1-6 Drawing the borderline: Diagnostic ambiguity, from Freud to DSM-5

1-6 Drawing the borderline: Diagnostic ambiguity, from Freud to DSM-5

No Comments

Celeste Pietrusza


Though the term “borderline” was once used to describe patients who, in a psychoanalytic context, could be described characterologically as being “in between” neurosis and psychosis, its current mainstream usage in the DSM-5 diagnosis of borderline personality disorder bears almost none of its original theoretical underpinnings.  Different approaches to diagnosis have, since the 1970s, sought to provide more nuanced detail to the category and concept of the borderline, however, many clinicians have also expressed concerns that “borderline” has become a “junk category” lacking in diagnostic precision.  In this paper, I look at the ways in which “borderlines” are perceived clinically and contextualize these perceptions within what I will call the “aetiology of a diagnosis.”  From this, I will argue that the descriptor “borderline,” whether describing a personality disorder, a subclinical organization, or particular clinical features, is a term that, as the word suggests, contains rather than merely classifies, and using this, move toward contemporary critiques as ways of rethinking these borders.


Stigma and the borderline

     “The beginning of wisdom,” George Valliant (1992) writes, “is never calling a patient borderline” (p. 117).  Reviewing the literature on borderline phenomena, a clinician can easily understand why Valliant would make such a claim.  Overall, the language used to describe “borderline” patients is either implicitly or explicitly pejorative.  Descriptions of patients with borderline personality organization range from psychodynamic understandings of individuals with “immature” (Valliant, 1992) or “primitive” (McWilliams, 1994) defenses who invoke “countertransference hatred” (Maltsberger & Buie, 1974) to complaints from hospital workers of patients who are “attention-seeking,” “a nuisance,” or “wasting time” (Arnold, 1995).  Other

characterizations view such patients as “fickle, egocentric, irresponsible” (Klein, 1977) individuals who “use hospitals to escape from responsibilities” (Houck, 1972).  Many articles describe patients whom clinicians describe as borderline as being “difficult” (Valliant, 1992) and a large-scale review of the literature on borderline personality disorder found that those with the diagnosis were routinely viewed as patients with “pernicious motivations” whom “few clinicians want to treat” (Gunderson, 2009).

Yet despite what Gunderson (2009) sees as clinicians’ recognition, in their descriptions of borderline phenomena, of a “stable instability,” precisely how this instability manifests itself is a question of much debate.  Valliant (1992) relates an informal survey he conducts yearly of his psychiatry residents in which he asks them to describe what they mean when they characterize a patient as “borderline”:  Almost unfailingly, he says, he gets a different response from each resident.  Taking the consistency of clinicians’ affective responses to such patients along with the inconsistency of their intellectual responses into consideration, Fromm (n.d.) speculates that the category of the borderline might not, in fact, be describing something within the patient, but rather an experience that arises between the patient and another—and, in the clinical situation, this other is the analyst or therapist.  Fromm suggests that the word borderline might, instead, describe the sensation a clinician has of meeting a certain type of patient, as if he or she were “on a frontier of some sort.”  What this boundary or frontier might be, however, has been hotly contested since the label’s appearance within psychoanalytic parlance, leading Spitzer and others in the American Psychiatric Task Force on Nomenclature and Statistics to claim that “[s]ome believe[d] that the borderline concept represents everything that is wrong with American psychiatry…[while] others believed that there is sufficient evidence of the utility” (Spitzer, Endicott & Gibbon, p. 17).


Psychoanalytic frontiers:  The proto-history of a diagnosis

The notion of the borderline as frontier reaches back to the early days of psychoanalysis with Freud’s 1924 paper “The Loss of Reality in Neurosis and Psychosis.”  While Freud did not use the term “borderline” in this paper, he wrote of the differences between ego functions in neurotic versus psychotic states and noted instances in which diagnostic clarity was uncertain.  Psychoanalytic practitioners following Freud noticed that this ambiguity in clinical presentation manifested itself most acutely in the unstructured situation of psychoanalysis.  In 1938 Stern described a tendency for some patients to have negative reactions to analysis and regress into what he called “borderline schizophrenia” characterized by low self-esteem and infantile defenses.  Though adumbrated by Emil Kraeplin’s early descriptions of “borderland” cases of schizophrenia, Stern’s description was most likely the earliest progenitor to a contemporary conceptualization what clinicians would now call borderline personality disorder (Dawson & Macmillan, 1993).

