affect. It is usually of a strongly hostile or depressed nature. The absence of flatness and pleasure and the presence of depersonalization may be useful in differential diagnosis.

2. A history of impulsive behavior. This may take many forms, including both episodic acts (e.g., self-mutilation, overdose of drugs) and more chronic behavior patterns (e.g., drug dependency, promiscuity). Often the result of these behaviors is self-destructive although their purpose is not.

3. Social adaptiveness. This may be manifested as good achievement in school or work, appropriate appearance and manners, and strong social awareness. However, this apparent strength may reflect a disturbed identity masked by mimicry, a form of rapid and superficial identification with others.

4. Brief psychotic experiences. These are likely to have a paranoid quality. It is felt that this potential is present even in the absence of such experiences. The psychoses may become evident during drug use or in unstructured situations and relationships.

5. Psychological testing performance. Borderline persons give bizarre, dereistic, illogical, or primitive responses on unstructured tests such as the Rorschach, but not on more structured tests such as the WAIS.

6. Interpersonal relationships. Characteristically, these vacillate between transient, superficial relationships and intense, dependent relationships that are marred by devaluation, manipulation, and demandingness.

These six features provide a rational basis for diagnosing borderline patients. The criteria can be readily assessed during an initial evaluation. Further research is under way to evaluate the relative frequency and discriminating value of each of these features. From these studies a reliable system of diagnostic criteria can develop. It is hoped that such prestated and standardized means of identifying borderline patients will permit better treatment planning and clinical research on these patients to proceed.