This is a response to “The validity of borderline personality disorder: Robins and Guze applied” by Mark Ruffalo originally published here
Arguing against the construct validity of BPD
The question of whether borderline personality disorder (BPD) is a valid diagnostic entity can’t be settled simply by showing that a particular cluster of difficulties can be reliably identified, researched, or responded to in treatment. Reliability and clinical utility are not unique to valid constructs; history is replete with categories that could be consistently applied and acted upon, but which we would no longer regard as capturing real entities. Familiarity and institutional entrenchment are not the same thing as validity. The real question is whether the diagnosis meaningfully captures something that exists as a coherent phenomenon in the world, rather than something we have learned to see by repeatedly naming it.
When you look closely at the evidence typically offered in support of BPD’s construct validity, it is far less conclusive than is often claimed. Much of it can equally be understood as inferring there is something distinct and specific about a broad, heterogeneous set of distress responses that are highly dependent on developmental history, relationships, power, threat, and context. In other words, we may be mistaking a convenient descriptive shorthand for a discrete disorder.
The limits of Robins and Guze
Robins and Guze’s framework is frequently invoked as though it were a neutral arbiter of diagnostic reality. It isn’t. Their criteria emerged from a period in psychiatry that assumed mental disorders would eventually come to resemble medical diseases: discrete, bounded entities with clear causes and trajectories that can be objectively tested for. Fifty years on, that assumption has not been borne out—not for personality disorders, and not for most of psychiatry more broadly.
More importantly, the Robins and Guze validators primarily assess internal coherence within an existing diagnostic system. They tell us whether a category behaves consistently once it has been defined, not whether the category itself is conceptually sound or necessary. A construct can show reliability, familial aggregation, apparent stability over time, and even treatment responsiveness, while still being theoretically muddled or overly inclusive. None of those features, on their own, establish that we are dealing with a specific group rather than a useful but imperfect clinical convention.
Symptom specificity and the problem of heterogeneity
One of the most persistent problems for BPD is the sheer breadth of presentations it accommodates. Two people can meet diagnostic criteria while sharing only a single feature. That alone should give us pause when we talk about a “coherent syndrome.”
The characteristics often described as central to BPD—emotional lability, interpersonal sensitivity, identity disturbance, impulsivity, fear of abandonment—are not unique to it, nor do they reliably cluster together outside the diagnostic framework itself. These features appear across trauma-related conditions, mood disorders, neurodevelopmental presentations, and in people navigating ongoing adversity.
Attempts to distinguish BPD on the basis of “reactive” mood shifts versus more episodic or endogenous mood changes rely heavily on retrospective interpretation and clinical judgement. In everyday practice, clinicians frequently struggle to differentiate BPD from bipolar spectrum conditions, PTSD and complex trauma, ADHD, autism, or chronic depression. This persistent diagnostic ambiguity suggests that discriminant validity is weaker than formal reliability studies would imply.
When statistical methods are applied, they tend to show overlap rather than clean boundaries. Where distinctions do emerge, they often map onto severity, developmental timing, or trauma exposure rather than discrete diagnostic entities. Framing cPTSD as “subsyndromal BPD” is telling here: it assumes BPD as the reference point, rather than questioning whether BPD itself is a product of a particular diagnostic lineage. A recent example illustrates the point. In an ACAMH Learn webinar, Michael Kaess described adolescents who self-harm. Those exposed to significantly greater adversity also presented with more severe difficulties—and were the ones more likely to be labelled with BPD. To be explicit: this was not framed as a more severe response to more severe adversity, but as evidence of a different disorder.
Genetics without specificity
Heritability estimates for BPD, often cited at around 40–50%, are frequently presented as strong evidence for diagnostic validity. In reality, they tell us very little that is specific to BPD. Most psychological traits relevant here—emotional reactivity, impulsivity, negative affect, interpersonal sensitivity—show similar levels of heritability across the population.
Family and twin studies point to shared vulnerability for broad dimensions of emotional and relational difficulty, not for BPD as a distinct syndrome. To date, there are no genetic markers, polygenic risk profiles, or endophenotypes that uniquely predict BPD. The genetic evidence is transdiagnostic and fits far more comfortably with dimensional models of vulnerability than with categorical diagnoses.