While the term “borderline” remained, for three decades after Stern’s writings, a “rarely and inconsistently used colloquialism within the psychoanalytic fraternity” (Gunderson, 2009, p. 2), various psychoanalytic writers and clinicians worked on concepts that would later prove important to the understanding and treatment of patients whose symptoms would later fall under the borderline rubric.  In the 1940s, Helene Deutsch (1942) described the “as if” personality as belonging to a person who was “socially appropriate,” yet plagued by disturbed relationships and a lack of identity.  The work of Donald Winncott in the 1960s, while not directly related to borderline phenomena, provided a developmental framework for later understandings of the aetiology of borderline personality disorder.  Madell (1963) would utilize Winnicott’s notion of the analytic space as akin to a maternal “holding environment” and the analyst as a “transitional object” that helped ease children’s processes of separation from their mothers in his work the following decade with patients described as borderline.

In 1967, Otto Kernberg began his highly influential lifelong work in developing the concept of what he called “borderline personality organization” as a distinct category within psychoanalysis.  In the United States at this time only patients considered to be firmly “neurotically-structured” were considered suitable for analysis.  Freudian ego psychologists in the wake of Anna Freud considered individuals on the borderline to be “unanalyzable” along with psychotics.  Whereas previously, psychoanalytic treatment with individuals with borderline characteristics was considered anti-productive, Kernberg proposed changes to the analytic process that could allow these patients to be treated with psychoanalytic psychotherapy (Gunderson, 2009). Using a post-object relations psychoanalytic framework highly influenced by the work of Melanie Klein, Ronald Fairbairn, and Winnicott, Kernberg added specificity to the borderline nomenclature and brought to it new constructs.  Kernberg characterized borderline personality organization as characterized by weak identity formation, “primitive” defenses such as splitting and projective identification, and lapses in reality testing under stressful conditions.  Kernberg (1968) would later describe those with borderline personality organization as experiencing feelings of emptiness arising out of insufficiencies of early childhood as well as a concomitant inability to self-soothe.  Along with Masterson (1972), however, Kernberg emphasized the possibilities for treating borderline patients with long-term therapy—and to this day, despite the emphasis on the pathology of borderline patients, research continues to show that given proper therapeutic treatment, they have the best prognoses among individuals with personality disorders.  As such, Kernberg’s analyses and understanding of the borderline pushed at the frontiers of American psychoanalysis and contributed to its widening scope beyond classical psychoanalysis into more contemporary psychodynamic applications.


Defining the perimeters:  Descriptive categorizations of the borderline

While psychoanalytic inquiries into borderline patients floundered between Stern’s observations in the 1930s and Kernberg’s work in the 1970s, descriptive psychologists and psychiatrists working with psychoanalysts in hospital settings looked for alternate ways to classify these patients.  In the 1950s, Robert Knight, noted, like Stern, that there were a subgroup of patients who did not fit the typical profiles of those diagnosed with neurotic or psychotic disorders, and were susceptible to developmental regression.  Observing inpatient units, he noticed how the failure to adequately provide for these patients lead to difficulties in the units as well as disagreements among staff members (Gunderson, 2008).  In his classificatory system, Knight separated the term “borderline” from schizophrenia and instead spoke of “borderline states.”  Later that decade, Schmideberg would add observations about enduring, life-long recurring patterns of borderline states in particular individuals, thus paving the way for the discussion of the “borderline” as a personality disorder come DSM-III (Mack, 1975).

In the 1960s, Roy Grinker emerged as an advocate for empirical validation of borderline diagnoses.  His co-written book The Borderline Syndrome:  A Behavioral Study of Ego Functions (Grinker, Werble & Drye, 1968) criticized psychiatry’s lack of empirical research on the concept and conducted a cluster analysis on patients described as borderline.  Via cluster analysis, the authors divided borderline patients into four categories:  A severely disturbed group that bordered on the psychotic, a core group characterized by negative affect and acting-out behaviors, “as if” persons (following Helene Deutsch) with a poor sense of self, and, finally, less severely disturbed individuals with neurotic features.  Using an ego psychology framework, they found that the commonalities between the patients seemed to be:  the predominance of anger in the realm of affect, issues with interpersonal relationships, a lack of self-identity, and depressive or anaclitic loneliness.  Grinker’s description of the “borderline syndrome” would later gone on to serve as a framework for structuring the criteria for the borderline personality disorder diagnosis in DSM-III (Spitzer, Endicott, & Gibbon, 1979).

After Grinker’s work, research into borderline constructs expanded tremendously throughout the 1970s.  By 1975, as Gunderson and Singer note, the adjective “borderline” could apply or refer to a personality organization, a character type, a pattern of behavior, a subset of schizophrenia, a condition, or a syndrome.  In an analysis of diagnostic criteria for borderline personality disorder in 1978, researchers Perry and Klerman wrote, “In one way or another, it seems as if the whole range of psychopathology of personality is represented” (as cited in Dawson & Macmillan, 1993, p. 150).  Often, a borderline label for a patient arose not from the individual’s presenting pathology, but instead, a treatment response:  Patients would be diagnosed as borderline due to chaotic relationships with hospital or psychological staff, “bizarre” performances on psychological testing or after reacting to treatment with intense affect or impulsive behavior (Dawson & Macmillan, 1993).  In the hopes of better identifying individuals presenting with borderline personality disorder and systematizing diagnosis, in conjunction with the publication of DSM-III, a structured diagnostic interview for borderline patients was finally developed (Gunderson, 2008).


Moving to the center:  BPD in DSM-III through DSM-5

Despite the inclusion of borderline personality disorder in the DSM-III in 1980, concerns still abounded regarding the lack of diagnostic clarity around the disorder.  As biological psychiatry was, at this time, just coming to the fore, Gunderson (2009) notes that “many felt that borderline personality disorder had been included in the DSM-III simply as a conciliatory gesture intended to placate the psychoanalytic plurality” (p. 3).  Nonetheless, as with other DSM disorders, its inclusion in the diagnostic manual fueled a proliferation of research from neurobiological, pharmacological, and developmental researchers.  Whereas before 1980, only fifteen research reports had been published on borderline personality disorder, over 275 articles were published on it in the decade from 1980 through 1990 (Gunderson, 2009).  While the research findings have been as diverse and heterogeneous as the criteria for the disorder itself, Gunderson (2009) argues that they have nonetheless tended to validate the integrity and clinical utility of the diagnosis.

Today, the notion of the disorder being on the “borderline” between neurosis and psychosis has been dropped, even in psychoanalytic parlance, as the course of the disorder has been established as distinctive from both those of psychotic and mood disorders (Gunderson, 2008).  Between DSM-III and DSM-IV, research instead focused significantly on the role of developmental factors in the disorder’s genesis.  Following Masterson’s (1972) psychoanalytic observations regarding inconsistent mothering in early childhood as a major contributing risk factor, attachment theorists such as Peter Fonagy began to look at how parental “mismatches” in empathetic responses could contribute to the later expression of borderline phenomena (Fonagy, Steele, Steele, Moran & Higgitt, 1991).  From variations on Ainsworth’s pivotal Strange Situation, experiment, researchers observed what they identified as a distinct fourth category of attachment—disorganized/disoriented attachment—that they found to be correlated with borderline personality features in later life (Main & Solomon, 1990).  Developmental factors such as unpredictable and high conflict family environments (Gunderson & Zanarini, 1989) as well as sexual abuse and other childhood traumas (Gunderson & Sabo, 1993) were also found to increase the probability that certain individuals would develop the disorder, leading renowned trauma researcher Judith Herman (1992) to claim that borderline personality presentation might in fact hide a more complex, underlying post-traumatic stress disorder.