Saying that BPD cannot be reduced to trauma alone is true, but unremarkable. It does not establish the existence of a discrete disorder; it simply reflects the well-established reality that human development is shaped by multiple interacting influences.
Course of illness: a problem for “personality disorder”
The commonly cited longitudinal course of BPD—substantial remission of self-harm, suicidality, and emotional instability over time—sits uneasily with its classification as a personality disorder. If the defining features remit relatively quickly for many people, what exactly is the enduring pathology supposed to be?
Ongoing difficulties in relationships and work are not unique to BPD. They are common consequences of chronic trauma, attachment disruption, social marginalisation, and long-term adversity. Interpreting improvement as evidence of validity, and persistence as confirmation of personality pathology, risks circular reasoning. Neither move resolves the deeper conceptual issue.
Neurobiology: correlation without discrimination
Neurobiological findings associated with BPD—such as heightened amygdala reactivity, frontolimbic differences, or HPA-axis dysregulation—are frequently cited as evidence that the disorder is “real.” But these patterns are not specific to BPD. They appear across PTSD, depression, anxiety disorders, dissociative conditions, and in people exposed to chronic stress without any psychiatric diagnosis.
What these findings demonstrate is that sustained emotional distress and threat leave biological traces. They do not demonstrate that BPD is a distinct neurobiological condition. Without specificity, neurobiology cannot do the conceptual work it is often asked to do here.
Treatment response does not prove the category
That there are effective treatments associated with BPD does not establish the validity of the diagnosis itself. Interventions such as DBT, MBT, schema therapy, and related approaches target processes that are clearly transdiagnostic: emotion regulation, mentalising, relational learning, and meaning-making.
These therapies are helpful because they address how people cope with intense emotional and interpersonal pain, not because they map neatly onto a discrete disorder. Their effectiveness supports the importance of these processes, not the existence of BPD as a natural kind.
Gender, power, and clinical function
Although population studies suggest similar levels of borderline traits across genders, the diagnosis continues to be applied disproportionately to women in crisis and acute settings. This is not a side issue. Diagnoses shape how distress is understood, how risk is managed, and how people are treated by systems.
Historically, BPD has often functioned as a relational diagnosis—used when distress is intense, care is difficult, or services feel under strain. That it can be applied reliably does not mean it is purely descriptive. It may also serve institutional and interpersonal functions that have little to do with identifying a discrete disorder.
Dimensional models and redundancy
Dimensional approaches don’t just “add nuance” to BPD; they raise the question of whether the category adds much explanatory value at all. When emotional dysregulation, attachment insecurity, impulsivity, identity disturbance, and trauma exposure are assessed directly, the BPD label often becomes redundant—a shorthand rather than an essential construct, but a shorthand that will bring significant disadvantage to those it is applied to.
Concerns that abandoning BPD would generate confusion echo earlier resistance to revising other problematic diagnoses. Psychiatry has revised diagnostic categories before when they were viewed as harmful and inaccurate—for example, the removal of homosexuality from diagnostic manuals. We are capable of similar shifts when the evidence and ethics demand it
Conclusion
The evidence commonly cited in support of BPD’s construct validity shows that certain patterns of emotional and interpersonal suffering are common, intelligible, and responsive to care. What it does not show is that these patterns constitute a single, coherent disorder that warrants categorical separation from trauma-related, neurodevelopmental, mood, or attachment-based formulations.
BPD is better understood as a historically contingent clinical shorthand—a map that has too often been mistaken for the territory. Substantially reframing, or retiring, the diagnosis would not erase clinical knowledge. It would allow for greater conceptual honesty, reduce circular reasoning, and bring psychiatric classification closer to what the evidence actually supports: dimensional vulnerability shaped by development, trauma, relationships, and context.
Marsha Linehan, the creator of DBT said “If you meet the criteria for Borderline personality disorder do not tell anybody. They’ll treat you differently if you do. And many mental health professionals will refuse to see you.”
We could adapt this advice for professionals. “If you see someone suffering, do not tell them and the world that their personality is disordered. You will make people who are supposed to care treat them worse, if they agree to see them at all”