While the DSM-IV for borderline personality disorder changed only modestly from DSM-III—for instance, the identity disturbance criterion was modified and “boredom” was removed from the diagnostic criteria as it was shown to be more characteristic of narcissistic personality disorder (American Psychiatric Association, 1980; American Psychiatric Association, 2000)—DSM-5 brings, with its new trait-based classification system for personality disorders, significant changes.  First, the general definition of a personality disorder has changed:  Instead of a pervasive pattern of thought, emotion and behaviors, a personality disorder instead now reflects “adaptive failure” involving “an impaired sense of self-identity” of “failure to develop effective interpersonal functioning.”  The essential features of a personality disorder are impairments in self and interpersonal functioning as well as the presence of pathological personality traits.  In an article supporting many of the then proposed DSM-5 revisions, Kernberg (2012) notes that while the inclusion of impaired self and interpersonal functioning to the definition of a personality disorder may seem like “common sense,” the “fact that this dimension could be explicated, operationalised, and clinically evaluated in terms of the degree of the disturbance” constitutes a “major innovative contribution and strength” that he believes will bring empirical dimensionalists and psychodynamic psychotherapists together around a shared construct (p. 235-236).

The two criteria related to self-pathology in the DSM-5 diagnosis of borderline personality disorder are identity instability (including excessive self-criticism, chronic emptiness, and dissociative states) and chaotic goal-setting.  Interpersonal disturbances for the disorder involve failures of empathy (hypersensitivity or negative bias) as well as problems with intimacy.  Pathological personality traits for borderline personality disorder may be in the following domains:  Negative affect (anxiousness, separation insecurity, and/or depressivity), disinhibition (impulsivity and risk-taking), and antagonism characterized by hostility.  Unlike in DSM-IV, to meet criteria for the diagnosis, the individual does not need to experience clinically significant distress or impairment—instead, it is more important, diagnostically, that the personality traits and function deviate from “the individual’s developmental stage or socio-cultural environment” (American Psychiatric Association, 2000; American Psychiatric Association, 2013).


Beyond the borderlines:  Debates in diagnosis and treatment

The notion of a personality disorder as a particular deviation from a socio-cultural norm is not a new one—the criterion was included, albeit via different wording, in the DSM-IV-TR (American Psychiatric Association, 2000).  Additionally, discussions about the social construction of borderline personality disorder date as far back as Grinker’s work in the 1960s.  Grinker suggested that borderline psychopathology might be becoming more noticeable and prevalent not merely due to physiological or even developmental reasons, but rather, social ones.  He cited the paradoxical burden placed on individuals by the social changes of the mid-20th century—ones that valued autonomy over collectivity, individual identity over collaboration.  Families became smaller and more isolated, many burdens of manual labor had been lifted from American workers and people of different class backgrounds had more time available than before for leisure, travel and communication.  These changes, Grinker hypothesized, necessitated a different skill set and sense of identity development not required of or even encouraged in many people before this time period (Gunderson, 2008).  There is discussion as to whether or not or how the expression of borderline features might vary cross-culturally and if the criteria looking at impulsivity and risk-taking might be more specific to the United States.  As an example of this, after the publication of DSM-III, researchers in the United Kingdom looked at inpatient units and, after not seeing similar behaviors as described by American psychiatrists, published an article in the Archives of General Psychiatry entitled “Are there borderlines in Britain?” that compared the DSM-III and ICD-9 diagnoses (Kroll, Carey, Sines & Sir, 1982).

Feminist critics have argued that the borderline personality disorder diagnosis is overused with female patients, particularly ones with histories of sexual abuse, and, as such, unduly pathologizes survivors (Shaw & Proctor, 2005).  Shaw & Proctor (2005) cite research in which 80% of people diagnosed with borderline personality disorder were found to have experienced some type of abuse.  From this and other research, they consider the ways that rather than being “adaptive failures” as described by DSM-5, the personality aspects of borderline personality disorder might be “better understood as adaptive reactions to early relational traumas” (p. 486).  Likewise, Judith Herman (1992), in her work on complex post-traumatic stress disorder, has suggested that a diagnosis for “borderline” patients that highlights the role of trauma rather than locating pathology as inherent to the individual could work to make significant changes in and improvements to mental health professionals’ perceptions and treatment of these patients.

Other researchers have argued that the term “borderline” should be dropped altogether, as it no longer serves a descriptive purpose with regard to diagnosis and treatment.  Gunderson (2010) suggested renaming the disorder “interpersonal regulatory disorder” and Quadrio (2005) put forth the concept of “post-traumatic personality organization,” however, both of these nomenclatures are problematic as they only describe a subset of individuals who fall under the current borderline personality disorder diagnosis.  Prior to the decision to eliminate the multi-axial diagnostic system in DSM-5, researchers New, Triebwasser and Charney (2012) suggested that borderline personality disorder be renamed “emotional dysregulation disorder,” and be reclassified on axis I with mood disorders rather than with personality pathologies on axis II, as  its symptoms, aetiology, phenomenology and course were more consistent with affective disruption.  The researchers found that patients diagnosed with borderline personality disorder had far better prognoses than those with other axis II disorders and, unlike those with other personality disorders, found their symptoms to be ego-dystonic rather than ego-syntonic or congruent with their sense of self.   In a similar vein, the latest version of the International Statistical Classification of Diseases and Related Health Problems, the ICD-10, set to be published by the World Health Organization in 2015, is retiring the diagnosis of borderline personality disorder in favor of “emotional unstable personality disorder” with modifiers for both “impulsive” and “borderline” subtypes (World Health Organization, n.d.).



Wherever the borderline is drawn and whatever it is called, patients will continue to fall within its boundaries:  Estimates of the prevalence of borderline personality disorder in the general population range from 2-3% and are estimated to constitute up to 25% of all inpatients and 15% of all outpatients in treatment (Guderson, 2008).  An acute awareness of developmental and neurobiological factors along with attentiveness to the social construction of the disorder will, hopefully, continue to increase clarity and improve treatments for the diagnosis.  Progress has been made:  Marsha Linehan’s development of dialectical-behavioral therapy has made management of what Knight may have once called “borderline states” more tolerable and easier for individuals suffering from them to negotiate.  Psychodynamic psychotherapies in which the analyst is no longer a “blank slate” but rather works with patients with possible borderline organization in an engaged and transferentially-aware aware have shown to be more efficacious and involve fewer negative therapeutic reactions than earlier treatments.  More mainstream books about borderline personality disorder have been being published in recent years and, in 2008, the National Institute of Mental Health established May as Borderline Personality Disorder Awareness month, seeking to decrease stigma around the diagnosis.  While diagnosis and treatment remain a priority, perhaps it is not so important what the borderlines are called so much as all the phenomena they demarcate.  As Nancy McWilliams writes, speaking of the early days of psychoanalysis, the borders between neurosis and psychosis may have been “somewhat arbitrary borders to begin with” (p. 81).

Pages: 1 2

About the author:

is a doctoral candidate at Duquesne University currently working on a qualitative dissertation investigating BDSM sexualities and alternative sexual communities. Her research interests include gender and sexuality, technology and embodiment, and developmental issues in adolescence and young adulthood. She tends to think both psychoanalytically and phenomenologically and, as such, is perpetually intrigued by both fantasy and the material world.


Editor-in-chief: Kai Bekkeli
Assistant editor: Erica Freeman
Assistant editor: Monica Lawson
Assistant editor: Christopher Bailey
Advisory editor: Celeste Pietrusza
Faculty editor: Elizabeth Fein, PhD

Mike Fosnaught, IC-Dev, web designer

Back to Top